1:
2: % ****** Start of file apssamp.tex ******
3: %
4: % This file is part of the APS files in the REVTeX 4 distribution.
5: % Version 4.0 of REVTeX, August 2001
6: %
7: % Copyright (c) 2001 The American Physical Society.
8: %
9: % See the REVTeX 4 README file for restrictions and more information.
10: %
11: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
12: % as well as the rest of the prerequisites for REVTeX 4.0
13: %
14: % See the REVTeX 4 README file
15: % It also requires running BibTeX. The commands are as follows:
16: %
17: % 1) latex apssamp.tex
18: % 2) bibtex apssamp
19: % 3) latex apssamp.tex
20: % 4) latex apssamp.tex
21: %
22: \voffset=1.5cm
23: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb,floatfix]{revtex4}
24: %\documentclass[showpacs,preprintnumbers,amsmath,amssymb,floatfix]{revtex4}
25:
26: \usepackage{graphicx}% Include figure files
27: \usepackage{dcolumn}% Align table columns on decimal point
28: \usepackage{bm}% bold math
29:
30:
31: \begin{document}
32:
33: \title{Phase separation and electron pairing in repulsive Hubbard clusters}
34:
35: \author{G. W. Fernando$^{1,4}$,
36: A. N. Kocharian$^2$, K. Palandage$^1$, Tun Wang$^1$, and J. W. Davenport$^3$}
37: \address{$^1$U-46, Physics Dept., University of Connecticut,
38: Storrs, CT 06269}
39: \address{$^2$Physics Dept., California State University, Northridge, CA 91330}
40: \address{$^3$Physics Dept., Brookhaven National Laboratory,
41: Upton, NY 11973}
42: \address {$^4$Institute of Fundamental Studies, Hantana Road, Kandy, Sri Lanka}
43: %\date{\today{}}
44: %\maketitle
45:
46:
47: \begin{abstract}
48:
49: Exact thermal studies of small (4-site, 5-site and 8-site)
50: Hubbard clusters with local electron repulsion
51: yield intriguing insight
52: into phase separation,
53: charge-spin separation, pseudogaps, condensation,
54: in particular, pairing fluctuations away from half filling
55: (near optimal doping).
56: These exact calculations, carried out in
57: canonical (i.e. for fixed electron number $N$)
58: and grand canonical
59: (i.e. fixed chemical potential $\mu$) ensembles,
60: monitoring variations in temperature
61: $T$ and magnetic field $h$, show rich phase diagrams in a
62: $T$-$\mu$ space
63: consisting of pairing fluctuations and signatures of
64: condensation. These electron pairing instabilities are seen
65: when the onsite Coulomb interaction $U$ is
66: smaller than a critical value $U_c(T)$ and they point to a possible
67: electron pairing mechanism.
68: The specific
69: heat, magnetization, charge pairing and spin pairing
70: provide strong support for the existence of competing
71: (paired and unpaired) phases near optimal doping in
72: these clusters as observed in recent experiments
73: in doped La$_{2-x}$Sr$_x$CuO$_{4+y}$ high T$_c$
74: superconductors.
75:
76:
77:
78: \end{abstract}
79: \pacs{65.80.+n, 73.22.-f, 71.10.Fd, 71.27.+a, 71.30.+h, 74.20.Mn}
80: % Classification Scheme.
81: \keywords{high T$_c$ superconductivity, particle pairing, phase
82: diagram, crossover, charge and spin pseudogaps}
83:
84: \maketitle
85: Since the discovery of the high temperature superconductors (HTSCs),
86: there has been an intense debate about a possible electron
87: (or hole) pairing mechanism. Early on,
88: P.~W.~Anderson~\cite{RVB}
89: suggested
90: that the large positive onsite Coulomb interaction in the Hubbard model
91: should contain the key to some of the perplexing physics observed in the
92: HTSCs. Although it is next to impossible to list every single
93: effort related to testing the above assertion, important progress has
94: been made in attempts to obtain a better
95: understanding of the physical
96: properties of these materials~\cite{Anderson,Nature,Timusk,Marshall,
97: Kivelson_Review, Andrea}.
