HAL Id: jpa-00246744
https://hal.archives-ouvertes.fr/jpa-00246744
Submitted on 1 Jan 1993
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Flux-quantization in one-dimensional copper-oxide models
C. Varma, A. Sudbø
To cite this version:
C. Varma, A. Sudbø. Flux-quantization in one-dimensional copper-oxide models. Journal de Physique
I, EDP Sciences, 1993, 3 (2), pp.585-589. �10.1051/jp1:1993125�. �jpa-00246744�
Classification
Physics
Abstracts05.30 74.70V 71.30
Flux-quantization in one-dimensional copper-oxide models
C. M. Varma and A.
Sudb#
AT&T Bell
Laboratories,
MurrayHill,
NewJersey
07974, U.S.A.(Received
19 October 1992,accepted
26 October1992)
Abstract. A calculation of the ground state energy
Eo(4l)
asa function of flux 4l in
a
one-dimensional'copper-oxide
model' onrings
of finite circumference L,including
on-site inter- actions, and a nearestneighbor
interaction V, ispresented.
For V of the order of thecharge-
transfer gap or
larger,
the modelextrapolated
tolarge
L exhibitsflux-quantization
withcharge
2e. The
extrapolated superfluid
stiffness appears,however,
finite only for not toolarge
V. These results suggest asuperconductive
state at V of order thecharge-transfer
gap of themodel,
buta
paired
andphase-separated
state atlarger
V.Three distinct electronic models have been
proposed
for thecopper-oxide
based metals withhigh-temperature superconductivity
and associatedphenomena.
Andersonsuggested
that the essentialphysical
features of the materials can be modelledby
the twc-dimensional(2D)
one-band Hubbard model
[Ii. Varma,
Schmitt-Rink and Abrahams(VSA) presented
a 2D modelconsisting
of three orbitals per unit cell[2], namely
the Cu3d~2-y2,
and oxygen2p~, 2py
withhopping
betweenthem,
arepulsive
interaction on Cusites, Ud,
and on oxygensites, Up,
as wellas a
nearest-neighbor
copper-oxygen interaction V.Independently, Emery presented
a three- band model[3],
in which theparameter
V was assumed toplay
no role. In reference[2],
it wasargued
that the three-band model with V included(hereafter
denotedas the VSA
model)
is the minimum model necessary to characterize the essentialproperties
of the materials. Inpartic- ular,
theparameter
V was shown to beimportant
ininducing low-energy charge-fluctuations
in the metallic state. It was also
argued
that such fluctuationspromote
asuperconducting instability
and morerecently
thatthey
may also lead to the anomalous normal metallic-stateproperties
of thehigh-Tc cuprates [4],
which at1/2-filling
arecharge-transfer
insulators. Herewe show that the
parameter
V is central ininducing pairing
at T= 0 in the one-dimensional
(ID)
version of the three-band model.The ID version of the VSA model may be written as H =
Ho
+Hu
+Hv,
whereHo
"T(16)
+ A~j [np,; nd,I],
;
Hu
=Ud~j
nd,;~ind~;~i +Up ~j
np~;,inp~;~i,(1)
; ;
586 JOURNAL DE
PHYSIQUE
I N°2Hv
=V~nd,I[np,;-i+np,;).
I
Here,
na,; =£~
n~~;~a=
£~ cl
; ~ca~;~a, where
~Y denotes an orbital index
(~Y E
(Cu, O)),
anda a
spin
index(a
E(i,1). Thr1~lghout, £;
is taken to run overunit-cells,
chosen such that the oxygen is to theright
of copper. We have introduced theparameter
A e(Ep Ed)/2,
with
Ep
>Ed
in a holenotation,
whereEp
andEd
are siteenergies
on the oxygen and coppersites, respectively,
and the zero of energy has been chosen at(Ep Ed)/2. Furthermore,
thequantity T(16)
is the kinetic energyoperator
of thesystem
in the presence of an external flux4l,
and will bespecified
below.The model of
equation (I)
withT(16)
= 0 isexactly
solvable forarbitrary
L in IDby
thetransfer-matrix method
[5].
For Udlarge compared
toUp
andA,
anddensity
away from1/2-filling, increasing
V totl(A)
leads to a combinedcharge-transfer
andphase-separation instability.
In thisinstability,
the averagecharge
on thep-orbitals
increases at the expense of thecharge
on d-orbitals. The former are eithernearly doubly occupied
orempty,
as reflectedby
thecompressibility tending
toinfinity.
Mean-field calculations also show such instabilities withfinite
kinetic energy [6], but in addition reveal aregion
ofparameters
where s-wavesuperconductivity
exists withoutbeing preempted by phase-separation.
A centralquestion
is whether one can see,through
exactcalculations,
that kinetic energy favourssuperconductivity
over the
charge-transfer/phase-separation
instabilities for anyfilling
N and in anyregion
of theparameter
space(Ud, Up,
V,A),
withpurely repulsive
bare interactions.To
explore
theseissues,
we have considered theground
state energyEo(16)
of the Hamiltonian ofequation (I),
in aring geometry
withL/2
number of Cu O unitcells, (L being
an eveninteger representing
the total number of sites in theproblem).
Thering
is threaded with adimensionless Bohm-Aharanov flux 16
by applying
a constant vectorpotential
ofmagnitude
A =
(hcle)16/L along
its circumference. The effect of the vectorpotential
A=
(hcle)4l/L
is included
by
a gauge tranformation(taking
h = e = c =I),
ca,m,a - c~~m~aexp(im4l/L).
Thus,
we haveT(4l)
=-tpd ~
(e~~/~ cl
; ~cp,;,a +
e~"/~c(
; ~cp,;-i,a +
h.cj (2)
I,a
From the
ground
state energyEo(4l)
one obtains information aboutflux-quantization
andpossible superconductivity:
In a normal one-dimensionalring,
theground
state energy is aperiodic,
even function of 4l withperiod
4lo +hcle.
If the flux 4l is measured in units of theflux-quantum 4lo,
we then haveEo(4l)
=Eo(4l
+n);
n =0,+1,+2,.. Hence,
for a normal metallicphase,
one then has stablephases
atparticular
values of theflux,
4l=
0,+1,+2.., leading
toflux-quantization
in units ofhcle [7].
In asuperconducting
state one further hasEo(4l)
=Eo(4l
+n/2)
in thethermodynamic
limit L - oo,showing
that new stablephases
appear also at 4l
=
+1/2, +3/2,.,
in addition to the ones at 4l=
0, +1,+2,..
in the normal metallic state. Insystems exhibiting superconductivity
flux is thusquantized
in units of4lo /2
=
hc/2e [7],
hereafter referred to as anomalousfluz-quantization.
Recently,
theground
state energyEo(4l)
for therepulsive
and attractive one-band Hubbardring
with finite circumference L and onsite interactionsonly,
was considered in detail[8].
Thismodel is
exactly integrable
forarbitrary
flux4l,
and exact results forground
stateproperties
are thus available for
arbitrary
L. For the attractive case, which is known to exhibitsinglet
superconductivity,
it was found that the results for thisparticular
case may be writtenFor the attractive
Hubbard-model, f
is alength
associated with a gap ofspin-excitations
in thesystem,
A =A(U, N)
is a coefficient whichdepends
on details of themodel,
U is the onsiteinteraction,
and N is the number ofparticles
on thering.
Anotherquantity
ofinterest,
which should be considered inconjunction
withEo(4l
+n/2) Eo(4l)
for the purpose ofdetecting
superconductivty,
is thesuperfluid
stiffness02[Eo(4~)/Li
(4) PS(~)
"d[4~/Lj2 '~"°'
This
quantity
shouldapproach
a finite constant as L - oo in asuperconducting
orperfectly conducting
metallic state.Equations (3)
and(4)
are usefuldiagnostic
tools whenlooking
forsuperconductivity
in the VSA modelequation (I),
with A =A(Up, Ud,
V,A, N).
We use
Lanczos-diagonalization
to obtain the T = 0eigenvalues
andeigenvectors
of the Hamiltonianequation (I).
Inconsidering
numericalapproaches
of thiskind,
whichby necessity
are limited in
system
sizeL,
one should bear in mind that states other thansuperconducting
ones may
produce
anomalousflux-quantization.
Forinstance,
it wasrecently
noted thatperi- odicity
inEo(4l),
withperiodicity 4lo/2,
is found in exactdiagonalization
studies of smallrings
of the
repulsive
one-band Hubbard model[9].
Wehave, however,
verified that no anomalousfluz-quantization
occurs in this model when states of agiven
totalspin
S are considered for all 4l. Thecrossing
in energy of states of differentspin
S occurs due todegeneracies
in finitesystems,
and associated Hund's rule effects which arenegligible
in thethermodynamic
limit L - oo. Paired states which arecrystalline (charge density waves)
orphase-separated
will also have atendency
to4lo/2-quantization,
but in these cases ps will decrease withL,
whereas ina
superconducting
orperfectly conducting
metallic state ps shouldapproach
afinite
constantas L
- oo.
Hence,
to test forsuperconductivity,
one must calculate bothEo(4l)
and ps, and consider their behavior withincreasing
L. In ourcomputations,
we have been limited insystem
sizes L < 12 sites forfillings
N >L/2
ofinterest,
I.e.systems doped beyond half-filling-
We now turn to the details of the calculations with the model of
equation (I).
The choice of theparameters
is made such that athalf-filling,
the model is acharge-transfer
insulatorwith a gap of about 2
(in
units oftpd),
with80$l
or more of thecharge residing
on the d- orbitals(corresponding
almost to aCu++
state in the Cu Omaterials).
WithUd
*6,
this is ensured for both A andV,
in a range between 0 and 4. In this range ofparameters,
asV
increases,
one obtains arapid
transfer ofcharge
fromcopper-orbitals
tooxygen-orbitals for systems doped beyond half-filling,
at the fewfillings
we can consider. This was also seen inmean field calculations
[6].
It is thisregion
ofparameters bordering
onphase-separation
which is theregion
of interest to us whensearching
forsuperconducting ground
states.Figure
I showsEo(4l)
obtained in the VSA-model withparameters
L =10,
N =8, tpd
"I,
Ud
"10, Up
=0,
and 2A +Ep Ed
" 2 for three values V =1.0,
V = 3.0 and V = 4.0.For the latter two
values,
thesystem
exhibits anomalousflux-quantization,
whereas it doesnot for V
= 1.0. A critical valve
of
V istherefore
needed to seepairing.
Thefilling
in thiscase
corresponds
to60$l doping beyond 1/2-filling.
For small V <I, Eo(4l)
is maximum at 4l =1/2,
and hasperiod I,
characteristic of a normal metallic state. As V isincreased,
alocal minimum
develops
at 4l =1/2
as a result of levelcrossing
between states of differenttotal momenta k =
2~m/L;
m=
0,+1,+2,...
It is found thatEo(1/2) Eo(0)
decreasesrapidly
as V isincreased,
consistent withequation (3):
As V increases the effective attractionon oxygen sites increases and
produces
moretightly
boundpairs [10].
Note that weexpect
that thepairing
inducedby
V will be morepronounced
when thedimensionality
D > I: the coordination number of the copper sites inHv, equation (I),
increases as D increases. The critical value of V necessary to inducepairing
will thus decrease when D increases. As the588 JOURNAL DE
PHYSIQUE
I N°2L=I$ N=8 O V
= I.O
0.05
Ud"10,U~=0
~ ~/"3.°A=2.0 ~/"4.0
j~
jJ j
j '
, j j
,
j O
i
. '
, o
.
, uJ
~
J
=
'
J
/
~
',
i
'~
/ l
J
u~
', ~__-~ A/ i
Q
i,
~, J
j
/ r"
x 0 125
',
j
p
Q ,
J /
',
, J
,~
, /J O
,,
'4~-O~~
J,'° °~
'UT'
o o.5 1.o
if 2n
Fig.I.
Ground state energyEo(4)
us. flux 4 for the 1D VSA model with L= 10, N
= 8, Up = 0, Ud = 10, 2a
= 2, and tpd = for various values of V. Note the
tendency
tohalfing
of theperiodicity
of
Eo(4),
characteristic offlux-quantization
withcharge
2e, when V= 3.0 and V
= 4.0 but not at
V = 1.0. Note also the
rapid
decrease of ED(1/2)
ED(0)
as V increases, consistent With a reduction of theCooper-pair
size v4thincreasing
V. The thin broken lines areguides
to the eye.onsite Coulomb
repulsion
on oxygen sites Up isincreased,
the effective attraction on oxygen due to V is reduced.Pairing
isconsequently suppressed,
and is notpresent
whenUp
mUd.
The I0-site chain with N
= 8 shows anomalous
flux-quantization
for alarge
range of V.Furthermore, by inspecting
thedispersion
of theground
state energyEo(k,
4l=
0),
it is seen that thesystem
maintainsquasi-particles
with mass within a factor of 2(I.e. Eo(k,4l
=0)
remains
dispersive)
from that at V = 0 in the entire range of 2 < V < 4 we haveconsidered,
and where
4lo/2-quantization
is observed. These areencouraging signs
of asuperconducting ground
state.However,
the 10-site chain is not convenient forcarrying
out the finite-sizescaling analysis
mentionedpreviously.
We therefore next consider a 6-site and a 12-sitechain,
which also allows twosystems
with identicalfillings
to beinvestigated.
For the 6-site chain we have used N =4,
while for the 12-sitechain,
we have used N =8,
whichcorresponds
to33$l doping beyond
halffilling.
Thefilling
is lower than in the case of the I0-sitechain;
observation of anomalousflux-quantization
thusrequires larger
values of V.For
Ud
"6.67, Up
=0,
2A =1.33,
and V=
2.67,
we haveAE~(L
=
6)/AE~(L
=12)
= 0.3 andps(L
=12)/ps(L
=6)
=0.67,
whereas for V=
4,
the former is 0.07 while the latter is 0.25.Here,
we have defined1hE~
eEo(16
=~) Eo(16
=0).
For asuperconductive ground
state and for L »f,
ps should scale to a constant whileAE~
should decreaseexponentially
withlarge
L. For V = 2.67 we findusing equation (2)
thatf
m6,
so we are not in theasymptotic
regime
and the small decrease of ps withincreasing
L is consistent with asuperconducting
state.
However,
for V=
4,
we havef
m2;
in this case thesharp
decrease of ps with L issuggestive
of apaired
butnon-superconducting
state.Furthermore,
fromEo(k,16
=
0)
forL =
12,
N =8,
the lowest energy state is seen to beessentially dispersionless,
consistent with apinned
CDW or apaired/phase-separated state,
the latter alsopreviously
indicatedby
mean-field calculations
[6].
In thiscontext,
we note that for L = 12 and V2 4,
we find that the chemicalpotential
~ =Eo(N
+I) Eo(N)
at N = 8 is close to its value athalf-filling,
asignature
ofphase-separation.
Thissuggests
asuperconducting
state forparameters locating
the
system
close to aphase-separation instability.
The above results were
presented by
one of us(C.
M.V.)
on the occasion of the memorialsymposium honoring
R. Rammal.They
arepart
of aninvestigation
carried out with collabc- ration of T.Giamarchi,
R. T.Scalettar,
and E. B. Stechel.They will, together
with furtherdetails,
also bepublished
elsewhere.Acknowledgements.
I
(C.
M.V.)
wish to thank Benoit Doucot forgiving
me theopportunity
to payhomage
to my friend Rammal. Igot
to know Rammal very wellduring
the year he was kindenough
toaccept
my invitation to visit the TheoreticalPhysics
ResearchGroup
at Bell Labs. Besides adeep
interest in all manners of
science,
we shared thespecial sensitivity
andperception
which comes fromliving
in ourrespective
societies as outsiders.Through
Rammal I learnt much about theacademic and scientific culture of France which he
deeply loved, although
in a bitter-sweet fashion.References
[1] P. W. Anderson, Science 235~ 1196,
(1987).
[2] C. M.
Varma,
S.Schmitt-Rink,
and E.Abrahams,
Solid StateCommun.,
62, 681(1987).
[3] V. J.
Emery, Phys.
Rev. Lett. 58, 2794(1987).
[4] A. E. Ruckenstein and C. M.
Varma, Physica
C 185-189, 134(1991).
[5] A.
Sudb@,
S. Schmitt-Rink, and C. M.Varma, Phys.
Rev. B 46, 5548(1992).
[6] P. B. Littlewood, Phys. Rev. 842, 10075
(1990);
S. N.Coppersmith
and P. B. Littlewood, ibid. B 41, 2646(1990)j
M. Grilli et. al., Phys. Rev. Lett. 67, 259(1991);
C. Castellani et. al., Phys. Rev.843,
8000(1991).
[7] N. Byers and C. N.
Yang, Phys.
Rev. Lett. 7, 46(1961).
[8] B. S.
Shastry
and B.Sutherland, Phys.
Rev. Lett. 65, 243(1990);
C. A.Stafford,
A. J. Millis, and B. S.Shastry, Phys.
Rev. B 44, 13660(1991).
[9] R. M.