8.3 Plasmons

Fundamentals of Plasmons

Plasmons are quantized collective oscillations of free electrons in conductive materials. They sit at the heart of how metals and other conductors interact with light, governing optical reflectivity, absorption, and transmission. Because plasmons concentrate electromagnetic energy into volumes far smaller than the wavelength of light, they've become central to nanophotonics, sensing, and a growing range of optoelectronic technologies.

This topic covers the main types of plasmons (bulk, surface, and localized), how they're excited and characterized experimentally, their dispersion relations under classical and quantum descriptions, and their applications from spectroscopy to quantum information.

Definition and Basic Properties

A plasmon is the quantum of a collective oscillation of free electrons in a conducting medium. Think of it this way: when the electron gas in a metal is displaced from equilibrium, the restoring force from the positive ion background causes the electrons to oscillate collectively. Quantizing that oscillation gives you a plasmon.

• Plasmons exhibit wave-like behavior with characteristic frequencies and wavelengths
• They occur in metals, doped semiconductors, and any material with a significant free carrier density
• They couple to electromagnetic fields, which is what makes them useful for manipulating light at the nanoscale

Collective Electron Oscillations

The key word is collective. A plasmon isn't a single electron bouncing around; it's a coordinated, in-phase motion of a huge number of conduction electrons responding to a perturbation.

• This coordinated displacement creates alternating regions of positive and negative charge density within the material
• In bulk, these oscillations propagate as longitudinal waves (the displacement is parallel to the propagation direction)
• Plasmons decay over time through several damping channels: electron-electron scattering, electron-phonon interactions, and radiation losses (for surface and localized modes)

Plasmon Frequency

The plasma frequency $\omega_p$ sets the fundamental energy scale for plasmonic behavior. It's determined entirely by the electron density and effective mass:

$\omega_p = \sqrt{\frac{ne^2}{\epsilon_0 m^*}}$

where $n$ is the free electron density, $e$ is the electron charge, $\epsilon_0$ is the vacuum permittivity, and $m^*$ is the electron effective mass.

• For most metals (Al, Ag, Au), $\omega_p$ falls in the ultraviolet range (typically 5–15 eV)
• Below $\omega_p$, the metal reflects incident light (the dielectric function is negative). Above $\omega_p$, the metal becomes transparent. This is why $\omega_p$ directly controls a material's optical response.

Types of Plasmons

Plasmons take different forms depending on the geometry of the system. The three main categories are bulk, surface, and localized surface plasmons, each with distinct properties and excitation requirements.

Bulk Plasmons

Bulk plasmons are longitudinal oscillations of the electron gas in the interior of a three-dimensional conductor.

• They oscillate at (or near) the plasma frequency $\omega_p$
• Because they are longitudinal, they cannot couple directly to transverse electromagnetic waves (light). There's a fundamental momentum mismatch: light at frequency $\omega_p$ carries negligible momentum compared to what the bulk plasmon requires.
• The primary experimental probe is electron energy loss spectroscopy (EELS), where a fast electron passing through a thin film loses energy in quanta of $\hbar\omega_p$
• Measuring the bulk plasmon energy gives direct information about the free electron density of the material

Surface Plasmons

Surface plasmon polaritons (SPPs) are electromagnetic excitations bound to the interface between a conductor and a dielectric. Unlike bulk plasmons, these are transverse magnetic modes with both longitudinal and transverse field components.

• Their fields decay evanescently into both the metal and the dielectric, confining energy to a region much thinner than the free-space wavelength
• This tight confinement is what enables subwavelength manipulation of light, beating the diffraction limit
• SPPs still can't be excited by light hitting a flat surface directly (momentum mismatch again). You need a coupling scheme: prism coupling (Kretschmann or Otto configuration), grating coupling, or scattering from surface defects
• Applications include biosensing, on-chip waveguiding, and enhanced spectroscopy

Localized Surface Plasmons

When the conductor is a nanoparticle or nanostructure smaller than the wavelength of light, you get localized surface plasmon resonances (LSPRs). These are non-propagating oscillations.

