1: \documentclass[floatfix,aps,prb,twocolumn,groupedaddress,showpacs]{revtex4}
2: %\documentclass[aps,prl,preprint,groupedaddress,showpacs]{revtex4}
3:
4: \usepackage{graphicx}% Include figure files
5:
6:
7: \begin{document}
8: \title{Acoustic Phonon Anomaly in MgB$_2$}
9: \author{R. F. Wood}
10: \author{Bo E. Sernelius}
11: \altaffiliation[Permanent address: ]{Department of Physics and
12: Measurement Technology, Link\"oping University, Sweden.}
13: \author{A. L. Chernyshev}
14: \altaffiliation[Also at ]{Institute of Semiconductor
15: Physics, Novosibirsk, Russia.}
16:
17: \affiliation{Solid State Division, Oak Ridge National Laboratory,
18: P.O. Box 2008, Oak Ridge, TN 37831}
19:
20: \date{\today}
21:
22: \begin{abstract}
23: Recent first principles calculations of the phonon dispersion curves in
24: MgB$_2$ have suggested the presence of anomalies in some of the curves,
25: particularly in the longitudinal acoustical (LA) branch
26: in the $\Gamma$ to $A$ direction. Similar behavior has been observed
27: in numerous other superconductors with T$_c$'s higher than those of
28: standard electron-phonon BCS superconductors.
29: Phenomenological calculations of the
30: $\Gamma\rightarrow A$ LA dispersion based on both an acoustical
31: plasmon and a ``resonant polarization'' mechanism
32: are given here to emphasize the importance of these similarities.
33: \end{abstract}
34: \pacs{63.20.-e, 63.20.Ls, 74.70.Ad, 74.25.Kc, 71.45.Gm}
35:
36: \maketitle
37:
38: Single crystals of the new superconductor MgB$_2$ large enough for
39: inelastic
40: neutron scattering experiments have not yet become available, but first
41: principles calculations of the phonon dispersion curves have appeared.
42: \cite{Bohnen,Kong}
43: In Ref.~\onlinecite{Kong},
44: a well-defined anomaly appears in the dispersion of the LA mode
45: in the $\Gamma$ to $A$ direction.
46: This anomalous behavior is remarkably similar to
47: that found in transition metal\cite{Nakagawa} and transition metal
48: carbide\cite{Smith}
49: superconductors with high transition temperatures. Indeed, it has often
50: been considered a signature of high-temperature electron-phonon
51: superconductivity in these systems. Here, we call attention to and
52: emphasize the potential importance of this similarity by carrying out model
53: calculations in which the anomalies are produced both by an acoustical
54: plasmon (AP) mechanism and by a parameterized, ${\bf q}$-dependent, interband
55: contribution to the dielectric response function.
56:
57: Many superconductors with even moderately high transition temperatures show
58: some type of anomaly in the phonon dispersion curves when compared to their
59: low-T$_c$ or nonsuperconducting analogues.
60: In this connection, the lattice dynamics in high-T$_c$ cuprate
61: materials\cite{Pintschovius} is currently being revisited.
62: It now appears that there are effects in the optical
63: modes\cite{McQueeney} and in the electronic
64: structure\cite{Lanzara} that correlate with
65: hole doping and that may be related to superconductivity.
66:
67: We are mainly interested in anomalies in the LA
68: dispersion curves because they are distinctive and prevalent, occurring in
69: systems that have no optical modes as well as those that do. They were
70: first observed and discussed for Nb (T$_c\sim 10$K)
71: more than thirty years ago\cite{Nakagawa}
72: and somewhat later in the transition metal carbides\cite{Smith}.
73: Figure \ref{fig_1} shows the LA
74: dispersion curves for NbC, TaC, and HfC, which have transition
75: temperatures of approximately 11K, 11K, and $<$1K, respectively. The
76: anomalies are obvious for the first two materials but do not occur in HfC.
77: Many other examples of this type of behavior could be given, notably the
78: elemental materials Nb, V, and Ta where only acoustical modes occur. Also
79: shown on Fig. \ref{fig_1}
80: are the results from Ref.~\onlinecite{Kong} of recent first-principles
81: calculations for MgB$_2$ (T$_c=38.5$K).
