Metallic Spin-Type Photonic Topological Insulator

• Metallic spin-type photonic topological insulators are defined by engineered pseudospin and magneto-electric coupling that opens a robust topological bandgap.
• The methodology employs 3D split-ring resonator lattices to hybridize electric and magnetic modes, yielding spin-filtered edge or surface states.
• Experimental tests demonstrate robust transport with over 20 dB bulk suppression and Dirac-like surface modes within a wide gap exceeding 25% bandwidth.

Searching arXiv for the core paper and closely related photonic spin-type/topological-insulator works to ground the article in the current literature. {"query":"arXiv metallic spin-type photonic topological insulator split-ring resonators three-dimensional photonic topological insulator", "max_results": 10} {"query":"(Yang et al., 2018) three-dimensional photonic topological insulator split-ring resonators", "max_results": 10} A metallic spin-type photonic topological insulator is a photonic topological phase realized in a metallic metamaterial or metallic waveguide platform, in which the relevant internal degree of freedom is not electron spin but a photonic pseudospin built from coupled electric and magnetic modes. In this class of systems, magneto-electric coupling or effective bi-anisotropy plays the role of a spin-orbit interaction, opening a topological gap and producing spin-filtered edge or surface transport. The most direct three-dimensional realization is a lattice of metallic split-ring resonators (SRRs) whose resonance-enhanced magneto-electric coupling generates a complete 3D topological bandgap and Dirac-like surface states, with the phase identified as a photonic “weak TI” (Yang et al., 2018).

1. Terminology and conceptual scope

The expression combines three distinct descriptors. It is metallic because the underlying photonic platform is built from metallic elements such as SRRs, metallic cylinders, or metallic plates, rather than from an ordinary dielectric photonic crystal. It is spin-type because the topological organization is carried by an effective pseudospin degree of freedom: depending on the platform, this pseudospin is encoded by left- and right-circular polarization, by TE/TM hybridization, or by phase-locked electric-magnetic superpositions. It is a photonic topological insulator because the bulk is gapped while interfaces support gapless or nearly gapless boundary transport protected by topology and symmetry (He et al., 2014).

This terminology is used across several related but not identical implementations. In the bi-anisotropic meta-waveguide, the platform is an all-metal parallel-plate waveguide loaded with a hexagonal array of metallic cylinders, and the topology is generated by TE-TM hybridization induced by asymmetric cylinder-plate gaps (Ma et al., 2014). In the uniaxial metacrystal waveguide, non-resonant meta-atoms are confined between metallic plates, and first-order waveguide modes generate an effective magnetoelectric coupling that yields a nontrivial $\mathbb{Z}_2$ phase with spin-filtered edge transport (Chen et al., 2014). In the 3D SRR metamaterial, the metallic character is literal and the pseudospin arises from in-phase and out-of-phase electric-magnetic dipole combinations, producing a three-dimensional weak topological insulator for photons (Yang et al., 2018).

A recurring source of confusion is the word “spin.” Photons do not carry the electronic spin degree of freedom relevant to fermionic topological insulators. In these systems, “spin” denotes a pseudospin built from electromagnetic mode structure. A second source of ambiguity is the word “metallic.” In most of the canonical spin-type photonic topological insulators, the bulk is still insulating in the topological sense, while the boundary is gapless. By contrast, a distinct line of work studies time-reversal-symmetric photonic topological metals, in which antihelical edge states coexist with bulk states and the phase is metallic in the band-structure sense (Xie et al., 2023).

2. Electromagnetic pseudospin and the photonic analogue of spin-orbit coupling

The central design problem is to create a pair of degenerate photonic channels and then gap them in a topologically nontrivial way. In the two-dimensional photonic analogue of the quantum spin Hall effect, the pseudospins are identified with left-circular polarization and right-circular polarization,

$$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$

and strong magneto-electric coupling supplies the pseudo-spin-orbit term that opens the bulk gap (He et al., 2014). The same work makes a sharper symmetry claim: the edge-state protection is attributed not to the ordinary bosonic time-reversal operator $T_b=\tau_z K$, for which $T_b^2=+1$, but to an effective fermionic-like operator

$T_f=i\tau_y K, \qquad T_f^2=-1,$

under which the two circular polarizations form a Kramers-like pair (He et al., 2014).