98: The bad metallic behavior and small correlation length of
99: dynamical spin fluctuations (different from conventional
100: superconductors) make HTSCs~\cite{Zha} a suitable platform to examine
101: the role of local interactions.
102: Exact studies of the Hubbard clusters placed in magnetic field $h$
103: are fundamental for understanding the nature of magnetism and
104: corresponding spin gaps in various cluster
105: geometries~\cite{Pastor,Sebold}.
106: Repulsive interactions can lead to phase separation,
107: electron pairing and ground state superconductivity in certain
108: mesoscopic structures~\cite{scalettar,white}.
109: In our opinion,
110: thermal properties of small clusters with strong correlations have
111: not been fully explored, although there have been numerous
112: exact calculations~\cite{Shiba,schumann}, and the present study is an
113: attempt to fill this void.
114:
115: Our recent work~\cite{PRB} indicates that
116: Hubbard clusters,
117: when connected to a particle
118: reservoir and a thermal bath,
119: possess a vivid variety
120: of interesting thermal and
121: physical properties, that could pave the way for a new class
122: of tunable materials. These inferences were drawn by not only
123: carrying out exact diagonalizations of the many-body Hamiltonian,
124: but also using these eigenvalues
125: in a statistical ensemble to study thermal
126: and other transitions,
127: by monitoring, for example, susceptibilities, {\sl i.e. fluctuations}.
128: The many-body nature of these correlated problems is at least
129: partly hidden in statistical fluctuations and it is no wonder
130: that these fluctuations give rise to intriguing results. The
131: crossovers and transitions between various phases,
132: that we identify at finite and zero temperatures,
133: are found by monitoring the corresponding
134: thermal and ground state properties
135: without taking the thermodynamic limit. The
136: results described in this paper provide new
137: insights into the physics of the 4-site as well as larger (repulsive)
138: Hubbard clusters.
139:
140:
141: These attempts may be questioned
142: since they appear not to comply with the standard
143: applications of
144: statistical mechanics with respect to the thermodynamic limit.
145: However, for finite
146: systems, it is necessary to re-evaluate these ideas
147: and a paradigm shift in our thinking
148: may be necessary.
149: We have already shown
150: that in such finite systems, one can define and identify
151: various transitions and phase
152: boundaries by monitoring maxima and minima
153: in susceptibilities~\cite{PRB}.
154: As synthesis techniques improve at a
155: rapid rate, it has become possible to synthesize isolated clusters and
156: hence it is clear that we need not
157: always look at the thermodynamic limit.
158: Finite, mesoscopic
159: structures (i.e. clusters containing a few atoms) in suitable
160: topological forms will be realistic enough to
161: synthesize and extract fascinating
162: physical properties.
163: Also, since the HTSCs
164: are known to consist of (stripes and
165: possibly other) inhomogeneities ~\cite{Tranquada},
166: it is possible that these cluster studies may be able to capture
167: some of the essential physics of the HTSCs.
168: The following is a list of properties, resulting from
169: our exact (4-site, 5-site and 8-site) Hubbard cluster studies,
170: that is shared with the HTSCs.
171:
172:
173: \begin{itemize}
174: \item Phase diagrams in a temperature-chemical potential (doping) plane
175: and the presence of a multitude of fascinating phases, including
176: Mott-Hubbard like paramagnetic and
177: antiferromagnetic phases~\cite{PRB}.
178:
179: \item Vanishing of a charge gap at a critical set of parameters
180: and thereby providing an effective
181: attraction leading to onset of electron
182: charge pairing (2e)
183: at a critical temperature $T^P_{c}$.
184:
185: \item Spin pairing at a lower temperature
186: ($T^P_{s}$) and hence the formation
187: of rigidly bound spin pairs in a narrow, critical region of
188: doping.
189:
190:
191: \item Low temperature specific heat peak, reminiscent of the
192: experimental,
193: low temperature specific heat behavior in the HTSCs~\cite{Tallon}.
194:
195: \item Temperature vs $U$
196: phase diagram, indicating the pressure effect
197: on the superconducting transition temperature as seen in recent
198: experiments~\cite{Xiao}.
