Theory of Superconductivity

Theory of Superconductivity
Carsten Timm
Wintersemester 2011/2012
TU Dresden
Institute of Theoretical Physics
Version: February 6, 2012 A L TEX & Figures: S. Lange and C. Timm
Contents
1 Introduction 1.1 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Books . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Basic experiments 2.1 Conventional superconductors . . . . . . . 2.2 Superuid helium . . . . . . . . . . . . . . 2.3 Unconventional superconductors . . . . . 2.4 Bose-Einstein condensation in dilute gases 3 Bose-Einstein condensation 5 5 5 6
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
7 . 7 . 9 . 10 . 12 13
4 Normal metals 18 4.1 Electrons in metals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 4.2 Semiclassical theory of transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 5 Electrodynamics of superconductors 5.1 London theory . . . . . . . . . . . . 5.2 Rigidity of the superuid state . . . 5.3 Flux quantization . . . . . . . . . . . 5.4 Nonlocal response: Pippard theory . 23 23 25 26 27 30 30 33 37 41 44
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
6 Ginzburg-Landau theory 6.1 Landau theory of phase transitions . . . . . . . 6.2 Ginzburg-Landau theory for neutral superuids 6.3 Ginzburg-Landau theory for superconductors . 6.4 Type-I superconductors . . . . . . . . . . . . . 6.5 Type-II superconductors . . . . . . . . . . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
7 Superuid and superconducting lms 52 7.1 Superuid lms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 7.2 Superconducting lms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 8 Origin of attractive interaction 8.1 Reminder on Green functions . . . . . 8.2 Coulomb interaction . . . . . . . . . . 8.3 Electron-phonon interaction . . . . . . 8.4 Eective interaction between electrons 70 70 73 77 79
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
9 Cooper instability and BCS ground state 82 9.1 Cooper instability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 9.2 The BCS ground state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 10 BCS theory 10.1 BCS mean-eld theory . . . . . . . . . . . . . 10.2 Isotope eect . . . . . . . . . . . . . . . . . . 10.3 Specic heat . . . . . . . . . . . . . . . . . . . 10.4 Density of states and single-particle tunneling 10.5 Ultrasonic attenuation and nuclear relaxation 10.6 Ginzburg-Landau-Gorkov theory . . . . . . . 11 Josephson eects 11.1 The Josephson eects in Ginzburg-Landau 11.2 Dynamics of Josephson junctions . . . . . 11.3 Bogoliubov-de Gennes Hamiltonian . . . . 11.4 Andreev reection . . . . . . . . . . . . . 89 89 95 95 97 102 106 107 107 109 113 114 120 120 123 130 132
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
theory . . . . . . . . . . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
12 Unconventional pairing 12.1 The gap equation for unconventional pairing . 12.2 Cuprates . . . . . . . . . . . . . . . . . . . . . 12.3 Pnictides . . . . . . . . . . . . . . . . . . . . 12.4 Triplet superconductors and He-3 . . . . . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
. . . .
Introduction
1.1 Scope
Superconductivity is characterized by a vanishing static electrical resistivity and an expulsion of the magnetic eld from the interior of a sample. We will discuss these basic experiments in the following chapter, but mainly this course is dealing with the theory of superconductivity. We want to understand superconductivity using methods of theoretical physics. Experiments will be mentioned if they motivate certain theoretical ideas or support or contradict theoretical predictions, but a systematic discussion of experimental results will not be given. Superconductivity is somewhat related to the phenomena of superuidity (in He-3 and He-4) and Bose-Einstein condensation (in weakly interacting boson systems). The similarities are found to lie more in the eective lowenergy description than in the microscopic details. Microscopically, superuidity in He-3 is most closely related to superconductivity since both phenomena involve the condensation of fermions, whereas in He-4 and, of course, Bose-Einstein condensates it is bosons that condense. We will discuss these phenomena briey. The course assumes knowledge of the standard material from electrodynamics, quantum mechanics I, and thermodynamics and statistics. We will also use the second-quantisation formalism (creation and annihilation operators), which are usually introduced in quantum mechanics II. A prior cours on introductory solid state physics would be usefull but is not required. Formal training in many-particle theory is not required, necessary concepts and methods will be introduced (or recapitulated) as needed.
1.2

Overview
basic experiments review of Bose-Einstein condensation electrodynamics: London and Pippard theories Ginzburg-Landau theory, Anderson-Higgs mechanism vortices origin of electron-electron attraction Cooper instability, BCS theory consequences: thermodynamics, tunneling, nuclear relaxation Josephson eects, Andreev scattering Bogoliubov-de Gennes equation cuprate superconductors pnictide superconductors topological superconductors 5
This is a maximal list of topics to be covered, not necessarily in this order:
1.3
Books
There are many textbooks on superconductivity and it is recommended to browse a few of them. None of them covers all the material of this course. M. Tinkhams Introduction to Superconductivity (2nd edition) is well written and probably has the largest overlap with this course.
Basic experiments
In this chapter we will review the essential experiments that have established the presence of superconductivity, superuidity, and Bose-Einstein condensation in various materials classes. Experimental observations that have helped to elucidate the detailed properties of the superconductors or superuids are not covered; some of them are discussed in later chapters.
2.1
Conventional superconductors
After H. Kamerlingh Onnes had managed to liquify Helium, it became for the rst time possible to reach temperatures low enough to achieve superconductivity in some chemical elements. In 1911, he found that the static resistivity of mercury abruptly fell to zero at a critical temperature Tc of about 4.1 K.
normal metal
superconductor 0 0 Tc T
In a normal metal, the resistivity decreases with decreasing temperature but saturates at a nite value for T 0. The most stringent bounds on the resisitivity can be obtained not from direct measurement but from the decay of persistant currents, or rather from the lack of decay. A current set up (by induction) in a superconducting ring is found to persist without measurable decay after the electromotive force driving the current has been switched o.
Assuming exponential decay, I (t) = I (0) et/ , a lower bound on the decay time is found. From this, an upper bound of 1025 m has been extracted for the resistivity. For comparison, the resistivity of copper at room temperature is Cu 1.7 108 m. The second essential observation was that superconductors not only prevent a magnetic eld from entering but actively expel the magnetic eld from their interior. This was oberserved by W. Meiner and R. Ochsenfeld in 1933 and is now called the Meiner or Meiner-Ochsenfeld eect.
B
B=0
From the materials relation B = H with the permeability = 1 + 4 and the magnetic suceptibility (note that we are using Gaussian units ) we thus nd = 0 and = 1/4 . Superconductors are diamagnetic since < 0. What is more, they realize the smallest (most diamagnetic) value of consistent with thermodynamic stability. The eld is not just diminished but completely expelled. They are thus perfect diamagnets. It costs energy to make the magnetic eld nonuniform although the externally applied eld is uniform. It is plausible that at some externally applied magnetic eld Hc (T ) Bc (T ) this cost will be so high that there is no advantage in forming a superconducting state. For typical conventional superconductors, the experimental phase diagram in the temperature-magnetic eld plane looks like this:
H normal metal Hc (T= 0)
1s to rd er
superconductor Tc
2nd order T
To specify which superconductors discovered after Hg are conventional, we need a denition of what we want to call conventional superconductors. There are at least two inequivalent but often coinciding denitions: Conventional superconductors show a superconducting state of trivial symmetry (we will discuss later what this means), result from an attractive interaction between electrons for which phonons play a dominant role. Conventional superconductivity was observed in quite a lot of elements at low temperatures. The record critical temperature for elements are Tc = 9.3 K for Nb under ambient pressure and Tc = 20 K for Li under high pressure. Superconductivity is in fact rather common in the periodic table, 53 pure elements show it under some conditions. Many alloys and intermetallic compounds were also found to show conventional superconductivity according to the above criteria. Of these, for a long time Nb3 Ge had the highest known Tc of 23.2 K. But it is now thought that MgB2 (Tc = 39 K) and a few related compounds are also conventional superconductors in the above sense. They nevertheless show some interesting properties. The rather high Tc = 39 K of MgB2 is interesting since it is max on the order of the maximum Tc 30 K expected for phonon-driven superconductivity. To increase Tc further, the interaction between electrons and phonons would have to be stronger, which however would make the material unstable towards a charge density wave. MgB2 would thus be an optimal conventional superconductor. 8
Superconductivity with rather high Tc has also been found in fullerites, i.e., compounds containing fullerene anions. The record Tc in this class is at present Tc = 38 K for b.c.c. Cs3 C60 under pressure. Superconductivity in fullerites was originally thought to be driven by phonons with strong molecular vibration character but there is recent evidence that it might be unconventional (not phonon-driven).
2.2
Superuid helium
In 1937 P. Kapiza and independently Allen and Misener discovered that helium shows a transition at Tc = 2.17 K under ambient pressure, below which it ows through narrow capilaries without resistance. The analogy to superconductivity is obvious but here it was the viscosity instead of the resistivity that dropped to zero. The phenomenon was called superuidity. It was also observed that due to the vanishing viscosity an open container of helium would empty itself through a ow in the microscopically thin wetting layer.
He
On the other hand, while part of the liquid ows with vanishing viscosity, another part does not. This was shown using torsion pendulums of plates submerged in helium. For T > 0 a temperature-dependent normal component oscillates with the plates.
He
Natural atmospheric helium consists of 99.9999 % He-4 and only 0.0001 % He-3, the only other stable isotope. The oberserved properties are thus essentially indistinguishable from those of pure He-4. He-4 atoms are bosons since they consist of an even number (six) of fermions. For weakly interacting bosons, A. Einstein predicted in 1925 that a phase transition to a condensed phase should occur (Bose-Einstein condensation ). The observation of superuidity in He-4 was thus not a suprisein contrast to the discovery of superconductivitybut in many details the properties of He-4 were found to be dierent from the predicted Bose-Einstein condensate. The reason for this is that the interactions between helium atoms are actually quite strong. For completeness, we sketch the temperature-pressure phase diagram of He-4:
p (10 6 Pa) 3 solid phases
2 normal liquid
superfluid 1 ambient 0
critical endpoint
gas
T (K)
The other helium isotope, He-3, consists of fermionic atoms so that Bose-Einstein condensation cannot take place. Indeed no superuid transition was observed in the temperature range of a few Kelvin. It then came as a big suprise when superuidity was nally observed at much lower temperatures below about 2.6 mK by D. Lee, D. Oshero, and R. Richardson in 1972. In fact they found two new phases at low temperatures. (They originally misinterpreted them as possible magnetic solid phases.) Here is a sketch of the phase diagram, note the temperature scale:
p (10 6 Pa) solid phases 3
superfluid phase A
superfluid phase B
normal liquid
1 ambient 0 0 1 2 3 T (mK)
Superuid He-3 shows the same basic properties as He-4. But unlike in He-4, the superuid states are sensitive to an applied magnetic eld, suggesting that the states have non-trivial magnetic properties.
2.3
Unconventional superconductors
By the late 1970s, superconductivity seemed to be a more or less closed subject. It was well understood based on the BCS theory and extensions thereof that dealt with strong interactions. It only occured at temperatures up to 23.2 K (Nb3 Ge) and thus did not promise widespread technological application. It was restricted to non-magnetic metallic elements and simple compounds. This situation started to change dramatically in 1979. Since then, superconductivity has been observed in various materials classes that are very dierent from each other and from the typical low-Tc superconductors known previously. In many cases, the superconductivity was unconventional and often Tc was rather high. We now give a brief and incomplete historical overview.
10
In 1979, Frank Steglich et al. observed superconductivity below Tc 0.5 K in CeCu2 Si2 . This material is not a normal metal in its normal state. Instead is is a heavy-fermion metal. The electrons at the Fermi energy have strong Ce f -orbital character. The very strong Coulomb repulsion between electrons in the f -shell leads to a high eective mass m me at the Fermi energy, hence the name. Since then, superconductivity has been found in various other heavy-fermion compounds. BCS theory cannot explain superconductivity in these highly correlated metals. Nuclear magnetic resonance (discussed below) and other experimental techniques have shown that many of these heavy-fermion superconductors show unconventional symmetry of the superconducting state. Also in 1979, D. Jrome et al. (Klaus Bechgaards group) observed superconductivity in an organic salt called (TMTSF)2 PF2 with Tc = 1.1 K. Superconductivity has since been found in various organic materials with a maximum Tc of about 18 K. The symmetry of the superconducting state is often unconventional. (We do not include fullerites under organic compounds since they lack hydrogen atoms.) While the previously mentioned discoveries showed that superconductivity can occur in unexpected materials classes and probably due to unconventional mechanisms, the Tc values did not surpass the Tc 23 K of Nb3 Ge. In 1986, J. G. Bednorz and K. A. Mller oberserved superconductivity in La2x Bax CuO4 (the layered perovskite cuprate La2 CuO4 with some Ba substituted for La) with Tc on the order of 35 K. In the following years, many other superconductors based on the same type of nearly at CuO2 planes sketched below were discovered. The record transition temperatures for cuprates and for all superconductors are Tc = 138 K for Hg0.8 Tl0.2 Ba2 Ca2 Cu3 O8+ at ambient pressure and Tc = 164 K for HgBa2 Ca2 Cu3 O8+ under high pressure. The high Tc values as well as many experimental probes show that the cuprates are unconventional superconductors. We will come back to these materials class below.
Cu
In 1991, A. F. Hebard et al. found that the fullerite K3 C60 = (K )3 C3 60 became superconducting below Tc = 18 K. Tc in this class has since been pushed to Tc = 33 K for Cs2 RbC60 at ambient pressure and Tc = 38 K for (b.c.c., while all the other known superconducting fullerites are f.c.c.) Cs3 C60 under high pressure. The symmetry of the superconducting state appears to be trivial but, as noted above, there is an ongoing debate on whether the pairing is phonon-mediated.
In 2001, Nagamatsu et al. reported superconductivity in MgB2 with Tc = 39 K. The high Tc and the layered crystal structure, reminiscant of cuprates, led to the expectation that superconductivity in MgB2 is unconventional. However, most experts now think that it is actually conventional, as noted above. The most recent series of important discoveries started in 2006, when Kamahira et al. (H. Hosonos group) observed superconductivity with Tc 4 K in LaFePO, another layered compound. While this result added a new materials class based on Fe2+ to the list of superconductors, it did not yet cause much excitement due to the low Tc . However, in 2008, Kamihara et al. (the same group) found superconductivity with Tc 26 K in LaFeAsO1x Fx . Very soon thereafter, the maximum Tc in this iron pnictide class was pushed to 55 K. Superconductivity was also observed in several related materials classes, some of them not containing oxygen (e.g., LiFeAs) and some with the pnictogen (As) replaced by a chalcogen (e.g., FeSe). The common structural element is a at, square Fe2+ layer with a pnictogen or chalcogen sitting alternatingly above and below the centers of the Fe squares. Superconductivity is thought to be unconventional.
11
Fe
As
2.4
Bose-Einstein condensation in dilute gases
An important related breakthrough was the realization of a Bose-Einstein condensate (BEC) in a highly diluted and very cold gas of atoms. In 1995, Anderson et al. (C. E. Wiman and E. A. Cornells group) reported condensation in a dilute gas of Rb-87 below Tc = 170 nK (!). Only a few months later, Davis et al. (W. Ketterles group) reported a BEC of Na-23 containing many more atoms. About a year later, the same group was able to create two condensates and then merge them. The resulting interference eects showed that the atoms where really in a macroscopic quantum state, i.e., a condensate. All observations are well understood from the picture of a weakly interacting Bose gas. Bose-Einstein condensation will be reviewed in the following chapter.
12
Bose-Einstein condensation
In this short chapter we review the theory of Bose-Einstein condensation. While this is not the correct theory for superconductivity, at least in most superconductors, it is the simplest description of a macroscopic quantum condensate. This concept is central also for superconductivity and superuidity. We consider an ideal gas of indistinguishable bosons. Ideal means that we neglect any interaction and also any nite volume of the particles. There are two cases with completely dierent behavior depending on whether the particle number is conserved or not. Rb-87 atoms are bosons (they consist of 87 nucleons and 37 electrons) with conserved particle number, whereas photons are bosons with non-conserved particle number. Photons can be freely created and destroyed as long as the usual conservation laws (energy, momentum, angular momentum, . . . ) are satised. Bosons without particle-number conservation show a Planck distribution, nP ( E ) = 1 eE 1 (3.1)
with := 1/kB T , for a grand-canonical ensemble in equilibrium. Note the absence of a chemical potential, which is due to the non-conservation of the particle number. This distribution function is an analytical function of temperature and thus does not show any phase transitions. The situation is dierent for bosons with conserved particle number. We want to consider the case of a given number N of particles in contact with a heat bath at temperature T . This calls for a canonical description (N, T given). However, it is easier to use the grand-canonical ensemble with the chemical potential given. For large systems, uctuations of the particle number become small so that the descriptions are equivalent. However, must be calculated from the given N . The grand-canonical partition function is Z= (
i
1 + e (i ) + e2 (i ) +
) =
1 , 1 e (i )
(3.2)
where i counts the single-particle states of energy i in a volume V . The form of Z expresses that every state can be occupied not at all, once, twice, etc. For simplicity, we assume the volume to be a cube with periodic boundary conditions. Then the states can be enumerated be wave vectors k compatible with these boundary conditions. Introducing the fugacity y := e , (3.3) we obtain Z= ln Z =
k k
1 1 yek ln ( ) 1 = ln 1 yek . k 1 ye
k
(3.4) (3.5)
13
The fugacity has to be chosen to give the correct particle number N=

!
nk =
1 e (k ) 1
1 y 1 ek 1
(3.6)
Since nk must be non-negative, must satisfy k k. (3.7)
For a free particle, the lowest possible eigenenergy is k = 0 for k = 0 so that we obtain 0. For a large volume V , the allowed vectors k become dense and we can replace the sums over k by integrals according to 2V d3 k 3 /2 = 3 (2m) d (3.8) V 3 (2 ) h
k 0
In the last equation we have used the density of states (DOS) of free particles in three dimensions. This replacement contains a fatal mistake, though. The DOS for = 0 vanishes so that any particles in the state with k = 0, k = 0 do not contribute to the results. But that is the ground state! For T = 0 all bosons should be in this state. We thus expect incorrect results at low temperatures. Our mistake was that Eq. (3.8) does not hold if the fraction of bosons in the k = 0 ground state is macroscopic, i.e., if N0 /N := n0 /N remains nite for large V . To correct this, we treat the k = 0 state explicitly (the same would be necessary for any state with macroscopic occupation). We write
k
2V (2m)3/2 h3
d
0
+ (k = 0 term).
(3.9)
Then 2V ln Z = 3 (2m)3/2 h with 2V N = 3 (2m)3/2 h Dening gn (y ) := 1 (n)

d
0
ln 1 ye
ln(1 y )
by parts
2V 2 (2m)3/2 3 h 3
d
0
3 /2 y 1 e 1
ln(1 y ) (3.10)
d
0
1 2V 1 + = 3 (2m)3/2 y e 1 y 1 1 h

d
0
y + . y 1 e 1 1 y
(3.11)
dx
0
y 1 ex
xn1 1
(3.12)
for 0 y 1 and n R, and the thermal wavelength := we obtain ln Z = V g5/2 (y ) ln(1 y ), 3 V y . N = 3 g3/2 (y ) + 1y
=: N =: N0
h2 , 2mkB T
(3.13)
(3.14) (3.15)
14
We note the identity gn (y ) = which implies gn (0) = 0, 1 gn (1) = = (n) kn

k=1 yk , kn
(3.16)
k=1
(3.17) for n > 1 (3.18)
with the Riemann zeta function (x). Furthermore, gn (y ) increases monotonically in y for y [0, 1[.
gn ( y) g3/2 2 g5/2 1 0 0 0.5 1 y

We now have to eliminate the fugacity y from Eqs. (3.14) and (3.15) to obtain Z as a function of the particle number N . In Eq. (3.14), the rst term is the number of particles in excited states (k > 0), whereas the second term is the number of particles in the ground state. We consider two cases: If y is not very close to unity (specically, if 1 y 3 /V ), N0 = y/(1 y ) is on the order of unity, whereas N is an extensive quantity. Thus N0 can be neglected and we get V N (3.19) = N = 3 g3/2 (y ). Since g3/2 (3/2) 2.612, this equation can only be solved for the fugacity y if the concentration satises (3/2) N . V 3 To have 1 y 3 /V in the thermodynamic limit we require, more strictly, N (3/2) < . V 3 (3.21) (3.20)
goo
Note that 3 T 3/2 increases with decreasing temperature. Hence, at a critical temperature Tc , the inequality is no longer fullled. From N ! (3/2) (3.22) =( ) 3 /2 V h2
2mkB Tc
we obtain kB Tc = 1 h2 2 / 3 [ (3/2)] 2m
N V
)2/3 . (3.23)
If, on the other hand, y is very close to unity, N0 cannot be neglected. Also, in this case we nd N = V V g3/2 (y ) = 3 g3/2 (1 O(3 /V )), 3 15 (3.24)
where O(3 /V ) is a correction of order 3 /V 1. Thus, by Taylor expansion, N = V V g3/2 (1) O(1) = 3 (3/2) O(1). 3 (3.25)
The intensive term O(1) can be neglected compared to the extensive one so that V N = 3 (3/2). This is the maximum possible value at temperature T . Furthermore, N0 = y/(1 y ) is solved by y= N0 1 = . N0 1 1 + 1/N0 (3.27) (3.26)
For y to be very close to unity, N0 must be N0 1. Since V N0 = N N = N 3 (3/2) must be positive, we require N (3/2) > V 3 T < Tc . We conclude that the fraction of particles in excited states is N V 3 ( Tc ) (3 / 2) = = = N N 3 3 ( T ) The fraction of particles in the ground state is then N0 =1 N In summery, we nd in the thermodynamic limit (a) for T > Tc : N = 1, N (b) for T < Tc : N = N ( T Tc ) 3 /2 , N0 1, N N0 =1 N ( T Tc )3/2 . (3.34) (3.33) ( T Tc ) 3 /2 . (3.32) ( T Tc )3/2 . (3.31) (3.29) (3.30) (3.28)
1 N0 / N N / N 0 0 0.5 1 T / Tc
16
We nd a phase transition at Tc , below which a macroscopic fraction of the particles occupy the same singleparticle quantum state. This fraction of particles is said to form a condensate. While it is remarkable that Bose-Einstein condensation happens in a non-interacting gas, the BEC is analogous to the condensate in strongly interacting superuid He-4 and, with some added twists, in superuid He-3 and in superconductors. We can now use the partition function to derive equations of state. As an example, we consider the pressure p= kB T =+ kB T ln Z = 3 g5/2 (y ) V V (3.35)
( is the grand-canonical potential). We notice that only the excited states contribute to the pressure. The term ln(1 y ) from the ground state drops out since it is volume-independant. This is plausible since particles in the condensate have vanishing kinetic energy. For T > Tc , we can nd y and thus p numerically. For T < Tc we may set y = 1 and obtain p= kB T (5/2) T 5/2 . 3 (3.36)
Remember that for the classical ideal gas at constant volume we nd p T. (3.37)
For the BEC, the pressure drops more rapidly since more and more particles condense and thus no longer contribute to the pressure.
17
Normal metals
To be able to appreciate the remarkable poperties of superconductors, it seems useful to review what we know about normal conductors.
4.1
Electrons in metals
Let us ignore electron-electron Coulomb interaction and deviations from a perfectly periodic crystal structure (due to defects or phonons) for now. Then the exact single-particle states are described by Bloch wavefunctions k (r) = uk (r) eikr , (4.1)
where uk (r) is a lattice-periodic function, is the band index including the spin, and k is the crystal momentum in the rst Brillouin zone. Since electrons are fermions, the average occupation number of the state |k with energy k is given by the Fermi-Dirac distribution function nF (k ) = 1 . e (k ) + 1 (4.2)
If the electron number N , and not the chemical potential , is given, has to be determined from N=
k
1 e (k ) +1
(4.3)
cf. our discussion for ideal bosons. In the thermodynamic limit we again replace d3 k V . (2 )3
k
(4.4)
Unlike for bosons, this is harmless for fermions, since any state can at most be occupied once so that macroscopic occupation of the single-particle ground state cannot occur. Thus we nd dk 3 N 1 = . (4.5) 3 e (k ) + 1 V (2 ) If we lower the temperature, the Fermi function nF becomes more and more step-like. For T 0, all states with energies k EF := (T 0) are occupied (EF is the Fermi energy), while all states with k > EF are empty. This Fermi sea becomes fuzzy for energies k EF at nite temperatures but remains well dened as long as kB T EF . This is the case for most materials we will discuss. The chemical potential, the occupations nF (k ), and thus all thermodynamic variables are analytic functions of T and N/V . Thus there is no phase transition, unlike for bosons. Free fermions represent a special case with only a single band with dispersion k = 2 k 2 /2m. If we replace m by a material-dependant eective mass, this 18
gives a reasonable approximation for simple metals such as alkali metals. Qualitatively, the conclusions are much more general. Lattice imperfections and interactions result in the Bloch waves k (r) not being exact single-particle eigenstates. (Electron-electron and electron-phonon interactions invalidate the whole idea of single-particle states.) However, if these eects are in some sense small, they can be treated pertubatively in terms of scattering of electrons between single-particle states |k.
4.2
Semiclassical theory of transport
We now want to derive an expression for the current in the presence of an applied electric eld. This is a question about the response of the system to an external perturbation. There are many ways to approach this type of question. If the perturbation is small, the response, in our case the current, is expected to be a linear function of the perturbation. This is the basic assumption of linear-response theory. In the framework of many-particle theory, linear-response theory results in the Kubo formula (see lecture notes on many-particle theory). We here take a dierent route. If the external perturbation changes slowly in time and space on atomic scales, we can use a semiclassical description. Note that the following can be derived cleanly as a limit of many-particle quantum theory. The idea is to consider the phase space distribution function (r, k, t). This is a classical concept. From quantum mechanics we know that r and p = k are subject to the uncertainty principle rp /2. Thus distribution functions that are localized in a phase-space volume smaller than on the order of 3 violate quantum mechanics. On the other hand, if is much broader, quantum eects should be negligible. The Liouville theorem shows that satises the continuity equation d = 0 +r +k t r k dt (4.6)
(phase-space volume is conserved under the classical time evolution). Assuming for simplicity a free-particle dispersion, we have the canonical (Hamilton) equations H p k = = , p m m 1 H 1 1 = 1p = = V = F k r = r with the Hamiltonian H and the force F. Thus we can write ( ) k F + + = 0. t m r k (4.7) (4.8)
(4.9)
This equation is appropriate for particles in the absence of any scattering. For electrons in a uniform and timeindependant electric eld we have F = eE. (4.10) Note that we always use the convention that e > 0. It is easy to see that ( ) eE (r, k, t) = f k + t
(4.11)
is a solution of Eq. (4.9) for any dierentiable function f . This solution is uniform in real space (/ r 0) and shifts to larger and larger momenta k for t . It thus describes the free acceleration of electrons in an electric eld. There is no nite conductivity since the current never reaches a stationary value. This is obviously not a correct description of a normal metal. Scattering will change as a function of time beyond what as already included in Eq. (4.9). We collect all processes not included in Eq. (4.9) into a scattering term S []: ) ( k F + + = S []. (4.12) t m r k 19
This is the famous Boltzmann equation. The notation S [] signies that the scattering term is a functional of . It is generally not simply a function of the local density (r, k, t) but depends on everywhere and at all times, to the extent that this is consistent with causality. While expressions for S [] can be derived for various cases, for our purposes it is sucient to employ the simple but common relaxation-time approximation. It is based on the observation that (r, k, t) should relax to thermal equilibrium if no force is applied. For fermions, the equilibrium is 0 (k) nF (k ). This is enforced by the ansatz (r, k, t) 0 (k) S [] = . (4.13) Here, is the relaxation time, which determines how fast relaxes towards 0 . If there are dierent scattering mechanisms that act independantly, the scattering integral is just a sum of contributions of these mechanisms, S [] = S1 [] + S2 [] + . . . (4.14) Consequently, the relaxation rate 1/ can be written as (Matthiessens rule) 1 1 1 = + + ... 1 2 There are three main scattering mechanisms: Scattering of electrons by disorder: This gives an essentially temperature-independant contribution, which dominates at low temperatures. Electron-phonon interaction: This mechanism is strongly temperature-dependent because the available phase space shrinks at low temperatures. One nds a scattering rate 1 e-ph T 3. (4.16) (4.15)
However, this is not the relevant rate for transport calculations. The conductivity is much more strongly aected by scattering that changes the electron momentum k by a lot than by processes that change it very little. Backscattering across the Fermi sea is most eective.
ky k = 2 kF kx
Since backscattering is additionally suppressed at low T , the relevant transport scattering rate scales as 1
trans e-ph
T 5.
(4.17)
Electron-electron interaction: For a parabolic free-electron band its contribution to the resistivity is actually zero since Coulomb scattering conserves the total momentum of the two scattering electrons and therefore does not degrade the current. However, in a real metal, umklapp scattering can take place that conserves the total momentum only modulo a reciprocal lattice vector. Thus the electron system can transfer momentum to the crystal as a whole and thereby degrade the current. The temperature dependance is typically 1
umklapp e-e
T 2.
(4.18)
20
We now consider the force F = eE and calulate the current density d3 k k (r, k, t). j(r, t) = e (2 )3 m To that end, we have to solve the Boltzmann equation ( ) k eE 0 + = . t m r k
(4.19)
(4.20)
We are interested in the stationary solution (/t = 0), which, for a uniform eld, we assume to be spacially uniform (/ r = 0). This gives eE 0 (k) (k) (k) = k eE (k) = 0 (k) + (k). k (4.21) (4.22)
We iterate this equation by inserting it again into the nal term: ( ) eE eE eE (k) = 0 (k) + 0 (k) + (k). k k k
(4.23)
To make progress, we assume that the applied eld E is small so that the response j is linear in E. Under this assumption we can truncate the iteration after the linear term, (k) = 0 (k) + eE 0 (k). k (4.24)
By comparing this to the Taylor expansion ( ) eE eE 0 k + = 0 (k) + 0 (k) + . . . k we see that the solution is, to linear order in E, ( ) eE (k) = 0 k + nF (k+eE / ).
(4.25)
(4.26)
Thus the distribution function is simply shifted in k-space by eE / . Since electrons carry negative charge, the distribution is shifted in the direction opposite to the applied electric eld.
E 0 eE /h
The current density now reads ( ) eE d3 k k d3 k k d3 k k eE 0 j = e k + e ( k ) e . = 0 0 (2 )3 m (2 )3 m (2 )3 m k

=0
(4.27)
21
The rst term is the current density in equilibrium, which vanishes. In components, we have e2 d3 k 0 by parts e2 d3 k k e2 d3 k j = E k = + E = E 0 . 0 m (2 )3 k m (2 )3 k m (2 )3
= =n
(4.28)
Here, the integral is the concentration of electrons in real space, n := N/V . We nally obtain j= so that the conductivity is e2 n . (4.30) m This is the famous Drude formula. For the resistivity = 1/ we get, based on our discussion of scattering mechanisms, ) ( 1 m m 1 1 (4.31) = 2 = 2 + transport + umklapp . e n e n dis e-ph e-e = e2 n ! E = E m (4.29)
large T : T 5 most due to electron phonon scattering residual resistivity due to disorder T
22
Electrodynamics of superconductors
Superconductors are dened by electrodynamic propertiesideal conduction and magnetic-eld repulsion. It is thus appropriate to ask how these materials can be described within the formal framework of electrodynamics.
5.1
London theory
In 1935, F. and H. London proposed a phenomenological theory for the electrodynamic properties of superconductors. It is based on a two-uid picture : For unspecied reasons, the electrons from a normal uid of concentration nn and a superuid of concentration ns , where nn + ns = n = N/V . Such a picture seemed quite plausible based on Einsteins theory of Bose-Einstein condensation, although nobody understood how the fermionic electrons could form a superuid. The normal uid is postulated to behave normally, i.e., to carry an ohmic current j n = n E governed by the Drude law e2 nn . (5.2) m The superuid is assumed to be insensitive to scattering. As noted in section 4.2, this leads to a free acceleration of the charges. With the supercurrent js = e ns vs (5.3) n = and Newtons equation of motion d F eE vs = = , dt m m we obtain (5.4) (5.1)
js e2 ns = E. (5.5) t m This is the First London Equation. We are only interested in the stationary state, i.e., we assume nn and ns to be uniform in space, this is the most serious restriction of London theory, which will be overcome in Ginzburg-Landau theory. Note that the curl of the First London Equation is e2 ns e2 ns B js = E= . t m mc t This can be integrated in time to give e2 ns B + C(r), (5.7) mc where the last term represents a constant of integration at each point r inside the superconductor. C(r) should be determined from the initial conditions. If we start from a superconducting body in zero applied magnetic js = 23 (5.6)
eld, we have js 0 and B 0 initially so that C(r) = 0. To describe the Meiner-Ochsenfeld eect, we have to consider the case of a body becoming superconducting (by cooling) in a non-zero applied eld. However, this case cannot be treated within London theory since we here assume the superuid density ns to be constant in time. To account for the ux expulsion, the Londons postulated that C 0 regardless of the history of the system. This leads to e2 ns js = B, (5.8) mc the Second London Equation. Together with Ampres Law 4 4 B= js + jn (5.9) c c (there is no displacement current in the stationary state) we get B= 4e2 ns 4e2 ns 4 4 B n E = n . B+ B 2 mc c mc2 c t (5.10)
We drop the last term since we are interested in the stationary state and use an identity from vector calculus, ( B) + 2 B = Introducing the London penetration depth L := this equation assumes the simple form 2 B = 1 B. 2 L (5.13) mc2 , 4e2 ns (5.12) 4e2 ns B. mc2 (5.11)
Let us consider a semi-innite superconductor lling the half space x > 0. A magnetic eld Bapl = Hapl = Bapl y is applied parallel to the surface. One can immediately see that the equation is solved by B(x) = Bapl y ex/L for x 0. (5.14)
The magnetic eld thus decresases exponentially with the distance from the surface of the superconductor. In bulk we indeed nd B 0.
B Bapl outside 0 L inside x
The Second London Equation js = and the continuity equation
c B 42 L
(5.15)
js = 0 can now be solved to give js (x) = c Bapl z ex/L 4L for x 0.
(5.16) (5.17)
24
Thus the supercurrent ows in the direction parallel to the surface and perpendicular to B and decreases into the bulk on the same scale L . js can be understood as the screening current required to keep the magnetic eld out of the bulk of the superconductor. The two London equations (5.5) and (5.8) can be summarized using the vector potential: js = e2 ns A. mc (5.18)
This equation is evidently not gauge-invariant since a change of gauge A A + (5.19)
changes the supercurrent. (The whole London theory is gauge-invariant since it is expressed in terms of E and B.) Charge conservation requires js = 0 and thus the vector potential must be transverse, A = 0. (5.20)
This is called the London gauge. Furthermore, the supercurrent through the surface of the superconducting region is proportional to the normal conponent A . For a simply connected region these conditions uniquely determine A(r). For a multiply connected region this is not the case; we will return to this point below.
5.2
Rigidity of the superuid state
F. London has given a quantum-mechanical justication of the London equations. If the many-body wavefunction of the electrons forming the superuid is s (r1 , r2 , . . . ) then the supercurrent in the presence of a vector potential A is [ ( ) ( ) ] 1 e e js (r) = e d3 r1 d3 r2 (r rj ) (5.21) s pj + A(rj ) s + s pj + A(rj ) s . 2m j c c Here, j sums over all electrons in the superuid, which have position rj and momentum operator pj = . i rj (5.22)
Making pj explicit and using the London gauge, we obtain [ ] e d3 r1 d3 r2 (r rj ) js (r) = s s s 2mi j rj rj s e2 A(r) d3 r1 d3 r2 (r rj ) s s mc j [ ] e e2 ns 3 3 = d r1 d r2 (r rj ) s s s s A(r). 2mi j rj rj mc
(5.23)
Now London proposed that the wavefunction s is rigid under the application of a transverse vector potential. More specically, he suggested that s does not contain a term of rst order in A, provided A = 0. Then, to rst order in A, the rst term on the right-hand side in Eq. (5.23) contains the unperturbed wavefunction one would obtain for A = 0. The rst term is thus the supercurrent for A 0, which should vanish due to Ampres law. Consequently, to rst order in A we obtain the London equation js = e2 ns A. mc (5.24)
The rigidity of s was later understood in the framework of BCS theory as resulting from the existence of a non-zero energy gap for excitations out of the superuid state. 25
5.3
Flux quantization
We now consider two concentric superconducting cylinders that are thick compared to the London penetration depth L . A magnetic ux = d2 rB (5.25)
penetrates the inner hole and a thin surface layer on the order of L of the inner cylinder. The only purpose of the inner cylinder is to prevent the magnetic eld from touching the outer cylinder, which we are really interested in. The outer cylinder is completetly eld-free. We want to nd the possible values of the ux .
y r
Although the region outside of the inner cylinder has B = 0, the vector potential does not vanish. The relation A = B implies ds A = d2 r B = . (5.26) By symmetry, the tangential part of A is . (5.27) 2r The London gauge requires this to be the only non-zero component. Thus outside of the inner cylinder we have, in cylindrical coordinates, A= = . (5.28) 2r 2 Since this is a pure gradient, we can get from A = 0 to A = (/2r) by a gauge transformation A = A A + with = (5.29)
. (5.30) 2 is continuous but multivalued outside of the inner cylinder. We recall that a gauge transformation of A must be accompanied by a transformation of the wavefunction, ( ) ie s exp (rj ) s . (5.31) c j This is most easily seen by noting that this guarantees the current in Sec. 5.2 to remain invariant under gauge transformations. Thus the wavefuncion at = 0 (A = 0) and at non-zero ux are related by ) ( ) ( e i e j 0 0 = exp i (5.32) = exp j s , s s c j 2 hc j 26
0 where j is the polar angle of electron j . For s as well as s to be single-valued and continuous, the exponential factor must not change for j j + 2 for any j . This is the case if
e Z hc
=n
hc e
with n Z.
(5.33)
We nd that the magnetic ux is quantized in units of hc/e. Note that the inner cylinder can be dispensed with: Assume we are heating it enough to become normal-conducting. Then the ux will ll the whole interior of the outer cylinder plus a thin (on the order of L ) layer on its inside. But if the outer cylinder is much thicker than L , this should not aect s appreciably, away from this thin layer. The quantum hc/e is actually not correct. Based in the idea that two electrons could form a boson that could Bose-Einstein condense, Onsager suggested that the relevant charge is 2e instead of e, leading to the superconducting ux quantum hc 0 := , (5.34) 2e which is indeed found in experiments.
5.4
Nonlocal response: Pippard theory
Experiments often nd a magnetic penetration depth that is signicantly larger than the Londons prediction L , in particular in dirty samples with large scattering rates 1/ in the normal state. Pippard explained this on the basis of a nonlocal electromagnetic response of the superconductor. The underlying idea is that the quantum state of the electrons forming the superuid cannot be arbitrarily localized. The typical energy scale of superconductivity is expected to be kB Tc . Only electrons with energies within kB Tc of the Fermi energy can contribute appreciably. This corresponds to a momentum range p determined by kB Tc ! = p p p = =
=EF
p m
=EF
= vF
(5.35) (5.36)
kB Tc vF
with the Fermi velocity vF . From this, we can estimate that the electrons cannot be localized on scales smaller than vF x = . (5.37) p kB Tc Therefore, Pippard introduced the coherence length 0 = vF kB Tc (5.38)
as a measure of the minimum extent of electronic wavepackets. is a numerical constant of order unity. BCS theory predicts 0.180. Pippard proposed to replace the local equation js = from London theory by the nonlocal 3 e2 ns js (r) = 40 mc with d3 r ( A(r )) /0 e 4 (5.40) e2 ns A mc (5.39)
= r r .
(5.41)
27
The special form of this equation was motivated by an earlier nonlocal generalization of Ohms law. The main point is that electrons within a distance 0 of the point r where the eld A acts have to respond to it because of the stiness of the wavefunction. If A does not change appreciably on the scale of 0 , we obtain 3 e2 ns ( A(r)) /0 3 e2 ns 3 ( A(r)) /0 d r e = d3 e . (5.42) js (r) = 40 mc 4 40 mc 4 The result has to be parallel to A(r) by symmetry (since it is a vector depending on a single vector A(r)). Thus js (r) = (r))2 ( A e/0 4 1 2 3 e2 ns cos2 /0 2 A(r) 2 d(cos ) d e = 4 40 mc 3 e2 ns A(r) 40 mc d3
1
3 e ns 2 e ns A(r) 2 0 = A(r). 40 mc 3 mc
0 2
(5.43)
We recover the local London equation. However, for many conventional superconductors, 0 is much larger than . Then A(r) drops to zero on a length scale of 0 . According to Pippards equation, the electrons respond to the vector potential averaged over regions of size 0 . This averaged eld is much smaller than A at the surface so that the screening current js is strongly reduced and the magnetic eld penetrates much deeper than predicted by London theory, i.e., L . The above motivation for the coherence length 0 relied on having a clean system. In the presence of strong scattering the electrons can be localized on the scale of the mean free path l := vF . Pippard phenomenologically generalized the equation for js by introducing a new length where 1 1 1 = + 0 l ( is a numerical constant of order unity) and writing 3 e2 ns ( A(r )) / js = d3 r e . 40 mc 4 (5.44)
(5.45)
Note that 0 appears in the prefactor but in the exponential. This expression is in good agreement with experiments for series of samples with varying disorder. It is essentially the same as the result of BCS theory. Also note that in the dirty limit l 0 , L we again recover the local London result following the same argument as above, but since the integral gives a factor of , which is not canceled by the prefactor 1/0 , the current is reduced to e2 ns e2 ns l js = A A. (5.46) = mc 0 mc 0 Taking the curl, we obtain e2 ns l B mc 0 4e2 ns l B= B mc2 0 4e2 ns l 2 B = B, mc2 0 js = (5.47) (5.48) (5.49)
in analogy to the derivation in Sec. 5.1. This equation is of the form 2 B = 1 B 2 (5.50)
28
0 0 = = L . l l Thus the penetration depth is increased by a factor of order 0 /l in the dirty limit, l 0 . mc2 4e2 ns
with the penetration depth
(5.51)
29
Ginzburg-Landau theory
Within London and Pippard theory, the superuid density ns is treated as given. There is no way to understand the dependence of ns on, for example, temperature or applied magnetic eld within these theories. Moreover, ns has been assumed to be constant in time and uniform and in spacean assumption that is expected to fail close to the surface of a superconductor. These deciencies are cured by the Ginzburg-Landau theory put forward in 1950. Like Ginzburg and Landau we ignore complications due to the nonlocal electromagnetic response. Ginzburg-Landau theory is developed as a generalization of London theory, not of Pippard theory. The starting point is the much more general and very powerful Landau theory of phase transitions, which we will review rst.
6.1
Landau theory of phase transitions
Landau introduced the concept of the order parameter to describe phase transitions. In this context, an order parameter is a thermodynamic variable that is zero on one side of the transition and non-zero on the other. In ferromagnets, the magnetization M is the order parameter. The theory neglects uctuations, which means that the order parameter is assumed to be constant in time and space. Landau theory is thus a mean-eld theory. Now the appropriate thermodynamic potential can be written as a function of the order parameter, which we call , and certain other thermodynamic quantities such as pressure or volume, magnetic eld, etc. We will always call the potential the free energy F , but whether it really is a free energy, a free enthalpy, or something else depends on which quantities are given (pressure vs. volume etc.). Hence, we write F = F (, T ), (6.1)
where T is the temperature, and further variables have been suppressed. The equilibrium state at temperature T is the one that minimizes the free energy. Generally, we do not know F (, T ) explicitly. Landaus idea was to expand F in terms of , including only those terms that are allowed by the symmetry of the system and keeping the minimum number of the simplest terms required to get non-trivial results. For example, in an isotropic ferromagnet, the order parameter is the three-component vector M. The free energy must be invariant under rotations of M because of isotropy. Furthermore, since we want to minimize F as a function of M, F should be dierentiable in M. Then the leading terms, apart from a trivial constant, are ( ) F (6.2) = M M + (M M)2 + O (M M)3 . 2 Denoting the coecients by and /2 is just convention. and are functions of temperature (and pressure etc.). What is the corresponding expansion for a superconductor or superuid? Lacking a microscopic theory, Ginzburg and Landau assumed based on the analogy with Bose-Einstein condensation that the superuid part is described by a single one-particle wave function s (r). They imposed the plausible normalization 2 d3 r |s (r)| = Ns = ns V ; (6.3) 30
Ns is the total number of particles in the condensate. They then chose the complex amplitude of s (r) as the order parameter, with the normalization 2 | | n s . (6.4) Thus the order parameter in this case is a complex number. They thereby neglect the spatial variation of s (r) on an atomic scale. The free energy must not depend on the global phase of s (r) because the global phase of quantum states is not observable. Thus we obtain the expansion F = | | +
2
4 6 | | + O(| | ). 2
(6.5)
Only the absolute value | | appears since F is real. Odd powers are excluded since they are not dierentiable at = 0. If > 0, which is not guaranteed by symmetry but is the case for superconductors and superuids, we can neglect higher order terms since > 0 then makes sure that F ( ) is bounded from below. Now there are two cases: If 0, F ( ) has a single minimum at = 0. Thus the equilibrium state has ns = 0. This is clearly a normal metal (nn = n) or a normal uid. If < 0, F ( ) has a ring of minima with equal amplitude (modulus) | | but arbitrary phase. We easily see F 3 = 2 | | + 2 | | = 0 | | = 0 (this is a maximum) or | | = (6.6) | | Note that the radicand is positive.
F >0 =0 <0 0 / Re
Imagine this gure rotated around the vertical axis to nd F as a function of the complex . F ( ) for < 0 is often called the Mexican-hat potential. In Landau theory, the phase transition clearly occurs when = 0. Since then T = Tc by denition, it is useful to expand and to leading order in T around Tc . Hence, = (T Tc ), const. = Then the order parameter below Tc satises | | = > 0, (6.7) (6.8)
(T Tc ) =
T Tc .
(6.9)
Tc
31
Note that the expansion of F up to fourth order is only justied as long as is small. The result is thus limited to temperatures not too far below Tc . The scaling | | (T Tc )1/2 is characteristic for mean-eld theories. All solutions with this value of | | minimize the free energy. In a given experiment, only one of them is realized. This eqilibrium state has an order parameter = | | ei with some xed phase . This state is clearly not invariant under rotations of the phase. We say that the global U(1) symmetry of the system is spontaneously broken since the free energy F has it but the particular equilibrium state does not. It is called U(1) symmetry since the group U(1) of unitary 1 1 matrices just contains phase factors ei . Specic heat Since we now know the mean-eld free energy as a function of temperature, we can calculate further thermodynamic variables. In particular, the free entropy is S= F . T (6.10)
Since the expression for F used above only includes the contributions of superconductivity or superuidity, the entropy calculated from it will also only contain these contributions. For T Tc , the mean-eld free energy is F ( = 0) = 0 and thus we nd S = 0. For T < Tc we instead obtain ( ) ( ) 1 2 2 2 F + S= = = T T 2 T 2 ( )2 ( )2 ( )2 (T Tc )2 = (T Tc ) = (Tc T ) < 0. = T 2 (6.11)
We nd that the entropy is continuous at T = Tc . By denition, this means that the phase transition is continuous, i.e., not of rst order. The heat capacity of the superconductor or superuid is C=T which equals zero for T Tc but is C= S , T (6.12)
( )2 T ( )2 Tc
(6.13)
for T < Tc . Thus the heat capacity has a jump discontinuity of C = (6.14)
at Tc . Adding the other contributions, which are analytic at Tc , the specic heat c := C/V is sketched here:
no
con rmal
tion tribu
Tc
Recall that Landau theory only works close to Tc .
32
6.2
Ginzburg-Landau theory for neutral superuids
To be able to describe also spatially non-uniform situations, Ginzburg and Landau had to go beyond the Landau description for a constant order parameter. To do so, they included gradients. We will rst discuss the simpler case of a superuid of electrically neutral particles (think of He-4). We dene a macroscopic condensate wave function (r), which is essentially given by s (r) averaged over length scales large compared to atomic distances. We expect any spatial changes of (r) to cost energythis is analogous to the energy of domain walls in magnetic systems. In the spirit of Landau theory, Ginzburg and Landau included the simplest term containing gradients of and allowed by symmetry into the free energy [ ] 4 2 (6.15) F [ ] = d3 r | | + | | + ( ) , 2 where we have changed the denitions of and slightly. We require > 0 so that the system does not spontaneously become highly non-uniform. The above expression is a functional of , also called the Landau functional. Calling it a free energy is really an abuse of language since the free energy proper is only the value assumed by F [ ] at its minimum. If we interpret (r) as the (coarse-grained) condensate wavefunction, it is natural to identify the gradient term as a kinetic energy by writing [ ] 2 1 2 4 F [ ] , (6.16) = d 3 r | | + | | + 2 2m i where m is an eective mass of the particles forming the condensate. From F [ ] we can derive a dierential equation for (r) minimizing F . The derivation is very similar to the derivation of the Lagrange equation (of the second kind) from Hamiltons principle S = 0 known from classical mechanics. Here, we start from the extremum principle F = 0. We write (r) = 0 (r) + (r), where 0 (r) is the as yet unknown solution and (r) is a small displacement. Then F [ 0 + ] = F [ 0 ] [ 3 + d r 0 + 0 + 0 0 0 + 0 0 0 + + O(, )2
by parts 2 2
(6.17)
2m
(0 ) +
2m
( ) 0
F [ 0 ] [ 3 + d r 0 + 0 + 0 0 0 + 0 0 0 + O(, )2 .
2m
( 0 )
2
2m
0
2
(6.18)
If 0 (r) minimizes F , the terms linear in and must vanish for any (r), (r). Noting that and are linearly independent, this requires the prefactors of and of to vanish for all r. Thus we conclude that 2m Dropping the subscript and rearranging terms we nd
2 0 + 0 0 0 2
2 0 = 0 + |0 | 0
2m
2 0 = 0 .
(6.19)
2 + + | | = 0. (6.20) 2m This equation is very similar to the time-independent Schrdinger equation but, interestingly, it contains a nonlinear term. We now apply it to nd the variation of (r) close to the surface of a superuid lling the half space x > 0. We impose the boundary condition (x = 0) = 0 and assume that the solution only depends on x. Then
2
2m
(x) + (x) + | (x)| (x) = 0.

2
(6.21)
33
Since all coecients are real, the solution can be chosen real. For x we should obtain the uniform solution lim (x) = . (6.22) x Writing (x) = we obtain
2
f (x)
(6.23)
2m
>0
f (x) + f (x) f (x)3 = 0.
(6.24)
This equation contains a characteristic length with 2 =

2
2m
2m (T
T)
> 0,
(6.25)
which is called the Ginzburg-Landau coherence length. It is not the same quantity as the the Pippard coherence length 0 or . The Ginzburg-Landau has a strong temperature dependence and actually diverges at T = Tc , whereas the Pippard has at most a weak temperature dependence. Microscopic BCS theory reveals how the two quantities are related, though. Equation (6.24) can be solved analytically. It is easy to check that x f (x) = tanh 2 (6.26)
is a solution satisfying the boundary conditions at x = 0 and x . (In the general, three-dimensional case, the solution can only be given in terms of Jacobian elliptic functions.) The one-dimensional tanh solution is sketched here:
(x)
0
Fluctuations for T > Tc
So far, we have only considered the state 0 of the system that minimizes the Landau functional F [ ]. This is the mean-eld state. At nonzero temperatures, the system will uctuate about 0 . For a bulk system we have (r, t) = 0 + (r, t) (6.27)
with uniform 0 . We consider the cases T > Tc and T < Tc separately. For T > Tc , the mean-eld solution is just 0 = 0. F [ ] = F [ ] gives the energy of excitations, we can thus write the partition function Z as a sum over all possible states of Boltzmann factors containing this energy, Z = D2 eF []/kB T . (6.28) The notation D2 expresses that the integral is over uncountably many complex variables, namely the values of (r) for all r. This means that Z is technically a functional integral. The mathematical details go beyond the 34
scope of this course, though. The integral is dicult to evaluate. A common approximation is to restrict F [ ] 4 to second-order terms, which is reasonable for T > Tc since the fourth-order term (/2) | | is not required to stabilize the theory (i.e., to make F for | | ). This is called the Gaussian approximation. It allows Z to be evaluated by Fourier transformation: Into [ ( ) ] 1 3 F [ ] = d r (r) (r) + (r) (r) (6.29) 2m i i we insert 1 ikr (r) = e k , V k [ ] 1 1 3 ikr+ik r d re F [ ] = k k + ( kk ) k k V 2m
kk
(6.30)
which gives
( = k k +
k
V kk
k k 2m k (
2 2
) =
(
k
k + 2m
2 2
k k .
(6.31)
Thus
) ) 2 2 ( k 1 Z d2 k exp + k k = kB T 2m k k )} ( ( ) 2 2 { 1 k 2 = . d k exp + k k kB T 2m
k
(6.32)
The integral is now of Gaussian type (hence Gaussian approximation) and can be evaluated exactly: Z = kB T 2 k2 . + 2m k (6.33)
From this, we can obtain the thermodynamic variables. For example, the heat capacity is C=T S kB T 2F 2 2 ln = T = T k T ln Z k T T = B B 2 k2 . T T 2 T 2 T 2 + 2m k (6.34)
We only retain the term that is singular at Tc . It stems from the temperature dependence of , not from the explicit factors of T . This term is ( ) 2 2 2 k Ccrit ln + = kB T 2 = kB T 2 2 2 k2 T 2m T + 2m k k ( )2 2 ( ) 1 2m = kB T 2 = kB T 2 ( )2 . (6.35) ( ( )2 2 2 k 2 )2 2 k + 2m2 + 2m k k Going over to an integral over k, corresponding to the thermodynamic limit V , we obtain ( )2 1 2m d3 k 2 2 ( ) V Ccrit = kB T ( )2 2 (2 )3 k 2 + 2m2 = 1 kB T 2 (m )3/2 2 1 kB T 2 (m )2 2 ( ) V = ( ) V 4 3 2 2m 2 2 2 1 . T Tc 35 (6.36)
Recall that at the mean-eld level, C just showed a step at Tc . Including uctuations, we obtain a divergence of the form 1/ T Tc for T Tc from above. This is due to superuid uctuations in the normal state. The system notices that superuid states exist at relatively low energies, while the mean-eld state is still normal. We note for later that the derivation has shown that for any k we have two uctuation modes with dispersion k = + k . 2m
2 2
(6.37)
The two modes correspond for example to the real and the imaginary part of k . Since = (T Tc ) > 0, the dispersion has an energy gap .
0 kx
In the language of eld theory, one also says that the superuid above Tc has two degenerate massive modes, with the mass proportional to the energy gap. Fluctuations for T < Tc Below Tc , the situation is a bit more complex due to the Mexican-hat potential: All states with amplitude |0 | = / are equally good mean-eld solutions.
F Im
Re ringshaped minimum
It is plausible that uctuations of the phase have low energies since global changes of the phase do not increase F . To see this, we write = (0 + ) ei , (6.38) where 0 and are now real. The Landau functional becomes [ F [ ] = d3 r (0 + )2 + (0 + )4 2 ) ( )] ( 1 i i i i + . ( ) e + ( + )( ) e ( ) e + ( + )( ) e 0 0 2m i i
(6.39)
36
As above, we keep only terms up to second order in the uctuations (Gaussian approximation), i.e., terms proportional to 2 , , or 2 . We get [ ( ) @ @ 1 @ @ 3 3 2 3 2 @ @ F [, ] d r 2 + 2 + + 3 + @ = 0 0 @@0 2m i i (( ) ] @ @ ) @@ @@ 1 1 @ @ @ + @ 0@ @ + ( 0 ) + ( 0 ) 0 (+ const). (6.40) @ @ 2m i i 2m @@@@ Note that the rst-order terms cancel since we are expanding about a minimum. Furthermore, up to second order there are no terms mixing amplitude uctuations and phase uctuations . We can simplify the expression: ) ] ( [ 2 1 3 2 () F [, ] = d r 2 + 2m i i 2m [( ) ] 2 2 2 k2 k = 2 + k , (6.41) k k 2m 2m k
k >0 >0
in analogy to the case T > Tc . We see that amplitude uctuations are gapped (massive) with an energy gap 2 = (Tc T ). They are not degenerate. Phase uctuations on the other hand are ungapped (massless ) with quadratic dispersion 2 k2 . (6.42) k = 2m The appearance of ungapped so-called Goldstone modes is characteristic for systems with spontaneously broken continuous symmetries. We state without derivation that the heat capacity diverges like C Tc T for T Tc , analogous to the case T > Tc .
6.3
Ginzburg-Landau theory for superconductors
To describe superconductors, we have to take the charge q of the particles forming the condensate into account. We allow for the possibility that q is not the electron charge e. Then there are two additional terms in the Landau functional: The canonical momentum has to be replaced by the kinetic momentum: i where A is the vector potential. The energy density of the magnetic eld, B 2 /8 , has to be included. Thus we obtain the functional F [, A] =
3
q A, c
(6.43)
1 4 d r | | + | | + 2 2m
2
) q A i c
] B2 + . 8
(6.44)
Minimizing this free-energy functional with respect to , we obtain, in analogy to the previous section, 1 2m ( q A i c )2 + + | | = 0.
2
(6.45)
37
To minimize F with respect to A, we write A(r) = A0 (r) + a(r) and inset this into the terms containing A: ([ [ ] ) ( 1 q q) 3 F [, A0 + a] = F [, A0 ] + d r A a 0 2m i c c ) ] ( 1 ( q ) q 1 + a A0 + ( A0 ) ( a) + O(a2 ) 2m c i c 4 ) q = F [, A0 ] + d r + a 2m c i i ] 2 q 1 2 + 2 | | A 0 a + ( B0 ) a + O(a2 ), m c 4
3
([
= a(A0 ) ((A0 )a)
(6.46)
where we have used A0 = B0 and Gauss theorem. At a minimum, the coecient of the linear term must vanish. Dropping the subscript we obtain i With Ampres Law we nd j= c q q2 2 B = i ([ ] ) | | A, 4 2m m c (6.48) q q2 1 2 ([ ] ) + 2 | | A + B = 0. 2m c m c 4 (6.47)
where we have dropped the subscript s of j since we assume that any normal current is negligible. Equations (6.45) and (6.48) are called the Ginzburg-Landau equations. In the limit of uniform (r), Eq. (6.48) simplies to j= This should reproduce the London equation j= which obviously requires e2 ns q 2 | | = . (6.51) m m As noted in Sec. 5.3, based on ux-quantization experiments and on the analogy to Bose-Einstein condensation, it is natural to set q = 2e. For m one might then insert twice the eective electron (band) mass of the metal. However, it turns out to be dicult to measure m independently of the superuid density ns and it is therefore common to set m = 2m 2me . All system-specic properties are thus absorbed into ns = m q 2 | | m e 2 | | 2 = 2 = 2 | | . 2 e m e 2m
2 2 2
q 2 | | A. m c e2 ns A, mc
(6.49)
(6.50)
(6.52)
With this, we can write the penetration depth as mc2 mc2 mc2 = = . = 2 2 4e ns 8e2 ( 8e2 | | ) We have found two characteristic length scales: : penetration of magnetic eld, : variation of superconducting coarse-grained wave function (order parameter).
(6.53)
38
Within the mean-eld theories we have so far employed, both quantities scale as , close to Tc . Thus their dimensionless ratio := ( T ) (T ) (6.55) 1 Tc T (6.54)
is roughly temperature-independent. It is called the Ginzburg-Landau parameter and turns out to be very important for the behaviour of superconductors in an applied magnetic eld. For elemental superconductors, is small compared to one. Fluctuations and the Anderson-Higgs mechanism When discussing uctuations about the mean-eld state of a superconductor, we have to take the coupling to the electromagnetic eld into account. We proceed analogously to the case of a neutral superuid and employ the Gaussian approximation throughout. For T > Tc , we note that the kinetic term 1 2m ( ) q A i c
2
(6.56)
is explicitly of second order in the uctuations = so that electromagnetic eld uctuations appear only in higher orders. Thus to second order, the order-parameter uctuations decouple from the electromagnetic uctuations, [ ] ) 2 2 2 ( B Bk 1 B2 k 3 k F [, B] = d r (r) (r) + = + . (6.57) + + k k 2m i 8 2m 8
k k
The superconducting uctuations consist of two degenerate massive modes with dispersion k = + 2 k 2 /2m , as for the neutral superuid. The case T < Tc is more interesting. Writing, as above, = (0 + ) ei , we obtain to second order [ ( ) @ @ 1 @ @ 2 2 2 3 3 @ @ 2 + 20 + + 30 + F [, , A] = d r @ @ @0 2m i i (( ) @ @ ) @@ 1 1 @@@ @ @ @ + + ( ) + ( 0 ) 0 @ 0 0 @ @ @ 2m i i 2 m @ @ @@ @ @@@ ( ) ( @ ) ( ) @ @ @ q q 1 @@ @ @ @ A 0 + A0 + 2m i @@@@@@ c c i @@ ( q) ( q ) ) 1 ( + ( 0 ) A0 + A0 0 2m c c ] ) ( ( ) q q 1 1 A 0 A 0 + ( A) ( A) (+ const) + 2m c c 8 [ ( ) 1 = d3 r 2 2 + 2m i i ] ( ) ( 2 q 1 q ) A A + ( A ) ( A ) . (6.58) 2m c c 8 Note that the phase of the macroscopic wave function and the vector potential appear in the combination
39
(q/ c)A. Physics is invariant under the gauge transformation A A + , 1 , c eiq/ c , (6.59) (6.60) (6.61)
where is the scalar electric potential and (r, t) is an arbitrary scalar eld. We make use of this gauge invariance by choosing c = . (6.62) q Under this tranformation, we get A A c =: A , q (6.63) (6.64)
= (0 + ) ei (0 + ) ei(+) = 0 + .
The macroscopic wave function becomes purely real (and positive). The Landau functional thus tranforms into ( ) [ 1 F [, A ] d3 r 2 2 + 2m i i ] 2 q 1 ( A ) A + ( A ) ( A ) (6.65) 2m c2 8 (note that A = A). Thus the phase no longer appears in F ; it has been absorbed into the vector potential. Furthermore, dropping the prime, ) ] 2 2 [( q2 1 k k k A Ak + (k Ak ) (k Ak ) F [, A] = 2 + 2m 2m c2 k 8 k ) ] 2 2 [( k q2 k2 1 = 2 + k k A Ak + A Ak (k Ak )(k Ak ) . (6.66) 2m 2m c2 k 8 k 8
k
Obviously, amplitude uctuations decouple from electromagnetic uctuations and behave like for a neutral superuid. We discuss the electromagnetic uctuations further. The term proportional to is due to superconductivity. Without it, we would have the free-eld functional Ffree [A] = 1 [ ] k 2 A k Ak (k Ak )(k Ak ) . 8
k
(6.67)
Decomposing A into longitudinal and transverse components, (k Ak ) + Ak k (k Ak ) Ak = k