In metallic waveguide implementations, the same logic is expressed in a TE/TM basis. The bi-anisotropic meta-waveguide begins from frequency-degenerate TE and TM Dirac cones at $K$ and $K'$, then introduces mirror-symmetry breaking by opening a gap between each cylinder and one metallic plate. This asymmetry induces bi-anisotropy, mixes TE and TM polarizations, and adds an effective mass term of the form

$\hat s_z \hat \tau_z \hat \sigma_z,$

which the authors identify as the photonic analogue of spin-orbit coupling in an electronic topological insulator (Ma et al., 2014). The resulting effective spin is a synthetic spin encoded in the phase-locked TE/TM hybridization.

The uniaxial metacrystal waveguide uses a related but distinct construction. There, first-order waveguide modes confined between metallic plates generate an effective magnetoelectric tensor with nonzero elements

$$\xi_{e,12}=\xi_{e,21}= \frac{i m \pi}{\omega d}, \qquad m\neq 0,$$

and the pseudo-spin variables are defined as

$$\mathbf{p}^{\pm} = \sqrt{\rho \varepsilon_0}\,\mathbf{e} \pm \sqrt{\mu_0}\,\mathbf{h}.$$

These two sectors are related by time-reversal symmetry and behave as opposite-spin copies of a quantum spin Hall system (Chen et al., 2014).

A later reduced-symmetry metallic platform on a rhombic lattice uses the same electromagnetic principle in a dual-surface form. There the pseudospin channels are

$$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$0

built from hybridized out-of-plane electric and magnetic fields. The authors explicitly interpret the coupling $$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$1 between the two metallic layers as the photonic analogue of Kane–Mele spin-orbit coupling (Davis et al., 26 Jul 2025).

3. Three-dimensional realization with metallic split-ring resonators

The three-dimensional SRR system is the clearest realization of a metallic spin-type photonic topological insulator in the literal sense of all three words. The structure is a 3D metallic metamaterial crystal built from split-ring resonators that support both electric and magnetic dipolar responses. The crystal is arranged on a triangular lattice in the $$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$2-$$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$3 plane and stacked along $$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$4. The starting point is a six-SRR unit cell with mirror symmetry

$$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$5

denoted $$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$6. In that symmetric configuration, the SRRs are arranged back-to-back, so the bianisotropy cancels at the $$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$7 and $$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$8 points, yielding frequency-isolated 3D Dirac points described as doubly-degenerate Weyl points with a four-fold degeneracy near the band crossings (Yang et al., 2018).

The topological-insulator phase is obtained by breaking that mirror symmetry. Concretely, the top three SRRs of the original six-SRR cell are removed and the $$\psi_{LCP}= \begin{pmatrix} 1\ i \end{pmatrix}, \qquad \psi_{RCP}= \begin{pmatrix} 1\ -i \end{pmatrix},$$9-periodicity is adjusted. This destroys the back-to-back cancellation and allows the electric and magnetic dipole modes, which in the symmetric crystal have opposite parity and remain decoupled for in-plane propagation at

$T_b=\tau_z K$0

to hybridize through a bianisotropic magneto-electric coupling. The resulting hybridized eigenmodes have electric and magnetic components either in phase or out of phase, and these two combinations are identified as pseudo-spin-up and pseudo-spin-down states (Yang et al., 2018).

The topological mechanism is therefore a photonic version of Dirac-point gapping by a spin-orbit mass term. The parent phase is a 3D Dirac-semimetal-like photonic crystal with symmetry-protected Dirac points. Breaking $T_b=\tau_z K$1 introduces resonance-enhanced magneto-electric coupling, which gaps those points. The realized large-gap structure is stated to be continuously deformable to the perturbative regime in which an infinitesimal gap is opened by weak mirror-symmetry breaking. This adiabatic continuity is the basis for assigning the realized wide-gap phase the same nontrivial topology as the Dirac-point-gapping mechanism (Yang et al., 2018).

The resulting phase is explicitly identified as a weak TI, namely a stack of 2D photonic quantum spin Hall layers with interlayer coupling. In that interpretation, the surface spectrum contains an even number of Dirac cones rather than the odd number associated with a strong TI. The complete 3D topological bandgap exceeds 25% bandwidth, which the paper describes as extremely wide for a topological photonic gap (Yang et al., 2018).