199:
200: \item The presence of a dormant magnetic state, lurking in the above narrow,
201: critical region of doping, that could be stabilized by either
202: applying a magnetic field, going above the spin pairing temperature,
203: or changing the chemical potential,
204: as seen in a recent, notable experiment~\cite{hashini}.
205:
206: \item The opening of a pseudogap above the pairing temperature,
207: as observed
208: in NMR experiments,
209: in both hole and electron doped cuprates~\cite{nmr}.
210:
211: \item Larger clusters with different topologies and
212: higher dimensionality illustrating how
213: the above properties get scaled with size.
214:
215: \end{itemize}
216:
217: In what follows, we will address the similarities
218: listed above using the many-body eigenvalues
219: of the Hubbard clusters ({\sl with energies measured in units
220: of $t$, the hopping parameter}) in combination with
221: statistical mechanics.
222: The grand partition function $Z$
223: (where the number of electrons $N$ and
224: the projection of spin $s^z$
225: can fluctuate) and its derivatives are calculated exactly without
226: taking the thermodynamic limit.
227: The response
228: functions related to electron or hole doping (i.e.
229: chemical potential $\mu$) or magnetic field $h$ demonstrate
230: clearly
231: observable, prominent
232: peaks paving the way for
233: strict definitions of Mott-Hubbard (MH), antiferromagnetic (AF),
234: spin pseudogaps and related crossover
235: temperatures~\cite{PRB,JMMM}.
236:
237:
238: %
239: Our exact studies of 4-site clusters indicate a
240: net electron attraction leading to the formation of bound electron
241: pairs and possible
242: condensation at finite temperature
243: for $U<U_c(T)$
244: (i.e. suggestive of
245: superconductivity)~\cite{PRB,JMMM}.
246: This pairing mechanism in the 4-site cluster, at 1/8 hole doping
247: ($\left\langle N\right\rangle\approx 3$)
248: away from half filling, exists when the onsite Coulomb interaction
249: $U$ is less
250: than an analytically obtained critical value,
251: $U_c(T=0) = 4.584$
252: (in units of the hopping parameter $t$). This critical value, first
253: reported in Ref.~\cite{JMMM},
254: is temperature
255: dependent and can be associated with an energy gap (order parameter)
256: which becomes negative below
257: $U_c(T)$
258: implying that it is more
259: energetically favorable to have a bound pair of
260: electrons (or holes) compared to two unpaired
261: ones at an optimal chemical potential
262: (or doping level) $\mu = \mu_P = 0.658$.
263: Above this critical
264: value $U_c(T)$,
265: there is a Mott-Hubbard like gap that exists when the average
266: particle number
267: $\left\langle N\right\rangle\approx 3$;
268: this gap decreases monotonically as $U$
269: decreases and vanishes at
270: $U_c(T)$.
271: The vanishing of the gap can be
272: directly linked to the {\sl onset of pair formation}.
273: There is an interval (width) around
274: $\mu_P$,
275: where the pairing phase competes with a phase (having a high
276: magnetic
277: susceptibility) that suppresses pairing at `moderate' temperatures.
278:
279:
280: \begin{figure} %[h]
281: \begin{center}
282: \includegraphics*[width=20pc]{Tvsmu_close.eps}
283: \hfill
284: \end{center}
285: \caption {The
286: $T$-$\mu$ phase diagram near $\mu_P=0.658$
287: ($\left\langle N\right\rangle\approx 3$)
288: at $U=4$ for the 4-site cluster.
289: The inset shows a
290: corresponding section (at a different scale) of the
291: $T$-$\mu$ phase diagram for $U=6$.