=: Ak
(6.68)
=: A k
:= k/k , we obtain with k Ffree [A] = ( )( )] 1 [ k 2 Ak Ak + k 2 A k Ak k Ak k Ak 8 k ] 1 1 [ $ $ $$ $$ = k 2$ Ak Ak + k 2 A k 2$ Ak Ak = k 2 A k Ak $ k Ak . $ 8 8

k k
(6.69)
Thus only the transverse components appearonly they are degrees of freedom of the free electromagnetic eld. They do not have an energy gap. 40
Including the superconducting contribution, we get ] [ q2 q2 1 2 F [A] = A Ak A Ak + k Ak Ak . 2m c2 k 2m c2 k 8

k
(6.70)
All three components of A appear now, the longitudinal one has been introduced by absorbing the phase (r). Even more importantly, all components obtain a term with a constant coecient, i.e., a mass term. Thus the electromagnetic eld inside a superconductor becomes massive. This is the famous Anderson-Higgs mechanism. The same general idea is also thought to explain the masses of elementary particles, although in a more complicated way. The Higgs bosons in our case are the amplitude-uctuation modes described by . Contrary to what is said in popular discussions, they are not responsible for giving mass to the eld A. Rather, they are left over when the phase uctuations are eaten by the eld A. The mass term in the superconducting case can be thought of as leading to the Meiner eect (nite penetration depth ). Indeed, we can write ) ] 2 2 ( 1 [1 k 2 2 + F [, A] = k k + A Ak + k Ak Ak (k Ak )(k Ak ) . (6.71) 2m 8 2 k
k k
The photon mass is proportional to 1/. (To see that the dispersion relation is 2 2 = m2 c4 + p2 c2 , we would have to consider the full action including a temporal integral over F and terms containing time-derivatives.) Elitzurs theorem One sometimes reads that in the superconducting state gauge symmetry is broken. This is not correct. Gauge symmetry is the invariance under local gauge transformations. S. Elitzur showed in 1975 that a local gauge symmetry cannot be spontaneously broken. Rather, superconductors and superuids spontaneously break a global U(1) symmetry in that the ordered state has a prefered macroscopic phase, as noted above.
6.4
Type-I superconductors
1 < . 2
Superconductors with small Ginzburg-Landau parameter are said to be of type I. The exact condition is (6.72)
It turns out that these superconductors have a uniform state in an applied magnetic eld, at least for simple geometries such as a thin cylinder parallel to the applied eld. The appropriate thermodyamic potential for describing a superconductor in an applied magnetic eld is the Gibbs free energy G (natural variable H) and not the Helmholtz free energy F (natural variable B), which we have used so far. They are related by the Legendre transformation HB G = F d3 r . (6.73) 4 For a type-I superconductor, the equilibrium state in the bulk is uniform (we will see later why this is not a trivial statement). The order parameter is then | | = / and the magnetic induction B as well as the vector potential (in the London gauge) vanish. Thus the Gibbs free-energy density is gs = fs = | | +
2
2 2 2 4 | | = + = . 2 2 2
(6.74)
On the other hand, in the normal state vanishes, but the eld penetrates the system and B H so that gn = fn H2 H2 H2 HB = = . 4 8 4 8 41 (6.75)
The system will only become superconducting if this reduces the Gibbs free energy, i.e., if gs gn = 2 H2 + < 0. 2 8 (6.76)
Thus in an applied magnetic eld of H Hc , with the critical eld 2 Hc (T ) := 4 , superconductivity does not occur. We can use the relation (6.77) and 2 ( T ) = mc2 8e2 | |
2
(6.77)
mc2 8e2
(6.78)
to express the phenomenological parameters and in terms of the measurable quantities and Hc : = 2e2 2 2 Hc , mc2 ( 2 )2 2e 2 = 4 4 Hc . mc2 (6.79) (6.80)
Domain-wall energy and intermediate states We can now calculate the energy per area of a domain wall between superconducting and normal regions. We assume that (r) is only a function of x and impose the boundary conditions { := / for x , (x) (6.81) 0 for x . What are reasonable boundary conditions for the local eld B (x)? In the superconductor we have B (x) 0 for x . (6.82)
At the other end, if we have limx B (x) > Hc , the bulk superconductor has a positive Gibbs free-energy density, as seen above. Thus superconductivity is not stable. For limx B (x) < Hc , the Gibbs free-energy density of the superconductor is negative and it eats up the normal phase. Thus a domain wall can only be stable for 2 B (x) Hc = 4 for x . (6.83) Technically, we have to solve the Ginzburg-Landau equations under these boundary conditions, which can only be done numerically. The qualitative form of the solution is clear, though:
B (x) ~ (x)
0 ~
42
The Gibbs free energy of the domain wall, per unit area, will be denoted by . To derive it, rst note that for the given boundary conditions, gs (x ) = gn (x ), i.e., the superconducting free-energy density deep inside the superconductor equals the normal free-energy density deep inside the normal conductor. This bulk Gibbs free-energy density is, as derived above, gs (x ) = gn (x ) = fn (x ) The additional free-energy density due to the domain wall is ( ) H2 H2 gs (x) c = gs (x) + c . 8 8 The corresponding free energy per area is
2 Hc H2 = c. 8 8
(6.84)
(6.85)
[ ] [ ] 2 2 Hc Hc B (x)Hc + dx gs (x) + = dx fs (x) 8 4 8
[ ( ) 1 q 2 4 dx | | + | | + A(x) 2 2m i x c
] 2 B2 BHc Hc + + . 8 4 8
= (B Hc )2 /8
(6.86)
We can simplify this by multiplying the rst Ginzburg-Landau equation (6.45) by and integrating over x:

1 dx | | + | | + 2m
2 4
)2 ] q A(x) i x c ( ) ( ) ] q q A(x) (x) A(x) i x c i x c (6.87)
by parts
[ 2 4 dx | | + | | +
1 2m (
1 dx | | + | | + 2m
2 4
) 2] q A(x) = 0. i x c
Thus

=
2 Hc 8
[ ] ( )2 ] [ 2 (B Hc )2 Hc B 4 4 = dx 2 | | + 1 dx | | + 2 8 8 Hc

[( dx 1
B Hc
)2
| | 4
] ,
(6.88)
2 where we have drawn out the characteristic energy density Hc /8 . The domain wall energy is given by the dierence of the energy cost of expelling the magnetic eld and the energy gain due to superconducting condensation. For strong type-I superconductors, ; there is a region of thickness > 0 in which the rst term is already large, while the second only slowly approaches its bulk value (see sketch above). Thus > 0 for type-I superconductors. They therefore tend to minimize the total area of domain walls. Note that even in samples of relatively simple shape, magnetic ux will penetrate for non-zero applied eld. It will do so in a way that minimizes the total Gibbs free energy, only one contribution to which is the domain-wall energy. For example in a very large slab perpendicular to the applied eld, the ux has to penetrate in some manner, since going around the large sample would cost much energy. A careful analysis shows that this will usually happen in the form of normal stripes separated by superconducting stripes, see Tinkhams book. Such a state is called an intermediate state. It should not be confused with the vortex state to be discussed later.
43
z x
6.5
Type-II superconductors
1 > . 2
Type-II superconductors are dened by a large Ginzburg-Landau parameter (6.89)
The analysis in the previous section goes through. But now domain walls have a region where the condensate is nearly fully developed but the ux is not completely expelled.
B (x) (x)
0 ~
Therefore, the domain-wall energy is negative, < 0. Hence, the system tends to maximize the total area of domain walls. This tendency should be counterbalanced by some other eect, otherwise the system would become inhomogeneous on microscopic (atomic) length scales and Ginzburg-Landau theory would break down. The eect in question is ux quantizationthe penetrating magnetic ux cannot be split into portions smaller than the ux quantum 0 = hc/2e, see Sec. 5.3. Fluxoid quantization We revisit the quantization of magnetic ux. Consider an arbitrary closed path S forming the edge of a surface S , where at least the edge S must lie inside a superconductor,
44
The magnetic ux through this loop is Stokes = da B = da ( A) =

S S S
d s A.
(6.90)
With the second Ginzburg-Landau equation j=i we obtain =

S
q q2 2 | | A ([ ] ) 2m m c }
2
(6.91)
{ ds
m c j q 2 | |
2
+i
c 2q | |
([ ] )
=
S
} { c mc ds 2 j + i (2i) e ns 2q (6.92)
mc = e
S
c d s vs 2e
S
ds ,
where vs := j/ens is the superuid velocity. The last term contains the charge of the phase of the macroscopic wave function along the path, which must be an integer multiple of 2 for (r) to be continuous: ds = 2n,
S
n Z.
(6.93)
(The minus sign is conventional.) Thus we nd for the so-called uxoid := mc e

S
d s vs =
c hc 2n = n = n 0 , 2e 2e
n Z.
(6.94)
We see that it is not the ux but the uxoid that is quantized. However, deep inside a superconducting region the current vanishes and equals . Vortices The smallest amount of uxoid that can penetrate a superconductor is one ux quantum 0 . It does so as a vortex (or vortex line ).
B vortex core
Following the above arguments, the phase of changes by 2 as one circles a vortex in the positive direction, where, by convention, the direction of a vortex is the direction of the magnetic eld B. Thus in the center of the vortex line (the vortex core ), the phase is undened. This is only consistent with a continuous if = 0 in the 45
vortex core. For a vortex along the z -axis we choose cylindrical coordinates , , z . We then have to solve the Ginzburg-Landau equations with the boundary conditions ( = 0) = 0, | ( )| = 0 , B( ) = 0. From symmetry, we have (r) = | | () ei = | | () ei , vs (r) = vs (), B(r) = z B () and we can choose A(r) = A () where B () = 1 A(). (6.101) (6.102) (6.98) (6.99) (6.100) (6.95) (6.96) (6.97)
Choosing a circular integration path of radius r centered at the vortex core, the enclosed uxoid is (r) = (r) 0 mc vs (r) = . A(r) e 2r mc e
S
ds vs = 2rA(r)
mc ! 2r vs (r) = 0 e (6.103)
This relation follows from uxoid quantization and thus ultimately from the second Ginzburg-Landau equation. To obtain j(r), (r), and A(r), one also has to solve the rst Ginzburg-Landau equation 1 4m and Ampres Law j= ( i + 2e A c )2 + + | | = 0 c B. 4
2 2
(6.104)
(6.105)
This cannot be done analytically because of the non-linear term | | . For small distances from the core, one can drop this term, thereby linearizing the Ginzburg-Landau equation. The solution is still complicated, the result is that | | increases linearly in . Numerical integration of the full equations gives the results sketched here for a cut through the vortex core:
~ (r)
B (r) 0 ~ x
46
Another useful quantity is the free energy per unit length of a vortex line (its line tension ). An analytical expression can be obtained in the strong type-II case of 1. We only give the result here: The vortex line tension is ( )2 0 v ln 4 H2 = c 4 2 ln . (6.106) 8 We can now calculate the eld for which the rst vortex enters the superconductor, the so-called lower critical eld Hc1 . This happens when the Gibbs free energy for a superconductor without vortices (Meiner phase) equals the Gibbs free energy in the presence of a single vortex. We assume the sample to have cross section A and thickness L, parallel to the applied eld.
B A L
Then we have the condition 0 = Gone vortex Gno vortex 1 =& F + L d3 r H B & F & & s v s 4 Hc1 d3 r B (r) = Lv 4 Hc1 L = Lv 0 . 4 Thus Hc1 = The line tension in Eq. (6.106) can also be written as v so that 0 Hc H ln ln = 0 c 4 2 4 2 Hc ln Hc1 = . 2 (6.109) 4v . 0
(6.107)
(6.108)
(6.110)
Recall that this expression only holds for 1. Hc is the thermodynamic critical eld derived above. In a type-II superconductor, nothing interesting happens at H = Hc . The Abrikosov vortex lattice We have considered the structure of an isolated vortex line. How does a nite magnetic ux penetrate a type-II superconductor? Based on Ginzburg-Landau theory, A. A. Abrikosov proposed in 1957 that ux should enter as a periodic lattice of parallel vortex lines carrying a single ux quantum each. He proposed a square lattice, which was due to a small mistake. The lowest-free-energy state is actually a triangular lattice.
47
j x
As noted above, the magnetic ux starts to penetrate the superconductor at the lower critical eld Hc1 . Furthermore, since ux expulsion in type-II superconductors need not be perfect, they can withstand stronger magnetic elds than type-I superconductors, up to an upper critical eld Hc2 > Hc , Hc1 .
H Hc2 (T ) normal metal Shubnikov (vortex) phase Hc1 (T ) Meiner phase
2nd ord e r
We will now review the basic ideas of Abrikosovs approach. Abrikosovs results are quantitatively valid only close to Hc2 since he assumed the magnetic ux density B to be uniform, which is valid for l, where l= 0 B (6.111)
is the typical distance between vortices (B/0 is the two-dimensional concentration of vortex lines). For B = Bz = const = H z we can choose the gauge A = y Hx. Then the rst Ginzburg-Landau equation becomes (note m = 2m) ( )2 1 2eH 2 + y x + + | | = 0. (6.113) 2m i c Slightly below Hc2 , | | is expected to be small (this should be checked!) so that we can neglect the non-linear term. Introducing the cyclotron frequency of a superconductor, c := we obtain (
2
) 1 2 2 i c x + m c x (r) = (r). 2m y 2
2 >0
This looks very much like a Schrdinger equation and from the derivation it has to be the Schrdinger equation for a particle of mass m and charge q = 2e in a uniform magnetic eld H . This well-known problem is solved by the ansatz (x, y ) = eiky y f (x). (6.116)
48
d 2n
2eH , m c
ord er
Tc
(6.112)
(6.114)
(6.115)
We obtain
) 1 2 2 iky y e i c x + m c x e f (x) 2m y 2 2 2 2 ky 1 d2 f 2 2 f i c x iky f + m c = 2 + x f 2m dx 2m ( 2 ) 2 2 2 ky d2 f ky 1 2 2 = 2 + m c x + 2 + f = f. 2 2m dx 2 m c (m )2 c

iky y 2 2
(6.117)
Dening x0 := ky /m c , we get
2
2m
1 d2 f 2 + m c (x x0 )2 f = f. dx2 2
(6.118)
This is the Schrdinger equation for a one-dimensional harmonic oscillator with shifted minimum. Thus we obtain solutions fn (x) as shifted harmonic-oscillator eigenfunctions for ( ) 1 = c n + , n = 0, 1, 2, . . . (6.119) 2 Solution can only exist if = (T Tc ) = (Tc T ) Keeping T Tc xed, a solution can thus only exist for H Hc2 Hc2 = Using 2 = 2 /2m , we obtain Hc2 (T ) = c 2e 2 (T ) m c . e 0 . 2 2 (T ) eH c = . 2 m c with the upper critical eld (6.120)
(6.121)
(6.122)
Note that since 1/ Tc T close to Tc , Hc2 (T ) sets in linearly, as shown in the above sketch. Hc2 should be compared to the thermodynamic critical eld Hc , ( )2 2 2 mc2 1 c 1 1 hc 1 2 Hc = 2 2 = = 2 2 2 2 2 2e 8e 8 2e 2 Hc2 Hc2 0 Hc = = = 2 2 2 2 Hc2 = 2 Hc . (6.123) For = 1/ 2, Hc2 equals Hc . This is the transition between a type-II and a type-I superconductor. So far, our considerations have not told us what the state for H Hc2 actually looks like. To nd out, one in principle has to solve the full, not the linearized, Ginzburg-Landau equation. (We have seen that the linearized equation is equivalent to the Schrdinger equation for a particle in a uniform magnetic eld. The solutions are known: The eigenfunctions have uniform amplitude and the eigenenergies form discrete Landau levels, very dierent from what is observed in the Shubnikov phase.) Abrikosov did this approximately using a variational approach. He used linear combinations of solutions of the linearized equation as the variational ansatz and assumed that the solution is periodic in two dimensions, up to a plane wave. We call the periods in the x- and y -directions ax and ay , respectively. The function n (x, y ) = eiky y fn (x) has the period ay in y if ky = 2 q, ay 49 q Z. (6.125) (6.124)
Then the harmonic oscillator is centered at x0 = 2 0 q= q. m c ay Hay (6.126)
Since the lowest-energy solution is obtained for n = 0 (the ground state of the harmonic oscillator), Abrikosov only considered the n = 0 solutions ( ]2 ) ( ) ( ) [ 1 m c 2y 2y 0 0 (x, y ) = exp iq f0 (x) = C exp iq exp x+ q . (6.127) ay ay 2 Hay
normalization x0
In the Gauss function we nd the quantity m c as long as H Hc2 . Thus 2y 0 (x, y ) = C exp iq ay ( ) =
2 c c = = 2 (T ), 2eH 2e Hc2 2m
(6.128)
[ ]2 ) 1 0 exp 2 x + q . 2 Hay
(6.129)
This is a set of functions enumerated by q Z. Abrikosov considered linear combinations ( ( ) [ ]2 ) 2y 1 0 (x, y ) = Cq exp iq exp 2 x + q . ay 2 Hay q =
(6.130)
To be periodic in x with period ax , up to a plane wave (the corresponding discussion in Tinkhams book is not fully correct), this ansatz has to satisfy ( ) [ ]2 ) ( 1 0 2y (x + ax , y ) = exp 2 x + ax + q Cq exp iq ay 2 Hay q = ( ( ) [ ( )]2 ) 2y 1 0 Hay = Cq exp iq exp 2 x + q+ ax ay 2 Hay 0 q = (x, y ) This requires Hax ay =: N. 0 (Note that this quantity is positive.) Then ( ) [ ]2 ) ( 2y 1 0 (x + ax , y ) = Cq exp iq exp 2 x + (q + ) ay 2 Hay q = ( ( ( ]2 ) ) ) [ 2y 2y 1 0 = exp i Cq exp iq exp 2 x + q . ay q= ay 2 Hay This equals (x, y ) up to a plane-wave factor if Cq = Cq q, (6.134) (6.132)
!
x, y.
(6.131)
(6.133)
i.e., if Cq is periodic. Abrikosov considered the case = 1, which leads to a square lattice. The lowest-free-energy solution is obtained for = 2 and C1 = iC0 , which gives a triangular lattice, sketched above. Note that = Hax ay 0 50 (6.135)
has a simple interpretation: It is the number of ux quanta passing through a rectangular unit cell of the vortex lattice. The vortex lattice is a rather complex system: It is a lattice of interacting lines with a line tension v . At non-zero temperatures, the vortices uctuate, which can lead to the melting of the vortex lattice. The resulting vortex liquid can be pictured as a pot of boiling spaghetti, with the constraint that the vortex lines must either terminate at the surface or form closed loops (!). Moving vortices lead to ohmic resistance, even though most of the sample is still superconducting. To complicate matters, interaction of vortices with defects (pinning) plays an important role. There is an extensive research literature on vortex matter, which we cannot review here.
H
51
Superuid and superconducting lms

Generally, uctuations are stronger in systems of lower dimension. Indeed, they change the properties of twodimensional superuid and superconducting lms qualitatively compared to three-dimensional samples. We will consider such lms within the Ginzburg-Landau theory.
7.1
Superuid lms
We start from the two-dimensional Landau functional [ ] 2 4 2 F [ ] = d r | | + | | + ( ) . 2
(7.1)
We consider temperatures T < Tc . As shown in Sec. 6.2, uctuations of the amplitude | | then have an energy gap, whereas uctuations of the phase are ungapped. Not too close to Tc , phase uctuations will thus dominate and we negelect amplitude uctuations, writing (r) = 0 ei(r) with 0 = /. (7.2) Thus, up to an irrelevant constant, F [ ] = ( ) ( ) d r () = k 2 k k .
2 k >0
(7.3)
We can now calculate the correlation function 2 2 1 (r) (0) = 0 ei(r) ei(0) = 0 D ei(r)+i(0) eF []/kB T . Z
(7.4)
52
This is a Gaussian average since F [] is bilinear in k . Thus we can write

2 (r) (0) = 0 1 (i)n ((r) (0))n n ! n=0 1 (i)n 1 3 5 (n 1) ((r) (0))2 n! n=0
2 = 0
n even 1 3 5 (2m 1) n=2m 2 = 0 (1)m ((r) 1 2 3 (2m) m=0 1 2 = 0 ((r) (0))2 (1)m 2 4 (2m) m=0 = 2m m!
(0))2
( ) ( )m 1 1 1 2 2 exp ((r) (0))2 . ((r) (0))2 = 0 = 0 m! 2 2 m=0 Herein, we have 1 ikr ((r) (0))2 = (e 1)(eik r 1) k k V kk 1 1 ikr kB T 1 = (e 1)(eik r 1) kk V 2 k 2 kk kB T 1 1 1 (2 2 cos k r) = . V 2 k 2
k
(7.5)
(7.6)
Going over to the continuum, we obtain kB T ((r) (0))2 = d2 k 1 (1 cos k r) (2 )2 k 2 2 kB T 1 1 = d dk (1 cos(kr cos )) (2 )2 k
0 0
kB T 1 2
1 dk (1 J0 (kr) ). k
Bessel function
(7.7)
The k -integral is cut o at , which is the inverse of some microscopic length scale, = 1/r0 . The idea is that physics at shorter length scales r < r0 have been integrated out to obtain the Landau functional. For large distances, r r0 , we approximate, rather brutally, { 1 for kr < 1 J0 (kr) (7.8) 0 for kr 1 and obtain ((r) (0))
2
kB T 1 = 2 (
1 /r0
dk kB T 1 r = ln k 2 r0 ) =
2 0
(7.9)
1/r
and thus
2 (r) (0) exp = 0
1 kB T 1 r ln 2 2 r0 53
r r0
) (7.10)
with =
1 kB T > 0. 4
(7.11)
Thus the correlation function of the order parameters decays like a power law of distance in two dimensions. We do not nd long-range order, which would imply limr (r) (0) = const. This agrees with the Mermin-Wagner theorem, which forbids long-range order for the two-dimensional superuid. The power-law decay characterizes so-called quasi-long-range order (short range order would have an even faster, e.g., exponential, decay). Isolated vortices We have argued that uctuations in the amplitude are less important because they have an energy gap proportional to 2 > 0. This is indeed true for small amplitude uctuations. However, there exist variations of the amplitude that are, while energetically costly, very stable once they have been created. These are vortices. In two dimensions, a vortex is a zero-dimensional object; the order parameter goes to zero at a single point at its center. The simplest form of a vortex at the origin can be represented by (r) = | (r)| ei(r) = | (r)| ei(0 ) , where r and are (planar) polar coordinates of r. An antivortex would be described by (r) = | (r)| ei(0 ) . (7.13) (7.12)
In both cases, limr0 | (r)| = 0. Note that we have changed the convention for the sign in the exponent compared to superconducting vortices. In the presence of vortices, the phase (r) of the order parameter is multivalued and, of course, undened at the vortex centers. On the other hand, the phase gradient v = is single-valued (but still undened at the vortex centers). For any closed loop C not touching any vortex cores, we have ds v = total change in phase along C = 2 NC ,
C
(7.14)
where NC Z is the enclosed winding number or vorticity. The vorticity can be written as the sum of the vorticities Ni = 1 of all vortices and antivortices inside the loop, Ni . (7.15) NC =
i
N1 = +1 N3 = 1
N2 = +1
Dening the vortex concentration by nv (r) :=
Ni (r Ri ),
(7.16)
where Ri is the location of the center of vortex i, we obtain v vy vx = 2 nv (r). x y 54 (7.17)
Note that in two dimensions the curl is a scalar. Now it is always possible to decompose a vector eld into a rotation-free and a divergence-free component, v = vph + vv with vph = 0, vv = 0 . (7.19) (7.20) (7.18)
Since the vortex concentation associated with vph vanishes, the component vph does not contain any vortices. Alternatively, note that the rst equation implies that there exists a single-valued scalar eld (r) so that vph = . (7.21)
is a single-valued component of the phase, which cannot be due to vortices. This is the contribution from small phase uctuations, which we have already discussed. Conversely, vv is the vortex part, for which vv = 2 nv , vv = 0. (7.22) (7.23)
This already suggests an electrodynamical analogy, but to formulate this analogy it is advantageous to rescale and rotate the eld vv : vv (r). E(r) := 2 z (7.24)
>0
Then the energy density far from vortex cores, where | | = w = Also, we nd
/ , is ( 2 )1 EE= 1 E E. 2 (7.25)
(v ) v = vv vv =
E = 2 ( vv ) = 2 2 nv E = 0.
(7.26)
and (7.27) These equations reproduce the fundamental equations of electrostatics if we identify the charge density with v = 2 nv . (7.28) (The factor in Gauss law is 2 instead of 4 since we are considering a two-dimensional system.) We can now derive the pseudo-electric eld E(r) for a single vortex, da E = 2r E = 2 2 (7.29) 2 E (r) = (7.30) r and thus E(r) = 2 r 55
r.
(7.31)
From this, we obtain the energy of a single vortex, 1 1 1 1 E1 = Ecore + d2 r E E = Ecore 2 d2 r 2 = Ecore 2 dr . 2 2 r r
2
(7.32)
Now we note that the derivation does not hold for small distances from the vortex center since there | | is not close to / . Thus we cut o the radial integral at the lower end at some vortex core radius r0 and put all energy contributions from the core into Ecore . r0 is on the order of the coherence length . But the integral still diverges; if our system has a characteristic size of L, we obtain E1 = Ecore 2 L
r0 >0
dr L = Ecore 2 ln . r r0
(7.33)
Thus the energy of a single, isolated vortex diverges logarithmically with the system size. This suggests that isolated vortices will never be present as thermal uctuations as long as < 0. This is not true, though. The probability of such vortices should be ( ) L 1 2 1 ln p1 2 eE1 /kB T = 2 eEcore /kB T exp r0 r0 kB T r0 ( ) k2T ( ) 21 L B L 1 1 = 2 eEcore /kB T = 2 eEcore /kB T . (7.34) r0 r0 r0 r0 The typical area per vortex is 1/p1 and the total number of vortices will be, on average, L2 Nv = = L2 p1 = eEcore /kB T 1/p1 ( L r0 )2 21 . (7.35)
For > 1/4, Nv diverges in the thermodynamic limit so that innitely many vortices are present. For < 1/4, Nv 0 for L and according to our argument, which is essentially due to Kosterlitz and Thouless, there are no vortices. It is plausible and indeed true that free vortices destroy quasi-long-range order and in this sense also superuidity. Note that 1 kB T ! 1 = = (7.36) 4 (T ) 4 is an equation for a critical temperature for the appearance of free vortices. We thus nd that the critical temperature in superuid lms should be reduced from the point where = 0 ( = ) to the one where = 1/4 due to vortices appearing as uctuations of the order parameter. While qualitatively true, our description is still incomplete, though, since we have so far neglected interactions between vortices. Vortex interaction The energy of two vortices with vorticities 1 can easily be obtained from the electrostatic analogy. We assume that core regions do not overlap, i.e., the separation is R 2r0 . The pseudo-electric eld of the two vortices, /2, is assumed to be located at R/2 = R x r R/2 r + R/2 E(r) = 2 2 2 |r R/2| |r + R/2|2 2 2 |r + R/2| (r R/2) |r R/2| (r + R/2) . (7.37) = 2 2 2 |r R/2| |r + R/2|
56
Thus the energy is E2 = 2Ecore + 1 EE 2 ( )2 2 2 1 |r + R/2| (r R/2) |r R/2| (r + R/2) 2 = 2Ecore 2 d r 2 2 2 |r R/2| |r + R/2| d2 r

= 2Ecore 2
4
dx
0
dy
2 4 2 2 2
|r + R/2| |r R/2| + |r R/2| |r + R/2| 2 |r + R/2| |r R/2| (r2 R2 /4) |r R/2| |r + R/2|
4 4
(7.38)
We introduce elliptic coordinates , according to x= R , 2 R 2 y= 1 1 2, 2 (7.39) (7.40)
where [1, [, [1, 1]. Then ( )2 2 2 R d d E2 = 2Ecore 2 2 2 1 1 2 ( R )4 ( )2 ( )4 ( )2 ( + )4 R ( )2 + R ( )4 R ( + )2 2 2 2 2 ( R )4 ( R )4 ( )4 2 ( + )4 2 ( )2 ( )2 ( )2 2 2 2 R ( + )2 R ( )2 R ( + ( 2 1)(1 2 ) 1) 2 2 2 = 2Ecore 2 1 ( + )2 + ( )2 2( 2 + 2 2) d d 2 2 2 1 1 2 1 4 = 2Ecore 2 d d 2 2 2 2 1 1 1 = 2Ecore 8 d . ( 2 1)
(7.41)
We have to keep in mind that the integrals in real space have a lower cuto r0 . The minimum separation from /2 is ( 1)R/2. For this separation to equal r0 , the lower cuto for must be the vortex at R x 0 = 1 + With this cuto, we obtain E2 = 2Ecore 8 which for R r0 becomes E2 = 2Ecore 4 1 (R + 2r0 )2 ln , 2 4r0 (R + r0 ) R ln . 4r0 (7.43) 2r0 . R (7.42)
(7.44)
We absorb an R-independent constant into Ecore and nally obtain E2 2Ecore + Vint (R) = 2Ecore 4
>0
R ln . r0
(7.45)
57
Note that the energy of a vortex-antivortex pair remains nite for large system sizes, L . Also, the vortexantivortex interaction Vint (R) increases with R, i.e., vortex and antivortex attract each other. Conversely, one can show that vortices with the same vorticity repel each other.
Vint
r0
One can also show that for arbitrary vorticities N1 , N2 Z, the interaction reads Vint (R) = 4 R N1 N2 ln . r0 (7.46)
<0
Since the equations of (pseudo-) electrostatics are linear, the superposition principle applies and we do not have additional 3-, 4-, etc. body interactions. The energy of a system of vortices is thus E= with vint (r) := 4
i
Ecore +
1 Ni Nj vint (|ri rj |) 2
ij,i=j
(7.47)
provided that i Ni = 0. If the total vorticity i Ni does not vanish, the energy diverges logarithmically with the system size, as we have seen. As long as the total vorticity is zero, the energy per vortex is nite and thus we expect a non-zero concentration of vortices for all temperatures T > 0. We now want to understand the consequences of their presence. Berezinskii-Kosterlitz-Thouless theory The vortices in a two-dimensional superuid behave like a Coulomb gas a gas of charged particles with Coulomb interaction, which is logarithmic in 2D. We have seen that the total vorticity (charge) has to vanish. It is therefore possible to group the vortices into vortex-antivortex pairs. We do this using the following simple algorithm: 1. nd the vortex and the antivortex with the smallest separation r, 2. mark this vortex and this antivortex as a pair, 3. repeat these steps for all remaining vortices until none are left.
+ + + +
r ln , r0
(7.48)
58
The energy of an isolated vortex-antivortex pair (we will use the term vortex pair from now on) of size r is E2 (r) = 2Ecore 4 Thus the probability density of such pairs is ) ( ( )2K0 1 2 r 1 E2 (r)/kB T 1 2 r = 4 y0 , p2 (r) = 4 e = 4 y0 exp 2K0 ln r0 r0 r0 r0 r0 where y0 := N0 eEcore /kB T
2 y0
r ln . r0
(7.49)
(7.50)
(7.51) is a vortex-pair fugacity, and (7.52)
is a vortex fugacity (N0 is a constant of order unity) or, more precisely, K0 := 1 2 > 0 kB T
is a dimensionless measure for the interaction strength in units of the thermal energy. K0 is called the stiness. The crucial idea is now that smaller pairs are polarized in the pseudo-electric eld of the vortex and antivortex forming a given pair. This leads to the screening of the vortex-antivortex interaction and thus reduces the energy of large pairs. Formally, this can be described by renormalization-group (RG) theory (Kosterlitz 1974), which we will summarize in the following. The grand-canonical partition function of the vortex-antivortex system is 2 2 1 d r 1 | r r | d r 2N i j 1 exp 2 Ni Nj K0 ln , (7.53) Z0 = y 2N 2 2 (N !)2 0 r0 r0 2 r0
N D1 D2N i=j
where we have already implemented the constraint that the number of vortices, N , equals the number of antivortices. The ranges of integration, Di , comprise the two-dimensional space R2 excluding disks of radius r0 centered at all vortices (and antivortices) with numbers j < i. This means that the minimum separation is r0 . The idea of RG theory is to perform the integrals for the smallest pairs of sizes between r0 and r0 + dr and rewrite the result (approximately) in a form identical to Z0 but with changed (renormalized) parameters. Physically, we thereby omit the smallest pairs and take their eect into account by renormalizing the parameters. In this way, the partition function, the fugacity, and the stiness become functions of the smallest length scale, which we now denote by r. These running quantities are written as Z , y , and K , respectively. Thus 2 2 1 | r r | d r d r 1 i j 1 2N Z= 2 Ni Nj K ln . (7.54) y 2N exp (N !)2 r2 r2 2 r
N D1 D2N i=j
We now perform the integrals over the smallest separations between r and r + dr. A crucial assumption is that this only involves vortex-antivortex pairs, not pairs of equal vorticity. This is plausible since a vortex and an antivortex attract each other. The integration starts by splitting the integrals, d2 r2N d2 r1 d2 r2N = d2 r1
D1 D2N
D1
1 Ni ,Nj 2
i=j
D2 N
d2 r1
D1
d2 r2N
D2 N
d2 rj
i,j D
d2 ri .
r |ri rj |<r +dr
(7.55)
excluding i,j i,j is the full R2 excluding The Di correspond to the Di but with the minimum separation increased to r + dr. D disks of radius r around all vortices n = i, j . Inserting this into the expression for Z , we obtain the integral, from
59
the second term, d2 rj

i,j D r |ri rj |<r +dr
d2 ri exp 2
n=i,j
Nn Nj K ln
|rn rj | r
|rn ri | |rj ri | 2 Nn Nj K ln 2 K ln r r n=i,j 2 ( r r ) ( r r ) r m j n j d2 rj 2r dr 1 + 2 K 2 Nm Nn = 2 2 | r r | | r r | m j n j m,n=i,j

i,j D
= 2r dr v + 2 K 2 Nm Nn r2
m,n=i,j
d2 rj
(7.56)
i,j D
(rm rj ) (rn rj ) 2 2 . |rm rj | |rn rj |
Herein, we have
d2 rj
i,j D
(rm rj ) (rn rj ) |rm rj | |rn rj |

2 2
= 2 ln
|rm rn | L 2 (1 mn ) ln . r r
(7.57)
The rst term diverges for L but drops out when the sum over m, n is performed, due to overall vanishing vorticity. The result is |rm rn | 3 2 2 = 2r dr N N ln v 2 K r m n r
m,n=i,j m=n
(7.58)
and with the sum over i, j : | r r | 1 m n (N 1)2 Nm Nn ln = 2r dr N 2 V 2 3 K 2 r2 , 2 r

i=j m=n
(7.59)
neglecting some terms of order N 0 . Inserting everything into Z , we obtain two terms: The rst corresponds to Z with Di replaced by Di and the second reads 1 ( y )2N 2 d r1 d2 r2N 2r dr (N !)2 r2
N
D1 D2 N
( N V 2 K r
2 3 2 2
2N 2 integrals
m=n
|rm rn | (N 1) Nm Nn ln r
2
) exp
1 |rm rn | 2K Ni Nj ln 2 r
i=j (2N 2)(2N 3) terms
) . (7.60)
We rename the summation index N as N + 1 in this second term. Then both terms contain 2N integrals under the sum over N . We also put all terms containing dr into the exponent using 1 + a dr = ea dr . We obtain ( ( ) ) y 2 1 ( y )2 N 2 Z = exp 2 2 r dr V d r d2 r2N 1 r (N !)2 r2
N
) ( | r r | dr 1 i j Ni Nj ln . 2K + 8 4 y 2 K 2 exp 2 r r
i=j
D1
D2 N
(7.61)
60
Next, we have to express Z in terms of the new length scale r := r + dr. This is only relevant in expressions not already linear in dr. This applies to, on the one hand, 1 + 2 dr 1 1 r = = , r2 (r dr)2 (r )2 and, on the other, 1 1 exp 2K Ni Nj ln r = exp 2K Ni Nj ln(r dr) 2 2 i=j i=j 1 1 dr = exp 2K Ni Nj ln r exp 2K Ni Nj 2 2 r i=j i=j )2N ( 1 dr = exp , 2K Ni Nj ln r 1 K 2 r
i=j
(7.62)
(7.63)
neglecting terms of order N 0 compared to N . The renormalized partition function is nally [ ( ] )2 ]2N 1 ( y )2N [ y dr Z = exp 2 r dr V 1 + (2 K ) (r )2 (N !)2 (r )2 r
N =: Zpair

D1
d2 r1
D2 N
d2 r2N
) ( dr 1 | r r | i j 4 2 2 . 2K + 8 y K Ni Nj ln exp 2 r r
i=j
(7.64)
Here, Zpair is the partition function of the small pair we have integrated out. It is irrelevant for the renormalization of y and K . Apart from this factor, Z is identical to Z if we set [ ] dr y (r ) = 1 + (2 K (r)) y (r) (7.65) r dr (7.66) K (r ) = K (r) 4 3 y 2 K 2 (r) . r Introducing the logarithmic length scale l := ln we obtain the Kosterlitz RG ow equations dy = (2 K ) y, dl dK = 4 3 y 2 K 2 . dl The initial conditions are y (l = 0) = y0 = eEcore /kB T , 1 K (l = 0) = K0 = 2 , kB T i.e., the parameters assume their bare values at r = r0 (l = 0). 61 (7.70) (7.71) (7.68) (7.69) r r0 dl = dr , r (7.67)
We will now discuss the physics encoded by the RG ow equations. First, note that the quantity C := 2 2 y 2 2 ln K K (7.72)
is invariant under the RG ow: dC dy 2 dK 1 dK = 4 2 y = 4 2 y 2 (2 K ) 8 2 y 2 + 4 3 y 2 K = 0. (7.73) 2 dl dl K dl K dl Thus C is a rst integral of the ow equations. We can calculate C from the initial values y0 , K0 and obtain ( ) 2 1 1 K 2 2 2 2 2 y = 2 y0 + + ln (7.74) K K0 K0 ( ) 1 1 1 1 K 2 y = y0 + 3 + 2 ln . (7.75) K K0 2 K0 The RG ow is along curves decribed by this expression, where the curves are specied by y0 , K0 . These parameters change with temperature as given in Eqs. (7.70) and (7.71). These initial conditions are sketched as a dashed line in the gure. We see that there are two distinct cases:
ix ratr epa
T = Tc
2/
For T < Tc , K ows to some nite value K (l ) > 2/ . This means that even innitely large pairs feel a logarithmic attraction, i.e., are bound. Moreover, the fugacity y ows to zero, y (l ) = 0. Thus large pairs are very rare, which is consistent with their (logarithmically) diverging energy. For T > Tc , K ows to K (l ) = 0. Thus the interaction between a vortex and an antivortex that are far apart is completely screened. Large pairs become unbound. Also, y diverges on large length scales, which means that these unbound vortices proliferate. This divergence is an artifact of keeping only the leading order in y in the derivation. It is cut o at nite y if we count vortex-antivortex pairs consistently. But the limit K 0 remains valid. At T = Tc we thus nd a phase transition at which vortex-antivortex pairs unbind, forming free vortices. It is called the Berezinskii-Kosterlitz-Thouless (BKT) transition. In two-dimensional lms, vortex interactions thus suppress c the temperature where free vortices appear and quasi-long-range order is lost from the point T = Tsingle vortex where 1 kB T 1 1 ! 1 2 = = = K0 = (7.76) 4 2 K0 4 to the one where K (l ) = 2/ and y0 , K0 lie on the separatrix between the two phases,
2 C = 2 2 y0
2 2 ! ln K0 = 0 1 ln K0
2 2 2 + ln K0 = 1 + ln + 2 2 y0 . K0
(7.77)
Clearly the two criteria agree if y0 = 0. This makes sense since for y0 = 0 there are no vortex pairs to screen the interaction. In addition, a third temperature scale is given by the mean-eld transition temperature TMF , where ! = 0. 62
quasilongrange order 0 K( l Tc oo ) = 2/
free vortices no condensate

c Tsingle vortex
TMF = 0
K 0 = 2/
In the low-temperature phase, the largest pairs determine the decay of the correlation function (r) (0) for large r. Thus we nd ( ) r 2 (r) (0) = 0 (7.78) r0 with 1 1 . (7.79) = 2 K (l ) We note that the exponent changes with temperature (one could say that the whole low-temperature phase is critical) but assumes a universal value at Tc : There, 2 1 (Tc ) = . (7.80) 4 Due to the screening of the vortex interaction, the condensate is less sti than it would be in the absence of vortices. This is described by the renormalization of K from K0 to K (l ). It is customary but somewhat misleading to express this as a renormalization of the superuid density ns , which is after all proportional to K0 . Assuming that the temperature dependence of and thus of the bare superuid density 2 n0 K0 (7.81) s 0 =
l
lim K =
is negligible close to Tc , the renormalized superuid density ns (T ) := K 0 n K0 s (7.82)
obtains its temperature dependence exclusively from K/K0 . (The argument l is implied here and in the following.) For T < Tc but close to the BKT transition we have K=
2 The invariant C = 2 2 y0 2 K0
2 + K.
(7.83)
ln K0 is an analytic function of temperature and its value at Tc is [ ] 2 2 Cc = 2 2 y 2 ln K = 1 ln . K T =Tc
(7.84)
Thus we can write, close to Tc ,
2 C = 1 ln + b (T Tc ) with some constant b := dC/dT |T =Tc . On the other hand, [ ] 2 C = 2 2 y 2 ln K K (T ) 2 2 ) ln = (2 + K + K ( ) 2 1 ln ln 1 + K = 1 + 2 K 2 & 2 2 & 1 2 K K 2 ln K+ K 2 = 1 + & & 2 4 & 2 2 4 & 2 2 K 2 . = 1 ln 8 63
(7.85)
(7.86)
Thus (K/K0 ) n0 s
2 K 2 = b ( T Tc ) 8
2 K 2b Tc T . =
const
(7.87)
Consequently, ns =
jumps to a nite value at Tc and then increases like a square root.
ns
Tc
This behavior was indeed measured in tortion-pendulum experiments on He-4 lms (Bishop and Reppy, 1980).
7.2
Superconducting lms
In this section we concentrate on what is dierent for charged superconductors compared to neutral superuids. Recall that for a superuid the gradient term in the Landau functional reads d2 r ( ) . (7.88) If we move around a vortex once, the phase has to change by 2 , regardless of the distance from the vortex core. This phase change leads to an unavoidable contribution to the gradient term of ( ) 1 i 1 i 2 d r ( ) = dr d r e e r r ( ) 2 1 1 dr = dr d = dr d = 2 , (7.89) r r r which diverges logarithmically with system size. On the other hand, for a superconducting lm the gradient term reads ) 2 ( 1 q d2 r A . (7.90) 2m i c Again, the phase of winds by 2 around a vortex, but the associated gradient can, in principle, be compensated by the vector potential. Pearl (1964) showed that this indeed leads to a nite energy of a single vortex in a superconducting lm. In addition, the free energy contains the magnetic-eld term B 2 (r) d3 r . (7.91) 8 Note that the integral is three-dimensional the eld is present in all space, also outside of the lm. We rst note that the lm has a new eective length scale: In the London gauge, Ampres law reads A = ( A) 2 A =
=0
4 j c
2 A =
4 j. c
(7.92)
On the other hand, from the London equation we have j= c A, 42 64 (7.93)
2 which is valid within the lm and away from vortex cores so that | | = / . However, the current is conned to the thin lm. We can write j(r) = K(x, y ) (z ) (7.94)
with a surface current density K. Thus

dz j(r) =
lm
K(x, y ) =
( c ) dz A(r). 42
(7.95)
If the thickness is d and A is approximately constant across the thickness, we have K(x, y ) = and nally d A (z ). (7.98) 2 This result exhibits the new length scale that controls the spatial variation of A and thus of the current j, 2 A = := 2 , d (7.99) c d A(x, y, 0) 4 2 c d A(r) (z ) j(r) = 4 2 (7.96) (7.97)
which is large, , for a thin lm. We see that in thin lms, assumes the role of the penetration depth . Since is large for thin lms, one could say that thin lms are always eectively of type II. We here do not discuss the full derivation of Pearl but only consider the far eld for large r and show that it does not lead to a diverging free energy for a single vortex. We assume that most of the magnetic ux of 0 penetrates the lm for = x2 + y 2 . Then the magnetic eld above (and below) the lm looks like a monopole eld for r .
This eld is easy to obtain from symmetry: d a B = 0

half space for z>0
2r2 B = 0 0 r for z > 0. 2r2
(7.100)
B (r ) = By symmetry,
0 2r2
and
B(r) =
(7.101)
B(r) = sgn z
0 r. 2r2
(7.102)
65
This gives a contribution to the eld energy of B2 2 0 d r = 8 32 3

3
1 2 0 d r 4 = 4 r 32 3
3
dr 2 0 1 = , 2 r 8 2
(7.103)
which is nite for an innity lm. The value of the lower cuto does not matter for this. The cuto has to be present, since the monopole eld is not a valid approximation for small r. We next obtain the current from Ampres law in integral form:
r d
4 dr B = da j c r +r 4 0 2 dr = r d j (r) 2 2 (r ) c r ( ) 4 0 1 1 0 r r d j (r) = = c r r + r r2 0 1 j (r) = 4 2 r2 d and with the vector character restored j(r) = Note that the sheet current is thus K(r) = 0 . 2 4 r2 d 0 . 2 4 r2
(7.104)
(7.105)
(7.106) (7.107)
(7.108)
(7.109)
For large r we have | | = 0 = / (note that typically ). Then the second Ginzburg-Landau equation gives ( q q2 2 q ) 2 2 q j = i (i i)0 0 A = 0 A . (7.110) 2m m c m c Thus we can rewrite the gradient term in the free energy as ( 2 2 1 q )2 m 2 2 0 (m ) d d2 r A = d d r j j = d d 2 r 2 2 j j. (7.111) 0 4 2 2m c 2m 0 q 2q 0 The contribution from large r is dm 2 2q 2 0 m 2 2 1 0 d r 04 4 2 = 2 d 2 16 r d 32 4 q 2 0
2
m 2 dr 1 0 = 2 d 2 , 3 r 8 3 q 2 0
(7.112)
which is also nite for an innite lm. We conclude that the free energy of an isolated vortex is nite in a superconducting lm. It is then plausible and indeed true that the interaction energy of a vortex-antivortex pair does not diverge for large separations r but saturates for r . Since for the far eld of a single vortex the magnetic-eld energy, Eq. (7.103), dominates over the energy due to the gradient term, we expect the large-r interaction to be dominated by the Coulomb-type
66
attraction of the magnetic monopoles in the upper and lower half spaces. This supposition is borne out by a proper analysis. Consequently, for large r the interaction behaves like ( Vint = const (0 /2 is the monopole strength, according to Eq. (7.102)). 0 2 )2 1 r (7.113)
N S
All this showes that, strictly speaking, there will be a non-zero concentration of free vortices at any temperature T > 0. Thus there is no quasi-long-range order. However, the relevant length scale is = 2 /d, which can be very large for thin lms, even compared to the lateral size L of the sample. In this case the large-r limit is experimentally irrelevant. But for vortex separations r , the magnetic-eld expulsion on the scale r is very weak since is the eective penetration depth. Then the fact that the condensate is charged is irrelevant and we obtain the same logarithmic interaction as for a neutral superuid. Thus for thin lms of typical size we can use the previously discussed BKT theory. For superconducting lms we even have the advantage of an additional observable, namely the voltage for given current. We give a hand-waving derivation of V (I ). The idea is that a current exerts a Magnus force on a vortex, in the direction perpendicular to the current. The force is opposite for vortices and antivortices and is thus able to break vortexantivortex pairs. As noted above, free vortices lead to dissipation. A vortex moving through the sample in the orthogonal direction between source and drain contacts leads to a change of the phase dierence by 2 . We will see in the chapter on Josephson eects why this corresponds to a non-zero voltage. Since free vortices act independently, it is plausible to assume that the resistance is R nv , (7.114)
where nv now denotes the concentration of free vortices. To nd it, note that the total potential energy due to vortex-antivortex attraction and Magnus force can be written as V = Vint 2FMagnus r with Vint = 2 kB T K ln r . r0 (7.116) (7.115)
67
Vint
rbarrier
There is a nite barrier for vortex-antivortex unbinding at a separation rbarrier determined from V /r = 0. This gives 2 kB T Kbarrier rbarrier = , (7.117) 2FMagnus where Kbarrier := K (rbarrier ). The barrier height is E := V (rbarrier ) V (r0 ) = V (rbarrier ) rbarrier = 2 kB T Kbarrier ln 2FMagnus rbarrier r0 ( ) rbarrier = 2 kB T Kbarrier ln 1 . r0 For small currents we have rbarrier r0 and thus rbarrier E . = 2 kB T K (l ) ln r0
K
(7.118)
(7.119)
The rate at which free vortices are generated is Rgen e E = The recombination rate of two vortices to form a pair is Rrec n2 v, since two vortices must meet. In the stationary state we have Rgen = Rrec and thus a resistance of R nv Since the Magnus force is FMagnus I we have rbarrier so that 1 1 FMagnus I (7.124) nv ( Rgen ( rbarrier r0 )K (7.122) (7.121) ( rbarrier r0 )2K . (7.120)
rbarrier r0
)K . (7.123)
( )K 1 R = I K . I 68
(7.125)
Finally, the voltage measured for a current I is V = RI I K I = I 1+K , where K K (l ) is the renormalized stiness. Since K= we nd for the exponent ) 2 2( + K 1 + 2b Tc T = 1 + K = 3 + 2 2b Tc T
= const
(7.126)
(7.127)
(7.128)
for T
Tc .
1+ K
3 2 1
ohmic Tc T
Above Tc we have K = 0 and thus ohmic resistance, V I , as expected. Below Tc , the voltage is sub-ohmic, i.e., the voltage is nite for nite current but rises more slowly than linearly for small currents. This behavior has been observed for thin superconducting lms.
69
Origin of attractive interaction

In the following chapters we turn to the microsopic theory of superconductivity. While the BCS theory is reasonably easy to understand if one assumes an attractive interaction between electrons in a superconductor, it is far from obvious where such an interaction should come from. The only fundamental interaction that is relevant for (non-radioactive) solids is the electromagnetic one, which naively gives a repulsive interaction of a typical strength of several eV between nearby electrons. How can this lead to an attraction at the low energy scale kB Tc 1 meV? We will see that the lattice of ion cores (nuclei with tightly bound inner electrons) plays an important role. We will use Feynman diagrams for Green functions to describe the physics. Unfortunately, we do not have the time to introduce these concepts rigorously; this is done in many good textbooks on many-particle physics as well as in the lecture notes on Vielteilchentheorie (in German), which are available online. For those familiar with Feynman diagrams, they rigorously represent mathematical expressions, for the others they should at least be useful as cartoons of the relevant processes.
8.1
Reminder on Green functions
Nevertheless, we start by briey summarizing some properties of Green functions. In many-particle physics, we usefully express the Hamiltonian in terms of an electronic eld operator (r, t), where = , is the electron spin. The eld operator can be expanded in any convenient basis of single-particle states characterized by wavefunctions (r), where represents all relevant quantum numbers, (r, t) = (r) c (t), (8.1) (r, t) =
(r) c (t).
(8.2)
c and c are annihilation and creation operators of electrons in the single-particle state, respectively. Note that we are working in the Heisenberg picture, i.e., the wavefunctions (r) are independent of time, whereas the time dependence is carried by the operators. For example, the Hamiltonian for free electrons reads 2 2 ( r ) H = d3 r (r). (8.3) 2m
70
In this case it is useful to expand into plane waves, 1 ikr (r) = e ck V k 1 H= d3 r eikr+ik r c k V =
k kk 2 2
(8.4)
2
(k )2 ck 2m (8.5)
k c ck . 2m k
Two types of Green functions are dened as follows: greater Green function : G> (rt, r t ) := i (r, t) (r , t ) ,
(8.6)
where . . . is the equilibrium average A = Tr Aeq = Tr AeH 1 Tr AeH , H Tr e Z (8.7)
where = 1/kB T is the inverse temperature and H is the many-particle Hamiltonian. For non-interacting electrons at temperature T EF /kB , eq describes the Fermi sea. G> describes the conditional probability amplitude for an electron created at point r with spin at time t to be found at point r with spin at time t. lesser Green function : G< (rt, r t ) := +i (r , t ) (r, t) ; (8.8)
it describes the propagation of a hole from time t to time t . These are not the most useful denitions since in quantum theory propagation of an electron forward in time cannot be separated from propagation of a hole backward in time. It is also useful to distiguish between the cases t > t and t < t . This is accomplished by these denitions: retarded Green function : GR (rt, r t ) := i (t t ) { } (r, t), ( r , t ) , (8.9)
where {A, B } := AB + BA is the anti-commutator (appropriate for fermions) and the step function { 1 for t > t (8.10) (t t ) = 0 for t < t selects only contributions with t > t (forward in time). advanced Green function : GA (rt, r t ) := +i (t t) { } (r, t), , (r , t ) (8.11)
this Green function analogously contains only contributions backward in time. The thermal averages in the Green functions introduce operators eH , whereas the time evolution of operators introduces time-evolution operators according to A(t) = eiHt/ A eiHt/ . (8.12)
71
Specically, we have GR (rt, r t ) =i (t t )
iH (t t)/ + eiHt / (r) eiHt/ (r ) e
( 1 iHt / Tr eH eiHt/ (r) eiH (tt )/ (r ) e Z )
(8.13)
In practice, we will want to write the Hamiltonian in the form H = H0 + V and treat V in perturbation theory. It is not surprising that this is complicated due to the presence of H in several exponential factors. However, these factors are of a similar form, only in some the prefactor of H is imaginary and in one it is the real inverse temperature. Can one simplify calculations by making all prefactors real? This is indeed possible by formally replacing t i , t i , which is the main idea behind the imaginary-time formalism. We cannot discuss it here but only state a few relevant results. It turns out to be useful to consider the Matsubara (or thermal) Green function G (rt, r t ) := T (r, t) ( r , t ) , (8.14) where for any operator and T is a time-ordering directive : A( ) := eH / A eH / . { A( )B ( ) for > T A( )B ( ) = B ( )A( ) for <
(8.15)
(8.16)
(upper/lower sign for bosonic/fermionic operators). For time-independent Hamiltonians, the Green function only depends on the dierence . One can then show that the resulting Green function G (r, r , ) is dened only for [ , ] and satises G (r, r , + ) = G (r, r , ) (2n + 1) , n Z. (8.17)
for fermions. This implies that the Fourier transform is a discrete sum over the fermionic Matsubara frequencies n := (8.18)
The imaginary-time formalism is useful mainly because G is easier to obtain or approximate than the other Green functions and these can be calculated from G based on the following theorem: The retarded Green function GR ( ) in Fourier space is obtained from G (in ) by means of replacing in by + i0+ , where i0+ is an innitesimal positive imaginary part, GR ( ) = G (in + i0+ ). (8.19) This is called analytic continuation. Analogously, GA ( ) = G (in i0+ ). (8.20)
G (i n)
in + i 0+
R G ( ) A G ( )
in
i 0+
72
It will be useful to write the electronic Green function in k space. In the cases we are interested in, momentum and also spin are conserved so that the Green function can be written as Gk ( ) = T ck ( ) c (8.21) k (0) . Analogously, the bosonic Matsubara Green function of phonons can be written as Dq ( ) = T bq ( ) b q (0) , where enumerates the three polarizations of acoustic phonons. We also note that for bosons we have D( + ) = +D( ), (8.23)
(8.22)
i.e., the opposite sign compared to fermions. Therefore, in the Fourier transform only the bosonic Matsubara frequencies 2n n := , nZ (8.24) occur.
8.2
Coulomb interaction
We now discuss the eect of the electron-electron Coulomb interaction. We will mainly do that at the level of Feynman diagrams but it should be kept in mind that these represent mathematical expressions that can be evaluated if needed. The Coulomb interaction can be written in second-quantized form as 1 Vint = d3 r1 d3 r2 (8.25) 1 (r1 )2 (r2 ) VC (|r2 r1 |) 2 (r2 )1 (r1 ) 2
1 2
with VC ( r ) = In momentum space we have
e2 . r
(8.26)
1 ikr (r) = e ck V k

