Band Theory: Checkpoint Review

Introduction

Band theory has several starting points that can look contradictory. A free-electron picture ignores the ions; nearly-free theory brings in a weak periodic potential; tight binding begins with electrons close to individual atoms. Bloch's theorem is the symmetry statement that organises the periodic models, not a rival choice. These approaches are useful limits of one larger problem. The purpose of this checkpoint is to connect their assumptions, recognise what each predicts well and decide when a simple band classification is insufficient.

Core explanation

In the free-electron model , the interior potential is constant and states have E = ħ²k²/(2m e). Counting spin-degenerate states gives a Fermi sphere and E F ∝ n^(2/3) in three dimensions. This estimates scales and explains why temperature changes occupation mainly near the Fermi surface. It does not generate crystal-specific gaps, directional effective masses or chemically distinctive d bands. Its k values fill a reciprocal-space continuum within a macroscopic limit, although finite boundary conditions quantise them densely.

Bloch's theorem assumes periodicity V(r+R)=V(r). It says states can be written ψ nk(r)=e^(ik·r)u nk(r), with u periodic. This does not dictate the actual E n(k); a periodic potential still has to be specified and solved. The first Brillouin zone holds one representative of k values modulo reciprocal vectors. At its boundaries, states that differ by a reciprocal vector can become degenerate in a simple free-wave picture. Coupling through a periodic Fourier component then opens a gap. The MIT quantum-physics notes derive both Bloch translation phases and tight-binding bands, making their shared symmetry especially clear.

The nearly-free-electron model is useful when electrons are delocalised and the lattice potential is a modest perturbation. It explains avoided crossings and gaps near Bragg planes, with local gap 2 V G in a stated Fourier convention. Tight binding is useful when local atomic orbitals retain strong identity and hopping to neighbours is a helpful parameter. A nearest-neighbour chain gives E(k)=ε₀+2t cos(ka), with width 4 t . Stronger overlap often broadens bands; directional s, p and d overlap gives different dispersions. Both models obey Bloch's theorem if their Hamiltonians are periodic.

Classification needs electron filling as well as allowed states. A partly filled band or overlapping bands suggest a metal in a weakly interacting model; a filled lower band separated from empty upper states suggests a gapped material. Gap size, temperature and dopants distinguish useful semiconductor behaviour from ordinary insulating behaviour, but no fixed universal gap cutoff settles every case. Group 14 diamond-type materials show filled bonding and empty antibonding manifolds with very different gaps. Peierls distortion shows that the lattice itself can change its period and open a gap at a half-filled one-dimensional Fermi level.

Transition-metal compounds supply the strongest caution: a nominal half-filled narrow d band can be insulating because on-site electron repulsion competes with hopping. In a simple Mott picture, large U relative to bandwidth W favours charge localisation. A ligand-to-metal charge-transfer gap or structural distortion may also be relevant. Band diagrams generated from one-electron methods are therefore a map of one approximation, not a direct proof of measured conductivity. MIT's solid-state chemistry band lecture connects bonding to material class, while the ACS analysis of oxide bandwidth illustrates the extra correlation question.

Finally, distinguish states , occupation and transport . DOS g(E) counts available states. Fermi–Dirac f(E) gives occupation probability. Their product gives occupied-state density; velocities and scattering then determine current. A flat band near the chemical potential may have many states yet poor transport. A semiconductor's conduction band may have states but very few electrons unless thermally excited or doped.

Step-by-step reasoning

1. State the crystal structure, electron count and relevant orbitals. 2. Choose a model based on whether wave-like or local-orbital behaviour dominates. 3. Use periodicity and reciprocal space to identify band labels and gaps. 4. Count occupation with Fermi–Dirac statistics and DOS. 5. Add scattering, defects, correlations or structural distortion before explaining actual properties.

Visual explanation

Draw a triangle with three corners labelled free electrons, periodic waves and atomic orbitals. Put Bloch symmetry above the triangle as the rule for both periodic descriptions. At the centre place an E–k graph and a separate DOS×occupation sketch. Arrows from defects and electron repulsion point into a measured-conductivity box, showing why band filling alone is not the final answer.

Real-world analogy

Studying traffic can start with cars moving freely on an open plain, a regular street grid, or cars parked at houses connected by roads. All describe different limits of movement. Counting parking spaces is not counting cars, and counting cars is not knowing traffic speed. Likewise electronic states, occupation and mobility answer separate questions.

Real-world example

Suppose a new oxide has a half-filled calculated d band but measured insulating behaviour. A sensible analysis checks the crystal symmetry and any distortion, the d-band width and ligand p character, magnetic order, sample stoichiometry and whether local repulsion is large. Declaring the calculation "wrong" without identifying which approximation fails would be less useful than testing these mechanisms against spectroscopy and transport.

Why?

Why are the free-electron and tight-binding models not mutually exclusive? They start from opposite limits of orbital delocalisation, but both produce allowed states E n(k) and can be compared with the same symmetry and filling framework. Intermediate materials require more detailed calculations, not a choice of one model as universally true.

Common misconception

"If a calculated band crosses the Fermi level, the sample must conduct well." Strong electron correlations, disorder or low mobility can suppress conduction. The plot may also represent an ideal stoichiometric structure unlike the actual sample.

Worked example

Question: A one-dimensional chain has one electron per site and one orbital per site. Uniform nearest-neighbour tight binding predicts a half-filled band. The atoms then alternate short and long separations. Explain two changes in the electronic description and one condition required for a stable distorted structure.

Reasoning: Alternation doubles the real-space repeat and halves the Brillouin-zone width. The new periodicity connects the original Fermi points and can split the uniform band into a filled lower and empty upper branch. The electronic energy gain from lowering occupied states must exceed the elastic energy of moving atoms. Whether it actually does depends on electron–lattice coupling and other material effects.

Answer: The unit cell doubles and a gap can open at the former Fermi level; stability requires net total-energy lowering.

Quick check

1. Which model directly introduces hopping t between local atomic orbitals? Answer: Tight binding; Bloch's theorem then organises its periodic eigenstates.

Exam focus

Compare assumptions before comparing formulas. Use free-electron E(k), tight-binding E(k) and Bloch form in the correct contexts. Explain why DOS, Fermi occupation and mobility are distinct. For a partly filled yet insulating d system, give a physically specific mechanism rather than contradicting the evidence.

Advanced insight

Band calculations often downfold complex orbital spaces into an effective tight-binding Hamiltonian. Fitted hopping parameters can reproduce calculated E–k curves, while interaction terms such as U are added to study correlation. The parameter values depend on chosen local orbitals and energy window, so they should not be treated as unique chemical constants independent of model construction.

Summary

Free-electron, nearly-free and tight-binding models describe different electronic limits. Bloch symmetry unites periodic treatments, and band filling plus DOS and Fermi occupation organise ideal behaviour. Real conductivity and gaps can also depend on lattice distortion, disorder, defects and many-electron effects.

Practice questions

1. What does Bloch's theorem require that the free-electron approximation does not use explicitly? Answer: A potential periodic under crystal lattice translations. 2. What does g(E)f(E)dE count? Answer: Occupied electron states in the energy interval dE, under the model's normalisation. 3. Which model gives a nearest-neighbour cosine band in a simple chain? Answer: Tight binding. 4. Name one reason a half-filled calculated band may correspond to an insulator. Answer: Strong correlation, a Peierls distortion, charge-transfer physics or disorder-induced localisation can prevent ordinary metallic transport.