Recent Advances in Theory and Applications of Electromagnetic Metamaterials
View this Special IssueReview Article  Open Access
Taming the Electromagnetic Boundaries via Metasurfaces: From Theory and Fabrication to Functional Devices
Abstract
As twodimensional metamaterials, metasurfaces have received rapidly increasing attention from researchers all over the world. Unlike threedimensional metamaterials, metasurfaces can be utilized to control the electromagnetic waves within one infinitely thin layer, permitting substantial advantages, such as easy fabrication, low cost, and high degree of integration. This paper reviews the history and recent development of metasurfaces, with particular emphasis on the theory and applications relating to the frequency response, phase shift, and polarization state control. Based on the current status of various applications, the challenges and future trends of metasurfaces are discussed.
1. Historical Remarks
Metasurface is a twodimensional (2D) analogy of metamaterials [1, 2], which are artificial threedimensional (3D) structures with subwavelength inclusions as well as effective constitutive parameters not occurring in natural materials. In the last years, metamaterials with negative, zerorefractiveindex, and other exotic phenomena have enabled novel functionalities, such as invisibility cloaking [3] and perfect imaging [4]. Although metamaterials have achieved great success in the entire spectrum, great challenges still exist in the largescale fabrication and the design of devices with broadband response, especially in the visible frequency range [5]. As alternatives, metasurfaces were proposed to tune the behavior of electromagnetic wave within one infinitely thin layer. Since metasurfaces are extremely thin and much easier to be fabricated than metamaterials, they have become one of the most promising researching areas in electromagnetics and optics [5]. As noted by Holloway et al., current optical metamaterials are actually often limited to construction of metasurfaces as a result of the huge fabrication obstacles [6].
Since James Clark Maxwell formulated his famous equations in 1863, it has been known that all electromagnetic problems can be solved if one knows the constitutive parameters, initial and boundary conditions [7]. Actually, the essence of metamaterials is the precise design of constitutive parameters, whereas metasurfaces can be treated as a modification of boundary condition since they possess theoretically vanishing thickness. From this point of view, one can conclude that both metamaterials and metasurfaces can provide complete control over the electromagnetic fields. As illustrated in Figure 1, metasurface could replace traditional metamaterials in many conditions, such as invisibility cloak, perfect lens, and highefficiency radiators [3, 4, 8, 9]. It should be noted that the interface between metamaterials and normal materials can be also regarded as a novel metasurface boundary condition, where many optical properties can be engineered without resorting to the bulk properties of the metamaterials [10]. In particular, exotic surface waves at these metasurfaces have also attracted great attention.
(a)
(b)
(c)
Although the upsurge in attention has been paid to metasurfaces only recently, the actual concept of metasurface is far from new. In fact, the electromagnetic theory of metasurface can be dated back to 1902, when Wood reported his notable anomaly [11] and LeviCivita gave the boundary relations for a metallic sheet with vanishing thickness [12]. The discoveries of Wood and LeviCivita have promoted the development of two areas, which are long treated as independent of each other. One research area of metasurface is about the surface plasmon polariton (SPP), which is a particular solution of surface wave at metaldielectric surface; the other includes frequency selective surface (FSS), impedance sheet, and various types of planar antenna arrays. In the optical regime, the two areas begin to overlap with each other [13], since most of the metals in the metasurfaces become plasmonic at this particular frequency range. Indeed, it will be shown that the SPP and ordinary electromagnetic wave in metasurfaces can be deduced from a unified theory.
In recent years, there are a lot of research articles and review papers regarding the topic of metasurfaces. As an attempt to distinguish the concept of metasurface from previous FSS, Holloway et al. suggested that the different concepts can be classified according to the length scales [6]. Nevertheless, it is difficult to separate FSS and metasurface from the perspective of electromagnetic interactions, since both of them can be well described by effective impedance, although FSS is only intended to modify the frequency dependent transmission/reflection.
In this review, we attempt to give a unified perspective to the development of metasurfaces. The mechanism and applications of these metasurfaces are discussed in detail with an outlook for future development. As shown in Figure 2, there are many categories in the applications of metasurfaces: spectrum filters (including FSS [14–16], color filters [17], Fano resonances [18–25], and modulators [26–29]), plasmonic components for nearfield optics [30–34], amplitude manipulation devices (absorbers [35–38] and antireflection coatings [39–42]), phase shifters (geometric phase [43, 44] and impedanceinduced [19, 45] and plasmoninduced [2, 46] phase shifter), antennas, and other electromagnetic devices [47–50].
2. Overview of Theoretical Approach
As implied by its name, metamaterial can be treated as an effective medium with complex permittivity and permeability . In the last several years, the effective medium theory (EMT) has been successfully applied in the analysis of metamaterials. Various methods, such as field averaging and parameters retrieving, have been proposed to obtain these effective parameters [51]. Although metasurface is an analogy of metamaterials with reduced dimensionality, the effective medium theory turns out to be less useful since the thickness of metasurface is approaching zero [6]. Indeed, the definition of permittivity and permeability is ambiguous in such a thin layer [52].
In this section, we would like to give an overview of the impedance theory for simple metasurfaces [2]. Instead of the susceptibilities, the electric impedance and magnetic impedance are adopted to connect the electric and magnetic fields at the boundaries of metasurfaces. For simplicity of discussion, the discussion in this paper is confined in uniaxial metasurfaces. First, the impedance boundary condition for metasurface is deduced from a general slab with given constitutive parameters and thickness. Second, some possible applications are discussed based on the properties of metasurfaces.
As depicted in Figure 3, the electromagnetic boundary condition requires that the tangent components of both electric fields and magnetic fields are continuous across the interface between two dielectric media. For both the transverse electric (TE) polarization and the transverse magnetic (TM) polarization, the boundary conditions can be written aswhere and for TE polarization and and for TM polarization. The horizontal admittance for each medium is defined as for TE polarization and for TM polarization (here is the index of layer). From Maxwell equations, one can obtain
(a)
(b)
Appling (2) to (1), the admittance can be obtained as
Here, is the horizontal wave vector in this system, which is continuous for all planar multilayers. If light is illuminated from the air side, the incidence angle is . From (1) and (3), the reflection and transmission coefficients can be then expressed as follows:
It should be noted that (3) and (4) are identical to the traditional Fresnel equations, except that the horizontal wave vector instead of the ambiguous incidence angle is used here. Our approach has also an advantage of simplicity since the reflection and transmission coefficients for both of the two polarization states keep the same.
As shown in Figure 3(b), for a thin slab sandwiched by two different media, the boundary conditions arewhere 1, 2, and 3 denote the three layers and , , , and are the coefficients for the counterpropagating waves inside the slab. By definition, there is
For extremely thin metasurface where , it follows that
From (7), one can derive that
Here, we focus on the case when the permittivity and/or permeability of the middle layer is larger than that of the surrounding space. For nonmagnetic layer, ; thus, , so (8) approximates as
Inserting (9) into (5), there is
Comparing with the electric impedance boundary conditionsone can derive the electric admittance for this slab:
When the permeability is much larger than , ; thus, , so (8) approximates as
As a result, there is
Since the magnetic impedance boundary conditions can be written asthus the magnetic impedance could be derived as
Clearly, the above assumption is valid even for oblique incidence as long as is much smaller than unit and the constitutive parameters are much larger than unit. The expressions in (12) and (16) are similar to the previous results obtained from radiation problem [53].
In the following, we will only discuss the case of thin structures without magnetic response. From (11), the reflection and transmission coefficients are
These equations can be considered as the revised Fresnel equations for metasurfaces. In principle, there are two kinds of waves: one with and the other with . When , waves will transmit from the upper space to the lower one. There is an abrupt change in the phase and amplitude in the reflection and transmission coefficients.
When , the surface plasmon modes can be derived by enforcing the reflection and transmission coefficients simultaneously to be infinite. For a single interface between dielectric and metal, this condition for (4) becomes , leading to a propagation constant for TM polarization:
Since the discussion is confined to the case of nonmagnetic response, there is no solution for TE polarization.
For an electric layer embedded in an infinite dielectric medium (insulatormetalinsulator, IMI), the guiding wave can be obtained similarly as . It follows that the propagation constant can be derived from
If one assumes , there is [54]
Considering the fact that , there is
According to (21), the wavelength of the coupled SPP approaches zero along with the decrease of the thickness [55].
For metalinsulatormetal (MIM) structure, the above approximation is not valid anymore. The accurate propagation constant should be derived by directly solving Maxwell’s equations with boundary conditions:
In the microwave frequencies, many metals behave as nearly perfect conductors, so that no SPP can be found in ideal planar metal surfaces. Instead, if the metal surface is perforated by subwavelength holes, the effective permittivity can be arbitrarily tuned to sustain spoof SPP [56], where
The plasmon frequency is the cutoff frequency:
As will be shown in Section 3.6, spoof SPPs have been successfully utilized in the design of groove antennas in the microwave frequencies.
3. Applications of Metasurfaces
As shown in the previous discussion, the waves in metasurfaces have many exotic properties, including the extremely short wavelength, abrupt phase change, and strong achromatic dispersion. Based on these properties, traditional electromagnetic laws and theory, such as the theory of diffraction limit, laws of reflection and refraction, the Fresnel equation, and absorption theory, should be revised correspondingly [2]. Practically, the revised theorem could find many applications as follows.
3.1. Frequency Selective Surfaces
Frequency selective surfaces (FSSs) are planar, periodic arrays of conducting patches on a substrate or a periodic array of apertures in a conducting sheet, which are used to filter electromagnetic waves [14]. The spectral responses of these structures are affected by various physical and electrical parameters, including the shape and type of their constituting elements. These FSSs can be designed to demonstrate lowpass, highpass, bandpass, or bandstop behavior. Over the years, they have been extensively used in a variety of applications ranging between radar radomes, radar absorbers, and so on [14].
Early in 1919, Marconi and Franklin granted a patent that the contribution of a parabolic reflector was mimicked by using halfwavelength metallic wires [14]. Owing to the great potential military applications, FSSs have been the subject of intensive study since the 1960s.
The performance of FSS can be interpreted by the impedance and equivalent circuit theory, combined with the boundary condition. For a nonmagnetic freestanding metasurface, the bandpass and bandstop FSSs correspond to parallel and series type LC circuits, respectively. One common feature of the traditional FSSs is that the size of the resonant elements and their spacing are comparable to half a wavelength at the desired frequency of operation [58]. Since the electromagnetic response is a collective effect, the finite surface must include a large number of the constituting elements and be illuminated by a planar phase front to obtain the desired frequency response. At some conditions, such as radomes at lower frequencies, FSSs of relatively small electrical dimensions that are less sensitive to incidence angle and can operate for nonplanar phase fronts are highly desirable. In order to achieve this purpose, Sarabandi and Behdad proposed a new type of FSS. The structure consists of a periodic array of small metallic patches printed on one side of a dielectric substrate and a wire grid structure printed on the other side, both with the same period. The periodicity of the printed patterns is much smaller than the wavelength of the operating wavelength. The validity of this concept was demonstrated by numerical simulations and experiments.
One of the recent challenges for FSS is to optimize the frequency spectra with respect to angle of incidence and polarization of the waves impinging on it. In principle, multilayered FSS can improve these performances, but with oversized profiles. By using cavity mode whose resonant frequency is angleindependent, it was recently demonstrated that annular apertures array can show angle and polarizationindependent transmission properties [15].
Ideal FSSs require that the transmission/reflection at the band edge changes sharply, which cannot be satisfied by most singlelayer FSSs. In 2011, we proposed a novel plasmonic FSS with sharp band edges. By making use of plasmon hybridization, the bandwidth of the filter is tunable over a large range from 56.6 to 182.2 THz with magnetic and electric couplings between adjacent unit cells, as shown in Figures 4(c)–4(f).
(a)
(b)
(c)
(d)
(e)
(f)
More recently, the unique spectral properties of FSS were utilized to mimic many quantum phenomena, such as electromagnetically induced transparency (EIT) [59–61] and Fano resonance [19, 20, 62]. The interactions of different modes in the metasurfaces were tuned independently to achieve the desired spectral distribution of reflection and transmission. In 2007, Fedotov et al. reported a resonance response with a very high quality factor in a planar metasurface with symmetry breaking in the shape of its structural elements [20]. It was shown that this symmetry breaking enabled the excitation of trapped modes, that is, modes that are weakly coupled to free space. The interference of trapped mode and normal mode makes the transmission spectrum be of the famous Fano shape.
In order to give a rigorous theory of Fano resonance in metasurface, we studied the electric and magnetic responses of a planar metasurface with perturbed periodicity [19]. Rigorous sheet impedance theory was given to analyze the electricmagnetic and magneticmagnetic coupling effect. Interestingly, perturbation of the widths of wire pairs makes the adjacent metallic particles couple with each other. Qfactor as large as 100 has been theoretically obtained accompanying huge field enhancement, as described in Figure 5. It was believed that this kind of periodicity perturbation can provide a general approach for Fano resonance with ultrastrong local field enhancement.
(a)
(b)
(c)
(d)
Since there are three resonant peaks, the magnetic impedance can be written as a combination of individual magnetic resonances, as described by equivalent LC circuit model:
Here, , , , , , and are the corresponding inductances and capacitances. By fitting (25) with the retrieved magnetic sheet impedance, the LC parameters can be obtained as nH, nH, nH, pF, pF, and pF. The corresponding resonance frequencies are 9.79 GHz, 10.77 GHz, and 10.87 GHz.
Similar to the magnetic sheet impedance, the magnetic sheet current is also a summation of all the individual sheet currents:where
Using (25), (26), and (27), the sheet currents for different resonators can be easily calculated. The current of the first resonator dominates at frequencies around 9.79 GHz. On the contrary, the second and third resonators dominate the frequency region around 10.8 GHz. Besides, one can find that the first resonator is out of phase with the other two resonators for 9.79 GHz < < 10.77 GHz. Also, for 10.77 GHz < < 10.87 GHz, the third one is out of phase with the others.
In fact, the phase shift between these resonators is the key of resonant enhancement of sheet current. As an example, at frequency 10.82 GHz, there are and . Thus, we have
It is interesting that the second and third sheet currents in (28) are out of phase. As a result, even higher enhancement factor and narrower bandwidth can be achieved by decreasing the difference between and .
Up to date, a great deal of previously proposed or demonstrated metasurfaces and nanostructures require the use of metallic inclusions, leading to large ohmic loss, a serious limitation to obtain ultrahigh Qfactor and huge field enhancement when Fano resonance occurs. In 2014, a new type of Fano resonance in silicon rods array was reported, and the maximum Qfactor was achieved at a rather large gap width between adjacent rods [63]. It was also different from the previously reported Fano resonance in photonics crystal slab [64], where the maximum Qfactor corresponds to a vanishing hole diameter.
In the optical regime, FSS can be utilized to realize color filtering [17, 65]. Xu et al. demonstrated plasmonic MIM nanoresonator structures to filter white light into individual colors [17]. The key concept of their method is to use the linear dispersion of plasmon to realize the photonplasmonphoton conversion efficiently at specific resonant wavelengths (Figure 6). Compared with the aforementioned colorfiltering methods, the new design significantly improved the absolute transmission, pass bandwidth, and compactness. In addition, the filtered light is naturally polarized, making it attractive for direct integration in liquid crystal displays (LCDs) without a separate polarizer layer.
(a)
(b)
(c)
(d)
(e)
(f)
FSS can be also utilized in selective solar absorbers to suppress infrared reradiation. Owing to the frequencyindependent absorption, the equilibrium temperature of an irradiated blackbody in vacuum is only 80°C. In order to realize frequency selective transmission and absorption, multilayered films were utilized. Nevertheless, hightemperature instability limits the highest operating temperatures. In 1974, Horwitz used an array of deep holes in a metal to realize selective solar absorbers with ratios of solar absorbance to thermal emittance of 30 : 1 at temperatures of about 200°C [66]. Recently, we proposed a multilayered pyramidal array made of tungsten to demonstrate the selective solar absorption [67]. The broadband absorber is made up of a periodic array of multilayered truncated pyramids. The absorbance is about 99% from 0.28 μm to 1.5 μm for both TE polarization and TM polarization at room temperature or high temperature (966 K).
3.2. PolarizationManipulating Metasurfaces
Polarization plays an important role in electromagnetic waves since a majority of phenomena are polarization sensitive. Manipulation of polarization has been a hot topic for quite a long time. Traditional birefringent and chiral medium have the ability of polarization modulation. However, these traditional media need large thickness to accumulate enough phase shift between the perpendicular components of the incident electric fields. Thus, the optical systems involving polarization applications have become complex and bulky.
With the help of metasurfaceassisted law of polarization conversion (MLPC), one can greatly improve the performance and enable high conversion efficiency and ultrathin thickness [2]. In general, metasurfaces for polarization manipulation include transmissive and reflective types and can be classified into anisotropic and chiral ones from the viewpoint of the structure properties.
3.2.1. Transmissive Anisotropic Metasurface Polarizer
It is well known that polarization states of light can be changed by naturally occurring anisotropic media as the two axes possess different refractive indexes ( and ), leading to a relative phase shift between the two axes. However, as the difference between and is typically small, the thickness of traditional polarizers is required to be much larger than the operational wavelength. Moreover, the working bandwidth is narrow because the phase shift is frequency dependent.
As an alternative, metasurfaces provide new opportunities to achieve polarization manipulation with ultrathin artificial structures, including strip gratings and meander lines [68, 69]. In particular, Wang et al. utilized a singlelayer metasurface with asymmetric crossshaped apertures and demonstrated an ultrathin () terahertz quarterwave plate [70].
The physical foundation of the polarization control with metasurface is the metasurfaceassisted polarization conversion (MLPC), that is, the revised Fresnel equations [2], which can also be written asBased on (29), we can control the reflection and transmission for both polarization states. One particular simple example of this kind of anisotropic metasurface is shown in Figure 7, where the gaps between the parallel strips can be taken as capacitance and the strips can effectively be equal to inductance. Therefore, the electromagnetic wave with electric perpendicular () or parallel () components to the strips would be modulated by the effective capacitance and inductance. Hence, the transmitted phases of and would also be, respectively, retarded and advanced. When the differential phase shift equals ±90°, circular polarized wave would emerge.
(a)
(b)
(c)
(d)
In 2010, Euler et al. [71] designed a singlelayer circularly polarized convertor using periodic split slot ring structure, and this polarization convertor can realize relative 3 dB axial ratio (AR) bandwidth of 11.75%. In 2011, a thick metasurface was employed to manipulate EM wave polarization with near 100% efficiency in transmission geometry [72].
Recently, Ma et al. reported a singlelayer metasurface to convert the incident linearly polarized wave into circularly polarized one in a broadband [45], as shown in Figure 8. This polarizer is composed of metallic periodic structures. Two basic patterns are the metallic annular ring and 135° orientated strip. When a linearly polarized plane wave with field along the axis or axis normally irradiates onto the polarizer, it can be converted into the circularly polarized wave. The simulated and measured results show that this polarizer can transform the incident linearly polarized light into circularly polarized one in the frequency range from 13.5 to 15.3 GHz, where the AR is less than 3 dB. The relative bandwidth of this polarizer is 17%, and its thickness is only ( is the central working wavelength).
