## Abstract

We experimentally studied the coupling between a double split ring resonator and a complementary split ring resonator. The greatest coupling occurs when the two resonators are separated by the average ring radius, and the dimensionless coupling is as large as 0.1, allowing a novel planar metamaterial based on this hybrid structure. The coupling strength can be varied up to a factor of 2 by changing the relative orientation of the split ring resonators. A 2×2 waveguide structure with −10 dB coupling factor can be achieved, and showing multi-mode plasmon-induced transparency. It can be considered one-dimensional metamaterials exhibiting negative permeability and permittivity simultaneously.

© 2017 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

## 1. Introduction

Recently artificial structures combining building blocks exhibiting electric and magnetic resonances emerge for realizing negative index material. [1,2] Among them, the split-ring resonators (SRR) and their counterpart, the complimentary split-ring resonators (CSRR) respectively providing negative permeability and permittivity within a narrow frequency window are widely used. [3–6] These resonators therefore enrich the engineering on planar microwave propagation [7], leading to miniature microwave circuits. [8] In addition to these first generation of meta-materials [5], a new class of meta-materials based on strongly coupled resonator emerges owing to their rich physics. As an example, it is found that strongly coupled resonators exhibit the interesting phenomena analogous to electromagnetically-induced transparency (EIT). [9] In the so-called plasmon-induced transparency(PIT), coherent interference between two strongly coupled resonators is enforced to induce transparency. [10–14] With the occurrence of PIT, a slow light behavior presents, promising vast applications. [15–17]

It has been proposed that many intriguing phenomena analogous to those in electronic systems such as Bloch oscillations [18], artificial gauge field [19,20] and nonlinear effects [21] can be realized in photonics by tailoring the photon hopping in coupled resonators. Recently, people have demonstrated artificial magnetic field for the hopping photon in two-dimension [22, 23], which leads to photonic topological insulators. [24–26] The importance of coupled SRRs can be envisaged in the viewing of the versatile optical properties empowered by the metamaterials constructed by strongly coupled resonators. Over the years intensive studies have explored how the coupling is governed by parameters such as spatial arrangement and orientation configurations. [27–29] Though much efforts have been put on this issue, unfortunately, previous works are mostly devoted to the same type of resonators. Indeed, “exotic” electronic behaviors typically occur in a lattice with diatomic basis, so the study of a “photonic diatomic basis” is mandatory. [30] Perhaps a structure combining a resonator and its complementary counterpart gets off to a good start because of their identical frequencies, mirror images in shape, and perfectly dual electromagnetic properties. As an example, a SRR/CSRR diatomic basis should provoke negative permeability and negative permittivity at the same frequency. We also like to note that previous experiments usually investigate a small quantity of samples, so very limited information on the coupling strength can be obtained. Here, by taking the advantage of a home-made moving platform, we in-situ change the relative position and orientation of a double split ring resonator (DSRR) and a CSRR to map out the shift of their resonances. In addition, hybrid DSRR/CSRR structure with strong inter-resonator coupling exhibit a multi-mode PIT effect and an efficient band-pass wave guiding.

## 2. Experimental methods

Coupled DSRRs (sample A), which is similar to common PIT device[14] were fabricated on 0.5 mm-thick glass substrate using standard photolithography and wet-etching of thermally evaporated 300 nm-thick Al film. The etching solution was diluted phorspher acid (phorspher acid: de-ionized water=4:1). The outer diameter of the DSRRs is 2*r* = 7.6 mm and the average ring radius *r _{a}* = 3.2 mm. The inter-resonator (center-to-center) distance

*d*varies from 7.7 to 8.2 mm. For probing the resonance frequency, a microstrip couples one of the DSRRs, which is referred as the “radiative” one. Fig. 1(a) illustrates an optical micrograph of a sample with “face-to-back” configuration. Here we refer the principle direction pointing the gap of the outer ring as the “face” side.

