NiO has a partially filled d-band (Ni²⁺ has 8 d-electrons in a 10-state d-shell). Band theory predicts it is a metal. Experiment shows it is an insulator with a gap of ~4 eV. What is the resolution?
ABand theory is wrong about the number of d-electrons
BThe strong Coulomb repulsion U ~ 8 eV between d-electrons on the same Ni site far exceeds the d-band width W ~ 3 eV. Electrons are prevented from hopping because double occupation costs too much energy. This Mott mechanism opens a correlation-driven gap between the lower Hubbard band (singly occupied states) and the upper Hubbard band (doubly occupied states), making NiO an insulator
CCrystal field splitting opens a band gap
DThe oxygen atoms in NiO donate electrons that fill the d-band
This is the textbook Mott insulator. In band theory, the d-band is split by the crystal field but remains partially filled — a metal. The Mott picture splits each d-state into two Hubbard bands: the lower Hubbard band (occupied, energy ~ε_d) and the upper Hubbard band (empty, energy ~ε_d + U). The gap between them is of order U - W. This was one of the earliest and most striking failures of independent-electron band theory, and it established that electron-electron interactions can qualitatively change the electronic ground state.
Question 2 Multiple Choice
The Mott transition can be driven by pressure in V₂O₃, which transitions from an antiferromagnetic insulator to a paramagnetic metal. Why does pressure favor the metallic state?
APressure destroys the crystal structure
BPressure decreases interatomic distances, increasing orbital overlap and thus the hopping integral t (and bandwidth W). When W becomes comparable to U, the kinetic energy gain from delocalization overcomes the Coulomb cost of double occupation, and the Mott gap closes. The Mott criterion (U/W ~ 1 for the transition) is reached from the insulating side by increasing W
CPressure increases the Coulomb repulsion U
DPressure aligns the electron spins, making the material metallic
V₂O₃ is the canonical system for studying the Mott transition. At ambient pressure, it is an antiferromagnetic insulator below 150 K. Applying ~25 kbar of pressure (or doping with Cr or Ti) drives it metallic. The transition is first-order at low temperatures (with a volume collapse — the metallic phase is denser) and ends at a critical point above which a smooth crossover connects the insulating and metallic phases. The entire phase diagram is governed by the ratio U/W.
Question 3 True / False
Mott insulators are often antiferromagnetic. However, Mott insulating behavior is not the same as antiferromagnetic ordering — the charge gap (Mott physics) and spin order (magnetism) are distinct phenomena.
TTrue
FFalse
Answer: True
The Mott gap arises from Coulomb repulsion suppressing charge fluctuations and exists even without magnetic order. The antiferromagnetism is a secondary consequence: once charges are localized (one electron per site), the residual exchange coupling J = 4t²/U between neighboring spins produces magnetic order. The Mott gap typically persists above the Neel temperature — the material remains insulating even when the spins are disordered (paramagnetic Mott insulator). In frustrated lattices (triangular, kagome), the magnetic ordering can be completely suppressed by geometric frustration, producing a Mott insulator with no long-range magnetic order — a quantum spin liquid candidate.
Question 4 Short Answer
Explain the difference between a Mott insulator and a band insulator, and describe how you could experimentally distinguish them.
Think about your answer, then reveal below.
Model answer: A band insulator has a gap due to the crystal potential (all bands fully occupied, gap between valence and conduction bands). A Mott insulator has a gap due to electron-electron interactions (partially filled band split into upper and lower Hubbard bands). Key experimental differences: (1) A Mott insulator has magnetic moments (from unpaired localized electrons) and shows Curie-Weiss susceptibility, while a band insulator is diamagnetic. (2) The Mott gap can be closed by pressure (increasing bandwidth) or doping (introducing mobile carriers), while a band gap is robust against these. (3) Spectroscopy (photoemission + inverse photoemission) shows spectral weight transfer between Hubbard bands as doping or temperature changes — a hallmark of correlated behavior absent in band insulators. (4) DFT calculations predict a metal — the disagreement between DFT and experiment is itself a diagnostic of Mott physics.
The classification has expanded beyond Mott vs. band: charge-transfer insulators (gap is between oxygen p-band and transition metal upper Hubbard band, as in many cuprates), Slater insulators (gap opens only with antiferromagnetic order), and topological Mott insulators (interaction-driven topological phases) are all variations on the theme.