292: For $U=4$, note how the paired states condense at
293: low temperature with a nonzero pair binding energy, while at
294: higher temperatures, unpaired states begin to appear. This
295: picture supports the idea that there is inhomogeneous, electronic phase
296: separation here. When $U$ is higher than $U_c(0)=4.584$,
297: these inhomogeneities disappear and a Mott-Hubbard like
298: stable paramagnetic insulating
299: region results around optimal doping. Note how the (low
300: temperature) region around optimal doping changes from a
301: pairing phase at $U=4$ to a paramagnetic phase at $U=6$ (inset) with
302: charge-spin separation as described in the text
303: and Ref.~\cite{PRB}.}
304: \label{fig:ph4_6}
305: \end{figure}
306:
307: An enlarged view of the
308: $T$-$\mu$ phase diagram, for the 4-site cluster near $\mu_P$, is shown in
309: Fig.~\ref{fig:ph4_6}. This exact phase diagram
310: (at $U=4$)
311: in the vicinity of the optimally doped ($N\approx 3$) regime has been
312: constructed using the ideas described in the text and
313: Ref.~\cite{PRB}. The electron pairing temperature,
314: $T_{c}^{P}$,
315: identifies the onset of charge pairing. As temperature
316: is further lowered, spin pairs begin to form at
317: $T_{s}^{P}$.
318: At this temperature (with zero
319: magnetic field),
320: spin susceptibilities become very weak indicating the
321: disappearance of the $\left\langle N\right\rangle\approx 3$
322: states. Below this spin pairing temperature, only paired states
323: are observed to exist having a
324: certain rigidity, so that a nonzero magnetic field or a finite
325: temperature is required to break
326: the pairs.
327: From a detailed analysis, it becomes evident that the system is on the
328: verge of an instability; the paired phase competing with
329: a phase that suppresses pairing
330: which has a high, zero-field magnetic susceptibility.
331: As the temperature is lowered,
332: the number of
333: $\left\langle N\right\rangle\approx 3$
334: (unpaired) clusters begins to decrease
335: while a mixture of (paired)
336: $\left\langle N\right\rangle\approx 2$ and
337: $\left\langle N\right\rangle\approx 4$
338: clusters appears.
339: Interestingly, the
340: critical doping $\mu_P$
341: (which corresponds to a filling
342: factor of $1/8$ hole-doping away from half filling),
343: where the above pairing
344: fluctuations take place when
345: $U<U_c(T)$,
346: is close to the doping level near which numerous intriguing
347: properties have been observed
348: in the hole-doped HTSCs.
349:
350:
351:
352:
353: \begin{figure} %[h]
354: \begin{center}
355: \includegraphics*[width=20pc]{cv_new.eps}
356: \end{center}
357: \caption {Specific heat vs temperature at $U=4$, calculated in the
358: grand canonical ensemble for the 4-site cluster at several doping
359: values near the critical doping,
360: $\mu_P\approx$ 0.658.
361: Note how the low
362: temperature peak shifts to higher temperatures when
363: the doping is changed from its
364: critical value.}
365: \label{fig:sp_ht}
366: \end{figure}
367:
368:
369: \begin{figure} %[h]
370: \begin{center}
371: \includegraphics*[width=20pc]{ph_TU.eps}
372: \hfill
373: \end{center}
374: \caption { $T$ vs $U$
375: phase diagram for the optimally doped
376: 4-site clusters, based on our exact calculations.
377: The inset shows
378: the charge gap
379: as a function of $U$ at zero temperature. A negative charge gap
380: implies charge pairing.}
381: \label{fig:ph_tu}
382: \end{figure}
383: %%%%%%%%%%%%%%%%
384: Specific heat
385: calculations (Fig.~\ref{fig:sp_ht} ),
386: associated with energy fluctuations,
387: also provide further support
388: for an electronic
389: phase change at low temperature.
390: There is strong evidence for pair condensation, from specific heat
391: calculations shown in the figure.
392: % The region
393: % between the peaks can be related to a pseudogap region
394: % indicating a suppression of
395: %density of
396: % states as the temperature is lowered from above.
397: As seen in this figure, there is a well separated, low temperature
398: peak at $\mu_P=0.658$ (around 40 K, if the hopping
399: parameter is set to 1 eV and $U$ to 4 eV).
400: This peak, which shifts to higher temperatures when
401: the doping level is different from crtical doping,
402: is due to fluctuations between paired states
403: ($\left\langle N\right\rangle = 2$ and
404: $\left\langle N\right\rangle = 4$).
