Debye Temperature and the T-cubed Law

Cool almost any crystal toward absolute zero and its heat capacity does not level off at the classical Dulong–Petit plateau — it collapses, and it collapses in a very particular way: as the cube of the temperature. Copper, diamond, argon ice, solid neon: measure $C_V$ at a few kelvin and you find $C_V \propto T^3$ , one of the cleanest power laws in all of solid-state physics. The value of $T$ at which a given crystal "notices" it is getting cold — the scale that separates the classical high-temperature world from the frozen quantum one — is a single number called the Debye temperature $\Theta_D$ . This article takes the phonon gas built in the Debye model of phonons and extracts its two exact thermal limits and the material scale that ties them together. We already have the ingredients: phonons are bosons of zero chemical potential with a linear dispersion $\omega = v_s k$ , a density of states $g(\omega) \propto \omega^2$ , and a hard cutoff at the Debye frequency $\omega_D$ fixed by the demand that the crystal carry exactly $3N$ vibrational modes. What remains is to integrate the Planck energy over that spectrum, differentiate, and read off the physics. Two limits fall out and neither depends on the messy details of the real phonon spectrum. When $T \gg \Theta_D$ every mode is classically active and $C_V \to 3Nk_B$ , the Dulong–Petit law. When $T \ll \Theta_D$ only the long-wavelength acoustic phonons — the ones whose energy quantum stays below $k_BT$ — remain excited, and their number grows as $T^3$ , giving the celebrated $T^3$ law with a universal coefficient $12\pi^4/5$ . Because the low-temperature physics samples only the linear-dispersion part of the spectrum, which the Debye model gets exactly right, the $T^3$ law is asymptotically exact even though the model is a crude approximation everywhere else.

Key takeaways

The Debye temperature $\Theta_D=\hbar\omega_D/k_B=(\hbar v_s/k_B)(6\pi^2n)^{1/3}$ is the characteristic temperature associated with the Debye cutoff frequency $\omega_D$ , which sets the upper limit of the acoustic-phonon spectrum in the Debye model.
For temperatures $T\gg\Theta_D$ , the heat capacity approaches the classical Dulong–Petit limit $C_V=3Nk_B$ , whereas for $T\ll\Theta_D$ it decreases rapidly toward zero.
For $T\ll\Theta_D$ , the Debye model predicts $C_V=\frac{12\pi^4}{5}Nk_B(T/\Theta_D)^3$ , with the universal coefficient $12\pi^4/5\approx233.8$ .
The lattice energy at low $T$ scales as $T^4$ — the solid-state twin of the photon-gas Stefan–Boltzmann law, differing only by a finite mode ceiling.
Because $C_V$ depends on $T$ only through $T/\Theta_D$ , all elemental solids collapse onto one universal Debye curve — a law of corresponding states for heat capacity.
Key idea: one number $\Theta_D$ sets where a crystal crosses from the classical $3Nk_B$ plateau to the quantum $T^3$ collapse, and the low- $T$ coefficient $12\pi^4/5$ is universal because only the $\omega \to 0$ acoustic modes survive the cold.

The core insight

In the Debye model, acoustic phonons obey a linear dispersion relation and a density of states proportional to $\omega^2$ , truncated at the Debye cutoff chosen to give $3N$ vibrational modes. Consequently, the heat capacity is a universal function of $T/\Theta_D$ , approaching $3Nk_B$ for $T\gg\Theta_D$ and $\frac{12\pi^4}{5}Nk_B(T/\Theta_D)^3$ for $T\ll\Theta_D$ .

Where this lives today

The Debye temperature and the low-temperature $T^3$ law are standard tools in cryogenic physics and materials characterization. Separating electronic and lattice contributions to the low-temperature heat capacity of metals and superconductors commonly begins with a plot of $C_V/T$ versus $T^2$ : the slope hands you $\Theta_D$ , the intercept the Sommerfeld coefficient $\gamma$ that measures the electronic density of states at the Fermi level. In the search for superconductors and heavy-fermion materials, an anomalous $\gamma$ or a $\Theta_D$ that shifts with pressure is a first diagnostic. The same physics governs the thermal budget of cryogenic detectors. Bolometers and superconducting sensors used in millikelvin astronomy and dark-matter searches rely on the fact that a crystal's heat capacity collapses as $T^3$ : a tiny absorbed energy produces a large, measurable temperature spike precisely because there is so little heat capacity left to soak it up. Materials with large Debye temperatures, such as silicon, sapphire, and diamond, have lower lattice heat capacities at cryogenic temperatures and are therefore widely used in sensitive low-temperature detectors.

The physical picture

Think of the crystal as a box of sound. Each of the crystal's $3N$ normal vibrational modes behaves as an independent quantum harmonic oscillator, with thermal energy determined by its phonon occupation number. At high temperature $k_BT$ dwarfs every mode's energy quantum $\hbar\omega$ , so all modes are freely excited and each contributes $k_B$ to the heat capacity — that is Dulong–Petit, $C_V = 3Nk_B$ , the classical plateau. The quantum world is invisible here. Now cool the crystal. A mode of frequency $\omega$ can only absorb a quantum if $k_BT$ can afford $\hbar\omega$ ; once $k_BT$ drops below $\hbar\omega$ the mode freezes into its ground state and stops taking up heat. The stiffest, shortest-wavelength modes freeze first. What survives to the lowest temperatures are the long-wavelength sound waves, whose frequency $\omega = v_s k$ runs continuously down to zero — there is always a band of them cheap enough to excite.

Figure

The heat-capacity story on one axis — classical plateau, quantum collapse, and the Debye scale between them

Why $T^3$ and not something else? The thermally active phonons are those with $\hbar\omega \lesssim k_BT$ , i.e. wavevectors $k \lesssim k_BT/\hbar v_s$ . In three dimensions the number of modes inside that little $k$ -sphere scales as its volume, $\propto (k_BT/\hbar v_s)^3 \propto T^3$ , and each active mode contributes about $k_B$ . Multiply and you get $C_V \propto T^3$ . The exponent is the spatial dimension, and the linear dispersion is what makes the accessible modes reach all the way to $\omega = 0$ . Einstein's earlier single-frequency model missed exactly this — with one gapped frequency there are no cheap modes, so his heat capacity dies exponentially instead.

Figure

Why the exponent is 3 — thermally active phonons fill a k-sphere whose volume grows as T³

The Debye temperature $\Theta_D=\hbar\omega_D/k_B$ is defined from the Debye cutoff frequency $\omega_D$ , where the cutoff wavevector is chosen so that the Debye model contains the correct total number of vibrational modes. A stiff, light material like diamond has a high sound speed and $\Theta_D \approx 2230$ K — at room temperature it is still in its quantum regime, which is why diamond's room-temperature heat capacity sits far below the classical value. A softer material such as lead has a much smaller Debye temperature, $\Theta_D\approx105\,\mathrm{K}$ , so at room temperature its heat capacity is already close to the classical limit. Thus $\Theta_D$ reflects the characteristic phonon frequencies of the material and depends on both its elastic properties and atomic mass.

The core equations

The whole thermodynamics of the Debye solid rides on one temperature-dependent quantity: the mean phonon energy summed over the spectrum. We collect the definitions inherited from the Debye model of phonons and then build the heat capacity from them. The dispersion is linear and the cutoff wavevector is fixed by counting: three acoustic polarizations must together supply exactly $3N$ modes.

\omega = v_s\,k,\qquad k_D=\left(6\pi^2\,\frac{N}{V}\right)^{1/3} \qquad (1)

where: $\omega$ — phonon angular frequency; $v_s$ — the Debye-averaged sound speed (the polarization average $3/v_s^3 = 1/v_L^3 + 2/v_T^3$ ); $k$ — wavevector magnitude; $k_D$ — Debye cutoff wavevector; $N/V$ — atomic number density. The cutoff is an effective feature of the Debye model, introduced so that the continuum approximation reproduces the correct total number of lattice vibrational modes. The single material scale is the Debye temperature, the cutoff frequency measured in kelvin.

\Theta_D \equiv \frac{\hbar\omega_D}{k_B} = \frac{\hbar v_s}{k_B}\left(6\pi^2\,\frac{N}{V}\right)^{1/3} \qquad (2)

where: $\Theta_D$ — Debye temperature; $\omega_D = v_s k_D$ — the maximum (Debye) frequency; $\hbar$ — reduced Planck constant; $k_B$ — Boltzmann constant. Read it physically: $\Theta_D$ is the characteristic temperature associated with the Debye cutoff frequency. For $T\gg\Theta_D$ nearly all phonon modes are thermally populated, whereas for $T\ll\Theta_D$ only the long-wavelength acoustic modes remain appreciably occupied. The Debye density of states is $g(\omega)\propto\omega^2$ , with the normalization chosen so that $\int_0^{\omega_D}g(\omega)\,d\omega=3N$ .

g(\omega)\,d\omega = \frac{9N}{\omega_D^3}\,\omega^2\,d\omega,\qquad 0 \le \omega \le \omega_D \qquad (3)

where: $g(\omega)$ — modes per unit angular frequency. The $\omega^2$ is pure three-dimensional mode counting with a linear dispersion; the prefactor $9N/\omega_D^3$ is whatever makes the total come out to $3N$ . This is the same shape as the photon spectrum — the phonon gas is a blackbody with a lid on it. The internal energy is the Planck energy per mode, integrated over the spectrum. With $x \equiv \hbar\omega/k_BT$ everything reduces to a dimensionless integral cut off at $\Theta_D/T$ .

C_V = 9Nk_B\left(\frac{T}{\Theta_D}\right)^{3}\int_0^{\Theta_D/T}\frac{x^4 e^x}{(e^x-1)^2}\,dx \qquad (4)

where: $x = \hbar\omega/k_BT$ ; the upper limit $\Theta_D/T$ is the cutoff in dimensionless units. This is the full Debye heat capacity: everything about temperature enters only through the ratio $\Theta_D/T$ , which is why one curve fits all elements. The integrand $x^4 e^x/(e^x-1)^2$ is the heat-capacity kernel of a single Planck oscillator, weighted by the $\omega^2$ density of modes. The two exact limits are the payoff, and both are read off equation (4) by pushing the cutoff to one extreme.

C_V \xrightarrow{\,T \ll \Theta_D\,} \frac{12\pi^4}{5}Nk_B\left(\frac{T}{\Theta_D}\right)^{3}, \qquad C_V \xrightarrow{\,T \gg \Theta_D\,} 3Nk_B \qquad (5)

where: $12\pi^4/5 \approx 233.78$ is the universal Debye low-temperature coefficient; $3Nk_B$ is the classical Dulong–Petit plateau. The first is the $T^3$ law that survives to $T \to 0$ ; the second is the classical value recovered when the discreteness of energy no longer matters. No material-specific number appears in either coefficient — only $\Theta_D$ , buried in the ratio.

Derivation ALow-temperature T-cubed law from the phonon energy

We derive the low-temperature asymptotic form of the Debye heat capacity. For $T\ll\Theta_D$ , the upper integration limit $\Theta_D/T$ may be replaced by $\infty$ with exponentially small error, reducing the problem to a standard Bose integral whose evaluation yields the universal coefficient $12\pi^4/5$ .

Step 1. Write the internal energy in dimensionless form. Insert $g(\omega) = 9N\omega^2/\omega_D^3$ into $U = \int_0^{\omega_D} \hbar\omega\,g(\omega)/(e^{\hbar\omega/k_BT}-1)\,d\omega$ and substitute $x = \hbar\omega/k_BT$ : $U = 9Nk_BT\,(T/\Theta_D)^3\int_0^{\Theta_D/T} x^3/(e^x-1)\,dx$ . Why: the $\omega^3$ from $\hbar\omega\cdot\omega^2$ becomes $x^3$ , and every factor of $\omega$ carries a $k_BT/\hbar$ , which is what builds the $(T/\Theta_D)^3$ out front.

Step 2. Send the cutoff to infinity and evaluate the Bose integral. For $T \ll \Theta_D$ , $\int_0^{\Theta_D/T} \to \int_0^\infty x^3/(e^x-1)\,dx = \Gamma(4)\zeta(4) = 6\cdot\pi^4/90 = \pi^4/15$ . Why: $\zeta(4) = \pi^4/90$ is exact, and the tail beyond $x \sim \Theta_D/T$ is exponentially suppressed — this is the whole reason the low- $T$ law is clean.

Step 3. Read off the $T^4$ energy. Substituting, $U \simeq \frac{3\pi^4}{5}Nk_B T (T/\Theta_D)^3 = \frac{3\pi^4}{5}\frac{Nk_B}{\Theta_D^3}T^4$ . Why: the lattice energy grows as the fourth power of temperature — precisely the Stefan–Boltzmann scaling of a photon gas, because both are linear-dispersion boson gases.

Step 4. Differentiate to get the heat capacity. $C_V = \partial U/\partial T = 4\cdot\frac{3\pi^4}{5}\frac{Nk_B}{\Theta_D^3}T^3 = \frac{12\pi^4}{5}Nk_B(T/\Theta_D)^3$ . Why: the factor of $4$ from differentiating $T^4$ times $3/5$ gives exactly $12/5$ ; equivalently the direct kernel integral $\int_0^\infty x^4 e^x/(e^x-1)^2\,dx = 4\pi^4/15$ and $9\cdot 4/15 = 12/5$ . The two routes agree, as they must.

Go deeper · Derivation B · 1 minHigh-temperature limit and the approach to Dulong-Petit

Derivation BHigh-temperature limit and the approach to Dulong-Petit

We want to show that the Debye heat capacity approaches the classical $3Nk_B$ from below as $T \to \infty$ , and to extract the leading quantum correction, which turns out to be $-\tfrac{1}{20}(\Theta_D/T)^2$ . The strategy is opposite to the low- $T$ case: now the upper limit $y = \Theta_D/T$ is small, so instead of extending the integral we Taylor-expand the kernel about $x = 0$ and integrate the polynomial term by term. The first term rebuilds Dulong–Petit; the next term is the first place quantum mechanics pokes through at high temperature.

Step 1. Expand the heat-capacity kernel near $x = 0$ . Using $e^x - 1 = x(1 + x/2 + x^2/6 + \cdots)$ one finds $x^4 e^x/(e^x-1)^2 = x^2\,[\,1 - x^2/12 + \cdots\,]$ . Why: the leading $x^2$ is what will integrate to the classical result; the $-x^2/12$ correction is the fingerprint of the discrete energy levels.

Step 2. Integrate term by term up to the small cutoff $y$ . $\int_0^y x^2(1 - x^2/12)\,dx = y^3/3 - y^5/60$ . Why: with $y = \Theta_D/T \ll 1$ only the first couple of powers matter, and each higher power is suppressed by another factor of $y^2$ .

Step 3. Assemble the heat capacity. $C_V = 9Nk_B\,y^{-3}(y^3/3 - y^5/60) = 3Nk_B(1 - y^2/20)$ . Why: the $9\cdot(1/3) = 3$ gives the Dulong–Petit plateau, and $9\cdot(1/60) = 3/20$ divided back through gives the $-1/20$ coefficient — the heat capacity sits just below the classical value and rises toward it as $T$ increases.

Limiting cases

The Debye heat capacity is a single universal function of $y \equiv \Theta_D/T$ :

case	result	meaning
High temperature $T \gg \Theta_D$	$C_V \simeq 3Nk_B\left[1 - \tfrac{1}{20}(\Theta_D/T)^2 + \cdots\right]$	Dulong–Petit: every mode is classically active. The leading correction is negative and quadratic — the crystal approaches the plateau from below.
Low temperature $T \ll \Theta_D$	$C_V \simeq \frac{12\pi^4}{5}Nk_B(T/\Theta_D)^3$	The $T^3$ law. Only the linear-dispersion acoustic branch matters, so this is asymptotically exact for real crystals, not just the model.
Metal at very low $T$	$\dfrac{C_V}{T} = \gamma + \beta T^2,\quad \beta = \frac{12\pi^4}{5}\frac{Nk_B}{\Theta_D^3}$	Electronic ( $\gamma T$ ) plus lattice ( $\beta T^3$ ). A plot of $C_V/T$ vs $T^2$ is a straight line: intercept gives $\gamma$ , slope gives $\Theta_D$ .
Lattice entropy at low $T$	$S = \tfrac{1}{3}C_V = \frac{4\pi^4}{5}Nk_B(T/\Theta_D)^3$	For any pure cubic $C_V = AT^3$ , integrating $C_V/T$ gives $S = \tfrac13 AT^3$ — a clean third-law check.

Worked examples

1 · The Debye temperature of copper from first principles (confidence builder)

Step 1. Get the cutoff wavevector from the atomic density. Copper has $n = N/V \approx 8.5\times10^{28}\ \mathrm{m}^{-3}$ , so $k_D = (6\pi^2 n)^{1/3} = (6\pi^2 \cdot 8.5\times10^{28})^{1/3} = (5.03\times10^{30})^{1/3} = 1.71\times10^{10}\ \mathrm{m}^{-1}$ .

Step 2. Convert to a Debye frequency with the sound speed. The relevant $v_s$ is the Debye-averaged speed, $3/v_s^3 = 1/v_L^3 + 2/v_T^3$ , which for copper works out near $v_s \approx 2600\ \mathrm{m/s}$ — notably slower than the bare longitudinal speed ( $\sim 4700\ \mathrm{m/s}$ ), because the two soft transverse branches dominate the average. Then $\omega_D = v_s k_D = 2600 \times 1.71\times10^{10} = 4.45\times10^{13}\ \mathrm{rad/s}$ .

Step 3. Convert to a temperature. $\Theta_D = \hbar\omega_D/k_B = (1.055\times10^{-34})(4.45\times10^{13})/(1.381\times10^{-23}) = \boxed{340\ \mathrm{K}}$ , in excellent agreement with the measured copper value $\Theta_D \approx 343$ K.

Step 4. Note the calibration trap. Using the naive longitudinal speed $4700$ m/s instead would have given $\Theta_D \approx 615$ K, nearly double the true value. The Debye average — weighting the two slow transverse branches twice — is essential; this is the single most common error in a first-pass Debye calculation.

2 · How frozen is copper at 20 K, and why does diamond stay quantum at room temperature (exam level)

Step 1. Copper at 20 K sits deep in the $T^3$ regime since $T/\Theta_D = 20/343 = 0.058 \ll 1$ . The molar lattice heat capacity is $C_V = \frac{12\pi^4}{5}R\,(T/\Theta_D)^3$ with $R = 8.314\ \mathrm{J\,mol^{-1}K^{-1}}$ : $C_V = 233.8 \times 8.314 \times (0.0583)^3 = 233.8 \times 8.314 \times 1.98\times10^{-4} = \boxed{0.385\ \mathrm{J\,mol^{-1}K^{-1}}}$ .

Step 2. Compare to the classical plateau. Dulong–Petit is $3R = 24.9\ \mathrm{J\,mol^{-1}K^{-1}}$ , so copper at 20 K carries only $0.385/24.9 = 1.5\%$ of its classical lattice heat capacity. The crystal is almost entirely frozen — nearly all its modes cost more than $k_BT$ .

Step 3. Now diamond at room temperature. Diamond's stiff bonds and light carbon atoms give $\Theta_D \approx 2230$ K, so at 300 K, $T/\Theta_D = 0.135$ — still firmly in the quantum regime. Evaluating the Debye function at $T/\Theta_D = 300/2230 \approx 0.135$ gives $C_V/3Nk_B \approx 0.24$ , so diamond has about one quarter of its classical Dulong–Petit heat capacity at room temperature.

Step 4. Contrast with copper at the same 300 K, where $T/\Theta_D = 0.87$ and $C_V/3Nk_B \approx 0.94$ — nearly classical. The universal Debye function, evaluated at different values of $T/\Theta_D$ , explains why diamond remains strongly quantum at room temperature whereas copper is already close to the classical limit. The difference arises from their very different Debye temperatures. This is exactly the puzzle — diamond's anomalously low heat capacity — that launched quantum theories of solids in 1907.

Origins · 1 mindebye temperature and how this was discovered

debye temperature

1884–1966 · Peter Debye, Maastricht → Zürich → Cornell

The Debye temperature carries the name of Peter Debye, who in 1912 replaced Einstein's single vibrational frequency with a spectrum of acoustic modes running down to zero frequency and so produced the correct $T^3$ low-temperature law. The scale $\Theta_D$ collapses every element's heat-capacity curve onto one universal function — a law of corresponding states for solids — and remains the standard single-number summary of a crystal's lattice dynamics. Debye's broader career ranged over dipole moments, X-ray scattering from powders, and light scattering; he received the 1936 Nobel Prize in Chemistry.

History

1819

— Dulong and Petit find that solid elements share a molar heat capacity near

3R

, the classical plateau.

1907

— Einstein models a solid as

3N

identical oscillators of one frequency, the first quantum theory of heat capacity, but predicts too steep a low-

T

falloff.

1912

— Debye replaces the single frequency by an acoustic spectrum with a cutoff, deriving the

T^3

law and the Debye temperature.

1928

— Sommerfeld's free-electron theory supplies the metallic linear

\gamma T

term that adds to the lattice

T^3

1930

s — low-temperature calorimetry confirms

C_V/T = \gamma + \beta T^2

for metals, letting

\Theta_D

and

\gamma

be read from one straight line.

Going further

Texts. Pathria & Beale, Statistical Mechanics, 3rd ed., §7.4 (Phonons and the Debye theory) · Kardar, Statistical Physics of Particles, Ch. 7 · Reif, Fundamentals of Statistical and Thermal Physics, §10.1–10.2 · Sethna, Statistical Mechanics: Entropy, Order Parameters, and Complexity, Ch. 7 · Ashcroft & Mermin, Solid State Physics, Ch. 23 · Landau & Lifshitz, Statistical Physics Part 1, §64–66. Key numerical benchmarks: the Bose integral $\int_0^\infty x^3/(e^x-1)\,dx = \pi^4/15$ ; the kernel integral $\int_0^\infty x^4 e^x/(e^x-1)^2\,dx = 4\pi^4/15$ ; the low- $T$ coefficient $12\pi^4/5 \approx 233.78$ ; the high- $T$ correction $-1/20$ ; $k_D = (6\pi^2 n)^{1/3}$ from mode counting; copper $\Theta_D \approx 340$ –343 K (with the Debye-averaged $v_s \approx 2600$ m/s); the low- $T$ entropy relation $S = \tfrac13 C_V$ . Representative $\Theta_D$ values: Cu 343 K, Al 428 K, Pb 105 K, diamond ~2230 K.

Prerequisites: Quantum Statistics of Ideal Gases: Occupation Numbers · Density of States for Free Particles. Leads to: Electronic Heat Capacity of Metals · Sommerfeld Expansion.

Where does the factor 9 in the density of states come from

It is the sum-rule normalization. Writing $g(\omega) = A\omega^2$ and demanding $\int_0^{\omega_D} A\omega^2\,d\omega = 3N$ gives $A\,\omega_D^3/3 = 3N$ , so $A = 9N/\omega_D^3$ . The $3$ inside the total ( $3N$ ) is the three acoustic polarizations per atom; the extra factor comes from integrating $\omega^2$ . So the $9$ is just $3\times 3$ : three polarizations, and a $3$ from the cubic integral of $\omega^2$ .

Why is the low-temperature coefficient exactly 12π⁴/5 and not material-dependent

Because at low $T$ only the linear-dispersion acoustic modes matter, and those are described exactly by the Debye model — the cutoff and the spectrum's high-frequency details have receded to irrelevance. The coefficient comes purely from the dimensionless Bose integral $\int_0^\infty x^3/(e^x-1)\,dx = \pi^4/15$ and the mode-count prefactor, neither of which knows about the material. All the material information is sequestered inside $\Theta_D$ ; the coefficient in front of $(T/\Theta_D)^3$ is a pure number.

How is the Debye temperature actually measured

The cleanest route in a metal is the $C_V/T$ versus $T^2$ plot at liquid-helium temperatures. Since $C_V = \gamma T + \beta T^3$ , dividing by $T$ gives a straight line $C_V/T = \gamma + \beta T^2$ . The slope $\beta = \frac{12\pi^4}{5}Nk_B/\Theta_D^3$ inverts to $\Theta_D$ ; the intercept $\gamma$ is the electronic term. In insulators there is no $\gamma$ term, so a $\log C_V$ versus $\log T$ plot showing slope 3 confirms the law and the prefactor fixes $\Theta_D$ directly.

Is the T-cubed law exact or just a model artifact

The $T^3$ law itself is asymptotically exact for any real crystal, not a Debye-model artifact — because as $T \to 0$ the only excited modes are long-wavelength acoustic phonons with a genuinely linear dispersion, which every crystal has. What is model-dependent is the value of $\Theta_D$ and everything at intermediate temperatures, where the true phonon spectrum with its Van Hove singularities differs from the smooth $\omega^2$ Debye form. That is why an "effective $\Theta_D(T)$ " that drifts slightly with temperature is often quoted.

How can phonons and photons share the same T⁴ energy law if solids are so different from a vacuum

Both are gases of bosons with zero chemical potential and a linear dispersion $\omega = v\,k$ — for photons $v = c$ , for phonons $v = v_s$ . The low-temperature energy of any such gas scales as $T^4$ , and the heat capacity as $T^3$ , entirely from the geometry of counting modes in a $T$ -dependent $k$ -sphere. The only structural difference is that phonons have a finite mode ceiling (there are exactly $3N$ of them, hence $\omega_D$ and $\Theta_D$ ), so their $T^4$ law saturates into Dulong–Petit once $T \gtrsim \Theta_D$ , whereas photons have no ceiling and their $T^4$ law holds at all temperatures.

Why is the correct Debye sound speed slower than the longitudinal speed

Because the Debye average weights the soft transverse branches twice: $3/v_s^3 = 1/v_L^3 + 2/v_T^3$ . A crystal supports one longitudinal acoustic branch (fast) and two transverse branches (slow), and the density of states counts all three. Since the average is over $1/v^3$ , the slow transverse speeds dominate — mathematically, the harmonic-like average is pulled toward the smaller values. Using the longitudinal speed alone typically overestimates $\Theta_D$ by a large factor, as the copper worked example shows.

Does the zero-point energy of the phonons affect the heat capacity

No. Each mode carries a ground-state energy $\tfrac12\hbar\omega$ , and summed over the spectrum this is a large but strictly temperature-independent contribution to $U$ . Since $C_V = (\partial U/\partial T)_V$ , any $T$ -independent piece differentiates to zero. The zero-point energy matters for the total lattice energy, the equation of state, and phenomena like the isotope effect on lattice constants, but it never appears in $C_V$ .