Least momentum space frustration as a condition for "high $T_c$ sweet spot" in the iron-based superconductors
Extracted source text ยท arXiv 1204.1717
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% INSTITUTE OF PHYSICS PUBLISHING %
% %
% `Preparing an article for publication in an Institute of Physics %
% Publishing journal using LaTeX' %
% %
% LaTeX source code `ioplau2e.tex' used to generate `author %
% guidelines', the documentation explaining and demonstrating use %
% of the Institute of Physics Publishing LaTeX preprint files %
% `iopart.cls, iopart12.clo and iopart10.clo'. %
% %
% `ioplau2e.tex' itself uses LaTeX with `iopart.cls' %
% %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
%
% First we have a character check
%
% ! exclamation mark " double quote
% # hash ` opening quote (grave)
% & ampersand ' closing quote (acute)
% $ dollar % percent
% ( open parenthesis ) close paren.
% - hyphen = equals sign
% | vertical bar ~ tilde
% @ at sign _ underscore
% { open curly brace } close curly
% [ open square ] close square bracket
% + plus sign ; semi-colon
% * asterisk : colon
% < open angle bracket > close angle
% , comma . full stop
% ? question mark / forward slash
% \ backslash ^ circumflex
%
% ABCDEFGHIJKLMNOPQRSTUVWXYZ
% abcdefghijklmnopqrstuvwxyz
% 1234567890
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
\documentclass[12pt]{iopart}
\newcommand{\gguide}{{\it Preparing graphics for IOP journals}}
%Uncomment next line if AMS fonts required
%\usepackage{iopams}
\usepackage{graphicx}
\begin{document}
\title{Least momentum space frustration as a condition
for ``high $T_c$ sweet spot'' in the iron-based superconductors}
\author{Hidetomo Usui$^1$, Katsuhiro Suzuki$^{1,2}$,
and Kazuhiko Kuroki$^{1,2}$}
\address{$^1$Department of Engineering Science, The University of Electro-Communications, Chofu, Tokyo 182-8585, Japan \\
$^2$JST, TRIP, Chiyoda, Tokyo 102-0075, Japan}
\ead{kuroki@vivace.e-one.uec.ac.jp}
\begin{abstract}
In the present paper, we describe how the band structure and the
Fermi surface of the iron-based superconductors vary as the
Fe-As-Fe bond angle changes. We discuss how these Fermi surface
configurations affect the superconductivity mediated by spin
fluctuations, and show that in several situations, frustration in the
sign of the gap function arises due to the repulsive pairing
interactions that requires sign change of the order parameter.
Such a frustration can result in nodes or very small gaps, and
generally works destructively against superconductivity.
Conversely, we propose that the optimal condition for superconductivity
is realized for the Fermi surface configuration that gives the
least frustration while maximizing the Fermi surface
multiplicity. This is realized when there are three hole
Fermi surfaces, where two of them have $d_{XZ/YZ}$ orbital
character and one has $d_{X^2-Y^2}$ {\it for all $k_z$} in the
three dimensional Brillouin zone. Looking at the band structures of
various iron-based superconductors, the
occurrence of such a ``sweet spot'' situation is limited to a narrow
window.
\end{abstract}
%Uncomment for PACS numbers title message
%\pacs{00.00, 20.00, 42.10}
% Keywords required only for MST, PB, PMB, PM, JOA, JOB?
%\vspace{2pc}
%\noindent{\it Keywords}: Article preparation, IOP journals
% Uncomment for Submitted to journal title message
%\submitto{\JPA}
% Comment out if separate title page not required
\maketitle
\section{Introduction}
The discovery of high temperature superconductivity in the
iron-based superconductors\cite{Hosono} has attracted much attention
in many aspects.
Not only the high $T_c$ itself, but also a number of experiments
indicating non-universality of the superconducting gap function,
such as sign reversing, anisotropy, or the presence of nodes,
suggest an unconventional pairing mechanism.
Most probable candidate for such an unconventional mechanism
is the pairing mediated by
spin fluctuations, where the superconducting gap changes
sign between the disconnected Fermi surfaces, namely,
the so-called $s\pm$ pairing\cite{Mazin,Kuroki1st}.
