Mid-infrared reconfigurable all-dielectric metasurface based on Ge2Sb2Se4Te1 phase-change material

Jinglin He; Zhuolin Shi; Sheng Ye; Minhua Li; Minhua Li; Jianfeng Dong; Jianfeng Dong

doi:10.1364/OE.471193

1. Introduction

The mid-infrared (MIR) device plays an important role in optical applications, such as metalens [1], holography [2], sensing [3] and infrared (IR) imaging [4], etc. However, due to the complexity and high cost of optics and optoelectronic devices in the MIR wavelength region, the development of MIR optical systems lags far behind visible and near-infrared (NIR) systems. Further exploration of the MIR high efficiency optical device is imminent.

The optical metasurfaces are composed of artificial subwavelength structures, which can achieve predictable manipulation of various electromagnetic responses by periodic arrangement of unit cell of the metasurface. The metasurfaces have high degree of freedom in manipulating the amplitude, phase and polarization of the incident light, showing advantages in realizing miniaturization and integration of the optical devices. Up to now, a variety of chiral metasurfaces have emerged to achieve circular dichroism (CD) [5–7] and circular polarization conversion [8]. For instance, Wang et al. [5] reported a chiral metasurface composed of achiral nanoholes, which has large CD. Li et al. [6] designed a chiral plasmonic metasurface with strong CD in the NIR band. Ouyang et al. [7] also designed and demonstrated chiral plasmonic metasurface absorbers to achieve large broadband infrared CD. Kaschke et al. [8] presented a circular polarization converter based on helical structure. However, these metasurfaces were mostly designed for only one function. From the practical application point of view, compact devices with multiple functions have better development prospects. Huang et al. [9] proposed a dual-functional metasurface which could generate linear dichroism (LD) and CD in the IR region. Cai et al. [10] also integrated quarter-wave plate and half-wave plate functions in a metasurface. Recently, multifunctional and reconfigurable metasurfaces which combine with phase-change material vanadium dioxide (VO₂) [11] and Ge₂Sb₂Te₅ (GST) [12] have attracted intense attention. Tripathi et al. [13] demonstrated a dielectric metasurface that could switch among transmission, reflection and absorption based on VO₂. Liu et al. [14] realized an active switch based on VO₂ that could switch from a chiral mirror to either a conventional mirror, a handedness-preserving mirror, or a chiral mirror of opposite handedness. Tian et al. [15] designed a multispectral optical switch and demonstrated that the GST material made the nanostructure have great active tunability. In addition, with the introduction of Pancharatnam-Berry (PB) phase [16], many wavefront manipulation studies [17] based on PB phase have emerged, such as abnormal refraction [18,19], orbital angular momentum (OAM) vortex beam [20,21], etc. For instance, the OAM vortex beam can improve communication capacity without increasing the bandwidth [22]. Considering the importance of improving channel quality and increasing transmission modes, it is of great significance to study abnormal refraction, OAM vortex beam and beam splitting, particularly in the MIR wavelength region. Combining with phase-change material GST, Zhang et al. [23] and Ding et al. [24] deeply studied the active tunable vortex generator based on PB phase. However, the exploration of tunable multifunctional/reconfigurable metasurfaces is still in its infancy.

In this work, we propose a MIR reconfigurable all-dielectric metasurface based on a novel phase-change material Ge₂Sb₂Se₄Te₁ (GSST). This novel phase-change material has low loss and large optical contrast between amorphous and crystalline states in the MIR band. The unit cell of the metasurface is composed of two pairs of elliptic nanorods that can break the in-plane inversion symmetry. When the circularly polarized waves are incident on the metasurface, inter-column interference phenomenon will appear, resulting in the different transmission for different polarization waves (such as dichroism). The introduction of phase-change material GSST is the key to switch from CD to circular polarization conversion in this work. Through optimization, when GSST nanorods in the metasurface are in amorphous state, this metasurface can transmit right-handed circular polarization (RCP) wave and reflect left-handed circular polarization (LCP) wave perfectly in the vicinity of 74 THz, and its CD value reaches 0.95. Under thermal excitation, GSST nanorods are transformed to crystalline state and the metasurface becomes a MIR circular polarization converter, achieving circular polarization conversion in the wideband range of 41 THz - 48 THz with efficiency of more than 90%. Moreover, because the metasurface has high transmission in the working frequency bands, we also specifically verify the multi-functionality of the proposed metasurface. Based on PB phase, three types of functional metasurfaces are simulated and characterized respectively: abnormal refraction, OAM vortex beam, and OAM vortex beam splitting.

2. Design and methods

GSST is often used to design reconfigurable metasurfaces due to its complex refraction index difference between amorphous and crystalline states [25,26]. The refraction index of GSST is shown in Appendix A (Fig. 10). Our study aims to design a MIR reconfigurable all-dielectric metasurface. Specifically, this metasurface can be switched from high CD to wideband circular polarization conversion when the GSST is changed from amorphous state to crystalline state. Our proposed metasurface consists of two pairs of GSST elliptic nanorods (m1 and m2) and a layer of magnesium fluoride (MgF₂) substrate. MgF₂ is a good substrate due to its low nondispersive refractive index (n = 1.35) and negligible extinction in the visible and infrared spectral ranges. The functional schematic of the metasurface, the three-dimensional and top views of the unit cell of metasurface are shown in Fig. 1 (P = 3.84 µm). Each elliptic nanorod is located at the center of its meta-atom of period P/2 (Fig. 1(a)). Here, m1 and m2 represent different nanorods with the same height (h = 3.12 µm) and different sizes. In addition, the elliptic nanorods m1 are tilted θ₁ with respect to the + x-axis, L₁= 0.6 µm, W₁= 0.24 µm, and the nanorods m2 are tilted θ₂ with respect to the + x-axis, L₂= 0.72 µm, and W₂= 0.48 µm.