• Unlike SPPs, LSPRs can be excited directly by light with no special coupling geometry, because the curved nanoparticle surface itself provides the necessary momentum
• The resonance frequency depends strongly on particle size, shape, composition, and the dielectric constant of the surrounding medium
• At resonance, the local electric field near the nanoparticle can be enhanced by factors of $10^2$ to $10^4$ relative to the incident field
• This enormous field enhancement is the basis for surface-enhanced Raman spectroscopy (SERS) and surface-enhanced infrared absorption (SEIRA)

Plasmon Excitation Methods

Different types of plasmons require different excitation strategies. The choice depends on which plasmon you're targeting and what information you need.

Optical Excitation

Light can directly excite localized surface plasmons in nanoparticles and, with the right coupling geometry, surface plasmon polaritons on flat films.

• For SPPs on flat surfaces, phase-matching is required to bridge the momentum gap between free-space photons and the SPP mode. Common approaches:
  1. Prism coupling (Kretschmann configuration): light enters through a high-index prism at an angle beyond total internal reflection, generating an evanescent wave that matches the SPP wavevector
  2. Grating coupling: a periodic surface structure adds discrete momentum increments $\Delta k = 2\pi/\Lambda$ (where $\Lambda$ is the grating period) to the incident photon
• Broadband sources allow spectroscopic characterization of resonance positions and linewidths; lasers provide high-intensity excitation for nonlinear studies

Electron Beam Excitation

A focused beam of high-energy electrons (typically 60–300 keV in a transmission electron microscope) can excite plasmons through inelastic scattering.

• The electron transfers both energy and momentum to the plasmon, so there's no momentum-matching problem. This means electron beams can excite all types of plasmons, including bulk modes that light cannot access.
• Spatial resolution reaches the sub-nanometer scale, allowing you to map plasmon modes across individual nanostructures
• The technique is called EELS (electron energy loss spectroscopy) when you measure the energy lost by the transmitted electron

Near-Field Coupling

Plasmons can also be excited through evanescent (near-field) interactions with nearby emitters or structures.

• In scanning near-field optical microscopy (SNOM), a sharp probe tip brought within nanometers of the surface generates evanescent fields that couple to plasmonic modes
• This approach achieves excitation and detection with spatial resolution well below the diffraction limit
• Near-field coupling also enables strong coupling between individual quantum emitters (quantum dots, molecules) and plasmonic modes, which is important for quantum plasmonics

Plasmon Dispersion Relations

Dispersion relations connect the energy (frequency) and momentum (wavevector) of plasmonic excitations. They tell you how plasmons propagate and at what frequencies they exist.

Drude Model

The Drude model is the simplest classical framework for the optical response of metals. It treats conduction electrons as a free gas that scatters with a characteristic damping rate $\gamma$.

The resulting dielectric function is:

$\epsilon(\omega) = 1 - \frac{\omega_p^2}{\omega^2 + i\gamma\omega}$

From this, you can derive the bulk plasmon dispersion (including the leading-order spatial dispersion correction):

$\omega^2 = \omega_p^2 + \frac{3}{5}v_F^2 k^2$

where $v_F$ is the Fermi velocity and $k$ is the wavevector. At $k = 0$, the bulk plasmon sits right at $\omega_p$.

The Drude model works well for simple metals like Al and the alkali metals. It breaks down for noble metals (Au, Ag, Cu) at frequencies where interband transitions contribute significantly to the dielectric function.

Dielectric Function Approach

For real materials, you use the full frequency-dependent dielectric function $\epsilon_m(\omega)$, which includes both free-electron (Drude) and interband contributions. This can come from experimental optical data or from more sophisticated models.

The surface plasmon polariton dispersion on a flat interface between a metal ($\epsilon_m$) and a dielectric ($\epsilon_d$) is:

$k_{sp} = \frac{\omega}{c}\sqrt{\frac{\epsilon_d \, \epsilon_m(\omega)}{\epsilon_d + \epsilon_m(\omega)}}$

Notice that $k_{sp}$ is always larger than the free-space photon wavevector $\omega/c$ (because $\epsilon_m < 0$ for frequencies below $\omega_p$). This is the mathematical origin of the momentum mismatch that prevents direct optical excitation of SPPs on flat surfaces.

The SPP frequency asymptotically approaches $\omega_{sp} = \omega_p / \sqrt{1 + \epsilon_d}$ at large wavevectors.

Quantum Mechanical Description

Classical models fail when structures shrink to a few nanometers, approaching the Fermi wavelength of the electrons (~0.5 nm in typical metals).