Other samples(B and C) were made on 1.6 mm-thick ordinary FR-4 printed circuit board(PCB), on which both sides covered by 36 *μ*m-thick Cu. A home-made moving platform was built for continuously probing the inter-coupling between a DSRR and a CSRR (sample B). Fig. 1(b) schematically illustrates the platform setup, which combines two layers of PCB: On the upper layer, a “radiative” DSRR was coupled to a microstrip for probing resonances. A CSRR was made on the lower layer, which could be moved by a 3-axis stage with the position resolution of 0.1 mm. The CSRR could be “darken” when it was moved away from the microstrip. The two PCBs were stacked in resemblance a structure with a DSRR and a CSRR on the two sides of a single PCB. Both the DSRR and the CSRR have the same 2*r* = 7.6 mm and the average ring radius *r _{a}* = 3.2 mm. The position of the CSRR can be described by a coordinate system, of which the origin is chosen at the center of DSRR. The 2.8 mm-wide microstrip is parallel to the

*x*direction, centered at

*y*= −5.4 mm.

To demonstrate the wave-guiding property of coupled resonators, we fabricated 3 samples(C1–C3), respectively employing short DSRR-, CSRR- and interleaved SRR-arrays to couple two microstrip transmission lines(TL). In C samples, center-to-center distance for DSRR (or CSRR) *d* = 10 mm and 2*r* = 7.6 mm. All samples were measured by a vector network analyzer(VNA, Agilent N5230A). For the measurement of *S*-parameters of sample C, ports not measured were terminated with a broadband 50 Ω load. Microwave simulations were performed by means of High Frequency Structure Simulator.

## 3. Results and discussions

#### 3.1. Coupled DSRRs

The microwave transmission amplitude of sample A’s with the face-to-face configuration are illustrated in Fig. 1(d). When the inter-resonator distance *d* is 8.2 mm, there is only one resonance at *f*_{0} =3.0 GHz, which is close to the resonance frequency of a single DSRR. As *d* is progressively reduced, the single resonance dip splits and the splitting increases, exhibiting a typical PIT spectrum. By noting the lower and the higher resonant frequency respectively as *f _{L}* and

*f*, the largest splitting

_{H}*f*−

_{H}*f*= 200 MHz when

_{L}*d*is the smallest,

*d*= 7.7 mm(gap=0.1 mm between resonators). Figure 1(e) shows how the splitting

*f*−

_{H}*f*changes on

_{L}*d*. In general, the splitting reduces rapidly as

*d*increases, though it depends on the relative orientation of the DSRRs. When

*d*< 8.2 mm, the coupling is larger when the two DSRRs are face-to-face than face-to-back.

When two resonators are close, they are capacitively and inductively coupled as a dimer. It has been found that the global excitation modes present a lower resonant frequency(*f _{L}*) for symmetric mode and higher frequency(

*f*) for anti-symmetric mode, and the splitting of resonance reflects the coupling strength. The splitting can be explained by the coupled-mode theory [31],

_{H}*κ*is the coupling coefficient. Here Ω = 2

*πf*

_{L(H)}, and

*ω*(

_{j}*j*= 1, 2) are intrinsic resonator frequencies. When

*ω*

_{1}=

*ω*

_{2}= 2

*πf*

_{0}, the result becomes

*g*= |

*κ*|/

*ω*

_{1}is the dimensionless coupling strength. In sample A with the face-to-face configuration, the largest dimensionless coupling strength achievable is 2

*g*= (

*f*−

_{H}*f*)/

_{L}*f*

_{0}= 6.6%. When the two DSRRs are very close, the coupling strength for the face-to-back one is reduced by a factor of 2. Ideally, the circuit analysis reveals that the face side can be modeled as a capacitor, while the back-side can be viewed as an inductor. [32] Previous works have pointed out that the face-to-face configuration provides a large capacitive coupling, while the coupling for face-to-back one is rather small. [33] However, when the DSRRs are separated as far as 8.2 mm, the face-to-back configuration surpass the face-to-face one, probably due to parallel electromagnetic polarization in the SRRs.

