Abstract
The functionality that is achievable with metasurfaces and metagratings depends heavily on the properties of the materials used to make their components. The key question is which loss level remains acceptable for a given target functionality. In this study, we numerically demonstrate the possible effects of loss-level restrictions for meta-atoms in supercell-based metasurfaces with phase-gradient-enabled functionality and metagratings. For metasurfaces, the loss level predetermines the range of the geometrical parameters needed to cover the full phase range. For metagratings composed of equal meta-atoms, loss-related restrictions can be mitigated compared to metasurfaces, provided that asymmetric diffractions are targeted. The parameter sets used are related to conventional dielectrics, very-high-permittivity composites, and unbiased ferroelectrics, which are suitable candidate materials for metasurfaces and metagratings in the THz range. The results directly demonstrate the effects of the dielectric loss tangent and a selected geometrical parameter.
1. Introduction
Metasurfaces, also known as quasi-planar metamaterials, have transformed the research landscape in photonics, microwaves, and materials engineering [1,2]. Phase gradients that are incorporated into the metasurface for reflected/transmitted waves enable the supercells that are needed for diverse wavefront manipulation functionalities, among which deflection and focusing are best known [3,4,5,6,7]. For coding and intelligent metasurfaces, a few discrete phase values are used [8,9]. In the above-mentioned cases, the required set of different phases (with nearly the same wave magnitudes) is obtained using dielectric (i.e., Mie-type) or quasi-planar metallic subwavelength resonators with slightly or moderately different sizes and/or geometries within supercells. For some functionalities, such as absorbers [10,11] or color filters [12], a phase gradient is not needed; therefore, the metasurface is usually composed of identical resonators. Notably, different wave phases can be created using identical but differently rotated resonant elements that generate a geometric phase (also known as the Pancharatnam–Berry phase) [7], whose polarization manipulation capability relates to PSHE [13,14]. Recently, much attention has focused on gradient and non-gradient metasurfaces that exploit advanced physics, such as BIC [15], topological effects [16], and non-Hermiticity [17,18]. In dielectric resonators, low-loss materials such as Si are commonly used across different parts of the electromagnetic spectrum. However, it remains unclear as to which loss level is still acceptable for targeted functionality. Notably, loss levels can strongly differ due to doping, impurities, or fabrication imperfections. Losses in materials, of which metasurfaces and metagratings are made, have been expected to play at least three roles: (1) they enable absorbers; (2) they enable non-Hermiticity-related effects in transmission, reflection, and wave guiding; and (3) they limit the functionality of gradient metasurfaces and diffraction metagratings. An ambiguous question remains as to whether higher material losses are always less suitable for the targeted functionality. Therefore, it is worth examining some examples to clarify which loss levels are still acceptable for the most common metasurface functionalities. Moreover, metagratings and metasurfaces were studied for their use in diffraction-enabled AT devices [19,20,21,22,23]. Photonic crystals, layered metamaterials, and gratings with special properties have previously been proposed for AT [24,25,26,27,28,29]. The Janus metasurface is a recent term for AT-enabling dielectric and metallic quasi-planar metastructures [30,31,32].
In this paper, we study how losses in selected high- and very-high-permittivity materials used for resonators in meta-atoms (unit cells) affect resulting functionalities. The main motivation is to clarify how losses limit functionality, such as deflection, focusing, coding, and AT. For selected sets of geometric and material parameters, we estimate which loss levels are acceptable and when a lower-loss material is needed. This should help to extend the range of applicable materials and indicate an appropriate way towards suitable design solutions. A particular task is to evaluate materials with very-high permittivity, which are associated with unbiased ferroelectrics and composites. Typically, they show low losses at microwave frequencies but may not retain an acceptably low loss level at THz frequencies [33,34]. At the same time, earlier results show that losses tend to increase when shifting from microwave to THz frequencies, and the dielectric loss tangent, tanδ, is usually larger than 0.01; see [35,36,37]. Despite this, few attempts to realize metastructures comprising such very-high-permittivity materials are known [35,38]. Moreover, very recent experimental results [39], and even earlier ones, demonstrated that tanδ can be significantly reduced for some materials. The presented numerical examples clarify which loss levels are sufficient for basic functionalities. Although the present study is restricted to THz frequencies, the proposed estimates can be useful for higher-frequency ranges. The results were obtained using the Frequency-Domain Solver module of CST Studio Suite 2020 [40]. The simulations used a tetrahedral mesh, a second-order solver, and an accuracy of 10−7 for the equation system solver.
2. Studied Meta-Atoms
Three types of meta-atoms are considered; see Figure 1. Each meta-atom contains (1) a cylindrical microresonator made of a high- or very-high-permittivity material, (2) a spacer, substrate, or half-space, atop which the resonators are placed, and (3) a back-side metallic reflector (for one of the three types). We assume that such meta-atoms will be used as building blocks for gradient metasurfaces and non-gradient metagratings.
Figure 1.
Meta-atoms comprising cylindrical resonators made of a (very)-high-ε material are (a) separated from the back-side (lower-side) reflector by a low-ε spacer, (b) placed atop the low-ε half-space, and (c) placed atop the low-ε finite-thickness substrate.
For gradient metasurfaces, phase coverage within the supercell, which comprises different meta-atoms of the same type, and the nearly constant high magnitude of the reflected or transmitted wave are the key functionality-enabling characteristics [1,2,3,4,5,6,7]. These predetermine how well the examined meta-atoms can perform the functionalities enabled by a phase gradient, i.e., deflection and focusing. The individual meta-atoms of gradient metasurfaces work in the subdiffraction regime. In turn, diffraction metagratings are built from identical meta-atoms and operate at higher frequencies [19,21]. Notably, the studied meta-atoms are asymmetric with respect to the cylinder’s vertical axis. This feature is common for many dielectric metasurfaces. Throughout the paper, rc and h stand for the radius and height of cylindrical microresonators, respectively; p denotes the transverse unit-cell size (this is equal to the period in the case of identical meta-atoms); ts is the spacer thickness; and εc and εs are the permittivities of the cylindrical resonator material and spacer/substrate/half-space material, respectively.