• Nonlocal effects: the dielectric response becomes wavevector-dependent, $\epsilon(\omega, k)$, because electrons at different positions respond to the average field over a region comparable to the Thomas-Fermi screening length
• Electron spill-out: electron density doesn't terminate sharply at the metal surface but decays over ~1 Å into the vacuum, shifting plasmon resonances
• Quantum tunneling: in sub-nanometer gaps between nanoparticles, electrons tunnel across the gap, dramatically altering the plasmonic response

These effects are captured by time-dependent density functional theory (TD-DFT) or the quantum hydrodynamic model. They become essential for accurately describing plasmons in ultrasmall nanoparticles, few-atom clusters, and angstrom-scale gaps.

Applications of Plasmons

Surface-Enhanced Spectroscopy

The intense local field enhancement near plasmonic nanostructures can boost spectroscopic signals by many orders of magnitude.

• SERS (surface-enhanced Raman spectroscopy): Raman scattering scales as the fourth power of the local field enhancement, $|E/E_0|^4$. Enhancement factors of $10^6$ to $10^{10}$ are routinely achieved, and single-molecule detection has been demonstrated in "hot spots" (nanogaps between particles).
• SEIRA (surface-enhanced infrared absorption): plasmonic antennas tuned to mid-infrared frequencies enhance vibrational absorption signals by factors of $10^3$ to $10^5$
• Both techniques are used for chemical identification, biological sensing, and trace detection

Plasmonic Sensors

Plasmon resonances shift when the local refractive index changes, making them natural transducers for detecting molecular binding events.

• SPR sensors: monitor the angle or wavelength shift of the surface plasmon resonance on a thin gold film. Widely used in pharmaceutical research to measure biomolecular binding kinetics in real time, without fluorescent labels.
• LSPR sensors: use the spectral shift of the localized plasmon resonance of nanoparticles. These can be miniaturized to single-nanoparticle detectors and are used in point-of-care diagnostics, environmental monitoring, and food safety testing.

Metamaterials and Cloaking

Engineered arrays of plasmonic elements can produce effective optical properties not found in nature.

• Negative refractive index materials bend light the "wrong" way and can be used to build superlenses that image below the diffraction limit
• Metasurfaces are single-layer arrays of plasmonic nanoantennas that control the phase, amplitude, and polarization of transmitted or reflected light. They enable flat lenses, beam steerers, and holograms.
• Plasmonic cloaking concepts use carefully designed shells or coatings to cancel the scattering from an object, rendering it less visible at specific frequencies

Experimental Techniques

Electron Energy Loss Spectroscopy

EELS in a transmission electron microscope (TEM) is the most direct way to probe plasmons with nanometer spatial resolution.

1. A focused electron beam (typically 60–300 keV) passes through or near the sample
2. Electrons that excite plasmons lose discrete amounts of energy
3. An energy-analyzing spectrometer measures the distribution of energy losses
4. By rastering the beam, you build a spatial map of plasmon modes across a nanostructure

EELS detects both bright (radiative) and dark (non-radiative) plasmon modes, which is an advantage over purely optical techniques.

Near-Field Optical Microscopy

SNOM (scanning near-field optical microscopy) maps the electromagnetic field distribution of plasmonic structures with resolution far below the diffraction limit (down to ~10–20 nm).

• A sharp probe tip (either an aperture tip or a scattering tip) is scanned nanometers above the sample surface
• The tip either collects or scatters the local evanescent field, converting it to propagating light that reaches a far-field detector
• Tip-enhanced variants (TERS, TENOM) use a plasmonic tip itself to enhance the local field, combining near-field imaging with spectroscopic sensitivity

Plasmon Resonance Spectroscopy

Far-field optical techniques characterize the resonance properties of plasmonic structures.

• UV-Vis extinction spectroscopy measures the combined absorption and scattering of nanoparticle ensembles in solution
• Dark-field scattering spectroscopy isolates the scattering spectrum of individual nanoparticles by illuminating at high angles and collecting only scattered light
• These measurements yield the resonance wavelength, linewidth (which gives the quality factor $Q$), and polarization dependence
• Correlating optical spectra with electron microscopy images of the same particle links geometry to plasmonic response

Plasmonics in Nanomaterials

Nanoscale geometry fundamentally shapes plasmonic behavior. Size, shape, and dimensionality all tune the resonance frequency, field enhancement, and propagation characteristics.

