Topological Matter
In mathematics, a coffee cup and a donut are “the same”: one can be smoothly deformed into the other without cutting or gluing. This equivalence might seem far removed from physics, yet topology now organizes many phases of condensed matter. Their invariants protect selected transport coefficients and boundary states against perturbations that preserve the gap and the required symmetries. This protection improves robustness, but it guarantees neither perfect conduction nor immunity to every kind of disorder.
This lesson builds on quantum mechanics (lessons 6-11), statistical physics (lessons 21-25), and band theory (lesson 68). The topological perspective connects to gauge theories (lesson 35) and to the fracton physics discussed in lesson 58.
What Topology Means for Physics
A topological invariant remains unchanged under continuous deformations that preserve the assumptions used to define it. For a given insulating phase, changing the invariant generally requires closing the gap or breaking the protecting symmetry. Hall quantization is robust to moderate disorder because of localization, but sufficiently strong perturbations can destroy the phase.
The Integer Quantum Hall Effect
Classical Hall Effect
When a current flows through a conductor in a perpendicular magnetic field, charges accumulate on one side, producing a transverse voltage. The Hall resistance is classically $R_H = B/(ne)$, proportional to the field and inversely proportional to the carrier density. This is a smooth, continuous function of the parameters.
The Quantum Hall Effect
In 1980, Klaus von Klitzing discovered something extraordinary in a two-dimensional electron gas at low temperature and high magnetic field. The Hall conductance is not a smooth function but is quantized to exact integer multiples of a fundamental unit:
The quantization is exact to better than one part in $10^9$, regardless of the sample geometry, impurity content, or precise value of the magnetic field (within a plateau). This extraordinary precision earned von Klitzing the Nobel Prize in 1985, and the value $e^2/h$ is now used to define the standard of electrical resistance: the von Klitzing constant $R_K = h/e^2 \approx 25{,}812.807$ $\Omega$.
Landau Levels
The physics behind the quantum Hall effect begins with Landau levels. In a magnetic field, the kinetic energy of a 2D electron is quantized into equally spaced levels:
where $\omega_c$ is the cyclotron frequency. Each Landau level has a macroscopic degeneracy: the number of states per level equals the number of flux quanta threading the sample, $N_\phi = BA/(h/e) = eBA/h$. When exactly $\nu$ Landau levels are filled ($\nu$ is the filling factor), the longitudinal resistance vanishes and the Hall conductance equals $\nu e^2/h$.
But why is the quantization so exact, even in the presence of disorder? This is where topology enters.
Berry Phase and Berry Connection
The Geometric Phase
In 1984, Michael Berry showed that a quantum state acquires a geometric phase when its parameters are varied adiabatically around a closed loop. Consider a Hamiltonian $H(\mathbf{R})$ depending on parameters $\mathbf{R}$. If $\mathbf{R}$ is slowly varied around a closed loop, the eigenstate $|n(\mathbf{R})\rangle$ returns to itself but acquires a phase:
where $\mathbf{A}_n(\mathbf{R}) = i\langle n(\mathbf{R})|\nabla_{\mathbf{R}}|n(\mathbf{R})\rangle$ is the Berry connection. This is formally identical to a gauge potential in electromagnetism. The associated Berry curvature is:
Why the Berry Phase Matters
The Berry phase is geometric: it depends only on the path in parameter space, not on the speed of traversal. It is gauge-invariant (up to $2\pi n$) and has measurable physical consequences. For electrons in a crystal, the relevant parameter space is the Brillouin zone, and the Berry curvature of the Bloch bands determines topological invariants that classify phases of matter.
TKNN Invariant: Topology Explains Quantization
In 1982, Thouless, Kohmoto, Nightingale, and den Nijs (TKNN) showed that the Hall conductance of a filled Landau level equals a topological invariant, the Chern number:
The Chern number $C_n$ is always an integer. This is a deep mathematical fact: the integral of the Berry curvature over a closed surface (the Brillouin zone is a torus) is always $2\pi$ times an integer, by the same theorem that says the total curvature of a sphere is $4\pi$.
Why the Chern Number Must Be an Integer
The Brillouin zone is a torus: Because $\mathbf{k}$ and $\mathbf{k} + \mathbf{G}$ are equivalent, the 2D Brillouin zone has the topology of a torus $T^2$.
Berry connection as a gauge field: The Berry connection $\mathbf{A}(\mathbf{k})$ is like an electromagnetic vector potential on this torus. The Berry curvature $\Omega(\mathbf{k})$ is the corresponding "magnetic field."
First Chern class: For a well-defined bundle of isolated bands over the Brillouin zone, the Berry-curvature integral is quantized: $\frac{1}{2\pi}\int \Omega \, d^2k \in \mathbb{Z}$. This integer cannot change while the bundle remains well defined, which in particular requires the relevant gap to stay open.
Physical consequence: The Hall conductance, being proportional to this integer, is exactly quantized. Disorder can broaden Landau levels but cannot change the Chern number as long as the gap remains open. This explains the extraordinary precision of the quantum Hall effect.
Topological Insulators
Insulators with Conducting Surfaces
The quantum Hall effect requires a strong magnetic field and low temperatures. A breakthrough came in 2005-2006 when Kane and Mele (and independently Bernevig, Hughes, and Zhang) predicted that certain materials could be topological even without any magnetic field. These topological insulators have:
- A full band gap in the bulk (they are insulators)
- Gapless, conducting states on their surface (or edge, in 2D)
- Surface states protected by time-reversal symmetry
The Bulk-Boundary Correspondence
The most profound principle in topological matter is the bulk-boundary correspondence: the topological invariant of the bulk determines the number and nature of gapless states at the boundary. If the bulk has a nontrivial topological invariant, the boundary must have gapless states. These states cannot be removed by any perturbation that preserves the protecting symmetry, because removing them would require changing the bulk topological invariant, which requires closing the bulk gap.
In a 2D topological insulator, the edge states are helical: spin-up electrons propagate in one direction and spin-down in the other. Backscattering would require flipping the spin, which is forbidden by time-reversal symmetry. This makes the edge transport dissipationless and robust against non-magnetic disorder.
In 3D, the surface states form a Dirac cone: a linear dispersion $E = \hbar v_F |\mathbf{k}|$ like massless relativistic fermions, but with spin locked perpendicular to momentum. Topological surface states have been experimentally observed in Bi$_2$Se$_3$, Bi$_2$Te$_3$, and many other materials using angle-resolved photoemission spectroscopy (ARPES).
The Z$_2$ Invariant
Unlike the quantum Hall effect, which is classified by the integer Chern number, time-reversal invariant topological insulators are classified by a $\mathbb{Z}_2$ invariant that takes only two values: 0 (trivial) or 1 (topological). The invariant counts the parity of the number of times the surface-state Fermi surface encloses a time-reversal invariant momentum. A material with the $\mathbb{Z}_2$ invariant equal to 1 must have an odd number of Dirac cones on its surface.
Chern Numbers and Topological Classification
The Chern number is the prototypical topological invariant. More broadly, topological phases of matter can be classified by examining which symmetries are present:
| Symmetry | Dimension | Invariant | Example |
|---|---|---|---|
| None (class A) | 2D | $\mathbb{Z}$ (Chern) | Quantum Hall effect |
| Time-reversal (class AII) | 2D | $\mathbb{Z}_2$ | HgTe quantum wells |
| Time-reversal (class AII) | 3D | $\mathbb{Z}_2$ | Bi$_2$Se$_3$ |
| Particle-hole (class D) | 1D | $\mathbb{Z}_2$ | Kitaev chain (Majorana) |
This classification, known as the "periodic table of topological insulators and superconductors," was developed by Kitaev and by Schnyder, Ryu, Furusaki, and Ludwig. It uses the mathematical theory of K-theory to systematically enumerate all possible topological phases in each dimension and symmetry class.
Majorana Fermions and Topological Quantum Computing
Majorana Fermions in Condensed Matter
In 1937, Ettore Majorana proposed that certain fermions could be their own antiparticles: $\gamma = \gamma^\dagger$. While no fundamental Majorana fermion has been confirmed (neutrinos are candidates), condensed matter physics offers a different realization. In certain topological superconductors, the quasiparticle excitations at boundaries or in vortex cores are Majorana modes: they satisfy $\gamma = \gamma^\dagger$ and have zero energy.
A key model is the Kitaev chain (Alexei Kitaev, 2001): a 1D chain of spinless fermions with p-wave superconducting pairing. In the topological phase, there is one Majorana zero mode at each end of the chain. These two Majorana modes together form one ordinary fermion, but it is highly nonlocal, being split across the entire length of the chain.
Non-Abelian Statistics
In 2D, Majorana zero modes bound to vortices obey non-abelian statistics: exchanging (braiding) two Majorana vortices does not simply multiply the wave function by a phase (as for bosons or fermions) but applies a unitary transformation to the degenerate ground state. The result depends on the order in which exchanges are performed. This is the property that makes Majorana modes potentially useful for quantum computing.
Topological Quantum Computing
The idea of topological quantum computing, proposed by Kitaev and developed by Freedman, Kitaev, Larsen, and Wang, exploits non-abelian anyons to perform fault-tolerant quantum computation. The quantum information is stored in the collective state of multiple Majorana modes and is manipulated by braiding them around each other. Because the computational result depends only on the topology of the braids (not on their precise geometry), topological qubits are inherently protected from local noise and decoherence.
Experimental realization has been pursued vigorously but remains challenging. Signatures consistent with Majorana zero modes have been reported in semiconductor nanowires coupled to superconductors (InSb and InAs nanowires with Al or NbTiN), but definitive proof, particularly of non-abelian statistics, is still lacking.
A Foundational Question
Topological protection sounds like a free lunch: quantum information that is automatically protected from errors. What is the catch? The protection is only against local perturbations. Processes that are topologically nontrivial (like creating a pair of anyons, braiding one around the system, and annihilating them) can still cause errors. As the system size grows, the energy splitting between degenerate ground states decreases exponentially, improving the protection. But at finite temperature, thermally excited quasiparticles provide a mechanism for errors. Topological protection reduces but does not eliminate the need for quantum error correction.
Connections to Other Areas
Connection: Fractons and Higher-Order Topology
Fracton phases (lesson 58) represent a frontier of topological matter that goes beyond the conventional classification. In fracton topological order, excitations have restricted mobility, ground state degeneracy depends on the geometry (not just the topology) of the system, and the mathematical framework of topological quantum field theory breaks down. Fractons show that our understanding of topological phases is still far from complete. The relationship between fracton order and gravity, through higher-rank gauge theories, hints at deep connections between condensed matter topology and spacetime structure.
Connection: Berry Phase and Gauge Theories
The Berry connection $\mathbf{A}(\mathbf{k})$ and Berry curvature $\Omega(\mathbf{k})$ are mathematically identical to a $U(1)$ gauge field and its field strength. The Chern number is the analog of a magnetic monopole charge. This is not a coincidence: the mathematical framework underlying both topological band theory and particle physics gauge theories (lesson 35) is fiber bundle theory. The same structures that classify topological insulators also classify instantons in QCD and magnetic monopoles in grand unified theories.
Experimental Highlights
The field of topological matter has grown explosively since the 2000s. Key experimental milestones include:
- 1980: Integer quantum Hall effect (von Klitzing). Nobel Prize 1985.
- 1982: Fractional quantum Hall effect (Tsui, Stormer, Gossard). Nobel Prize 1998 (Laughlin, Stormer, Tsui).
- 2007: Quantum spin Hall effect observed in HgTe quantum wells (Konig et al.), confirming the prediction of Bernevig, Hughes, and Zhang.
- 2008: 3D topological insulator surface states observed in Bi$_{1-x}$Sb$_x$ by ARPES (Hsieh et al.).
- 2012: Signatures of Majorana zero modes in InSb nanowires (Mourik et al.), though interpretation remains debated.
- 2016: Nobel Prize in Physics to Thouless, Haldane, and Kosterlitz "for theoretical discoveries of topological phase transitions and topological phases of matter."
Exercises
- Landau levels: For an electron in a magnetic field $B = 10$ T, calculate the cyclotron frequency $\omega_c$, the Landau level spacing $\hbar\omega_c$ in meV, and the corresponding temperature $T^* = \hbar\omega_c/k_B$. Use $m = 0.067 m_e$ (GaAs effective mass). Why must experiments be done at low temperature?
- Filling factor: A 2D electron gas has density $n = 3 \times 10^{15}$ m$^{-2}$ in a magnetic field $B = 6$ T. Calculate the filling factor $\nu = nh/(eB)$. What is the expected quantized Hall conductance?
- Berry phase of a spin-1/2: A spin-1/2 particle in a magnetic field $\mathbf{B}$ that slowly rotates around a cone of half-angle $\theta$ acquires a Berry phase $\gamma = -\pi(1 - \cos\theta)$. Calculate the Berry phase when $\theta = \pi/2$ (field rotates in a plane). What is the geometric interpretation of this result?
- Chern number argument: The Chern number is an integer. Explain why this means that the quantum Hall conductance cannot change smoothly as disorder is introduced. What must happen for the conductance to change from $\nu e^2/h$ to $(\nu-1)e^2/h$?
- Topological protection: In a topological insulator, backscattering of edge states requires spin flip. Explain why non-magnetic impurities cannot backscatter helical edge states, while magnetic impurities can. What does this predict about the stability of edge conductance in the presence of different types of disorder?
- Topology provides a classification of quantum phases based on integer-valued invariants (like the Chern number) that cannot change under smooth deformations of the Hamiltonian, explaining the extraordinary precision of the quantum Hall effect.
- The Berry phase is a geometric phase acquired by quantum states during adiabatic evolution; the Berry curvature of Bloch bands plays the role of a gauge field in momentum space, and its integral over the Brillouin zone gives the Chern number.
- Topological insulators are materials that are insulating in the bulk but have robust, gapless conducting states on their surface, protected by symmetry and the bulk-boundary correspondence.
- The periodic table of topological phases classifies all possible topological insulators and superconductors by dimension and symmetry class, using the mathematical framework of K-theory.
- Majorana fermions, predicted to appear in topological superconductors, obey non-abelian statistics and could enable intrinsically fault-tolerant topological quantum computing, though experimental confirmation remains a major challenge.
- Fracton phases (lesson 58) go beyond the conventional topological classification, showing that our understanding of topological matter continues to evolve with deep connections to gravity and quantum information.