1. Introduction

Metasurfaces, as a kind of 2D version of metamaterials [1–4], have become one of the hottest research areas in the past decade due to their flexible ability to manipulate the properties of electromagnetic (EM) waves (e.g., amplitude, phase, and polarization), which take the advantages of easy fabrication, minimal material loss, and a low profile compared to 3D bulky metamaterials. Numerous wireless communicated applications of metasurfaces are reported [5–22], such as meta-lens antennas [7–12], vortex beam generators [13–14], polarization converters [16–18], and direction-controlled metasurfaces [19–20].

In order to make metasurfaces more widely used in the wireless communication filed, it is necessary that metasurfaces should not only work at multiple frequencies, but also has diversified functionalities. Therefore, many investigations use the independent polarizations [23–25] or combine both propagation and Pancharatnam-Berry phases to produce two different spin-dependent linear phase gradients [26–29] to achieve additional functionalities and working frequencies. But most of these metasurfaces are limited to half-space (reflection or transmission space), without effectively manipulating the EM wave of full-space. Full-space metasurfaces not only can work in both reflection and transmission space simultaneously, but also can manipulate the reflected and transmitted EM wave independently [30–33]. Some metasurfaces that achieve full-space EM wave manipulation through multilayer dielectric cascaded structures have been reported [30–31]. Full-space MSs have greatly broadened the prospect of wave control, which laid the foundation for further multifunctionality integration [32–33]. However, multilayer dielectric cascading could require more complicated processing procedures (such as more photolithography procedures, hot briquetting and metallization process), which will greatly increase the manufacturing difficulty and costs. Meanwhile, a single-layer metasurface could be much more attractive and preferable in many applications. The reflective structures [34–36] and the transmissive metasurfaces [37–39] using only a single substrate have been investigated, but full-space metasurface with a single substrate are rarely reported. Therefore, it is necessary to conduct research on single-layer metasurface which can independently control EM wave in transmission and reflection modes with high-performance.

In the study, we propose a general strategy to realize high-performance dual-frequency metasurfaces with independent CP wave control in full-space by using a single substrate layer. The meta-atom is composed of two different metallic layers printed on both sides of a single substrate. According to the PB (Pancharatnam-Berry) principle [16], the CP wave can be independently controlled in both transmission and reflection modes by rotating different metal structures on both sides. To prove the concept, two operating frequencies are set as f₁= 8.3 GHz and f₂= 12.8 GHz, the crosstalk among two frequencies is almost completely suppressed. Our metasurface acts as a CP beam splitter at the frequency of f₁ and behaves as a meta-lens at the frequency of f₂ for the incident CP waves as shown in Fig. 1(a). The full-wave simulation and experimental results show that our metasurface can efficiently manipulate the CP waves in both reflection and transmission geometries. Our design finds wide applications in wireless communication systems and radar systems.

 

Fig. 1. Schematic diagram of working mode and meta-atom composition of our metasurface. (a) At f₁= 8.3 GHz, the metasurface operates in the reflection mode and realizes a CP beam splitter. Beam deflection angle can be determined by the gradients on the metasurface. At f₂= 12.8 GHz, the metasurface behaves as a focusing lens in the transmission mode under illumination of the LCP wave. (b) Photograph of the designed MS. (c), (d) Top and bottom views of the designed metasurface. Specifically, p = 10 mm, r₁ = 4.8 mm, r₂ = 3.8 mm, r₃ = 3.4 mm, r₄ = 4.1 mm, r₅ = 2.7 mm, w₁ = 0.2 mm, w₂ = 0.4 mm, w₃ = 0.15 mm, w₄ = 0.9 mm, l₁ = 2.3 mm, l₂ = 4.5 mm, and h = 3 mm.

Download Full Size | PDF

2. Meta-atom design and theoretical analysis

In this work, we aim to achieve full-space EM control based on working frequencies and helicities simultaneously. The designed meta-atom must maintain high reflectivity and high transmittance in its own operation frequency band, and be able to accurately control its reflection and transmission phases, respectively. The crosstalk among the two working states should be suppressed, because crosstalk directly determines the performance of metadevices, and may make it impossible to independently manipulate the two working states.

In order to explain the efficient principle of the designed meta-atom, we use a Jones matrix to describe its characteristics. ${R_{lin}} = \left( {\begin{array}{cc} {{r_{xx}}}&{{r_{xy}}}\\ {{r_{yx}}}&{{r_{yy}}} \end{array}} \right)$ and ${T_{lin}} = \left( {\begin{array}{cc} {{t_{xx}}}&{{t_{xy}}}\\ {{t_{yx}}}&{{t_{yy}}} \end{array}} \right)$ are the reflection and transmission matrix under linear polarization incidence. Here the subscripts x and y are defined as x and y polarized waves, respectively. According to the principle of geometric phase [40–41], nearly 100% efficiency of PB metasurfaces in full-space should satisfy the following conditions:

After combining the above conditions, the results of Eqs. (3) and (4) can be obtained at frequencies of f₁ and f₂, respectively. Where the subscripts l and r are defined as LCP and RCP waves, and θ is the rotation angle of the structure.

At this time, not only the reflection circular polarization amplitude $|{{r_{ll,{f_1}}}} |= |{{r_{rr,{f_1}}}} |= 1$ and the transmission circular polarization amplitude $|{{t_{rl,{f_2}}}} |= |{{t_{lr,{f_2}}}} |= 1$ at f₁ and f₂ should be satisfied, but also the phase of the meta-atom can be manipulated from 0° to 360°. Therefore, we can efficiently manipulate the phase of metasurface at two frequencies by rotating the structure according to the PB theory [16].

Considering all the above, we designed our full-space meta-atom with high efficiency and great isolation, as shown in Fig. 1(b). The meta-atom is composed of two different metallic layers (I and II) printed on both sides of FR4 substrate with a dielectric constant of ε_r = 4.3 and a loss tangent of tanδ = 0.003. The metallic I layer consists of a circle ring, two split-ring resonators (SRRs) and an electric field coupled resonator (ELCR). The metallic II layer consists of a circle ring and an ELCR that is identical with the I layer. Compared with the bandpass filter composed of a single circular frequency selective surface (FSS), the bandpass filter formed by the circle rings of the I layer and the II layer has a better rectangle factor, higher transmittance and reflectivity. This filter lays a prerequisite for the efficient control of reflected and transmitted CP waves.

For the reflection mode, the two SRRs in the I layer are used to manipulate the reflected CP waves, and they work in magnetic resonance mode. The inner SRR introduces a reflection resonance at low frequencies. The direction of the outer SRR is orthogonal to the inner SRR, which improves the reflectivity at the resonance. And another reflection resonance is introduced, thereby increasing the bandwidth of the reflection mode. To validate the performance of the proposed meta-atom, simulations is carried out by using finite-difference time-domain (FDTD) method. First, the metasurface is illuminated by a LP wave propagating along -z direction, and Fig. 2(a) show the reflection spectra of meta-atom. It can be clearly seen that two resonances are introduced at 8 GHz and 8.6 GHz respectively in accordance with the expected design. And the reflection amplitudes of the structure along the two linear polarizations are larger than 0.96 while their reflection phases exhibit a π difference at the target frequency of f₁ = 8.3 GHz, which satisfies well the PB theoretical value required by Eq. (1). Then shining LCP wave normally onto the metasurface, Fig. 2(c) illustrates that the amplitude of reflection |r_ll| > 0.95 in the 8.2 - 8.6GHz frequency band with other EM modes (|r_rl|, |t_ll|, and |t_rl|) totally suppressed. If the two split resonant rings are not placed orthogonally, it will have a great influence on the transmission performance and greatly increase the crosstalk of the two operating modes. For the transmission mode, the ELCR of both I layer and II layer are designed to manipulate the transmitted CP wave, and it works in electric resonance mode. This structure needs to be carefully designed. Its resonance mode should different from the reflective structure and has some central symmetry, so as to obtain better isolation. Figure 2(b) shows the transmission spectra of meta-atom for LP incidence, the transmission amplitudes of |t_xx| and |t_yy| are better than 0.9, and the transmission phases satisfy φ_xx – φ_yy ≈ 180° at frequency f₂ = 12.8 GHz. Those EM responses are very close to the theoretical value required by Eq. (2). Therefore, as shows in Fig. 2(d) when the LCP wave is normally incident, a transparent window (|t_rl| > 0.8) appears in the 11.65–11.95GHz frequency band with other EM modes (|r_rl|, |r_ll|, and |t_ll|) suppressed.

 

Fig. 2. EM responses and current distribution of the unit cell when the incident wave is along the -z direction. (a), (b) EM responses of reflection and transmission modes under linear polarization. (c), (d) Amplitude of reflection and transmission modes under circular polarization. (e) The current distribution of meta-atom under LCP incident wave cut at xoy plane at f₁ = 8.3 GHz. (f) The current distribution of meta-atom under LCP incident wave cut at xoy plane at f₂ = 12.8 GHz.

