12.20.03 · quantum / condensed-matter

Magnetism, Mott physics, and topological phases of matter

shipped3 tiersLean: none

Anchor (Master): Bernevig Topological Insulators and Topological Superconductors; Auerbach Interacting Electrons; Chaikin & Lubensky Principles of Condensed Matter Physics

Intuition Beginner

Every electron carries spin, a tiny quantum arrow that doubles as a magnetic north–south pole. In most solids the arrows point every which way and cancel, so the material shows no magnetism. In a few — iron, nickel, cobalt — the arrows lock together and the whole chunk becomes a magnet.

What locks them is not ordinary magnetism acting between distant arrows; that force is far too weak. It is the exchange interaction, a purely quantum effect that comes from the Pauli exclusion principle plus electrical repulsion. Exchange makes neighbouring electrons prefer either parallel or opposite spins, with no classical analogue at all.

Some materials choose alternating up–down–up–down patterns instead of all-aligned. These antiferromagnets have perfect magnetic order yet zero net magnet — order without a fridge magnet. The same exchange physics, when electron repulsion is strong enough, can split a metal into an insulator (a Mott insulator) that band theory cannot explain.

The deepest surprise is topology. In a topological insulator the bulk is insulating but the surface or edge conducts, and nothing short of destroying the material can remove those conducting channels — they are protected by a global topological invariant of the electron bands, the way the number of holes in a donut cannot change without cutting it.

Visual Beginner

A two-dimensional slab whose inside (bulk) is insulating while its one-dimensional rim carries an electric current; beside it, a plot of Hall resistance versus magnetic field showing flat exact plateaus at fixed quantum values.

The edge channel survives defects and impurities because it is enforced by the topology of the bulk band, not by any fine-tuned detail.

Worked example Beginner

In the integer quantum Hall effect, a 2D electron sheet in a strong perpendicular magnetic field conducts along its edge with a Hall resistance locked to exact values

R_{x y} = \frac{R _{K}}{ν}, R_{K} = \frac{h}{e ^{2}} \approx 25.8 k Ω,

where $ν$ is an integer (the filling factor), $h = 6.626 \times 1 0^{- 34}$ J·s is Planck's constant, and $e = 1.602 \times 1 0^{- 19}$ C is the electron charge. The constant $R_{K}$ is the von Klitzing constant.

Step 1. Square the charge: $e^{2} = (1.602 \times 1 0^{- 19})^{2} = 2.566 \times 1 0^{- 38}$ C².

Step 2. Divide Planck's constant by it: $R_{K} = 6.626 \times 1 0^{- 34} /2.566 \times 1 0^{- 38} = 2.58 \times 1 0^{4} Ω \approx 25.8 k Ω$ .

Step 3. At the $ν = 2$ plateau, $R_{x y} = R_{K} /2 \approx 12.9 k Ω$ . At $ν = 1$ , $R_{x y} \approx 25.8 k Ω$ .

What this tells us: the Hall resistance is exact to a few parts in a billion and depends only on fundamental constants — a macroscopic, countable sign of quantum topology. This precision is now used worldwide to define the SI unit of resistance.

Check your understanding Beginner

Formal definition Intermediate+

Magnetic exchange. For two electrons at sites $i, j$ the Coulomb repulsion, combined with the antisymmetry of the fermionic wavefunction, produces an effective spin–spin coupling called the exchange interaction ^{[Auerbach 1994]}:

H_{ex} = - ⟨ ij ⟩ \sum J_{ij} S_{i} \cdot S_{j} .

This is the Heisenberg Hamiltonian. The sign of $J_{ij}$ decides the order: $J > 0$ favours parallel spins (ferromagnetism); $J < 0$ favours antiparallel spins on neighbouring sites (antiferromagnetism). Ferrimagnets have antiparallel sublattices of unequal moment, giving a reduced net magnetisation.

Hund's rules determine the spin, orbital, and total angular momentum of a magnetic ion in an atom from the Coulomb and exchange energies of its open shell; the third rule fixes whether $J = L + S$ or $J = ∣ L - S ∣$ minimises the energy according to the shell's fill.

