Topological Superconductors

A topological superconductor is a superconducting phase whose gapped bulk Bogoliubov quasiparticles cannot be deformed to a topologically trivial superconducting reference while the relevant symmetries and bulk gap are preserved. Its bulk obstruction is expressed through a Bogoliubov–de Gennes (BdG) invariant, and a boundary or topological defect can consequently support Majorana modes.

This definition has three important qualifications.

1. The topology belongs to a bulk phase, not to an isolated zero-bias feature.
2. The standard classification discussed here is a classification of gapped quadratic BdG Hamiltonians. Nodes, strong interactions, crystalline symmetries, and disorder can require refined tools.
3. A Majorana zero mode is a self-conjugate emergent quasiparticle operator. It is not evidence that a fundamental relativistic Majorana particle has been discovered.

A near-zero state, a gapped topological phase, and a controlled non-Abelian operation are different achievements. This page explains the physics that relates them. Anyons and Braiding develops the operations, while Topological Qubits explains encoding and control.

Required background. BCS Mean-Field Theory supplies the pairing saddle, Bogoliubov Quasiparticles supplies Nambu branch counting, and Topology in Quantum Matter explains why a gap and specified symmetries constrain continuous deformations.

Helpful background. Symmetry Classification Preview supplies tenfold-way language, the Kitaev Chain supplies the minimal class-D model, and the superconductivity gateway supplies the material and proximity-effect preparation.

From Pairing to Topology

A conventional band insulator is described by occupied electron bands separated from empty electron bands. A superconductor is different: its condensate mixes electron annihilation and creation operators, so a quasiparticle generally carries neither a definite electron number nor a definite electric charge. The natural single-particle description is therefore written in Nambu space, where particles and holes appear together.

The mean-field Hamiltonian does not conserve the electron number operator $\hat N$. It does preserve fermion parity,

$$\hat P_f = (-1)^{\hat N},$$

because pairing creates or annihilates electrons in pairs. In an isolated exact system, fermion parity is a genuine superselection label. In an open device, quasiparticle poisoning can exchange a single electron with the environment and spoil the effective parity lifetime.

Topology requires a specified gap. For a fully gapped superconductor, the positive BdG spectrum obeys

$$E_n(\mathbf k) \geq \Delta_{\mathrm{qp}} > 0$$

throughout the Brillouin zone. A nodal superconductor can possess topological charges attached to points or lines of nodes, as well as protected surface flat bands or arcs, but it is not classified by simply inserting its Hamiltonian into the table for fully gapped phases.

The condensate itself need not have an exotic microscopic origin. Conventional $s$-wave pairing can produce an effective topological superconductor when combined with spin–orbit coupling, Zeeman splitting, a topological-insulator surface, or an appropriately structured junction. Conversely, unconventional pairing does not automatically imply nontrivial topology. Pairing symmetry, normal-state band structure, dimensionality, and protecting symmetries must be considered together.

Bogoliubov–de Gennes Hamiltonians

Let $h(\mathbf k)$ be an $N\times N$ normal-state Bloch Hamiltonian, including orbital and spin labels, and let $\Delta(\mathbf k)$ be its pairing matrix. For the Nambu spinor

$$\Psi_{\mathbf k} = \begin{pmatrix} c_{\mathbf k1} \\ \vdots \\ c_{\mathbf kN} \\ c^\dagger_{-\mathbf k1} \\ \vdots \\ c^\dagger_{-\mathbf kN} \end{pmatrix},$$

the quadratic mean-field Hamiltonian can be written

$$\hat H_{\mathrm{MF}} = \frac{1}{2} \sum_{\mathbf k} \Psi_{\mathbf k}^\dagger \mathcal H_{\mathrm{BdG}}(\mathbf k) \Psi_{\mathbf k} + E_0,$$

with

$$\mathcal H_{\mathrm{BdG}}(\mathbf k) = \begin{pmatrix} h(\mathbf k)-\mu & \Delta(\mathbf k) \\ \Delta^\dagger(\mathbf k) & -h^T(-\mathbf k)+\mu \end{pmatrix}.$$

The factor $1/2$ compensates for Nambu doubling. Counting every eigenvalue of the doubled matrix as an independent excitation would count the same physical degree of freedom twice.

Fermi statistics impose

$$\Delta(\mathbf k) = -\Delta^T(-\mathbf k).$$

This is the compact form of the antisymmetry of a Cooper pair under exchange. A spin-singlet pair can be even in momentum because its spin tensor is antisymmetric; a single-component or spin-polarized pair must be odd in momentum.

