Debye Model of Phonons

Chill a diamond toward absolute zero and its ability to store heat collapses — not gently, but as the cube of the temperature. Warm it back up and the heat capacity climbs to a flat classical ceiling, the same 3R per mole that Dulong and Petit measured in 1819 for almost every solid. Between those two limits lies one of the cleanest triumphs of quantum statistical mechanics: Peter Debye's 1912 model of a solid as a gas of sound quanta. The idea is to stop tracking individual atoms and instead track the $3N$ collective vibrations — the normal modes — of a crystal of $N$ atoms. Quantize each mode and you get phonons: bosonic quanta whose number is not conserved, so their chemical potential is exactly zero. The lattice is then thermodynamically a photon gas with light replaced by sound. All the physics lives in one function, the density of modes $g(\omega)$ , and Debye's masterstroke was to approximate it by the long-wavelength acoustic law $\omega = c_s k$ carried all the way up to a sharp cutoff frequency $\omega_D$ chosen so the modes number exactly $3N$ . That single approximation reproduces both ends of the curve. At low temperature only the slow, long-wavelength modes are awake, and because their count grows as $\omega^3$ , the excited energy scales as $T^4$ and the heat capacity as $T^3$ — the celebrated Debye law. At high temperature every mode equipartitions and you recover Dulong–Petit. One material number, the Debye temperature $\Theta_D = \hbar\omega_D/k_B$ , sets the crossover and collapses the heat capacity of every insulator onto a single universal curve. This is where the Bose statistics you built for the ideal gas become the everyday thermal physics of the solid world.

Key takeaways

A crystal of $N$ atoms has $3N$ vibrational normal modes; quantizing each gives phonons — bosons with zero chemical potential, so their occupation is the pure Planck factor.
Debye approximates the mode spectrum by a linear acoustic dispersion $\omega = c_s k$ for all three polarizations, filling a sphere in $k$ -space up to a cutoff $\omega_D$ .
The cutoff is fixed by demanding the total mode count equal $3N$ , giving $\omega_D = c_s(6\pi^2 n)^{1/3}$ and the Debye temperature $\Theta_D = \hbar\omega_D/k_B$ .
The density of states is $g(\omega) \propto \omega^2$ — the same power as the photon gas, but with three polarizations and a hard frequency ceiling instead of an unbounded spectrum.
At low $T$ the heat capacity follows the universal $T^3$ law; at high $T$ it saturates at the classical Dulong–Petit value $3Nk_B$ .
Key idea: a solid is a Bose gas of sound quanta — replace its true vibrational spectrum by linear acoustic modes filling a sphere up to $\omega_D$ (fixed by the $3N$ mode count), and its heat capacity becomes a universal function of $T/\Theta_D$ that is $\propto T^3$ at low $T$ and $3Nk_B$ at high $T$ .

The core insight

A crystal's lattice vibrations form a gas of phonons — Bose quanta with $\mu = 0$ . Debye replaces the true mode spectrum by linear acoustic modes filling a $k$ -space sphere up to a cutoff $\omega_D$ set by the total count $3N$ , making the heat capacity a universal function of $T/\Theta_D$ .

Where this lives today

The Debye model is not a museum piece — it is the working baseline of every low-temperature laboratory. When experimentalists measure the specific heat of a new material, they plot $C/T$ against $T^2$ : the intercept gives the electronic Sommerfeld coefficient $\gamma$ and the slope gives the Debye $T^3$ term, cleanly separating the electron and phonon contributions that aq-sm-electronic-heat-capacity-metals and this article respectively own. The Debye temperature extracted this way is tabulated for essentially every crystalline solid and cross-checked against elastic constants. Its reach extends far past heat capacity. The same phonon gas sets the lattice thermal conductivity $\kappa = \tfrac13 C_V v_s \ell$ that governs how heat flows through a semiconductor chip, and phonon engineering — nanostructuring materials to scatter phonons while passing electrons — is how modern thermoelectrics squeeze more waste heat into electricity. The Debye zero-point energy feeds the Lindemann melting criterion and the Grüneisen theory of thermal expansion, and the phonon quasiparticle concept Debye's model helped establish is the template for magnons, plasmons, and every other collective excitation that populates a contemporary condensed-matter paper.