Magnons. A ferromagnetic ground state has all spins aligned. A spin wave is a slow, coherent tilt of the spin direction that propagates through the lattice; its quantum is the magnon, a bosonic excitation with dispersion $ω (k)$ that is quadratic ( $ω \propto k^{2}$ ) in a ferromagnet and linear ( $ω \propto k$ ) in an antiferromagnet at long wavelength.

Mott transition. The single-band Hubbard model,

H = - t ⟨ ij ⟩, σ \sum (c_{iσ}^{†} c_{j σ} + h.c.) + U i \sum n_{i ↑} n_{i ↓},

interpolates between a metal (small $U / t$ ) and a Mott insulator (large $U / t$ at one electron per site), in which electron repulsion localises the carriers and produces a moment on every site even though band theory predicts a half-filled conducting band ^{[Mott 1949]}.

Berry phase and curvature. For a Bloch band $n$ with cell-periodic states $∣ u_{n k} ⟩$ (the building blocks of 12.20.01), the Berry connection and Berry curvature are

A_{n} (k) = i ⟨ u_{n k} ∣ \nabla_{k} ∣ u_{n k} ⟩, Ω_{n} (k) = \nabla_{k} \times A_{n} (k) .

The first Chern number of the band,

C_{n} = \frac{1}{2 π} \int_{BZ} Ω_{n} (k) d^{2} k \in Z,

is an integer topological invariant: it cannot change under smooth deformations of the band that keep the gap open. A topological insulator is a gapped material whose occupied bands have a nontrivial topological invariant; the bulk-boundary correspondence states that the boundary then hosts a number of protected conducting modes equal to the jump in that invariant across the interface ^{[Bernevig 2013]}.

Counterexamples to common slips

Magnetism is not pair alignment of orbital currents. Atomic moments come from spin (and orbital) angular momentum of localized electrons; the alignment is enforced by the exchange interaction, not by magnetic dipole forces, which are roughly $1 0^{- 3}$ times too weak.
A Mott insulator is not a band insulator. Both are insulating, but a band insulator has an even integer number of electrons filling bands (band theory suffices), whereas a Mott insulator is driven by electron–electron repulsion $U$ localising carriers on sites, often at half-filling where band theory predicts a metal.
Topological protection is not invulnerability. Edge states are robust against non-magnetic disorder, but a perturbation that closes the bulk gap, or (for some invariants) breaks time-reversal symmetry, can destroy them.

Core model Intermediate+

The Heisenberg model captures low-energy magnetism once the charge degrees of freedom are frozen into a Mott or band insulator with one localised spin per site. For a nearest-neighbour ferromagnet on a lattice with coordination $z$ ,

H = - J ⟨ ij ⟩ \sum S_{i} \cdot S_{j}, J > 0.

The fully aligned state $∣ ↑↑ \dots ↑ ⟩$ is an exact ground state with energy $E_{0} = - J N z S^{2} /2$ . Low-energy excitations are spin flips delocalised as waves. Applying the Holstein–Primakoff transformation $S_{i}^{+} \approx 2 S a_{i}$ , $S_{i}^{z} = S - a_{i}^{†} a_{i}$ (with $a_{i}$ a boson) and Fourier transforming gives, to leading order in $1/ S$ , a gas of non-interacting magnons ^{[Auerbach 1994]}:

H \approx E_{0} + k \sum ℏ ω_{k} a_{k}^{†} a_{k}, ℏ ω_{k} = J S z (1 - γ_{k}), γ_{k} = \frac{1}{z} δ \sum e^{i k \cdot δ} .

At long wavelength $ℏ ω_{k} \approx D k^{2}$ with spin stiffness $D = J S a^{2}$ , so ferromagnetic magnons have a quadratic dispersion and contribute a $T^{3/2}$ correction to the magnetisation (Bloch's law), in sharp agreement with experiment on iron and nickel well below the Curie temperature $T_{C} \approx 1043$ K for iron.

The antiferromagnet ( $J < 0$ ) is subtler: linear spin-wave theory on two interpenetrating sublattices yields magnons with linear dispersion $ω \propto k$ at small $k$ , analogous to acoustic phonons, and a divergent magnon density that suppresses long-range order in one and two dimensions (consistent with the Mermin–Wagner theorem). The Hubbard model at large $U / t$ reduces, via second-order perturbation theory in $t / U$ , to an antiferromagnetic Heisenberg model with $J = 4 t^{2} / U$ — the microscopic origin of antiferromagnetism in Mott insulators such as the undoped parent compounds of the cuprate superconductors of 12.20.02.