Fig. 1. The schematic diagram of proposed metasurface. (a) Top and (b) 3D views of the unit cell. (c) The metasurface exhibits high CD when the GSST layer is in amorphous state and (d) exhibits circular polarization conversion when the GSST layer is in crystalline state.

Download Full Size | PDF

The proposed structure may appear in the specific device preparation processes. The substrate is firstly deposited with GSST films of corresponding thickness by magnetron sputtering (the substrate is held near room temperature throughout the film deposition process), and then electron beam lithography and dry etching are carried out to obtain the required unit patterns. After optical characterization of the GSST layer in the amorphous (as-deposited) state, the phase transition of GSST from amorphous to crystalline state can be achieved by applying light pulses or thermal annealing. Then the sample is transitioned to the crystalline state by hot-plate annealing at 250 °C for 30 min. Similar manufacturing technologies are used in refs. [1,27].

In our simulation, the plane wave excitation is set along the + z-axis, incident on the bottom surface of the MgF₂ substrate. To perform the periodic structures, we set the unit cell boundary conditions along the ± x and ± y directions and apply two floquet ports along the ± z directions. The related optical properties are simulated based on finite integration technique (FIT) method [28].

The transmission for the LCP and RCP waves incidence, and circular dichroism (CD) in transmission are defined as:

(1)$$\left\{ \begin{array}{l} {T_{RR}} = |{t_{RR}}{|^2},{T_{LR}} = |{t_{LR}}{|^2},{T_{LL}} = |{t_{LL}}{|^2},{T_{RL}} = |{t_{RL}}{|^2}\\ CD = |{T_{RCP}} - {T_{LCP}}|= |{(}{T_{RR}} + {T_{LR}}{)} - {(}{T_{LL}} + {T_{RL}}{)}|\end{array} \right.$$

where the t_RR (t_LL) and t_LR (t_RL) are the co-polarized and cross-polarized transmission coefficients, respectively. Here the first subscript and second subscript represent transmitted wave and incident wave respectively. The relationship of the transmission coefficients between circular polarization and linear polarization waves can be expressed as [29]:

(2)$${T_{cir}} = \left( {\begin{array}{{ll}} {{t_{RR}}}&{{t_{RL}}}\\ {{t_{LR}}}&{{t_{LL}}} \end{array}} \right) = \frac{1}{2}\left( {\begin{array}{*{20}{l}} {\left({{t_{xx}} + {t_{yy}}} \right)+ i\left({{t_{xy}} - {t_{yx}}} \right)}&{\left({{t_{xx}} - {t_{yy}}} \right)- i\left({{t_{xy}} + {t_{yx}}} \right)}\\ {\left({{t_{xx}} - {t_{yy}}} \right)+ i\left({{t_{xy}} + {t_{yx}}} \right)}&{\left({{t_{xx}} + {t_{yy}}} \right)- i\left({{t_{xy}} - {t_{yx}}} \right)} \end{array}} \right)$$

where subscripts x and y denote the x-polarized and y-polarized components for linear polarization waves, respectively. The T_cir represents the circularly polarized transmission matrix. If the m1 or m2 is rotated by an angle θ and the amplitudes of transmission close to 1, the linearly polarized transmission Jones matrices of m1 and m2 can be described by:

(3)$$\begin{array}{l} T_{lin}^1 = \left( {\begin{array}{{cc}} {\textrm{ }\cos ( - {\theta_1})}&{\sin ( - {\theta_1})}\\ { - \sin ( - {\theta_1})}&{\textrm{ }\cos ( - {\theta_1})} \end{array}} \right)\left( {\begin{array}{{cc}} {{e^{i{\varphi_{f1}}}}}&0\\ 0&{{e^{i{\varphi_{s1}}}}} \end{array}} \right)\left( {\begin{array}{{cc}} {\textrm{ }\cos {\theta_1}}&{\sin {\theta_1}}\\ { - \sin {\theta_1}}&{\cos {\theta_1}} \end{array}} \right)\\ T_{lin}^2 = \left( {\begin{array}{{cc}} {\textrm{ }\cos ( - {\theta_2})}&{\sin ( - {\theta_2})}\\ { - \sin ( - {\theta_2})}&{\cos ( - {\theta_2})} \end{array}} \right)\left( {\begin{array}{{cc}} {{e^{i{\varphi_{f2}}}}}&0\\ 0&{{e^{i{\varphi_{s2}}}}} \end{array}} \right)\left( {\begin{array}{{cc}} {\textrm{ }\cos {\theta_2}}&{\sin {\theta_2}}\\ { - \sin {\theta_2}}&{\cos {\theta_2}} \end{array}} \right) \end{array}$$

where θ₁ and θ₂ represent orientation angles of m1 and m2 respectively. The φ_f and φ_s are phase retardations along the fast and slow axes of elliptic nanorods. To convert the linearly polarized transmission matrix to the circularly polarized transmission matrix, a practical transformation matrix $\Lambda = \frac{1}{{\sqrt {2} }}\left( {\begin{array}{{cc}} 1&1\\ i&{ - i} \end{array}} \right)$ is applied here: ${T_{cir}} = {\varLambda ^{ - 1}}{T_{lin}}\varLambda $. Combining the transformation matrices, Eqs. (2) and (3), the transmission Jones matrices of circular polarization wave of m1 and m2 can be expressed as (here assume θ₂= 0) [30]:

(4)$$\begin{array}{l} T_{cir}^1 = \frac{1}{2}\left( {\begin{array}{{cc}} {{e^{i{\varphi_{f1}}}} + {e^{i{\varphi_{s1}}}}}&{({{e^{i{\varphi_{f1}}}} - {e^{i{\varphi_{s1}}}}} ){e^{ - i2{\theta_1}}}}\\ {({{e^{i{\varphi_{f1}}}} - {e^{i{\varphi_{s1}}}}} ){e^{i2{\theta_1}}}}&{{e^{i{\varphi_{f1}}}} + {e^{i{\varphi_{s1}}}}} \end{array}} \right)\\ T_{cir}^2 = \frac{1}{2}\left( {\begin{array}{{cc}} {{e^{i{\varphi_{f2}}}} + {e^{i{\varphi_{s2}}}}}&{{e^{i{\varphi_{f2}}}} - {e^{i{\varphi_{s2}}}}}\\ {{e^{i{\varphi_{f2}}}} - {e^{i{\varphi_{s2}}}}}&{{e^{i{\varphi_{f2}}}} + {e^{i{\varphi_{s2}}}}} \end{array}} \right) \end{array}$$

where the ±2θ₁ refers to the PB phase caused by the rotation of the elliptic nanorods [31]. If relative rotation angles of the two pairs of elliptic nanorods are properly designed, the incident waves could experience constructive interference or destructive interference and lead to CD when the circular polarization waves are incident on the metasurface [32,33]. To realize complete interference of the transmitted waves from m1 and m2, the polarization of the transmitted wave must be consistent. Hence, the transmitted waves from m1 and m2 should contain only cross-polarized component. At this point, both m1 and m2 should have the function of half-wave plate, which convert the incident circularly polarized wave to the transmitted circularly polarized wave with opposite spin direction. The transmitted wave from the unit cell can be regarded as a coherent superposition of the transmitted wave from the elliptic nanorods: T = (T₁+ T₂)/2. Then, the theoretical calculations of the transmission of circularly polarized incident waves are as follows [30]:

(5)$$\begin{array}{l} {T_{LCP}} = \frac{{1 + \cos \Delta {\phi _L}}}{2}\\ {T_{RCP}} = \frac{{1 + \cos \Delta {\phi _R}}}{2} \end{array}$$

where Δϕ_L = (φ_f1 − φ_f2) −2θ₁ and Δϕ_R = (φ_f1 − φ_f2) +2θ₁ represent the phase difference of the transmitted waves between m1 and m2 under LCP or RCP incidence, respectively. It can be clearly seen from the above equation that T_LCP and T_RCP are determined by phase difference Δϕ_L and Δϕ_R, respectively. And they all have the propagation phase difference and the PB phase difference, in which the propagation phase difference is independent of spin, while the PB phase difference is related to spin. To generate high CD (T_LCP= 0 and T_RCP= 1), the phase difference of the transmitted waves between m1 and m2 must be satisfied: Δϕ_L =±180°, Δϕ_R =0 (i.e, the propagation phase difference between m1 and m2 satisfies φ_f1 − φ_f2 =−90° and θ₁= 45°). In other words, so long as the rotation angle difference between m1 and m2 reaches 45°: |θ₁-θ₂|=45°, the desired CD can be obtained.

The reflection for the LCP and RCP waves incidence, the CD in reflection mode and the reflection coefficients matrix of circularly polarized waves can be expressed by the following formulas [34]:

(6)$$\left\{ \begin{array}{l} {R_{RR}} = |{r_{RR}}{|^2},{R_{LR}} = |{r_{LR}}{|^2},{R_{LL}} = |{r_{LL}}{|^2},{R_{RL}} = |{r_{RL}}{|^2}\\ CD = |{R_{RCP}} - {R_{LCP}}|= |{(}{R_{RR}} + {R_{LR}}{)} - {(}{R_{LL}} + {R_{RL}}{)}|\end{array} \right.$$

(7)$${R_{cir}} = \left( {\begin{array}{{ll}} {{r_{RR}}}&{{r_{RL}}}\\ {{r_{LR}}}&{{r_{LL}}} \end{array}} \right) = \frac{1}{2}\left( {\begin{array}{{ll}} {\left({{r_{xx}} - {r_{yy}}} \right)+ i\left({{r_{xy}} + {r_{yx}}} \right)}&{\left({{r_{xx}} + {r_{yy}}} \right)- i\left({{r_{xy}} - {r_{yx}}} \right)}\\ {\left({{r_{xx}} + {r_{yy}}} \right)+ i\left({{r_{xy}} - {r_{yx}}} \right)}&{\left({{r_{xx}} - {r_{yy}}} \right)- i\left({{r_{xy}} + {r_{yx}}} \right)} \end{array}} \right)$$

where r_RR (r_LL) and r_LR (r_RL) are the co-polarized and cross-polarized reflection coefficients for incident circularly polarized waves, respectively. The r_xx (r_yy) and r_yx (r_xy) are the co-polarized and cross-polarized reflection coefficients for linearly polarized waves incidence, respectively.

3. Results and discussions

3.1 Mid-infrared bifunctional metasurface

We optimize the unit at first and obtain the final unit size as shown in Figs. 1(a, b). The transmission (T), reflection (R), and field distributions of the metasurface are shown in Figs. (2) and (3). Figure 2 shows the simulation results of the proposed metasurface when the GSST layer is in amorphous state. When the RCP and LCP waves are incident, the cross-polarized transmission of RCP wave (T_LR) at 74 - 78 THz is much higher than T_LL, T_RL and T_RR, and reaches more than 80% in the range of 73 - 75 THz (Fig. 2(a)). The T_RCP and T_LCP (Fig. 2(b)) are also calculated by formula (1). It can be clearly seen that the metasurface has high transmission for RCP wave and reflects most of LCP wave. The CD can be calculated by the formula (1) and reaches over 0.90 near 74 - 75 THz. Meanwhile, considering that the transmission of LCP wave is very low, we also discuss the reflection of the metasurface. When the LCP wave is incident, the incident wave is reflected by the metasurface (Fig. 2(d)). Figure 2(e) shows the CD is also very high in reflection mode. So our designed metasurface can produce CD both in transmission mode and reflection mode [35–37]. The transmission electric field distributions (Fig. 2(c)) also confirm the asymmetric transmission of the two circularly polarized waves by the metasurface.

Fig. 2. The performance of the proposed metasurface. Transmission (a, b), reflection (d, e) under the circularly polarized waves incidence when the GSST layer is in amorphous state. The electric field distributions (c) and magnetic field distributions (f) of this metasurface under the circularly polarized waves incidence at 74 THz.

Download Full Size | PDF

Then we discuss the magnetic field distributions at 74 THz in Fig. 2(f). There are strong fields in the two pairs of ellipse nanorods. It is the resonance between the two pairs of nanorods that result in almost total polarization-dependent interference (the orthogonally polarized wave incident on the metasurface experiences polarization-dependent interference, one of the polarized waves will experience constructive interference that cause complete transmission, orthogonally polarized wave will experience destructive interference leading to complete reflection). Begin with two separate nanorods (m1 and m2), we discuss in depth the conditions of polarization-dependent interference occur in the designed unit in Appendix B. And the results in Figs. 2(b) and 2(e) also demonstrate this inference.

However, when the amorphous GSST (A-GSST) layer is changed to crystalline GSST (C-GSST) layer, the CD working frequency band become very narrow in the frequency range of 70-78 THz, just same as Ref. [33]. Nevertheless, we surprisingly find that the effective working band of the metasurface is shifted when GSST changes from amorphous to crystalline state due to the difference of refraction index between two states. Then we consider designing a functional device with wider working band: different phase states correspond to different working bands, and the two bands are superimposed to achieve the effect of broadening working band. When the amorphous GSST (A-GSST) layer is changed to crystalline GSST (C-GSST) layer, we also analyze the transmission (T), reflection (R), and field distributions of the metasurface. The working frequency range of the metasurface has been changed from 70 - 78 THz (Fig. 2) to 40 - 50 THz (Fig. 3), and the previously realized CD also disappears. There is almost no reflection, only transmission under the RCP and LCP waves incidence in the working band (Figs. 3(a, d)). However, the most transmitted waves are cross-polarized waves, which means that the metasurface has no selective transmission for the incident RCP and LCP waves. In other words, the metasurface only has one function of circular polarization conversion at this point. And the maximum value of transmitted wave is nearly 100% at 42.5 THz. The peak of CD (Fig. 3(b)) may result from the constructive interference for RCP wave and destructive interference for LCP wave at 48.3 THz, but the peak of CD (Fig. 3(e)) in reflection mode is opposite. By observing the magnetic field distributions (Fig. 3(f)) when the GSST layer is in the crystalline state, we find that the field strength is mostly concentrated on m2, which is different from A-GSST attributed to the change of the refraction index after the phase-state changes of GSST. Therefore, we believe that the total polarization conversion of the unit (Fig. 3(a)) is superimposed by the conversion of a pair of m2 nanorods when the GSST layer is in the crystalline state.

Fig. 3. The performance of the proposed metasurface. Transmission (a, b), reflection (d, e) under the circularly polarized waves incidence when the GSST layer is in crystalline state. The electric field distributions (c) and magnetic field distributions (f) of this metasurface under the circularly polarized waves incidence at 42 THz.

Download Full Size | PDF

3.2 Mid-infrared beam deflector

The phase and transmission characteristics (Fig. 4) of the metasurface indicate that our metasurface can support many optical applications based on Pancharatnam-Berry phase. Here we have simulated and characterized three types of functional metasurfaces. The first functional metasurface is the beam deflector. The incident CP (f = 74 THz) waves (RCP and LCP) are selected for the A-GSST and the incident CP (f = 42 THz) waves (RCP and LCP) for the C-GSST. The deflection angles of different unit arrays are calculated by using the generalized Snell's law [38]:

(8)$${n_t}sin ({\theta _t}) - {n_i}sin ({\theta _i}) = \frac{{{\lambda _0}}}{{2\pi }}\frac{{d\varphi }}{{dx}}$$

where n_t and n_i are the refractive index of the transmitted and incident layer, respectively. θ_t and θ_i are the transmission angle and incident angle, λ₀ is the working wavelength, dφ and dx are the phase and distance difference between adjacent units along the ± x direction.

Fig. 4. The designed units covering 2π phase arrangement and transmission diagram when the GSST layer is in two states.

Download Full Size | PDF

Arranging a set of A-GSST units covering 2π phases (the phase of the units in Fig. 4 is varied from 0° to 360° in steps of 36° along the + x direction). For LCP incidence along the + z direction, we predesign the deflection angle θ=−6° of the unit arrays (due to the presence of Pancharatnam-Berry phase, deflection angle of incident RCP wave is 6°). Using the time-domain solver, open boundaries are applied in ± x and ± z directions, and the periodic boundaries are applied in ± y directions. The results are shown in Fig. 5. As predicted, the incident RCP wave is perfectly converted to LCP wave and deflected by 6° when the GSST layer is in amorphous state (Fig. 5(a)). The calculated deflection efficiency (defined as the ratio between the deflected intensity of the predesigned angle to incident intensity) for RCP wave incident is 38.8%. However, there is weak electric field (Fig. 5(b)) in the transmission field under the LCP wave incidence. The normalized scattering intensity of the transmission wave becomes very low, and the deflection efficiency is only 1.87%. It indicates that the metasurface can effectively block the LCP wave in A-GSST. When the GSST layer is changed to a crystalline state with the refractive index changes, we recalculate the deflection angle θ=12° of the transmission wave in the crystalline state according to the generalized Snell's law. The simulation results are consistent as Figs. 5(c–d) (different incident waves (RCP, LCP) cause different deflection angles (+12°, −12°), the deflection efficiency is 45.5% for RCP wave incidence and 47.6% for LCP wave incidence). Therefore, we have successfully realized two MIR beam deflectors that work in different frequency ranges.

Fig. 5. The electric field and normalized scattering intensity (a, b) under the CP (f = 74 THz) (RCP and LCP) incidence when the GSST layer is in amorphous state. The electric field and normalized scattering intensity (c, d) under the CP (f = 42 THz) (RCP and LCP) incidence when the GSST layer is in crystalline state.

Download Full Size | PDF

3.3 Mid-infrared OAM vortex beam generator

In addition, since OAM vortex beam plays an important role in improving the quality of communication transmission, a MIR metasurface that generates vortex beams has been proposed as shown in Fig. 6(b). The simulation results of the MIR beam deflector have shown that the metasurface allows RCP wave transmission in both states, here we choose the incident RCP (f = 74 THz) wave for the A-GSST and the incident RCP (f = 42 THz) wave for the C-GSST. OAM vortex beam with different topological charges can be generated by phase distribution design. Starting from the center of the metasurface, the phase of each unit position (x, y) on the metasurface can be expressed as [20]:

(9)$$\varphi ({x,y} )= L \cdot arctan ({y/x} )$$

where φ (x, y) represents the phase of the unit (x, y) position in the metasurface, and L represents the topological charge. Vortex beams with different topological charges can be generated by changing the value of L. In this paper, a transmissive metasurface with a topological charge L = 1 and a size of 24 × 24 units is designed, and eight cross-dimensions with nearly equal phase steps of π/4 to cover the full 2π range are selected. Figure 6(a) shows the phase distribution of the vortex generator. To find out the optimal OAM vortex beam generated by the vortex generator in the amorphous and crystalline states, we simulate the vortex beam generation on this metasurface at 10 equally spaced frequency points in 41-43 THz, 73-75 THz under the open boundary conditions. As the three-dimensional (3D) far-field scattering patterns shown in Figs. 6(c)–(d), the metasurface has generated great OAM vortex beams in both states, reaching the best at 42 THz and 74 THz, respectively. Figures 7(a)–(b) respectively show the phase distribution and normalized electric field intensity of the OAM vortex beam generated in the A-GSST (f_RCP= 74 THz) and C-GSST (f_RCP= 42 THz).

Fig. 6. The partial phase distribution (a) and 3D far-field scattering pattern (b) of designed metasurface. The 3D far-field scattering patterns of the metasurface (c) with L = 1 from 73 to 75 THz in steps of 0.5 THz under the RCP incidence when the GSST layer is in amorphous state and (d) with L = 1 from 41 to 43 THz in steps of 0.5 THz under the RCP incidence when the GSST layer is in crystalline state.

Download Full Size | PDF

Fig. 7. The phase and normalized electric field intensity (a), 2D electric field intensity distributions (c) and purity of vortex beam scattering(e) under the RCP (f = 74 THz) incidence when the GSST layer is in amorphous state. The phase and normalized electric field intensity (b), 2D electric field intensity distributions (d) and purity of vortex beam scattering (f) under the RCP (f = 42 THz) incidence when the GSST layer is in crystalline state.

Download Full Size | PDF

To analyze the purity of the vortex beam when the metasurface in the A-GSST and C-GSST, the 2D vortex beam electric field obtained by the above simulation is shown in Figs. 7(c)–(d). The phase singularities of each beam are taken as the center of the circle, and one ring electric field data is taken along the main beam for Fourier transform. Then the OAM spectrum analysis results of the corresponding vortex beam can be obtained. The calculation formula can be expressed as follows:

(10)$${A_l} = \frac{1}{{2\pi }}\int {_0^{2\pi }} E(\varphi ){e^{ - il\varphi }}d\varphi$$

where E(φ) is the selected electric field data, and A_l is the amplitude of the corresponding mode. The purity of vortex beam mode can be expressed by the relative magnitude of the components of each mode in the total energy [39]. It can be seen that the main mode of vortex electromagnetic wave generated by the metasurface occupies the highest energy from the normalized OAM spectrum analysis (Figs. 7(e)–(f)), the efficiency (defined as the ratio between the main mode purity to the total mode purity) of vortex generator in the A-GSST and C-GSST can be calculated as 86.3% and 86.7%, respectively. It shows that the metasurface can generate high-purity OAM vortex beams in two working frequency ranges (they correspond to A-GSST and C-GSST respectively).

3.4 Mid-infrared OAM vortex beam splitter

Then we consider whether the beam splitting and OAM vortex beam can be combined to achieve OAM vortex beam splitter. The simulation results demonstrate that the vortex beam splitting is realized. We decided to use superposition and convolution operation [40,41] to effectively combine the two arrays. In this way, the vortex beam splitting metasurface can be obtained by simple convolution operation of the array of vortex beams and the array of beam splitting (010101…) along the x-axis. We can see that the unit array of vortex beam splitter is generated after superposition in Fig. 8(a) and the designed vortex beam splitter consists of 24 × 24 units. Open boundary conditions are set in x, y, and z directions,9 the 3D far-field scattering pattern (Figs. 8(b)–(c)), phase distribution and normalized electric field intensity (Figs. 9(a)–b)) of this metasurface in amorphous and crystalline states can be obtained by simulation. The purities of vortex beam splitting mode are shown in Figs. 9(c)–(d). And the efficiency in the two states (A-GSST and C-GSST) are 58.8% and 83.2%, respectively. Vortex beam splitting plays an important role in realizing multi-channel communication and encryption. We have successfully achieved MIR vortex beam splitter in different bands.

Fig. 8. Designed metasurface with simply arranging the units (a) for dual-vortex beam generation. The 3D far-field scattering pattern (b).

Download Full Size | PDF

Fig. 9. The phase and normalized electric field intensity (a), purity of vortex beam scattering (c) under the RCP (f = 74 THz) incidence when the GSST layer is in amorphous state. The phase and normalized electric field intensity (b), purity of vortex beam scattering (d) under the RCP (f = 42 THz) incidence when the GSST layer is in crystalline state.

Download Full Size | PDF

4. Conclusion

In summary, we have proposed a reconfigurable all-dielectric metasurface, which changes its response from high circular dichroism (CD) to wideband circular polarization conversion under the incidence of circularly polarized waves. And novel phase-change material Ge₂Sb₂Se₄Te₁ which has low loss and large optical contrast between amorphous and crystalline states in the MIR band has been applied. In addition, under the premise of maintaining high transmission, we have demonstrated three kinds of functional metasurfaces based on Pancharatnam-Berry phase, and implemented abnormal refraction, orbital angular momentum (OAM) vortex beam and OAM vortex beam splitting. Using tunable highly transparent phase-change material GSST provides an excellent approach for achieving more efficient MIR optical integration systems.

Appendix A: Optical properties of GSST

Fig. 10. The optical properties of GSST in two phase states (A-GSST, C-GSST), respectively. The optical properties are obtained from measured results in Ref. [25].

Download Full Size | PDF

Appendix B: Analysis of m1 and m2

To further investigate the physical mechanism of CD, we analyze the transmission and phase distribution of m1 and m2 elliptic nanorods when the GSST layer is in amorphous state, respectively. It can be observed that both m1 and m2 nanorods realize polarization conversion at the working frequency range, as shown in Fig. 11(a). However, we find that the transmission efficiency of cross-polarized wave is not high, and the transmittance of co-polarized wave is higher than the cross-polarized wave, while the metasurface unit composed of m1 and m2 only basically transmits the cross-polarized wave (Fig. 2(a)). Here we think the reason is the intrinsic chirality of the metasurface. Due to asymmetric transmission (chiral structures is often accompanied by asymmetric transmission), a large part of the co-polarized wave reverses under the action of the chiral structure (since the unit with broken rotational symmetry significantly enhance spin-selection performance). Moreover, Fig. 11(b) shows that the phase difference Δφ of cross-polarized waves between the nanorods is 90° around 74 THz. Referring to previous studies [32,33], when a circularly polarized wave is normally incident to the nanorod (m1 or m2), the transmitted wave can be expressed by the following formula:

(11)$$\cos \frac{\delta }{2}{\textrm{exp} ^{i\varphi }}\left[ {\begin{array}{{c}} 1\\ { - i\sigma } \end{array}} \right] - i\sin \frac{\delta }{2}{\textrm{exp} ^{i( - 2\sigma \xi + \varphi )}}\left[ {\begin{array}{{c}} 1\\ {i\sigma } \end{array}} \right]$$

where δ=φ_f - φ_s is the phase difference of the fast and slow axes of elliptic nanorods, φ is the propagation phase determined by the nanorod size, ξ is the rotation angle of the nanorod (ξ = θ₁ for m1, ξ = θ₂ = θ₁+45° for m2). −2σξ is the Pancharatnam-Berry (PB) phase, and σ = ±1 represents the spin state of incident circular polarization waves. It indicates that our elliptic nanorods both have the propagation phase and the PB phase, in which the propagation phase is independent of spin, while the PB phase is related to spin. Therefore, the phase difference Δψ between m1 and m2 under the RCP or LCP wave incident can be calculated as follows:

(12)$$\Delta \psi = \Delta \varphi + 2\sigma \alpha$$

Fig. 11. Simulated transmission of the ellipse nanorods (a), simulated phase of the cross-polarized waves of the ellipse nanorods (b). Simulated transmission of the m1 ellipse nanorod (c), simulated transmission of the m2 ellipse nanorod (d).

Download Full Size | PDF

In this case, Δφ = 90° and α = 45° provides phase difference of Δψ =180° for LCP wave and Δψ = 0° for RCP wave, respectively. According to the interference theory, LCP incidence will experience destructive interference and be reflected, while RCP incidence experiences constructive interference and allow transmission of converted LCP, then CD is implemented (Fig. 2(a)). The transmission electric field distribution (Fig. 2(c)) further demonstrates our inference, which only allows RCP wave transmission and reflects LCP wave in the amorphous state.

However, different from the amorphous state, the phase difference Δφ of cross-polarized waves between the C-GSST nanorods (m1 and m2) is far from 90° in the working frequency range, which cannot meet the conditions of polarization-dependent interference. The polarization conversion efficiency of m2 is relatively high, while that of m1 is opposite, as shown in Fig. 11(c). Nevertheless, from the magnetic field distribution (Fig. 3(f)), we can find that m2 plays a major role in the work at this time, so the polarization conversion efficiency of the metasurface unit consisting of m1 and m2 nanopillars is still high. In other words, the unit composed of a pair of m2 nanorods lead to the polarization conversion. The total polarization conversion efficiency (Fig. 3(a)) of the unit is superimposed by the conversion efficiency of a pair of m2 nanorods when the GSST layer is in the crystalline state.

Funding

National Natural Science Foundation of China (62175119).

Disclosures

The authors declare no conflicts of interest.

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

References

1. M. Y. Shalaginov, S. An, Y. F. Zhang, F. Yang, P. Su, V. Liberman, J. B. Chou, C. M. Roberts, and C. Rios, “Reconfigurable all-dielectric metalens with diffraction-limited performance,” Nat. Commun. 12(1), 1225 (2021). [CrossRef]

2. Y. Huang, T. X. Xiao, Z. W. Xie, J. Zheng, Y. R. Su, W. D. Chen, K. Liu, M. J. Tang, P. Muller-Buschbaum, and L. Li, “Single-layered reflective metasurface achieving simultaneous spin-selective perfect absorption and efficient wavefront manipulation,” Adv. Opt. Mater. 9(5), 2001663 (2021). [CrossRef]

3. A. V. Muraviev, V. O. Smolski, Z. E. Loparo, and K. L. Vodopyanov, “Massively parallel sensing of trace molecules and their isotopologues with broadband subharmonic mid-infrared frequency combs,” Nat. Photonics 12(4), 209–214 (2018). [CrossRef]

4. M. N. Julian, C. Williams, S. Borg, S. Bartram, and H. J. Kim, “Reversible optical tuning of GeSbTe phase-change metasurface spectral filters for mid-wave infrared imaging,” Optica 7(7), 746–754 (2020). [CrossRef]

5. Z. Wang, B. H. Teh, Y. Wang, G. G. Adamo, J. H. Teng, and H. D. Sun, “Enhancing circular dichroism by super chiral hot spots from a chiral metasurface with apexes,” Appl. Phys. Lett. 110(22), 221108 (2017). [CrossRef]

6. Z. G. Li, D. Rosenmann, D. A. Czaplewski, X. D. Yang, and J. Gao, “Strong circular dichroism in chiral plasmonic metasurfaces optimized by micro-genetic algorithm,” Opt. Express 27(20), 28313–28323 (2019). [CrossRef]

7. L. X. Ouyang, D. Rosenmann, D. A. Czaplewski, J. Gao, and X. D. Yang, “Broadband infrared circular dichroism in chiral metasurface absorbers,” Nanotechnology 31(29), 295203 (2020). [CrossRef]

8. J. Kaschke and M. Wegener, “Optical and infrared helical metamaterials,” Nanophotonics 5(4), 510–523 (2016). [CrossRef]

9. Y. J. Huang, X. Xie, M. B. Pu, Y. H. Guo, M. F. Xu, X. L. Ma, X. Li, and X. G. Luo, “Dual-functional metasurface toward giant linear and circular dichroism,” Adv. Opt. Mater. 8(11), 1902061 (2020). [CrossRef]

10. Z. R. Cai, Y. D. Deng, C. Wu, C. Meng, Y. T. Ding, S. I. Bozhevolnyi, and F. Ding, “Dual-functional optical waveplates based on gap-surface plasmon metasurfaces,” Adv. Opt. Mater. 9(11), 2002253 (2021). [CrossRef]

11. H. W. Liu, J. P. Lu, and X. R. Wang, “Metamaterials based on the phase transition of VO₂,” Nanotechnology 29(2), 024002 (2018). [CrossRef]

12. F. Ding, Y. Q. Yang, and S. I. Bozhevolnyi, “Dynamic metasurfaces using phase-change chalcogenides,” Adv. Opt. Mater. 7(14), 1801709 (2019). [CrossRef]

13. A. Tripathi, J. John, S. Kruk, Z. Zhang, H. S. Nguyen, L. Berguiga, P. R. Romeo, R. Orobtchouk, S. Ramanathan, Y. Kivshar, and S. Cueff, “Tunable Mie-resonant dielectric metasurfaces based on VO₂ phase-transition materials,” ACS Photonics 8(4), 1206–1213 (2021). [CrossRef]

14. M. Liu, E. Plum, H. Li, S. Duan, S. X. Li, Q. Xu, X. Q. Zhang, C. H. Zhang, C. W. Zou, B. B. Jin, J. G. Han, and W. L. Zhang, “Switchable chiral mirrors,” Adv. Opt. Mater. 8(15), 2000247 (2020). [CrossRef]

15. J. Y. Tian, H. Luo, Y. Q. Yang, F. Ding, Y. R. Qu, D. Zhao, M. Qiu, and S. I. Bozhevolnyi, “Active control of anapole states by structuring the phase-change alloy Ge₂Sb₂Te₅,” Nat. Commun. 10(1), 396 (2019). [CrossRef]

16. L. Zhang, S. Liu, L. L. Li, and T. J. Cui, “Spin-controlled multiple pencil beams and vortex beams with different polarizations generated by Pancharatnam-Berry coding metasurfaces,” ACS Appl. Mater. Interfaces 9(41), 36447–36455 (2017). [CrossRef]

17. Y. B. Zhang, H. Liu, H. Cheng, J. G. Tian, and S. Q. Chen, “Multidimensional manipulation of wave fields based on artificial microstructures,” Opto-Electron. Adv. 3(11), 200002 (2020). [CrossRef]

18. H. Cheng, S. Q. Chen, P. Yu, W. W. Liu, Z. C. Li, J. X. Li, B. Y. Xie, and J. G. Tian, “Dynamically tunable broadband infrared anomalous refraction based on graphene metasurfaces,” Adv. Opt. Mater. 3(12), 1744–1749 (2015). [CrossRef]

19. A. J. Ollanik, J. A. Smith, M. J. Belue, and M. D. Escarra, “High-efficiency all-dielectric Huygens metasurfaces from the ultraviolet to the infrared,” ACS Photonics 5(4), 1351–1358 (2018). [CrossRef]

20. Y. H. Guo, S. C. Zhang, M. B. Pu, Q. He, J. J. Jin, M. F. Xu, Y. X. Zhang, P. Gao, and X. G. Luo, “Spin-decoupled metasurface for simultaneous detection of spin and orbital angular momenta via momentum transformation,” Light: Sci. Appl. 10(1), 63 (2021). [CrossRef]

21. B. Y. Liu, Y. J. He, S. W. Wong, and Y. Li, “Multifunctional vortex beam generation by a dynamic reflective metasurface,” Adv. Opt. Mater. 9(4), 2001689 (2021). [CrossRef]

22. J. S. Li and J. Z. Chen, “Multi-beam and multi-mode orbital angular momentum by utilizing a single metasurface,” Opt. Express 29(17), 27332–27339 (2021). [CrossRef]

23. F. Zhang, X. Xie, M. B. Pu, Y. H. Guo, X. L. Ma, X. Li, J. Luo, Q. He, H. L. Yu, and X. G. Luo, “Multistate switching of photonic angular momentum coupling in phase-change metadevices,” Adv. Mater. 32(39), 1908194 (2020). [CrossRef]

24. Z. Cai, C. Wu, J. Jiang, Y. T. Ding, Z. W. Zheng, and F. Ding, “Phase-change metasurface for switchable vector vortex beam generation,” Opt. Express 29(26), 42762–42771 (2021). [CrossRef]

25. Y. F. Zhang, J. B. Chou, J. Y. Li, H. S. Li, Q. Y. Du, A. Yadav, S. Zhou, M. Y. Shalaginov, Z. R. Fang, H. K. Zhong, C. Roberts, P. Robinosn, B. Bohlin, C. Rios, H. T. Lin, M. K. Kong, T. Gu, J. Warner, V. Liberman, K. Richardson, and J. J. Hu, “Broadband transparent optical phase change materials for high-performance nonvolatile photonics,” Nat. Commun. 10(1), 4279 (2019). [CrossRef]

26. Q. H. Zhang, Y. F. Zhang, J. Y. Li, R. Soref, T. Gu, and J. J. Hu, “Broadband nonvolatile photonic switching based on optical phase change materials: beyond the classical figure-of-merit,” Opt. Lett. 43(1), 94–97 (2018). [CrossRef]

27. S. Abdollahramezani, O. Hemmatyar, M. Taghinejad, H. Taghinejad, A. Krasnok, A. A. Eftekhar, S. Deshmukh, M. A. EI-Sayed, E. Pop, M. Wutting, A. Alu, W. S. Cai, and A. Adibi, “Electrically driven reprogrammable phase-change metasurface reaching 80% efficiency,” Nat. Commun. 13(1), 1696 (2022). [CrossRef]

28. L. L. Wang, X. J. Huang, M. H. Li, and J. F. Dong, “Chirality selective metamaterial absorber with dual bands,” Opt. Express 27(18), 25983–25993 (2019). [CrossRef]

29. C. L. Zhen, J. Li, S. L. Wang, J. T. Li, M. Y. Li, H. L. Zhao, X. R. Hao, H. P. Zang, and J. Q. Yao, “Optically tunable all-silicon chiral metasurface in terahertz band,” Appl. Phys. Lett. 118(5), 051101 (2021). [CrossRef]

30. H. Huang, S. Qin, K. Q. Jie, J. P. Guo, Q. F. Dai, H. Z. Liu, H. Y. Meng, F. Q. Wang, and Z. C. Wei, “Dynamic generation of giant linear and circular dichroism via phase-change metasurface,” Opt. Express 29(25), 40759–40769 (2021). [CrossRef]

31. D. M. Lin, P. Y. Fan, E. Hasman, and M. L. Brongersma, “Dielectric gradient metasurface optical elements,” Science 345(6194), 298–302 (2014). [CrossRef]

32. F. Zhang, M. B. Pu, X. Li, P. Gao, X. L. Ma, J. Luo, H. L. Yu, and X. G. Luo, “All-dielectric metasurfaces for simultaneous giant circular asymmetric transmission and wavefront shaping based on asymmetric photonic spin-orbit interactions,” Adv. Funct. Mater. 27(47), 1704295 (2017). [CrossRef]

33. Y. J. Huang, T. X. Xiao, Z. W. Xie, J. Zheng, Y. R. Su, W. D. Chen, K. Liu, M. J. Tang, J. Q. Zhu, and L. Li, “Single-layered phase-change metasurfaces achieving efficient wavefront manipulation and reversible chiral transmission,” Opt. Express 30(2), 1337–1350 (2022). [CrossRef]

34. B. Q. Lin, J. X. Guo, L. T. Lv, J. Wu, Y. H. Ma, B. Y. Liu, and Z. Wang, “Ultra-wideband and high-efficiency reflective polarization converter for both linear and circular polarized waves,” Appl. Phys. A 125(2), 76 (2019). [CrossRef]

35. L. Q. Jing, Z. J. Wang, R. K. Maturi, B. Zheng, H. Q. Wang, Y. H. Yang, and H. S. Chen, “Gradient chiral metamirrors for spin-selective anomalous reflection,” Laser Photonics Rev. 11(6), 1700115 (2017). [CrossRef]

36. X. Q. Luo, F. R. Hu, and G. Y. Li, “Dynamically reversible and strong circular dichroism based on Babinet-invertible chiral metasurfaces,” Opt. Lett. 46(6), 1309–1312 (2021). [CrossRef]

37. Y. Chen, J. Gao, and X. D. Yang, “Direction-controlled bifunctional metasurface polarizers,” Laser Photonics Rev. 12(12), 1800198 (2018). [CrossRef]

38. Q. H. Zhu, S. Y. Shi, J. J. Wang, Q. H. Fang, M. H. Li, and J. F. Dong, “Linear optical switch metasurface composed of cross-shaped nano-block and Ge₂Sb₂Te₅ film,” Opt. Commun. 498, 127222 (2021). [CrossRef]

39. G. Q. Li, H. Y. Shi, K. Liu, B. L. Li, J. J. Yi, A. X. Zhang, and X. Zhuo, “Multi-beam multi-mode vortex beams generation based on metasurface in terahertz band,” Acta Phys. Sin. 70(18), 188701 (2021). [CrossRef]

40. W. M. Pan and J. S. Li, “Diversified functions for a terahertz metasurface with a simple structure,” Opt. Express 29(9), 12918–12929 (2021). [CrossRef]

41. M. Zhong and J. S. Li, “Terahertz vortex beam and focusing manipulation utilizing a notched concave metasurface,” Opt. Commun. 511, 127997 (2022). [CrossRef]

Mid-infrared reconfigurable all-dielectric metasurface based on Ge₂Sb₂Se₄Te₁ phase-change material

Abstract

1. Introduction

2. Design and methods

3. Results and discussions

3.1 Mid-infrared bifunctional metasurface

3.2 Mid-infrared beam deflector

3.3 Mid-infrared OAM vortex beam generator

3.4 Mid-infrared OAM vortex beam splitter

4. Conclusion

Appendix A: Optical properties of GSST

Appendix B: Analysis of m1 and m2

Funding

Disclosures

Data availability

References

Data availability

Cited By

Figures (11)

Equations (12)

Optics Express