Gauge conventions

Under a global electron rephasing $c\mapsto e^{i\chi/2}c$, the Nambu spinor transforms as

$$\Psi \mapsto e^{i\chi\tau_z/2}\Psi,$$

and the pairing matrix changes by $\Delta\mapsto e^{i\chi}\Delta$. A uniform condensate phase can therefore be moved between the off-diagonal matrix elements and the Nambu basis. Relative phases across a junction and phase winding around a vortex remain physical.

The Pauli matrices $\tau_i$ below act in particle–hole space. Spin Pauli matrices are written $s_i$. In a multiorbital effective model, the meaning of either set must be stated; a symbol appearing like physical spin can instead label a pseudospin doublet.

Particle–Hole Redundancy

Every BdG Hamiltonian has an antiunitary particle–hole relation

$$\mathcal C \mathcal H_{\mathrm{BdG}}(\mathbf k) \mathcal C^{-1} = -\mathcal H_{\mathrm{BdG}}(-\mathbf k).$$

In the canonical Nambu basis one may choose

$$\mathcal C = \tau_xK, \qquad \mathcal C^2=+1,$$

where $K$ denotes complex conjugation. Thus, if

$$\mathcal H_{\mathrm{BdG}}(\mathbf k) \Phi_{n\mathbf k} = E_{n\mathbf k} \Phi_{n\mathbf k},$$

then

$$\tau_x\Phi_{n\mathbf k}^*$$

is an eigenvector at $-\mathbf k$ with energy $-E_{n\mathbf k}$. In second-quantized language, the quasiparticle associated with the negative-energy partner is the creation operator for the positive-energy mode:

$$\alpha_{-E,-\mathbf k} = \alpha^\dagger_{E,\mathbf k}.$$

This relation is often called particle–hole symmetry, but it is first of all a redundancy introduced by the Nambu description. It is not ordinary charge conservation, not a transformation that maps the laboratory to a distinct state of equal positive energy, and not relativistic charge conjugation. Additional physical symmetries, such as time reversal, inversion, mirror symmetry, or spin rotation, constrain the BdG Hamiltonian independently.

At exactly zero energy, the particle–hole partner lies in the same eigenspace. A real basis for that eigenspace can be chosen so that each basis vector satisfies

$$\mathcal C\Phi_0 = \Phi_0.$$

The corresponding quasiparticle operator obeys

$$\gamma^\dagger = \gamma.$$

After normalization, this self-adjoint operator is a Majorana zero mode. A finite closed system has an even number of Majorana operators; “one at an end” describes spatial separation, not an odd total number. Particle–hole redundancy permits such an operator, but it does not force every superconductor to have one. Existence and protection require topology, symmetry, or fine tuning.

Symmetry Classes and Dimensions

The stable free-fermion classification asks which gapped BdG Hamiltonians can be continuously deformed into one another after adding trivial bands. Three antiunitary or unitary constraints organize the problem:

$$\begin{aligned} \mathcal T \mathcal H(\mathbf k) \mathcal T^{-1} &= \mathcal H(-\mathbf k), \\ \mathcal C \mathcal H(\mathbf k) \mathcal C^{-1} &= -\mathcal H(-\mathbf k), \\ \mathcal S \mathcal H(\mathbf k) \mathcal S^{-1} &= -\mathcal H(\mathbf k). \end{aligned}$$

Here $\mathcal T$ is physical time reversal, $\mathcal C$ is BdG particle–hole conjugation, and chiral symmetry $\mathcal S$ can arise from their product. The most common symmetry classes for topological superconductors are:

Class	Time reversal	1D	2D	3D
D	absent	$\mathbb Z_2$	$\mathbb Z$	$0$
DIII	$\mathcal T^2=-1$	$\mathbb Z_2$	$\mathbb Z_2$	$\mathbb Z$
BDI	$\mathcal T^2=+1$	$\mathbb Z$	$0$	$0$

All three rows use $\mathcal C^2=+1$; DIII and BDI also have chiral symmetry. These are useful cases of the free classification of Schnyder and colleagues and Kitaev. Class D describes a generic superconductor without time reversal or extra spin-rotation constraints. Class DIII includes spinful time reversal without full spin-rotation symmetry. Spin-rotation-invariant systems can instead require classes C or CI after reduction to irreducible blocks; the canonical Nambu redundancy alone does not select a row.

A real Kitaev chain has an additional effective antiunitary symmetry squaring to $+1$ and lies in BDI. This effective symmetry need not be the microscopic time reversal of a spinful electron. Complex perturbations that break it while retaining particle–hole conjugation reduce the robust distinction to the class-D parity.

