1: %%%%%%%%%% espcrc2.tex %%%%%%%%%%
2: %
3: % $Id: espcrc2.tex,v 1.2 2004/02/24 11:22:11 spepping Exp $
4: %
5: \documentclass[fleqn,twoside]{article}
6: % \usepackage{espcrc2}
7: % Use the option 'headings' if you want running headings
8: \usepackage[headings]{espcrc2}
9:
10: % identification
11: \readRCS
12: $Id: espcrc2.tex,v 1.2 2004/02/24 11:22:11 spepping Exp $
13: \ProvidesFile{espcrc2.tex}[\filedate \space v\fileversion
14: \space Elsevier 2-column CRC Author Instructions]
15:
16: % change this to the following line for use with LaTeX2.09
17: % \documentstyle[twoside,fleqn,espcrc2]{article}
18:
19: % if you want to include PostScript figures
20: \usepackage{graphicx}
21: % if you have landscape tables
22: \usepackage[figuresright]{rotating}
23:
24: % put your own definitions here:
25: % \newcommand{\cZ}{\cal{Z}}
26: % \newtheorem{def}{Definition}[section]
27: % ...
28: \newcommand{\ttbs}{\char'134}
29: \newcommand{\AmS}{{\protect\the\textfont2
30: A\kern-.1667em\lower.5ex\hbox{M}\kern-.125emS}}
31:
32:
33: \def\prb{Phys. Rev. B}
34: \def\prl{Phys. Rev. Lett.}
35: \def\pla{Phys. Lett. A}
36: \def\pr{Phys. Rev.}
37: \def\MgB2{MgB$_{2}$}
38: \def\cm-1{cm$^{-1}$\,}
39: \def\cmT-1{cm$^{-1}$/T\,}
40: \def\E2g{$E_{2g}$}
41: \def\A1g{$A_{1g}$}
42: \def\2DS{$2\Delta_{S}^{E}$}
43: \def\DL{$2\Delta_{L}^{E}$}
44: \def\2DA{$2\Delta^{A}$}
45: \def\D0{$2\Delta_{0}$}
46:
47: % set the starting page if not 1
48: % \setcounter{page}{17}
49:
50: % declarations for front matter
51: \title{Multi-Gap Superconductivity in \MgB2: Magneto-Raman Spectroscopy}
52:
53: \author{G.~Blumberg\address[BL]{Bell Laboratories, Lucent
54: Technologies, Murray Hill, NJ 07974, USA}\thanks{Corresponding
55: author. E-mail: girsh@bell-labs.com},
56: A. Mialitsin\addressmark\thanks{Department of Physics and Astronomy,
57: Rutgers University, Piscataway, NJ 08854, USA},
58: B. S. Dennis\addressmark,
59: N.~D.~Zhigadlo\address[ETH]{Solid State Physics Laboratory, ETH,
60: CH-8093 Z\"urich, Switzerland}
61: and
62: J.~Karpinski\addressmark}
63:
64: % If you use the option headings,
65: % the title is also used as the running title,
66: % and the authors are also used as the running authors.
67: % You can change that by using \runtitle and \runauthor.
68:
69: \runtitle{Multi-Gap Superconductivity in \MgB2: Magneto-Raman Spectroscopy}
70: \runauthor{G.~Blumberg {\it et. al.}}
71:
72: \begin{document}
73:
74: \begin{abstract}
75: Electronic Raman scattering studies on \MgB2 single crystals as a
76: function of excitation
77: and polarization have revealed three distinct superconducting
78: features: a clean gap below 37~\cm-1 and
79: two coherence peaks at 109~\cm-1 and 78~\cm-1 which we identify as
80: the superconducting gaps in $\pi$- and $\sigma$-bands and as the
81: Leggett's collective mode arising from the
82: fluctuation in the relative phase between two superconducting
83: condensates residing on corresponding bands.
84: The temperature and field dependencies of the superconducting features
85: have been established.
86: A phononic Raman scattering study of the \E2g boron stretching mode
87: anharmonicity
88: and of superconductivity induced self-energy effects is presented.
89: We show that anharmonic two phonon decay is mainly responsible for the
90: unusually large linewidth of the \E2g mode.
91: We observe $\sim 2.5\%$
92: hardening of the \E2g phonon frequency upon cooling into the
93: superconducting state and estimate the electron-phonon coupling
94: strength associated with this renormalization.
95: \vspace{1pc}
96: \end{abstract}
97:
98: \maketitle
99:
100: \section{INTRODUCTION}
101:
102: The multi-gap nature of superconductivity in \MgB2 was predicted
103: theoretically \cite{Liu} and has been
104: experimentally established by a number of spectroscopies.
105: A double-gap structure in the quasi-particle energy spectra was
106: determined from tunneling spectroscopy \cite{Iavarone,Sza30}.
107: The two gaps have been assigned to distinctive quasi-two-dimensional
108: $\sigma$-bonding states of the boron $p_{x,y}$ orbitals and
109: three-dimensional $\pi$-states of the boron $p_z$ orbitals Fermi
110: surface (FS) sheets by means of ARPES~\cite{Tsuda,Souma}:
111: $\Delta_{\sigma} = 5.5 - 6.5$ and $\Delta_{\pi} = 1.5 -
112: 2.2$~meV.
113: Scanning tunneling microscopy (STM) has provided a reliable fit for
114: the smaller gap, $\Delta_{\pi} = 2.2\,{\rm meV}$ \cite{Esklidsen}.
115: This value manifests in the absorption threshold
116: energy at 31~\cm-1 obtained from
117: magneto-optical far-IR studies~\cite{Perucchi}.
118: The nominal upper critical field $H^{\pi}_{c2}$ deduced from the
119: coherence length $\xi_{\pi} = 49.6\,{\rm nm}$ by vortex imaging is
120: $H^{\pi}_{c2} \approx 0.13\,{\rm T}$~\cite{Esklidsen} which is
121: much smaller than the critical field $H^{opt}_{c2} \approx
122: 5\,{\rm T}$ found by magneto-optical measurements~\cite{Perucchi}.
123:
124: Electronic Raman studies on \MgB2 have explored the
125: superconducting (SC) energy gap and changes in phonon lineshapes, starting
126: with the work of \cite{Chen,Goncharov} and followed thereafter by
127: \cite{Quilty,TajimaCax}.
128: The dependence of the Raman response on scattering geometry allowed
129: an observation of the pairing gap on
130: the two-dimensional $\sigma$ bands and the 3D $\pi$ bands.
131: By orienting the light polarizations along the c-axis of \MgB2
132: (perpendicular to the hexagonal planes) the weakly dispersing
133: $\sigma$ bands cannot be probed and thus only the $\pi$ bands
134: are projected out, giving an observed threshold at
135: $2\Delta_{\pi}=29$~\cm-1 \cite{TajimaCax}.
