LETTERS
In situ doping control of the surface of
high-temperature superconductors
M. A. HOSSAIN1*, J. D. F. MOTTERSHEAD1*, D. FOURNIER1, A. BOSTWICK2, J. L. MCCHESNEY2,
E. ROTENBERG2, R. LIANG3, W. N. HARDY1,3, G. A. SAWATZKY1,3, I. S. ELFIMOV3, D. A. BONN1,3
AND A. DAMASCELLI1,3†
1 Department of Physics and Astronomy, University of British Columbia, Vancouver, British Columbia V6T 1Z1, Canada
2 Advanced Light Source, Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA
3 AMPEL, University of British Columbia, Vancouver, British Columbia V6T 1Z4, Canada
*These authors contributed equally to this work
† e-mail: damascelli@physics.ubc.ca
Published online: 22 June 2008; doi:10.1038/nphys998
Central
to
the
understanding
of
high-temperature
2
a
b
π
Vacuum
superconductivity is the evolution of the electronic structure
FS
as doping alters the density of charge carriers in the CuO
B
2 planes.
FSAB
Superconductivity emerges along the path from a normal metal
FSCh
on the overdoped side to an antiferromagnetic insulator on the
underdoped side. This path also exhibits a severe disruption of
0
the overdoped normal metal’s Fermi surface1–3. Angle-resolved
photoemission spectroscopy (ARPES) on the surfaces of easily
cleaved materials such as Bi
b
2Sr2CaCu2O8+δ shows that in zero
magnetic field the Fermi surface breaks up into disconnected
a
arcs4–6. However, in high magnetic field, quantum oscillations7 at
Y
Ba
Cu
O
–2π
0
2π
low temperatures in YBa2Cu3O6.5 indicate the existence of small
Fermi surface pockets8–18. Reconciling these two phenomena
through ARPES studies of YBa
Figure 1 The surface of cleaved YBCO. a, Crystal structure of oxygen-ordered
2Cu3O7−δ
(YBCO) has been
hampered by the surface sensitivity of the technique19–21. Here,
YBCO6.5 (with alternating oxygen-full and -empty chains), showing the BaO-plane
we show that this difficulty stems from the polarity and resulting
and CuO-chain terminations of the cleaved surface. Electronic reconstruction takes
self-doping of the YBCO surface. Through in situ deposition
place at these polar surfaces, similar to the prototypical case of a polar catastrophe
of potassium atoms on cleaved YBCO, we can continuously
in ionic insulators with a |1 + |1 − |1 + |1 − |... layer-by-layer charge25. b, This
control the surface doping and follow the evolution of the Fermi
leads to larger-than-expected Fermi surface features for the surface topmost layers:
surface from the overdoped to the underdoped regime. The
the one-dimensional CuO-chain band (FSCh) and especially ‘overdoped-like’
present approach opens the door to systematic studies of high-
CuO2-plane bonding and antibonding bands (FSB and FSAB).
temperature superconductors, such as creating new electron-
doped superconductors from insulating parent compounds.
In the heavily overdoped regime, angular magnetoresistance
and correlated character of the electronic structure of YBCO6.5
oscillation1
and
angle-resolved
photoemission
spectroscopy
(ref. 11), the determination of the nature of these pockets and
(ARPES) experiments2,3 on Tl2Ba2CuO6+δ have arrived at a
their generality to the underdoped cuprates requires connecting
quantitative consensus in observing a large hole-like Fermi surface.
transport and single-particle spectroscopy information on the
On reducing the number of holes in the CuO2 planes, the Fermi
same underdoped system. The study of YBCO6.5 by ARPES is
surface volume decreases in the manner expected by Luttinger’s
thus crucial. Unfortunately, this material is complicated by the
theorem, but below optimal doping the single-particle Fermi
lack of a natural [001] cleavage plane (Fig. 1a) and also by the
surface seems to reduce to four disconnected nodal Fermi arcs. This
presence of CuO-chain layers. More specifically, YBCO cleaves
scenario was suggested from ARPES studies of Bi cuprates4,5 and
between the CuO-chain layer and the BaO layer, leaving on the
Ca2−xNaxCuO2Cl2 (ref. 6), and is thought to be connected to the
cleaved surface relatively large regions (>100 ˚A) of either CuO
existence of the pseudogap. The detection of quantum oscillations
or BaO terminations. Scanning tunnelling microscopy shows that
in oxygen-ordered ortho-II YBa2Cu3O6.5 (YBCO6.5) suggests a
the CuO-chain terraces are characterized by prominent surface
different scenario involving Fermi surface reconstruction into hole
density waves22 and differ substantially from the bulk: as also seen
and/or electron pockets7,9,10. These two pictures are derived from
in ARPES19, they exhibit surface states and unavoidable doping
quite different measurement techniques, carried out on different
disorder. Recent ARPES studies of nearly optimally doped YBCO
materials, and under conditions of high magnetic field in one
indicated that surfaces terminating in either a CuO and BaO layer
case, but zero field in the other. Owing to the complex multiband
give different contributions to the total photoemission intensity21,
nature physics VOL 4 JULY 2008 www.nature.com/naturephysics
527
© 2008 Macmillan Publishers Limited. All rights reserved.