Download Full Size | PDF

Then, in order to intuitively observe the great isolation of the two operating modes, Figs. 2(e) and 2(f) illustrate the current distribution at the two operating frequencies. In the reflection mode, the current distribution (Fig. 2(e)) (cut at the xoy plane) shown that main working sections were the SRRs, and the ELCR of I layer and II layer generate a little induced current. In the transmission mode, Fig. 2(f) indicates the current distribution of meta-atom (cut at the xoy plane) at f₂. Obviously, the ELCR of I layer and II layer is working, whereas the SRRs structures of I layer almost have no reaction. These phenomena clearly illustrate the high isolation of the designed unit in the different operation modes.

At last, we shine a CP wave normally on the MS and discuss the EM responses of the meta-atom when different structures rotate at different angles θ₁/θ₂. When the ELCR structures are rotated, the MS works in the transmission mode, the transmission amplitude |t_rl| is better than 0.85 at 12.8 GHz and the phase φ^t_rl also satisfies the linear relationship of 2θ₁ (show in Fig. 3(d)). Simultaneously, the meta-atom exhibits stable high reflection amplitudes |r_ll| (Fig. 3(a)) and the reflection phase φ^r_ll hardly changes (Fig. 3(b)). This further illustrates that when the transmission performance is manipulated, the reflection performance is not affected. For the reflection mode, when rotating the SRR structures, it basically has no effect on the transmission amplitude (Fig. 3(e)) and phase (Fig. 3(g)) of the meta-atom. The meta-atom exhibits very high reflection amplitudes (|r_ll| > 0.95), and their phase vary as 2θ₂ (Fig. 3(h)). Therefore, the designed full-space meta-atom can not only maintain high reflectance and transmittance in reflection and transmission modes, but also have good isolation between the two working modes, the phase can be independently manipulated of each working mode according to geometric phase theory. By using the designed meta-atom, we can easily and independently control the CP waves in the full-space with only a single substrate layer and may lead to some potential applications in communication systems.

 

Fig. 3. Meta-atom design properties against rotation angle and polarizations when the incident wave is along the -z direction. (a-d) EM response of the meta-atom when rotating ELCR structures under LCP. (e-h) EM response of meta-atom when rotating SRR structures under LCP. (a) Reflection amplitude |r_ll| and transmission amplitude |t_rl| variation with respect to rotation angle θ₁. (b) Reflection phase φ^r_ll variation with respect to rotation angle θ₁. (c) Transmission phase φ^t_rl variation with respect to rotation angle θ₁. (d) Transmission amplitude |t_rl| and phase φ^t_rl as a change of θ₁ at 8.3GHz. (e) Reflection amplitude |r_ll| and transmission amplitude |t_rl| variation with respect to rotation angle θ₂. (f) Reflection phase φ^r_ll variation with respect to rotation angle θ₂. (g) Transmission phase φ^t_rl variation with respect to rotation angle θ₂. (h) Reflection amplitude |r_ll| and phase φ^r_ll as a change of θ₁ at 12.8 GHz.

Download Full Size | PDF

3. Multifunctional full-space metasurface design and experiment

Specifically, we design and fabricated a multifunctional full-space meta-device, which consists of 19 × 19 meta-atoms and occupies a total area of 190 × 190 mm², as shown in Fig. 4(a). For the reflection mode, the meta-device can act as a CP beam splitter under a LP incident wave. Figure 4(c) illustrates the FDTD simulated reflection phase and reflection amplitude of each meta-atom within three supercells. Note that these meta-atoms exhibit high reflection amplitudes (|r_ll| > 0.95), meanwhile the reflection phase φ^r_ll agree well with the geometric phase theory. For the CP wave incidence, the phase gradient is set as ξ = dφ / dx = 0.602k₀, therefore the rotation angle of the SRR structures is calculated as Δθ₂= 30°. For the transmission mode, the meta-device acts as a focusing lens for LCP wave and the focal length is set as F = 60mm. The phase distribution should satisfy the following equation: ${\varphi ^t}_{rl} = {k_0}(\sqrt {{F^2} + {x^2} + {y^2}} - F)$, k₀= w/c is the propagation constant. Figure 4(d) shows that the phase value calculated by FDTD which agrees well with the theoretical phase value of the meta-lens.

 

Fig. 4. Design of multifunctional full-space metasurface. (a) Schematic diagram of fabricated sample, inset shows the structure of the supercell. (b) Setup of the measurement environment. (c) The theoretical phase and simulation values of designed CP beam splitter vary with the position under illumination of LCP wave. (d) Designed phase distribution of φ^t_rl (x,y) for meta-lens.