(a)
(b)
(c)
(d)
As depicted in Figures 8(b) and 8(d), the circular polarizer was integrated into a conventional linearly polarized patch antenna to realize high gain and circular polarized radiation. The gain of conventional patch antenna is sharply increased from 7 dB to about 11 dB at 13.9 GHz. In addition, nearperfect circularly polarized radiation pattern is realized in the frequency band ranging from 13.55 GHz to 13.9 GHz. This metasurfacebased circularly polarized antenna solved the difficulty problem of antenna design, and the radiation character was greatly improved.
The above mentioned anisotropic metasurfaces modulate the phase shift in two perpendicular directions, in a way similar to the traditional dielectric polarizers. However, this kind of singlelayer metasurface suffers the half power loss; namely, the incident power is 50% reflected by the metasurface, and only half of the incident linearly polarized wave can be converted into the circularly polarized wave. To solve this problem, more layers of metallic structures are needed according to the theory proposed by Markovich et al. [73]. It is proved that the efficiency can reach 100% in the two layered metasurfaces. However, the perfect polarization transformation is limited to a single frequency, and the transformation effect deteriorates drastically away from the central frequency.
3.2.2. Reflective Anisotropic Metasurface Polarizer
In order to increase the polarization conversion efficiency and working bandwidth of the metasurfaces, reflective metasurfaces (or metamirrors) have been proposed by various groups [74–79]. In 2007, Hao et al. [75] proposed a circular polarizer based on an ultrathin metasurface reflector. Hao et al. [80] and Pors et al. [74] also obtained similar phenomenon at optical frequencies using orthogonally oriented electrical dipoles. Compared to the transmissive polarization transformers, the metamirrors have much smaller thickness due to the large anisotropy as well as higher energy efficiency since no complicated antireflection technique is applied.
Owing to the intrinsic resonance, metamirrors are often realized in narrow frequency band. In order to overcome this problem, we designed a dispersive ultrathin metamirror to extend the bandwidth [81]. In general, the impedance sheet for both the  and directions can be frequency dependent. As a first attempt, we set as being constant () and as being highly dispersive. Since the dielectric spacer is chosen as air (), the reflected phase shift can be calculated directly from the wellestablished transfer matrix method [81]:where is the wave vector in free space and is the thickness of dielectric spacer. The optimal impedance for a certain relative phase shift can be calculated aswhere and . Subsequently, the optimal impedance required for the perfect polarization control can be obtained. The optimal impedance for , that for , and that for were calculated, which vary slightly with working frequency. In particular, the required impedance for is mainly capacitive within the whole frequency range. For , however, the curve is separated to one capacitive region and the other inductive region at the two sides of the central frequency , where is the speed of light. For , the requirement of dispersion becomes mainly inductive within the entire frequency range.
Using anisotropic metasurface, the optimal impedance can be approximated. Then, metamirror with Ishaped unit cell was constructed, as shown in Figure 9. The geometrical parameters were optimized using finite element method (FEM) at near normal incidence. In order to prove the above numerical results experimentally, a sample with 40 × 40 unit cells was fabricated using PCB technology. The measured reflection coefficient was below −15 dB in the frequency range between 5.5 and 16.5 GHz. This large relative bandwidth further confirmed the design of broadband polarization convertor. The reflective metasurface ensures the energy efficiency.
(a)
(b)
(c)
(d)
Based on the theory of broadband polarization conversion, Guo et al. [82] improved the performance of the broadband polarization convertor. In the design, it was proposed that the operation bandwidth and frequency selectivity of metasurfaces can be increased significantly with fully released dispersion management capability in two dimensions. Multiple resonance mechanism was employed to match the effective impedance of the metamirror with the ideal impedance and significantly broaden the operating bandwidth. Experimental results demonstrated that this metamirror worked well from 3.2 to 16.4 GHz with polarization conversion efficiency higher than 85%. This converter was also superior to the previous devices in the frequency band selectivity because the operation band approximates an ideal rectangle. The rectangular coefficient, defined as the bandwidth ratio between high (>80%) and low (<20%) conversion efficiency, was high up to 0.94.
In 2013, Grady et al. experimentally demonstrated a reflective metasurfacebased terahertz polarization converter that is capable of rotating a linear polarization state into its orthogonal one [77]. As depicted in Figure 9(d), the device is able to rotate the linear polarization by 90°, with a conversion efficiency exceeding 50% from 0.52 to 1.82 THz, with the highest efficiency of 80% at 1.04 THz. They also created multilayered structures capable of realizing nearperfect anomalous refraction.
More recently, Jiang et al. proposed an equivalent theoretical description of metasurface polarizer [78]. As examples to apply this concept, a broadband quarterwave plate and a halfwave plate were demonstrated. Once again, this approach validated the importance of chromatic dispersion in metasurface polarization control. In a similar work, broadband metasurface polarizers were demonstrated in the visible frequencies with highefficiency, angleinsensitive polarization transformation over an octavespanning bandwidth [79]. Nanofabricated reflective halfwave and quarterwave plates designed using this approach have measured polarization conversion ratios and reflection magnitudes greater than 92% over a broad wavelength range from 640 to 1290 nm and a wide field of view up to ±40°.
3.2.3. Chiral Metasurface Polarizer
Beside the anisotropic polarizers, there is another type of material suitable for polarization manipulation, which is called chiral material. Chirality refers to the materials or structures that cannot superpose with their mirror structures by only transferring or rotating the materials or structures. As depicted in Figure 10, chirality can be observed in naturally occurring media, including aminophenol, DNA molecule, and crystals.
(a)
(b)
(c)
(d)
(e)
Due to the lack of mirror symmetry, the crosscoupling between electric field and magnetic field exists in chiral materials, which is the physical original of these special electromagnetic properties. The strength of the coupling can be represented by the chirality parameter , and thus the constitutive relation in the chiral materials can be stated aswhere and are the permittivity and permeability of vacuum, and are the relative permittivity and permeability, and is the strength of the coupling between the electric and magnetic fields. Negative refraction index can be obtained in chiral metamaterials while it is not necessary to realize negative permittivity and negative permeability at the same time. Chiral materials have attracted great attention for their special electromagnetic properties such as circular dichroism and optical activity. Right circularly polarized (RCP) wave and left circularly polarized (LCP) wave would encounter different transmission coefficients at the resonances and exhibit polarization manipulation property, such as circular dichroism (CD) and optical rotation (OR). The refraction indices of the LCP wave () and RCP wave () are related to the chiral parameter and can be stated as . Therefore, negative refraction index can be obtained if the chiral parameter is large enough. As analyzed above, chiral materials can act as a kind of polarizer for their selection of circularly polarized waves. However, the crosscoupling between the electric and magnetic fields is rather weak in these natural materials, and very thick materials are often needed to produce useful CD and OR properties.
The appearance of artificial chiral structures makes it possible to realize giant CD and OR with the thickness of ~ at the operating frequency. Thus, due to their unique properties, chiral metamaterials promise to tailor the polarization state of electromagnetic wave and be functionalized as polarization rotators, circular polarizers, and polarization spectrum filters with ripplefree isolated transmission peaks. Furthermore, chiral metamaterials are more suitable to achieve multiband and multipolarization conversion compared to anisotropic ones.
In the extreme case, chiral material can be used to achieve circular polarization selective surface (CPSS), which was thought to be an important polarizer [83]. The ideal CPSS should have thickness much larger than the working wavelength; thus, it cannot be approximated as metasurface. In 2009, Gansel et al. [84] designed a broadband CPSS based on the periodic gold helix structure, which splits the incident circularly polarized waves in an octave bandwidth at midinfrared frequency. Later, this helix shaped chiral metamaterial was scaled to other frequencies [85, 86].
To obtain chiral metasurfaces, the bandwidth needs to be compromised. In recent years, a great number of chiral metasurfaces have been reported in the form of twisted Ushape [87] and Lshape [88], and so forth. Ma et al. proposed chiral metasurfaces based on planar spiral structures to achieve multiband and multipolarization conversion [89].
Firstly, a multiband circular polarizer was proposed by using three layered planar spiral metasurface structure in analogy with classic spiral antenna [86]. At three distinct resonant frequencies, as shown in Figures 11(a) and 11(b), the incident linearly polarized wave with electric field parallel to one specific direction was transformed into the left/righthanded circularly polarized waves.
(a)
(b)
(c)
(d)
(e)
Subsequently, a dualband asymmetric chiral metasurface was constructed by using two pairs of planar spiral structures with a certain twisted angle [89], as shown in Figure 11(c). Two planar spiral structures with different radii in each layer were adopted to achieve more resonant frequencies. The incident linearly polarized wave can be converted into the circularly polarized waves with different rotation directions at four distinct resonances. This multiband chiral metasurface greatly decreased the loss of chiral metamaterials, compared to the results reported before by combining the chiral and anisotropic properties in the unit cells. This design of multiband response by putting together structures with different size into one unit cell can also be extended to other metasurfaces and metamaterials.
In order to extend the operational bandwidth, Huang et al. [91] proposed a dualband wideband polarization rotator which is proposed by using chiral metasurface composed of two pairs of two layered twisted split ring resonators (SRRs) in each unit. This chiral metasurface transformed the incident linearly polarized wave into its crosspolarized one at two distinct frequencies with high efficiency.
In some satellite communication systems, different circular polarization types are required, that is, LCP wave for the uplink and RCP wave for the downlink at Ku band. The usual method for the above requirement is to design two antennas with different circular polarization types as transmitter and receiver, respectively. Another solution is integrating dielectric polarizers or orthomode transducers (OMT) into the horn antennas. Though broadband polarization splitting can be realized, these methods may increase the size of the antennas.
Owing to the giant polarization conversion efficiency, the chiral metasurfaces are suitable for the construction of high performance circularly polarized antennas with multipolarization and multiband performance. As shown in Figure 12(e), a dualband dual circularly polarized horn antenna was designed [92] based on the planar spiral chiral metamaterial. The chiral metasurface converts the incident linearly polarized wave into the LCP and RCP waves at two resonance frequencies, with insertion loss less than 0.6 dB. The measured 3 dB AR bandwidths of the antenna were, respectively, 0.8% (12.4 GHz–12.5 GHz) and 1.4% (14.2 GHz–14.4 GHz). The radiation patterns of the proposed antenna for the LCP and RCP wave depicted great crosspolarization ratio, which confirmed that the dominant circular polarization was lefthanded at the lower frequency and was righthanded at the higher resonance. The gain of the antenna composite was only degraded by 0.6 dB around these two resonant frequencies, in comparison with the horn antenna without the chiral metasurface. The designed antenna has the advantages of low cost and simple structure and thus could be utilized in satellite communication systems.
(a)
(b)
(c)
(d)
(e)
(f)
In order to extend the operational bandwidth, Huang et al. [91] proposed a dualband wideband polarization rotator by using chiral metasurface composed of two pairs of two layered twisted split ring resonators (SRRs) in each unit, as shown in Figures 13(a) and 13(b). The sizes of the two pairs of SRRs are different. Therefore, this chiral metasurface transformed the incident linearly polarized wave into its crosspolarized one at two distinct frequencies with high efficiency. Pure optical activity was observed at two distinct resonant frequencies. The designed planar polarization rotator had a high transmission in the polarization rotation frequency bands and can be also easily extended for transforming polarization state at the multiplefrequency bands, which may have many potential applications in the microwave domain.
(a)
(b)
(c)
(d)
(e)
(f)
Lately, Ma et al. proposed a doublelayer twisted Yshape structure to achieve obvious circular dichroism and giant optical rotation at different frequencies [93]. The schematic geometry of the proposed metasurface is depicted in Figure 13(d). The unit cell is composed of two Yshaped metallic structures at a certain twisted angle printed on two sides of a dielectric lamina. The angle between the neighboring two branches of this Yshaped structure is designed to be 120°. When a polarized wave is incident to this chiral metamaterial, circular dichroism with a great difference of 25 dB between the transmission coefficients for the circularly polarized waves is obtained at 12.28 GHz. Meanwhile, 90° optical rotation is observed at 12.70 GHz, where the incident polarized wave is transformed into its crosspolarization with a transmission coefficient of −1.15 dB.
3.3. Antireflection Metasurfaces
In many optical devices and electrooptical equipment, the efficiency of light transmission determines to a large extent the overall performance. For example, it was reported that a normal solar panel absorbs and reflects approximately 25% and 33% of the incident solar radiation, respectively. For lens systems, the reflection would seriously reduce the imaging quality.
The first theoretical frame for analyzing antireflection coating (ARC) is Fresnel equation, which was developed in 1823 by AugustinJean Fresnel. Before this time, some antireflection phenomena have been already discovered by Lord Rayleigh and Joseph Von Fraunhofer [94].
In general, there are two ways to create broadband antireflection surfaces, which are based on either coating multilayer thin films or texturing the substrate surfaces with subwavelength structures. The first method is based on multilayer material with gradient refractive indices, and the latter needs only a singlelayer material but gradient morphology. Due to the scarcity of optical materials with refractive indices close to air, multilayer thin films for broadband antireflection surfaces were not realizable until recently [95]. However, the mismatch in thermal properties of different materials hinders multilayer thin film applications.
One of the inspirations of ARC design comes from the natural worlds. In the eyes of some moths, broadband antireflection was observed and the physical mechanism is attributed to the effective gradient refractive index distribution. However, it was found that, in motheyelike structures, highreflection frequency bands alternate with lowreflection frequency bands. It is of great interest for us to see if these highreflection bands can be suppressed and hence obtain a broadband reflection approaching 0.1%. As an effort to achieve this object, it was proposed that a hybrid motheye structure could be used to enhance broadband antireflection properties. An ultralow average reflectance down to 0.11% over the solar spectral range has been achieved, showing a 50% enhancement in broadband antireflection capability as compared with corresponding uniform motheye structures.
The largest challenge faced by traditional antireflection proposals is the complex fabrication process. In 2007, biomimetic silicon nanostructures were demonstrated by Huang et al. with improved broadband and quasiomnidirectional antireflection properties [97]. In 2014, Hong et al. used laser direct writing to create microstructures on Si surfaces that reduce light reflection by light trapping [42]. By decoration of the Si nanowires with metallic nanoparticles, surface plasmon resonance can be used to further control the broadband reflections, reducing the reflection to about 0.8% across the bandwidth from 300 to 1200 nm.
Although structured surfaces with gradient dielectric constant (e.g., pyramidal rods and moth’s eyes) provide rather good antireflection performance, their thicknesses and mechanical performances are not well balanced. In order to reduce both the thickness and the reflectance, the metallic metasurfaces are recently investigated for this purpose [39–41, 96, 98].
In 2008, Thoman et al. demonstrated that nanostructured gold films can serve as broadband impedancematching coatings for substrates in the terahertz frequency range [39]. They showed in theory and experiment that the internally reflected electric field amplitude of a broadband terahertz pulse in a silicon substrate can be suppressed to below 1% of that without coating, which is at least a factor of 5 better than the best suppression achievable with bulk gold layers (see Figure 14). Subsequently, Zhou et al. examined the potential of stacked multilayer graphene as broadband terahertz (THz) antireflection coating [40]. The reflected pulses from the quartz and silicon substrates were observed to change with the layer number and doping concentration of the graphene coating. Remarkable broadband impedance matching was achieved due to optimized THz conductivity.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
In a similar work, Teng et al. demonstrated broadband antireflection coatings using deep subwavelength periodic thin metallic lamellar grating [98]. As shown in Figures 14(a)–14(c), the measured reflectance from Cr grating coated sample was reduced over 99% compared with bare Si sample in the whole bandwidth of 0.06–3 THz. Recently, Zhang et al. proposed a metasurface optical antireflection coating [96]. By tuning the crossshaped metallic patches on a dielectric spacer, the reflection coefficient can be dramatically reduced (Figures 14(d)–14(g)). The bandwidth and incidence angle range were comparable to a quarterwave antireflection coating.
3.4. Perfect Absorbing Metasurfaces
3.4.1. SinglePort Perfect Absorbing Metasurface
Absorption phenomena of electromagnetic wave can be found anywhere in our surrounding world. The colorful leafs and flowers are related to the frequency selective absorption of light. It was also well known that black object, such as carbon, absorbs most of the visible light. In 1860, Kirchhoff proposed the ideal blackbody absorber as “bodies can be imagined which, for infinitely small thicknesses, completely absorb all incident rays, and neither reflect nor transmit any.” Since then, the pursuit of high efficient absorbers never stops.
Along with the development of Maxwell equations and invention of radar, the electromagnetic theory of absorbers was brought out after World War II. In the review given by Hans Severin in 1956 [35], the theory of the socalled Salisbury absorber, Jaumann absorber, Dallenbach absorber, pyramidal absorber, and dipole absorber was given in detail. The dipole absorber is just the precursor of the circuit analogy absorber (CAA) [99], which is also one kind of lossy metasurface. Severin also noted that there is a conflict between the thickness and bandwidth of these absorbers. Almost half a century later, Rozanov developed a rigorous theory for the bandwidthtothickness ratio [100], illustrating that there is a fundamental limit for this value. Remarkably, this conclusion consists with the prediction of Planck and Masius with respect to the thickness of blackbody [101].
In recent years, broadband absorbers with optimal thickness were widely researched [38, 102–111]. Except for broadband absorbers, narrow band absorbers are also of particular importance in frequency selective applications. In 2008, Landy et al. proposed the concept of “Perfect Metamaterial Absorber” with thickness of only , where is the working wavelength [36]. They interpreted the principle of this absorber as the simultaneous control of electric and magnetic response such that the impedance is matched to the free space. However, it is now widely known that such impedance match does not ensure that the transmission and reflection can be simultaneously reduced to zero. As stated by Vora et al., it is ambiguous to define the and in such complex structures, since the metamaterial perfect absorber cannot be strictly considered as homogeneous bulk media [52].
In 2008, we extended the metaabsorber into the visible frequencies. By converting the propagating wave into surface wave, perfect absorption of light wave was demonstrated. Antisymmetric surface plasmons coupling formed by subwavelength hole array (SHA) and reflecting layer was proved to dominate the multiorders nearperfect absorption. Although only SHA was investigated in this case, the design idea, realizing perfect absorption based on structured surface combined with thick metal layers, can also be extended to other cases, such as split ring resonator (SRR) combined with metal black plate. The most fascinating potential application of SHA combined with reflecting layer is the introduction of functional material into center dielectric region to realize imaging and detecting. In 2009, we further investigated the loss mechanism of perfect absorber. It was demonstrated that the ohmic loss and the dielectric loss both contributed to the absorption. The energy exchange was recently studied to enhance the efficiency of solar cells [52]. In addition, subwavelength high performance metamaterial absorber was also demonstrated by Hao et al. for optical frequencies [112]. Experimental results show that an absorption peak of 88% is achieved at the wavelength of ~1.58 μm. At almost the same time, Liu et al. demonstrated an infrared perfect absorber and verified its application as plasmonic sensor [113].
As shown in Figure 15, the reflection and transmission problem can be rewritten as a scattering and reradiation problem of the bounded wave [2]. The propagating wave and bounded wave (Mwave) could exchange with each other with the help of metasurface. In the far fields, we only care about propagating wave. As such, the total electromagnetic fields can be written asHere, is the scattered electric fields, which is determined by the incident electric field and the structure function . To realize complete absorption, the function should be designed with nobackward scattering.
(a)
(b)
(c)
In practical applications, metallic reflecting layers are often used as a ground plane for these absorbers. As illustrated in Figure 16, we proposed a circuit model to interpret the electromagnetic interaction in these absorbers [37]. The magnetic response was mathematically treated using a modified equivalence circuit model. The relation between the reflection and impedance can be written aswhere , , and are the intrinsic admittance of vacuum, dielectric spacer, and metal and and are permittivities of dielectric and metal. and are the wave vector in the vacuum and dielectric spacer. is the reflectivity of thick metal layer. Equation (34) is the basis of the metasurfaceassisted absorption theory (MAT) [2].
(a)
(b)
(c)
(d)
(e)
(f)
(g)
Subsequently, Feng et al. showed that the bandwidth can be dramatically increased by tailoring the dispersion of metasurfaces [114]. With a thin layer of structured nichrome, a polarizationindependent absorber with absorption larger than 97% was numerically demonstrated over larger than one octave bandwidth. It was shown that the bandwidth enhancement is related to the transformation of Drude model of free electron gas in metal film to Lorentz oscillator model of bound electron in the structured metallic surface. It should be noted that the Lorentz form is just one kind of dispersion that mimics that of ideal absorbing metasurface. Thus, further engineering of the dispersion may lead to even larger bandwidth.
In addition, to meet the requirement of bandwidth enhancement, metasurface absorbers are expected to be designed to reduce the fabrication complexity of traditional absorbers. Many works have been devoted to this design. For instance, a simple method was proposed by randomly adsorbing chemically synthesized silver nanocubes onto a nanoscale thick polymer spacer layer on a gold film [115]. The filmcoupled nanocubes provide a reflectance spectrum that can be tailored by varying the geometry (the size of the cubes and/or the thickness of the spacer).
In general, the maximal absorption bandwidth is limited by the optical thickness as indicated by the thicknessbandwidth ratio. For absorbers working at terahertz and higher frequencies, the physical thickness is very small even for quite large optical thickness. As a result, the thickness is not a big problem for broadband absorption at these frequencies. In contrast, the fabrication technique becomes a challenge since most of the broadband absorbers require multilayer thin films or complicated structures. Recently, we focused on the design of broadband absorber based on doped silicon, which has been considered as a new kind of metamaterials [116]. When the working frequency was larger than the plasmon frequency, doped silicon behaves as a highly lossy dielectric material. Yet a doped silicon slab by itself was not a good absorber due to the impedance mismatch between the silicon slab and free space. The power reflection at the interface was larger than 28% for refractive index , which was similar with nondoped silicon. In order to reduce the reflection and enhance transmission (the case for nondoped silicon) or absorption (the case for doped silicon), antireflection techniques should be used.
In the framework of transmission enhancement, the period of the grating structure should be in deep subwavelength scale to suppress nonzeroorder diffractions. However, for high efficient absorber, we showed that the absorption bandwidth can be dramatically increased by utilizing both the zero and firstorder diffractions [117], as illustrated in Figure 17. Compared with previous broadband absorbers, the structure we proposed was mechanically stable and much easier to be fabricated. This idea was further extended by other researchers [104, 118]. The experimental results showed that more than 95% absorption can be obtained from 1 to 2 THz. It was also interesting to investigate whether secondorder or higherorder diffraction can be utilized to increase the absorption bandwidth. In order to do this, the diffraction at higher frequencies was calculated for a twolayer grating. The relative absorption bandwidth for becomes larger than 150%. Further increase of layers may lead to larger bandwidth. Nevertheless, the fabrication process will become more complex and the thickness will become larger.
(a)
(b)
(c)
(d)
(e)
(f)
By stacking metasurfaces in multilayers, the absorption bandwidth could be further increased. As illustrated in Figures 18(a) and 18(b), an excellent absorber is designed with a very large bandwidth (3.26–34.65 GHz) [110]. It is shown that the total thickness of the design (14.5 mm) is only a little thicker than the minimum possible thickness dictated by the physical bound. Nevertheless, the multilayered absorber is difficult to be fabricated since the resistance of each metasurface should be accurately controlled. More recently, we fabricated a broadband absorber using magnetic controlled sputtering and optical lithography technique [2]. Since both the substrates and conducting material are transparent in the optical spectra, such absorbers could be utilized in many areas, such as smart windows and cockpits. As can be seen in Figures 18(c) and 18(d), both the optical transmittance and the microwave absorption are large enough for practical applications.
(a)
(b)
(c)
(d)
3.4.2. Coherent Perfect Absorber
Recently, coherent perfect absorption of light was proposed and demonstrated in a planar intrinsic silicon slab when illuminated by two beams with equal intensities and proper relative phase [119, 120]. Such a device is termed a “coherent perfect absorber” (CPA) and a “timereversed laser.” Compared with the perfect absorbers based on metasurfaces or plasmonic structures, the new device provides additional tunability of absorption through the interplay of absorption and interference. The coherent control of absorption is potentially useful in transducers, modulators, or optical switches [121, 122]. However, as a timereversed process of laser, CPA is characterized by narrow bandwidth and thought to be not applicable in solarcell and stealth technology.
As shown in Figure 19, it was demonstrated that the bandwidth of a thin film CPA can be rather large if the thickness is thin enough and the corresponding material has a specific chromatic dispersion resembling that of metal [38]. Two different regimes of metallic thin film CPA were derived based on the general CPA condition, characterized by extremely broad and moderately narrow bandwidth, respectively.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
For nonmagnetic material, the CPA condition for normal incidence can be obtained:The sign is corresponding to the symmetrical or antisymmetrical inputs. In the previous discussion, an infinite number of discrete solutions of (35) have been found for . The bandwidth is defined as the frequency width between the maximum absorption and adjacent minimum absorption and is characterized by . In this case, the CPA is rather narrowband and referred to as a timereversed process of laser.
In fact, the bandwidth of CPA can be very large if is extremely thin (, ). In this case, the left side of CPA equation becomes , and the right side can be approximated as . As is very small, only plus sign in the right side (symmetric mode) should be chosen and the real and imaginary parts of the refractive index ( and ) become equal with
In this case, becomes ; thus, the working regime can be approximated as and the required refractive index must be much larger than unit (). According to the definition of Yu and Capasso, such kind of thin film can also be thought of as a metasurface [1].
Different from the general CPA condition, the CPA condition for ultrathin film is explicit. Obviously, Drude metal could be used as natural material for CPA, with extremely broadband response. We obtained the thickness for CPA at this frequency regime:where is the speed of light in vacuum. This characteristic length is also called the Woltersdorff thickness [12], which quantifies the thickness of a metallic film with maximum absorption for incoherent beam input in the low frequency regime. When , the maximum absorption is 0.5, while the reflection and transmission are 0.25, respectively. For , most of the energy is transmitted; for , most of it is reflected.
Since the Woltersdorff thickness is independent of frequency, the absorption is very broadband. Generally, this frequency range covers all the low frequencies up to terahertz. In particular, a 0.3 nm thick tungsten film can absorb almost all the microwave and even terahertz energy under coherent condition, thus breaking the thickness limit proposed by Planck and Rozanov. Compared with the original CPA [120], the bandwidth increased more than 10^{10} times. In the optical CPA, the thickness of tungsten should be increased to 17 nm, which is still neglectable compared to the operational wavelength. The concept of ultrabroadband CPA was recently experimentally demonstrated in the microwave frequency. As shown in Figure 19(d), the absorption coefficients reach 100% at frequencies ranging from 6 to 18 GHz.
Subsequently, a customized CPA was also proposed based on the electric and magnetic resonances in a three layered metalinsulatormetal structure [123]. These kinds of resonances were attributed to the plasmon hybridization effect due to the coupling of individual resonators. Most importantly, it was found that the antisymmetrical absorption associated with magnetic resonance is almost independent of the polarization of light and angle of incidence; thus, the structure is highly suitable for coherent absorption of divergent beams (Figure 19(f)). To interpret the interaction of magnetic and electric fields with the metasurface, effect constitutive parameters are retrieved and general CPA condition for material with both electric and magnetic responses was given:where is the wave vector in free space and is the total thickness of the effective slab. The signs are corresponding to the symmetrical and antisymmetrical inputs, respectively. This equation provided a general approach to the design of CPA operating at arbitrary frequencies.
The coherent control method can also be utilized to realize other functionalities. For example, Zheludev et al. proposed the application of CPA in signal processing [124]. Potential applications include but are not limited to (a) a pulse restoration (clock recovery) device to restore the form of distorted signal pulses according to that of a clock (control) pulse, (b) a coherence filter that improves the coherence of light beams by absorbing incoherent components, and (c) a coherent lightbylight modulator wherein a digital or analogue intensity or phasemodulated control input governs signal channel output. In addition, they showed that reflection and refraction effects on phase gradient metasurfaces can be coherently controlled by a second wave [125]. In addition, Cao et al. also demonstrated the possibility of coherent subwavelength focus. Broadband operation was made possible by engineering the dispersion of the complex dielectric function, similar to the previous methods [38]. The local enhancement can be significantly improved compared to the standard plane wave illumination of a metallic nanoparticle. Their numerical simulation showed that an optical pulse as short as 6 fs can be focused to an 11 nm region.
More recently, coherent perfect rotation (CPR) of electromagnetic polarization states has been proposed by utilizing the odd time reversal symmetry in Faraday rotation [126]. Similar to coherent perfect absorption, the bandwidth of CPR is limited. The main advantage over other polarization transformers is that the polarization states can be dynamically tuned by the phase retardation. Nevertheless, it is pointed out that only Faraday rotation, but not optical activity or anisotropy, is capable of coherent perfect rotation. In our recent work, a dynamic polarization transformer was constructed by taking advantage of the time reversal symmetry of polarization conversion in anisotropic metasurface [76]. It should be noted that no Faraday rotation effect is utilized. The working bandwidth covers the entire microwave and terahertz bands. More interestingly, it is demonstrated that the output polarization states can be easily tuned between linear polarization and arbitrary circular polarization through phase modulation.
3.5. Plasmonic Metasurfaces
3.5.1. Superlens and Hyperlens
As discussed in Section 2, some particular metasurfaces can support surface waves, such as SPPs. The last several decades have witnessed the rapid development of SPPs and related areas. One of the largest advantages of SPPs is their ability to construct superlens and hyperlens to break the diffraction limit, which prevents the imaging of subwavelength features [34]. As we demonstrated recently, the electromagnetic waves on metasurfaces are characterized by three distinct properties [2]. The most important one is the short wavelength [55], which is actually the basis to break the diffraction limit.
In 2000, Pendry firstly proposed the concept of perfect lens [4], in which the amplification of evanescent waves and perfect imaging could be obtained by a negative refraction index slab. The short wavelength characteristic of SPPs on a sliver film was firstly demonstrated in 2003 [127, 128], and surface plasmon resonant interference nanolithography technique (SPRINT) was proposed to achieve resolution of halfpitch 50 nm (~1/9 wavelength) in 2004 [31], as shown in Figures 20(a) and 20(b). Using exposure recording scheme, Fang et al. experimentally demonstrated in 2005 that a 35 nm thick silver superlens was capable of subdiffraction imaging of halfpitch 60 nm objects (~1/6 wavelength) [30]. Figure 20(c) illustrates the experimental scheme of subwavelength imaging through a silver lens, where the objects were located 40 nm away from the silver superlens. Due to the superlens effect, the object of “NANO” characters was imaged on the photoresist and the average line width of images is 89 nm (Figure 20(d)). By utilizing a similar scheme, the superlens effect was also confirmed by other groups [129, 130].
(a)
(b)
(c)
(d)
(e)
(f)
The metasurfaceassisted diffraction theory (MDT) can be utilized in plasmonic lithography [2], where the silver lens has practical limits, such as optical loss, surface roughness, and film thickness. To address these issues, some important studies were reported, including active lens and plasmonic reflective slab [32, 131]. Based on the reflective amplification of evanescent wave, Wang et al. experimentally achieved the deep subwavelength imaging lithography for nanocharacters with about 50 nm line width [32], as depicted in Figures 20(e) and 20(f). It was also reported that the plasmonic cavity lens can be utilized to achieve high aspect profile [132]. The profile depth of halfpitch 32 nm resist patterns was enhanced to be about 23 nm, which is much larger than the previously reported results (sub10 nm). The resolution of SPPs imaging lithography was also extended to be halfpitch of 22 nm, whereas the quality needs to be further increased.
The subdiffraction imaging model of a planar plasmonic lens can be written in the form ofwhere , , and are the illumination function, transmission function of object, and optical transfer function (OTF) of plasmonic lens, respectively. The OTF can be obtained through transfer matrix. If there is reflection plane in the plasmonic lens, the OTF should be determined by using the fields in the photoresist layer, implying that the images can also be modulated by the reflecting layer [32].
Although the superlens provided images of nanoobjects well beyond the diffraction limit, the inherent disadvantage of nearsighted mode impedes its widespread application. It is highly desirable to utilize a plasmonic lens system which could resolve subwavelength details of objects in far field. In 2006, Jacob et al. proposed a hyperlens composed of cylindrical metaldielectric multilayers to magnify subwavelength details of objects so that the subwavelength features are above the diffraction limit at the hyperlens output [133]. Then, conventional microscopy can be utilized to capture the output of hyperlens to achieve farfield super resolution imaging.
3.5.2. Plasmonic Metasurface Lens
Based on Snell’s law, traditional optical lenses must have curved surfaces to focus light. In other words, the phase retardation was accumulated by height variation much larger than the wavelength. In recent years, metasurfaces were widely utilized to achieve planar lenses, based on the metasurfaceassisted law of reflection and refraction (MLRR) [2, 134]:where is the phase gradient in the metasurface plane, which is determined by the geometric structure and distribution, and may be changed by external stimuli, leading to the adaptive tuning of the law of refraction and reflection. and are the refractive index of media at the incident and transmit sides. , , and are the angles for incident, refracted, and reflected light.
In particular, plasmonic lenses have become one widely researched approach, which has now been demonstrated in both onedimensional and twodimensional cases [46, 135, 136]. These devices offered an alternative to conventional refraction microlenses. The planar and compact designs are beneficial for obtaining short focal distances on the order of several micrometers.
Early in the 21st century, a novel method was proposed to manipulate the phase retardation of light at nanoscale, providing a new approach to develop various nanooptical devices, such as flat lenses, collimators, and splitters. Figure 21(a) illustrates the principle of planar lens based on widthtuned plasmonic slits. The slits transport electromagnetic energy in the form of SPPs in nanometric waveguides and provide desired phase retardations with variant phase propagation constant. Under plane wave illumination, the light energy was focused at the focal plane μm, as shown in Figure 21(b). According to Fermat’s Principle, such devices could also achieve arbitrary angle of light deflecting such as that shown in Figure 21(c).
(a)
(b)
(c)
(d)
(e)
(f)
Although the plasmonic lens is easy to be fabricated, it was only until 2008 that the experimental demonstration of such device was reported [46]. For plane wave illumination, the light was focused at the focal length of 5.3 μm. The excellent agreement between experiment and simulation validated the design principle of this approach. It should be noted that the plasmonic version of abnormal refraction is similar to that in other metasurfaces, as shown in Figures 21(e) and 21(f).
More recently, as shown in Figure 22, Chen et al. demonstrated that the farfield focusing pattern of planar metal lens could be modulated by slits filled with phase change material (Ge_{2}Sb_{2}Te_{5}, GST) [135]. By varying the crystallization level of GST from 0% to 90%, the FabryPérot resonance supported inside each slit can be spectrally shifted across the working wavelength at 1.55 μm, which results in a transmitted electromagnetic phase modulation as large as 0.56π. Based on this geometrically fixed platform, different phase fronts can be constructed spatially on the lens plane by assigning the designed GST crystallization levels to the corresponding slits, achieving various farfield focusing patterns.
In 2007, Min et al. investigated a type of metallic nanooptic lens consisting of slits with variant widths, filled with Kerr nonlinear media [139]. As shown in Figure 23(a), each slit is designed to transmit light with specific phase retardation controlled by the intensity of incident light, owing to the nonlinear response. This new lens can actively control the deflection angle and the focus length of output beam. Very recently, Hu et al. proposed an active terahertz (THz) plasmonic lens tuned by an external magnetic field [140]. Different from traditional tunable devices, the proposed active lens is tuned by changing the cyclotron frequency through manipulating magnetoplasmons (MPs). It was shown that THz wave propagating through the designed structure could be focused to a small size spot via the control of MPs. The tuning range of the focal length under the applied magnetic field (up to 1 T) is ~31, about 50% of the original focal length.
(a)
(b)
(c)
(d)
(e)
It should be noted that the plasmonic slit lens is polarization selective; that is, only one particularly polarized light beam could be transmitted and modulated in phase, since the SPP is intrinsically polarized. This polarization selectivity has two effects. Firstly, this effect is attractive for some cases where only one polarization type is needed. Secondly, it may become undesired for many other applications. To eliminate this problem, rectangular or circular plasmonic holes were adopted to achieve phase modulation [136, 141]. As illustrated in Figure 24, Ishii et al. experimentally demonstrated a polarizationindependent holeymetal lens [136]. They showed that by changing the radii of subwavelength holes in a metallic film, which act as singlemode waveguide elements, they can control the phase of light transmitted through the holes. Finally, it was demonstrated that the focal distance of the lens can be controlled by changing the incident wavelength.
(a)
(b)
(c)
More recently, Litchinitser et al. demonstrated that the change of hole radius along the azimuthal direction is able to generate orbital angular momentum (OAM) on the nanoscale. They proposed and experimentally demonstrated that a nanowaveguide array milled in a metal film can be used to control the wavefront of a light beam and that an optical vortex at 532 nm was produced by using such an array (see Figures 24(b) and 24(c)).
3.5.3. Plasmonic Metasurfaces for Sensing Applications
Owing to the strong lightmatter interaction, metasurfaces can provide a robust and efficient platform for biosensing. In particular, the surface plasmon resonance (SPR) at planar surfaces or localized surface plasmon resonance (LSPR) for nanometersized metallic structures is accompanied with dramatic local field enhancement [142]. The implication of the local field enhancement is twofold. Firstly, the spectra of the SPR and LSPR are highly dependent on the environment; thus, a little change of refractive index would result in a considerable spectrum shift. Secondly, the local field enhancement can be exploited to achieve surfaceenhanced Raman spectroscopy (SERS) [143]. Both of the two mechanisms can be utilized for clinic detection.
As shown in Figures 25(a)–25(d), Wu et al. introduced an infrared plasmonic surface based on a Fanoresonant metasurface which exhibits sharp resonances caused by the interference between subradiant and superradiant plasmonic resonances [144]. Owing to the asymmetry, the frequency of the subradiant resonance can be precisely determined and matched to the molecule’s vibrational fingerprints. In these Fano resonances, the nearfield coupling of plasmonic modes is crucial to enhance the sensitivity [145, 146]. It was also demonstrated that Young’s interference can be observed in plasmonic structures when two or three nanoparticles with separation on the order of the wavelength are illuminated simultaneously by a plane wave. This effect leads to the formation of intermediatefield hybridized modes with a character distinct of those mediated by nearfield and/or farfield radiative effects [147].
(a)
(b)
(c)
(d)
(e)
(f)
In a similar work [148], by exploiting extraordinary light transmission phenomena through high quality factor subradiant dark modes, Yanik et al. experimentally demonstrated high figures of merits (FOMs as high as 162) for intrinsic detection limits surpassing those of the gold standard prism coupled surface plasmon sensors, as illustrated in Figures 25(e) and 25(f).
Over the past several years, the absorption effect of plasmonic metasurface has also been exploited in sensing applications. In 2011, Tittl et al. presented a simple design of plasmonic absorber based on palladium nanowires [149]. Due to hydrogen incorporation, palladium undergoes a phase transition from a metal to a metal hydride which leads to an expansion of the palladium lattice. The fabricated structure showed a reflectance of 0.5% which in combination with a complete suppression of transmission yields an absorbance of . As shown in Figures 26(a)–26(d), they utilized the absorber structure for hydrogen sensing and were able to reliably detect concentration down to 0.5% H_{2} in air with response time in the range of seconds.
(a)
(b)
(c)
(d)
(e)
(f)
More recently, it was demonstrated that graphene monolayers transferred on arrays of split ring resonators (SRRs) could exhibit resonances in the visible range. As illustrated in Figures 26(e) and 26(f), Raman enhancement factors per area of graphene up to 75 were measured, demonstrating the strong plasmonic coupling between graphene and the metasurface resonances.
Recently, an important sensing technique via multifrequency antennas was also demonstrated [151–155]. Such antennas can efficiently detect different vibrational modes of molecular species in a window of several micrometers, where informative fingerprint spectra of many molecules are present.
Since there have already been many reviews on this topic [34, 142, 143, 156], we only give a small part of examples here. To completely understand the role of metasurface in sensing applications, we would like to suggest that the readers find more information in these literatures.
3.6. Surface PlasmonInspired Metasurfaces
In previous discussion, metasurfaces refer to arrays of subwavelength structures on thin films. More generally speaking, any kinds of subwavelength structures on such films can be regarded as metasurfaces, with novel applications in optics and electromagnetics. Early in 1998 [157], it was found that extraordinary light transmission can be observed when it passes through the subwavelength metal pinhole arrays (Figure 27). Also, adding a surface groove structure around the subwavelength aperture can break the traditional diffraction limit and achieve highly directional energy transfer with ultrahigh transmission [158].
(a)
(b)
(c)
(d)
Subsequently, Poujet et al. firstly demonstrated an experimental extraordinary optical transmission up to 90% in the subwavelength annular apertures [159]. In spite of the metal loss, the transmission of incident light has been greatly enlarged by employing the guided mode propagating through the annular aperture array. The subwavelength hole arrays can also act as color filters as the wavelength selectivity is directly related to the period of the holes [65]. An array of dimples is prepared by focusedionbeam milling an Ag film. Some of the dimples were milled through to the other side so that light can be transmitted. When this structure is illuminated with white light, the transmitted colour is determined by the period of the array. In this case, the periods were chosen to be 550 and 450 nm, respectively, to achieve the red and green colours.
Afterwards, several research groups including MartinMoreno et al. [160] analyzed the underlying physics principle of generation of beaming effect in groove surface and pointed out that transmission enhancement of the groove is attributed to three factors as follows: surface plasmon resonance modes in the grooves, phase match of groove secondary radiation, and waveguide modes in the slot. Considering the SPs in metallic structure at the visible range, Luo and Yan established a physics model, namely, quasiperfect conductor model (QPCM), to optimize the directed radiation characteristics of SPPs [34]. In addition to beaming effect, the multidirectional radiation and abnormal homogeneous radiation phenomenon of subwavelength groove structure were also systematically discussed based on the diffraction theory.
More recently, the extraordinary transmission and beaming effect based on cycle groove structures were also expanded to other frequency bands. Pendry et al. demonstrated that even a perfect electric conductor can support surface plasmon [56]. In order to distinguish it from traditional SPs, this surface plasmon supported by groove structures in low frequency is named spoof surface plasmon (SSP). Lockyear et al. [161] demonstrated that microwave energy could be coupled into SSPs through the reflective spectrum experiment of grating structure. Lately, AkarcaBiyikli et al. [162] experimentally validated the abnormal transmission and beaming effect in the 1D groove structure. With the beaming effect and extraordinary transmission, groove structures have wide potential applications in various frequency regions, as shown in the following discussion.
3.6.1. Surface PlasmonInspired Metasurfaces for Microwave Antennas
The directional energy transfer realized in groove structure meets the requirements of highdirectivity antennas and a variety of antennas have been proposed in the microwave band. For example, Huang et al. [163] used periodic groove structure in linearly polarized waveguide end slot array antenna. As shown in Figures 28(a) and 28(b), the beam angles in plane and plane were compressed to be 6 and 13 degrees, respectively. The groove shaped metasurfaces also have great contribution in reducing the level of sidelobes of slot antennas. Huang et al. proposed a method of integrating groove structures with two different periods into slot antenna to obtain low sidelobe radiation (Figures 28(c) and 28(d)). The sidelobe of the antenna with grooves was greatly decreased compared to the slot array without groove structures. Furthermore, after adding the artificial electromagnetic softsurface structure to onedimensional subwavelength outer groove, the antenna backward radiation was significantly reduced by 10 dB. However, the antenna gain was only increased by 0.2 dB due to long distance between the edge groove and the central feed.
(a)
(b)
(c)
(d)
(e)
(f)
For the characteristics of low cost, light weight, and low profile, microstrip patch antenna has gotten a great prospect on the area of military and civilian applications. However, due to the deterioration of surface wave, the antenna’s radiation efficiency is very low, and this restricts its development. Ying and Kildal [164] improved the radiation performance of the antenna through an artificial electromagnetic softsurface groove onto a ground.
The groove structures can also be loaded between slot elements to achieve low coupling between the slots and thus obtain high aperture efficiency of the slot antenna arrays. The crosscoupling between the two slots was sharply decreased by 30 dB after loading the groove structure, as shown in Figures 28(e) and 28(f). Furthermore, after introducing groove structure, the gain of slot antenna array was enhanced by 4 dB, and the sidelobe level on plane is low and main lobe on plane is compressed.
The period of a traditional beaming effect based groove structure is approximate to the wavelength, and the beaming effect usually requires several periods to implement. Thus, they make the size of the whole antenna very large and bring about the low utilization of aperture. Therefore, with no sacrifice of the antenna performance, how to reduce the groove cycle has formed a research direction.
Huang et al. have still taken account for the loading of high dielectric constant medium between the grooves to reduce the groove cycle equivalently. Both single slot antenna and twodimensional slot antenna array loaded dielectric medium with high permittivity were analyzed. The operating frequency of this antenna is 14.5 GHz, and the rectangular waveguide end slots were used as the exciting source. The dielectric constant loaded between the two neighboring grooves was 11.9, and the thickness of the loaded media silicon was 1.1 mm. The introduction of the traditional cyclegroove structure has no effect on antenna’s resonant frequency, while the mediumloading groove structure makes the center resonant point of the antenna shift to the low frequency, and it has a certain impact on antenna input matching. However, the slot antenna can be kept at < − dB by tuning the slot length in the working frequency, meeting the needs of practical engineering. The gain of a traditional cyclegroove slot antenna is 15.11 dB, increasing by about 9 dB compared to the flatpanel slot antenna. However, the antenna gain has further improved by 3 dB after the mediumloading groove structure is used, reaching 18.29 dB. It is notable that the physical diameter of the designed antenna is smaller than the size of a traditional cyclegroove slot antenna, only threequarters of it, but the gain has improved more significantly instead. From the radiation patterns, we can also find that the plane of a traditional cyclegroove antenna shows an obvious beaming effect.
This mediumloading method can also increase the performance of twodimensional slot antenna array. This idea broke through the limit that the traditional array element interval is less than wavelength to widen applications of the periodic groove structure. Afterwards, media with high dielectric constant were loaded between the grooves to reduce the groove cycle equivalently without sacrificing of the antenna performance. This method featured in smaller periods compared to the traditional periodic groove structures, which results in reduced antenna aperture and inspiring more surface wave to improve further antenna gain in the performance of the antenna radiation. Furthermore, it compresses the beam angles of sides and to illustrate prominent directed radiation ability. So it can be considered that the application of this new periodic groove structure was a development direction of highdirectionality antenna. It was the first time to use a periodic groove structure as a secondary source and load it between the slot array elements with interval greater than wavelength to avoid sidelobe. This idea broke through the limit that the traditional array element interval is less than wavelength to widen the applications of the periodic groove structure.
3.6.2. Surface PlasmonInspired Metasurfaces for Applications in Terahertz and MidInfrared Regions
One of the disadvantages of the grooves for directive radiation is that the working bandwidth is limited. Nevertheless, this is not a series problem for some applications. For instance, almost all kinds of lasers are narrowband in spectrum. In this regard, the groove structures can be integrated to lasers in order to improve their performances. By defining a metallic subwavelength slit and groove array on the facet of quantum cascade lasers (QCLs) [165], the divergence angle in the laser polarization direction is only 2.4 degrees as shown in Figure 29. Compared to the original 9.9 μm wavelength laser without a groove array, a reduction in beam spread by a factor of 25 is achieved, without significant reductions in output power. Taking advantage of SSPs, groove array integrated terahertz QCLs were also demonstrated [166]. The beam divergence of the lasers was reduced from ~180 degrees to ~10 degrees, the directivity was improved by over 10 decibels, and the power collection efficiency was increased by a factor of about six compared to the original unpatterned devices. In addition, multibeam QCLs and MIDIR QCLs with integrated plasmonic polarizers were reported through combining the groove structures with QCLs by Capasso’s group [167].
(a)
(b)
(c)
(d)
(e)
(f)
3.6.3. Surface PlasmonInspired Metasurfaces for Applications in the Visible and NearInfrared Regime
The intriguing properties of plasmonic waves at metasurfaces can also be utilized to enhance the directivity of classically noncoherent or notdirective radiations. For example, spontaneous emission of fluorescent molecules or quantum dots is radiated along all directions when emitters are diluted in a liquid solution, which severely limits the amount of collected light. Making use of the directional manipulation of the SPinspired metasurface, Aouani et al. [168] successfully controlled the radiation properties for nanoemitters, as shown in Figure 30. Enhancing the fluorescence intensity and narrowing the emission directivity were simultaneously obtained, and a full control of fluorescence was achieved. However this control is static. In 2011, active control over fluorescent emission was also reported by electrically pumping the similar lightemitting device [169]. This type of device facilitates the realization of a new class of active manipulation for use in new optical sources and a wide range of nanoscale optical spectroscopy applications.
(a)
(b)
(c)
(d)
(e)
(f)
In addition to the directional manipulating and transmission enhancing, the applications of groove type SPinspired metasurface were greatly expanded recently. Orbital angular momentum (OAM) transfer [170] and wavefront shaping [171] were experimentally demonstrated. Chiral plasmonic grooves defined on a metallic film would transform a normal plane wave to the vortex beams with tunable topological charge, as shown in Figures 31(a)–31(c). By surfacewaveholography method, Chen et al. shaped the wavefront of the incident nearinfrared light into predesignated complex patterns such as Latin letters, after passing through a 180 nm radius hole that is surrounded by welldesigned groove patterns, as depicted in Figures 31(d)–31(g).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
More recently, we provided a scheme to guide and collimate OAM at the micro and nanolevels [172]. The coaxial plasmonic waveguide is exploited as a naturally occurring medium for light carrying OAM to transfer. The guided OAM wave is coupled to free space through corrugated grooves surrounding the coaxial waveguide, where coherent scattering of spiral surface plasmon was demonstrated to be responsible for the huge enhancement of beam directivity. Experimental results at nm validated the nearfield transportation of OAM beams, where the topological charge is tunable via the modulation of a liquid crystal spatial light modulator (SLM).
3.7. Passive Metasurface Antennas and Generalized Snell’s Law
In many cases, metasurfaces can be considered as passive antennas array. By tuning the geometries and material parameters, the amplitude, phase, and polarization state of the reradiation of these antennas can be fully controlled. In this section, we first review two kinds of phasetype metasurfaces, where only the phase is tuned. Then, the possibility of full control over the phase, polarization, and amplitude is discussed.
3.7.1. Passive Antenna Array Based on Impedance Transition
The impedance boundary conditions of metasurfaces can be utilized to tailor the phase delay within one single layer regardless of its thickness. This means that, theoretically, there exists no thickness limit for this problem. Based on the rigorous form of Huygens’ principle developed by Love and Schelkunoff [173, 174], it has been demonstrated that the transmitted or reflected phase shift can vary between 0 and to provide complete phase coverage by adjusting the magnitude of the impedance [19, 175, 176].
As depicted in Figure 32, the concept of Huygens’ metasurface was proposed by Pfeiffer and Grbic in 2013 [176] and realized with twodimensional arrays of subwavelength structures that provide both electric and magnetic polarization currents to generate the prescribed wavefronts. The applications of the metasurfaces in a beamrefracting surface and a GaussiantoBessel beam transformer were discussed.
(a)
(b)
(c)
(d)
(e)
(f)
As an attempt to increase the energy efficiency and bandwidth, we presented an efficient approach, as depicted in Figure 33, to extend the bandwidth of phase modulation by utilizing the broadband characteristic of lowquality factor (Qfactor) metasurface in the reflection mode [177]. The dispersion of the metasurface was engineered to achieve phase modulation within . Anomalous nearly perfect reflection with relative bandwidth near 40% was demonstrated in the microwave regime. Similar designs in reflection geometry have also been validated by other authors [178–180].
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
3.7.2. Passive Antenna Array Based on Ultrathin Resonators
In the phaseengineering regime, one interesting but sometimes neglected phenomenon is that a rotation of electric field of circular polarized electromagnetic wave results in a corresponding phase shift. As early as 1955, the rotatory phase shifter was proposed by Sichak and Levine [182]. They showed that the phase of the output voltage of a circularly polarized antenna was proportional to the angle of rotation of the antenna about its longitudinal axis. Two circularly polarized antennas with a proper rotation angle between each other were mounted in a circular waveguide to construct a phase shifter. In 1961, this rotatory phase was utilized in the construction of large radio telescope in the University of Illinois [183]. Subsequently, similar phase shifting results were demonstrated for passive reflectarray antennas [184].
In 1984, Berry proposed that an adiabatic polarization can introduce a phase shift [43], which was further experimentally demonstrated in 1988 [185]. Since the phase is associated with circular polarization, it can be termed as one kind of optical spinorbit interaction. In 1999, polarization gratings were proposed [186]. This idea was further developed by Hasman et al., with various kinds of applications, such as beam splitter, optical vortex generation, and focusing lens [44, 187–189]. In 2006, the famous qplates were proposed to realize spintoorbital angular momentum conversion [190]. These qplates can also be used as polarization converters [191]. In recent years, nanoantennas were widely adopted to control circularly polarized waves [192–195].
In 2011, Yu et al. proposed another phase shifting mechanism with linearly polarized illumination. As shown in Figure 34, by varying the angle and length of Vshaped nanoantennas, arbitrary modulation of phase shift and amplitude transmission were demonstrated. The operational principle of such Vshaped antennas relies on the asymmetric transmission enabled by the nanoantennas. Similar to the case of circular polarization, such antennas have low energy efficiency because there always exists one component with the same polarization as the incident beam. Based on this abrupt phase change, they proposed the socalled “generalized Snell’s law.” By properly tuning the geometrical parameters of each antenna, broadband phase change could be achieved, although the phase was not rigorously achromatic [134, 138, 196].
After the pioneering work by Capasso, much work was devoted to the design of phasediscontinuity metasurfaces. Many optical devices were subsequently designed and experimentally demonstrated. For example, Zhang et al. proposed that Cshape metallic elements can also be utilized to tune the amplitude and phase of crosspolarized scattering for linear polarization incidence [196]. In addition to the simple phaseengineering, other applications, such as spiral phase plates and wave plates, were also demonstrated [197, 198].
Based on the locality of the phase shift [199], Lin et al. extended the Vshaped antennas in the nearinfrared bands. The focus length was also decreased to a large extent to obtain a focus spot close to the diffraction limit. Jiang et al. demonstrated similar effect in the terahertz regime [200]. More recently, Ma et al. used a metasurface to focus vortex beam into a nanodoughnut [201]. As shown in Figures 35(a)–35(d), the proposed metasurface has good performance close to the ideal case. The metasurface tuned the phase distribution of the incident beam based on the concept of discontinuous phase modulation. Consequently, the proposed metasurface has two main functions. Firstly, with a circularly polarized incident beam, the metasurface converts it into a beam carrying OAM. Secondly, the incident field can be focused into a point with a certain focus length. The total energy coefficient of this metasurface exceeds 40%. By properly designing the phase modulation distribution, the copolarized component of the incident field does not participate in the focusing process. Therefore, only optical vortex can be observed in the focused field. The thickness of this planar chiral metasurface is much less than the operational wavelength and thus has potential application in compact integrated optical systems.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
The geometric phase in aperture array has direct application in holography. As demonstrated in Figures 35(e)–35(h), metasurface with metallic square apertures was constructed in our lab to form a hologram image of a China map based on the concept of discontinuous phase modulation in visible band.
Since only part of the transmitted light takes geometric phase, single layered metasurfaces are often characterized by low efficiency. In principle, the efficiency could be increased dramatically with the help of a reflective layer. Taking the circularly polarized geometric phase as an example, the combination of an anisotropic metasurface with a reflective layer would form a reflective wave plate. By rotating the main axis of the embedded antennas, gradient phase could be directly obtained. This technique was utilized in reflectarray to increase the gain of antenna while keeping the reflective surface to be flat [184]. In the optical regime, the reflective nanoantennas array was used as building blocks of high efficient holograms [202]. The metasurface has an ultrathin and uniform thickness of 30 nm and is compatible with scalar diffraction theory, even with subwavelength pixel sizes, thus simplifying the design of holograms.
Another important application of the metasurface relies on the fact that phase shifting metasurfaces placed at the optical pupil plane can increase the resolution of telescopes, which means that the farfield diffraction limit can be overcome [203]. As shown in Figure 36, the designed metasurface has several rings, which exert phase shift of 0 and on the incoming light and change the corresponding spatial frequency [2]. To characterize the performance, we measured the images of two circular and triangular holes with and without the metasurface. After the insertion of the metasurface, the resolution is decreased to 0.625 times of the original value, implying that a 10meter telescope could have the same resolution as that of a 16meter one (i.e., a 1.6 times improvement in the resolution power).
(a)
(b)
(c)
Very recently [204], we proposed and demonstrated the concept of broadband virtual shaping at the visible, infrared, and microwave spectrum by tailoring the spatiotemporal property of spinorbit interaction in cascaded metasurfaces. When electromagnetic waves impinge on the designed metasurface, they are reflected to predefined directions to avoid being detected by unwanted antennas. Resorting to the dispersion engineering techniques in metasurfacebased polarizers, the bandwidth was dramatically enhanced, although the thicknesstobandwidth ratio may be analogous to that of broadband absorber. The design principle provided a new route for the control of electromagnetic wave for applications ranging from laser beam shaping to 3D holographic display and conformal camouflage [205].
As shown in Figures 37(a) and 37(b), we measured the specular reflection of an inhomogeneous metasurface. Since the reflected light was designed to depart from its original direction via the metasurfaceassisted law reflection, only a small reflection echo was measured. Owing to the flexibility of the metasurface, this approach could be utilized in complex objects. For example, the RCS of a cylinder can be dramatically reduced by covering our metasurfaces on it (Figure 37(c)). The metallic cylinder in our design has a radius of = 90 mm and height of = 360 mm. The geometric phase distribution on the metasurface was designed to be , where rad/m. As shown in Figure 37(d), the RCS reductions for TE polarization and TM polarization were calculated under normal incidence along the direction. We noted that the reduction amount is a bit smaller than the planar case, which possibly stems from the nonoptimized phase profile and the fact that the RCS of a cylinder itself is smaller than its planar counterpart. Nevertheless, the conformal metasurfaces provide a new sight into the virtual shaping of nonplanar objects. In addition, this approach can be exploited to increase or decrease the RCS.
(a)
(b)
(c)
(d)
In general, the metallic metasurfaces suffer from low efficiency due to the intrinsic ohmic loss and the limited scattering cross sections of the antennas, which is not anticipated in many realms. Recently, dielectric metasurfaces based on geometric phase were presented to control the phase [44]. Nevertheless, the discrete phases lead to degradation of the performance. To overcome this issue, we proposed a semicontinuous structure to increase the continuity of phase shift. The metasurfaces consist of highrefractiveindex and lowloss silicon annual ring gratings. Ohmic loss is avoided compared to its metallic counterpart. The continuity of annual ring grating ensures the continuity of the transmitted field, thus suppressing high order diffraction effects that may arise from discontinuity.
At the end of this section, we would like to comment that metasurface with geometric phase could enable many applications in distinct disciplines. In general, traditional refractive, reflective, and diffractive devices could be replaced by metasurfaces. These elements include the spherical lenses, aspherical lenses, parabolic reflectors, fanout grating, and laser beam shapers. In the microwave regime, metasurfaces could also be utilized to substitute traditional phasedarray antenna. Furthermore, this kind of metasurface is also able to shift the momentum of scanning beam to extend the scanning range of phased array antennas, when it is integrated in radar radomes [206].
3.7.3. Full Control of Phase, Polarization, and Amplitude
In many applications, the phase, polarization state, and amplitude are all important for the manipulation of electromagnetic waves. For example, a low sidelobe phased array antenna needs the full control of phase and amplitude [207], while a planar lens with large longitudinal fields in the focus spots requires that the polarization and phase be simultaneously controlled [208].
In order to achieve complete modulation of phase and polarization, Pfeiffer and Grbic proposed a fourlayer cascaded metasurface lens that can focus light and control its polarization [209]. By adjusting the dimensions of structure, one can independently adjust the inductance and capacitance in the  and directions so that the phase and polarization can be completely controlled.
Kim et al. demonstrated that the local reflection coefficients of gapplasmon resonators can be independently tailored in both magnitude and phase [181]. Using an array of nanorods closely placed on a metallic layer, it was shown that the magnitude and phase of the surface’s local reflection coefficients can be simultaneously tuned, with a magnitude range of and a phase range of 360°. The power efficiency can reach 45% for an operational wavelength of 800 nm. This method was utilized in the design of a DolphTschebyscheff array, with equalamplitude sidelobes of prescribed magnitude, as represented in Figures 38(a)–38(d).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
More recently, Li et al. proposed and validated another metasurface that can provide simultaneous manipulation of the phase and polarization of the transmitted light [210]. SPPs were adopted to obtain a phase shift when the top and bottom nanoapertures are laterally translated by a distance in the direction. Hence, the phase difference can cover 0 to by varying the parameters of nanoaperture for normally incident light polarized in the direction. The metasurface enabled anomalous refraction with efficiency of about 21.4% for linearly polarized incident light (Figures 38(c)–38(g)).
It also should be noted that the phase over the surface of micro and nanoantennas can also be measured with the help of scattering nearfield techniques [211, 212], which provide a way to further understand the physical processes of these antennas.
3.8. Active Metasurfaces
Although passive metasurfaces have been proved to be useful in many applications, the dynamic tuning of performance is always desired in practical applications. Most of the metasurfaces have fixed electromagnetic properties when the structures are designed or fabricated. Nevertheless, there is a type of “reconfigurable” metasurface which is able to tune the electromagnetic response with an outside stimulus [213–217]. The general way to obtain the tunable phase modulation is by loading active devices in its unit cells, such as varactor diodes [218], liquid crystals, and ferroelectric and microelectromechanical switch (MEMS) [219], as can be seen in Figure 39.
(a)
(b)
(c)
(d)
These reconfigurable or tunable metasurfaces efficiently broaden the application range of metamaterial and simplify the optical and communication systems. According to the particular applications, active metasurfaces can be classified into three main subdomains. In the following, we would give a brief review on these novel metasurfaces.
3.8.1. Active Phase Modulation
Aiming at satisfying the requirement of modern communication systems and solving the disadvantages of traditional phased array antenna, metasurfaces with phasetuned ability have been introduced to construct novel beam scanning antennas. The principle of beam steering for this metasurface is illustrated in Figure 40(a). The designed active metasurface is divided into regions, which can be regarded as a linear array with a period of . When a plane wave transmits through these regions, the phase difference between adjacent regions would be . As a result, the angle of beaming direction is deviated from the normal direction, which is determined by .
(a)
(b)
(c)
(d)
(e)
(f)
In Figure 40(a), the transmission phase of region can be shown aswhere is a constant and is an integer. The phase shift and the phase difference can be expressed aswhere is the wave number. The beam steering angle can be then written as
Clearly, (43) is one particular case of the metasurfaceassisted Snell’s law, based on which the beaming direction can be tuned by changing the phase difference between adjacent elements. In 2012, Jiang et al. [206] introduced an active cascaded metasurface for phase manipulation by tuning varactordiodes loaded in the unit cell. The tunable metasurface was employed as the radome to tune the plane radiation pattern of the antenna array. However, the loss of this active phase modulation metasurface is relatively large due to its eight layered structure.
Based on the phase modulation property of the metasurface, we have proposed an active phase modulation metasurface to manipulate the phase of the incident wave and to achieve beam steering of a horn antenna [221], as depicted in Figure 40(d). The unit is composed of a metallic rectangular ring and a patch, with a pair of microwave varactor diodes inserted in between along incident electric field polarization direction. Transmission phase of the emitted wave can be tuned by changing the bias voltage applied to the varactor diodes, and 360° phase difference can be obtained at 5.3 GHz as the capacitance value is tuned from 0.65 pF to 2.6 pF.
Through different configurations of the bias voltages, one can obtain the gradient phase distribution of the emitted wave along plane and plane. This metasurface loaded antenna can steer the directive beam with an angle of ±30° in both plane and plane at 5.3 GHz with a bandwidth of 180 MHz. Using an improved approach, the scanning range can also be increased to ±360° in both plane and plane [2].
In previous works, the polarization state and beam direction are not simultaneously controlled. In order to achieve this goal, a novel transmitarray element was proposed, which achieved 1bit phase shifting at two orthogonal linear polarization modes. The phase tuning and polarization reconfiguration can be independently controlled based on PIN diodes. The unit cell of the dual linearly polarized metasurface consists of twolayer metallic patterns connected by a metalized viahole. Onelayer metallic pattern is a rectangular patch with two PIN diodes loaded in Oslot along electric field polarization direction, which is utilized as a receiverantenna to achieve 1bit phase tuning. The other metallic pattern is a dual linearly polarized transmitterantenna that adopts a square ring patch with two PIN diodes distributed at the crosspolarization directions. The simulation results show that the designed antenna can achieve 1bit phase tuning and linear polarization reconfiguration at 10.5 GHz with insertion loss of about 1.1 dB. This kind of the metasurfacebased antenna has the advantages of low cost and simple implementation, which could be developed for possible applications in some communication and radar systems, especially in the area where highgain beam steering and the antiinterference capabilities are urgently required.
3.8.2. Active Polarization Modulation
The polarization states can also be dynamically tuned by active metasurfaces. Zhang et al. [215] proposed a reconfigurable chiral metasurface, which switched its handedness when the semiconductor region was illuminated by external light, as shown in Figure 41. This reconfigurable or tunable property efficiently broadens the application range of metasurface and simplifies the optical and communication systems.
(a)
(b)
(c)
(d)
(e)
(f)
(g)
In order to achieve polarization conversion with simple metasurface, Ma et al. [92] demonstrated a doublelayered active metasurface which changes the structure between chiral and isotropic by tuning the working states of the PIN diodes in its unit cells. When it presented chirality, the handedness could also be tuned. As shown in Figures 42(a) and 42(b), the designed metasurface could convert the incident linearly polarized wave into RCP or LCP one. Furthermore, it could also be switched to be isotropic and keep the polarization state of the incident wave unchanged, by controlling the voltages applied on these diodes.
(a)
(b)
(c)
(d)
(e)
(f)
We have also employed this concept of active polarization manipulation into the design of antennas. As it is well known, the dipole cannot radiate efficiently near a perfect conductor due to its opposite image current; in addition, the single dipole can only produce linearly polarized waves, which cannot satisfy the requirement in many communication fields. To address the above issues, the anisotropic metasurface was proposed as the ground plane for the dipole. This novel reflective metasurface not only reduces the profile of the dipole but also assists it to achieve emission of circularly polarized wave. Through adjusting the PIN diodes loaded on the unit cell of the metasurface, the polarization states of the emitted wave from the dipole, including RHCP, LHCP, and LP waves, can be dynamically tuned.
The tunable metasurface can make the dipole antenna work at four states. For states 1 and 2, as seen in Figure 42(d), the emitted waves with left and righthand circular polarization are produced at 4.6 GHz, and the AR is 1.2 and 0.3 dB, respectively. For states 3 and 4, the dipole antenna radiates the linearly polarized waves with the gains of about 6 dB at 4.6 and 4.25 GHz (see Figure 42(f)), respectively. This kind of the polarization reconfigurable antenna has a strong environmental adaptability, which may find many potential applications in the communication field.
Recently, Wang et al. demonstrated a coherent perfect polarization transformer [76]. It was proved that the phase shift between two coherent inputs could be utilized to dynamically tune the polarization states of output waves. Depending on the phase difference, the outputs could be polarization, LCP, and RCP for polarized inputs.
3.8.3. Active Amplitude Modulation
Besides the phase property of the incident electromagnetic wave, the amplitude can also be modulated. As analyzed in the effective impedance theory, the transmitted amplitude and the phase of the incident wave are modulated by the effective impedance of the metasurface. Therefore, the element loading method in metasurface allows the control of the amplitude and hence provides new applications in microwave communication systems, such as controlling the sidelobe and beamwidth of the radiation antennas.
Lim et al. [222] had designed a kind of composite right/lefthanded microstrip antenna in which varactors were used to tune the amplitude of each element for achieving dynamic control of the beamwidth. Recently, a partially reflective surface (PRS) was utilized to actively change the beamwidth of the emitted waves by tuning its reflection through the varactors mounted on each element of the PRS [223].
Antennas with low sidelobe level (SLL) are urgently required since it can efficiently reduce the electromagnetic interference and improve the ability of signal capture. However, traditional approaches to achieve low SLL are mainly realized by modifying the elements weighting or spacing distribution in antenna array. In comparison, metasurfaces with amplitude tunability were employed to reduce the SLL of antennas since the radiation pattern can be reformed by metasurfaces.
Our team proposed an approach to reduce the SLL of antenna array to a desired value by utilizing amplitude modulated metasurface, as shown in Figure 43. By manipulating the structure parameters and the PIN diodes loaded in the unit cells, the transmission amplitude of each unit can be distinctly controlled. Then, the metasurface was utilized as the superstrate of the horn antenna array, and the superstrate was divided into several regions with different transmission amplitudes along the electric field polarization direction. With a plane wave illustrated on the superstrate, the transmission amplitudes were designed as Taylor distribution while the transmission phases were equal, which resulted in a tapered amplitude distribution of the output beam. An plane sectorial horn array is adopted as a plane wave source to verify the performance of the SLL reduction. The SLL was reduced from −12.4 to −25.9 dB at the desired frequency, while not influencing the gain of the antenna array.
(a)
(b)
(c)
(d)
(e)
(f)
Wang et al. proposed a kind of metasurface superstrate to control the beamwidth of horn antenna array based on the concept of amplitude modulation [225], as depicted in Figure 44. The unit cell of the metasurface was composed of twolayer periodic structures loaded with varactors. By tuning the capacitance of the varactors, the unit cell can produce two different states which are transparent and absorptive to the incident wave at a certain frequency, respectively. The whole metasurface can be divided into several groups, and each group is composed of some unit cells with the same state. By switching the state of each group, the transparent window of the metasurface can be changed. Consequently, the emitted waves with variable beamwidth can be expected, as validated in Figure 44(c), when this metasurface was placed above the horn antenna array as superstrate.
(a)
(b)
(c)
(d)
(e)
(f)
As another method to modulate the amplitude, metasurfaces partially absorbing ability has been introduced and employed in the antenna realm. Compared to the traditional means, such as shaping of target surface and adopting radar absorbing material, such metasurface may efficiently reduce the RCS while not decreasing the radiation performance of the antenna. We proposed a design to reduce the RCS and enhance the gain of a patch antenna by using partially reflecting surface (PRS) that consists of two layers of metallic structure [58], as demonstrated in Figure 44(d). The top absorbing layer is utilized to reduce RCS, while the bottom layer was combined with the ground plane to construct the FP cavity to achieve high gain. The antenna gain was enhanced by about 6.5 dB at 11.5 GHz, and its RCS is dramatically reduced in a broad frequency range from 6 to 14 GHz. It was fully verified that the designed PRS did not influence the radiation performance of the antenna within the bandwidth of the RCS reduction, but it also obviously improved the antenna gain. This new design provides a good method to solve the conflict between the gain enhancement and the RCS reduction.
In the above amplitude modulation metasurface, the normalized amplitude can be controlled in the range of 01. However, the amplitude of the incident wave can be gained above unit by utilizing amplifier in the unit cell of the metasurfaces. Recently, we also proposed an amplifying reconfigurable metasurface to achieve simultaneous control of the amplitude and phase of the incident wave. The varactor diodes were adopted to construct the reflectiontype phase shifters, which achieved over 400° of phase shift extending the operation bandwidth beyond 10%. Most importantly, the amplifier integrated in the unit cell of the metasurface made the amplitude of the whole element simultaneously gain larger than 7 dB. In addition, the overall thickness of the proposed metasurface element is only in free space, which is more compact compared to the existing metallic elements. This kind of amplifying and phase modulation metasurface could be developed to construct highgain antenna for beam steering.
3.8.4. Active Absorption Modulation
In order to satisfy the requirement of modern military and civil applications, novel materials with smart absorbing characteristics are needed. In general, active metasurface absorbers have two advantages over traditional absorbers. Firstly, since traditional metamaterials absorbers are limited in the operational bandwidth, active tuning may increase the effective bandwidth in a timediving way. Secondly, the absorption coefficients can be dynamically tuned to mimic the electromagnetic spectrum of environment, which is not obtained through traditional approaches.
The intrinsic circuit property of metasurface absorbers permits an easy way to realize active tuning. As shown in many references, the metasurface can be characterized by its equivalent resistances, inductances, and capacitances [99, 114]. When lumped circuit devices, such as resistors, inductors, capacitors, and various kinds of diodes, are integrated in the metasurface, the electromagnetic response can be greatly improved.
Early in 2004, Tennant and Chambers presented an adaptive absorber by controlling the bias voltage of PIN diodes loaded between adjacent metallic bow ties [213]. The absorber was similar with the topology of a Salisbury screen, but in which the conventional resistive layer was replaced by an active metasurface controlled by PIN diodes. The resulting structure had superior absorption characteristics compared to conventional passive absorbers of corresponding thickness. Measured results showed that the reflectivity response of the absorber can be controlled over the frequency band from 9 to 13 GHz. In a similar way, other circuit elements, such as varactors [226, 227], were introduced to change impedance of the metasurfaces.
In 2012, an electrical tunable Lband absorbing material for two polarization types was presented by Wang et al. [228]. As shown in Figures 45(a)–45(c), the proposed absorber consisted of a metal resonator and four surrounding metal lines which were connected by pin diodes and exhibited the tunable range of reflectivity reaching −40 dB for both polarization types. Later [227], Zhao et al. demonstrated a tunable absorber with a measured bandwidth of 1.5 GHz (or relative bandwidth of 30%). Since the units are placed along two orthogonal directions, the absorber is insensitive to the polarization of incident waves.
In principle, an active absorber should have the ability of controlling echo characteristics in both aspects of frequency and intensity. However, few active absorbers simultaneously considered the two issues in a single device. As an attempt to accomplish this purpose, Wu et al. demonstrated an electrically active absorber in which the working frequency and absorbing intensity could be separately controlled by the PIN and varactor diodes [229]. The structures and measured results are presented in Figures 45(d) and 45(e). Compared with those metasurfaces proposed before, although at the expense of increased complexity, the interesting characteristic of the absorber designed here is that it contains a special designed active magnetic resonator in which both the resonant frequency and amplitude can be, respectively, tuned and therefore shows separable modulation of absorbing frequency and peak intensity. Both simulation and experimental results revealed that the active absorber simultaneously contains such two working models. Further theoretical analysis based on LC theory also confirmed this dual ability.
Along with the increase of working frequency, sealed semiconductor devices become unsuitable due to the limited frequency response. In millimeter wave and terahertz, absorption behaviors can be tuned by modulating the relative position of cells on the top layer [230]. In 2013, Shrekenhamer et al. presented an active absorber in the terahertz regime [231]. By incorporation of liquid crystal into strategic locations within the unit cell, they were able to modify the absorption by 30% at 2.62 THz and also tune the resonant absorption over 4% in bandwidth. Other tunable metasurface absorbers adopted materials, such as vanadium oxide [232] and semiconductors [233].
Recently, we constructed selftuning metasurface absorbers with capability of selfadaptively tracing and absorbing the incident wave with relatively narrow bandwidth. As depicted in Figure 46, the operational principle includes two main steps. Firstly, the central frequency of the incident wave was detected by a detection circuit. Secondly, a voltage bias is added to the active metasurface to achieve high efficient absorption of the incident wave. The experimental results show that the echo attenuation within the frequency region of 2.8~3.2 GHz is all larger than 10 dB; for some frequency it even reaches 35 dB.
(a)
(b)
(c)
3.9. Flexible and Stretchable Metasurfaces
One advantage of metasurfaces over traditional 3D materials is its flexibility which is stemming from the ultrathin thickness. Flexible metasurfaces also provide a route to actively tune the performances. In traditional active metasurfaces, the electromagnetic response can be dynamically manipulated by tuning the constitutive parameters of the related optical materials. However, the mechanical properties, such as shape and geometrical size of these metasurfaces, are invariable during the tuning process. Even for MEMSbased metasurface, the macroscopic geometry is not changed in practical operation.
In recent years, the concept of flexible and stretchable metasurfaces has accepted particular attention, as inspired by the development of flexible electronic devices. For these metasurfaces, the key properties are determined by the substrate. For most metasurfaces except for some freestanding structures, the substrate not only provides a mechanical support for the metallic or dielectric structures but also offers an additional degree of freedom for the design of metasurface. Up to date, the most commonly used flexible substrates for metasurfaces are polydimethylsiloxane (PDMS), polyimide, metaflex, polyethylene naphthalene (PEN) [234–236], polyethylene terephthalate (PET), polymethylmethacrylate (PMMA), and polystyrene [237].
One particular application of flexible metasurface is the conformal devices such as conformal antennas, conformal lenses, and conformal absorbers [238–240]. For nonplanar metasurface absorbers, different areas experience different incident angles. As a result, the largeangle stability is required. Fortunately, it was found that metasurface absorber can have good absorption for large incidence angle by proper design [241]. As a result, conformal absorbers may maintain its absorption. However, earlier designs of metasurface absorbers suffer from the strong polarization dependence [36]. In 2011, we designed and fabricated a polarizationindependent wideangle absorber, following the design principle we established [37]. As shown in Figure 47(a), the absorption does not change dramatically when the metasurface is bent as a cylinder.
(a)
(b)
(c)
(d)
In 2013, a conformal absorber was designed by utilizing nonuniform metasurface [239]. The authors designed three different unit cells, which are optimized at 0°, 30°, and 45° incident angles, respectively. As illustrated in Figure 47(b), the proposed concept was demonstrated by EM simulation and experiments. The RCS reduction with nonuniform unit cells was compared with that of the conventional planar metamaterial absorber with uniform unit cells and superior results were demonstrated.
Regarding the wavefront engineering device (Figure 48), Aieta et al. designed a conformal metasurface lens to eliminate the offaxis aberration [240]. It was found that aberrationfree focusing is possible under axial illumination but offaxis aberrations appear when the excitation is not normal to the interface. An alternative design for an aplanatic metasurface on a curved substrate was proposed to focus light without coma and spherical aberrations.
(a)
(b)
The flexible property of PDMS has direct applications in other optical systems. For example, when a metallic grating was fabricated on PDMS, the period could be changed dynamically by stretching the substrate. As illustrated in Figure 49(a), the fabricated metasurface in our experiment looks orange and green, respectively, when stretched properly. Figure 49(b) shows the diffraction angle versus the overall length of the substrate when illuminated by a laser source at = 405 nm. The diffraction angle can be dynamically steered between 70 and 46 degrees.
(a)
(b)
(c)
(d)
In a similar way, Shen et al. proposed a color generation mechanism that produces colors by the Fano resonance effect on thin PMMA metasurface. The metasurface consists of a periodic array of cylinder of nanorods and the resonant frequency was tuned from 400 to 600 nm, as shown in Figures 49(c) and 49(d).
3.10. Graphene Metasurfaces
Current metasurfaces are often limited to dielectric and metallic structured thin film. In recent years, graphene has attracted special attention owing to its exotic electronic and optical properties. Not surprisingly, graphene has been widely utilized as one kind of metasurfaces or as a key component of functional metasurfaces.
Graphene is sp^{2}hybridized monolayer of carbon atoms densely packed into a honeycomb lattice, which was firstly experimentally isolated from graphite in the lab in 2003 [247]. In 2004, Geim et al. demonstrated that the carrier density in the graphene sheet can be controlled by a gate voltage, paving the way of graphenebased electronic devices. Subsequently, after the magnetic field response of graphene was validated, the studies on graphene’s electrooptics magnetooptics features increased dramatically. Moreover, the approach of growing largescale pattern graphene film makes the practical applications of such devices possible.
Indeed, the unique structural element of graphene gives rise to many excellent electromagnetic properties, permitting various important applications. For example, a single sheet of homogeneous graphene was found to be able to absorb only a little (~2.3%) fraction of incident white light at Dirac point, making it possible to be used as transparent conducing electrodes [248]. Besides, graphene exhibits ultrahigh electron mobility, anomalous quantum Hall effect, ambipolar electric field effect, massless relativistic carriers, over micrometerscale spin coherence length, highly confined plasmonic propagation, and large nonlinear Kerr effect. Thanks to these properties, graphene, up to now, has been widely used to fabricate fieldeffect transistors, electrochemical biosensor, solar cells, super capacitors, lithium ion batteries, catalyst carriers, electromagnetic absorber, and so on [247].
Surface conductivity is usually adopted to describe the electromagnetic characteristic of 2D graphene. If there is no external magnetostatic bias going through the graphene, the local conductivity is isotropic and can be approximately calculated using the Kubo formula on condition that , [249]:where is the thermal energy, is the chemical potential, is the scattering rate, and , , and are electron charge, Boltzmann constant, and reduced Plank constant (Dirac constant), respectively. Two inequivalent pairs of cones with apex at the Brillouin zone corners constitute graphene’s band structure; thus, two kinds of electron transition exist: interband and intraband transition. The first term in (44) is due to the contribution of interband transition and the second term arises for intraband transition. From the equation we know that the chemical potential governs the transition frequency of electrons. Further, if the frequency satisfies the condition of , the interband transitions can be neglected and (44) is simplified as Drude formula: where is a Drude weight. Owing to the mapping relation between gate voltage and chemical potential, the two equations are fairly enough to describe the conductivity of gate voltagebiased graphene.
Moreover, if the graphene sheet is present in a perpendicular magnetostatic bias, owing to the cyclotron motion of charge carriers, a nondiagonal component is introduced, namely, Hall conductivity. In low frequency (<10 THz usually), graphene is represented by full tensor conductivity:withwhere is the cyclotron frequency, is the amplitude of magnetostatic bias, and is the Fermi velocity of the Dirac fermions in graphene. Thus, the conductivities for the RCP light and the LCP light can be derived as , where sign “+” stands for RCP light and “−” stands for LCP light.
More efforts have been devoted to the investigation of graphene in recent years [250]. Some works were focused on enhancing the Faraday rotation angle in the microwave, THz, and infrared region. Meanwhile, graphene as a component of ultrathin electromagnetic absorber has drawn great attention. It is reported that the monolayer graphene is an effective saturable absorber for modelocking fiber lasers. As illustrated in Figures 50(a) and 50(b), it was demonstrated that complete optical absorption can take place in a single patterned sheet of doped graphene [251]. One interesting phenomenon shows that monolayer graphene shows highly directive comblike thermal radiation at nearinfrared frequency [249]. With the layer number increases, the radiation angle and beamwidth would be changed, as shown in Figures 50(c) and 50(d).
(a)
(b)
(c)
(d)
(e)
(f)
To increase the freedom degree of design, subwavelength structures have been proposed to change of electromagnetic response of graphene. Consequently, the effective conductivity of pattered graphene was investigated. In 2012, effective impedance for a patch patterned graphene sheet was proposed. It can be considered as a combination of two parts: impedance of unpatterned graphene produced by a scaling factor and capacitance caused by the gaps. The accuracy of the proposed impedance is debatable, because the capacitance is originated from the electronic motion which is relative to the magnetic responsible conductivity of graphene. In 2014, we [252] presented an approach for calculating the effective impedance for a holepatterned graphene sheet based on equivalent surface RLC impedance model. The impedance caused by periodic hole was obviously relative to the magnetostatic bias. Utilizing the model, authors well explained the magnetic circular dichroism of the holepatterned graphene absorber in a varying bias within 0~7 Tesla.
Graphene can also be utilized in plasmonic metasurface to increase the performance of traditional plasmonic devices. It is well known that traditional plasmonic materials do not work well in the infrared band, owing to the significant loss. In 2011, Engheta’s theoretical study showed that one can engineer the patch of surface wave on graphene by varying the chemical potential, making it a new platform of transformation optics [54]. Owing to its complex conductivity, graphene supports surface plasmon modes, accompanied with extremely short effective wavelength. Based on the exotic electronic transfer properties, graphenebased plasmonic waveguide devices and even Luneburg lens were constructed (see Figures 51(a) and 51(b)).
(a)
(b)
(c)
(d)
(e)
(f)
(g)
Very recently, we demonstrated that spiral surface plasmon could be utilized to transfer orbital angular momentum (OAM) [253]. Graphenebased microtube was demonstrated to be able to accomplish this goal perfectly. When a graphene tube is coated on the dielectric cylinder, the whole structure can also steadily support the steady propagation of light carrying OAM with different topological charges, as shown in Figures 51(c)–51(g). Furthermore, as the effective wavelength of the SPP induced in the graphene is much shorter than that in vacuum, the graphenebased microtube is suitable for the propagation of subwavelength of OAM.
3.11. Nonlinear Metasurfaces
Metasurfaces can also be utilized to enhance the nonlinear effect [254]. Different from other broadband applications, the nonlinear process requires strong increase of the local fields, which is only obtainable for resonant metasurface. One of the most important applications of nonlinear metasurface is the harmonic generation, such as second harmonic generation (SHG) and third harmonic generation (THG).
One way to enhance SHG is to engineer the metasurface so that they resonate at both the fundamental frequency (FF) and the harmonic frequency (HF) [255–257]. In 2014, Lee et al. proposed and experimentally realized metasurfaces with a recordhigh nonlinear response based on the coupling of electromagnetic modes in plasmonic metasurfaces with quantum engineered electronic intersubband transitions in semiconductor heterostructures [258]. It was stated that the obtained susceptibility is many orders of magnitude larger than any secondorder nonlinear response in optical metasurfaces measured up to then.
To obtain stronger nonlinear response, it is critical to understand the intrinsic relation between the linear and nonlinear processes. Fortunately, it was also demonstrated that the nonlinear response can also be predicted by using the information of linear response [259].
4. Fabrication Techniques of Metasurface
Up to date, there are many fabrication techniques available for the physical realization of metasurface. In microwave frequencies, since the line width of metasurface is larger than ten microns, traditional optical or mechanical approaches could be used, such as printed circuit board (PCB) technologies based on mechanical milling, chemical etching, and laser ablation. In terahertz to midinfrared region, the typical feature size of the metasurface structure is around 1~100 μm. Traditional optical lithography and laser direct writing are commonly used for the structure fabrication in this frequency region.
However, when the working frequency range goes into the nearinfrared and visible spectra, the fine features of metasurfaces generally could not be realized by traditional mechanical or optical methods. Instead, electron beam or iron beams are required to accomplish the fabrication goal. Electron beam lithography (EBL) and focused ion beam milling (FIB) are frequently adopted to make the related structures working in this wavelength range, as shown in Figure 52. Nevertheless, the fabrication processes of both of these techniques are based on the moving of the beam or stage, which makes them of slow speed and high cost, thus limiting their practical applications in metasurfaces fabrication in a certain extent.
(a)
(b)
(c)
(d)
Recently, some novel techniques are proposed in the fabrications of specific design of metasurfaces. Interference lithography is a very useful strategy in the fabrication of regular periodic patterns, such as grating and dot array. Periodic structures with feature size of hundred nanometers are easily realized by using UV laser. In order to further shrink the fabrication size of interference lithography, surface plasmon interference technique was developed. Subdiffraction limit patterns were obtained by employing the interference effect of SPPs [31, 260, 261]. The surface plasmon interference technique is considered to be the most promising plasmonic lithography technique for fabricating largearea simple periodic structures. However, it is difficult to generate complex and nonperiodic structures due to the limited propagating space for evanescent waves. More recently, reflective plasmonic slab or SPP cavity was proposed to improve the aspect profile and contrast of imaging pattern [32, 132], with experimentally reported 22 nm line width. As shown in Figures 53(a) and 53(b), it was shown that surface plasmon could be utilized to achieve line width of 50 nm. Very recently, hyperbolic metasurfaces composed of SiO_{2}/Al films are explored to squeeze out bulk plasmon polaritons (BPPs) to produce largearea and uniform deep subwavelength interference patterns. As examples, two and four bulk plasmon polaritons (BPPs) interference lithography with a half pitch of 45 nm (~) were demonstrated in experiments. Much deeper resolution up to 22.5 nm (~) and variety of BPPs interference patterns are feasible.
(a)
(b)
(c)
(d)
With the purpose of achieving high efficient, largearea fabrication of the nanoscale metasurfaces with arbitrary shape, we propose to utilize surface plasmon imaging lithography to transfer the patterns on mask. As shown in Figures 54(a) and 54(b), a Babinetinverted plasmonic gradient metalens was fabricated. The metalens consists of anisotropic nanohole arrays with parabolic phase distributions. The size of approximate elliptic nanohole is 75 nm × 140 nm. We characterized the focusing property of metalens at visible light with wavelengths of 633 nm and 532 nm. The results of simulations show good agreement with the measured results. The results of experiment demonstrate that surface plasmonic lithography not only offers the ability of transfer pattern with nanoscale resolution but also could be applied in the fabrication of planar metasurface devices.
(a)
(b)
(c)
(d)
Nanosphere lithography (NSL) is another promising technology in fabrication of metasurface with nanoscale feature size. The NSL was primely developed in the 1995, which was also referred to colloidal lithography. When NSL was just invented, only simple periodic nanoparticles could be produced. In recent years, the power of NSL was improved greatly. As shown in Figures 54(c) and 54(d), Nemiroski et al. showed that the parameter space of shadowderived shapes could enable a substantial expansion of the power of NSL [262]. They used customdesigned software to engineer compositions of shadows that guide multiangled deposition of one or multiple materials through a plasmaetched monolayer colloidal crystal (MCC). Meanwhile, Zhao et al. demonstrated multiple repetitions of holemask colloidal nanolithography to create singlelayer metasurfaces with complex, multishape plasmonic nanostructures that exhibit desired optical functionalities [263]. This fabrication method is particularly suited for the creation of largearea, singlelayer C3rotationally symmetric, 3D chiral metasurfaces. It should be noted that a similar glancing angle deposition technique with shadowing effect is able to produce 3D film structures with fine control on a scale less than 10 nm [264].
In addition to the fabrication technologies aforementioned, some other methods are introduced in some special situations. The capacity of nanoimprinting, multiple photons polymerization, microsphere lithography, and so forth [265] are also demonstrated in making the structures of metasurfaces. To overview, one should choose the proper fabrication method based on the corresponding requirement of specific structures which are designed.
5. Discussion
5.1. Unified Bandwidth Limit of Broadband Metasurfaces
As shown in the above discussion, the metasurfaces have many different applications depending on their constitutive materials and geometry. Nevertheless, all these devices can be classified as broadband, wideband, and narrowband according to their frequency responses. In this section, we would give a unified theory for the bandwidth limit of these metasurfaces, especially regarding the absorbers and polarizers. For the simplification of the discussion, we would like to consider only the narrowband and broadband cases.
The typical case of narrowband application is the Fano resonance. One of the most important aspects of Fano resonance is the sharp resonance peak accompanied with strong local field enhancement, which is the basis of many applications, such as biological sensing and nonlinear photonics [21]. Unfortunately, ultrahigh Qfactor is difficult to be realized for twodimensional metasurfaces due to the lack of large volume confinement of electromagnetic fields and strong coupling to free space [20]. Recently, we showed that there is fundamentally no limit for the bandwidth when no absorption loss is considered and the metasurface is infinite in the horizontal plane [19]. However, a further study is needed to show the limit for the finite metasurface.
For broadband and multiband applications, the dispersion of metasurfaces can be exploited. As illustrated in Figure 55, it is shown that the dispersion of metasurface can enable superresolution imaging beyond the diffraction limit, high performance color filter, broadband absorbers, and polarizers.
(a)
(b)
(c)
(d)
Most importantly, the results on coherent perfect absorber and coherent perfect rotation have shown that there is also no bandwidth limit for these applications [38, 76]. However, for noncoherent absorbers and polarizers, the classic limits are still applicable [100, 266], if we do not consider the case of nonFoster and active devices [267]. It is important to note that the bandwidth limits for the absorbers and polarizers have the same origin, and both of them can be deduced from the concept of artificial magnetic conductor (AMC). For AMC, there is a fundamental limit on the bandwidth [266]where is the reflective phase shift, which should be 0 for perfect magnetic conductor and for perfect electric conductor.
For perfect absorber comprised of a resistive sheet on a lossless AMC, the relation between the absorption and phase shift can be written as
The 10 dB and 20 dB copolarized reflection coincides with and . The thickness limit can be written as and . From Rozanov’s theory, this limit should be written aswhere is the reflectance in dB. For 10 dB and 20 dB reflectance, the thickness limits are and , agreeing well with the above results. As shown in Table 1, the thickness ratios of all the current available absorbers except the thin film CPA are smaller than the ideal case.

For perfect reflective halfwave plate, the copolarized reflection can be written as
The 10 dB and 20 dB copolarized reflection coincides with and . The thickness limit can be written as and . In this regard, our recent results have given an approach to overcome this limitation [81, 82].
In addition, we would like to comment that the thicknessbandwidth limitation of perfect absorbers can also be extended to the broadband antireflection coatings. Considering light passing from a medium with high refractive index to air, the energy can be nearly perfectly absorbed if a resistive layer with frequencyindependent resistance is added at the boundary [39]. As a result, the maximum thickness for the antireflection coating is that of the perfect absorbers.
5.2. Future Trends of Metasurfaces
In summary, the area of metasurfaces has become a cuttingedge and promising researching direction in recent years, although long histories can be found in almost any of their branches. In general, there is an obvious trend to extend the researching frequency regime from microwave to optical regime or the new emerging terahertz region. One may wonder that the extension of frequency range is straightforward and thus not so physically meaningful. However, we must note that the materials properties would change dramatically at different frequencies. The fabrication techniques and measurement equipment will also vary. At optical frequencies, even quantum phenomena would play an important role [21].
As illustrated in Figure 56, in the next few years, the researches in metasurfaces will be focused on the aspects including but not limited to [2] the following.
Functional Metasurfaces with Better Performances. For example, current flat lenses suffered from the strong chromatic dispersion. Although these devices are thin, lightweight, and capable of working at a wide range of frequencies, the focus lengths would vary as the change of operational wavelength. Recently, some works have been devoted to solving this problem, but the performance is still far from perfect [268]. Another drawback of these metasurfaces is that efficiencies are typically very low, especially for the devices based on geometric phase [134, 193]. As a result, much work should be paid to enhance the efficiency [204].
Fabrication Technologies for LargeArea Metasurfaces. One of the huge challenges for the successful application of metasurfaces is how to fabricate these metasurfaces fast and lowcostly. With these fabrication techniques [269], a bright future of metasurfaces and metasurfacebased devices can be envisioned.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
This work was supported by the 973 Program of China under Grant 2013CBA01700 and by the National Natural Science Funds under Grant 61138002. The authors also thank Z. Y. Zhao, C. T. Wang, C. Huang, C. G. Hu, Y. Q. Wang, and P. Gao for their helpful suggestions.
References
 N. Yu and F. Capasso, “Flat optics with designer metasurfaces,” Nature Materials, vol. 13, no. 2, pp. 139–150, 2014. View at: Publisher Site  Google Scholar
 X. Luo, “Principles of electromagnetic waves in metasurfaces,” Science China Physics, Mechanics & Astronomy, vol. 58, no. 9, Article ID 594201, pp. 1–18, 2015. View at: Publisher Site  Google Scholar
 J. B. Pendry, D. Schurig, and D. R. Smith, “Controlling electromagnetic fields,” Science, vol. 312, no. 5781, pp. 1780–1782, 2006. View at: Publisher Site  Google Scholar  Zentralblatt MATH
 J. B. Pendry, “Negative refraction makes a perfect lens,” Physical Review Letters, vol. 85, no. 18, pp. 3966–3969, 2000. View at: Publisher Site  Google Scholar
 N. Meinzer, W. L. Barnes, and I. R. Hooper, “Plasmonic metaatoms and metasurfaces,” Nature Photonics, vol. 8, no. 12, pp. 889–898, 2014. View at: Publisher Site  Google Scholar
 C. L. Holloway, E. F. Kuester, J. A. Gordon, J. O'Hara, J. Booth, and D. R. Smith, “An overview of the theory and applications of metasurfaces: the twodimensional equivalents of metamaterials,” IEEE Antennas and Propagation Magazine, vol. 54, no. 2, pp. 10–35, 2012. View at: Publisher Site  Google Scholar
 J. D. Jackson, Classical Electrodynamics, Wiley, Hoboken, NJ, USA, 3rd edition, 1999.
 D. Schurig, J. J. Mock, B. J. Justice et al., “Metamaterial electromagnetic cloak at microwave frequencies,” Science, vol. 314, no. 5801, pp. 977–980, 2006. View at: Publisher Site  Google Scholar
 T. Ergin, N. Stenger, P. Brenner, J. B. Pendry, and M. Wegener, “Threedimensional invisibility cloak at optical wavelengths,” Science, vol. 328, no. 5976, pp. 337–339, 2010. View at: Publisher Site  Google Scholar
 A. V. Kildishev, A. Boltasseva, and V. M. Shalaev, “Planar photonics with metasurfaces,” Science, vol. 339, no. 6125, Article ID 1232009, 2013. View at: Publisher Site  Google Scholar
 R. W. Wood, “On a remarkable case of uneven distribution of light in a diffraction grating spectrum,” Proceedings of the Physical Society of London, vol. 18, no. 1, article 325, 1901. View at: Publisher Site  Google Scholar
 T. Senior, “Approximate boundary conditions,” IEEE Transactions on Antennas and Propagation, vol. 29, no. 5, pp. 826–829, 1981. View at: Publisher Site  Google Scholar
 K. Yao and Y. Liu, “Plasmonic metamaterials,” Nanotechnology Reviews, vol. 3, no. 2, pp. 177–210, 2014. View at: Publisher Site  Google Scholar
 B. A. Munk, Frequency Selective Surfaces, Wiley, New York, NY, USA, 2000.
 D. Van Labeke, D. Gérard, B. Guizal, F. I. Baida, and L. Li, “An angleindependent Frequency Selective Surface in the optical range,” Optics Express, vol. 14, no. 25, pp. 11945–11951, 2006. View at: Publisher Site  Google Scholar
 A. H. Abdelrahman, A. Z. Elsherbeni, and F. Yang, “Transmission phase limit of multilayer frequencyselective surfaces for transmitarray designs,” IEEE Transactions on Antennas and Propagation, vol. 62, no. 2, pp. 690–697, 2014. View at: Publisher Site  Google Scholar
 T. Xu, Y.K. Wu, X. Luo, and L. J. Guo, “Plasmonic nanoresonators for highresolution colour filtering and spectral imaging,” Nature communications, vol. 1, p. 59, 2010. View at: Google Scholar
 A. E. Miroshnichenko, S. Flach, and Y. S. Kivshar, “Fano resonances in nanoscale structures,” Reviews of Modern Physics, vol. 82, no. 3, pp. 2257–2298, 2010. View at: Publisher Site  Google Scholar
 M. Pu, C. Hu, C. Huang et al., “Investigation of Fano resonance in planar metamaterial with perturbed periodicity,” Optics Express, vol. 21, no. 1, pp. 992–1001, 2013. View at: Publisher Site  Google Scholar
 V. A. Fedotov, M. Rose, S. L. Prosvirnin, N. Papasimakis, and N. I. Zheludev, “Sharp trappedmode resonances in planar metamaterials with a broken structural symmetry,” Physical Review Letters, vol. 99, no. 14, Article ID 147401, 2007. View at: Publisher Site  Google Scholar
 B. Luk'yanchuk, N. I. Zheludev, S. A. Maier et al., “The Fano resonance in plasmonic nanostructures and metamaterials,” Nature Materials, vol. 9, no. 9, pp. 707–715, 2010. View at: Publisher Site  Google Scholar
 M. Rahmani, B. Luk'yanchuk, and M. Hong, “Fano resonance in novel plasmonic nanostructures,” Laser and Photonics Reviews, vol. 7, no. 3, pp. 329–349, 2013. View at: Publisher Site  Google Scholar
 M. Rahmani, T. Tahmasebi, Y. Lin, B. Lukiyanchuk, T. Y. F. Liew, and M. H. Hong, “Influence of plasmon destructive interferences on optical properties of gold planar quadrumers,” Nanotechnology, vol. 22, no. 24, Article ID 245204, 2011. View at: Publisher Site  Google Scholar
 M. Rahmani, B. Lukiyanchuk, B. Ng, K. G. A. Tavakkoli, Y. F. Liew, and M. H. Hong, “Generation of pronounced Fano resonances and tuning of subwavelength spatial light distribution in plasmonic pentamers,” Optics Express, vol. 19, no. 6, pp. 4949–4956, 2011. View at: Publisher Site  Google Scholar
 J. Yang, M. Rahmani, J. H. Teng, and M. H. Hong, “Magneticelectric interference in metaldielectricmetal oligomers: generation of magnetoelectric fano resonance,” Optical Materials Express, vol. 2, no. 10, pp. 1407–1415, 2012. View at: Publisher Site  Google Scholar
 C. S. Lim, M. H. Hong, Z. C. Chen, N. R. Han, B. Luk'Yanchuk, and T. C. Chong, “Hybrid metamaterial design and fabrication for terahertz resonance response enhancement,” Optics Express, vol. 18, no. 12, pp. 12421–12429, 2010. View at: Publisher Site  Google Scholar
 Z. C. Chen, M. H. Hong, H. Dong et al., “Parallel laser microfabrication of terahertz metamaterials and its polarizationdependent transmission property,” Applied Physics A: Materials Science and Processing, vol. 101, no. 1, pp. 33–36, 2010. View at: Publisher Site  Google Scholar
 D. Wang, C. Qiu, and M. Hong, “Coupling effect of spiralshaped terahertz metamaterials for tunable electromagnetic response,” Applied Physics A: Materials Science and Processing, vol. 115, no. 1, pp. 25–29, 2014. View at: Publisher Site  Google Scholar
 D. Wang, Q. Huang, C. Qiu, and M. Hong, “Selective excitation of resonances in gammadion metamaterials for terahertz wave manipulation,” Science China Physics, Mechanics & Astronomy, vol. 58, no. 8, Article ID 084201, 2015. View at: Publisher Site  Google Scholar
 N. Fang, H. Lee, C. Sun, and X. Zhang, “Subdiffractionlimited optical imaging with a silver superlens,” Science, vol. 308, no. 5721, pp. 534–537, 2005. View at: Publisher Site  Google Scholar
 X. Luo and T. Ishihara, “Surface plasmon resonant interference nanolithography technique,” Applied Physics Letters, vol. 84, no. 23, pp. 4780–4782, 2004. View at: Publisher Site  Google Scholar
 C. Wang, P. Gao, Z. Zhao et al., “Deep subwavelength imaging lithography by a reflective plasmonic slab,” Optics Express, vol. 21, no. 18, pp. 20683–20691, 2013. View at: Publisher Site  Google Scholar
 Z. Liu, H. Lee, Y. Xiong, C. Sun, and X. Zhang, “Farfield optical hyperlens magnifying subdiffractionlimited objects,” Science, vol. 315, no. 5819, p. 1686, 2007. View at: Publisher Site  Google Scholar
 X. Luo and L. Yan, “Surface plasmon polaritons and its applications,” IEEE Photonics Journal, vol. 4, no. 2, pp. 590–595, 2012. View at: Publisher Site  Google Scholar
 H. Severin, “Nonreflecting absorbers for microwave radiation,” IRE Transactions on Antennas and Propagation, vol. 4, no. 3, pp. 385–392, 1956. View at: Publisher Site  Google Scholar
 N. I. Landy, S. Sajuyigbe, J. J. Mock, D. R. Smith, and W. J. Padilla, “Perfect metamaterial absorber,” Physical Review Letters, vol. 100, no. 20, Article ID 207402, 2008. View at: Publisher Site  Google Scholar
 M. Pu, C. Hu, M. Wang et al., “Design principles for infrared wideangle perfect absorber based on plasmonic structure,” Optics Express, vol. 19, no. 18, pp. 17413–17420, 2011. View at: Publisher Site  Google Scholar
 M. Pu, Q. Feng, M. Wang et al., “Ultrathin broadband nearly perfect absorber with symmetrical coherent illumination,” Optics Express, vol. 20, no. 3, pp. 2246–2254, 2012. View at: Publisher Site  Google Scholar
 A. Thoman, A. Kern, H. Helm, and M. Walther, “Nanostructured gold films as broadband terahertz antireflection coatings,” Physical Review B, vol. 77, no. 19, Article ID 195405, 2008. View at: Publisher Site  Google Scholar
 Y. Zhou, X. Xu, F. Hu et al., “Graphene as broadband terahertz antireflection coating,” Applied Physics Letters, vol. 104, no. 5, Article ID 051106, 2014. View at: Publisher Site  Google Scholar
 L. Ding, Q. Y. S. Wu, and J. H. Teng, “Polarization independent broadband terahertz antireflection by deepsubwavelength thin metallic mesh,” Laser & Photonics Reviews, vol. 8, no. 6, pp. 941–945, 2014. View at: Publisher Site  Google Scholar
 J. Yang, F. Luo, T. S. Kao et al., “Design and fabrication of broadband ultralow reflectivity black Si surfaces by laser micro/nanoprocessing,” Light: Science and Applications, vol. 3, article e185, 2014. View at: Publisher Site  Google Scholar
 M. V. Berry, “Quantal phase factors accompanying adiabatic changes,” Proceedings of the Royal Society of London. Series A. Mathematical, Physical and Engineering Sciences, vol. 392, no. 1802, pp. 45–57, 1984. View at: Publisher Site  Google Scholar  MathSciNet
 D. Lin, P. Fan, E. Hasman, and M. L. Brongersma, “Dielectric gradient metasurface optical elements,” Science, vol. 345, no. 6194, pp. 298–302, 2014. View at: Publisher Site  Google Scholar
 X. Ma, C. Huang, M. Pu, C. Hu, Q. Feng, and X. Luo, “Singlelayer circular polarizer using metamaterial and its application in antenna,” Microwave and Optical Technology Letters, vol. 54, no. 7, pp. 1770–1774, 2012. View at: Publisher Site  Google Scholar
 L. Verslegers, P. B. Catrysse, Z. Yu et al., “Planar lenses based on nanoscale slit arrays in a metallic film,” Nano Letters, vol. 9, no. 1, pp. 235–238, 2009. View at: Publisher Site  Google Scholar
 R. C. Hadarig, M. E. De Cos, and F. LasHeras, “Microstrip patch antenna bandwidth enhancement using AMC/EBG structures,” International Journal of Antennas and Propagation, vol. 2012, Article ID 843754, 6 pages, 2012. View at: Publisher Site  Google Scholar
 H.X. Xu, G.M. Wang, and C.X. Zhang, “Fractalshaped metamaterials and applications to enhancedperformance devices exhibiting high selectivity,” International Journal of Antennas and Propagation, vol. 2012, Article ID 515167, 14 pages, 2012. View at: Publisher Site  Google Scholar
 F. Yang, Z. L. Mei, and T. J. Cui, “Control of the radiation patterns using homogeneous and isotropic impedance metasurface,” International Journal of Antennas and Propagation, vol. 2015, Article ID 917829, 7 pages, 2015. View at: Publisher Site  Google Scholar
 M. Tauseef Asim and M. Ahmed, “Metamaterial inspired microstrip antenna investigations using metascreens,” International Journal of Antennas and Propagation, vol. 2015, Article ID 236136, 9 pages, 2015. View at: Publisher Site  Google Scholar
 X. Chen, T. M. Grzegorczyk, B.I. Wu, J. Pacheco Jr., and J. A. Kong, “Robust method to retrieve the constitutive effective parameters of metamaterials,” Physical Review E—Statistical, Nonlinear, and Soft Matter Physics, vol. 70, no. 1, Article ID 016608, 2004. View at: Publisher Site  Google Scholar
 A. Vora, J. Gwamuri, N. Pala, A. Kulkarni, J. M. Pearce, and D. O. Güney, “Exchanging ohmic losses in metamaterial absorbers with useful optical absorption for photovoltaics,” Scientific Reports, vol. 4, article 4901, 2014. View at: Publisher Site  Google Scholar
 A. Karlsson, “Approximate boundary conditions for thin structures,” IEEE Transactions on Antennas and Propagation, vol. 57, no. 1, pp. 144–148, 2009. View at: Publisher Site  Google Scholar
 A. Vakil and N. Engheta, “Transformation optics using graphene,” Science, vol. 332, no. 6035, pp. 1291–1294, 2011. View at: Publisher Site  Google Scholar
 X. Luo and T. Ishihara, “Subwavelength photolithography based on surfaceplasmon polariton resonance,” Optics Express, vol. 12, no. 14, pp. 3055–3065, 2004. View at: Publisher Site  Google Scholar
 J. B. Pendry, L. MartínMoreno, and F. J. GarciaVidal, “Mimicking surface plasmons with structured surfaces,” Science, vol. 305, no. 5685, pp. 847–848, 2004. View at: Publisher Site  Google Scholar
 X. Li, L. Yang, C. Hu, X. Luo, and M. Hong, “Tunable bandwidth of bandstop filter by metamaterial cell coupling in optical frequency,” Optics Express, vol. 19, no. 6, pp. 5283–5289, 2011. View at: Publisher Site  Google Scholar
 K. Sarabandi and N. Behdad, “A frequency selective surface with miniaturized elements,” IEEE Transactions on Antennas and Propagation, vol. 55, no. 5, pp. 1239–1245, 2007. View at: Publisher Site  Google Scholar
 Y. Yang, I. I. Kravchenko, D. P. Briggs, and J. Valentine, “Alldielectric metasurface analogue of electromagnetically induced transparency,” Nature Communications, vol. 5, article 5753, 2014. View at: Publisher Site  Google Scholar
 J. Gu, R. Singh, X. Liu et al., “Active control of electromagnetically induced transparency analogue in terahertz metamaterials,” Nature Communications, vol. 3, article 1151, 2012. View at: Publisher Site  Google Scholar
 J. Kim, R. Soref, and W. R. Buchwald, “Multipeak electromagnetically induced transparency (EIT)like transmission from bull'seyeshaped metamaterial,” Optics Express, vol. 18, no. 17, pp. 17997–18002, 2010. View at: Publisher Site  Google Scholar
 S. H. Mousavi, A. B. Khanikaev, and G. Shvets, “Optical properties of Fanoresonant metallic metasurfaces on a substrate,” Physical Review B, vol. 85, no. 15, Article ID 155429, 2012. View at: Publisher Site  Google Scholar
 M. Pu, M. Song, H. Yu et al., “Fano resonance induced by mode coupling in alldielectric nanorod array,” Applied Physics Express, vol. 7, no. 3, Article ID 032002, 2014. View at: Publisher Site  Google Scholar
 S. Fan and J. D. Joannopoulos, “Analysis of guided resonances in photonic crystal slabs,” Physical Review B, vol. 65, no. 23, Article ID 235112, 8 pages, 2002. View at: Publisher Site  Google Scholar
 C. Genet and T. W. Ebbesen, “Light in tiny holes,” Nature, vol. 445, no. 7123, pp. 39–46, 2007. View at: Publisher Site  Google Scholar
 C. M. Horwitz, “A new solar selective surface,” Optics Communications, vol. 11, no. 2, pp. 210–212, 1974. View at: Publisher Site  Google Scholar
 M. Song, H. Yu, C. Hu et al., “Conversion of broadband energy to narrowband emission through doublesided metamaterials,” Optics Express, vol. 21, no. 26, pp. 32207–32216, 2013. View at: Publisher Site  Google Scholar
 L. Young, L. A. Robinson, and C. Hacking, “Meanderline polarizer,” IEEE Transactions on Antennas and Propagation, vol. 21, no. 3, pp. 376–378, 1973. View at: Publisher Site  Google Scholar
 S. L. Wadsworth and G. D. Boreman, “Broadband infrared meanderline reflective quarterwave plate,” Optics Express, vol. 19, no. 11, pp. 10604–10612, 2011. View at: Publisher Site  Google Scholar
 D. Wang, Y. Gu, Y. Gong, C.W. Qiu, and M. Hong, “An ultrathin terahertz quarterwave plate using planar babinetinverted metasurface,” Optics Express, vol. 23, no. 9, pp. 11114–11122, 2015. View at: Publisher Site  Google Scholar
 M. Euler, V. Fusco, R. Cahill, and R. Dickie, “325 GHz single layer submillimeter wave FSS based split slot ring linear to circular polarization convertor,” IEEE Transactions on Antennas and Propagation, vol. 58, no. 7, pp. 2457–2459, 2010. View at: Publisher Site  Google Scholar
 W. Sun, Q. He, J. Hao, and L. Zhou, “A transparent metamaterial to manipulate electromagnetic wave polarizations,” Optics Letters, vol. 36, no. 6, pp. 927–929, 2011. View at: Publisher Site  Google Scholar
 D. L. Markovich, A. Andryieuski, M. Zalkovskij, R. Malureanu, and A. V. Lavrinenko, “Metamaterial polarization converter analysis: limits of performance,” Applied Physics B: Lasers and Optics, vol. 112, no. 2, pp. 143–152, 2013. View at: Publisher Site  Google Scholar
 A. Pors, M. G. Nielsen, G. D. Valle, M. Willatzen, O. Albrektsen, and S. I. Bozhevolnyi, “Plasmonic metamaterial wave retarders in reflection by orthogonally oriented detuned electrical dipoles,” Optics Letters, vol. 36, no. 9, pp. 1626–1628, 2011. View at: Publisher Site  Google Scholar
 J. Hao, Y. Yuan, L. Ran et al., “Manipulating electromagnetic wave polarizations by anisotropic metamaterials,” Physical Review Letters, vol. 99, no. 6, Article ID 063908, 2007. View at: Publisher Site  Google Scholar
 Y. Wang, M. Pu, C. Hu, Z. Zhao, C. Wang, and X. Luo, “Dynamic manipulation of polarization states using anisotropic metasurface,” Optics Communications, vol. 319, pp. 14–16, 2014. View at: Publisher Site  Google Scholar
 N. K. Grady, J. E. Heyes, D. R. Chowdhury et al., “Terahertz metamaterials for linear polarization conversion and anomalous refraction,” Science, vol. 340, no. 6138, pp. 1304–1307, 2013. View at: Publisher Site  Google Scholar
 S.C. Jiang, X. Xiong, Y.S. Hu et al., “Controlling the polarization state of light with a dispersionfree metastructure,” Physical Review X, vol. 4, no. 2, Article ID 021026, 2014. View at: Publisher Site  Google Scholar
 Z. H. Jiang, L. Lin, D. Ma et al., “Broadband and wide fieldofview plasmonic metasurfaceenabled waveplates,” Scientific Reports, vol. 4, article 7511, 2014. View at: Publisher Site  Google Scholar
 J. Hao, Q. Ren, Z. An et al., “Optical metamaterial for polarization control,” Physical Review A: Atomic, Molecular, and Optical Physics, vol. 80, no. 2, Article ID 023807, 2009. View at: Publisher Site  Google Scholar
 M. Pu, P. Chen, Y. Wang et al., “Anisotropic metamirror for achromatic electromagnetic polarization manipulation,” Applied Physics Letters, vol. 102, no. 13, Article ID 131906, 2013. View at: Publisher Site  Google Scholar
 Y. Guo, Y. Wang, M. Pu et al., “Dispersion management of anisotropic metamirror for superoctave bandwidth polarization conversion,” Scientific Reports, vol. 5, p. 8434, 2015. View at: Publisher Site  Google Scholar
 J. E. Roy and L. Shafai, “Reciprocal circularpolarizationselective surface,” IEEE Antennas and Propagation Magazine, vol. 38, no. 6, pp. 18–32, 1996. View at: Publisher Site  Google Scholar
 J. K. Gansel, M. Thiel, M. S. Rill et al., “Gold helix photonic metamaterial as broadband circular polarizer,” Science, vol. 325, no. 5947, pp. 1513–1515, 2009. View at: Publisher Site  Google Scholar
 J. K. Gansel, M. Wegener, S. Burger, and S. Linden, “Gold helix photonic metamaterials: a numerical parameter study,” Optics Express, vol. 18, no. 2, pp. 1059–1069, 2010. View at: Publisher Site  Google Scholar
 X. Ma, C. Huang, M. Pu, C. Hu, Q. Feng, and X. Luo, “Multiband circular polarizer using planar spiral metamaterial structure,” Optics Express, vol. 20, no. 14, pp. 16050–16058, 2012. View at: Publisher Site  Google Scholar
 M. Mutlu, A. E. Akosman, A. E. Serebryannikov, and E. Ozbay, “Asymmetric chiral metamaterial circular polarizer based on four Ushaped split ring resonators,” Optics Letters, vol. 36, no. 9, pp. 1653–1655, 2011. View at: Publisher Site  Google Scholar
 C. Menzel, C. Helgert, C. Rockstuhl et al., “Asymmetric transmission of linearly polarized light at optical metamaterials,” Physical Review Letters, vol. 104, no. 25, Article ID 253902, 2010. View at: Publisher Site  Google Scholar
 X. Ma, C. Huang, M. Pu et al., “Dualband asymmetry chiral metamaterial based on planar spiral structure,” Applied Physics Letters, vol. 101, no. 16, Article ID 161901, 2012. View at: Publisher Site  Google Scholar
 X. Ma, C. Huang, W. Pan, B. Zhao, J. Cui, and X. Luo, “A dual circularly polarized horn antenna in Kuband based on chiral metamaterial,” IEEE Transactions on Antennas and Propagation, vol. 62, no. 4, pp. 2307–2311, 2014. View at: Publisher Site  Google Scholar
 C. Huang, X. Ma, M. Pu, G. Yi, Y. Wang, and X. Luo, “Dualband 90° polarization rotator using twisted split ring resonators array,” Optics Communications, vol. 291, pp. 345–348, 2013. View at: Publisher Site  Google Scholar
 X. Ma, W. Pan, C. Huang et al., “An active metamaterial for polarization manipulating,” Advanced Optical Materials, vol. 2, no. 10, pp. 945–949, 2014. View at: Publisher Site  Google Scholar
 X. Ma, C. Huang, M. Pu, W. Pan, Y. Wang, and X. Luo, “Circular dichroism and optical rotation in twisted Yshaped chiral metamaterial,” Applied Physics Express, vol. 6, no. 2, Article ID 022001, 2013. View at: Publisher Site  Google Scholar
 H. K. Raut, V. A. Ganesh, A. S. Nair, and S. Ramakrishna, “Antireflective coatings: a critical, indepth review,” Energy and Environmental Science, vol. 4, no. 10, pp. 3779–3804, 2011. View at: Publisher Site  Google Scholar
 J.Q. Xi, M. F. Schubert, J. K. Kim et al., “Optical thinfilm materials with low refractive index for broadband elimination of Fresnel reflection,” Nature Photonics, vol. 1, no. 3, pp. 176–179, 2007. View at: Publisher Site  Google Scholar
 B. Zhang, J. Hendrickson, N. Nader, H.T. Chen, and J. Guo, “Metasurface optical antireflection coating,” Applied Physics Letters, vol. 105, no. 24, Article ID 241113, 2014. View at: Publisher Site  Google Scholar
 Y.F. Huang, S. Chattopadhyay, Y.J. Jen et al., “Improved broadband and quasiomnidirectional antireflection properties with biomimetic silicon nanostructures,” Nature Nanotechnology, vol. 2, no. 12, pp. 770–774, 2007. View at: Publisher Site  Google Scholar
 L. Ding, Q. Y. S. Wu, J. F. Song, K. Serita, M. Tonouchi, and J. H. Teng, “Perfect broadband Terahertz antireflection by deepsubwavelength, thin, lamellar metallic gratings,” Advanced Optical Materials, vol. 1, no. 12, pp. 910–914, 2013. View at: Publisher Site  Google Scholar
 E. F. Knott, J. F. Shaeffer, and M. T. Tuley, Radar Cross Section, SciTech Publishing, Raleigh, NC, USA, 2nd edition, 2004. View at: Publisher Site
 K. N. Rozanov, “Ultimate thickness to bandwidth ratio of radar absorbers,” IEEE Transactions on Antennas and Propagation, vol. 48, no. 8, pp. 1230–1234, 2000. View at: Publisher Site  Google Scholar
 M. Planck and M. Masius, The Theory of Heat Radiation, P. Blakiston's Son & Co, Philadelphia, Pa, USA, 1914.
 F. Ding, Y. Cui, X. Ge, Y. Jin, and S. He, “Ultrabroadband microwave metamaterial absorber,” Applied Physics Letters, vol. 100, no. 10, Article ID 103506, 2012. View at: Publisher Site  Google Scholar
 Y. Cui, K. H. Fung, J. Xu et al., “Ultrabroadband light absorption by a sawtooth anisotropic metamaterial slab,” Nano Letters, vol. 12, no. 3, pp. 1443–1447, 2012. View at: Publisher Site  Google Scholar
 X. Zang, C. Shi, L. Chen, B. Cai, Y. Zhu, and S. Zhuang, “Ultrabroadband terahertz absorption by exciting the orthogonal diffraction in dumbbellshaped gratings,” Scientific Reports, vol. 5, article 8901, 2015. View at: Publisher Site  Google Scholar
 J. A. Bossard, L. Lin, S. Yun, L. Liu, D. H. Werner, and T. S. Mayer, “Nearideal optical metamaterial absorbers with superoctave bandwidth,” ACS Nano, vol. 8, no. 2, pp. 1517–1524, 2014. View at: Publisher Site  Google Scholar
 L. K. Sun, H. F. Cheng, Y. J. Zhou, and J. Wang, “Broadband metamaterial absorber based on coupling resistive frequency selective surface,” Optics Express, vol. 20, no. 4, pp. 4675–4680, 2012. View at: Publisher Site  Google Scholar
 K. V. Nerkararyan, S. K. Nerkararyan, and S. I. Bozhevolnyi, “Plasmonic blackhole: broadband omnidirectional absorber of gap surface plasmons,” Optics Letters, vol. 36, no. 22, pp. 4311–4313, 2011. View at: Publisher Site  Google Scholar
 J. Grant, Y. Ma, S. Saha, A. Khalid, and D. R. S. Cumming, “Polarization insensitive, broadband terahertz metamaterial absorber,” Optics Letters, vol. 36, no. 17, pp. 3476–3478, 2011. View at: Publisher Site  Google Scholar
 E. E. Narimanov and A. V. Kildishev, “Optical black hole: broadband omnidirectional light absorber,” Applied Physics Letters, vol. 95, no. 4, Article ID 041106, 2009. View at: Publisher Site  Google Scholar
 A. Kazemzadeh, “Nonmagnetic ultrawideband absorber with optimal thickness,” IEEE Transactions on Antennas and Propagation, vol. 59, no. 1, pp. 135–140, 2011. View at: Publisher Site  Google Scholar
 A. K. Zadeh and A. Karlsson, “Capacitive circuit method for fast and efficient design of wideband radar absorbers,” IEEE Transactions on Antennas and Propagation, vol. 57, no. 8, pp. 2307–2314, 2009. View at: Publisher Site  Google Scholar
 J. Hao, J. Wang, X. Liu, W. J. Padilla, L. Zhou, and M. Qiu, “High performance optical absorber based on a plasmonic metamaterial,” Applied Physics Letters, vol. 96, no. 25, Article ID 251104, 2010. View at: Publisher Site  Google Scholar
 N. Liu, M. Mesch, T. Weiss, M. Hentschel, and H. Giessen, “Infrared perfect absorber and its application as plasmonic sensor,” Nano Letters, vol. 10, no. 7, pp. 2342–2348, 2010. View at: Publisher Site  Google Scholar
 Q. Feng, M. Pu, C. Hu, and X. Luo, “Engineering the dispersion of metamaterial surface for broadband infrared absorption,” Optics Letters, vol. 37, no. 11, pp. 2133–2135, 2012. View at: Publisher Site  Google Scholar
 A. Moreau, C. Ciracì, J. J. Mock et al., “Controlledreflectance surfaces with filmcoupled colloidal nanoantennas,” Nature, vol. 491, no. 7427, pp. 86–89, 2012. View at: Publisher Site  Google Scholar
 A. J. Hoffman, L. Alekseyev, S. S. Howard et al., “Negative refraction in semiconductor metamaterials,” Nature Materials, vol. 6, no. 12, pp. 946–950, 2007. View at: Publisher Site  Google Scholar
 M. Pu, M. Wang, C. Hu et al., “Engineering heavily doped silicon for broadband absorber in the terahertz regime,” Optics Express, vol. 20, no. 23, pp. 25513–25519, 2012. View at: Publisher Site  Google Scholar
 C. Shi, X. Zang, Y. Wang, L. Chen, B. Cai, and Y. Zhu, “A polarizationindependent broadband terahertz absorber,” Applied Physics Letters, vol. 105, no. 3, Article ID 031104, 2014. View at: Publisher Site  Google Scholar
 Y. D. Chong, L. Ge, H. Cao, and A. D. Stone, “Coherent perfect absorbers: timereversed lasers,” Physical Review Letters, vol. 105, no. 5, Article ID 053901, 2010. View at: Publisher Site  Google Scholar
 W. Wan, Y. Chong, L. Ge, H. Noh, A. D. Stone, and H. Cao, “Timereversed lasing and interferometric control of absorption,” Science, vol. 331, no. 6019, pp. 889–892, 2011. View at: Publisher Site  Google Scholar
 S. Li, J. Luo, S. Anwar et al., “An equivalent realization of coherent perfect absorption under single beam illumination,” Scientific Reports, vol. 4, p. 7369, 2014. View at: Publisher Site  Google Scholar
 G. Nie, Q. Shi, Z. Zhu, and J. Shi, “Selective coherent perfect absorption in metamaterials,” Applied Physics Letters, vol. 105, no. 20, Article ID 201909, 2014. View at: Publisher Site  Google Scholar
 M. Pu, Q. Feng, C. Hu, and X. Luo, “Perfect absorption of light by coherently induced plasmon hybridization in ultrathin metamaterial film,” Plasmonics, vol. 7, no. 4, pp. 733–738, 2012. View at: Publisher Site  Google Scholar
 J. Zhang, K. F. MacDonald, and N. I. Zheludev, “Controlling lightwithlight without nonlinearity,” Light: Science & Applications, vol. 1, no. 7, article e18, 2012. View at: Publisher Site  Google Scholar
 J. Shi, X. Fang, E. T. F. Rogers, E. Plum, K. F. MacDonald, and N. I. Zheludev, “Coherent control of Snell's law at metasurfaces,” Optics Express, vol. 22, no. 17, pp. 21051–21060, 2014. View at: Publisher Site  Google Scholar
 M. Crescimanno, N. J. Dawson, and J. H. Andrews, “Coherent perfect rotation,” Physical Review A—Atomic, Molecular, and Optical Physics, vol. 86, no. 3, Article ID 031807, 2012. View at: Publisher Site  Google Scholar
 H. Yao, G. Yu, P. Yan, X. Chen, and X. Luo, “Patterning sub 100 nm isolated patterns with 436 nm lithography,” in Proceedings of the International Microprocesses and Nanotechnology Conference, Tokyo, Japan, 2003. View at: Publisher Site  Google Scholar
 X. Luo and I. Teruya, “Sub 100 nm lithography based on plasmon polariton resonance,” in Proceedings of the International Microprocesses and Nanotechnology Conference, pp. 138–139, Tokyo, Japan, 2003. View at: Publisher Site  Google Scholar
 C. Park, J.H. Park, C. Rodriguez et al., “Fullfield subwavelength imaging using a scattering superlens,” Physical Review Letters, vol. 113, Article ID 113901, 2014. View at: Publisher Site  Google Scholar
 K. Aydin, I. Bulu, and E. Ozbay, “Subwavelength resolution with a negativeindex metamaterial superlens,” Applied Physics Letters, vol. 90, no. 25, Article ID 254102, 2007. View at: Publisher Site  Google Scholar
 M. D. Arnold and R. J. Blaikie, “Subwavelength optical imaging of evanescent fields using reflections from plasmonic slabs,” Optics Express, vol. 15, no. 18, pp. 11542–11552, 2007. View at: Publisher Site  Google Scholar
 P. Gao, N. Yao, C. Wang et al., “Enhancing aspect profile of halfpitch 32 nm and 22 nm lithography with plasmonic cavity lens,” Applied Physics Letters, vol. 106, no. 9, Article ID 093110, 2015. View at: Publisher Site  Google Scholar
 Z. Jacob, L. V. Alekseyev, and E. Narimanov, “Optical hyperlens: farfield imaging beyond the diffraction limit,” Optics Express, vol. 14, no. 18, pp. 8247–8256, 2006. View at: Publisher Site  Google Scholar
 N. Yu, P. Genevet, M. A. Kats et al., “Light propagation with phase discontinuities: generalized laws of reflection and refraction,” Science, vol. 334, no. 6054, pp. 333–337, 2011. View at: Publisher Site  Google Scholar
 Y. Chen, X. Li, Y. Sonnefraud et al., “Engineering the phase front of light with phasechange material based planar lenses,” Scientific Reports, vol. 5, p. 8660, 2015. View at: Publisher Site  Google Scholar
 S. Ishii, V. M. Shalaev, and A. V. Kildishev, “Holeymetal lenses: sieving single modes with proper phases,” Nano Letters, vol. 13, no. 1, pp. 159–163, 2013. View at: Publisher Site  Google Scholar
 T. Xu, C. Wang, C. Du, and X. Luo, “Plasmonic beam deflector,” Optics Express, vol. 16, no. 7, pp. 4753–4759, 2008. View at: Publisher Site  Google Scholar
 X. Ni, N. K. Emani, A. V. Kildishev, A. Boltasseva, and V. M. Shalaev, “Broadband light bending with plasmonic nanoantennas,” Science, vol. 335, no. 6067, p. 427, 2012. View at: Publisher Site  Google Scholar
 C. Min, P. Wang, X. Jiao, Y. Deng, and H. Ming, “Beam manipulating by metallic nanooptic lens containing nonlinear media,” Optics Express, vol. 15, no. 15, pp. 9541–9546, 2007. View at: Publisher Site  Google Scholar
 B. Hu, Q. J. Wang, S. W. Kok, and Y. Zhang, “Active focal length control of terahertz slitted plane lenses by magnetoplasmons,” Plasmonics, vol. 7, no. 2, pp. 191–199, 2012. View at: Publisher Site  Google Scholar
 J. Sun, X. Wang, T. Xu, Z. A. Kudyshev, A. N. Cartwright, and N. M. Litchinitser, “Spinning light on the nanoscale,” Nano Letters, vol. 14, no. 5, pp. 2726–2729, 2014. View at: Publisher Site  Google Scholar
 K. A. Willets and R. P. Van Duyne, “Localized surface plasmon resonance spectroscopy and sensing,” Annual Review of Physical Chemistry, vol. 58, pp. 267–297, 2007. View at: Publisher Site  Google Scholar
 P. L. Stiles, J. A. Dieringer, N. C. Shah, and R. P. Van Duyne, “Surfaceenhanced Raman spectroscopy,” Annual Review of Analytical Chemistry, vol. 1, no. 1, pp. 601–626, 2008. View at: Publisher Site  Google Scholar
 C. Wu, A. B. Khanikaev, R. Adato et al., “Fanoresonant asymmetric metamaterials for ultrasensitive spectroscopy and identification of molecular monolayers,” Nature Materials, vol. 11, no. 1, pp. 69–75, 2012. View at: Publisher Site  Google Scholar
 M. Rahmani, D. Y. Lei, V. Giannini et al., “Subgroup decomposition of plasmonic resonances in hybrid oligomers: modeling the resonance lineshape,” Nano Letters, vol. 12, no. 4, pp. 2101–2106, 2012. View at: Publisher Site  Google Scholar
 M. Rahmani, B. Lukiyanchuk, T. T. V. Nguyen et al., “Influence of symmetry breaking in pentamers on Fano resonance and nearfield energy localization,” Optical Materials Express, vol. 1, no. 8, pp. 1409–1415, 2011. View at: Publisher Site  Google Scholar
 M. Rahmani, A. E. Miroshnichenko, D. Y. Lei et al., “Beyond the hybridization effects in plasmonic nanoclusters: diffractioninduced enhanced absorption and scattering,” Small, vol. 10, no. 3, pp. 576–583, 2014. View at: Publisher Site  Google Scholar
 A. A. Yanik, A. E. Cetin, M. Huang et al., “Seeing protein monolayers with naked eye through plasmonic Fano resonances,” Proceedings of the National Academy of Sciences of the United States of America, vol. 108, no. 29, pp. 11784–11789, 2011. View at: Publisher Site  Google Scholar
 A. Tittl, P. Mai, R. Taubert, D. Dregely, N. Liu, and H. Giessen, “Palladiumbased plasmonic perfect absorber in the visible wavelength range and its application to hydrogen sensing,” Nano Letters, vol. 11, no. 10, pp. 4366–4369, 2011. View at: Publisher Site  Google Scholar
 G. Sarau, B. Lahiri, P. Banzer et al., “Enhanced Raman scattering of graphene using arrays of split ring resonators,” Advanced Optical Materials, vol. 1, no. 2, pp. 151–157, 2013. View at: Publisher Site  Google Scholar
 H. Aouani, H. Šípová, M. Rahmani et al., “Ultrasensitive broadband probing of molecular vibrational modes with multifrequency optical antennas,” ACS Nano, vol. 7, no. 1, pp. 669–675, 2013. View at: Publisher Site  Google Scholar
 M. König, M. Rahmani, L. Zhang et al., “Unveiling the correlation between nanometerthick molecular monolayer sensitivity and nearfield enhancement and localization in coupled plasmonic oligomers,” ACS Nano, vol. 8, no. 9, pp. 9188–9198, 2014. View at: Publisher Site  Google Scholar
 M. Rahmani, B. Lukiyanchuk, T. Tahmasebi, Y. Lin, T. Y. F. Liew, and M. H. Hong, “Polarizationcontrolled spatial localization of nearfield energy in planar symmetric coupled oligomers,” Applied Physics A: Materials Science and Processing, vol. 107, no. 1, pp. 23–30, 2012. View at: Publisher Site  Google Scholar
 H. Aouani, M. Rahmani, H. Šípová et al., “Plasmonic nanoantennas for multispectral surfaceenhanced spectroscopies,” The Journal of Physical Chemistry C, vol. 117, no. 36, pp. 18620–18626, 2013. View at: Publisher Site 