11.1 Phonons in solids

Lattice vibrations in solids

Lattice vibrations are the collective oscillations of atoms in crystalline materials. They form the foundation for understanding how solids store and transport thermal energy, conduct heat, and interact with light and electrons.

Crystal structure basics

Crystals are built from atoms arranged in a periodic, repeating pattern across three dimensions. The unit cell is the smallest building block that, when repeated, reconstructs the entire crystal.

• Bravais lattices classify all possible 3D crystal structures into 14 unique types
• Symmetry operations (translations, rotations, reflections) describe how the structure maps onto itself
• Miller indices $(hkl)$ label specific crystal planes and directions within the lattice

Normal modes of vibration

When atoms in a crystal vibrate, they don't move independently. Instead, they oscillate collectively in patterns called normal modes, each characterized by a specific frequency $\omega$ and wave vector $\mathbf{k}$.

Think of it like a standing wave on a string, but extended to three dimensions across billions of atoms. Each normal mode has a polarization vector that describes the direction atoms move:

• Longitudinal modes: atoms displace parallel to the wave's propagation direction
• Transverse modes: atoms displace perpendicular to the wave's propagation direction

Normal mode frequencies come from solving the equations of motion for the coupled atoms in the lattice. For a 1D monatomic chain with spring constant $C$ and mass $m$, the dispersion relation is $\omega = 2\sqrt{C/m}\,|\sin(ka/2)|$, where $a$ is the lattice spacing.

Phonon dispersion relations

A dispersion relation plots phonon frequency $\omega$ versus wave vector $\mathbf{k}$, and it tells you almost everything about how vibrations behave in a given material.

• Acoustic branches start from $\omega = 0$ at the Brillouin zone center (the $\Gamma$ point) and rise with increasing $|\mathbf{k}|$. These correspond to atoms moving roughly in phase.
• Optical branches have non-zero frequency at $\Gamma$. These appear in crystals with more than one atom per unit cell, where sublattices vibrate against each other.
• The group velocity $v_g = d\omega/dk$ is the slope of the dispersion curve and determines how fast energy propagates.
• Van Hove singularities occur at critical points where the dispersion curve flattens ($v_g \to 0$), producing peaks in the density of states.
• The first Brillouin zone contains all unique wave vectors; anything outside maps back in via a reciprocal lattice vector.

Quantum theory of phonons

Classical mechanics describes lattice vibrations as waves, but a full treatment requires quantum mechanics. Quantizing these vibrations produces phonons, quasiparticles that carry energy and momentum through the lattice.

Phonon quantization

Phonons are to lattice vibrations what photons are to electromagnetic waves. Each phonon mode behaves like a quantum harmonic oscillator.

• Energy of a single phonon: $E = \hbar\omega$
• Crystal momentum: $p = \hbar k$
• The total energy of a mode with $n$ phonons is $E_n = \left(n + \tfrac{1}{2}\right)\hbar\omega$

That $\tfrac{1}{2}\hbar\omega$ term is the zero-point energy: even at absolute zero, atoms still vibrate due to the uncertainty principle. This has measurable consequences, such as zero-point motion preventing helium from solidifying at atmospheric pressure.

Creation and annihilation operators

The quantum harmonic oscillator formalism uses ladder operators to move between phonon number states:

• Creation operator $a^\dagger$: adds one phonon to a mode, $a^\dagger|n\rangle = \sqrt{n+1}\,|n+1\rangle$
• Annihilation operator $a$: removes one phonon, $a|n\rangle = \sqrt{n}\,|n-1\rangle$
• They satisfy the bosonic commutation relation: $[a, a^\dagger] = 1$
• The number operator $\hat{N} = a^\dagger a$ gives the phonon count in a mode: $\hat{N}|n\rangle = n|n\rangle$

These operators make it straightforward to compute matrix elements for phonon transitions and to build the Hamiltonian for interacting phonon systems.

Phonon statistics

Phonons are bosons, so any number of them can occupy the same quantum state. Their thermal occupation follows the Bose-Einstein distribution:

$n(\omega) = \frac{1}{e^{\hbar\omega / k_B T} - 1}$

At high temperatures ($k_B T \gg \hbar\omega$), this reduces to $n \approx k_B T / \hbar\omega$, recovering the classical equipartition result. At low temperatures, high-frequency modes freeze out because there isn't enough thermal energy to excite them.

The Debye temperature $\theta_D = \hbar\omega_D / k_B$ sets the scale: below $\theta_D$, quantum effects dominate phonon behavior. For diamond, $\theta_D \approx 2230\,\text{K}$; for lead, it's only about $105\,\text{K}$.

Thermodynamic properties of phonons

Phonons are the main contributors to the thermal properties of insulators and semiconductors. Getting the phonon physics right is essential for predicting heat capacity, thermal conductivity, and thermal expansion.

Phonon density of states

The density of states $g(\omega)$ counts how many phonon modes exist per unit frequency interval. You calculate it by integrating over all wave vectors in the Brillouin zone.

• The Debye approximation assumes a linear dispersion ($\omega = v|\mathbf{k}|$), giving $g(\omega) \propto \omega^2$ up to a cutoff frequency $\omega_D$
• Van Hove singularities appear as sharp features in $g(\omega)$ where the dispersion curve is flat
• At low temperatures, only low-frequency acoustic modes are thermally populated, so they dominate the effective density of states
• At higher temperatures, optical modes become populated and contribute significantly

Heat capacity of solids

The heat capacity tells you how much energy a solid absorbs per degree of temperature increase. Phonons are the dominant contribution in non-metals.

• Classical limit (Dulong-Petit): At high $T$, each atom contributes $3k_B$ to the heat capacity, giving $C_V = 3Nk_B = 3R$ per mole. This works well above $\theta_D$.
• Low-temperature behavior: Heat capacity drops as phonon modes freeze out. For $T \ll \theta_D$, acoustic phonons give a characteristic $C_V \propto T^3$ dependence.
• In metals, there's an additional electronic contribution $C_{\text{el}} \propto T$ that dominates at very low $T$, so the total is $C_V = \gamma T + \alpha T^3$.

Debye model vs Einstein model

These two models take different approaches to approximating the phonon spectrum:

The Debye model gets the low-temperature physics right because it correctly captures the linear acoustic dispersion near $\mathbf{k} = 0$. The Einstein model fails there because it assumes a gap in the spectrum, but it's useful for describing the contribution of optical branches. For real materials with complex phonon spectra, a combined Debye-Einstein model often fits experimental data best.

Phonon interactions

In a perfectly harmonic crystal, phonons would propagate forever without scattering. Real crystals are anharmonic, and phonons scatter off each other, off electrons, and off defects. These interactions determine thermal conductivity and set phonon lifetimes.

Phonon-phonon scattering

Anharmonic terms in the interatomic potential (cubic, quartic, etc.) cause phonons to interact.

• Normal (N) processes: Total crystal momentum is conserved. These redistribute momentum among phonon modes but don't directly create thermal resistance.
• Umklapp (U) processes: The total wave vector exceeds the first Brillouin zone and gets folded back by a reciprocal lattice vector $\mathbf{G}$. These do create thermal resistance because net momentum is transferred to the lattice.
• At low temperatures, Umklapp scattering is exponentially suppressed ($\propto e^{-\theta_D/bT}$) because few phonons have large enough wave vectors. This is why thermal conductivity of pure crystals peaks at low $T$.
• Three-phonon processes dominate at moderate temperatures. Higher-order (four-phonon) processes become relevant at high $T$.
• The relaxation time approximation assigns a scattering rate $1/\tau$ to each phonon mode, simplifying transport calculations.

Electron-phonon coupling

In metals and semiconductors, electrons and phonons exchange energy and momentum.

• This interaction is the primary source of electrical resistivity in metals above the Debye temperature (resistivity $\propto T$)
• In BCS superconductivity, phonon exchange provides the attractive interaction that binds electrons into Cooper pairs
• The coupling constant $\lambda$ quantifies the interaction strength; $\lambda > 1$ indicates strong coupling
• A polaron forms when an electron distorts the surrounding lattice and becomes "dressed" by a cloud of virtual phonons, increasing its effective mass
• The Fröhlich Hamiltonian is the standard model for electron-phonon coupling in polar semiconductors

Phonon-impurity interactions

Real crystals contain defects, and phonons scatter off them. This scattering contributes to thermal resistance and is especially important at intermediate temperatures where Umklapp scattering is weak.

• Rayleigh scattering ($\propto \omega^4$) dominates when impurities are much smaller than the phonon wavelength
• Resonant scattering occurs when the impurity's local vibrational frequency matches the phonon frequency
• Isotope scattering matters in elements with multiple stable isotopes. Natural germanium, for example, has five isotopes, and isotopically purified $^{70}\text{Ge}$ shows measurably higher thermal conductivity.
• Extended defects (dislocations, grain boundaries) scatter phonons across a broad frequency range

Experimental techniques

Several complementary methods probe phonon properties. Each has different strengths in terms of the frequency range, momentum range, and types of modes it can access.

Neutron scattering

Neutron scattering is the most direct way to measure full phonon dispersion relations because thermal neutrons have wavelengths and energies comparable to interatomic spacings and phonon energies.

• Inelastic neutron scattering measures both the energy and momentum transferred between neutrons and phonons
• Triple-axis spectrometry targets specific $(q, \omega)$ points in the dispersion, giving high precision
• Time-of-flight spectroscopy maps out broad regions of the dispersion simultaneously
• Energy loss (Stokes) and energy gain (anti-Stokes) events correspond to phonon creation and annihilation, respectively
• Coherent scattering probes collective (phonon) modes; incoherent scattering probes individual atomic dynamics

The main limitation is that neutron sources (reactors, spallation sources) are large-scale facilities, so access is limited.

Raman spectroscopy

Raman spectroscopy is a lab-based optical technique that probes phonons through inelastic light scattering.

• Incident photons scatter off phonons, shifting in frequency by $\pm\omega_{\text{phonon}}$
• Stokes shift (photon loses energy): phonon created. Anti-Stokes shift (photon gains energy): phonon absorbed.
• Selection rules restrict which modes are Raman-active, based on symmetry of the phonon mode
• Because visible light has very small momentum compared to the Brillouin zone, Raman spectroscopy probes only phonons near $\mathbf{k} \approx 0$ (zone center optical modes)
• Temperature-dependent measurements reveal anharmonic frequency shifts and linewidth broadening

Raman is non-destructive, requires minimal sample preparation, and can be done with a benchtop instrument.

X-ray techniques for phonons

X-rays complement neutrons and Raman by offering high spatial resolution and access to small samples.

• Thermal diffuse scattering (TDS): the diffuse background in X-ray diffraction patterns arises from phonons. Its intensity increases with temperature as phonon populations grow.
• Inelastic X-ray scattering (IXS): using high-brilliance synchrotron sources, IXS measures phonon dispersions with energy resolution of a few meV. It's especially useful for small crystals and materials that absorb neutrons strongly.
• Huang scattering arises from long-range strain fields around point defects
• X-ray absorption fine structure (XAFS) probes local atomic vibrations around specific elements

Phonons in different materials

Phonon behavior varies dramatically depending on the material class. The bonding type, crystal structure, and dimensionality all shape the phonon spectrum and transport properties.

Metals vs insulators

• In metals, strong electron-phonon coupling shortens phonon lifetimes and modifies dispersion curves. Kohn anomalies are dips or kinks in the phonon dispersion at wave vectors $\mathbf{q} = 2\mathbf{k}_F$ (spanning the Fermi surface), caused by electronic screening.
• In insulators, phonons travel with longer mean free paths because there are no conduction electrons to scatter off. This is why diamond (an insulator) has one of the highest thermal conductivities of any material (~2000 W/m·K).
• Acoustic phonons dominate thermal transport in both classes at low temperatures.
• Debye temperatures tend to be higher in materials with strong, stiff bonds (diamond: ~2230 K) and lower in soft metals (lead: ~105 K).

Nanostructures and low-dimensional systems

When a material's dimensions shrink to the scale of phonon wavelengths (typically nanometers), confinement effects reshape the phonon spectrum.

• Quantum dots: 3D confinement produces discrete phonon modes, similar to a particle in a box
• Nanowires and nanotubes: phonon modes are quantized in the transverse directions but continuous along the wire axis
• 2D materials like graphene have distinctive in-plane modes and a unique quadratic flexural (out-of-plane) acoustic branch, $\omega \propto k^2$, which strongly affects thermal transport
• Surface and interface phonons become increasingly important as the surface-to-volume ratio grows
• Phonon engineering exploits these confinement effects to tune thermal conductivity, for example reducing it in thermoelectric materials by introducing nanostructured boundaries

Phonons in superconductors

Phonons play a central role in conventional superconductivity.

• In BCS theory, two electrons near the Fermi surface exchange virtual phonons, creating an attractive interaction that binds them into Cooper pairs
• The isotope effect ($T_c \propto M^{-1/2}$) provided early evidence that phonons mediate pairing: heavier isotopes lower the critical temperature
• Soft phonon modes (modes whose frequency drops toward zero) often signal structural phase transitions in superconducting materials
• Eliashberg theory extends BCS to the strong-coupling regime, accounting for the retarded nature of the phonon-mediated interaction and the competition with Coulomb repulsion (captured by the Coulomb pseudopotential $\mu^*$)
• In high-temperature superconductors, the pairing mechanism is still debated, but phonons likely play at least a partial role alongside electronic correlations

Applications of phonon theory

Understanding phonon physics has direct technological payoffs, from keeping computer chips cool to converting waste heat into electricity.

Thermal conductivity

In electrically insulating materials, phonons carry virtually all the heat. The kinetic theory expression for lattice thermal conductivity is:

$\kappa = \frac{1}{3} C v \ell$

where $C$ is the volumetric heat capacity, $v$ is the average phonon group velocity, and $\ell$ is the phonon mean free path.

Each scattering mechanism limits $\ell$ differently:

1. Umklapp scattering dominates at high $T$, giving $\kappa \propto 1/T$
2. Boundary scattering dominates at low $T$ in pure crystals, where $\ell$ is set by the sample size
3. Impurity/defect scattering is important at intermediate temperatures

Phonon engineering strategies for thermal management include nanostructuring (reducing $\ell$), alloying (introducing mass disorder), and using materials with intrinsically anharmonic bonding.

Thermoelectric effects

Thermoelectric devices convert temperature differences into voltage (Seebeck effect) or use voltage to pump heat (Peltier effect). Their efficiency is governed by the dimensionless figure of merit:

$ZT = \frac{S^2 \sigma T}{\kappa}$

where $S$ is the Seebeck coefficient, $\sigma$ is electrical conductivity, and $\kappa$ is total thermal conductivity (electronic + lattice).

• The phonon-glass electron-crystal concept aims to minimize $\kappa_{\text{lattice}}$ (scatter phonons like a glass) while keeping $\sigma$ high (conduct electrons like a crystal)
• Nanostructuring introduces boundaries that scatter phonons more than electrons, selectively reducing $\kappa_{\text{lattice}}$
• Phonon drag (phonons pushing electrons along) enhances the Seebeck coefficient at low temperatures but diminishes above $\sim\theta_D/5$

Phononic crystals

Phononic crystals are engineered periodic structures that control phonon propagation, analogous to how photonic crystals control light.

• Periodic variations in density and elastic constants create phononic band gaps: frequency ranges where no phonon modes can propagate
• Introducing defects into the periodic structure creates localized modes, enabling waveguides and cavities for sound/heat
• Applications include acoustic filters, vibration isolation, and thermal management devices
• Hypersonic phononic crystals operate at GHz-THz frequencies, relevant for on-chip thermal control
• Optomechanical crystals combine phononic and photonic band gaps, coupling light and mechanical vibrations for sensing and quantum information applications

Advanced topics

Anharmonic effects

The harmonic approximation treats interatomic potentials as perfectly parabolic. Real potentials are asymmetric, and this anharmonicity has several important consequences:

• Thermal expansion: the asymmetry of the potential means the average atomic position shifts outward as temperature increases
• Finite phonon lifetimes: anharmonic terms couple phonon modes, causing them to decay and scatter
• The Grüneisen parameter $\gamma = -\frac{V}{\omega}\frac{\partial \omega}{\partial V}$ quantifies how phonon frequencies shift with volume, connecting phonon physics to thermal expansion and the equation of state
• Higher-order anharmonic effects enable phenomena like phonon frequency conversion and parametric amplification
• Accurate modeling of thermal transport requires going beyond the harmonic approximation, especially near phase transitions

Optical vs acoustic phonons

For crystals with more than one atom per unit cell, the phonon spectrum splits into acoustic and optical branches.

• Acoustic phonons: neighboring atoms move roughly in phase. Frequency $\to 0$ as $k \to 0$. Three acoustic branches exist in 3D (1 longitudinal, 2 transverse).
• Optical phonons: neighboring atoms in different sublattices move out of phase. Frequency is finite at $k = 0$. For $p$ atoms per unit cell, there are $3(p-1)$ optical branches.
• In polar materials (like GaAs or NaCl), the longitudinal optical (LO) and transverse optical (TO) modes split at $k = 0$ due to long-range Coulomb interactions. This LO-TO splitting is described by the Lyddane-Sachs-Teller relation: $\omega_{LO}^2 / \omega_{TO}^2 = \varepsilon_0 / \varepsilon_\infty$.
• Surface phonon polaritons form when optical phonons couple to electromagnetic waves at a material surface, creating hybrid modes that propagate along the interface.

Surface and interface phonons

Surfaces and interfaces break translational symmetry, creating vibrational modes that don't exist in the bulk.

• Rayleigh waves are surface acoustic modes with elliptical particle motion, decaying exponentially into the bulk. Their velocity is slightly below the bulk transverse sound speed.
• Fuchs-Kliewer phonons are surface optical modes in polar materials, detectable by electron energy loss spectroscopy
• Interface phonons at boundaries between dissimilar materials have frequencies and character intermediate between the two bulk materials
• Phonon focusing near surfaces arises from elastic anisotropy, causing phonon energy flux to concentrate along preferred crystallographic directions
• These modes become increasingly important in thin films and nanostructures, where a large fraction of atoms sit near surfaces or interfaces