6.3 BCS theory

BCS theory explains how electrons pair up in superconductors, overcoming their mutual repulsion to form Cooper pairs. This pairing, mediated by lattice vibrations (phonons), opens an energy gap that gives superconductors their zero-resistance and diamagnetic properties.

Developed by Bardeen, Cooper, and Schrieffer in 1957, BCS theory was the first successful microscopic explanation of superconductivity. It quantitatively predicts phenomena like the isotope effect and the critical temperature, and it laid the conceptual groundwork for technologies ranging from SQUIDs to superconducting qubits.

Foundations of BCS theory

BCS theory provides a quantum mechanical description of how electrons in a crystal lattice can form bound pairs and condense into a single coherent quantum state. This microscopic picture explains why certain materials, when cooled below a critical temperature, lose all electrical resistance.

Cooper pair formation

Two electrons with opposite spins and opposite momenta can form a bound state called a Cooper pair, even though electrons normally repel each other. The binding is indirect: one electron distorts the positively charged ion lattice as it moves through, and a second electron is attracted to that distortion. The net effect is a weak attractive interaction that wins out over the screened Coulomb repulsion at low energies.

• The binding energy of a Cooper pair is small (on the order of meV), but any net attraction, no matter how weak, is enough to form a bound state at the Fermi surface. This is Cooper's key result.
• Each pair has zero total momentum and zero total spin (a singlet state).
• Because Cooper pairs carry integer spin, they obey Bose statistics. This allows all pairs to condense into the same quantum ground state, much like a Bose-Einstein condensate.

Electron-phonon interaction

The attractive glue holding Cooper pairs together is the electron-phonon interaction. Here's how it works step by step:

1. An electron moves through the lattice and attracts nearby positive ions toward it, creating a region of slightly higher positive charge density.
2. The lattice distortion persists for a short time (set by the phonon frequency) after the first electron has moved on.
3. A second electron, arriving in that region, feels the residual positive polarization and is effectively attracted toward the first electron's former position.
4. This exchange of a virtual phonon produces a net attractive interaction between the two electrons.

The strength of this coupling is characterized by the electron-phonon coupling constant $\lambda$. In weak-coupling BCS theory, $\lambda \ll 1$; materials with larger $\lambda$ require strong-coupling extensions like Eliashberg theory.

Energy gap in superconductors

The formation of Cooper pairs opens a superconducting energy gap $\Delta$ around the Fermi energy. This gap represents the minimum energy needed to break a Cooper pair into two unpaired quasiparticles.

• Typical gap values are on the order of 0.1–3 meV (for comparison, $k_B T$ at room temperature is about 25 meV, so thermal fluctuations easily destroy pairing at high temperatures).
• The gap is largest at $T = 0$ and shrinks continuously to zero as $T \to T_c$.
• Because there are no available low-energy single-particle excitations, scattering processes that cause electrical resistance are frozen out. This is the microscopic origin of zero resistance.

Key concepts in BCS theory

Coherent ground state

In the superconducting state, all Cooper pairs occupy the same quantum state, described by a macroscopic wave function $\Psi = |\Psi| e^{i\phi}$. The amplitude $|\Psi|^2$ is proportional to the superfluid density, and the phase $\phi$ is coherent across the entire sample.

• This long-range phase coherence is what produces zero resistance and the Meissner effect.
• The state spontaneously breaks the $U(1)$ gauge symmetry associated with charge conservation, which has deep consequences for the electromagnetic response and for topological properties of the order parameter.
• The analogy to Bose-Einstein condensation is useful but not exact: Cooper pairs overlap heavily in real space (their size $\xi_0$ is typically hundreds of nanometers), so they are very different from tightly bound bosons.

Quasiparticle excitations

When a Cooper pair breaks, the resulting excitations are Bogoliubov quasiparticles, which are quantum superpositions of electron-like and hole-like states. Their energy dispersion is:

$E_k = \sqrt{\xi_k^2 + \Delta^2}$

where $\xi_k$ is the normal-state energy measured from the Fermi level. The minimum excitation energy is $\Delta$, not zero. This gapped spectrum is what suppresses low-temperature dissipation and gives superconductors their distinctive thermodynamic behavior.

Density of states

The quasiparticle density of states in a BCS superconductor is:

$N_s(E) = N(0) \frac{|E|}{\sqrt{E^2 - \Delta^2}} \quad \text{for } |E| > \Delta$

and zero for $|E| < \Delta$. Two features stand out:

• A hard gap of width $2\Delta$ centered on the Fermi energy, where no states exist.
• Sharp coherence peaks at $E = \pm\Delta$, where the density of states diverges.

These features are directly visible in tunneling spectroscopy and serve as a fingerprint of BCS superconductivity.

Mathematical formalism

BCS wave function

The BCS ground state is written as:

$|\Psi_{\text{BCS}}\rangle = \prod_k (u_k + v_k \, c_{k\uparrow}^\dagger c_{-k\downarrow}^\dagger) |0\rangle$

Here $c_{k\uparrow}^\dagger$ creates an electron with momentum $k$ and spin up, while $u_k$ and $v_k$ are variational parameters satisfying $|u_k|^2 + |v_k|^2 = 1$. The coefficient $|v_k|^2$ gives the probability that the pair state $(k\uparrow, -k\downarrow)$ is occupied. These parameters are determined by minimizing the total energy of the system.

Mean-field approximation

The full many-body problem of interacting electrons is intractable, so BCS theory uses a mean-field decoupling. The idea is to replace the four-fermion interaction term with an effective pairing field proportional to the order parameter $\Delta$:

$\Delta = -V \sum_k \langle c_{-k\downarrow} c_{k\uparrow} \rangle$

where $V$ is the pairing interaction strength. This reduces the Hamiltonian to a quadratic (solvable) form. The approximation becomes exact in the thermodynamic limit for weak-coupling superconductors.

Gap equation

The self-consistency condition on $\Delta$ yields the BCS gap equation:

$1 = V \sum_k \frac{1}{2E_k} \tanh\left(\frac{E_k}{2k_B T}\right)$

This equation determines $\Delta(T)$ implicitly. At $T = 0$, it gives the zero-temperature gap $\Delta(0)$. As $T$ increases, $\Delta(T)$ decreases and vanishes continuously at $T = T_c$, signaling a second-order phase transition. Near $T_c$, the gap vanishes as $\Delta(T) \propto \sqrt{1 - T/T_c}$.

Predictions of BCS theory

Critical temperature

BCS theory predicts that the critical temperature is related to the zero-temperature gap by a universal ratio:

$\frac{2\Delta(0)}{k_B T_c} \approx 3.53$

This ratio holds well for weak-coupling superconductors. $T_c$ itself depends on the Debye frequency $\omega_D$, the density of states at the Fermi level $N(0)$, and the coupling constant $V$:

$k_B T_c \approx 1.13 \, \hbar\omega_D \, \exp\left(-\frac{1}{N(0)V}\right)$

The exponential dependence explains why $T_c$ is so sensitive to material parameters. For conventional superconductors, $T_c$ typically ranges from below 1 K up to about 20 K.

Isotope effect

Because the pairing is phonon-mediated, $T_c$ depends on the ion mass $M$ through the Debye frequency ($\omega_D \propto M^{-1/2}$). This gives:

$T_c \propto M^{-\alpha}, \quad \alpha \approx 0.5$

Replacing lattice atoms with heavier isotopes lowers $T_c$, and this has been confirmed experimentally in mercury, lead, tin, and many other conventional superconductors. Deviations from $\alpha = 0.5$ occur in strong-coupling materials or when non-phonon mechanisms contribute.

Meissner effect

BCS theory accounts for the Meissner effect: the complete expulsion of magnetic flux from a superconductor's interior (for fields below $H_c$). Screening supercurrents flow in a thin surface layer of thickness $\lambda_L$ (the London penetration depth), canceling the applied field inside.

• The superfluid density, derived from the BCS ground state, determines $\lambda_L$.
• BCS theory connects smoothly to the London equations in the long-wavelength limit, providing a microscopic justification for those phenomenological results.

Experimental evidence

Tunneling experiments

Tunneling spectroscopy through superconductor-insulator-normal metal (SIN) or superconductor-insulator-superconductor (SIS) junctions directly maps out the quasiparticle density of states. These experiments reveal:

• The energy gap $\Delta$ as a region of suppressed conductance.
• Sharp coherence peaks at the gap edges, exactly as BCS predicts.
• The temperature dependence $\Delta(T)$, which follows the BCS curve closely in conventional superconductors.

Giaever's tunneling experiments in the early 1960s provided some of the most direct confirmations of BCS theory.

Specific heat measurements

The electronic specific heat of a superconductor behaves very differently from a normal metal:

• At $T_c$, there is a discontinuous jump in the specific heat (no latent heat, consistent with a second-order transition). BCS theory predicts $\Delta C / \gamma T_c \approx 1.43$.
• Below $T_c$, the specific heat drops exponentially as $\sim \exp(-\Delta / k_B T)$, reflecting the energy gap that suppresses thermal excitations.

Both features match BCS predictions quantitatively in conventional superconductors.

Electromagnetic absorption

Superconductors are transparent to photons with energy below $2\Delta$ because there are no available excitations to absorb them. Above the threshold $\hbar\omega = 2\Delta$, photons can break Cooper pairs, and absorption sets in. This sharp onset at $2\Delta$ has been observed in far-infrared and microwave experiments and provides another direct measure of the gap.

Limitations of BCS theory

High-temperature superconductors

Standard BCS theory cannot explain superconductivity in materials with $T_c$ well above 30 K. The cuprate superconductors (e.g., YBCO with $T_c \approx 93$ K) and iron-based superconductors exhibit strong electronic correlations and pairing mechanisms that go beyond simple phonon exchange. Proposed alternative theories include resonating valence bond (RVB) states and spin-fluctuation-mediated pairing, but a consensus microscopic theory for these materials remains elusive.

Unconventional superconductors

Some superconductors display order parameters with symmetries other than the isotropic s-wave assumed in BCS theory:

• d-wave pairing in cuprates, where the gap has nodes (zeros) along certain directions in momentum space.
• p-wave pairing in candidates like $\text{Sr}_2\text{RuO}_4$, which would carry nontrivial topological properties.

These materials often show power-law (rather than exponential) temperature dependences in thermodynamic quantities, signaling the presence of gap nodes. Describing them requires extensions of BCS theory that allow for anisotropic pairing interactions.

Beyond BCS theory

Several theoretical frameworks extend or go beyond the original BCS formulation:

• Eliashberg theory accounts for strong electron-phonon coupling and retardation effects, giving more accurate $T_c$ predictions for materials like lead and niobium.
• Spin-fluctuation models replace phonons with magnetic excitations as the pairing glue.
• Approaches incorporating quantum criticality and competing order parameters aim to capture the rich phase diagrams seen in unconventional superconductors.

Applications of BCS theory

Josephson junctions

A Josephson junction consists of two superconductors separated by a thin insulating barrier. Cooper pairs tunnel coherently across the barrier, producing effects that follow directly from the macroscopic phase coherence of the BCS ground state:

• DC Josephson effect: A supercurrent flows with zero applied voltage, with magnitude $I = I_c \sin(\Delta\phi)$, where $\Delta\phi$ is the phase difference across the junction.
• AC Josephson effect: A constant voltage $V$ across the junction produces an oscillating current at frequency $f = 2eV/h$.

Josephson junctions are used in ultra-precise voltage standards and form the basis of SQUIDs and superconducting qubits.

SQUID devices

A SQUID (Superconducting Quantum Interference Device) consists of a superconducting loop interrupted by one or two Josephson junctions. The critical current of the loop oscillates as a function of the magnetic flux threading it, with a period of one flux quantum $\Phi_0 = h/2e \approx 2.07 \times 10^{-15}$ Wb.

• SQUIDs achieve magnetic field sensitivities on the order of femtotesla ($10^{-15}$ T).
• Applications include magnetoencephalography (brain imaging), geophysical surveying, and materials characterization.

Superconducting qubits

Superconducting circuits containing Josephson junctions can behave as artificial quantum two-level systems (qubits). The Josephson junction acts as a nonlinear, dissipationless inductance, which is essential for creating discrete energy levels in a circuit that would otherwise have a continuous spectrum.

• Major qubit designs include the transmon, flux qubit, and charge qubit.
• These are currently among the leading platforms for scalable quantum computing, with coherence times that have improved by orders of magnitude over the past two decades.

BCS theory vs other models

Ginzburg-Landau theory

Ginzburg-Landau (GL) theory is a phenomenological framework that describes superconductivity near $T_c$ using a complex order parameter $\Psi$. It is built on Landau's general theory of second-order phase transitions and does not assume a specific microscopic mechanism.

• GL theory excels at describing spatially inhomogeneous situations: vortices, interfaces, and the distinction between Type I and Type II superconductors.
• Gor'kov showed in 1959 that GL theory can be derived from BCS theory as a limiting case near $T_c$, connecting the phenomenological order parameter to the microscopic gap function.

London theory

The London equations (1935) describe the electrodynamics of superconductors at a macroscopic level. They predict the Meissner effect and introduce the penetration depth $\lambda_L$, but they do not explain why superconductivity occurs.

• London theory can be derived from BCS theory in the long-wavelength, low-frequency limit.
• It also predicts flux quantization in superconducting rings, with the flux quantum $\Phi_0 = h/2e$ reflecting the charge $2e$ of Cooper pairs.

Microscopic vs macroscopic theories

BCS theory is a microscopic theory: it starts from electrons and their interactions and derives superconducting properties from first principles. Ginzburg-Landau and London theories are macroscopic (phenomenological): they describe observable electromagnetic and thermodynamic behavior without specifying the underlying mechanism.

The power of BCS theory is that it justifies and unifies the macroscopic theories. GL and London descriptions remain extremely useful in practice because they are simpler to work with for many problems involving spatial structure and electromagnetic response. The three frameworks are complementary, each most natural at its own length and energy scale.