Polarization-Selective Photon-Magnon Coupling

• Polarization-selective photon-magnon coupling is a mechanism where light and spin interactions depend on polarization, mode symmetry, and spatial overlap.
• It spans diverse platforms such as cavity magnonics, optomagnonics, and planar microwave circuits, revealing bright and dark mode behaviors.
• Selective hybridization through engineered selection rules and overlap integrals offers promising applications in photonic isolation, signal transduction, and polarization filtering.

Polarization-selective photon-magnon coupling (PMC) denotes hybrid light-spin dynamics in which the photon-magnon matrix element depends on optical or microwave polarization, mode symmetry, propagation geometry, or the polarization character of the magnon itself. In the literature, PMC appears in several non-equivalent forms: cross-polarized optomagnonic three-wave mixing in magnetic dielectrics, bright/dark antiferromagnetic magnon superpositions in cavities, electric-field coupling to electric-active magnons in polar antiferromagnets, orientation-controlled switching between orthogonal microwave photon modes, and superconductivity-enhanced selection between antiferromagnetic branches (Bittencourt et al., 2021, Yuan et al., 2017, Maurya et al., 6 May 2026).

1. Physical origin and conceptual scope

PMC is not a single microscopic mechanism. In cavity magnonics based on yttrium iron garnet, the interaction is the magnetic dipole term, effectively governed by the overlap between the cavity microwave magnetic field and the transverse dynamical magnetization. In the YIG/Pt cavity experiment, the relevant structure is the vector overlap $$g_{\mathrm{eff}} \propto \left|\int d^3 r\, \mathbf{m}_\perp(\mathbf{r}) \cdot \mathbf{h}_\perp(\mathbf{r})\right|$$, so only the component of the microwave field perpendicular to the equilibrium magnetization couples efficiently (Maier-Flaig et al., 2016). This already implies polarization selectivity, even when only one cavity mode is used.

In optical optomagnonics, the coupling is instead a three-particle interaction between one magnon and two photons mediated by the magneto-optical dielectric tensor. For a cubic magnetic solid with $\mathbf{M}_0 \parallel [001]$, the off-diagonal dielectric-tensor elements couple transverse magnetization to cross-polarized optical field components, and the resulting coupling coefficients $g_{mm'n}^{(\pm)}$ depend on overlap integrals $G_{44,mm'n}^{(\pm)}$ and $K_{mm'n}^{(\pm)}$ built from TE/TM mode profiles (Liu et al., 2016). In that setting, PMC is fundamentally a polarization-conversion problem.

A third route is magnetoelectric. In Fe$_2$Mo$_3$O$_8$, the relevant excitation is an electric-active magnon at $\omega_0/(2\pi) \approx 1.25\,\text{THz}$, and the dominant interaction is $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$, not $\mathbf{M}_0 \parallel [001]$0. The magnetic excitation possesses in-plane oscillation of electric polarization, proposed to be induced by the inverse Dzyaloshinskii-Moriya interaction and/or single-site anisotropy, so the coupling strength is determined by the polarization of the THz electric field in the $\mathbf{M}_0 \parallel [001]$1 plane (Shi et al., 2020).

2. Hamiltonian structure and mode descriptions

A general dispersive optomagnonic formulation was developed for magnetized epsilon-near-zero media, where the permittivity tensor is fully dispersive and linear in magnetization. In that framework, magnons are quantized fluctuations about a saturated magnetization $\mathbf{M}_0 \parallel [001]$2, and the interaction Hamiltonian for non-degenerate optical modes takes the three-wave form

$\mathbf{M}_0 \parallel [001]$3

with $\mathbf{M}_0 \parallel [001]$4 determined by both $\mathbf{M}_0 \parallel [001]$5, $\mathbf{M}_0 \parallel [001]$6, and the frequency dependence of optical polarizations $\mathbf{M}_0 \parallel [001]$7 (Bittencourt et al., 2021). For degenerate Voigt modes, the same theory yields

$\mathbf{M}_0 \parallel [001]$8

so the magnon mediates polarization conversion inside a degenerate optical manifold.

In antiferromagnetic cavity magnonics, the mode structure is different but the selection principle is analogous. For a bipartite antiferromagnet, the Bogoliubov magnons $\mathbf{M}_0 \parallel [001]$9 couple to a circularly polarized photon mode with equal matrix elements, giving the $g_{mm'n}^{(\pm)}$0 matrix

$g_{mm'n}^{(\pm)}$1

At $g_{mm'n}^{(\pm)}$2, the branches are degenerate and reorganize into a bright combination $g_{mm'n}^{(\pm)}$3 and a dark combination $g_{mm'n}^{(\pm)}$4; only the symmetric combination couples to the photon (Yuan et al., 2017). In this case, PMC is realized as symmetry-selected hybridization within a degenerate magnon doublet.

Planar microwave implementations generalize the same idea to multiple photon modes. In the planar cavity-magnonic system with an electric-LC resonator and a YIG film, the effective three-mode Hamiltonian contains two photon operators $g_{mm'n}^{(\pm)}$5, one magnon operator $g_{mm'n}^{(\pm)}$6, and couplings $g_{mm'n}^{(\pm)}$7 and $g_{mm'n}^{(\pm)}$8, with the waveguide-resonator orientation controlling which channel is active (Maurya et al., 6 May 2026).

3. Selection rules, dark modes, and polarization filtering

A defining feature of PMC is the existence of bright and dark channels enforced by symmetry or dispersion. In the dispersive ENZ theory, Faraday and Voigt geometries behave qualitatively differently. For Faraday modes, the overlap integrals vanish identically: $g_{mm'n}^{(\pm)}$9 and $G_{44,mm'n}^{(\pm)}$0, so the Kittel mode does not couple the circular Faraday modes at all, in either the degenerate or non-degenerate case. In Voigt geometry, by contrast, the $G_{44,mm'n}^{(\pm)}$1 modes are generically coupled, but the non-degenerate amplitude has exact zeros when

$G_{44,mm'n}^{(\pm)}$2

which coincide with the points where the Voigt $G_{44,mm'n}^{(\pm)}$3 mode becomes purely right-handed or purely left-handed circular. These are the “dispersion-tuned selection rules” of the ENZ model (Bittencourt et al., 2021).

In antiferromagnets, the canonical zero-field example is the bright/dark pair. At $G_{44,mm'n}^{(\pm)}$4, the two magnon branches are degenerate, but only the symmetric combination overlaps with the cavity polarization. The eigenfrequencies are

$G_{44,mm'n}^{(\pm)}$5

so the third eigenfrequency remains exactly at the bare magnon frequency and is completely uncoupled (Yuan et al., 2017). A common misconception is that both degenerate branches necessarily hybridize once they are frequency-matched to the cavity; in this model, polarization and sublattice symmetry prohibit that.

In optomagnonics, the dominant selection rule is cross-polarized scattering. For YIG with $G_{44,mm'n}^{(\pm)}$6, transitions between different polarizations are strongly favored, especially TE$G_{44,mm'n}^{(\pm)}$7TM, while TE$G_{44,mm'n}^{(\pm)}$8TE and TM$G_{44,mm'n}^{(\pm)}$9TM are much smaller (Liu et al., 2016). In the planar microwave ELCR platform, the same logic appears as orientation-controlled switching between orthogonal photon modes: at $K_{mm'n}^{(\pm)}$0, only mode-1 near $K_{mm'n}^{(\pm)}$1 GHz is bright; at $K_{mm'n}^{(\pm)}$2, mode-1 is dark and mode-2 near $K_{mm'n}^{(\pm)}$3 GHz is bright (Maurya et al., 6 May 2026).

4. Coupling regimes and quantitative scales

Dispersive ENZ media provide an especially explicit scaling law. For degenerate Voigt modes,

$K_{mm'n}^{(\pm)}$4

and near the ENZ frequency,

$K_{mm'n}^{(\pm)}$5

For YIG-inspired parameters $K_{mm'n}^{(\pm)}$6 kA/m, $K_{mm'n}^{(\pm)}$7, $K_{mm'n}^{(\pm)}$8, $K_{mm'n}^{(\pm)}$9, $_2$0, and $_2$1 THz, the estimate is $_2$2 GHz, comparable to typical GHz magnon frequencies (Bittencourt et al., 2021).

In the zero-field antiferromagnetic cavity model, the anticrossing gap is $_2$3, with $_2$4. Using $_2$5, $_2$6, $_2$7, $_2$8, $_2$9, and $_3$0, the gap is $_3$1. The maximum damping compatible with resolvable double peaks is $_3$2 in the example considered, whereas NiO can have $_3$3 (Yuan et al., 2017).

Electric-active THz PMC can also reach large normalized couplings. In Fe$_3$4Mo$_3$5O$_3$6, the split peaks at $_3$7 and $_3$8 THz around the $_3$9 THz magnon imply $_8$0 and $_8$1, which the work describes as an ultra-strong spin-photon coupling effect. The measured thickness dependence, $_8$2 for $_8$3, is close to the expected $_8$4 (Shi et al., 2020).

