of 6
January 2009
EPL,
85
(2009) 17004
www.epljournal.org
doi:
10.1209/0295-5075/85/17004
Observation of vortices and hidden pseudogap from scanning
tunneling spectroscopic studies of the electron-doped cuprate
superconductor Sr
0
.
9
La
0
.
1
CuO
2
M. L. Teague
1
,A.D.Beyer
1
,M.S.Grinolds
1
,S.I.Lee
2
and
N.-C. Yeh
1(a)
1
Department of Physics, California Institute of Technology - Pasadena, CA 91125, USA
2
National Creative Research Initiative Center for Superconductivity, Department of Physics, Sogang University
Seoul, Korea 121-742
received 2 September 2008; accepted in final form 27 November 2008
published online 7 January 2009
PACS
74.50.+r
– Tunneling phenomena; point contacts, weak links, Josephson effects
PACS
74.25.Op
– Mixed states, critical fields, and surface sheaths
PACS
74.72.Jt
– Other cuprates, including Tl- and Hg-based cuprates
Abstract
– We present the first demonstration of vortices in an electron-type cuprate super-
conductor, the highest
T
c
(=43K) electron-type cuprate Sr
0
.
9
La
0
.
1
CuO
2
. Our spatially resolved
quasiparticle tunneling spectra reveal a hidden low-energy pseudogap inside the vortex core and
unconventional spectral evolution with temperature and magnetic field. These results cannot
be easily explained by the scenario of pure superconductivity in the ground state of high-
T
c
superconductivity.
Copyright
c
©
EPLA, 2009
Introduction. –
The physical properties of cuprate
superconductors exhibit dramatic contrasts between the
electron- and hole-type systems [1,2]. Various anomalous
phenomena observed in underdoped and optimally doped
hole-type cuprates above the superconducting transition
T
c
, such as the low-energy pseudogap (PG) and Fermi
arcs occurring at
T
c
<T<T
, where
T
denotes the
“pseudogap” (PG) temperature [3–5], are conspicuously
absent in the electron-type [1,2,6–8]. Among various
theoretical attempts to account for these unconventional
and non-universal physical properties of the cuprates, one
schoolofthoughtisbasedonthe“one-gap”or“preformed
pair”scenario[3,9].Thisscenarioassertsthattheonsetof
Cooperpairformationoccursat
T
andthatthePGstate
at
T
c
<T<T
is a disordered pairing state with strong
superconducting phase fluctuations. In this context,
superconductivity below
T
c
would represent a phase
coherent state of the preformed pairs. The other school
of thought entertains the possibility of competing orders
(COs) coexisting with superconductivity (SC) [10–13]
so that both order parameters are responsible for the
low-energy quasiparticle excitations. The COs may be
established at a temperature different from
T
c
, thereby
accounting for the PG phenomena. To date numerous
(a)
E-mail:
ncyeh@caltech.edu
experimentalfindings[14–18]andquantitativeanalysesof
the experimental data [2,17–19] seem to favor the “two-
gap” scenario. Nonetheless, the debate continues because
the phenomenology of coexisting COs and SC does not
directly provide a microscopic pairing mechanism that
involves the interplay of COs and SC. On the other hand,
there is no obvious theoretical reason to preclude the
occurrenceofCOinstabilitiesinthepresenceofpreformed
pairs. In fact, if COs coexist with coherent Cooper pairs
in the SC state, they may also coexist with preformed
pairs above
T
c
. In this context, the microscopic theory of
high-
T
c
superconductivity may not depend on either the
one-gaportwo-gapscenariobeingthesufficientcondition.
Regardless of the theoretical views, a feasible empirical
approach to verify the validity of the “two-gap” scenario
istoconductvortex-statequasiparticletunnelingspectro-
scopicstudies.Vorticesinducedbyappliedmagneticfields
in type-II superconductors are known to form periodic
regions within which the SC order parameter is locally
suppressed [20], leading to continuous bound states [21]
and an enhanced local density of states (LDOS) at zero
energy (referenced to the Fermi level) inside vortices [22].
Consequently, if a CO coexists with SC in the ground
state, it should be revealed inside the vortex core as a
gapped feature at a CO energy
V
CO
upon local suppres-
sion of SC. This behavior would be in sharp contrast to
17004-p1
M. L. Teague
et al.
the enhanced zero-energy LDOS near the center of vortex
cores had SC been the sole ground-state order parameter.
