The magnetic resonance in high-temperature superconductors: Evidence for an extended s-wave pairing symmetrysymmetry
Extracted source text ยท arXiv cond-mat/0302566
%\documentstyle[preprint,aps]{revtex}
\documentstyle[prl,aps]{revtex}
\topmargin-1.3cm
\textheight24cm
\input BoxedEPS.tex
\SetepsfEPSFSpecial
\HideDisplacementBoxes
\begin{document}
%\draft{}
\twocolumn[\hsize\textwidth\columnwidth\hsize
\csname @twocolumnfalse\endcsname
\title{The magnetic resonance in high-temperature
superconductors: Evidence for an extended s-wave
pairing symmetry}
\author{Guo-meng Zhao$^{*}$}
\address{ Department of
Physics and Astronomy, California State University at Los Angeles, Los
Angeles, CA 90032, USA}
\maketitle
\widetext
\begin{abstract}
We have identified several important features in the neutron
scattering data of
cuprates, which are difficult to be explained in terms of d-wave and
isotropic s-wave
order parameters. Alternatively, we show that the neutron data are
in
quantitative agreement with an order parameter that has an extended
s-wave ($A_{1g}$) symmetry and opposite sign in
the bonding and antibonding electron bands formed within the
Cu$_{2}$O$_{4}$ bilayers. The extended s-wave has eight line nodes
and change signs when a node is crossed. This $A_{1g}$ pairing
symmetry may be compatible with a charge fluctuation mediated pairing
mechanism.
\end{abstract}
\narrowtext
\vspace{0.3cm}
%\newpage
]
The microscopic pairing mechanism responsible for high-temperature
superconductivity in copper-based perovskite oxides is still a
subject of intense debate despite tremendous experimental and
theoretical efforts for over 15 years. The debate has centered
around
the role of antiferromagnetic spin fluctuations in high-temperature
superconductivity and the symmetry of superconducting condensate
(order parameter). Extensive inelastic neutron scattering
experiments
have accumulated a great deal of important data that should be
sufficient to address these central issues. Of particular interest
is
the magnetic resonance peak that has been observed in double-layer
cuprate superconductors such as YBa$_{2}$Cu$_{3}$O$_{y}$
(YBCO) \cite{Bourges95,Fong,Bourges96,Bourges98,Bourges99,Dai} and
Bi$_{2}$Sr$_{2}$CaCu$_{2}$O$_{8+y}$ (BSCCO) \cite{FongNature,He}, and
in a
single-layer compound Tl$_{2}$Ba$_{2}$CuO$_{6+y}$ (Tl-2201)
\cite{He2002}. A
number of theoretical models \cite{Demler,Abanov,Morr,Brinckmann}
have
been proposed to explain the magnetic resonance peak in terms of
d-wave
magnetic pairing mechanisms. These
theories can qualitatively explain some features of neutron data but
are
particularly difficult to account for an important feature: The
magnetic
resonance in optimally and overdoped double-layer cuprates is much
more
pronounced in the odd channel than in the even channel
\cite{Fong,Bourges96,Bourges98}. In order to overcome this
difficulty, Mazin and co-worker \cite{Mazin} proposed an order
parameter that has isotropic s-wave symmetry and opposite sign in
the
bonding and antibonding electron bands formed within the
Cu$_{2}$O$_{4}$ bilayers \cite{Mazin}. However, this model predicts
that the resonance energy is larger than twice the magnitude of the
superconducting gap along the Cu-O bonding direction, in disagreement
with experiment. Further, the nodeless s-wave gap symmetry is
inconsistent with the measurements of the penetration depth, thermal
conductivity and specific heat, which consistently suggest the
existence of line nodes in the gap function of hole-doped cuprates
\cite{Hardy,Chiao}.
Here we identify several important features in the neutron scattering
data of
cuprates. These features are inconsistent with d-wave order
parameter (OP). Alternatively, we show that the
neutron data are in
quantitative agreement with an order parameter that has an extended
s-wave ($A_{1g}$) symmetry \cite{Zhao,Brandow} and opposite sign in
the
\begin{figure}[htb]
\ForceWidth{7cm}
\centerline{\BoxedEPSF{Fig1a.eps}}
\vspace{0.2cm}
\ForceWidth{7cm}
\centerline{\BoxedEPSF{Fig1b.eps}}
\vspace{0.3cm}
\caption[~]{a) The imaginary part of dynamic spin susceptibility as a
function of excitation energy for YBa$_{2}$Cu$_{3}$O$_{6.92}$ in the
normal and superconducting states. The figure was reproduced from
Ref.~\cite{Bourges99}. b) $q_{l}$-scan at $E $= 40 meV in
YBa$_{2}$Cu$_{3}$O$_{6.92}$ displaying a modulation typical of odd
excitation. The figure was reproduced from Ref.~\cite{Bourges98}.
The background (lower line and open squares) was obtained from q-scans
across the magnetic line. The upper solid curve is a fit to
$a+bF^{2}(\vec{Q})\sin^{2}(\pi zq_{l})$, where $F(\vec{Q})$ is the Cu
magnetic form factor. The modulation is not complete as even
excitations are sizable at $q_{l}$= 3.5, 7 but with a magnitude 5
times smaller \cite{Bourges98}. }
\protect\label{Fig1}
\end{figure}
\noindent
bonding and antibonding electron bands formed within the
Cu$_{2}$O$_{4}$ bilayers \cite{Mazin}.
The magnetic resonance peak observed in optimally doped and
overdoped double-layer compounds YBa$_{2}$Cu$_{3}$O$_{y}$
\cite{Bourges95,Fong,Bourges96,Bourges98,Bourges99} and
Bi$_{2}$Sr$_{2}$CaCu$_{2}$O$_{8+y}$ \cite{FongNature,He} is a
sharp collective mode that occurs at an energy of 38-43 meV and at
the
two-dimensional wavevector $\vec{Q}_{AF}$ = ($\pi/a$, $\pi/a$), where
$a$ is the nearest-neighbor Cu-Cu distance. Fig.~1a shows the
imaginary part of the odd channel spin susceptibility as a function
of
excitation energy for a slightly underdoped double-layer compound
YBa$_{2}$Cu$_{3}$O$_{6.92}$. The figure is reproduced from
Ref.\cite{Bourges99}. There are several striking features in the
data: (a) The resonance peak at $E_{r}$= 41 meV appears below $T_{c}$
and this resonance peak intensity in the superconducting state
$I_{odd}^{S}$($E_{r}$) is larger than the normal-state intensity
$I_{odd}^{N}$($E_{r}$) by a factor of about 3.6, i.e.,
$I_{odd}^{S}(E_{r})$/$I_{odd}^{N}(E_{r})$ = 3.6; (b) A spin gap
feature is seen below $T_{c}$ and there is a small shoulder that
occurs at an energy slightly above the spin gap energy $E_{g}$ =32
meV. Fig.~1b shows a $q_{l}$-scan at $E $= 40 meV in
YBa$_{2}$Cu$_{3}$O$_{6.92}$ displaying a modulation typical of odd
excitation. The figure is reproduced from Ref.\cite{Bourges98}.
