Topological Phases of Matter: When Topology Becomes Physics

In 1980, Klaus von Klitzing was measuring the Hall resistance of a two-dimensional electron gas at low temperature and high magnetic field.

He found something unexpected.

The Hall conductance, instead of varying continuously with the magnetic field, was quantized.

It took only integer values, in units of \(e^2/h\):

\[\sigma_{xy} = \nu \frac{e^2}{h}, \quad \nu \in \mathbb{Z}\]

The quantization was extraordinary: not approximately integer. Exactly integer. To one part in \(10^9\).

More precise than any other measurement in condensed matter. Independent of sample geometry. Independent of impurities. Independent of the details of the material.

Von Klitzing received the Nobel Prize in 1985.

The question: why so precise?

The answer, worked out over the following years, was shocking:

Because topology.

The quantization is not approximate. It is exact, because it is a topological invariant. And topological invariants are integers. Integers are exact.

This was the beginning of a revolution.

I. The Landau Paradigm and Its Limits

To understand the revolution, we need to understand what it overturned.

The Landau paradigm of phases of matter:

Lev Landau (1937) established the framework that dominated condensed matter physics for fifty years.

Phases of matter are classified by symmetry breaking and order parameters.

Water to ice: translational symmetry broken → crystal. Paramagnet to ferromagnet: rotational symmetry broken → magnetization. Normal metal to superconductor: \(U(1)\) broken → Cooper pair condensate.

Every phase transition involves the breaking of some symmetry. Every phase is characterized by an order parameter, a local quantity that is zero in the symmetric phase and nonzero in the broken phase.

This framework is extraordinarily successful. It classifies an enormous range of phenomena.

Its failure:

The integer quantum Hall effect (IQHE) does not fit the Landau paradigm.

Different quantum Hall states, with different values of \(\nu\), have the same symmetries. No symmetry is broken going from \(\nu = 1\) to \(\nu = 2\).

There is no local order parameter that distinguishes them.

By the Landau classification, they should be the same phase.

They are not. They are topologically distinct.

You cannot continuously deform the \(\nu = 1\) state into the \(\nu = 2\) state without closing the bulk energy gap, without a phase transition.

The distinction is topological, not symmetry-based.

A new classification principle was needed.

II. The TKNN Invariant: Topology Enters

In 1982, Thouless, Kohmoto, Nightingale, and den Nijs (TKNN) computed the Hall conductance of a two-dimensional Bloch Hamiltonian.

Their result:

\[\sigma_{xy} = \frac{e^2}{h} \sum_n C_n\]

where the sum is over filled bands and \(C_n\) is the Chern number of band \(n\):

\[C_n = \frac{1}{2\pi}\int_{BZ} d^2k \,\Omega_n(\mathbf{k})\]

The Berry curvature:

\[\Omega_n(\mathbf{k}) = -2\,\text{Im} \langle\partial_{k_x}u_{n\mathbf{k}}|\partial_{k_y}u_{n\mathbf{k}}\rangle\]

where \(|u_{n\mathbf{k}}\rangle\) is the periodic part of the Bloch wavefunction.

The Chern number is a topological invariant: it is an integer, unchanged by smooth deformations of the Hamiltonian that do not close the gap.

The geometric origin:

The Berry curvature \(\Omega_n(\mathbf{k})\) is the curvature of a connection on a fiber bundle, the Bloch wavefunction bundle over the Brillouin zone (BZ).

The Brillouin zone of a 2D crystal is a torus \(T^2\).

The Chern number is the integral of the Berry curvature over this torus, the total flux of curvature through the Brillouin zone.

By Chern’s theorem, this flux, the integral of the first Chern class, is necessarily an integer.

The physical meaning:

The Chern number counts the number of times the wavefunction “winds” around the Brillouin zone.

It is a winding number, a topological property that cannot be changed without tearing the wavefunction.

This is why the Hall conductance is quantized: it is proportional to a topological invariant. Topological invariants are integers. Therefore the conductance is an integer times \(e^2/h\).

Exactly. Not approximately.

III. The Bulk-Boundary Correspondence

The topological invariant of the bulk, the Chern number, has a remarkable physical consequence at the surface.

The bulk-boundary correspondence:

A topological phase with Chern number \(C\) has exactly \(|C|\) chiral edge modes, one-dimensional channels at the boundary that propagate in one direction only.

These edge modes are topologically protected: they cannot be localized or gapped without destroying the bulk topology.

Why this is remarkable:

In an ordinary metal, impurities scatter electrons backward. Resistance arises from backscattering.

In a quantum Hall state, the edge modes are chiral: they can only propagate forward. There is no backward-moving mode to scatter into.

Backscattering is impossible, not because the impurities are weak but because of topology.

The edge conduction is perfect, resistance exactly zero, regardless of impurities.

This is the physical basis of the exact quantization of the Hall conductance.

The formal statement:

For a system with open boundary conditions, the number of edge states crossing the Fermi level equals the bulk topological invariant.

This is a deep result: the bulk topology is manifest at the boundary. The interior of the material determines what happens at its surface.

IV. The BKT Transition: Topology in Classical Systems

Before the quantum Hall revolution, Berezinskii, Kosterlitz, and Thouless discovered topological order in classical 2D systems.

The 2D XY model:

Spins in a plane, on a 2D lattice. The Hamiltonian:

\[H = -J\sum_{\langle i,j\rangle}\cos(\theta_i - \theta_j)\]

The Mermin-Wagner theorem: no long-range order at any finite temperature, continuous symmetry cannot break in 2D.

But something happens at a critical temperature \(T_{BKT}\).

Vortices:

The key excitations are vortices, topological defects where the spin angle winds by \(\pm 2\pi\) around a point.

A vortex has logarithmically divergent energy: \(E_{vortex} = \pi J \ln(L/a)\)

where \(L\) is the system size and \(a\) is the lattice spacing.

At low temperature: vortices and antivortices are bound in pairs, the energy cost confines them. The system has quasi-long-range order: power-law correlations.

At high temperature: the entropy gain from free vortices overcomes the energy cost. Vortices proliferate. Correlations decay exponentially.

The BKT transition is a transition between the vortex-bound phase and the vortex-unbound phase.

It is driven by topological defects, not by a local order parameter.

The Landau paradigm has no description of it.

Kosterlitz and Thouless received the Nobel Prize in 2016 for this work (done in the early 1970s).

V. Symmetry-Protected Topological Phases

The quantum Hall effect requires an external magnetic field: it breaks time-reversal symmetry.

Can topological phases exist without breaking time-reversal symmetry?

Yes. This led to one of the richest developments in modern condensed matter:

Symmetry-Protected Topological (SPT) phases.

The quantum spin Hall effect (2005-2007):

Kane and Mele (2005) predicted a new topological phase in graphene, the quantum spin Hall insulator, protected by time-reversal symmetry.

Unlike the IQHE, it exists without a magnetic field.

The key: spin-orbit coupling creates a topological band structure.

The bulk is insulating. The edges carry helical edge modes: spin-up propagates right, spin-down propagates left.

These are protected by time-reversal symmetry: to scatter a right-mover into a left-mover requires flipping the spin, which requires breaking time-reversal.

Impurities that preserve time-reversal cannot backscatter these modes.

The \(\mathbb{Z}_2\) invariant:

The quantum spin Hall insulator is characterized not by an integer (Chern number) but by a \(\mathbb{Z}_2\) invariant: either 0 (trivial) or 1 (topological).

This invariant, developed by Kane and Mele, counts the parity of the number of Kramers pairs of edge states.

3D topological insulators:

Fu, Kane, and Mele extended this to 3D (2007).

A 3D topological insulator has:

Insulating bulk
Metallic surface states

The surface states form a single Dirac cone, a band structure that cannot exist in a purely 2D system (the fermion doubling theorem).

The surface Dirac fermion is topologically protected: it cannot be gapped by any perturbation preserving time-reversal symmetry.

Experimental realization:

Bi\(_2\)Se\(_3\) and Bi\(_2\)Te\(_3\) were predicted (2009) and confirmed to be 3D topological insulators.

ARPES (Angle-Resolved Photoemission Spectroscopy) directly observed the surface Dirac cone.

This was a triumph of topological band theory, a theoretical prediction confirmed by direct experimental observation.

VI. The Tenfold Way: Classification of All Topological Phases

Ryu, Schnyder, Furusaki, and Ludwig (2008) and Kitaev (2009) completed the classification of free-fermion topological phases.

Their result: the tenfold way, ten symmetry classes, each with its own topological classification in each spatial dimension.

The ten classes arise from combinations of:

Time-reversal symmetry \(\mathcal{T}\)
Particle-hole symmetry \(\mathcal{C}\)
Chiral symmetry \(\mathcal{S} = \mathcal{T}\mathcal{C}\)

Each symmetry can be absent, present with \(\mathcal{T}^2 = +1\), or present with \(\mathcal{T}^2 = -1\) (similarly for \(\mathcal{C}\)).

This gives \(3 \times 3 + 1 = 10\) classes (the Altland-Zirnbauer classification).

For each class in each dimension, the possible topological invariants are: \(\mathbb{Z}\), \(\mathbb{Z}_2\), or trivial (0).

The complete classification organizes all possible topological insulators and superconductors in a periodic table, the table of topological phases.

The Bott periodicity:

The classification follows an 8-fold pattern: Bott periodicity in K-theory.

The mathematics: the classification of topological phases is the classification of vector bundles over spheres, the same problem that arises in homotopy theory and K-theory.

Physics rediscovered a deep result of 20th-century mathematics.

VII. Topological Superconductors and Majorana Fermions

Some of the most exciting topological phases are topological superconductors: bulk superconductors with topologically protected boundary modes.

The Kitaev chain (2001):

Alexei Kitaev proposed a 1D model, a chain of spinless fermions with p-wave superconducting pairing:

\[H = -t\sum_i(c_i^\dagger c_{i+1} + h.c.) - \mu\sum_i c_i^\dagger c_i + \Delta\sum_i(c_i c_{i+1} + h.c.)\]

In the topological phase (\(|\mu| < 2t\)): the chain has Majorana zero modes at its ends.

Majorana fermions:

A Majorana fermion is its own antiparticle: \(\gamma = \gamma^\dagger\)

In condensed matter, Majorana modes arise as half-fermion excitations: the fermion is split, with half at each end of the chain.

The two Majorana modes form a non-local qubit:

\[\tilde{c} = \frac{1}{2}(\gamma_1 + i\gamma_2)\]

This qubit is non-local: it is not at any specific position but distributed between the two ends.

Local perturbations at one end cannot affect the qubit state: they cannot “see” the other half.

This is topological protection: the qubit is protected from decoherence by non-locality.

Experimental realizations:

Semiconductor nanowires (InAs, InSb) with strong spin-orbit coupling, proximitized by an s-wave superconductor, in a magnetic field: this engineered system is predicted to be a topological superconductor.

Signatures of Majorana zero modes, zero-bias conductance peaks, have been observed by multiple groups (Mourik et al., Science 2012; many subsequent experiments).

The interpretation remains debated: zero-bias peaks can arise from non-topological mechanisms.

Definitive demonstration of non-Abelian braiding remains an outstanding experimental challenge.

VIII. Non-Abelian Anyons and Topological Quantum Computation

We have discussed anyons in the BEC essay. Here we go deeper, into the non-Abelian case.

Non-Abelian anyons:

For Abelian anyons: exchanging two particles gives a phase \(e^{i\theta}\). The phase commutes with everything. The order of exchanges does not matter.

For non-Abelian anyons: the state space of \(n\) anyons is a multi-dimensional degenerate subspace. Exchanging particles performs a unitary matrix operation on this space.

Different sequences of exchanges give different unitary operations: they do not commute.

The exchanges form a representation of the braid group \(B_n\), not the symmetric group \(S_n\).

The Moore-Read state:

The fractional quantum Hall state at \(\nu = 5/2\) is believed to be described by the Moore-Read (Pfaffian) state, which supports non-Abelian anyons.

The quasiparticles are Ising anyons. They fuse according to the rules: \(\sigma \times \sigma = 1 + \psi\) \(\sigma \times \psi = \sigma\) \(\psi \times \psi = 1\)

The fusion of two \(\sigma\) anyons can give either the vacuum (1) or a fermion (\(\psi\)). Before measurement, both outcomes are possible.

Topological quantum computation (Kitaev, 2003):

The degenerate ground state subspace of non-Abelian anyons can be used as a topological qubit.

Gates are implemented by braiding: moving anyons around each other.

The braiding unitary depends only on the topology of the braiding path, not on the speed, the exact path, or small perturbations.

This gives inherent fault tolerance: the computation is protected by the topology of the anyon worldlines.

Local errors cannot corrupt the computation without creating detectable anyon pairs.

The challenge:

Demonstrating non-Abelian braiding requires:

A system that supports non-Abelian anyons
The ability to create, move, and fuse anyons
The ability to read out the topological qubit state

None of these has been fully demonstrated.

Microsoft’s topological qubit program, targeting Majorana-based systems, is the major industrial effort.

Progress has been slow. The physics is hard. The reward, a fundamentally decoherence-resistant qubit, would be transformative.

IX. Topological Semimetals

Beyond topological insulators: topological semimetals, materials where the topological protection manifests in the bulk band structure.

Weyl semimetals:

In a Weyl semimetal, the conduction and valence bands touch at isolated points in the Brillouin zone: Weyl points.

Near a Weyl point, the Hamiltonian takes the form:

\[H(\mathbf{k}) = \chi \hbar v_F \boldsymbol{\sigma}\cdot\mathbf{k}\]

where \(\chi = \pm 1\) is the chirality, the topological charge.

Weyl points are topological monopoles of Berry curvature:

\[\oint_{S^2} \mathbf{\Omega} \cdot d\mathbf{S} = 2\pi\chi\]

They always come in pairs (Nielsen-Ninomiya theorem). They cannot be removed individually, only by annihilating a pair of opposite chirality.

Fermi arcs:

The surface of a Weyl semimetal has unusual Fermi arc surface states, open arcs in the surface BZ connecting the projections of Weyl points.

These are topologically protected: they cannot be removed without destroying the bulk Weyl points.

Fermi arcs were experimentally observed in TaAs (2015), the first Weyl semimetal discovered.

The chiral anomaly:

Weyl fermions in parallel electric and magnetic fields exhibit the chiral anomaly: the non-conservation of chirality:

\[\partial_\mu j^\mu_5 = \frac{e^2}{4\pi^2\hbar^2}\mathbf{E}\cdot\mathbf{B}\]

This is a quantum field theory anomaly (originally discovered in particle physics) manifested in a condensed matter system.

The physical consequence: negative magnetoresistance, resistance decreases with parallel magnetic field.

Observed in multiple Weyl semimetals.

Condensed matter as a quantum field theory simulator, directly accessing phenomena usually confined to high-energy physics.

X. The K-Theory Classification and Beyond

The tenfold way classifies free-fermion topological phases.

What about interacting systems?

Symmetry-Protected Topological (SPT) order in interacting systems:

For free fermions: topological phases classified by K-theory: \(\mathbb{Z}\) or \(\mathbb{Z}_2\).

With interactions: some of these phases collapse: the group structure of topological phases changes.

In 1D: the Haldane phase (spin-1 antiferromagnetic chain) is a bosonic SPT phase: it has no free-fermion realization.

In 3D: interacting bosonic SPT phases are classified by group cohomology:

\[H^{d+1}(G, U(1))\]

where \(G\) is the symmetry group and \(d\) is the spatial dimension.

This is a mathematical structure, group cohomology, that appears in topology and algebra, now classifying phases of matter.

Topological order:

Beyond SPT phases: topological order, phases with long-range entanglement that cannot be described by any local order parameter or any SPT classification.

Topological order is characterized by:

Ground state degeneracy on topologically nontrivial manifolds
Anyonic excitations
Topological entanglement entropy

The toric code (Kitaev, 2003) is the canonical example: a \(\mathbb{Z}_2\) topologically ordered phase with Abelian anyons, exact solvability, and topological quantum error correction.

The classification of topologically ordered phases, beyond free fermions, beyond SPT, is a major open problem.

Tensor category theory, modular tensor categories, topological quantum field theory: these are the mathematical tools. The physics is not yet fully understood.

XI. What Topology Tells Us About Matter

The topological revolution in condensed matter has revealed something profound:

The ground state wavefunction carries topological information that is invisible to local probes.

Traditional condensed matter physics focused on local order parameters: quantities you can measure at a point.

Topological phases are characterized by global, non-local properties of the wavefunction, properties that cannot be detected by any local measurement.

This is new. It requires new tools: mathematical tools from topology, physical tools that access global quantum correlations.

Entanglement as the key:

The topological entanglement entropy:

\[S(A) = \alpha|\partial A| - \gamma\]

where the universal constant \(\gamma = \ln\mathcal{D}\) is determined by the total quantum dimension \(\mathcal{D}\) of the topological phase.

For a trivial phase: \(\gamma = 0\). For the toric code: \(\gamma = \ln 2\).

The topological phase is characterized by the structure of its long-range entanglement: the topological entanglement entropy is a universal signature.

This connects topological order to quantum information: topological phases are phases with specific patterns of quantum entanglement.

The classification program:

The full classification of topological phases, including interactions, all symmetries, all spatial dimensions, is one of the great open problems in condensed matter physics.

It sits at the intersection of:

Condensed matter physics
Algebraic topology
K-theory
Tensor category theory
Topological quantum field theory
Quantum information

It is a problem where physics and mathematics are genuinely co-evolving, each driving the other forward.

XII. From von Klitzing to the Future

In 1980, von Klitzing measured an extraordinarily precise integer.

He did not know he was measuring topology.

Over the following forty years, we learned that:

The integer is a Chern number, a topological invariant of the Bloch wavefunction bundle.

The precision is exact because topological invariants are integers.

The phenomenon is the first example of topological order in nature, the first instance of quantum matter organized not by symmetry breaking but by the global topology of quantum states.

This led to:

Topological insulators and superconductors
Non-Abelian anyons and topological quantum computation
Weyl and Dirac semimetals
SPT phases and their classification
A new understanding of quantum entanglement

And it raised new questions:

What is the full classification of topological phases? Can non-Abelian anyons be realized in the laboratory? Is topological quantum computation feasible? What role does topology play in strongly correlated systems, in high-temperature superconductors, in heavy fermion materials, in frustrated magnets?

We are at the beginning of the topological era in condensed matter physics, not at its end.

The von Klitzing measurement of 1980 was the first word.

We are still working out what it means.

Essential reading:

von Klitzing, Dorda & Pepper, “New Method for High-Accuracy Determination of the Fine-Structure Constant” (1980) Phys. Rev. Lett. 45, 494 (the original discovery)

Thouless, Kohmoto, Nightingale & den Nijs, “Quantized Hall Conductance in a Two-Dimensional Periodic Potential” (1982) Phys. Rev. Lett. 49, 405 (the TKNN paper)

Haldane, “Model for a Quantum Hall Effect Without Landau Levels” (1988) Phys. Rev. Lett. 61, 2015 (the Chern insulator)

Kane & Mele, “Quantum Spin Hall Effect in Graphene” (2005) Phys. Rev. Lett. 95, 226801

Fu, Kane & Mele, “Topological Insulators in Three Dimensions” (2007) Phys. Rev. Lett. 98, 106803

Kitaev, “Unpaired Majorana Fermions in Quantum Wires” (2001) Phys.-Usp. 44, 131

Ryu, Schnyder, Furusaki & Ludwig, “Topological insulators and superconductors: tenfold way and dimensional hierarchy” (2010) New J. Phys. 12, 065010

For reviews:

Hasan & Kane, “Colloquium: Topological insulators” (2010) Rev. Mod. Phys. 82, 3045 (the essential review)

Qi & Zhang, “Topological insulators and superconductors” (2011) Rev. Mod. Phys. 83, 1057

For the mathematically serious:

Bernevig & Hughes, Topological Insulators and Topological Superconductors (2013) Princeton University Press (the graduate textbook)

Have a question about this piece? Ask →