For pure
d
x
2
y
2
-wave pairing symmetry, additional four-
fold “star-shape” orientational dependent conductance is
expected due to the existence of nodal quasiparticles [23].
Earlier studies of the vortex-state scanning tunnel-
ing spectroscopy (STS) on two hole-type cuprates
YBa
2
Cu
3
O
7
δ
(Y-123) [4,24] and Bi
2
Sr
2
CaCu
2
O
8+
x
(Bi-2212) [25,26] have shown no evidence for either
the zero-energy conductance peak or the star-shape
conductance pattern inside the vortex cores. In Bi-2212
it was found that by integrating the excess field-induced
spectral weight up to a finite energy 12meV, the result-
ing spatially resolved conductance map revealed weak
checkerboard-like modulations inside each vortex [26].
These LDOS modulations have been attributed to the
presence of a coexisting competing order (CO) such as
CDW [11] or SDW [13] in Bi-2212 upon the suppression
of superconductivity inside the vortices. Addition-
ally, recent zero-field spatially resolved STS studies
of a one-layer electron-type cuprate superconductor
Pr
0
.
88
LaCe
0
.
12
CuO
4
(PLCCO) have shown supporting
evidence for a bosonic mode related to spin excitations in
the SC state [27]. However, to date there have not been
any spatially resolved
vortex-state
STS studies reported
on the electron-type cuprates.
Given various contrasts in the physical properties of
electron- and hole-type cuprates, we expect the investi-
gation of vortex-state quasiparticle tunneling spectra of
electron-type cuprates to provide new insights into the
feasibilityoftheCOscenario.Inthisletter,weprovidethe
first detailed vortex-state investigation on an optimally
doped electron-type cuprate Sr
0
.
9
La
0
.
1
CuO
2
(La-112).
Experimental. –
The unit cell of La-112 is nearly
cubic, with in-plane and
c
-axis lattice constants being
0.395nmand0.341nm,respectively.Thesuperconducting
coherence length in the CuO
2
plane is
ξ
ab
4
.
86nm and
along the
c
-axis is
ξ
c
0
.
52nm, which is longer than
the
c
-axis lattice constant [28]. Various bulk properties
such as the anisotropic upper critical fields and irre-
versibility fields of this system have been characterized
previously [29].
The spatially resolved differential tunneling conduc-
tance(d
I/
d
V
)-
vs.
-energy(
ω
)spectraforthequasiparticle
LDOS maps were obtained with our homemade cryogenic
scanning tunneling microscope (STM). The STM has a
base temperature of 6K, variable temperature range up
to room temperature, magnetic field range up to 7tesla,
andultra-highvacuumcapabilitydowntoabasepressure
<
10
9
torr at 6K. Given its highly three-dimensional
chemical structure, the surface of La-112 cannot be
cleaved as in the case of Bi-2212, and must be chemically
etched [30]. The chemically etched surface is found to
be stoichiometric from X-ray photoemission spectroscopy
(XPS) [31], which also yields high-quality and repro-
ducible zero-field STS at cryogenic temperatures [7]. For
the STS data reported in this work, the tunnel junction
resistance has always been kept at
1GΩ to ensure
high-quality junctions.
For each constant temperature (
T
) and magnetic
field (
H
), the experiments were conducted by tunneling
currents along the
c
-axis of a single-crystal grain of
La-112 under a range of bias voltages at a given location.
The finite-field tunneling spectra were always taken
under the zero-field–cool condition. Current (
I
)-
vs.
-bias
voltage (
V
) measurements were repeated pixel-by-pixel
over an extended area of the sample, and a typical
two-dimensional map consisting of (128
×
128) pixels.
Therefore, the spatial resolution was determined by
the area of spectral measurements divided by the total
numberofpixels.Toremoveslightvariationsinthetunnel
junction resistance from pixel to pixel, the differential
conductanceateachpixelisnormalizedtoitshigh-energy
background, as specified in ref. [7].