Nanoparticle Plasmons

Metal nanoparticles (Au, Ag, Cu, Al) support localized surface plasmon resonances whose properties are highly tunable.

• Size dependence: for particles much smaller than the wavelength, the resonance is governed by the quasistatic (Mie) limit. As particle size increases, retardation effects red-shift and broaden the resonance.
• Shape dependence: nanorods, nanotriangles, nanostars, and other anisotropic shapes split the resonance into multiple modes and push resonances into the near-infrared
• Collective effects: ordered arrays of nanoparticles support lattice plasmon modes (also called surface lattice resonances) with much narrower linewidths than individual particles. Fano resonances arise from interference between broad and narrow modes.
• Photothermal applications: resonant absorption converts light to heat efficiently, enabling photothermal cancer therapy and plasmon-driven catalysis

Nanowire Plasmons

Metal nanowires act as subwavelength plasmonic waveguides.

• SPP modes propagate along the wire axis with mode areas far below $(\lambda/2)^2$
• Propagation lengths range from a few micrometers (visible) to tens of micrometers (near-infrared), limited by ohmic losses in the metal
• Quantum emitters placed near a nanowire can couple their emission into the guided plasmon mode, enabling directional single-photon sources
• Nanowire networks are being explored as interconnects for nanoscale photonic circuits

Graphene Plasmonics

Graphene supports two-dimensional plasmons with properties quite different from those in metals.

• The plasmon frequency in graphene scales as $\omega \propto (E_F \, q)^{1/2}$, where $E_F$ is the Fermi energy and $q$ is the in-plane wavevector. This means the resonance is electrically tunable via a gate voltage that shifts $E_F$.
• Graphene plasmons operate in the terahertz to mid-infrared range, complementing the visible/UV range of metal plasmons
• They achieve extreme spatial confinement: the plasmon wavelength can be 10–100 times shorter than the free-space wavelength at the same frequency
• Relatively low losses (compared to metals at these frequencies) give longer propagation lengths
• Applications include tunable infrared modulators, detectors, and reconfigurable metasurfaces

Plasmon-Phonon Interactions

When plasmon and phonon energies overlap, the two excitations hybridize. This coupling is especially important in polar semiconductors and ionic materials, where optical phonons carry significant oscillator strength.

Coupled Modes

• Plasmon-phonon coupling produces hybrid modes that share characteristics of both excitations
• Near the crossing point of the uncoupled plasmon and phonon dispersions, you see anticrossing (avoided crossing) behavior: the modes repel each other, opening an energy gap proportional to the coupling strength
• The resulting upper and lower branches are called coupled plasmon-phonon modes (or, at surfaces, surface phonon-plasmon polaritons)

Energy Transfer Mechanisms

Plasmon-phonon coupling provides pathways for energy to flow between electronic and vibrational degrees of freedom.

• When a plasmon decays non-radiatively, it generates hot electrons. These hot electrons thermalize by emitting phonons, transferring energy to the lattice on a timescale of ~1 ps.
• Conversely, phonon-assisted processes contribute to plasmon damping, broadening the plasmon linewidth
• In SERS, the coupling between plasmons and molecular vibrations is what produces the enhanced Raman signal
• Managing plasmon-phonon energy transfer is critical for thermal design in plasmonic devices, where localized heating can be either useful (photothermal therapy) or detrimental (device degradation)

Phonon-Plasmon Polaritons

In doped polar semiconductors (e.g., doped GaAs, InP, or SiC), free-carrier plasmons and transverse optical phonons can both couple to light, forming phonon-plasmon polaritons.

• The dispersion relation shows multiple branches with anticrossing at both the phonon and plasmon frequencies
• These hybrid modes exist in the mid-infrared to terahertz range
• They enable tunable control of light-matter interactions by adjusting the carrier concentration (which shifts $\omega_p$ relative to the phonon frequency)
• Applications include tunable thermal emitters, infrared sensors, and THz optoelectronic devices

Nonlinear Plasmonics

Plasmonic field enhancement doesn't just boost linear signals. Because nonlinear optical processes scale as high powers of the local field, even modest field enhancements translate into enormous nonlinear signal boosts.