4. Topological indices, bulk–boundary correspondence, and surface spectra

The topological characterization of metallic spin-type photonic topological insulators varies with dimension and symmetry class, but a common structure recurs: opposite pseudospin sectors carry opposite topological charge, while the full system remains time-reversal symmetric or reciprocity-preserving.

In the two-dimensional photonic quantum spin Hall analogue, the low-energy Hamiltonian maps to the BHZ model, and the phase is characterized by a nontrivial $T_b=\tau_z K$2 invariant. With inversion symmetry, the invariant is computed from parity eigenvalues at the four time-reversal-invariant momenta through

$T_b=\tau_z K$3

and the reported result is

$T_b=\tau_z K$4

showing that the phase is topologically nontrivial rather than merely edge-guiding in an accidental sense (He et al., 2014).

The uniaxial metacrystal waveguide likewise combines a $T_b=\tau_z K$5 description with a spin-Chern description. The nontrivial phase has

$T_b=\tau_z K$6

while the trivial photonic ordinary insulator has

$T_b=\tau_z K$7

Within the nontrivial region, the spin Chern number can be

$T_b=\tau_z K$8

so the system can remain $T_b=\tau_z K$9-nontrivial while reversing the propagation direction of the spin-filtered edge state (Chen et al., 2014).

For the 3D SRR metamaterial, the main-text characterization is the weak-TI interpretation. The bulk gap binds two-dimensional gapless surface states on an interface or domain wall where the effective topological mass changes sign, and those states take the form of massless Dirac cones. Near the surface Dirac point, the measured dispersion is conical and consistent with

$T_b^2=+1$0

The paper reports one observed Dirac point near the projection of $T_b^2=+1$1 and another near $T_b^2=+1$2, consistent with the even-number-of-cones structure of a weak TI (Yang et al., 2018).

Later metallic platforms extend the diagnostic toolkit. The rhombic-lattice metallic spin-type PTI uses spin-projected Berry curvature and non-Abelian Wilson-loop spectra to demonstrate that the reduced-symmetry system remains in the same $T_b^2=+1$3-nontrivial class as the ordinary Kane–Mele quantum spin Hall phase; for representative tight-binding parameters the Wilson-loop branches wind oppositely and together span the full Brillouin-zone phase range, indicating

$T_b^2=+1$4

even though the high hexagonal symmetry is absent (Davis et al., 26 Jul 2025).

5. Experimental signatures: bulk gaps, Dirac states, and robust transport

The 3D SRR realization was experimentally established through direct measurements of both the gapped bulk and the Dirac-like surface states. In the bulk transmission experiment, the measured transmittance drops by about 20 dB over approximately 4.3–6.0 GHz, directly evidencing a 3D bulk bandgap. Along an internal domain wall, however, transmission remains high across the same frequency window. Fourier-transformed field scans map the projected bulk dispersion and confirm a complete bulk bandgap, while surface-state measurements reveal a cone-like linear dispersion with a surface Dirac point at $T_b^2=+1$5 along the $T_b^2=+1$6-$T_b^2=+1$7 line of the surface Brillouin zone. The surface modes are tightly localized to the domain wall with a penetration depth of about 10.3 mm (Yang et al., 2018).

That work also provided a specifically three-dimensional robustness test. The authors built a sharply twisted non-planar internal domain wall with two $T_b^2=+1$8 corners. Surface waves launched into this wall propagated around the bends with transmission comparable to that along a straight wall of the same length, and a full 3D field map at 4.7 GHz showed the wave confined to the surface along the twisted route. The same behavior was observed after decomposing measured fields into several $T_b^2=+1$9 components,

$T_f=i\tau_y K, \qquad T_f^2=-1,$0

which is why the experiment was interpreted as a genuine 2D topological mode living on a 3D surface rather than a planar edge mode embedded in a leaky environment (Yang et al., 2018).