(8.27)
so that Vint = = where 1 2V 2 1 2V 2
d3 r1 d3 r2 eikr1 ik r2 +ik d3 R d3 ei(kk +k

r2 +ik r1 ck1 c k 2
VC (|r2 r1 |) ck 2 ck 1 VC () ck 2 ck 1 , (8.28)
kk k k 1 2
+k )R i(kk +k k )/2 e ck1 c k 2
kk k k 1 2
R=
r1 + r2 , 2 = r2 r1 .
(8.29) (8.30)
We can now perform the integral over R: 1 d3 k+k k k ,0 ei(kk +k k )/2 c Vint = k1 ck 2 VC () ck 2 ck 1 2V kk k k 1 2 1 = d3 ei(k k ) VC () c k1 ck 2 ck 2 ck+k k ,1 . 2V
kk k
1 2
(8.31)
73
Substituting new momentum variables k1 = k + k k , k2 = k , q = k k , we obtain Vint = with VC (q) = 1 VC (q) c k1 q,1 ck2 +q,2 ck2 2 ck1 1 2V
k1 k2 q
1 2
(8.32) (8.33) (8.34)
(8.35)
d3 eiq VC ().
(8.36)
The interaction Vint is a sum over all processes in which two electrons come in with momenta k1 and k2 , a momentum of q is transfered from one to the other through the Coulomb interaction, and the electrons y out with momenta k1 q and k2 + q.
k1 q,1
Vc (q)
k2 + q,2
k11
k 2 2
VC (q) is obviously the Fourier transform of the Coulomb interaction. It is most easily obtained by Fourier transforming the Poisson equation for a point charge, 2 (r) = 4 (r) = 4Q (r) d3 r eiqr 2 (r) = 4Q by parts d3 r (iq)2 eiqr (r) = 4Q q 2 (q) = 4Q Q (q) = 4 2 q so that VC (q) = 4 Screening and RPA The bare Coulomb interaction VC is strongly repulsive, as noted above. But if one (indirectly) measures the interaction between charges in a metal, one does not nd VC but a reduced interaction. First of all, there will be a dielectric function from the polarizability of the ion cores, VC (q) 4 e2 , q2 (8.43) e2 . q2 (8.37) (8.38) (8.39) (8.40) (8.41)
(8.42)
but this only leads to a quantitative change, not a qualitative one. We absorb the factor 1/ into e2 from now on. More importantly, a test charge in a metal is screened by a cloud of opposite charge so that from far away full the eective charge is strongly reduced. Diagramatically, the eective Coulomb interaction VC (q, in ) is given by the sum of all connected diagrams with two external legs that represent the Coulomb interaction. With the representations VC 74 (8.44)
(the minus sign is conventional) and

full VC
(8.45)
as well as G0 for the bare electronic Green function one would nd for the non-interacting Hamiltonian H0 , we obtain = + + + (8.46)
(8.47)
This is an expansion in powers of e2 since VC contributes a factor of e2 . We have exhibited all diagrams up to order e6 . If we try to evaluate this sum term by term, we accounter a problem: At each vertex , momentum and energy (frequency) must be conserved. Thus in partial diagrams of the form
k k q=0
the Coulomb interaction carries momentum q = 0. But VC (0) = 4 e2 02 (8.48)
is innite. However, one can show that the closed G loop corresponds to the average electron density so that the diagram signies the Coulomb interaction with the average electronic charge density (i.e., the Hartree energy ). But this is compensated by the average charge density of the nuclei. Thus we can omit all diagrams containing the tadpole diagram shown above. Since we still cannot evaluate the sum in closed form we need an approximation. We rst consider the limiting full cases of small and large q in VC (q, in ). For large q, corresponding to small distances, the rst diagram is proportional to 1/q 2 , whereas all the others are at least of order 1/q 4 and are thus suppressed. We should recover the bare Coulomb interaction for large q or small distances, which is plausible since the polarization of the electron gas cannot eciently screen the interaction between two test charges that are close together. For small q we nd that higher-order terms contain higher and higher powers of 1/q 2 and thus become very large. This is alarming. The central idea of our approximation is to keep only the dominant term (diagram) at each order in e2 . The dominant term is the one with the highest power in 1/q 2 . Only the VC lines forming the backbone of the diagrams (drawn horizontally) carry the external momentum q due to momentum conservation at the vertices. Thus the dominant terms are the ones with all VC lines in the backbone:
q k q k q
k+q
k+ q
75
Summing up these dominant terms we obtain the approximate eective interaction

RPA VC :=
+ + (8.49)
+ +
This is called the random phase approximation (RPA) for historical reasons that do not concern us here, or the Thomas-Fermi approximation. The most important part of RPA diagrams is clearly the bubble diagram 0 which stands for 0 (q, in ) = = 1 1 0 0 Gk+q, (in + in ) Gk (in ) i V k n 2 1 1
in k
(8.50)
in + in k+q in k
(8.51)
(the factor of 2 is due to the spin). We can write the diagramatic series also as
RPA VC (q, in ) = VC (q) + VC (q) 0 (q, in ) VC (q) VC (q) 0 (q, in ) VC (q) 0 (q, in ) VC (q) + [ ] = VC (q) 1 0 (q, in ) VC (q) + 0 (q, in ) VC (q) 0 (q, in ) VC (q) . (8.52)
This is a geometric series, which we can sum up,

RPA VC (q, in ) =
VC ( q ) . 1 + VC (q) 0 (q, in )
(8.53)
0 can be evaluated, it is essentially 1 times the susceptibility of the free electron gas. We cannot explain this here, but it is plausible that the susceptibility, which controls the electric polarization of the electron gas, should RPA enter into a calculation of the screened Coulomb interaction. Note that VC has a frequency dependence since 0 (or the susceptibility) has one. In the static limit i 0 + i0+ and at low temperatures T EF /kB , one can show that 0 (q, 0) = const = N (EF ), (8.54)
where N (EF ) is the electronic density of states at the Fermi energy, including a factor of two for the spin. Thus we obtain
RPA VC (q)
e 4 q 2 e 1 + 4 q 2 0
2
= 4
e2 q 2 + 4e2 0
e2 q 2 + 4e2 N (EF ) e2 = 4 2 q + 2 s = 4 with s := 4e2 N (EF ).
(8.55)
(8.56)
76
Note that summing up the more and more strongly diverging terms has led to a regular result in the limit q 0. The result can be Fourier-transformed to give es r (8.57) r (not too much confusion should result from using the same symbol e for the elementary charge and the base of the exponential function). This is the Yukawa potential, which is exponentially suppressed beyond the screening length 1/s . This length is on the order of 108 m = 10 nm in typical metals. While we thus nd a strong suppression of the repulsive interaction at large distances, there is no sign of it becoming attractive. On the other hand, for very high frequencies EF / the electron gas cannot follow the perturbation, the susceptibility and 0 go to zero, and we obtain the bare interaction,
RPA VC (r) = e2
e2 RPA VC (q, ) = VC (q) = 4 2 . q
(8.58)
8.3
Electron-phonon interaction
The nuclei (or ion cores) in a crystal oscillate about their equilibrium positions. The quanta of these lattice vibrations are the phonons. The many-particle Hamiltonian including the phonons has the form H = Hel + Hph + Hel-ph , where we have discussed the electronic part Hel before, ( ) 1 q b b + Hph = q q 2
q
(8.59)
(8.60)
is the bare Hamiltonian of phonons with dispersion q , and ( ) 1 Hel-ph = gq c k+q, ck bq + bq, V
k q
(8.61)
describes the electron-phonon coupling. gq is the coupling strength. Physically, an electron can absorb (bq ) or emit (b q, ) a phonon under conservation of momentum. Hence, electrons can interact with one another by exchanging phonons. Diagrammatically, we draw the simplest possible process as
k1 q,1
0 Dq
k2+ q,2 k 22
k11
In detail, we dene, in analogy to the Coulomb interaction VC (q) the interaction due to phonon exchange, ,
(8.62)
1 2 0 |gq | Dq . (in ) V We quote the expression for the bare phonon Green function, i.e., the one obtained from Hph alone: Vph (q, , in )
0 Dq (in ) =
(8.63)
1 2 q 1 + = . in q in q (in )2 2 q
(8.64)
The phonon-mediated interaction is thus frequency-dependent, whereas the Coulomb interaction is static. However, we could write the Coulomb interaction in a very similar form as the exchange of photons. Since the speed of light is so much larger than the speed of sound, we can neglect the dynamics for photon exchange but not for phonon exchange. 77
Jellium phonons For our discussions we need a specic model for phonons. We use the simplest one, based on the jellium approximation for the nuclei (or ion cores). In this approximation we describe the nuclei by a smooth positive charge density + (r, t). In equilibrium, this charge density is uniform, + (r, t) = 0 + . We consider small deviations + (r, t) = 0 + + + (r, t). Gauss law reads E = 4 + (r, t), (8.66) since 0 + is compensated by the average electronic charge density. The density of force acting on + is f = + E = 0 + E to leading order in + . Thus (8.67) f = 4 0 + + . The conservation of charge is expressed by the continuity equation + + + v = 0. t
= j+
(8.65)
(8.68)
To leading order this reads + + 0 + v = 0 t 2 Newton f + v = 0 , = 0 + + t2 t m (8.69) (8.70)
where m is the mass density of the nuclei. With the nuclear charge Ze and mass M we obtain Ze Ze 2 + f = 4 0 = + + , 2 t M M which is solved by with = + (r, t) = + (r) eit Ze 0 4 = M + 4 Z 2 e2 0 n+ , M (8.71)
(8.72)
(8.73)
where n0 + is the concentration of nuclei (or ions). We thus obtain optical phonons with completely at dispersion, i.e., we nd the same frequency for all vibrations. One can also obtain the coupling strength gq . It is clear that it will be controlled by the Coulomb interaction between electrons and uctuations + in the jellium charge density. We refer to the lecture notes on many-particle theory and only give the result: 1 2 |gq | = VC (q). (8.74) V 2 Consequently, the electron-electron interaction due to phonon exchange becomes Vph (q, in ) = 1 2 2 2 0 |gq | Dq (in ) = VC (q) = V ( q ) . C V 2 (in )2 2 (in )2 2 (8.75)
It is thus proportional to the bare Coulomb interaction, with an additional frequency-dependent factor. The retarded form reads
R Vph (q, ) = Vph (q, in + i0+ ) = VC (q)
2 2 = V ( q ) , C ( + i0+ )2 2 2 2 + i0+ sgn
(8.76)
where we have used 2 i0+ = i0+ sgn and have neglected the square of innitesimal quantities. 78
8.4
Eective interaction between electrons
Combining the bare Coulomb interaction and the bare interaction due to phonon exchange, calculated for the jellium model, we obtain the bare eective interaction between electrons, Ve (q, in ) := VC (q) + Vph (q, in ) = VC (q) + VC (q) = VC (q) The retarded form is
R Ve (q, ) = Ve (q, in + i0+ )
2 (in )2 2 (8.77)
(in )2 . (in )2 2
= VC (q)
2 . 2 2 + i0+ sgn
(8.78)
R This expression is real except at = and has a pole there. Moreover, Ve is proportional to VC with a negative prefactor as long as < .
R Veff
V c( q ) 0
The eective interaction is thus attractive for < . The exchange of phonons overcompensates the repulsive Coulomb interaction. On the other hand, for 0, the eective interaction vanishes. This means that in a quasi-static situation the electrons do not see each other at all. What happens physically is that the electrons polarize the (jellium) charge density of the nuclei. The nuclei have a high inertial mass, their reaction to a perturbation has a typical time scale of 1/ or a frequency scale of . For processes slow compared to , the nuclei can completely screen the electron charge, forming a polaron, which is charge-neutral. For frequencies > 0, we have to think in terms of the response of the system to a test electron oscillating with frequency . The jellium acts as an oscillator with eigenfrequency . At the present level of approximation it is an undamped oscillator. The jellium oscillator is excited at the frequency . For 0 < < , it is driven below its eigenfrequency and thus oscillates in phase with the test electron. The amplitude, i.e., the jellium polarization, is enhanced compared to the = 0 limit simply because the system is closer to the resonance at . Therefore, the oscillating electron charge is overscreened. On the other hand, for > the jellium oscillator is driven above its eigenfrequency and thus follows the test electron with a phase dierence of . Thus the electron charge is not screened at all but rather enhanced and the interaction is more strongly repulsive than the pure Coulomb interaction. Screening of the eective interaction From our discussion of the Coulomb interaction we know that the real interaction between two electrons in a metal is strongly screened at all except very short distances. This screening is well described within the RPA. We now apply the RPA to the eective interaction derived above. We dene Ve = 79 + (8.79)
and
RPA Ve
:=
(8.80)
or
RPA Ve (q, in ) = Ve (q, in ) + Ve (q, in ) 0 (q, in ) Ve (q, in ) Ve (q, in ) 0 (q, in ) Ve (q, in ) 0 (q, in ) Ve (q, in ) +
(8.81)
As above, we can sum this up,

RPA (q, in ) Ve
Ve (q, in ) = = VC (q) 1 + Ve (q, in )0 (q, in ) 1 + VC (q) = VC (q) )2 2
(in )2 (in )2 2 (in )2 (in )2 2
0 (q, in )
(in )2 (in + (in )2 VC (q)0 (q, in ) VC (q) (in )2 + (in )2 VC (q)0 (q, in ) = 1 + VC (q)0 (q, in ) (in )2 2 + (in )2 VC (q)0 (q, in )
RPA (q,i ) = VC n
RPA = VC (q, in )
(in )2 (in )2
2 1+VC (q)0 (q,in ) (in )2 2 (i ) (in )2 q n
RPA = VC (q, in )
(8.82)
with the renormalized phonon frequency . q (in ) := 1 + VC (q)0 (q, in )

RPA To see that this is a reasonable terminology, compare Ve to the bare eective interaction
(8.83)
Ve (q, in ) = VC (q)
(in )2 . (in )2 2
(8.84)
RPA Evidently, screening leads to the replacements VC VC and q . For small momenta and frequencies, we have 0 N (EF ), the density of states at EF . In this limit we thus obtain q = q. (8.85) = = 2 = 2 2 s s s e 1 + 4 q 1 + q2 2 N (EF ) 2 q
Due to screening we thus nd an acoustic dispersion of jellium phonons. This is of course much more realistic than an optical Einstein mode. Beyond the low-frequency limit it is important that 0 and thus q obtains a sizable imaginary part. It smears RPA out the pole in the retarded interaction Ve (q, ) or rather moves it away from the real-frequency axisthe lattice vibrations are now damped. The real part of the retarded interaction is sketched here for xed q:
80
RPA, R Re Veff
RPA V (q ) c
Re q
Note that the interaction still vanishes in the static limit 0, the interaction is attractive for 0 < < Re q , where Re q q . It is important that the static interaction is not attractive but zero. Hence, we do not expect static bound states of two electrons. To obtain analytical results, it is necessary to simplify the interaction. The main property required for superconductivity is that the interaction is attractive for frequencies below some typical phonon frequency. The typical phonon frequency is the material specic Debye frequency D . We write the eective RPA interaction in terms of the incoming and transferred momenta and frequencies,
RPA RPA Ve = Ve (k, in ; k , in ; q, in ).
(8.86)
k q, i n i n
q, i n
k+ q , i n + i n k, i n
k, i n
We then approximate the interaction (very crudely) by a constant V0 < 0 if both incoming frequencies are smaller than D and by zero otherwise, { V0 for |in | , |in | < D , RPA (8.87) Ve 0 otherwise.
81
Cooper instability and BCS ground state

In this chapter we will rst show that the attractive eective interaction leads to an instability of the normal state, i.e., of the Fermi sea. Then we will dicuss the new state that takes its place.
9.1
Cooper instability
Let us consider the scattering of two electrons due to the eective interaction. A single scattering event is represented by the diagram
k, i n
k+ q , i n + i n
k, i n
Electrons can also scatter multiple times: +
k q, i n i n
(9.1)
An instability occurs if this series diverges since then the scattering becomes innitely strong. In this case the perturbative expansion in the interaction strength V0 represented by the diagrams breaks down. This means that the true equilibrium state cannot be obtained from the equilibrium state for V0 = 0, namely the noninteracting Fermi gas, by perturbation theory. A state that is perturbatively connected to the free Fermi gas is called a Landau Fermi liquid. It is an appropriate description for normal metals. Conversely, a scattering instability signals that the equilibrium state is no longer a Fermi liquid. Like in the RPA, it turns out to be sucient to consider the dominant diagrams at each order. These are the ladder diagrams, which do not contain crossing interaction lines. Moreover, the instability occurs rst for the scattering of two electrons with opposite momentum, frequency, and spin. We thus restrict ourselves to the
82
diagrams describing this situation. We dene the scattering vertex by
k in

p i n
k in k in
:=
p i n p i n k p, i n i n + k in k k1 , i n i 1 n k in k1 i 1 n k1 i 1 n p i n k1 p, + i 1 n i n p i n
(9.2)
k in
p i n
This is a geometric series, which we can sum up,
1 +
(9.3)
1
With our approximation
RPA Ve
{ V0 0
for |in | < D , otherwise, +V0

k1
(9.4)
we obtain (in ) 1 V0
1
1 , |i 1 | < in D n
1 V
0 (i 1 ) G 0 1 Gk n k1 (in ) 1
(9.5)
for |in | < D and zero otherwise. Thus V0 1 1 0 0 1 1 1 , |i 1 | < (in ) 1 V0 in k1 Gk1 (in ) Gk1 (in ) D V n 0 We see that the scattering vertex diverges if V0 1
1 in 1 |in | < D
for |in | < D , otherwise.
(9.6)
1 0 1 0 1 Gk1 (in ) G k1 (in ) = 1. V

k1
(9.7)
This expression depends on temperature. We now evaluate it explicitly: V0 kB T |

1 in 1 in < D
1 1 1 = V0 kB T 1 1 V in k1 in k1
k1
| |
d D( + )
1 in 1 in < D
1 1 )2 + 2 (n
(9.8)
83
with the density of states per spin direction and per unit cell, D(). Assuming the density of states to be approximately constant close to the Fermi energy we get (with = 1) V0 kB T
1 |in | < D 1 in
D(EF )
1 )2 (n
d + 2
1| = /|n
D /2
= V0 kB T D(EF ) 2
1
(2n+1)
kB T D(EF ) = V0
D /2
[ ( )] D . = V0 D(EF ) + ln 4 2
n=0
n=0
1 n+
1 2
(9.9)
In the last step we have used an approximation for the sum over n that is valid for D 1, i.e., if the sum has many terms. Since Tc for superconductors is typically small compared to the Debye temperature D /kB (a few hundred Kelvin), this is justied. 0.577216 is the Euler constant. Altogether, we nd V0 ( ) D 1 V0 D(EF ) + ln 2 (9.10)
for |in | < D . Coming from high temperatures, but still satisfying kB T D , multiple scattering enhances . diverges at T = Tc , where ( ) 2D V0 D(EF ) + ln =1 (9.11) kB Tc 2e D 1 ln = (9.12) kB Tc V0 D(EF ) ( ) 2e 1 kB Tc = D exp . (9.13) V0 D(EF )
1.13387 1
This is the Cooper instability. Its characteristic temperature scale appears to be the Debye temperature of a few hundred Kelvin. This is disturbing since we do not observe an instability at such high temperatures. However, the exponential factor tends to be on the order of 1/100 so that we obtain Tc of a few Kelvin. It is important that kB Tc is not analytic in V0 at V0 = 0 (the function has an essential singularity there). Thus kB Tc cannot be expanded into a Taylor series around the non-interacting limit. This means that we cannot obtain kB Tc in perturbation theory in V0 to any nite order. BCS theory is indeed non-perturbative.
9.2
The BCS ground state
We have seen that the Fermi sea becomes unstable due to the scattering of electrons in states |k, and |k, . Bardeen, Cooper, and Schrieer (BCS) have proposed an ansatz for the new ground state. It is based on the idea that electrons from the states |k, and |k, form (so-called Cooper) pairs and that the ground state is a superposition of states built up of such pairs. The ansatz reads ) ( |BCS = uk + vk c (9.14) k ck, |0 ,
k
84
where |0 is the vacuum state without any electrons and uk , vk are as yet unknown complex coecients. Normalization requires 1 = BCS |BCS ) ( + vk c c = 0| ( u + v c c ) u k , k k k k k k , |0 = 0| = (
k k k
|u k | +
2
u k vk ck ck, 2
+ uk vk ck, ck + |vk |
|0 (9.15)
(
k
|uk | + |vk |
2
) .
2
This is certainly satised if we demand |uk | + |vk | = 1 for all k, which we will do from now on. Note that the occupations of |k, and |k, are maximally correlated; either both are occupied or both are empty. Also, |BCS is peculiar in that it is a superposition of states with dierent total electron numbers. (We could imagine the superconductor to be entangled with a much larger electron reservoir so that the total electron number in superconductor and reservoir is xed.) As a consequence, expressions containing unequal numbers of electronic creation and annihilation operators can have non-vanishing expectation values. For example, c BCS c k ck, k ck, BCS BCS ) ( = 0| ( u + v c c ) c c u + v c c k k k k k , k k k, k k , |0
k
uk 0| vk
(
k =k
|uk | + |vk |
=1
uk . |0 = vk
(9.16)
The coecients uk , vk are chosen so as to minimize the expectation value of the energy, BCS |H | BCS , under 2 2 the constraint |uk | + |vk | = 1 for all k. |BCS is thus a variational ansatz. We write the Hamiltonian as H= k c (9.17) k ck + Vint
k
and, in the spirit of the previous section, choose the simplest non-trivial approximation for Vint that takes into account that 1. only electrons with energies |k | This leads to Vint = with Vkk D relative to the Fermi energy are important and 2. the instability is due to the scattering between electrons in single-particle states |k and |k, . 1 Vkk c k ck, ck , ck N
kk
(9.18)
{ V0 = 0
for |k | < D and |k | < D , otherwise.
(9.19)
We assume that scattering without momentum transfer, k = k, contributes negligibly compared to k = k since there are many more scattering channels for k = k. We also assume that uk = uk , vk = vk , which is, at worst,
85
a restriction of our variational ansatz. Then we obtain ) ( ( ) BCS |H | BCS = k 0| u uq + vq c q + vq cq, cq ck ck q cq , |0 ) ( ( ) 1 + Vkk 0| u + v c c c c c c u + v c c q q q q , |0 q q q, q k k, k k N q q kk 2 2 = k 0| |vk | ck, ck c k 0| |vk | ck ck, c k ck ck ck, |0 + k ck ck, ck |0 1 + Vkk 0| vk uk uk vk ck, ck c k ck, ck , ck ck ck , |0 N kk 1 2 2k |vk | + Vkk vk uk u = k vk =: EBCS . N
k kk k k k q q
(9.20)
This energy should be minimized with respect to the uk , vk . For EBCS to be real, the phases of uk and vk must be the same. But since EBCS is invariant under uk uk eik , vk vk eik , (9.21)
2 we can choose all uk , vk real. The constraint from normalization then reads u2 k + vk = 1 and we can parametrize the coecients by uk = cos k , vk = sin k . (9.22)
Then EBCS = 1 Vkk sin k cos k sin k cos k N k kk 1 Vkk k (1 cos 2k ) + = sin 2k sin 2k . N 4 2k sin2 k +
k kk
(9.23)
We obtain the minimum from EBCS 1 Vqk 1 Vkq = 2q sin 2q + cos 2q sin 2k + sin 2k cos 2q q N 2 N 2 k k 1 ! Vqk cos 2q sin 2k = 0. = 2q sin 2q + N
k
(9.24)
We replace q by k and parametrize k by k sin 2k =: 2 k + 2 k and write k cos 2k = 2 . k + 2 k (9.25)
(9.26)
The last equality is only determined by the previous one up to the sign. We could convince ourselves that the other possible choice does not lead to a lower EBCS . Equation (9.24) now becomes 2 1 k k k k + =0 Vkk 2 2 2 2 2 N k + k k + 2 k k + k k 1 k k = . Vkk 2 N 2 k + 2 k
k
(9.27) (9.28)
86
This is called the BCS gap equation for reasons to be discussed below. When we have solved it, it is easy to obtain the original variational parameters in terms of k , ( ) 1 k 2 uk = 1+ 2 , (9.29) 2 k + 2 k ( ) 1 k 2 vk = 1 2 , (9.30) 2 k + 2 k k uk vk = 2 . 2 k + 2 k (9.31)
The relative sign of uk and vk is thus the sign of k . The absolute sign of, say, uk is irrelevant because of the invariance of EBCS under simultaneous phase rotations of uk , vk (consider a phase factor of ei = 1). For our special interaction { V0 for |k | , |k | < D , Vkk = (9.32) 0 otherwise, the BCS gap equation becomes 1 N k = 0 which can be solved with the ansatz
k , |k | < D
k (V0 ) 2 2 k + 2 k
for |k | < D , (9.33) otherwise,
{ 0 > 0 for |k | < D , k = 0 otherwise.
(9.34)
We obtain 0 = V0 N
k |k | < D
0 2 + 2 2 k 0 = V0 2 + 2 2 k 0 1
D
(9.35)
V0 1= N
k |k | < D
1 d D( + ) , 2 2 + 2 0
(9.36)
where D() is the density of states per spin direction and per unit cell. If the density of states is approximately constant within D of the Fermi energy, we obtain 1 1 = V0 D(EF ) 2
D
d 2 + 2 0
= V0 D(EF ) Arsinh
D 0
(9.37) (9.38) (9.39)
1 D sinh = V0 D(EF ) 0 1 0 = D . sinh V0 D1 (EF ) In the so-called weak-coupling limit of small V0 D(EF ), this result simplies to ( ) 1 0 = 2D exp . V0 D(EF ) 87
(9.40)
Interestingly, apart from a numerical factor, the value of 0 agrees with kB Tc for the Cooper instability. We will return to this observation below. The non-analyticity of the function 0 (V0 ) means that we cannot obtain 0 and thus |BCS within perturbation theory in V0 . We can now nd the energy gain due to the superconducting state, i.e., the condensation energy. For this, we insert uk , vk into EBCS , ( ) 1 1 k k k + . (9.41) 1 2 EBCS = 2 Vkk 2 k 2 2 2 N 2 + + 2 4 k k k k k + k kk k We use the simple form of Vkk and assume that D() is constant within D of the Fermi energy but not outside of this interval. This gives
D D
( d D( + )
EBCS = N
d D( + ) 2 + N
D
1 2 + 2 0
+N
D D D ( ) D D 2 2 2 d D( + ) + N D(EF ) D D + 0 + 0 Arsinh = 2N 0
2 N V0 D2 (EF ) 2 0 Arsinh
D D 2 0 d D ( + ) 0 + N d d D( + )D( + )(V0 ) 2 + 2 4 ( )2 + 2
0 0
D . 0 1 = V0 D(EF ) Arsinh D 0 2 + 2 . D 0
(9.42)
With the gap equation
(9.43)
this simplies to EBCS = 2N
d D( + ) N D(EF ) D
(9.44)
The normal-state energy should be recovered by taking 0 0. The energy dierence is 2 + 2 + N D (E ) 2 . EBCS := EBCS EBCS |0 0 = N D(EF ) D D F D 0 Since for weak coupling we have 0 D , we can expand this in 0 /D , ( )2 0 1 2 2 EBCS = N D(EF ) D 1 + + N D(EF ) D = N D(EF ) 2 0. D 2 The condensation-energy density is thus (counted positively) 1N D(EF ) 2 eBCS = 0. 2V
(9.45)
(9.46)
(9.47)
Type-I superconductivity is destroyed if eBCS equals the energy density required for magnetic-eld expulsion. At 2 H = Hc , this energy is Hc /8 , as we have seen above. We thus conclude that N Hc = 4 D(EF ) 0 . (9.48) V This prediction of BCS theory is in reasonably good agreement with experiments for simple superconductors.
88
10
BCS theory
The variational ansatz of Sec. 9.2 has given us an approximation for the many-particle ground state |BCS . While this is interesting, it does not yet allow predictions of thermodynamic properties, such as the critical temperature. We will now consider superconductors at non-zero temperatures within mean-eld theory, which will also provide a new perspective on the BCS gap equation and on the meaning of k .
10.1
BCS mean-eld theory

(10.1)
We start again from the Hamiltonian 1 Vkk c k c H= k ck, ck , ck . k ck + N

k kk
A mean-eld approximation consists of replacing products of operators A, B according to AB = A B + A B A B . Note that the error introduced by this replacement is AB A B A B + A B = (A A)(B B ), (10.3) (10.2)
the help of a dierent choice, namely A = c k ck, , B = ck , ck . This leads to the mean-eld BCS Hamiltonian ( ) 1 V c c HBCS = k c c c + c c + c c c c c c c . kk k , k k k, k , k k k k k, k , k k k, N kk k (10.4) We dene 1 Vkk ck , ck (10.5) k := N k
c k, ck , . However, Bardeen, Cooper, and Schrieer realized that superconductivity can be understood with
i.e., it is of second order in the deviations of A and B from their averages. A well-known mean-eld approximation is the Hartree or Stoner approximation, which for our Hamiltonian amounts to the choice A = c k ck , B =
so that k =
1 Vkk c k ck , . N
k
(10.6)
At this point it is not obvious that the quantity k is the same as the one introduced in Sec. 9.2 for the special case of the ground state. Since this will turn out to be the case, we nevertheless use the same symbol from the start. We can now write HBCS = k c k c (10.7) k ck, ck k ck k ck, + const.
k k k
89
The constant is irrelevant for the following derivation and is omitted from now on. Since HBCS is bilinear in c, c it describes a non-interacting eective system. But what is unusual is that HBCS contains terms of the form cc and c c , which do not conserve the electron number. We thus expect that the eigenstates of HBCS do not have a sharp electron number. We had already seen that the BCS ground state has this property. This is a bit strange since superpositions of states with dierent electron numbers are never observed. One can formulate the theory of superconductivity in terms of states with xed electron number, but this formulation is cumbersome and we will not pursue it here. To diagonalize HBCS , we introduce new fermionic operators, which are linear combinations of electron creation and annihilation operators, ( ) ( )( ) ck k uk vk = . (10.8) vk uk c k, k, This mapping is called Bogoliubov (or Bogoliubov-Valatin ) transformation. Again, it is not clear yet that uk , vk are related to the previously introduced quantities denoted by the same symbols. For the to satisfy fermionic anticommutation relations, we require { } k , k = k k + k k
= u k uk ck ck uk vk ck ck, vk uk ck, ck + vk vk ck, ck, + uk u k ck ck uk vk ck ck, vk uk ck, ck + vk vk ck, ck, { } { } { } 2 2 c = |uk | ck , c k, , ck, k uk vk {ck , ck, } vk uk ck, , ck + |vk | =1 =0 =0 =1
= |uk | + |vk | = 1. Using this constraint, we nd the inverse transformation, ( ) ( )( ) ck k uk vk = . vk u c k k, k,
2 !
(10.9)
(10.10)
Insertion into HBCS yields )( ) ( )( ) { ( HBCS = k u + v u + v + v + u v + u k k k k, k k k k k, k k k k, k k k k, ( )} )( )( ) ( k + u uk k + vk vk k vk k + uk k, k k, k, k uk k + vk k, ) {( 2 2 = k |uk | k |vk | + k vk uk + k uk vk k k ) ( 2 2 + k |vk | + k |uk | + u v + v u k k k k k k k, k, ( ) 2 2 + k u k k vk + k uk vk + k vk k (uk ) k, ( ) } 2 2 + k vk uk + k vk u k u + ( v ) + const. k k , k k k k The coecients uk , vk should now be chosen such that the and terms vanish. This requires
2 2k u k vk + k vk k (uk ) = 0. 2 k k
(10.11)
(10.12)
Writing k = |k | eik , u k = |u k | e vk = |vk | e 90

ik
(10.13) (10.14) (10.15)
ik
we obtain
( ) 2 2 2k |uk | |vk | ei(k k ) + |k | |vk | ei(2k k ) |uk | ei(k 2k ) = 0.
(10.16)
A special solution of this equation (we do not require the general solution) is given by k = 0, k = k , ( ) 2 2 2k |uk | |vk | + |k | |vk | |uk | = 0. From the last equation we obtain ( ) 2 2 2 4 2 2 4 2 |uk | |vk | = |k | |vk | 2 |vk | |uk | + |uk | 4k ( ) ( ) ( )2 2 2 2 2 4 2 2 4 2 2 2 2 2 4 k + |k | |uk | |vk | = |k | |vk | + 2 |vk | |uk | + |uk | = |k | |vk | + |uk | = | k | so that |uk | |vk | = Together with |uk | + |vk | = 1 we thus nd |u k | =
2 2 2 2 2
(10.17) (10.18) (10.19)
(10.20) (10.21) (10.22)
| k | |uk | |vk | = 2 + | |2 2 k k k 2k |uk | |vk | = . | k | 2 2 k + |k |
(10.23)
k 1 , 1+ 2 2 2 k + |k | 1 k 2 . |vk | = 1 2 2 2 k + |k |
(10.24)
(10.25)
Restoring the phases in Eq. (10.22), we also conclude that k uk vk = . 2 2 2 k + |k | The BCS Hamiltonian now reads, ignoring a constant, 2 ( ) 2 |k | k k + k, HBCS = + k k, 2 + | |2 2 + | |2 k k k k k ( ) 2 + | |2 k + . = k k k , k k,
k
(10.26)
(10.27)
Using k = k and the plausible assumption |k | = |k |, we obtain the simple form HBCS = Ek k k
k
(10.28)
with the dispersion Ek :=
2 + | |2 . k k
(10.29)
91
It is instructive to rst consider the normal state, for which k ( ) { 0 1 k 2 |u k | = 1+ = 2 |k | 1 ( ) { 1 1 k 2 |vk | = 1 = 2 |k | 0
0. Then for k < 0, for k > 0, for k < 0, for k > 0. (10.30)
(10.31)
We see that the Bogoliubov quasiparticles described by , , are holes for energies below the Fermi energy (k < 0) and electrons for energies above (k > 0). Their dispersion is Ek = |k |. For a parabolic normal dispersion k :
Ek
hole excitations 0 kF
electron excitations k
The excitation energies Ek are always positive except at the Fermi surfaceit costs energy to create a hole in the Fermi sea and also to insert an electron into an empty state outside of the Fermi sea. Superconductivity changes the dispersion to Ek =
2 + | | : k k 2
Ek
kF 0 kF k
Superconductivity evidently opens an energy gap of magnitude |kF | in the excitation spectrum. We should recall that in deriving HBCS we have ignored a constant, which we now reinsert, HBCS = EBCS + Ek k k .
k
(10.32)
The energy of the system is EBCS if no quasiparticles are present and is increased (by at least |kF |) if quasiparticles are excited. The state without any quasiparticles is the pure condensate. Since EBCS depends on temperature through ck, ck , the condensate is not generally the BCS ground state discussed previously. However, one can show that it agrees with the ground state in the limit T 0.
92
uk
vk 0
2 2
kF
We also nd 0 < |uk | < 1 and 0 < |vk | < 1, i.e., the Bogoliubov quasiparticles are superpositions of particles and holes. Deep inside the Fermi sea, the quasiparticles are mostly hole-like, while far above EF they are mostly electron-like. But right at the Fermi surface we nd, for example for spin = , 1 1 k = ck eik c k, . 2 2 (10.33)
The quasiparticles here consist of electrons and holes with the same amplitude. This means that they are electrically neutral on average. So far, we have not determined the gap function k . This can be done by inserting the Bogoliubov transformation into the denition 1 k = Vkk ck , ck , (10.34) N
k
which yields k = ( )( ) 1 Vkk vk k uk k + vk + uk k , k , N k } { 1 2 . (10.35) Vkk vk uk k vk k + u2 = k k , k k , k + uk vk k , k , N

k
For selfconsistency, the averages have to be evaluated with the BCS Hamiltonian HBCS . This gives k = nF (Ek ), k k = 0, k , k , k = 0, k , k , = 1 nF (Ek ),
(10.36) (10.37) (10.38) (10.39)
and we obtain the BCS gap equation, now at arbitrary temperature, k = 1 1 k Vkk uk vk [1 2 nF (Ek )] = Vkk [1 2 nF (Ek )] N N 2 + | |2 2 k k k k k 1 Vkk [1 2 nF (Ek )] . N 2Ek
k
(10.40)
We see that nF (Ek ) 0 for Ek > 0 and T 0 so that the zero-temperature BCS gap equation (9.28) is recovered as a limiting case. For our model interaction and assuming a k-independent real gap, we obtain, in analogy to the ground-state derivation, ) ( D D 2 + 2 1 2 nF 0 2 + 2 tanh 0 2 1 = V0 d D( + ) = V . (10.41) d D ( + ) 0 2 + 2 2 2 + 2 2 0 0
D D
93
Note that this only works for 0 = 0 since we have divided by 0 . If the density of states is approximately constant close to EF , the equation simplies to D tanh 2 + 2 0 2 V0 D(EF ) 1= d . (10.42) 2 2 + 2 0
D
The integral is easily evaluated numerically, leading to the temperature dependence of 0 :
0 0 (0)
0
For weak coupling we have already seen that ( 0 (0) = 2D exp
Tc
) .
1 V0 D(EF )
(10.43)
We can also obtain an analytical expression for Tc : If T approaches Tc from below, we can take the limit 0 0 in the gap equation, 1 = V0 D(EF )
D
D by parts
tanh 2 | | d = V0 D(EF ) 2 | | { D D tanh 2 2
tanh 2 d = V0 D(EF ) ln x . cosh2 x }
D /2
dx
0
tanh x x
D /2
V0 D(EF ) ln
dx
0
(10.44)
In the weak-coupling limit we have D = D /kB T 1 (this assertion should be checked a-posteriori ). Since the integrand of the last integral decays exponentially for large x, we can send the upper limit to innity, { } ) ( ln x D D D V0 D(EF ) ln dx 1= tanh + ln , (10.45) V D ( E ) ln = 0 F 2 2 2 4 cosh2 x
=1 0
where is again the Euler gamma constant. This implies ( ) 1 2D exp e = V0 D(EF ) ( ) 2e 1 kB Tc = D exp . V0 D(EF )
(10.46) (10.47)
This is exactly the same expression we have found above for the critical temperature of the Cooper instability. Since the approximations used are quite dierent, this agreement is not trivial. The gap at zero temperature and the critical temperature thus have a universal ratio in BCS theory, 20 (0) 2 = 3.528. kB Tc e (10.48)
This ratio is close to the result measured for simple elementary superconductors. For example, for tin one nds 20 (0)/kB Tc 3.46. For superconductors with stonger coupling, such as mercury, and for unconventional superconductors the agreement is not good, though. 94
10.2
Isotope eect
How can one check that superconductivity is indeed governed by a phonon-mediated interaction? BCS theory predicts ) ( 1 kB Tc , 0 D exp . (10.49) V0 D(EF ) It would be ideal to compare kB Tc or 0 for superconductors that only dier in the Debye frequency D , not in V0 of D(EF ). This is at least approximately possible by using samples containing dierent isotopes (or dierent fractions of isotopes) of the same elements. The eigenfrequency of a harmonic oscillator scales with the mass like 1 . (10.50) m The entire phonon dispersion, and thus in particular the Debye frequency, also scales like 1 1 q , D (10.51) M M with the atomic mass M for an elementary superconductor. The same scaling has been found above for the jellium model, see Eq. (8.73). Consequently, for elementary BCS weak-coupling superconductors, 1 . (10.52) 2 This is indeed found for simple superconductors. The exponent is found to be smaller or even negative for materials that are not in the weak-coupling regime V D(EF ) 1 or that are not phonon-mediated superconductors. In particular, if the relevant interaction has nothing to do with phonons, we expect = 0. This is observed for optimally doped (highest Tc ) cuprate high-temperature superconductors. kB Tc , 0 M with =
10.3
Specic heat
We now discuss further predictions following from BCS theory. We start by revisiting the heat capacity or specic heat. The BCS Hamiltonian HBCS = EBCS + Ek k (10.53) k
k
with Ek = leads to the internal energy
2 + | |2 k k
k
(10.54) (10.55)
U = HBCS = EBCS + 2
Ek nF (Ek ).
However, this is inconvenient for the calculation of the heat capacity C = dU/dT since the condensate energy EBCS depends on temperature through ck, ck . We better consider the entropy, which has no contribution from the condensate. It reads S = kB [(1 nF ) ln(1 nF ) + nF ln nF ], (10.56)
k
where nF nF (Ek ). From the entropy, we obtain the heat capacity C=T dS dS = dT d d dnF [(1 nF ) ln(1 nF ) + nF ln nF ] = 2kB [ ln(1 nF ) 1 + ln nF + 1] = 2kB d d
k k = ln
= 2kB 2
nF 1nF
= ln eEk = Ek
Ek
d nF (Ek ) . d
(10.57)
95
Note that nF (Ek ) depends on or temperature both explicitly and through the temperature dependence of 2 + | (T )|2 : Ek = k k C = 2kB
2
( Ek
=
nF
Ek nF Ek
nF 1 d |k | + Ek 2Ek d
) = 2kB
nF
k
(
2 Ek
Ek
1 d | k | + 2 d
) . (10.58)
The rst term is due to the explicit dependence, i.e., to the change of occupation of quasiparticle states with temperature. The second term results from the temperature dependence of the quasiparticle spectrum and is 2 absent for T Tc , where |k | = 0 = const. The sum over k contains the factor eEk nF 1 = = 2 = nF (Ek )[1 nF (Ek )] Ek Ek eEk + 1 (eEk + 1) so that C = 2kB
2
(10.59)
( nF (1 nF )
2 Ek
1 d |k | + 2 d
) . (10.60)
Here, nF (1 nF ) is exponentially small for Ek kB T . This means that for kB T min , where min is the minimum superconducting gap, all terms in the sum are exponentially suppressed since kB T min Ek . Thus the heat capacity is exponentially small at low temperatures. This result is not specic to superconductorsall systems with an energy gap for excitations show this behavior. For the simple interaction used above, the heat capacity can be obtained in terms of an integral over energy. The numerical evaluation gives the following result:
C C
l norma state
Tc
We nd a downward jump at Tc , reproducing a result obtained from Landau theory in section 6.1. The jump occurs for any mean-eld theory describing a second-order phase transition for a complex order parameter. Since BCS theory is such a theory, it recovers the result. 2 The height of the jump can be found as follows: The Ek term in Eq. (10.60) is continuous for T Tc (0 0). Thus the jump is given by C = 1
2 T3 kB c
nF (k ) [1 nF (k )]
d2 0 d
.
T Tc
(10.61)
To obtain 0 close to Tc , we have to solve the gap equation

D
1 = V0 D(EF )
D
D tanh tanh 2 + 2 2 + 2 0 0 2 2 = V D ( E ) d d 0 F 2 + 2 2 2 + 2 0 0
0
(10.62)
for small 0 . Writing =
1 1 = kB T kB (Tc T )
(10.63)
96
and expanding for small T and small 0 , we obtain 1 = V0 D(EF )

D
d
0
tanh 2kB Tc
V0 D(EF ) + T 2 2kB Tc
d cosh
2 2kB Tc
= 1 (gap equation)
V0 D(EF ) 2 + 0 4kB Tc
d 2
1 cosh2
2k B T c
2kB Tc tanh 2kB Tc
) . (10.64)
In the weak-coupling limit, D kB Tc , we can extend the integrals to innity, which yields 0 = V0 D(EF ) V0 D(EF ) 2 2 T 7 (3) 2 0 T 2 k T = V D ( E ) V D ( E ) B c 0 F 0 F 0 2 2kB Tc 4kB Tc 2kB Tc Tc 8 2 (kB Tc )2 Thus d 2 0 d =
T Tc
7 (3)
(10.65) (10.66)
2 8 k 2 Tc T. 2 0 = 7 (3) B
dT d2 0 d dT
1 8 2 3 3 8 2 C = 2 3 nF (k ) [1 nF (k )] k T = kB N D(EF ) d nF ( ) [1 nF ( )] kB Tc 7 (3) B c 7 (3) k = k B Tc
T 0
2 = kB Tc
d 2 0 d T
T 0
8 2 3 3 k T = 7 (3) B c
(10.67)
8 2 2 = N D(EF ) kB Tc , 7 (3) where N is the number of unit cells. The specic-heat jump is c = C 8 2 2 = d(EF ) kB Tc , V 7 (3)
(10.68)
(10.69)
where d(EF ) = D(EF ) N/V is the density of states per volume.
10.4
Density of states and single-particle tunneling
Detailed experimental information on the excitation spectrum in a superconductor can be obtained from singleparticle tunneling between a superconductor and either a normal metal or another superconductor. We discuss this in the following assuming, for simplicity, that the density of states D(E ) in the normal state is approximately constant close to the Fermi energy. We restrict ourselves to single-electron tunneling; pair tunneling, which leads to the Josephson eects, will be discussed later. Quasiparticle density of states The density of states per spin of the Bogoliubov quasiparticles created by is easily obtained from their dispersion: ( ) 1 1 2 + | |2 . Ds (E ) = (E Ek ) = E k (10.70) k N N
k k
97
The subscript s of Ds (E ) stands for superconducting. For the case of approximately constant normal-state density of states Dn and constant gap 0 , we nd

Ds (E ) =
) ( ) 2 2 2 2 d Dn ( ) E + 0 = Dn (EF ) d E + 0
( ( ) ) 2 2 2 2 + E E 0 0 = Dn (EF ) d + | 2 2 | 2 2 Dn (EF ) 2E E 2 2 = 0 0
+0 +0
for E > 0 , for E < 0 .
(10.71)
Ds (E ) Dn (EF)
There are of course no states in the energy gap. Importantly, we nd a divergence at the gap edge at E = 0 . In the normal-state limit, 0 0, we obtain Ds (E )/Dn (EF ) 2, deviating from the result of unity given in Tinkhams book. The origin is that E is an excitation energy relative to the Fermi sea, i.e., both electron and hole excitations contribute to the density of states at positive E . Single-particle tunneling It is plausible that the density of states Ds (E ) can be mapped out by tunneling experiments, for example using a normal-metal/insulator/superconductor structure. However, we have to keep in mind that the particles tunneling out of or into a normal metal are real electrons, whereas the quasiparticles in a superconductor are superpositions of electrons and holes. To study tunneling eects theoretically, we employ a tunneling Hamiltonian of the form H = HL + HR + HT , (10.72)
where HL and HR describe the materials to the left and right of the tunneling region. Either can be a normal metal or a superconductor and is assumed to be unaected by the presence of the tunneling region. For example, while translational invariance is necessarily broken in a tunneling device, we nevertheless write HL,R in terms of lattice-momentum states. HT describes the tunneling between the two materials, HT = tkq c (10.73) k dq + h.c.
kq
Here, c and d are electronic operators referring to the two sides of the tunneling barrier and tkq is a tunneling matrix element, which might depend on the momenta of the incoming and outgoing electron. We assume a non-magnetic tunneling barrier so that the electron spin is conserved and tkq does not depend on it. Note that in the presence of interactions HT is an approximation valid for weak tunneling. We treat the two bulk materials in the mean-eld approximation so that HL,R are eectively non-interacting mean-eld Hamiltonians. We further assume constant normal-state densities of states and momentum-independent tunneling matrix elements t = tkq . 98
For tunneling between two normal metals, we can calculate the current for an applied voltage V using the Landauer formula : Inn = ILR IRL (10.74)
left to right right to left
with ILR 2e h 2e h
n n d DL ( + ) DR ( + + eV ) |t| nF ( ) [1 nF ( + eV )], n n d DL ( + ) DR ( + + eV ) |t| nF ( + eV ) [1 nF ( )], 2 2
(10.75)
IRL
(10.76)
n n is the density of states per spin direction in the left/right normal metal. The factors of DL ( + ) nF ( ) where DL,R etc. can be understood as the probabilities that the relevant initial states exist and are occupied and the nal states exist and are empty. We get
Inn
2e h
n n d DL ( + ) DR ( + + eV ) |t| [nF ( ) nF ( + eV )] 2
2e n 2 n D (EF ) DR (EF ) |t| = h L =
d [nF ( ) nF ( + eV )] (10.77)
2e2 n 2 n DL (EF ) DR (EF ) |t| V. h Inn = Gnn V.
We thus nd ohmic behaviour, (10.78) We next consider the case that one material is normal while the other (without loss of generality the left) is superconducting. Now there are additional factors because the quasiparticles in the superconductor are not pure electrons or holes. Let us say an electron is tunneling out of the superconductor with energy > 0. First of all, this is only possible if 0 because of the energy gap. Now the electron can either come from an electron-like quasiparticle (k > kF ), which contains an electron portion of ) ( ( ) 2 2 1 k 1 2 0 |u k | = 1+ 2 = 1+ , (10.79) 2 2 2 2 k + 2 0
k = +0
or from a hole-like quasiparticle (k < kF ) with an electron portion of ( ) ( ) 2 2 1 k 1 2 0 |uk | = 1+ 2 = 1 . 2 2 2 2 k + 2 0

k = +0
(10.80)
On the other hand, an electron tunneling out with energy < 0 is best described as a hole tunneling in with energy > 0. The relevant factors are ( ) 2 2 1 2 0 1 (10.81) |vk | = 2 | | and 1 |vk | = 2
2
) 2 2 0 1+ . | |
(10.82)
99
Thus the current owing from left to right is ILR

2e 2 s n d DL ( ) DR ( + + eV ) |t| nF ( ) [1 nF ( + eV )] h 0 ) ( )} { ( 2 2 2 2 1 1 0 0 1+ + 1 2 2 =1
2e 2 s n d DL (| |) DR ( + + eV ) |t| nF ( ) [1 nF ( + eV )] h { ( ) ( )} 2 2 2 2 1 1 0 0 1 + 1+ 2 | | 2 | |
=1 s n d DL (| |) DR ( + + eV ) |t| nF ( ) [1 nF ( + eV )], 2
2e h
(10.83)
s where DL ( ) now is the superconducting density of states neither containing a spin factor of 2 nor the factor of 2 due describing both electrons and holes as excitations with positive energypositive and negative energies are here treated explicitly and separately. We see that the electron-hole mixing does not lead to any additional factors beyond the changed density of states. With the analogous expression
IRL we obtain Isn
2e h
n s ( + + eV ) |t| nF ( + eV ) [1 nF ( )] (| |) DR d DL 2
(10.84)
2e h
s n d DL (| |) DR ( + + eV ) |t| [nF ( ) nF ( + eV )] 2
2e n 2 n D (EF ) DR (EF ) |t| = h L where

s DL (| |) n DL (EF )
s DL (| |) [nF ( ) nF ( + eV )], n DL (EF )
(10.85)
| | 2 2 = 0 0
for | | > 0 , for | | < 0 .
(10.86)
Thus Isn Gnn = e
s DL (| |) [nF ( ) nF ( + eV )]. n DL (EF )
(10.87)
It is useful to consider the dierential conductance Gsn dIsn = Gnn := dV = Gnn

s DL (| |) d n DL (EF )
) ( nF ( + eV ) (10.88)
s DL (| |) nF ( + eV )[1 nF ( + eV )]. n DL (EF )
100
In the limit kB T 0 this becomes

Gsn = Gnn
s DL ( | |) Ds (|eV |) ( + eV ) = Gnn L n n (E ) . DL (EF ) DL F
(10.89)
Thus low-temperature tunneling directly measures the superconducting density of states. At non-zero temperatures, the features are smeared out over an energy scale of kB T . In a band picture we can see that the Fermi energy in the normal material is used to scan the density of states in the superconductor.
2 0
+ eV I N
For a superconductor-insulator-superconductor contact, we only present the result for the current without derivation. The result is very plausible in view of the previous cases: Iss Gnn = e

s s DL (| |) DR (| + eV |) [nF ( ) nF ( + eV )]. n n (E ) DL (EF ) DR F
(10.90)
2 L
2 R
+ eV
Now we nd a large change in the current when the voltage V is chosen such that the lower gap edge at one side is aligned with the upper gap edge at the other side. This is the case for |eV | = L + R . This feature will remain sharp at non-zero temperatures since the densities of states retain their divergences at the gap edges as long as superconductivity is not destroyed. Eectively, we are using the density-of-states singularity of one superconductor to scan the density of states of the other. A numerical evaluation of Iss gives the following typical behavior (here for L = 2kB T , R = 3kB T ):
3.5
3.0
2.5
2.0
Iss
1.5 1.0
0.5
|1 2 |
0 1 2
1 + 2
3 4
0.0
eV
101
10.5
Ultrasonic attenuation and nuclear relaxation
To conclude the brief survey of experimental consequences of BCS theory, we discuss the eects of time-dependent perturbations. They will be exemplied by ultrasonic attenuation and nuclear relaxation, which represent two distinct ways in which quasiparticle interference comes into play. Quite generally, we write the perturbation part of the Hamiltonian as (10.91) H1 = Bk k c k ck ,
kk
where Bk k are matrix elements of the perturbation between single-electron states of the non-interacting system. In the superconducting state, we have to express c, c in terms of , ,
ck = uk k + vk k, , ck = uk k vk k, ,
(10.92) (10.93)
where we have assumed uk = uk , vk = vk . Thus, with , = 1, ( )( ) H1 = Bk k u uk k + vk k k + vk k , k, =

kk kk
( Bk k u k uk k k + uk vk k k, (10.94)
) + vk uk k , k + vk vk k , k, .
It is useful to combine the terms containing Bk k and Bk,,k , since both refer to processes that change momentum by k k and spin by . If the perturbation couples to the electron concentration, which is the case for ultrasound, one nds simply Bk,,k , = Bk k . (10.95) This is often called case I. Furthermore, spin is conserved by the coupling to ultrasound, thus Bk k = Bk k . Adding the two terms, we obtain Hultra = ( 1 Bk k u k uk k k + uk vk k k, 2
kk + vk uk k , k + vk vk k , k, + uk uk k, k , ) u k vk k, k vk uk k k , + vk vk k k .
(10.96)
(10.97)
Assuming uk , vk , k R for simplicity, we get, up to a constant, Hultra = [ ( ) 1 Bk k (uk uk vk vk ) k k + k, k , 2 kk ( )] + (uk vk + vk uk ) k . k, + k , k
(10.98)
We thus nd eective matrix elements Bk k (uk uk vk vk ) for quasiparticle scattering and Bk k (uk vk + vk uk ) for creation and annihilation of two quasiparticles. Transition rates calculated from Fermis golden rule contain the absolute values squared of matrix elements. Thus the following two coherence factors will be impor-
102
tant. The rst one is

2
1 (uk uk vk vk ) = 4
) k 1+ 2 k + 2 k )} ( )( k k k k 2 2 + 1 2 1 2 2 + 2 k + 2 k k + 2 k + 2 k k k k ( ) 1 k k k k = 1+ 2 Ek Ek Ek Ek k 1+ 2 k + 2 k
{(
)(
(10.99)
2 + 2 . If the matrix elements B and the normal-state density of states are even functions of the with Ek = k k normal-state energy relative to the Fermi energy, k , the second term, which is odd in k and k , will drop out under the sum kk . Hence, this term is usually omitted, giving the coherence factor ( ) 1 k k F (k, k ) := 1 (10.100) 2 Ek Ek relevant for quasiparticle scattering in ultrasound experiments. Analogously, we obtain the second coherence factor ( ) k k =: F+ (k, k ) (10.101) 2 (uk vk + vk uk )2 = 1 + Ek Ek
=1
for quasiparticle creation and annihilation. We can gain insight into the temperature dependence of ultrasound attenuation by making the rather crude approximation that the matrix element B is independent of k, k and thus of energy. Furthermore, typical ultrasound frequencies satisfy 0 and kB T . Then only scattering of quasiparticles by phonons but not their creation is important since the phonon energy is not sucient for quasiparticle creation. The rate of ultrasound absorption (attenuation) can be written in a plausible form analogous to the current in the previous section:

d Ds (| |) Ds (| + |) |B | F (, + ) [nF ( ) nF ( + )] , ( 1 ) .
(10.102)
where now 1 F (, ) = 2
2 0 | | | |
(10.103)
With the approximations introduced above we get

2 2 Dn (EF ) |B |
Ds (| |) Ds (| + |) 1 d Dn (EF ) Dn (EF ) 2
( 1
2 0 | | | + |
) [nF ( ) nF ( + )] . (10.104)
The normal-state attentuation rate is found by letting 0 0:

2 n Dn (EF ) |B | 2
1 2 nF ( ) nF ( + ) 2 = Dn (EF ) |B | 2 2
(10.105)
s 1 = n
Ds (| |) Ds (| + |) | | | + | 2 0 [nF ( ) nF ( + )] . Dn (EF ) Dn (EF ) | | | + |
(10.106)
103
Since is small, we can expand the integral up to linear order in , s 1 = n ]2 2 nF Ds (| |) 2 0 () d Dn (EF ) 2 [ ]2 0 0 2 2 | | n nF F 0 = = + d + d 2 2 2 0

0 0
= nF () + nF (0 ) nF (0 ) + nF () = nF (0 ) + 1 nF (0 )
=0 =1
= 2 nF (0 ) =
2 e 0 + 1
(10.107)
Inserting the BCS prediction for 0 (T ), we can plot s /n vs. temperature:
s n 1
Tc
We now turn to the relaxation of nuclear spins due to their coupling to the electrons. We note without derivation that the hyperne interaction relevant for this preocess is odd in momentum if the electron spin is not changed but is even if the electron spin is ipped, i.e., Bk,,k , = Bk k (recall = 1). This is called case II. The perturbation Hamiltonian now reads ( 1 HNMR = Bk k u k uk k k + uk vk k k, 2
kk + vk uk k , k + vk vk k , k, uk uk k, k , ) + u k vk k, k + vk uk k k , vk vk k k .
(10.108)
(10.109)
Assuming uk , vk , k R, we get, up to a constant, [ ( ) 1 HNMR = Bk k (uk uk + vk vk ) k k k, k , 2 kk ( )] + (uk vk vk uk ) k . k, k , k
(10.110)
Compared to ultrasonic attenuation (case I) there is thus a change of sign in both coherence factors. An analogous derivation now gives the interchanged coherence factors ( ) 1 k k F+ (k, k ) = 1+ (10.111) 2 Ek Ek for quasiparticle scattering and 1 F (k, k ) = 2
( ) k k 1 Ek Ek
(10.112)
104
for quasiparticle creation and annihilation. The relevant energy is the Zeeman energy of a nuclear spin in the applied uniform magnetic eld and is small compared to the gap 0 . Thus we can again restrict ourselves to the small- limit. The derivation is initially analogous to case I, but with F replaced by F+ . The nuclear-spin relaxation rate is

d Ds (| |) Ds (| + |) |B | F+ (, + ) [nF ( ) nF ( + )] ( 1+ ) .
(10.113)
with 1 F+ (, ) = 2
2 0 | | | |
(10.114)
Thus
2 s Dn (EF ) |B | 2
Ds (| |) Ds (| + |) 1 Dn (EF ) Dn (EF ) 2
( 1+
2 0 | | | + |
) [nF ( ) nF ( + )] (10.115)
s 1 = n
Ds (| |) Ds (| + |) | | | + | + 2 0 [nF ( ) nF ( + )] . Dn (EF ) Dn (EF ) | | | + |
(10.116)
If we now expand the integral for small as above, we encounter a problem: s 1 = 2 n

[ d
| | 2 2 0
]2
2 + 2 nF 0 () = 2 2
2 + 2 0 nF . 2 2 0
(10.117)
This integral diverges logarithmically at the lower limit. Keeping a non-zero but realistically small removes the divergence. However, the calculated s /n is still too large compared to experiments. The origin of this problem is the strong singularity in the superconducting density of states. A k-dependent gap k removes the problem; in a realistic theory k always has a k dependence since it cannot have higher symmetry than the underlying normal dispersion k . Introducing some broadening of the density of states by hand, we numerically nd the following temperature dependence:
s n case II (NMR) 1 case I 0 Tc T
There is a large maximum below the transition temperature, called the Hebel-Slichter peak. It results from the 2 factor Ds (| |) (for 0) in the integrand,
2 2 Ds (| |) (FE ) = Dn
2 , 2 2 0
(10.118)
which for nuclear relaxation is not canceled by the coherence factor F+ , whereas for ultrasonic attenuation it is canceled by F . Physically, the strong enhancement below Tc of the density of states of both initial and nal states at 0 leads to increased nuclear relaxation. 105
10.6
Ginzburg-Landau-Gorkov theory
We conclude this chapter by remarking that Lev Gorkov managed, two years after the publication of BCS theory, to derive Ginzburg-Landau theory from BCS theory. The correspondence is perfect if the gap is suciantly small, i.e., T is close to Tc , and the electromagnetic eld varies slowly on the scale of the Pippard coherence length 0 (see Sec. 5.4). These are indeed the conditions under which Ginzburg and Landau expected their theory to be valid. Gorkov used equations of motion for electronic Green functions, which he decoupled with a mean-eld-like approximation, which allowed for spatial variations of the decoupling term (r). The derivation is given in Schrieers book and we omit it here. Gorkov found that in order to obtain the Ginzburg-Landau equations, he had to take q = 2e, m = 2m, as anticipated, and (using our conventions) 7 (3) 0 (r) (r) = ns , 4 kB Tc where n0 s := ns T 1 T c .
T Tc
(10.119) (10.120)
(10.121)
(10.122)
Recall that ns 1 T /Tc close to Tc in Ginzburg-Landau theory. The spatially dependent gap is thus locally proportional to the Ginzburg-Landau condensate wavefunction or order parameter (r). Since we have already found that London theory is a limiting case of Ginzburg-Landau theory, it is also a limiting case of BCS theory. But London theory predicts the two central properties of superconductors: Ideal conduction and ux expulsion. Thus Gorkovs derivation also shows that BCS theory indeed describes a superconducting state. (Historically, this has been shown by BCS before Gorkov established the formal relationship between the various theories.)
106
11
Josephson eects
Brian Josephson made two important predictions for the current owing through a tunneling barrier between two superconductors. The results have later been extended to various other systems involving two superconducting electrodes, such as superconductor/normal-metal/superconductor heterostructures and superconducting weak links. Rather generally, for vanishing applied voltage a supercurrent Is is owing which is related to the phase dierence of the two condensates by ( ) 2 Is = Ic sin ds A . (11.1) 0 We will discuss the critical current Ic presently. We consider the case without magnetic eld so that we can choose the gauge A 0. Then the Josephson relation simplies to Is = Ic sin . (11.2)
It should be noted that this DC Josephson eect is an equilibrium phenomenon since no bias voltage is applied. The current thus continues to ow as long as the phase dierence is maintained. Secondly, Josephson predicted that in the presence of a constant bias voltage V , the phase dierence would evolve according to 2e d = V (11.3) dt (recall that we use the convention e > 0) so that an alternating current would ow, ( ) 2e Is (t) = Ic sin 0 Vt . (11.4) This is called the AC Josephson eect. The frequency J := 2eV (11.5)
of the current is called the Josephson frequency. The AC Josephson eect relates frequencies (or times) to voltages, which makes it important for metrology.
11.1
The Josephson eects in Ginzburg-Landau theory
We consider a weak link between two identical bulk superconductors. The weak link is realized by a short wire of length L and cross section A made from the same material as the bulk superconductors. We choose this setup since it is the easiest to treat in Ginzburg-Landau theory since the parameters and are uniform, but the only property that really matters is that the phase of the order parameter (r) only changes within the weak
107
link. We employ the rst Ginzburg-Landau equation for A 0 assuming (r) to depend only on the coordinate x along the wire, 2 f (x) + f (x) f 3 (x) = 0 (11.6) with f (x) = (x) = | ()| (x). (11.7)
We assume the two bulk superconductors to be uniform and to have a relative phase of . This allows us to write { 1 for x 0, f (x) = (11.8) i e for x L. For the wire we have to solve Eq. (11.6) with the boundary conditions f (0) = 1, f (L) = ei . (11.9)
Since L , the rst term in Eq. (11.6) is larger than the other two by a factor of order 2 /L2 , unless = 0, in which case the solution is trivially f 1. It is thus sucient to solve f (x) = 0, which has the solution f (x) = Lx x + ei . L L (11.10)
Inserting f (x) into the second Ginzburg-Landau equation (with A 0), we obtain ( ) q e js = i [( ) ] = i [(f ) f f f ] 2m 2m [( )( ) ( )( )] e 1 1 Lx x Lx x 1 1 = i 2ns + ei + ei + ei + ei 2m L L L L L L L L [ ] [ ] ( ) ( ) e ns x L x i e ns x Lx i = i 2 ei ei e e = 2 sin sin m L L2 m L2 L2 e ns = 2 sin . (11.11) mL The current is obviously obtained by integrating over the cross-sectional area, Is = so that we get 2e ns A . (11.13) m L The negative sign is due to the negative charge 2e of the Cooper pairs. The amplitude of the current-phase relation is clearly |Ic |. Ic = 2e ns A sin , m L (11.12)
108
Ginzburg-Landau theory also gives us the free energy of the wire. Since we have neglected the and terms when solving the Ginzburg-Landau equation, we must for consistency do the same here, L F =A
0 2
dx
4m
| (x)| = A
0
( ) 1 1 dx + ei 4m L L
2
2ns =A 4m L2
L dx 2 (1 cos )
0
A 2 ns = (1 cos ) . L m
(11.14)
The free energy is minimal when the phases of the two superconductors coincide. Thus if there existed any mechanism by which the phases could relax, they would approach a state with uniform phase across the junction, a highly plausible result. We can now also derive the AC Josephson eect. Assuming that the free energy of the junction is only changed by the supercurrent, we have d F = Is V, (11.15) dt i.e., the electrical power. This relation implies that F d = Is V dt A 2 ns d 2e ns A sin = sin V L m dt m L 2e d = V, dt (11.16) (11.17) (11.18)
as stated above. Physically, if a supercurrent is owing in the presence of a bias voltage, it generates power. Since energy is conserved, this power must equal the change of (free) energy per unit time of the junction.
11.2
Dynamics of Josephson junctions
For a discussion of the dynamical current-voltage characteristics of a Josephson junction, it is crucial to realize that a real junction also 1. permits single-particle tunneling (see Sec. 10.4), which we model by an ohmic resistivity R in parallel to the junction, 2. has a non-zero capacitance C . This leads to the resistively and capacitively shunted junction (RCSJ) model represented by the following circuit diagram:
junction
109
The current through the device is the sum of currents through the three branches, I= V dV +C Ic sin , R dt (11.19)
where we take Ic > 0 and have made the sign explicit. With d 2e = V dt we obtain I= We introduce the plasma frequency p := and the quality factor Q := p RC of the junction. This leads to 1 d2 1 d I sin = 2 2 Ic p dt Qp dt and with := p t nally to d2 1 d I + + sin = . d 2 Q d Ic (11.25) (11.24) (11.23) C d2 d Ic sin . 2eR dt 2e dt2 2eIc C (11.20)
(11.21)
(11.22)
Compare this equation to the Newton equation for a particle moving in one dimension in a potential Vpot (x) with Stokes friction, mx + x = dVpot dx 1 dVpot x + x = . m m dx (11.26) (11.27)
This Newton equation has the same form as the equation of motion of if we identify t , x , 1 , m Q 1 I Vpot (x) cos . m Ic (11.28) (11.29) (11.30) (11.31)
Thus the time dependence of ( ) corresponds to the damped motion of a particle in a tilted-washboard potential
I Ic Vpot m
0
110
Equation (11.25) can be used to study a Josephson junction in various regimes. First, note that a stationary solution exists as long as |I | Ic . Then sin = const = I Ic and V 0. (11.32)
This solution does not exist for |I | > Ic . What happens if we impose a time-independent current that is larger than the critical current? We rst consider a strongly damped junction, Q 1. Then we can neglect the acceleration term and write 1 Q 1 Q d I + sin = d Ic I d = sin d Ic d I = Q d Ic + sin
I Ic
(11.33) (11.34) (11.35) 2 = ( ) I 2

Ic 1 + I tan 2 arctan (Ic ) I 2 1 1 Ic
Q ( 0 ) =
0
d + sin
I > Ic
.
0
(11.36)
We are interested in periodic solutions for ei or mod 2 . One period T is the time it takes for to change from 0 to 2 (note that d /d < 0). Thus
2
Qp T =
0
I Ic
d + sin
I > Ic
2 ( ) I 2
Ic
(11.37) 1 (11.38)
2 1 1 1 ( ) ( ) T = = 2 = . 2 2 2 2 Qp 2eIc R eR I Ic I I 1 1
Ic Ic
The voltage V d /dt is of course time-dependent but the time-averaged voltage is simply = 1 V T T
0
1 dt V (t) = 2e T
T dt
0
1 d 1 2 = [(T ) (0)] = = R I 2 Ic dt 2e T e T
= 2
(11.39)
= R for I > Ic . By symmetry, V current thus look like this:
2 for I < I . The current-voltage characteristics for given direct I 2 Ic c
Ic 0 Ic
V/R V
For |I | Ic , the current ows without resistance. At Ic , non-zero DC and AC voltages set in gradually. For |I | Ic , the DC voltage approaches the ohmic result for a normal contact.
111
The solution for general Q requires numerical calculation but we can analyze the opposite case of weak damping, Q 1. The stationary solution = const, V = 0 still exists for I Ic . The mechanical analogy suggests that the time-dependent solution with periodic ei will be a very rapid slide down the washboard, overlaid by a small-amplitude oscillation, (11.40) = t + , where p and is periodic in time and small. Inserting this ansatz into the equation of motion we nd I p p , Ic 2 p = 2 sin t. =Q
1
(11.41) (11.42)
Thus
2 p 2 sin . (11.43) p p We convince ourselves that this is a good solution for Q 1: Inserting it into Eq. (11.25), we obtain for the left-hand side ( ) 2 p 1 1 p sin cos sin + 2 sin p Q p Q p p p ( ) ( )2 ( ) & I & 1 Ic Ic 1 1 & & 2 cos sin cos cos +O . (11.44) = sin 2 & & Ic Q I p p Q I p Q4 & p & p To leading order in 1/Q this is just I/Ic , which agrees with the right-hand side. We thus nd an averaged voltage of ) T ( 1 d I 2eIc I V = dt = = Q p = RC = RI, (11.45) T 2e dt 2e 2e Ic 2e C Ic
i.e., the ohmic behavior of the normal junction. Note that the time-dependent solution exists for all currents, = RI . If we would not just for |I | > Ic . Thus for |I | Ic there are now two solutions, with V 0 and with V change the imposed current we could expect hysteretic behavior. This is indeed observed.
I V/R Ic 0 Ic V
If we instead impose a constant voltage we obtain, for any Q, 2e d = V = const dt and thus d2 = 0 dt2 (11.46)
( ) 1 2e 2e I (t) V + sin V t + 0 = Qp Ic ( ) Ic 2e 2e I (t) = V + Ic sin V t 0 . Qp 112
(11.47) (11.48)
The averaged current is just = Ic 2e V = C Ic 2e V = V . I Qp 2eIc RC R (11.49)
Note that this result holds for any damping. It is evidently important to carefully specify whether a constant current or a constant voltage is imposed.
11.3
Bogoliubov-de Gennes Hamiltonian
It is often necessary to describe inhomogeneous systems, Josephson junctions are typical examples. So far, the only theory we know that is able to treat inhomogeneity is the Ginzburg-Landau theory, which has the disadvantage that the quasiparticles are not explicitly included. It is in this sense not a microscopic theory. We will now discuss a microscopic description that allows us to treat inhomogeneous systems. The essential idea is to make the BCS mean-eld Hamiltonian spatially dependent. This leads to the Bogoliubov-de Gennes Hamiltonian. It is useful to revert to a rst-quantized description. To this end, we introduce the condensate state |BCS as the ground state of the BCS Hamiltonian HBCS = k c k c (11.50) k ck, ck k ck k ck, + const.
k k k
|BCS agrees with the BCS ground state dened in Sec. 9.2 in the limit T 0 (recall that k and thus HBCS is temperature-dependent). We have HBCS |BCS = EBCS |BCS , (11.51) where EBCS is the temperature-dependent energy of the condensate. Since HBCS is bilinear, it is sucient to consider single-particle excitations. Many-particle excitations are simply product states, or more precisely Slater determinants, of single-particle excitations. We rst dene a two-component spinor ( ) ( ) |k1 ck |k := |BCS . (11.52) |k2 ck, It is easy to show that
[HBCS , c k ] = k ck k ck, ,
(11.53) (11.54)
[HBCS , ck, ] = k ck, k c k . With these relations we obtain ( ) HBCS |k1 = HBCS c | = c c + c H |BCS BCS k k , BCS k k k k = (EBCS + k ) |k1 k |k2 and ( ) HBCS |k2 = HBCS ck, |BCS = k ck, k c k + ck, HBCS |BCS = (EBCS k ) |k2 k |k1 . Thus for the basis {|k1 , |k2 } the Hamiltonian has the matrix form ( ) EBCS + k k . EBCS k k
(11.55)
(11.56)
(11.57)
This is the desired Hamiltonian in rst-quantized form, except that we want to measure excitation energies relative to the condensate energy. Thus we write as the rst-quantized Hamiltonian in k space ( ) k k HBdG (k) = . (11.58) k k 113
This is the Bogoliubov-de Gennes Hamiltonian for non-magnetic superconductors. Its eigenvalues are 2 + | |2 = E k k k with corresponding eigenstates ( ) uk |k1 vk |k2 = uk c v c |BCS = k k , k k |BCS and ( ) vk |k1 + u k |k2 = vk ck + uk ck, |BCS = k, |BCS
(11.59)
(11.60)
(11.61)
with uk , vk dened as above. (A lengthy but straightforward calculation has been omitted.) We can now understand why the second eigenvalue comes out negative: The corresponding eigenstate contains a quasiparticle annihilation operator, not a creation operator. Hence, HBdG (k) reproduces the excitation energies we already know. The next step is to Fourier-transform the Hamiltonian to obtain its real-space representation, which we write as ( ) 1 ikr H0 (r) (r) HBdG (r) := e HBdG (k) = , (11.62) (r) H0 (r) N
k
where we expect 2 + V (r) (11.63) 2m as the free-electron Hamiltonian. But in this form it becomes easy to include spatially inhomogeneous situations: Both V (r) and (r) can be chosen spatially dependent (and not simply lattice-periodic). The corresponding Schrdinger equation HBdG (r) (r) = E (r) (11.64) with (r) = ( ) 1 (r) 2 (r) (11.65) H0 (r) =
2
is called the Bogoliubov-de Gennes equation. Note that in this context the gap (r) is usually dened with the opposite sign, which is just a phase change, so that the explicit minus signs in the o-diagonal components of HBdG are removed. Furthermore, in Bogoliubov-de Gennes theory, the gap function is typically not evaluated selfconsistantly from the averages ck, ck . Rather, (r) is treated as a given function characterizing the tendency of superconducting pairing.
11.4
Andreev reection
As an application of the Bogoliubov-de Gennes approach, we study what happens to an electron that impinges on a normal-superconducting interface from the normal side. We model this situation by the Bogoliubov-de Gennes Hamiltonian ( 2 ) 2m 2 0 (x) HBdG = (11.66) 2 2 0 (x) 2m + (note the changed sign of (r)) so that HBdG (r) = E (r). (11.67) In the normal region, x < 0, the two components 1 (r), 2 (r) are just superpositions of plane waves with wave vectors k1 , k2 that must satisfy
2 2 k1 = 2m( + E ) = kF + 2mE, 2 k2
(11.68) (11.69)
= 2m( E ) = 114
2 kF
2mE,
where
= 1. In the superconductor, x > 0, we have ( ) 1 2 1 (r) + 0 2 (r) = E 1 (r), 2m ( ) 1 2 + 2 (r) + 0 1 (r) = E 2 (r) 2m
2 E + 21 m + 2 (r) = 1 (r) 0 ( )2 1 2 + 1 (r) = (E 2 2 0 ) 1 (r) 2m
(11.70) (11.71) (11.72) (11.73)
and analogously ( )2 1 2 + 2 (r) = (E 2 2 0 ) 2 (r). 2m (11.74)
If the energy is above the gap, |E | > 0 , the solutions are again plain wave vectors q1 , q2 , where now ( )2 2 q1 ,2 = E 2 2 0 >0 2m ) ( 2 2 2 q1,2 = 2m + E 0 ,
(11.75) (11.76)
and amplitudes coupled by Eq. (11.72). We are here interested in the more surprising case |E | < 0 . Since the solution must be continuous across the interface and is plane-wave-like in the normal region, we make the ansatz 1 (r) = ei(k1y y+k1z z) 1 (x), from which ( 1 d2 + + 2m 2m dx2
2 k1
(11.77)
)2 1 (x) = (E 2 2 0 ) 1 (x) (11.78)
with k1 := (k1y , k1z ). Since this equation is linear with constant coecients, we make an exponential ansatz 1 (x) = ex+iqx with , q R. This leads to ( 2 )2 k1 ( + iq )2 + + = E 2 2 0 <0 2m 2m ( 2 )2 ( 2 )2 k1 + q 2 k1 + q 2 iq 2 2 iq ++ + = 2m m 2m 2m 2m m ( 2 ) ( ) 2 2 2 k1 + q 2 2 iq k1 + q 2 2 q 2 = 2 = E 2 2 0. 2m 2m m2 m 2m 2m Since the right-hand side is real, we require 2 =0 2m 2m 2 2 2 = k1 + q 2m. 115
2 2 k1 +q
(11.79)
(11.80)
(11.81)
(11.82) (11.83)
For the real part it follows that 2 q 2 = E 2 2 0 m2 ( ) 2 2 2 k1 + q 2m q 2 q 2 2 = = 2 0E >0 2 m2 ( m ) 2 2 2 2 2 q 4 + k1 2m q m (0 E ) = 0 ( 2 2 )2 2m k1 2m k1 2 2 q = + m2 (2 0E ) 2 2 [ ] ( )2 1 2 2 2 2 2 k k1 = kF k1 + 4m2 (0 E 2 ) . 2 F (11.84) (11.85) (11.86)
(11.87)
Both solutions are clearly real but the one with the minus sign is negative so that q would be imaginary, contrary to our assumption. Thus the relevant solutions are ( )2 1 2 2 2 k2 2 kF k1 + kF q = q1 := + 4m2 (2 (11.88) 0 E ). 1 2 From this we get =
2 2 k1
2 kF
[ 1 2 2 = k kF + 2 1 and
] [ ( )2 1 2 2 2 k2 2 k k1 + kF + 4m2 (2 + 0E ) 1 2 F ] ( )
2 k2 kF 1 2 2 + 4m2 (2 0E )
(11.89)
( )2 1 2 2 2 k2 2 (11.90) = 1 := k1 kF + kF + 4m2 (2 0 E ). 1 2 The positive root exists but would lead to a solution that grows exponentially for x . For 2 (r) the derivation is completely analogous. However, 1 (r) and 2 (r) are related be Eq. (11.72), which for exponential functions becomes a simple proportionality. Therefore, we must have k1 = k2 , which already implies q1 = q2 and 1 = 2 . Then we have [ ] 2 2 k1 E + 21 1 (1 iq1 )2 m + 2 (r) = 1 (r) = E+ + 1 (r) 0 0 2m 2m [ ] 2 2 k1 1 1 q1 2 +q 1 = i E ++ 1 (r). (11.91) 0 2m 2m m
=0
Since we already know that

2 2 1 q1 2 = 2 0E m2
(11.92)
and 1 , q1 have been dened as positive, we get
2 E i 2 0E 1 (r). 2 (r) = 0
(11.93)
We now write down an ansatz and show that it satises the Bogoliubov-de Gennes equation and the continuity conditions at the interface. The ansatz reads ) ( eik1 r + r eik1 r for x 0, (11.94) (r) = a eik2 r 116
1 := (k1x , k1y , k1z ) and k2 = (k2x , k1y , k1z ) with k2x > 0, where k 2 + k 2 = k 2 + 2mE and k 2 + k 2 = with k 1x 2x F 1 1 2 kF 2mE , and ( ) + ei(q1 x+k1y y+k1z z) + ei(q1 x+k1y y+k1z z) (r) = e1 x for x 0. (11.95) + ei(q1 x+k1y y+k1z z) + ei(q1 x+k1y y+k1z z) 1 is the wave vector of a specularly reected electron. From Eq. (11.93) we get Note that k 2 E i 2 0E . = 0 From the continuity of 1 , 2 , and their x-derivatives we obtain 1 + r = + + , a = + + , ik1x r ik1x = + (1 + iq1 ) + (1 iq1 ), a ik2x = + (1 + iq1 ) + (1 iq1 ). (11.97) (11.98) (11.99) (11.100)
(11.96)
We thus have six coupled linear equations for the six unknown coecients r, a, + , , + , . The equations are linearly independent so that they have a unique solution, which we can obtain by standard methods. The six coecients are generally non-zero and complex. We do not give the lengthy expressions here but discuss the results physically. The solution in the superconductors decays exponentially, which is reasonable since the energy lies in the superconducting gap. In the normal region there is a secularly reected electron wave (coecient r), which is also expected. So far, the same results would be obtained for a simple potential step. However, explicit evaluation shows that 2 in general |r| < 1, i.e., not all electrons are reected. There is also a term 2 (r) = a eik2 r for x 0. (11.101) Recall that the second spinor component was dened by |k2 = ck, |BCS . (11.102)
2 Hence, the above term represents a spin-down hole with wave vector k2 . Now k1 = k2 and k1 x = 2 2 2 2 2 kF k1 + 2mE and k2x = kF k1 2mE . But the last terms 2mE are small since
|E | < 0 =
2 kF 2m
(11.103)
in conventional superconductors. Thus |k2x k1x | is small and the hole is traveling nearly in the opposite direction compared to the incoming electron wave. This phenomenon is called Andreev reection.
N electron ~ k1
y,z
S electronlike quasiparticle (evanescent)
electron k1 k2 hole x
117
holelike quasiparticle (evanescent)
Since not all electrons are reected and in addition some holes are generated, where does the missing charge go? The quasiparticle states in the superconductor are evanescent and thus cannot accommodate the missing charge. The only possible explanation is that the charge is added to the superconducting condensate, i.e., that additional Cooper pairs are formed. (The whole process can also run backwards, in which case Cooper pairs are removed.) Recall that the condensate does not have a sharp electron number and can therefore absorb or emit electrons without changing the state. But it can only absorb or emit electrons in pairs. The emerging picture is that if an incoming electron is not specularly reected, a Cooper pair is created, which requires a second electron. This second electron is taken from the normal region, creating a hole, which, as we have seen, travels in the direction the original electron was coming from.
e h Cooper pair
Andreev bound states An interesting situation arises if a normal region is delimited by superconductors on two sides. We here only qualitatively consider a superconductor-normal-superconductor (SNS) hetero structure. Similar eects can also occur for example in the normal core of a vortex. If no voltage is applied between the two superconductors, an electron in the normal region, with energy within the gap, is Andreev reected as a hole at one interface. It is then Andreev reected as an electron at the other interface. It is plausible that multiple reections can lead to the formation of bound states. The real physics is somewhat more complicated since the electron is also partially specularly reected as an electron. It is conceptually clear, though, how to describe Andreev bound states within the Bogoliubov-de Gennes formalism: We just have to satisfy continuity conditions for both interfaces.
N electron
2 0
hole
2 0
If Andreev reection dominates, as assumed for the sketch above, a Cooper pair is emitted into the right superconductor for every reection at the right interface. Conversely, a Cooper pair is absorbed from the left superconductor for every reection at the left interface. This corresponds to a supercurrent through the device. Andreev bound states thus oer a microscopic description of the Josephson eect in superconductor-normalsuperconductor junctions. If we apply a voltage V , the situation changes dramatically: If an electron moving, say, to the right, increases its kinetic energy by eV due to the bias voltage, an Andreev reected hole traveling to the left also increases its kinetic energy by eV since it carries the opposite charge. An electron/hole Andreev-reected multiple times can thus gain arbitrarily high energies for any non-vanishing bias voltage.
118
2 0 2 eV eV
In particular, an electron-like quasiparticle from an occupied state below the gap in, say, the left superconductor can after multiple reections emerge in a previously unoccupied state above the gap in the right superconductor. A new transport channel becomes available whenever the full gap 20 is an odd integer multiple of eV : 20 = (2n + 1) eV, n = 0, 1, . . . 0 eV = , n = 0, 1, . . . n+ 1 2 (11.104) (11.105)
The case n = 0 corresponds to direct quasiparticle transfer from one superconductor to the other, similar to quasiparticle tunneling in a superconductor-insulator-superconductor junction. The opening of new transport channels for n = 0, 1, . . . , i.e., at 2 2 2 (11.106) eV = 0 , 0 , 0 , . . . 3 5 7 leads to structures in the current-voltage characteristics below the gap, specically to peaks in the dierential conductance dI/dV .
dI dV
2 0
2 3 0 2 5 0
2 0
eV
119
12
Unconventional pairing
In this chapter we rst discuss why interactions dierent from the phonon-mediated one might lead to unconventional pairing, that is to a gap function k with non-trivial k dependence. Then we will briey consider the origin of such interactions.
12.1
The gap equation for unconventional pairing
We will still use the BCS gap equation even when discussing unconventional superconductors. While the BCS mean-eld theory is not quantitatively correct in such cases, it will give a clear understanding of why the gap k can have non-trivial symmetry. To get started, we briey review results from BCS theory. The screened eective interaction was derived in Sec. 8.4,
RPA RPA Ve (q, in ) = VC (q, in )
(in
)2
(in )2 , 2 (i ) q n
(12.1)
RPA where VC is the screened Coulomb interaction and q is the renormalized phonon dispersion. The retarded interaction at small frequencies is R Ve (q, in ) = 4
e2 2 , 2 R 2 + q 2 + 2 s q ( ) + i0 sgn
(12.2)
where s is the inverse screening length. The bevavior at small distances r and small but non-zero frequency R is determined by Ve at large q , where q can by approximated by the Debye frequency. The interaction is thus attractive and decays like 1/r for small r. The interaction is strongest at the same site in a tight-binding model. In order to understand the physics, it makes sense to replace the interaction by a simplied one that is completely local (attractive Hubbard model) or, equivalently, constant in k space, as we have done above. However, the BCS gap equation 1 k k = Vkk [1 nF (Ek )] (12.3) N 2Ek
k
is in fact much more general. In the gap equation, Vkk describes the amplitude for scattering of two electrons with momenta k and k and opposite spins into states with momenta k and k. Let us rst consider the case that the interaction is local in real space (at in k space) but repulsive. This would apply if the phonons were for some reason ineective in overscreening the Coulomb interaction. Then we obtain 1 k [1 nF (Ek )] (12.4) k = V0 N 2Ek
k
with V0 > 0. The right-hand side is clearly independent of k so that we have k = 0 and can cancel a factor of 0 if it is non-zero: V0 1 nF (Ek ) . (12.5) 1= N 2Ek
k
120
But now the right-hand side is always negative. Consequently, there is no non-trivial solution and thus no superconductivity for a k-independent repulsion. Now let us look at a strong interaction between nearest-neighbor sites. We consider a two-dimensional square lattice for simplicity and since it is thought to be a good model for the cuprates. In momentum space, a nearestneighbor interaction is written as
Vkk = 2V1 [cos (kx kx )a + cos (ky ky )a],
(12.6)
where V1 > 0 (V1 < 0) for a repulsive (attractive) interaction. For our discussion of momentum-dependent interactions it is crucial to realized which terms in the k sum in the gap equation k = 1 1 nF (Ek ) Vkk k N 2Ek
k
(12.7)
are most important. The factor [1 nF (Ek )]/2Ek is largest on the normal-state Fermi surface, where Ek = |k |, and is exponentially suppressed on an energy scale of kB T away from it. Thus only the vicinity of the Fermi surface is important. Let us rst consider the repulsive case V1 > 0. Then Vkk is most strongly repulsive (positive) for momentum transfer k k 0. But k should be a smooth function of k, thus for k and k close together, k and k are also similar. In particular, k will rarely change its sign between k and k . Consequently, the right-hand side of the gap equation always contains a large contribution with sign opposite to that of k , coming from the sum over k close to k. Hence, a repulsive nearest-neighbor interaction is unlikely to lead to superconductivity. For the attractive case, V1 < 0, Vkk is most strongly attractive for k k 0 and most strongly repulsive for k k (/a, /a) and equivalent points in the Brillouin zone. The attraction at small q = k k is always favorable for superconductivity. However, we also have an equally strong repulsion around q (/a, /a). A critical situation thus arises if both k and k lie close to the Fermi surface and their dierence is close to (/a, /a). The central insight is that this can still help superconductivity if the gaps k and k at k and k , respectively, have opposite sign. In this case the contribution to the right-hand side of the gap equation from such k has the same sign as k since Vkk > 0 and there is an explicit minus sign. This eect is crucial in the cuprates, which do have an eective attractive nearest-neighbor interaction and have a large normal-state Fermi surface shown here for a two-dimensional model:
ky
Q 0 kx
The vector Q in the sketch is Q = (/a, /a). Following the previous discussion, k close to the Fermi surface should have dierent sign between points separated by Q. On the other hand, the small-q attraction favors gaps k that change sign as little as possible. By inspection, these conditions are met by a gap changing sign on the diagonals:
121
ky
Q 0 kx
This type of gap is called a dx2 y2 -wave (or just d-wave) gap since it has the symmetry of a dx2 y2 -orbital (though in k-space, not in real space). The simplest gap function with this symmetry and consistent with the lattice structure is k = 0 (cos kx a cos ky a). (12.8) Recall that the gap function away from the Fermi surface is of limited importance. The d-wave gap k is distinct from the conventional, approximately constant s-wave gap in that it has zeroes on the Fermi surface. These zeroes are called gap nodes. In the present case they appear in the (11) and equivalent directions. The quasiparticle dispersion in the vicinity of such a node is 2 + | |2 2 (12.9) Ek = k = (k )2 + 2 k 0 (cos kx a cos ky a) . ( ) kF 1 k0 = . 2 1 Writing k = k0 + q and expanding for small q, we obtain 0 0 2 Ek0 +q = (vF q)2 + 2 0 [(sin kx a)qx a + (sin ky a)qy a] , One node is at where k vF := k vF = 2 ( ) 1 1 (12.10)
(12.11)
(12.12)
k=k0
is the normal-state Fermi velocity at the node. Thus [( ) ]2 kF a kF a Ek0 +q a sin , a sin q = (vF q)2 + (vqp q)2 , = (vF q)2 + 2 0 2 2 where
(12.13)
( ) kF a 1 vqp := 0 a sin vF . (12.14) 1 2 Thus the quasiparticle dispersion close to the node is a cone like for massless relativistic particles, but with dierent velocities in the directions normal and tangential to the Fermi surface. Usually one nds vF > vqp . The sketch shows equipotential lines of Ek . (12.15)
ky
kx
122
The fact that the gap closes at some k points implies that the quasiparticle density of states does not have a gap. At low energies we can estimate it from our expansion of the quasiparticle energy, Ds (E ) = 1 (E Ek ) N k ( ) 4 E (vF q)2 + (vqp q)2 = N q ( ) d2 q 2 + (v 2 E ( v q ) q ) = 4auc F qp (2 )2 ) 4auc d2 u ( 2 = E u2 x + uy 2 vF vqp (2 ) 2auc 2auc du u (E u) = = E, vF vqp 0 vF vqp
(12.16)
where auc is the area of the two-dimensional unit cell. We see that the density of states starts linearly at small energies. The full dependence is sketched here:
Ds (E ) Dn (EF)
max
An additional nice feature of the d-wave gap is the following: The interaction considered above is presumably not of BCS (Coulomb + phonons) type. However, there should also be a strong short-range Coulomb repulsion, which is not overscreened by phonon exchange. This repulsion can again be modeled by a constant V0 > 0 in k-space. This additional interaction adds the term 1 1 nF (Ek ) V0 k N 2Ek
k
(12.17)
to the gap equation. But since Ek does not change sign under rotation of k by /2 (i.e., 90 ), while k does change sign under this rotation, the sum over k vanishes. d-wave pairing is thus robust against on-site Coulomb repulsion.
12.2
Cuprates
Estimates of Tc based on phonon-exchange and using experimentally known values of the Debye frequency, the electron-phonon coupling, and the normal-state density of states are much lower than the observed critical temperatures. Also, as we have seen, such an interaction is at in k-space, which favors an s-wave gap. An s-wave gap is inconsistant with nearly all experiments on the cuprates that are sensitive to the gap. The last section has shown that dx2 y2 -wave pairing in the cuprates is plausible if there is an attractive interaction for momentum transfers q (/a, /a). We will now discuss where this attraction could be coming from.
123
AFM
SC
hole doping x
A glance at typical phase diagrams shows that the undoped cuprates tend to be antiferromagnetic. Weak hole doping or slightly stronger electron doping destroy the antiferromagnetic order, and at larger doping, superconductivity emerges. At even larger doping (the overdoped regime), superconductivity is again suppressed. Also in many other unconventional superconductors superconductivity is found in the vicinity of but rarely coexisting with magnetic order. This is true for most pnictide and heavy-fermion superconductors. The vicinity of a magnetically ordered phase makes itself felt by strong magnetic uctuations and strong, but short-range, spin correlations. These are seen as an enhanced spin susceptibility. At a magnetic second-order phase transition, the static spin susceptibility q diverges at q = Q, where Q is the ordering vector. It is Q = 0 for ferromagnetic order and Q = (/a, /a) for checkerboard (Nel) order on a square lattice. Even some distance from the transition or at non-zero frequencies , the susceptibility q ( ) tends to have a maximum close to Q. Far away from the magnetic phase or at high frequencies this remnant of magnetic order becomes small. This discussion suggests that the exchange of spin uctuations, which are strong close to Q, could provide the attractive interaction needed for Cooper pairing. The Hubbard model The two-dimensional, single-band, repulsive Hubbard model is thought (by many experts, not by everyone) to be the simplest model that captures the main physics of the cuprates. The Hamiltonian reads, in real space, H= tij c ci ci c (12.18) i cj + U i ci
ij i
with U > 0 and, in momentum space, H=

k
k c k ck +
U c c ck ck . N k+q, k q,
kk q
(12.19)
Also, the undoped cuprate parent compounds have, from simple counting, an odd number of electrons per unit cell. Thus for them the single band must be half-lled, while for doped cuprates it is still close to half lling. The underlying lattice in real space is a two-dimensional square lattice with each site i corresponding to a Cu+ ion.
Cu+
The transverse spin susceptibility is dened by
O 2
+ (q, ) = T S + (q, ) S (q, 0) , 124
(12.20)
where is the imaginary time, T is the time-ordering directive, and S (q, ) := S x (q, ) iS y (q, ) with 1 S (q, ) := ck+q, ( ) ck ( ) 2 N k (12.21)
(12.22)
are electron-spin operators. = ( x , y , z ) is the vector of Pauli matrices. The susceptibility can be rewritten as ) ) ( 1 ( (12.23) + (q, ) = T ck+q, ck ( ) c k q, ck (0) . N
kk
In the non-interacting limit of U 0, the average of four fermionic operators can be written in terms of products of averages of two operators (Wicks theorem). The resulting bare susceptibility reads 1 + T c T ck ( ) c 0 (q, ) = k+q, ( ) ck+q, (0) k (0) N 0 0 k 1 = T ck+q, (0) c T ck ( ) c k+q, ( ) k (0) N 0 0 k 1 0 0 = Gk (12.24) +q, ( ) Gk ( ). N
k
The Fourier transform as a function of the bosonic Matsubara frequency in is

+ 0 (q, in )
=
0
d ein + 0 (q, )
1 = N
d ein
0
1 0 1 0 ein ( ) Gk ein Gk +q, (in ) (in ) i

k in
n
1 1 = N 2
k
in ,in
0 0 , G n +n k+q, (in ) Gk (in ) = n
1 1 1 1 = . N i in in k+q in k
k
n
1 1 0 0 G (in in ) Gk (in ) N i k+q,

k
n
(12.25)
This expression can be written in a more symmetric form by making use of the identity k = k and replacing the summation variables k by k q and in by in + in . The result is
+ 0 (q, in ) =
1 1 1 1 . N i in k in + in k+q
k
n
(12.26)
The Matsubara frequency sum can be evaluated using methods from complex analysis. We here give the result without proof, 1 nF (k ) nF (k+q ) + . (12.27) 0 (q, in ) = N in + k k+q
k
Note that the same result is found for the bare charge susceptibility except for a spin factor of 2. Diagramatically, the result can be represented by the bubble diagram
+ 0 (q, in )
1 = 0 (q, in ) = 2
k , , in
. (12.28)
k + q , , in + i n
125
What changes when we switch on the Hubbard interaction U ? In analogy with the RPA theory for the screened Coulomb interaction we might guess that the RPA spin susceptibility is given by
+ RPA = ?
(12.29)
but the second and higher terms vanish since they contain vertices at which the Hubbard interaction supposedly ips the spin,
(12.30)
which it cannot do. On the other hand, the following ladder diagrams do not vanish and represent the RPA susceptibility:
+ RPA =
(12.31)
Since the Hubbard interaction does not depend on momentum, this series has a rather simple mathematical form,
+ + + + + + + RPA (q, in ) = 0 (q, in ) + 0 (q, in ) U 0 (q, in ) + 0 (q, in ) U 0 (q, in ) U 0 (q, in ) + + + + = + 0 (q, in )[1 + U 0 (q, in ) + U 0 (q, in ) U 0 (q, in ) + ] + 0 (q, in ) 1 U + 0 (q, in )
(12.32)
[the signs in the rst line follow from the Feynman rules, in particular each term contains a single fermionic loop, which gives a minus sign, which cancels the explicit one in Eq. (12.31)]. The RPA spin susceptibility can be evaluated numerically for given dispersion k . It is clear that it predicts an instability of the Fermi liquid if the static RPA spin susceptibility
,R + + + RPA (q, 0) = RPA (q, in + i0 )| 0 ,R 1 U + (q, 0) 0 ,R (q, 0) + 0
(12.33)
diverges at some q = Q, i.e., if
,R (Q, 0) = 1. U + 0
(12.34)
Since the spin susceptibility diverges, this would be a magnetic ordering transition with ordering vector Q. Note that the RPA is not a good theory for the antiferromagnetic transition of the cuprates since it is a resummation of a perturbative series in U/t, which is not small in cuprates. t is the typical hopping amplitude. Nevertheless it gives qualitatively reasonable results in the paramagnetic phase, which is of interest for superconductivity. The numerical evaluation at intermediate doping and high temperatures gives a broad and high peak in ,R + RPA (q, 0) centered at Q = (/a, /a). This is consistant with the ordering at Q observed at weak doping.
, R (( q, q), 0) + RPA
a
126
2 a
At lower temperatures, details become resolved that are obscured by thermal broadening at high T . The RPA and also more advanced approaches are very sensitive to the electronic bands close to the Fermi energy; states with |k | kB T have exponentially small eect on the susceptibility. Therefore, the detailed susceptibility at low T strongly depends on details of the model Hamiltonian. Choosing nearest-neighbor and next-nearest-neighbor hopping in such a way that a realistic Fermi surface emerges, one obtains a spin susceptiblity with incommensurate peaks at a (1, 1 ) and a (1 , 1).
qy a
qx
These peaks are due to nesting : Scattering is enhanced between parallel portions of the Fermi surface, which in turn enhances the susceptibility [see M. Norman, Phys. Rev. B 75, 184514 (2007)].
ky a (1 , 1) a (1, 1 ) a a
kx
The results for the spin susceptibility are in qualitative agreement with neutron-scattering experiments. However, the RPA overestimates the tendency toward magnetic order, which is reduced by more advanced approaches. Spin-uctuation exchange The next step is to construct an eective electron-electron interaction mediated by the exchange of spin uctuations. The following diagrammatic series represents the simplest way of doing this, though certainly not the only one:
U
Veff
:=
(12.35)
127
Note that the external legs do not represent electronic Green functions but only indicate the states of incoming and outgoing electrons. The series is very similar to the one for the RPA susceptibility. Indeed, the eective interaction is
+ + Ve (q, in ) = U + U + 0 (q, in ) U + U 0 (q, in ) U 0 (q, in ) + + + = U + U 2 [+ 0 (q, in ) + 0 (q, in ) U 0 (q, in ) + ] = U + U 2 + RPA (q, in ).
(12.36)
Typically one goes beyond the RPA at this point by including additional diagrams. In particular, also charge uctuations are included through the charge susceptibility and the bare Green function G 0 is replaced by a selfconsistent one incorporating the eect of spin and charge uctuations on the electronic self-energy. This leads to the uctuation-exchange approximation (FLEX ). One could now obtain the Cooper instability due to the FLEX eective interaction in analogy to Sec. 9.1 and use a BCS mean-eld theory to describe the superconducting state. However, since the system is not in the weak-coupling limitthe typical interaction times the electronic density of states is not smallone usually employs a strong-coupling generalization of BCS theory known as Eliashberg theory. Since the eective interaction is, like the spin susceptibility, strongly peaked close to (/a, /a), it favors dx2 y2 -wave pairing, as we have seen. The numerical result of the FLEX for Tc and for the superuid density ns are sketched here:
Tc
ns
doping x
The curve for Tc vs. doping does not yet look like the experimentally observed dome. There are several aspects that make the region of weak doping (underdoping) dicult to treat theoretically. One is indicated in the sketch: ns is strongly reduced, which indicates that superconductivity may be in some sense fragile in this regime. Furthermore, the cuprates are nearly two-dimensional solids. In fact we have used a two-dimensional model so far. If we take this seriously, we know from chapter 7 that any mean-eld theory, which Eliashberg theory with FLEX eective interaction still is, fails miserably. Instead, we expect a BKT transition at a strongly reduced critical temperature. We reinterpret the FLEX critical termperature as the mean-eld termperature TMF and the FLEX superuid density as the unrenormalized superuid density n0 s . We have seen in chapter 7 that the bare stiness K0 is proportional to n0 . The BKT transition temperature T c is dened by K (l ) = 2/ . It is thus s , corresponding to a small initial value K (0) K reduced by small n0 0 . This is physically clear: Small stiness s makes it easy to create vortex-antivortex pairs. Tc is of course also reduced by TMF and can never be larger than TMF . A BKT theory on top of the FLEX gives the following phase diagram, which is in qualitative agreement with experiments:
T TMF vortex fluctuations Tc 0 superconductor x ns
This scenario is consistent with the Nernst eect (an electric eld measured normal to both an applied magnetic eld and a temperature gradient) in underdoped cuprates, which is interpreted in terms of free vortices in a broad temperature range, which we would understand as the range from Tc to TMF . 128
Also note that spin uctuations strongly aect the electronic properties in underdoped cuprates up to a temperature T signicantly higher than TMF . For example, below T the electronic density of states close to the Fermi energy is suppressed compared to the result of band-structure calculations. The FLEX describes this eect qualitatively correctly. This suppression is the well-known pseudogap. Quantum critical point The previously discussed approach relies on a resummation of a perturbative series in U/t. This is questionable for cuprates, where U/t is on the order of 3. Many dierent approaches have been put forward that supposedly work in this strong-coupling regime. They emphasize dierent aspects of the cuprates, showing that it is not even clear which ingredients are the most important for understanding the phase diagram. Here we will review a lign of thought represented by Chandra Varma and Subir Sachdev, among others. Its starting point is an analysis of the normal region of the phase diagram.
T T*
strange metal
~T
pseudo gap AFM superconductor
Fermi liquid
0
Roughly speaking, there are three regimes in the normal-conducting state:
a pseudo-gap regime below T at underdoping, in which the electronic density of states at low energies is suppressed, a strange-metal regime above the superconducting dome, without a clear suppression of the density of states but with unusual temperature and energy dependencies of various observables, for example a resistivity linear in temperature, T , an apparently ordinary normal-metal (Fermi-liquid) regime at overdoping, with standard const + T 2 dependence. The two crossover lines look very much like what one expects to nd for a quantum critical point (QCP), i.e., a phase transition at zero temperature.
T quantum critical affected by phase I affected by phase II
phase I QCP c
phase II tuning parameter
The regions to the left and right have a characteristic energy scale ( ) inherited from the ground state (T = 0), which dominates the thermal uctuations. The energy scales go to zero at the QCP, i.e., for c from both sides. Right at the QCP there then is no energy scale and energy or temperature-dependent quantities have to be power laws. The emerging idea is that the superconducting dome hides a QCP between antiferromagnetic (spindensity-wave) and paramagnetic order at T = 0. The spin uctuations associated with this QCP become stronger 129
as it is approached. Consequently, the superconductivity caused by them is strongest and has the highest Tc right above the QCP. Compare the previous argument: There, the spin uctuations are assumed to be strongest close to the (nite-temperature) antiferromagnetic phase and one needs to envoke small ns and vortex uctuations to argue why the maximum Tc is not close to the antiferromagnetic phase.
12.3
Pnictides
The iron pnictide superconductors are a more heterogeneous group than the cuprates. However, for most of them the phase diagram is roughly similar to the one of a typical cuprate in that superconductivity emerges at nite doping in the vicinity of an antiferromagnetic phase. This antiferromagnetic phase is metallic, not a Mott insulator, though, suggesting that interactions are generally weaker in the pnictides.
T 140 K CeFeAsO1x Fx
al on ic rag mb tet horo ort
AFM 0 0.05
superconductor doping x
The crystal structure is quasi-two-dimensional, though probably less so than in the cuprates. The common structural motif is an iron-pnictogen, in particular Fe2+ As3 , layer with Fe2+ forming a square lattice and As3 sitting alternatingly above and below the Fe2+ plaquettes.
As 3 Fe 2+
While the correct unit cell contains two As and two Fe ions, the glide-mirror symmetry with respect to the Fe plane allows to formulate two-dimensional models using a single-iron unit cell. The fact that dierent unit cells are used leads to some confusion in the eld. Note that since the single-iron unit cell is half as large as the two-iron unit cell, the corresponding single-iron Brillouin zone is twice as large as the two-iron Brillouin zone. A look at band-structure calculations or angular resolved photoemission data shows that the Fermi surface of pnictides is much more complicated than the one of cuprates. The kz = 0 cut typically shows ve Fermi pockets.
130
ky X M holelike
Q2 electronlike Q1 X kx
singleiron Brillouin zone

The (probably outer) hole pocket at and the electron pockets at X and X are well nested with nesting vectors Q1 = (/a, 0) and Q2 = (0, /a), respectively. Not surprisingly, the spin susceptibility is peaked at Q1 and Q2 in the paramagnetic phase and in the antiferromagnetic phase the system orders antiferromagnetically at either of three vectors. Incidentally, the antiferromagnet emerges through the formation and condensation of electron-hole pairs (excitons ), described by a BCS-type theory. The same excitonic instability is for example responsible for the magnetism of chromium. Assuming that the exchange of spin fulctuations in the paramagnetic phase is the main pairing interaction, the gap equation 1 1 nF (Ek ) k = Vkk k (12.37) N 2Ek
k
with Vkk large and positive for k k = Q1 or Q2 , favors a gap function k changing sign between k and k + Q1 and between k and k + Q2 , see Sec. 12.1. This is most easily accomodated by a nodeless gap changing sign between the electron and hole pockets:
ky
kx
131
This gap is said to have s-wave symmetry in that it does not have lower symmetry than the lattice, unlike d-wave. To emphasize the sign change, it is often called an s -wave gap. The simplest realization would be k = 0 cos kx a cos ky a. (12.38)
12.4
Triplet superconductors and He-3
So far, we have assumed that Cooper pairs are formed by two electrons with opposite spin so that the total spin of pair vanishes (spin-singlet pairing). This assumption becomes questionable in the presence of strong ferromagnetic interactions, which favor parallel spin alignment. If superconductivity is possible at all in such a situation, we could expect to nd spin-1 Cooper pairs. Since they would be spin triplets, one is talking of triplet superconductors. This scenario is very likely realized in Sr2 RuO4 (which is, interestingly, isostructural to the prototypical cuprate La2 CuO4 ), a few organic salts, and some heavy-fermion compounds. It is even more certain to be responsible for the superuidity of He-3, where neutral He-3 atoms instead of charged electrons form Cooper pairs, see Sec. 2.2. Formally, we restrict ourselves to a BCS-type mean-eld theory. We generalize the eective interaction to allow for an arbitrary spin dependence, H=
k
k c k ck +
1 V (k, k ) c k ck, ck , ck , N
kk
(12.39)
where , , , =, are spin indices. In decomposing the interaction, we now allow the averages ck, ck to be non-zero for all , . Thus the mean-eld Hamiltonian reads ( ) 1 k c HMF = V (k, k ) c k ck + k ck, ck , ck + ck ck, ck , ck + const. N k kk (12.40) We dene 1 (k) := V (k, k ) ck , ck (12.41) N
k
so that (k) = Here, we have used that
1 V (k , k) c k ck , . N
k
(12.42)
[ ] V (k, k ) = V (k , k),
(12.43)
which follows from hermiticity of the Hamiltonian H . Then (k) c HMF = k c (k) ck, ck k ck, + const. k ck
k k k
(12.44)
The gap function (k) is now a matrix in spin space, ( ) (k) (k) k) = ( . (k) (k)
(12.45)
k) has an important symmetry property that follows from the symmetry of averages ck, ck : The function ( It is clear that ck ck, = ck, ck . (12.46) Furthermore, by relabeling , , k k, k k in the interaction term of H , we see that the interaction strength must satisfy the relation V (k, k ) = V (k, k ). 132 (12.47)
Thus (k) = or, equivalently, (12.49) We now want to write (k) in terms of singlet and triplet components. From elementary quantum theory, a spin-singlet pair is created by c k ck, ck ck, s := , (12.50) k 2 while the m = 1, 0, 1 components of a spin-triplet pair are created by
t k1 := ck ck, ,
1 1 V (k, k ) ck ck , = + V (k, k ) ck , ck = (k) N N k k (12.48) k) = T (k). (
(12.51) (12.52) (12.53)
t k0 := t k,1
c k ck, + ck ck, , 2 := c k ck, ,
respectively. Alternatively, we can transform onto states with maximum spin along the x, y , and z axis. This is analogous to the mapping from (l = 1, m) eigenstates onto px , py , and pz orbitals for the hydrogen atom. The new components are created by t kx t ky t kz
c k ck, + ck ck, := , 2 c k ck, + ck ck, , := i 2 c k ck, + ck ck, := = t k0 , 2
(12.54) (12.55) (12.56)
which form a vector t k . The term in HMF involving (k) can now be expressed in terms of the new operators, [ ] t t it s + t s + t kx itky k x k y kz + (k) (k)ck ck, = (k) + (k) k kz + (k) k 2 2 2 2 k k ] [ ! = 2 k + d(k) tk (12.57) (the factor
k
2 is conventional), which requires (k) (k) , 2 (k) + (k) dx (k) = , 2 (k) + (k) , dy (k) = i 2 (k) + (k) dz (k) = . 2 k = (12.58) (12.59) (12.60) (12.61)
The term in HMF involving (k) is just the hermitian conjugate of the one considered. We have now identied the singlet component of the gap, k , and the triplet components, k(k). Since ( ) dx (k) + idy (k) k + dz (k) (k) = , (12.62) k + dz (k) dx (k) + idy (k) 133
we can write the gap matrix in a compact form as k) = (k + d(k) ) i y , ( (12.63)
where is the vector of Pauli matrices. The mean-eld Hamiltonian HMF is diagonalized by a Bogoliubov k) is obtained selfconsistently from a gap equation in complete analogy transformation and the gap function ( k) is now a matrix and that we require four coecients to the singlet case discussed in Sec. 10.1, except that ( uk , uk , vk , vk . We do not show this here explicitly. k) = T (k) implies A few remarks on the physics are in order, though. The symmetry ( ( ) (k + d(k) ) i y = i ( y )T k + d(k) T (12.64) k + d(k) = ) k + d(k) T y y x T y ( ) = k + d(k) y ( y )T y y ( z )T y y x y = k + d(k) y y y y z y
y i y
= k d(k) . Thus we conclude that k is even, k = k , whereas d(k) is odd, d(k) = d(k).
(12.65) (12.66) (12.67)
Hence, the d -vector can never be constant, unlike the single gap in Sec. 10.1. Furthermore, we can expand k into even basis functions of rotations in real (and k) space, k = s s (k) + dx2 y2 dx2 y2 (k) + d3z2 r2 d3z2 r2 (k) + dxy dxy (k) + dyz dyz (k) + dzx dzx (k) + . . . , (12.68) and expand (the components of) d(k) into odd basis functions, d(k) = dpx px (k) + dpy py (k) + dpz pz (k) + . . . (12.69)
For the singlet case we have already considered the basis functions s (k) = 1 for conventional superconductors, s (k) = cos kx a cos ky a for the pnictides, and dx2 y2 (k) = cos kx a cos ky a for the cuprates. A typical basis function for a triplet superconductor would be px = sin kx a. He-3 in the so-called B phase realized at now too high pressures (see Sec. 2.2) has the d -vector px (k) + y py (k) + z pz (k), d(k) = x (12.70)
where the basis functions are that standard expressions for px , py , and pz orbitals (note that there is no Brillouin zone since He-3 is a liquid), 3 px (k) = sin k cos k , (12.71) 4 3 py (k) = sin k sin k , (12.72) 4 3 pz (k) = cos k . (12.73) 4 This is the so-called Balian-Werthamer state. 134
It is plausible that in a crystal the simultaneous presence of a non-vanishing k-even order parameter k and a non-vanishing k-odd order parameter d(k) would break spatial inversion symmetry. Thus in an inversionsymmetric crystal singlet and triplet superconductivity do not coexist. However, in crystals lacking inversion symmetry (noncentrosymmetric crystals), singlet and triplet pairing can be realized at the same time. Moreover, one can show that spin-orbit coupling in a noncentrosymmetric superconductor mixes singlet and triplet pairing. In this case, k and d(k) must both become non-zero simultaneously below Tc . Examples for such superconductors are CePt3 Si, CeRhSi3 , and Y2 C3 .
135

Theory of Superconductivity

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Theory of Superconductivity

Загружено:

Авторское право:

Доступные форматы

Theory of Superconductivity

Institute of Theoretical Physics

Version: February 6, 2012 A L TEX & Figures: S. Lange and C. Timm

This is a maximal list of topics to be covered, not necessarily in this order:

p (10 6 Pa) 3 solid phases

Bose-Einstein condensation in dilute gases

The fugacity has to be chosen to give the correct particle number N=

Since nk must be non-negative, must satisfy k k. (3.7)

Then 2V ln Z = 3 (2m)3/2 h with 2V N = 3 (2m)3/2 h Dening gn (y ) := 1 (n)

We note the identity gn (y ) = which implies gn (0) = 0, 1 gn (1) = = (n) kn

(3.17) for n > 1 (3.18)

gn ( y) g3/2 2 g5/2 1 0 0 0.5 1 y

Semiclassical theory of transport

The current density now reads ( ) eE d3 k k d3 k k d3 k k eE 0 j = e k + e ( k ) e . = 0 0 (2 )3 m (2 )3 m (2 )3 m k

The Second London Equation js = and the continuity equation

js = 0 can now be solved to give js (x) = c Bapl z ex/L 4L for x 0.

This equation is evidently not gauge-invariant since a change of gauge A A + (5.19)

Rigidity of the superuid state

Nonlocal response: Pippard theory

with the penetration depth

Landau theory of phase transitions

Ginzburg-Landau theory for neutral superuids

(x) + (x) + | (x)| (x) = 0.

f (x) + f (x) f (x)3 = 0.

This equation contains a characteristic length with 2 =

Ginzburg-Landau theory for superconductors

= a(A0 ) ((A0 )a)

Decomposing A into longitudinal and transverse components, (k Ak ) + Ak k (k Ak ) Ak = k

:= k/k , we obtain with k Ffree [A] = ( )( )] 1 [ k 2 Ak Ak + k 2 A k Ak k Ak k Ak 8 k ] 1 1 [ $ $ $$ $$ = k 2$ Ak Ak + k 2 A k 2$ Ak Ak = k 2 A k Ak $ k Ak . $ 8 8

Including the superconducting contribution, we get ] [ q2 q2 1 2 F [A] = A Ak A Ak + k Ak Ak . 2m c2 k 2m c2 k 8

[ ] [ ] 2 2 Hc Hc B (x)Hc + dx gs (x) + = dx fs (x) 8 4 8

)2 ] q A(x) i x c ( ) ( ) ] q q A(x) (x) A(x) i x c i x c (6.87)

Type-II superconductors are dened by a large Ginzburg-Landau parameter (6.89)

The magnetic ux through this loop is Stokes = da B = da ( A) =

With the second Ginzburg-Landau equation j=i we obtain =

(The minus sign is conventional.) Thus we nd for the so-called uxoid := mc e

) 1 2 2 iky y e i c x + m c x e f (x) 2m y 2 2 2 2 ky 1 d2 f 2 2 f i c x iky f + m c = 2 + x f 2m dx 2m ( 2 ) 2 2 2 ky d2 f ky 1 2 2 = 2 + m c x + 2 + f = f. 2 2m dx 2 m c (m )2 c

Then the harmonic oscillator is centered at x0 = 2 0 q= q. m c ay Hay (6.126)

Superuid and superconducting lms

We start from the two-dimensional Landau functional [ ] 2 4 2 F [ ] = d r | | + | | + ( ) . 2

This is a Gaussian average since F [] is bilinear in k . Thus we can write

Dening the vortex concentration by nv (r) :=

where Ri is the location of the center of vortex i, we obtain v vy vx = 2 nv (r). x y 54 (7.17)

We introduce elliptic coordinates , according to x= R , 2 R 2 y= 1 1 2, 2 (7.39) (7.40)

(7.51) is a vortex-pair fugacity, and (7.52)

the second term, d2 rj

|rn ri | |rj ri | 2 Nn Nj K ln 2 K ln r r n=i,j 2 ( r r ) ( r r ) r m j n j d2 rj 2r dr 1 + 2 K 2 Nm Nn = 2 2 | r r | | r r | m j n j m,n=i,j

(rm rj ) (rn rj ) 2 2 . |rm rj | |rn rj |

(rm rj ) (rn rj ) |rm rj | |rn rj |

and with the sum over i, j : | r r | 1 m n (N 1)2 Nm Nn ln = 2r dr N 2 V 2 3 K 2 r2 , 2 r

free vortices no condensate

is negligible close to Tc , the renormalized superuid density ns (T ) := K 0 n K0 s (7.82)

ln K0 is an analytic function of temperature and its value at Tc is [ ] 2 2 Cc = 2 2 y 2 ln K = 1 ln . K T =Tc

Thus we can write, close to Tc ,

jumps to a nite value at Tc and then increases like a square root.

On the other hand, from the London equation we have j= c A, 42 64 (7.93)

with a surface current density K. Thus

This eld is easy to obtain from symmetry: d a B = 0

2r2 B = 0 0 r for z > 0. 2r2

This gives a contribution to the eld energy of B2 2 0 d r = 8 32 3

Origin of attractive interaction

Reminder on Green functions

where . . . is the equilibrium average A = Tr Aeq = Tr AeH 1 Tr AeH , H Tr e Z (8.7)

Specically, we have GR (rt, r t ) =i (t t )