---
title: 'Mie-tronics: Resonant Multipolar Nanophotonics'
url: https://www.emergentmind.com/topics/mie-tronics
type: topic
---

# Mie-tronics: Resonant Multipolar Nanophotonics

Mie-tronics, also called Mie-resonant metaphotonics or Mietronics, is the branch of subwavelength optics that exploits low-loss, high-index dielectric nanoparticles and their arrays to achieve strong, tunable resonant light–matter interactions via Mie (multipolar) modes; in a narrower TMDC formulation, it denotes exciton–Mie coupled nonlinear nanophotonics in which a Mie mode enhances the local pump field at $\omega$ and an excitonic resonance enhances $\chi^{(2)}(2\omega)$ or $\chi^{(3)}(\omega)$ [2409.13631][2105.04985]. The term is used across isolated nanoparticles, structured surfaces, Mie voids, surface-lattice resonances, nonlocal metasurfaces, and finite moiré arrays, with recurring emphasis on magnetic dipole and electric dipole resonances, anapoles, bound states in the continuum, Purcell enhancement, and multipolar interference as design primitives [2409.13631][2507.11995].

## 1. Definition and scope

Mie-tronics is defined in the broad literature as a branch of subwavelength optics that uses low-loss dielectric resonators rather than metallic plasmonic structures. In this usage, the central objects are isolated high-index dielectric subwavelength particles and metasurfaces whose optical response is organized by electric and magnetic multipoles of comparable strength, with design strategies based on magnetic dipole (MD), electric dipole (ED), anapole, Friedrich–Wintgen BIC, and supercavity modes [2409.13631]. Key advantages stated for this platform are low dissipation and compatibility with CMOS, as well as large near-field enhancements and strong Purcell effects without metal losses.

A narrower but influential use of the term arises in TMDC nonlinear nanophotonics, where Mie-tronics is described as an emerging paradigm in nano-optics that marries the low-loss, high-$Q$ dielectric Mie resonances of subwavelength architectures with the strong excitonic nonlinearities of transition-metal dichalcogenides such as MoS$_2$ and WS$_2$ [2105.04985]. In that formulation, the basic device concept is a TMDC nanoresonator engineered so that a Mie mode enhances the local field at the pump frequency while an excitonic resonance enhances the nonlinear susceptibility at the generated frequency.

The scope has subsequently expanded to inverted-index cavities (“Mie voids”), collective Mie exciton-polaritons in atomically thin semiconductors, passive “nonreciprocal Mie-surfaces,” and finite-array or moiré supermodes in nonlocal metasurfaces [2601.19420][2007.15313][2605.04549][2511.03560]. This suggests that the term now functions less as a label for a single device class than as a unifying multipolar framework for resonant light localization, nonlinear conversion, emission control, sensing, and symmetry-engineered wave manipulation.

## 2. Electromagnetic foundations

The canonical starting point is the Mie solution for a sphere of radius $R$ and relative refractive index $m \equiv n_{\rm particle}/n_{\rm medium}$. Expanding the fields in vector spherical harmonics gives the electric and magnetic scattering coefficients
$$
a_n  =  \frac{m \psi_n(m x) \psi_n'(x)  -  \psi_n(x) \psi_n'(m x)}
{ m \psi_n(m x) \xi_n'(x)  -  \xi_n(x) \psi_n'(m x)},
$$
$$
b_n  =  \frac{\psi_n(m x) \psi_n'(x)  -  m \psi_n(x) \psi_n'(m x)}
{ \psi_n(m x) \xi_n'(x)  -  m \xi_n(x) \psi_n'(m x)},
$$
with $x \equiv kR = 2\pi R/\lambda$, $\psi_n(z)=z j_n(z)$, and $\xi_n(z)=z h_n^{(1)}(z)$. The total scattering cross section is
$$
\sigma_{\rm scat}(\lambda)  =  \frac{2\pi}{k^2}  \sum_{n=1}^\infty  (2n+1)  (|a_n|^2 + |b_n|^2).
$$
For subwavelength particles, the dominant contributions are usually the lowest multipoles, especially $n=1$ and $n=2$, corresponding to ED/MD and quadrupolar channels [2409.13631].

High-index dielectric particles support strong MD and ED modes at $x<1$. In the overview of isolated dielectric resonators, the first MD resonance occurs roughly when $\mathrm{Re}[b_1]$ peaks, typically at $x \approx 0.8$–$1$ for $m \approx 3$–$4$, while the first ED occurs at $x \approx 1.2$–$1.5$ [2409.13631]. In the TMDC-disk formulation, the magnetic-dipole resonance is associated with the $\ell=1$, $b_1$ coefficient peaking when the denominator is minimized, approximately when $m kR \approx \pi$, and the internal field enhancement is summarized by
$$
L(\omega) \equiv |E_{\rm loc}(\omega)/E_0(\omega)| \simeq Q/V_m^{1/2},
$$
with
$$
V_m = \frac{\int \epsilon(r)|E(r)|^2 dV}{\max[\epsilon|E|^2]}.
$$
For TMDC disks of $R \sim 275$ nm and height $H=110$ nm, FDTD yields $Q \approx 10$–$20$ and $V_m \lesssim 0.02$ $\mu{\rm m}^3$, giving $L \sim 20$–$30$ [2105.04985].

An inverted version of the same multipolar physics appears in Mie voids, where an air cavity is embedded in a high-index medium. For a spherical air void of radius $a$ in silicon, the size parameter becomes $x \equiv n_h k_0 a$ and $m \equiv n_{\rm void}/n_h = 1/n_h$, but resonances still occur when the denominators of $a_\ell$ or $b_\ell$ are minimized [2601.19420]. In that setting, the excitation enhancement and emission modification are expressed through
$$
F_{\rm ex}(\omega_{\rm exc}) =
\frac{\int_V |E(r,\omega_{\rm exc})|^2 dV}{\int_V |E_0(r,\omega_{\rm exc})|^2 dV},
$$
$$
F_p(r_0,\omega_{\rm em}) \equiv \gamma_{\rm tot}/\gamma_0,
$$
and
$$
F_{qe}(r_0) = F_{\rm ex}(r_0,\omega_{\rm exc}) \cdot F_q(r_0,\omega_{\rm em}),
$$
which separate local-field enhancement from LDOS-mediated quantum-yield control [2601.19420]. In sensing-oriented void platforms, the same resonant picture is recast through the refractive-index sensitivity
$$
S=\frac{\Delta \lambda}{\Delta n},
$$
and the detection limit
$$
\Delta n_{\min}\approx\frac{\lambda_0}{2\,Q\,S\,\mathrm{SNR}},
$$
emphasizing that a well-defined sensing volume and full access to the modal field are part of the device concept, not merely by-products of fabrication [2407.02331].

## 3. Resonant mechanisms, interference states, and collective modes

A major part of Mie-tronic design consists of engineering interference among multipolar radiation channels. The anapole is the canonical example: in Cartesian multipole language, an anapole occurs when the far-field contributions of the Cartesian electric dipole and the dominant type-I toroidal dipole destructively interfere,
$$
p^{(c,e)} + (i k_0 \epsilon_d/c) T^{(1)}_{(c,e)} = 0,
$$
so that the spherical coefficient $a_1 \to 0$ and $\sigma_{\rm scat}$ dips sharply even though energy is trapped inside the particle [2605.04549]. In an amorphous-Si hemisphere, finite-element simulations show that this anapole appears only for backward illumination, not for forward illumination, and the backward ED scattering cross section plunges to near zero at $\lambda \approx 863$ nm.

When such hemispheres are repeated in a square lattice, the single-particle anapole governs a direction-dependent reflectance feature. The reported reflection isolation ratio,
$$
R\text{-iso}(\lambda)=10\log_{10}[R_f(\lambda)/R_b(\lambda)],
$$
reaches 45–50 dB in air at normal incidence near the anapole, while $T$-iso remains approximately $0$ dB because $T_f=T_b$ in the lossless-air configuration [2605.04549]. The paper explicitly states that the individual materials are Lorentz reciprocal, but the current nonreciprocity is due to interference. A recurring misconception is therefore addressed directly in the literature: the structure is called a “Nonreciprocal Mie-surface,” yet the mechanism is passive asymmetric Mie scattering associated with the anapole, rather than magneto-optic nonreciprocity or nonlinearity.

Collective resonances in periodic arrays are another central mechanism. In Si nanoparticle arrays, coherent coupling via Rayleigh anomalies produces narrow dispersive Mie surface-lattice resonances (Mie-SLRs), including an e-SLR associated with in-plane electric dipoles and an m-SLR associated with out-of-plane magnetic dipoles [2007.15313]. More generally, Mie-tronics extends classical single-particle scattering to finite arrays through long-range multipole–multipole couplings, allowing Fabry–Pérot, whispering-gallery, band-edge, and bound-state-in-continuum concepts to be treated in a unified multipolar language [2507.11995].

Recent work on finite arrays and nonlocal metasurfaces makes this unification explicit. In a 17$\times$17 hole array, a supercavity tuned by photonic-crystal mirrors reaches $F_P \gtrsim 500$ and $Q \sim 10^4$, while twisted hexagonal moiré arrays exhibit multiple “photonic magic angles” with $Q_{\max} \approx 5\times10^4$ at $\theta_M \approx 1.6^\circ$ and $F_P > 10^3$ [2507.11995]. In a related finite-array analysis, symmetry breaking in T-shaped metasurfaces enhances light trapping by strengthening in-plane nonlocal coupling pathways, and the reported trend is that finite arrays can show $Q$-factor enhancement driven by redistributed radiation channels, reversing the trend predicted by infinite-lattice theory [2511.03560]. This is significant because it replaces the common assumption that symmetry breaking necessarily degrades confinement with a finite-system criterion based on channel redistribution.

## 4. Hybrid light–matter, nonlinear, and optomagnetic regimes

In TMDC nanoresonators, nonlinear Mie-tronics is formulated through the second-harmonic polarization
$$
P_i(2\omega) = \epsilon_0 \sum_{j,k} \chi_i^{jk}(2\omega;\omega,\omega) E_j(\omega) E_k(\omega),
$$
with a resonant excitonic susceptibility
$$
\chi^{(2)}(2\omega) = \frac{\chi_0}{1 - (2\omega/\omega_{\rm exc})^2 - i (2\omega/\omega_{\rm exc})/Q_{\rm exc}}.
$$
The far-field SH field depends on the overlap integral between the nonlinear polarization and the second-harmonic mode, and the SHG enhancement is written as
$$
\eta_{\rm SHG} \equiv I_{2\omega}^{\rm disk}/I_{2\omega}^{\rm mono}
\simeq |L(\omega)|^4 |\chi^{(2)}(2\omega)|^2 (V_{\rm eff}/V_{\rm ref})^2.
$$
Experimentally, MoS$_2$ nanodisks show $\eta_{\rm SHG}\approx 23$ when the MD resonance at 900 nm overlaps the C-exciton at 450 nm, and $\eta_{\rm SHG}\approx 5$ when only the MD mode at 800 nm overlaps the exciton tail [2105.04985]. SHG rotational anisotropy further yields a sixfold pattern $I_{2\omega}(\phi)\propto \sin^2(3\phi)$, with near-zero isotropic background, consistent with preservation of in-plane crystal orientation and SH generation from the top $\sim 10$ nm layer.

Strong-coupling Mie-tronics appears in atomically thin semiconductors coupled to dielectric arrays. In monolayer WS$_2$ integrated with a poly-Si nanodisk lattice, the exciton and Mie-SLR are modeled by a non-Hermitian $2\times2$ coupled-oscillator Hamiltonian, and the polariton branches obey
$$
\omega_\pm(k_\parallel)=\frac{1}{2}[\omega_{\rm SLR}(k_\parallel)+\omega_{\rm ex}]
\pm \frac{1}{2}\sqrt{[\omega_{\rm SLR}(k_\parallel)-\omega_{\rm ex}]^2+4g^2}.
$$
At zero detuning the dispersion of the electric Mie-SLR exhibits a clear anti-crossing with a Rabi splitting $\Omega_R=32$ meV, whereas the magnetic Mie-SLR nearly crosses the exciton band because its field is dominated by out-of-plane components that do not efficiently couple with the in-plane excitonic dipoles of monolayer WS$_2$ [2007.15313]. The reported e-SLR linewidth $\Gamma_{e\text{-SLR}}\approx 18$ meV gives $Q\approx 120$, which is central to the room-temperature strong-coupling interpretation.

Optomagnetic Mie-tronics uses the inverse Faraday effect (IFE) at Mie resonances. In a magnetic dielectric sphere, the IFE field is
$$
H^{\rm IFE}(r,t)  = - \frac{g}{16\pi M_s}\,\mathrm{Im}\{E(r,t)\times E^*(r,t)\},
$$
and the internal optical field is expanded in vector spherical harmonics with coefficients $c_q(\lambda)$ and $d_q(\lambda)$ [2404.04569]. Distinct resonance orders generate distinct $H^{\rm IFE}$ patterns: a 400 nm-diameter Bi-iron-garnet sphere supports MD at $\sim 1107$ nm, ED at $\sim 907$ nm, MQ at $\sim 735$ nm, EQ at $\sim 603$ nm, MO at $\sim 552$ nm, and EO at $\sim 505$ nm. By tuning wavelength, one switches among one, two, or four localized IFE hotspots, and the transverse components form vortex patterns that the paper identifies as natural seeds for magnetic skyrmions. Reported operating conditions include pump intensities of order 1 mJ/cm$^2$ in 100 fs pulses, yielding local $H^{\rm IFE}\sim 10$–100 Oe inside hotspots [2404.04569].

## 5. Materials platforms, fabrication routes, and inverse design

The experimental platforms grouped under Mie-tronics are diverse, but they remain tied to a common multipolar vocabulary. The following examples summarize representative systems and reported figures of merit.

| Platform | Resonant mechanism | Reported result |
|---|---|---|
| MoS$_2$ nanodisk | MD resonance at 900 nm overlapping the C-exciton at 450 nm | $\eta_{\rm SHG}\approx 23$ [2105.04985] |
| WS$_2$ on Si nanoparticle array | e-SLR strong coupling with the A-exciton | $\Omega_R=32$ meV; $Q\approx120$ [2007.15313] |
| Silicon Mie void | Air-defined cavity with independently tunable excitation and quantum-yield enhancement | Maximum $F_p \approx 2.6$ near $R \simeq 430$ nm at $\lambda_{\rm em} \simeq 520$ nm [2601.19420] |
| Single Mie void sensor | Resonant confinement of light in air with full access to the modal field | $S\approx 400$–$520$ nm/RIU; $\Delta n_{\min}\approx 6.9\times10^{-4}$ in 850 aL [2407.02331] |
| CsPbBr$_3$ nanocube | Single-particle Mie-resonant all-dielectric nanolaser | $\lambda_{\rm las}=535$ nm; $\Delta E\approx 3.5$ meV [1905.08646] |
| Sb$_2$S$_3$ array | Phase-change tuning of visible Mie resonances | $\Delta\lambda=127$ nm; $Q\approx30$ [2107.07081] |

Fabrication routes are correspondingly heterogeneous. Single-crystal MoS$_2$ flakes of thickness 110 nm are exfoliated onto Si/SiO$_2$ substrates, patterned by electron-beam lithography with negative resist ARN 7520.18, and etched in SF$_6$ to form nanodisks with diameters 300–550 nm [2105.04985]. The WS$_2$ strong-coupling platform uses a square lattice of poly-Si nanodisks with pitch $P=420$ nm, height $h=90$ nm, and diameter $D=130\pm4$ nm defined by e-beam lithography and reactive-ion etching on fused silica and capped by a PDMS superstrate [2007.15313]. Silicon Mie-void arrays are fabricated by focused-ion-beam milling into a 200 nm-thick amorphous silicon film on glass, with a 30 kV beam and 100 pA current, while sensing-oriented Mie voids are described as cylindrical cavities in high-index hosts produced by e-beam or deep-UV lithography followed by dry etching [2601.19420][2407.02331]. Visible phase-change Mie-tronics in Sb$_2$S$_3$ uses RF sputtering, EBL, Pd hard masks, ICP-RIE, and Si$_3$N$_4$ encapsulation, with crystallization by hot-plate anneal at 300 $^\circ$C for 1 h in Ar and amorphization by a 780 nm, 100 fs, 80 MHz raster-scanned laser [2107.07081]. Single-particle nanolasing in CsPbBr$_3$ uses monocrystalline nanocubes produced by solution conversion from PbBr$_2$ to CsPbBr$_3$ on sapphire, followed by dark-field scattering and femtosecond-pump lasing measurements [1905.08646].

A separate design direction replaces brute-force geometry sweeps with machine learning. In TiO$_2$ meta-atoms of fixed height $H=320$ nm, Li et al. constructed a dataset of $N=36\,300$ distinct (shape, $\lambda$) combinations using COMSOL and trained three models: a forward prediction model mapping a 128$\times$128 shape mask and wavelength to six multipolar scattering channels, an inverse design model implemented as a tandem network, and an electric-field prediction model reconstructing the complex 3D field on three orthogonal slices [2305.18589]. The reported performance metrics are a validation $\mathrm{MSE}_{\rm Loss}\approx 6.5\times10^{-4}$ $\mu{\rm m}^2$ for the forward model, an average multipolar reconstruction error of $\approx 7\times10^{-3}$ $\mu{\rm m}^2$ for the inverse design model, and field error $<5\%$ RMS for the electric-field predictor. This suggests that multipolar objectives such as strong magnetic octupole response or super-scattering can be posed as inverse problems directly in the Mie-tronic basis rather than indirectly through geometric heuristics.

## 6. Devices, interpretive issues, and future directions

The device landscape is unusually broad because the same multipolar framework supports distinct observables. Reported implementations include on-chip nonlinear frequency converters, active metasurfaces, and quantum light emitters in TMDC nanoresonators [2105.04985]; passive linear nonreciprocal photonic devices based on asymmetric Mie scattering and one-sided anapoles [2605.04549]; spin-wave sources and nanoscopic skyrmion writers driven by vortex-like IFE fields [2404.04569]; programmable, high-density multimodal displays based on silicon Mie voids [2601.19420]; attoliter refractive-index sensors with defined sensing volumes [2407.02331]; and single-particle, all-dielectric nanolasers operating at room temperature [1905.08646].

Two interpretive issues recur in the literature. The first concerns reciprocity. The “Nonreciprocal Mie-surface” exhibits forward/backward differences in reflection and transmission that can differ by tens of decibels, but the paper also states that the individual materials are Lorentz reciprocal, and the current nonreciprocity is due to interference [2605.04549]. The second concerns symmetry breaking. Infinite-lattice intuition often associates symmetry breaking with weaker confinement, yet finite-array Mie-tronics reports the opposite trend: controlled symmetry breaking can enhance light trapping by strengthening in-plane nonlocal coupling pathways and redistributing radiation channels [2511.03560]. These points do not negate the device proposals, but they sharpen the physical meaning of the terminology.

Application-specific performance figures further show how the field spans different operating regimes. Silicon Mie voids were used to encode a bimodal 48$\times$48-void pattern with 0.8 $\mu$m-pitch pixels that reveals the “EPFL” logo in the bright field and the “SJTU” logo in both dark field and photoluminescence micrographs, with potential pixel densities $\gtrsim 10^6$ dpi [2601.19420]. Sb$_2$S$_3$ color pixels with pitch $\sim 280$ nm correspond to a potential 90 000 dpi and reversible tuning over a resonance shift of 127 nm, with reported endurance of more than 10 full amorphous–crystalline cycles [2107.07081]. In sensing, single voids reach well-defined volumes down to approximately 100 attoliters, sensitivities of approximately 400–520 nm per refractive index unit, and detection of refractive-index changes as small as $6.9\times10^{-4}$ in a defined volume of 850 attoliters [2407.02331]. In active emission, the smallest reported non-plasmonic single-mode nanolaser is a 420 nm CsPbBr$_3$ cube operating at 535 nm with linewidth approximately 3.5 meV [1905.08646].

The main open directions are also consistent across subfields. Broad reviews emphasize advanced nanofabrication for sub-10 nm features and low-loss high-$\epsilon$ materials, active control via phase-change materials, 2D materials, and gain media, non-Hermitian physics and exceptional points, machine-learning–assisted inverse design, and array-scale engineering of flatbands and topological edge modes [2409.13631]. Platform-specific outlooks add loss reduction at $2\omega$ in bulk TMDCs, electrical or optical control of exciton linewidth and oscillator strength, extension to higher-order nonlinearities and multi-exciton processes, scalable wafer-level fabrication of TMDC nanoresonator arrays with sub-10-nm precision, BICs in void arrays for ultra-high $Q$, and hybrid integration with 2D emitters such as WSe$_2$ [2105.04985][2407.02331][2601.19420]. A plausible implication is that future Mie-tronic systems will be judged less by a single resonance label and more by how effectively they combine local multipolar enhancement, collective interference, and material functionality within a quantitatively controlled open photonic environment.

Source: https://www.emergentmind.com/topics/mie-tronics