1 Introduction

Superconductivity = vast research area, here:

• mainly electronic transport phenomena

• focus on current research topics

Literature

• Tinkham, Introduction to Superconductivity

Good introduction to superconductivity, many aspects of nanostructures (was
Tinkham’s research field)
• current research topics often not covered in textbooks → review articles cited
in the script

For background on some topics:

• Parks, Superconductivity (2 volumes)

Detailed review articles for some topics
• De Gennes, Superconductivity of Metals and Alloys
Good description of Bogoliubov-de Gennes theory

But first, a brief phenomenological description of superconductivity:

Electric resistance. Superconductivity was discovered in 1911 by Heike Kamer-

lingh Onnes.1 Superconductors have zero resistance (within the experimental accu-
racy), i.e.,, they are perfect conductors, hence the name (see slide 1). Experimental
test: resistance measurements or persistent currents (more accurate, see slide 2).

Critical temperature. Superconductivity appears in many metals (or compounds)

below a critical temperature 𝑇c . Typical values (see slide 3):

material class discovered 𝑇c

elementary metals 1911 1 − 10 K
cuprate high-temperature superconductors 19862 up to ∼ 140 K
superhydrides 20153 > 200 K at ∼ 1 Mbar

1
H. Kamerlingh Onnes, Commun. Phys. Lab. Univ. Leiden 120b (1911), Nobel prize 1913
2
J. G. Bednorz and K. A. Müller, Z. Phys. B 64, 189 (1986), Nobelpreis 1987
3
A. P. Drozdov et al., Nature 525, 73 (2015), B. Lilia et al., J. Phys.: Condens. Matter 34, 183002
(2022)

1
As every phase transition, superconductivity implies an energy gain over the normal
phase, i.e., the free energy in the superconducting (S) phase is lower than in the
normal (N) phase
𝑓S (𝑇 ) < 𝑓N (𝑇 ) for 𝑇 < 𝑇c . (1.1)
At zero magnetic field, the phase transition is second order.

Meissner effect and critical field. Superconductors completely expel magnetic

fields from their interior (perfect diamagnetism, see slide 4).4 When a magnetic
field is applied below 𝑇c , one might expect screening currents generated by induc-
tion, which would then flow forever due to perfect conductivity. A superconductor,
however, generates screening currents even when cooled below 𝑇c in a static field,
where no induction occurs. Therefore, perfect diamagnetism is more than just per-
fect conductivity. The energy required to expel the magnetic field comes from the
energy gain of forming the superconducting state. Therefore, superconductivity is
destroyed when the energy gain is expended. This argument leads to the definition
of the thermodynamic critical field 𝐵c,th (see slide 5)
2 (𝑇 )
𝐵c,th
= [𝑓N (𝑇 ) − 𝑓S (𝑇 )]𝐵=0 . (1.2)
2𝜇0
Typical values: 𝐵c,th ∼ 10 − 100 mT for metals. At nonzero field, the phase tran-
sition is first order (jump in magnetization). A simple empirical approximation of
the temperature dependence is given by
2
𝐵c,th (𝑇 ) 𝑇
≈1−( ) . (1.3)
𝐵c,th (0) 𝑇c
Eq. (1.2) is the critical field for a homogeneous macroscopic superconductor in the
absence of demagnetization effects. The experimentally observed critical field can
deviate considerably depending on sample geometry and material properties. Also,
the field penetrates into a thin surface layer, see London theory below.

Critical current. A current-carrying wire creates a magnetic field (Ørsted field,

see slide 6). When the Ørsted field at the surface of a superconductor reaches the
critical field, the supercurrent can not increase any further (Silsbee criterion). The
critical field is therefore connected to the critical current 𝐼c (or, more accurately, the
critical current density 𝑗c ). One can also make the reverse argument: the maximum
screening current density at the critical field is the critical current density.

London theory (1935). The London theory5 is a simple phenomenological model

to describe the properties of superconductors (essentially the material equations for
the electrodynamics of superconductors, replacing Ohm’s law for normal metals).
𝜕𝒋
1. 𝛬 = 𝑬 (undamped acceleration)
𝜕𝑡 (1.4)
2. 𝛬∇ × 𝒋 = −𝑩 (Meissner effect)

4
W. Meissner and R. Ochsenfeld, Naturwissenschaften 21, 787 (1933)
5
F. London and H. London, Proc. Royal Soc. A 149, 71 (1935)

2
with the material constant 𝛬 = 𝑚/𝑛S 𝑒2 . Here, 𝑛S is the density of superconducting
charge carriers (also called the superfluid density). Inserting the second London
equation into the Maxwell equation ∇ × 𝑩 = 𝜇0 𝒋 yields

1
∇2 𝑩 = 𝑩 (1.5)
𝜆2L

with the London penetration depth

𝑚
𝜆L = √ . (1.6)
𝜇0 𝑛S 𝑒2

The solutions of (1.5) have the form

𝑩 = 𝑩0 𝑒−|𝒓|/𝜆L , (1.7)

i.e., the field penetrates into the superconductor over the length scale 𝜆L at the
surface. Re-inserting eq. (1.7) into the Maxwell equation shows that supercurrents
are also limited to the surface, i.e., the interior of superconductors is free of fields
and currents. Typical values: 𝜆L ∼ 20 − 50 nm.