From
Fig.~1b, we find feature (c): The odd-channel magnetic resonance
intensity within the Cu$_{2}$O$_{4}$ bilayers is larger than the
even-channel one by a factor of about 5 (Ref.\cite{Bourges98}), i.e.,
$I^{S}_{odd}(E_{r})/I^{S}_{even}(E_{r})$ = 5. From the neutron
studies on different double-layer compounds with different doping
levels, one identifies feature (d): The resonance energy $E_{r}$ does
not increase with doping in the overdoped range but is proportional
to
$T_{c}$ as $E_{r}/k_{B}T_{c}$ $\simeq$ 5.2 in both underdoped and overdoped
ranges \cite{He}.
On the other hand, the neutron data for a single-layer Tl-2201 are
different from those for double-layer compounds. For the
single-layer
Tl-2201, we show in Fig.~2a the difference spectrum of the neutron
intensities at $T$ = 27 K ($<$$T_{c}$) and $T$ = 99 K ($>$$T_{c}$),
at
a wavevector of $\vec{Q}$ = (0.5, 0.5, 12.25). The figure is
reproduced from Ref.~\cite{He2002}. Although a sharp peak feature is
clearly seen at $E_{r}$ = 46 meV, it is remarkable that the peak
intensity in the superconducting state is close to the magnetic
neutron intensity in the normal state, that is,
$I^{S}(E_{r})$/$I^{N}(E_{r})$$\simeq$ 1. This is in sharp contrast
with the above result for the
double-layer compound YBa$_{2}$Cu$_{3}$O$_{6.92}$ where we found
$I_{odd}^{S}(E_{r})$/$I_{odd}^{N}(E_{r})$ = 3.6. Thus we identify
feature (e) as $I^{S}(E_{r})$/$I^{N}(E_{r})$ $\simeq$ 1 for
the single-layer compound.
We would like to mention that feature (e) identified for the
single-layer Tl-2201
is valid only if the nonmagnetic background has a negligible
temperature dependence below 100 K. It is known that nuclear
contributions predominantly constitute the main part of the
neutron signal \cite{Bourges96}. Any nuclear contributions
(nonmagnetic backgrounds) are only
expected to obey the following standard temperature dependence,
$1/[1-\exp(-E/k_{B}T)]$ (see detailed discussions in
Ref.~\cite{Bourges96}).
In the energy range of interest ($E$ $>$ 40 meV), this temperature
dependence factor is close to unity and nearly independent of
temperature for $T < $ 150 K. Indeed, the normal-state neutron
intensities of both
YBCO \cite{Bourges96} and BSCCO \cite{FongNature,He} show a
negligible
temperature dependence for $T_{c}$ $<$ $T$ $<$ 150 K. For ~~$E$ =
\begin{figure}[htb]
\ForceWidth{7cm}
\centerline{\BoxedEPSF{Fig2a.eps}}
\ForceWidth{7cm}
\vspace{0.3cm}
\centerline{\BoxedEPSF{Fig2b.eps}}
\vspace{0.3cm}
\caption[~]{a) The difference spectrum of the neutron intensities
of single-layer Tl-2201 crystals (with a total volume of 0.11
cm$^{3}$) at
$T$ =27 K ($<$$T_{c}$) and $T$ = 99 K ($>$$T_{c}$), and at wavevector
$\vec{Q}$ = (0.5, 0.5, 12.25). The figure is reproduced from
Ref.~\cite{He2002}. The solid line is guide to the eye. The
difference spectrum tends to zero at about 56 meV (about 10 meV
higher than the resonance energy). This suggests that the
nonmagnetic background at 56 meV has negligible temperature
dependence below 100 K. b) The difference spectrum of the neutron
intensities of optimally doped BSCCO crystals (with a total volume
of 0.06
cm$^{3}$) at $T$ =10 K ($<$$T_{c}$) and $T$ = 100 K ($>$$T_{c}$), and
at wavevector $\vec{Q}$ = (0.5, 0.5, -14). The figure is reproduced
from Ref.~\cite{FongNature}. The
difference spectrum goes to zero at 60 meV (about 16 meV
higher than the resonance energy). This suggests that the
nonmagnetic background at 60 meV has negligible temperature
dependence below 100 K. }
\protect\label{Fig2}
\end{figure}
\noindent
$E_{r}$ = 46 meV in Tl-2201, the factor $1/[1-\exp(-E/k_{B}T)]$
decreases by 0.4$\%$ when
the temperature is lowered from 99 K to 27 K. If we take the
nonmagnetic
background of 2500 counts per two hours \cite{He2002}, the
nonmagnetic
background decreases by 10 counts per two hours when the temperature
goes from 99 K to 27 K, and by 60 counts per two hours when the
temperature goes from 150 K to 99 K. The variation of the
nonmagnetic
background at $E$ = 46 meV below 100 K is negligibly small compared
with the magnetic resonance intensity ($\sim$150 counts per two
hours). Therefore, feature (e) identified for the single-layer
Tl-2201 is well justified.
In order to further justify feature (e), we compare the ratios of the
magnetic to the nonmagnetic background signals in those neutron
experiments on different compounds. From the neutron data, we find
that the
signal-to-background ratio is about 6$\%$ for Tl-2201
(Ref.~\cite{He2002}), about 12$\%$ for an
optimally doped
BSCCO \cite{FongNature}, about 6$\%$ for an overdoped BSCCO
\cite{He}, and about 100$\%$ for an overdoped YBCO \cite{Bourges96}.
The
signal-background-ratio in the overdoped BSCCO is similar to that
for
Tl-2201. But the overdoped BSCCO shows a difference spectrum similar
to that for the overdoped YBCO where the signal-background-ratio is
larger than that for the overdoped BSCCO by a factor of 20. This
suggests that, if feature (e) identified for Tl-2201 were an artifact
caused by a small signal-background-ratio, one would not have
observed
the intrinsic difference spectra for the overdoped BSCCO. The fact that the
difference spectra for the overdoped BSCCO are similar to the one for YBCO suggests
that one indeed finds the intrinsic magnetic difference spectra for
the overdoped BSCCO even though the signal-background-ratio is only 6$\%$. There
are no
reasons to
believe
that only the difference spectrum for Tl-2201 is an artifact.
Therefore, the
pronounced difference between the difference spectrum of the
single-layer Tl-2201 and that of the double-layer BSCCO (compare Fig.~2a
and Fig.~2b) is due to the fact that
$I_{odd}^{S}(E_{r})$/$I_{odd}^{N}(E_{r})$ $>$$>$ 1 for the
double-layer BSCCO while $I^{S}(E_{r})$/$I^{N}(E_{r})$ $\sim$ 1 for
the
single-layer Tl-2201. Moreover, we will show below that, for the
intraband scattering (even channel) in YBa$_{2}$Cu$_{3}$O$_{6.92}$,
$I_{even}^{S}(E_{r})$/$I_{even}^{N}(E_{r})$ = 0.72, in good agreement
with feature (e) for the single-layer Tl-2201. This consistency
gives
further support to the thesis that feature (e) is intrinsic.
