A preview of this full-text is provided by Springer Nature.
Content available from Nature Physics
This content is subject to copyright. Terms and conditions apply.
Letters
https://doi.org/10.1038/s41567-019-0509-5
1Laboratoire Matériaux et Phénomènes Quantiques (UMR 7162 CNRS), Université Paris Diderot-Paris 7, Bat.Condorcet, Paris, France. 2Service de
Physique de l’État Condensé, DSM/IRAMIS/SPEC (UMR 3680 CNRS), CEA Saclay, Gif sur Yvette Cedex, France. 3Laboratoire National des Champs
Magnétiques Intenses, CNRS-Université Grenoble Alpes-Université Paul Sabatier-Institut National des Sciences Appliquées, European Magnetic
Field Laboratory, Grenoble, France. 4Laboratoire de Physique des Solides, CNRS, Université Paris-Sud, Université Paris-Saclay, Orsay Cedex, France.
*e-mail: alain.sacuto@univ-paris-diderot.fr
The cuprate high-temperature superconductors develop
spontaneous charge density wave (CDW) order below a tem-
perature TCDW and over a wide range of hole doping (p). An
outstanding challenge in the field is to understand whether
this modulated phase is related to the more exhaustively stud-
ied pseudogap and superconducting phases1,2. To address this
issue, it is important to extract the energy scale ΔCDW asso-
ciated with the CDW order, and to compare it with the pseu-
dogap ΔPG and with the superconducting gap ΔSC. However,
while TCDW is well characterized from earlier work3, little is
currently known about ΔCDW. Here, we report the extraction
of ΔCDW for several cuprates using electronic Raman spec-
troscopy. We find that on approaching the parent Mott state
by lowering p, ΔCDW increases in a manner similar to the dop-
ing dependence of ΔPG and ΔSC. This reveals that these three
phases have a common microscopic origin. In addition, we find
that ΔCDW ≈ ΔSC over a substantial doping range, which sug-
gests that CDW and superconducting phases are intimately
related; for example, they may be intertwined or connected by
an emergent symmetry1,4–9.
In recent years, many experiments and different techniques have
established the ubiquity of CDW order in cuprates3. In particular,
these works have determined TCDW(p), which displays a dome-like
shape in the temperature–doping (T–p) phase diagram, in a fashion
reminiscent of the superconducting dome TSC(p), even though the
CDW order is present over a narrower p range, and mostly below
optimal doping. The CDW was found to compete with supercon-
ductivity10–16 but there are indications that the interplay between the
two phenomena might be more complex than a simple competition.
The energy scale ΔCDW associated with the CDW has attracted
much less experimental attention, even though knowledge of this
quantity is crucial to address several important questions. First,
can the CDW be understood within a scenario of weakly interact-
ing electrons for which mean-field theory is adequate? A telltale
signature of it would be if TCDW(p) ∝ ΔCDW(p). On the other hand,
a more intricate physics is implied if their doping trends are dif-
ferent, as is famously the case for high-Tc superconductivity, where
TSC(p) displays a dome whereas ΔSC(p) increases with lower doping.
Here we show that this is also the case for CDW, and that ΔCDW(p)
and TCDW(p) have very different doping trends. Second, what is the
relationship between the CDW, superconducting and pseudogap
phases that can be understood through a comparison of the magni-
tudes and the doping dependencies of their energy scales ΔCDW(p),
ΔSC(p) and ΔPG(p)? We show that these three energy scales have sim-
ilar doping evolutions, implying that it is likely that they originate
from the same electronic interaction. Moreover, we find that the
magnitude of ΔCDW(p) and that of ΔSC(p) are nearly the same over
a significant doping range. This is a signature either of competing
superconducting and CDW orders connected by an approximate
emergent symmetry, or of these orders cooperating via a micro-
scopic intertwining1,4–9.
A typical signature of a density wave in Raman spectroscopy
is the loss of spectral weight of the electronic continuum at low
energy, followed by a recovery of spectral weight at higher energy,
as the order sets in as a function of temperature17. However, in the
cuprates, TCDW is below T*, the characteristic pseudogap tempera-
ture, so one technical challenge is to distinguish CDW from the loss
of spectral weight due to the pseudogap itself.
This difficulty can be overcome by studying the B2g Raman
response, which preferentially probes the nodal regions of the
Brillouin zone (Fig. 1a), and where pseudogap effects are known to
be minimal. This intuition is further aided by the fact that the Bloch
states that are expected to show the largest reconstruction due to
the CDW are in between the nodal and the anti-nodal regions, as
indicated by the CDW ordering vectors Qx and Qy on Fig. 1a3,18.
Therefore, the B2g Raman geometry should be able to capture the
signature of the CDW.
Consequently, we have performed electronic Raman scatter-
ing measurements on cuprates in the B2g geometry. We have first
investigated HgBa2Ca2Cu3O8+δ (Hg-1223) single crystals grown
with a single-step synthesis19. From a materials point of view,
its inner CuO2 plane is likely to be the cleanest CuO2 plane of all
cuprates because it is homogeneously doped and screened from
out-of-plane disorder by the outer planes, as demonstrated by the
analysis of the 63Cu-NMR linewidth20,21. Having obtained promis-
ing results in Hg-1223 (see below), we also measured single crystals
from HgBa2CuO4+δ (Hg-1201) and YBa2Cu3O7−δ (Y-123) to dem-
onstrate that the CDW signature in the Raman response is present
in different cuprate families. The experimental details are given in
Supplementary Section A.
Our first central observation is that the B2g Raman response
χ ω′′ T(, )
Bg2 of an underdoped Hg-1223 crystal displays a well-
Intimate link between charge density wave,
pseudogap and superconducting energy scales
in cuprates
B. Loret1, N. Auvray1, Y. Gallais1, M. Cazayous1, A. Forget2, D. Colson2, M.-H. Julien3, I. Paul1, M. Civelli4
and A. Sacuto 1*
NATURE PHYSICS | VOL 15 | AUGUST 2019 | 771–775 | www.nature.com/naturephysics 771
Content courtesy of Springer Nature, terms of use apply. Rights reserved