For gradient metasurfaces capable of wavefront manipulation, the supercells are composed of meta-atoms with closely spaced resonance frequencies. In the ideal case, the magnitude of the reflected or transmitted wave is unity for each meta-atom. To deflect the transmitted and reflected waves, the required phase variation, ϕ, along the surface is governed by the generalized Snell’s law [3,7]:
where λ is the electromagnetic wavelength; ni = nr and nt are the refractive indices of the transmission and incidence/reflection half-spaces, respectively; and αi = αr and αt are the incidence/reflection and transmission angles, respectively. For focusing, a parabolic phase distribution is needed. In the one-dimensional case, it is given by [5,6,7].
where F denotes the focal length. The energy balance yields:
where T, R, and A are, respectively, transmittance, reflectance, and absorptance. If the structure comprises a back-side reflector whose thickness is chosen to completely suppress the transmission, the balance is written as R + A = 1. Absorption can be significantly enhanced at resonances, so the trade-off must be addressed at the design stage to achieve the targeted phase variations and avoid unwanted absorption. In order to recall, forward and backward (i.e., upper- and lower-side illumination) transmittances are the same in the subdiffraction regime due to Lorentz reciprocity but are typically unequal in the diffraction regime due to propagating higher diffraction orders [24,28,41].
ntsinαt − nisinαi = (λ/2π)dϕ/dx,
nrsinαr − nisinαi = (λ/2π)dϕ/dx,
T + R + A = 1,
3. Results
In this section, we study the microscale meta-atoms’ ability to exhibit the functionalities attributed to gradient metasurfaces for different values of tanδ and various resonator materials. Notably, absorption is generally not proportional to the amount of lossy material [42,43], whereas the effect of tanδ can be significant for narrow resonances [44]. Consideration is restricted to normal incidence. The spacer/substrate/lower half-space permittivity, εs, is taken as 2.25.
3.1. Effect of Losses in Phase-Gradient-Enabled Functionality
Figure 2 presents the magnitudes and phases (ϕ) of the reflected wave as a function of frequency, f, for the structures that comprise the back-side metallic reflector; see Figure 1a. The results are presented for three sets of rc, h, εc, ts, and p.
Figure 2.
Magnitude (left panel) and phase in degrees (right panel) of the reflected wave for meta-atoms with a back-side reflector (see Figure 1a) at: (a) rc = h = 90 μm, Reεc = 9, ts = 200 μm, p = 275 μm; (b) rc = h = 110 μm, Reεc = 9, ts = 200 μm, and p = 275 μm; and (c) rc = h = 50 μm, Reεc = 35, ts = 200 μm, and p = 150 μm; for different values of tanδ: red lines—0.00001; green lines—0.0001; blue lines—0.0005; orange lines—0.001; violet lines—0.005; brown lines—0.01; black lines—0.05; and gray lines—0.1. All curves are shown for the entire f-range; however, all or some of them overlap.
The magnitude dependence reflects alternating high- and low-transmission ranges (the latter due to absorption). A key feature of phase dependence is the existence of branching points. As a result, the resonances can create strong phase jumps (e.g., Δϕ = 360°) within a particular range of tanδ. Moreover, the observed behavior changes from one resonance to another. For example, for the resonance observed at f = 0.67 THz in Figure 2a, branching occurs between tanδ = 0.005 and tanδ = 0.01. All phase curves corresponding to tanδ ≤ 0.005 overlap when f > 0.67 THz. For the resonance observed near f = 0.58 THz in Figure 2b, opposite signs of Δϕ at the phase jumps are obtained when tanδ = 0.01 and tanδ = 0.05. Similarly, for the resonance arising at f = 0.35 THz in Figure 2c, the phase jumps have opposite signs when tanδ = 0.05 and tanδ = 0.1. In turn, at f = 0.4 THz, branching occurs between tanδ = 0.001 and tanδ = 0.005. For the geometric and material parameters used in Figure 2c, all phase lines corresponding to tanδ ≤ 0.001 overlap if 0.4 < f < 0.53 THz. Therefore, the resulting functionality should not be affected by tanδ in this range. As observed, the loss level governs the behavior of ϕ, which changes abruptly at a critical value, tanδ = tanδc. This feature is especially important for the targeted phase-gradient-enabled functionality. Clearly, the tanδ-related changes in ϕ can correspond to different magnitude behaviors, as observed in Figure 2a–c. The minima depths relate to absorption (see Equation (4)); therefore, they represent the second restriction on phase-gradient-enabled functionality. Phase jumps are required to achieve a desired phase gradient, but they can also lead to strong absorption. Therefore, the design must balance this trade-off.
Similar features to those in Figure 2 appear in the transmitted and reflected waves of the metastructure with a finite-thickness low-ε substrate; see Figure 1c. Figure 3 presents examples of the transmission magnitude and phase dependences on f for two selected sets of geometrical parameters. Figure 3a presents the results of the lower-ε but larger cylinders. A branching point appears in the phase dependence near f = 0.6 THz (jumps show opposite signs for tanδ = 0.05 and tanδ = 0.1), whereas sharp but smaller jumps occur at 0.76 < f < 1 THz. The above-mentioned features are demonstrated in Figure 3a at tanδ ≤ 0.05, while more branching points appear at f > 1 THz, where opposite signs of the jumps are obtained for tanδ = 0.01 and tanδ = 0.05. Transmission magnitude is highly sensitive to variations in tanδ > 0.001, exhibiting a marked decrease, whereas it remains near unity at tanδ = 0.00001. Some transmission features, such as the magnitude maxima at f = 0.6 THz, disappear when tanδ > 0.01, leading to deep minima at the frequencies where the maxima appear at a smaller tanδ.
Figure 3.
Magnitude (left panel) and phase in degrees (right panel) of the transmitted wave for meta-atoms with a finite-thickness low-ε substrate (see Figure 1c) at (a) rc = h = 110 μm, Reεc = 9, ts = 200 μm, and p = 275 μm; (b) rc = h = 50 μm, Reεc = 35, ts = 200 μm, and p = 150 μm; for different values of tanδ: red lines—0.00001; green lines—0.0001; blue lines—0.0005; orange lines—0.001; violet lines—0.005; brown lines—0.01; black lines—0.05; gray lines—0.1. All curves are shown for the entire f-range; however, all or some of them overlap.