Key derivation Intermediate+

The integer quantum Hall effect is the cleanest bridge from band topology to a measurable number. Consider a filled, isolated band $n$ in two dimensions. Semiclassical dynamics in crossed electric and magnetic fields gives an anomalous velocity proportional to the Berry curvature, and the Kubo formula for the Hall conductivity of a single occupied band reduces to ^{[Thouless, Kohmoto, Nightingale & den Nijs 1982]}

σ_{x y} = - \frac{e ^{2}}{ℏ} \int_{BZ} \frac{d ^{2} k}{( 2 π ) ^{2}} Ω_{n} (k) .

Pulling the constants out and using the definition of the Chern number,

σ_{x y} = \frac{e ^{2}}{h} \frac{1}{2 π} \int_{BZ} Ω_{n} (k) d^{2} k = \frac{e ^{2}}{h} C_{n} .

Summing over all occupied bands gives the TKNN formula

σ_{x y} = \frac{e ^{2}}{h} n \in occ \sum C_{n} = ν \frac{e ^{2}}{h}, ν \in Z .

Because each $C_{n}$ is an integer, $ν$ is an integer: the Hall conductance is quantised. Crucially, $C_{n}$ cannot change unless the band gap closes, so disorder that does not close the gap cannot shift $σ_{x y}$ — this is why the Hall plateaus are exact to parts in $1 0^{10}$ . The Hall resistance is $R_{x y} = 1/ σ_{x y} = R_{K} / ν$ with $R_{K} = h / e^{2}$ , exactly the plateau law of the worked example. The number of chiral edge channels equals $ν$ , the bulk-boundary correspondence: a bulk topological invariant fixes a boundary transport property.

Bridge. The TKNN invariant builds toward 12.13.01, where the Berry connection is the adiabatic holonomy of a line bundle over the Brillouin torus in the Fock-space formalism, and appears again in 12.20.02, whose flux quantisation $Φ_{0} = h / (2 e)$ is the same single-valuedness of a macroscopic quantum phase. The foundational reason the Hall conductance is quantised is that a Chern number is an integer counting the winding of the Bloch wavefunction; this is exactly the bridge from the band structure of 12.20.01 to a topologically protected, disorder-immune measurement, and the pattern generalises to the fractional quantum Hall effect and to the $Z_{2}$ topological insulators protected by time-reversal symmetry. Putting these together, the central insight is that bulk topology dictates boundary physics: the bridge is that a global invariant of the occupied bands enforces exact, countable edge conductance.

Exercises Intermediate+

Exercise 3 (medium, symbolic).

Explain why a ferromagnetic magnon has a quadratic dispersion $ω \propto k^{2}$ at long wavelength while an acoustic phonon has a linear one $ω \propto k$ .

Hint

Compare the restoring symmetry: ferromagnetic ground states conserve total spin, so the energy cost of a long-wavelength tilt vanishes quadratically.

Answer

The ferromagnetic ground state spontaneously breaks continuous spin-rotation symmetry, but total magnetisation $S_{tot}^{z}$ is conserved, so the Goldstone mode (a uniform rotation of all spins, $k = 0$ ) costs zero energy and a slowly varying tilt costs energy proportional to gradients squared, giving $ω = D k^{2}$ . Acoustic phonons break continuous translation symmetry with a conserved momentum-like elastic law, and elasticity theory gives $ω = c_{s} k$ linearly. The different dispersion reflects the different symmetry being broken and the absence of a restoring force linear in displacement for spin rotations.

Exercise 5 (medium, symbolic).

Sketch why the first Chern number $C_{n} = (2 π)^{- 1} \int_{BZ} Ω_{n} d^{2} k$ must be an integer.

Hint

Rewrite the curvature as the curl of the Berry connection and use that the Brillouin zone is a torus with periodic gauge patches.