---

LETTERS
a
b
c
Max
2π
2π
2π
)
)
)
k y
k y
k y
0
0
0
Momentum (
Momentum (
Momentum (
–2π
–2π
–2π
YBCO6.5 surface as-cleaved
YBCO6.5 light K-deposition
YBCO6.5 heavy K-deposition
Min
0
2π
0
2π
0
2π
Momentum (k )
Momentum (k )
Momentum (k )
x
x
x
Figure 2 YBCO Fermi surface evolution on electron doping. a, ARPES Fermi surface of as-cleaved YBCO6.5, exhibiting an effective hole doping p = 0.28 per planar Cu
atom, as determined from the average area of bonding and antibonding CuO2-bilayer Fermi surfaces. b,c, By evaporating potassium on the same sample (<1 monolayer),
electrons are transferred to the topmost CuO2 bilayer and the corresponding Fermi surfaces become progressively more hole-underdoped; for heavy K-deposition (p = 0.11
as estimated from the area of the chain Fermi surface), the EF ARPES intensity reduces to the 1D CuO-chain Fermi surface and four disconnected nodal CuO2 Fermi arcs (c).
and that the doping of the topmost CuO2 planes is p = 0.3, almost
the self-doping of the YBCO6.5 polar surface is that of a heavily
irrespective of the nominal bulk doping. This corresponds to heavy
overdoped, non-superconducting cuprate.
overdoping, all the way into the non-superconducting regime
The next step is that of actively controlling the self-doping
(Fig. 4), and is actually not achievable in bulk, fully oxygenated
of the surface, to reduce its hole content to that of underdoped,
YBCO7.0 for which p = 0.194 (ref. 23). Similar problems have
bulk YBCO6.5. This can be achieved by in situ evaporation of
been encountered in the ARPES study of YBa2Cu4O8 (ref. 24).
potassium onto the cleaved surface (see the Methods section):
Overcoming these problems requires, first of all, recognizing
owing to the low ionization potential, K1+ ions are adsorbed on
that the cleaved surface of YBCO is actually polar. This can lead
the surface and electrons are doped into the topmost layers. As
to overdoped-like Fermi surfaces (Fig. 1b) due to reconstruction
a consequence, we would anticipate the evolution of all detected
of the electronic states25. A wide momentum-distribution map of
features towards the underdoped regime of hole-doped cuprates,
the Fermi surface from ‘as-cleaved’ YBCO6.5 is shown in Fig. 2a
which is precisely what can be observed in Fig. 2, from a to c,
(see the Methods section). The ARPES data are a superposition of
on increasing the K1+ concentration (decreasing the hole doping).