In the planar ELCR-YIG system, the extracted microwave couplings are explicitly angle dependent: $_8$5 at $_8$6, while $_8$7 at the same angles. The corresponding cooperativities are $_8$8 for mode-1 coupling at $_8$9, and $\omega_0/(2\pi) \approx 1.25\,\text{THz}$0 for mode-2 coupling at $\omega_0/(2\pi) \approx 1.25\,\text{THz}$1 (Maurya et al., 6 May 2026). Superconducting heterostructures extend the accessible scale further: S/AF/S systems were predicted to reach $\omega_0/(2\pi) \approx 1.25\,\text{THz}$2 GHz, exceeding $\omega_0/(2\pi) \approx 1.25\,\text{THz}$3 of the antiferromagnetic resonant frequency (Gordeeva et al., 28 Oct 2025).

5. Experimental manifestations and diagnostics

The standard signature of PMC is an avoided crossing in transmission or reflection, but the literature shows that this is not exhaustive. In the antiferromagnetic cavity model, strong coupling yields two hybridized transmission peaks near resonance, whereas the dark mode does not appear in transmission at all (Yuan et al., 2017). In the YIG/Pt cavity experiment, simultaneous microwave reflection and electrically detected spin pumping showed that the hybridized uniform mode can produce a weaker spin pumping signal than weakly coupled standing spin-wave modes, because the magnon component of the polariton is reduced in the strong-coupling regime (Maier-Flaig et al., 2016).

Time-domain THz work demonstrates an additional diagnostic. In Fe$\omega_0/(2\pi) \approx 1.25\,\text{THz}$4Mo$\omega_0/(2\pi) \approx 1.25\,\text{THz}$5O$\omega_0/(2\pi) \approx 1.25\,\text{THz}$6, the magnon-polariton splitting is visible as long-lived beating in the time trace and as two peaks in a delayed-window Fourier transform, but the splitting is absent in standard FTIR frequency-domain spectra, which show only a single absorption peak at $\omega_0/(2\pi) \approx 1.25\,\text{THz}$7 THz (Shi et al., 2020). This directly addresses the misconception that strong coupling must always be visible in ordinary frequency-domain absorption.

Spatiotemporal imaging in BiFeO$\omega_0/(2\pi) \approx 1.25\,\text{THz}$8 shows that hybridization can be inferred from propagation dynamics. The magnon-polariton associated with the $\omega_0/(2\pi) \approx 1.25\,\text{THz}$9 electromagnon appears around $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$0 and $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$1, with representative group velocity $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$2, lifetime $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$3, and propagation length $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$4. The corresponding pure magnon propagation length was estimated as $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$5 for $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$6 (Kainuma et al., 2023). In such systems, PMC is manifested not only as spectral splitting but also as a dramatic reconfiguration of transport.

6. Architectures, control knobs, and broader significance

Several control knobs recur across the field. In dispersive optomagnonics, operating near ENZ, choosing Voigt rather than Faraday geometry, and tuning to the zeros of $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$7 provide direct control over which polarization channel is bright or dark. The same framework extends to Fabry-Pérot cavities, magnetized plasmas, metal-gyroelectric stacks, and on-chip ENZ metamaterials (Bittencourt et al., 2021). In planar microwave circuits, the geometric rotation angle $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$8 acts as a polarization knob, with a measured transition near $$H_{\mathrm{int}} \sim - \mathbf{P}(t)\cdot \mathbf{E}_{\mathrm{THz}}(t)$$9 and a symmetry-related model-predicted transition near $\mathbf{M}_0 \parallel [001]$00 separating radiative-channel dominance between the two photon modes (Maurya et al., 6 May 2026).

Antiferromagnets supply a distinct form of selectivity because they naturally support multiple magnon branches with different symmetry. In the two-sublattice cavity model, zero field yields one bright and one dark mode, while finite field removes the degeneracy and activates both branches (Yuan et al., 2017). In superconducting S/AF/S heterostructures, the same principle was elevated to the THz ultrastrong-coupling regime: at zero magnetic field only the lower-frequency antiferromagnetic mode couples to the photon, while an applied field activates both modes, and the resulting magnon-polaritons attain group velocities amounting to several tenths of the speed of light (Gordeeva et al., 28 Oct 2025).

The broader significance of PMC is therefore not merely stronger hybridization. It is selective hybridization: one can choose which polarization, which magnon branch, which scattering direction, or which frequency-conversion pathway remains bright. This suggests applications in polarization-selective magnon readout, Stokes/anti-Stokes discrimination, photonic isolation, reconfigurable multimode hybridization, and microwave-to-THz or optical transduction. It also suggests a unifying view of apparently different platforms: ENZ optomagnonics, antiferromagnetic cavities, electromagnon polaritons, planar resonator circuits, and superconducting heterostructures all realize PMC by arranging that only specific photon and magnon polarizations have a nonzero matrix element.