Abstract
1 Introduction
In the past decade, optical metasurfaces have attracted considerable attention due to their superior capabilities of arbitrarily molding light, thereby challenging conventional bulky optics with ultrathin planar metadevices exhibiting compact footprints and advanced functionalities.1
Capitalizing on the concepts of conventional MIM metasurfaces and Babinet-inspired slot-antenna-based metasurfaces,10
2 Results and Discussion
2.1 Design of the Meta-atom
The designed MCSA unit cell consists of a rectangular-shaped slot antenna milled in a subwavelength-thick Au film, a middle spacer layer, and an optically thick Au mirror, forming a modified MIM meta-atom [Fig. 1(a)]. The periodicity of MCSA unit cells is , which is much smaller than the design wavelength of to eliminate any free propagating diffraction orders. The thickness of the topmost slot antenna is , and the thickness of the bottom Au layer is , which is thick enough to block the transmission. To study the optical response of the MCSA unit cell, we performed full-wave simulations with the commercially available finite-element software COMSOL Multiphysics (version 5.6). In simulations, the refractive index of the spacer layer is assumed to be 1.45, whereas the relative permittivity of Au is described by the Drude model fitted with the Johnson and Christy data,30 with the plasma frequency being and the damping frequency being . The medium above and within slot antennas is chosen to be air with the refractive index of 1. The simulation domain of the unit cell is truncated at the air side with a perfectly matched layer to minimize the reflection. With this model, we calculated the complex reflection coefficients on the top of the slot antenna as a function of and for normally incident transverse magnetic (TM) polarized excitation (the electric field is polarized along the axis) at .
Sign up for Advanced Photonics Nexus TOC. Get the latest issue of Advanced Photonics Nexus delivered right to you!Sign up now
Figure 1.(a) Schematic of the basic MCSA unit cell consisting of a slot antenna on top of an
It is seen [Fig. 1(b)] that the variation of both and results in the phase variation for the reflected TM-polarized light in a wide range (up to 320 deg) at a relatively large spacer thickness . This variation occurs because of tuning the meta-atom in the vicinity of the mode resonance formed by hybridization between the lateral plasmonic resonance confined in the slot antenna and vertical Fabry–Pérot (FP) resonance between the slot antenna and the bottom Au mirror. In contrast to the well-studied freestanding slot antenna that exhibits limited phase coverage of only 180 deg for the copolarized electric fields, the bottom Au film functions as a reflecting mirror of the FP cavity (Fig. S1 in the Supplementary Material). Therefore, the incident light will bounce back and forth before being reflected with a sufficiently large phase shift. Close to the plasmonic resonance that is indicated by the pronounced absorption of the TM-polarized incident light, the phase varies strongly, with a range of [Fig. 1(b)]. Away from the resonance region, the reflection phase is gradually varied with the lateral dimensions. For example, the reflection contribution of the top slot-antenna layer will be dominating when the slot antennas are narrow in the polarization direction, leading to the incident wave being mostly reflected by the top slot-antenna layer that approximates a continuous Au film [Figs. 1(b) and 1(c) and Figs. 2(a) and 2(c)]. When the slot antennas are wide, the top Au layer becomes more transparent, and the light is mainly reflected by the bottom Au layer. Therefore, the reflection phase changes with the thickness of the middle spacer [Fig. 1(b) and Figs. 2(a) and 2(b)]. According to the theory of Babinet-inspired plasmonic metasurfaces,10
Figure 2.(a), (b) Numerically calculated complex reflection coefficients as a function of the slot antenna’s lateral dimensions at the design wavelength of
Owing to the hybrid plasmonic/FP resonance, the phase modulation for the reflected polarized light thereby depends heavily on the spacer thickness , which influences the accessible phase coverage together with the absorption around the resonance of the slot antenna. Figure 2 depicts the phase dependence on the slot dimensions for two different spacer thicknesses. For a thickness of , the phase space of around 320 deg is still accessible [Fig. 2(a)]. In contrast to the case with a thicker spacer of 130 nm, we observe a steeper phase variation in the region of resonance. Specifically, the phase contours become crammed close to each other for a smaller spacer thickness, different from Fig. 1(b). Meanwhile, the absorption becomes stronger with a more pronounced and narrower resonance. Therefore, this thickness is not ideal for a practicable accurate selection of slot-antenna candidates, even though similar phase contours can be found in other regions away from the resonance. The field distribution of the -component of the electrical field is still strongly localized within the hole volume, as shown in Fig. 2(c). In comparison to the field localization at a 130-nm spacer thickness in Fig. 1(c), a decrease in field strength is observable. Further decreasing the spacer thickness to 10 nm causes a change in the phase distribution and reflectivity as well as in the field localization [Figs. 2(b) and 2(d)]. The resulting phase space is confined to a limited interval of only [Fig. 2(b)]. Additionally, the resonant structure does not show a strong localization, where the component of the electrical field is only confined at the edges of the slot antenna for this 10-nm- configuration [Fig. 2(d)]. As a result, we choose a 130-nm spacer thickness for the following numerical and experimental discussions, since it provides a sufficiently large phase coverage with more spreading phase contours and higher reflection amplitudes.
2.2 Functional Phase-Gradient MCSA Metasurfaces for Beam Steering
To demonstrate the applicability of this kind of MCSA meta-atom, we design a phase-gradient metasurface that functions as a beam steerer. Under the excitation of a TM-polarized wave at normal incidence, the reflected light will be redirected to a specific direction, which is well described by the generalized Snell’s law8 and may be also viewed as a blazed grating where the angle of reflection reassembles the first-order diffraction.34 By marking the reflection phase through constant phase-contour lines in the complex reflection coefficient map, we select three meta-atoms with a nearly constant phase shift between neighbors and considerably high reflectivities according to realistic fabrication sizes. (The selected dimensions are shown in Table S1 and Fig. S2 in the Supplementary Material) Specifically, the chosen MCSA meta-atoms have a constant phase interval of between each other. The selected meta-atoms are then placed at a center-to-center distance of to form a supercell. However, the supercell design is nontrivial, since two important factors should be considered. First, the supercell period should be larger than the incident wavelength to avoid coupling to surface plasmons.35
Figure 3.Theoretical performance of the beam-steering MCSA metasurfaces that reflect the normal incident TM-polarized light into the +1 diffraction order for supercells consisting of (a) pair, (b) triple, and (c) quadruple identical MCSA meta-atoms. The left columns display the simulated diffraction efficiencies of different orders as a function of wavelength. The right columns show the electrical field distributions at the
As shown in Figs. 3(b) and 3(c), the supercells consisting of 9 and 12 elements perform well for steering the incoming light to the first diffraction order (green solid line), each reaching a diffraction efficiency of at . Still, the performance of the 12-element supercell can be considered more advantageous, since the reflectivity into the unwanted diffraction orders, mainly the diffraction order (violet solid line), is more suppressed compared to the 9-element design. For the six-element design [Fig. 3(a)], the specular reflection is dominating while the desired diffraction order is relatively low, which can be explained by plotting the electric field distributions within the top plane cut at the center of the slot antenna (right column in Fig. 3). It is clearly observable that the fields within the identical meta-atoms become more similar when the tuple size of the supercell is increased. As a final comment, it should be mentioned that the influence of near-field coupling between neighboring meta-atoms with significantly varied dimensions could also be decreased by inserting more meta-atoms with intermediate reflection phases (instead of the meta-atom multiplication considered above) to provide a smoother phase gradient and better performance (Fig. S3 in the Supplementary Material). However, this design is rather challenging to faithfully reproduce during the fabrication, since the neighbor slot dimensions are too close.
2.3 Experimental Demonstration of Functional MCSA Metasurfaces
Following the design, we fabricated the beam-steering MCSA metasurface composed of 12-element supercells on a silicon substrate using thin-film deposition, electron beam lithography, lift-off, and ion-beam etching techniques. First, a 2-nm chromium (Cr) adhesion layer and a 130-nm Au layer are evaporated on a silicon substrate using electron beam evaporation. Subsequently, a 140-nm-thick spin-on dielectric is applied (IC1-200, Futurrex, nominal refractive index of 1.34, hard-baked at 200 deg), which serves as the dielectric spacer layer with an equal optical thickness to the original design. Next, a 2-nm Cr adhesion and a 50-nm Au layer are deposited using electron beam evaporation. We use electron beam lithography (Raith Voyager, 50 kV acceleration voltage, beam current 2.14 nA) to define the structures in a positive tone resist (AR-P 6200.13, Allresist, area dose ). After development, the resist mask is transferred to the underlying Au layer by Ar ion-beam etching (Technics Plasma). The remaining resist layer is removed in NEP at 80 °C (N-Ethyl-2-pyrrolidone). The fabricated metasurface has a size of , whose scanning electron microscopy (SEM) images are displayed in Fig. 4(b). Generally, the fabricated slot antennas resemble very well the desired shapes and dimensions, despite the inevitable surface roughness and rounded corners. The measurements were performed using a custom-built optical setup (Fig. S4 in the Supplementary Material) within the wavelength interval of 700 to 850 nm in steps of 10 nm. The inset optical image of the diffraction spots in Fig. 4(c) illustrates the beam-steering ability of the fabricated metasurface at the wavelength of , nearest to the design wavelength. To determine the overall performance quantitatively, we measured the corresponding diffraction efficiencies of different orders from to +2. In the measurement, four fabricated metasurfaces with identical recipes were measured independently to calculate the mean value and corresponding standard errors of the mean. From Fig. 4(c), it is evident that the total reflectivity is close to 40%, and the diffraction efficiency of the order is () at . Further, results show that beam-steering capability is pursued in a broadband spectrum ranging from 700 to 800 nm with zeroth order below 10% and higher diffraction orders suppressed below 5%. Compared to the averaged reflectance of the three selected MCSA meta-atoms () at , the measured diffraction efficiency is reduced by half, which is mainly ascribed to the increased dissipation loss from the 2-nm-thick Cr adhesion layer as well as the surface scattering and grain boundary effects of the deposited Au layers. To validate our assumptions, we increased the damping rate of Au by a factor of 3 in the simulation and obtained reasonably matched simulation results with the measurements.
Figure 4.(a) Schematic illustration of the MCSA reflective metasurface for beam steering. (b) SEM images of the fabricated beam-steering metasurface. The lower image depicts the fabricated supercells. (c) Simulated (solid lines) and experimental (dots with error bars) diffraction efficiencies of different orders as a function of wavelength for TM incident light. The error bars denote the standard deviation of the measured data of four metasurface samples. The inset image shows the diffracted light spots at
For further validation of the MCSA concept, an additional metasurface was designed and fabricated that enables splitting the incoming TM wave into ±1 diffraction orders with equal intensity, which is made up of a quadruple arrangement of the two slot antennas with the phase difference of (Fig. 5). This leads to a supercell consisting of eight elements with a diffraction angle of at the design wavelength of .8 The SEM images of part of the fabricated sample are shown in Fig. 5(b). Similarly, the fabricated beam-splitting metasurface samples were characterized by measuring the diffraction efficiencies into different orders in Fig. 5(c). We again observe a reasonable agreement between the simulated and measured results in a wide wavelength range. At the measured wavelength of , () of the incident wave is split into the diffraction orders, whereas the total reflectivity almost reaches . Due to increased ohmic losses in metallic layers (as compared to the tabulated values) especially at shorter wavelengths, the total reflectivity in the simulation and experiment exhibits different dependences on the optical wavelength. As a final comment, we emphasize that the efficiencies of MCSA metasurfaces could be further improved using single-crystalline metallic materials in the fabrication process to reduce the ohmic loss.39,40
Figure 5.(a) Schematic illustration of the MCSA reflective metasurface for beam splitting. (b) SEM images of the fabricated beam-splitting metasurface. The lower image shows the designed and fabricated supercells. (c) Simulated (solid lines) and experimental (dots with error bars) diffraction efficiencies of different orders as a function of wavelength for TM incident light. The error bars denote the standard deviation of the measured data of four metasurface samples. The inset image shows the diffracted light spots at
3 Conclusion
We have demonstrated a new type of reflection optical metasurface based on MCSAs, which combine the advantages of high efficiencies and ease of fabrication from conventional MIM metasurfaces as well as the electrical connection from Babinet-inspired slot-antenna-based metasurfaces. By tuning the lateral dimensions of the slot antennas and optimizing the spacer thickness, a phase coverage of 320 deg can be achieved for the copolarized reflected light. With this platform, we have implemented functional phase-gradient metasurfaces that perform beam steering and beam splitting in the near-infrared wavelength range. Reasonably good working performance, in terms of efficiencies and operation bandwidths, was achieved by alleviating the near-field coupling between adjacent differently sized meta-atoms. We believe that this newly established reflective metasurface configuration with electrically connected individual elements offers an alternative and promising platform for electrically controlled reflective metasurfaces.
Sven Ebel received his BS degree from Kiel University in 2020. He is currently a master student at the Institute for Experimental and Applied Physics, Kiel University, Germany. He holds one issued patent. His research interests include metasurfaces and electron–light interactions in ultrafast electron microscopy.
Yadong Deng received his bachelor’s degree in electronic science and technology from Xuzhou University of Technology, China, in 2017 and his master’s degree in physics from Zhejiang Normal University, China, in 2020. After 1-year research assistant at the Center for Nano Optics, University of Southern Denmark (SDU Nano Optics), Denmark, he is currently one full-time PhD student at SDU Nano Optics since 2021. His fields of interest include nanophotonics, plasmonics, and metasurfaces.
Mario Hentschel studied physics in Bonn, finishing his diploma thesis with Manfred Fiebig in 2009. He moved to the group of Harald Giessen (of Stuttgart and MPI for Solid University State Research), completing his PhD thesis in 2013. Being awarded a Feodor Lynen postdoctorial scholarship (Alexander von Humboldt Foundation), he joined the group of A. Paul Alivisatos at UC Berkeley. Since 2015, he has been a head of nanofabrication at the Fourth Physics Institute at University of Stuttgart.
Chao Meng received his BS and PhD degrees in optical engineering from Zhejiang University, Hangzhou, China, in 2008 and 2013, respectively. During 2014 to 2017, he was a team leader of Shanghai Silight Technology for commercializing silicon photonics. He is currently a postdoctoral fellow at the Centre for Nano Optics, University of Southern Denmark, Odense, Denmark. His current research interests include nanophotonics, plasmonics, and metasurfaces. He was a recipient of the Marie S.-Curie MULTIPLY fellowship in 2019.
Sören im Sande received his bachelor’s degree from TU Lübeck and his master’s degree from the University of Southern Denmark. He is currently a PhD research fellow at the Centre for Nano Optics of the University of Southern Denmark. His work is focused on plasmonic metasurfaces, integrated photon sources, as well as quantum metasurfaces.
Harald Giessen is a full professor and holds the chair for ultrafast nano-optics in the Department of Physics at the University of Stuttgart. He is also co-chair of Stuttgart Center of Photonics Engineering. He is a fellow of the Optical Society of America. His research interests include ultrafast nano-optics, plasmonics, metamaterials, 3D-printed micro- and nano-optics, medical micro-optics, miniature endoscopy, innovative MID-IR ultrafast laser sources, applications in microscopy, biology, and sensing.
Fei Ding received his PhD in optical engineering from Zhejiang University in 2015. He is currently an assistant professor at the University of Southern Denmark. His current research interests include metasurfaces, plasmonics, and quantum nanophotonics. He has published more than 65 articles in peer-reviewed journals. He won the Wang Daheng Optical prize, the Young Scientist Award from the Progress in Electromagnetics Research Symposium, Villum Young Investigator, and the DOPS award.
Sergey I. Bozhevolnyi is a professor and head of the Centre for Nano Optics at the University of Southern Denmark and Chair of Technical Science at the Danish Institute of Advanced Study. His current research interests include linear and nonlinear nano-optics and plasmonics, including plasmonic interconnects, quantum plasmonics and metasurfaces. He is a fellow of Optical Society of America (2007), an elected member of Danish Academies of Natural Sciences (2010) and Technical Sciences (2019).
References
[2] N. Yu, F. Capasso. Flat optics with designer metasurfaces. Nat. Mater., 13, 139-150(2014).
[7] A. M. Shaltout, V. M. Shalaev, M. L. Brongersma. Spatiotemporal light control with active metasurfaces. Science, 364, eaat3100(2019).
[20] D. Lin et al. Dielectric gradient metasurface optical elements. Science, 345, 298-302(2014).
[23] A. Kuznetsov et al. Optically resonant dielectric nanostructures. Science, 354, aag2472(2016).
Set citation alerts for the article
Please enter your email address