82: While only four points were calculated
83: in the $\Gamma$ to $A$ direction, they clearly show that the behavior is
84: anomalous in the sense being used here.
85: %-----------------------------------------------
86: \begin{figure}[b]
87: \includegraphics[angle=270,width=7cm,clip=true]{Fig1}
88: \caption{Phonon dispersion curves along $(q,0,0)$ for NbC, TaC, and HfC; the
89: anomalous behavior in the first two is evident. Data from the recent
90: first-principles calculations for MgB$_2$ (Ref.~\onlinecite{Kong})
91: are also shown. Lines are guides to the eye, r.l.u. means reciprocal
92: lattice units.}
93: \label{fig_1}
94: \end{figure}
95: %-----------------------------------------------
96:
97: Nakagawa and Woods\cite{Nakagawa}
98: found they could reproduce their phonon data for Nb with a Born-von
99: Karman model utilizing
100: numerous force constants. The need for such a complex
101: model was thought to be indicative of a strong electron-phonon interaction.
102: Ganguly and Wood\cite{Ganguly}
103: observed that the anomalies may be produced by the interaction
104: of acoustical plasmons and
105: phonons. While it is difficult to reconcile this suggestion with the
106: electronic band structure of
107: Nb, as pointed out by Ruvalds\cite{Ruvalds},
108: the notion that there is a connection
109: between AP-like
110: electronic modes and phonons persisted. The resonant
111: electronic polarization arising from the double shell model of Weber
112: {\it et al.}\cite{Weber}
113: is somewhat related to the acoustical plasmon mechanism in that an
114: additional mode which interacts with the ``bare'' phonons is introduced.
115: First principles calculations\cite{Weber1,Sinha}
116: that began to appear subsequently could
117: replicate the anomalies but the assignments to specific mechanisms
118: were often in conflict. However, the important roles of
119: electronic polarization and details of the band structure were clear.
120:
121: The acoustical plasmon concept was introduced by Pines\cite{Pines}
122: in 1958 and
123: subsequently applied to superconductivity by a number of
124: authors\cite{Radhakrishnan,Gelikman,Frohlich,Rothwarf}.
125: It is thought that APs may occur in systems in which heavy and
126: light mass carriers coexist and the light carriers screen the heavy ones to
127: produce an acoustical ($\omega\rightarrow 0$ as ${\bf q}\rightarrow
128: 0$) excitation. Transition metal
129: systems with the $d$ electrons playing the role of heavy particles and
130: the $s$
131: electrons the light ones have been long seen as good candidates.
132: In other
133: systems, holes may be the heavy particles and electrons the light ones. In
134: the simplest acoustical plasmon treatment of superconductivity in
135: transition metals, the
136: plasmons either replace entirely or simply complement the
137: phonons in a standard BCS-type theory. An extensive criticism of
138: AP-mediated
139: superconductivity has been given based on system stability and
140: other restrictive requirements\cite{Ginsburg,Uspenskii}.
141: Nevertheless, while to our
142: knowledge APs
143: have not been shown to exist in any bulk material (see
144: also Ref.~\onlinecite{Ruvalds}), the
145: concept is an intriguing one and based on sound theoretical arguments
146: regardless of any role it may have in superconductivity. We note that
147: acoustical plasmons have been proposed\cite{Voelker}
148: as existing in MgB$_2$, although not in the context of
149: the present lattice dynamical discussion and in a much higher frequency
150: range.
151:
152: Magnesium diboride has the AlB$_2$
153: structure consisting of hexagonal layers of
154: Mg interspersed between graphite-like B layers.
155: Several calculations\cite{Kortus,Yildirim,Ravindran} establish
156: that the occupied bands consist of $\sigma$
157: bands formed from boron $sp^2$ hybridized orbitals and $\pi$
158: bands formed from the $2p_z$ orbitals directed along the $c$
159: axis. Interactions with the Mg $3s$
160: orbitals result in the $\sigma$ bands not being fully occupied, as in
161: graphite, which creates holes at the top of the bands. The various
162: calculations are in good agreement in showing a weakly dispersing band
163: in the $\Gamma$ to $A$ direction lying just above the Fermi surface.