405: This low temperature peak is in agreement with specific heat
406: experiments carried out for the HTSCs~\cite{Tallon}, and is a
407: manifestation of the near degeneracy of the states in the neighborhood of
408: critical doping $\mu_P$.
409:
410:
411:
412: Our calculations may also be used to reproduce the
413: variation of
414: $T_c (p)$
415: vs pressure $p$ in the HTSCs,
416: assuming that $U$ decreases
417: with pressure~\cite{Chen}. Fig.~\ref{fig:ph_tu} shows condensation
418: of electron charge below $T\leq T_P(U)$ with bound charge $2e$ and
419: decoupled spin
420: $\left\langle s^z\right\rangle = 1/2$
421: for the 4-site cluster.
422: Below the lower curve $T_s(U)$, the spin degrees are bounded and a
423: finite applied magnetic field is needed to break them
424: ~\cite{Sebold,PRB}.
425: Thus below $T\leq T_s(U)$, both the charge and spin
426: are condensed and the spin degrees can
427: follow those of charge. The inset in Fig.~\ref{fig:ph_tu} shows
428: the variation of the charge gap, $E(2) + E(4) - 2E(3)$,
429: as a function of $U$ where $E(N)$
430: refers to the canonical energies for $N$ electrons at $T=0$.
431: When this gap is negative, pairing is
432: favored as discussed in Ref.~\cite{PRB}.
433: In addition, the increase of
434: $T_s(U)$ reproduces
435: $T_c$ (superconducting transition temperature) {\it vs}
436: pressure $p$ in the optimally and nearly
437: optimally doped HTSC materials~\cite{Xiao}, indicating a significant
438: role of pair binding in enhancing $T_c(p)$.
439:
440:
441: Exact results for
442: $\left\langle N\right\rangle\approx 3$
443: in Fig.~\ref{fig:ph_tu} suggest that
444: the enhancement of $T_c$ in the
445: optimally doped HTSCs may be due to an increase of pairing with
446: decreasing $U$ under pressure rather than an increase of the
447: pressure-induced hole concentration. Thus it
448: appears that the 4-site cluster near
449: $\left\langle N\right\rangle\approx 3$ indeed
450: captures the essential physics of
451: %the pressure effect on condensation
452: the electron condensation under pressure.
453: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
454:
455: \begin{figure} %[h]
456: \begin{center}
457: \includegraphics*[width=20pc]{n_s_h_new.eps}
458: \hfill
459: \end{center}
460: \caption {Variation of electron number
461: $\left\langle N\right\rangle$
462: and magnetization
463: $\left\langle s^z\right\rangle$
464: as a function of external magnetic field for
465: several values close to the critical doping $\mu_P = 0.658$
466: at $T=0.002$ and U=4 for the 4-site cluster. Note how
467: the $\left\langle N\right\rangle = 3$ clusters get stabilized in a
468: nonzero magnetic field at low temperature. These results support
469: the recent observation of a dormant magnetic state near optimal doping in
470: hole-doped La cuprates~\cite{hashini}.}
471: \label{fig:degree_freedom_1}
472: \end{figure}
473:
474: %%%%%%%%%%%%%%%%
475: \begin{figure} %[h]
476: \begin{center}
477: \includegraphics*[width=20pc]{gap8_5.eps}
478: \hfill
479: \end{center}
480: \caption {(a) Charge gaps for the $2{\rm x}4$ cluster at $T=0$
481: for various couplings $c$ between the squares. (b) Charge
482: gaps for the 5-site cluster
483: illustrate how the coupling $c$ to a 5th site
484: above the square is responsible for narrowing the window of
485: $U$ for which pairing exists. In both (a) and (b), the doping level
486: is one electron off half-filling and the couplings $t$
487: within the squares are set to $1$. Compare these with the inset of
488: Fig.~\ref{fig:ph_tu}.}
489: \label{fig:gap0204}
490: \end{figure}
491:
492:
493:
494:
495:
496: Another intriguing fact emerging from the exact thermal studies of
497: the 4-site clusters is the existence of a magnetic state (unpaired
498: states with $\left\langle N\right\rangle = 3$) with a large
499: magnetic susceptibility.
500: At rather
501: low temperature $T\le T^{P}_{s}$, this state is
502: dormant.
503: However, a small magnetic field or a
504: change in chemical
505: potential can stabilize it over the paired states
506: $\left\langle N\right\rangle\approx 2,4$ as seen from
507: Fig.~\ref{fig:degree_freedom_1}
508: and the calculated grand canonical probabilities (not shown).
509: The variation of the magnetic field mimics the
510: doping to some extent here. Small changes in doping
511: (at zero field) can also switch
512: the system from one state to
513: another with a different
514: $\left\langle N\right\rangle$.
515: These are useful for understanding
516: some recent experimental results reported in Ref~\cite{hashini},
517: where a magnetic (and non-superconducting) state has been observed
518: near 1/8 hole-doping in La$_{2-x}$Sr$_x$CuO$_{4+y}$. This system is
519: said to be on the verge of an instability, surprisingly similar to
520: what we observe in these clusters at
521: $\left\langle N\right\rangle\approx 3$
522: (i.e. near optimal doping away from half-filling).
523:
524:
525: In order to monitor the size effects on the properties described above
526: for the 4-site cluster, we have carried out a series of
527: numerical
528: calculations for
529: clusters with different topologies and sizes.
530: %Figs.~\ref{fig:num}(b) and
531: Fig.~\ref{fig:gap0204}
532: illustrates one such set of calculations
533: of charge gaps done on a 8-site cluster
534: ($2{\rm x}4$ ladder),
535: where the hopping term or coupling $c$ between the two squares was allowed
536: to be different from the coupling within a given square. The pairing
537: fluctuations that are seen for the 4-site cluster exists even for these
538: ladders near half filling
539: ($\left\langle N\right\rangle\approx$ 7),
540: and most of the trends observed for the 4-site clusters,
541: such as the MH like charge gaps and
542: vanishing of such gaps at critical $U$ values,
543: remain valid here. The fluctuations that occur here at optimal
544: doping are among
545: the states with $\left\langle N\right\rangle\approx$ 6, 7 and 8
546: electrons.
547: Clearly,
548: the dormant magnetic state corresponds to
549: $\left\langle N\right\rangle\approx$ 7.
550: In addition, a 5-site
551: pyramid with a square base shows a pairing gap when the coupling $c$ to the
552: fifth site (i.e. the site above the square) is weak (up to about
553: 0.4$t$ where $t$ is the hopping parameter in the plane)
554: and disappears above this coupling strength.
555:
556:
557:
558: All of the above, from the 4-site and larger cluster calculations,
559: points to a pair binding instability
560: near optimal doping at relatively low temperature.
561: Thermal and quantum
562: fluctuations in the density of holes between the clusters (for
563: $U<U_c(0)$) make it energetically more favorable to form pairs.
564: In this case, snapshots of the system at relatively low
565: temperatures and at a critical
566: doping level (such as $\mu_P$ in Fig.~\ref{fig:ph4_6})
567: would reveal phase separation and equal probabilities of finding
568: in the ensemble of hole-rich or hole-poor clusters only.
569:
570: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
571:
572:
573:
574: In summary, the above results
575: demonstrate
576: the importance of the {\it many-body
577: interactions} in microscopic clusters.
578: Our exact Hubbard cluster calculations show the existence of
579: charge and spin pairing, electronic phase separation, pseudogaps
580: and condensation and hence demonstrate a rich variety of
581: properties which can be tuned by doping.
582: Furthermore, it is quite surprising to see the number of
583: properties that these
584: exact clusters share with the HTSCs. This may be, at least in part,
585: due to the fact that in all these `bad' metallic high $T_c$
586: materials, short-range correlations play a key role.
587:
588:
589:
590:
591: This research was supported in part by the U.S. Department of
592: Energy under Contract No. DE-AC02-98CH10886.
593: \begin{thebibliography}{0}
594:
595: \bibitem{RVB}
596: P.~W.~Anderson, Science {\bf 235}, 1196 (1987).