Results. –
A representative zero-field tunneling spec-
trum of the normalized differential conductance (d
I/
d
V
)
relative to the quasiparticle energy (
ω
=
eV
,
V
being
the bias voltage) taken at 6K is shown in fig. 1(a),
together with three different theoretical fitting curves to
be discussed later. Here we denote the energy associ-
ated with the spectral peaks as ∆
eff
. With increasing
temperature, ∆
eff
(
T
) steadily decreases, as shown in the
main panel of fig. 1(b) and summarized in the inset of
fig. 1(b). Moreover, we find relatively homogeneous zero-
field quasiparticle tunneling spectra, as manifested by the
eff
histograminfig.1(c).Thisfindingisinstarkcontrast
tothenano-scalespectralvariationsinoptimalandunder-
doped Bi-2212 [32].
Upon applying magnetic fields (
H
) with the condition
0
<H

H
ab,c
c
2
(see fig. 1(d)), the quasiparticle tunneling
spectra exhibit strong spatial inhomogeneity, as exem-
plified in fig. 2(a) for
H
=1T over a (64
×
64)nm
2
area
and fig. 2(c) for
H
=2T over a (65
×
50)nm
2
area.
Given that we are primarily interested in achieving high
spatial resolution in order to investigate the inter- to
intra-vortex spectral evolution, we first focus on smaller
spatial maps in finite fields. Even for the smaller maps,
vortices are still clearly visible at both 1T and 2T with
averaged vortex lattice constants
a
B
=52nm and 35nm,
respectively, which are comparable to the theoretical
value
a
B
=1
.
075(Φ
0
/B
)
1
/
2
. Specifically, we identify the
location of vortices by plotting the spatial map of the
conductancepowerratio,definedasthevalueof(d
I/
d
V
)
2
atthespectralpeakenergy
ω
=∆
eff
relativetothatatthe
zero energy
ω
=0. We find that the presence of vortices
is associated with the local minimum of the conductance
powerratio,whichisconsistentwithenhancedzero-energy
quasiparticle density of states inside the vortex core. We
further note that the average radius of the vortices is
comparable to the superconducting coherence length
ξ
ab
.
In addition to the high-resolution vortex maps in
fig. 2(a) and (c), we illustrate a larger area vortex map
17004-p2
Observation of vortices and hidden pseudogap in Sr
0
.
9
La
0
.
1
CuO
2
Fig. 1: (Colour on-line) Zero-field tunneling spectra and crit-
ical fields of La-112: (a) A (d
I/
d
V
)-
vs.
-energy (
ω
) tunneling
spectrum (circles) at
T
=6K normalized to its high-energy
background [7], together with three different theoretical fitting
curves that assume pure
s
-waveSC(green),pure
d
x
2
y
2
-wave
SC (red), and coexisting
d
x
2
y
2
-wave SC and commensurate
SDW (blue) with fitting parameters ∆
SC
=(12
.
0
±
0
.
2)meV,
V
CO
=
V
SDW
=(8
.
0
±
0
.
2)meV, and SDW wave vector
Q
=
(
π,π
). The inset shows the zoom-in comparison of the data
and three different theoretical fittings. As elaborated in the
Discussionanddetailedinrefs.[17,18],coexistingSCandSDW
provides the best fitting, and the spectral peak is associated
with ∆
eff
. (b) Evolution of quasiparticle spectra with temper-
ature from
T
=6Kto
T
=49K
>T
c
=43K, showing absence
of PG above
T
c
. The solid lines are theoretical fitting curves
using temperature Green’s function [17,18] that yield the
correct empirical ∆
eff
(
T
) given in the inset. The temperature-
dependent (∆
SC
,
V
CO
) values in units of meV are (12.0, 8.0)
for6K,(9.5,2.0)for17K,(7.5,0)for33K,(0,0)for49K.The
zero-field ∆
eff
-
vs.
-
T
data (red circles) in the inset of fig. 1(b)
largely follow the BCS temperature dependence (solid line).
(c) A histogram of the zero-field effective gap ∆
eff
at
T
=
6K, showing relatively homogeneous gap values ∆
eff
=12
.
2
±
0
.
8meV over an area of (64
×
64)nm
2
. (d) The experimental
(
H,T
) variables investigated in this work are shown in refer-
encetotheuppercriticalfields
H
ab
c
2
for
H
ˆ
c
and
H
c
c
2
for
H

ˆ
c
ofLa-112[29].Ourexperimentalconditionsareconsistentwith
the
T

T
c
and
H

H
c
2
limit, as shown in the shaded area.
for
H
=1
.
5T with reduced spatial resolution in fig. 2(e),
which shows disordered vortices over a (160
×
152)nm
2
area. Despite the disorder, we find that the total flux
is still conserved within the area studied. That is, the
total number of vortices multiplied by the flux quantum
is equal to the magnetic induction multiplied by the
area, within experimental errors. We obtain an averaged
vortex lattice constant
a
B
=42nm, comparable to the
theoretical value of
a
B
=40nm. For comparison, we show
in fig. 2(f) the zero-field conductance power ratio map
over the same area as in fig. 2(e). The conductance ratio
Fig. 2: (Colour on-line) Vortex maps of La-112 at
T
=6Kand
for
H

c
-axis:(a)Aspatialmapoftheconductancepowerratio
(in log scale) taken over a (64
×
64)nm
2
area with
H
=1T,
showing a zoom-in view of vortices separated by an average
vortex lattice constant
a
B
=52nm, which compares favorably
with the theoretical value of 49nm. The average radius of the
vortices(indicatedbytheradiusofthecircles)is(4
.
7
±
0
.
7)nm,
comparable to the SC coherence length
ξ
ab
=4
.
9nm.Herethe
conductance power ratio is defined as the ratio of (d
I/
d
V
)
2
at
|
ω
|
=∆
eff
and that at
ω
=0. (b) Spatial evolution of the
conductance (d
I/
d
V
) along the black dashed line cutting
through two vortices in (a) for
H
=1T, showing significant
modulations in the zero-bias conductance and slight modula-
tionsinthepeak-to-peakenergygap.Thesemodulationsgener-
ally follow a periodicity of
a
B
, and the zero-bias conductance
(the peak-to-peak gap value) reaches a maximum (minimum)
inside the vortex core. (c) A spatial map of the conductance
power ratio (in log scale) taken over a (65
×
50)nm
2
area with
H
=2T, showing a zoom-in view of vortices with an average
vortex lattice constant
a
B
=35nm, which is consistent with
the theoretical value. The averaged vortex radius (indicated
bytheradiusofthecircles)is(5
.
0
±
1
.
3)nm. (d) Spatial
evolution of the conductance is shown along the black dashed
line cutting through three vortices in (b) for
H
=2T.(e)A
spatialmapoftheconductancepowerratio(inlogscale)taken
over a larger (160
×
152)nm
2
area with
H
=1
.
5T, showing
disordered vortices. Nonetheless, the average vortex lattice
constant
a
B
=42nm remains consistent with Arbikosov’s
theory. (f) A spatial map of the conductance power ratio
(in log scale) taken at
H
=0 over the same (160
×
152)nm
2
area as in (e), showing a relatively homogeneous map of the
conductanceratioincontrasttoallothermapstakenat
H>
0.
map at
H
=0 clearly demonstrates significant contrast to
the maps at finite fields, further verifying our observation
of vortices in La-112. However, we note that the shape
17004-p3
M. L. Teague
et al.
Fig. 3: Evolution of the inter- and intra-vortex quasiparticle
tunnelingspectrawithmagneticfieldinLa-112for(a)
H
=1T
and
T
=6K, (b)
H
=2T and
T
=6K, (c)
H
=3
.
5T and
T
=6K, (d)
H
=6Tand
T
=11K,wherethePGspectraat
thecenterofvortexcoresaregivenbythethicklinesandthose
exteriortovorticesaregivenbythethinlines.Wenotethatthe
peak features associated with the inter-vortex spectra broaden
with increasing
H
, and the zero-bias conductance of the inter-
vortex spectra increases with increasing
H
.
of all vortices observed in our experiments is generally
irregular. Possible cause for the irregular vortex shape
may be due to microscopic disorder in the sample as well
as surface roughness after chemical etching. Additionally,
possible interaction between the STM tip and vortices
may have also contributed to the irregular vortex shape.
Tobetterevaluatethespatialevolutionofthetunneling
spectra, we take a line cut through multiple vortices and
illustratethecorrespondingconductance(d
I/
d
V
)spectra
in fig. 2(b) for
H
=1T and in fig. 2(d) for
H
=2T.We
find that the tunneling spectra inside vortices exhibit
PG-like features rather than a peak at the zero bias,
similar to the findings in hole-type cuprates of Y-123
and Bi-2212 [4,24,25]. The PG energy is smaller than the
zero-field ∆
eff
(
H
=0) and comparable to the zero-field
competing order energy
V
CO
(
H
=0) derived from the
Green’sfunctionanalysis(seeDiscussionandrefs.[17,18]).
Both the zero-bias conductance and the peak-to-peak
energy gap exhibit modulations with a periodicity of
a
B
; the zero-bias conductance shows a maximum inside
vortices, whereas the peak-to-peak energy gap reaches a
minimum inside vortices. Upon increasing magnetic field,
the energy associated with the peak features at ∆
eff
(
H
)
outside
vortices decreases slightly and the linewidth of
thepeaksbroadens,whereasthePG-energyat
V
CO
inside
vortices remains constant, as shown in figs. 3(a)–(d)
for comparison of representative inter- and intra-vortex
spectra taken at
H
=1
,
2
,
3
.
5 and 6T.
To fully account for the statistics of the spectral evolu-
tion with increasing fields, we illustrate the histograms
of the characteristic energies identified from all spectra
taken over spatial maps varying from (50
×
50)nm
2
to
(100
×
100)nm
2
in fig. 4(b). The characteristic energy
Fig. 4: (Colour on-line) Quasiparticle spectral evolution in
La-112 as a function of magnetic field (
H
): (a) Magnetic-field
dependence of the characteristic energies ∆
eff
,∆
SC
and
V
CO
for
H
=1, 2, 3, 3.5, 4.5 and 6T. For each field
H
the
corresponding energy histogram is obtained by identifying
one-half of the quasiparticle spectral peak-to-peak energy
separation as ∆
pk
pk
over (50
×
50)nm
2
to (100
×
100)nm
2
areas. Each histogram can be fit with a Lorentzian functional
form, and the peak position of the Lorentian is identi-
fied as ∆
eff
(
H
). The low-energy cutoff of the histogram
is identified as
V
CO
,andtheSCgapforagivenfield
is defined as ∆
SC
(
H
)=
{
[∆
2
eff
(
H
)
V
2
CO
]
1
/
2
/
cos(2
φ
AF
)
}
,
where
φ
AF
25
is an angle associated with the antiferro-
magnetic “hot spots” [8]. Empirically, we find that
V
CO
is
nearly constant, whereas ∆
SC
(
H
) decreases slightly with
increasing
H
. (b) Energy histograms of La-112 determined
from our quasiparticle tunneling spectra of La-112, showing
the spectral evolution with
H
. Note that there is no zero-bias
conductance peak in the vortex-state, and that a low-energy
cutoff at nearly a constant value
V
CO
=(8
.
5
±
0
.
6)meV exists
for all fields. (c) Schematic illustration of the theoretical
histogramsinaconventionaltype-IIsuperconductorunderthe
conditions
T

T
c
and
H

H
c
2
. With increasing magnetic
field, the spectral weight in the vortex state of a conventional
type-II superconductor shifts to lower energies and peaks at
ω
=0. The downshifted spectral weight for a given field
H
relative to the total spectral weight in
H
=0 is approximated
by the ratio of the vortex core area relative to the Abrikosov
vortex unit cell, (
πξ
2
ab
/
2)
/
(
3
a
2
B
/
4).
of each spectrum is defined as one-half of the peak-to-
peak energy ∆
pk
-
pk
. We find that each histogram can
be fit by a Lorentzian functional form, and we associate
17004-p4
Observation of vortices and hidden pseudogap in Sr
0
.
9
La
0
.
1
CuO
2
the peak energy with ∆
eff
(
H
), which decreases slightly
with increasing magnetic field, as summarized in fig. 4(a).
Additionally, there is an apparent low-energy “cutoff” at
V
CO
=(8
.
5
±
0
.