The physical picture

Picture the solid not as a lattice of balls-and-springs but as a box full of standing sound waves. A long-wavelength wave rocks whole planes of atoms slowly — low frequency, easy to excite thermally. A short-wavelength wave wiggles neighbouring atoms against each other — high frequency, and it stays frozen until $k_B T$ climbs to meet $\hbar\omega$ . So at any temperature $T$ , only modes up to a "thermal frequency" $\omega \sim k_B T/\hbar$ are awake, and the rest sit in their quantum ground state contributing nothing to the heat capacity. Why $T^3$ ? Because in three dimensions the number of modes with frequency below $\omega$ grows as the volume of a sphere in $k$ -space, $\propto k^3 \propto \omega^3$ (linear dispersion). The count of thermally awake modes therefore scales as $(k_B T/\hbar c_s)^3 \propto T^3$ ; each carries roughly $k_B T$ of energy, so the vibrational energy above the ground state scales as $T^4$ and its temperature derivative — the heat capacity — as $T^3$ . This is the exact analogue of the photon gas, whose energy density famously goes as $T^4$ . Unlike photons, phonons live on a discrete lattice, and a crystal cannot support a wave shorter than the distance between atoms. That gives a hard ceiling: there are only $3N$ vibrational degrees of freedom, not infinitely many. Debye enforces this by filling exactly enough of $k$ -space to hold $3N$ modes — replacing the true Brillouin zone (an awkward polyhedron) with a sphere of the same volume, cut off at the Debye frequency $\omega_D$ .

Figure

Two Bose gases side by side — the photon spectrum climbs forever, the Debye phonon spectrum is cut off at ω_D

The Debye temperature $\Theta_D$ is the single knob that tunes a material between quantum and classical behaviour. Stiff, light-atom solids have a high sound speed and pack their modes to high frequency, so $\Theta_D$ is large and they stay quantum (sub-Dulong–Petit) far above room temperature: diamond has $\Theta_D \approx 2230$ K. Soft, heavy solids have low $\Theta_D$ and reach the classical ceiling near room temperature: lead has $\Theta_D \approx 105$ K. Copper sits between at $\Theta_D \approx 343$ K.

Figure

The universal Debye curve — every solid's heat capacity collapses onto one line versus T/Θ_D

The core equations

Every phonon lives in a single vibrational mode of the crystal, and because phonon number is not conserved the mode occupation is the pure Planck factor with no chemical-potential term. This is exactly the boson occupation of the ideal quantum gas evaluated at $\mu = 0$ , the same simplification the photon gas enjoys.

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

where: $\langle n\rangle$ — mean number of phonons in a mode of frequency $\omega$ ; $\hbar$ — reduced Planck constant; $k_B$ — Boltzmann constant; $T$ — temperature. The vanishing chemical potential is the signature of a gas whose particles can be created and destroyed at will: at fixed $T,V$ equilibrium sets $\mu = (\partial F/\partial N)_{T,V} = 0$ , exactly as for blackbody photons in aq-sm-photon-gas-planck-radiation. To turn occupations into thermodynamics we need the density of modes. Counting wavevectors in $k$ -space — one mode per volume $(2\pi)^3/V$ from periodic boundary conditions, times three acoustic polarizations, converted to a frequency variable through the linear dispersion $\omega = c_s k$ — gives a spectrum that rises as $\omega^2$ .

g(\omega)\,\mathrm{d}\omega = \frac{3V}{2\pi^2 c_s^3}\,\omega^2\,\mathrm{d}\omega,\qquad 0\le\omega\le\omega_D \qquad (2)

where: $g(\omega)$ — number of modes per unit frequency; $V$ — crystal volume; $c_s$ — polarization-averaged sound speed; $\omega_D$ — Debye cutoff; the factor $3$ counts the three polarizations. The $\omega^2$ power is the fingerprint of a three-dimensional gas with linear dispersion — sharply different from a free particle, whose $\varepsilon \propto k^2$ gives $g(\varepsilon) \propto \sqrt{\varepsilon}$ (the machinery owned by aq-sm-density-of-states-free-particles). Same box-counting geometry, different dispersion, different power. The effective sound speed folds the one longitudinal and two transverse branches into a single number through a harmonic-cube average, because it is $c_s^{-3}$ that appears in the mode count.

\frac{1}{c_s^3} = \frac{1}{3}\left(\frac{1}{c_L^3}+\frac{2}{c_T^3}\right) \qquad (3)

where: $c_L$ — longitudinal sound speed; $c_T$ — transverse sound speed (two branches). This is not cosmetic: using the longitudinal speed alone badly overestimates $\Theta_D$ , because the slower transverse branches dominate the mode count (a lesson we return to in the worked example). The cutoff frequency is fixed by the one physical constraint Debye imposes — that the modes number exactly $3N$ , the true count of vibrational degrees of freedom.

\int_0^{\omega_D} g(\omega)\,\mathrm{d}\omega = 3N \;\Rightarrow\; \omega_D = c_s\left(6\pi^2 \frac{N}{V}\right)^{1/3} \qquad (4)

where: $N$ — number of atoms; $n = N/V$ — atomic number density; $\omega_D$ — Debye frequency. The corresponding Debye wavevector is $k_D = (6\pi^2 n)^{1/3}$ , comparable to the Brillouin-zone radius $\pi/a$ — confirmation that the sphere is a sensible replacement for the true zone. Dividing by $k_B$ turns the cutoff into a temperature — the single material parameter of the model.

\Theta_D \equiv \frac{\hbar\omega_D}{k_B} = \frac{\hbar c_s}{k_B}\left(6\pi^2 n\right)^{1/3} \qquad (5)

where: $\Theta_D$ — Debye temperature, the crossover scale between the quantum ( $T^3$ ) and classical (Dulong–Petit) regimes. The internal energy follows by summing $\hbar\omega\langle n\rangle$ over the spectrum, plus a temperature-independent zero-point term $U_0 = \tfrac{9}{8}Nk_B\Theta_D$ .

U(T) = \frac{9}{8}Nk_B\Theta_D + \int_0^{\omega_D}\frac{\hbar\omega}{e^{\hbar\omega/k_B T}-1}\,g(\omega)\,\mathrm{d}\omega \qquad (6)

where: the first term is the ground-state energy of all the oscillators; the integral is the thermal energy that produces the heat capacity. Differentiating with respect to $T$ and substituting $x = \hbar\omega/k_B T$ collapses everything into the Debye master formula.

C_V = 9N k_B\left(\frac{T}{\Theta_D}\right)^3\int_0^{x_D}\frac{x^4 e^x}{(e^x-1)^2}\,\mathrm{d}x,\quad x_D=\frac{\Theta_D}{T} \qquad (7)

where: $x = \hbar\omega/k_B T$ ; $x_D = \Theta_D/T$ is the dimensionless cutoff. This one equation carries the entire heat-capacity curve; its two limits are the payoff, derived below.

Derivation AThe low-temperature T-cubed law

Our goal is to extract the universal cubic heat capacity that governs every insulator for $T \ll \Theta_D$ , showing that it depends on the material only through $\Theta_D$ and computing the exact numerical coefficient. The strategy is to recognize that at low temperature the dimensionless cutoff $x_D = \Theta_D/T$ runs to infinity because the integrand decays exponentially, so the finite Debye integral becomes a standard $\Gamma\zeta$ integral over the whole half-line. The one subtlety to watch is the integration by parts that converts the heat-capacity integrand into the simpler Bose form — dropping the resulting factor of four is the classic error that turns $12\pi^4/5$ into $3\pi^4/5$ .

Step 1. Start from the master formula (7), $C_V = 9Nk_B (T/\Theta_D)^3 \int_0^{x_D} x^4 e^x/(e^x-1)^2\,\mathrm{d}x$ . Why: this is the exact Debye expression; everything below is the evaluation of its low-temperature limit.

Step 2. Extend the upper limit to infinity. For $T \ll \Theta_D$ we have $x_D = \Theta_D/T \to \infty$ , and the integrand behaves as $x^4 e^{-x}$ for large $x$ , so pushing the limit to $\infty$ introduces only exponentially small error. Why: physically, the cutoff modes near $\omega_D$ are frozen out at low $T$ — only the long-wavelength acoustic modes matter, and those know nothing of the ceiling.

Step 3. Integrate by parts and evaluate. With $u = x^4$ and $\mathrm{d}v = e^x/(e^x-1)^2\,\mathrm{d}x$ (so $v = -1/(e^x-1)$ ), the boundary terms vanish and

\int_0^\infty\frac{x^4 e^x}{(e^x-1)^2}\,\mathrm{d}x = 4\int_0^\infty\frac{x^3}{e^x-1}\,\mathrm{d}x = 4\,\Gamma(4)\zeta(4) = 4\cdot 6\cdot\frac{\pi^4}{90} = \frac{4\pi^4}{15}.

Why: the identity $\int_0^\infty x^{s-1}/(e^x-1)\,\mathrm{d}x = \Gamma(s)\zeta(s)$ (expand $1/(e^x-1) = \sum_{n\ge1}e^{-nx}$ and integrate term by term) with $s=4$ , $\Gamma(4)=6$ , $\zeta(4)=\pi^4/90$ . This is the very same Bose integral the photon gas uses for the Stefan–Boltzmann law.

Step 4. Collect the coefficient.

C_V = 9Nk_B\left(\frac{T}{\Theta_D}\right)^3\cdot\frac{4\pi^4}{15} = \frac{12\pi^4}{5}Nk_B\left(\frac{T}{\Theta_D}\right)^3 \approx 233.8\,Nk_B\left(\frac{T}{\Theta_D}\right)^3.

Why: $9 \times 4/15 = 36/15 = 12/5$ , giving the famous prefactor $12\pi^4/5 \approx 233.78$ . The full derivation of this coefficient and its material collapse is owned by aq-sm-debye-temperature-t-cubed-law; here we establish it as the low- $T$ limit of the model. *Verified in verify_debye-model-phonons.py: $\int_0^\infty x^3/(e^x-1)\,\mathrm{d}x = \Gamma(4)\zeta(4) = 6\pi^4/90 = \pi^4/15$ (symbolic and numeric), the integration-by-parts identity giving $4\pi^4/15$ for the $C_V$ integrand, and $9\cdot 4/15 = 12/5$ .*

Go deeper · Derivation B · 1 minHigh-temperature limit and Dulong–Petit

Derivation BHigh-temperature limit and Dulong–Petit

Our goal is to show that the Debye model correctly recovers the classical equipartition heat capacity $3Nk_B$ at high temperature, and to compute the leading quantum correction that first pulls $C_V$ below the classical ceiling. The strategy is opposite to the low- $T$ case: now $T \gg \Theta_D$ makes $x = \hbar\omega/k_B T \ll 1$ over the entire integration range $[0, x_D]$ , so we expand the integrand as a power series in small $x$ and integrate term by term. The payoff is not just the leading $3Nk_B$ but the specific $-(\Theta_D/T)^2/20$ correction, which tells experimenters how close to classical their sample is at a given temperature.

