Superconductivity, Josephson Junctions and Superfluidity
CSIR NET Part C: Meissner effect, London depth, type I and II superconductors, BCS gap, flux quantum, Josephson effects, SQUIDs and superfluid helium.
- 18 min
On this page · 16 sections
- Prerequisites
- The Hook: Zero Resistance Is Not the Whole Story
- Phenomenology of the Superconducting State
- London Theory
- Coherence Length, Type I and Type II
- BCS Theory in Brief
- Flux Quantisation
- Josephson Effects
- Superfluidity
- Memory Tricks
- Don’t Confuse
- Common Mistakes
- Practice
- Quick Revision Table
- Key Takeaways
- What’s Next
Prerequisites
- Band Theory: Metals, Insulators and Semiconductors: the Fermi surface and the idea of an energy gap.
- Classical and Quantum Statistics: Ideal Bose and Fermi Gases: Bose–Einstein condensation, the starting picture for superfluid helium.
- First- and Second-Order Phase Transitions: order parameters, Landau theory and the heat-capacity jump at a second-order transition.
- Maxwell’s Equations, Potentials and Gauge Invariance: the vector potential, which enters the London equation and flux quantisation.
The Hook: Zero Resistance Is Not the Whole Story
Phenomenology of the Superconducting State
| Property | Statement |
|---|---|
| Zero DC resistance | below \( T_c \), persistent currents decay immeasurably slowly |
| Meissner effect | \( \mathbf{B} = 0 \) in the bulk, whether cooled in field or not: perfect diamagnetism, \( \chi = -1 \) |
| Critical field | superconductivity destroyed above \( H_c(T) \approx H_c(0)[1 - (T/T_c)^2] \) |
| Critical current | a transport current whose own field reaches \( H_c \) at the surface destroys it (Silsbee rule) |
| Energy gap | heat capacity falls as \( e^{-\Delta/k_BT} \); photons absorbed only above \( 2\Delta \) |
| Heat capacity jump | finite discontinuity at \( T_c \) in zero field: a second-order transition; first order in a field |
| Isotope effect | \( T_c \propto M^{-\alpha} \), \( \alpha \approx 1/2 \) for many simple metals: phonons are involved |
| Flux quantisation | flux through a superconducting ring is \( n\Phi_0 \), \( \Phi_0 = h/2e \) |
Perfect conductor versus superconductor
A perfect conductor obeys \( \partial\mathbf{B}/\partial t = 0 \) inside (from \( \mathbf{E} = 0 \) and Faraday’s law): it keeps whatever field it had. Cool it in a field and switch the field off, and it holds the flux, with surface currents. A superconductor ends with \( \mathbf{B} = 0 \) inside in both orders of operation. That path independence is why the superconducting state is a true thermodynamic phase with a free energy.
Thermodynamics
Since the superconductor expels the field, magnetising it against \( H \) costs \( \mu_0H^2/2 \) per unit volume. At \( H_c \) this equals the free-energy gain of the superconducting state:
$$\boxed{g_n(T) - g_s(T) = \frac{\mu_0H_c^2(T)}{2}}$$the condensation energy. For \( B_c = \mu_0H_c = 0.08 \) T it is \( B_c^2/2\mu_0 \approx 2.5\times10^3 \) J m\( ^{-3} \). Differentiating with respect to \( T \) gives the entropy difference (zero at \( T_c \), so no latent heat in zero field) and the heat capacity jump. Since \( dH_c/dT < 0 \) everywhere, \( S_s < S_n \): the superconducting state is more ordered.