Answer

Write $Ω_{n} = \nabla_{k} \times A_{n}$ . The Brillouin zone is a torus $T^{2}$ , but the Bloch state $∣ u_{n k} ⟩$ need not be globally single-valued, so $A_{n}$ exists only in patches. By Stokes' theorem the integral of the curvature equals the circulation of $A_{n}$ around the boundary of a fundamental patch, which is the mismatch in the Berry phase across the two gauge patches. That mismatch is $2 π$ times the winding number of the transition function, an integer. Hence $C_{n} \in Z$ , unchanged under any smooth deformation that keeps the band gap open.

Exercise 6 (medium, symbolic).

State the bulk-boundary correspondence for the integer quantum Hall effect and explain why disorder does not destroy the quantised Hall conductance.

Hint

Edge channel count equals the bulk Chern number; disorder can shift states but not change an integer unless the gap closes.

Answer

The bulk-boundary correspondence says a 2D sample with total Chern number $ν = \sum_{n} C_{n}$ of occupied bands carries exactly $ν$ chiral (one-way) edge modes at its boundary, and the Hall conductance is $σ_{x y} = ν e^{2} / h$ . Non-magnetic disorder broadens and shifts Landau levels but cannot change the integer $ν$ unless it closes the bulk mobility gap. Because the edge modes have no available backscattering partner (they run one way only), they are immune to impurities, and the Hall conductance stays pinned to its plateau value to extraordinary precision.

Exercise 7 (hard, symbolic).

Derive the antiferromagnetic exchange $J = 4 t^{2} / U$ of the large- $U / t$ half-filled Hubbard model by treating hopping in second-order perturbation theory.

Hint

A single hop creates a doubly occupied site costing energy $U$ ; a second hop can restore one electron per site. Track the spin dependence of the two-hop amplitude.

Answer

At half-filling and $U ≫ t$ the ground manifold has one electron per site, a vast degeneracy over spin configurations. A single hop $c_{i ↑}^{†} c_{j ↑}$ creates a virtual state with a doubly occupied site costing energy $U$ . Second-order perturbation theory gives an effective Hamiltonian within the ground manifold, $H_{eff}^{(2)} = - P H_{t} \frac{1}{H _{U}} H_{t} P$ , where $P$ projects onto no-double-occupancy states. For a bond $⟨ ij ⟩$ the two-hop return amplitude depends on whether the two spins are parallel or antiparallel: parallel spins are Pauli-blocked from double occupancy, while antiparallel spins can virtually hop and back. The resulting spin dependence is $- (4 t^{2} / U) S_{i} \cdot S_{j}$ plus a constant, i.e. an antiferromagnetic Heisenberg coupling $J = 4 t^{2} / U > 0$ (with the convention $H = + J \sum S_{i} \cdot S_{j}$ ). This is the microscopic origin of antiferromagnetism in Mott insulators.

Exercise 8 (hard, symbolic).

Show that a two-band Hamiltonian $H (k) = d (k) \cdot σ$ on a 2D Brillouin zone has Chern number equal to the winding number of the map $\hat{d} : T^{2} \to S^{2}$ , and compute it for $H = sin k_{x} σ_{x} + sin k_{y} σ_{y} + (m + cos k_{x} + cos k_{y}) σ_{z}$ as a function of $m$ .

Hint

The lower band's Berry curvature is $Ω = - \frac{1}{2} \hat{d} \cdot (\partial_{k_{x}} \hat{d} \times \partial_{k_{y}} \hat{d})$ ; integrate and count preimages of a generic point on $S^{2}$ .

Answer

For a two-level Hamiltonian the eigenstates are determined by the unit vector $\hat{d} = d /∣ d ∣$ , and the Berry curvature of the lower band is $Ω (k) = - \frac{1}{2} \hat{d} \cdot (\partial_{k_{x}} \hat{d} \times \partial_{k_{y}} \hat{d})$ . Integrating over the Brillouin torus gives the degree (winding number) of the map $\hat{d} : T^{2} \to S^{2}$ , an integer counting how many times the image wraps the sphere. For the given model the gap closes (and the Chern number can change) only where $sin k_{x} = sin k_{y} = 0$ and $m + cos k_{x} + cos k_{y} = 0$ , i.e. at the high-symmetry points $Γ (0, 0)$ , $X (π, 0)$ , $Y (0, π)$ , $M (π, π)$ , with critical masses $m = - 2, - 0, 0, + 2$ respectively. Adding the local Dirac contributions $\pm \frac{1}{2}$ gives $C = + 1$ for $0 < m < 2$ , $C = - 1$ for $- 2 < m < 0$ , and $C = 0$ for $∣ m ∣ > 2$ . This is the lattice-regularised Haldane/Dirac model, the prototype Chern insulator.

Lean formalization Intermediate+

lean_status: none. The two formal pillars of this unit — second-quantised many-body Hilbert space and the differential geometry of band bundles — are both beyond present Mathlib. The Heisenberg and Hubbard Hamiltonians require the Fock-space and creation-annihilation machinery of 12.13.01; the Holstein–Primakoff reduction and the Mott transition need operator-algebra infrastructure that does not yet exist. On the topological side, the Berry connection, its curvature, and the first Chern number of a complex line bundle over the Brillouin torus touch complex-vector-bundle and characteristic-class theory that Mathlib develops only in pieces, and the TKNN theorem $σ_{x y} = (e^{2} / h) C$ and the bulk-boundary correspondence have no formal statement. The present unit's correctness gate is therefore derivation consistency and experimental evidence, not formal proof.

Advanced results Master

Integer quantum Hall effect (IQHE). Klaus von Klitzing's 1980 experiment measured the Hall resistance of a 2D electron gas at low temperature in a strong magnetic field and found plateaus at $R_{x y} = R_{K} / ν$ with $ν$ integer and $R_{K} = h / e^{2} \approx 25.81 k Ω$ to ten significant figures ^{[von Klitzing, Dorda & Pepper 1980]}. The TKNN paper ^{[Thouless, Kohmoto, Nightingale & den Nijs 1982]} explained the integer $ν$ as the sum of Chern numbers of occupied Landau levels, establishing topology as a principle of condensed matter.

Fractional quantum Hall effect (FQHE). Two years later Tsui, Störmer, and Gossard observed a plateau at the fractional filling $ν = 1/3$ ^{[Tsui, Störmer & Gossard 1982]}, impossible for non-interacting electrons. Robert Laughlin explained it with the variational wavefunction

Ψ_{m} (z_{1}, \dots, z_{N}) = i < j \prod (z_{i} - z_{j})^{m} exp (- \frac{1}{4 ℓ _{B}^{2}} i \sum ∣ z_{i} ∣^{2}), m odd,

describing an incompressible quantum liquid at filling $ν = 1/ m$ ^{[Laughlin 1981]}. Its quasiparticle excitations carry fractional charge $e / m$ and obey anyonic exchange statistics, the first laboratory realisation of particles outside the Bose/Fermi dichotomy. The FQHE is the paradigmatic strongly correlated topological phase, requiring many-body entanglement beyond band theory.

$Z_{2}$ topological insulators and the Kane–Mele model. Charles Kane and Eugene Mele showed in 2005 that spin–orbit coupling in graphene realises a quantum spin Hall phase: two time-reversed copies of the QHE, one per spin, whose edge carries counter-propagating spin-filtered modes ^{[Kane & Mele 2005]}. Time-reversal symmetry forbids backscattering, so the edge is protected even though the net Chern number vanishes — the protecting invariant is a $Z_{2}$ index. Three-dimensional $Z_{2}$ topological insulators (predicted and observed in Bi $_{2}$ Se $_{3}$ and related compounds) host conducting 2D surface states with a single Dirac cone.

Bulk-boundary correspondence and the Haldane model. Haldane's 1988 model — a honeycomb lattice with complex next-nearest-neighbour hopping — was the first Chern insulator: it has $σ_{x y} = (e^{2} / h) C$ with nonzero $C$ at zero net magnetic field, proving that topology, not field strength, drives the QHE. The general principle is the bulk-boundary correspondence: the number of protected boundary modes equals the difference in the bulk topological invariant across the interface, a theorem of index-theoretic flavour.

Synthesis. Magnetism and topology are two faces of electron correlation: the exchange interaction and the Hubbard model build toward 12.13.01, where the Heisenberg and Hubbard Hamiltonians live in second-quantised Fock space, and appear again in 12.20.02, whose Cooper-pair physics shares the same many-body exchange machinery and whose flux quantisation mirrors the single-valuedness that fixes the Chern number. The foundational reason the quantum Hall conductance is quantised is that a Chern number is an integer winding of the Bloch bundle over the Brillouin torus; this is exactly the bridge from the band theory of 12.20.01 to a disorder-immune measurement known to ten significant figures, and the pattern generalises to fractional Hall liquids, $Z_{2}$ topological insulators, and topological superconductors hosting Majorana modes. Putting these together, the central insight is that global topological invariants of bulk bands dictate protected boundary physics; the bridge is that the bulk-boundary correspondence turns an abstract integer into a countable, measurable edge current.

Full proof set Master

Proposition (TKNN quantisation of the Hall conductance). For a two-dimensional band insulator with isolated occupied bands, the zero-temperature Hall conductance is

σ_{x y} = \frac{e ^{2}}{h} C, C = n \in occ \sum \frac{1}{2 π} \int_{BZ} Ω_{n} (k) d^{2} k \in Z .

Proof. Start from the Kubo formula for the Hall response of a filled band $n$ in a weak electric field $E$ , with current operator $j = - (e /ℏ) \nabla_{k} H$ . Summing the interband matrix elements and using $⟨ u_{n k} ∣ \nabla_{k} H ∣ u_{m k} ⟩ = (ε_{n} - ε_{m}) ⟨ \nabla_{k} u_{n k} ∣ u_{m k} ⟩$ for $m \neq = n$ gives the Kubo–Berry identity

σ_{x y}^{(n)} = - \frac{e ^{2}}{ℏ} \int_{BZ} \frac{d ^{2} k}{( 2 π ) ^{2}} 2 Im ⟨ \nabla_{k_{x}} u_{n k} ∣ \nabla_{k_{y}} u_{n k} ⟩ .

The factor $2 Im ⟨ \nabla_{k_{x}} u_{n k} ∣ \nabla_{k_{y}} u_{n k} ⟩$ is exactly the $z$ -component of the Berry curvature $Ω_{n} (k) = \nabla_{k} \times A_{n}$ (a short verification uses the gauge-covariant derivative of the projector $∣ u_{n} ⟩ ⟨ u_{n} ∣$ ). Hence

σ_{x y}^{(n)} = - \frac{e ^{2}}{ℏ} \cdot \frac{1}{( 2 π ) ^{2}} \int_{BZ} Ω_{n} d^{2} k = \frac{e ^{2}}{h} C_{n},

using $h = 2 π ℏ$ and the sign convention for $Ω_{n}$ . To see that $C_{n}$ is an integer, cover the torus $T^{2}$ by two gauge patches on which $∣ u_{n k} ⟩$ is single-valued. By Stokes' theorem the curvature integral over the torus equals the line integral of the Berry connection around the equator between the patches, which is the change in the Berry phase across the gauge transition function $e^{i φ (k)}$ . That change is $2 π$ times the winding number of $φ$ , an integer. Therefore each $C_{n} \in Z$ , the sum $C$ is an integer, and $σ_{x y}$ is quantised in units of $e^{2} / h$ . Because a smooth perturbation cannot change an integer without first closing the band gap, the quantisation is robust against disorder, completing the proof. $□$

Connections Master

Condensed-matter foundations 12.20.01. Band structure, Bloch states, and the reciprocal-lattice Brillouin zone are the substrate on which the Berry connection and Chern numbers are built: the topological invariants of this unit are integrals over the same momentum-space torus introduced there.
Superconductivity 12.20.02. Flux quantisation $Φ_{0} = h / (2 e)$ and the Hall quantum $h / e^{2}$ share a common origin in the single-valuedness of a macroscopic quantum phase; topological superconductors extend the bulk-boundary correspondence to Majorana edge modes, the superconducting analogue of the chiral Hall edge channel.
Fock spaces and second quantisation 12.13.01. The Heisenberg and Hubbard Hamiltonians and the magnon Holstein–Primakoff bosons are operators on fermionic and bosonic Fock spaces; the many-body Laughlin state of the fractional quantum Hall effect is a highly entangled vector in that same formalism.
Many-body quantum mechanics 12.09.01. The exchange interaction derives from the antisymmetry of identical fermions, and Hund's rules and ferromagnetic ordering are direct consequences of the identical-particle statistics developed there.
Spontaneous symmetry breaking 08.02.02. A ferromagnet breaks continuous spin-rotation symmetry and supports a Goldstone magnon, the same Landau symmetry-breaking framework that classifies superconductors and superfluids; topological phases, by contrast, evade Landau's classification and require the topological invariants introduced here.

Historical & philosophical context Master

Werner Heisenberg identified the exchange interaction as the origin of ferromagnetism in 1928, resolving a puzzle that had defeated classical physics: the magnetic dipole force between atomic moments is far too weak to align them at room temperature, and only the quantum exchange energy, an offshoot of Pauli antisymmetry, could account for the Curie temperature of iron ^{[Heisenberg 1928]}. Hund's empirical rules for atomic spectra, codified in the same period, gave the recipe for the localised moments that exchange then orders.

In 1949 Nevill Mott argued that nickel oxide is insulating not because of band filling but because strong electron–electron repulsion localises the carriers, a state of affairs with no description in single-particle band theory ^{[Mott 1949]}. The Hubbard model, introduced a decade later, made Mott's picture precise and remains the central testbed for strongly correlated physics, with applications from the cuprate superconductors of 12.20.02 to modern ultracold-atom experiments.

The topological era opened on the night of 4–5 February 1980, when Klaus von Klitzing measured the Hall resistance of a silicon field-effect transistor at the Grenoble high-field magnet and found it pinned to $R_{K} / ν$ with extraordinary precision ^{[von Klitzing, Dorda & Pepper 1980]}. The TKNN paper of 1982 ^{[Thouless, Kohmoto, Nightingale & den Nijs 1982]} interpreted the integer $ν$ as a Chern number, importing differential geometry into solid-state physics. The discovery of the $ν = 1/3$ fractional state by Tsui, Störmer, and Gossard later that year ^{[Tsui, Störmer & Gossard 1982]}, explained by Laughlin's wavefunction ^{[Laughlin 1981]}, revealed quasiparticles of fractional charge and anyonic statistics.

The theoretical framework broadened with Haldane's 1988 Chern insulator and the 2005 prediction by Kane and Mele of the $Z_{2}$ topological insulator in graphene ^{[Kane & Mele 2005]}, soon generalised to three dimensions and confirmed experimentally in bismuth-based compounds. The 2016 Nobel Prize in Physics, awarded to David Thouless, Duncan Haldane, and Michael Kosterlitz, recognised this entire programme: phases of matter classified not by broken symmetry but by topology. Philosophically, the quantum Hall effect is the most striking known example of a macroscopic quantity fixed exactly by an abstract integer — a laboratory confrontation between physics and pure mathematics.

Bibliography Master

@book{Bernevig2013,
  author = {Bernevig, B. Andrei},
  title = {Topological Insulators and Topological Superconductors},
  publisher = {Princeton University Press},
  year = {2013},
}

@book{Auerbach1994,
  author = {Auerbach, Assa},
  title = {Interacting Electrons and Quantum Magnetism},
  publisher = {Springer},
  year = {1994},
}

@book{ChaikinLubensky1995,
  author = {Chaikin, P. M. and Lubensky, T. C.},
  title = {Principles of Condensed Matter Physics},
  publisher = {Cambridge University Press},
  year = {1995},
}

@book{Kittel2005,
  author = {Kittel, Charles},
  title = {Introduction to Solid State Physics},
  edition = {8},
  publisher = {Wiley},
  year = {2005},
}

@book{Simon2013,
  author = {Simon, Steven H.},
  title = {The Oxford Solid State Basics},
  publisher = {Oxford University Press},
  year = {2013},
}

@article{Heisenberg1928,
  author = {Heisenberg, Werner},
  title = {Zur Theorie des Ferromagnetismus},
  journal = {Zeitschrift f{\"u}r Physik},
  volume = {49},
  pages = {619--636},
  year = {1928},
}

@article{Mott1949,
  author = {Mott, N. F.},
  title = {The basis of the electron theory of metals, with special reference to the transition metals, the metal-insulator transition and brass-like alloys},
  journal = {Proceedings of the Physical Society. Section A},
  volume = {62},
  pages = {416--422},
  year = {1949},
}

@article{Klitzing1980,
  author = {von Klitzing, K. and Dorda, G. and Pepper, M.},
  title = {New Method for High-Accuracy Determination of the Fine-Structure Constant Based on Quantized Hall Resistance},
  journal = {Physical Review Letters},
  volume = {45},
  pages = {494--497},
  year = {1980},
}

@article{TKNN1982,
  author = {Thouless, D. J. and Kohmoto, M. and Nightingale, M. P. and den Nijs, M.},
  title = {Quantized Hall Conductance in a Two-Dimensional Periodic Potential},
  journal = {Physical Review Letters},
  volume = {49},
  pages = {405--408},
  year = {1982},
}

@article{TsuiStormerGossard1982,
  author = {Tsui, D. C. and St{\"o}rmer, H. L. and Gossard, A. C.},
  title = {Two-Dimensional Magnetotransport in the Extreme Quantum Limit},
  journal = {Physical Review Letters},
  volume = {48},
  pages = {1559--1562},
  year = {1982},
}

@article{Laughlin1981,
  author = {Laughlin, R. B.},
  title = {Quantized Hall conductivity in two dimensions},
  journal = {Physical Review B},
  volume = {23},
  pages = {5632--5633},
  year = {1981},
}

@article{Laughlin1983,
  author = {Laughlin, R. B.},
  title = {Anomalous Quantum Hall Effect: An Incompressible Quantum Fluid with Fractionally Charged Excitations},
  journal = {Physical Review Letters},
  volume = {50},
  pages = {1395--1398},
  year = {1983},
}

@article{Haldane1988,
  author = {Haldane, F. D. M.},
  title = {Model for a Quantum Hall Effect without Landau Levels: Condensed-Matter Realization of the ``Parity Anomaly''},
  journal = {Physical Review Letters},
  volume = {61},
  pages = {2015--2018},
  year = {1988},
}

@article{KaneMele2005,
  author = {Kane, C. L. and Mele, E. J.},
  title = {$\mathbb{Z}_2$ Topological Order and the Quantum Spin Hall Effect},
  journal = {Physical Review Letters},
  volume = {95},
  pages = {146802},
  year = {2005},
}

Prerequisites

12.20.01

Tier anchors

beginner: Kittel Introduction to Solid State Physics Ch. 11 (diamagnetism, paramagnetism, ferromagnetism in plain terms); Simon The Oxford Solid State Basics Ch. on topology (intuition: bulk topology forces protected conducting edges)
intermediate: Simon The Oxford Solid State Basics Ch. on topology; Kittel Ch. 11–15 (magnetism, exchange, spin waves); Auerbach Interacting Electrons Ch. 1–5
master: Bernevig Topological Insulators and Topological Superconductors; Auerbach Interacting Electrons; Chaikin & Lubensky Principles of Condensed Matter Physics

References

Bernevig — Topological Insulators and Topological Superconductors · Ch. 1–5: Berry phase, Chern number, integer and fractional quantum Hall effects, the bulk-boundary correspondence, and the Haldane and Kane–Mele models.
Auerbach — Interacting Electron Theory · Ch. 1–5: the exchange interaction, the Hubbard model and the Mott transition, spin waves and magnons, and strongly correlated magnetism.
Kittel — Introduction to Solid State Physics · Ch. 11–15: diamagnetism and paramagnetism, Hund's rules, the Heisenberg model of ferromagnetism, spin waves, and the classical picture of magnetic order.
Thouless, Kohmoto, Nightingale & den Nijs — Quantized Hall Conductance (TKNN, 1982) · The Diophantine equation linking the Hall conductance to the first Chern number of the occupied magnetic subbands: $σ_{x y} = (e^{2} / h) C$ .
von Klitzing, Dorda & Pepper — New Method for High-Accuracy Determination of Fine-Structure Constant (1980) · Discovery of the integer quantum Hall effect: Hall resistance plateaus at $R_{K} / ν$ with $R_{K} = h / e^{2} \approx 25.8 k Ω$ .

Estimated time

beginner: 15 min
intermediate: 50 min
master: 110 min