Step 1. Expand the integrand for small $x$ . Using $e^x/(e^x-1)^2 = 1/[4\sinh^2(x/2)] = x^{-2}(1 - x^2/12 + \cdots)$ ,

\frac{x^4 e^x}{(e^x-1)^2} = x^2\left(1 - \frac{x^2}{12} + \cdots\right).

Why: at high $T$ every mode is nearly classical; the small- $x$ expansion is the systematic quantum correction to equipartition.

Step 2. Integrate term by term up to the cutoff $x_D$ .

\int_0^{x_D}\left(x^2 - \frac{x^4}{12}\right)\mathrm{d}x = \frac{x_D^3}{3} - \frac{x_D^5}{60}.

Why: elementary polynomial integration; the cutoff now matters because at high $T$ all modes are excited, right up to $\omega_D$ .

Step 3. Multiply by the prefactor $9Nk_B x_D^{-3}$ .

C_V = 9Nk_B\,x_D^{-3}\left(\frac{x_D^3}{3} - \frac{x_D^5}{60}\right) = 3Nk_B\left(1 - \frac{x_D^2}{20}\right), \qquad x_D = \frac{\Theta_D}{T}.

Why: the leading term $3Nk_B$ is exactly Dulong–Petit — each of the $3N$ modes contributes $k_B$ (equipartition: $\tfrac12 k_B$ kinetic $+\tfrac12 k_B$ potential). The correction $-(\Theta_D/T)^2/20$ is the first sign that the highest-frequency modes are not yet fully thawed. *Verified in verify_debye-model-phonons.py: the $\sinh$ small- $x$ expansion $x^4 e^x/(e^x-1)^2 = x^2 - x^4/12 + O(x^6)$ (symbolic series), and that the leading heat capacity equals $3Nk_B$ with the $-x_D^2/20$ correction.*

Limiting cases

The Debye heat capacity is a universal function of $x_D = \Theta_D/T$ ; its two ends and its ground-state energy are exactly computable.

Case	Result	Meaning
Low temperature $T\ll\Theta_D$	$C_V \simeq \dfrac{12\pi^4}{5}\,N k_B\left(\dfrac{T}{\Theta_D}\right)^3$	Only linear acoustic modes excited; the Debye integral runs to $\infty$ giving $4\pi^4/15$ . The hallmark $T^3$ law.
High temperature $T\gg\Theta_D$	$C_V \simeq 3N k_B\left[1 - \dfrac{1}{20}\left(\dfrac{\Theta_D}{T}\right)^2 + \cdots\right]$	All $3N$ modes equipartition: the Dulong–Petit ceiling $3Nk_B$ , with the leading quantum correction.
Zero-point energy $T\to 0$	$U(0) = \dfrac{9}{8}N k_B \Theta_D$	Sum of $\tfrac12\hbar\omega$ over all modes with Debye $g(\omega)$ ; a real, finite ground-state energy that contributes nothing to $C_V$ .
Excited phonon number, low $T$	$N_{\rm ph} \propto T^3$ , coefficient $2\zeta(3)$	$N_{\rm ph} = \int g\langle n\rangle\,\mathrm{d}\omega \propto T^3\int_0^\infty x^2/(e^x-1)\,\mathrm{d}x = 2\zeta(3)\,T^3$ ; the same cube that drives the heat capacity.
Einstein-model contrast	$C_V^{\rm E} \propto (\Theta_E/T)^2 e^{-\Theta_E/T}$	All modes at one frequency give an exponential low- $T$ freeze-out — too fast. Debye's $T^3$ is the correct low- $T$ law (owned by aq-sm-einstein-model-heat-capacity).

Common trapThe Debye temperature is a fixed constant of the material

It is not — the "effective" $\Theta_D$ that best fits data drifts with temperature, and this is expected, not a failure. The Debye parabola $g \propto \omega^2$ is only an approximation to the real phonon density of states, which carries van Hove singularities and optical branches the model ignores. A fit at each temperature probes a different part of the spectrum, so the extracted $\Theta_D$ typically dips at intermediate $T$ before settling. The physically meaningful Debye temperature is the one read from the genuine $T \to 0$ , $T^3$ regime; it agrees with the value computed from elastic constants. Two more traps: the phonon chemical potential is exactly zero, not merely small; and the Debye sphere (one cutoff capturing all $3N$ modes) is a completely different construction from the Einstein model (all modes crammed at one frequency).