Back in 2001, one of the present authors proposed that
spin fluctuation mediated pairing in systems with nested
disconnected Fermi surfaces may give rise to a very high $T_c$
superconductivity\cite{KA,KA2}.
The idea is that the repulsive pairing interaction
mediated by spin fluctuations can be fully exploited without
introducing nodes of the superconducting gap on the Fermi surfaces.
Although the Fermi surface of the iron-based superconductors
does resemble the proposed Fermi surface configuration, there are
some important differences. One is that the Fermi surface in the
iron-based superconductors has multiple orbital characters, and
second is that there can be frustrations in the sign of the gap
function, which can give rise to gap nodes on the Fermi surface.
In the present paper, we focus on this gap-sign frustration problem,
and propose that the Fermi surface configuration that gives the
least frustration provides the optimal condition for high $T_c$
in the iron-based superconductors. In ref.\cite{Usui}, two of the
present authors pointed out that maximizing the Fermi surface
multiplicity leads to the optimization for $T_c$, but considering
the frustration problem studied in the present paper, the
``sweet spot'' for high $T_c$ is further limited to a narrow window
of the Fermi surface configuration.
\section{Typical band structure and Fermi surfaces}
Let us first describe the band structure of LaFeAsO.
LaFeAsO takes a layered structure, where Fe atoms form a
square lattice in each layer,
sandwiched by As atoms with tetrahedral coordination.
We use the band structure obtained from first principles\cite{pwscf}
to construct the maximally localized
Wannier functions\cite{MaxLoc}.
These Wannier functions
have five orbital symmetries ($d_{3Z^2-R^2}$,
$d_{XZ}$, $d_{YZ}$, $d_{X^2-Y^2}$,
$d_{XY}$), where $X, Y, Z$ refer to those for this unit cell
with two Fe sites as shown in Fig.\ref{fig1}(a).
The two Wannier orbitals in
each unit cell are equivalent in that each Fe atom has the same
local arrangement of other atoms.
We can then take a unit cell that
contains only one orbital per symmetry by
unfolding the Brillouin zone,
and an effective five-band model on a
square lattice is obtained, where
$x$ and $y$ axes are rotated by
45 degrees from $X$-$Y$.
\begin{figure}
\begin{center}
\includegraphics[width=8cm,clip]{fig1.eps}
\end{center}
\caption{(a) The original (dashed lines) and
reduced (solid) unit cells with $\bullet$ (Fe), $\nabla$ (As
below the Fe plane) and $\triangle$ (above Fe).
(b) The band structure (left) of the five-band model for LaFeAsO, and the
Fermi surface (right) at $k_z=0$ for 10 percent electron doping.
The main orbital
characters of some portions of the bands and the Fermi surface
are indicated.
The dashed horizontal line in the band structure indicates the Fermi level
for 10 percent electron doping.
The short arrow in the band structure indicates the
position of the Dirac cone closest to the Fermi level.
The gray areas in the Fermi surface
around the zone corners represent the $\gamma$ Fermi surface,
which is barely absent for 10 percent electron doping.
(c) The portion of the band that has mainly the $d_{X^2-Y^2}$
orbital character.
\label{fig1}}
\end{figure}
In Fig.\ref{fig1}(b) (right),
the Fermi surface for 10 percent electron doping
is shown in the two-dimensional unfolded Brillouin zone.
The Fermi surface consists of four pieces:
two concentric hole pockets (denoted here as $\alpha_1$, $\alpha_2$)
centered around $(k_x, k_y)=(0,0)$, two electron pockets
around $(\pi,0)$ $(\beta_1)$ or $(0,\pi)$ $(\beta_2)$, respectively.
Besides these pieces of the Fermi surface, there is a portion of the band
near $(\pi,\pi)$ that
touches the $E_F$, so that
the portion acts as a ``quasi Fermi surface $(\gamma)$'' around $(\pi,\pi)$.
As for the orbital character, $\alpha$ and portions of $\beta$ near
Brillouin zone edge have mainly $d_{XZ}$ and $d_{YZ}$ character,
while the portions of
$\beta$ away from the Brillouin zone edge and $\gamma$ have
mainly $d_{X^2-Y^2}$ orbital character (see also Fig.\ref{fig1}(c)).