164: Voelker {\it et al.},\cite{Voelker}
165: have provided analytical fits to the bands they calculated,
166: with the holes in the $\sigma$ bands having
167: high effective mass for motion along the $k_z$ direction. This led those
168: authors to investigate the possible existence of acoustical plasmons. They
169: found evidence that they do exist but the validity of their calculation was
170: subsequently questioned by Ku {\it et al.}\cite{Ku}
171: We emphasize that in any case the APs
172: considered in these calculations are in a far different frequency range
173: than those needed here, as will be seen in the following.
174:
175: We write the ${\bf q}$- and $\omega$-dependent dielectric
176: function as
177: \begin{eqnarray}
178: \label{1}
179: \varepsilon({\bf q},\omega) = 1 + \alpha^0 + \alpha^l + \alpha^h
180: + \alpha^L \ ,
181: \end{eqnarray}
182: and look for its zeros in a standard approach.
183: $\alpha^l$ and $\alpha^h$ are the polarizabilities of
184: the two distinct groups of light and heavy carriers and
185: $\alpha^0$ contains all other contributions to the
186: electronic polarizability, e.g., from interband transitions;
187: $\alpha^L$ is the lattice contribution.
188: Electronic polarizabilities are calculated in the free electron
189: approximation:
190: \begin{eqnarray}
191: \label{1a}
192: \alpha^i_{\bf q}(\omega) = \frac{8\pi e^2}{q^2}\ \sum_{\bf k}
193: \frac{n^i_{{\bf q}+{\bf k}}-n^i_{\bf k}}{\omega+E^i_{\bf k}
194: -E^i_{{\bf q}+{\bf k}}+i0}\ ,
195: \end{eqnarray}
196: where $E^i_{\bf k}$ and $n^i_{\bf k}$ are the free-electron
197: energy and Fermi
198: function for the $i=l,h$ particles, respectively.
199:
200: Neglecting $\alpha^L$
201: for the moment, an approximate expression for the dispersion of the
202: APs can be obtained in the long-wavelength limit.
203: Using the first terms of the expansions given by Lindhard,\cite{Lindhard}
204: \begin{eqnarray}
205: \label{2}
206: \alpha^i_{\bf q}(\omega) \simeq \left\{ \begin{array}{cc}
207: 3\omega_{0i}^2 /q^2v_i^2 \ & \ \ q v_i > \omega\\
208: -\omega_{0i}^2 /\omega^2 \ & \ \ \ q v_i < \omega \ .
209: \end{array} \right.
210: \end{eqnarray}
211: $v_{i}$ is the Fermi velocity of the $i$th particle, and
212: $\omega_{0i}^2= 4\pi e^2n_i /m^*_i$
213: is the plasmon frequency,
214: with $m^*_i$ the effective mass and $n_i$ the particle density; the
215: imaginary parts are neglected.
216: Substituting these in Eq. (\ref{1}) and defining
217: $1+\alpha^0=\varepsilon_0$
218: gives for the acoustical plasmon frequency, $\omega_{p}(q)$,
219: \begin{eqnarray}
220: \label{5}
221: \omega_{p}(q)=\frac{v_p q}{\sqrt{1+q^2\lambda^2}}
222: \ , \ \ qv_h \ll \omega_p(q) \ll qv_l\ ,
223: \end{eqnarray}
224: where $v_p=v_h\sqrt{N_h/3N_l}$ is the acoustical plasmon sound velocity,
225: $N_i$ is the density of states of the $i$th particle
226: at the Fermi level, and
227: $\lambda=\sqrt{\frac{\varepsilon_0\hbar^2\pi}{4e^2k_{F,l}m^*_l}}$ is
228: the screening length of light particles.
229: The Landau damping of acoustical plasmons due to the continuum of
230: light particles is neglected.
231:
232: In a more complete calculation, we use Eqs. (\ref{1}) and
233: the full Lindhard expressions for $\alpha^i$.
234: To include the phonon contribution we use a simple,
235: approximate expression for the lattice polarizability
236: $\alpha^L$
237: \begin{eqnarray}
238: \label{6}
239: \alpha^L_{\bf q}(\omega)= -\omega^0_{ph}(q)^2
240: (\varepsilon^0+\alpha^l_{\bf q}(\omega))/\omega^2 \ ,
241: \end{eqnarray}
242: and take in the $\Gamma$ to $A$ direction
243: \begin{eqnarray}
244: \label{7}
245: \omega^0_{ph}(q)=c_0 \sin\left(\frac{\pi}{2}\frac{q}{q_0}\right) \ ,
246: \end{eqnarray}
247: where $q/q_0$ is the fractional distance to the BZ
248: boundary and $c_0$
249: is adjusted to give the frequency of the LA phonon in MgB$_2$
250: at the zone boundary, as taken from Refs.~\onlinecite{Bohnen,Kong}.
251: This implies that the ``bare'' phonon frequency
252: $\omega^0_{ph}$ already contains the screening due to $\varepsilon^0$
253: and the light carriers, hence this screening
254: is taken out in Eq. (\ref{6}).
255:
256: The results of such a calculation are shown in Fig. \ref{fig_2}a.
257: The unperturbed or bare phonon
258: dispersion curve is given by the dashed line and the
259: acoustical plasmon by the dotted solid curve. The latter is
260: Landau damped as it enters the heavy particle continuum at $q_c\sim 0.25$.
261: Note again that the AP
262: mode shown here is calculated {\it without} taking into account the
263: phonon contribution. Also the bare
264: phonon is {\it before} the interaction with heavy particles is
265: included.
266: The result of the full calculation
267: ($\varepsilon=\varepsilon_0+\alpha^l+\alpha^h+\alpha^L$) are shown by
268: the solid line. After the phonon
269: and heavy particles are allowed to interact, the phonon frequency
270: rises above the
271: unperturbed curve and continues to be affected even after the plasmon is
272: damped, causing the anomalous behavior to develop.
273: Interestingly,
274: the perturbed phonon remains well-defined until it enters the heavy
275: particle
276: continuum where its width increases rapidly in the region of the anomaly
277: due to the continuum of heavy particle
278: contribution to the imaginary part of $\varepsilon({\bf q},\omega)$.
279: Neutron scattering experiments should readily pick up this feature if it
280: exists. We note that this is also the region where the electron-phonon
281: interaction was found to be strong in
282: Ref.~\onlinecite{Kong}. Some of these effects will be discussed in
283: more detail in a subsequent paper.
284: %-----------------------------------------------
285: \begin{figure}[t]
286: \includegraphics[angle=270,width=7cm,clip=true]{Fig2a}\vskip 0.5cm
287:
288: \includegraphics[angle=270,width=7cm,clip=true]{Fig2b}
289: \caption{(a) An example of the calculation of the anomaly given by the
290: acoustical plasmon mechanism, as described in the text. (b) The
291: anomaly as a function of the mass of the heavy particle.}
292: \label{fig_2}
293: \end{figure}
294: %-----------------------------------------------
295:
296: To obtain the curves on Fig. \ref{fig_2}a we took the
297: mass ratio $m^*_h/m^*_l=20$, and Fermi momentum ratio $k_{F,l}/k_{F,h}=2.4$.
298: $n_l=10^{21}$ cm$^{-3}$,
299: $n_h=0.75\times 10^{20}$ cm$^{-3}$, $v_l=2.4\times 10^7$ cm/s,
300: and $v_h=5\times 10^5$ cm/s. This corresponds to the choice of light mass
301: $m^*_l= 1.5 m_0$. Also a value of $\varepsilon_0=30$ was assigned.
302: As was
303: the case with Nb, it is difficult to reconcile these parameter values with
304: the published band structure calculations and this remains a serious
305: drawback for the application of the acoustical plasmon concept to
306: lattice dynamics.
307:
308: Additional dispersion curves, obtained by changing the effective mass of
309: the heavy particles over a limited range, are shown on
310: Fig. \ref{fig_2}b. By varying
311: other parameters, a wide variety of behavior can be obtained, which may be
312: of interest for other materials.
313:
314: We followed Fr\"ohlich and Rothwarf a bit further by calculating T$_c$
315: from a
316: modified BCS-like expression,
317: $kT_c=\hbar\omega_{pm}\exp(-1/F)$. $\omega_{pm}$
318: was taken to be the
319: frequency at which the plasmon enters the heavy hole continuum and $F$ was
320: calculated as described in Refs.~\onlinecite{Frohlich,Rothwarf}.
321: Values of T$_c$ quite close to the
322: experimental value could be found but they were very sensitive to the
323: relevant parameters which could be changed over a fairly wide range without
324: greatly changing the dispersion curves. Having in mind the known
325: restrictions on the acoustical plasmon
326: mechanism\cite{Ginsburg,Uspenskii},
327: we do not consider
328: these results of direct, fundamental significance for the high transition
329: temperatures in MgB$_2$, that is to say, we do not argue that
330: superconductivity is produced by
331: acoustical plasmons. However,
332: the results do suggest that the electronic
333: structure of this material leads to some resonance-like interaction of
334: the electrons or holes with ``bare''
335: phonons. How this might come about from another viewpoint is illustrated
336: by the following calculation.
337:
338: In Eq. (\ref{1}), $\alpha^h$,
339: the polarizability of the heavy carriers, which gives rise
340: to the acoustical plasmons,
341: was replaced by a ${\bf q}$-dependent polarization term,
342: $\alpha^{lh}$, ascribed
343: here to interband transitions that were previously included in the
344: constant $\varepsilon^0$. $\alpha^{lh}(q)$
345: was chosen to be a ``skewed Lorentzian'' of the form
346: \begin{eqnarray}
347: \label{8}
348: \alpha^{lh}(q)=\frac{a}{1+b^2(q-q_0)^2} \ .
349: \end{eqnarray}
350: We initially fixed $q_0$ at $0.7$
351: to locate the maximum in the vicinity of the
352: dip in the phonon dispersion curve. Modification of the bare phonon
353: dispersion and trial and error with the parameters $a$ and $b$
354: readily provided
355: a satisfactory representation of the anomalous behavior. The results of
356: three such calculations for different values of $q_0$
357: are shown by the upper
358: curves on Fig. \ref{fig_3}.
359: We note that the importance of interband transitions has
360: been emphasized in recent electronic structure calculations and
361: particularly those of Refs.~\onlinecite{Voelker,Ku}.
362: %-----------------------------------------------
363: \begin{figure}[t]
364: \includegraphics[angle=270,width=7cm,clip=true]{Fig3}
365: \caption{Results for the resonant polarization mechanism as described
366: in the text. The upper curves
367: are for the phonon and the lower ones are for the supershell frequencies.}
368: \label{fig_3}
369: \end{figure}
370: %-----------------------------------------------
371:
372: An interesting aspect of this calculation is that the $q$-dependence
373: of $\alpha^{lh}$ correlates
374: closely with that of the ``supershell frequency'', $\omega_{ss}(q)$,
375: in Weber's double-shell model\cite{Weber,Weber1}.
376: Smith {\it et al.}\cite{Smith1} calculated $\omega_{ss}(q)$ for TaC by
377: assigning
378: the free electron mass to the supershell.
379: The supershell motion was then
380: retained explicitly in the dynamical matrix rather than being transformed
381: out as is usually done in the shell model. They found that
382: $\omega_{ss}(q)$ for the
383: longitudinal mode had a pronounced, asymmetrical dip centered at the
384: position of the anomaly. Loosely speaking, the function of the supershell
385: is to provide a ``resonant polarization'' mechanism to screen the relevant
386: part of the dynamical matrix. This implies that the shell frequency should
387: vary inversely with the square root of this polarization. Calculating the
388: $q$-dependence of $\omega_{ss}$ in this way from Eq. (\ref{8})
389: and comparing it with the $\omega_{ss}$
390: extracted from Ref.~\onlinecite{Smith1}
391: gives the results shown in the lower part of Fig. \ref{fig_3};
392: the agreement is obviously quite good. The supershell frequencies have been
393: normalized to one another at $q=0$
394: and then shifted into the range of phonon
395: frequencies. Of course, the introduction of a distinct, well-defined mode
396: for the supershell in this region of $({\bf q},\omega)$
397: space is a rather drastic
398: simplification whose implications are not clear. However, it is remarkable
399: that the calculations of Ref.~\onlinecite{Ku}
400: also point to a sharp
401: resonance in the vicinity of 2.5~eV attributed to interband
402: transitions. In any case, the simple form in Eq. (\ref{8}) does seem
403: to capture much the same physics contained in the double shell model,
404: and perhaps also that in less phenomenological treatments.
