Band Theory of Solids

⚡ Key Takeaways

Band Theory of Solids explains why solids are conductors, insulators, or semiconductors by treating electrons as quantum waves in a periodic crystal potential. Key result: the allowed electron energies group into continuous bands separated by forbidden gaps; whether a material conducts depends on how these bands are filled. Applies to crystalline solids (metals, semiconductors, insulators). Developed primarily 1928–1931 by Felix Bloch, Rudolf Peierls, Léon Brillouin, and Alan Wilson.

💡 The Core Insight

A perfectly periodic potential does not localize electrons. Instead, the translational symmetry of the lattice forces every electronic eigenstate to be a Bloch wave — a plane wave modulated by a function with the periodicity of the lattice. The eigenvalues $E(\mathbf{k})$ , viewed as a function of the crystal wavevector $\mathbf{k}$ , form continuous bands. The interplay between the number of electrons and the number of available band states (set by the number of unit cells) determines whether the highest occupied band is partly full (metal) or completely full with a gap above it (insulator/semiconductor).

DefinitionPeriodic Potential and Bravais Lattice

A crystalline solid is modeled by a one-electron potential $U(\mathbf{r})$ invariant under all lattice translations $\mathbf{R}$ of a Bravais lattice:

1

\quad U(\mathbf{r}+\mathbf{R}) = U(\mathbf{r}), \qquad \mathbf{R} = n_1\mathbf{a}_1 + n_2\mathbf{a}_2 + n_3\mathbf{a}_3,\ \ n_i \in \mathbb{Z}

Derivation hint: $\mathbf{a}_i$ are the primitive lattice vectors; the integers $n_i$ generate the full set of translation symmetry operations of the crystal. Here $U(\mathbf{r})$ is the effective potential felt by a single electron (combining nuclear attraction and a mean-field average of the other electrons), and $\mathbf{R}$ ranges over the discrete set of lattice points.

TheoremBloch's Theorem

The eigenstates $\psi$ of the one-electron Hamiltonian $H = -\frac{\hbar^2}{2m}\nabla^2 + U(\mathbf{r})$ with periodic $U$ can be chosen to have the form

2

\quad \psi_{n\mathbf{k}}(\mathbf{r}) = e^{i\mathbf{k}\cdot\mathbf{r}}\, u_{n\mathbf{k}}(\mathbf{r}), \qquad u_{n\mathbf{k}}(\mathbf{r}+\mathbf{R}) = u_{n\mathbf{k}}(\mathbf{r})

Derivation hint: the lattice translation operators $T_\mathbf{R}$ commute with $H$ and with each other, so simultaneous eigenstates exist; their eigenvalues must be phase factors $e^{i\mathbf{k}\cdot\mathbf{R}}$ . Equivalently, translating by a lattice vector multiplies the wavefunction by a phase:

3

\quad \psi_{n\mathbf{k}}(\mathbf{r}+\mathbf{R}) = e^{i\mathbf{k}\cdot\mathbf{R}}\,\psi_{n\mathbf{k}}(\mathbf{r})

Derivation hint: substitute the Bloch form [2] and use the periodicity of $u_{n\mathbf{k}}$ to extract the pure phase $e^{i\mathbf{k}\cdot\mathbf{R}}$ . The label $n$ is the band index; $\mathbf{k}$ is the crystal wavevector, restricted to the first Brillouin zone. The eigenvalues $E_n(\mathbf{k})$ are the band energies.

ProofBloch's Theorem via Translation Operators

Step 1 — Translation operators commute with H. Define $T_\mathbf{R}f(\mathbf{r}) = f(\mathbf{r}+\mathbf{R})$ . Since $U$ is periodic and the kinetic term is translation-invariant, $[T_\mathbf{R}, H] = 0$ for every lattice vector $\mathbf{R}$ .

Step 2 — Translations commute with each other. $T_\mathbf{R}T_{\mathbf{R}'} = T_{\mathbf{R}+\mathbf{R}'} = T_{\mathbf{R}'}T_\mathbf{R}$ . A commuting set of operators (including $H$ ) admits simultaneous eigenfunctions.

Step 3 — Eigenvalues are phases. Let $T_\mathbf{R}\psi = c(\mathbf{R})\psi$ . Because $T_\mathbf{R}$ is unitary, $|c(\mathbf{R})|=1$ . From $T_\mathbf{R}T_{\mathbf{R}'}=T_{\mathbf{R}+\mathbf{R}'}$ we get $c(\mathbf{R})c(\mathbf{R}') = c(\mathbf{R}+\mathbf{R}')$ , so $c$ is a one-dimensional representation: $c(\mathbf{R}) = e^{i\mathbf{k}\cdot\mathbf{R}}$ for some real vector $\mathbf{k}$ .

Step 4 — Construct the Bloch form. Write $\psi(\mathbf{r}) = e^{i\mathbf{k}\cdot\mathbf{r}}u(\mathbf{r})$ . Then

\psi(\mathbf{r}+\mathbf{R}) = e^{i\mathbf{k}\cdot\mathbf{R}}\psi(\mathbf{r}) \ \Rightarrow\ u(\mathbf{r}+\mathbf{R}) = u(\mathbf{r}),

so $u$ is lattice-periodic, completing equation [2]. $\qquad\blacksquare$

∎

ProofOrigin of Gaps — Nearly Free Electron Model

Step 1 — Start free. With $U=0$ , energies are $E^0(\mathbf{k}) = \hbar^2 k^2/2m$ , a single parabola.

Step 2 — Add a weak periodic potential. Expand $U(\mathbf{r}) = \sum_\mathbf{G} U_\mathbf{G}\,e^{i\mathbf{G}\cdot\mathbf{r}}$ over reciprocal lattice vectors $\mathbf{G}$ . The Fourier component $U_\mathbf{G}$ couples plane waves $|\mathbf{k}\rangle$ and $|\mathbf{k}+\mathbf{G}\rangle$ .

Step 3 — Degenerate perturbation at the zone boundary. When $|\mathbf{k}| = |\mathbf{k}+\mathbf{G}|$ (the Bragg condition, satisfied at Brillouin-zone boundaries), the two free states are degenerate. Diagonalizing the $2\times2$ coupling matrix gives

5

\quad E_\pm = \frac{E^0_\mathbf{k}+E^0_{\mathbf{k}+\mathbf{G}}}{2} \pm |U_\mathbf{G}|

Derivation hint: solve $\det\!\begin{pmatrix} E^0_\mathbf{k}-E & U_\mathbf{G} \\ U_\mathbf{G}^* & E^0_{\mathbf{k}+\mathbf{G}}-E\end{pmatrix}=0$ at the degeneracy point.

Step 4 — A gap opens. At the zone boundary the splitting is $E_+ - E_- = 2|U_\mathbf{G}|$ . No allowed states exist between $E_-$ and $E_+$ : this is the band gap. $\qquad\blacksquare$

∎

FB

Felix Bloch

1905–1983

A Swiss-American physicist who, as Heisenberg's first doctoral student in Zurich, proved in his 1928 thesis the theorem that bears his name — establishing that electrons in a perfect crystal propagate as undamped waves. This dissolved the paradox of why metallic conduction electrons have enormous mean free paths despite the dense lattice. Bloch later pioneered nuclear magnetic resonance (the Bloch equations), for which he shared the 1952 Nobel Prize in Physics with Edward Purcell. His 1928 work is the cornerstone on which the entire edifice of solid-state band theory rests.

🧠 Intuition

Think of two complementary pictures that meet in the middle.

Tight-binding picture (atoms first): Bring $N$ identical atoms from far apart. Each isolated atom has a sharp energy level (say the 3s level). When atoms approach, neighboring orbitals overlap, and the single level splits — first into bonding/antibonding pairs, then, for $N$ atoms, into $N$ closely spaced levels. Because $N\sim10^{23}$ , these $N$ levels are spaced so finely they form a quasi-continuous band. Each atomic level becomes one band; the gaps between bands are the energy ranges that no overlapping atomic state reaches.

Nearly-free picture (waves first): Start with free electrons (a smooth parabola $E\propto k^2$ ) and switch on a weak lattice. The lattice acts like a diffraction grating; electrons whose wavelength satisfies the Bragg condition are reflected and cannot propagate. At those special $\mathbf{k}$ -values (zone boundaries) the parabola splits open, carving the smooth spectrum into bands separated by gaps. Both pictures predict the same thing — bands and gaps — and meet quantitatively for intermediate coupling. The deep lesson: periodicity, not the strength of the potential, is what creates the band structure.

Figure

Energy bands and the band gap: the free-electron parabola (gold) splits open at the Brillouin-zone boundary

k=\pm\pi/a

, producing allowed bands separated by a forbidden gap.

ExampleCounting States: Is Sodium a Metal?

Given: Sodium crystallizes in a BCC lattice with one atom per primitive cell. Its valence configuration is $[\text{Ne}]\,3s^1$ — one valence electron per atom. A crystal has $N$ primitive cells.

Step 1 — States per band. Each band holds $2N$ electrons (factor 2 from spin, $N$ from the $N$ allowed $\mathbf{k}$ -values, equation [4]).

Step 2 — Electrons to place. Sodium contributes 1 valence electron per cell $\Rightarrow N$ valence electrons total.

Step 3 — Fill the band. $N$ electrons in a band of capacity $2N$ fills it exactly half full.

Step 4 — Conclusion. A half-filled band has empty states immediately above the highest occupied level (the Fermi level lies inside a band), so electrons respond freely to an applied field.

\boxed{\text{Sodium: highest band half-filled} \Rightarrow \textbf{metal}}

ExampleThe Band Gap of Silicon and Optical Absorption

Given: Silicon has an indirect band gap $E_g = 1.12\ \text{eV}$ at room temperature. Find the longest-wavelength photon that can excite an electron across the gap.

Step 1 — Energy–wavelength relation.

E_g = \frac{hc}{\lambda} \ \Rightarrow\ \lambda = \frac{hc}{E_g}

Step 2 — Insert numbers. Using the convenient constant $hc = 1240\ \text{eV}\,\cdot\,\text{nm}$ :

\lambda = \frac{1240\ \text{eV}\,\cdot\,\text{nm}}{1.12\ \text{eV}} = 1107\ \text{nm}

Step 3 — Interpret. Photons with $\lambda < 1107\ \text{nm}$ (energy $> 1.12$ eV, visible and near-IR) are absorbed by promoting electrons across the gap; longer-wavelength photons pass through.

\boxed{\lambda_{\max} \approx 1.11\ \mu\text{m}\ \ (\text{near-infrared})}

This is exactly why silicon is opaque to visible light but transparent in the far infrared, and sets the upper-wavelength limit of silicon photodetectors and solar cells.

⚠️ Common mistake

"Band gaps come from strong potentials." False. A gap opens for an arbitrarily weak periodic potential — equation [5] shows the gap is $2|U_\mathbf{G}|$ , which is nonzero whenever the relevant Fourier component is nonzero, no matter how small. It is the periodicity, acting through Bragg reflection, that creates the gap, not the magnitude of $U$ .

"The crystal wavevector $\mathbf{k}$ is the electron's true momentum." No. $\hbar\mathbf{k}$ is crystal momentum, conserved only up to a reciprocal lattice vector $\hbar\mathbf{G}$ . The actual momentum expectation value is generally not $\hbar\mathbf{k}$ because Bloch states are superpositions of plane waves $|\mathbf{k}+\mathbf{G}\rangle$ .

"Band theory requires interacting electrons." The standard band picture is a one-electron (mean-field) theory; electron–electron correlations are subsumed into the effective potential $U(\mathbf{r})$ . Strongly correlated materials (Mott insulators) can violate band-theory predictions entirely.

CounterexampleWhen Band Theory Fails: NiO

Simple band-filling counting predicts that any solid with an odd number of electrons per primitive cell must be a metal (half-filled band). Nickel monoxide (NiO) has the rocksalt structure and, by this rule, should be a good conductor. In reality NiO is a robust insulator with a gap of about 4 eV. The failure arises because the strong on-site Coulomb repulsion $U$ between $d$ -electrons localizes them — a Mott insulator. This shows that the one-electron band picture breaks down when electron–electron correlations dominate, motivating extensions like the Hubbard model and dynamical mean-field theory.

connection

Band theory sits at the crossroads of several pillars of physics. It is a direct application of Bloch's Theorem and the representation theory of the Translation Group of a crystal. The geometry of allowed $\mathbf{k}$ -values is governed by the Reciprocal Lattice and the Brillouin Zone. The occupation of bands at finite temperature follows Fermi–Dirac Statistics, with the Fermi Surface separating filled from empty states in metals. Quantitative band structures are computed via the Tight-Binding Model and the Nearly Free Electron Model, and at research level by Density Functional Theory. The dynamics of electrons in bands under external fields is captured by the Semiclassical Model of Electron Dynamics and the concept of Effective Mass, which in turn underlies the Drude Model's quantum successor, the Sommerfeld theory.

🔬 Where this appears

⏳ History

- 1900 — Paul Drude treats metallic electrons as a classical gas, explaining Ohm's law but failing on heat capacity.

- 1927 — Arnold Sommerfeld adds Fermi–Dirac statistics to the free-electron gas, fixing the heat-capacity problem but still ignoring the lattice.

- 1928 — Felix Bloch proves Bloch's theorem in his Zurich PhD thesis, showing electrons in a periodic potential form propagating waves with banded energy spectra.

- 1928 — Hans Bethe and others develop early lattice-electron calculations; Léon Brillouin formalizes the Brillouin zone construction.

- 1929 — Rudolf Peierls explains positive Hall coefficients via nearly-filled bands and introduces the concept that later becomes the hole.

- 1930 — Brillouin clarifies the role of Bragg reflection at zone boundaries in opening gaps.

- 1931 — Alan H. Wilson publishes the decisive papers distinguishing metals, insulators, and (intrinsic/extrinsic) semiconductors purely by band filling — the birth of semiconductor theory.

- 1933 — Eugene Wigner and Frederick Seitz develop a practical cellular method for computing band structures (the Wigner–Seitz cell).

- 1937 — John Slater and others advance augmented-plane-wave methods, foreshadowing modern computational band-structure theory.

open-problem

Band theory in its strict one-electron form is a mean-field approximation, and its boundaries remain active research territory:

Strong correlations. When the on-site Coulomb energy $U$ exceeds the bandwidth $W$ , band theory's filling rule fails (Mott insulators, high- $T_c$ cuprates). A fully predictive theory of strongly correlated electrons is still open.
The band-gap problem. Standard density functional theory (the workhorse for computing bands) systematically underestimates gaps, sometimes predicting metals where experiments find insulators. Many-body $GW$ corrections improve this but at large computational cost.
Disorder and Anderson localization. Real crystals are imperfect; sufficient disorder can localize Bloch states entirely, turning a would-be metal into an insulator — the precise mobility-edge behavior in three dimensions is subtle.
Topology of bands. The realization (post-2005) that band structures carry global topological invariants (Chern numbers, $\mathbb{Z}_2$ indices) has redrawn the metal/insulator classification, and a complete catalog of topological materials is an ongoing project.

Formal Statement

Why It Holds / Derivation

Building Intuition

Worked Examples

Common Misconceptions

Connections

Real-World Applications

Historical Development

Subtleties & Limits

See Also