Modern Physics · Advanced Topics

Superconductivity & BCS Theory

Superconductors carry current with zero resistance and expel magnetic fields (Meissner effect). BCS theory explains conventional superconductivity through phonon-mediated Cooper pairing, producing a finite-temperature phase transition into a macroscopic coherent state.

PrerequisitesSolid-state physics (Ch. SS) · Quantum mechanics (Ch. 22) · Statistical mechanics (Ch. S) · Second quantization
Learning Goals
  • Derive the London penetration depth λL\lambda_{L} from the London equations and explain the Meissner effect.
  • State the BCS gap equation and compute Δ(0)\Delta(0) and TcT_{c} for a given coupling constant N(0)V.
  • Apply the DC and AC Josephson effects to calculate junction current and frequency-to-voltage ratio.
  • Classify Type I and Type II superconductors using the Ginzburg-Landau parameter κ=λ/ξ\kappa = \lambda/\xi.
  • Describe the Abrikosov vortex lattice and derive the upper critical field Hc2H_{c2} in GL theory.

SC.1 Phenomenology

Perfect conductivity: resistivity ρ → 0 below TcT_c. Record high-TcT_c: HgBa₂Ca₂Cu₃O₈ at 138 K (at ambient pressure), HₓS at 203 K (at 150 GPa), LaH₁₀ at 250 K (under pressure). Room-temperature superconductivity remains elusive.

Meissner effect: below TcT_c, B = 0 inside a superconductor regardless of whether it was cooled in a field or not. Not merely perfect diamagnetism — it is an active expulsion of flux. The London penetration depth λL\lambda_L characterizes the exponential decay of B from the surface: B(x)=B0e−x/λLB(x) = B_0 e^{-x/\lambda_L}.

Definition SC.1 — London Equations
F. and H. London (1935) proposed: ∂Js/∂t=(nse2/m)E\partial J_{s}/\partial t = (n_{s} e^{2}/m) E and Js=−(nse2/m)AJ_{s} = -(n_{s} e^{2}/m) A (London gauge ∇⋅A=0\nabla\cdot A = 0). The second equation gives the Meissner effect: ∇2B=B/λL2\nabla^{2}B = B/\lambda_{L}^{2} where λL2=m/(μ0nse2)\lambda_{L}^{2} = m/(\mu_{0} n_{s} e^{2}). Typical values: λL≈50nm(Al),500nm(YBCO)\lambda_{L} \approx 50 nm (Al), 500 nm (YBCO).

Type I vs Type II: Type I (Hg, Al, Pb) have a single critical field HcH_c. Type II (Nb, YBCO, MgB₂) have Hc1H_{c1} and Hc2H_{c2} — between them, magnetic flux enters in quantized vortices (Abrikosov lattice). Flux quantum: Φ₀ = h/(2e) ≈ 2.07×10⁻¹⁵ Wb. The factor of 2e is direct evidence for Cooper pairs.

SC.2 Cooper Pairs and the BCS Ground State

L. Cooper (1956): two electrons near the Fermi surface, interacting via an attractive potential V (mediated by phonons), form a bound state (Cooper pair) for any V > 0 — no matter how weak. This is special to 2D Fermi surface; no bound state in 3D free space for weak attraction.

Ebinding=−2ℏωD e−2/N(0)V(Cooper pair binding energy, N(0)=density of states at EF)E_{\text{binding}} = -2\hbar\omega_D\, e^{-2/N(0)V} \qquad \text{(Cooper pair binding energy, } N(0) = \text{density of states at } E_F\text{)}(SC.1)

The BCS ground state (Bardeen-Cooper-Schrieffer, 1957):

∣BCS⟩=∏k(uk+vk ck↑†c−k↓†)∣0⟩(BCS wavefunction)|\text{BCS}\rangle = \prod_k \left(u_k + v_k\, c^\dagger_{k\uparrow} c^\dagger_{-k\downarrow}\right) |0\rangle \qquad \text{(BCS wavefunction)}(SC.2)

where ∣uk∣2+∣vk∣2=1|u_k|^2 + |v_k|^2 = 1, and vk2=12(1−ξk/Ek)v_k^2 = \tfrac{1}{2}(1 - \xi_k/E_k) with Ek=ξk2+Δ2E_k = \sqrt{\xi_k^2 + \Delta^2}. The order parameter (gap) Δ=V∑k⟨c−k↓ck↑⟩\Delta = V \sum_k \langle c_{-k\downarrow} c_{k\uparrow}\rangle satisfies the BCS gap equation.

Theorem SC.1 — BCS Gap Equation
The superconducting gap Δ satisfies:1N(0)V=∫0ℏωDdξξ2+Δ2\dfrac{1}{N(0)V} = \int_0^{\hbar\omega_D} \frac{d\xi}{\sqrt{\xi^2 + \Delta^2}}→ Δ(0)≈2ℏωD e−1/(N(0)V)\Delta(0) \approx 2\hbar\omega_D\, e^{-1/(N(0)V)} (weak coupling). Temperature dependence: Δ(T)→0\Delta(T) \to 0 as T→TcT \to T_{c} wherekBTc=1.13 ℏωD e−1/(N(0)V)=Δ(0)/1.76k_B T_c = 1.13\, \hbar\omega_D\, e^{-1/(N(0)V)} = \Delta(0)/1.76. The ratio 2Δ(0)/(kBTc)=3.522\Delta(0)/(k_{B} T_{c}) = 3.52 is a universal BCS prediction.
Example SC.1 — Aluminum as a BCS Superconductor

AlhasTc=1.2KAl has T_{c} = 1.2 K, Debye temperature ΘD=428K(ℏωD=kB×428K)\Theta_{D} = 428 K (\hbar\omega_{D} = k_{B} \times 428 K). Estimate Δ(0),N(0)V\Delta(0), N(0)V, and the Fermi velocity from the coherence length.

Gap from Tc:Δ(0)=1.76kBTc=1.76×(8.617×10−5eV/K)(1.2K)=1.76×1.03×10−4eV=1.82×10−4eV\Delta(0) = 1.76 k_{B} T_{c} = 1.76 \times (8.617\times 10^{-5} eV/K)(1.2 K) = 1.76 \times 1.03\times 10^{-4} eV = 1.82\times 10^{-4} eV = 0.18 meV. This matches tunneling spectroscopy measurements on Al junctions.
Coupling constant N(0)V:From Tc=1.13ΘDexp⁡(−1/(N(0)V)):N(0)V=1/ln⁡(1.13ΘD/Tc)=1/ln⁡(1.13×428/1.2)=1/ln⁡T_{c} = 1.13 \Theta_{D} \exp(-1/(N(0)V)): N(0)V = 1/\ln(1.13 \Theta_{D}/T_{c}) = 1/\ln(1.13 \times 428/1.2) = 1/\ln(402.8) = 1/5.999 ≈ 0.167. This is weak coupling (N(0)V ≪ 1) — BCS applies well.
Coherence length:ξ0=ℏvF/(πΔ)—\xi_{0} = \hbar v_{F}/(\pi\Delta) — the characteristic size of a Cooper pair. ForAl:vF=2.03×106m/s.ξ0=(1.055×10−34×2.03×106)/(π×1.82×10−4×1.6×10−19)=2For Al: v_{F} = 2.03\times 10^{6} m/s. \xi_{0} = (1.055\times 10^{-34} \times 2.03\times 10^{6})/(\pi \times 1.82\times 10^{-4} \times 1.6\times 10^{-19}) = 2.14×10−28/9.15×10−23≈2.3×10−6m=2300nm14\times 10^{-28}/9.15\times 10^{-23} \approx 2.3\times 10^{-6} m = 2300 nm. This is much larger than the lattice spacing — Cooper pairs are highly overlapping, not tightly bound molecules.
GL ratio:λL/ξ0≈50nm/2300nm≈0.022≪1/2≈0.707\lambda_{L}/\xi_{0} \approx 50nm/2300nm \approx 0.022 ≪ 1/\sqrt 2 \approx 0.707. Al is a Type I superconductor (Ginzburg-Landau parameter κ=λ/ξ<1/2\kappa = \lambda/\xi < 1/\sqrt 2). Forκ>1/2For \kappa > 1/\sqrt 2 (Type II): Abrikosov vortex lattice forms above Hc1H_{c1}.

SC.3 Josephson Effect

Two superconductors separated by a thin insulator (tunnel junction). The macroscopic wavefunction is Ψ=ns eiθ\Psi = \sqrt{n_s}\, e^{i\theta}. The phase difference δ = θ₁ − θ₂ drives:

I=Icsin⁡δ(DC Josephson effect)dδdt=2eVℏ(AC Josephson effect)I = I_c \sin\delta \quad \text{(DC Josephson effect)} \qquad \frac{d\delta}{dt} = \frac{2eV}{\hbar} \quad \text{(AC Josephson effect)}(SC.3)

DC Josephson: supercurrent flows with no voltage (V = 0). Critical current Ic∝ΔI_c \propto \Delta.AC Josephson: with DC voltage V, the phase oscillates at 2eV/ℏ — microwave radiation at f = 2eV/h ≈ 484 GHz/mV. The Josephson relation defines the voltage-frequency ratio 2e/h — used in metrology (voltage standard).

SQUID (Superconducting Quantum Interference Device): two junctions in parallel. Total current I=2Iccos⁡(πΦ/Φ0) sin⁡δavgI = 2I_c \cos(\pi\Phi/\Phi_0)\, \sin\delta_{\text{avg}} — periodic in the enclosed flux Φ. Sensitivity: δΦ ~ 10⁻⁶ Φ₀ — the most sensitive magnetometer ever built. Applications: brain imaging (MEG), gravitational wave detection, dark matter searches.

SC.4 Ginzburg-Landau Theory

Near TcT_c, expand the free energy in the order parameter Ψ (complex):

F=Fn+α∣Ψ∣2+β2∣Ψ∣4+12m∗∣(−iℏ∇−2eA)Ψ∣2+B22μ0F = F_n + \alpha|\Psi|^2 + \frac{\beta}{2}|\Psi|^4 + \frac{1}{2m^*}\left|(-i\hbar\nabla - 2e\mathbf{A})\Psi\right|^2 + \frac{B^2}{2\mu_0}(SC.4)

where α=α0(T−Tc)\alpha = \alpha_0(T - T_c) changes sign at TcT_c. Minimizing δF/δΨ∗=0\delta F/\delta\Psi^* = 0 gives the Ginzburg-Landau (GL) equation. Two characteristic lengths emerge: penetration depth λ(T)=λL/1−T/Tc\lambda(T) = \lambda_L/\sqrt{1 - T/T_c} and coherence length ξ(T)=ξ0/1−T/Tc\xi(T) = \xi_0/\sqrt{1 - T/T_c}. GL parameter κ = λ/ξ: κ < 1/√2 → Type I; κ > 1/√2 → Type II.

Vortex structure (Abrikosov, 1957): in Type II, each vortex carries one flux quantum Φ₀ = h/(2e). Order parameter |Ψ| → 0 at vortex core (radius ξ). Magnetic field decays over λ from the core. The upper critical fieldHc2=Φ0/(2πξ2)H_{c2} = \Phi_0/(2\pi\xi^2) — at Hc2H_{c2} vortex cores overlap and superconductivity is destroyed. For YBCO: Hc2∼100H_{c2} \sim 100 T, enabling high-field magnet applications (MRI, LHC).

Definition SC.2 — Common Traps
  • Zero resistance is not the whole story: the Meissner effect distinguishes superconductors from perfect conductors.
  • Cooper pairs are correlated states: they are not tiny molecules moving through the lattice.
  • Type I and type II behave differently in fields: vortices appear only in type II superconductors.
  • Critical current and field matter: exceeding either destroys superconductivity.
Exercises — SC.1–SC.4 Superconductivity
1.
Derive the London penetration depth λL\lambda_{L} from the London equations. Estimate λL\lambda_{L} for niobium (density ns≈5×1028m−3n_{s} \approx 5\times 10^{28} m^{-3}).
nm
Straightforward
2.
Derive the BCS gap equation at T=0T = 0 and show Δ(0)=2ℏωDexp⁡(−1/(N(0)V))\Delta(0) = 2\hbar\omega_{D} \exp(-1/(N(0)V)). Find the ratio 2Δ(0)/(kBTc)=3.522\Delta(0)/(k_{B} T_{c}) = 3.52. Apply to lead (Pb,Tc=7.2K,ΘD=96KPb, T_{c} = 7.2 K, \Theta_{D} = 96 K).
meV
Intermediate
3.Explain the AC Josephson effect. Show that the frequency-to-voltage ratio is 2e/h=483.6GHz/mV2e/h = 483.6 GHz/mV. How is this used to define the volt in the modern SI?
Intermediate
4.Derive the upper critical field Hc2H_{c2} in GL theory and the Abrikosov vortex lattice spacing. Estimate Hc2H_{c2} for YBCO(ξ≈1.5nm,λ≈150nm)YBCO (\xi \approx 1.5 nm, \lambda \approx 150 nm) and explain why vortex pinning matters for applications.
Challenging
Key Takeaways
  • London equations: Js=−(nse2/m)AJ_{s} = -(n_{s} e^{2}/m)A → Meissner effect, field decays over λL=m/(μ0nse2)\lambda_{L} = \sqrt{m/(\mu_{0}n_{s} e^{2})}.
  • Cooper pairs: phonon attraction binds electrons near EF;anyV>0E_{F}; any V>0 gives a bound state (special to Fermi surface).
  • BCSgap:Δ(0)=2ℏωDexp⁡(−1/(N(0)V))BCS gap: \Delta(0) = 2\hbar\omega_{D} \exp(-1/(N(0)V)), ratio 2Δ(0)/(kBTc)=3.522\Delta(0)/(k_{B} T_{c}) = 3.52 universal.
  • DC Josephson: I=Icsin⁡(δ)I = I_{c} \sin(\delta), supercurrent with no voltage. AC Josephson: f=2eV/hf = 2eV/h (voltage standard).
  • Type II: flux enters as Abrikosov vortices above Hc1H_{c1}. Each carries Φ0=h/(2e)\Phi_{0} = h/(2e). Hc2=Φ0/(2πξ2)H_{c2} = \Phi_{0}/(2\pi\xi^{2}).
  • GL theory: two lengths λ(T)\lambda(T) and ξ(T)\xi(T), parameter κ=λ/ξ.κ<1/2\kappa = \lambda/\xi. \kappa < 1/\sqrt 2: Type I;κ>1/2I; \kappa > 1/\sqrt 2: Type II.