Two-dimensional metallic platforms established analogous robustness in earlier settings. In the bi-anisotropic meta-waveguide, domain walls between regions of opposite bi-anisotropy support edge modes that survive spatial disorder in bi-anisotropy strength, cavity defects, and zigzag interfaces with $T_f=i\tau_y K, \qquad T_f^2=-1,$1 turns. Quantitatively, in the cavity test the trivial defect mode exceeds $T_f=i\tau_y K, \qquad T_f^2=-1,$2 transmission over only $T_f=i\tau_y K, \qquad T_f^2=-1,$3 of the gap frequency range, whereas the topological edge mode exceeds $T_f=i\tau_y K, \qquad T_f^2=-1,$4 transmission over 100\% of the gap frequency range. In the bend test the defect mode exceeds $T_f=i\tau_y K, \qquad T_f^2=-1,$5 transmission over only $T_f=i\tau_y K, \qquad T_f^2=-1,$6 of the gap range, while the topological edge mode does so over $T_f=i\tau_y K, \qquad T_f^2=-1,$7 (Ma et al., 2014).

The uniaxial metacrystal waveguide added direct phase-sensitive evidence for spin filtering. At the PTI–POI interface, a rightward mode excited from the left showed a nearly constant phase plateau

$T_f=i\tau_y K, \qquad T_f^2=-1,$8

across approximately 2.68–2.92 GHz, while the leftward mode measured after swapping source and detector showed

$T_f=i\tau_y K, \qquad T_f^2=-1,$9

The $K$0 contrast between these plateaus was interpreted as direct evidence that the two propagation directions correspond to opposite pseudo-spin sectors; transmission remained high even after replacing five unit cells near the edge with a local obstacle (Chen et al., 2014).

6. Extensions, variations, and conceptual boundaries

The metallic spin-type photonic topological insulator has diversified in several directions without abandoning its central ingredients. One direction lowers crystalline symmetry. The rhombic-lattice implementation replaces the customary hexagonal unit cell by a reduced-symmetry rhombus formed by dual metallic electromagnetic surfaces, preserving spin-topological transport while enabling straight-line interfaces that are easier to integrate into microwave systems than the zig-zag interfaces of hexagonal layouts (Davis et al., 26 Jul 2025). This suggests that high hexagonal symmetry is not always essential, provided electromagnetic duality and bianisotropic hybridization are maintained.

A second direction studies loss and non-Hermiticity rather than only Hermitian band topology. In the $K$1-symmetric reciprocal parallel-plate waveguide, Maxwell’s equations decompose into two conserved pseudospin sectors, and the Hamiltonian is described as an infinite direct sum of Kane–Mele type Hamiltonians with a common band gap. The spin-Chern phases survive material dissipation up to a critical loss level, with a topological phase transition at

$K$2

Because of particle-hole symmetry, the gap charge becomes the ill-defined alternating series

$K$3

and the physical interpretation offered is that the common band gap hosts an infinite number of pseudospin-polarized edge states, one for each modal sector (Câmara et al., 2023).

A third direction clarifies the conceptual boundary of the term “metallic.” In the two-dimensional photonic topological metal, the phase is not insulating in the band-structure sense. Instead, it is a time-reversal-symmetric topological metal with antihelical edge states, characterized by opposite spin-Chern numbers,

$K$4

and controlled by a metal-insulator phase transition at

$K$5

This phase shares the spin-resolved boundary transport of spin-type PTIs but differs from the insulating metallic-platform realizations because edge states coexist with bulk states rather than traversing a full bulk gap (Xie et al., 2023).

A final distinction concerns relation to other three-dimensional bosonic topological phases. The 3D photonic crystal with a single surface Dirac cone is a bosonic topological crystalline insulator protected by glide reflection and enabled by time-reversal breaking; its surface Dirac-point degeneracy is protected by a nonsymmorphic glide, not by the pseudospin constructions typical of metallic spin-type PTIs (Lu et al., 2015). This contrast is useful: not every 3D photonic topological insulator with Dirac surface states is spin-type in the SRR or TE/TM-hybridization sense.

Taken together, these developments indicate a coherent research program rather than a single architecture. Across metallic SRR crystals, metallic waveguides, dual electromagnetic surfaces, and non-Hermitian impedance-guided systems, the defining pattern remains the same: an engineered photonic pseudospin, a magneto-electric or duality-mediated coupling that acts as a spin-orbit term, a topological bulk phase, and boundary transport whose direction is locked to that pseudospin (Yang et al., 2018).