136: The larger $2\Delta_{\sigma}$ gap has been demonstrated by Raman
137: experiments as a SC coherence peak at 105~\cm-1\,\cite{Quilty}.
138:
139: For multi-band superconductors collective modes
140: associated with fluctuations of the relative phase and amplitudes of
141: coupled condensates \cite{Leggett,Griffin,Marel} as well as distinctive
142: self energy effects associated with intra- and inter-band
143: interactions \cite{Liu,Mazin} were expected.
144: It has been suggested from STM vortex imaging that the
145: superconductivity in the $\pi$-band is induced by
146: superconductivity in the $\sigma$-band \cite{Esklidsen}, however, the
147: coupling mechanism remained unclear.
148: Previous phononic Raman spectroscopy has identified a broad
149: $\Gamma$-point phonon centered around $620-640$~\cm-1
150: \cite{Chen,Goncharov,Quilty} consistent with the calculated frequency
151: of the anharmonic \E2g boron stretching
152: mode~\cite{Mazin,Yildirim}.
153: The phononic dispersion has been studied by
154: inelastic x-ray scattering \cite{Shukla,Baron}.
155: However, the expected self energy effects \cite{Liu,Mazin} have not
156: been demonstrated.
157:
158: \subsection{Experimental}
159:
160: Polarized Raman scattering can probe excitations around the
161: Brillouin zone (BZ) center that belong to different symmetry
162: representations within the space group of the crystal structure.
163: The point group associated with ${\rm MgB_2}$ is $D_{6h}$.
164: We denote by
165: $(\textbf{e}_{in} \textbf{e}_{out})$ a configuration in which the
166: incoming/outgoing photons are polarized along the
167: $\textbf{e}_{in}$/$\textbf{e}_{out}$ directions. The vertical
168: ($V$) or horizontal ($H$) directions were chosen
169: perpendicular or parallel to the crystallographic $a$-axis. The
170: ''right-right'' ($RR$) and ''right-left'' ($RL$)
171: notations refer to circular polarizations:
172: $\textbf{e}_{in} = (H - i V) / \sqrt{2}$, with
173: $\textbf{e}_{out} = \textbf{e}_{in}$ for the $RR$ and
174: $\textbf{e}_{out} = \textbf{e}_{in}^{*}$ for the $RL$ geometry.
175: For the $D_{6h}$ point group the $RR$ and $H\!H$
176: polarizations select correspondingly $A_{1g}$ and $A_{1g}$ +
177: $E_{2g}$ symmetries while both $RL$ and $V\!H$ select
178: the $E_{2g}$ representation.
179:
180: Raman scattering was performed in back scattering geometry from the
181: $ab$ surface of \MgB2 single crystals grown as described in
182: \cite{Karpinski} using less than 2\,mW of incident power focused to
183: a $100 \times 200\, \mu$m spot.
184: The data in magnetic field was acquired with a continuous flow
185: cryostat inserted into the horizontal bore of a SC magnet.
186: The sample temperatures quoted have been corrected for laser heating.
187: We used the excitation lines of a Kr$^{+}$ laser and a
188: triple-grating spectrometer for analysis of the scattered light.
189: The data were corrected for the spectral response of the
190: spectrometer and the CCD detector and for the optical properties of
191: the material at different wavelengths as described in Ref.
192: \cite{Blumberg94}.
193:
194: \subsection{Raman response}
195:
196: \begin{figure}[t]
197: \includegraphics[width=1\columnwidth]{Fig1.eps}
198: \caption{
199: The Raman response spectra of an \MgB2 crystal in
200: the normal (red) and SC (blue) states for the \E2g (top) and
201: $A_{1g}$ (bottom) scattering channels.
202: The data is acquired at 8\,K.
203: The normal state has been achieved by applying a 5\,T
204: magnetic field parallel to the $c$-axis.
205: Decomposition into SC coherence peaks, \E2g phonon and the fits are
206: shown by solid lines.
207: }
208: \label{fig:1}
209: \end{figure}
210: In Fig.\,\ref{fig:1} we show the Raman response from an \MgB2
211: single crystal for the \E2g and \A1g scattering channels
212: in the normal and SC states.
213: The $E_{2g}$ scattering channel is accessed by $RL$ scattering
214: polarization geometries and the $A_{1g}$
215: channel by $RR$ geometry.
216:
217: The response comprises electronic and phononic contributions.
218: The electronic Raman response at low frequencies in the SC state is
219: decomposed into a sum (solid lines) of a
220: gapped normal state continuum with temperature
221: broadened $2\Delta_{0}= 37$\,\cm-1 gap cutoff (threshold at 33\,\cm-1),
222: the SC coherence peak at $2\Delta_{l} = 109$\,\cm-1 (black solid line),
223: and a novel collective mode at $\omega_{L} = 76$\,\cm-1 (green solid line).
224: The latter is present only in the $A_{1g}$ scattering channel.
225: To fit the observed shapes the theoretical coherence peak singularity
226: $\chi^{\prime\prime} \sim 4 \Delta_{l}^{2}/(\omega \sqrt{\omega^{2} -
227: 4 \Delta_{l}^{2}})$ is broadened by convolution with a Lorentzian with
228: HWHM = 12\% of $2\Delta_{l}$ for the \E2g channel and 20\% for
229: the \A1g channel.
230: The collective mode $\omega_{L}$ is broadened to HWHM = 18\,\cm-1.
231:
232: For the high energy part of the spectra the broad \E2g band \emph{I}
233: centered at
234: about $630-640$\,\cm-1 corresponds to the boron stretching mode which is
235: the only Raman active phonon for the \MgB2 structure.
236: It is also the only phononic mode demonstrating renormalization
237: below the SC transition \cite{Mialitsin}.
238: All the other high frequency modes (\emph{II-VI}) in the \A1g and
239: \E2g channels correspond to twice the
240: energy of distinctive flat portions in the phonon dispersions
241: measured by inelastic x-ray scattering \cite{Shukla,Baron} and we
242: assign them to two phonon scattering.
243: The `E$_{1u}$-branch' and a two-fold degenerate low energy acoustic
244: phonon branch have a coinciding minimum in the A-point of the BZ
245: thus delivering a large Raman response for the two-phonon peak \emph{II}.
246: Peak \emph{III} is due to flat portions of low energy acoustic phonon
247: branches when they approach the M-point.
248: Peak \emph{IV} is at twice the frequency of a distinctive
249: saddle point of a high energy acoustic phonon branch in A-point.
250: The `A$_{2u}$ branch' is mostly flat all the way along the $\Gamma$-A
251: line at around $400$\,\cm-1.
252: This might explain the peculiar
253: symmetry indifferent behavior of peak \emph{V}.
254: Finally the \E2g optical branch has a minimum in the A-point at about
255: 530\,\cm-1 resulting in the two-phonon scattering peak \emph{VI}.