6)meV for all histograms. Following the
analysis described in the Discussion, we may attribute
the decrease in ∆
eff
(
H
)tothatin∆
SC
(
H
), as shown
in fig. 4(a). This behavior is in stark contrast to the
histogram expected for a conventional type-II supercon-
ductor in a magnetic field
H
if SC were the sole order
parameter. In the latter situation and for
H

H
c
2
and
T

T
c
,afractionofthetotalspectralweightontheorder
of (
πξ
2
ab
/
2)
/
(
3
a
2
B
/
4) in the histogram would have been
downshifted to energies
ω<
SC
(0) due to the suppres-
sionofSCinsidevortices.Thedownshiftedspectralweight
should have been concentrated at
ω
=0, as schematically
illustratedinfig.4(c).Thezero-biasspectralweightwould
have been linearly proportional to
H
and become readily
observable in our experiments,
e.g.
25% for
H
=6T.
Discussion. –
The complete absence of a zero-bias
conductancepeakinthevortex-statequasiparticlespectra
of La-112, the occurrence of PG-like behavior revealed
inside the vortex core, and the existence of a low-energy
cutoff at
V
CO
, cannot be easily explained by the scenario
of pure SC in the ground state. On the other hand, these
anomalous experimental findings may be compared with
the scenario of coexisting CO and SC by means of the
Green’s function analysis detailed in refs. [17,18] together
with realistic bandstructures [8]. In fig. 1(a), we compare
threedifferenttheoreticalfittingcurveswiththezero-field
tunneling spectrum of La-112 at 6K. The fittings assume
pure
s
-wave SC (green), pure
d
x
2
y
2
-wave SC (red), and
coexisting
d
x
2
y
2
-waveSCandcommensurateSDW(blue)
with the following fitting parameters: SC gap ∆
SC
=
(12
.
0
±
0
.
2)meV, CO energy
V
CO
=(8
.
0
±
0
.
2)meV, and
CO wave vector
Q
=(
π,π
) for the SDW. Here we note
that the consideration of commensurate SDW [33] as
the relevant CO is consistent with neutron scattering
data from electron-type cuprate superconductors [34]. As
shown in refs. [17,18], the coexisting SC and CO scenario
yields only one set of spectral peaks at
ω
=
±
eff
as long
asthecondition
V
CO

SC
issatisfied,and∆
eff
[∆
2
SC
+
V
2
CO
]
1
/
2
[17,18]. We further note that attempts to fit the
spectra with the Dynes model [35] that assumes pure
SC together with extra quasiparticle lifetime-broadening
would have led to substantial increase in the zero-energy
LDOS, inconsistent with the empirical observation of
vanishing zero-field LDOS at
ω
0.
The analysis shown in fig. 1(a) suggests that the CO
scenario best accounts for the zero-field spectral details
at
T

T
c
. Interestingly, the CO energy derived from the
zero-field analysis (
V
CO
=(8
.
0
±
0
.
2)meV) is consistent
with the PG energy observed inside the vortex core,
implying that the intra-vortex PG has the same physical
origin as the zero-field CO. Additionally, the CO scenario
can account for the temperature dependence of the zero-
field tunneling spectra, as shown in the main panel of
fig.1(b), wherethetheoretical fitting curves(lines) tothe
experimental data (symbols) are obtained by using the
temperatureGreen’sfunctionandtemperature-dependent
SC
and
V
CO
values that yield the correct empirical
eff
(
T
) given in the inset of fig. 1(b) [17,18]. Given that
neither the intra-vortex PG nor the spectral evolution
of ∆
eff
with magnetic field may be explained by pure
d
x
2
y
2
-wave SC, we find that the CO scenario with the
relevant CO being a commensurate SDW better accounts
for our experimental findings in the optimally doped
La-112 system.
As an interesting comparison, our recent spatially
resolved vortex-state quasiparticle tunneling studies on
a hole-type optimally doped cuprate YBa
2
Cu
3
O
7
δ
(Y-123) also revealed PG-like features and field-induced
LDOS modulations inside vortices [36], except that the
PG energy in Y-123 is
larger
than ∆
SC
. This finding
may also be interpreted as a CO being revealed upon the
suppression of SC. Furthermore, the larger energy associ-
ated with the PG-like features inside vortices of Y-123 is
consistent with the presence of a zero-field PG tempera-
ture
higher
than
T
c
. These findings from the vortex-state
quasiparticle spectra of Y-123 are in contrast to those of
La-112; in the latter no zero-field PG exists above
T
c
and
the field-induced PG energy is
smaller
than ∆
SC
.
However, it is worth discussing two alternative scenar-
ios for the absence of zero-bias conductance peak inside
vortices. First, in the case of
s
-wave SC with parabolic
bandstructures, it has been shown that the bound states
inside the vortex core can lead to a zero-bias conductance
peak everywhere inside the vortex core under finite ther-
mal smearing [37]. On the other hand, in the limit of
small thermal smearing, it is found numerically [37] that
for bound states of high angular momenta, the LDOS at
0
<r<ξ
SC
inside the vortex core (
r
: the distance from
the center of a vortex core) may acquire a dip-like feature
in the conductance [37] rather than a peak at zero bias,
whichseemtoexplainourexperimentalfindings.Nonethe-
less, for pure
d
x
2
y
2
-wave superconductors such as the
cuprates, no true bound states exist inside the vortex
core [23]. Moreover, in contrast to the theoretical predic-
tionsofazero-biasconductancepeakat
r
0andPG-like
features at continuously varying energies 0
<ω<
SC
for
0
<r<ξ
SC
and
T

T
c
[37], our spectral studies revealed
a nearly constant PG energy (8
.
5
±
0
.
6)meV everywhere
inside the vortex core, with approximate (0
.
5
×
0
.
5)nm
2
spatial resolution for a vortex core radius
5nm. There-
fore, the field-induced PG-like features in La-112 cannot
be easily explained by the scenario of high–angular-
momentum bound states in an
s
-wave superconductor.
Additionally,ourobservationofafield-inducedPGenergy
larger than ∆
SC
in Y-123 [36] cannot be reconciled with
thenotionofboundstatesinsidethevortexcoreofapure
superconductor, because the bound-state energy inside a
vortex core of a pure SC system cannot exceed ∆
SC
.
Second, recent theoretical studies [38,39] have demon-
strated that the inclusion of quantum fluctuations for a
17004-p5
M. L. Teague
et al.
singlevortexinapure
d
x
2
y
2
superconductorcangiverise
tosuppressionofthezero-energyLDOSpeakatthevortex
center,whichcorroboratesourpreviousempiricalfindings
of strong field-induced quantum fluctuations in cuprate
superconductors [40]. Nonetheless, quantitative discrep-
ancies remain between experimental results and theory
that considers quantum fluctuations alone for a single
vortex without including CO’s and vortex-vortex corre-
lations [38,39]. Further investigation appears necessary to
fully account for the experimental observation.
Finally, we attribute the difficulties previously encoun-
tered with directly identifying vortices in electron-type
cuprate superconductors to the presence of a PG feature
inside the vortex core and the decreasing contrast
between the intra- and inter-vortex spectra with increas-
ing magnetic field, as exemplified in figs. 3(a)–(d). In
particular, the SC gap values of other electron-type
cuprates are much smaller than our La-112 system,
rendering the task of distinguishing inter-vortex spectra
from intra-vortex spectra even more difficult. We further
note that our finding of a hiddent PG inside vortex
cores in the infinite-layer La-112 system is consistent
with a previous bulk break-junction study of one-layer
electron-type cuprate superconductors Pr
2
x
Ce
x
CuO
4
y
and La
2
x
Ce
x
CuO
4
y
[41], where a spatially averaged
field-induced PG with an energy smaller than the SC gap
has been observed for a range of doping levels.
Conclusion. –
We have demonstrated in this letter
the first-time observation of vortices in an electron-doped
cuprate superconductor. We have also revealed from
spatially resolved quasiparticle tunneling spectroscopy
a hidden pseudogap inside vortices and unconventional
spectral evolution with temperature and magnetic field.
None of these results can be easily explained by the
scenario of pure superconductivity in the ground state of
the cuprates, thereby imposing important constraints on
the theory of high-
T
c
superconductivity.
∗∗∗
The work at Caltech was jointly supported by the
Moore Foundation and the Kavli Foundation through
the Kavli Nanoscience Institute at Caltech, and the NSF
Grant DMR-0405088. The work at Sogang University
was supported by the Center of Superconductivity from
the program of Acceleration Research of MOST/KOSEF
of Korea and Special fund of Sogang University. ADB
acknowledges the support of Intel Graduate Fellowship.
REFERENCES
[1]
Yeh N.-C.
et al.
,
Int. J. Mod. Phys. B
,
19
(2005) 285.
[2]
Yeh N.-C.
et al.
,
Chin.J.Phys.
,
45
(2007) 263.
[3]
Lee P. A.
et al.
,
Rev. Mod. Phys.
,
78
(2006) 17.
[4]
Fischer Ø.
et al.
,
Rev. Mod. Phys.
,
79
(2007) 353.
[5]
Damascelli A.
et al.
,
Rev. Mod. Phys.
,
75
(2003) 473.
[6]
Kleefisch S.
et al.
,
Phys. Rev. B
,
63
(2001) 100507.
[7]
Chen C.-T.
et al.
,
Phys. Rev. Lett.
,
88
(2002) 227002.
[8]
Matsui H.
et al.
,
Phys. Rev. Lett.
,
95
(2005) 017003.
[9]
Gros C.
et al.
,
Proc. Natl. Acad. Sci. U.S.A.
,
103
(2006)
14298.
[10]
Vojta M.
et al.
,
Phys. Rev. B
,
62
(2000) 6721.
[11]
Kivelson S. A.
et al.
,
Rev. Mod. Phys.
,
75
(2003)
1201.
[12]
Demler E.
et al.
,
Rev. Mod. Phys.
,
76
(2004) 909.
[13]
Polkovnikov A.
et al.
,
Phys. Rev. B
,
65
(2002)
220509(R).
[14]
Krasnov V. M.
et al.
,
Phys. Rev. Lett.
,
84
(2000) 5860.
[15]
LeTacon M.
et al.
,
Nat. Phys.
,
2
(2006) 537.
[16]
Boyer M. C.
et al.
,
Nat. Phys.
,
3
(2007) 802.
[17]
Chen C.-T.
et al.
,
Solid State Commun.
,
143
(2007)
447.
[18]
Beyer A. D.
et al.
,
Physica C
,
468
(2008) 471.
[19]
Yu B.-L.
et al.
, to be published in
Solid State Commun.
(2008) doi: 10.1016/j.ssc.2008.12.018.
[20]
Abrikosov A. A.
,
Sov. Phys. JETP
,
5
(1957) 1174.
[21]
Caroli C.
et al.
,
J. Phys. Lett.
,
9
(1964) 307.
[22]
Hess H.
et al.
,
Phys. Rev. Lett.
,
64
(1990) 2711.
[23]
Franz M.
and
Tesanovic Z.
,
Phys. Rev. Lett.
,
80
(1998)
4763.
[24]
Maggio-Aprile I.
et al.
,
Phys. Rev. Lett.
,
75
(1995)
2754.
[25]
Pan S. H.
et al.
,
Phys. Rev. Lett.
,
85
(2000) 1536.
[26]
Hoffman J. E.
et al.
,
Science
,
295
(2002) 466.
[27]
Niestemski F. C.
et al.
,
Nature
,
450
(2007) 1058.
[28]
Kim M. S.
et al.
,
Phys. Rev. B
,
66
(2002) 214509.
[29]
Zapf V. S.
et al.
,
Phys. Rev. B
,
71
(2005) 134526.
[30]
Vasquez R. P.
,
J. Electron Spectrosc.
,
66
(1994) 209.
[31]
Vasquez R. P.
et al.
,
J. Phys.: Condens. Matter
,
13
(2001) 7977.
[32]
McElroy K.
et al.
,
Phys. Rev. Lett.
,
94
(2005) 197005.
[33]
Schrieffer J. R.
et al.
,
Phys. Rev. B
,
39
(1989) 11663.
[34]
Motoyama E. M.
et al.
,
Phys. Rev. Lett.
,
96
(2006)
137002.
[35]
Dynes R. C.
et al.
,
Phys. Rev. Lett.
,
41
(1978) 1509.
[36]
Beyer A. D.
et al.
, arXiv:0808.3016 (2008).
[37]
Shore J. D.
et al.
,
Phys. Rev. Lett.
,
62
(1989) 3089.
[38]
Balents L.
et al.
,
Phys. Rev. B
,
71
(2005) 144508.
[39]
Nikolic P.
et al.
,
Phys. Rev. B
,
74
(2006) 144516.
[40]
Beyer A. D.
et al.
,
Phys. Rev. B
,
76
(2007) 140506(R).
[41]
Alff L.
et al.
,
Nature
,
422
(2003) 698.
17004-p6