Skip to main content

Full text of "Unconventional particle-hole mixing in the systems with strong superconducting fluctuations"

See other formats

Unconventional particle-hole mixing in the systems 
with strong superconducting fluctuations 

















T. Domanski 

Institute of Physics, M. Curie Sklodowska University, 20-031 Lublin, Poland 

(Dated: October 7, 2008) 

Development of the STM and ARPES spectroscopies enabled to reach the resolution level sufficient 
for detecting the particle-hole entanglement in superconducting materials. On a quantitative level 
one can characterize such entanglement in terms of the, so called, Bogoliubov angle which determines 
to what extent the particles and holes constitute the spatially or momentum resolved excitation 
spectra. In classical superconductors, where the phase transition is related to formation of the 
Cooper pairs almost simultaneously accompanied by onset of their long-range phase coherence, the 
Bogoliubov angle is slanted all the way up to the critical temperature T c . In the high temperature 
superconductors and in superfluid ultracold fermion atoms near the Feshbach resonance the situation 
is different because of the preformed pairs which exist above T c albeit loosing coherence due to the 
strong quantum fluctuations. We discuss a generic temperature dependence of the Bogoliubov angle 
in such pseudogap state indicating a novel, non-BCS behavior. For quantitative analysis we use a 
two-component model describing the pairs coexisting with single fermions and study their mutual 
feedback effects by the selfconsistent procedure originating from the renormalization group approach. 


Such vastly distinct systems as the classical and/or 
high T c cuprate superconductors, the ultracold superfluid 
fermion atoms as well as certain cosmological (superfluid 
neutron stars) and even subatomic objects (odd-odd nu- 
clei) reveal signatures of ideally coherent pairs consisting 
of particles from a vicinity of the Fermi surface. Obvi- 
ously, what differs one case from another is an under- 
lying mechanism and energy scale engaged in the pair- 
ing. They all however share the universal feature re- 
lated to the effective Bogoliubov quasiparticlcs represent- 
ing a superposition of the fermion particles and their ab- 
sence. This emerging particle-hole (p-h) mixing [1] has a 
purely quantum nature (imposed by the structure of the 
BCS wave function) which to some extent resembles the 
corpuscular-wave dualism. One of its spectacular mani- 
festations is the mechanism of subgap Andreev reflection 
where an incident fermion-particlc can convert into the 
pair with a simultaneous reflection of the fermion-hole 
what is indeed observed experimentally in superconduc- 
tors [2] , for the relativistic-like particles [3] and in quan- 
tum dots attached to superconducting electrodes [4, 5]. 

In the recent papers A. Balatsky and coworkers have 
emphasized that p-h mixing can be quantitatively probed 
by the present-day STM [1] and ARPES spectroscopies 
[6] . These techniques are capable to determine either the 
spatially [7] or momentum resolved [8] single particle ex- 
citation spectra of superconductors. In principle also the 
simultaneous k- and r-space measurements are feasible 
by means of the Fourier transformed quasiparticlc inter- 
ference imaging [9]. Roughly speaking, the p-h mixing 
manifests itself in the single particle spectra by appear- 
ance of two peaks around the Fermi level separated by 
twice the (pseudo)gap and whose spectral weigths yield 
the information on particle/hole contributions to the Bo- 
goliubov quasiparticles. 

Usually for conventional superconductors these con- 

tributions are given by the BCS coefficients u k and 
= 1 — tt£, so it is convenient to define the Bogoliubov 
angle [10] 

2 arctan f 7—^ 7 

2 V k 


as a measure of the particle-hole mixing. Its magni- 
tude can vary between —tt/2 and 7r/2 depending on a 
momentum and indirectly on temperature. #k has a 
particularly clear interpretation in the pseudospin rep- 
resentation S k ,z = i (l- 4:t^T _6 -k4.C-kJ.)> $M(») = 
2^7 (^k|^-ki+ ( — )c-kj.Ckf^ introduced by P.W. Ander- 
son [11], where it denotes an azimuthal angle of the vec- 
tor (§k) ■ Restricting to the part of Hilbert space where 
(c k |Ck|) = (cL k |C_ki) the pseudospin eventually points 
down (up) when effective quasiparticles are represented 
by particles (holes). The upper and bottom panels of fig- 
ure 1 illustrate such behavior well known for the normal 
and superconducting states [11]. 

In general, pseudospins obey the non-trivial dynam- 
ics governed by the Bloch-type equations of motion [11]. 
This aspect has a particular importance in the context 
of ultracold atoms where traversing through the Fesh- 
bach resonance can lead to the soliton-like solutions [12]. 
On the other hand, in the highly inhomogeneous cuprate 
superconductors with pairing on a local (interactomic) 
distance both the excitation spectrum [7] and the Bogoli- 
ubov angle are strongly varying in space. Such issue has 
been already explored within the Bogoliubov de Gcnncs 
approach and results were confronted with the available 
STM data [1]. 

Since the Bogoliubov angle (1) is sensitive to existence 
of the paired fermions one may ask if any signatures of the 
p-h mixing would be able to appear above T c . ARPES 
studies [13] confirm that the superconducting state of 
cuprates obeys roughly the usual BCS behavior but there 




T > T 

T <T<T 

c p 


T < T 


FIG. 1: Variation of the Anderson's pseudospin (the left h.s. column) and the Bogoliubov angle #k (the right h.s. column) 
against momentum in the normal, pseudogap and superconducting states. Notice that particle-hole mixing is present in the 
superconducting and pseudogap states, however above T c the Bogoliubov angle becomes discontinuous at kF. 

is still no firm agreement on the nature of pseudogap state 
and its relation to superconductivity [14]. Nevertheless, 
various experimental data [15, 16] seem to indicate that 
preformed fermions' pairs are present already in the nor- 
mal state (at least in the underdoped samples) at temper- 
atures up to dozen Kelvin above T c . Transition temper- 
ature might there correspond to the onset of long-range 
phase coherence [17]. Another evidence of the preexist- 
ing pairs above T c is known for the ultracold atoms of Li 6 
and K 40 . Near the Feshbach resonance the weakly bound 
boson molecules are scattered into the Cooper-like pairs 
and such unitary limit is in a crossover between the BCS 
and BEC regimes beeing influenced by strong quantum 
fluctuations [18]. 

Our purpose here is to explore the impact of preformed 
pairs on the Bogoliubov angle in the pseudogap state. 
In particular, we address the question whether p-h mix- 
ing can at all show up above T c and if so, then how it 
would manifest itself. For the considerations we use a 
phenomenological two-component model [19] where itin- 
erant fermions and their paired counterparts arc intro- 
duced without referring to any specific microscopic mech- 

anism. From the selfconsistcnt treatment of interactions 
between the paired and single fermions we find the evi- 
dence of particle-hole mixing signified by |#k| 7 z ^ 7r /2. Fur- 
thermore, lack of the phase coherence above T c leads to 
a discontinuity of 6>k at hp. We will show that in the 
pseudogap state the Bogoliubov angle behaves in a man- 
ner which partly resembles the normal and partly the 
superconducting phases (see figure 1). 

In the next section we briefly introduce the model and 
discuss its main properties. Methodological details are 
presented in section III and the essential part on the p- 
h mixing for the pseudogap state is described in section 
IV. We finally summarize our results and point out some 
related unresolved problems. 


For modelling the pseudogap state we use the following 
Hamiltonian [19] 

H = ^(£k-A*)cLck CT + ^(£ q -2/i)fe q 6 q + -L X]( 5k -^q 6t i-klCkT+5k,q4TC q -k^q) 

k,<7 q V/V k,q 


where operators refer to annihilation (creation) of to the local pairs of energy _E q . Potential of the interac- 
single fermions with the energy £k and tt^ correspond 


tion between the single and paired fermions is denoted by 
5k. q- For simplicity, we shall assume that concentration 
of pairs per lattice site is small enough so that S q ^ obey 
the usual bosonic commutation relations (we neglect the 
hard-core effect). 

This model (2) has been invented [19] and explored by 
J. Ranninger with coworkers [20] and independently by 
T.D. Lee et al [21] as well as some other groups. Start- 
ing from various microscopic models several authors [22] 
have also concluded that the relevant physics of strongly 
correlated cuprates is well captured by the fermion and 
boson degrees of freedom expressed by the Hamiltonian 
(2). Moreover, such model well describes the ultracold 
fermion atoms interacting with the Feshbach resonance 
[18, 23]. 

In the simplest mean-field approach one can linearize 
the interaction term so that the decoupled boson and 
fermion parts become exactly solvable [19]. The re- 
sulting spectrum of fermions has then BCS structure 
A MF (k, ui) = ul_S(uj - E k ) + vIS(uj + Ek) with the usual 
quasiparticle energy Ek — yj (e k — A*) 2 + A 2 , and coher- 
ence factors u k ,f k = % [1 ± (e k — fj)/Ek] which lead to 
the standard Bogoliubov angle. Energy gap of the sin- 
gle particle excitation spectrum is effectively given by 
Ak = gk,o\/ (tIq )• This means that fermions undergo 
transition to the superconducting state if and only if the 
Bose-Einstein condensation of bosons takes place [19]. 
Actually, the latter property is valid exactly [24] without 
limitations to any approximation. 

The mean-field treatment does not take into account 
the quantum fluctuations whose efficiency increases upon 
approaching T c and above of it. In the next section we 
present the method which enables a selfconsistent study 
of the boson-fermion feedback effects. In particular, we 
will analyze the remnants of superconducting correlations 
above T c and study their effect on the Bogoliubov angle. 


For studying the model (2) we use the selfconsistent, 
non-perturbative procedure based on a canonical trans- 
formation H — ► e s ^ He~ s ^ with a continuous formal 
parameter I [25]. The main idea is to eliminate the inter- 
action part <7k,q through a sequence of infinitesimal steps 
I — > 1 + 51. Proceeding along the lines of the Renormaliza- 
tion Group (RG) technique one starts from renormalizing 
the high energy sector and subsequently turns to the low 
energy sector (by latter we mean the fermion states close 
to \i and boson states near 2/x). We briefly describe some 
technicalities in order to clarify how the particle and hole 
spectral contributions can be evaluated within this pro- 

Practically we start by setting H(l) = e^He~ S{ - l \ 
where H (0) corresponds to the initial Hamiltonian, and 
then construct the flow equation diH(l) — [fj(l), H(l)] 
with the generating operator fj(l) = diS(l). Following 


FIG. 2: The single particle excitation spectrum of fermions 
in the pseudogap regime. Besides the long-lived quasiparticle 
at o; = e k — /i there emerges its mirror reflection corresponding 
to the damped Bogoliubov shadow branch whose presence 
has been confirmed by the recent ARPES measurements [16] . 
Both branches are separated by the pseudogap which vanishes 
at Tp ^> Tc 

the original proposal of Wegner [25] we choose fj(l) = 
[H (l), Hi n t (I)], where H (l) denotes the total kinetic 
energy of fermions and bosons whereas H int {l) stands 
for their interaction. From a straightforward algebra 
we obtain fj(l) = £ kq a k , q (0 (& q c q _ ki c kT - h.c.) 

with ak, q (0 = (e k (0 + £q-k(0 - Eq(l)) ffk, q (0- One can 
prove analytically [26] that such antihermitean operator 
f)(l) indeed guaranties an asymptotic disappearance of 
the boson-fermion coupling lin^oo 3k, q (0 = 0- 

Applying this scheme to the boson-fermion Hamilto- 
nian (2) we obtain the following set of coupled flow equa- 
tions [26] 

%k, q (0 = -a k , q (%k, q (0 (3) 

dl£ ^ ^ |E^. q (0lffk, q (0| 2 < B) (4) 

wo = !E a k- q ,k(oi.9k- q , k (oi 2 


x (-l + 4- ) qi +^T ) -) (5) 

We have solved them numerically considering fermions 
coupled with bosons on a lattice avoiding thus any need 
for the infrared cutoffs. The fixed point values 

lim Sk{l) = ik, lim £7 q (/) = £ q (6) 

turned out to reveiled the following features: 

(a) for T <T C the renormalized fermion dispersion e k 
develops a true gap at \x which evolves into a pseu- 
dogap for T c <T<T p , 


(b) the effective boson dispersion _E q shows the long- 
wavelength Goldstone mode for T <T C and its rem- 
nants are preserved even in the pseudogap state 

For a complete information about the fermion and bo- 
son spectra we need to proceed with transformations 

e s(i) d W e -s(i) and 

for the individual operators c k ^(Z) 

bQ\l) = e^Wb^e ^ which is a rather difficult task be- 
cause S(l) is not known explicitly. Since our primary 
interest is in estimating the particle-hole mixing for the 
single particle fermion excitations we focus on the flow 

equation <9;c k ^(0 — [fj, c k T j(/)]- The generating operator 
fj(l) chosen according to Wegner's prescription [25] yields 
the following ansatz for fermion operators [27] 

Ckt(i) = u k (l) c kT + v k (l) cl ki (7) 





X(0 CkT c -ki 

^]&Y1 [~ W k,q(0 & q C q+kT + <,q(0 & 


q C q-k| 

where u k (0) = 1 and all other coefficients are vanishing 
at 1 = 0. The ^-dependent coefficients must be determined 
from the following set of flow equations [27] 

diUk(l) = a_k,o(0 v k (l) 





a k ,o(0 «k(0 (10) 

J! "k,q(0 « + <+k T ) «k,q(0. 

a-k,q(0 Wk(0. (11) 

- a k , q (0«k(0- (12) 

We explored them numerically along with the equations 
dis k (l), diEq(l), dig k ,q(l) on the 2-dimensional square 
lattice with the initial (I = 0) tight-binding dispersion 
£k(0) = — 2t (cos(k x a) + cos(k y a)) and the localized bo- 
son energy -E q (0) = Eq. Moreover, we imposed g k ,q(0) = 
g (cos(fc x a) — cos(fc y a)) to obtain the d-wave symmetry of 
energy gap (pseudogap) below (above) T c . We solved the 
coupled flow equations iteratively by the Rungc-Kutta 
method for E (0) = 0.2t keeping a fixed charge con- 
centration ntot = 2 when the concentration of fermions 
n F = 1 + x yield the realistic value x <~ 0.1. In figures 
2-4 we present the results obtained along the antinodal 
direction (0,0) <-> (tt,0) i.e. for k y = 0. 

Our ansatz (7,8) generalizes the standard Bogoliubov- 
Valatin transformation by including the effect of scatter- 
ing on finite momentum preformed pairs. Influence of 
such scattering shows up in the effective single particle 
spectral function which takes the following form 

A(k,uj) = \u k \ 2 5(u)+n-s k ) + j^Yl ( n q + n q+kt) l"k,q| 2 £(w+M-£q+k+-E , q) 


+ \i k \ 2 5 (w-/i+£_ k ) + ( n q + n q-k|) \v k ^\ 2 5(uj-fi+i q - k -E cl ), 



where u k , v k and Uk,q, ^k,q denote the asymptotic I — ► oo 
values. We have determined them numerically solving 
the flow equations (9-12) for the fixed total charge con- 
centration ntot = 2 J2q U q + J2 k ( n kT + n k|) • 

The structure of spectral function (13) indicates 
that besides the narrow peaks (long-lived states) there 
also forms a background of the damped (finite life- 
time) states. If we neglected ?ik, q and v k <i then 
the flow equations (9,10) would simplify to diu k (l) = 


mean-field solution u k ,v k = . .. ^ 

V V( e k-M) 2 +"o lflk,o| 2 / 

In order to go beyond this BCS solution we need to take 
into account the effect of scattering on the finite momen- 
tum pairs affecting the spectral function (13) through the 
coefficients Uk, q and «k, q - We shall do it for T>T C . 

y^qU a_ k ,o(0wk(0 and div k (l) = - ^n~^ a k , (l)u k (l) IV. PARTICLE-HOLE MIXING ABOVE T c 

yielding the invariance |w k (0l + l w k(0l = 1- By 

rewriting the first equation as f " k /„°° 7 Uk , dUk( -^ — = 
^ 5 H J« k (o)=i ^/i-K(i)|2 

/ °° ee- ki o(l)dl we then right away reproduce the 


Preformed pairs occupy in the normal phase only the 
finite momenta states (in other words (&q=o) = 0) there- 
fore the equations (10, 11) imply v k (l) = and u k ^(l) = Q. 




FIG. 3: Spectral function A(k,uj) which consists of the long- 
lived states (we have artificially broadened the delta peak into 
Lorentzian using the units marked on the left axis) and the 
damped fermion states (labels on the right h.s. axis) slightly 
below k.F for T = 0.004D. The weight of particle peak 
|wk F | 2 — 0.47 whereas the hole weight |«k F | 2 — 0.19 (the 
shaded area) is estimated by subtracting the high tempera- 
ture background. 

The ansatz (7,8) is thus above T c simplified to 

Ckt(0 = Uk(0PkT + "7= ^2 v k,q{l) Mq-14 ( 14 ) 
cLki(0 = -^53 «£,„(/) & q Cq + k T + <(0C-kl 



and the corresponding spectral function becomes 

A(k, u) = |u k | 2 5 (w+Ai-£k) (16) 

+ ("q + n q-kl) |Wk,q| 2 ^(w-M + £q-k--Bq)- 


The first term in (16) represents the long-lived states at 
the renormalized energies e^—fi whose spectral weight is 
|t2k| 2 < 1- Remaining part of the spectrum is distributed 
among the damped fermion states. Most of them are al- 
most insensitive to temperature and can be regarded as 
an incoherent background. However, there is a certain 
fraction (very important to us) of a different character - 
these states emerge around u>~ — (ek^ A*) near the Fermi 
surface as shown in figure 2. Such partly broadened ex- 
citation branch, being sort of a mirror reflection of the 
quasiparticle dispersion £k — corresponds to the hole 
(particle) contribution for momenta below (above) kp. 
These ingredients allow us to estimate the Bogoliubov 
angle (1) in the pseudogap state and our procedure for 
determining the particle and hole weights is illustrated 
in figure 3. 

To support this treatment we recall some analytical ar- 
gument explaining appearance of the Bogoliubov shadow 
branch upon approaching T c from above. For a decreas- 
ing temperature the preformed pairs start populating the 
lower and lower energies so that is dominated by the 
states located just above E^. The resulting spectral 




(k-k F )a 


FIG. 4: Variation of the Bogoliubov angle estimated in the 
pseudogap state for temperatures T = 0.004 (the solid line), 
0.007 (the short-dashed curve) and 0.012 (the long-dashed 
line). For comparison we plot by open circles the character- 
istics of superconducting state for T = 0. 

function (16) reduces then to 

A(k,u>) ~ |u k | 2 <5 (w + /i-e k ) + |w k | 2 
+ A inc (k,u)) 

r k /7T 

(u;- M +£k) 2 r k 

where the last term describes solely the structureless in- 
coherent background. We obtained |«k| 2 by integrating 
the spectral function with respect to u> for a given T and 
subtracting from it the integrated spectral function for 
high temperatures. 

We notice that near kp the long-lived state and its 
mirror reflection (a shadow) do not merge because of a 
finite value of the pseudogap A pg . Figure 4 shows the 
calculated Bogoliubov angle as a function of momentum 
measured with respect to the Fermi surface. In the pseu- 
dogap region the shadow branch has a substantial effect 
on the Bogoliubov angle leading to the p-h mixing near 
kp. Yet, exactly at the Fermi surface the Bogoliubov 
angle is discontinuous. The BCS-typc behavior is finally 
recovered at temperatures T<T C as marked by the open 
circles in figure 4. Since the magnitude of superconduct- 
ing gap does not much change [28] the Bogoliubov angle is 
below T c practically frozen (temperature-independent). 


We have analyzed the effect of strong superconducting 
fluctuations above the transition temperature [29] where 
the single fermions coexist and interact with the pre- 
formed pairs. Influence of pairs on the single particle 
excitation spectrum has been studied within the selfcon- 
sistent RG-like method [25]. We have found that near 
kp the renormalized dispersion e k is depleted above T c 
and additionally there appears a shadow branch in the 
fermion spectrum responsible for the particle-hole mix- 
ing. We have estimated the particle and hole spectral 
weights thereby determining the Bogoliubov angle for the 
normal state with preformed pairs. 


We have found that momentum dependence of the Bo- 
goliubov angle in the pseudogap regime differs qualita- 
tively from its behavior for the normal and supercon- 
ducting states. In the normal state (where particle-hole 
mixing is absent) #k changes abruptly at from —n/2 
to ir/2 whereas in the superconducting state (below T c ) 
the Bogoliubov angle continuously evolves between these 
extreme values over an energy regime ~ A sc , so that par- 
ticle and hole excitations are mixed with one another. In 
the pseudogap regime we find that |6* k | ^ ir/2 but still 

at the Fermi surface the Bogoliubov angle is discontinu- 
ous. We hope that STM and ARPES techniques would 
be able to detect such unconventional relation between 
the particle and hole weights predicted for the systems 
with strong pairing fluctuations. 

Author acknowledges valuable discussions with J. Ran- 
ninger, R. Micnas and F. Wegner. This work is partly 
supported by the Ministry of Science and Education un- 
der the grant NN202187833. 

K. Fujita, I. Grigorenko, J. Lee, M. Wang, J.X. Zhu, J.C. 
Davis, H. Eisaki, S. Uchida, and A.V. Balatsky, Phys. 
Rev. B 78, 054510 (2008). 

G. Deutscher, Rev. Mod. Phys. 77, 109 (2005); G.E. 
Blonder, M. Tinkham, and T.M. Klapwijk, Phys. Rev. 
B 25, 4515 (1982). 

C.W.J. Beenakker, Rev. Mod. Phys. (2008) in print. 
T. Domahski and A. Donabidowicz, Phys. Rev. B 78, 
073105 (2008); T. Domahski, A. Donabidowicz, and K.I. 
Wysokihski Phys. Rev. B 76, 104514 (2007). 

G. E. Graber, M. Tinkham, and T.M. Klawijk, Notech- 
nology 15, 479 (2004). 

A.V. Balatsky, W.S. Lee, and Z.X. Shen, cond- 
mat/0807.1893 (preprint). 

O. Fisher, M. Kugler, I. Maggio-Aprile, Ch. Berthod and 
Ch. Renner, Rev. Mod. Phys. 79, 353 (2007). 
A. Damascelli, A. Hussain, and Z.X. Shen, Rev. Mod. 
Phys. 75, 473 (2003). 

Y. Kohsaka, C. Taylor, P. Wahl, A. Schmidt, J. Lee, K. 
Fujita, J W. Alldredge, K. McElroy, Jinho Lee, H. Eisaki, 
S. Uchida, D.-H. Lee, and J.C. Davis, Nature 454, 1072 

In reference [1] the Bogolubov angle is defined as # k = 

arctan ^ but we use a different notation adopted from 

the Anderson's pseudospin representation [11]. 

PW. Anderson, Phys. Rev. 112, 1900 (1958). 

R.A. Barankov and L.S. Levitov, Phys. Rev. A 73, 

033614 (2006); A.V. Andreev, V. Gurarie and L. Radzi- 

hovsky, Phys. Rev. Lett. 93, 130402 (2004). 

H. Matsui, T. Sato, T. Takahashi, S.-C. Wang, H.-B. 
Yang, H. Ding, T. Fujii, T. Watanabe, and A. Matsuda, 
Phys. Rev. Lett. 90, 217002 (2003). 

P.A. Lee, N. Nagaosa, X.G. Wen, Rev. Mod. Phys. 78, 
17 (2006); M. Eschrig, Adv. Phys. 55, 47 (2006); M. 
Norman, D. Pines, and C. Kallin, Adv. Phys. 54, 715 

Z.A. Xu, N.P. Ong, Y. Wang, T. Takeshita, S. Uchida, 
Nature 406, 486 (2000); J. Corson, R. Mallozzi, J. Oren- 
stein, J.N. Eckstein, I. Bozovic, Nature 398, 221 (1999). 

[16] A. Kanigel, U. Chatterjee, M. Randeria, MR. Norman, 
G. Koren, K. Kadowaki, and J. Campuzano, Phys. Rev. 
Lett. 1001, 137002 (2008). 

[17] V.J. Emery and S.A. Kivelson, Nature 374, 434 (1995). 

[18] W. Ketterle and MW. Zwierlein, Making, probing and 
understanding ultracold Fermi gases, in Proc. Int. School 
Phys. "Enrico Fermi", course CLXIV, edited by M. In- 
guscio, W. Ketterle, and C. Salomon, IOS Press, Ams- 
terdam (2008). 

[19] J. Ranninger and S. Robaszkiewicz, Physica B 135, 468 

[20] R. Micnas, J. Ranninger and S. Robaszkiewicz, Rev. 
Mod. Phys. 62, 113 (1990); for a recent review see also 
the paper by J. Ranninger and A. Romano, Phys. Rev. 
B 78, 054527 (2008). 

[21] R. Friedberg and T.D. Lee, Phys. Rev. B 40, 423 (1989). 

[22] J.A. Wilson, Philos. Mag. 84, 2183 (2004); E. Altman 
and A. Auerbach, Phys. Rev. B 65, 104508 (2002); M. 
Franz and Z. Tesanovic, Phys. Rev. Lett. 87, 257003 
(2001); C. Lannert, M.P.A. Fisher and T. Senthil, Phys. 
Rev. B 64, 014518 (2001); V.B. Geshkenbein, L.B. Ioffe 
and A.I. Larkin, Phys. Rev. B 55, 3173 (1997). 

[23] Q. Chin, J. Stajic, S. Tan, and K. Levin, Phys. Rep. 412, 
1 (2005). 

[24] T. Kostyrko and J. Ranninger, Phys. Rev. B 54, 13105 
(1996); Y. Ohashi and A. Griffin, Phys. Rev. Lett. 89, 
130402 (2002). 

[25] F. Wegner, Ann. Physik 3, 77 (1994); S. Kehrein, Flow 
Equation Approach to Many Particle Systems, Springer 
Tracts in Modern Physics 217, Berlin (2006). 

[26] T. Domahski and J. Ranninger, Phys. Rev. B 63, 134505 

[27] T. Domahski and J. Ranninger, Phys. Rev. Lett. 91, 

255301 (2003); T. Domahski and J. Ranninger, Phys. 

Rev. B 70, 184503 (2004). 
[28] A. Kanigel, U. Chatterjee, M. Randeria, M.R. Norman, 

S. Souma, M. Shi, Z.Z. Li, H. Raffy, and J. Campuzano, 

Phys. Rev. Lett. 99, 157001 (2007). 
[29] A. Schmid, Z. Phys. 231, 324 (1970).