\section{Fermi surface appearance/disappearance against the bond angle}
\label{FSvsangle}
The band structure and the Fermi surfaces of the iron-based
superconductors are sensitive to the
lattice structure. In this section, we consider
hypothetical lattice structures of LaFeAsO, where
we fix the bond length at its original length
and vary the bond angle $\alpha$(Fig.\ref{fig2}(a))\cite{Usui}.
We first neglect the three dimensionality (out-of-plane hoppings), and
consider a two dimensional model.
The Fermi surface is obtained for 10 percent electron doping.
When the bond angle is large, two hole Fermi surfaces,
$\alpha_1$ and $\alpha_2$
are present around the wave vector (0,0), while the
$\gamma$ around $(\pi,\pi)$ is missing. As $\alpha$ decreases,
the $\gamma$ Fermi surface appears around $(\pi,\pi)$, and there are now
three hole Fermi surfaces. This appearance of the additional Fermi surface
has been noticed as an effect of increasing the pnictogen height
\cite{Singh,Vildosola,KKprb,Lebegue,OKAndersen}.
When $\alpha$ is further reduced, the $\alpha_1$ Fermi surface
disappears, and again the
Fermi surface multiplicity reduces to two, but in this
case one around $(0,0)$ and another around $(\pi,\pi)$.
Such a disappearance of the $\alpha_1$ hole Fermi surface was
first noticed in the band calculation of Ca$_2$Al$_4$O$_6$Fe$_2$As$_2$
\cite{Shirage} by Miyake {\it et al.} in refs.\cite{Miyake,Kosugi}.
This material indeed has very small bond angle of about 102 degrees.
\begin{figure}
\begin{center}
\includegraphics[width=8cm,clip]{fig2.eps}
\end{center}
\caption{ The band structure and the Fermi surface of hypothetical lattice
structures of LaFeAsO with varying the Fe-As-Fe bond angle
while fixing the Fe-As bond length to its original value.
Position of the bands having $X^2-Y^2$ and $XZ/YZ$ orbital characters
at (0,0) is indicated.}
\label{fig2}
\end{figure}
Fig.\ref{fig3} explains schematically the band structure/Fermi surface
variation against the
bond angle reduction\cite{Miyake}.
As the bond angle is reduced, the $X^2-Y^2$ band below
the Fermi level around $(\pi,\pi)$ rises up, while another $X^2-Y^2$ portion
of the band above the Fermi level around $(0,0)$ comes down.
This band deformation can be understood from Fig.\ref{fig1}(c), where the
$X^2-Y^2$ portion of the band for the original LaFeAsO is extracted.
In the tightbinding picture, the energy difference between
the wave vectors $(0,0)$ and $(\pi,\pi)$ is roughly equal to $8t_{X^2-Y^2}$,
where $t_{X^2-Y^2}$
is the nearest neighbor hopping of the $X^2-Y^2$ orbital. As the bond angle
is reduced, the contribution to $t_{X^2-Y^2}$ from the
Fe$\rightarrow$As$\rightarrow$Fe path decreases, while that from the direct
Fe$\rightarrow$ Fe path increases. The two contributions have opposite signs,
so that the reduction of the bond angle results
in a decrease of $t_{X^2-Y^2}$\cite{Miyake}.
When the bottom of the upper $X^2-Y^2$ portion
sinks below the $XZ/YZ$ bands around $(0,0)$,
a band structure reconstruction takes place, and now the two bands that are
degenerate at $(0,0)$ repel with
each other (one going up, the other going down)
as the wave vector moves away from $(0,0)$ (Fig.\ref{fig3}(c)).
In this situation, the band below these two bands
has $X^2-Y^2$ character near $(0,0)$, and changes its character to
$XZ/YZ$ as the wave number increases. Therefore, just before the
inner hole ($\alpha_1$) Fermi surface disappears,
it has strong $X^2-Y^2$ character.
\begin{figure}
\begin{center}
\includegraphics[width=8cm,clip]{fig3.eps}
\end{center}
\caption{A schematic figure for the band structure variation against
the bond angle $\alpha$. The black (red) portions indicate the bands
with strong $X^2-Y^2$ $(XZ/YZ)$ orbital character.}
\label{fig3}
\end{figure}
\section{Frustration in the superconducting gap
in the absence of $\gamma$ Fermi surface : a brief review}
Having understood the Fermi surface variation against the bond angle,
we now consider how this should affect the gap function of the
superconductivity mediated by spin fluctuations.
The pairing interaction mediated by spin fluctuations is repulsive,
so the superconducting gap has an tendency of changing its sign
between the initial and final wave vectors of the pair scattering.
Another important point is that the spin fluctuations develop at
wave vectors that bridge the portions of the Fermi surface having the
same orbital character\cite{KKprb}.
In the presence of the $X^2-Y^2$ originated $\gamma$ and $XZ/YZ$ originated
$\alpha_1$ and $\alpha_2$ Fermi surfaces (case (b) in Fig.\ref{fig3}) ,
it is known that the $\gamma$-$\beta$ interaction among portions having
$X^2-Y^2$ character and the $\alpha$-$\beta$ interactions among portions with
$XZ/YZ$ character dominate, and the superconducting gap is fully open on
all the Fermi surfaces while changing its sign as $+$, $-$, $+$ along
$\alpha$, $\beta$, $\gamma$ Fermi surfaces. This is the fully gapped
$s\pm$ gap\cite{KKprb,DHLee,Thomale}.
On the other hand, in the absence of the $\gamma$ Fermi surface
(case (a) in Fig.\ref{fig3}), the pairing
interaction between the
$X^2-Y^2$ portions of the $\beta$ Fermi surfaces and that
between the $XZ/YZ$ portions of the $\alpha$ and $\beta$ Fermi surfaces
results in a frustration in the sign of the gap function.
This situation was studied
in several previous papers\cite{Graser,KKprb,DHLee,Thomale}.
This can result in
either nodal $s\pm$-wave or $d$-wave pairings, where the nodes of the gap
go into the $\beta$ Fermi surface in the former, and $\alpha$ in the latter.
A schematic figure summarizing the above
is shown in Fig.\ref{fig4}\cite{KKprb}.
As was shown in ref.\cite{KKprb}, this frustration effect degrades
$T_c$ of the superconductivity, so that the lattice structure acts as
a switch between high $T_c$ nodeless and low $T_c$ nodal superconductivity.
\begin{figure}
\begin{center}
\includegraphics[width=10cm,clip]{fig4.eps}
\end{center}
\caption{The fully-gapped $s\pm$ wave, the nodal
$s\pm$ wave and $d$ wave gap
are schematically shown. The solid red (dashed blue) curves
represent positive (negative) sign of the gap. The arrows
indicate the dominating nesting vectors.}
\label{fig4}
\end{figure}
\section{Frustration in the case of nearly vanishing $\alpha_1$ Fermi surface}
Here we discuss another situation where the frustration arises in the
sign of the superconducting gap function.
As mentioned in section\ref{FSvsangle},
the $X^2-Y^2$ orbital character strongly
mixes into the $\alpha_1$ Fermi surface
just before the Fermi surface vanishes as the
bond angle is reduced. In this situation, there are now $X^2-Y^2$
components on $\alpha_1$, $\beta$, and $\gamma$ (if present) Fermi surfaces.
Since these Fermi surfaces interact with repulsive pairing interactions,
once again, a frustration arises in the sign of the superconducting gap
(Fig.\ref{fig5}(a)).
In addition to this, there can also be
some $XZ/YZ$ component remaining in the $\alpha_1$ Fermi surface,
and this portion tends to change the sign from the $\beta$ Fermi
surfaces (Fig.\ref{fig5}(b)) so this can be another factor for the
frustration.
To actually see this frustration effect,
here we consider
hypothetical lattice structures of
Ca$_4$Al$_2$O$_6$Fe$_2$As$_2$\cite{Shirage},
where we vary the bond angle from 108 to 109 degrees,
and contrust a two dimensional five orbital model in the unfolded
Brillouin zone.
In this angle regime, the $\alpha_1$ Fermi surface is barely
present, and it is
indeed composed of mixed $X^2-Y^2$ and $XZ/YZ$ orbital components.
We apply fluctuation exchange approximation to this model\cite{Bickers},
and obtain the eigenfunction (gap function) of the linearized
Eliashberg equation as was done in ref.\cite{Usui}.
In Fig.\ref{fig6}, we show the gap function for the
two angles for 10 percent electron doping and temperature $T=0.01$eV.
It can be seen that the magnitude of the gap on the
$\alpha_1$ Fermi surface is very small, and its sign actually changes as the
bond angle is varied, reflecting the frustration.
\begin{figure}
\begin{center}
\includegraphics[width=8cm,clip]{fig5.eps}
\end{center}
\caption{Dominating pairing interactions for the (a)$X^2-Y^2$ and (b) $XZ/YZ$
portions of the Fermi surface in the case where the inner hole Fermi surface
($\alpha_1$) is barely present. In this case, $\alpha_1$ is a mixture of
$X^2-Y^2$ and $XZ/YZ$.}
\label{fig5}
\end{figure}
\begin{figure}
\begin{center}
\includegraphics[width=12cm,clip]{fig6.eps}
\end{center}
\caption{The gap function obtained by FLEX for the
hypothetical lattice structures of Ca$_4$Al$_2$O$_6$Fe$_2$As$_2$.
The bond angle $\alpha$ is varied to 108 or 109 degrees, while the
bond length is fixed to the original value.}
\label{fig6}
\end{figure}
\section{Effect of three dimensionality}
When the systems exhibit some three dimensionality , the
above mentioned variation of the Fermi surface configuration
against the bond angle depends on $k_z$. This is shown
schematically in Fig.\ref{fig7}. In systems with moderate three
dimensionality, the orbital character change of the $\alpha_1$
Fermi surface and its disappearance first starts at $k_z=\pi$
(around Z point)
as the bond angle is reduced, and ends at $k_z=0$ ($\Gamma$ point).
Namely, the
$\alpha_1$ Fermi surface becomes three dimensional before it
disappears completely. In this case,
the above mentioned frustration effect should be present
around the top and the bottom portions of the three dimensional
Fermi surface.
In fact, the above mentioned three dimensionality of the band structure is
rather common for the iron-based superconductors.
This has been discussed in detail in ref.\cite{OKAndersen}.
In Fig.\ref{fig8}, we show the band structure of various
iron-based superconductors for $k_z=0$ and $k_z=\pi$ planes
(in the original folded Brillouin zone) calculated by
using the Wien2k package\cite{Wien2k}. One can see that, except for
LaFeAsO, the two bands degenerate at Z point $(0,0,\pi)$
repel with each other
as the wave vector moves away from Z in the $k_z=\pi$ plane.
This means that configurations (c) or (d) in Fig.\ref{fig3}
are realized for a certain range of $k_z$. On the other hand,
in FeSe, LiFeAs and BaFe$_2$As$_2$ the band structure near the
$k_z=0$ plane takes configuration (b), so that in these materials,
the band structure near the Fermi level is indeed three dimensional.
Conversely, the 1111 systems can be considered as somewhat exceptional
in that, although the three dimensional $X^2-Y^2$ band does exist,
it does not come down too rapidly before the $X^2-Y^2$ originated $\gamma$
Fermi surface appears. In fact, NdFeAsO, one of the
materials having the highest $T_c$, seems to have the least frustration
from the above mentioned viewpoint. Namely, as seen in the band structure
shown in Fig.\ref{fig9} in the unfolded Brillouin zone\cite{KKprb},
the $\gamma$
Fermi surface is present but the three dimensional $X^2-Y^2$ band
still lies above the $XZ/YZ$ bands for all $k_z$.
On the other hand, from the comparison to the band structure of LaFeAsO
(Fig.\ref{fig1}(b)),
it can be seen that
as a trade-off for the appearance of the $\gamma$
Fermi surface around $(\pi,\pi)$, the three dimensional $X^2-Y^2$ band
along $(0,0,0)$-$(0,0,\pi)$
has certainly come down very close to the $XZ/YZ$ band,
and for smaller bond angle or higher pnictogen position,
the band reconstruction and thus the frustration starts to take place.
Therefore, from the present viewpoint, the ``sweet spot'' for high $T_c$
is restricted to a narrow window,
which may explain the experimental observations\cite{Lee,Takano}.
\begin{figure}
\begin{center}
\includegraphics[width=10cm,clip]{fig7.eps}
\end{center}
\caption{A schematic figure representing the band structure and
the hole Fermi surface variation against the bond angle
for moderately three dimensional systems. (a) to (d) correspond to
the configurations shown in Fig.\ref{fig3}. The ``sweet spot''
is restricted to the regime where the $X^2-Y^2$ originated
$\gamma$ Fermi surface
is effective and the $\alpha_1$ Fermi surface has $XZ/YZ$ character
for all $k_z$. }
\label{fig7}
\end{figure}
\begin{figure}
\begin{center}
\includegraphics[width=10cm,clip]{fig8.eps}
\end{center}
\caption{First principles band structure for various iron-based
superconductors. $k_z=0$ and $k_z=\pi$ planes are shown.
The red solid circles (or elipse) indicates the portion
where the two bands degenerate at wave vector $(k_x,k_y)=(0,0)$
repel with each other as the wave number increases, while the
dashed blue cirles are the portions where the two bands degenerate
at (0,0) both form a hole Fermi surface. For LaFeAsO, a typical
example for 1111 systems, the three dimensional band along $\Gamma$-Z
having $X^2-Y^2$ character (green dashed elipse)
stays above the degenerate $XZ/YZ$ bands (red solid).}
\label{fig8}
\end{figure}
\begin{figure}
\begin{center}
\includegraphics[width=10cm,clip]{fig9.eps}
\end{center}
\caption{The band structure of NdFeAsO in the unfolded Brillouin zone.
The Fermi level is for 10 percent electron doping.
The two conditions for least
frustration is satisfied: (i) the $X^2-Y^2$ Fermi surface is present
around the wave vector $(\pi,\pi)$, and (ii) the three dimensional
$X^2-Y^2$ band
from $(0,0,0)$ to $(0,0,\pi)$ does not
intersect the $XZ/YZ$ bands for all $k_z$.}
\label{fig9}
\end{figure}
\section{Conclusion}
As described in the present paper, the multiplicity of the
Fermi surfaces and even their orbital characters change
as the lattice structure is varied. Because of the presence of
multiple Fermi surfaces and the repulsive pairing interaction
mediated by the spin fluctuations, frustration can arise in the
sign of the gap on the disconnected Fermi surfaces. This
can result in nodal structures in the gap function, and should generally
work destructively against superconductivity. Conversely, from the present
viewpoint, the optimal situation
for the spin fluctuation mediated superconductivity
is when the $X^2-Y^2$ $\gamma$ Fermi surface is effective and the $\alpha_1$
Fermi surface has $XZ/YZ$ orbital character for {\it all} $k_z$.
As far as the first principles band calculations are concerned,
this situation is realized in limited materials like
NdFeAsO and SmFeAsO, where the
highest $T_c$ among the iron based superconductors is observed experimentally.
From this viewpoint, further band structure calculation studies
may give useful information
for obtaining new related superconductors with higher $T_c$.
\section*{Acknowledgment}
We acknowledge C.H. Lee,
T. Miyake, O.K. Andersen, H. Mukuda, H. Kinouchi, and Y. Kitaoka
for illuminating discussions.
\section*{References}
\begin{thebibliography}{99}
\bibitem{Hosono}
Y.Kamihara, T.Watanabe, M.Hirano, and H. Hosono, J. Am. Chem. Soc.
{\bf 130}, 3296 (2008).
\bibitem{Mazin}I.I. Mazin, D.J. Singh, M.D. Johannes, and M.H. Du,
Phys. Rev. Lett. {\bf 101}, 057003 (2008).
\bibitem{Kuroki1st} K.Kuroki, S.Onari, R.Arita, H.Usui,
Y.Tanaka, H.Kontani, and H.Aoki, Phys. Rev. Lett. {\bf 101}, 087004 (2008).
\bibitem{KA} K. Kuroki and R. Arita, Phys. Rev. B {\bf 64}, 024501 (2001).
\bibitem{KA2} K. Kuroki, T. Kimura and R. Arita,
Phys. Rev. B {\bf 66}, 184508 (2002).
\bibitem{Usui} H. Usui and K. Kuroki, Phys. Rev. B {\bf 84}, 024505
(2011).
\bibitem{pwscf}
S. Baroni,
A. Dal Corso, S. de Gironcoli, P. Giannozzi,
C. Cavazzoni, G. Ballabio, S. Scandolo, G. Chiarotti, P. Focher,
A. Pasquarello, K. Laasonen, A. Trave, R. Car, N. Marzari and
A. Kokalj,http://www.quantum-espresso.org/.
\bibitem{MaxLoc} N. Marzari and D. Vanderbilt, Phys. Rev. B
{\bf 56}, 12847 (1997);
I. Souza, N. Marzari and D. Vanderbilt,
Phys. Rev. B {\bf 65}, 035109 (2001).
The Wannier functions are generated by the code developed by
A. A. Mostofi,
J. R. Yates, N. Marzari, I. Souza and D. Vanderbilt,
(http://www.wannier.org/).
\bibitem{Singh} D.J.Singh and M.H. Du, Phys. Rev. Lett. {\bf 100},
237003 (2008).
\bibitem{Vildosola} V. Vildosola, L. Pourovskii, R. Arita, S. Biermann, and
A. Georges, Phys. Rev. B {\bf 78}, 064518 (2008).
\bibitem{Lebegue} S. Lebegue, Z.P. Yin, and W.E. Pickett, New J. Phys.
{\bf 11}, 025004 (2009).
\bibitem{OKAndersen} O.K. Andersen and L. Boeri, Annalen der Physik, {\bf 1}
8 (2011).
\bibitem{Shirage} P.M. Shirage, K. Kihou, C.H. Lee, H. Kito, H. Eisaki,
and A. Iyo, Applied Physics Letters {\bf 97} 172506 (2010).
\bibitem{Miyake} T. Miyake, T. Kosugi, S. Ishibashi, and K. Terakura,
J. Phys. Soc. Jpn. {\bf 79}, 123713 (2010).
\bibitem{Kosugi} T. Kosugi, T. Miyake, S. Ishibashi, and K. Terakura,
J. Phys. Soc. Jpn. {\bf 81}, 014701 (2012).
\bibitem{KKprb} K. Kuroki, H. Usui, S. Onari, R. Arita,
and H. Aoki, Phys. Rev. B {\bf 79}, 224511 (2009).
\bibitem{DHLee} F. Wang, H. Zhai, and D.-H. Lee,
Phys. Rev. B {\bf 81}, 184512 (2010).
\bibitem{Thomale} R. Thomale, C. Platt, W. Hanke, and B.A. Bernevig,
Phys. Rev. Lett. {\bf 106}, 187003 (2011).
\bibitem{Graser} S. Graser, T. A. Maier, P. J. Hirschfeld, and
D. J. Scalapino, New J. Phys. {\bf 11}, 025016 (2009).
\bibitem{Bickers} N.E. Bickers, D.J. Scalapino, and S.R. White,
Phys. Rev. Lett. {\bf 62}, 961 (1989).
\bibitem{Wien2k}
P. Blaha, K. Schwarz, G. K. H. Madsen, D. Kvasnicka, J. Luitz
In WIEN2K, An Augmented Plane Wave + Local Orbitals Program
for Calculating Crystal Properties,
(Karlheinz Schwarz/ Techn. Universit$\ddot{\rm a}$t Wien,
Wien, Austria, 2001).
\bibitem{Lee} C.H. Lee, A.Iyo, H. Eisaki, H. Kito,
M.T. Fernandez-Diaz, T. Ito, K. Kihou, H. Matsuhata,
M. Braden, and K. Yamada, J. Phys. Soc. Jpn.{\bf 77}, 083704 (2008).
\bibitem{Takano} Y. Mizuguchi and Y. Takano, J. Phys. Soc. Jpn. {\bf 79},
102001 (2010) and references therein.
\end{thebibliography}
\end{document}