The entries in the table are stable, noninteracting, internal-symmetry classifications. Interactions can identify phases that are distinct in the quadratic table; for example, the one-dimensional BDI integer reduces to a $\mathbb Z_8$ distinction and the three-dimensional DIII integer reduces to a $\mathbb Z_{16}$ distinction under suitable symmetry-preserving interactions. Crystalline symmetry can add weak, mirror, rotation, hinge, or higher-order indices. Disorder can preserve a real-space or scattering invariant even after momentum ceases to be a good quantum number.

Whatever invariant is used, it cannot change under a continuous deformation unless at least one defining condition fails. Typically the quasiparticle gap closes:

$$\Delta_{\mathrm{qp}} \longrightarrow 0.$$

A gap closing is therefore necessary at a generic quadratic topological phase transition, but a measured gap closing by itself does not identify which invariant changed. Topological Phase Transitions derives the ideal nanowire criterion and separates quasiparticle closings from first-order, disordered, and interacting routes.

One-Dimensional Class D

For a translation-invariant one-dimensional class-D superconductor with $\mathcal C=\tau_xK$, define at the particle–hole-invariant momenta $k_\star=0,\pi$

$$B(k_\star) = \mathcal H_{\mathrm{BdG}}(k_\star)\tau_x.$$

Particle–hole conjugation makes $B(k_\star)$ antisymmetric in this basis. Using one consistent Majorana-basis orientation at both momenta, and a convention in which these Pfaffians are real and nonzero, write

$$\mathcal M = \operatorname{sgn} \left[ \operatorname{Pf}B(0) \operatorname{Pf}B(\pi) \right].$$

The phase is topological for $\mathcal M=-1$ and trivial for $\mathcal M=+1$. A change of sign requires a Pfaffian to vanish, which is precisely a bulk gap closing at one of the invariant momenta. Gauge-covariant formulations or scattering-matrix invariants should be used when the convenient real Pfaffian convention is unavailable.

Kitaev-chain diagnosis

For the real, nearest-neighbor Kitaev-chain convention

$$\mathcal H(k) = \left( -\mu-2t\cos k \right)\tau_z + 2\Delta\sin k\,\tau_y,$$

the pairing term vanishes at $k=0,\pi$. The invariant reduces to

$$\mathcal M = \operatorname{sgn} \left[ (\mu+2t)(\mu-2t) \right].$$

Hence

$$\lvert\mu\rvert < 2\lvert t\rvert \quad\Longrightarrow\quad \mathcal M=-1.$$

This conclusion assumes $t\ne0$ and $\Delta\ne0$. The full dispersion is

$$E(k)^2=(\mu+2t\cos k)^2+4\Delta^2\sin^2k.$$

At $\Delta=0$ and $|\mu|<2|t|$, the expression for the endpoint signs is still negative, but there are gapless Fermi points between them. It does not then classify a gapped superconductor. With nonzero pairing, a sufficiently long open chain in the topological interval has one approximately zero-energy Majorana mode at each end. At $\lvert\mu\rvert=2\lvert t\rvert$, the bulk gap closes and the localization length diverges. The Kitaev Chain entry develops the lattice Hamiltonian and solvable limits; the point here is the general correspondence between a bulk Pfaffian parity and an odd number of protected boundary Majoranas.

Majorana Boundary Modes

A normalized Majorana operator localized near a boundary has the form

$$\gamma_a = \sum_j \left( u_{aj}c_j + u_{aj}^*c_j^\dagger \right),$$

with $\sum_j|u_{aj}|^2=1$ and mutually orthogonal real mode vectors. Its algebra is

$$\gamma_a^\dagger=\gamma_a, \qquad \{\gamma_a,\gamma_b\} = 2\delta_{ab}.$$

Two separated Majoranas form one ordinary complex fermion:

$$f = \frac{\gamma_L+i\gamma_R}{2}, \qquad i\gamma_L\gamma_R = 2f^\dagger f-1.$$

The occupation $n_f=0,1$ changes the fermion parity associated with the pair. A finite overlap produces the effective Hamiltonian

$$\hat H_{\mathrm{split}} = \frac{i\varepsilon}{2} \gamma_L\gamma_R = \varepsilon \left( f^\dagger f-\frac{1}{2} \right).$$

For a clean long wire, the splitting commonly behaves as

$$\varepsilon(L) \sim A e^{-L/\xi} \cos(k_FL+\varphi),$$

where $\xi$ is the Majorana localization length. The oscillation is not itself a topological invariant; it reflects microscopic phase accumulation and can be strongly modified by disorder, multiple bands, and smooth confinement.

Local perturbations near one end cannot directly implement the nonlocal bilinear $i\gamma_L\gamma_R$ when the two wavefunctions do not overlap. This is the origin of exponential protection in an idealized wire. It is not absolute protection:

• finite overlap splits the pair;
• a local perturbation can close the protecting gap;
• accidental low-energy Andreev states can couple to an end mode;
• quasiparticle poisoning changes total fermion parity;
• measurement leads can broaden or decohere the state;
• a single pair does not provide two usable states within a fixed total-parity sector.

