Tunable Plasmon Interactions

• Tunable plasmon interactions are phenomena where in situ adjustments (e.g., polarization, gating, deformation) modulate collective charge oscillations.
• Control methods span localized hybridization, coupled-mode interference, and spatial reconfiguration, enabling shifts in resonance energy and Fano lineshapes.
• Applications include reconfigurable sensors, tunable optical filters, and dynamic plasmonic circuits, with challenges from damping and fabrication precision.

Tunable plasmon interactions are plasmon-coupling phenomena in which the strength, phase, symmetry, spectral position, spatial localization, or topology of collective charge oscillations is varied in situ by controllable parameters. In the recent literature, this scope includes localized and propagating plasmon coupling, plasmon–molecule Fano interference, plasmon–photon and plasmon–exciton polaritons, plasmon–phonon hybridization, hyperbolic and topological plasmon transitions, and phase-controlled coupled plasmons in anisotropic multilayers (Osley et al., 2012, Yanai et al., 2010, Xie et al., 2022, Giri et al., 15 Jun 2025).

1. Fundamental interaction classes

At the most basic level, tunable plasmon interactions arise in three recurrent settings. The first is localized hybridization, in which near fields couple two or more localized surface plasmon resonances (LSPRs), shifting bright bonding and antibonding branches, modifying hot-spot intensities, and creating angle-, gap-, or shape-dependent resonances. The second is plasmon coupling to external continua or discrete resonances, producing Fano lineshapes, avoided crossings, or polaritonic mode splitting. The third is collective propagation in periodic or anisotropic media, where Coulomb interactions, lattice harmonics, or conductivity anisotropy reshape dispersion, band topology, and polarization response (Sahu et al., 7 Apr 2025, Osley et al., 2012, Weick et al., 2014, Xie et al., 2022).

These classes share a common electrodynamic core. For propagating metal–dielectric surface waves, the reference dispersion is

$$k_{\mathrm{spp}} = k_0 \sqrt{\frac{\varepsilon_m(\omega)\varepsilon_d}{\varepsilon_m(\omega)+\varepsilon_d}},$$

while localized resonances in anisotropic particles or cavities are set by geometry-dependent depolarization factors through conditions such as

$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$

Tuning then acts by changing $\varepsilon(\omega)$, the effective mode volume, the momentum-matching condition, the dipolar interaction tensor, or the discrete–continuum detuning (Yanai et al., 2010, Arcangeli et al., 2018, Maksymov et al., 2017).

A second unifying feature is that tunability rarely targets only one observable. Depending on the platform, the same control parameter can simultaneously modify resonance energy, linewidth, oscillator strength, asymmetry parameter, and modal composition. This is explicit in two-continuum Fano systems, plasmon–exciton polaritons, and anisotropic hyperbolic materials, where a single knob can redistribute spectral weight among several interfering channels rather than merely shift a peak (Osley et al., 2012, Lee et al., 2016, Xie et al., 2022).

2. Control parameters and tuning modalities

Polarization, phase, and symmetry are among the most direct tuning knobs. In the mid-IR plasmonic metamolecule of an asymmetric cruciform aperture, rotating the incident polarization angle $\theta$ changes the excitation amplitudes of two orthogonal plasmonic continua as $T_{a,E}\propto \sin\theta$ and $T_{b,E}\propto \cos\theta$, which continuously tunes the Fano asymmetry of the PMMA carbonyl resonance. In periodically corrugated Ag plates, the lateral grating shift $\Delta x$ sets the Fourier-harmonic phase $\phi_m=2\pi m\Delta x/\Lambda$, thereby switching same-symmetry and opposite-symmetry mode couplings on and off; in the tunable-filter configuration, reflectivity changes from approximately $0.1$ to $0.9$, and the reflection dip shifts from $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$0 nm to $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$1 nm as $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$2 varies from $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$3 to $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$4. In all-metal pyramidal metasurfaces, varying the angle of incidence from $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$5 to $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$6 produces a double anticrossing among two LSPRs and an SPP, together with a strong field enhancement in a blue-shifted red-spectral LSPR branch (Osley et al., 2012, Yanai et al., 2010, Marques et al., 2024).

