4.2 Einstein and Debye models

Heat Capacity in Solids

Heat capacity quantifies how much heat you need to add to raise a material's temperature by one degree. In solids, this thermal energy is stored in the vibrations of atoms around their equilibrium positions. These quantized lattice vibrations are called phonons, and they're the central object in both the Einstein and Debye models.

The classical prediction (Dulong-Petit law) says every solid should have a molar heat capacity of $C = 3Nk_B$ regardless of temperature. Experiments show this works well at high temperatures but fails badly at low temperatures, where heat capacity drops toward zero. The Einstein and Debye models were developed to explain that drop using quantum mechanics.

Einstein Model of Heat Capacity

Assumptions of the Einstein Model

The Einstein model is the simpler of the two. It treats a solid containing $N$ atoms as $3N$ independent quantum harmonic oscillators, all vibrating at a single frequency $\omega_E$ (the Einstein frequency). Each oscillator has quantized energy levels spaced by $\hbar\omega_E$.

Key simplifications:

• Every atom vibrates at the same frequency, independent of temperature
• There's no coupling between different vibrational modes
• No distinction is made between acoustic and optical phonons

Einstein Temperature

The Einstein temperature is defined as:

$\Theta_E = \frac{\hbar\omega_E}{k_B}$

This sets the energy scale for the model. The heat capacity per atom comes out to:

$C_V = 3Nk_B \left(\frac{\Theta_E}{T}\right)^2 \frac{e^{\Theta_E/T}}{\left(e^{\Theta_E/T} - 1\right)^2}$

Two limiting cases matter most:

• High temperature ($T \gg \Theta_E$): All oscillators are fully excited, and $C_V \to 3Nk_B$, recovering the classical Dulong-Petit value.
• Low temperature ($T \ll \Theta_E$): The vibrational modes "freeze out" because there isn't enough thermal energy to excite them. $C_V$ drops toward zero exponentially as $e^{-\Theta_E/T}$.

Limitations of the Einstein Model

The exponential decay at low $T$ is the model's main failure. Experiments show that heat capacity actually vanishes as $T^3$, which is a much slower approach to zero than an exponential. The problem is that a single frequency can't capture the fact that real solids have low-frequency acoustic modes that remain active even at very low temperatures. Those low-frequency modes are exactly what the Debye model adds.

Debye Model of Heat Capacity

Assumptions of the Debye Model

Instead of a single frequency, the Debye model treats the solid as a continuous elastic medium supporting sound waves with a linear dispersion relation $\omega = v|\mathbf{k}|$, where $v$ is the speed of sound. To keep the total number of modes equal to $3N$, Debye introduced a maximum cutoff frequency $\omega_D$ (the Debye frequency).

Key assumptions:

• Linear dispersion for all acoustic branches up to $\omega_D$
• A single average sound velocity for all polarizations
• Optical phonons are neglected
• The density of states goes as $g(\omega) \propto \omega^2$ up to the cutoff

Debye Temperature

The Debye temperature is defined analogously to the Einstein temperature:

$\Theta_D = \frac{\hbar\omega_D}{k_B}$

Materials with stiff bonds and light atoms (like diamond, $\Theta_D \approx 2230$ K) have high Debye temperatures. Soft, heavy-atom materials (like lead, $\Theta_D \approx 105$ K) have low ones. Above $\Theta_D$, essentially all phonon modes are thermally excited.

Low-Temperature Limit

At $T \ll \Theta_D$, only the lowest-frequency acoustic phonons carry significant energy. The Debye model predicts:

$C_V = \frac{12\pi^4}{5}Nk_B\left(\frac{T}{\Theta_D}\right)^3$

This is the Debye $T^3$ law, and it matches experimental data very well at low temperatures. The $T^3$ behavior comes directly from the $\omega^2$ density of states combined with the Bose-Einstein distribution at low $T$.

High-Temperature Limit

At $T \gg \Theta_D$, every mode is fully excited and $C_V \to 3Nk_B$, just like the Einstein model and the classical Dulong-Petit result. Both models converge here because the quantum details stop mattering once thermal energy far exceeds the phonon energy scale.

Comparison to the Einstein Model

The Debye model wins at low temperatures because it includes the low-frequency acoustic modes that dominate there. Neither model is perfect for intermediate temperatures in real materials, where the actual phonon density of states can be quite complex.

Phonon Density of States

Acoustic vs. Optical Phonons

If your unit cell has more than one atom, the phonon spectrum splits into acoustic and optical branches:

• Acoustic phonons: Neighboring atoms move roughly in phase. These have low frequencies that go to zero as $k \to 0$. They carry sound through the material. For a 3D crystal, there are 3 acoustic branches (1 longitudinal, 2 transverse).
• Optical phonons: Neighboring atoms move out of phase. These have higher frequencies and a nonzero frequency at $k = 0$. They're called "optical" because in ionic crystals they can couple to light.

The Debye model only captures the acoustic branches. For materials with a significant optical phonon contribution (like polar semiconductors), this is a real limitation.

Debye Approximation

The Debye approximation replaces the true dispersion relation with a simple linear one, $\omega = v|k|$, and cuts off at $\omega_D$. This gives a density of states:

$g(\omega) = \frac{9N}{\omega_D^3}\omega^2 \quad \text{for } \omega \leq \omega_D$

and $g(\omega) = 0$ for $\omega > \omega_D$. The $\omega^2$ form is exact for a 3D isotropic elastic medium at low frequencies. At higher frequencies near the Brillouin zone boundary, the real dispersion curves flatten out and deviate significantly from linearity, so the Debye approximation becomes less accurate there.

Phonon Dispersion Relation

The full phonon dispersion relation $\omega(k)$ maps out how phonon frequency depends on wavevector across the Brillouin zone. It reveals:

• The number and character of phonon branches
• Group velocities ($v_g = d\omega/dk$) that determine heat transport
• Band gaps between acoustic and optical branches
• Van Hove singularities in the density of states

Experimentally, dispersion relations are measured using inelastic neutron scattering (the standard technique) or inelastic X-ray scattering. Raman and infrared spectroscopy probe optical phonons near $k = 0$ but don't map the full dispersion.

Thermal Conductivity

Phonon Scattering Mechanisms

Thermal conductivity in insulating solids is carried almost entirely by phonons. The kinetic theory expression is:

$\kappa = \frac{1}{3}Cv\ell$

where $C$ is the heat capacity per unit volume, $v$ is the average phonon velocity, and $\ell$ is the phonon mean free path. What limits $\ell$ is phonon scattering, which comes from several sources:

• Phonon-phonon scattering: Phonons interact with each other through lattice anharmonicity. This dominates at high temperatures.
• Phonon-boundary scattering: Phonons scatter off the surfaces of the sample. This dominates at very low temperatures in small or nanostructured samples where $\ell$ becomes comparable to the sample size.
• Phonon-defect scattering: Point defects, isotopic mass disorder, and dislocations scatter phonons. This is important at intermediate temperatures and in alloys.

Temperature Dependence of Thermal Conductivity

The thermal conductivity of a crystalline insulator follows a characteristic pattern:

1. Low $T$: $\kappa$ rises steeply (roughly as $T^3$) because the heat capacity is increasing while the mean free path is limited mainly by boundaries (temperature-independent).
2. Intermediate $T$: $\kappa$ reaches a peak. Here the rising heat capacity and the onset of phonon-phonon scattering compete.
3. High $T$: $\kappa$ decreases (roughly as $1/T$) because Umklapp scattering intensifies and shortens the mean free path faster than the heat capacity saturates.

The peak typically occurs somewhere around $\Theta_D/10$ to $\Theta_D/20$.

Umklapp Processes

Not all phonon-phonon collisions resist heat flow. You need to distinguish two types:

• Normal (N) processes: Two phonons collide and produce a third whose wavevector stays inside the first Brillouin zone. Total crystal momentum is conserved, so these don't directly resist heat flow.
• Umklapp (U) processes: The resulting wavevector lands outside the first Brillouin zone and gets folded back by a reciprocal lattice vector $\mathbf{G}$. This effectively reverses the direction of phonon momentum and directly resists thermal transport.

Umklapp processes require phonons with large enough wavevectors to reach the zone boundary. At low temperatures, such high-energy phonons are rare, so Umklapp scattering freezes out exponentially as $\sim e^{-\Theta_D/bT}$ (where $b$ is a constant of order unity). At high temperatures, plenty of large-$k$ phonons exist, making Umklapp scattering the dominant source of thermal resistance.

Applications of Heat Capacity Models

Specific Heat of Metals

In metals, both phonons and conduction electrons contribute to the heat capacity. At low temperatures, the total specific heat takes the form:

$C = \gamma T + AT^3$

• The $\gamma T$ term is the electronic contribution, arising from electrons near the Fermi surface. It dominates at the lowest temperatures.
• The $AT^3$ term is the lattice (Debye) contribution.

By plotting $C/T$ vs. $T^2$, you get a straight line whose intercept gives $\gamma$ (related to the electronic density of states at the Fermi level) and whose slope gives $A$ (related to $\Theta_D$). This is a standard experimental technique.

Thermal Expansion of Solids

The Debye model by itself assumes perfectly harmonic vibrations, which produce no thermal expansion. Real thermal expansion comes from anharmonicity in the interatomic potential: the potential well is asymmetric, so as atoms vibrate more vigorously at higher temperatures, their average position shifts outward.

The Grüneisen parameter $\gamma_G$ connects thermal expansion to the heat capacity:

$\alpha = \frac{\gamma_G C_V}{3B V}$

where $\alpha$ is the volumetric thermal expansion coefficient, $B$ is the bulk modulus, and $V$ is the volume. The Grüneisen parameter measures how phonon frequencies shift with volume ($\gamma_G = -\frac{d \ln \omega}{d \ln V}$) and is roughly constant for many materials (typically 1-3).

Thermal Properties of Semiconductors

At low temperatures, the specific heat of semiconductors follows the Debye $T^3$ law just like insulators, since the lattice contribution dominates and there are very few free carriers. At higher temperatures, the lattice contribution saturates toward $3Nk_B$.

Thermal conductivity in semiconductors tends to be lower than in simple metals for several reasons:

• More complex crystal structures (multiple atoms per unit cell) create optical phonon branches that carry less heat
• Phonon-impurity scattering from dopants reduces the mean free path
• In compound semiconductors, mass disorder between different atomic species adds scattering

These thermal properties directly affect device performance. Heat dissipation is a major design constraint in integrated circuits, and materials like silicon carbide ($\Theta_D \approx 1200$ K, high thermal conductivity) are chosen partly for their superior thermal management.