256:
257: \subsection{Resonant Raman excitation profile}
258: \begin{figure}[t]
259: \includegraphics[width=1\columnwidth]{Fig2.eps}
260: \caption{
261: Raman response function at 8\,K in the SC (blue) and
262: normal (red) states for the \E2g ($RL$)
263: and the \A1g ($RR$ polarization) scattering channels as a function of
264: excitation wavelengths.
265: }
266: \label{fig:2}
267: \end{figure}
268: Light can couple to electronic and phononic excitations \emph{via}
269: resonant or non-resonant Raman processes \cite{DevereauxRMP}.
270: The Raman scattering cross-section can be substantially
271: enhanced when the incident
272: photon energy is tuned into resonance with optical interband
273: transitions.
274: The resonance Raman excitation profile (RREP) provides information
275: about the scattering probabilities seen in the Raman spectra.
276: For \MgB2 the interband contribution to the in-plane optical conductivity
277: $\sigma_{ab}(\omega)$ contains strong IR peaks with a tail
278: extending to the red part of the visible range and
279: a pronounced band around 2.6~eV \cite{Kuz'menko,Guritanu}
280: (Fig.\,\ref{fig:3}).
281: The IR peaks are associated with transitions between two
282: $\sigma$-bands while the 2.6\,eV peak is associated with the $\pi
283: \rightarrow \sigma$ electronic transitions in the vicinity of the
284: $\Gamma$ point and $\sigma \rightarrow \pi$ transitions in the
285: vicinity of the $M$ point of the BZ
286: \cite{Mazin,Antropov,Kortus}.
287:
288: \begin{figure}[t]
289: \includegraphics[width=0.95\columnwidth]{Fig3.eps}
290: \caption{
291: Comparison of $ab$-plane optical conductivity
292: (Ref.\cite{Guritanu}) and resonant Raman excitation profiles for
293: phononic and electronic excitations.
294: The empty symbols show the 340~\cm-1 $A_{1g}$ and the 640~\cm-1 $E_{2g}$
295: phonon intensities and the solid symbols show the SC coherence peaks
296: intensities.
297: All dashed lines are guides for the eyes.
298: }
299: \label{fig:3}
300: \end{figure}
301: To explore the resonance conditions we analyze Raman spectra as a
302: function of excitation energy.
303: In Fig.\,\ref{fig:2} we show Raman spectra in the SC and normal
304: states for the $E_{2g}$ and
305: $A_{1g}$ scattering channels for four excitation energies.
306: The normal state has been achieved by applying a 5\,T
307: magnetic field parallel to the $c$-axis.
308: All spectra show a relatively strong electronic Raman continuum that
309: extends beyond our measurement range.
310: The electronic scattering intensity in the fully symmetric $A_{1g}$
311: channel is not much weaker than in the $E_{2g}$ channel indicating
312: cancellation of screening that
313: could be due to multi-band contributions with opposite sign of the
314: effective mass near the FS \cite{Cardona}.
315:
316: \section{ELECTRONIC RAMAN RESPONSE}
317:
318: The low frequency part of the
319: electronic Raman continuum changes in the SC state
320: (Figs.\,\ref{fig:1}-\ref{fig:2}), reflecting renormalization of electronic
321: excitations resulting in three new features in the spectra:
322: (i) a threshold of Raman intensity at $2\Delta_{0}= 37$\,\cm-1,
323: (ii) a SC coherence peak at $2\Delta_{l} = 109$\,\cm-1, and
324: (iii) a new mode at 76\,\cm-1, which is in-between the $2\Delta_{0}$ and
325: $2\Delta_{l}$ energies.
326: The observed energy scales of the fundamental gap $\Delta_{0}$
327: and the large gap $\Delta_{l}$ are consistent with
328: $\Delta_{\pi}$ and $\Delta_{\sigma}$ as assigned by one-electron
329: spectroscopies \cite{Tsuda,Souma,Esklidsen}.
330: The features (i-ii) are seen in all scattering geometries while
331: mode (iii) contributes only to the $A_{1g}$ scattering channel.
332:
333: The Raman coupling to the $2\Delta_{l}$ electronic coherence peak in
334: the SC state is provided by density fluctuations in
335: the $\sigma$-band.
336: For the \E2g channel the peak intensity is enhanced by about an order
337: of magnitude when
338: the excitation photon energy is in resonance with the 2.6~eV
339: $\sigma \rightarrow \pi$ inter-band transitions (Fig.\,\ref{fig:3}).
340: In contrast, for the fully symmetric \A1g channel the integrated
341: intensity of the $2\Delta_{l}$ coherence peak does not follow the optical
342: conductivity and is about five times weaker than for the resonant
343: excitation in the \E2g channel.
344: Nonetheless, due to relative charge density fluctuations between two
345: coupled $\sigma$- and $\pi$-bands the intensity in the fully
346: symmetric channel is only partially screened.
347: The integrated intensity of the $\omega_{L}$ collective mode in the
348: \A1g channel shows excitation dependence similar to one
349: for the $2\Delta_{l}$ coherence peak in the same channel.
350:
351: \subsection{The fundamental gap}
352: At the fundamental gap value $2\Delta_{0}$ the spectra for all
353: symmetry channels show a threshold without a coherence peak.
354: This threshold appears cleanest for the spectra with lower energy
355: excitations for which the low-frequency contribution of
356: multi-phonon scattering from acoustic branches is suppressed (see
357: Fig.\,\ref{fig:2}).
358: The absence of the coherence peak above the threshold is consistent
359: with the expected
360: behavior for a dirty superconductor \cite{DevereauxRMP}.
361: Thus the $\pi$-bands show signatures of strong intrinsic
362: scattering leading to the observed Raman continuum.
363:
364: The ratio $2\Delta_{0}/k_{B}T_{c}$ is only 1.2 which makes the
365: $\pi$-band contribution to the two band superconductivity quite
366: tenuous.
367: That is in agreement with rapid suppression of the threshold
368: by a relatively weak magnetic field applied along the $c$-axis.
369:
370: \subsection{Large gap in $\sigma$ bands}
371: The $2\Delta_{l}$ coherence peak is seen for all scattering
372: geometries.
373: For the \E2g channel it appears as a sharp singularity with
374: continuum renormalization extending to high energies, which is in
375: agreement with expected behavior for clean
376: superconductors \cite{DevereauxRMP}.
377: The $2\Delta_{l}$ coherence peak frequency shows a BCS-like temperature
378: dependence with the $2\Delta^{E}/k_{B}T_{c}$ ratio of about 4
379: indicating a moderately strong coupling limit.
380:
381: Coulomb screening suppresses the scattering intensity
382: for the fully symmetric \A1g channel.
383: The $2\Delta_{l}$ coherence peak intensity does not follow the
384: optical conductivity.
385: The Raman intensity in the fully
386: symmetric channel is governed by the difference in light coupling to
387: the the $\pi$- and $\sigma$-bands which explains the intensity
388: enhancement seen for the pre-resonant excitations (Fig.\,\ref{fig:3}).
389: Also, the $2\Delta_{l}$ coherence peak in the \A1g channel
390: is broader than in the \E2g channel due to stronger cross
391: relaxational coupling to the $\pi$-band quasiparticles.
392:
393: \subsection{Leggett's collective mode}
394: \begin{figure}[t]
395: \includegraphics[width=1\columnwidth]{Fig4.eps}
396: \caption{
397: Evolution of low-frequency Raman response as a
398: function of temperature at zero field and field at 8~K.
399: The \E2g scattering channel $(RL)$ with
400: 482.5~nm excitation and the
401: \A1g channel $(RR)$ with 752.5~nm excitation are shown.
402: }
403: \label{fig:4}
404: \end{figure}
405: The novel mode at 76\,\cm-1 contributes only to the \A1g
406: scattering channel.
407: This mode is more pronounced for off-resonance excitation for which
408: the electronic continuum above the fundamental threshold
409: $2\Delta_{0}$ is weaker.
410: We attribute this feature to the collective mode proposed by
411: Leggett \cite{Leggett}:
412: If a system contains two coupled superfluid liquids a simultaneous
413: cross-tunneling of a pair of electrons becomes possible.
414: Leggett's collective mode is caused by dynamic oscillations of
415: Cooper pairs between the two superfluids leading to
416: small fluctuations of the relative phase of two condensates while the
417: total electron density at every spatial point of the superconductor
418: is conserved.
419: Such excitation couples to the \A1g Raman scattering channel.
420: If the energy of this mode is below the pair-breaking gap, the mode
421: dissipation is suppressed and the excitation is expected to be long-lived.
422: In the case of \MgB2 the two coupled SC condensates reside at the
423: $\sigma$- and $\pi$-bands.
424:
425: The excitation of Leggett's mode is gapped with a dispersion
426: relation for small momentum $q$ given by \cite{Leggett,Griffin,Sharapov}
427: \begin{equation}
428: \Omega_{L}(q)^{2} = \omega_{L}^{2} + v^{2}q^{2},
429: \label{dispersion}
430: \end{equation}
431: where in the low frequency limit the excitation gap can be expressed
432: \emph{via} intra- and
433: inter-band pairing potentials $V_{\sigma\sigma}$, $V_{\pi\pi}$ and
434: $V_{\sigma\pi}$,
435: the gaps $\Delta_{\sigma}$ and $\Delta_{\pi}$ and the density of
436: states $N_{\sigma}$ and $N_{\pi}$ in corresponding bands
437: \begin{equation}
438: \omega_{L}^{2} = \frac{N_{\sigma} + N_{\pi}}{N_{\sigma}
439: N_{\pi}} \frac{4 V_{\sigma\pi} \Delta_{\sigma}
440: \Delta_{\pi}}{V_{\sigma\sigma} V_{\pi\pi} - V_{\sigma\pi}^{2}}.
441: \label{LeggettMode}
442: \end{equation}
443: Leggett's mode exists only if
444: $V_{\sigma\sigma} V_{\pi\pi} > V_{\sigma\pi}^{2}$.
445: The estimates of the coupling constants by first principal
446: computations \cite{Liu,Mazin,Choi} show that for the \MgB2
447: superconductor this condition is satisfied and the estimate for the
448: mode energy is in between 60 - 85\,\cm-1 which is consistent with the
449: observed mode at 76\,\cm-1.
450: Because the collective mode energy is in between the two-particle
451: excitation thresholds
452: for the $\pi$- and $\sigma$-bands, $2\Delta_{\pi} < \omega_{L} <
453: 2\Delta_{\sigma}$, Leggett's excitation rapidly relaxes into
454: $\pi$-band quasiparticles.
455: Indeed, the measured $Q$-factor for this mode is about two:
456: the Cooper pair tunneling energy relaxes
457: into $\pi$-band quasiparticle continuum within a couple of tunneling
458: oscillations.
459: Despite being short lived, Leggett's mode in \MgB2 couples
460: to light and is observed by Raman spectroscopy.
461:
462: \subsection{Effects of temperature and field}
463: In Fig.\,\ref{fig:4} the evolution of the $2\Delta_{l}$ coherence
464: peak and Leggett's collective mode $\omega_{L}$
465: across the SC transition is
466: displayed for two cases: varying temperature at zero magnetic
467: field (\textit{a},\,\textit{c}) and varying magnetic field at
468: 8~K (\textit{b},\,\textit{d}).
469: The coherence peaks lose their intensity and move to lower energies
470: by either increasing temperature or field.
471: The intensity threshold $2\Delta_{0}$ is
472: already smeared out at magnetic fields as weak as 0.2\,T,
473: consistent with $H^{\pi}_{c2}$ deducted from vortex imaging
474: \cite{Esklidsen}.
475: Leggett's collective mode $\omega_{L}$ persists up
476: to 0.6\,T while the SC coherence peak $2\Delta_{l}$ is suppressed beyond 2\,T.
477: $2\Delta_{l}(T, H)$ is shown in the insets of
478: Fig.\,\ref{fig:5}.
479: It exhibits a BCS-like temperature dependence and a linear reduction
480: in field with a rapid slope of about -15~\cmT-1.
481: A linear extrapolation for the $2\Delta_{l}$ gap collapse leads to
482: 7\,T, a field that is higher than $H^{opt}_{c2}$ \cite{Perucchi},
483: while the coherence peak intensity survives only up to 2\,T.
484:
485:
486: \section{PHONONIC RAMAN RESPONSE}
487:
488: High-$T_c$ superconductivity in \MgB2 is known to be promoted
489: mainly due to the boron layers \cite{Kortus},
490: thus the high frequency
491: lattice vibrations of light boron atoms beneficially increase the
492: electron-phonon coupling.
493: The \E2g Raman active in-plane boron vibrational mode contributes
494: significantly to
495: superconductivity; this fact is reflected by the Eliashberg
496: function $\alpha^2\,F(\omega)$ peaking in the
497: same frequency range where a high phononic density of states is
498: accounted for by Van Hove singularities of the \E2g branch in
499: the $\Gamma$ and $A$ points of the BZ
500: \cite{Yildirim,DagheroPhC}.
501: The reason the \E2g mode plays a prominent role in the SC
502: mechanism is that the mode strongly couples to the
503: $\sigma$-type states of the boron plane as can be seen from the
504: basic geometry of the electronic configuration~\cite{Choi}.
505:
506: \begin{figure}[t]
507: \includegraphics[width=0.98\columnwidth]{Fig5.eps}
508: \caption{
509: Evolution of the \E2g phonon with temperature (a) and field (b)
510: using 482.5\,nm excitation and ($RL$) polarization.
511: The phonon frequency $\omega(T, H)$ and the damping constant
512: $\Gamma(T, H)$ are drawn as functions of temperature (c) and field
513: (d) for two crystal ${\mathcal A}$ (black) and ${\mathcal B}$ (red
514: symbols).