597: \bibitem{Anderson} P.~W.~Anderson, Adv.~Rev. {\bf 46}, 3 (1997).
598:
599: \bibitem{Nature}
600: V. Emery and S. A. Kivelson, Nature (London), {\bf 374}, 434
601: (1995).
602: \bibitem{Timusk} T.~Timusk and B.~Statt, Rep.~Prog.~Phys. {\bf 62}, 61
603: (1999).
604: \bibitem{Kivelson_Review} S.~A.~Kivelson {\it et al.}, Rev.~Mod.~Phys.
605: {\bf 75}, 1201 (2003).
606:
607: \bibitem{Marshall} D.~S.~Marshall {\it et al.}, Phys.~Rev.~Lett.
608: {\bf 76}, 4841 (1996).
609:
610: \bibitem{Andrea} Andrea Damascelli, Zahid Hussain,
611: Zhi-Xun Shen, Rev. Mod. Phys. {\bf 75}, 473 (2003).
612:
613: \bibitem{Zha} Y. Zha, V. Barzykin and D. Pines, Phys.~Rev. B{\bf 54}, 7561 (1996).
614:
615: % NEW REFERENCES ON DEXACT CALCULATIONS IN CLUSTERS
616: \bibitem{Pastor} F.
617: Lopez-Urias and G. M. Pastor, Phys.~Rev. B {\bf 59}, 5223 (1999).
618: \bibitem{Sebold} A.~N.~Kocharian and Joel~H.~Sebold, Phys.~Rev. B {\bf 53}, 12804 (1996).
619: \bibitem{scalettar} R.~M. Fye, M.~J. Martins, and R.~T. Scalettar,
620: Phys.~Rev. B {\bf 42}, R6809 (1990).
621: \bibitem{white} Steven R.
622: White, Sudip Chakravarty, Martin P. Gelfand, and Steven A.
623: Kivelson, Phys.~Rev. B {\bf 45}, 5062 (1992).
624: \bibitem{Shiba} H. Shiba and P. A.
625: Pincus, Phys. Rev. B {\bf 5}, 1966 (1972).
626: \bibitem{schumann} R.~Schumann, Ann.~Phys. {\bf 11}, 49 (2002).
627:
628:
629: \bibitem{PRB} A.~N.~Kocharian, G.~W.~Fernando, K. Palandage and
630: J.~W.~Davenport, Phys.~Rev. B {\bf 74}, 024511 (2006).
631:
632: \bibitem{Tranquada} J. M. Tranquada, B. J. Sternlieb, J. D. Axe,
633: Y. Nakamura, S. Uchida, Nature, {\bf 375}, 561 (1995).
634:
635: \bibitem{Tallon} G. V. M. Williams, J.~L.~Tallon and J.~W.~Loram,
636: Phys.~Rev. B {\bf 58}, 15053 (1998).
637:
638: \bibitem{Xiao}
639: Xiao-Jia Chen, Viktor V. Struzhkin, Russell J. Hemley, Ho-kwang
640: Mao, and Chris Kendziora, Phys.~Rev. B{\bf 70}, 214502 (2004).
641:
642: \bibitem{hashini} H. E. Mohottala, B. O. Wells, J. I. Budnick,
643: W. A. Hines, C. Niedermayer,
644: L. Udby, C. Bernard, A. R. Moodenbaugh and Fang-Cheng Chou, Nature
645: (London), {\bf 5}, 377 (2006).
646:
647: \bibitem{nmr} Y. Itoh, M. Matsumara and H. Yamagata, J. Phys. Soc. Jpn,
648: {\bf 66}, 3383 (1997).
649:
650: \bibitem{JMMM} A.~N.~Kocharian, G.~W.~Fernando, K. Palandage and
651: J.~W.~Davenport, J.~Mag.~Mag.~Mater., {\bf 300}, e585 (2006).
652:
653: \bibitem{Chen} X. J. Chen, H. Q. Lin, C. D. Gong, Phys. Rev. Lett.
654: {\bf 85}, 2180 (2000).
655:
656:
657: \end{thebibliography}
658: \end{document}
659: