Density of States in k-Space

A free quantum particle confined to a box cannot carry arbitrary momentum: the boundary conditions select a discrete set of allowed wavevectors $\mathbf{k}$ . Counting these modes is the foundational bookkeeping of quantum statistical mechanics, because nearly every thermodynamic quantity of an ideal gas — pressure, energy, particle number — is a sum over single-particle modes weighted by an occupation factor. The observation that makes the sum tractable is geometric and strikingly simple: the allowed $\mathbf{k}$ sit on a uniform cubic lattice in reciprocal space, one state per cell of volume $(2\pi/L)^3$ . We take the microscopic spectrum from quantum mechanics. In aq-qm-particle-in-3d-box, solving the free Schrödinger equation in a cube gives product standing waves $\psi_{n_1 n_2 n_3} \propto \sin(n_1\pi x/L)\cdots$ whose energy "depends only on the sum $n_1^2 + n_2^2 + n_3^2$ ," so that "states of equal energy therefore lie on a sphere of fixed radius $\sqrt{N}$ " — the "degeneracy" that the QM article identifies as "the central new phenomenon of the three-dimensional problem," where "counting degeneracy becomes counting lattice points on a sphere." That article even notes that the count below $E$ "scales as the volume, giving a density of states $\propto V\sqrt{E}$ — the entry point to the free-electron model of metals." What this statistical-mechanics article owns is the bulk mode-counting rule: the uniform reciprocal-space density $V/(2\pi)^3$ , and the replacement $\sum_{\mathbf{k}} \to \frac{V}{(2\pi)^3}\int d^3k$ that turns quantum discreteness into the continuous integrals of the grand canonical ensemble. Because the cells are uniform and tiny for macroscopic $L$ , the spectrum is effectively a continuum. This single substitution underlies the Bose–Einstein and Fermi–Dirac thermodynamics, the blackbody spectrum, the Debye theory of solids, and the Sommerfeld theory of metals. The energy-space companion $g(\varepsilon)$ — obtained by folding in a dispersion relation — is owned by aq-sm-density-of-states-energy; here we build the raw $k$ -space lattice, its universal density, and the shell $4\pi k^2\,dk$ that converts it into an energy count.

Key takeaways

We take from QM (aq-qm-particle-in-3d-box): the box spectrum, standing waves $\sin(n_i\pi x/L)$ , and the degeneracy in which "states of equal energy lie on a sphere of radius $\sqrt{N}$ ."
A particle in a box of volume $V = L^3$ has wavevectors quantized so that (with periodic boundary conditions) allowed $\mathbf{k}$ form a uniform cubic lattice with spacing $2\pi/L$ .
Each mode occupies a $k$ -space cell $(2\pi/L)^3 = (2\pi)^3/V$ , so the density of modes is the constant $V/(2\pi)^3$ , independent of $\mathbf{k}$ and of the dispersion relation.
This converts sums into integrals: $\sum_{\mathbf{k}} \to \frac{V}{(2\pi)^3}\int d^3k = \frac{V}{h^3}\int d^3p$ — the workhorse of ideal quantum gases.
An isotropic energy count uses a spherical shell $4\pi k^2\,dk$ ; spin (or polarization) multiplies by a degeneracy factor $g_s = 2s+1$ .
Key idea: in a box of volume $V$ , allowed wavevectors form a uniform lattice with one mode per cell $(2\pi)^3/V$ ; the density of modes is therefore the constant $V/(2\pi)^3$ , and $\sum_{\mathbf{k}} \to \frac{V}{(2\pi)^3}\int d^3k$ — the sum-to-integral bridge for every quantum gas.

The core insight

In a box of volume $V$ , allowed wavevectors form a uniform lattice in $k$ -space with one mode per cell $(2\pi)^3/V$ ; hence the density of modes is the constant $V/(2\pi)^3$ , and $\sum_{\mathbf{k}} \to \frac{V}{(2\pi)^3}\int d^3k$ . The microscopic box spectrum is owned by aq-qm-particle-in-3d-box; SM owns the bulk counting rule and its continuum limit.

Where this lives today

The uniform $k$ -space density is the first line of essentially every ideal-gas calculation in condensed matter and statistical physics. Filling the lattice with electrons up to the Fermi wavevector defines the Fermi sphere that fixes the electronic properties of every metal; wrapping the shell around photon modes gives the blackbody spectrum and the cosmic microwave background; the same counting with a Debye cutoff yields the low-temperature $T^3$ heat capacity of solids. In modern nanoscience the dimensional version matters directly: quantum wells (2D, constant density of states), quantum wires (1D, van Hove singularities), and quantum dots (0D, discrete levels) engineer the density of states by confining electrons in fewer dimensions, and the whole discussion starts from the mode-counting rule built here.

The physical picture

Imagine reciprocal space tiled by a grid of tiny boxes, each of side $2\pi/L$ . Every quantum state of the particle sits at exactly one grid point, and the grid is perfectly uniform — no region of $k$ -space is privileged. Counting states then reduces to measuring a volume and dividing by the cell size. This uniformity is the whole content of the $k$ -space density of states, and it is why the seemingly hard problem of summing over a quantum spectrum becomes an easy integral. The factor $(2\pi)^3/V$ per state is the $k$ -space analogue of the phase-space cell $h^3$ per state in classical mechanics: multiplying by $\hbar^3 = (h/2\pi)^3$ converts the $k$ -space measure $d^3k$ into the momentum measure $d^3p/h^3$ , recovering the familiar phase-space counting. Quantum mechanics thus reproduces — and justifies — the classical $h^3$ cell that Sackur and Tetrode had to postulate. The energy dependence people care about lives entirely in the conversion: states of energy near $\varepsilon$ lie on a constant-energy surface, and a spherical shell of that surface holds a $k$ -space volume $4\pi k^2\,dk$ , which is where the eventual $\sqrt{\varepsilon}$ growth of the energy density of states will come from — worked out in aq-sm-density-of-states-energy.

Figure

The allowed wavevectors form a uniform cubic lattice in k-space with spacing 2π/L (periodic boundary conditions). Each lattice point is one quantum state and occupies a cell of volume (2π/L)³, so the density of modes is the constant V/(2π)³ everywhere. An isotropic energy count sums the states in a thin spherical shell of k-space volume 4πk² dk.

The picture that carries the whole subject: allowed modes are a uniform lattice, so counting is just volume ÷ cell. An isotropic energy count wraps a spherical shell $4\pi k^2\,dk$ around the origin and multiplies by the constant density $V/(2\pi)^3$ .

Figure

Periodic versus hard-wall boundary conditions give different lattices — periodic packs points every 2π/L over all of k-space including negative components, hard-wall packs them every π/L but only in the positive octant. The denser hard-wall grid (×8 in 3D) is exactly cancelled by the one-eighth octant restriction, so both give the identical bulk density V/(2π)³.

The payoff of universality: the choice of boundary condition changes the lattice but not the bulk count. The factor-of-eight denser hard-wall grid is exactly undone by restricting to the positive octant.

The core equations

Periodicity over the box side quantizes each wavevector component to integer multiples of $2\pi/L$ . This is the mode set we count.

\psi(\mathbf{r}+L\hat{e}_i)=\psi(\mathbf{r})\ \Rightarrow\ k_i=\frac{2\pi n_i}{L},\quad n_i\in\mathbb{Z} \qquad (1)

where: $L$ — box side, $V=L^3$ ; $k_i$ — Cartesian component of $\mathbf{k}$ ; $n_i$ — integer quantum number. Physical reading: with periodic boundary conditions each component is an integer step of $2\pi/L$ , so the allowed $\mathbf{k}$ tile reciprocal space on a simple cubic lattice. Each allowed mode therefore owns a small cube of reciprocal space, and the inverse of that cube's volume is the number of modes per unit $k$ -space volume.

\Delta V_k=\left(\frac{2\pi}{L}\right)^3=\frac{(2\pi)^3}{V},\qquad \rho(\mathbf{k})=\frac{1}{\Delta V_k}=\frac{V}{(2\pi)^3} \qquad (2)

where: $\Delta V_k$ — $k$ -space volume per state; $\rho(\mathbf{k})$ — modes per unit $k$ -space volume. Physical reading: $\rho(\mathbf{k})$ is a constant, the same everywhere in reciprocal space — the single fact that makes quantum sums tractable. Because the density is uniform, a sum over modes becomes a simple integral; writing $\mathbf{p}=\hbar\mathbf{k}$ turns it into the classical phase-space measure with its $h^3$ cell.

\sum_{\mathbf{k}}\ \longrightarrow\ \frac{V}{(2\pi)^3}\int d^3k\ =\ \frac{V}{h^3}\int d^3p \qquad (3)

where: $h=2\pi\hbar$ — Planck constant; $\mathbf{p}=\hbar\mathbf{k}$ — momentum. Physical reading: quantum mode counting reproduces the classical phase-space cell $h^3$ per state — the microscopic justification of a rule Sackur and Tetrode had to assume. Internal degeneracy (spin, polarization) simply multiplies the count, and an isotropic energy count wraps a spherical shell around the origin.

\sum_{\text{states}}\to g_s\frac{V}{(2\pi)^3}\int d^3k,\qquad dN=g_s\frac{V}{(2\pi)^3}\,4\pi k^2\,dk \qquad (4)

where: $g_s$ — internal degeneracy ( $2s+1$ for spin- $s$ , $2$ for photon polarizations); $4\pi k^2\,dk$ — volume of a thin spherical shell in 3D $k$ -space. Physical reading: this shell is the object that, together with a dispersion relation $\varepsilon(\mathbf{k})$ , produces the energy density of states $g(\varepsilon)$ in aq-sm-density-of-states-energy.

Derivations

Derivation AFrom box boundary conditions to the uniform k-space density

The goal is to derive the constant $k$ -space mode density $V/(2\pi)^3$ from the box spectrum, and to show it is the geometry of the quantization lattice alone — independent of the particle's mass or dispersion relation. We take the single-particle spectrum from aq-qm-particle-in-3d-box (the standing waves and the energy pattern), impose periodic boundary conditions to obtain the lattice, and read off the density as one over the cell volume. The payoff is the sum-to-integral rule that every ideal-quantum-gas calculation begins with.

Step 1. Take the free-particle eigenstates. In a box the energy eigenstates are plane waves $\psi_{\mathbf{k}} = V^{-1/2}e^{i\mathbf{k}\cdot\mathbf{r}}$ (periodic) with energy $\hbar^2 k^2/2m$ ; the equivalent standing-wave form $\sin(n_i\pi x/L)$ and the degeneracy on spheres of radius $\sqrt{N}$ are owned by aq-qm-particle-in-3d-box. Why: the mode set is inherited from the quantum problem; SM only counts it.

Step 2. Impose periodicity. Requiring $e^{i k_i L}=1$ gives $k_i = 2\pi n_i/L$ with $n_i \in \mathbb{Z}$ , so the allowed $\mathbf{k}$ are the points of a simple cubic lattice of spacing $2\pi/L$ . Why: the wavefunction must close on itself around the box, quantizing each component independently.

Step 3. Find the cell volume. Each lattice unit cell contains exactly one allowed $\mathbf{k}$ and has volume $\Delta V_k = (2\pi/L)^3 = (2\pi)^3/V$ . Why: one point per cube of side $2\pi/L$ is the definition of the lattice spacing.

Step 4. Read off the density. The number of modes per unit $k$ -space volume is $\rho(\mathbf{k}) = 1/\Delta V_k = V/(2\pi)^3$ , a constant. Multiplying by $\hbar^3$ turns $d^3k$ into $d^3p/h^3$ , recovering the phase-space cell. Why: the density is one over the cell volume and carries no $\mathbf{k}$ dependence, which is precisely why the discrete sum becomes an integral with a constant prefactor. This matches aq-qm-density-of-final-states, whose eq. (1) states $dN = \frac{V}{(2\pi)^3}d^3k$ . Verified: the cell volume $(2\pi/L)^3 = (2\pi)^3/V$ , the density $V/(2\pi)^3$ , and the $\hbar^3$ conversion to the $h^3$ phase-space cell are checked symbolically in verify_density-of-states-k-space.py.

Go deeper · Derivation B · 1 minEquivalence of periodic and hard-wall boundary conditions

Derivation BEquivalence of periodic and hard-wall boundary conditions

The goal is to prove that hard-wall (Dirichlet, standing-wave) boundary conditions yield the same bulk density of states $V/(2\pi)^3$ as periodic ones, despite a different lattice spacing — so that the convenient periodic choice costs nothing physical. Hard walls give sine standing waves with spacing $\pi/L$ and only positive $k_i$ , an eightfold-denser grid restricted to one octant; we show the two changes cancel exactly, leaving the bulk density unchanged and boundary terms subleading in the thermodynamic limit (Weyl's law).

Step 1. Write the hard-wall modes. Vanishing at the walls forces $\psi \propto \sin(k_1 x)\sin(k_2 y)\sin(k_3 z)$ with $k_i = \pi n_i/L$ , $n_i = 1,2,3,\dots$ — spacing $\pi/L$ , half the periodic spacing, and only positive integers. Why: a standing wave pinned to zero at both walls fits an integer number of half-wavelengths, and $\sin$ with negative $n_i$ is not an independent state.

Step 2. Restrict to the octant. Since $\sin(-k_i x) = -\sin(k_i x)$ describes the same physical state, only the positive octant of $k$ -space is counted — a factor $1/8$ in 3D. Why: the standing wave has no directional sign, so negative components double-count.

Step 3. Count states in a shell. The hard-wall density is $1/(\pi/L)^3 = V/\pi^3$ , applied over $1/8$ of the shell: $dN = \frac{V}{\pi^3}\cdot\frac{1}{8}\cdot 4\pi k^2\,dk = \frac{V}{2\pi^2}k^2\,dk$ . Why: density times the fraction of the shell that lies in the allowed octant.

Step 4. Compare with periodic. The periodic count is $\frac{V}{(2\pi)^3}\cdot 4\pi k^2\,dk = \frac{V}{2\pi^2}k^2\,dk$ — identical. The eightfold-denser grid ( $2^3$ from halving the spacing) is exactly cancelled by the $1/8$ octant, and boundary corrections scale as surface/volume $\sim V^{-1/3} \to 0$ . Why: the two competing factors of $8$ cancel, so the bulk density of states is boundary-condition-independent in the thermodynamic limit — Weyl's law. Verified: the identity $\frac{V}{\pi^3}\cdot\frac{1}{8}\cdot 4\pi k^2 = \frac{V}{(2\pi)^3}\cdot 4\pi k^2 = \frac{V}{2\pi^2}k^2$ and the eightfold cancellation are checked symbolically in verify_density-of-states-k-space.py.

Limiting cases

The $k$ -space density is universal; only the energy conversion (owned by 1212) depends on dimension and dispersion. The shell factor $k^{d-1}$ makes this concrete.

Case	$k$ -space count	Consequence for $g(\varepsilon)$ (SM 1212)
3D, periodic BC	$\frac{V}{(2\pi)^3}\,4\pi k^2\,dk$	$g(\varepsilon)\propto\varepsilon^{1/2}$ for $\varepsilon\propto k^2$
2D	$\frac{A}{(2\pi)^2}\,2\pi k\,dk$	$g(\varepsilon)=$ const
1D	$\frac{L}{2\pi}\,2\,dk$	$g(\varepsilon)\propto\varepsilon^{-1/2}$ (van Hove edge)
Photons/ultrarelativistic 3D	same shell, $\varepsilon=\hbar c k$	$g(\varepsilon)\propto\varepsilon^{2}$ (Stefan–Boltzmann)
Hard-wall BC	$\frac{V}{\pi^3}\cdot\frac18\cdot4\pi k^2\,dk$	identical bulk density (Weyl)

Worked examples

Example 1 — Count the states in a small k-shell for a real box.

Step 1. Setup. Take a cube of side $L = 1\,\text{cm} = 0.01\,$ m, so $V = 10^{-6}\,\text{m}^3$ . Count the modes with $|\mathbf{k}|$ between $k = 1.00\times10^{9}\,\text{m}^{-1}$ and $k + dk$ with $dk = 1.0\times10^{6}\,\text{m}^{-1}$ (spin-0, $g_s = 1$ ).

Step 2. Cell and density. The cell volume is $(2\pi/L)^3 = (628.3)^3 = 2.48\times10^{8}\,\text{m}^{-3}$ ; the density is $V/(2\pi)^3 = 10^{-6}/248.05 = 4.03\times10^{-9}\,\text{m}^3$ . (Check: $1/2.48\times10^{8} = 4.03\times10^{-9}$ . ✓)

Step 3. Shell volume. $4\pi k^2\,dk = 4\pi(10^{9})^2(10^{6}) = 4\pi\times10^{24} = 1.257\times10^{25}\,\text{m}^{-3}$ .

Step 4. State count. $dN = \frac{V}{(2\pi)^3}\,4\pi k^2\,dk = (4.03\times10^{-9})(1.257\times10^{25}) = \mathbf{5.1\times10^{16}}$ modes. A macroscopic box packs astronomically many states into a thin shell — which is exactly why the discrete sum is safely replaced by an integral.

Example 2 — Fill the box with electrons: recover the Fermi wavevector.

Step 1. Setup. Put $N$ electrons ( $g_s = 2$ for spin) into the box, filling all $k$ -states up to $k_F$ . Count them: $N = g_s\frac{V}{(2\pi)^3}\cdot\frac{4}{3}\pi k_F^3$ .

Step 2. Solve for $k_F$ . Rearranging, $k_F^3 = \frac{3\pi^2 N}{V} = 3\pi^2 n$ , so $k_F = (3\pi^2 n)^{1/3}$ .

Step 3. Number for copper. With $n = 8.5\times10^{28}\,\text{m}^{-3}$ : $3\pi^2 n = 29.6\times8.5\times10^{28} = 2.52\times10^{30}$ , so $k_F = (2.52\times10^{30})^{1/3} = \mathbf{1.36\times10^{10}\,m^{-1}}$ .

Step 4. Meaning. This $k_F$ (and the Fermi energy $E_F = \hbar^2 k_F^2/2m = 7.0\,$ eV it implies) come purely from the uniform $k$ -space density and the electron count — the mode-counting rule of this article, filled with electrons up to the Fermi surface that aq-qm-particle-in-3d-box names as "the entry point to the free-electron model of metals."

Origins · 1 minreciprocal space and how this was discovered

reciprocal space

k-space mode counting · Reciprocal space — the space of allowed wavevectors

\mathbf k

— is where mode counting is cleanest: quantization by the box places one mode in every cell of volume

(2\pi/L)^3

, so the density of modes is the uniform constant

V/(2\pi)^3

regardless of the particle's dispersion. The counting was made concrete by Debye's 1912 lattice-vibration model, sharpened by Weyl's 1911 theorem that the asymptotic mode count is boundary-condition independent, and applied to electrons by Sommerfeld's 1927 free-electron theory of metals. The physics of the specific particle enters only when the dispersion

\varepsilon(\mathbf k)

converts this uniform

\mathbf k

-space density into an energy density of states.

History

1911

— Weyl proves that the asymptotic count of eigenmodes in a cavity depends only on its volume, not its shape or boundary condition

1912

— Debye counts lattice-vibration modes in k-space to derive the T³ heat capacity of solids

1924

— Bose recounts photon modes in phase space to derive Planck's law, the first indistinguishable-particle mode count

1927

— Sommerfeld fills the k-space lattice with electrons up to the Fermi surface, founding the free-electron model of metals

Going further

Texts: Pathria & Beale, Statistical Mechanics (3rd ed.), §1.4, §6.1, §7.1; Kardar, Statistical Physics of Particles, Ch. 7; Reif, Fundamentals of Statistical and Thermal Physics, §9.9–9.10; Ashcroft & Mermin, Solid State Physics, Ch. 2; Landau & Lifshitz, Statistical Physics Part 1, §53.
Microscopic owner: the box spectrum, standing waves $\sin(n_i\pi x/L)$ , and the degeneracy ("states of equal energy lie on a sphere of radius $\sqrt{N}$ ") are owned by aq-qm-particle-in-3d-box; the $dN = \frac{V}{(2\pi)^3}d^3k$ measure is shared with aq-qm-density-of-final-states.
Builds on: Occupation Number Representation · Thermal de Broglie Wavelength · aq-qm-particle-in-3d-box.
Leads to: Density of States in Energy (aq-sm-density-of-states-energy, folds in the dispersion) · Grand Canonical Ensemble for Ideal Quantum Gases · Bose–Einstein Distribution · Fermi–Dirac Distribution · Maxwell–Boltzmann Classical Limit.
Provenance: AllQFT `sm` export (Pathria & Beale), aq-sm-density-of-states-k-space, ref STAT.QSTAT.DOSK.1211. All counts (cell volume, $V/(2\pi)^3$ , the $h^3$ conversion, the hard-wall cancellation, the $k$ -shell number, $k_F$ for copper) python-verified (verify_density-of-states-k-space.py, all checks PASS).

Why use periodic boundary conditions when real boxes have hard walls?

Because they are far more convenient and cost nothing in the bulk. Periodic conditions give plane-wave eigenstates $e^{i\mathbf{k}\cdot\mathbf{r}}$ that carry momentum and make integrals simple, whereas hard walls give standing waves that are clumsier to sum. The second derivation proves the two give the identical bulk density of states $V/(2\pi)^3$ , with any difference confined to boundary terms that scale as surface/volume $\sim V^{-1/3}$ and vanish in the thermodynamic limit (Weyl's law). So for extensive quantities the periodic choice is exact.

Is V/(2π)³ really independent of the dispersion relation?

Yes. The $k$ -space density counts allowed wavevectors, which are fixed by the boundary conditions alone — it does not know whether the particle is massive ( $\varepsilon \propto k^2$ ), a photon ( $\varepsilon = \hbar c k$ ), or anything else. The dispersion relation enters only when you convert to an energy count $g(\varepsilon)$ , through the shell $4\pi k^2$ and the Jacobian $dk/d\varepsilon$ . That is why this article can be dispersion-agnostic and the energy-DOS article (1212) cannot.

Where does the factor h³ per state come from physically?

It is the quantum cell of phase space. Writing $\mathbf{p} = \hbar\mathbf{k}$ turns the $k$ -space measure into $\frac{V}{(2\pi)^3}d^3k = \frac{V}{h^3}d^3p$ , so each quantum state occupies a phase-space volume $h^3$ (one factor of $h$ per position–momentum pair). This is the microscopic origin of the $1/h^{3N}$ that Sackur and Tetrode had to insert by hand into the classical partition function to get absolute entropies right — quantum mode counting derives it.

Why a spherical shell and not a cube?

The shell is used because the energy of a free particle depends only on the magnitude $k = |\mathbf{k}|$ , so all states with energy near $\varepsilon$ lie on a sphere of radius $k(\varepsilon)$ . Counting them means measuring the $k$ -space volume between two nearby spheres, which is the shell $4\pi k^2\,dk$ . If the dispersion were anisotropic (as in a crystal), the constant-energy surface would not be a sphere and one would integrate over its actual shape — but for a free particle the sphere is exact.

How does spin enter the counting?

Spin (or any internal degree of freedom) simply multiplies every mode count by a degeneracy factor $g_s$ . For spin- $s$ particles $g_s = 2s+1$ , so electrons contribute $g_s = 2$ and photons $g_s = 2$ (two transverse polarizations). It appears as an overall prefactor in $dN = g_s\frac{V}{(2\pi)^3}4\pi k^2\,dk$ and carries straight through to $g(\varepsilon)$ ; for electrons it is why a single spatial $k$ -mode holds two particles and why the Fermi sphere holds $N/2$ spatial modes.

Does this article compute g(ε)?

Only in passing, to preview it. This article owns the $k$ -space lattice, the constant density $V/(2\pi)^3$ , and the shell $4\pi k^2\,dk$ ; the conversion to the energy density of states $g(\varepsilon) = \frac{g_s V}{4\pi^2}(2m/\hbar^2)^{3/2}\sqrt{\varepsilon}$ , including the dispersion Jacobian and the dimensional scaling $g(\varepsilon) \propto \varepsilon^{d/2-1}$ , is owned by aq-sm-density-of-states-energy. Here the shell is where the story stops; there it becomes the object plugged into the Bose and Fermi occupation sums.