features from the BaO- and CuO-terminated regions: because of the
The doping is indeed changing according to an increase in electron
few ˚angstr¨om electron escape depth at these photon energies, the
filling (all data were acquired on the same sample after subsequent
ARPES intensity from the BaO-terminated regions is dominated
K evaporations). This is demonstrated by the continuous FSCh area
by the CuO
=
2-bilayer bands and that from the CuO-terminated
increase (counting electrons), which evolves from FSCh
13.8%
regions by the chain band. The comparison with Fig. 1b enables
for the as-cleaved surface to FSK1 = 14.7% and FSK2 = 16.6%
Ch
Ch
us to identify the Fermi surface features originating from the
for the increasingly K-deposited YBCOK1 (Fig. 2b) and YBCOK2
CuO2-plane bonding and antibonding bands (FSB and FSAB) and
(Fig. 2c). By comparing the carrier concentration per chain Cu
the one-dimensional (1D) CuO-chain band (FSCh). Note that the
measured by the FSCh area with the results of ab initio local
strong momentum-dependent intensity modulation of the ARPES
density approximation (LDA) band-structure calculations11, we can
features, which seems inconsistent with the sample symmetry, is
estimate the corresponding hole doping per planar Cu. This way
simply a manifestation of the matrix element effects associated
we find good agreement with the value p = 0.28 already estimated
with the photon/crystal/electron geometry changing across the field
for as-cleaved YBCO6.5 from the average of FSB and FSAB. Most
of view (no symmetrization was carried out). In addition, this
importantly, however, we obtain a hole doping pK1 = 0.20 and
particular ortho-II sample happened to be twinned, in the bulk
pK2 = 0.11 for YBCOK1 and YBCOK2, which means that the
and not just on the surface, as confirmed by X-ray diffraction
surface of YBCOK2 is very close to the doping level p = 0.097 of
(this has an effect on FSCh but not on the discussion of the
bulk oxygen-ordered ortho-II YBCO6.5.
four-fold symmetric Fermi arcs). Most importantly, the fit of
The most interesting aspect of the data in Fig. 2 is the
the 2D ARPES Fermi surfaces over multiple zones returns the
evolution of the CuO2-plane features. For heavy K deposition
following areas, counting electrons, relative to the Brillouin zone
(Fig. 2c), the LDA-like CuO2-bilayer bonding and antibonding
area A
=
=
BZ
4π2/ab: the bonding Fermi surface area FSB
46.2%,
Fermi surfaces of overdoped YBCO have collapsed into four nodal
the antibonding FS =
=
AB
26.0% and the chain surface FSCh
13.8%.
Fermi arcs, consistent with other underdoped cuprates4–6. This
From the average of bonding and antibonding Fermi surface areas,
is accompanied by the complete suppression of CuO2 antinodal
we can calculate the hole doping p = 0.28 ± 0.01 per planar copper
spectral intensity as well as nodal bilayer splitting, which in contrast
(p = 0 for the 1/2-filled Mott insulator with 1 hole per Cu atom).
were clearly resolved for as-cleaved YBCO6.5 (Fig. 2a). Their
As summarized in the phase diagram of Fig. 4, this indicates that
disappearance with K deposition suggests a severe loss of coherence
528
nature physics VOL 4 JULY 2008 www.nature.com/naturephysics
© 2008 Macmillan Publishers Limited. All rights reserved.

---

LETTERS
Max
a
c
YBCO6.5 surface as-cleaved
ANAB
ANB
E
NAB
F
NB
CuO2
y (eV)
B–AB
500
0
0.5
0.46 eV
Binding energ
YBCO6.5 as-cleaved
BaO
Photoemission intensity (a.u.)
1.0
d
AN
N
Min
(–π,–2π)
(–π,–π)
(–π,0)
(–π,π)
(–π,2π)
Momentum (k = –π;k )
x
y
Max
b
YBCO6.5 heavy K-deposition
YBCO6.5 light-K
EF
e
AN
y (eV)
CuO
chain
N
0.5
0.56 eV
Binding energ
BaO
1.0
YBCO6.5 heavy-K
Min
(–π,–2π)
(–π,–π)
(–π,0)
(–π,π)
(–π,2π)
500
250
0
Momentum (k = –
)
Binding energy (meV)
π;k
x
y
Figure 3 YBCO dispersion and EDC evolution on electron doping. a,b, The BaO-band high-energy shift in YBCOK2 as compared with YBCO6.5 reveals that electrons are
added on K deposition; in addition, bonding (B) and antibonding (AB) CuO2 features vanish at the antinodes and only the CuO-chain band is detected. c–e, Nodal (N) and
antinodal (AN) kF EDCs showing the progressive opening of an antinodal gap and loss of CuO2-plane quasiparticle coherence on underdoping (∆ ∼ 10 and 80 meV for
YBCOK1 and YBCOK2); the bonding (c) and antibonding (c, inset) splitting is resolved only for as-cleaved YBCO6.5.
on underdoping. Correspondingly, the CuO2 nodal Fermi wave
for YBCOK1 and YBCOK2, respectively. This is consistent with
vectors, relative to the Brillouin zone diagonal (0,0)–(π,π), have
YBCO being superconducting for p = 0.11 and 0.20 but not 0.28
evolved from kAB = 0.29 and kB = 0.36 for ‘overdoped’ as-cleaved
(Fig. 4—the experiments were all carried out at T = 20 K), and
F
F
YBCO6.5, to a single kK2 = 0.40 for ‘underdoped’ YBCOK2. These
enables us to conclude that the properties of the K-deposited YBCO
F
numbers compare well to what has been observed on other cuprates
surfaces are representative of bulk YBCO.
at similar dopings (note that the following are both single CuO2-
We now summarize our findings, illustrated by the phase
layer systems): k =
F
0.36 and 0.41, respectively, for overdoped
diagram and symmetrized Fermi surface data for YBCOK2 and
(p = 0.26) Tl2Ba2CuO6+δ (ref. 2) and underdoped (p = 0.12)
as-cleaved YBCO6.5 in Fig. 4. Our study demonstrates that the
Ca2−xNaxCuO2Cl2 (ref. 6).
self-hole-doping of the cleaved polar surfaces of YBCO can be
The transfer of electrons to the surface of YBCO on
controlled by the in situ evaporation of alkali metals, in the present
K deposition is also confirmed by an inspection of the
case K. This novel approach paves the way for the study of this
electronic dispersions along the (−π, −2π)–(−π, 2π) direction
important material family—-across the whole phase diagram—
(corresponding momentum-distribution curves are shown in the
by single-particle spectroscopies. As this material has been the
Supplementary Information). On p = 0.28 as-cleaved YBCO6.5
gold standard in a number of seminal bulk-sensitive studies of
(Fig. 3a), we detect well-defined bonding and antibonding CuO2
the normal and superconducting properties, the direct connection
bands crossing EF at the antinodes, as well as the BaO band
with single-particle spectroscopy can lead to an understandable
with a maximum binding energy at the zone corners (−π, ±π).
underdoped anchor point, analogous to Tl2Ba2CuO6+δ in the
On underdoped p = 0.11 YBCOK2 (Fig. 3b), the BaO band is
overdoped regime3. The results obtained for p = 0.11 YBCOK2
located ∼100 meV deeper in energy. This indicates a shift of
establish that the ARPES Fermi surface of underdoped YBCO
the chemical potential consistent with the transfer of electrons
consists of the superposition of 1D CuO-chain Fermi surface and
from adsorbed K atoms to the topmost BaO-plane, CuO-chain
CuO2-plane-derived nodal Fermi arcs. It is thus consistent, with the
and CuO2-plane layers. On YBCOK2, the only coherent feature
extra complication of the chains, to what has already been observed
detected at the antinodes is the 1D CuO-chain band; the antinodal
in oxychloride6 and Bi cuprates4,5. In this sense, the disruption of
CuO2-plane spectral weight has now become fully incoherent.
the large hole-like coherent Fermi surface in underdoped cuprates
The nodal and antinodal energy-distribution curves (EDCs) in
is a truly universal phenomenon.
Fig. 3c–e clearly show that the progressive loss of quasiparticle
Having obtained the first momentum-resolved Fermi surface
coherence with K deposition is very similar to what is observed
data for underdoped YBCO, it becomes crucial to understand the
on other cuprates on hole underdoping2,4–6. Most importantly,
connection between ARPES and quantum oscillation results7,10.
whereas metallic behaviour is observed on as-cleaved YBCO6.5,
First, the detection of the BaO-band maximum at ∼0.5 eV below
a leading-edge antinodal gap ∆ ∼ 10 and 80 meV is detected
EF rules out the scenario coming from LDA band-structure
nature physics VOL 4 JULY 2008 www.nature.com/naturephysics
529
© 2008 Macmillan Publishers Limited. All rights reserved.

---

LETTERS
YBa Cu O
bulk YBCO6.5 (refs 7,10). However, these are hole, not electron
2
3 7–δ
6.0
6.5
7.0
pockets. Thus, the interpretation of high field measurements in
250
terms of electron and hole pockets differs markedly from single-
particle spectroscopy on the same material, suggesting that the high
magnetic field might be inducing a state different from that being
200
studied in zero field.
Whatever the solution to the puzzle outlined above, it should be
emphasized that the present approach, based on the in situ alkaline
metal evaporation on freshly cleaved surfaces, opens the door to
150
this type of manipulation of other cuprates and complex oxides,
ture (K)
not only to control the self-doping of polar surfaces but also to
p = 0.11
p = 0.28
reach doping levels otherwise precluded in the bulk. For instance,
tor
T
empera
100
we could try to underdope the surface of Tl2Ba2CuO6+δ, which
Max
grows naturally overdoped3, or even to obtain an electron-doped
superconductor starting from the insulating parent compounds.
gnetic insula
50
High-Tc
METHODS
superconductor
Antiferroma
SAMPLE PREPARATION
Min
YBa
0
2Cu3O6+x single crystals were grown in non-reactive BaZrO3 crucibles
0
0.1
0.2
0.3
using a self-flux technique. The CuOx-chain oxygen content was set to x = 0.51
Hole doping (p)
by annealing in a flowing O2:N2 mixture and homogenized by further annealing
in a sealed quartz ampoule, together with ceramic at the same oxygen content.
After mounting for the cleave required in an ARPES measurement, the samples
were cooled from 100 ◦C to room temperature over several days to establish
Figure 4 Phase diagram of YBCO by ARPES. Schematic temperature-doping phase
the ortho-II superstructure ordering of the CuO
diagram of YBCO adapted from ref. 7. The hole doping p per planar copper (p
x -chain layer28. The particular
= 0 for
sample used here was twinned, as confirmed by X-ray diffraction after the
the 1/2-filled Mott insulator with 1 hole per Cu atom), and the corresponding oxygen
ARPES measurements.
content (7 − δ), are indicated on the bottom and top axes23. The ARPES Fermi
surface for under- and overdoped YBCO is also shown; the momentum-distribution
ARPES EXPERIMENTS
maps have been two-fold and four-fold symmetrized for p = 0.11 and 0.28,
ARPES measurements were carried out on the Electronic Structure Factory
respectively. Similar to the data in Fig. 2, the doping levels were determined for
endstation at Beamline 7.01 of the Advanced Light Source. The data were
p = 0.11 from the area of FSCh, and for p = 0.28 from the area of FSB and FSAB.
measured with linearly polarized 110 eV photons and a Scienta R4000 electron
analyser in angle-resolved mode. YBCO6.5 single crystals were cleaved in situ
at a base pressure better than 2.5 × 10−11 torr and then oriented by taking
calculations, which suggested that the small Fermi surface found
fast Fermi surface scans. Several procedures were tried on these samples in
in the quantum oscillation measurements might originate from
an attempt to suppress the surface contribution to the total photoemission
intensity and gain direct insight into the bulk electronic structure. For instance,
small pockets produced by BaO–Cuchain bands at (±π, ±π)11,12.
samples were temperature-cycled between 20 and 100 K or were cleaved at
We also did not observe any signature of CuO2-derived band
higher temperature (∼80 K) and then cooled to 20 K (ref. 29), to age and/or
folding arising from the ortho-II oxygen-ordering of the chains11,12,
vary the characteristics of the cleaved surfaces. Whereas both procedures have
which is possibly consistent with the loss of three-dimensional
proved successful in measuring the bulk dispersion and Fermi surface of layered
coherence demonstrated by the suppression of bilayer band
Sr2RuO4 (ref. 29), no effect was observed in the case of YBCO6.5. Successful
splitting on underdoping. Recent measurements of the Hall
control of the self-doping of the cleaved surfaces was achieved by in situ
resistance in high magnetic field have noted a sign change with
deposition of submonolayers of potassium, with a commercial SAES getter
decreasing temperature9, suggesting that the quantum oscillations
source30, on freshly cleaved YBCO6.5. In this last case, the samples were kept at
seen on top of a negative Hall coe
20 K at all times during the cleaving, K-deposition and ARPES measurements
fficient might come from small
electron pockets, rather than the hole pockets originally proposed7.
(temperature-dependent measurements could not be carried out because of
However, there is no sign of such electron pockets in our ARPES
the need to maintain the most stable experimental conditions over extended
time, during and between subsequent K evaporations). All ARPES data shown
data from underdoped YBCOK2, nor are there signs of extra zone-
in Figs 2–4 were obtained on the same sample, that is, as-cleaved, and after
folding due to the kinds of density wave instabilities that might give
two subsequent K evaporations. Energy and angular resolutions were set to
rise to such a Fermi surface reconstruction8,9,15.
∼ 30 meV and 0.2◦ (±15◦ angular window) for the data in Figs 2 and 3, and to
If any pocket had to be postulated on the basis of the present
∼ 30 meV and 0.1◦ (±7◦ angular window) for the higher-quality Fermi surface
ARPES data, the most obvious possibility would be that the Fermi
mappings in Fig. 4. The Fermi surface maps of Figs 2 and 4 were obtained
arcs are in fact hole-like nodal Fermi pockets, as obtained for
by integrating the ARPES intensity over a 30 meV energy window about
light doping of the antiferromagnetic parent compound in self-
EF, and then by normalizing the intensity maps relative to one another for
consistent Born calculations26 and already speculated from the
display purposes. A more quantitative comparison of the ARPES intensity for
study of other underdoped cuprates4–6. The lack of a finite ARPES
as-cleaved and K-deposited surfaces is given in the Supplementary Information
intensity on the outer side of the pockets would be consistent
on the basis of the analysis of momentum-distribution curves, which shows
an overall intensity suppression by a factor of 4.4 and 7.2 for low and high K
with the strong drop in the quasiparticle coherence Zk expected
coverage, with respect to the as-cleaved surface.
beyond the antiferromagnetic zone boundary13,27. To estimate an
area for these ostensible nodal pockets, we can fold the detected arc
Received 5 January 2008; accepted 16 May 2008; published 22 June 2008.
profile, either with respect to the antiferromagnetic zone boundary
References
or the end points of the arc itself, obtaining values of either 2.6%
1. Hussey, N. E., Abdel-Jawad, M., Carrington, A., Mackenzie, A. P. & Balicas, L. A coherent
or 1.3%, relative to the full Brillouin zone area A
=
BZ
4π2/ab.
three-dimensional Fermi surface in a high-transition-temperature superconductor. Nature 425,
These numbers compare relatively well to the pocket area of 1.9%
814–817 (2003).
2. Plat´e, M. et al. Fermi surface and quasiparticle excitations of overdoped Tl2Ba2CuO6+δ by ARPES.
suggested on the basis of quantum oscillation experiments on
Phys. Rev. Lett. 95, 077001 (2005).
530
nature physics VOL 4 JULY 2008 www.nature.com/naturephysics
© 2008 Macmillan Publishers Limited. All rights reserved.

---

LETTERS
3. Peets, D. C. et al. Tl2Ba2CuO6+δ brings spectroscopic probes deep into the overdoped regime of the
20. Lu, D. H. et al. Superconducting gap and strong in-plane anisotropy in untwinned YBa2Cu3O7−d.
high-Tc cuprates. New J. Phys. 9, 1–32 (2007).
Phys. Rev. Lett. 86, 4370–4373 (2001).
4. Norman, M. R. et al. Destruction of the Fermi surface in underdoped high-Tc superconductors.
21. Zabolotnyy, V. B. et al. Momentum and temperature dependence of renormalization effects in the
Nature 392, 157–160 (1998).
high-temperature superconductor YBa2Cu3O7−d. Phys. Rev. B 76, 064519 (2007).
5. Kanigel, A. et al. Evolution of the pseudogap from Fermi arcs to the nodal liquid. Nature Phys. 2,
22. Derro, D. J. et al. Nanoscale one-dimensional scattering resonances in the CuO chains of
447–451 (2006).
YBa2Cu3O6+x. Phys. Rev. Lett. 88, 097002 (2002).
6. Shen, K. M. et al. Nodal quasiparticles and antinodal charge ordering in Ca2−xNaxCuO2Cl2. Science
23. Liang, R., Bonn, D. A. & Hardy, W. N. Evaluation of CuO2 plane hole doping in YBa2Cu3O6+x single
307, 901–904 (2005).
crystals. Phys. Rev. B 73, 180505 (2006).
7. Doiron-Leyraud, N. et al. Quantum oscillations and the Fermi surface in an underdoped high-Tc
24. Kondo, T. et al. Dual character of the electronic structure of YBa2Cu4O8: The conduction bands of
superconductor. Nature 447, 565–568 (2007).
CuO2 planes and CuO chains. Phys. Rev. Lett. 98, 157002 (2007).
8. Balakirev, F. F. et al. Fermi surface reconstruction at optimum doping in high-Tc superconductors.
25. Hesper, R., Tjeng, L. H., Heeres, A. & Sawatzky, G. A. Photoemission evidence of electronic
Preprint at <http://arxiv.org/abs/0710.4612> (2007).
stabilization of polar surfaces in K3C60. Phys. Rev. B 62, 16046–16055 (2000).