405:
406: The calculations of the phonon spectra of MgB$_2$ in
407: Ref.~\onlinecite{Kong}, which show the anomalies,
408: used a density functional theory (DFT), full-potential linear response,
409: LMTO approach. The
410: calculations of Ref.~\onlinecite{Bohnen},
411: which show no anomaly, used a mixed
412: basis pseudopotential method and DFT \cite{elastic_c}.
413: While the results of the two
414: calculations are in general agreement, they differ strongly in the
415: dispersion of interest here. Kong {\it et al.}\cite{Kong}
416: also calculated
417: the electron-phonon interaction and found it to be
418: particularly large in the vicinity of the anomaly. Since the calculations
419: of Ref.~\onlinecite{Kong}
420: obtained the anomalies, their origin must be embedded in and, in
421: principle, can be extracted from those calculations. Conversely, since the
422: anomalies were not obtained in Ref.~\onlinecite{Bohnen},
423: a comparison of the two approaches
424: is important from both a computational and a physical standpoint. As
425: discussed above, attempts to identify a well-defined mechanism producing
426: the anomalies in other materials in the 1970s were
427: inconclusive and MgB$_2$ may now provide an impetus to reconsider their
428: origins. It will indeed be interesting to see how well these calculations
429: fit the experimental lattice dynamical data when crystals large enough
430: for neutron scattering become
431: available. Of course, any hint of acoustical plasmon effects also would be
432: of great interest.
433:
434: To summarize, one recent first-principles calculation of the phonon
435: dispersion curves in MgB$_2$ shows anomalies in the LA branch similar
436: to those
437: found in other high-T$_c$ electron-phonon
438: BCS superconductors; however, another calculation
439: did not give them. These anomalies can be calculated using an acoustical
440: plasmon mechanism but the band parameters involved seem unrealistic. A
441: resonant polarization mechanism in which interband transitions give a
442: ${\bf q}$-dependent
443: contribution to the dielectric response function is probably
444: more realistic. A detailed analysis of the new MgB$_2$ lattice-dynamical
445: calculations and comparison with earlier calculations for transition metal
446: materials should be invaluable in giving insight into the underlying
447: mechanism.
448:
449: We would like to acknowledge valuable discussions with R.~Fishman,
450: I.~Mazin, and D.~Singh. This research was supported by
451: ORNL, managed by UT-Battelle, LLC, for the U.S. DOE under
452: contract DE-AC05-00OR22725.
453:
454:
455: %-------------------------------------------------------
456: \begin{thebibliography}{}
457:
458: \bibitem{Bohnen} K.-P. Bohnen, R. Heid, and B. Renker,
459: Phys. Rev. Lett. {\bf 86}, 5771 (2001).
460:
461: \bibitem{Kong} Y. Kong, O. V. Dolgov, O. Jepsen, and O. K. Andersen,
462: Phys. Rev. B {\bf 64}, 20501 (2001).
463:
464: \bibitem{Nakagawa} Y. Nakagawa and A. D. B. Woods,
465: Phys. Rev. Lett. {\bf 11}, 271 (1963).
466:
467: \bibitem{Smith} H. G. Smith and W. Gl\"aser, Phys. Rev. Lett. {\bf 25},
468: 1611 (1970).
469:
470: \bibitem{Pintschovius} L. Pintschovius, N. Pyka, W. Reichardt,
471: A. Yu. Rumiantsev, N. L. Mitrofanov, A. S. Ivanov, G. Collin,
472: and P. Bourges, Physica C {\bf 185-189}, 156 (1991).
473:
474: \bibitem{McQueeney} R. J. McQueeney, J. L. Sarrao, P. G. Pagliuso,
475: P. W. Stephens, and R. Osborn, Phys. Rev. Lett. {\bf 87}, 077001 (2001).
476:
477: \bibitem{Lanzara} A. Lanzara, P. V. Bogdanov, X. J. Zhou, S. A. Kellar,
478: D. L. Feng, E. D. Lu, T. Yoshida, H. Eisaki, A. Fujimori, K. Kishio,
479: J.-I. Shimoyama, T. Noda, S. Uchida, Z. Hussain, and Z.-X.
480: Shen, Nature {\bf 412}, 510 (2001).
481:
482: \bibitem{Ganguly} B. N. Ganguly and R. F. Wood, Phys. Rev. Lett. {\bf 28},
483: 681 (1972).