Flux quantization and macroscopic wave function. The magnetic flux inside a
hollow superconducting cylinder is quantized in units of the flux quantum 𝛷0 =
ℎ/2𝑒 (see slide 7). The reason for this is the presence of a macroscopic wave func-
tion (combined with gauge invariance).6 The charge 2𝑒 in the flux quantum indi-
cates that the charge carriers in superconductors are pairs of electrons.

Ginzburg-Landau theory (1950). The Ginzburg-Landau theory7 (short GL the-

ory) is a phenomenological theory that combines Landau’s general phenomeno-
logical theory of phase transitions with the idea of a macroscopic wave function
𝜓 = ∣𝜓∣ exp(𝑖𝜃). GL theory obeys the laws of thermodynamics and gauge in-
variance. The density of superconducting charge carriers is given by 𝑛S = ∣𝜓∣2 ,
while the supercurrents are related to the gradient of the macroscopic phase 𝜃. GL
theory predicts the same penetration depth 𝜆 as London theory. Due to gauge in-
variance, GL theory also implies flux quantization. Unlike London theory, where
𝑛S is constant, GL theory can describe spatial variations of 𝑛S . Spatial variations
imply an energy cost, and therefore GL theory introduces a second length scale, the
Ginzburg-Landau coherence length 𝜉GL , the smallest scale on which 𝑛S can vary.
In GL-theory, both 𝜆 and 𝜉GL are temperature-dependent (see slide 8). Their ratio
𝜅 = 𝜆/𝜉GL (GL parameter) is constant.

6
The existence of a macroscopic wave function (and flux quantization) was first postulated by Fritz
London (however, with the charge 𝑒 rather than 2𝑒), see F. London, Phys. Rev. 74, 562 (1948).
7
V. L. Ginzburg and L. D. Landau, Zh. Eksp. Teor. Fiz. 20, 1064 (1950), english translation
L. D. Landau, Collected Papers of L. D. Landau, Pergamon Press, Oxford (1965). Nobel prize
1962/2003

3
Type II superconductors. Some superconductors show perfect Meissner screen-
ing only up to a lower critical field 𝐵c1 . Between 𝐵c1 and the upper critical field 𝐵c2 ,
the magnetic field penetrates the superconductor in the form of flux lines. Each flux
line carries one flux quantum.8 The reason is that the interface energy between the
normal and superconducting phase becomes negative for 𝜅 > 1/√2, and therefore
the superconductor forms as much interface area as possible, which is only limited
by flux quantization. 𝐵c2 can be much larger than 𝐵c,th , and therefore all technical
superconductors are type II (NbTi: 𝑇c ∼ 10 K, 𝐵c2 ≳ 13 T, Nb3 Sn: 𝑇c ∼ 18 K,
𝐵c2 ≳ 20 T, more recently also high temperature superconductors). Problem: when
a transport current is applied, flux lines move sideways due to the Lorentz force on
them, which in turn leads to an induction voltage and dissipation. Solution: pinning
centers.

Cooper pairs, BCS theory. The BCS theory (Bardeen-Cooper-Schrieffer)9 gives

the microscopy explanation of superconductivity. The conduction electrons are
bound to Cooper pairs (usually singlets with opposite momentum, see slide 11).
In the classical low-temperature superconductors, the pairing interaction is the re-
tarded electron-phonon interaction. Cooper pairs occupy a macroscopic wave func-
tion. BCS theory defines the BCS coherence length 𝜉0 , which can be interpreted
as the size of a Cooper pair. The order of magnitude can be estimated from the un-
certainty relation, if ones assumes that mostly electrons in the energy range ±𝑘B 𝑇c
around the Fermi energy contribute to superconductivity: 𝜉0 ∼ ℏ𝑣F /𝑘B 𝑇c . Typical
values: 𝜉0 ∼ 10 − 1000 nm. GL theory can be derived as a limiting case of BCS
theory10 , and for 𝑇 = 0 one finds 𝜉GL ≈ 0.74𝜉0 .

Quasiparticles, energy gap. Cooper pairs have a pairing energy. When they are
broken, fermionic single particle excitations are created (quasiparticles, see slide
12). The minimum excitation energy is the energy gap 𝛥. The minimal energy to
break a Cooper pair is 2𝛥 (creating two quasiparticles). Therefore, sometimes 2𝛥 is
referred to as energy gap. Typical values: 𝛥 ∼ 𝑘B 𝑇c ∼ 0.1 − 1 meV. Experimental
tests: e.g., specific heat, electromagnetic absorption, tunnel spectroscopy.

Superconductor nanostructures. Superconductors have two intrinsic length

scales, the coherence length 𝜉 and the penetration depth 𝜆. Both are ∼ 100 nm.
The physics of nanostructures on these length scales is called mesoscopic super-
conductivity. Of particular interest are hybrid structures, which are characterized
by the competition of superconductivity and other interactions or ground states (e.g.,
magnetism, Coulomb repulsion).

Examples:

• proximity effect: Cooper pairs penetrate into a normal metal (slide 13)
8
A. A. Abrikosov, Sov. Phys. JETP 5, 1174 (1957), Nobel prize 2003
9
J. Bardeen et al., Phys. Rev. 108, 1175 (1957), Nobelpreis 1972
10
L. P. Gor’kov, Sov. Phys. JETP 9, 1364 (1959)

4
• triplet proximity effect: dissipationless spin currents (slide 14)

• charge qubit: coherent dynamics (slide 15)

• Cooper pair splitter: entangled electron pairs (slide 16)