Second-Harmonic Generation

Second-harmonic generation (SHG) converts two photons at frequency $\omega$ into one photon at $2\omega$. It's a second-order nonlinear process that requires broken inversion symmetry.

• Metal surfaces and nanoparticle surfaces inherently break inversion symmetry, making them SHG-active
• Plasmonic enhancement of the local field at $\omega$ boosts the SHG signal roughly as $|E(\omega)/E_0|^4$
• SHG is used as a sensitive probe of surface symmetry, plasmonic near-field distributions, and ultrafast dynamics
• Nanostructures with intentionally broken symmetry (e.g., L-shaped or split-ring resonators) show particularly strong SHG

Four-Wave Mixing

Four-wave mixing (FWM) is a third-order nonlinear process involving the interaction of four photons. It doesn't require broken symmetry and occurs in all materials.

• Plasmonic nanostructures enhance FWM efficiency through local field concentration
• Applications include wavelength conversion, all-optical switching, and coherent anti-Stokes Raman spectroscopy (CARS)
• FWM in plasmonic systems is being explored for on-chip nonlinear signal processing at scales far below conventional photonic components

Plasmonic Hot Electrons

Non-radiative plasmon decay generates hot electrons with energies above the Fermi level, distributed in a non-thermal (non-Fermi-Dirac) fashion.

• These hot electrons can have enough energy to overcome Schottky barriers at metal-semiconductor interfaces or to drive chemical reactions that would otherwise require UV photons
• Photocatalysis: hot electrons injected into adsorbed molecules or a semiconductor catalyst can drive reactions like water splitting or $\text{CO}_2$ reduction using visible light
• Hot-electron photodetectors: devices that harvest sub-bandgap photons by injecting plasmon-generated hot electrons across a metal-semiconductor junction
• The challenge is that hot electrons thermalize within ~100 fs to 1 ps, so efficient extraction requires very short transport distances (nanometer-scale junctions)

Quantum Plasmonics

Quantum plasmonics investigates what happens when plasmonic structures shrink to the point where classical electrodynamics breaks down, and when individual quanta of the plasmonic field become relevant.

Quantum Size Effects

When nanoparticle dimensions approach the Fermi wavelength (~0.5 nm for noble metals), the continuous band structure gives way to discrete electronic states.

• Plasmon resonances blue-shift and broaden in particles below ~2 nm diameter, deviating from classical Mie theory predictions
• In sub-nanometer gaps between particles, quantum tunneling of electrons across the gap creates a conductive channel that shorts out the classical field enhancement
• These effects set fundamental limits on how small you can make a plasmonic structure and still get useful field enhancement
• Accurate modeling requires TD-DFT or quantum-corrected classical models (e.g., the quantum-corrected model, QCM)

Single-Plasmon Devices

The goal here is to control plasmonic excitations at the level of individual quanta.

• A single quantum emitter (quantum dot, nitrogen-vacancy center, single molecule) coupled to a plasmonic nanostructure can emit single plasmons, which then radiate as single photons
• This provides a route to nanoscale single-photon sources with high emission rates (thanks to the Purcell effect) and directional output
• Single-plasmon detection remains challenging because of the high losses in metals, but progress is being made using superconducting nanowire detectors and correlation measurements
• Potential applications include quantum key distribution, quantum sensing, and on-chip quantum information processing

Quantum Emitter-Plasmon Coupling

The interaction between a quantum emitter and a plasmonic mode can range from weak to strong coupling, depending on the ratio of the coupling rate to the loss rates.

• Weak coupling (Purcell regime): the emitter's spontaneous emission rate is enhanced by the Purcell factor $F_P \propto Q/V$, where $Q$ is the mode quality factor and $V$ is the mode volume. Plasmonic cavities have very small $V$, so even with modest $Q$ (~10–50), Purcell factors of $10^2$ to $10^3$ are achievable.
• Strong coupling: when the coupling rate exceeds both the plasmon decay rate and the emitter dephasing rate, the system enters the strong coupling regime. The plasmon and exciton hybridize into plexcitons (plasmon-exciton polaritons), visible as anticrossing in the spectrum.
• Strong coupling has been observed at room temperature in single-nanoparticle systems, which is remarkable since most strong-coupling experiments in photonic cavities require cryogenic temperatures
• Applications include ultrafast single-photon nonlinearities, room-temperature quantum electrodynamics, and nanoscale light sources with controlled emission properties