Download Full Size | PDF

Then we experimentally measured the transmission and reflection properties of the fabricated MS in a microwave anechoic chamber. For the transmission mode, illustrating an LCP wave via a standard CP horn antenna normally onto the sample, and scan the E field distribution through a monopole probe (≈ 20 mm long). Both the CP horn antennas and the probe are connected to a vector-field network analyzer (Agilent E8362C PNA) to collect the local electric field, as schematically shown in Fig. 4(b). The LCP wave is irradiated on the designed lens, and the focal length of the meta-lens is defined as the point with maximum intensity in |E|² ∼ z curve (Fig. 5(d)), which is obtained according to the field distribution. Both the FDTD simulated focal length (59mm) and the measured focal length (57mm) agrees very well with the theoretical value (60 mm). This slight difference can be explained by the finite-size effect of our sample, and the incident wave is not a plane wave. And the electric field distribution also shows the focusing effect very well. Figures 5(a) and 5(b) depict the measured Re(E) distributions at xoz plane and the |E|² distributions on both yoz and xoz plane at f₂ = 12.8 GHz. From these two figures, we can see that the incident wave energy has been indeed concentrated at the focal point. The focal spot size is defined by the full width at half maximum (FWHM) of the focal spot on the focal plane, and find it is about 20 mm. In order to quantitatively evaluate the focusing effect of the transmitted wave, η_f = P_f / P_t is defined as the relative efficiency of the meta-lens, indicating the ratio of the focused energy to the total transmitted energy, where P_f represents the transmitted focused energy and P_t represents the total transmitted energy. Figure 5(c) shows the field distribution on the measured focal plane, the red dotted line area is the focused energy. As the result, the measured relative efficiency of the meta-lens lies in ≈ 82.7% at a frequency of 12.8 GHz. The efficiency of the meta-lens can be improved by fabricating a larger meta-lens or using a high-gain horn antenna to improve the plane wave characteristics.

 

Fig. 5. Simulated and measured results of the fabricated full-space metasurface. (a-d) Performance of designed transmissive meta-lens under illumination of LCP wave at 12.8 GHz. (a) Re(E) distribution at xoz plane. (b) |E|² distribution on both yoz and xoz plane. (c) The measured |E|² distribution at focal plane. (d) The simulated and measured |E|² vary with the axis z. (e-h) Performance of designed reflective CP beam splitter. (e), (f) Normalized scattering intensity versus frequency and angle of detection. (g), (h) Efficiencies of the CP beam splitter. Insets show the simulated and measured scattering patterns of the beam deflecting results at 8.3 GHz.

Download Full Size | PDF

For the reflection mode, we experimentally characterized the CP beam splitter performance under excitation of LP wave. Figures 5(g) and 5(h) depict the measure scattered-field intensity under illumination of LP wave, it is clear that most of the reflected LCP/RCP wave energy is distributed in the 8 - 8.6 GHz frequency band. The measured LCP wave was deflected to -37°, while the RCP wave was deflected to +37° at 8.3 GHz. The insets to Figs. 5(g) and 5(h) show that the simulated deflection angles are θ_r = ± 36.5°, the left/right-handed reflected angles were symmetrical along a normal line. The simulated and measured deflection angles are in good agreements with the theoretical value (± 36.4°) based on generalized Snell’s laws of reflection [3]. And it is clearly observed that the reflected energy is mainly distributed on the anomalous beams and the energy in the other direction is very little. The scattered-field intensity is normalized by a metal plate with the same sample size. Then we calculate the efficiency of designed CP beam splitter (shown in Fig. 5(g) and 5(h)) by defining the efficiency as the ratio of the power of the anomalous reflection to the power of incident wave (obtained through the reflection of the same size metal plate). The maximum efficiency reach values of 84.9% for LCP wave and 84.1% for RCP wave. The missing power is partly lost to the transmission and partly to absorption. The efficiency of the CP wave splitter can be improved by increasing the reflectivity and improving the accuracy of the phase.

4. Conclusions

We proposed a simple and low-cost method to manipulate the full-space CP wave properties. Both the reflection and transmission wave can be controlled independently and efficiently at two frequencies. A multifunctional full-space MS is designed by using the single-layer meta-atom, and experimentally demonstrated its performance. For the reflection mode, CP beam splitter are realized, and the maximum efficiency reach values of 84.9% for LCP and 84.1% for RCP at f₁ = 8.3 GHz. For the transmission mode, designed metasurface acts as a meta-lens for the incident LCP waves, and the measured relative efficiency of the meta-lens lies in ≈ 82.7% at a frequency of f₂ = 12.8 GHz. Compare to other reported full-space metadevices, our full-space metadevices show great advantage of low profile, easy fabrication and high performance, which yield many potential applications in multifunctional communication systems.