One may also argue that the glue that glues about 300 small
crystals of Tl-2201 on Al-plates would cause a substantial decrease
of the nonmagnetic
background
below 100 K. If this argument were relevant, one would have observed
a similar effect in overdoped
Y$_{0.9}$Ca$_{0.1}$Ba$_{2}$Cu$_{3}$O$_{7-y}$ (YBCO-Ca) because
60 larger crystals of YBCO-Ca are also glued on
Al-plates \cite{Pai}. Because the total volume of YBCO-Ca crystals
is
larger than that of Tl-2201 crystals by a factor of 3.2, one can
readily show
that the amount of glue for YBCO-Ca crystals is comparable with or
even larger than that for Tl-2201. This suggests that the effect of
glue on the nonmagnetic background in Tl-2201 is similar to that in
YBCO-Ca. From the data for YBCO-Ca (Fig.~2a and Fig.~3a of
Ref.~\cite{Pai}), one can clearly see that the difference spectrum at
50 meV is very close to zero. Because the magnetic signal at an
energy that is about 10 meV higher than the resonance energy is
independent of temperature below 100 K (see Fig.~1a), the nearly zero
value of the difference spectrum at 50 meV implies that the nonmagnetic
background at 50 meV has negligible temperature dependence below 100 K
in the case of YBCO-Ca. Even in the case of Tl-2201 (see Fig.~2a ),
the difference spectrum tends to zero at about 56 meV (about 10 meV
higher than the resonance energy) although one needs more data points
in the vicinity of 56 meV to definitively address this issue. This
indicates that the nonmagnetic background at 56 meV has negligible
temperature dependence below 100 K in the case of Tl-2201.
We would like to mention that the normal-state magnetic
intensities in Fig.~4 of Ref.~\cite{He2002} are significantly
underestimated.
This is because the authors assume that the $q$-width of the
magnetic peak
in the normal state is the same as that in the superconducting state
\cite{He2002}. This assumption is unphysical. From Fig.~19d and
Fig.~19h
of Ref.~\cite{Dai}, one can clearly see that the $q$-width of the
magnetic peak in the normal state is a factor of 1.6 larger than
that in the superconducting state in the case of underdoped
YBa$_{2}$Cu$_{3}$O$_{6.8}$. Further, the $q$-width increases with
the increase of doping. With a much broader magnetic peak in the normal
state, the $q$ range (0.35 rlu - 0.65 rlu) for the normal-state
$q$-scan spectrum (see Fig.~2B of Ref.\cite{He2002}) is too narrow to get meaningful estimates
of the nonmagnetic background and the normal-state magnetic intensity.
Moreover, the normal-state $q$-scan spectrum of
YBa$_{2}$Cu$_{3}$O$_{6.8}$ is
nearly featureless for the same narrow $q$-range (0.35 rlu - 0.65 rlu),
similar to the data of Fig.~2B of Ref.~\cite{He2002}. If one would fit the
normal-state $q$-scan data only in this narrow $q$-range (0.35 rlu - 0.65
rlu) for YBa$_{2}$Cu$_{3}$O$_{6.8}$ and assume the same $q$-width as
that in the superconducting state, one would find that
the normal-state magnetic intensity is underestimated by a factor of
4. Such a significant underestimate may be also true for the
normal-state magnetic intensities of Tl-2201. The authors of
Ref.~\cite{He2002} should have extended their measurements to a wider
$q$ range to get reliable estimates of the nonmagnetic background and
the normal-state magnetic intensities.
Since magnetic signals in both normal and
superconducting states strongly depend on doping, as clearly seen in
YBCO \cite{Bourges98}, the comparison of the resonance peak
intensities and spectral weights among different compounds should be
carefully made. For this Tl-2201, there is some indication of a broad
peak centered around $\vec{Q}_{AF}$ even above T$_{c}$ (Ref.
\cite{He2002}), as observed in underdoped YBCO \cite{Bourges98,Dai}.
This suggests that the Tl-2201 is slightly underdoped so that the
magnetic intensity in the normal state should be close to that for
slightly underdoped YBa$_{2}$Cu$_{3}$O$_{6.92}$. This is because both
compounds have a similar ratio T$_{c}$/T$_{cm}$ and thus a similar
doping level (where T$_{cm}$ is the superconducting transition at
optimal doping). As seen from Fig.~1a, the normal-state magnetic
intensity at 40 meV is about 100 $\mu_{B}^{2}$/eV per Cu in slightly
underdoped YBa$_{2}$Cu$_{3}$O$_{6.92}$. The resonance peak intensity
for the Tl-2201 can be estimated to be about 120 $\mu_{B}^{2}$/eV per
Cu from the resonance spectral weight of 0.7$\mu_{B}^{2}$ per Cu
(Ref.~\cite{He2002}). Therefore, the resonance peak intensity for the
Tl-2201 is close to the normal-state magnetic intensity for the
slightly underdoped YBa$_{2}$Cu$_{3}$O$_{6.92}$. Since the normal-state magnetic intensity
for Tl-2201 should be similar to
that for the
slightly underdoped YBa$_{2}$Cu$_{3}$O$_{6.92}$, then $I^{S}(E_{r})$/$I^{N}(E_{r})$ $\simeq$ 1.2 for this
single-layer Tl-2201. This is in good agreement with feature (e)
deduced independently from the difference spectrum (Fig.~2a).
Moreover, we find that the resonance peak intensity for the Tl-2201
($\sim$120 $\mu_{B}^{2}$/eV per Cu) is a factor of about three smaller
than the resonance peak intensity for the slightly underdoped
YBa$_{2}$Cu$_{3}$O$_{6.92}$ ($\sim$330 $\mu_{B}^{2}$/eV per Cu).
Similarly, the resonance spectral weight for the Tl-2201 is also a
factor of three smaller than that for the slightly underdoped
YBa$_{2}$Cu$_{3}$O$_{6.92}$.
These important features we have identified above should place strong
constraints on theories for high-temperature
superconductivity in cuprates. Any correct theories should be able
to
explain all the magnetic resonance features in a consistent and
quantitative way.
In a more exotic approach \cite{Demler}, the
neutron data are interpreted in terms of a collective mode in the
spin-triplet particle-particle channel, which couples to the
particle-hole channel in the superconducting state with d-wave OP.
This model predicts that the
resonance peak energy $E_{r}$ is proportional to the doping level
$p$, in disagreement with feature (d): $E_{r}$ does not increase
with
increasing $p$ in the overdoped range but is proportional to $T_{c}$
as $E_{r}/k_{B}T_{c}$ $\simeq$ 5.2 (Ref.~\cite{He}). Moreover, this model predicts \cite{Norman}
that $E_{r} > 2\Delta_{M}$ (where $\Delta_{M}$ is the maximum d-wave
gap),
which contradicts experiment.
Other theories based on spin-fermion interactions also show
that $E_{r}$ increases monotonically with increasing $p$
\cite{Abanov,Morr}, in disagreement with feature (d).