London Theory
Treat a density \( n_s \) of “superelectrons” that accelerate freely: \( m\,d\mathbf{v}_s/dt = -e\mathbf{E} \), with \( \mathbf{J}_s = -n_se\mathbf{v}_s \). This gives the first London equation \( \partial\mathbf{J}_s/\partial t = (n_se^2/m)\mathbf{E} \). London’s step is the stronger postulate
$$\nabla\times\mathbf{J}_s = -\frac{n_se^2}{m}\mathbf{B} \qquad \left(\text{or } \mathbf{J}_s = -\frac{n_se^2}{m}\mathbf{A} \text{ in the London gauge}\right)$$With Ampère’s law \( \nabla\times\mathbf{B} = \mu_0\mathbf{J}_s \):
$$\nabla^2\mathbf{B} = \frac{\mathbf{B}}{\lambda_L^2}, \qquad \boxed{\lambda_L = \sqrt{\frac{m}{\mu_0n_se^2}}}$$Outside field \( B_0 \) parallel to a flat surface decays as \( B(x) = B_0e^{-x/\lambda_L} \). For \( n_s = 10^{28} \) m\( ^{-3} \), \( \lambda_L \approx 53 \) nm; \( \lambda_L \propto n_s^{-1/2} \). Near \( T_c \), \( n_s \to 0 \) and \( \lambda \) diverges; an empirical form is \( \lambda(T) = \lambda(0)/\sqrt{1 - (T/T_c)^4} \).
Coherence Length, Type I and Type II
The coherence length \( \xi \) is the distance over which the superconducting order parameter can change. Ginzburg–Landau theory introduces a complex order parameter \( \psi(\mathbf{r}) \) (with \( |\psi|^2 \propto n_s \)) and the ratio
$$\kappa = \frac{\lambda}{\xi}$$- Type I (\( \kappa < 1/\sqrt2 \)): the normal–superconductor interface has positive surface energy. Complete Meissner effect up to \( H_c \), then abruptly normal. Mostly pure elemental metals (Hg, Pb, Sn, Al).
- Type II (\( \kappa > 1/\sqrt2 \)): negative interface energy, so the material lets field in as many thin tubes. Full Meissner state below \( H_{c1} \); mixed (vortex) state between \( H_{c1} \) and \( H_{c2} \); normal above \( H_{c2} \). Alloys, compounds (Nb–Ti, Nb\( _3 \)Sn) and the cuprates; elemental niobium is also type II.
| Field range | Type I | Type II |
|---|---|---|
| \( H < H_{c1} \) | Meissner | Meissner |
| \( H_{c1} < H < H_{c2} \) | (normal above \( H_c \)) | vortices, each carrying one \( \Phi_0 \) |
| \( H > H_{c2} \) | normal | normal |
Each vortex has a normal core of radius about \( \xi \) and circulating currents out to about \( \lambda \). Vortices repel and form a triangular (Abrikosov) lattice. The upper critical field is
$$\mu_0H_{c2} = \frac{\Phi_0}{2\pi\xi^2}$$so a short coherence length gives a huge \( H_{c2} \): \( \xi = 10 \) nm gives about 3.3 T, \( \xi = 2 \) nm about 82 T. Also \( H_{c1}H_{c2} \approx H_c^2\ln\kappa \), so \( H_{c1} < H_c < H_{c2} \). Pinning of vortices by defects lets type II wires carry large currents in high fields: the basis of superconducting magnets.
In a field \( B \), the vortex density is \( B/\Phi_0 \), about \( 4.8\times10^{13} \) m\( ^{-2} \) at 0.1 T, spaced roughly 0.14 μm apart.
BCS Theory in Brief
- Cooper pairs. Any attraction, however weak, between two electrons above a filled Fermi sea produces a bound pair. In metals the attraction comes from the lattice: one electron distorts the ions, and the distortion attracts another (exchange of virtual phonons), which is why the isotope effect exists.
- Pairs have opposite momenta and opposite spins in conventional superconductors (\( \mathbf{k}\uparrow, -\mathbf{k}\downarrow \)), total spin zero, charge \( 2e \). The ground state is a coherent condensate of overlapping pairs.
- Gap. Excitations cost at least \( \Delta \) each; breaking a pair costs \( 2\Delta \).
with \( N(0) \) the density of states per spin at \( E_F \) and \( V \) the pairing interaction. Near \( T_c \), \( \Delta(T) \approx 1.74\Delta(0)(1 - T/T_c)^{1/2} \) (mean-field exponent 1/2). The heat capacity jumps by \( \Delta C = 1.43\gamma T_c \).