484:
485: \bibitem{Ruvalds} J. Ruvalds, Adv. Phys. {\bf 30}, 677 (1981).
486:
487: \bibitem{Weber} W. Weber, H. Bilz, and U. Schr\"oder,
488: Phys. Rev. Lett. {\bf 28}, 600 (1972).
489:
490: \bibitem{Weber1} W. Weber, Phys. Rev. B {\bf 8}, 5082 (1973).
491:
492: \bibitem{Sinha} S. K. Sinha and B. N. Harmon, Phys. Rev. Lett. {\bf 35},
493: 1515 (1975); W. Hanke, J. Hafner, and H. Bilz, Phys. Rev. Lett. {\bf
494: 37}, 1560 (1976); C. M. Varma and W. Weber, Phys. Rev. Lett. {\bf 39}, 1094
495: (1977).
496:
497: \bibitem{Pines} D. Pines, Can. J. Phys. {\bf 34}, 1379 (1956).
498:
499: \bibitem{Radhakrishnan} V. Radhakrishnan, Phys. Lett. {\bf 16}, 247
500: (1965).
501:
502: \bibitem{Gelikman} B. T. Gelikman, Sov. Phys. JETP {\bf 21}, 796 (1965).
503:
504: \bibitem{Frohlich} H. Fr\"ohlich, J. Phys. C (Proc. Phys. Soc.) {\bf 1},
505: 548 (1967).
506:
507: \bibitem{Rothwarf} A. Rothwarf, Phys. Rev. B {\bf 2}, 3560 (1970).
508:
509: \bibitem{Ginsburg} V. L. Ginsburg, Sov. Phys. Usp. {\bf 13}, 335
510: (1970);
511: V. L. Ginsburg and D. A. Kirzhnits, {\it
512: High-Temperature Superconductivity}, (Consultants Bureau, New York,
513: 1982).
514:
515: \bibitem{Uspenskii} Yu. A. Uspenskii, Sov. Phys. JETP {\bf 49}, 822 (1979).
516:
517: \bibitem{Voelker} K. Voelker, V. I. Anisimov, and T. M. Rice, cond-mat/0103082.
518:
519: \bibitem{Kortus} J. Kortus, I. I. Mazin, K. D. Belashchenko,
520: V. P. Antropov, and L. L. Boyer, Phys. Rev. Lett. {\bf 86},
521: 4656 (2001).
522:
523: \bibitem{Yildirim} T. Yildirim, O. G\"ulseren, J. W. Lynn, C. M. Brown,
524: T. J. Udovic, Q. Huang, N. Rogado, K. A. Regan, M. A. Hayward,
525: J. S. Slusky, T. He, M. K. Haas, P. Khalifah, K. Inumaru, and R. J. Cava,
526: Phys. Rev. Lett. {\bf 87}, 037001 (2001).
527:
528: \bibitem{Ravindran} P. Ravindran, P. Vajeeston, R. Vidya,
529: A. Kjekshus, and H. Fjellv\aa g, Phys. Rev. B {\bf 64}, 224509 (2001).
530:
531: \bibitem{Ku} W. Ku, W. E. Pickett, R. T. Scalettar, and A. G. Eguiluz,
532: Phys. Rev. Lett. {\bf 88},
533: 057001 (2002).
534:
535: \bibitem{Lindhard} J. Lindhard, Dan. Mat. Fys. Medd. {\bf 28}, 1
536: (1954).
537:
538: \bibitem{Smith1} H. G. Smith, N. Wakabayashi, and M. Mostroller,
539: in {\it Superconductivity in $d$-
540: and $f$-band Metals}, (ed. by D. H. Douglas, Plenum Publishing
541: Co., New York, 1976).
542:
543: \bibitem{elastic_c} First principles calculations of the elastic
544: constants seem to agree with ones extracted from the phonon dispersion
545: in Ref.~\onlinecite{Bohnen}: see A. K. M. A. Islam and F. N. Islam,
546: Physica C {\bf 363}, 189 (2001) and Ref.~\onlinecite{Ravindran}.
547:
548: \end{thebibliography}
549:
550:
551: \end{document}
552:
553:
554:
555:
556:
557:
558:
559:
560: