high‐efficiency metasurfaces: principles, realizations, and applications · 2019-03-18 ·...

Review

1800415 (1 of 23) © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advopticalmat.de

High-Efficiency Metasurfaces: Principles, Realizations, and Applications

Qiong He, Shulin Sun, Shiyi Xiao, and Lei Zhou*

DOI: 10.1002/adom.201800415

application demands in different aspects, such as information communications, national defenses, security and sensing, and energy utilizations. A key scientific issue, commonly requested in all those applications, is how to efficiently control EM waves at will based on devices/systems with physical sizes as miniaturized as pos-sible. Conventional devices (such as lens, mirror, polarizer, etc.) to control EM waves are made by naturally existing dielectric materials. However, since these materials possess very limited variation range of permittivity, conventional devices typically exhibit bulky sizes and curved shapes to ensure enough propagating phases accu-mulated to achieve certain functionality. In addition, conventional devices do not nec-essarily exhibit high working efficiencies, due to the impedance-mismatch issues caused by lacking magnetic responses in natural materials.

Metamaterials (MTMs), 3D artificial materials composed by array of sub-wavelength microstructures (e.g., “meta-

atoms”) with tailored EM responses, provide the possibility to solve the above grand challenges. Through careful structural tuning, MTMs can in principle be designed to exhibit arbi-trary values of permittivity ε and permeability μ, and thus they exhibit significantly strengthened capabilities to control EM waves, generating many fascinating wave-manipulation effects not achievable with naturally existing materials.[1–6] However, despite of great successes already achieved, applications of the new concepts proposed based on 3D MTMs are hampered by the device sizes, losses in metallic structures, and fabrication challenges of complex 3D MTM structures.

Metasurfaces are probably the best candidate to solve the issues of 3D MTMs. Simply speaking, metasurfaces can be viewed as the 2D version of MTMs, which are constructed by planar meta-atoms with purposely selected EM responses arranged in some specific orders. However, the working princi-ples of metasurfaces are quite different from that of 3D MTMs in controlling EM waves. Whereas typically 3D MTMs still reply on manipulating the propagating phases inside the bulk media to realize certain wave-control effects, metasurfaces largely utilize the abrupt phase changes on the structure surfaces for transmitted or reflected waves. In general, tailoring the spatial inhomogeneity of metasurfaces to make them exhibit certain predetermined phase distributions for transmitted or reflected wave, one can use them to efficiently reshape the wave-fronts of

Metasurfaces are planar metamaterials exhibiting certain inhomogeneous phase distributions for transmitted or reflected waves, which can efficiently reshape the wave-fronts of incident beams in desired manners based on the Huygens principle. Due to their exotic abilities to freely manipulate electro-magnetic (EM) waves on a flat and ultrathin platform, metasurfaces have attracted intensive attention recently, resulting in numerous new concepts and effects that could possibly find applications in many different aspects. In this article, the key achievements in this field are briefly summarized, with focus put on the efficiency issue of metasurfaces. After introducing the basic concept of metasurfaces and their early realizations, successively the mechanisms and realizations of high-efficiency metasurfaces are described: for manipulating linearly polarized EM waves in reflection and transmission geometries, and for manipulating circularly polarized EM waves based on the geometric-phase concept, using plasmonic materials as well as dielectric ones. Then, several important applications of high-efficiency metasurfaces are summarized, including polarization control, metaholograms, metalenses, and surface wave couplers. Finally, this review is concluded with personal perspectives on the future directions of this rapidly growing research field.

Prof. Q. He, Prof. L. ZhouState Key Laboratory of Surface Physics and Key Laboratory of Micro and Nano Photonic Structures (Ministry of Education)Fudan UniversityShanghai 200433, ChinaE-mail: [email protected]. Q. He, Prof. L. ZhouCollaborative Innovation Center of Advanced MicrostructuresNanjing 210093, ChinaProf. S. SunEngineering Research Center of Ultra-Precision Optical ManufacturingGreen Photonics and Department of Optical Science and EngineeringFudan UniversityShanghai 200433, ChinaProf. S. XiaoKey Laboratory of Specialty Fiber Optics and Optical Access NetworksJoint International Research Laboratory of Specialty Fiber OpticsShanghai UniversityShanghai 200444, China

Metasurfaces

1. Introduction

The arbitrary control on electromagnetic (EM) wave within the whole frequency domain is highly desired, not only due to scientific curiosities but also because of tremendous

Adv. Optical Mater. 2018, 6, 1800415

http://crossmark.crossref.org/dialog/?doi=10.1002%2Fadom.201800415&domain=pdf&date_stamp=2018-07-29

www.advancedsciencenews.com

© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1800415 (2 of 23)


incident beams based on the Huygens principle. Such devices are typically thin, flat, easy to fabricate, and can avoid a large portion of losses, all being highly desired in realistic applica-tions. Significant achievements have been made in this field after the pioneering work of Capasso and co-workers,[7] on both fundamental science and diversified applications of metasur-faces. As this field is developing very rapidly, it is highly desired at this stage to have a concise review, which can guide readers catch the state-of-the-art development and assist newcomers to jump into the field without many difficulties.

In this article, we present such a review to briefly summarize the milestone achievements in this field, focusing particularly on the efficiency issue of metasurfaces. In our opinion, effi-ciency is the most crucial issue for applications of metasur-faces, and discussing it in a coherent way is very helpful for potential readers, which also makes our review article distinct from other existing ones.[8–17] To achieve this end, we organize our article in the following way. After introducing the basic concept of metasurfaces and the low-efficiency issues in early realizations (Section 2), we describe successively the working principles and practical realizations of high-efficiency metas-urfaces for controlling linearly polarized (LP) waves in reflec-tion (Section 3) and transmission (Section 4) geometries, and for controlling circularly polarized (CP) waves based on the geometric-phase concept (Section 5). We then review the effi-cient EM-wave manipulations achieved with dielectric metasur-faces (Section 6). Understanding these principles help people realize a wide-range of applications based on high-efficiency metasurfaces, such as polarization control, lensing, hologram, and surface wave couplers, which are summarized in Section 7. We finally conclude this review in Section 8, together with our perspectives on the future directions of this fast-developing research field.

2. Generalized Snell’s Law and Metasurfaces Realized with V-Shaped Antennas

Reflection/refraction at an interface is probably the simplest approach to control the propagation direction of EM wave. At a planar interface between two homogeneous optical media, reflections and refractions of EM waves are governed by Snell’s law, which states that the reflection angle θr should be identical to the incident angle θi, while the refraction angle θt is related to θi via n2 sinθt = n1 sinθi with n1 and n2 being the refrac-tion indices of two media (see left panel in Figure 1a). The core physics underlying this law is the conservation of parallel momentums of reflected/refracted light beams with respect the incident one

k k k

||t

||r

||i= = (1)

which can be derived from the boundary conditions of EM fields, thanks to the translation-invariance at the planar inter-face. Such a law can be alternatively understood based on the Huygens principle. As shown in Figure 1a, reflected/refracted beams can be viewed as the reconstructions of waves radiated from different local positions at the interface, which serve as the second sources when shined by the incident beam. At a

translation-invariance interface, all second sources are identical except the initial phases given by the excitation field ik xxexp( )i ⋅ . It is this difference that determines the travelling directions of reflected/refracted beams, based on reconstructing the equal-phase planes of two beams (see left panel in Figure 1a).

It is thus clear that the original Snell’s law does not hold for an interface without translation-invariance symmetry. However, if the

Qiong He received his PhD degree in Physics from Paris Institute of Optics in Paris-Sud University (Orsay, France) in 2008. From 2009 to 2013, he was postdoctoral fellow in Physics Department of Fudan University. He is cur-rently an associate professor at Physics Department of Fudan University (Shanghai, China). His research inter-

ests focus on metamaterials and plasmonics. He has coauthored more than 40 publications in scientific journals.

Shulin Sun received his PhD degree in Physics Department of Fudan University in 2009. From 2010 to 2013, he was postdoc-toral fellow of the Physics Division of NCTS in National Taiwan University. In 2013, he joined Department of Optical Sciences & Engineering of Fudan University as an associate professor. He has

coauthored more than 40 publications in international journals. His research interest includes metamaterials/metasurfaces, plasmonics, and photonic crystals.

Lei Zhou received his PhD in Physics from Fudan University, China, in 1997. From 1997 to 2000, he was a postdoctoral fellow of Institute for Material Research in Tohoku University (Sendai, Japan). In 2000–2004, he was a visiting scholar in Physics Department of the Hong Kong University of Science and Technology. He

joined Physics Department of Fudan University in 2004 as a professor, and became a “Xi-De” chair professor since 2013. He has been working in the fields of magnetism, metamaterials, photonic crystals, and plasmonics, and has published over 140 papers in scientific journals.





second source at a position x of an inhomogeneous interface can carry an additional phase ϕ(x) linearly depending on x so that its total initial phase becomes k xx

i ϕ( )+ ∇ ⋅ , Yu et al. derived the following generalized Snell’s laws for reflection k xsin sin d /dr i 0

1θ θ− = Φ− and refraction n n k xsin sin d /dt t i i 0

1θ θ− = Φ− , simply by repeating the phase matching arguments, as shown in Figure 1a (right panel). In a more general form in which the phase gradient direction can be different from that of the in-plane k vector of incident wave, the generalized Snell’s law can be rewritten as[7]

k k

k k

||r

||i

||t

||i

ϕϕ

= + ∇

= + ∇

(2)

indicating that the phase gradient essentially provides an effective k vector to both reflected and refracted waves. Equation (2) is of profound physics, which not only generalizes the conventional Snell’s law but also implies lots of unusual wave-manipulation effects. For example, one can control the


Figure 1. Generalized Snell’s law and metasurfaces realized with V-shape antennas. a) Understanding the (generalized) Snell’s Law at interfaces with and without translation-invariance symmetry based on Huygens’ Principle. b) Analytically calculated amplitude and phase shift of the cross-polarized scattered light for V-antennas consisting of gold rods with a circular cross-section and with various length h and angle between the rods Δ at λ = 8 μm. The four circles indicate the values of h and Δ used in experiments. Inset: Symmetric and antisymmetric modes excited, respectively, by components of the incident field along s and â axes of V-antenna. The angle between the incident polarization and the antenna symmetry axis is 45°. c) Left panel: Experimental far-field scans showing the ordinary and extraordinary refraction generated by different metasurfaces at different wavelengths, given normally incident light. Right panel: Angle of refraction versus angle of incidence for the ordinary (black curve and triangles) and anomalous refraction (red curve and dots) for one sample. Inset: Scanning electron microscope (SEM) image of a representative antenna array fabricated on a silicon wafer. d) Left panel: Transverse polarization splitting induced by a metasurface with a strong gradient of phase retardation along the x-direction. The rapid phase retardation refracts light in a skewing direction and results in the PSHE. Inset: SEM image of a metasurface with a rapid phase gradient in the horizontal (x) direction. Right panel: Observation of a giant SHE: the helicity of the anomalously refracted beam with normally incident LP light along the x-direction along the phase gradient. Red and blue represent the right and left circular polarizations, respectively. e) Left-top inset: Schematic view of a Babinet-inverted, complementary plasmonic V-antenna and experimentally realized hologram image of letter “P”. Left-bottom inset: Holographic images of word “PURDUE” produced by the experimental metasurface holograms. Right panel: SEM image of a fabricated metasurface for generating a hologram of the Letter “P”. f) Left-top inset: Photograph of part of the fabricated cylindrical terahertz lens. Right panel: FDTD simulated intensity distribution of the cross-polarized light for the designed cylindrical lens along the z-direction. Left-bottom inset: Intensity distributions along the white dashed lines shown in right panel. (b),(c) Reproduced with permission.[7] Copyright 2011, the American Association for the Advancement of Science (AAAS). (d) Reproduced with permission.[21] Copyright 2013, AAAS. (e) Reproduced with permission.[22] Copyright 2013, Nature Publication Group (NPG). (f) Reproduced with permission.[23] Copyright 2013 Wiley-VCH.




propagating directions of reflected and transmitted light beams nearly arbitrarily, inside or even outside the incident-plane, simply by engineering the phase gradient of the interface.

Yu et al. designed the first metasurface (working at λ = 8 μm) to experimentally demonstrate the generalized Snell’s law. According to the arguments presented above, a crucial thing is to find a set of planar meta-atoms that can yield different phases (covering the full 2π range) for scattered (reflected or transmitted) waves, but with nearly identical scattering ampli-tudes.[7,18] In MTM community, one typically utilizes the dispersions of resonators to achieve desired responses at a given frequency. However, a single electric resonator (say, a nanorod[19]) exhibits a Lorentz-type response with maximum phase varia-tion π (the lossless case) and with scattering amplitude tightly linked with the phase variation. Understanding these inherent difficulties, Yu et al. proposed to use V-shaped nanoantennas, containing two resonance modes with orthogonal EM responses at different frequencies, as their meta-atoms. Specifically, the V-shaped antenna consists of two nanorods with equal length (h) jointed together forming a certain angle Δ. Such a plasmonic antenna can support a symmetric resonant mode for the electric (E) field polarized along its diagonal axis and an antisymmetric mode for the E field polarized along the orthogonal direction (inset of Figure 1b). When the incident E field is neither perpen-dicular nor parallel to the symmetry axis of the V-shape antenna (say, at an angle 45° with respect to the symmetry axis), both resonant modes can be excited but with substantially different amplitudes and phases, due to their distinct resonance condi-tions dictated by the geometrical parameters. As the result, the scattered light has a component polarized orthogonal to that of incident light, roughly proportional to the response difference between two excited modes. The unique feature of combining two out-of-phase resonators makes the cross-polarized scattered light exhibit a phase dictated by two geometrical parameters h and Δ (see left panel in Figure 1b) with variation range cov-ering 2π, breaking the strict π limit imposed on a single electric resonator. One should notice that the copolarized scattered light does not exhibit such a unique feature, since it is contributed by two in-phase resonators and thus it behaves in the similar way with a single electric resonator. The scattering amplitude can also be tuned by h and Δ (see right panel Figure 1b). Via careful structural tuning, Yu et al. successfully found eight different antennas, which can provide nearly equal scattering amplitude and π/4 phase increments for cross-polarized light. With all meta-atoms at hand, the authors then fabricated their metasur-face sample and experimentally characterized its scattering prop-erties under illuminations at different angles. Results shown in Figure 1c,d indicate that, apart from normal modes, there indeed appear anomalously reflected and refracted signals, with reflection/refraction angles satisfying the generalized Snell’s law. Soon after the pioneering work of Yu et al., Ni et al. extended the concept to near-infrared (NIR),[20] and demonstrated that the metasurface constructed by V-shaped antennas can exhibit a very broadband performance. Such broadband property is due to the low quality factors of the nanoantenna adopted in constructing the metasurfaces. In addition, since meta-atoms are not required to work at their resonance frequencies, the absorption due to resonance is largely suppressed, which is again in sharp contrast with a usual resonance-type MTM.

Based on metasurfaces constructed by V-shaped meta-atoms, many exciting wave-control phenomena were demon-strated in different frequency domains.[21–26] For example, Yin et al. experimentally demonstrated a strong photonic spin Hall effect (PSHE) at a metasurface based on V-shaped antennas (see Figure 1d).[21] The rapidly varying phase discontinui-ties along a metasurface, breaking the axial symmetry of the system, significantly enhances the spin–orbital coupling of light enabling the direct observation of large transverse motion of circularly polarized light even at normal incidence. Ni et al. experimentally demonstrated ultrathin metasurface holograms based on complementary V-shaped nanoantennas, which can provide both amplitude and phase modulation in the visible wavelength range to generate high-resolution low-noise images (Figure 1e).[22] Hu et al. fabricated various ultrathin terahertz devices based on metasurfaces with complementary V-shaped antennas, and experimentally demonstrated their focusing and imaging performances (Figure 1f).[23] It is worth mentioning that metasurfaces are conceptually related to transmit-arrays and reflect-arrays demonstrated in the microwave range,[27,28] but are quite different from frequency selective surfaces,[29] which usually do not exhibit spatial variations of EM responses.

Strictly speaking, the derivation of the generalized Snell’s law requests that the interface exhibits a continuously varying phase profile. Unfortunately, such a system is extremely difficult to realize in practice. Alternatively, people usually designed/fab-ricated metasurfaces consisting of superperiodicity to truncate the otherwise diverging phase responses. In the latter case, one can find an alternative yet inspiring interpretation of the gener-alized Snell’s law based on the diffraction theory.[30,31] Indeed, employing the diffraction theory, Smith et al.[32,33] and Lalanne et al.[34] proposed and realized planar blazed gratings con-structed by uniform-thickness layers with effective refraction-index varying linearly in space, and demonstrated that they can efficiently steer the incident wave to the anomalous directions dictated by one of the diffraction modes. These analyses rec-oncile the pictures based on the generalize Snell’s law and the diffraction theory, providing an alternative yet inspiring way to understand the inherent physics.

Despite of the great successes achieved, metasurfaces based on V-shaped meta-atoms suffer from several issues that need to be mentioned. First, these metasurfaces naturally support multiple modes (i.e., normal reflection/transmission modes and anomalous transmission/reflection modes), and thus the working efficiency of the desired anomalous reflection/refraction must be low,[35,36] which represents a significant obstacle for practical applications. Second, in this configuration, the anoma-lous modes exhibit a polarization perpendicular to that of the incident light, which is also unfavorable for real applications. In what follows, we discuss the efforts from different groups to build high-efficiency metasurfaces in different geometries and for EM waves with different polarizations.

3. High-Efficiency Metasurfaces in Reflection Geometry

The key issue of the V-shape meta-atom is that it can radiate to both sides of the metasurface once being excited, leading to





multimode generations degrading the working efficiency of a metasurface. To maximize the working efficiency, ideally one requires meta-atoms which are either purely reflective or purely transparent, yet exhibiting fully controllable phases.

In 2011, Sun et al. proposed the first reflection-type gra-dient metasurface with nearly 100% efficiency, based on meta-atoms in metal–insulator–metal (MIM) configuration.[37] Such MIM meta-atom typically consists of a metallic subwavelength resonator and a continuous metal film separated by a dielec-tric spacer (left inset to Figure 2a). The presence of a metallic ground plane prohibits any transmission through the system.

Via appropriately tuning its geometrical parameters, an MIM meta-atom can reflect EM waves with very high amplitudes (equal to 100% in ideal lossless case) and phases varying in the range of [−π, π] depending on frequency (see Figure 2a for a typical example). Such unique feature makes MIM meta-atom ideal candidate to build high-efficiency metasurfaces in reflection geometry. Based on MIM meta-atoms with H-shape antennas as top resonators, Sun et al. designed and fabricated a series of microwave metasurfaces exhibiting different values of phase gradients ξ = ∇ϕ, and experimentally demonstrated that they can reflect EM waves following the generalized Snell’s law


Figure 2. High-efficiency metasurfaces in reflection geometry. a) Computed spectra of reflection amplitude and phase of a typical MIM meta-atom consisting of a 130 nm thick gold substrate coupled with a gold nanobrick (30 nm thick) through a 50 nm thick MgF2 spacer (left inset). Right inset: Smith curves of the reflection coefficient r obtained with the CMT model with different ratios of Γi/Γr. b) Absorption as a function of Γi and Γr calculated with the CMT for the MIM systems. Perfect absorption (A = 1) happens when Γi = Γr, while the system is an underdamped/overdamped resonator when Γi < Γr (Γi > Γr). c) Top panel: Photograph of a fabricated microwave gradient MIM metasurface using H-shaped antennas with varying arm-lengths to manipulate the phase response. Inset: Sketch of an MIM meta-atom consisting of a metallic “H” and a metal plate separated by a dielectric spacer. Medium panel: reflection phase distributions based on analytical and simulation results for the metasurface with ∇ϕ = 0.8k0. Bottom panel: Measured and simulated far-field intensity profiles for a microwave gradient metasurface with ∇ϕ = 0.8k0. d) Top panel: Schematic of a NIR beam deflector based on MIM metasurface. Left inset: SEM image of part of one fabricated metasurface. Bottom panel: Measured normalized scattered electric field intensity for the gradient metasurface under illuminations a y-polarized light at λ = 850 nm at different incident angles. e) Left panel: SEM image of part of one fabricated flat lens based on MIM metasurface. Right panel: intensity maps near the focus for NA = 0.8. f) Left panel: Schematics of high-efficiency polarization-dependent metahologram in the NIR regime. Inset: SEM image of part of one fabricated sample. Right panel: Absolute efficiency of the metahologram versus incident angle θ at the wavelengths of 780, 632.8, and 405 nm. (b) Reproduced with permission.[64] Copyright 2015, APS. (c) Reproduced with permission.[37] Copyright 2012 Macmillan Publishers Limited. (d) Reproduced with permission.[38] Copyright 2012, American Chemical Society (ACS). (e) Reproduced with permission.[40] Copyright 2013, ACS. (f) Reproduced with permission.[45] Copyright 2014, ACS.




with nearly 100% efficiencies (see Figure 2c for one example).[37] Sun et al. soon extended the idea to high frequencies and exper-imentally demonstrated highly efficient (≈80%) anomalous reflection of NIR light within a broadband (750–900 nm), based on a MIM metasurface (Figure 2d).[38]

Various high-efficiency metadevices were subsequently real-ized based on MIM structures.[39–48] For example, Li et al. and Pors et al. independently realized metasurfaces exhibiting para-bolic phase profiles in microwave[39] and NIR regimes,[40] and experimentally demonstrated that they can focus EM waves with very high efficiencies (see Figure 2e). In addition to high working efficiency, such MIM metasurfaces possess another merit of preserving the polarization of reflected light. This prop-erty not only makes MIM metasurfaces easy to use in practical applications, but offers the possibility to construct polarization-dependent multifunctionalities.[45,47,49–53] Chen et al. experi-mentally demonstrated that an optical metasurface designed with anisotropic MIM meta-atoms can reconstruct polarization-controlled dual hologram images within a broad operating band (≈800 nm) and under a wide range of incident angles (15°–45°) (see Figure 2f).[45] Boroviks et al. demonstrated a multifunc-tional MIM metasurface that can efficiently split orthogonal LP light and focuses the reflected beams with different polari-zations into different focal spots. The working efficiency of realized metadevice can reach ≈65% with polarization extinction ratio of ≈30 dB. Such device also features broadband response with working wavelength covering the range of 750–950 nm.[51]

The working principle of MIM metasurfaces is usually inter-preted as follows. Strong near-field couplings between two metallic layers induce electrical currents flowing in opposite directions on them, generating a magnetic resonance inside the structure. At optical frequencies, such a mode is often termed as the gap-surface plasmon (GSP) mode.[54–56] In sharp contrast to a single electric resonator, the presence of a magnetic reso-nance can efficiently tune the impedance of the entire MIM structure, making its reflection phase easily cover the full 2π range with slowly varying reflection-amplitude as frequency changes (see Figure 2a). However, this simple picture cannot explain why MIM structures, upon careful design, can also behave as perfect light absorbers with strongly varying reflec-tance, or as reflectors with very small phase-tuning ranges. Without a coherent theoretical framework to understand all these diversified responses of MIM structures, people have to rely on brute-force simulations to try different structures in order to find those fitting their own application requests (phase modulation,[42,43,57] perfect absorber,[58,59] or color printing[60–62]).

In a series of papers, Pors and Bozhevolnyi provided an intuitive physical explanation on the above issue, highlighting the importance of the spacer thickness. They noticed that MIM structures constitute cavities for GSP modes whose phase and group velocities are controlled by the spacing between the ground plane and the patch. This spacing can control the way that the GSP modes are reflected on the edges of the patch. When the gap is extremely thin, one can get very efficient, miniaturized absorbers. However, as the gap is increased, the decreased impedance mismatch between the GSP mode and the outside medium can decrease the reflection coefficient allowing the cavity to better exchange light with the outside, explaining why MIM structures with larger gaps do not

efficiently absorb light.[55,63] In 2015, Qu et al.[64] extended the arguments of the works of Pors and Bozhevolnyi to a more gen-eral scenario and provided a complete phase diagram to guide researchers design MIM metasurfaces with desired functionali-ties. Through a careful analysis based on coupled mode theory (CMT),[65] the authors established a complete phase diagram in which the EM properties of MIM structures fully controlled the competitions between two parameters—the intrinsic loss Γi and the radiation loss Γr of the resonance mode supported. According to CMT, the reflection coefficient r of such a system can be generally expressed as

ri f f

12

( )r

0 i r

= − +Γ

− − + Γ + Γ (3)

with f0 being the resonance frequency of the mode. Equation (3) indicates that an MIM metasurface can behave as a perfect absorber at Γi = Γr (dashed line in Figure 2b), but can strongly reflect EM waves (i.e., |r|→1) when the ratio Γi/Γr is far away from 1. Smith curves of the reflection coefficient r with dif-ferent values of Γi/Γr (right inset to Figure 2a) further reveal that the MIM structure with Γi < Γr represents an underdamped resonator with phase variation covering 2π as frequency passes through the resonance (blue line), while that with Γi > Γr rep-resents an overdamped resonator with small variation range of reflection phase (red line). Such a generic phase diagram can help people design their own MIM structures with desired functionalities. Since the two loss parameters are closely related to the geometrical/material parameters of the MIM structure,[66] one can simply vary their structural (say, the spacer thickness or the resonator’s structure) and materials parameters to make the two loss parameters of a given MIM system fall into the desired phase region. In a parallel line, Bowen et al. established a theory of perfect absorption in MIM systems through a semianalytical calculation based on the CMT,[67,68] which yielded similar con-clusions as previous theories[64,66] and further explained the off-axis perfect-absorption behaviors of MIM structures under TM excitations.

4. High-Efficiency Metasurfaces in Transmission Geometry

While reflective MIM metasurfaces can easily achieve very high working efficiencies, they suffer from the issue of wave inter-ferences and thus are inconvenient for realistic applications in some occasions. To construct high-efficiency metasurfaces in transmission mode, the crucial task is to find meta-atoms that can be (nearly) perfectly transparent for EM waves but exhibit transmission phases covering the full 2π range. This is again not achievable by either a single electric resonator or a V-shape multi-mode electric nanoantenna, because their impedances do not match with that of free space. In 2013, Pfeiffer and Grbic proposed a new scheme to design transparent metasurfaces (usually called Huygens’ surfaces), simply through tailoring both electric and magnetic responses within a single meta-atom.[69] As schematically shown in Figure 3a, to generate desired field distributions in two regions, fictitious surface electric and magnetic currents ( J

s and M

s ) are needed on the





surface S separating two regions which satisfy the boundary conditions, J n H H

ˆs 2 1( )= × − , M n E E

ˆs 2 1( )= × − , in order to com-pensate the field discontinuities. These surface currents fully dictate the electric and magnetic polarizabilities of the meta-atom at a particular position. With all meta-atoms determined, a metasurface can then be constructed.[69] Based on this con-cept, the authors designed the first high-efficiency transmissive metasurface in the microwave regime, constructed by spatially different meta-atoms composed by electric resonators and magnetic split-ring-resonators (SRR) with tailored properties (Figure 3b). The fabricated metasurface yields an efficiency of 86% for anomalous beam deflection in transmission geometry (Figure 3b). Other interesting applications were also demon-strated by appropriately designed Huygens’ surfaces, including polarization controller,[70] metalens,[35] generations of Bessel beams (Figure 3c).[71,72] Recently, Chen et al. presented a micro-wave reconfigurable metalens constructed based on the con-cept of Huygens’s surface, which can control the focal spot position dynamically with fast response time (control speed ≈105 switches s−1) (Figure 3d).[73] However, we note that these Huygens’ surfaces require complicated 3D structures (vertical SRRs) which are difficult to realize at high frequencies. In addi-tion, the nonflatness also adds other issues (such as mechan-ical instability) to the fabricated devices, making them less favorable for on-chip integration-optics applications. This also explains why transmission-mode metasurfaces of this type are rarely seen at frequencies other than microwave.

These issues motivated researchers to use multilayer meta-atoms, which are more fabrication-friendly and compatible with flat-optics, to construct transmission-mode high-efficiency metasurfaces. The core idea of this scheme is that, instead of adding vertical SRRs, one can also utilize the interlayer cou-plings to generate effective “magnetic responses” to match the impedance. The working principles of such a scheme can be understood from Figure 3e where the generic phase dia-gram of an ABA-type meta-atom is depicted. Consider an ABA trilayer structure with A and B representing two ultrathin MTM layers (thicknesses dA, dB, permittivity εA, εB, and permeability μA = μB = 1), a simple transfer-matrix-method calculation[74] shows that the system can support perfect transmissions[75–77] when the following condition

Z k d k dZ

Z

k d

k d

Z

Z

k d k dZ

Z

k d

k d

Z

Z

2 tan tantan

tan

2 tan tantan

tan

A2

A A B BA

B

B B

A A

B

A

A A B BB

A

B B

A A

A

B

( ) ( ) ( )( )

( ) ( ) ( )( )

− +

= − + (4)

is satisfied where k ki i0 ε= and Zi i1/ ε= , (i = A, B). Equation (4) can be met with different combinations of materials and geo-metrical parameters, yielding distinct transmission phase ϕ covering the whole 2π range (see Figure 3e). Here, perfect transparencies in an ABA structure are governed by different mechanisms in different parameter regions. For example, in the region of εA < 0, εB > 0 and εA > 0, εB < 0,[76] the transparency is induced by the scattering-cancellation mechanism, while the governing physics becomes the Fabry–Perot resonance in the regions of εA > 0, εB > 0. A large ϕ is naturally realized when both εA and εB are positively large. Such unique properties offer

researchers wide-range design freedoms to choose their meta-atoms for different purpose. For example, Sun et al. chose five meta-atoms in different phase regions of Figure 3e to construct a highly efficient surface plasmon polariton (SPP) coupler in the microwave regime,[78] which we will introduce with more details in Section 7.4. Pfeiffer et al. proposed an isotropic, broadband Huygens’ metasurface by cascading three iden-tical, patterned, metallic sheets together, to efficiently refract normally incident light with arbitrary polarization state in tel-ecommunication wavelength (Figure 3f).[79]

Stacking multiple anisotropic MTM layers with principle axes rotated, such meta-atoms can also control the polari-zation of transmitted waves, yielding more versatile wave-manipulation effects.[53,75,80–85] In 2013, Grady et al. fabricated a freestanding THz metasurface made by spatially modulated trilayer meta-atoms, each consisting of a cut-wire-like resonator with principle axis tilted at an angle of 45° sandwiched by a pair of perpendicularly aligned gratings (Figure 3g). Such uniquely designed meta-atoms can convert a linearly polarized incident THz wave into its cross-polarization over a wide bandwidth with working efficiency up to ≈80%.[82] The underlying physics can be understood as follows. While the two orthogonal gratings can largely cut off copolarized transmission and cross-polarized reflection, the remaining channel of copolarized reflection can be engineered to zero based scattering cancellations from dif-ferent layers. Due to energy conversion, the cross-polarized transmission is the only allowed channel which must exhibit a high efficiency. Such mechanism is closely related to that of metamaterial antireflection,[86] perfect absorbers,[87] and the ABA high transmission.[76] Through tuning the structures of inner resonators, the authors found a set of meta-atoms to con-struct a gradient metasurface, based on which the anomalous refraction of THz wave was experimentally demonstrated.[82] In a closely related work, Liu et al. designed a free-standing trans-missive metasurface and experimentally demonstrated that it can realize the functions of anomalous refraction and polari-zation conversion simultaneously.[71] The adopted meta-atom is again a trilayer structure but composed by three SRRs with successively varying orientations (Figure 3h), which provides a magnetoelectric coupling responsible for the 90° polarization conversion.

5. Pancharatnam–Berry Phase Metasurfaces

All metasurfaces discussed in last sections are designed for LP waves, which unfortunately cannot be applied to control CP waves. The latter are of great importance in many appli-cations such as chiral-molecule manipulations. Recently, Pancharatnam–Berry (PB) metasurfaces attracted intense attention due to their strongly capabilities in controlling CP waves. The design principle of PB metasurfaces is based on the geometric-phase concept. Consider two identical arbitrary-shape scatters placed on the xy-plane, with the second one rotated by an angle θ with respect to the first one. Shining the two scatters by the same CP light, one can find through simple calculations that the spin-flipped components of waves scattered by the two meta-atoms differ only with a phase factor ei2θ, independent of the scatter details and the frequency. Such a robust additional





phase, usually termed as the PB phase, is of a geometric-phase origin, which exactly equals to the solid angle of the area sur-rounded by two different paths connecting north and south poles on Poincare’s sphere (Figure 4a).[88] It is thus clear that the PB mechanism differs from usual resonant mechanism in two aspects: it can generate dispersionless phase shift and works for CP lights.

In 2011, Hasman and co-workers constructed a PB meta-surface by arranging circular apertures along a specific curve on a metal film.[89] Shine the metasurface by a CP light, the PB phases carried by waves scattered by those apertures exhibit linear dependence on lateral position, providing an additional k vector to the transmitted beam, generating a spin-dependent anomalous deflection of light (called as PSHE) (Figure 4b).[90,91] Zhang and co-workers further experimentally demonstrated the spin-dependent generalized Snell’s law based on a PB meta-surface composed by metallic nanorods with spatially rotating orientation angles (see Figure 4c,d).[92] Using Z-shaped apertures as meta-atoms to design a metasurface, Li et al. experimentally demonstrated transverse angular splitting of transmitted optical signals with opposite spins, due to significantly enhanced

spin–orbit interactions of light on the metasurface (Figure 4e).[93] Since geometric phases are solely determined by the orienta-tion angles of meta-atoms, PB metasurfaces exhibiting complex phase distributions can be easily designed, which can achieve various functionalities, including flat-lens imaging,[94–96] Bessel beam generation,[71,97] 2D or even 3D optical holograms,[98–101] and vortex beam generation.[93,102–105] For example, 3D optical holography was realized to project an on-axis 3D image with a wide field of view evaluated as ±40° (Figure 4f).[98] Shalaev and co-workers experimentally demonstrated that an optically active metasurface with thickness of λ/50 can rotate LP light by 45° with a broad bandwidth in the NIR region.[106] By carefully designing the locations and orientations of nanoantennas in the superunits, Shaltout et al. achieved an effective chiral meta-surface with a collection of nonchiral antennas (Figure 4g).[106] However, although PB metasurfaces exhibited powerful capa-bilities to manipulate light, most of early realizations are of very low efficiencies. The inherent physics is that an arbitrary meta-atom can also generate spin-conserved scatterings which do not have PB phases and thus do not contribute to or even can destroy the desired wave-manipulation effects. In fact, it


Figure 3. High-efficiency transmissive metasurfaces. a) Understanding the working mechanism of Huygens’ surface based on surface equivalence principle. b) Left-top panel: Sketch of a transmissive meta-atom consisting of an electric resonator and a magnetic one. Left-bottom panel: Measured H-field distribution for the beam-deflecting metasurface shined by normally incident wave at 15 GHz. Right-up panel: photograph of a fabricated microwave Huygens’ metasurface consisting of a stack of identical circuit-board strips. Right-lower panel: copper traces on the top and bottom sides of the boards provide the necessary electric and magnetic polarization currents, respectively. c) Left-top inset: Fabricated metasurface that converts a CP Gaussian beam into a Bessel beam. Right-top inset: Schematic of a section of the polarization rotator. Bottom panel: Measured transmitted H-field in the xy plane when a z-propagating, LCP Gaussian beam is incident on the metasurface. d) Active Huygens’ metalens for dynamic EM wave focusing, of which meta-atoms are biased by computer-controlled multichannel voltage sources. Inset: Capacitance-dependent phase and amplitude responses of EM wave transmission at 6.9 GHz for incident electric and magnetic fields which are parallel to y- and x-axes, respectively. e) Generic εA −εB phase diagram of an ABA-type meta-atom. |t = 1| line (black line) and five equal-φ lines (red: φ1 = 22°, orange: φ2 = 94°, green: φ3 = 166°, blue: φ4 = 238°, purple: φ5 = 310°) of the ABA structure (with μA = μB = 1, dA = 1.37 mm, and dB = 0.26 mm) f) Schematic of a transmissive metasurface that efficiently refracts a normally incident beam at telecommunication wavelengths. Right inset: Simulated E-field distribution for the beam-deflecting metasurface under normal-incidence illumination. g) Cross-polarized transmittance Spectrum obtained through experimental measurements, numerical simula-tions, and analytical calculations. Inset: Schematic of a unit cell for transmitted THz HWP. h) Measured transmission amplitude spectrum through free-standing transmissive THz metasurface, composed by three SRRs with successively varying orientations. Inset: sketch of the transmission-type coding metasurface to refract the normal incidence in anomalous direction with cross-polarization conversion. (a),(b) Reproduced with permission.[69] Copyright 2014, APS. (c) Reproduced with permission.[72] Copyright 2014, APS. (d) Reproduced with permission.[73] Copyright 2017 Wiley-VCH. (e) Reproduced with permission.[78] Copyright 2016 NPG. (f) Reproduced with permission.[82] Copyright 2013 AAAS. (g) Reproduced with permission.[71] Copyright 2017 ACS. (h) Reproduced with permission.[79] Copyright 2014 ACS.




was theoretically proved that the maximum working efficiency of a single-layer PB metasurface is bounded by 25%,[35] which was later experimentally demonstrated in the microwave regime (Figure 4h).[107]

In 2015, Luo et al. derived the general criteria for designing PB metasurfaces exhibiting 100% efficiency.[108] Consider a planar meta-atom with transmission/reflection properties given by two Jones matrices T(0) = diag(tuu,tvv) and R(0) = diag(ruu,rvv) in linear-polarization bases, where u and v denote its two prin-ciple axes. The Jones matrices of the meta-atom, rotated by an angle φ with respect to the z axis (see left inset to Figure 5a), can be expressed in terms of tuu, tvv, and ruu, rvv. Specifically, the transmission matrix of the rotated meta-atom in circular-polarization bases is found as

r r Ii

r r r r e e

ir r e e

uu vv uv vu uu vvi i

uv vui i

RR( )12

ˆ2

ˆ 12

ˆ ˆ

2ˆ ˆ

32 2

2 2

φ σ σ σ

σ σ

( )( )

( ) ( ) ( )

( )

= + + − + − +

+ + − +

φ φ

φ φ

−+ −

−+ −

(5)

where I is the identify matrix, { ˆ , ˆ , ˆ }1 2 3σ σ σ are three Pauli matrices and iˆ ( ˆ , ˆ )/ 21 2σ σ σ=± are two spin-flip operators.

Obviously, the first two terms in Equation (5), with spin main-tained, do not acquire a PB phase, while the last two terms do gain the PB phases e±i2ϕ. The transmission Jones matrix TT ( )φ has a similar form with RR ( )φ only with {ruu,ruv,rvu,rw} modified to {tuu,tuv,tvu,tvv}. When a PB metasurface is built up with such meta-atom, naturally it can generate four beams, i.e., normal and anomalous transmission/reflection modes (see Figure 5a), with efficiencies determined by the coefficients of different terms in Equation (5) and TT ( )φ .

In reflection geometry, the criterion to design a 100%-effi-ciency PB device is simply ruu + rvv = ruv − rvu = 0, indicating that the designed meta-atom should function as a perfect half wave plate. Following the pioneering work of refs. [57,109], anisotropic MIM structures working in the underdamping region (see Figure 2b) were widely used to control EM wave polarizations, at frequencies ranging from microwave to visible. Motivated by these works, Luo et al. designed two different types of PB metasurfaces, constructed by symmetric or asymmetric meta-atoms satisfying the above criterion, and experimentally dem-onstrated that both can achieve nearly 100%-efficiency PSHE in the microwave regime (Figure 5c,d).[108] Independently, Zheng et al. proposed a similar criterion to design high-efficiency PB


Figure 4. PB metasurfaces based on single-layer metallic structures. a) Geometrical original of the Pancharatnam–Berry phase illustrated by the Poin-caré sphere. b) Top panel: SEM image of a curved chain with local orientation varying linearly along the x axis. Bottom panel: Measured PSHE at the wavelength of 780 nm. Red and blue lines stand for incident right-handed and left-handed CP light, respectively. c) Right-top inset: SEM image of gold nanorods fabricated on an ITO coated glass substrate. Left inset: Schematic illustration of normal and anomalous refraction by dipole arrays when illuminated with LCP and RCP light, respectively. d) Reflection-angle versus incident angle for the PB metasurface shined by LCP light at the wavelength of 810 nm. e) PSHE at a PB metasurface with associated orbital rotation observed in experiments (left panel) and simulations (right panel). Top inset: A meta-atom containing a Z-shape aperture on gold film. Bottom inset: SEM image of the fabricated PB metasurface constructed by such meta-atoms. f) 3D holography consisting of subwavelength metallic nanorods with spatially varying orientations generates an on-axis “jet” image upon normal incidence of CP light.[98] g) Top panel: schematic view of polarization-controlled metasurface using two PB-phase subarrays (blue and red) rotating in opposite directions. Bottom panel: SEM image of part of fabricated sample with supercell highlighted by the dashed rectangle. Reproduced with permission.[106] Copyright 2014 ACS. h) Cross-polarization transmission-coefficient spectrum of a PB meta-atom depicted in inset, under the illumina-tion of a LCP wave. The maximum cross-polarization transmission coefficient is 0.5, implying a maximum 25% efficiency for PSHE. (a) Reproduced with permission.[88] Copyright 2002, OSA. (b) Reproduced with permission.[89] Copyright 2011, ACS. (c),(d) Reproduced with permission.[92] Copyright 2014, ACS. (e) Reproduced with permission.[93] Copyright 2014 ACS. (g) Reproduced with permission.[106] Copyright 2014, ACS. (h) Reproduced with permission.[107] Copyright 2015, Wiley-VCH.




metasurfaces, based on which they further demonstrated a metahologram of Einstein’s portrait at the working wavelength of 825 nm with nearly 80% efficiency (see Figure 5e,f).[110] Here, Ohmic losses of plasmonic metals, inevitably existing at optical frequencies, are responsible for the degraded efficiency.

However, 100% efficiency ultrathin PB metasurfaces in transmission geometry appear only very recently. Based on similar Jones-matrix arguments, the criterion to design 100% efficiency transmissive PB metasurfaces are derived as ruu = rvv = 0, |tuu| = |tvv| = 1, t tuu vvarg( ) arg( ) π= ± , meaning that the meta-atom should not only be perfectly transparent, but also behave as a half wave plate (HWP). Since an electric

dipole radiates symmetrically to its both sides, a single-layer PB meta-atom exhibiting purely electric response can never fulfill the above requirements. Luo et al. then argued that the desired meta-atom should contain both electric and magnetic responses, resulting in constructive interference at the trans-mission side and destructive interference at the reflection side. Further adding anisotropy to the meta-atom design, a trans-parent HWP can then be obtained. Based on this strategy, Luo et al. successfully designed such a meta-atom with ABA geom-etry in the microwave regime, and experimentally demonstrated that a PB metasurface constructed by such meta-atoms can achieve PSHE in transmission mode with an efficiency ≈91%


Figure 5. High-efficiency PB metasurfaces. a) Anomalous/normal transmissions and reflections (with efficiencies of Ta,Ra,Tn,Rn) generated by a generic PB metasurface consisting of meta-atoms with spatially varying orientation angles (0, φ, 2φ, 3φ, ⋅⋅⋅), under illumination of a LP beam. Inset: Jones matrices of a meta-atom with local axes rotated by an angle φ. b) Schematics of a φ-orientated meta-atom with transmission/reflection characteristics described by two Jones’ matrices T and R, and the E fields radiated from the electric and magnetic currents ( Je

and Jm

) generated on the meta-atom. c) Measured and simulated spectra of reflection phases (Φuu and Φvv) for the meta-atom shown in the inset. d) Measured normalized scattered-field intensities (color map) versus frequency and detecting angle for PB metasurface illuminated by normally incident linearly polarized beams, with receivers chosen as a LCP antenna. Inset: Picture of part of fabricated microwave sample. e) Simulated cross-polarization and copolarization reflectivity with normal light incidence with MIM meta-atom of the nanorod-based hologram shown by the inset. f) Schematic illustration of the working principle of a highly efficient reflective hologram employing a PB metasurface. Inset: SEM image of the fabricated metasurface. g) Absolute efficiencies of the four modes versus frequency for the PB metasurface obtained through analyzing experimental data. Inset: Sketch of the designed meta-atom. h) Measured scattered-field intensities (color map) in the transmission sides versus the frequency and the detecting angle for the PB metasurface illuminated by normally incident LP beams. Inset: Picture of part of the fabricated sample. i) Top panel: Pictures of vortex-beam generators with topological charge q = 1,2, constructed with the ABA meta-atom studied in Figure 5g. Measured E field distributions on an xy-plane 50 cm below two metasurfaces shined by normally incident RCP waves at the working frequency of 10.5 GHz (middle panel) and an off-working frequency of 14.5 GHz (bottom panel). (a),(b),(g),(h),(i) Reproduced with permission.[111] Copyright 2017, APS. (c),(d) Reproduced with permission.[108] Copyright 2015, Wiley-VCH. (e),(f) Reproduced with permission.[110] Copyright 2015, NGP.




(see Figure 5g,h).[111] In addition to high working efficiency, PB metadevices made by such meta-atoms can realize desired wave manipulations, without suffering from interferences with the normal modes. Figure 5i compares the measured patterns of vortex-beams with different topological charges, generated by PB metasurfaces made by the designed meta-atoms at their working frequency (thus exhibiting high efficiency) or at an off-working frequency.

6. All-Dielectric Metasurfaces

The inevitable Ohmic losses of metals at optical frequencies significantly degrade the performances of metasurfaces made by plasmonic structures, which stimulated the developments of dielectric metasurfaces.[112] In optical regime, low-loss dielectric resonators can also support both electric and magnetic reso-nances in analogy with their plasmonic counterparts, but with mechanism governed by the Mie resonances.[113–115] While the existence of electric resonance in dielectric resonators is easy

to expect, the magnetic resonance originates from the strong coupling of light with circulating displacement currents within the resonators, generating a strong magnetic dipole resonance inside (Figure 6a,b). In general, when the electric and mag-netic resonances of dielectric resonators are located at different frequencies, the transmission phase through the resonators can only cover up to π phase variation. By overlapping the two types of resonances in frequency via adjusting the resonator’s geom-etry, however, it is possible to achieve a phase variation cov-ering the entire 2π range with high transmission,[116,117] which allows one to achieve an optimal manipulation on transmitted light based on the concept of Huygens’ metasurfaces.[118–120] High transmission in dielectric resonators appears when the electric and magnetic resonances exhibit similar amplitudes and phases, resulting in the constructive interference in the forward direction and the destructive interference in the back-ward one. Huygens’ metasurfaces[118–120] of dielectric resona-tors were experimentally demonstrated by enlarging the overlap regimes of two types of resonances through tailoring the reso-nators’ geometries.[121,122] A metasurface composed of dielectric


Figure 6. Dielectric metasurfaces. a) Schematic of dielectric Huygens’ metasurface using arrays of nanodisks represented as electric and magnetic dipoles under x-polarized illumination. b) Electric (colored arrows) and magnetic (plain color) field distributions for the magnetic (top) and electric (bottom) mode of periodic Si nanodisks. c) A dielectric gradient metasurface optical element (DGMOE) serving as a blazed grating. Upper panel: theo-retical (continuous line) and experimental (solid squares) spectra of diffraction efficiency of the DGMOE. Inset: SEM image of the fabricated DGMOE. Bottom panel: measured diffraction patterns from the DGMOE under illumination with right circular polarization (left) and left circular polarization (right) at λ = 550 nm. d) Top panel: Top-view SEM image of a gradient metasurface used for beam-deflection experiments. Bottom panel: images taken in the back focal plane of the collection objective at 665, 715, and 775 nm (top to bottom, respectively) normalized to the maximum intensity in each image. It is seen that at a working wavelength of 715 nm most of the power is coupled into −1 order corresponding to the expected beam-deflection behavior. e) Sketch of hexagonal unit cell consisting of amorphous silicon nanoposts and the working principle of birefringent metasurface where x- and y-polarized incident light experience phase gradients of opposite sign. f) Simulation and measurement results of the device in (e), which separates x- and y-polarized light and focuses them to two different points. Inset: Side-view of an SEM image of part of the fabricated sample. g) Measured focusing efficiency of metalens consisting of high-aspect ratio TiO2 with PB-phase rotational morphology. Left inset: SEM image of part of a fabricated sample. Right inset: Schematic of dielectric metalens. h) High transmission and diffraction efficiency metaholograms with inset showing an SEM image of the fabricated Si nanopillar arrays. (a),(b) Reproduced with permission.[121] Copyright 2015, Wiley-VCH. (c) Reproduced with permission.[124] Copyright 2015, AAAS. (d) Reproduced with permission.[116] Copyright 2015, Wiley-VCH. (e),(f) Reproduced with permission.[125] Copyright 2015, NPG. (g) Reproduced with permission.[126] Copyright 2016, AAAS. (h) Reproduced with permission.[138] Copyright 2016, Optical Society of America.




nanoparticles with nearly perfect transmission over the broad spectral range, while exhibiting a controllable transmissive phase covering 2π, opens a novel avenue to efficiently manip-ulate light beam in NIR and optical regime. Many fascinating and highly efficient wave-control effects were demonstrated in both reflection and transmission geometries for both LP and CP beams.[118,120,123–140] For example, Brongersma and co-workers fabricated dielectric gradient metasurfaces based on silicon nanobeams with appropriate geometric dimension to realize high performance transmissive optical elements in the visible regime, such as ultrathin metagratings, metalenses, and meta-axicons.[124] As shown in Figure 6c, the fabricated dielec-tric gradient metasurface blazed grating can efficiently convert an incident right CP (RCP) beam into left CP (LCP) light, and steer it into the refracted angle determined by the phase gra-dient induced by the PB phases. Such devices operate over a broadband wavelength range (490–700 nm) with maximal theoretical working efficiency of 95% (experimental relative efficiency of 75% at 500 nm). Yu et al. experimentally demon-strated a beam deflection with ≈45% efficiency for incident light with arbitrary polarization at visible frequency (665–775 nm) based on a gradient metasurface with a superunit consisting of 8 elements of transparent silicon nanodisk (Figure 6d).[116] Recently, Arbabi et al. designed several metadevices utilizing elliptical high-contrast dielectric nanoposts as building blocks, which can provide complete control on polarization and phase in the NIR regime (Figure 6e). Their experiments demonstrated that such a metasurface platform can function as polarization beam splitter, wave-plates, and lenses with measured efficien-cies larger than 72%, through tailoring the ellipticiy, size, as well as orientation of nanoposts.[125] Figure 6f illustrates the incident-polarization-dependent focusing effect. In a parallel line, Capasso and co-workers experimentally demonstrated that metasurfaces made by high-aspect-ratio titanium dioxide can be used to realize high numerical aperture (NA) and high-efficiency metalens (Figure 6g),[126] broadband achro-matic metalens,[127,128] Bessel beam generator,[129] and so on.[11] Shalaev et al. experimentally demonstrated two high-efficiency polarization-sensitive dielectric metasurfaces (a vortex converter and a beam deflector) at telecom frequency.[120] Kivshar and co-workers designed transparent metaholograms composed of sub-diffraction lattices of silicon nanopillars that allow encoding high-resolution grayscale images (Figure 6h). The experimen-tally realized holograms exhibit very high (over 90%) with an operating spectrum bandwidth of 375 nm.[138] Very recently, Tsai and co-workers experimentally demonstrated a GaN-based achromatic metalens with focal length remaining unchanged as the incident wavelength is varied from 400 to 660 nm.[139,140] The authors demonstrated that the fabricated metalens can achieve full-color imaging, with measured efficiency reaching 40% over the whole visible spectrum.

7. Applications

The extraordinary abilities of metasurfaces in controlling EM waves, as demonstrated in last sections, offer people many possibilities to realize flat, highly efficient, and miniaturized metadevices fitting different application demands. In what

follows, we briefly review some of exciting progresses along this line, focusing on polarization control, metalenses, meta-holograms, and SPP couplers.

7.1. Polarization-Control Metadevices

Polarization is an important characteristic of EM waves and its full and efficient control is highly desired. In recent years, many different types of polarization-control metadevices were demonstrated in transmission[69,75,81,83,85,141–143] and reflection geometries,[42,43,55,57,64,82,109,141,142,144–153] at frequencies ranging from microwave to visible. For example, Yu et al. proposed a background-free broadband (5–12 μm) quarter-wave plate (QWP)[152] based on the metasurface concept (see Figure 7a). The supercell of the designed metasurface contains two lines of gradient-phased V-shaped antennas, which can scatter incident light to the same direction with same amplitudes and orthog-onal polarizations, but with a π/2 phase difference introduced by the off-set between the two subunits. Shining the device by a LP light at normal incidence, a CP light beam propagating at an off-normal angle is generated, which does not interfere with the incident and normally reflected lights. However, the working efficiency of such a device is quite low (around 5%), due to the multimode generations in such metasurfaces. To boost up the efficiency, Ding and co-workers proposed a broadband and background-free HWP based on reflective PB metasurface in the NIR regime by carefully designing a superunit with spatial phase gradient (Figure 7b).[149] The experimentally achieved polarization conversion efficiency can reach 95% with over 20% reflectance for the anomalously reflected cross-polarized wave at the working wavelength of 1000 nm. Based on the PB-phase modulation method, Wu et al. experimentally demonstrated a reflective metasurface capable of producing light beams with any polarizations, generated from a single LP light source.[153] In addition, such Al-based metasurfaces allow broadband operations covering the entire visible spectrum. Recently, polarization-control devices based on dielectric-metasur-faces[112,133,135,154] were also experimentally demonstrated with better performances compared to their plasmonic counterparts.

Besides polarization control, polarization determination and detection are also important in realistic applications. However, conventional approaches usually require multiple measure-ments to determine the Stokes parameters. PB Metasurfaces can deflect light with different CP components to different directions, offering an efficient approach to reduce the com-plexity of detection process.[108,155–157] As shown in Figure 7c, by measuring the intensities of spatially separated LCP and RCP light beams reflected from a PB metasurface shined by an arbitrarily polarized incident beam,[155] the complete informa-tion on the incident polarization can be retrieved but with the polarization azimuth angle undetermined in optical regime. Recently, Pors et al. demonstrated simultaneous determination of Stokes’ parameters of incident light based on a plasmonic metagratings with three sub-metasurface integrated at the wavelength of 800 nm,[158] as shown in Figure 7d. Very recently, Wu et al. proposed an on-chip polarimeter based on a reflec-tive PB metasurface (utilizing Al as plasmonic metal), which can resolve the complete Stokes parameters across the entire





visible spectrum.[159] Furthermore, other parameters of EM waves could also be detected with metasurface-based devices, for example, the orbital angular momentum.[133,160,161]

7.2. Metalenses

Conventional lenses rely on propagating-phase accumu-lations to refract light and therefore they must exhibit curved shapes. Metasurfaces utilized the local abrupt phase changes at the interfaces, which provide the possibility to make ultrathin and flat lenses suitable for future on-chip applications. To design a flat lens, a metasurface should exhibit

a phase profile x y x y F F( , )2 2 2 2φ πλ ( )= + + − for transmitted or

reflected waves, with λ being the wavelength and F the focal length. Aieta et al. demonstrated a metalens at telecom wave-length using V-shaped antennas (Figure 8a).[24] However, its efficiency is low and can only focus the cross-polarized light beam. Using complementary V-shaped apertures to build metalenses[162] did not solve the efficiency issue. Li et al. experimentally realized a reflective metalens in the microwave regime, which exhibits nearly 100% efficiency due to both neg-ligible losses in this frequency regime and the high-efficiency

MIM meta-atoms adopted.[39] Similarly, metalenses in MIM geometries were constructed in the NIR[40] (Figure 2e) and mid-infrared regimes[163] to achieve high-efficiency focusing. Utilizing the unique spin-dependent properties of PB phases, Huang and co-workers presented a flat lens that can realize helicity-dependent real or virtual image in visible and NIR regimes (Figure 8b).[94]

Knowing that plasmonic metalenses suffer from signifi-cant Ohmic losses at high frequencies, many recent efforts were devoted to dielectric metalenses.[11,131,132,164–166] In 2014, Lin et al. realized a transmissive metalens based on the PB mechanism employing silicon nanobeams as meta-atoms (Figure 8c).[134] Via careful structural tuning, the lens was designed to exhibit a focal length 100 μm with an NA = 0.43 at wavelength 500 nm. Shined by RCP light under normal incidence, the lens can concentrate the transmitted LCP light into a focal spot of size 670 nm, close to diffraction limit. Later, Arbabi et al. demonstrated that a silicon nano-post can serve as an efficient meta-atom, providing 2π phase coverage with subwavelength resolution and high transmis-sion, which can be used to construct high-efficiency metal-enses. A polarization-insensitive and high-NA transmissive metalens was experimentally reported with the measured focusing efficiency over 42% and a focal-spot size 0.57λ at


Figure 7. Polarization-control devices based on metasurfaces. a) Left panel: Schematic of a background-free metasurface-based QWP consisting of two V-shaped antenna subunits that generate two copropagating waves with orthogonal linear polarizations, equal amplitudes, and a λ/2 phase difference upon LP incidence. Right panel: Calculated degree of circular polarization and intensity of the extraordinary beam as a function of wavelength. Inset: SEM image of a fabricated sample. b) Left panel: Schematics of a reflective background-free HWP. Right panel: Measured normalized reflection for the gradient metasurface under the illumination of x-polarized light at wavelength of 1000 nm with different incident angles. Inset: SEM image of part of one fabricated metasurface with a supercell highlighted by light blue color. c) Experimental results of reflected power for LCP and RCP incident beams at different wavelengths as a function of reflected angle showing discrimination of LCP and RCP spectra. Top inset: Illustration of the metasurface used as a circular dichroism spectrometer using the PSHE. Bottom inset: SEM image of part of fabricated sample.[155] d) Left bottom inset: Illustration of the metagrating’s working principle.[158] An incoming beam with an unknown polarization state is on reflection represented in the six bases, thereby allowing one to determine the Stokes parameters and, hence, the state of polarization without the need of multiple measurements or an interferometric setup. Left-top inset: SEM image of part of fabricated sample. Right panel: Experimental verification of the metasurface polarization splitters: Normal-ized optical images of the diffraction spots for the two orthogonal polarizations that the metasurfaces are designed to split. The incident polarization state is denoted in each panel. (a) Reproduced with permission.[152] Copyright 2012 ACS, (b) Reproduced with permission.[149] Copyright 2015 ACS.




the wavelength of 1500 nm (Figure 8d).[167] These metalenses based on high contrast transmit-array were fabricated in one lithographic step, allowing the realization of more complex optical systems with new functionalities, such as planar ret-roreflector.[168] However, optical losses of silicon still sub-stantially degrade the metasurfaces’ efficiency in the visible regime, especially at shorter wavelengths. Recently, based on a new process to fabricate high-aspect-ratio titanium dioxide (TiO2) nanostructures, Capasso and co-workers successfully made large-NA (0.8) metalenses based on the PB mechanism,

which exhibit focusing efficiencies as high as 86% across the whole visible spectrum (Figure 6g). They further demon-strated polarization-independent lenses based on TiO2 posts with circular cross-section, exhibiting NA = 0.85 with diffrac-tion-limited focusing efficiency over 60% at the wavelengths of 532 and 660 nm.[138,165]

The intrinsic resonance dispersions in metasurfaces can cause undesired chromatic aberrations in metadevices, making their phase distributions deviating from the designed one at frequencies other than the working frequency. A cylindrical


Figure 8. Metalenses realized with metasurfaces. a) Flat focusing lens consisting of V-shaped nanoantenna array. Bottom left inset: zoom-in SEM image of the fabricated lens. Bottom right inset: 3D plots of intensity distribution of the lens on the focal plane. b) Dual-polarity metalens for vis-ible light realized with PB phase metasurface.[94] Bottom panel shows the SEM image of part of fabricated sample and its phase distribution for LCP light. c) A dielectric metalens based on Si nanobeams. Upper panel: SEM image of a fabricated dielectric metalens with a focal length of 100 mm at λ = 550 nm. (Inset) 2D intensity profile in the focal plane. Bottom panel: Measured intensity profile generated behind the dielectric metalens on the xz plane. The intensity distributions along the optical axis and through the focus are shown along the vertical and horizontal axes. d) A metalens with high NA and large efficiency based on Si high contrast transmit-arrays (HCTAs). Top-left panel: Measured 2D intensity profile at the focal plane for a microlens with d = 50 μm. Top-right panel: Top-view and side-view SEM images of fabricated Si transmission arrays. Bottom panel: Measured FWHM of focus spot size, transmission and focusing efficiency of the HCTA microlenses as a function of their focusing distance. e) Multi-wavelength dielectric metasurface cylindrical lens. Upper left panel: False colored side-view SEM image of the metalens with each unit cell identified by a different color. Upper right panel: Measured intensity distributions on the plane perpendicular to the silicon ridges. Bottom panel: Measured intensity profiles across the focal plane for three wavelengths of 1300, 1550, and 1800 nm. Intensity profiles are extracted from the measurement results in the upper right panel. f) Achromatic converging metalens based on MIM metasurface.[170] Top panel: Zoom-in SEM image of fabricated metalens (top left panel) and measured light intensity of focal spot at incident wavelength λ = 1500 nm (top right panel). Bottom panel: Experimental intensity profiles of broadband achromatic metalens along axial planes at incident wavelengths of 1200, 1500, and 1650 nm. (a) Reproduced with permission.[24] Copyright 2012, ACS. (c) Reproduced with permission.[134] Copyright 2015, AAAS. (d) Reproduced with permission.[167] Copyright 2015, NGP. (e) Reproduced with permis-sion.[164] Copyright 2015, ACS.




metalens exhibiting achromatic functionality at three dis-crete wavelengths (1300, 1550, and 1800 nm) was realized (Figure 8e),[164] by using meta-atoms containing multiple dielec-tric resonators thus possessing more design freedoms. Other approaches based on spatial multiplexing[137,138] and multilayer stacking[169] were also proposed to realize achromatic focusing effect with broadband property. Recently, Wang et al. estab-lished a scheme to design broadband achromatic metalenses employing both low-Q resonance-induced phases and geo-metric phases that can largely compensate each other. They experimentally demonstrated an achromatic metalens with an efficiency of ≈12% and NA = 0.26, which can focus light at the same focal plane within a large wavelength window of 1200–1680 nm (Figure 8f).[170]

7.3. Metaholograms

Due to its exotic controllability on local amplitude, phase, and polarization of light beam, metasurface can serve as an ideal platform to realize arbitrary computer-generated holo-grams (CGHs), opening up new application possibilities. In 2012, Walther et al. presented a fishnet-like binary-phase metadevice (Figure 9a), consisting of apertures of two dif-ferent types yielding required phases for transmitted waves, that can reconstruct two distinctive holograms (i.e., “META” and “CGH”) at two infrared wavelengths.[171] However, the working efficiency of the device is extremely low (0.1%), since the transmission through these subwavelength apertures must be weak. To enhance the working efficiency, reflective


Figure 9. Holograms realized with metasurfaces. a) Left panel: SEM image of part of a metasurface hologram consisting of nanoaperture antennas. Different colors represent pixels with distinctive transmission coefficients. Right top inset: Light intensity transmitted through the metasurface recorded in the far-field at λ1 = 905 nm and λ2 = 1385 nm, respectively. Right bottom inset: Schematic of the metasurface. b) Schematic of multicolor holograms using Al nanorods as the basic building block. Inset: SEM image of one pixel of the hologram. c) Schematic of a helicity-multiplexed hologram real-ized by a reflective PB metasurface. Inset: SEM image of part of the fabricated sample. d) Top panel: Schematic of the off-axis illumination method for metahologram (left), SEM image (top right), and schematic of the metaholograms (bottom right). Bottom panel: experimental results of a flower RGB image.[178] e) A polarization switchable phase hologram based on Si nanoposts, generating two patterns for x- and y-polarized light. Top panel: Schematic illustration and SEM image of the devices. Bottom panel: Simulated and measured intensity profiles, with “Caltech” generated for the case of x-polarization input and an icon for y-polarization input. f) Dispersion engineering and hologram design. Top panel: SEM image of the device. Middle and bottom panels: images generated by the hologram in the visible range. (a) Reproduced with permission.[171] Copyright 2012, Wiley-VCH. (b) Reproduced with permission.[172] Copyright 2015, ACS. (c) Reproduced with permission.[175] Copyright 2015, NPG. (e) Reproduced with permission.[125] Copyright 2015, NPG. (f) Reproduced with permission.[176] Copyright 2017, AAAS.




plasmonic metasurfaces were frequently adopted. Tsai and co-workers reported a multicolor metahologram design based on wavelength-multiplexing technique, using MIM meta-atoms resonant at different wavelengths within the visible regime.[172] Via careful design, the metadevice can encode phase profiles for three different holograms at three wavelengths (corre-sponding to red, green, and blue colors) (Figure 9b). Diversi-fied polarization-multiplexing metaholograms have also been realized in reflection and transmission geometries at different frequencies.[45,173,174] For example, still based on MIM meta-atoms, Chen et al. experimentally demonstrated a broadband metahologram with polarization-dependent dual images, exploring the polarization-sensitive responses of MIM resona-tors with in-plane anisotropy,[45] as we already mentioned early (Figure 2f). However, the working efficiencies of these metahol-ogram devices for LP light remain relatively low. The highest efficiency for LP plasmonic metahologram realized by Yifat et al. is around 50%, achieved with patch-dipole nanoantenna reflect-arrays at the wavelength of 1550 nm.[46]

Due to the significantly simplified design process, PB meta-surfaces were widely used to achieve metaholograms for CP light.[98,99,102,110,140,174–181] Very recently, Li et al. proposed an off-axis illumination method to overcome the limitation of cross-talking in multicolor metahologram design, and achieved multicolor metaholograms with remarkable image qualities employing a single type of plasmonic resonator. Multi-image hologram, seven-color metahologram, and full-color 3D meta-holograph were demonstrated experimentally in the visible regime (Figure 9d).[178]

Using MIM meta-atoms designed based on the high-efficiency criterion (Section 3), Zhang and co-workers dem-onstrated a PB metahologram with a maximum 80% at 825 nm,[110] which is the high efficiency so far achieved (Figure 5c). Based on such configuration, many helicity-multi-plexed metaholograms have been demonstrated.[99,102,178,182,183] For instance, Wen et al. experimentally demonstrated that a reflective metahologram combining two series of PB phase profiles can independently and efficiently project two encoded images into different sides of incident CP beams (Figure 9c). The measured conversion efficiency could reach 40% covering the working wavelength range from 620 to 1020 nm.[175]

Despite of exciting progresses in plasmonic metaholo-grams,[22,98,171,173,181,184] the low-efficiency issue of trans-missive plasmonic metasurfaces for visible light still hinders their practical applications. It is this reason that stimulates the rapid developments of all-dielectric metaholograms.[129,130,135,176,179,185–190] Recently, Faraon and co-workers reported that a polarization-switchable phase hologram can generate two different patterns for two orthogonally polarized LP lights.[125] The hologram was realized on a carefully designed dielectric metasurface where amorphous-silicon nanoposts with elliptical cross-sections provide complete control of amplitude and phase in transmission (Figure 9e). At the wavelength of 915 nm, the measured efficiencies of metahologram can reach to 84% and 91% for x- and y-polarized incident light. Khorasa-ninejad et al. experimentally demonstrated high-efficiency holo-grams with broadband and chirality-dependent functionalities, based on the concept of detour phase hologram.[176,185,186] The realized high-performance metaholograms exhibit working

efficiency as high as 75% and operating bandwidth covering from visible to NIR regime (Figure 9f).[176]

7.4. SPP Metacouplers

In addition to manipulating propagating waves (PWs), metasur-faces also exhibit extraordinary capabilities to couple PWs with surface waves (SWs), including SPPs and their low-frequency equivalents (i.e., spoof SPPs). SWs can find many applications in practice, but conventional devices to excite them (such as prisms and gratings) all suffer from the issues of bulky size and/or low excitation efficiency. In 2012, Sun et al. proposed that gradient metasurfaces can be used as efficient SPP cou-plers. Normally shining a metasurface with phase gradient ∇φ larger than free-space wave-vector k0, Sun et al. argued that the wave “reflected” by the metasurface is no longer a PW, but must be a SW since its parallel wave-vector is larger than k0 according to Equation (2). Distinct from the prism or grating couplers, here the phase gradient of metasurface can compensate the momentum mismatch between PW and SW and achieve a nearly perfect PW–SW conversion at any incidence angle larger than a critical value.[37,191,192] The authors then fabricated a reflective MIM metasurface with ∇φ > k0 in the microwave regime, and employed near-field scanning technique to map out the field pattern of the converted SW (Figure 10a,b) on the metasurface shined by a normally incident PW. Understanding that the generated SW is essentially a “driven-mode” not the eigen SPP, the authors further placed a purposely designed periodic metasurface supporting eigen-SPP-mode at the side of the gradient metasurface to couple out the driven SW to SPP. Such a side-coupling geometry was found rather insensitive to excitation angles, different from the resonant coupling mecha-nism adopted in prism or grating couplers. However, a later analysis showed that such a side-coupling configuration only works well for excitation light with small beam size.[191] While this is certainly good for on-chip applications, it also presents a fundamental issue for other applications.

Such a PW–SW coupling concept was soon experimentally demonstrated in NIR regime by Bozhevolnyi and co-workers. Pors et al. proposed a polarization-controlled unidirectional SPP coupler based on a reflective MIM metasurface in side-coupling geometry.[193] In particular, arrays of MIM meta-atoms can generate two independent orthogonal phase-gradients (being larger than k0) upon reflections by incident lights with two linear polarizations. Therefore, incident lights with dif-ferent polarizations can be efficiently (up to 25%) converted into SPPs propagating along different directions determined by the phase gradients (see Figure 10c for leakage-radiation-microscopy images obtained at the wavelength of 1500 nm).[193]

A different set of SPP couplers are formed by arrays of anisotropic rectangular apertures in opaque metal films, which can convert CP light to SPPs.[194–199] Lin et al. experimentally demonstrated a spin-dependent directional SPP coupler con-taining arrays of rectangular apertures with principle axes perpendicular to each other (see Figure 10d).[195] A rectan-gular aperture in a metal film can selectively scatter incident light with polarization parallel to its short axis, behaving as an equivalent electric dipole and generating SPPs. When such





aperture array is arranged in a column, the induced SPPs are plane waves propagating toward two sides of the column. As two columns with orthogonally oriented apertures are placed at a distance of quarter of SPP wavelength, the interference between launched SPPs yields directional SPP excitations dic-tated solely by the helicity of incident CP light (Figure 10d).[194] While such concept is already intimately linked with the PB concept, an alternative approach is to directly use PB metasur-faces.[198,199] Arranging anisotropic apertures in a lattice with their principle axes rotating successively, the designed PB metasurface can provide spin-dependent phase gradients to

compensate the k-mismatch between incident CP light and the eigen SPPs on the metal film. Zhang and co-workers experi-mentally demonstrated helicity-selective unidirectional excita-tions of SPPs at the optical regime (Figure 10e). However, the excitation efficiencies in these schemes are quite low (3–5%), since the apertures support normal transmissions and reflec-tions, which do not satisfy the 100% efficiency design criterion as discussed in Section 5.

Very recently, Zhou and co-workers proposed two approaches to solve the low-efficiency issues in SPP couplers for LP[78] and CP[200] excitations, respectively. Understanding that the key


Figure 10. SPP metacouplers. a) Schematic of the PW–SW coupling process on a gradient metasurface. b) Ez distributions (with phase information included) on part of the ξ = 1.14k0 metasurface under illumination of a normally incident x-polarized wave, obtained by near-field scanning measure-ments (top) and simulations (bottom). c) Polarization-sensitive SPP coupler fabricated in the NIR regime. Left-top panel: SEM image of the SPP coupler composed by 9 MIM meta-atoms arranged in a 3 × 3 array. LRM-images showing SPP excitation along y-direction for y-polarized light (right-top panel) and x-direction for x-polarized light (bottom panel) at working wavelength of λ = 1500 nm. d) Left panel: SEM image of a helicity-controlled SPP coupler composed by subwavelength apertures. Inset: Zoom-in view of part of the fabricated sample. Right panel: Near-field images of SPP generations as the coupler is illuminated by lights with different polarizations. e) Working principle (upper) and experimental demonstrations (bottom) of a spin-selective unidirectional SPP coupler based on a PB metasurface. Inset: SEM image of part of fabricated PB metasurface. f) Upper panel: Simulated E field pattern representing a Cherenkov surface-plasmon wake, generated by a 1D metamaterial containing rotated slits on a metal film, shined by S-polarized light. Middle panel: SEM picture of the metamaterial generating SPP wakes. Bottom panel: Near-field scanning optical microscopy image of swatching of propagation direction of the SPP wakes by changing spin properties of incident light. g) Upper left panel: A high-efficiency SPP metacoupler containing a transparent gradient metasurface placed at an optimum distance above the target plasmonic metal. Upper right panel: Picture of part of the fabri-cated sample and experimental setup where monopole antenna is used to probe the spoof SPPs field distribution. Bottom panel: Measured spectra of integrated reflection (blue stars, left axis) and intensities of the excited spoof SPPs (right axis) flowing to the right side (black squares) and left side (red circles). (a),(b) Reproduced with permission.[37] Copyright 2012, NPG. (c) Reproduced with permission.[193] Copyright 2014, NPG. (d) Reproduced with permission.[195] Copyright 2013, AAAS. (e) Reproduced with permission.[198] Copyright 2013, NGP. (f) Reproduced with permission.[203] Copyright 2015, AAAS. (g) Reproduced with permission.[78] Copyright 2016, NGP.




issues degrading the SPP excitations in previous schemes are the direct reflections at the devices’ surfaces (e.g., prism cou-plers) and the decoupling of SPPs to PWs by the scatterings at the inhomogeneous metasurfaces (side-coupling geometry), the authors proposed a new configuration for SPP excitation, which contains a transparent gradient metasurface put at an optimized distance above the plasmonic metal (Figure 10g).[78] Shining the metacoupler by a PW, the incident beam is first efficiently con-verted into a driven SW bounded on the metasurface without reflections, and then resonantly coupled to the eigen SW mode supported by the plasmonic metal. By suppressing the surface reflections and the decouplings from SPP to PW, such a new SPP coupler exhibits inherently high efficiency (94% in simula-tions). The excitation efficiency of spoof SPPs in the microwave regime, measured through near-field and far-field experiments, reaches ≈73%, several times higher than that realized by other available approaches. Later, the group proposed another scheme to solve the low-efficiency issue in SPP couplers for CP excitation. Choosing a 100% efficiency PB meta-atom with MIM configuration and designing carefully the guiding-out structure to make it support SPPs with both transvers-magnetic and transvers-electric polarizations, the authors experimentally demonstrated a direction-controllable spoof-SPP excitation with measured conversion efficiency ≈78% (92% for simulations) in the microwave regime.[200]

In addition to coupling SWs with PWs, metasurfaces were also applied to directly control SWs, which is also of great importance in photonics research.[100,201–203] In 2011, Li et al. experimentally realized a plasmonic Airy beam in the optical regime, based on a carefully designed metasurface con-sisting of nanocave array. SPP waves diffracted by nanocaves can interfere with each other in a predetermined way, which ultimately helps build two SPP Airy beams, exhibiting non-spreading, self-bending, and self-healing properties.[201] Based on similar SPP phase matching strategy, Li et al. further real-ized a broadband (bandwidth ≈100 nm) SPP lens and a SPP de-multiplexer (resolution ≈12 nm) in the visible regime.[202] Recently, Genevet et al. argued that a MTM composed by an array of rotated nanoslits in a metallic film can freely control the phase velocity of running wave of polarization; the latter is determined by the spin angular momentum of the incident photon and the incidence angle of the input beam (upper panel in Figure 10f). Therefore, such an MTM can steer the direc-tion of Cherenkov surface plasmon wakes generated on the metallic film, with physics again governed by the interferences between SPP waves scattered by nanoapertures under incident radiations (Figure 10f).[203] Xiao et al. demonstrated the inde-pendent generations of SPP patterns exhibiting different spins on a plasmonic metasurface, which is consisting of an array of nanoaperture with carefully designed orientation profiles.[100] Such flexible control enables the two spins to work together in a coherent way to make motion pictures played by varying the phase difference between the two spins. Similarly, many other SPP manipulation effects were achieved based on phase matching between SPPs scattered by different nanostructures such as grooves,[204,205] slits,[100,206–208] and V-antennas.[209,210]

However, these phase-matching approaches for controlling SWs still suffer from low-efficiency issue. Very recently, Dong et al. proposed to use metawalls, consisting of gradient-index

dielectric slab and a metal film, placed on a plasmonic surface to efficiently reshape the wave fronts of incident SPPs. Anoma-lous reflection of SW was experimentally demonstrated in the microwave regime, which satisfies the generalized Snell’s law and exhibits a high efficiency (≈70%).[211] Cui and co-workers proposed to use digital coding metasurfaces to control SWs.[212] With elaborately designed ellipse-shaped coding particles, var-ious coding metasurfaces were designed, which exhibit many high-efficiency SW-manipulation functionalities in the micro-wave regime.[213]

8. Conclusions and Perspectives

To summarize, we have briefly reviewed the key achievements in developing high-efficiency metasurfaces, including the core con-cept of wave-control based on flat optical devices, working mech-anisms, and physical realizations of high-efficiency metasurfaces in different geometries and for lights with different polarizations, and a few representative applications based on metasurfaces. As the field is growing very rapidly, it is impossible for us to cover all important aspects in its developments within this short review. For example, we are unfortunately not able to discuss several important sub-branches in metasurface research, such as nonlinear metasurfaces,[214–218] tunable metasurfaces,[15,219–227] parity-time metasurfaces,[228,229] hyperbolic metasurfaces,[230,231] and many other applications of metasurfaces, such as orbital angular momentum generations,[7,120,135,232–234] mathematical operations,[235,236] optical cloaking,[237] etc.

Before concluding this review, we would like to mention sev-eral important future directions in this field, based on our own perspectives:

[1] Active/Tunable Metasurfaces: Active control on metasurfaces is highly desired in practical applications, as it adds a fur-ther degree of freedom to control EM waves. In principle, tunable metasurfaces can be realized by adding external-knob-controllable elements to the meta-atoms or dielectric environments. Various materials[238,239] are lined up for ac-tive-device designs at different frequencies, such as positive intrinsic-negative (PIN) diodes,[224] phase-changing materi-als,[223,240] 2D materials,[219,241–243] micro-electromechanical systems,[244,245] liquid crystals,[246,247] and nonlinear optical materials.[248] For example, graphene-based metasurfaces have recently been demonstrated to exhibit wide-range phase modulation abilities in the terahertz regime[219] and multiple switchable functionalities in the mid-IR regime.[242] Howev-er, still lots of fundamental and technical issues should be solved before such devices can be really used in practice.

[2] Multifunctional Metasurfaces: Facing the increasing de-mands on data-storage capacity and information processing speed in modern science and technology, EM integrations has intrigued intensive attention with remarkable appli-cations. Metasurfaces could be the best candidate to integrate diversified multiple functionalities into one single ultrathin device. Different approaches, such as wave-length multiplexing,[127,179,234,249] polarization multiplex-ing,[45,49,101,125,172,250,251] and helicity multiplexing,[99,100,161,175] have been developed to achieve this goal. Just mention two





of them. Cai et al. experimentally demonstrated several high-efficiency polarization-dependent metadevices exhibiting two distinct functionalities in different geometries.[49,250] Very re-cently, Faraon and co-workers experimentally demonstrated that metasurfaces consisting of U-shaped dielectric meta-atoms can realize different holographic images under illumi-nations at incident angles in reflection geometry.[252]

[3] New Mechanisms for Constructing Metasurfaces with 100% Efficiencies: The challenge of large-angle beam-bending with 100% efficiency was considered as an inherent issue of gradi-ent metasurfaces constructed with the Huygens principle.[253] Recently, several groups proposed different strategies to over-come this issue, by engineering metasurfaces with impedance profiles based on exact boundary conditions imposed by the local fields including both the incident and the desired out-going beams.[254,255] Based on this approach, steering an inci-dent wave into an arbitrary direction[253–255] or even transform-ing it to surface waves[253] with unitary efficiency are proposed based on active/lossy metasurfaces or strongly nonlocal meta-surfaces. For practical considerations, people relax the mate-rial requirement by utilizing the evanescent[256] or leaky[257,258] modes, as such passive metasurface can exhibit the required nonlocal effects and achieve very high efficiencies. Very re-cently, Ra’di and co-workers proposed a new concept to design metasurfaces achieving 100% efficiency large-angle diffrac-tion,[253] and verified it in the microwave regime. We expect that this will be a very active sub-branch in metasurface re-search, and more different approaches will be established.

AcknowledgementsThis work was funded by National Key Research and Development Program of China (Grant Nos. 2017YFA0303504 and 2017YFA0700201), National Natural Science Foundation of China (Grant Nos. 11734007, 11474057, 11674068, and 11704240), and Natural Science Foundation of Shanghai (Grant Nos. 16ZR1445200, 16JC1403100, 18ZR1403400, 18QA1401800, and 17ZR1409500).

Conflict of InterestThe authors declare no conflict of interest.

Keywordselectromagnetic wave manipulations, geometric phases, high-efficiency, metasurfaces, MIM metasurfaces

Received: March 30, 2018Revised: May 28, 2018

Published online: July 29, 2018

[1] N. Engheta, R. W. Ziolkowski, Metamaterials: Physics and Engineering Explorations, John Wiley & Sons, Inc., Hoboken, NJ, USA 2006.

[2] W. Cai, V. Shalaev, Optical Metamaterials: Fundamentals and Appli-cations, Springer New York, New York, NY 2010.

[3] Y. Liu, X. Zhang, Chem. Soc. Rev. 2011, 40, 2494.

[4] D. R. Smith, J. B. Pendry, M. C. K. Wiltshire, Science 2004, 305, 788.

[5] C. M. Soukoulis, M. Wegener, Nat. Photonics 2011, 5, 523.[6] N. I. Zheludev, Y. S. Kivshar, Nat. Mater. 2012, 11, 917.[7] N. Yu, P. Genevet, M. a. Kats, F. Aieta, J.-P. Tetienne, F. Capasso,

Z. Gaburro, Science 2011, 334, 333.[8] N. Yu, F. Capasso, Nat. Mater. 2014, 13, 139.[9] H. Chen, A. J. Taylor, N. Yu, Rep. Prog. Phys. 2016, 79, 076401.

[10] H.-H. Hsiao, C. H. Chu, D. P. Tsai, Small Methods 2017, 1, 1600064.

[11] M. Khorasaninejad, F. Capasso, Science 2017, 358, eaam8100.[12] N. Meinzer, W. L. Barnes, I. R. Hooper, Nat. Photonics 2014, 8,

889.[13] F. Ding, A. Pors, S. I. Bozhevolnyi, Rep. Prog. Phys. 2018, 81,

026401.[14] S. B. Glybovski, S. A. Tretyakov, P. A. Belov, Y. S. Kivshar,

C. R. Simovski, Phys. Rep. 2016, 634, 1.[15] S. Liu, T. J. Cui, Adv. Opt. Mater. 2017, 5, 1700624.[16] P. Genevet, F. Capasso, F. Aieta, M. Khorasaninejad, R. Devlin,

Optica 2017, 4, 139.[17] C. L. Holloway, E. F. Kuester, J. A. Gordon, J. O’Hara, J. Booth,

D. R. Smith, IEEE Antennas Propag. Mag. 2012, 54, 10.[18] R. Blanchard, G. Aoust, P. Genevet, N. Yu, M. A. Kats, Z. Gaburro,

F. Capasso, Phys. Rev. B 2012, 85, 155457.[19] N. Yu, P. Genevet, F. Aieta, M. a. Kats, R. Blanchard, G. Aoust,

J.-P. Tetienne, Z. Gaburro, F. Capasso, IEEE J. Sel. Top. Quantum Electron. 2013, 19, 4700423.

[20] X. Ni, N. K. Emani, A. V. Kildishev, A. Boltasseva, V. M. Shalaev, Science 2012, 335, 427.

[21] X. Yin, Z. Ye, J. Rho, Y. Wang, X. Zhang, Science 2013, 339, 1405.[22] X. Ni, A. V Kildishev, V. M. Shalaev, Nat. Commun. 2013, 4,

2807.[23] D. Hu, X. Wang, S. Feng, J. Ye, W. Sun, Q. Kan, P. J. Klar, Y. Zhang,

Adv. Opt. Mater. 2013, 1, 186.[24] F. Aieta, P. Genevet, M. a. Kats, N. Yu, R. Blanchard, Z. Gaburro,

F. Capasso, Nano Lett. 2012, 12, 4932.[25] F. Aieta, P. Genevet, N. Yu, M. A. Kats, Z. Gaburro, F. Capasso,

Nano Lett. 2012, 12, 1702.[26] A. V. Kildishev, A. Boltasseva, V. M. Shalaev, Science 2013, 339,

1232009.[27] D. M. Pozar, S. D. Targonski, H. D. Syrigos, IEEE Trans. Antennas

Propag. 1997, 45, 287.[28] N. Gagnon, A. Petosa, D. A. McNamara, IEEE Antennas

Propag. Mag. 2013, 55, 29.[29] B. A. Munk, Frequency Selective Surfaces, John Wiley & Sons, Inc.,

Hoboken, NJ, USA 2000.[30] S. Larouche, D. R. Smith, Opt. Lett. 2012, 37, 2391.[31] D. R. Smith, Y.-J. Tsai, S. Larouche, IEEE Antennas Wireless

Propag. Lett. 2011, 10, 1605.[32] D. R. Smith, J. J. Mock, A. F. Starr, D. Schurig, Phys. Rev. E 2005,

71, 036609.[33] Y.-J. Tsai, S. Larouche, T. Tyler, G. Lipworth, N. M. Jokerst,

D. R. Smith, Opt. Express 2011, 19, 24411.[34] P. Lalanne, S. Astilean, P. Chavel, E. Cambril, H. Launois,

Opt. Lett. 1998, 23, 1081.[35] F. Monticone, N. M. Estakhri, A. Alù, Phys. Rev. Lett. 2013, 110,

203903.[36] A. Arbabi, A. Faraon, Sci. Rep. 2017, 7, 43722.[37] S. Sun, Q. He, S. Xiao, Q. Xu, X. Li, L. Zhou, Nat. Mater. 2012, 11,

426.[38] S. Sun, K. Y. Yang, C. M. Wang, T. K. Juan, W. T. Chen, C. Y. Liao,

Q. He, S. Xiao, W.-T. Kung, G.-Y. Guo, L. Zhou, D. P. Tsai, Nano Lett. 2012, 12, 6223.

[39] X. Li, S. Xiao, B. Cai, Q. He, T. J. Cui, L. Zhou, Opt. Lett. 2012, 37, 4940.






[40] A. Pors, M. G. Nielsen, R. L. Eriksen, S. I. Bozhevolnyi, Nano Lett. 2013, 13, 829.

[41] A. Pors, S. I. Bozhevolnyi, Opt. Express 2013, 21, 2942.[42] A. Pors, M. G. Nielsen, S. I. Bozhevolnyi, Opt. Lett. 2013, 38, 513.[43] S. Ma, X. Wang, W. Luo, S. Sun, Y. Zhang, Q. He, L. Zhou, EPL

2017, 117, 37007.[44] W. Ma, D. Jia, X. Yu, Y. Feng, Y. Zhao, Appl. Phys. Lett. 2016, 108,

071111.[45] W. T. Chen, K.-Y. Yang, C.-M. Wang, Y.-W. Huang, G. Sun,

I.-D. Chiang, C. Y. Liao, W.-L. Hsu, H. T. Lin, S. Sun, L. Zhou, A. Q. Liu, D. P. Tsai, Nano Lett. 2014, 14, 225.

[46] Y. Yifat, M. Eitan, Z. Iluz, Y. Hanein, A. Boag, J. Scheuer, Nano Lett. 2014, 14, 2485.

[47] H. F. Ma, G. Z. Wang, G. S. Kong, T. J. Cui, Sci. Rep. 2015, 5, 9605.[48] W.-L. Hsu, P. C. Wu, J.-W. Chen, T.-Y. Chen, B. H. Cheng,

W. T. Chen, Y.-W. Huang, C. Y. Liao, G. Sun, D. P. Tsai, Sci. Rep. 2015, 5, 11226.

[49] T. Cai, S. Tang, G. Wang, H. Xu, S. Sun, Q. He, L. Zhou, Adv. Opt. Mater. 2017, 5, 1600506.

[50] F. Ding, R. Deshpande, S. I. Bozhevolnyi, R. Deshpande, Light: Sci. Appl. 2017, 6, e17027.

[51] S. Boroviks, R. A. Deshpande, N. A. Mortensen, S. I. Bozhevolnyi, ACS Photonics 2017, 5, 1648

[52] H.-X. Xu, S. Tang, X. Ling, W. Luo, L. Zhou, Ann. Phys. 2017, 529, 1700045.

[53] S. Liu, T. J. Cui, Q. Xu, D. Bao, L. Du, X. Wan, W. X. Tang, C. Ouyang, X. Y. Zhou, H. Yuan, H. F. Ma, W. X. Jiang, J. Han, W. Zhang, Q. Cheng, Light: Sci. Appl. 2016, 5, e16076.

[54] J. M. Hao, L. Zhou, C. T. Chan, Appl. Phys. A 2007, 87, 281.[55] A. Pors, S. I. Bozhevolnyi, Opt. Express 2013, 21, 27438.[56] A. Pors, S. I. Bozhevolnyi, Opt. Mater. Express 2015, 5, 2448.[57] J. Hao, Y. Yuan, L. Ran, T. Jiang, J. Kong, C. Chan, L. Zhou,

Phys. Rev. Lett. 2007, 99, 63908.[58] N. Liu, M. Mesch, T. Weiss, M. Hentschel, H. Giessen, Nano Lett.

2010, 10, 2342.[59] J. Hao, J. Wang, X. Liu, W. J. Padilla, L. Zhou, M. Qiu,

Appl. Phys. Lett. 2010, 96, 1.[60] X. Zhu, C. Vannahme, E. Højlund-Nielsen, N. A. Mortensen,

A. Kristensen, Nat. Nanotechnol. 2016, 11, 325.[61] X. Duan, S. Kamin, N. Liu, Nat. Commun. 2017, 8, 14606.[62] M. Miyata, H. Hatada, J. Takahara, Nano Lett. 2016, 16, 3166.[63] S. I. Bozhevolnyi, T. Søndergaard, Opt. Express 2007, 15, 10869.[64] C. Qu, S. Ma, J. Hao, M. Qiu, X. Li, S. Xiao, Z. Miao, N. Dai,

Q. He, S. Sun, L. Zhou, Phys. Rev. Lett. 2015, 115, 235503.[65] S. Fan, W. Suh, J. D. Joannopoulos, J. Opt. Soc. Am. A 2003, 20, 569.[66] S. Ma, S. Xiao, L. Zhou, Phys. Rev. B 2016, 93, 045305.[67] P. T. Bowen, A. Baron, D. R. Smith, Phys. Rev. A 2016, 93, 063849.[68] P. T. Bowen, D. R. Smith, Phys. Rev. B 2014, 90, 195402.[69] C. Pfeiffer, A. Grbic, Phys. Rev. Lett. 2013, 110, 197401.[70] C. Pfeiffer, C. Zhang, V. Ray, L. Jay Guo, A. Grbic, Optica 2016, 3, 427.[71] S. Liu, A. Noor, L. L. Du, L. Zhang, Q. Xu, K. Luan, T. Wang,

Z. Tian, W. Tang, J. Han, W. Zhang, X. Zhou, Q. Cheng, T. Cui, ACS Photonics 2016, 3, 1968.

[72] C. Pfeiffer, A. Grbic, Phys. Rev. Appl. 2014, 2, 44012.[73] K. Chen, Y. Feng, F. Monticone, J. Zhao, B. Zhu, T. Jiang, L. Zhang,

Y. Kim, X. Ding, S. Zhang, A. Alù, C. Qiu, Adv. Mater. 2017, 29, 1606422.

[74] J. Hao, L. Zhou, Phys. Rev. B 2008, 77, 094201.[75] W. Sun, Q. He, J. Hao, L. Zhou, Opt. Lett. 2011, 36, 927.[76] L. Zhou, W. Wen, C. T. Chan, P. Sheng, Phys. Rev. Lett. 2005, 94,

243905.[77] Z. Song, Q. He, S. Xiao, L. Zhou, Appl. Phys. Lett. 2012, 101, 181110.[78] W. Sun, Q. He, S. Sun, L. Zhou, Light: Sci. Appl. 2016, 5, e16003.[79] C. Pfeiffer, N. K. Emani, A. M. Shaltout, A. Boltasseva,

V. M. Shalaev, A. Grbic, P. Carl, Naresh, E. K., S. Amr M.,

B. Alexandra, S. Vladimir M., G. Anthony, Nano Lett. 2014, 14, 2491.

[80] Y. Li, J. Zhang, S. Qu, J. Wang, L. Zheng, Y. Pang, Z. Xu, A. Zhang, J. Appl. Phys. 2015, 117, 044501.

[81] L. Cong, N. Xu, J. Gu, R. Singh, J. Han, W. Zhang, Laser Photonics Rev. 2014, 8, 626.

[82] N. K. Grady, J. E. Heyes, D. R. Chowdhury, Y. Zeng, M. T. Reiten, A. K. Azad, A. J. Taylor, D. A. R. Dalvit, H.-T. Chen, Science 2013, 340, 1304.

[83] L. Cong, W. Cao, Z. Tian, J. Gu, J. Han, W. Zhang, New J. Phys. 2012, 14, 115013.

[84] C. Pfeiffer, A. Grbic, Phys. Rev. Appl. 2014, 2, 44011.[85] R.-H. Fan, Y. Zhou, X.-P. Ren, R.-W. Peng, S.-C. Jiang, D.-H. Xu,

X. Xiong, X.-R. Huang, M. Wang, Adv. Mater. 2015, 27, 1201.[86] H.-T. Chen, J. Zhou, J. F. O’Hara, F. Chen, A. K. Azad, A. J. Taylor,

Phys. Rev. Lett. 2010, 105, 73901.[87] H.-T. Chen, Opt. Express 2012, 20, 7165.[88] Z. Bomzon, G. Biener, V. Kleiner, E. Hasman, Opt. Lett. 2002, 27, 1141.[89] N. Shitrit, I. Bretner, Y. Gorodetski, V. Kleiner, E. Hasman, Nano

Lett. 2011, 11, 2038.[90] C. Leyder, M. Romanelli, J. P. Karr, E. Giacobino, T. C. H. Liew,

M. M. Glazov, A. V. Kavokin, G. Malpuech, A. Bramati, Nat. Phys. 2007, 3, 628.

[91] K. Y. Bliokh, A. Niv, V. Kleiner, E. Hasman, Nat. Photonics 2008, 2, 748.[92] L. Huang, X. Chen, H. Mühlenbernd, G. Li, B. Bai, Q. Tan, G. Jin,

T. Zentgraf, S. Zhang, Nano Lett. 2012, 12, 5750.[93] G. Li, M. Kang, S. Chen, S. Zhang, E. Y. B. Pun, K. W. Cheah, J. Li,

Nano Lett. 2013, 13, 4148.[94] X. Chen, L. Huang, H. Mühlenbernd, G. Li, B. Bai, Q. Tan, G. Jin,

C. Qiu, S. Zhang, T. Zentgraf, Nat. Commun. 2012, 3, 1198.[95] D. Wen, S. Chen, F. Yue, K. Chan, M. Chen, M. Ardron, K. F. Li,

P. W. H. Wong, K. W. Cheah, E. Y. B. Pun, G. Li, S. Zhang, X. Chen, Adv. Opt. Mater. 2016, 4, 321.

[96] X. Chen, M. Chen, M. Q. Mehmood, D. Wen, F. Yue, C.-W. Qiu, S. Zhang, Adv. Opt. Mater. 2015, 3, 1201.

[97] Y. Y. Liu, Y. Ke, J. Zhou, Y. Y. Liu, H. Luo, S. Wen, D. Fan, Sci. Rep. 2017, 7, 44096.

[98] L. Huang, X. Chen, H. Mühlenbernd, H. Zhang, S. Chen, B. Bai, Q. Tan, G. Jin, K. Cheah, C. Qiu, J. Li, T. Zentgraf, S. Zhang, Nat. Commun. 2013, 4, 2808.

[99] L. Huang, H. Mühlenbernd, X. Li, X. Song, B. Bai, Y. Wang, T. Zentgraf, Adv. Mater. 2015, 27, 6444.

[100] S. Xiao, F. Zhong, H. Liu, S. Zhu, J. Li, Nat. Commun. 2015, 6, 8360.

[101] X. Li, H. Ren, X. Chen, J. Liu, Q. Li, C. Li, G. Xue, J. Jia, L. Cao, A. Sahu, B. Hu, Y. Wang, G. Jin, M. Gu, Nat. Commun. 2015, 6, 6984.

[102] M. Q. Mehmood, S. Mei, S. Hussain, K. Huang, S. Y. Siew, L. Zhang, T. Zhang, X. Ling, H. Liu, J. Teng, A. Danner, S. Zhang, C.-W. W. Qiu, Adv. Mater. 2016, 28, 2533.

[103] Y. Guo, M. Pu, Z. Zhao, Y. Wang, J. Jin, P. Gao, X. Li, X. Ma, X. Luo, ACS Photonics 2016, 3, 2022.

[104] E. Karimi, S. A. Schulz, I. De Leon, H. Qassim, J. Upham, R. W. Boyd, Light: Sci. Appl. 2014, 3, e167.

[105] F. Yue, D. Wen, C. Zhang, B. D. Gerardot, W. Wang, S. Zhang, X. Chen, Adv. Mater. 2017, 29, 1603838.

[106] A. Shaltout, J. Liu, V. M. Shalaev, A. V. Kildishev, Nano Lett. 2014, 14, 4426.

[107] X. Ding, F. Monticone, K. Zhang, L. Zhang, D. Gao, S. N. Burokur, A. de Lustrac, Q. Wu, C.-W. Qiu, A. Alù, Adv. Mater. 2015, 27, 1195.

[108] W. Luo, S. Xiao, Q. He, S. Sun, L. Zhou, Adv. Opt. Mater. 2015, 3, 1102.[109] J. Hao, Q. Ren, Z. An, X. Huang, Z. Chen, M. Qiu, L. Zhou,

Phys. Rev. A 2009, 80, 23807.[110] G. Zheng, H. Mühlenbernd, M. Kenney, G. Li, T. Zentgraf,

S. Zhang, Nat. Nanotechnol. 2015, 10, 308.





[111] W. Luo, S. Sun, H.-X. Xu, Q. He, L. Zhou, Phys. Rev. Appl. 2017, 7, 044033.

[112] S. Jahani, Z. Jacob, Nat. Nanotechnol. 2016, 11, 23.[113] S. Person, M. Jain, Z. Lapin, J. J. Sáenz, G. Wicks, L. Novotny,

Nano Lett. 2013, 13, 1806.[114] Y. H. Fu, A. I. Kuznetsov, A. E. Miroshnichenko, Y. F. Yu,

B. Luk’yanchuk, Nat. Commun. 2013, 4, 1527.[115] J. C. Ginn, I. Brener, D. W. Peters, J. R. Wendt, J. O. Stevens,

P. F. Hines, L. I. Basilio, L. K. Warne, J. F. Ihlefeld, P. G. Clem, M. B. Sinclair, Phys. Rev. Lett. 2012, 108, 097402.

[116] Y. F. Yu, A. Y. Zhu, R. Paniagua-Domínguez, Y. H. Fu, B. Luk’yanchuk, A. I. Kuznetsov, Laser Photonics Rev. 2015, 9, 412.

[117] Z. Zhou, J. Li, R. Su, B. Yao, H. Fang, K. Li, L. Zhou, J. Liu, D. Stellinga, C. P. Reardon, T. F. Krauss, X. Wang, ACS Photonics 2017, 4, 544.

[118] M. Decker, I. Staude, M. Falkner, J. Dominguez, D. N. Neshev, I. Brener, T. Pertsch, Y. S. Kivshar, Adv. Opt. Mater. 2015, 3, 813.

[119] A. I. Kuznetsov, A. E. Miroshnichenko, M. L. Brongersma, Y. S. Kivshar, B. Luk’yanchuk, B. Luk’yanchuk, B. Luk’yanchuk, Sci-ence 2016, 354, aag2472.

[120] M. I. Shalaev, J. Sun, A. Tsukernik, A. Pandey, K. Nikolskiy, N. M. Litchinitser, Nano Lett. 2015, 15, 6261.

[121] I. Staude, A. E. Miroshnichenko, M. Decker, N. T. Fofang, S. Liu, E. Gonzales, J. Dominguez, T. S. Luk, D. N. Neshev, I. Brener, Y. Kivshar, ACS Nano 2013, 7, 7824.

[122] B. S. Luk’yanchuk, N. V. Voshchinnikov, R. Paniagua-Domínguez, A. I. Kuznetsov, ACS Photonics 2015, 2, 993.

[123] Y. Yang, W. Wang, P. Moitra, I. I. Kravchenko, D. P. Briggs, J. Valentine, Nano Lett. 2014, 14, 1394.

[124] D. Lin, P. Fan, E. Hasman, M. L. Brongersma, Science 2014, 345, 298.[125] A. Arbabi, Y. Horie, M. Bagheri, A. Faraon, Nat. Nanotechnol. 2015,

10, 937.[126] M. Khorasaninejad, W. T. Chen, R. C. Devlin, J. Oh, A. Y. Zhu,

F. Capasso, Science 2016, 352, 1190.[127] M. Khorasaninejad, Z. Shi, A. Y. Zhu, W. T. Chen, V. Sanjeev,

A. Zaidi, F. Capasso, Nano Lett. 2017, 17, 1819.[128] W. T. Chen, A. Y. Zhu, V. Sanjeev, M. Khorasaninejad, Z. Shi,

E. Lee, F. Capasso, Nat. Nanotechnol. 2018, 13, 220.[129] W. T. Chen, M. Khorasaninejad, A. Y. Zhu, J. Oh, R. C. Devlin,

A. Zaidi, F. Capasso, Light: Sci. Appl. 2017, 6, e16259.[130] W. Zhao, B. Liu, H. Jiang, J. Song, Y. Pei, Y. Jiang, Opt. Lett. 2016,

41, 147.[131] K. E. Chong, I. Staude, A. James, J. Dominguez, S. Liu,

S. Campione, G. S. Subramania, T. S. Luk, M. Decker, D. N. Neshev, I. Brener, Y. S. Kivshar, Nano Lett. 2015, 15, 5369.

[132] E. Maguid, I. Yulevich, M. Yannai, V. Kleiner, M. L. Brongersma, E. Hasman, Light: Sci. Appl. 2017, 6, e17027.

[133] D. Lin, A. L. Holsteen, E. Maguid, G. Wetzstein, P. G. Kik, E. Hasman, M. L. Brongersma, Nano Lett. 2016, 16, 7671.

[134] Y. Nagasaki, M. Suzuki, J. Takahara, Nano Lett. 2017, 17, 7500.[135] B. Wang, F. Dong, Q.-T. T. Li, D. Yang, C. Sun, J. Chen, Z. Song,

L. Xu, W. Chu, Y.-F. F. Xiao, Q. Gong, Y. Li, Nano Lett. 2016, 16, 5235.[136] H. Zhang, X. Zhang, Q. Xu, C. Tian, Q. Wang, Y. Xu, Y. Li,

J. Gu, Z. Tian, C. Ouyang, X. Zhang, C. Hu, J. Han, W. Zhang, Adv. Opt. Mater. 2018, 6, 1700773.

[137] M. Lawrence, D. R. Barton, J. A. Dionne, Nano Lett. 2018, 18, 1104.[138] L. Wang, S. Kruk, H. Tang, T. Li, I. Kravchenko, D. N. Neshev,

Y. S. Kivshar, Optica 2016, 3, 1504.[139] S. Wang, P. C. Wu, V.-C. Su, Y.-C. Lai, M.-K. Chen, H. Y. Kuo,

B. H. Chen, Y. H. Chen, T.-T. Huang, J.-H. Wang, R.-M. Lin, C.-H. Kuan, T. Li, Z. Wang, S. Zhu, D. P. Tsai, Nat. Nanotechnol. 2018, 13, 227.

[140] B. H. Chen, P. C. Wu, V. C. Su, Y. C. Lai, C. H. Chu, I. C. Lee, J. W. Chen, Y. H. Chen, Y. C. Lan, C. H. Kuan, D. P. Tsai, Nano Lett. 2017, 17, 6345.

[141] J. Shao, J. Li, Y.-H. Wang, J.-Q. Li, Q. Chen, Z.-G. Dong, J. Appl. Phys. 2014, 115, 243503.

[142] X. Gao, X. Han, W. P. Cao, H. O. Li, H. F. Ma, T. J. Cui, IEEE Trans. Antennas Propag. 2015, 63, 3522.

[143] T. Li, H. Liu, S.-M. Wang, X.-G. Yin, F.-M. Wang, S.-N. Zhu, X. Zhang, Appl. Phys. Lett. 2008, 93, 21110.

[144] Y. Z. Cheng, W. Withayachumnankul, A. Upadhyay, D. Headland, Y. Nie, R. Z. Gong, M. Bhaskaran, S. Sriram, D. Abbott, Appl. Phys. Lett. 2014, 105, 181111.

[145] Z. Chen, C. Wang, F. Xu, Y. Lou, B. Cao, X. Li, Photonics Nano-struct. 2014, 12, 189.

[146] S.-C. Jiang, X. Xiong, Y.-S. Hu, Y.-H. Hu, G.-B. Ma, R.-W. Peng, C. Sun, M. Wang, Phys. Rev. X 2014, 4, 21026.

[147] Y. Guo, Y. Wang, M. Pu, Z. Zhao, X. Wu, X. Ma, C. Wang, L. Yan, X. Luo, Sci. Rep. 2015, 5, 8434.

[148] Y. Dai, W. Ren, H. Cai, H. Ding, N. Pan, X. Wang, Opt. Express 2014, 22, 7465.

[149] F. Ding, Z. Wang, S. He, V. M. Shalaev, A. V. Kildishev, ACS Nano 2015, 9, 4111.

[150] Q. Lévesque, M. Makhsiyan, P. Bouchon, F. Pardo, J. Jaeck, N. Bardou, C. Dupuis, R. Haïdar, J.-L. L. Pelouard, Appl. Phys. Lett. 2014, 104, 2013.

[151] T. Ribaudo, A. Taylor, B.-M. Nguyen, D. Bethke, E. A. Shaner, Opt. Express 2014, 22, 2821.

[152] N. Yu, F. Aieta, P. Genevet, M. a. Kats, Z. Gaburro, F. Capasso, Nano Lett. 2012, 12, 6328.

[153] P. C. Wu, W.-Y. Tsai, W. T. Chen, Y.-W. Huang, T.-Y. Chen, J.-W. Chen, C. Y. Liao, C. H. Chu, G. Sun, D. P. Tsai, Nano Lett. 2017, 17, 445.

[154] S. Kruk, B. Hopkins, I. I. Kravchenko, A. Miroshnichenko, D. N. Neshev, Y. S. Kivshar, APL Photonics 2016, 1, 030801.

[155] A. Shaltout, J. Liu, A. Kildishev, V. Shalaev, Optica 2015, 2, 860.[156] D. Wen, F. Yue, S. Kumar, Y. Ma, M. Chen, X. Ren, P. E. Kremer,

B. D. Gerardot, M. R. Taghizadeh, G. S. Buller, X. Chen, Opt. Express 2015, 23, 10272.

[157] J. P. Balthasar Mueller, K. Leosson, F. Capasso, Optica 2016, 3, 42.[158] A. Pors, M. G. Nielsen, S. I. Bozhevolnyi, Optica 2015, 2, 716.[159] P. C. Wu, J.-W. Chen, C.-W. Yin, Y.-C. Lai, T. L. Chung, C. Y. Liao,

B. H. Chen, K.-W. Lee, C.-J. Chuang, C.-M. Wang, D. P. Tsai, ACS Photonics 2018, 5, 2568.

[160] P. Genevet, J. Lin, M. a. Kats, F. Capasso, Nat. Commun. 2012, 3, 1278.[161] J. P. Balthasar Mueller, N. A. Rubin, R. C. Devlin, B. Groever,

F. Capasso, Phys. Rev. Lett. 2017, 118, 113901.[162] X. Ni, S. Ishii, A. V. Kildishev, V. M. Shalaev, Light Sci. Appl. 2013,

2, e72.[163] S. Zhang, M.-H. Kim, F. Aieta, A. She, T. Mansuripur, I. Gabay,

M. Khorasaninejad, D. Rousso, X. Wang, M. Troccoli, N. Yu, F. Capasso, Opt. Express 2016, 24, 18024.

[164] M. Khorasaninejad, F. Aieta, P. Kanhaiya, M. A. Kats, P. Genevet, D. Rousso, F. Capasso, Nano Lett. 2015, 15, 5358.

[165] M. Khorasaninejad, A. Y. Zhu, C. Roques-Carmes, W. T. Chen, J. Oh, I. Mishra, R. C. Devlin, F. Capasso, Nano Lett. 2016, 16, 7229.

[166] B. Groever, W. T. Chen, F. Capasso, Nano Lett. 2017, 17, 4902.[167] A. Arbabi, Y. Horie, A. J. Ball, M. Bagheri, A. Faraon, Nat. Commun.

2015, 6, 7069.[168] A. Arbabi, E. Arbabi, Y. Horie, S. M. Kamali, A. Faraon, Nat. Photo-

nics 2017, 11, 415.[169] O. Avayu, E. Almeida, Y. Prior, T. Ellenbogen, Nat. Commun. 2017,

8, 14992.[170] S. Wang, P. C. Wu, V.-C. Su, Y.-C. Lai, C. Hung Chu, J.-W. Chen,

S.-H. Lu, J. Chen, B. Xu, C.-H. Kuan, T. Li, S. Zhu, D. P. Tsai, Nat. Commun. 2017, 8, 187.

[171] B. Walther, C. Helgert, C. Rockstuhl, F. Setzpfandt, F. Eilenberger, E.-B. Kley, F. Lederer, A. Tünnermann, T. Pertsch, Adv. Mater. 2012, 24, 6300.





[172] Y. W. Huang, W. T. Chen, W. Y. Tsai, P. C. Wu, C. M. Wang, G. Sun, D. P. Tsai, Nano Lett. 2015, 15, 3122.

[173] Z. Xie, T. Lei, G. Si, X. Wang, J. Lin, C. Min, X. Yuan, ACS Photonics 2017, 4, 2158.

[174] W. Wan, J. Gao, X. Yang, ACS Nano 2016, 10, 10671.[175] D. Wen, F. Yue, G. Li, G. Zheng, K. Chan, S. Chen, M. Chen,

K. F. Li, P. W. H. Wong, K. W. Cheah, E. Yue Bun Pun, S. Zhang, X. Chen, Nat. Commun. 2015, 6, 8241.

[176] M. Khorasaninejad, A. Ambrosio, P. Kanhaiya, F. Capasso, Sci. Adv. 2016, 2, e1501258.

[177] X. Zhang, J. Jin, Y. Wang, M. Pu, X. Li, Z. Zhao, P. Gao, C. Wang, X. Luo, Sci. Rep. 2016, 6, 19856.

[178] X. Li, L. Chen, Y. Li, X. Zhang, M. Pu, Z. Zhao, X. Ma, Y. Wang, M. Hong, X. Luo, Sci. Adv. 2016, 2, e1601102.

[179] B. Wang, B. Quan, J. He, Z. Xie, X. Wang, J. Li, Q. Kan, Y. Zhang, Sci. Rep. 2016, 6, 35657.

[180] S. Choudhury, U. Guler, A. Shaltout, V. M. Shalaev, A. V. Kildishev, A. Boltasseva, Adv. Opt. Mater. 2017, 5, 1700196.

[181] R. C. Devlin, M. Khorasaninejad, W. T. Chen, J. Oh, F. Capasso, Proc. Natl. Acad. Sci. USA 2016, 113, 10473.

[182] C. Zhang, F. Yue, D. Wen, M. Chen, Z. Zhang, W. Wang, X. Chen, ACS Photonics 2017, 4, 1906.

[183] X. Zhang, M. Pu, J. Jin, X. Li, P. Gao, X. Ma, C. Wang, X. Luo, Ann. Phys. 2017, 529, 1700248.

[184] W. Ye, F. Zeuner, X. Li, B. Reineke, S. He, C.-W. Qiu, J. Liu, Y. Wang, S. Zhang, T. Zentgraf, Nat. Commun. 2016, 7, 11930.

[185] B. C. Kress, P. Meyrueis, Applied Digital Optics, John Wiley & Sons, Ltd, Chichester, UK 2009.

[186] A. W. Lohmann, D. P. Paris, Appl. Opt. 1967, 6, 1739.[187] M. Khorasaninejad, W. T. Chen, A. Y. Zhu, J. Oh, R. C. Devlin,

D. Rousso, F. Capasso, Nano Lett. 2016, 16, 4595.[188] K. E. Chong, L. Wang, I. Staude, A. R. James, J. Dominguez,

S. Liu, G. S. Subramania, M. Decker, D. N. Neshev, I. Brener, Y. S. Kivshar, ACS Photonics 2016, 3, 514.

[189] R. Paniagua-Domínguez, Y. F. Yu, E. Khaidarov, S. Choi, V. Leong, R. M. Bakker, X. Liang, Y. H. Fu, V. Valuckas, L. A. Krivitsky, A. I. Kuznetsov, Nano Lett. 2018, 18, 2124.

[190] V. V. Kotlyar, A. G. Nalimov, S. S. Stafeev, C. Hu, L. O’Faolain, M. V. Kotlyar, D. Gibson, S. Song, Opt. Express 2017, 25, 8158.

[191] C. Qu, S. Xiao, S. Sun, Q. He, L. Zhou, EPL 2013, 101, 54002.[192] S. Xiao, Q. He, C. Qu, X. Li, S. Sun, L. Zhou, Opt. Express 2013, 21,

27219.[193] A. Pors, M. G. Nielsen, T. Bernardin, J.-C. Weeber, S. I. Bozhevolnyi,

Light Sci. Appl. 2014, 3, e197.[194] F. J. Rodriguez-Fortuno, G. Marino, P. Ginzburg, D. O’Connor,

A. Martinez, G. A. Wurtz, A. V. Zayats, Science 2013, 340, 328.[195] J. Lin, J. P. B. Mueller, Q. Wang, G. Yuan, N. Antoniou, X.-C. Yuan,

F. Capasso, Science 2013, 340, 331.[196] A. E. Miroshnichenko, Y. S. Kivshar, Science 2013, 340, 283.[197] J. Yang, S. Zhou, C. Hu, W. Zhang, X. Xiao, J. Zhang, Laser Photo-

nics Rev. 2014, 8, 590.[198] L. Huang, X. Chen, B. Bai, Q. Tan, G. Jin, T. Zentgraf, S. Zhang,

Light Sci. Appl. 2013, 2, e70.[199] H. Mühlenbernd, P. Georgi, N. Pholchai, L. Huang, G. Li,

S. Zhang, T. Zentgraf, ACS Photonics 2016, 3, 124.[200] J. Duan, H. Guo, S. Dong, T. Cai, W. Luo, Z. Liang, Q. He, L. Zhou,

S. Sun, Sci. Rep. 2017, 7, 1354.[201] L. Li, T. Li, S. M. Wang, C. Zhang, S. N. Zhu, Phys. Rev. Lett. 2011,

107, 126804.[202] L. Li, T. Li, S. Wang, S. Zhu, X. Zhang, Nano Lett. 2011, 11, 4357.[203] P. Genevet, D. Wintz, A. Ambrosio, A. She, R. Blanchard,

F. Capasso, Nat. Nanotechnol. 2015, 10, 804.[204] Y.-G. Chen, Y.-H. Chen, Z.-Y. Li, Opt. Lett. 2014, 39, 339.[205] D. Wintz, P. Genevet, A. Ambrosio, A. Woolf, F. Capasso, Nano

Lett. 2015, 15, 3585.

[206] S.-Y. Lee, K. Kim, S.-J. Kim, H. Park, K.-Y. Kim, B. Lee, Optica 2015, 2, 6.[207] S. S. Kou, G. Yuan, Q. Wang, L. Du, E. Balaur, D. Zhang, D. Tang,

B. Abbey, X.-C. Yuan, J. Lin, Light Sci. Appl. 2016, 5, e16034.[208] J. Lin, Q. Wang, G. Yuan, L. Du, S. S. Kou, X.-C. Yuan, Sci. Rep.

2015, 5, 10529.[209] D. Wintz, A. Ambrosio, A. Y. Zhu, P. Genevet, F. Capasso, ACS

Photonics 2017, 4, 22.[210] E. Y. Song, S. Y. Lee, J. Hong, K. Lee, Y. Lee, G. Y. Lee, H. Kim,

B. Lee, Laser Photonics Rev. 2016, 10, 299.[211] S. Dong, Y. Zhang, H. Guo, J. Duan, F. Guan, Q. He, H. Zhao,

L. Zhou, S. Sun, Phys. Rev. Appl. 2018, 9, 014032.[212] T. J. Cui, S. Liu, L. Zhang, J. Mater. Chem. C 2017, 5, 3644.[213] S. Liu, T. J. Cui, A. Noor, Z. Tao, H. C. Zhang, G. D. Bai, Y. Yang,

X. Y. Zhou, Light Sci. Appl. 2018, 7, 18008.[214] S. Keren-Zur, O. Avayu, L. Michaeli, T. Ellenbogen, ACS Photonics

2016, 3, 117.[215] G. Li, S. Chen, N. Pholchai, B. Reineke, P. W. H. Wong, E. Y. B. Pun,

K. W. Cheah, T. Zentgraf, S. Zhang, Nat. Mater. 2015, 14, 607.[216] A. E. Minovich, A. E. Miroshnichenko, A. Y. Bykov, T. V. Murzina,

D. N. Neshev, Y. S. Kivshar, Laser Photonics Rev. 2015, 9, 195.[217] G. Li, S. Zhang, T. Zentgraf, Nat. Rev. Mater. 2017, 2, 17010.[218] N. Nookala, J. Lee, M. Tymchenko, J. Sebastian Gomez-Diaz,

F. Demmerle, G. Boehm, K. Lai, G. Shvets, M.-C. Amann, A. Alu, M. Belkin, Optica 2016, 3, 283.

[219] Z. Miao, Q. Wu, X. Li, Q. He, K. Ding, Z. An, Y. Zhang, L. Zhou, Phys. Rev. X 2015, 5, 41027.

[220] Y.-W. Huang, H. W. H. Lee, R. Sokhoyan, R. A. Pala, K. Thyagarajan, S. Han, D. P. Tsai, H. A. Atwater, Nano Lett. 2016, 16, 5319.

[221] A. She, S. Zhang, S. Shian, D. R. Clarke, F. Capasso, Sci. Adv. 2018, 4, eaap9957.

[222] S. M. Kamali, A. Arbabi, E. Arbabi, Y. Horie, A. Faraon, Nat. Commun. 2016, 7, 11618.

[223] Q. Wang, E. T. F. Rogers, B. Gholipour, C.-M. Wang, G. Yuan, J. Teng, N. I. Zheludev, Nat. Photonics 2016, 10, 60.

[224] H.-X. Xu, S. Ma, W. Luo, T. Cai, S. Sun, Q. He, L. Zhou, Appl. Phys. Lett. 2016, 109, 193506.

[225] H.-X. Xu, S. Sun, S. Tang, S. Ma, Q. He, G.-M. Wang, T. Cai, H.-P. Li, L. Zhou, Sci. Rep. 2016, 6, 27503.

[226] B. O. Zhu, K. Chen, N. Jia, L. Sun, J. Zhao, T. Jiang, Y. Feng, Sci. Rep. 2015, 4, 4971.

[227] L. Li, T. Jun Cui, W. Ji, S. Liu, J. Ding, X. Wan, Y. Bo Li, M. Jiang, C. Qiu, S. Zhang, Nat. Commun. 2017, 8, 197.

[228] M. Lawrence, N. Xu, X. Zhang, L. Cong, J. Han, W. Zhang, S. Zhang, Phys. Rev. Lett. 2014, 113, 093901.

[229] F. Monticone, C. A. Valagiannopoulos, A. Alù, Phys. Rev. X 2016, 6, 041018.

[230] J. S. T. Smalley, F. Vallini, S. A. Montoya, L. Ferrari, S. Shahin, C. T. Riley, B. Kanté, E. E. Fullerton, Z. Liu, Y. Fainman, Nat. Commun. 2017, 8, 13793.

[231] O. V. Kotov, Y. E. Lozovik, Phys. Rev. B 2017, 96, 235403.[232] J. A. H. Van Nieuwstadt, M. Sandtke, R. H. Harmsen,

F. B. Segerink, J. C. Prangsma, S. Enoch, L. Kuipers, Phys. Rev. Lett. 2006, 97, 1.

[233] J. He, X. Wang, D. Hu, J. Ye, S. Feng, Q. Kan, Y. Zhang, Opt. Express 2013, 21, 20230.

[234] Z. Shi, M. Khorasaninejad, Y.-W. Huang, C. Roques-Carmes, A. Y. Zhu, W. T. Chen, V. Sanjeev, Z. Ding, M. Tamagnone, K. Chaudhary, R. C. Devlin, C. Qiu, F. Capasso, Nano Lett. 2018, 18, 2420.

[235] A. Silva, F. Monticone, G. Castaldi, V. Galdi, A. Alu, N. Engheta, Science 2014, 343, 160.

[236] A. Pors, M. G. Nielsen, S. I. Bozhevolnyi, Nano Lett. 2015, 15, 791.[237] X. Ni, Z. J. Wong, M. Mrejen, Y. Wang, X. Zhang, Science 2015,

349, 1310.[238] A. Y. Zhu, A. I. Kuznetsov, B. Luk’yanchuk, N. Engheta, P. Genevet,

Nanophotonics 2017, 6, 1.





[239] A. M. Shaltout, N. Kinsey, J. Kim, R. Chandrasekar, J. C. Ndukaife, A. Boltasseva, V. M. Shalaev, Proc. IEEE 2016, 104, 2270.

[240] D. Y. Lei, K. Appavoo, F. Ligmajer, Y. Sonnefraud, R. F. Haglund, S. A. Maier, ACS Photonics 2015, 2, 1306.

[241] Y. Yao, R. Shankar, M. A. Kats, Y. Song, J. Kong, M. Loncar, F. Capasso, Nano Lett. 2014, 14, 6526.

[242] N. K. Emani, A. V. Kildishev, V. M. Shalaev, A. Boltasseva, Nano-photonics 2015, 4, 214.

[243] C. E. Petoukhoff, M. B. M. Krishna, D. Voiry, I. Bozkurt, S. Deckoff-Jones, M. Chhowalla, D. M. O’Carroll, K. M. Dani, ACS Nano 2016, 10, 9899.

[244] L. Cong, P. Pitchappa, Y. Wu, L. Ke, C. Lee, N. Singh, H. Yang, R. Singh, Adv. Opt. Mater. 2017, 5, 1600716.

[245] F. Ma, Y. Qian, Y.-S. Lin, H. Liu, X. Zhang, Z. Liu, J. Ming-Lin Tsai, C. Lee, Appl. Phys. Lett. 2013, 102, 161912.

[246] J. J. Sautter, I. Staude, M. Decker, E. Rusak, D. N. Neshev, I. Brener, Y. S. Kivshar, ACS Nano 2015, 9, 4308.

[247] M. Parry, A. Komar, B. Hopkins, S. Campione, S. Liu, A. E. Miroshnichenko, J. Nogan, M. B. Sinclair, I. Brener, D. N. Neshev, Appl. Phys. Lett. 2017, 111, 053102.

[248] F. Walter, G. Li, C. Meier, S. Zhang, T. Zentgraf, Nano Lett. 2017, 17, 3171.

[249] E. Arbabi, A. Arbabi, S. M. Kamali, Y. Horie, A. Faraon, Sci. Rep. 2016, 6, 32803.

[250] T. Cai, G. Wang, S. Tang, H. Xu, J. Duan, H. Guo, F. Guan, S. Sun, Q. He, L. Zhou, Phys. Rev. Appl. 2017, 8, 034033.

[251] X. Wan, X. Shen, Y. Luo, T. J. Cui, Laser Photonics Rev. 2014, 8, 757.

[252] S. M. Kamali, E. Arbabi, A. Arbabi, Y. Horie, M. Faraji-Dana, A. Faraon, Phys. Rev. X 2017, 7, 041056.

[253] Y. Ra’di, D. L. Sounas, A. Alù, Phys. Rev. Lett. 2017, 119, 067404.[254] V. S. Asadchy, M. Albooyeh, S. N. Tcvetkova, A. Díaz-Rubio,

Y. Ra’di, S. A. Tretyakov, Phys. Rev. B 2016, 94, 075142.[255] A. Epstein, G. V. Eleftheriades, J. Opt. Soc. Am. B 2016, 33, A31.[256] A. Epstein, G. V. Eleftheriades, Phys. Rev. Lett. 2016, 117,

256103.[257] A. Díaz-Rubio, V. S. Asadchy, A. Elsakka, S. A. Tretyakov, Sci. Adv.

2017, 3, e1602714.[258] V. S. Asadchy, A. Wickberg, A. Díaz-Rubio, M. Wegener, ACS Photo-

nics 2017, 4, 1264.

high‐efficiency metasurfaces: principles, realizations, and applications · 2019-03-18 ·...

Documents