#### 3.2. Coupling between a DSRR and a CSRR

Next we discuss the hybrid dimer formed by a CSRR and a DSRR by presenting the data of sample B. Figure 2(a) illustrates the measured microwave transmission amplitude of sample B for various CSRR positions. When the SRRs are exactly aligned(*x* =*y*=0 mm), two resonances respectively at 2.92 GHz and 3.27 GHz can be found. The lower one is attributed to CSRR because it only appears when the CSRR is in proximity of the microstrip. This can be further confirmed that we only see the higher resonance when replacing the lower layer by a metal ground. Although having the same size, the SRRs do not produce the exact resonance frequencies because the CSRR is overlaid by PCBs from both sides.

The DSRR and CSRR resonance frequencies *f _{D}* and

*f*are mapped out in the range of 20.0 mm × 20.0 mm with a pitch of 1.0 mm. As illustrated in Fig. 2(b),

_{C}*f*displays the rotational symmetry and its minimum appears at

_{D}*d*= 3.0 mm. Figure 2(c) illustrates the cross section of

*f*and

_{D}*f*mappings at

_{C}*y*= 0 mm, and reveals two important findings: When DSRR and CSRR are separated by a distance of

*d*= 3.0 mm, both resonance frequencies red-shifts mostly. When they are widely separated(

*d*> 7.0 mm), or perfectly aligned(

*d*= 0 mm), their frequencies are largest. Later we will discuss how the coupling strength between the SRRs affect the red-shifts. In regarding the movement in

*y*-direction,

*f*red-shifts when CSRR moves toward the microstrip, that is confirmed by an experiment using a upper layer without a DSRR. The red-shift in

_{C}*f*owing to the impact of microstrip provides us a guidance how to strongly couple a CSRR to a microstrip, and is useful in designing sample C.

_{C}The red-shifted resonant frequencies go beyond the prediction of Eq. (1), which always describes the repulsion of the resonant frequencies when coupling presents. Nevertheless we can analyze this problem using an effective circuit model as that described in Appendix A. From the observation that both resonance frequencies reduce as the resonators are in proximity, we expect that the inter-coupling contains capacitance. One can easily judge the capacitance(C)-coupling strength by using the result that

*are intrinsic and shifted resonator frequencies, and*

_{j}*δω*= Ω

_{j}*−*

_{j}*ω*.

_{j}*C*and

_{j}*L*are capacitance and inductance of the resonator.

_{j}*α*=

_{j}*C*/

_{p}*C*, in which ${C}_{p}^{-1}={C}_{1}^{-1}+{C}_{2}^{-1}+{C}_{x}^{-1}~{C}_{x}^{-1}$. For extracting the mutual inductance(M)-coupling strengths, we can use the following expression by noting that

_{j}*β*=

_{j}*M*/

*L*,

_{j}*γ*≡

*C*

_{1}/

*C*

_{2}=

*α*

_{2}/

*α*

_{1}, and ${\beta}_{1}=\gamma {\beta}_{2}{\omega}_{1}^{2}/{\omega}_{2}^{2}$. Equation (3) provides a possible way to determine

*β*

_{2}if

*γ*is known,

*α*’s and

*β*’s can also be found by numerical methods.

The dimensionless coupling strengths for C-coupling and M-coupling can be respectively defined by 2*g _{C}* = (

*α*

_{1}+

*α*

_{2})/2 and $2{g}_{M}=\sqrt{{\beta}_{1}{\beta}_{2}}$. As one can see in Eq. (2), when

*α*’s and

*β*’s are not large, the sum of the red-shifts is linearly proportional to 2

*g*. From above analysis, we conclude that the largest coupling in DSRR/CSRR hybrid dimer occurs at

_{C}*d*= 3.0 mm ∼

*r*, and the corresponding 2

_{a}*g*values are 2.0%. 2

_{C}*g*is also exactly calculated by assuming

_{M}*γ*= 0.2. Their position dependences are illustrated in Fig. 2(d). We note that experimental 2

*g*is lessen by the impact of the lower PCB and metal stage in our platform. Simulations reveal that in an ideal single PCB case, a larger coupling 2

_{C}*g*= 7.7% is recovered

_{C}Here we try to discuss why the optimum coupling occurs at *d* ∼ *r _{a}*. When

*d*≫

*r*, the DSRR and CSRR can be respectively viewed as a single magnetic dipole and electric dipole. By assuming that the two dipoles always oscillates in-phase to reduce the energy, we found that the coupling energy reduces as

*d*

^{−2}when

*d*is smaller than wavelength 2

*πc*/

*ω*. In the case

*d*is much larger than the wavelength, the expression for radiation zone should be imposed and

*U*∝

*d*

^{−1}.

When *d* < ∼ *r*, the structure of the resonators are important in determining the coupling energy. According to coupled-mode theory, the coupling energy between the normal modes of two structures separated by *d⃗* can be estimated by the following integral on total space,

*dτ*is the volume element.

*E⃗*and

_{j}*B⃗*are the mode electric and magnetic fields produced by resonator

_{j}*j*(= 1, 2).

*r⃗*are the position vectors of the element

_{j}*dτ*from the resonator centers, and

*d⃗*=

*r⃗*

_{2}−

*r⃗*

_{1}. Ideally, two complementary structures produce dual electromagnetic fields. When their distance is small

*r⃗*

_{2}∼

*r⃗*

_{1},

*E*(or

*B*) fields generated by two structures are always perpendicular to each other, so the coupling energy becomes minimal. Perhaps by taking the two trends into account, we should expect the maximum

*U*and maximal coupling occurs at

*d*∼

*r*∼

*r*.

_{a}#### 3.3. Orientation dependence

From the coupled DSRR case, we see that 2*g* for face-to-face configuration is roughly twice than that for face-to-back one. In light of this, we expect that the orientation of the SRRs may have a strong impact to the coupling strength. For systematical study, we parametrize the orientation alignment using two angles *θ _{D}* and

*θ*, representing the angles between their principal (or “face”) directions to the center-to-center vector as illustrated in Fig. 3(a). We performed experiments using our movable system with selected orientation configurations and all showed that the condition

_{C}*d*∼

*r*gives the largest frequency shift in

_{a}*f*and

_{D}*f*. For convenience, we attribute the frequency results at

_{C}*d*= 4.0 mm and

*d*> 11.0 mm as the fully shifted (Ω

*) and original frequencies(*

_{j}*ω*) for the dimer.

_{j}Figures 3(b) and 3(c) shows the deduced 2*g _{C}* and 2

*g*for 10 major configurations using movable-stage experiment, and simulations for 1-layer and 2-layer PCB settings. Again, 2

_{M}*g*values obtained for the 3 methods follow the order that 1-layer value > 2-layer value > experiment value. Yet the data points agree well to each other if the later two are re-scaled by a common factor. In general, when

_{C}*θ*is fixed, the coupling strength is smaller when

_{D}*θ*∼

_{C}*π*. In addition, the maximal value can be as 2 times as the minimal value. Similar factor was found in the coupled DSRR case as mentioned before. If looking C-coupling strength only, we found the following order

*g*(

_{C}*π*, 0) >

*g*(0, 0) =

_{C}*g*(

_{C}*π*,

*π*) >

*g*(0,

_{C}*π*).

The above observation reveals that when *θ _{D}* =

*π*/2, the coupling strength can vary greatly as

*θ*changes. For a close look of

_{C}*θ*dependence when

_{C}*θ*=

_{D}*π*/2, simulations with 1-layer PCB setting were performed. Presented as a polar plot in Fig. 3(d), the coupling strength 2

*g*shows a maximum value at

_{C}*θ*= 0 and minimal one at

_{C}*θ*=

_{C}*π*. The data points can be roughly described by a function form of 2

*g*= 0.088 + 0.014 cos

_{C}*θ*as marked by the red curve. The maximum and minimum of 2

_{C}*g*occur at

_{M}*θ*= 3

_{C}*π*/4 and 5

*π*/4, respectively. Though the 2

*g*data cannot be described by a simple function.

_{M}#### 3.4. One-dimensional array based on SRRs

Taking the advantage of strong coupling between resonators, we can build signal couplers between two TLs using hybrid DSRR/CSRR structure. Based on our orientation study, we chose the orientation (*π*/2, *π*/2) owing to the large coupling and good symmetry. Figures 4(a) and 4(b) respectively show the *S*-parameters *S*_{21}, *S*_{31} and *S*_{41} of the coupler using 2 coupled DSRRs (C1) and 3 CSRRs (C2). Their coupling factor are very poor, except for C2, the coupling factors is ∼ −40 dB at the first resonance at 3.05 GHz but increases to −20 dB at second resonance 6.40 GHz. Unlike C1 and C2, coupled DSRR/CSRR (C3) provides a large coupling factor up to −10 dB in a wider frequency range from 2.6 − 3.3 GHz as illustrated in Fig. 4(c). At the second resonance of 6.0 GHz, it also maintains a −12 dB coupling factor. Furthermore, the coupling is symmetric in regarding to port 3 and 4. Intriguingly one can see the transmission coefficient shown in Fig. 5(a) and 5(b) exhibit a structure splits into 4 resonances dips, featuring the multi-mode PIT spectrum.

To understand the wave guiding property, we consider the following Lagrangian for a one-dimensional(1D) chain of *N* SRRs. Without loss of generality, pure C-coupling is under consideration,

*q*’s and

_{j}*q*

_{x,j}’s are charge variables of SRRs and coupling capacitors. By noting that all DSRR(

*j*=even) and all CSRR(

*j*=odd) are identical and using the notation of ${\alpha}_{1}={C}_{p}{C}_{1}^{-1}$ and ${\alpha}_{2}={C}_{p}{C}_{2}^{-1}$, one gets

Here *q _{j}* (

*t*) =

*q̃*. In degenerate case

_{j}e^{iωt}*ω*

_{1}=

*ω*

_{2}, the 1D system forms a two-band structure with a dispersion relation,

*α*

_{1},

*α*

_{2}≪ 1, each subband has a bandwidth about min (

*α*

_{1},

*α*

_{2})

*ω*

_{1}, and centered at $\sqrt{1-{\alpha}_{1}-{\alpha}_{2}\pm \left({\alpha}_{1}^{2}+{\alpha}_{2}^{2}\right)}{\omega}_{1}$.

The dips in Fig. 5(a) are roughly equal spaced by 150 MHz with a span of 584 MHz and centered at 3.03 GHz, showing a large coupling strength of *α*_{1} ∼ *α*_{2} ∼ 2*g _{C}* ∼ 10% between DSRR and CSRR. The experiment results clearly demonstrate that a hybrid structure with interleaved DSRRs and CSRRs allows in a high efficient band-pass waveguiding. In comparison to the arrays composed of single type of resonators which can only have one negativity, the hybrid structure has simultaneous negative permeability and permittivity. In addition, The DSRR/CSRR unit automatically forms a diatomic basis and paves the way to the photonic band structure with a phonon-like dispersion. We leave some of the properties for future study.

## 4. Conclusion

In conclusion, we have demonstrated an efficient planar wave-guiding using hybrid structure of interleaved DSRRs and CSRRs at microwave frequency. It provides a −10 dB coupling between two TLs and a bandwidth of about 600 MHz. To understand the origin of the strong coupling between a DSRR and CSRR, we systematically investigated how the resonance frequencies of the two resonators change with their separation, and found that the coupling is greatest when *d* ∼ *r _{a}*. We also studied the coupling strengths when the two resonators are in various orientation, and found that a factor-of-2 reduction can be experimentally observed. Compared to chains built with single type of resonators, the overlapping structure yields a greater coupling between the resonators and greatly enlarges the transmission amplitude and bandwidth. When used in constructing large lattice structures, the DSRR/CSRR basis enables the engineering of the photonic band structure and could find vast applications. By utilizing state-of-the-art micro fabrication, one can reduce the resonator size and increase the operation frequency upto THz and optical one.

## A. Circuit model for coupled SRRs

To analyze the observed shift in resonances, we used a model circuit as shown in Fig. 1(c). Each resonator is modeled by a LC circuit and they are coupled by a capacitance *C _{x}* and a mutual inductance

*M*. The system Lagrangian reads,

*q*

_{1}and

*q*

_{2}are charge variables producing current passing through inductor

*L*

_{1}and

*L*

_{2}, respectively.

*q*is charge accumulate on the coupling capacitor

_{x}*C*. The equations of motions for

_{x}*q*

_{1}and

*q*

_{2}are

Here we used ${q}_{x}\approx {C}_{p}\left({C}_{1}^{-1}{q}_{1}-{C}_{2}^{-1}{q}_{2}\right)$ and ${C}_{p}^{-1}\equiv {C}_{1}^{-1}+{C}_{2}^{-1}+{C}_{x}^{-1}$. ${\omega}_{j}=1/\sqrt{{L}_{j}{C}_{j}}$ are the intrinsic resonance frequencies of the two SRRs. In the last equation, notations ${\alpha}_{j}\equiv {C}_{p}{C}_{j}^{-1}$ and ${\beta}_{j}\equiv M{L}_{j}^{-1}$ are used. The eigen-frequency Ω satisfies,

For degenerate SRRs *ω*_{1} = *ω*_{2}, the resonances are

*α*’s and

*β*’s are omitted since the coupling strengths are usually much smaller than 1. One can clear see the frequency shift in pure capacitance(C)-coupling and mutual inductance(M)-coupling cases are different. For the former, Ω =

*ω*

_{1}and ${\omega}_{1}\sqrt{1-{\alpha}_{1}-{\alpha}_{2}}$, giving a repulsion of resonance ΔΩ = Ω

_{1}− Ω

_{2}∼

*ω*

_{1}(

*α*

_{1}+

*α*

_{2})/2 = 2

*ω*

_{1}

*g*. For the later, $\mathrm{\Omega}={\omega}_{1}\sqrt{1\pm \sqrt{{\beta}_{1}{\beta}_{2}}}$, and repulsion of resonance $\mathrm{\Delta}\mathrm{\Omega}~{\omega}_{1}\sqrt{{\beta}_{1}{\beta}_{2}}=2{\omega}_{1}{g}_{M}$. If the SRR are identical, namely

_{C}*α*

_{1}=

*α*

_{2}and

*β*

_{1}=

*β*

_{2}, the result can be simplified further as

For non-degenerate case, a simple expression can be used to extract the C-coupling strengths only,

*δω*are not large,

_{j}*g*.

_{C}## Funding

Ministry of Science and Technology, Taiwan (102-2628-M-005-001-MY4, 106-2112-M-005-007).

## Acknowledgments

We are grateful to the National Center for High-performance Computing for computer time and facilities. Fruitful discussions with C. C. Huang and C. S. Wu are acknowledged. This work is financially supported by the Ministry of Science and Technology, Taiwan under grant Nos. 102-2628-M-005-001-MY4 and 106-2112-M-005-007, and Research Center for Sustainable Energy and Nanotechnology, NCHU.

## Disclosures

The authors declare that there are no conflicts of interest related to this article.

## References and links

**1. **D. R. Smith, J. B. Pendry, and M. C. K. Wiltshire, “Metamaterials and negative refractive index,” Science **305**, 788–792 (2004). [CrossRef] [PubMed]

**2. **V. M. Shalaev, “Optical negative-index metamaterials,” Nature Photon. **1**, 41–48 (2007). [CrossRef]

**3. **R. Shelby, D. Smith, and S. Schultz, “Experimental verification of a negative index of refraction,” Science **292**, 77–79 (2001). [CrossRef] [PubMed]

**4. **S. Linden, C. Enkrich, M. Wegener, J. Zhou, T. Koschny, and C. M. Soukoulis, “Magnetic response of metamaterials at 100 terahertz,” Science **306**, 1351–1353 (2004). [CrossRef] [PubMed]

**5. **C. Caloz and T. Itoh, *Electromagnetic metamaterials: transmission line theory and microwave applications* (John Wiley & Sons, 2005).