9. LeBoeuf, D. et al. Electron pockets in the Fermi surface of hole-doped high-Tc superconductors.
26. Sushkov, O. P., Sawatzky, G. A., Eder, R. & Eskes, H. Hole photoproduction in insulating copper
Nature 450, 533–536 (2007).
oxide. Phys. Rev. B 56, 11769–11776 (1997).
10. Jaudet, C. et al. de Haas-van Alphen oscillations in the underdoped cuprate YBa2Cu3O6.5. Phys. Rev.
27. Eskes, H. & Eder, R. Hubbard model versus t–J model: The one-particle spectrum. Phys. Rev. B 54,
Lett. 100, 187005 (2008).
14226–14229 (1996).
11. Elfimov, I. S., Sawatzky, G. A. & Damascelli, A. Fermi pockets and correlation effects in underdoped
28. Liang, R., Bonn, D. A. & Hardy, W. N. Preparation and x-ray characterization of highly ordered
YBa2Cu3O6.5. Phys. Rev. B 77, 060504(R) (2008).
ortho-II phase YBa2Cu3O6.50 single crystals. Physica C 336, 57–62 (2000).
12. Carrington, A. & Yelland, E. A. Band-structure calculations of Fermi-surface pockets in ortho-II
29. Damascelli, A. et al. Fermi surface, surface states, and surface reconstruction in Sr2RuO4. Phys. Rev.
YBa2Cu3O6.5. Phys. Rev. B 76, 140508(R) (2007).
Lett. 85, 5194–5197 (2000).
13. Harrison, N., McDonald, R. D. & Singleton, J. Cuprate Fermi orbits and Fermi arcs: The effect of
30. Ohta, T., Bostwick, A., Seyller, T., Horn, K. & Rotenberg, E. Controlling the electronic structure of
short-range antiferromagnetic order. Phys. Rev. Lett. 99, 206406 (2007).
bilayer graphene. Science 313, 951–954 (2006).
14. Chen, W.-Q., Yang, K.-Y., Rice, T. M. & Zhang, F. C. Quantum oscillations in magnetic field induced
antiferromagnetic phase of underdoped cuprates : Application to ortho-II YBa
Supplementary Information accompanies this paper on www.nature.com/naturephysics.
2 Cu3 O6.5 . Europhys.
Lett. 82, 17004 (2008).
15. Millis, A. J. & Norman, M. Antiphase stripe order as the origin of electron pockets observed in
Acknowledgements
1/8-hole-doped cuprates. Phys. Rev. B 76, 220503(R) (2007).
We gratefully acknowledge J. van den Brink for insightful discussions. This work was supported by the
16. Chakravarty, S. & Kee, H.-Y. Fermi pockets and quantum oscillations of the Hall coefficient in high
Alfred P. Sloan Foundation (A.D.), an ALS Doctoral Fellowship (M.A.H.), the CRC Program (A.D. and
temperature superconductors. Preprint at <http://arxiv.org/abs/0710.0608> (2007).
G.A.S), NSERC, CFI, CIFAR Quantum Materials and BCSI. The Advanced Light Source is supported
17. Alexandrov, A. S. Theory of quantum magneto-oscillations in underdoped cuprate superconductors.
by the Director, Office of Science, Office of Basic Energy Sciences, of the US Department of Energy
J. Phys. Condens. Matter 20, 192202 (2008).
under Contract No. DE-AC02-05CH11231.
18. Melikyan, A. & Vafek, O. Quantum oscillations in the mixed state of d-wave superconductor. Preprint
at <http://arxiv.org/abs/0711.0776> (2007).
Author information
19. Schabel, M. C. et al. Angle-resolved photoemission on untwinned YBa2Cu3O6.95. I. Electronic
Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions.
structure and dispersion relations of surface and bulk bands. Phys. Rev. B 57, 6090–6106 (1998).
Correspondence and requests for materials should be addressed to A.D.
nature physics VOL 4 JULY 2008 www.nature.com/naturephysics
531
© 2008 Macmillan Publishers Limited. All rights reserved.

---

Document Outline

• In situ doping control of the surface of high-temperature superconductors
• Methods
  • Sample preparation
  • ARPES experiments
• Figure 1 The surface of cleaved YBCO.
• Figure 2 YBCO Fermi surface evolution on electron doping.
• Figure 3 YBCO dispersion and EDC evolution on electron doping.
• Figure 4 Phase diagram of YBCO by ARPES.
• References
• Acknowledgements
• Author information

---