Encoding a parity-conserving qubit requires at least four Majoranas or an equivalent multi-island architecture, plus controlled measurement and initialization. Braiding separated Majorana defects implements noncommuting operations only when the evolution remains adiabatic relative to the smallest excitation gap encountered, including vortex-core or other subgap levels, fast relative to poisoning, and insensitive to residual splittings. The time dependence of the Hamiltonian also matters; a bulk gap value alone does not set a safe operation time. Those operations are developed in Anyons and Braiding.

A local low-energy state versus separated end modes

Any ordinary fermionic level can be decomposed into two self-adjoint operators. The important question is where their wavefunctions live and what sets their coupling. The figure compares two open Kitaev chains with $t=\Delta>0$: a uniform topological chain with $\mu_j=t$, and a chain with $\mu_j=t$ on the first five sites but $\mu_j=3t$ everywhere else. The second system contains a short topological segment at its left end, while its long bulk is trivial. Its two low-energy Majorana components remain near that same end.

Two components of the lowest positive-energy state in a 24-site chain. Each curve is normalized to unit total weight. In the uniform chain, increasing total length separates the components and reduces their coupling. In the second chain, extending the trivial region leaves their separation nearly unchanged. These are calculated model wavefunctions, not conductance traces; self-adjointness or an end-localized signal alone does not determine the topology of the long bulk.

For example, increasing the chain from 12 to 24 sites changes its lowest positive energy from $3.66\times10^{-4}t$ to $8.94\times10^{-8}t$ in the uniform case, but only from $0.03717t$ to $0.03714t$ in the end-segment case. Enlarging the short segment could suppress the latter energy too. Length dependence must therefore identify which physical region is being extended. Dot and confinement models provide related ways for Andreev states to imitate local Majorana spectroscopy (Liu and colleagues).

Two- and Three-Dimensional Phases

Chiral superconductors

A spin-polarized continuum model with upper-right pairing proportional to $k_x+ik_y$ is

$$\mathcal H(\mathbf k)=\xi_{\mathbf k}\tau_z +\hbar v_\Delta(k_x\tau_x-k_y\tau_y),$$ $$\xi_{\mathbf k}=\frac{\hbar^2k^2}{2m^*}-\mu.$$

Here $m^*>0$, $v_\Delta>0$ is a fixed pairing velocity, and $\mathbf k$ is a wavevector. Using a fixed velocity avoids referring to a Fermi wavevector that would cease to exist for $\mu<0$.

Use $A_i=+i\langle u|\partial_{k_i}u\rangle$ and oriented area $dk_x\wedge dk_y$. For a lattice model, integrate the negative-energy bands over the Brillouin zone:

$$C=\frac{1}{2\pi}\sum_{E_n<0}\int_{\mathrm{BZ}}\Omega_{n,xy}\,d^2k.$$

For the displayed continuum model the domain is instead $\mathbb R^2$. Its quadratic kinetic term makes the unit vector $\widehat{\mathbf d}$ approach $+\widehat z$ in every direction at infinity, compactifying the momentum plane. When $\mu>0$, $\widehat{\mathbf d}$ starts at $-\widehat z$ at the origin and covers the sphere once, giving $C=+1$ for the stated chirality. When $\mu<0$, it starts and ends at the north pole and $C=0$. At $\mu=0$ the spectrum closes at $\mathbf k=0$. Dropping the quadratic term and integrating an unregularized Dirac model would lose this integer conclusion.

This is the Read–Green weak- versus strong-pairing distinction: two phases with the same broken symmetries can have different BdG topology.

Place the class-D material at $x<0$ and trivial vacuum at $x>0$, and count positive velocity along $+y$ as right-moving. Bulk–boundary correspondence gives $N_{\mathrm R}-N_{\mathrm L}=-C$ for Majorana branches at this boundary. Since a Majorana branch has central charge $1/2$, the signed chiral central charge is

$$c_- = -\frac{C}{2}.$$

Define the transport heat tensor by $J_x^Q=-\kappa_{xy}\partial_yT$ (with the longitudinal gradient set to zero), and $\kappa_{yx}=-\kappa_{xy}$. In the ideal low-temperature limit its thermal Hall response is

$$\frac{\kappa_{xy}}{T} = c_-\frac{\pi^2k_B^2}{3h} = -C\frac{\pi^2k_B^2}{6h}.$$

The factor of one half relative to a chiral complex fermion is characteristic of a Majorana channel. As a check, embed a charge-conserving Chern insulator with occupied-band invariant $C_{\mathrm{el}}$ in BdG form without pairing. Its negative-energy electron bands and conjugated unoccupied hole bands each contribute $C_{\mathrm{el}}$, so $C=2C_{\mathrm{el}}$. The thermal formula reduces to

$$\begin{aligned} \frac{\kappa_{xy}}{T} &=-C_{\mathrm{el}}\frac{\pi^2k_B^2}{3h} \\ &=\frac{\pi^2k_B^2}{3e^2}\sigma_{xy}, \end{aligned}$$

which is the normal-state Wiedemann–Franz result in the same tensor convention. BdG quasiparticles do not carry a conserved $U(1)$ charge, and the electromagnetic response also includes the condensate. Thus $C$ fixes the ideal chiral-Majorana thermal response, but it does not fix an analogous universal quantized electric Hall conductance.

Time-reversal-invariant superconductors

Class DIII preserves spinful time reversal with $\mathcal T^2=-1$. Its protected boundaries appear in dimension-dependent forms:

• a one-dimensional phase has a Kramers pair of Majorana end modes;
• a two-dimensional phase has a helical Majorana edge;
• a three-dimensional phase can have a gapless Majorana surface cone.

After flattening a three-dimensional chiral-symmetric BdG Hamiltonian into an off-diagonal unitary block $q(\mathbf k)$, define

$$\omega_i=q^{-1}\partial_iq.$$

One convention for the integer winding number is

$$\nu_3=\frac{1}{24\pi^2}\int_{\mathrm{BZ}}d^3k\, \epsilon^{ijk}\operatorname{Tr}(\omega_i\omega_j\omega_k).$$

Sign conventions depend on the orientation and block definition, but $\lvert\nu_3\rvert$ counts the stable surface-cone multiplicity in the minimal free description. Superfluid $^3\mathrm{He}$-B is a benchmark class-DIII topological neutral superfluid. It demonstrates the phase structure without being an electronic superconducting material.

For a centrosymmetric, fully gapped, time-reversal-invariant superconductor, in the weak-pairing regime, odd-parity pairing and an odd total number of enclosed time-reversal-invariant momenta give a sufficient criterion for a nontrivial time-reversal-invariant phase (Fu and Berg). It is not a substitute for determining the pairing symmetry: uncertainty about the order parameter propagates directly into uncertainty about the topological label.

Vortices and Zero Modes

A vortex carries condensate phase winding

$$\oint\nabla\phi\cdot d\boldsymbol\ell = 2\pi n.$$

The order parameter is suppressed in its core, so subgap states are expected even in a conventional superconductor. In the semiclassical regime $E_F\gg\Delta$, an ordinary singly quantized vortex has Caroli–de Gennes–Matricon levels of order

$$E_m \simeq \left( m+\frac{1}{2} \right)\omega_0, \qquad \omega_0 \sim \frac{\Delta^2}{E_F},$$

with integer $m$ for a simple isotropic conventional vortex; spin can supply additional copies. Here $\omega_0$ denotes an energy spacing, despite its frequency-like name. These levels can lie very close to zero when $\Delta/E_F$ is small. A near-zero vortex peak is therefore not automatically a Majorana mode.

In a two-dimensional chiral topological superconductor with odd Chern parity, an odd-vorticity defect can bind an unpaired Majorana zero mode. Spectral flow effectively removes the half-level offset for one branch, leaving a state pinned to zero by particle–hole conjugation unless it couples to another Majorana or the protecting assumptions fail. A proximitized topological-insulator surface provides another route: spin-momentum locking plus ordinary $s$-wave pairing yields an effective spinless pairing structure, and a vortex can bind a Fu–Kane Majorana mode.

For a finite collection of vortices, zero modes hybridize pairwise:

$$\hat H_{\mathrm{M}} = \frac{i}{2} \sum_{a<b} t_{ab}\gamma_a\gamma_b.$$

The amplitudes $t_{ab}$ depend exponentially and oscillatory on separation in a clean system. Disorder, anisotropy, nearby surfaces, other vortex-core levels, and three-dimensional vortex-line dispersion complicate the spectrum. A resolved zero-energy peak must be compared with these other core states and with the appropriate bulk or defect invariant.

Engineered Platforms

Spin–orbit nanowires

For this nanowire model, change from the canonical Nambu order above to the time-reversed-hole basis

$$\widetilde\Psi_k = \begin{pmatrix} c_{k\uparrow} \\ c_{k\downarrow} \\ c^\dagger_{-k\downarrow} \\ -c^\dagger_{-k\uparrow} \end{pmatrix}.$$

Here $s_i$ and $\tau_i$ act on spin and the two particle–hole blocks, respectively. Choose the uniform induced pairing $\Delta_{\mathrm{ind}}$ real. An ideal single-band Rashba nanowire with Zeeman energy $V_Z$ then has

$$\mathcal H_{\mathrm{NW}}(k) ={} \left( \frac{\hbar^2k^2}{2m^*}-\mu \right)\tau_z + \alpha k\,s_y\tau_z + V_Zs_x + \Delta_{\mathrm{ind}}\tau_x.$$

In this basis the same particle–hole redundancy is represented by $\widetilde{\mathcal C}=\tau_y s_y K$, with $\widetilde{\mathcal C}^2=+1$ and

$$\widetilde{\mathcal C}\mathcal H_{\mathrm{NW}}(k) \widetilde{\mathcal C}^{-1} =-\mathcal H_{\mathrm{NW}}(-k).$$

The earlier $\tau_xK$ and Pfaffian formulas apply in the canonical Nambu basis; they must be transformed along with the Hamiltonian before use here.

At $k=0$, the ideal bulk gap closes when

$$V_Z^2 = \mu^2+\Delta_{\mathrm{ind}}^2.$$

The single-band topological regime lies on the side

$$V_Z^2 > \mu^2+\Delta_{\mathrm{ind}}^2,$$

provided a reopened gap survives. Spin–orbit coupling converts the spin-singlet proximity effect into an effective odd-momentum pairing channel in the helical band.

This criterion is a model result, not a device certification. Real devices have several transverse bands, electrostatic inhomogeneity, orbital magnetic effects, field-dependent parent pairing, interface disorder, smooth quantum dots, finite length, dissipative leads, and interactions. A large Zeeman field can satisfy the algebraic inequality while simultaneously destroying the induced gap.

Interpreting spectroscopy and transport

The nanowire mechanism proposed by Lutchyn, Sau and Das Sarma and independently by Oreg, Refael and von Oppen is one route. A topological-insulator surface paired by an ordinary superconductor is another (Fu and Kane). Intrinsic odd-parity pairing provides a third possibility when its symmetry and gap satisfy the relevant criterion. Each requires its own Hamiltonian and gap analysis; a material name does not select an invariant.

Measurement	What it can probe	What else matters
End tunneling	Local low-energy spectrum	Andreev states and broadening
Nonlocal conductance	Extended excitations	Contacts and charge conversion
Vortex tunneling	Core levels	Ordinary core spacing
Josephson response	Parity-dependent branches	Relaxation and drive dynamics
Thermal transport	Chiral energy channels	Phonons and equilibration
Fusion or braiding	Non-Abelian operations	Initialization and control errors

For an isolated Majorana coupled to a single effective normal-lead channel, resonant Andreev reflection gives the ideal zero-temperature line shape

$$G_0(V)=\frac{2e^2}{h}\frac{\Gamma^2}{(eV)^2+\Gamma^2},$$

where $\Gamma$ is the tunnel broadening (Law, Lee and Ng). Thermal broadening averages this curve over an energy window of order $k_BT$ and lowers a narrow peak. Finite overlap, extra channels, poisoning and dissipation change the simple model. Trivial Andreev states can also give near-zero peaks, including heights close to $2e^2/h$ in appropriate regimes. A peak height is therefore a test of a resonance model, not a direct measurement of a bulk invariant.

Gap closing and reopening, end localization, and nonlocal transport constrain complementary parts of the interpretation. A local probe can miss an extended bulk closing; correlated peaks at both ends can also arise from extended trivial states. Compare parameter evolution with a model of the whole device and its contacts. For experimental non-Abelian protocols and their limits, see Anyons and Non-Abelian Statistics.

Fractional Josephson effect and 4π periodicity

The fractional Josephson effect concerns a parity-resolved spectral branch. It must be distinguished from the equilibrium free energy.

For a topological Josephson junction with two coupled Majoranas, the fixed-parity energies can take the form

$$E_\pm(\phi) = \pm E_M\cos\frac{\phi}{2}.$$

Following one parity branch through $\phi\mapsto\phi+2\pi$ changes its sign, so that branch returns only after $4\pi$. In thermal equilibrium, however, the system can switch parity at the crossing and recover a $2\pi$-periodic free energy. The phrase “$4\pi$ Josephson effect” must therefore specify the drive frequency, parity lifetime, occupation dynamics, and alternative nonequilibrium mechanisms.

Common pitfalls

Double-counting BdG states. The $E$ and $-E$ partners describe annihilation and creation of the same quasiparticle. Use positive energies for independent excitations and treat the zero-mode subspace separately.

Confusing pairing symmetry with topology. An unusual order parameter is not enough: identify dimensionality, the full quasiparticle gap, and the symmetries allowed during deformation. A nodal phase requires different invariants.