515: Insets show temperature and field
516: dependencies of the $2\Delta_{l}$ energy.
517: }
518: \label{fig:5}
519: \end{figure}
520: \begin{table*}[t]
521: \caption{
522: Comparison of $T_{c}$ and the \E2g oscillator parameters for
523: crystals ${\mathcal A}$ and ${\mathcal B}$.}
524: \begin{tabular}{cccccccc}
525: \hline
526: & $T_{c}$
527: & $\omega_0^{N}$ & $\omega_0^{SC}$ & $\Gamma_{0}$ &
528: $\Gamma_{3}$ & $\Gamma_{4}$ & $\kappa$ \\
529: Crystal & (K) & (\cm-1) & (\cm-1) & (\cm-1) & (\cm-1) & (\cm-1) & (\%) \\
530: \hline\hline ${\mathcal A}$ & 308 & 640 & 659 &
531: $32\!\pm\!12$ &
532: $253\!\pm\!10$ & small & $23\!\pm\!4$\\
533: ${\mathcal B}$ & 311 & 630 & 649 & small &
534: $185\!\pm\!6$ &
535: $23\!\pm\!3$ & $19\!\pm\!2$\\
536: \hline
537: \end{tabular}
538: \label{tab:AnhDecay}
539: \end{table*}
540: Raman spectra exhibit an unusually broad linewidth of the \E2g boron
541: stretching mode \cite{Goncharov,Quilty,Mialitsin,RenkerJLTP} which
542: has been the subject of numerous speculations.
543: While high impurity scattering in earlier low
544: quality samples has been suggested as one of the possible reasons,
545: this mechanism can be readily excluded with recent high quality
546: single crystals.
547: The two remaining contributions to the \E2g phonon rapid decay are
548: (i) strong electron-phonon coupling and
549: (ii) multiphononic decay (subsequently referred to as
550: \emph{anharmonicity}).
551: The relative importance of the electron-phonon coupling and
552: anharmonicity in this matter is still under debate.
553: On one hand a density functional theory
554: calculation asserts that the anharmonic contribution to
555: the \E2g phonon linewidth is negligible ($\sim 10$\,\cm-1) \cite{Shukla}.
556: On the other hand analysis of the phonon self-energy in the long
557: wavelength limit shows that the $\sigma$-band contribution to the
558: phonon decay is vanishing \cite{Calandra}.
559: Thus, even when
560: contributions of the spectral weight of
561: $\alpha^2\,F(\omega)|_{\omega < \omega_{E_{2g}}}$
562: to the damping of the \E2g phonon are accounted for \cite{Cappelluti},
563: the experimentally observed linewidth of
564: $200-280$\,\cm-1 at low temperatures \cite{Quilty,Mialitsin,RenkerJLTP} cannot
565: be explained with electron-phonon coupling alone whose part in
566: the \E2g mode linewidth at low temperatures amounts to about
567: 50\,\cm-1 even in such an elaborate scenario as that in Ref.
568: \cite{Cappelluti}.
569: %
570:
571: Raman scattering experiments have shown that the frequency of the \E2g mode
572: in single crystals at room temperature is around
573: 635\,\cm-1 \cite{Quilty,RenkerJLTP,Mialitsin} whereas theoretical
574: calculations systematically underestimate this value by about
575: 80\,\cm-1 \cite{Yildirim,Shukla}.
576: It has been suggested that if the
577: \E2g band around the $\Gamma$-point is anharmonic then the \E2g mode
578: frequency is increased by the missing amount to match the
579: experimentally observed value \cite{Liu,Boeri}.
580: In addition, the experimentally observed $T_c$ and the reduced isotope
581: effect \cite{HinksPhC} can only be reconciled within anisotropic
582: strong coupling theory if the \E2g mode anharmonicity is explicitly
583: included \cite{Choi,ChoiPhysRevB}.
584: % \begin{figure}[t]
585: % \includegraphics[width=0.98\columnwidth]{Fig5a.eps}
586: % \caption{
587: % The phonon dispersion for the charachteristic directions of the BZ is
588: % shown with red solid lines as calculated from first principles
589: % [Shukla et al.].
590: % Small circles (black) denote single points of the dispersion relation
591: % as measured by inelastic x-ray scattreing [Shukla et al.].
592: % The bigger blue circle at Gamma shows the energy of the E2g phonon
593: % mode as measured by Raman [Quilty, Bohnen, self reference].
594: % This point is not accessed by inelastic x-ray or neutron scattering.
595: % }
596: % \label{fig:5a}
597: % \end{figure}
598:
599:
600: \subsection{Excitation dependence}
601: The Raman intensities for phononic modes are in resonance with
602: the 2.6~eV optical transitions.
603: The resonance is more distinct for the \E2g phonon mode that
604: reduces by an order of magnitude for adjacent violet and red
605: excitations and almost vanishes in the infra-red
606: (see Fig.\,\ref{fig:3}) inferring that the Raman coupling to this
607: phonon is realized only via $\pi \leftrightarrow \sigma$ interband
608: transitions.
609: In contrast, the two-phonon scattering in the \A1g channel remains
610: visible even for pre-resonance excitations.
611:
612: \subsection{Dependence on temperature and field}
613: In Fig.\,\ref{fig:5}\,(\textit{a-b}) we show the temperature
614: dependence of the \E2g Raman response measured on cooling in zero
615: field and as a function of field at 8\,K.
616: The data (dots) are fitted with two phononic oscillators and a SC
617: coherence peak (solid lines) on an electronic continuum
618: (decompositions for the lowest spectra are shown).
619: In Fig.\,\ref{fig:5}\,(\textit{c-d}) we evaluate the temperature and
620: field dependencies of the \E2g phonon frequency $\omega(T, H)$ and
621: the damping constant $\Gamma(T, H)$ for two crystals ${\mathcal A}$
622: and ${\mathcal B}$ where we distinguish between the respective values
623: for the SC and normal states measured at zero field cooling (solid
624: symbols) and 8\,T cooling (empty symbols).
625: The solid line in Fig.\,3\,(\emph{c}) is a fit of the
626: damping constant $\Gamma(T)$ in the normal state to a model of
627: anharmonic two and three phonon decay at one-half and one-third
628: frequencies:
629: \begin{eqnarray}
630: \Gamma(T) = & \Gamma_{0} +
631: \Gamma_{3}[1 + 2n(\Omega(T)/2)] + \nonumber \\
632: & \Gamma_{4}[1 +
633: 3n(\Omega(T)/3)+3n^2(\Omega(T)/3)].