**6. **F. Falcone, T. Lopetegi, M. A. G. Laso, J. D. Baena, J. Bonache, M. Beruete, R. Marques, F. Martin, and M. Sorolla, “Babinet principle applied to the design of metasurfaces and metamaterials,” Phys. Rev. Lett. **93**, 197401 (2004). [CrossRef] [PubMed]

**7. **F. Martin, J. Bonache, F. Falcone, M. Sorolla, and R. Marques, “Split ring resonator-based left-handed coplanar waveguide,” Appl. Phys. Lett. **83**, 4652–4654 (2003). [CrossRef]

**8. **M. Gil, J. Bonache, and F. Martin, “Metamaterial filters: A review,” Metamaterials **2**, 186–197 (2008). [CrossRef]

**9. **M. Fleischhauer, A. Imamoglu, and J. P. Marangos, “Electromagnetically induced transparency: Optics in coherent media,” Rev. Mod. Phys. **77**, 633–673 (2005). [CrossRef]

**10. **Q. Xu, S. Sandhu, M. L. Povinelli, J. Shakya, S. Fan, and M. Lipson, “Experimental realization of an on-chip all-optical analogue to electromagnetically induced transparency,” Phys. Rev. Lett. **96**, 123901 (2006). [CrossRef] [PubMed]

**11. **M. Parvinnezhad Hokmabadi, E. Philip, E. Rivera, P. Kung, and S. M. Kim, “Plasmon-induced transparency by hybridizing concentric-twisted double split ring resonators,” Sci. Rep. **5**, 15735 (2015). [CrossRef] [PubMed]

**12. **S. Han, L. Cong, H. Lin, B. Xiao, H. Yang, and R. Singh, “Tunable electromagnetically induced transparency in coupled three-dimensional split-ring-resonator metamaterials,” Sci. Rep. **6**, 20801 (2016). [CrossRef] [PubMed]

**13. **J. Wu, B. Jin, J. Wan, L. Liang, Y. Zhang, T. Jia, C. Cao, L. Kang, W. Xu, and J. Chen, “Superconducting terahertz metamaterials mimicking electromagnetically induced transparency,” Appl. Phys. Lett. **99**, 161113 (2011). [CrossRef]

**14. **N. Papasimakis, Y. H. Fu, V. A. Fedotov, S. L. Prosvirnin, D. P. Tsai, and N. I. Zheludev, “Metamaterial with polarization and direction insensitive resonant transmission response mimicking electromagnetically induced transparency,” Appl. Phys. Lett. **94**, 211902 (2009). [CrossRef]

**15. **K. Totsuka, N. Kobayashi, and M. Tomita, “Slow light in coupled-resonator-induced transparency,” Phys. Rev. Lett. **98**, 213904 (2007). [CrossRef] [PubMed]

**16. **F. Miyamaru, H. Morita, Y. Nishiyama, T. Nishida, T. Nakanishi, M. Kitano, and M. W. Takeda, “Ultrafast optical control of group delay of narrow-band terahertz waves,” Sci. Rep. **4**, 4346 (2014). [CrossRef] [PubMed]

**17. **C. Wu, A. B. Khanikaev, and G. Shvets, “Broadband slow light metamaterial based on a double-continuum fano resonance,” Phys. Rev. Lett. **106**, 107403 (2011). [CrossRef] [PubMed]

**18. **S. Alexander and N. Stefan, “Discrete optics in femtosecond-laser-written photonic structures,” J. Phys. B At. Mol. Opt. Phys. **43**, 163001 (2010). [CrossRef]

**19. **R. O. Umucalilar and I. Carusotto, “Artificial gauge field for photons in coupled cavity arrays,” Phys. Rev. A **84**, 043804 (2011). [CrossRef]

**20. **K. Fang, Z. Yu, and S. Fan, “Realizing effective magnetic field for photons by controlling the phase of dynamic modulation,” Nature Photon. **6**, 782–787 (2012). [CrossRef]

**21. **I. Carusotto and C. Ciuti, “Quantum fluids of light,” Rev. Mod. Phys. **85**, 299–366 (2013). [CrossRef]

**22. **M. Hafezi, E. A. Demler, M. D. Lukin, and J. M. Taylor, “Robust optical delay lines with topological protection,” Nature Phys. **7**, 907–912 (2011). [CrossRef]

**23. **M. C. Rechtsman, J. M. Zeuner, A. Tunnermann, S. Nolte, M. Segev, and A. Szameit, “Strain-induced pseudomagnetic field and photonic landau levels in dielectric structures,” Nature Photon. **7**, 153–158 (2013). [CrossRef]

**24. **R. O. Umucalilar and I. Carusotto, “Fractional quantum hall states of photons in an array of dissipative coupled cavities,” Phys. Rev. Lett. **108**, 206809 (2012). [CrossRef] [PubMed]

**25. **M. C. Rechtsman, J. M. Zeuner, Y. Plotnik, Y. Lumer, D. Podolsky, F. Dreisow, S. Nolte, M. Segev, and A. Szameit, “Photonic floquet topological insulators,” Nature **496**, 196–200 (2013). [CrossRef] [PubMed]

**26. **M. Hafezi, S. Mittal, J. Fan, A. Migdall, and J. M. Taylor, “Imaging topological edge states in silicon photonics,” Nature Photon. **7**, 1001–1005 (2013). [CrossRef]

**27. **F. Hesmer, E. Tatartschuk, O. Zhuromskyy, A. A. Radkovskaya, M. Shamonin, T. Hao, C. J. Stevens, G. Faulkner, D. J. Edwards, and E. Shamonina, “Coupling mechanisms for split ring resonators: Theory and experiment,” Phys. Status Solidi B **244**, 1170–1175 (2007). [CrossRef]

**28. **A. Radkovskaya, O. Sydoruk, E. Tatartschuk, N. Gneiding, C. J. Stevens, D. J. Edwards, and E. Shamonina, “Dimer and polymer metamaterials with alternating electric and magnetic coupling,” Phys. Rev. B **84**, 125121 (2011). [CrossRef]

**29. **E. Tatartschuk, N. Gneiding, F. Hesmer, A. Radkovskaya, and E. Shamonina, “Mapping inter-element coupling in metamaterials: Scaling down to infrared,” J. Appl. Phys. **111**, 094904 (2012). [CrossRef]

**30. **O. Sydoruk, O. Zhuromskyy, E. Shamonina, and L. Solymar, “Phonon-like dispersion curves of magnetoinductive waves,” Appl. Phys. Lett. **87**, 072501 (2005). [CrossRef]

**31. **H. A. Haus and W. Huang, “Coupled-mode theory,” Proc. IEEE. **79**, 1505–1518 (1991). [CrossRef]

**32. **J. D. Baena, J. Bonache, F. Martin, R. M. Sillero, F. Falcone, T. Lopetegi, M. A. G. Laso, J. Garcia-Garcia, I. Gil, M. F. Portillo, and M. Sorolla, “Equivalent-circuit models for split-ring resonators and complementary split-ring resonators coupled to planar transmission lines,” IEEE Trans. Microwave Theory Tech. **53**, 1451–1461 (2005). [CrossRef]

**33. **R. S. Penciu, K. Aydin, M. Kafesaki, T. Koschny, E. Ozbay, E. N. Economou, and C. M. Soukoulis, “Multi-gap individual and coupled split-ring resonator structures,” Opt. Express **16**, 18131–18144 (2008). [CrossRef] [PubMed]