Worked examples

1 · The Debye temperature of copper — and a base-model trap (confidence builder)

Step 1. Copper has atomic number density $n = 8.47\times10^{28}\ \mathrm{m^{-3}}$ . Its Debye wavevector is $k_D = (6\pi^2 n)^{1/3} = (6\pi^2 \times 8.47\times10^{28})^{1/3} = 1.71\times10^{10}\ \mathrm{m^{-1}}$ .

Step 2. The right sound speed is the polarization average of (3), not the longitudinal speed. Copper has $c_L = 4760\ \mathrm{m/s}$ and $c_T = 2325\ \mathrm{m/s}$ , so

c_s = \left[\tfrac13\left(c_L^{-3} + 2c_T^{-3}\right)\right]^{-1/3} = 2612\ \mathrm{m/s}.

The transverse branches, being slower, dominate the average and pull $c_s$ well below $c_L$ .

Step 3. The Debye frequency and temperature follow:

\omega_D = c_s k_D = 2612 \times 1.71\times10^{10} = 4.47\times10^{13}\ \mathrm{rad/s},\qquad \Theta_D = \frac{\hbar\omega_D}{k_B} = \frac{(1.055\times10^{-34})(4.47\times10^{13})}{1.381\times10^{-23}} = \boxed{341\ \mathrm{K}}.

This matches the measured copper value $\Theta_D \approx 343$ K to better than 1%.

Step 4. The trap — and a corrected base error. If you had (as some references loosely do) plugged a single "typical" sound speed of $3600\ \mathrm{m/s}$ into $\Theta_D = \hbar c_s(6\pi^2 n)^{1/3}/k_B$ , you would get $\Theta_D = 471\ \mathrm{K}$ , badly wrong. The base draft of this article carried exactly that slip — it quoted $v_s \approx 3600\ \mathrm{m/s}$ and claimed $\approx 340$ K, which is internally inconsistent (3600 m/s gives 471 K). The resolution is that $c_s$ in the Debye formula must be the transverse-dominated polarization average, $2612\ \mathrm{m/s}$ for copper — which does reproduce the measured 343 K.

2 · Low-temperature heat capacity of diamond versus copper (exam level)

Step 1. Use the $T^3$ law $C_V = (12\pi^4/5)Nk_B(T/\Theta_D)^3$ with $12\pi^4/5 = 233.8$ . Take one mole, $N = N_A = 6.022\times10^{23}$ , so $Nk_B = R = 8.314\ \mathrm{J\,mol^{-1}K^{-1}}$ .

Step 2. Copper at $T = 20\ \mathrm{K}$ , $\Theta_D = 343\ \mathrm{K}$ : $(T/\Theta_D)^3 = (20/343)^3 = 1.98\times10^{-4}$ , so

C_V = 233.8 \times 8.314 \times 1.98\times10^{-4} = \boxed{0.385\ \mathrm{J\,mol^{-1}K^{-1}}}.

Step 3. Diamond at the same $T = 20\ \mathrm{K}$ , $\Theta_D = 2230\ \mathrm{K}$ : $(20/2230)^3 = 7.2\times10^{-7}$ , so

C_V = 233.8 \times 8.314 \times 7.2\times10^{-7} = \boxed{1.4\times10^{-3}\ \mathrm{J\,mol^{-1}K^{-1}}}.

Step 4. The ratio is $(\Theta_D^{\rm Cu}/\Theta_D^{\rm dia})^3 = (343/2230)^3 = 3.6\times10^{-3}$ , i.e. diamond stores about 270 times less vibrational heat than copper at 20 K. Its stiff, light-atom lattice pushes almost every mode above $k_B T$ — diamond is "classically frozen" at temperatures where copper is already stirring. This single factor is why $\Theta_D$ is the organizing number of low-temperature solid-state thermal physics.

Origins · 1 minPhonon and how this was discovered

Phonon

quasiparticle of lattice vibration · Tamm 1932, Frenkel

The word phonon — a quantum of lattice vibration, by analogy with the photon — was coined by Igor Tamm around 1932 and popularized by Yakov Frenkel, though the physics traces to Einstein's 1907 quantized oscillators and Debye's 1912 acoustic spectrum. A phonon is a collective excitation, not a particle you can isolate: it is a single quantum of energy $\hbar\omega$ shared across a normal mode that ripples through the whole crystal. Treating these quasiparticles as a Bose gas turns the forbidding many-body problem of $3N$ coupled atoms into the tractable statistics of independent modes — the same conceptual move that makes the photon gas manageable. Phonons carry heat and sound, scatter electrons (electrical resistance), and pair them into Cooper pairs (conventional superconductivity); the quasiparticle idea they pioneered underlies nearly all of modern condensed-matter physics.

History

1819

— Dulong and Petit measure the molar heat capacity of solids as ≈ 3R, temperature-independent — the classical law the quantum theory must eventually recover.

1907

— Einstein builds the first quantum model of a solid: 3N identical oscillators at one frequency, explaining qualitatively why C_V drops at low T.

1912

— Debye replaces the single frequency with a linear acoustic spectrum cut off at ω_D, deriving the correct T³ law.

1912

— Born and von Kármán formulate the full lattice-dynamics theory with periodic boundary conditions, the rigorous framework Debye approximated.

1932

— Tamm introduces the term "phonon" for a quantum of lattice vibration; Frenkel spreads the quasiparticle language.

Going further

Texts. Pathria & Beale, Statistical Mechanics, 3rd ed., §7.4 · Kardar, Statistical Physics of Particles, Ch. 7 · Reif, Fundamentals of Statistical and Thermal Physics, §10.1–10.2 · Ashcroft & Mermin, Solid State Physics, Ch. 23 · Kittel, Introduction to Solid State Physics, Ch. 5.

Provenance. All standard results: Bose occupation with $\mu = 0$ , Debye DOS $g \propto \omega^2$ with three polarizations, $\omega_D = c_s(6\pi^2 n)^{1/3}$ , the $C_V$ master integral, the low- $T$ coefficient $12\pi^4/5 \approx 233.8$ , the Dulong–Petit limit $3Nk_B$ with the $-(\Theta_D/T)^2/20$ correction, and the zero-point energy $\tfrac98 Nk_B\Theta_D$ . A verifier should recompute: (i) $\int_0^\infty x^3/(e^x-1)\,\mathrm{d}x = \Gamma(4)\zeta(4) = \pi^4/15$ and the integration-by-parts giving the $C_V$ integral $= 4\pi^4/15$ ; (ii) the small- $x$ expansion of $x^4 e^x/(e^x-1)^2$ yielding the high- $T$ correction; (iii) the mode-count normalization fixing $\omega_D$ . The copper $\Theta_D = 341$ K uses the polarization-averaged $c_s = 2612$ m/s from $c_L = 4760$ , $c_T = 2325$ m/s; a base draft's quoted $v_s = 3600$ m/s is internally inconsistent (it yields 471 K) and is corrected here. Tabulated Debye temperatures (diamond 2230 K, Cu 343 K, Pb 105 K) are from Ashcroft & Mermin Table 23.3.

Prerequisites. Quantum Statistics: Occupation Numbers · Density of States for Free Particles · Bose-Einstein Condensation in the Ideal Gas. Leads to. Electronic Heat Capacity of Metals · Sommerfeld Expansion.

Why is the phonon density of states proportional to ω² when the free-particle one goes as √ε?

Both come from the same $k$ -space counting — the number of modes in a shell is $\propto k^2\,\mathrm{d}k$ — but they use different dispersion relations. Phonons have a linear dispersion $\omega = c_s k$ , so $k \propto \omega$ and $\mathrm{d}k \propto \mathrm{d}\omega$ , giving $g(\omega) \propto \omega^2$ . A free particle has $\varepsilon \propto k^2$ , so $k \propto \sqrt{\varepsilon}$ and $\mathrm{d}k \propto \mathrm{d}\varepsilon/\sqrt{\varepsilon}$ , giving $g(\varepsilon) \propto \sqrt{\varepsilon}$ . Same geometry, different power because energy scales differently with wavevector.

What exactly is the Debye approximation approximating?

It replaces the true phonon dispersion — which bends over and flattens at the Brillouin-zone boundary and splits into acoustic and optical branches — with a single linear branch $\omega = c_s k$ carried to a sphere of the same total volume as the real zone. Two idealizations: the dispersion stays linear at all $k$ (wrong near the zone edge), and the Brillouin zone becomes a sphere (wrong in shape). Both are excused because the thermodynamically important low-frequency modes are exactly the long-wavelength acoustic ones the model gets right, and the total count is enforced to be $3N$ .

Why does only the transverse-averaged sound speed give the right Debye temperature?

Because it is $1/c_s^3$ that enters the mode count, and a solid has two transverse branches for every one longitudinal branch. Transverse waves are slower, so $1/c_T^3$ is larger, and the harmonic-cube average $c_s^{-3} = \tfrac13(c_L^{-3} + 2c_T^{-3})$ is dominated by the transverse contribution. Using $c_L$ alone overestimates $c_s$ and hence $\Theta_D$ ; for copper the difference is $4760$ versus the correct $2612$ m/s, which is the difference between a wrong 471 K and the measured 343 K.

Where does the zero-point energy ⅞... actually 9/8 Nk_B Θ_D come from, and why doesn't it affect C_V?

Each mode has a ground-state energy $\tfrac12\hbar\omega$ even at $T = 0$ . Summing over the Debye spectrum, $U_0 = \int_0^{\omega_D}\tfrac12\hbar\omega\,g(\omega)\,\mathrm{d}\omega$ , and with $g = (9N/\omega_D^3)\omega^2$ this integrates to $\tfrac98 N\hbar\omega_D = \tfrac98 Nk_B\Theta_D$ . It is a real, measurable energy (it shifts the lattice's cohesive energy), but because it is temperature-independent it drops out of $C_V = \partial U/\partial T$ entirely. The heat capacity sees only the thermal part of $U$ .

How is the Debye model related to the blackbody photon gas?

They are mathematically almost identical Bose gases with $\mu = 0$ and an $\omega^2$ density of states. The differences are three: phonons have three polarizations (two transverse, one longitudinal) versus the photon's two; phonons travel at the sound speed $c_s$ rather than $c$ ; and — crucially — phonons have a hard cutoff $\omega_D$ because the lattice is discrete, whereas the photon spectrum is unbounded. That cutoff is why the phonon energy saturates at high $T$ (Dulong–Petit) while the photon energy grows forever as $T^4$ . Below the cutoff, at low $T$ , both give the same $T^4$ energy and $T^3$ behaviour in their respective heat contents.

If I fit real data, why does the Debye temperature I extract change with temperature?

Because the real phonon density of states is not a clean $\omega^2$ parabola — it has van Hove peaks, optical branches, and a non-spherical zone. Each temperature weights different frequencies (low $T$ probes the acoustic tail, high $T$ the whole spectrum), so the single-parameter Debye fit compromises differently at each $T$ , and the "effective" $\Theta_D(T)$ dips and rises. The genuine, temperature-independent Debye temperature is the one from the true $T^3$ regime as $T \to 0$ ; it is the value that matches the elastic-constant estimate.