634: \label{damping}
635: \end{eqnarray}
636: %
637: %
638: Here $\Omega(T)={h\,c\, \omega_{h}}/{k_{B}T}$, with the harmonic
639: frequency $\omega_{h} = 540$\,\cm-1
640: \cite{Mazin,Shukla,Kortus}, $n(x)$ is
641: the Bose-Einstein distribution function, $\Gamma_{0}$ is the
642: internal temperature independent linewidth of the phonon, and
643: $\Gamma_{3,4}$ are broadening coefficients due to the cubic and
644: quartic anharmonicity. The results of the fit to this anharmonic
645: decay model are collected in Table \ref{tab:AnhDecay}.
646: For both
647: crystals the broadening coefficients $\Gamma_{3} + \Gamma_{4} \gg
648: \Gamma_{0}$ and therefore the anharmonic decay is primarily
649: responsible for the large damping constant of the \E2g phonon.
650: We identify the reason for this rapid phononic decay in the phononic
651: density of states (PDOS) peaking at 265\,\cm-1, half of the harmonic
652: \E2g phonon frequency $\omega_h$ (Refs. \cite{Yildirim,Osborn}),
653: which corresponds to the Van-Hove
654: singularity of the lower acoustic branch,
655: almost dispersionless along the $\Gamma-{\rm K}-{\rm M}$ direction.
656: In this context the narrowing of the \E2g mode
657: with Al substitution observed in Refs.~\cite{RenkerJLTP,Bohnen} can
658: be readily explained with the \E2g phonon branch moving to energies
659: above 100\,meV with increased Al concentration
660: whereas the acoustic modes that provide the decay channels stay close
661: to their original energies with high PDOS in the energy range of
662: $200-320$\,\cm-1 \cite{Bohnen}.
663: In short, the fast decay of the \E2g mode
664: is due to the unique combination of its harmonic frequency
665: in the $\Gamma$ point corresponding to high PDOS at half of
666: this frequency.
667: The residual
668: linewidth $\Gamma_0$ that we obtain from the fit to the
669: anharmonic decay model, while small, is not in contradiction with the
670: theoretical estimates \cite{Calandra,Cappelluti} of the
671: electron-phonon decay contribution to the \E2g phonon linewidth.
672:
673: It is worth noting that individual $\Gamma_i$ parameters differ for
674: the two single crystals despite the fact that both
675: samples were grown in the same batch.
676: The \E2g mode for crystal ${\mathcal A}$ is broader by about 10\,meV
677: than for crystal ${\mathcal B}$.
678: With $\Gamma_0^{\mathcal A}$
679: somewhat higher than $\Gamma_0^{\mathcal B}$ and
680: $\Gamma_3^{\mathcal A}$ substantially higher than
681: $\Gamma_3^{\mathcal B}$ (see Table\,\ref{tab:AnhDecay}) the \E2g
682: mode in crystal ${\mathcal A}$ is more anharmonic than in crystal
683: ${\mathcal B}$. Accordingly the crystal ${\mathcal A}$ mode is
684: pushed to about 10\,\cm-1 higher frequency at low temperatures. We
685: note a correlation between the larger anharmonicity and slightly lower
686: $T_{c}$\,in the case of crystal ${\mathcal A}$.
687:
688: \subsection{Pressure and Al substitution}
689: %
690: \begin{figure}[t]
691: \begin{center}
692: \includegraphics[width=0.64\columnwidth]{Fig6.eps}
693: \end{center}
694: \vspace{-30pt}
695: \caption{
696: The phononic Raman intensity in \E2g channel as function of
697: substitution, pressure and temperature.
698: (a) \MgB2 vs. AlB$_{2}$ at room temperature from Ref.\,\cite{Bohnen};
699: (b) The pressure dependence from Ref.\,\cite{Goncharov};
700: (c) The temperature dependence at ambient pressure.
701: Spectra are shifted vertically for clarity.
702: }
703: \label{fig:6}
704: \end{figure}
705: The boron stretching \E2g phonon has been found to be the most
706: sensitive mode to structural changes upon substitution of Mg sites with Al.
707: The Raman spectra of gradual substitution Al$_{x}$Mg$_{1-x}$B$_{2}$
708: are quite complicated with nonuniform transfer of spectral weight from the
709: 640\,\cm-1 mode as observed in pure \MgB2 to the 980\,\cm-1 AlB$_{2}$
710: \E2g mode \cite{RenkerJLTP}.
711: Upon complete substitution the change in shape and frequency is
712: striking as the \E2g mode in AlB$_{2}$ has stiffened
713: more than 300\,\cm-1 and its line width has narrowed from 400\,\cm-1
714: to about 50\,\cm-1 (see Fig 6\,(\emph{a})).
715: With Al substitution the large anharmonicity of the
716: \E2g phonon mode is reduced.
717:
718: The pressure dependence of the \E2g phonon frequency links its
719: frequency shift to the variation of the lattice parameters
720: \cite{Goncharov}.
721: In the pressure range up to 15\,GPa the mode frequency shifts linearly with
722: pressure (Fig.\,6\,(\emph{b})).
723: An unusually large Gr\"ueneisen parameter of 3.9 has been deduced from
724: this frequency shift \cite{GoncharovPhysicaC}.
725: The larger Gr\"ueneisen parameters are usually related to increased
726: anharmonicity of the mode \cite{Zallen} which fits in the overall
727: picture of anharmonicity as discussed above.
728: Also interesting is the link to the linear decrease of $T_{c}$ with
729: pressure which in the above mentioned pressure range
730: suppresses $T_{c}$ \cite{Tomita,Loa}
731:
732: The temperature dependence between room temperature and $T_{c}$ is a
733: smooth but nonmonotonic function peaking at 200 K (fit results to
734: the spectra are shown in Fig.5\,(\emph{c})).
735: Its particular functional shape reflects the variation of
736: anharmonicity of the \E2g as a function temperature as both the \E2g
737: band and the corresponding decay channels react to lattice expansion
738: with increased temperature.
739:
740: \subsection{Phononic self-energy effects}
741: To describe the superconductivity induced self-energy effect we
742: refer to Fig.\,\ref{fig:5}\,(\emph{c}).
743: Upon cooling in zero field the \E2g phonon frequency exhibits
744: nonmonotonic but smooth behavior down to $T_c$.
745: Then at $T_c$ it displays
746: abrupt hardening with $\omega_0^{SC}(T)$ scaling to the functional
747: form of the SC gap magnitude $2\Delta_{l}(T)$.
748: For in-field cooling the \E2g phonon frequency $\omega_0^N(T)$ remains
749: unrenormalized.
750: The differences between the phonon frequencies in
751: the normal and SC states at 8\,K are
752: $18 \pm 3$ and $15 \pm 1.6$\,\cm-1 for crystals ${\mathcal A}$
753: and ${\mathcal B}$ respectively.
754: To quantify the relative hardening of the \E2g mode
755: we obtain the superconductivity induced renormalization
756: constant $\kappa = (\omega_0^{SC}/{\omega_0^{N}}) - 1 \approx 2.5\%$
757: (see Table\,\ref{tab:AnhDecay}) which
758: is much smaller than the theoretically predicted
759: $\kappa \approx 12$\% \cite{Liu}.