- BCS coherence length \( \xi_0 = \hbar v_F/\pi\Delta(0) \): tens of nm to over a micrometre in clean elemental metals, only a few nm or less in cuprates.
- For \( T_c = 9.2 \) K, \( 2\Delta(0) \approx 2.8 \) meV, matching a photon of about 680 GHz (far infrared): a superconductor absorbs only above this frequency at \( T \ll T_c \).
Flux Quantisation
The condensate has a single macroscopic wavefunction \( \psi = |\psi|e^{i\theta} \). The supercurrent of particles of charge \( q = -2e \) and mass \( 2m \) is
$$\mathbf{J}_s \propto \hbar\nabla\theta - q\mathbf{A}$$Deep inside a thick ring \( \mathbf{J}_s = 0 \), so integrating round a loop: \( \oint\nabla\theta\cdot d\mathbf{l} = (q/\hbar)\oint\mathbf{A}\cdot d\mathbf{l} = q\Phi/\hbar \). Since \( \theta \) must be single valued, \( \oint\nabla\theta\cdot d\mathbf{l} = 2\pi n \):
$$\boxed{\Phi = n\Phi_0, \qquad \Phi_0 = \frac{h}{2e} = 2.07\times10^{-15}\text{ Wb}}$$The measured value \( h/2e \), not \( h/e \), was direct evidence that the carriers are pairs.
Josephson Effects
Two superconductors separated by a thin barrier (an insulator about 1–2 nm thick, a normal metal or a constriction) form a Josephson junction. Pairs tunnel coherently, and the junction is described by the phase difference \( \varphi = \theta_2 - \theta_1 \):
$$\boxed{I = I_c\sin\varphi \qquad \text{(DC Josephson)}}$$$$\boxed{\frac{d\varphi}{dt} = \frac{2eV}{\hbar} \qquad \text{(AC Josephson)}}$$| Effect | Condition | Observation |
|---|---|---|
| DC | \( V = 0 \) | a supercurrent up to \( I_c \) flows with no voltage |
| AC | DC voltage \( V \) | current oscillates at \( f = 2eV/h = 483.6 \) MHz per μV |
| Shapiro steps | microwaves at frequency \( f \) applied | constant-voltage steps in the I–V curve at \( V_n = nhf/2e \) |
At \( V = 10 \) μV the AC Josephson frequency is 4.84 GHz. Under 10 GHz irradiation, Shapiro steps are spaced by \( hf/2e = 20.7 \) μV. Because the relation involves only fundamental constants and a frequency, Josephson junctions are used as voltage standards.
For a tunnel junction at \( T = 0 \), the Ambegaokar–Baratoff result \( I_cR_n = \pi\Delta/2e \) links the critical current to the normal-state resistance.
SQUIDs
A DC SQUID is a superconducting loop containing two junctions in parallel. The flux \( \Phi \) through the loop shifts the phases of the two arms; the maximum supercurrent is
$$I_{\max}(\Phi) = 2I_c\left|\cos\frac{\pi\Phi}{\Phi_0}\right|$$for identical junctions and negligible loop inductance. \( I_{\max} \) oscillates with period \( \Phi_0 \) in flux: the two-slit interference pattern of the superconducting wave. A loop of area \( 10^{-8} \) m\( ^2 \) (100 μm square) changes by one period when \( B \) changes by \( \Phi_0/A \approx 0.21 \) μT, and electronics can resolve a tiny fraction of a period: SQUIDs are the most sensitive magnetometers.
Superfluidity
Helium-4
Liquid \( ^4 \)He, made of bosons, undergoes the lambda transition at about 2.17 K (at saturated vapour pressure): the heat capacity has a logarithmic, lambda-shaped peak. Below it, He II shows:
- flow through fine capillaries with no measurable viscosity;
- a film that creeps up and over container walls;
- the fountain (thermomechanical) effect: heating one side of a fine-pored plug drives superfluid towards the heat;
- second sound, a temperature (entropy) wave rather than a pressure wave;
- quantised circulation \( \oint\mathbf{v}_s\cdot d\mathbf{l} = nh/m_4 \approx n\times9.97\times10^{-8} \) m\( ^2 \) s\( ^{-1} \), and quantised vortices.
Link to BEC. An ideal Bose gas at the density of liquid helium (about 145 kg m\( ^{-3} \)) would condense at about 3.1 K, close to 2.17 K. Interactions are strong, so only a small fraction of atoms are in the zero-momentum state even at \( T = 0 \), yet the superfluid fraction reaches 1.
Two-fluid model
Below \( T_\lambda \) the liquid behaves as two interpenetrating fluids: a superfluid component (density \( \rho_s \), zero entropy, zero viscosity) and a normal component (\( \rho_n \), carries all the entropy and viscosity), with \( \rho = \rho_s + \rho_n \). \( \rho_s/\rho \) rises from 0 at \( T_\lambda \) to 1 at \( T = 0 \). Andronikashvili’s rotating-disc experiment measured \( \rho_n \).
Landau criterion
A body moving through the liquid at speed \( v \) can create an excitation of momentum \( p \) and energy \( \varepsilon(p) \) only if \( v \geq \varepsilon(p)/p \). So flow is dissipationless below
$$v_c = \min_p\frac{\varepsilon(p)}{p}$$A free-particle spectrum \( p^2/2m \) gives \( v_c = 0 \): an ideal Bose gas is not a superfluid by this criterion. The excitation spectrum of He II is linear (phonons, \( \varepsilon = cp \)) at small \( p \) and has a roton minimum \( \varepsilon \approx \Delta + (p - p_0)^2/2\mu \) at larger \( p \). The roton minimum fixes \( v_c \approx \Delta/p_0 \): with \( \Delta/k_B = 8.6 \) K and \( p_0/\hbar = 1.9\times10^{10} \) m\( ^{-1} \), about 60 m s\( ^{-1} \). Observed critical velocities in channels are usually far lower, because vortices are created first.
Helium-3
\( ^3 \)He atoms are fermions, so they must pair first, like electrons in BCS theory. They do, but only in the millikelvin range, and the pairs have spin 1 and orbital angular momentum 1 (p-wave), giving several superfluid phases with anisotropic properties.
Memory Tricks
- “483.6 MHz per microvolt.” AC Josephson: \( f = 2eV/h \).
- “Gap is 3.5 \( k_BT_c \).” \( 2\Delta(0) = 3.52k_BT_c \); jump \( \Delta C = 1.43\gamma T_c \).
- “One over root two” separates type I (\( \kappa \) below) from type II (above).
- “Flux comes in \( h/2e \), circulation in \( h/m \).” Pairs of charge \( 2e \); helium atoms of mass \( m_4 \).
- “Below \( H_{c1} \) none, between vortices, above \( H_{c2} \) all.”
Don’t Confuse
| This | vs | That |
|---|---|---|
| Superconductor: expels \( B \) whatever the history | Perfect conductor: freezes the \( B \) it had | |
| London depth \( \lambda \): field decay length | Coherence length \( \xi \): order-parameter healing length | |
| Type I: \( \kappa < 1/\sqrt2 \), single \( H_c \) | Type II: \( \kappa > 1/\sqrt2 \), \( H_{c1} < H_{c2} \) | |
| DC Josephson: current at zero voltage | AC Josephson: oscillating current at fixed voltage | |
| Flux quantum \( h/2e \) | Quantum of circulation \( h/m_4 \) (He-4) | |
| Zero-field transition: second order | Transition in a field: first order, latent heat | |
| He-4 (bosons): superfluid at about 2.17 K | He-3 (fermions): superfluid only at mK, via p-wave pairing |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Use \( h = 6.626\times10^{-34} \) J s, \( e = 1.602\times10^{-19} \) C, \( k_B = 1.381\times10^{-23} \) J K\( ^{-1} \), \( m_e = 9.11\times10^{-31} \) kg, \( \mu_0 = 4\pi\times10^{-7} \) H m\( ^{-1} \).
Level 1: Foundation
- What is the magnetic susceptibility of a bulk superconductor in the Meissner state?
- What is the value of \( \kappa \) separating type I and type II superconductors?
- What is the frequency of the AC Josephson current at a DC bias of 1 μV?
Answers
- \( \chi = -1 \) (perfect diamagnet).
- \( \kappa = 1/\sqrt2 \approx 0.71 \).
- 483.6 MHz (\( 2eV/h \)).
Level 2: Part B Style
Practice Q1. A type I superconductor has \( T_c = 7.2 \) K and \( B_c(0) = 0.080 \) T. The critical field at 3.6 K is closest to:
(a) 0.020 T (b) 0.040 T (c) 0.060 T (d) 0.080 T
Answer: (c) 0.060 T
\( B_c = 0.080[1 - (0.5)^2] = 0.060 \) T. (b) uses a linear dependence on \( T \).
Practice Q2. The London penetration depth for a superelectron density of \( 1.0\times10^{28} \) m\( ^{-3} \) (charge \( e \), mass \( m_e \)) is closest to:
(a) 5.3 nm (b) 17 nm (c) 53 nm (d) 530 nm
Answer: (c) 53 nm
\( \lambda_L = \sqrt{m_e/(\mu_0n_se^2)} = \sqrt{9.11\times10^{-31}/(4\pi\times10^{-7}\times10^{28}\times2.566\times10^{-38})} = 5.31\times10^{-8} \) m. Using pairs (\( n_s/2 \), charge \( 2e \), mass \( 2m_e \)) gives the same \( \lambda \), since \( m/(nq^2) \) is unchanged.
Practice Q3. A Josephson junction is irradiated with 10 GHz microwaves. The spacing of the Shapiro steps on the voltage axis is closest to:
(a) 10.3 μV (b) 20.7 μV (c) 41.4 μV (d) 4.84 μV
Answer: (b) 20.7 μV
\( \Delta V = hf/2e = 6.626\times10^{-34}\times10^{10}/(2\times1.602\times10^{-19}) = 2.07\times10^{-5} \) V. (c) uses \( h/e \).
Practice Q4. In BCS theory, the minimum photon frequency absorbed at \( T \ll T_c \) by a superconductor with \( T_c = 9.2 \) K is closest to:
(a) 34 GHz (b) 190 GHz (c) 340 GHz (d) 680 GHz
Answer: (d) 680 GHz
\( 2\Delta = 3.52k_BT_c = 3.52\times1.381\times10^{-23}\times9.2 = 4.47\times10^{-22} \) J; \( f = 2\Delta/h = 6.8\times10^{11} \) Hz. (c) uses \( \Delta \) instead of \( 2\Delta \).
Level 3: Part C Style
Practice Q5. A superconductor has penetration depth \( \lambda = 50 \) nm and coherence length \( \xi = 10 \) nm. Consider:
- It is a type II superconductor.
- Its upper critical field \( \mu_0H_{c2} \) is about 3.3 T.
- In an applied field between \( H_{c1} \) and \( H_{c2} \), each vortex carries flux \( h/e \).
- At \( B = 0.1 \) T inside the sample, the vortex density is about \( 5\times10^{13} \) m\( ^{-2} \).
Which are correct?
(a) 1 and 2 only (b) 1, 2 and 4 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 4 only
- \( \kappa = 5 > 0.71 \). True.
- \( \Phi_0/(2\pi\xi^2) = 2.07\times10^{-15}/(2\pi\times10^{-16}) = 3.29 \) T. True.
- Each vortex carries \( \Phi_0 = h/2e \). False.
- \( B/\Phi_0 = 0.1/2.07\times10^{-15} = 4.8\times10^{13} \) m\( ^{-2} \). True.
Practice Q6. A DC SQUID with two identical junctions (critical current \( I_c \) each, negligible loop inductance) encloses an area of \( 1.0\times10^{-8} \) m\( ^2 \). Consider:
- The maximum supercurrent at zero flux is \( 2I_c \).
- The maximum supercurrent vanishes when the enclosed flux is \( \Phi_0/2 \).
- The pattern repeats each time the field changes by about 0.21 μT.
- If the carriers were single electrons, the period in field would be about 0.10 μT.
Which are correct?
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2 and 4 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 3 only
- \( 2I_c|\cos0| = 2I_c \). True.
- \( \cos(\pi/2) = 0 \). True.
- \( \Delta B = \Phi_0/A = 2.07\times10^{-15}/10^{-8} = 2.07\times10^{-7} \) T. True.
- With charge \( e \) the flux period would be \( h/e = 2\Phi_0 \), so the field period doubles to about 0.41 μT, not halves. False.
Practice Q7. Liquid helium-4 has density about 145 kg m\( ^{-3} \). Consider:
- An ideal Bose gas of \( ^4 \)He atoms at this density would condense at about 3.1 K.
- In the two-fluid model, the normal component carries all the entropy.
- For an excitation spectrum \( \varepsilon = p^2/2m \), the Landau critical velocity is zero.
- Taking a roton minimum at \( \Delta/k_B = 8.6 \) K and \( p_0/\hbar = 1.9\times10^{10} \) m\( ^{-1} \), the Landau critical velocity is about 600 m s\( ^{-1} \).
Which are correct?
(a) 1, 2 and 3 only (b) 2 and 3 only (c) 1 and 4 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 3 only
- \( n = 145/(6.646\times10^{-27}) = 2.18\times10^{28} \) m\( ^{-3} \); \( T_c = \dfrac{2\pi\hbar^2}{mk_B}\left(\dfrac{n}{2.612}\right)^{2/3} = 3.1 \) K. True.
- The superfluid component has zero entropy. True.
- \( \min(p/2m) = 0 \) as \( p \to 0 \). True.
- \( v_c = \Delta/p_0 = (8.6\times1.381\times10^{-23})/(1.9\times10^{10}\times1.055\times10^{-34}) = 59 \) m s\( ^{-1} \), not 600. False.
Quick Revision Table
| Idea | Formula or fact |
|---|---|
| Critical field | \( H_c(T) = H_c(0)[1 - (T/T_c)^2] \) |
| Condensation energy | \( \mu_0H_c^2/2 \) per volume |
| London depth | \( \lambda_L = \sqrt{m/\mu_0n_se^2} \); \( B = B_0e^{-x/\lambda} \) |
| Type I / II | \( \kappa = \lambda/\xi \lessgtr 1/\sqrt2 \) |
| Upper critical field | \( \mu_0H_{c2} = \Phi_0/2\pi\xi^2 \) |
| BCS | \( 2\Delta(0) = 3.52k_BT_c \); \( \Delta C = 1.43\gamma T_c \); \( \Delta \propto (1 - T/T_c)^{1/2} \) |
| Isotope effect | \( T_c \propto M^{-1/2} \) (simple metals) |
| Flux quantum | \( h/2e = 2.07\times10^{-15} \) Wb |
| Josephson | \( I = I_c\sin\varphi \); \( f = 2eV/h = 483.6 \) MHz μV\( ^{-1} \); steps \( nhf/2e \) |
| SQUID | \( I_{\max} = 2I_c\vert \cos(\pi\Phi/\Phi_0)\vert \) |
| He II | \( T_\lambda \approx 2.17 \) K; two fluids; circulation \( h/m_4 \); \( v_c = \min\varepsilon/p \) |
Key Takeaways
What’s Next
- Defects, Dislocations, Liquid Crystals and Quasicrystals: imperfect crystals, ordered liquids and order without periodicity.
- First- and Second-Order Phase Transitions, for Landau theory of the superconducting and lambda transitions.
- Condensed Matter Physics chapter overview and Quick Revision.