For the reflected wave, multiple strong jumps in ϕ (up to 360°) appear, corresponding mainly to the magnitude maxima. In this case, all phase jumps are down jumps, except for tanδ = 0.1. The reflection magnitude decreases by 50% as tanδ increases from 0.001 to 0.1. For the studied designs, the total number of well-pronounced branching points is larger than for transmission. Stronger variations in the reflection wave phase often co-exist with weaker variations in the transmitted wave phase.
Figure 3b presents the results for smaller cylinders made of a very-high-ε material. For the transmitted wave phase, the first branching point appears at f = 1 THz, when the phase shows a strong down-jump at tanδ ≤ 0.05 and a weak up-jump at tanδ = 0.1. The weak jumps appear near f = 0.65 THz and f = 0.97 THz, corresponding to the ranges where transmission is suppressed. Interestingly, they either correspond (at f = 0.65 THz) or do not correspond (near f = 0.97 THz) to significant variations in the reflected wave phase; therefore, we cannot state a general rule. Note that not only passbands but also stopbands degrade as tanδ increases; therefore, a larger tanδ does not necessarily lead to stronger reflection. Both transmission and reflection magnitudes/phases are weakly sensitive to the variations in tanδ near f = 0.82 THz because there are no resonances there.
Next, Figure 4 presents the magnitude and phase of the transmitted wave when the cylindrical resonators are placed atop a semi-infinite low-ε half-space; see Figure 1b. As expected, the magnitude behavior is similar to that in Figure 3 over either part of the frequency range (as seen in Figure 4a) or the whole range (as seen in Figure 4b). Both differences and similarities in magnitude behavior can be interpreted by using input impedance; however, doing so goes beyond the scope of this paper. For instance, the differences occur at f > 0.75 THz for the first parameter set; compare Figure 4a to Figure 3a. The effect exerted by tanδ on magnitude is similar to that seen in Figure 3. Phase dependence in Figure 3a seems more sensitive to the applied variations in tanδ than those in Figure 4a. Although the branching point with a strong phase jump at f ≈ 0.6 THz remains, the other significant jump at f ≈ 1 THz disappears, as do the small jumps that occurred between 0.76 THz and 1 THz.
Figure 4.
Magnitude (left panel) and phase in degrees (right panel) of the transmitted wave in case of meta-atoms with a semi-infinite low-ε half-space (see Figure 1b) at (a) rc = h = 110 μm, Reεc = 9, and p = 275 μm; (b) rc = h = 50 μm, Reεc = 35, and p = 150 μm; for different values of tanδ: red lines—0.00001; green lines—0.0001; blue lines—0.0005; orange lines—0.001; violet lines—0.005; brown lines—0.01; black lines—0.05; gray lines—0.1. All curves are shown for the entire f-range; however, all or some of them overlap.
In turn, comparing Figure 3b and Figure 4b shows that all basic features are identical, except in the f-range around 1 THz, where the branching becomes more complicated in Figure 4b. We observe three different branches of ϕ-dependence as follows: one for tanδ = 0.1, one for tanδ = 0.00001, and one for all the remaining (intermediate) values of tanδ. The small jumps observed for both metastructures near 0.7 THz and 1 THz are similar. In comparison, the magnitude dependences shown in Figure 3b and Figure 4b show no significant differences. Notably, each plot in Figure 2, Figure 3 and Figure 4 shows the results for a given rc; they allow estimation of the capability of a phase-gradient-enabled functionality, and their role is to prove whether the desired phase jumps are possible.
Now, let us demonstrate how the results obtained can be used to assess which functionality is achievable while gradually sweeping rc at tanδ = const. To do so, the phase and magnitude results can be presented on the frequency–geometric parameter plane, according to ref. [42]. Figure 5 presents an example of the color maps of the reflected wave’s phase as a function of frequency and microcylinder radius for metastructures with a back-side metallic reflector; see Figure 1a. The other geometric parameters are kept the same as seen in Figure 2a. The necessary phase condition that must be fulfilled is the passing, such as from red (ϕ = 180°) to red (ϕ = −180°) via all other colors at f = const, while varying rc; this predetermines a possible range of rc variations, which may yield the required full coverage of ϕ. Clearly, this is a necessary condition but not a sufficient condition. Indeed, not all such rc ranges are applicable because the magnitude should not be smaller than a preset value close to unity [42]. It is evident that there is no significant difference between tanδ = 0.0005 and tanδ = 0.005; therefore, they offer nearly the same f-ranges for phase coverage. The situation changes when tanδ = 0.05. Therefore, it can be concluded that the tanδ ≤ 0.005 condition is suitable for the parameters used in Figure 5 when full phase coverage is needed. It should be noted that sharper resonances, which may enable narrower ranges of rc variation, are less desirable because fabrication imperfections in the considered frequency and size ranges are expected to be within ±1 μm [45].
Figure 5.
Phase of reflected wave (in degrees) on (f,rc)-plane for the structures with a back-side reflector (see Figure 1a) at rc = h = 90 μm, Reεc = 9, ts = 200 μm, and p = 275 μm: (a) tanδ = 0.0005, (b) tanδ = 0.005, and (c) tanδ = 0.05. Dashed white lines indicate the possible ranges of rc that are suitable for full phase coverage.
Figure 6 presents colored maps of the reflected wave phase as a function of frequency and cylinder radius for metastructures of the same type as in Figure 5; however, now the same parameters as those in Figure 2c are used. At tanδ = 0.0005, full phase coverage appears. At tanδ = 0.05, such rc ranges that enable full coverage are not found. Notably, it is still practical to cover half of the entire ϕ-range; i.e., from 0 to 180° (as required, for instance, for 1-bit coding), which corresponds to a transition, such as from a red to a cyan region at f = const. On the other hand, a wider phase range can be obtained while using a wider rc range than in Figure 6, even when tanδ = 0.05. Compared to Figure 5, the structure in Figure 6 can yield sharper resonances at tanδ = 0.0005, but this makes the required rc values less equidistant if ∆ϕ ≈ const is targeted while forming a discrete set of radii for a supercell. Therefore, whether the parameter set used in Figure 6 is more suitable remains a question for discussion. Its only possible advantage, as compared to Figure 5, is smaller meta-atoms (unit cells).
Figure 6.
Phase of reflected wave (in degrees) on (f,rc)-plane for the structures with a back-side reflector (see Figure 1a) at rc = h = 50 μm, Reεc = 35, ts = 200 μm, and p = 150 μm: (a) tanδ = 0.0005 and (b) tanδ = 0.05. Dashed white lines in (a) indicate three ranges of rc that are formally suitable to cover the full phase range.
3.2. Effect of Losses in Metagratings Enabling Asymmetric Diffraction
In this section, we study the effects of the variations in the resonator material’s tanδ on asymmetric diffraction—one of the basic mechanisms of AT—for the selected metagratings. The main goal is to demonstrate the potential of the studied asymmetric metagratings to generate asymmetric diffraction at different tanδ values. However, finding regimes with maximal asymmetry in reflection and/or transmission goes beyond the scope of this paper. Notably, diverse aspects of modern metagratings have been studied in the recent literature; for instance, see refs. [19,21,46].
Figure 7 presents the magnitudes and phases of the transmitted and reflected waves arising from the ±1-diffraction orders for structures that represent a periodic array of dielectric resonators placed at the interface between air and low-ε half-spaces; see Figure 1b. This structural type is selected among the three types introduced in Figure 1 because it can ensure diffraction asymmetry while changing the incidence direction for the opposite one. Indeed, it is more common in AT studies when the upper and lower half-spaces comprise the same material, e.g., see refs. [19,24,26,28]; however, a periodic interface between two different uniform media offers reasonable alternatives in many cases. Since the zero incidence angle is retained, we have |t−1| = |t+1| for the transmitted waves and |r−1| = |r+1| for the reflected waves in all considered examples. The results are presented for the three sets of parameters rc, εc, h, and p. The first set is similar to that in Figure 4a, except for the rc value, which is the same as in Figure 2a; the second set is the same as in Figure 4a; and the third set is similar to that in Figure 4b but corresponds to larger values of rc and p, which are taken to enhance diffraction.
Figure 7.
Magnitude of (a,c,d) ±1-order co-polarized transmitted wave and (b) ±1-order co-polarized reflected wave for meta-atoms with a semi-infinite low-ε half-space (see Figure 1b) at (a,b) rc = h = 90 μm, Reεc = 9, and p = 275 μm; (c) rc = h = 110 μm, Reεc = 9, and p = 275 μm; (d) rc = h = 90 μm, Reεc = 35, and p = 275 μm,; for different values of tanδ: red lines—0.00001; green lines—0.0001; blue lines—0.0005; orange lines—0.001; violet lines—0.005; brown lines—0.01; black lines—0.05; gray lines—0.1; thick lines—forward-case illumination (denoted by fw); thin lines—backward-case illumination (denoted by bw). R and T stand for reflection and transmission, respectively. All curves are shown for the entire f-range, but some of them overlap.
In line with the theory of diffraction gratings [47], placing a periodic array at the interface of two uniform half-spaces immediately creates asymmetry in coupling conditions for forward-case and backward-case illuminations, if f1 < f < f2, where f1 = c/(pεs1/2) and f2 = c/p are threshold frequencies, starting from which ±1-orders may propagate in the half-space with ε = εs and in the air half-space, respectively. As a result, asymmetric diffractions do appear, while Lorentz reciprocity is preserved, and zero-order transmission is the same at forward and backward illumination [19,41,48]. Nonzero (higher) diffraction orders are expected to appear in the lower half-space (under forward illumination), starting at f = f1. In contrast, in the upper air half-space (under backward illumination), they can only start at f = f2. The results presented in Figure 7 agree with the qualitative explanation provided above.
In Figure 7a, diffractions are observed in the transmission regime, starting from f ≈ 0.6 THz under forward illumination and from f ≈ 1.07 THz under backward illumination. The contrast |t−1fw|/|t−1bw| is not strongly affected here by tanδ—at least if it is less than or equal to 0.01. For example, |t−1fw|/|t−1bw| ≈ 65 at f = 1.005 THz. Figure 7b presents the complementary results for |r−1|. The contrast for reflected waves, |r−1bw|/|r−1fw|, is of the same order as |t−1fw|/|t−1bw|, provided that f1 < f < f2. The observed effects of tanδ on the reflected and the transmitted diffraction orders are similar. For the set of parameters used in Figure 7a,b, the condition tanδ ≤ 0.01 can be adopted since it preserves the basic features of asymmetric diffraction. In Figure 7c, the regime with |t−1fw| = 0.488 occurs at f = 0.828 THz, while the contrast |t−1fw|/|t−1bw| ≈ 3170. In Figure 7d, we obtain |t−1fw| ≈ 0.55 at f ≈ 0.939 THz, which yields the contrast |t−1fw|/|t−1bw| ≈ 406. The behavior of |r−1fw| and |r−1bw| at the parameters used in Figure 7c,d is similar to that observed in Figure 7b. At a larger εc, sharper resonances may appear that are more sensitive to variations in tanδ. In all cases discussed above, the magnitudes and contrasts were evaluated at tanδ = 0.00001.
The results show that the microcylinder parameters, i.e., h, εc, and rc, are important for diffraction efficiency and must be carefully adjusted, as well as p. The role of tanδ seems to be less important than for phase-gradient-enabled functionalities.
Asymmetric diffractions enable not only the co-polarized waves but also the cross-polarized outgoing waves. Figure 8 shows the asymmetric cross-polarization effects for two of the three metagratings taken from Figure 7. The magnitudes of the cross-polarized −1 diffraction orders, for forward and backward illumination, are denoted by |t−1fw(cross)| and |t−1bw(cross)|, respectively. For instance, |t−1fw(cross)| ≈ 0.485 and |t−1fw(cross)|/|t−1bw(cross)| ≈ 1033 at f ≈ 0.87 THz in Figure 8a, whereas |t−1fw(cross)| ≈ 0.6 and |t−1fw(cross)|/|t−1bw(cross)|≈ 50 at f ≈ 1.015 THz in Figure 8b. In these cases, and for most of the considered range of f variation, the results are almost the same when tanδ ≤ 0.01. Moreover, the diffraction magnitudes can be insensitive to the choice of tan δ, as observed in Figure 8a at 0.75 < f < 0.79 THz and in Figure 8b at f = 0.8 THz; however, this does not correspond to a maximum of |t−1fw(cross)|. Such behavior is typical in cases with relatively weak field localization in the resonators, but they still yield high-to-moderate-efficiency nonzero diffraction orders. As seen in Figure 7, the magnitudes and contrasts were evaluated at the lowest value of tanδ.
Figure 8.
Magnitude of ±1-order cross-polarized transmitted wave for the meta-atoms with a semi-infinite low-ε half-space (see Figure 1b) at (a) rc = h = 90 μm, Reεc = 9, and p = 275 μm; (b) rc = h = 110 μm, Reεc = 9, and p = 275 μm,; for different values of tanδ: red lines—0.00001; green lines—0.0001; blue lines—0.0005; orange lines—0.001; violet lines—0.005; brown lines—0.01; black lines—0.05; gray lines—0.1; thick lines—forward-case illumination (denoted by fw); thin lines—backward-case illumination (denoted by bw). All curves are shown for the entire f-range, but some of them overlap.
Figure 9 presents the results for free-standing asymmetric metagratings comprising a finite-thickness substrate, as seen in Figure 1c. In this case, the media above and below the metagrating are air; therefore, the f-ranges where higher diffraction orders may propagate are the same under forward and backward illumination (i.e., f1 = f2). In turn, the f-ranges where diffraction asymmetry is strong are rather accidental. In contrast to the metagratings placed at the interface, as seen in Figure 7 and Figure 8, strong diffraction occurs when the noncorrugated side of the structure is illuminated; i.e., under backward illumination. For instance, |t−1bw| = 0.48 at f = 1.1315 THz and |t1bw|/|t−1fw| ≈ 10.7 in Figure 9a.
Figure 9.
Magnitude of (a) ±1-order co-polarized transmitted wave and (b) ±1-order cross-polarized transmitted wave for the meta-atoms with a low-ε finite-thickness substrate (see Figure 1c) at (a) rc = h = 90 μm, Reεc = 9, ts = 200 μm, and p = 275 μm; (b) rc =h = 90 μm, Reεc =35, ts = 200 μm, and p = 275 μm; for different values of tanδ: red lines—0.00001; green lines—0.0001; blue lines—0.0005; orange lines—0.001; violet lines—0.005; brown lines—0.01; black lines—0.05; gray lines—0.1; thick lines—forward-case illumination (denoted by fw); thin lines—backward-case illumination (denoted by bw). All curves are shown for the entire f-range, but some of them overlap.
In Figure 9b, moderate transmission asymmetry appears at f = 1.093 THz. In this case, |t−1bw(cross)| = 0.61 and |t1bw(cross)|/|t−1fw(cross)| ≈ 5.86. One more feature should be noted, i.e., |t1bw(cross)| ≈ |t−1fw(cross)| ≈ 10.7 at f = 1.0963 THz in Figure 9b. It can be viewed as an anti-asymmetric case because despite the strong structural asymmetry, cross-polarized diffractions are nearly symmetric. Moreover, the rigorous symmetric case may be achievable by optimizing geometrical parameters. All features observed in Figure 9b are preserved, at least if tanδ ≤ 0.001. As above, the magnitudes and contrasts were calculated at tanδ = 0.00001. The studied metagratings with a finite-thickness substrate are more sensitive to variations in tanδ than those in Figure 7 and Figure 8, but they do not achieve significantly higher contrasts.
4. Discussion
In this work, we studied meta-atoms based on Mie-type (volumetric) resonators. We assumed that the joint effect of different meta-atoms was responsible for the phase gradient. In contrast, a recent study (ref. [49]) used volumetric dielectric resonators to design the geometric-phase effect in THz multilayer gradient metasurfaces, which allow spin-dependent amplitude and phase modulation. In that case, using Si as a resonator material has enabled moderate-to-high performance, similar to many other metasurfaces operating in different frequency ranges.
Moreover, quasi-planar metallic resonant components are often used in metasurfaces, and losses should be considered at the design stage. Two main approaches to their incorporation are similar to those for Mie-type resonators. In the first approach, the phase gradient is achieved by using resonators that differ slightly or moderately. In the second approach, the same resonant elements are applied to create a geometric-phase effect. For instance, the latter has recently been used to generate vectorial beams [50]. In turn, the former has been used to achieve THz focusing with 80% efficiency [51]. Although materials in metasurfaces in refs. [49,50,51] have been chosen—most likely based on earlier results or even intuitively—it remains unclear whether the targeted functionality is preserved when substituting some of those materials with more lossy ones. Despite this, the results in refs. [49,50,51] can be used, along with those presented in this paper, to clarify which material choice is possible.
The same is true for non-gradient metasurfaces comprising quasi-planar metallic components; e.g., see ref. [52]. Notably, earlier studies of structures comprising very-high-permittivity materials have focused on the primary possibility of achieving a specific functionality when the material loss is zero. In particular, this has been performed in various studies on components for optical nano-circuits [53], scattering suppression [54], AT-enabling photonic-crystal gratings [41], and polarization-preserving diffraction-free metagratings [55], to name a few.
For AT devices, breaking spatial inversion symmetry is insufficient. They need a periodic array at the interface to create new (in our case, nonzero-diffraction-order-enabled) transmission channels [19,24,28,41,48]. Diffraction metagratings can be less sensitive to tanδ than gradient metasurfaces because the targeted functionality is achievable even at larger tanδ values, despite lower efficiency and AT contrast. Generally, phase-gradient-enabled functionalities of metasurfaces are possible (or impossible), depending on the tanδ and the specifics of the Mie-type resonance used.
At normal incidence and for the meta-atoms in Figure 1, we always have |t−1| = |t+1| and |r−1| = |r+1|; therefore, the maximum magnitude of either the transmitted or reflected wave is limited to 0.7, which is consistent with energy conservation. Compared to polarization-preserving structures (e.g., see refs. [41,55]), the studied metagratings may enable cross-polarized transmission and reflection in the diffraction regime. The nonoptimized structure in Figure 9b yields |t-1bw(cross)|2 = |t+1bw(cross)|2 = 0.84 near f = 1.09 THz; i.e., more than 80% of the incident-wave energy is transferred to the cross-polarized transmitted waves arising from the ±1 orders. A similar regime with |t-1bw| = |t+1bw| has been obtained for co-polarized transmission in the photonic-crystal gratings [48], where it was attributed to unidirectional splitting. By comparison, a nearly perfect AT with |t−1fw| ≈ 1 and |t−1bw| ≈ 0 can be obtained in photonic-crystal gratings at oblique incidence [41]. Obtaining arbitrary polarization for ±1-order waves seems realistic but needs further study. To compare, the conventional approach to polarization manipulation, either combined [4] or not combined [44] with diffraction, invokes few-layer metasurfaces. Notably, at microwave frequencies, we may have more options for loss control in metasurfaces; for instance, by using a tunable load [55].
5. Concluding Remarks
The results demonstrate how the loss tangent of the material used to make cylindrical resonators affects the achievable functionality of metasurfaces and metagratings. They show that strong phase variations with frequency, i.e., phase jumps required for phase-gradient-enabled functionality, may depend on the resonance choice, for given values of tanδ, and geometrical parameters. In turn, for each resonance and parameter set, there is a critical value of tanδ, above which the desired phase jump is not achieved. A universal condition for phase-gradient-related functionalities cannot be introduced, but some estimates for tanδ can be obtained. The performed study makes evident the facts that tanδ predetermines (i) whether a functionality requiring the full phase coverage can be achieved, and (ii) which range of the resonator radius is needed and, hence, which fabrication accuracy is still acceptable. Although this study is restricted to the THz range, it can provide useful guidelines for choosing materials in a much wider frequency range, including near-infrared and, probably, the visible frequencies. Metagratings enabling asymmetric diffractions are generally less sensitive to the choice of tanδ and its possible variations. Although diffraction efficiency and forward(backward)-to-backward(forward) diffraction contrast still depend on tanδ, the asymmetric functionality does not disappear. The results show that transferring incident-wave energy into higher diffraction orders is possible for diverse materials, which can differ strongly in permittivity and losses. In particular, materials with tanδ ≤ 0.01, as expected for many composites and unbiased ferroelectrics at THz frequencies, can still be suitable for use in diffraction metagratings. Asymmetric diffractions may create not only co-polarized but also cross-polarized components for higher diffraction orders. Diffraction-enabled asymmetric polarization manipulation will be a subject of one of our forthcoming studies. Notably, both metastructure optimization for higher efficiency and resonance identification were beyond the scope of this paper. Nevertheless, this study clarifies which materials are appropriate and which are inappropriate for a selected functionality.
Funding
The contribution by A.E.S. was funded by Narodowe Centrum Nauki, Project UMO-2020/39/I/ST3/02413.
Data Availability Statement
The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.
Conflicts of Interest
The author declares no conflicts of interest.
Abbreviations
The following abbreviations are used in this manuscript:
| BIC | Bound States in the Continuum |
| PSHE | Photonic Spin Hall Effect |
| AT | Asymmetric Transmission |
References
- Chen, H.T.; Taylor, A.J.; Yu, N. A review of metasurfaces: Physics and applications. Rep. Prog. Phys. 2016, 79, 076401. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Ding, F.; Pors, A.; Bozhevolnyi, S.I. Gradient metasurfaces: A review of fundamentals and applications. Rep. Prog. Phys. 2016, 81, 026401. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Yu, N.; Genevet, P.; Kats, M.A.; Aieta, F.; Tetienne, J.P.; Capasso, F.; Gaburro, Z. Light propagation with phase discontinuities: Generalized laws of reflection and refraction. Science 2011, 334, 333–337. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Grady, N.K.; Heyes, J.E.; Chowdhury, D.R.; Zeng, Y.; Reiten, M.T.; Azad, A.K.; Taylor, A.J.; Dalvit, D.A.R.; Chen, H.-T. Terahertz metamaterials for linear polarization conversion and anomalous refraction. Science 2013, 340, 1304–1307. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Jiang, X.; Chen, H.; Li, Z.; Yuan, H.; Cao, L.; Luo, Z.; Zhang, K.; Zhang, Z.; Wen, Z.; Zhu, L.-G.; et al. All-dielectric metalens for terahertz wave imaging. Opt. Express 2018, 26, 14132–14142. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Khorasaninejad, M.; Capasso, F. Metalenses: Versatile multifunctional photonic components. Science 2017, 358, eaam8100. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Kang, M.; Feng, T.; Wang, H.T.; Li, J. Wave front engineering from an array of thin aperture antennas. Opt. Express 2012, 20, 15882–15890. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Li, L.; Zhao, H.; Liu, C.; Li, L.; Cui, T.J. Intelligent metasurfaces: Control, communication and computing. Elight 2022, 2, 7. [Google Scholar] [CrossRef] [Scilit]
- Liu, S.; Cui, T.J. Concepts, working principles, and applications of coding and programmable metamaterials. Adv. Opt. Mater. 2017, 5, 1700624. [Google Scholar] [CrossRef] [Scilit]
- Abdulkarim, Y.I.; Mohanty, A.; Acharya, O.P.; Appasani, B.; Khan, M.S.; Mohapatra, S.K.; Muhammadsharif, F.F.; Dong, J. A review on metamaterial absorbers: Microwave to optical. Front. Phys. 2022, 10, 893791. [Google Scholar] [CrossRef] [Scilit]
- Alaee, R.; Albooyeh, M.; Rockstuhl, C. Theory of metasurface based perfect absorbers. J. Phys. D Appl. Phys. 2017, 50, 503002. [Google Scholar] [CrossRef] [Scilit]
- Heydari, E.; Sperling, J.R.; Neale, S.L.; Clark, A.W. Plasmonic color filters as dual-state nanopixels for high-density microimage encoding. Adv. Func. Mater. 2017, 27, 1701866. [Google Scholar] [CrossRef] [Scilit]
- Yin, X.; Ye, Z.; Rho, J.; Wang, Y.; Zhang, X. Photonic spin Hall effect at metasurfaces. Science 2013, 339, 1405–1407. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Liu, Y.; Ke, Y.; Luo, H.; Wen, S. Photonic spin Hall effect in metasurfaces: A brief review. Nanophotonics 2017, 6, 51–70. [Google Scholar] [CrossRef] [Scilit]
- Kang, M.; Liu, T.; Chan, C.T.; Xiao, M. Applications of bound states in the continuum in photonics. Nat. Rev. Phys. 2023, 5, 659–678. [Google Scholar] [CrossRef] [Scilit]
- Song, Q.; Odeh, M.; Zúñiga-Pérez, J.; Kanté, B.; Genevet, P. Plasmonic topological metasurface by encircling an exceptional point. Science 2021, 373, 1133–1137. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Yan, Q.; Zhao, B.; Zhou, R.; Ma, R.; Lyu, Q.; Chu, S.; Hu, X.; Gong, Q. Advances and applications on non-Hermitian topological photonics. Nanophotonics 2023, 12, 2247–2271. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Shou, Y.; Wang, D.; Wang, Y.; Huang, Q.K.L.; Chen, H.; Yu, W.; Ju, R.; Chen, H.; Li, Y. Resonant and scattering exceptional points in non-Hermitian metasurfaces. npj Nanophotonics 2025, 2, 29. [Google Scholar] [CrossRef] [Scilit]
- Serebryannikov, A.E.; Lakhtakia, A.; Ozbay, E. Thermally mediated transmission-mode deflection of terahertz waves by lamellar metagratings containing a phase-change material. Opt. Mater. Express 2024, 14, 745–758. [Google Scholar] [CrossRef] [Scilit]
- Sung, J.; Lee, G.Y.; Choi, C.; Hong, J.; Lee, B. Polarization-dependent asymmetric transmission using a bifacial metasurface. Nanoscale Horiz. 2020, 5, 1487–1495. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Ra’di, Y.; Sounas, D.L.; Alù, A. Metagratings: Beyond the limits of graded metasurfaces for wave front control. Phys. Rev. Lett. 2017, 119, 067404. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Liu, W.; Huang, L.; Ding, J.; Xie, C.; Luo, Y.; Hong, W. High-performance asymmetric optical transmission based on a dielectric–metal metasurface. Nanomaterials 2021, 11, 2410. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Shang, G.; Li, G.; Guan, C.; Zhao, W.; Li, H.; Zhang, K.; Liu, J.; Wu, Q.; Ding, X.-M.; Burokur, S.N.; et al. Metasurface with directional-controlled asymmetric transmissions. Adv. Photonics Res. 2023, 4, 2200359. [Google Scholar] [CrossRef] [Scilit]
- Rodríguez-Ulibarri, P.; Beruete, M.; Navarro-Cía, M.; Serebryannikov, A.E. Wideband unidirectional transmission with tunable sign-switchable refraction and deflection in nonsymmetric structures. Phys. Rev. B 2013, 88, 165137. [Google Scholar] [CrossRef] [Scilit]
- Serebryannikov, A.E.; Ozbay, E. Isolation and one-way effects in diffraction on dielectric gratings with plasmonic inserts. Opt. Express 2009, 17, 278–292. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Habib, M.; Serebryannikov, A.E.; Caglayan, H.; Vandenbosch, G.A.; Ozbay, E. Connection of collimation, asymmetric beaming, and independent transmission-reflection processes in concentric-groove gratings supporting spoof surface plasmons. Plasmonics 2019, 14, 721–729. [Google Scholar] [CrossRef] [Scilit]
- Barkathulla, A.; Khan, F.; Lin, M.; Wang, Q.; Shen, S.; Liu, Q.; Natesan, Y.; Ouyang, Z. Demonstration of Asymmetric Electromagnetic Mode Conversion Induced by Spatial Symmetry Breaking in Linear Asymmetric Photonic Crystals. Laser Photonics Rev. 2026, 20, e02904. [Google Scholar] [CrossRef] [Scilit]
- Lockyear, M.J.; Hibbins, A.P.; White, K.R.; Sambles, J.R. One-way diffraction grating. Phys. Rev. E—Stat. Nonlinear Soft Matter Phys. 2006, 74, 056611. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Cicek, A.; Yucel, M.B.; Kaya, O.A.; Ulug, B. Refraction-based photonic crystal diode. Opt. Lett. 2012, 37, 2937–2939. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Chen, K.; Ding, G.; Hu, G.; Jin, Z.; Zhao, J.; Feng, Y.; Jiang, T.; Alù, A.; Qiu, C.-W. Directional Janus metasurface. Adv. Mater. 2020, 32, 1906352. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Chang, Y.C.; Lin, C.Y.; Su, H.T.; Wu, Y.C.; Hsu, C.; Chang, S.; Huang, Y. Directionally asymmetric holographic imaging for arbitrary polarization states using Janus metasurfaces. Adv. Opt. Mater. 2025, 13, e01546. [Google Scholar] [CrossRef] [Scilit]
- Yang, W.; Chen, K.; Dong, S.; Wang, S.; Qu, K.; Jiang, T.; Zhao, J.; Feng, Y. Direction-duplex Janus metasurface for full-space electromagnetic wave manipulation and holography. ACS Appl. Mater. Interfaces 2023, 15, 27380–27390. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Hoorfar, A.; Perrotta, A. An experimental study of microstrip antennas on very high permittivity ceramic substrates and very small ground planes. IEEE Trans. Antennas Propag. 2001, 49, 838–840. [Google Scholar] [CrossRef] [PubMed]
- Savelev, R.S.; Filonov, D.S.; Petrov, M.I.; Krasnok, A.E.; Belov, P.A.; Kivshar, Y.S. Resonant transmission of light in chains of high-index dielectric particles. Phys. Rev. B 2015, 92, 155415. [Google Scholar] [CrossRef] [Scilit]
- Lan, C.; Zhu, D.; Gao, J.; Li, B.; Gao, Z. Flexible and tunable terahertz all-dielectric metasurface composed of ceramic spheres embedded in ferroelectric/elastomer composite. Opt. Express 2018, 26, 11633–11638. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Zeng, Y.; Wang, W.; Ling, F.; Yao, J. Terahertz wave modulation properties of thermally processed BST/PZT ferroelectric photonic crystals. Photonics Res. 2020, 8, 1002–1010. [Google Scholar] [CrossRef] [Scilit]
- Dupas, C.; Guillemet-Fritsch, S.; Geffroy, P.M.; Chartier, T.; Baillergeau, M.; Mangeney, J.; Roux, J.-F.; Ganne, J.-P.; Marcellin, S.; Degiron, A.; et al. High permittivity processed SrTiO3 for metamaterials applications at terahertz frequencies. Sci. Rep. 2018, 8, 15275. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Gaillot, D.P.; Croënne, C.; Lippens, D. An all-dielectric route for terahertz cloaking. Opt. Express 2008, 16, 3986–3992. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Djemmah, D.A.; Gourdonnaud, D.; Bouamrane, F.; Roux, J.F.; Geffroy, P.M.; Akmansoy, E. Impact of sintering conditions on the dielectric properties of TiO2 ceramics for metamaterials applications at terahertz frequencies. arXiv 2026, arXiv:2606.25930. [Google Scholar]
- Electromagnetic Simulation Solvers. Available online: https://www.3ds.com/products-services/simulia/products/cst-studio-suite/solvers/ (accessed on 25 May 2026).
- Serebryannikov, A.E.; Cakmak, A.O.; Ozbay, E. Multichannel optical diode with unidirectional diffraction relevant total transmission. Opt. Express 2012, 20, 14980–14990. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Serebryannikov, A.E.; Fataliyev, K.; Cakmak, A.O.; Colak, E. Capability of dielectric resonator based meta-atoms with VO2 Components for switchable coding and wavefront-manipulating THz metasurfaces. Materials 2026, 19, 2449. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Rodríguez-Ulibarri, P.; Beruete, M.; Serebryannikov, A.E. One-way quasiplanar terahertz absorbers using nonstructured polar dielectric layers. Phys. Rev. B 2017, 96, 155148. [Google Scholar] [CrossRef] [Scilit]
- Gokkavas, M.; Gundogdu, T.F.; Ozbay, E.; Serebryannikov, A.E. Few-layer bifunctional metasurfaces enabling asymmetric and symmetric polarization-plane rotation at the subwavelength scale. Sci. Rep. 2024, 14, 13636. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Lee, W.S.; Kaltenecker, K.; Nirantar, S.; Withayachumnankul, W.; Walther, M.; Bhaskaran, M.; Fischer, B.M.; Sriram, S.; Fumeaux, C. Terahertz near-field imaging of dielectric resonators. Opt. Express 2017, 25, 3756–3764. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Liu, F.; Kwon, D.H.; Tretyakov, S. Reflectarrays and metasurface reflectors as diffraction gratings: A tutorial. IEEE Antennas Propag. Mag. 2023, 65, 21–32. [Google Scholar] [CrossRef] [Scilit]
- Maystre, D.R. Selected Papers on Diffraction Gratings; SPIE: Bellingham, WA, USA, 1993. [Google Scholar]
- Serebryannikov, A.E.; Alici, K.B.; Magath, T.; Cakmak, A.O.; Ozbay, E. Asymmetric Fabry-Perot-type transmission in photonic-crystal gratings with one-sided corrugations at a two-way coupling. Phys. Rev. A—At. Mol. Opt. Phys. 2012, 86, 053835. [Google Scholar] [CrossRef] [Scilit]
- Li, H.; Zhao, C.; Li, J.; Xu, H.; Xu, W.; Tan, Q.; Song, C.; Shen, Y.; Yao, J. Spin-dependent amplitude and phase modulation with multifold interferences via single-layer diatomic all-silicon metasurfaces. Opto-Electron. Sci. 2025, 4, 240025. [Google Scholar] [CrossRef] [Scilit]
- Wang, Z.; Pan, W.; He, Y.; Zhu, Z.; Jin, X.; Liu, M.; Ma, S.; He, Q.; Sun, S.; Zhou, L. Generation of vectorial terahertz beams using surface-wave excited metasurfaces. Opto-Electron. Sci. 2025, 4, 240024. [Google Scholar] [CrossRef] [Scilit]
- You, X.; Ako, R.T.; Sriram, S.; Withayachumnankul, W. 3D terahertz confocal imaging with chromatic metasurface. Laser Photonics Rev. 2025, 19, 2401011. [Google Scholar] [CrossRef] [Scilit]
- Zhang, Y.; Cao, X.; Yang, L.; Cai, B.; Wang, D.; Chen, F.; Luo, H.; Cheng, Y.; Li, X. Dynamic tunable terahertz broadband metasurface filter with independent frequency and amplitude modulations via vanadium dioxide and photosensitive silicon. Physica B Condens. Matter 2026, 729, 418423. [Google Scholar] [CrossRef] [Scilit]
- Alù, A.; Engheta, N. Optical ‘shorting wires’. Opt. Express 2007, 15, 13773–13782. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Serebryannikov, A.E.; Usik, P.V.; Ozbay, E. Non-ideal cloaking based on Fabry-Perot resonances in single-layer high-index dielectric shells. Opt. Express 2009, 17, 16869–16876. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Kwon, D.H.; Ptitcyn, G.; Díaz-Rubio, A.; Tretyakov, S.A. Transmission magnitude and phase control for polarization-preserving reflectionless metasurfaces. Phys. Rev. Appl. 2018, 9, 034005. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.