Treating a pair as a qubit. Two Majoranas form one fermion whose two occupations have opposite parity. A fixed-parity qubit needs more modes; suppressing overlap alone does not suppress poisoning or control errors.

Exercises

1. A negative Pfaffian product without a gapped phase

Take $\mu=0$, $t>0$ in the Kitaev model. Find the minimum positive bulk energy for $0<|\Delta|\leq t$. What happens at $\Delta=0$, and why is checking only $k=0,\pi$ insufficient there?

Solution

The dispersion becomes

$$E(k)^2=4t^2\cos^2k+4\Delta^2\sin^2k.$$

For $|\Delta|\leq t$, the minimum occurs at $k=\pm\pi/2$, giving $\Delta_{\mathrm{qp}}=2|\Delta|$. At zero pairing, those points are gapless although the endpoint Pfaffian product is still $-4t^2$. The gapped class-D invariant presupposes a gap everywhere in momentum space.

2. Normalize a self-conjugate mode

Let a unit-normalized zero-energy Nambu vector be $\Phi=(u,u^*)^T$. For its corresponding operator $\eta=\Phi^\dagger\Psi=\sum_j(u_j^*c_j+u_jc_j^\dagger)$, find $\{\eta,\eta\}$ and the rescaling needed for the convention $\gamma^2=1$.

Solution

The conjugates follow from projecting $\Psi$ onto $\Phi$ with $\Phi^\dagger$; using $\Phi^T$ would generally select a different operator when $u$ is complex. Nambu normalization gives $2\sum_j|u_j|^2=1$. The fermion anticommutators therefore give $\{\eta,\eta\}=2\sum_j|u_j|^2=1$, or $\eta^2=1/2$. Set $\gamma=\sqrt2\eta$. The coefficients of this normalized Majorana satisfy $\sum_j|\sqrt2u_j|^2=1$. A self-conjugate Nambu vector and a Majorana operator use different normalization conventions.

3. Build one fermion from two Majoranas

Given $\{\gamma_a,\gamma_b\}=2\delta_{ab}$, prove that

$$f = \frac{\gamma_1+i\gamma_2}{2}$$

obeys the canonical fermion algebra and derive $i\gamma_1\gamma_2=2f^\dagger f-1$.

Solution

Self-adjointness gives

$$f^\dagger = \frac{\gamma_1-i\gamma_2}{2}.$$

Using $\gamma_1^2=\gamma_2^2=1$ and $\gamma_1\gamma_2=-\gamma_2\gamma_1$,

$$\{f,f^\dagger\} = 1, \qquad f^2=(f^\dagger)^2=0.$$

The number operator is

$$\begin{aligned} f^\dagger f &= \frac{1}{4} \left( \gamma_1-i\gamma_2 \right) \left( \gamma_1+i\gamma_2 \right) \\ &= \frac{1}{2} \left( 1+i\gamma_1\gamma_2 \right). \end{aligned}$$

Rearranging yields

$$i\gamma_1\gamma_2 = 2f^\dagger f-1.$$

The bilinear measures the parity of the complex fermion, up to the stated convention.

4. Explain the half thermal channel

A two-dimensional class-D superconductor has $C=3$. Use the material-left, vacuum-right edge orientation and heat tensor defined above. Find its signed chiral central charge and ideal low-temperature $\kappa_{xy}/T$. Why is the magnitude half that of three chiral complex-fermion channels?

Solution

At this boundary $N_{\mathrm R}-N_{\mathrm L}=-3$. Each chiral Majorana branch contributes central-charge magnitude $1/2$, so

$$c_- = -\frac{C}{2} = -\frac{3}{2}.$$

Thus

$$\frac{\kappa_{xy}}{T} = c_- \frac{\pi^2k_B^2}{3h} = -\frac{\pi^2k_B^2}{2h}.$$

Three chiral complex fermions would have twice this thermal-conductance magnitude. A complex channel contains two real Majorana channels, so it carries twice the thermal central charge of one Majorana mode.

5. Derive the nanowire phase boundary

Use the time-reversed-hole Nambu basis defined in the nanowire section. At $k=0$, the Rashba term vanishes. Show that the positive excitation energies are

$$E_\pm(0) = \left| V_Z \pm \sqrt{\mu^2+\Delta_{\mathrm{ind}}^2} \right|$$

and find the gap-closing condition.

Solution

At $k=0$,

$$\mathcal H(0) = -\mu\tau_z + \Delta_{\mathrm{ind}}\tau_x + V_Zs_x.$$

The spin operator $s_x$ commutes with the Nambu matrices. In an $s_x$ eigenstate with eigenvalue $\sigma=\pm1$, the Nambu block has eigenvalues

$$E = \sigma V_Z \pm \sqrt{\mu^2+\Delta_{\mathrm{ind}}^2}.$$

Taking positive magnitudes gives the two displayed branches. The smaller branch reaches zero when

$$\lvert V_Z\rvert = \sqrt{\mu^2+\Delta_{\mathrm{ind}}^2},$$

or

$$V_Z^2 = \mu^2+\Delta_{\mathrm{ind}}^2.$$

The derivation locates an ideal bulk transition; it does not prove that a finite disordered device has entered a robust topological regime.

Continue studying

The Kitaev Chain derives the lattice model and its solvable limits. Proximity and Andreev Physics explains ordinary subgap spectroscopy, while Vortex Matter discusses vortex motion and material response. Bulk–Boundary Correspondence connects the invariant to boundary modes, and Symmetry-Protected Topological Phases explains what interactions change.

References

• C. Caroli, P. G. de Gennes, and J. Matricon, “Bound Fermion States on a Vortex Line in a Type II Superconductor,” Physics Letters 9, 307–309 (1964), doi:10.1016/0031-9163(64)90375-0.
• L. Fu and C. L. Kane, “Superconducting Proximity Effect and Majorana Fermions at the Surface of a Topological Insulator,” Physical Review Letters 100, 096407 (2008), doi:10.1103/PhysRevLett.100.096407.
• L. Fu and E. Berg, “Odd-Parity Topological Superconductors: Theory and Application to $\mathrm{Cu_xBi_2Se_3}$,” Physical Review Letters 105, 097001 (2010), doi:10.1103/PhysRevLett.105.097001.
• A. Y. Kitaev, “Unpaired Majorana Fermions in Quantum Wires,” Physics-Uspekhi 44, 131–136 (2001), doi:10.1070/1063-7869/44/10S/S29.
• A. Kitaev, “Periodic Table for Topological Insulators and Superconductors,” AIP Conference Proceedings 1134, 22–30 (2009), doi:10.1063/1.3149495.
• K. T. Law, P. A. Lee, and T. K. Ng, “Majorana Fermion Induced Resonant Andreev Reflection,” Physical Review Letters 103, 237001 (2009), doi:10.1103/PhysRevLett.103.237001.
• C.-X. Liu, J. D. Sau, T. D. Stanescu, and S. Das Sarma, “Andreev Bound States versus Majorana Bound States in Quantum Dot–Nanowire–Superconductor Hybrid Structures: Trivial versus Topological Zero-Bias Conductance Peaks,” Physical Review B 96, 075161 (2017), doi:10.1103/PhysRevB.96.075161.
• R. M. Lutchyn, J. D. Sau, and S. Das Sarma, “Majorana Fermions and a Topological Phase Transition in Semiconductor–Superconductor Heterostructures,” Physical Review Letters 105, 077001 (2010), doi:10.1103/PhysRevLett.105.077001.
• Y. Oreg, G. Refael, and F. von Oppen, “Helical Liquids and Majorana Bound States in Quantum Wires,” Physical Review Letters 105, 177002 (2010), doi:10.1103/PhysRevLett.105.177002.
• X.-L. Qi, T. L. Hughes, S. Raghu, and S.-C. Zhang, “Time-Reversal-Invariant Topological Superconductors and Superfluids in Two and Three Dimensions,” Physical Review Letters 102, 187001 (2009), doi:10.1103/PhysRevLett.102.187001.
• N. Read and D. Green, “Paired States of Fermions in Two Dimensions with Breaking of Parity and Time-Reversal Symmetries and the Fractional Quantum Hall Effect,” Physical Review B 61, 10267–10297 (2000), doi:10.1103/PhysRevB.61.10267.
• A. P. Schnyder, S. Ryu, A. Furusaki, and A. W. W. Ludwig, “Classification of Topological Insulators and Superconductors in Three Spatial Dimensions,” Physical Review B 78, 195125 (2008), doi:10.1103/PhysRevB.78.195125.

Further reading

• J. Alicea, “New Directions in the Pursuit of Majorana Fermions in Solid State Systems,” Reports on Progress in Physics 75, 076501 (2012), doi:10.1088/0034-4885/75/7/076501.
• C. W. J. Beenakker, “Search for Majorana Fermions in Superconductors,” Annual Review of Condensed Matter Physics 4, 113–136 (2013), doi:10.1146/annurev-conmatphys-030212-184337.
• B. A. Bernevig and T. L. Hughes, Topological Insulators and Topological Superconductors, Princeton University Press, 2013.
• M. Sato and Y. Ando, “Topological Superconductors: A Review,” Reports on Progress in Physics 80, 076501 (2017), doi:10.1088/1361-6633/aa6ac7.