Alternatively, feature (d) is consistent with a simple particle-hole
excitation across the
superconducting gap within an itinerant magnetism model. This is
because the particle-hole excitation
energy increases with the superconducting gap which in turn should be
proportional to $T_{c}$
at least in the overdoped range. This itinerant magnetism model is
also
supported by very
recent Fourier transform scanning
tunnelling spectroscopic (FT-STS) studies on a nearly optimally doped
BSCCO
\cite{Davis}, which show that the quasiparticles in the
superconducting state exhibit particle-hole mixing similar to that of
conventional Fermi-liquid superconductors. These FT-STS results thus
provide evidence for a Fermi-liquid behavior in the superconducting
state of optimally doped and overdoped cuprates. Here we
quantitatively
explain all these neutron data
\cite{Bourges95,Fong,Bourges96,Bourges98,Bourges99,Dai,FongNature,He,He2002}
in terms of an order parameter (OP) that has an extended s-wave
symmetry \cite{Zhao,Brandow} and opposite sign in the bonding and
antibonding electron bands formed within the Cu$_{2}$O$_{4}$ bilayers
\cite{Mazin}. In our model, the neutron resonance peak is due to
excitations of electrons from the extended saddle points below the
Fermi level to the superconducting gap edge above the Fermi level.
Within the itinerant magnetism model, neutron scattering intensity at
a wavevector
$\vec{q}$ and an energy $E$ is proportional to the imaginary part of
the dynamic
electron spin susceptibility, $\chi^{\prime\prime}(\vec{q}, E)$.
Qualitatively,
with the neglect of the Bardeen-Cooper-Schrieffer (BCS) coherence
factor,
the bare imaginary part of spin susceptibility,
$\chi^{\prime\prime}_{\circ}(\vec{q}, E)$, is proportional to the
joint density of states $A(\vec{q}, E)= \sum_{\vec{k}}{\delta (E -
E_{\vec{k}+\vec{q}}-E_{\vec{k}})}$, where $E_{\vec{k}}
=\sqrt{\epsilon_{\vec{k}}^{2}+\Delta^{2}_{\vec{k}}}$ is the
quasiparticle dispersion law below $T_{c}$, $\epsilon_{\vec{k}}$ is
the electronic band dispersion, and $\Delta_{\vec{k}}$ is the order
parameter \cite{Mazin}. The two particle energy
$E_{2}(\vec{k},\vec{q})=E_{\vec{k}+\vec{q}}+E_{\vec{k}}$ has a
minimum
and several saddle points for a fixed $\vec{q}$. The minimum defines
the threshold energy, or spin gap energy $E_{g}$, which is achieved
at
vector $\vec{k}$ and $\vec{k}+\vec{q}$ such that both $\vec{k}$ and
$\vec{k}+\vec{q}$ belong to the Fermi surface and thus $E_{g} =
2\Delta_{\vec{k}}$ (Ref.~\cite{Mazin}).
The joint density of
states has divergences at the extended saddle
points \cite{Mazin}. The divergent peak in
$\chi^{\prime\prime}_{\circ}(\vec{q},
\omega)$ occurs because of transitions between the occupied states
located in the extended saddle points below $E_{F}$ and empty
quasiparticle states at the superconducting gap edge above $E_{F}$.
The saddle points
produce Van Hove singularities in
the quasiparticle density of states in the superconducting state at
an energy of
$\sqrt{(\epsilon^{VH}_{\vec{k}})^{2}+\Delta_{\vec{k}}^{2}}$
and the superconducting condensate creates a sharp coherence peak in
the quasiparticle density of states at
the gap edge. Thus, the
divergence in the joint density of states in the superconducting
state is then located at the energy
$E^{*}
=\Delta_{\vec{k}+\vec{q}}+\sqrt{(\epsilon^{VH}_{\vec{k}})^{2}+\Delta_{\vec{k}}^{2}}$. This simple expression for $E^{*}$ has been verified by numerical
calculations in the case of an isotropic s-wave order parameter
\cite{Mazin}.
In Fig.~3, we plot the Fermi surface for a slightly underdoped
BSCCO,
which is inferred by extending a portion of the Fermi surface
determined by
angle-resolved photoemission spectroscopy (ARPES) \cite{White}. There
are extended saddle points that are located near ($\pm$$\pi$, 0) and
(0, $\pm$$\pi$), as shown by ARPES \cite{Dessau}. A solid line in
Fig.~3
represents a segment of the extended saddle points very close to the
Fermi surface.
Other portions of the saddle points [e.g., along the line from (0,
0)
to ($\pi$, 0) ] are significantly away from the
Fermi surface \cite{Dessau} and not shown in the figure. If we only
consider the
magnetic excitations at a fixed wavevector that corresponds to the
antiferromagnetic wavevector $\vec{Q}_{AF}$, only four electron
wavevectors at the Fermi surface are connected by $\vec{Q}_{AF}$ as
indicated by arrow 1 in Fig.~3. Each of these vectors forms an angle
of $\theta_{r}$ with respect to the Cu-O bonding direction. The
electron transitions from the occupied states located at the extended
saddle points below $E_{F}$ to empty quasiparticle states at the
superconducting gap edge above $E_{F}$ are indicated by arrow 2.
Because the quasiparticle densities of states at the gap edge and the
saddle points are divergent at zero temperature, such transitions will
produce a sharp resonance peak at $E_{r}
=\Delta_{\vec{k}+\vec{Q}_{AF}}+\sqrt{(\epsilon^{VH}_{\vec{k}})^{2}+\Delta_{\vec{k}}^{2}}$ (Ref.~\cite{Mazin}).
\begin{figure}[htb]
\ForceWidth{7cm}
\centerline{\BoxedEPSF{Fig3.eps}}
\vspace{0.3cm}
\caption[~]{The Fermi surface for a slightly underdoped BSCCO
with $T_{c}$ = 88 K. This Fermi-surface is extrapolated from a part of the Fermi
surface determined by ARPES \cite{White} using symmetry arguments. Arrow 1
indicates electron transitions from the occupied states in the
superconducting gap edge below $E_{F}$ to empty quasiparticle states
at the gap edge above $E_{F}$. Arrow 2 marks electron transitions
from the occupied states located in the extended saddle points below
$E_{F}$ to empty quasiparticle states in the superconducting gap edge
above $E_{F}$. }
\protect\label{Fig3}
\end{figure}
In terms of $\theta_{r}$, both $E_{g}(\vec{Q}_{AF})$ and
$E_{r}(\vec{Q}_{AF})$ can be rewritten as
\begin{equation}\label{Ne1}
E_{g} (\vec{Q}_{AF})= 2\Delta (\theta_{r})
\end{equation}
and
\begin{equation}\label{Ne2}
E_{r} (\vec{Q}_{AF})= \Delta (\theta_{r}) + \sqrt{[\Delta
(\theta_{r})]^{2} +
(\epsilon^{VH}_{r})^{2}}.
\end{equation}
Here $\sqrt{[\Delta (\theta_{r})]^{2} +(\epsilon^{VH}_{r})^{2}}$ is
the energy of a saddle point below the
Fermi
level along the $\theta_{r}$ direction. One should note that
Eq.~\ref{Ne2} is valid only if the saddle points are very close to the
Fermi surface, as is the case.
We now consider the BCS coherence factor that has been ignored in the
above discussions. The BCS coherence factor is given by \cite{Book}:
\begin{equation}\label{coh1}
\xi (\vec{q},\vec{k})=\frac{1}{2}(1-\frac{\epsilon_{\vec{k}}
\epsilon_{\vec{k}+\vec{q}}+\Delta_{\vec{k}+\vec{q}}\Delta_{\vec{k}}}
{E_{\vec{k}}E_{\vec{k}+\vec{q}}}).
\end{equation}
For $\epsilon_{\vec{k}}$$<$$<$ $\Delta_{\vec{k}}$
and $\epsilon_{\vec{k}+\vec{q}}$ $<$$<$ $\Delta_{\vec{k}+\vec{q}}$,
the
BCS coherence factor is close to 1 when $\Delta_{\vec{k}+\vec{q}}$ and
$\Delta_{\vec{k}}$ have opposite sign, but is close to zero when
$\Delta_{\vec{k}+\vec{q}}$ and $\Delta_{\vec{k}}$ have the same sign.
The coherence factor is not negligible for
$\epsilon_{\vec{k}}$ $\simeq$ $\Delta_{\vec{k}}$ and
$\epsilon_{\vec{k}+\vec{q}}$ = 0 even if $\Delta_{\vec{k}+\vec{q}}$
and $\Delta_{\vec{k}}$ have the same sign.
For a single-layer compound
with d-wave order parameter symmetry, $\Delta_{\vec{k}+\vec{Q}_{AF}}$
and $\Delta_{\vec{k}}$ have opposite sign so that
the BCS coherence factor $\simeq$ 1 and thus
$I^{S}(E_{r})$/$I^{N}(E_{r})$$>$$>$ 1 (see Ref.~\cite{Brinckmann}).
Therefore, the experimental observation of
$I^{S}(E_{r})$/$I^{N}(E_{r})$ $\sim$ 1.0 in the single-layer Tl-2201
\cite{He2002} rules out the d-wave order parameter. Alternatively,
for a single-layer compound with an s-wave symmetry,
$\Delta_{\vec{k}+\vec{Q}_{AF}}$ and $\Delta_{\vec{k}}$ have the same
sign, so the BCS coherence factor could be much less than 1. This may
lead to $I^{S}(E_{r})$/$I^{N}(E_{r})$$\sim$ 1, in agreement with feature (e). Hence, only the intralayer (intraband) s-wave symmetry is
compatible with feature (e): $I^{S}(E_{r})$/$I^{N}(E_{r})$ $\sim$
1.0. On the other hand, if extended saddle points are significantly
below the Fermi level, the BCS coherence factor will be substantial
even for the intralayer (intraband) magnetic scattering in the case of
s-wave gap symmetry. Then $I^{S}(E_{r})$/$I^{N}(E_{r})$ could be much
larger than 1 in this special case. Therefore, the observation of
$I^{S}(E_{r})$/$I^{N}(E_{r})$ $>>$ 1 for the intralayer (intraband)
magnetic scattering is consistent with either s-wave or d-wave gap
symmetry.
For a double-layer compound, interactions within Cu$_{2}$O$_{4}$
bilayers yield bonding
and antibonding bands. Transitions between electronic states
of the same type (bonding-to-bonding or antibonding-to-antibonding)
and those of
opposite types are characterized by even or odd symmetry,
respectively, under exchange
of two adjacent CuO$_{2}$ layers. As a result, the magnetic
excitations between different bands correspond to odd channel
excitations
while the excitations within the same band to even channel
excitations
\cite{Bourges98}. If the order parameters in the bonding and
antibonding bands have the same sign, the magnetic resonance
intensities in both channels should be similar for an intraband
pairing symmetry of either s-wave or
d-wave. This is because the BCS
coherence factors for both channels are the same in this case. On
the
other hand, if the order parameter has an s-wave symmetry and
opposite
sign in the bonding and antibonding electron bands, the magnetic
resonance intensity in the odd channel will be much larger than that
in the even channel due to the large difference in their BCS
coherence
factors. Therefore, only if the order parameter has an s-wave
symmetry and opposite sign in the bonding and antibonding electron
bands, can the observed feature (c):
$I^{S}_{odd}(E_{r})/I^{S}_{even}(E_{r})$ = 5 for the slightly
underdoped YBCO and $I^{S}_{odd}(E_{r})/I^{S}_{even}(E_{r})$ $>$10
for the overdoped YBCO \cite{Bourges96} be explained within the
itinerant magnetism approach.
Based on the t-J model and d-wave pairing symmetry, Brinckmann and
Lee \cite{Lee} have recently
attempted to account for feature (c) using a very unrealistic
parameter: $J_{\perp}$/$J_{\parallel}$ = 0.6
(where $J_{\parallel}$ and $J_{\perp}$ are the effective
intralayer and interlayer antiferromagnetic exchange energies,
respectively). For undoped YBCO, neutron experiments \cite{Bourges98}
show that
$J_{\perp}$/$J_{\parallel}$ = 0.1, which is far less than 0.6 used
in the calculation \cite{Lee}. Further, the value of
$J_{\perp}$/$J_{\parallel}$ will be negligible
if $J_{\parallel}$ remains substantial and the optical magnon gap
($\propto$ $\sqrt{J_{\parallel}J_{\perp}}$) goes to zero, which
should be the case
for optimally doped and overdoped YBCO. Neutron
data for underdoped YBCO \cite{Fong2000} show that the optical
magnon gap is about 50 meV for YBa$_{2}$Cu$_{3}$O$_{6.5}$ and is
reduced to about 25 meV for YBa$_{2}$Cu$_{3}$O$_{6.7}$. If we
linearly extrapolate the optical magnon gap
with the oxygen content, the gap will tend to zero in
YBa$_{2}$Cu$_{3}$O$_{y}$
for $y$ $>$ 0.9, implying that $J_{\perp}$/$J_{\parallel}$ $<<$ 0.1
for
optimally doped and overdoped cuprates.
Using a more realistic parameter of $J_{\perp}$ and the d-wave
pairing symmetry, Millis and
Monien \cite{Millis} appear to be able to explain feature (c). They
showed that the resonance spectral weights in the odd
and even channels are related to the difference between the resonance
energy $E_{r}$ and the particle-hole spin excitation energy 2$\Delta
(\theta_{r}$) at $\vec{Q}_{AF}$, i.e., $W^{odd}/W^{even}$ = $[2\Delta
(\theta_{r})-E_{r}^{odd}]/[2\Delta (\theta_{r})-E_{r}^{even}]$. Here
2$\Delta (\theta_{r})$ must be larger than both $E_{r}^{even}$ and
$E_{r}^{odd}$ (Ref.\cite{Millis}). The INS data of
Y$_{0.9}$Ca$_{0.1}$Ba$_{2}$Cu$_{3}$O$_{7-y}$ would be consistent with
this
theoretical model if one would choose an unrealistic parameter
2$\Delta
(\theta_{r})$ = 49 meV (Ref.~\cite{Pai}). As discussed below,
$\theta_{r}$
is about 15$^{\circ}$ for slightly overdoped cuprates so that
2$\Delta
(\theta_{r})$ =1.73$\Delta_{M}$ (where $\Delta_{M}$ is the maximum
d-wave gap). The intrinsic tunneling spectra, which are
unsusceptible
to surface deterioration, can provide the most reliable
determination of the bulk superconducting gap \cite{Kras}. From the
intrinsic tunneling spectra, one finds that 2$\Delta_{M}/k_{B}T_{c}$
=
8.09 for the optimally doped BSCCO with T$_{c}$ = 94 K
(Ref.~\cite{Kras}), 2$\Delta_{M}/k_{B}T_{c}$ = 6.73 for a slightly overdoped BSCCO
with
\widetext
\begin{table}[htb]
\caption[~]{Comparison of experiments with extended s-wave,
d-wave, and isotopic s-wave. In both extended and isotropic s-wave
models, the order parameters are assumed to have opposite signs in the
bonding and antibonding electron bands formed within the
Cu$_{2}$O$_{4}$ bilayers. Here DA = definitive agreement, A =
agreement, QA = qualitative agreement, PA = possible agreement, D =
disagreement and DD = definitive disagreement.}
\begin{center}
\begin{tabular}{llll}
INS data &extended s-wave&d-wave&isotropic s-wave\\
\hline
Feature (a): $I^{S}$($E_{r}$)/$I^{N}$($E_{r}$) = 3.6 for YBCO & DA&
DA & DA\\
Feature (c): $I^{S}_{odd}(E_{r})/I^{S}_{even}(E_{r})$ = 5 for YBCO &
DA& DD & DA\\
Feature (e): $I^{S}$($E_{r}$)/$I^{N}$($E_{r}$) $<$1 for Tl-2201 & DA& DD & DA\\
The magnitudes of $E_{r}$ and $E_{g}$ & DA& DD & DD\\
Spin gap in La$_{2-x}$Sr$_{x}$CuO$_{4}$ & DA& DD & DD\\
\hline
Other data &extended s-wave&d-wave&isotropic s-wave\\
\hline
Penetration depth &DA&QA&DD\\
Thermal conductivity & DA& QA&DD\\
Specific heat &DA& QA& DD\\
Nonlinear Meissner effect &DA& QA & DD\\
ARPES (underdoped)& A & A& DD\\
ARPES (overdoped) & DA& DD& DD\\
Quasiparticle tunneling (underdoped)& PA& PA& DD\\
Quasiparticle tunneling (overdoped) & DA& DD& DD\\
Raman scattering (underdoped)& PA & PA & DD\\
Raman scattering (overdoped)&DA& DD& PA\\
NMR/NQR & A & A &DD\\
Andreev reflection& A & PA& DD\\
Pb/c-axis YBCO Josephson junction& A& D& A\\
c-axis BSCCO twist Josephson junction& DA& DD& PA\\
Corner SQUID/Josephson junction& PA & PA &PA\\
Tricrystal/Tetracrystal Josephson junctions&PA &PA & PA\\
%\hline \hline
\end{tabular}
\end{center}
%\protect\label{tab:OIS}
\end{table}
\narrowtext
\noindent
T$_{c}$ = 89 K (Ref.~\cite{Kras}), and 2$\Delta_{M}/k_{B}T_{c}$
= 5.37 for an overdoped BSCCO with T$_{c}$ = 80 K (Ref.~\cite{Yu}).
It
is apparent that 2$\Delta_{M}/k_{B}T_{c}$ decreases almost linearly
with T$_{c}$ in the overdoped range. Using the fitted curve of
2$\Delta_{M}/k_{B}T_{c}$ versus T$_{c}$, we
estimate 2$\Delta_{M}/k_{B}T_{c}$ = 6.04 and 2$\Delta (\theta_{r})$
=
38.6 meV for Y$_{0.9}$Ca$_{0.1}$Ba$_{2}$Cu$_{3}$O$_{7-y}$ with
T$_{c}$ =
85.5 K. Similarly
we can obtain 2$\Delta_{M}/k_{B}T_{c}$ = 5.66 and 2$\Delta
(\theta_{r})$ =
35.1 meV for overdoped BSCCO with T$_{c}$ = 83 K. INS experiment on
Y$_{0.9}$Ca$_{0.1}$Ba$_{2}$Cu$_{3}$O$_{7-y}$ (Ref.~\cite{Pai})
indicates $E_{r}^{even}$ = 43 meV $>$
2$\Delta (\theta_{r})$, in contradiction with the theoretical model
\cite{Millis}. For overdoped BSCCO with T$_{c}$ = 83 K,
$E_{r}^{odd}$ =
38 meV (Ref.~\cite{He}) and thus $E_{r}^{even}$ = 45 meV by analogy
with
the case of Y$_{0.9}$Ca$_{0.1}$Ba$_{2}$Cu$_{3}$O$_{7-y}$.
Both $E_{r}^{even}$ and
$E_{r}^{odd}$ in this overdoped BSCCO are far larger than 2$\Delta
(\theta_{r})$, in contradiction with any theoretical models based
on the d-wave pairing symmetry.
For the underdoped YBa$_{2}$Cu$_{3}$O$_{6.7}$, the even-channel
magnetic intensity in the
normal state is a factor of 2.0-2.5 lower than the odd-channel one
for
$E$ = 40 meV, which is about 15 meV above the optical magnon gap
\cite{Fong2000}. This implies that the
interlayer antiferromagnetic correlation does not influence magnetic
excitations
well above the optical magnon gap \cite{Fong2000}. Since
the optical magnon gaps for optimally doped
and overdoped YBCO are much smaller than that for the underdoped
YBa$_{2}$Cu$_{3}$O$_{6.7}$, one should expect that
$I^{N}_{odd}(E)/I_{even}^{N}(E)$ $\simeq$ 1 for $E$ $\simeq $ 40 meV
in optimally doped and overdoped
\newpage
\noindent
YBCO. Thus the observed feature
(c): $I^{S}_{odd}(E_{r})$/$I^{S}_{even}(E_{r})$ $>$$>$ 1 for optimally
doped and overdoped YBCO can only be explained by an OP that has
s-wave symmetry and opposite sign in the bonding and antibonding
electron bands.
For the slightly underdoped YBa$_{2}$Cu$_{3}$O$_{6.92}$, we can also
deduce the value of $I_{even}^{S}(E_{r})$/$I_{even}^{N}(E_{r})$ using the measured
$I_{odd}^{S}(E_{r}) /I_{odd}^{N}(E_{r})$ = 3.6, $I^{S}_{odd}(E_{r})/I_{even}^{S}(E_{r})$ = 5, and the inferred
$I^{N}_{odd}(E_{r})/I_{even}^{N}(E_{r})$ $\simeq$ 1 (see above
discussion). From the measured
$I^{S}_{odd}(E_{r})/I_{even}^{S}(E_{r})$ = 5, we have
$I_{even}^{S}(E_{r}) = 0.2 I_{odd}^{S}(E_{r})$ and thus
$I_{even}^{S}(E_{r})$/$I_{even}^{N}(E_{r})$ $\simeq$
0.2$I_{odd}^{S}(E_{r})$/$I_{odd}^{N}(E_{r})$ = 0.72. This value is
close to that deduced for the single-layer Tl-2201 ($\simeq$ 1).
These results consistently suggest that the intraband neutron
intensity at $E_{r}$ in the superconducting state is close to that in
the normal state.
This unique and important feature we
have identified rules out d-wave order parameter symmetry because
d-wave symmetry predicts \cite{Brinckmann,Lee} that
$I_{even}^{S}(E_{r})/I_{even}^{N}(E_{r})$$>$$>$1 for bilayer
compounds
and $I^{S}(E_{r})/I^{N}(E_{r})$$>$$>$1 for single-layer compounds.
From Eqs.~\ref{Ne1} and \ref{Ne2}, it is easy to calculate
$E_{g}$ and $E_{r}$ if
one knows the gap function $\Delta (\theta) $ and the $\theta_{r}$
value. From the measured Fermi surface, one can readily determine
$\theta_{r}$. For example, we find $\theta_{r}$ =
18.4$^{\circ}$ for a slightly underdoped BSCCO from Fig.~3. For
~optimally ~doped cuprates, we get $\theta_{r}$ $\simeq$
16.0$^{\circ}$. If we would use an isotropic s-wave gap function
$\Delta
(\theta )$ = 28 meV for a slightly overdoped YBCO, we would have
$E_{g}$ = 56 meV and $E_{r}$ $>$ 56 meV, which are far larger than
the
measured $E_{g}$ = 33 meV and $E_{r}$ = 40 meV
(Ref.~\cite{Bourges96}). If we would use a d-wave gap function
$\Delta
(\theta )$= 28$\cos 2\theta$ meV, we would have $E_{g}$ = 47.5 meV
and
$E_{r}$ $>$ 47.5 meV, which are also far larger than the measured
values.
Thus, one cannot quantitatively explain the neutron data in terms of
d-wave
and isotropic s-wave symmetries.
Alternatively, an extended s-wave with eight line nodes ($A_{1g}$
symmetry) is in quantitative agreement with two-thirds of the
experiments
that were designed to test the order-parameter symmetry for
hole-doped
cuprates \cite{Zhao}. The remaining one-third (e.g.,
tricrystal grain-boundary Josephson junction experiments) are
explained
qualitatively by Zhao \cite{Zhao} and by Brandow \cite{Brandow}. In
Table 1, we compare nearly all the experiments used to test the OP
symmetry with the extended s-wave, d-wave, and isotropic s-wave
models.
In both extended s-wave and isotropic s-wave models, the order
parameters are assumed to have opposite sign in the bonding and
antibonding
electron bands formed within the Cu$_{2}$O$_{4}$ bilayers
\cite{Mazin}. The
detailed comparisons with other experiments are made in
Ref.~\cite{Zhao}. From Table 1, one can see that the neutron data
alone provide a definitive answer to the intrinsic, bulk OP symmetry
because INS is a bulk, phase and angle sensitive technique. Other
bulk and non-phase sensitive experiments provide complementary
support
to the present conclusions. The phase and surface sensitive
experiments cannot definitively
determine the intrinsic bulk OP symmetry because the surface OP might
be different from the bulk
one \cite{Zhao}.
For a slightly overdoped YBCO, more than six independent experiments
consistently suggest that \cite{Zhao} the gap function is $\Delta
(\theta ) = 24.5(\cos 4\theta + 0.225)$ meV. Substituting
$\theta_{r}$ = 16$^{\circ}$ into the gap function, we get $\Delta
(\theta_{r})$ = 16.3 meV and thus $E_{g}$ = 32.6 meV, in quantitative
agreement with the measured one (32-33 meV).
In our model, the position of the extended saddle point
along the $\theta_{r}$ direction is located at $E_{r}-E_{g}/2$ in the
superconducting state (see Eqs.~\ref{Ne1} and \ref{Ne2}). For
optimally doped YBCO with $E_{r}$ = 41 meV and $E_{g}$ = 32 meV, we
find that the saddle point in the superconducting state is located at
an energy of 25 meV below the Fermi level. This is in agreement with
the ARPES studies \cite{King} which suggest that the Fermi level in
the superconducting state for optimally doped cuprates is $\leq$ 30
meV above the extended saddle points that have the same energy over a
large momentum space \cite{Dessau}. Further, electronic Raman
scattering spectra in YBa$_{2}$Cu$_{4}$O$_{8}$ have been used to
determine the energy of the extended saddle points more accurately
\cite{Sherman}. At 10 K (well below $T_{c}$), the energy of the
extended saddle points is found to be 24.3 meV below the Fermi level,
in excellent agreement with the result predicted by our model from
the
INS data ($\simeq$ 25 meV).
We have identified the unique and important
feature: $I^{S}(E_{r})$/$I^{N}(E_{r})$ $\simeq$ 1.0 for the
intralayer
and intraband magnetic scattering. This feature unambiguously rules
out
the d-wave OP symmetry. The unambiguous determination of the
intrinsic extended s-wave pairing symmetry for hole-doped cuprates
places strong constraints on the pairing mechanism for
high-temperature superconductivity. A recent calculation suggests
that high energy Cu-O charge fluctuations can
lead to an attractive interaction between conduction electrons and
the
pairing symmetry may be of extended s-wave ($A_{1g}$) \cite{Miklos}.
Indeed, precise thermal-difference reflectance
spectra of several cuprate superconductors ($T_{c}$ =105-120 K)
exhibit
pronounced features at photon energies of about 2.0 eV, which may be
related to the Cu-O charge fluctuations (Ref.~\cite{Little}). These
features can be well described within
Eliashberg theory with an electron-boson coupling constant
$\lambda_{ch}$ of
about 0.40. In order to explain a superconducting transition
temperature of 105-120 K,
the authors simulated \cite{Little} an electron-phonon coupling
feature at 50 meV with the coupling constant $\lambda_{ph}$ of about
1.0. This Eliashberg model's simulated electron-phonon coupling
agrees
well with the results obtained from first principle calculations
\cite{Kra}.
The contribution of this 2 eV component can be
equivalent to a {\em negative} Coulomb pseudopotential $\mu^{*}$
within Eliashberg theory
\cite{Zhao1}. By using a realistic electron-phonon coupling
spectral
weight deduced from tunneling spectra and a $\mu^{*}$ = -0.15, and
taking into a polaronic effect, Zhao
{\em et al. } \cite{Zhao1} are able to explain the negligible isotope
effect
on $T_{c}$, and the magnitudes of $T_{c}$ and the superconducting gap
for
optimally doped 90 K superconductors.
In summary, we have identified several important features in the
neutron scattering data of
cuprates, which are difficult to be explained in terms of d-wave and
isotropic s-wave
order parameters. Alternatively, we show that the neutron data are
in
quantitative agreement with an order parameter that has an extended
s-wave ($A_{1g}$) symmetry and opposite sign in
the bonding and antibonding electron bands formed within the
Cu$_{2}$O$_{4}$ bilayers. This $A_{1g}$ pairing
symmetry may be compatible with a charge fluctuation mediated pairing
mechanism. High-temperature superconductivity in cuprates may be due
to
the combination of strong electron-phonon coupling and substantial
electron-charge-fluctuation coupling \cite{Little,Zhao1}.
~\\
~\\
$^{*}$Correspondence should be addressed to gzhao2@calstatela.edu
\begin{thebibliography}{99}
\bibliographystyle{prsty}
\bibitem{Bourges95} P. Bourges, Y. Sidis, B. Hennion, R. Villeneuve,
G. Collin, and J. F. Marucco, Physica B, \textbf{213-214}, 48 (1995).
\bibitem{Fong} H. F. Fong, B. Keimer, P. W. Anderson, D.
Reznik,
F. Dogan, and I. A. Aksay, Phys. Rev. Lett. \textbf{75}, 316
(1995).
\bibitem{Bourges96} P. Bourges, L. P. Regnault, Y. Sidis, and
C. Vettier, Phys. Rev. B {\bf 53}, 876 (1996).
\bibitem{Bourges98}P. Bourges, {\em The gap Symmetry and
Fluctuations
in High Temperature Superconductors}, Edited by J. Bok, G.
Deutscher,
D. Pavuna, and S. A. Wolf. S. A. (Plenum Press, 1998) p 349 (Vol.
371
in NATO ASI series, Physics).
\bibitem{Bourges99}P. Bourges, Y. Sidis, H. F. Fong, B. Keimer,
L. P.
Regnault, J. Bossy, A. S. Ivanov, D. L. Milius, and I. A. Aksay,
{\em High Temperature Superconductivity }, Edited by Barnes S. E.
{\it et al} p207 (CP483 American Institute of Physics, Amsterdam,
1999)
\bibitem{Dai} P. C. Dai, H. A. Mook, R. D. Hunt, and F. Dogan,
Phys. Rev. B {\bf 63}, 054525 (2001).
\bibitem{FongNature} H. F. Fong, P. Bourges, Y. Sidis, L. P.
Regnault,
A. Ivanov, G. D. Guk, N. Koshizuka, and B. Keimer,
Nature(London)
\textbf{398}, 588 (1999).
\bibitem{He} H. He, Y. Sidis, P. Bourges, G. D. Gu, A. Ivanov,
N.
Koshizuka, B. Liang, C. T. Lin, L. P. Regnault, E. Schoenherr,
and B. Keimer, Phys. Rev. Lett. \textbf{86}, 1610 (2001)
\bibitem{He2002}H. He, P. Bourges, Y. Sidis, C. Ulrich, L. P.
Regnault,
S. Pailhes, N. S. Berzigiarova, N. N. Kolesnikov, and B. Keimer,
Science {\bf 295}, 1045 (2002).
\bibitem{Demler} E. Demler and S. C. Zhang, Phys.
Rev. Lett. \textbf{75}, 4126 (1995).
\bibitem{Abanov}A. Abanov and
A. V. Chubukov, Phys. Rev. Lett. \textbf{83}, 1652 (1999).
\bibitem{Morr} D. K. Morr and D. Pines, Phys. Rev. Lett.
\textbf{81}, 1086 (1998).
\bibitem{Brinckmann} J. Brinckmann andP. A. Lee, Phys. Rev.
Lett.\textbf{82}, 2915 (1999).
\bibitem{Mazin}I. I. Mazin and V. M. Yakovenko, Phys. Rev.
Lett.
\textbf{75}, 4134 (1995).
\bibitem{Hardy}W. N. Hardy, D. A. Bonn, D. C.
Morgan, Ruixing Liang, and K. Zhang, Phys. Rev. Lett.
\textbf{70},
3999 (1993).
\bibitem{Chiao}M. Chiao, R. W. Hill, C. Lupien, L.
Taillefer, P. Lambert, R. Gagnon, and P. Fournier, Phys. Rev. B
\textbf{62}, 3554 (2000).
\bibitem{Zhao} G. M. Zhao, Phys. Rev. B {\bf 64}, 024503 (2001).
\bibitem{Brandow} B. H. Brandow, Phys. Rev. B {\bf 65}, 054503
(2002).
\bibitem{Pai} S. Pailhes, Y. Sidis, P. Bourges, C. Ulrich, V. Hinkov,
L.P. Regnault,
A. Ivanov, B. Liang, C.T. Lin, C. Bernhard, and B. Keimer,
cond-mat/0308394.
\bibitem{Norman}O. Tchernyshyov, M. R. Norman, and A. V. Chubukov,
Phys. Rev. B {\bf 63}, 144507 (2001).
\bibitem{Davis} K. McElroy, R. W. Simmonds, J. E. Hoffman, D.-H.
Lee,
J. Orenstein, H. Eisaki, S. Uchida, and J. C. Davis, J. C.
Nature
(London) {\bf 422}, 592 (2003).
\bibitem{White} P. J. White, Z.-X. Shen, C. Kim, J. M. Harris,
A. G. Loeser, P. Fournier, and A. Kapitulnik, Phys. Rev. B
\textbf{54}, R15 669 (1996).
\bibitem{Dessau} D. S. Dessau, Z.-X. Shen, D. M. King, D. S.
Marshall,
L. W. Lombardo, P. H. Dickinson, A. G. Loeser, J. DiCarlo,
C.-H.
Park, A. Kapitulnik, and W. E. Spicer, Phys. Rev. Lett. {\bf
71},
2781 (1993).
\bibitem{Book} J. R. Schrieffer, {\em Theory of Superconductivity},
(Frontiers in Physics (20), Addison Wesley, 1988).
\bibitem{Lee} J. Brinckmann and P. Lee, Phys. Rev. B {\bf 65},
014502 (2001).
\bibitem{Fong2000} H. F. Fong, P. Bourges, Y. Sidis, L. P.
Regnault,
J. Bossy, A. Ivanov, D. L. Milius, I. A. Aksay, and B. Keimer,
Phys. Rev. B {\bf 61}, 14773 (2000).
\bibitem{Millis} A. J. Millis and H. Monien, Phys. Rev. B {\bf 54},
16172 (1996).
\bibitem{Kras}V.M. Krasnov, A. Yurgens, D. Winkler, P. Delsing, and
T. Claeson,
Phys. Rev. Lett. {\bf 84}, 5860 (2000).
\bibitem{Yu} A. Yurgens, D.Winkler, T.Claeson, S. J. Hwang, J. H.
Choy, Int. J. Mod. Phys. B {\bf 13}, 3758 (1999).
\bibitem{King}D. M. King, Z.-X. Shen, D. S. Dessau, D. S.
Marshall, C. H. Park, W. E. Spicer, J. L. Peng, Z. Y. Li,
and R. L. Greene, Phys. Rev. Lett. \textbf{73}, 3298 (1994).
\bibitem{Sherman}E. Ya. Sherman and O. V. Misochko, Phys. Rev.
B {\bf 59}, 195 (1999).
\bibitem{Miklos} M. Gulacsi and R. J. Chan, J. Supercond. {\bf
14}, 651 (2001), and private communication with M. Gulacsi.
\bibitem{Little}M. J. Holcomb, J. P. Collman, and W. A. Little,
Phys. Rev. Lett. {\bf 73}, 2360 (1994); M. J. Holcomb, C. L.
Perry, J. P. Collman, and W. A. Little, Phys. Rev. B {\bf 53},
6734 (1996); M. J. Holcomb, Phys. Rev. B {\bf 54}, 6648 (1996).
\bibitem{Kra} W. E. Pickett, Rev. Mod. Phys. {\bf 61}, 434
(1989). \bibitem{Zhao1} G. M. Zhao, V. Kirtikar, and D. E.
Morris, Phys. Rev. B {\bf 63}, R220506 (2001).
\end{thebibliography}
\end{document}