Electrical tuning acts either on the plasmon directly or on a coupled matter resonance. In nanostructured graphene, gating $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$7 tunes the localized plasmon frequency $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$8 and its third harmonic; for a $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$9 nm nanodisk in vacuum with $\varepsilon(\omega)$0 eV, the paper reports $\varepsilon(\omega)$1 eV and $\varepsilon(\omega)$2 eV, enabling resonant nonlinear driving of near-IR emitters. In a borophene–perovskite system, tuning the sheet density from $\varepsilon(\omega)$3 to $\varepsilon(\omega)$4 brings the borophene guiding plasmon into zero detuning with the perovskite exciton, yields a Rabi splitting of about $\varepsilon(\omega)$5 meV, and enables active reflective phase modulation over a $\varepsilon(\omega)$6 range. In graded InAs nanowires, a back gate shifts a localized plasmon resonance hotspot by about $\varepsilon(\omega)$7 nm over $\varepsilon(\omega)$8 V, with measured slopes of $\varepsilon(\omega)$9–$\theta$0 nm/V. In hybrid TMD plasmonic systems, gating can instead reweight excitonic channels: in MoS$\theta$1 gap-mode metasurfaces it tunes the exciton/trion balance with little change in exciton energy, whereas in WSe$\theta$2/gap-plasmon nanocavities a DC bias from $\theta$3 V to $\theta$4 V changes the nanogap, tunes the lower polariton by about $\theta$5 nm, and yields an ON/OFF intensity ratio up to $\theta$6 (Cox et al., 2019, Yan et al., 2022, Arcangeli et al., 2018, Ni et al., 2019, Darlington et al., 2023).

Mechanical deformation, nanogap variation, and thickness are equally powerful. In elastomer-embedded Ga droplet assemblies, stretching the substrate from $\theta$7 mm to $\theta$8 mm blue-shifts the bright bonding mode from $\theta$9 nm to $T_{a,E}\propto \sin\theta$0 nm; Lorentzian fits give an energy shift from $T_{a,E}\propto \sin\theta$1 eV to $T_{a,E}\propto \sin\theta$2 eV, while the quality factor decreases from $T_{a,E}\propto \sin\theta$3 to $T_{a,E}\propto \sin\theta$4. In liquid-metal droplets, capillary oscillations modulate shape and depolarization factors directly; for $T_{a,E}\propto \sin\theta$5 nm and the $T_{a,E}\propto \sin\theta$6 mode, the reported oscillation frequency ranges from about $T_{a,E}\propto \sin\theta$7 MHz to $T_{a,E}\propto \sin\theta$8 MHz as $T_{a,E}\propto \sin\theta$9 varies from $T_{b,E}\propto \cos\theta$0 to $T_{b,E}\propto \cos\theta$1, and resonant electrical drive requires roughly $T_{b,E}\propto \cos\theta$2–$T_{b,E}\propto \cos\theta$3 V across $T_{b,E}\propto \cos\theta$4–$T_{b,E}\propto \cos\theta$5. In thick aligned CNT films, film thickness and resonator length jointly set the plasmon frequency through the empirical fit $T_{b,E}\propto \cos\theta$6 with $T_{b,E}\propto \cos\theta$7; experimentally the resonance is tunable over $T_{b,E}\propto \cos\theta$8–$T_{b,E}\propto \cos\theta$9, with peak attenuation up to $\Delta x$0, $\Delta x$1 up to $\Delta x$2, and a frequency tuning factor of $\Delta x$3 under electrostatic or chemical charging (Sahu et al., 7 Apr 2025, Maksymov et al., 2017, Chiu et al., 2017).