760:
761: We estimate the electron-phonon coupling constant
762: $\lambda^{\Gamma}_{E_{2g}}$ around the BZ center using
763: approximations adopted in Refs.~\cite{Zeyher,Rodriguez}: $\lambda
764: = -\kappa \, {\mathcal Re}\,(\frac{\sin u}{u})$, where $u \equiv
765: \pi + 2 i \cosh^{-1}({\omega^{N}}/{2\Delta_{\sigma}})$, and obtain
766: $\lambda^{\Gamma}_{E_{2g}} \approx 0.3$.
767: This estimate of the coupling constant is consistent with the fit to
768: a phenomenological model \cite{Zeyher2} where the direct coupling of
769: light to the $\sigma$ bands is neglected but it is
770: smaller than the values predicted by the first principal
771: computations \cite{Liu,Mazin,Kortus,Choi,Golubov}.
772:
773: We note that all the other modes contributing to two-phonon
774: scattering do not exhibit any measurable renormalization upon
775: cooling into the SC state (see Figs.\,\ref{fig:1}-\ref{fig:2}),
776: thereby the 635\,\cm-1 \E2g boron stretching
777: mode is the only phonon that exhibits renormalization below the SC
778: transition.
779:
780: \section*{SUMMARY}
781: We have measured the polarization resolved Raman response as a function of
782: temperature, field and excitation energy for \MgB2 single crystals.
783:
784: The electronic scattering data revealed three superconductivity
785: induced spectroscopic features:
786: a clean threshold below $2\Delta_{0}=37$\,\cm-1 corresponding to the
787: fundamental gap,
788: a coherence peak at $2\Delta_{l}=109$\,\cm-1 corresponding to the gap
789: on the $\sigma$-bands FS,
790: and the Leggett's collective mode at $\omega_{L}=78$\,\cm-1 arising from the
791: fluctuation in the relative phase between two coupled SC
792: condensates residing on two bands.
793: Altogether the electronic Raman spectra show signatures for
794: superconductivity in the clean limit for quasi-two-dimensional
795: $\sigma$-bands and dirty limit for three-dimensional
796: $\pi$-bands.
797: The ratio $2\Delta_{0}/k_{B}T_{c}$ is only 1.2 which makes the
798: $\pi$-band contribution to the two band superconductivity quite
799: tenuous, in agreement with rapid suppression of the threshold
800: frequency by a relatively weak magnetic field.
801: The large gap shows a BCS-like temperature
802: dependence with the $2\Delta_{l}/k_{B}T_{c}$ ratio of about 4
803: indicating a moderately strong coupling limit.
804: The $2\Delta_{l}$
805: gap magnitude is suppressed by an external magnetic field at the
806: rapid rate of -15~\cmT-1.
807:
808: From the
809: temperature dependence of the \E2g boron stretching phonon we
810: conclude that anharmonic decay is primarily responsible for the
811: anomalously large damping constant of this mode.
812: For this phonon we observe a SC induced self-energy effect and
813: estimate the electron-phonon coupling constant.
814:
815: \subsection*{Acknowledgments}
816: The authors thank M.\,V.~Klein, A.\,A. Kuz'menko, D.\,van\,der\,Marel,
817: I.\,I.~Mazin and W.\,E.~Pickett for valuable discussions.
818: AM was supported by the Lucent-Rutgers Fellowship program.
819: NDZ was supported by the Swiss National Science Foundation through
820: NCCR pool MaNEP.
821:
822:
823: \begin{thebibliography}{9}
824:
825: \bibitem{Liu}
826: A. Y. Liu, I. I. Mazin, and J. Kortus, Phys. Rev. Lett. {\bf 87},
827: 087005 (2001).
828:
829: \bibitem{Iavarone}
830: M. Iavarone {\it et. al.},
831: Phys. Rev. Lett. {\bf 89}, 187002 (2002).
832:
833: \bibitem{Sza30}
834: P. Szab\'{o} {\it et. al.},
835: Phys. Rev. Lett. {\bf 87}, 137005 (2002).
836:
837: \bibitem{Tsuda}
838: S. Tsuda {\it et. al.},
839: % and T. Yokoya and T. Kiss and Y. Takano,
840: % K. Togano, H. Kito, H. Ihara, and S. Shin,
841: Phys. Rev. Lett. {\bf 87}, 177006 (2001).
842:
843: \bibitem{Souma}
844: S. Souma {\it et. al.},
845: % Y. Machida, T. Sato, T. Takahashi, H.
846: % Matsui, S.-C. Wang, H. Ding, A. Kaminski, J. C.
847: % Campuzano, S. Sazaki, and K. Kadowaki,
848: Nature {\bf 423}, 65 (2003).
849:
850: \bibitem{Esklidsen}
851: M.R. Eskildsen {\it et. al.},
852: % M. Kugler, S. Tanaka, J. Jun, S.
853: % M. Kazakov, J. Karpinski, and \O. Fischer,
854: Phys. Rev. Lett. {\bf 89}, 187003 (2002).
855:
856: \bibitem{Perucchi}
857: A. Perucchi {\it et. al.},
858: % L. Degiorgi, J. Jun, M. Angst and J. Karpinski,
859: Phys. Rev. Lett. {\bf 89}, 097001 (2002).
860:
861: % \bibitem{Gorsh1}
862: % B. Gorshunov {\it et. al.}, Eur. Phys. J. B {\bf 21}, 159 (2001).
863: %
864: \bibitem{Chen}
865: X. K. Chen {\it et. al.},
866: % M. J. Konstantinovi\'{c}, J. C.
867: % Irwin, D. D. Lawrie, and J. P. Franck,
868: Phys. Rev. Lett. {\bf 87}, 157002(R) (2001).
869:
870: \bibitem{Goncharov}
871: A. Goncharov {\it et. al.},
872: % V. V. Struzhkin, E. Gregoryanz,
873: % J. Hu, R. J. Hemley, and Ho-k. Mao,
874: Phys. Rev. B {\bf 64}, 100509 (2001).
875:
876: \bibitem{Quilty}
877: J.W. Quilty {\it et. al.},
878: % S. Lee, Yamamoto, and S. Tajima,
879: Phys. Rev. Lett. {\bf 88}, 087001 (2002).
880:
881: \bibitem{TajimaCax}
882: J.W. Quilty {\it et. al.},
883: % S. Lee, S. Tajima, and A. Yamanaka,
884: Phys. Rev. Lett. {\bf 90}, 207006 (2003).
885:
886: \bibitem{Leggett}
887: A. J. Leggett,
888: Prog. Theor. Phys. {\bf 36}, 901 (1966).
889:
890: \bibitem{Griffin}
891: W.-C. Wu and A. Griffin,
892: Phys. Rev. Lett. {\bf 74}, 158 (1995).
893:
894: \bibitem{Marel}
895: For a review see: D. van der Marel,
896: Journal of Superconductivity \textbf{17}, 559 (2004).
897:
898: \bibitem{Mazin}
899: I. I. Mazin and V. Antropov, Physica C {\bf 385}, 49 (2003).
900:
901: \bibitem{Yildirim}
902: T. Yildirim {\it et. al.},
903: % O. G\"{u}lseren, J.W. Lynn, P. Kalifah, K. Inumaru, and R.J. Cava,
904: Phys. Rev. Lett. {\bf 87}, 037001 (2001).
905:
906: \bibitem{Shukla}
907: A. Shukla {\it et. al.},
908: Phys. Rev. Lett. {\bf 90}, 095506 (2003).
909:
910: \bibitem{Baron}
911: A.Q.R. Baron {\it et. al.},
912: % H. Uchiyama, Y. Tanaka, K.-P.
913: % Bohnen, S. Tajima, and T. Ishikawa,
914: Phys. Rev. Lett. {\bf 92}, 197004 (2004).
915:
916: \bibitem{Karpinski}
917: J. Karpinski {\it et. al.}, Supercond. Sci. Tech. {\bf 16}, 221
918: (2003).
919:
920: \bibitem{Blumberg94}
921: G. Blumberg {\it et. al.},
922: Phys. Rev. B {\bf 49}, 13 295 (1994).
923:
924: \bibitem{Mialitsin}
925: A. Mialitsin, B.S. Dennis, N.D. Zhigadlo, J. Karpinski, G. Blumberg,
926: % Anharmonicity and self-energy effects of the \E2g phonon in \MgB2.
927: Phys. Rev. B {\bf 86}, 020509(R) (2007).
928:
929: \bibitem{DevereauxRMP}
930: T. P. Devereaux, R. Hackl,
931: % Inelastic light scattering from correlated electrons.
932: Rev. Mod. Phys. {\bf 78}, 175 (2007).
933:
934: \bibitem{Kuz'menko}
935: A. A. Kuz'menko {\it et. al.}, Sol. St. Comm. {\bf 121}, 175
936: (1990).
937:
938: \bibitem{Guritanu}
939: V.~Guritanu {\it et. al.},
940: % , A.~Kuzmenko, D.~van der Marel, S.~Kazakov, N.~Zhigadlo, and J.~Karpinski,
941: Phys. Rev. B \textbf{73}, 104509 (2006).
942:
943: \bibitem{Antropov}
944: V. P. Antropov {\it et. al.}, cond-mat/0107123.
945:
946: \bibitem{Kortus}
947: J. Kortus, {\it et. al.},
948: Phys. Rev. Lett. {\bf 86}, 4656 (2001).
949:
950: \bibitem{Cardona}
951: T. Strohm and M. Cardona,
952: % Electronic Raman scattering in YBa$_{2}$Cu$_{3}$O$_{7}$ and other superconducting cuprates.
953: Phys. Rev. B {\bf 55}, 12725 (1997).
954:
955: \bibitem{Sharapov}
956: S. G. Sharapov, V. P. Gusynin, H. Beck,
957: % Effective action approach to the Leggett's mode in two-band superconductors.
958: Eur. Phys. J. B {\bf 30}, 45 (2002).
959:
960: \bibitem{Choi}
961: H.~Choi {\it et. al.},
962: % , D.~Roundy, S.~Hong, M.~Cohen, and S.~Loule,
963: \newblock Nature \textbf{418}, 758 (2002).
964:
965: \bibitem{DagheroPhC}
966: D.~Daghero {\it et. al.},
967: % , R.~Gonnelli, G.~Ummarino, O.~Dolgov, J.~Kortus,
968: % A.~Golubov, and S.~Shulga,
969: Physica C \textbf{408-410}, 353 (2004).
970:
971: \bibitem{RenkerJLTP}
972: B.~Renker {\it et. al.},
973: % , H.~Schober, P.~Adelmann, P.~Bohnen, D.~Ernst, R.~Heid,
974: % P.~Schweiss, and T.~Wolf,
975: J. Low Temp. Phys. \textbf{131}, 411 (2003).
976:
977: \bibitem{Calandra}
978: M.~Calandra and F.~Mauri,
979: Phys. Rev. B \textbf{71}, 064501 (2005).
980:
981: \bibitem{Cappelluti}
982: E.~Cappelluti,
983: Phys. Rev. B \textbf{73}, 140505 (2006).
984:
985: \bibitem{Boeri}
986: L.~Boeri, E.~Cappelluti, and L.~Pietronero,
987: Phys. Rev. B \textbf{71}, 012501, (2005).
988:
989: \bibitem{HinksPhC}
990: D.~Hinks and J.~Jorgensen,
991: Physica C \textbf{385}, 98, (2003).
992:
993: \bibitem{ChoiPhysRevB}
994: H.~Choi {\it et. al.},
995: % , D.~Roundy, H.~Marvin, L.~Cohen, and G.~Steven,
996: Phys. Rev. B, \textbf{66}, 020513 (2002).
997:
998: \bibitem{Osborn}
999: R. Osborn {\it et. al.},
1000: % , E. Goremychkin, A. Kolesnikov, and D.~Hinks,
1001: Phys. Rev. Lett. \textbf{87}, 017005 (2001).
1002:
1003: \bibitem{Bohnen}
1004: K.~Bohnen, R.~Heid, and B.~Renker,
1005: Phys. Rev. Lett. \textbf{86} 5771 (2001).
1006:
1007: \bibitem{GoncharovPhysicaC}
1008: A. F. Goncharov and V. V. Struzhkin,
1009: Physica C \textbf{385}, 117 (2003).
1010:
1011: \bibitem{Zallen}
1012: R. Zallen,
1013: Phys. Rev. B \textbf{9}, 4485 (1974).
1014:
1015: \bibitem{Tomita}
1016: T. Tomita {\it et. al.},
1017: Phys. Rev. B \textbf{64}, 092505 (2001).
1018:
1019: \bibitem{Loa}
1020: I. Loa and K. Syassen,
1021: Physica C \textbf{118}, 279, (2001).
1022:
1023: \bibitem{Zeyher}
1024: R. Zeyher and G. Zwicknagl, Z. Phys. B {\bf 78}, 479 (2002).
1025:
1026: \bibitem{Rodriguez}
1027: C. R. Rodriguez {\it et. al.}, Phys. Rev. B {\bf 42}, R2692 (1990).
1028:
1029: \bibitem{Zeyher2}
1030: R. Zeyher,
1031: Phys. Rev. Lett. {\bf 90}, 107002 (2003).
1032:
1033: \bibitem{Golubov}
1034: A. Golubov {\it et. al.}, J. Phys. Cond. Mat. {\bf 14}, 1353
1035: (2002).
1036:
1037: \end{thebibliography}
1038:
1039: \end{document}
1040: