Band theory of solids
Band theory of solids: periodic potential → energy bands and gaps, E–k, Brillouin zones, effective mass, electrical classification (metal/semiconductor/insulator), and related formulas and models.
Drafted with Aria, reviewed by the AiCanCode.org team · Updated . How we write and check our notes. Spotted an error? Use Give Feedback at the bottom of the page.
Why it matters
Key ideas
- Bloch’s theorem (exact statement): In a periodic potential V(r + R) = V(r), the single-electron wavefunction solutions of the Schrödinger equation can be written as ψ_k(r) = u_k(r)·e^{i k·r}, where u_k(r) has the periodicity of the lattice: u_k(r + R) = u_k(r). Consequence: energy eigenvalues are functions of wavevector k, written E_n(k), forming bands labeled by band index n.
- Energy band: a continuous range of allowed electron energies E_n(k) that arises because atomic orbitals from many atoms split and spread when atoms come together in a periodic array. Adjacent energy ranges with no allowed states are band gaps (forbidden energies).
- Brillouin zone (BZ): the Wigner–Seitz cell in reciprocal (k) space. The first Brillouin zone is the set of k points closer to the origin than to any other reciprocal lattice point. It contains all unique k due to periodicity in k-space.
- Reciprocal lattice and zone boundaries: reciprocal lattice vectors G map equivalent k-points: E(k) = E(k + G). Zone boundaries occur where k lies on Bragg planes; energy gaps open at these boundaries due to Bragg reflection of electron waves.
- Origin of band gaps (qualitative): When the electron de Broglie wavelength satisfies the Bragg condition (2k = G), forward-traveling and back-scattered waves interfere to form standing waves with different charge densities at ion cores, giving two different energies separated by a gap. Assumptions: weak periodic potential (nearly free electron picture) and elastic scattering from the periodic lattice.
- Nearly free electron (NFE) model: starts from free electrons with E = ħ^2 k^2/(2m_e) and treats V(r) as a small perturbation. Predicts parabolic E–k except near zone boundaries, where gaps open and E–k curves are flattened.
- Tight-binding (TB) model: starts from localized atomic orbitals and considers small overlap with neighbors. Predicts bands with widths proportional to overlap integrals; band edges occur at k where bonding/antibonding combinations give extrema.
- Kronig–Penney model: a 1D exactly solvable periodic potential (array of square wells/barriers or δ-barriers) that explicitly yields allowed and forbidden energy regions. It demonstrates how band gaps arise from periodicity. Commonly examined because it is analytically tractable in 1D.
- Statement of the main Kronig–Penney dispersion relation (1D, square-well form): For a periodic potential of period a with well width b and barrier width (a − b), the allowed energies E satisfy cos(k a) = cos(α b) cosh(β (a − b)) − (P/2) ( (α/β) + (β/α) ) sin(α b) sinh(β (a − b)), where α = √(2mE)/ħ (inside wells), β = √(2m(V0 − E))/ħ (inside barriers), and P depends on barrier strength. For the δ-barrier limit (thin, high barriers), it simplifies to cos(k a) = cos(α a) + (P/α a) sin(α a). Bands occur where |cos(k a)| ≤ 1; gaps where |cos(k a)| > 1. Assumptions: 1D, periodic piecewise-constant potential, single electron, stationary states.
- Filling of bands and Fermi level at 0 K: Electrons fill the lowest available energy states up to the Fermi energy E_F (highest occupied state at 0 K). If the highest filled band is only partially filled or overlaps with the next band, the solid is a metal; if a filled valence band is separated by a finite gap Eg from an empty conduction band, it is an insulator or semiconductor (small Eg ≲ ~3 eV for semiconductors). Note: full treatment of Fermi level vs temperature belongs to “Semiconductors and the Fermi level.”
- Effective mass: Near an extremum of a band (minimum for electrons, maximum for holes), E(k) can be approximated by a parabola. Define the scalar effective mass m* by 1/m* = (1/ħ^2) d^2E/dk^2 (for 1D along the principal direction). It modifies Newton-like dynamics: m* replaces the free electron mass m_e in acceleration and density-of-states expressions. At a band maximum, m* is negative for electrons; we treat missing electrons as positively charged quasiparticles called holes with positive mass.
- Group velocity: The electron wavepacket velocity is v_g = (1/ħ) ∇_k E(k). In 1D, v_g = (1/ħ) dE/dk. Flat bands (small slope) imply low velocity and high effective mass; steep bands imply high mobility (other scattering factors being similar).
- Semiclassical dynamics in bands: Under a uniform electric field E (vector), ħ dk/dt = −q E for an electron (q = +e magnitude, electron charge −e). Under a magnetic field B, ħ dk/dt = −q (E + v × B). Assumptions: slowly varying fields, no interband transitions.
- Density of states (DOS): Number of states per unit energy per unit volume. In 3D for a parabolic band minimum at energy Ec with effective mass m* (electrons), g_c(E) = (1/2π^2) (2m*/ħ^2)^{3/2} √(E − Ec) for E ≥ Ec, and zero below. Units: J⁻¹·m⁻³. This is used with Fermi–Dirac statistics in semiconductor analyses (covered in another topic).
- Diagram to draw and label (energy bands and gaps): Draw E on the vertical axis and k on the horizontal axis from −π/a to +π/a (first BZ). Sketch a parabolic band centered at k = 0 and show a gap at k = ±π/a where the parabola splits. Shade an occupied valence band and show an empty conduction band above with a gap Eg between their edges at k = 0 (direct gap) or at k ≠ 0 (indirect gap). Mark “First Brillouin zone,” “Valence band,” “Conduction band,” and “Band gap Eg.”
- Diagram to draw and label (Kronig–Penney criterion): Plot y = cos(ka) vs ka and superpose y = f(E) from the right-hand side of the Kronig–Penney equation as a horizontal band for each E (or annotate). Indicate that allowed energies satisfy |cos(ka)| ≤ 1. Label regions as “Allowed bands” and “Forbidden gaps.”
- Diagram to draw and label (reciprocal lattice and BZ): For 1D, draw k-axis with points 0, ±2π/a, ±4π/a. Shade the interval −π/a to +π/a as the first BZ. Mark Bragg planes at k = ±π/a and note “Gaps open here.”
Derivation sketches commonly asked
- Formation of bands from atomic levels (tight-binding picture)
- Start with an isolated atom: discrete atomic energy level E0.
- Bring N identical atoms to form a crystal. Due to overlap, the atomic level splits into N closely spaced levels spanning a bandwidth W ≈ 4|t| in simple 1D TB (t = hopping integral). As N → large, levels form a quasi-continuum = a band.
- Repeat for other atomic orbitals to get multiple bands separated by gaps when symmetry forbids mixing or when energy ranges do not overlap.
- NFE derivation of a gap at the zone boundary (qualitative math)
- Start from free-electron states |k⟩ and |k − G⟩ that are degenerate when k is at the boundary (k = G/2).
- Include a single Fourier component of the periodic potential V_G.
- Write the 2×2 secular equation: |E_k^0 − E V_G; V_G* E_{k−G}^0 − E| = 0. At degeneracy E_k^0 = E_{k−G}^0, eigenvalues split by ≈ 2|V_G|, forming an energy gap Δ ≈ 2|V_G| at k = G/2.
- Conclude: gaps open at Bragg planes, flattening E–k near boundaries and creating forbidden regions.
- Kronig–Penney (1D, outline)
- Consider piecewise-constant periodic potential of period a.
- Solve Schrödinger equation in well and barrier regions, get sinusoidal/exponential solutions with wavevectors α and β.
- Apply boundary conditions of continuity of ψ and dψ/dx at interfaces in one cell; apply Bloch condition ψ(x + a) = e^{ika} ψ(x).
- Eliminate coefficients to obtain the dispersion relation (as stated above). Identify allowed energies from |cos(ka)| ≤ 1.
- Effective mass relation
- Expand E(k) near an extremum k0: E(k) ≈ E0 + (1/2)(d^2E/dk^2)_{k0} (k − k0)^2.
- Match with parabolic form E = E0 + ħ^2 (k − k0)^2 / (2 m*), yielding 1/m* = (1/ħ^2)(d^2E/dk^2)_{k0}.
Classification from band picture (at 0 K)
- Metal (conductor): Fermi level lies within a band (partially filled) or valence and conduction bands overlap → many states at E_F, easy acceleration → high conductivity.
- Insulator: Valence band completely filled and separated by a large gap Eg (typically > ~3 eV) from an empty conduction band → negligible carriers at 0 K.
- Semiconductor: Similar to insulator but with a small gap (few eV). At finite temperature or with doping, carriers are thermally/chemically excited into the conduction band. Full carrier statistics are in “Semiconductors and the Fermi level.”
Optical transitions and direct/indirect gap (brief)
- Direct gap: Conduction band minimum (CBM) and valence band maximum (VBM) occur at the same k; optical transitions conserve k and are strong (useful for LEDs/lasers).
- Indirect gap: CBM and VBM at different k; phonon involvement needed for momentum conservation → weaker optical absorption/emission near the edge.
Formulas
ψ_k(r) = u_k(r) · e^{i k·r} — Bloch function; k: wavevector (m⁻¹); r: position (m); u_k(r): lattice-periodic part; holds for periodic potentials.
E_n(k) = E_n(k + G) — Band energy periodicity in reciprocal space; G: reciprocal lattice vector (m⁻¹); valid in crystals.
E_free(k) = ħ² k² / (2 m_e) — Free-electron dispersion; ħ: J·s; k: m⁻¹; m_e: kg; applies without periodic potential.
cos(k a) = cos(α a) + (P / (α a)) sin(α a) — Kronig–Penney (δ-barrier limit); a: m; α = √(2 m E)/ħ (m⁻¹); P: dimensionless barrier strength; allowed bands for |cos(ka)| ≤ 1.
Δ ≈ 2 |V_G| — Energy gap at a zone boundary in NFE model; V_G: Fourier component of periodic potential (J); applies near Bragg planes.
1 / m* = (1 / ħ²) · d²E/dk² — Effective mass (scalar, 1D principal direction); m*: kg; E in J; k: m⁻¹; valid near band extrema where E–k is parabolic.
E(k) ≈ E₀ + ħ² (k − k₀)² / (2 m*) — Parabolic approximation near an extremum; E₀: J; k₀: m⁻¹; m*: kg; valid sufficiently close to extrema.
v_g = (1/ħ) · dE/dk — Group velocity; v_g: m·s⁻¹; E: J; k: m⁻¹; holds for wavepackets in a band.
ħ · dk/dt = − q E — Semiclassical equation under electric field; q = +e = 1.602×10⁻¹⁹ C (magnitude); E: V·m⁻¹; k: m⁻¹; for electrons the force is −qE; valid without interband transitions.
g_c(E) = (1 / (2 π²)) · (2 m*/ħ²)^{3/2} · √(E − E_c) — 3D DOS for a parabolic conduction band; g_c: J⁻¹·m⁻³; m*: kg; E and E_c: J; valid for E ≥ E_c.
|G| = 2 π / a — Magnitude of the smallest reciprocal lattice vector in 1D; a: m; defines first BZ edges at k = ±π/a.
Worked examples
Example 1: First Brillouin zone width in 1D and location of gaps Given: 1D lattice constant a = 0.300 nm.
- Formula:
|G| = 2 π / a. First zone edges atk = ±π/a. Zone width =2π/a. - Substitute: a = 0.300 nm = 0.300 × 10⁻⁹ m = 3.00 × 10⁻¹⁰ m.
- π/a = 3.1416 / (3.00 × 10⁻¹⁰) = 1.0472 × 10¹⁰ m⁻¹.
- 2π/a = 2.0944 × 10¹⁰ m⁻¹.
- Result: First BZ extends from k = −1.047 × 10¹⁰ m⁻¹ to +1.047 × 10¹⁰ m⁻¹. Gaps (in NFE) open at k = ±π/a. Final answer: Edges at ±1.047 × 10¹⁰ m⁻¹; zone width 2.094 × 10¹⁰ m⁻¹.
Example 2: Effective mass and group velocity near a band minimum
Given: Near k₀ = 0, band is E(k) = E_c + ħ² k² / (2 m*) with m* = 0.20 m_e. For k = 5.00 × 10⁸ m⁻¹, find v_g and the acceleration under E = 1.00 × 10⁵ V·m⁻¹.
Constants: m_e = 9.109 × 10⁻³¹ kg; ħ = 1.054 × 10⁻³⁴ J·s; q = 1.602 × 10⁻¹⁹ C.
- Group velocity formula:
v_g = (1/ħ) dE/dk. For the parabola,dE/dk = ħ² k / m*. Thusv_g = (ħ k)/ m*. - Compute m*: m* = 0.20 m_e = 0.20 × 9.109 × 10⁻³¹ = 1.8218 × 10⁻³¹ kg.
- Compute v_g: v_g = (1.054 × 10⁻³⁴ × 5.00 × 10⁸) / (1.8218 × 10⁻³¹) = (5.270 × 10⁻²⁶) / (1.8218 × 10⁻³¹) = 2.893 × 10⁵ m·s⁻¹.
- Acceleration a under E using semiclassical dynamics and effective mass: Force on electron is −qE (direction opposite to E). Magnitude of acceleration
a = q E / m*.- a = (1.602 × 10⁻¹⁹ × 1.00 × 10⁵) / (1.8218 × 10⁻³¹) = (1.602 × 10⁻¹⁴) / (1.8218 × 10⁻³¹) = 8.799 × 10¹⁶ m·s⁻² (directed opposite to E). Final answers: v_g = 2.89 × 10⁵ m·s⁻¹, a = 8.80 × 10¹⁶ m·s⁻² (opposite to E).
Example 3: Classify material from band picture at 0 K Given descriptions of three solids: A) Uppermost filled band overlaps with next band by 0.5 eV. B) A filled valence band is separated from the conduction band by Eg = 1.1 eV (CBM and VBM at same k). C) A filled valence band is separated by Eg = 6 eV from the conduction band. Questions: Identify metal/semiconductor/insulator and note any optical implication.
- Rule: Partially filled or overlapping bands → metal; small finite gap → semiconductor; large gap → insulator. Direct gap enables strong optical transitions.
- Apply:
- A: Overlap means partially filled at E_F → Metal (good conductor).
- B: Small gap 1.1 eV → Semiconductor; direct gap → strong optical transitions near Eg (good for LEDs/lasers if other conditions are met).
- C: Large gap 6 eV → Insulator; requires high energy photons (UV) for transitions. Final answer: A: Metal; B: Direct-gap semiconductor; C: Insulator.
Common mistakes
- Treating m* as always positive; near a band maximum, electron m* is negative and the correct picture is holes with positive effective mass and positive charge.
- Using free-electron mass m_e in formulas near band edges where m* is required.
- Confusing Fermi energy (0 K concept) with Fermi level at finite temperature; detailed temperature dependence belongs to semiconductor physics.
- Drawing E–k diagrams without labeling the first Brillouin zone or band edges, losing marks for incomplete diagrams.
- Unit errors in DOS: forgetting that g(E) is per Joule per cubic metre (J⁻¹·m⁻³), not per eV unless converted.
For your exam
- Derivations: Outline of Kronig–Penney dispersion relation; NFE gap opening at zone boundary via 2×2 matrix; effective mass from Taylor expansion of E(k). State boundary conditions and Bloch condition clearly.
- Diagrams: E–k plot over the first BZ showing valence/conduction bands and gaps; reciprocal lattice with first BZ; Kronig–Penney allowed/forbidden regions (|cos(ka)| ≤ 1). Label axes, k-points (±π/a), band edges (Ec, Ev), and Eg.
- Numericals: First BZ width for a given a; compute group velocity v_g and acceleration using given m* and k; simple DOS evaluation at a specified E (ensure units). Show formula first, substitute with units, keep significant figures, and box the final result.
- Presentation: Start with assumptions, define symbols and units, keep steps logical, and annotate diagrams. Quote constants you use.
Quick check
- What does Bloch’s theorem say about the form of electron wavefunctions in a periodic crystal?
- Where do band gaps open in the nearly free electron model, and why?
- Write the relation connecting effective mass and the curvature of the E–k curve in 1D. Answers: 1) ψ_k(r) = u_k(r) e^{i k·r} with u_k periodic like the lattice. 2) At Brillouin zone boundaries (Bragg planes) due to Bragg reflection that couples k and k − G, splitting degenerate levels. 3) 1/m* = (1/ħ²) d²E/dk² (near a band extremum).
Finished this topic? Mark it so your progress, study plan and readiness keep up.
Stuck on something here?