Composition, strain, temperature, and thickness tune plasmon interactions at the material level. In Mo-doped WTe$\Delta x$4, the optical topological transition shifts from $\Delta x$5 in pure WTe$\Delta x$6 to $\Delta x$7 at $\Delta x$8 Mo and $\Delta x$9 K; temperature drives the same boundary from about $\phi_m=2\pi m\Delta x/\Lambda$0 at $\phi_m=2\pi m\Delta x/\Lambda$1 K to about $\phi_m=2\pi m\Delta x/\Lambda$2 at $\phi_m=2\pi m\Delta x/\Lambda$3 K and suppresses the far-IR hyperbolic regime above $\phi_m=2\pi m\Delta x/\Lambda$4 K. In monolayer $\phi_m=2\pi m\Delta x/\Lambda$5-WTe$\phi_m=2\pi m\Delta x/\Lambda$6, strain and doping tune the hyperbolic windows from the undoped values $\phi_m=2\pi m\Delta x/\Lambda$7–$\phi_m=2\pi m\Delta x/\Lambda$8 eV and $\phi_m=2\pi m\Delta x/\Lambda$9–$0.1$0 eV to a broader range of $0.1$1–$0.1$2 eV. In large-area WTe$0.1$3 thin films, the plasmon resonance increases by about $0.1$4 between $0.1$5 and $0.1$6 K and follows $0.1$7 over $0.1$8–$0.1$9 nm thickness. In atomic-chain nanoarrays, changing composition moves the main plasmon from $0.9$0 eV in Cu–Cu to $0.9$1 eV in Au–Au, while Au–Pd fragmentation creates peaks at $0.9$2, $0.9$3, $0.9$4, and $0.9$5 eV (Xie et al., 2022, Torbatian et al., 2020, Wang et al., 2021, Conley et al., 2018).

3. Theoretical descriptions

One canonical description is the discrete-state–continuum framework. In the PMMA-on-metamolecule system, the Hamiltonian

$0.9$6

couples a discrete vibrational state $0.9$7 to two orthogonal plasmonic continua with Lorentzian densities of states. The absorption takes the form

$0.9$8

with the Fano function $0.9$9. The distinctive two-continuum consequence is that even when $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$00, a finite background remains, so the absorption dip cannot cancel completely as it can in the single-continuum case. This gives non-binary control of dip depth and asymmetry (Osley et al., 2012).

A second major framework is symmetry-governed coupled-mode theory for grating-mediated plasmonics. In corrugated double-metal stacks, each grating harmonic carries a phase $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$01, and the effective coupling elements obey

$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$02

Avoided crossings then follow

$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$03

This formulation makes the tuning mechanism explicit: $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$04 does not merely perturb resonance energies; it changes the parity of the effective perturbation and therefore which mode pairs are allowed to hybridize (Yanai et al., 2010).

For plasmon–exciton and plasmon–photon polaritons, the dominant language is the non-Hermitian coupled-oscillator Hamiltonian. In MoS$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$05 nanoantenna lattices and borophene–perovskite structures, the relevant model is

$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$06

with strong coupling identified by resolved anti-crossing and criteria such as $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$07 or, in the borophene case,

$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$08

More complex hybrids require three-mode Hamiltonians, as in WS$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$09–J-aggregate–SPP systems and WSe$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$10 gap-plasmon nanocavities, where Hopfield coefficients quantify redistribution among plasmonic and excitonic fractions (Lee et al., 2016, Yan et al., 2022, Morales et al., 13 Mar 2025, Darlington et al., 2023).

First-principles and full-wave methods provide the complementary microscopic side. RCWA/Fourier-modal calculations validate polarization-selective field profiles and reflection/transmission spectra in metamaterials and corrugated plates; FDTD and FEM track nanogap fields and hybrid cavity modes in gap plasmonics and metasurfaces; TDDFT with transition-contribution maps resolves how collective and single-particle channels coexist in atomic-chain arrays; and RPA-type conductivity formalisms describe anisotropic and hyperbolic plasmons in WTe$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$11 and semi-Dirac systems (Osley et al., 2012, Yanai et al., 2010, Conley et al., 2018, Xie et al., 2022, Giri et al., 15 Jun 2025).

4. Spectral signatures and coupled-mode phenomena

The most recognizable weak-to-intermediate-coupling signature is the Fano profile. In the mid-IR metamolecule–PMMA structure, the carbonyl line evolves from a narrow vibrational feature into a pronounced dip-then-peak resonance as $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$12, including absorption-induced transparency on the long-wavelength side of $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$13m. In reflection, the same interference reverses sign with polarization and yields absorption-induced reflectivity as large as about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$14 at $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$15 (Osley et al., 2012).

In stronger-coupling regimes, the defining signature is anti-crossing. All-metal pyramidal metasurfaces show a clear double anticrossing when two LSPRs interact with a grating-launched SPP. In planar Ag/Al$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$16O$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$17/WS$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$18/TDBC stacks, strong coupling is evidenced by double Rabi splitting at the two excitonic resonances: $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$19 meV in both stacks, while $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$20 increases from $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$21 meV to $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$22 meV as molecular participation rises. In borophene–perovskite gratings, zero detuning produces a reported splitting of about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$23 meV. In WSe$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$24 gap-plasmon nanocavities, force- and voltage-tuned spectra show upper and lower polariton branches with room-temperature vacuum Rabi splitting up to about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$25 meV. In monolayer MoS$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$26 nanoantenna lattices, electrostatic doping continuously moves the neutral-exciton–plasmon channel from strong coupling with $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$27 meV at $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$28 to near-zero coupling above $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$29 V, while the trion–plasmon channel grows to about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$30 meV at high electron density (Marques et al., 2024, Morales et al., 13 Mar 2025, Yan et al., 2022, Darlington et al., 2023, Lee et al., 2016).

A distinct manifestation is topological or phase switching of collective modes. In WTe$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$31, the sign structure of the in-plane optical response changes the isofrequency contour from elliptic to hyperbolic, which directly alters propagation directionality, canalization, and polarization selection; in skew ribbons the angle of maximal absorption obeys

$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$32

In bilayer semi-Dirac systems, the second plasmon branch is an acoustic mode, and rotating the upper layer controls whether coupled oscillations are predominantly in-phase or out-of-phase. This is a form of tunable plasmon interaction in which the controlled variable is the relative phase symmetry of the collective excitation rather than only its energy (Xie et al., 2022, Giri et al., 15 Jun 2025, Torbatian et al., 2020).

5. Spatial reconfiguration and collective-media effects

Some of the most distinctive realizations of tunable plasmon interactions are spatial rather than purely spectral. In the graded InAs nanowire, the local plasma frequency varies along the axis, so a fixed mid-IR excitation creates a single localized resonance where $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$33; gating shifts the entire carrier profile and thereby moves the hotspot itself. The measured localized response extends over about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$34 nm and can be translated by about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$35 nm, suggesting a route to spatially reconfigurable plasmonic circuits and scanning near-field sensors (Arcangeli et al., 2018).

Collective arrays add another layer of tunability through interaction-mediated renormalization. Thick aligned CNT films support Fabry–Pérot plasmons whose frequency depends not only on segment length but also on film thickness through inter-tube coupling; experimentally, increasing thickness blueshifts the resonance into the near-IR, reaches wavelengths as short as about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$36m, and preserves an optical response that is about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$37 linearly polarized along the nanotube axis. In simple-cubic metallic nanoparticle arrays, near-field dipole–dipole coupling turns on-site Mie oscillators into collective plasmons and plasmon polaritons; the resulting polaritonic stop band scales as $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$38, and for $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$39 the predicted gap is about $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$40, with modulation of roughly $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$41 under polarization rotation. In mixed metallic atomic chains, geometry-dependent confinement of $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$42 electrons and chemistry-dependent $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$43 hybridization jointly tune plasmon position, peak splitting, and broadband absorption (Chiu et al., 2017, Weick et al., 2014, Conley et al., 2018).

Spatial disorder does not eliminate tunability if the relevant control variable is the local nanogap. Random Ga droplet ensembles embedded in PDMS suppress lattice resonances and polarization selection, yet still exhibit a continuous, reversible blue shift of the bright bonding mode under macroscopic strain. This provides a non-periodic counterpart to lithographic nanogap tuning: the interaction is read out from ensemble spectra, but its microscopic origin remains local dimer hybridization (Sahu et al., 7 Apr 2025).

6. Applications, limitations, and open directions

The application space is broad because tunability changes not only resonance frequency but the balance between absorption, scattering, LDOS, and modal directionality. Mid-IR plasmon–molecule Fano control enables polarization-dependent surface-enhanced IR absorption and chemical sensing. Corrugated double-metal plates function as tunable reflection and transmission filters. Graphene nonlinear near fields provide electrically tunable resonant excitation, EIT, and temporal control of quantum emitters without placing the emitter on a strong plasmon resonance at its own frequency. TMD gap metasurfaces and plexcitonic nanocavities operate as actively tunable emitters and modulators. WTe$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$44 hyperbolic plasmons support canalization, focusing, and routing. Pyramidal all-metal metasurfaces deliver angle-tunable near-field enhancement for SERS and label-free biosensing, including a reported sensitivity of $$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$45 for the Pb27r antigen. CNT films extend tunable plasmonics into the telecom band (Osley et al., 2012, Yanai et al., 2010, Cox et al., 2019, Ni et al., 2019, Darlington et al., 2023, Xie et al., 2022, Marques et al., 2024, Chiu et al., 2017).

A recurring misconception is that tunability is synonymous with simple peak displacement. Reported systems contradict this. In MoS$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$46 metasurfaces, gating primarily modulates oscillator strengths and the exciton/trion balance while the cavity design fixes the plasmon near the A-exciton. In two-continuum Fano systems, tuning can chiefly alter asymmetry and dip depth. In semi-Dirac bilayers, rotation controls the relative phase and branch character of collective modes. In natural WTe$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$47, hyperbolicity is intrinsic and does not require metasurface patterning, so the tunable object is the isofrequency topology itself rather than only a resonance line (Ni et al., 2019, Osley et al., 2012, Giri et al., 15 Jun 2025, Xie et al., 2022).

The limiting factors are equally diverse but follow a consistent pattern: damping, disorder, and imperfect control of the tuning knob. Dephasing and linewidths restrict Fano contrast and polariton splitting; lithographic tolerances and alignment errors perturb symmetry-selective couplings; trap screening produces hysteresis in semiconductor nanowires; metal losses and quenching limit gap-mode metasurfaces; polycrystallinity masks the intrinsic anisotropy of large-area WTe$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$48 films; high continuous-wave intensities can cause heating in nonlinear graphene schemes; and liquid-metal approaches must contend with oxide-skin and fluid-mechanical stability. These constraints do not negate tunability, but they determine whether it appears primarily as spectral selectivity, large field enhancement, topological switching, or robust low-contrast modulation (Osley et al., 2012, Yanai et al., 2010, Arcangeli et al., 2018, Ni et al., 2019, Wang et al., 2021, Cox et al., 2019, Maksymov et al., 2017).

A plausible implication of the present literature is that the most versatile future platforms will combine several knobs simultaneously: geometric or polarization control to select interaction channels, electrical gating to move detuning or oscillator strength, and material-level tuning to reshape anisotropy or topology. The proposal to remotely modulate molecule–plasmon strong coupling through WS$$\mathrm{Re}[\varepsilon(\omega_{\mathrm{res}})] = -\frac{1-L}{L}\varepsilon_{\mathrm{env}}.$$49 gating, the demonstration of voltage- and force-tunable room-temperature plexcitons, and the phase-selective control of bilayer semi-Dirac plasmons all point in that direction (Morales et al., 13 Mar 2025, Darlington et al., 2023, Giri et al., 15 Jun 2025).