1: \documentstyle[preprint,floats,aps,psfig,epsfig]{revtex}
2:
3: \begin{document}
4:
5: \title{Instability of the Fermi-liquid fixed point in an extended Kondo model}
6:
7: \author{M. Lavagna$^{1,*}$, A. Jerez$^{2}$ and D. Bensimon$^{1,3}$}
8: \address{$^1$Commissariat \`a l'Energie Atomique, DRFMC/SPSMS,
9: 17, rue des Martyrs, 38054 Grenoble Cedex 9, France}
10: \address{$^2$European Synchrotron Radiation Facility, 6, rue Jules Horowitz,
11: 38043 Grenoble Cedex 9, France}
12: \address{$^3$Department of Appled Physics, Hongo 7-3-1, University of Tokyo,
13: Tokyo 113-8656, Japan}
14:
15: \maketitle
16:
17: \begin{abstract}
18: We study an extended SU(N) single-impurity Kondo model in
19: which the impurity spin is described by a combination of
20: Abrikosov fermions and Schwinger bosons. Our aim is to describe
21: both the quasiparticle-like excitations and the locally critical
22: modes observed in various physical situations, including
23: non-Fermi liquid (NFL) behavior in heavy fermions in the
24: vicinity of a quantum critical point and anomalous transport
25: properties in quantum wires. In contrast with models with
26: either pure bosonic or pure fermionic impurities, the strong
27: coupling fixed point is unstable against the conduction
28: electron kinetic term under certain conditions. The stability
29: region of the strong coupling fixed point coincides with the
30: region where the partially screened, effective impurity repels
31: the electrons on adjacent sites. In the instability region, the
32: impurity tends to attract $(N-1)$ electrons to the neighboring
33: sites, giving rise to a double-stage Kondo effect with additional
34: screening of the impurity.
35: \end{abstract}
36:
37: \bigskip
38:
39: \section{Introduction}
40:
41: One of the most striking properties discovered in Heavy-Fermion systems these
42: last years is the existence of a non-Fermi-liquid behavior
43: \cite{vl96,steglich96,mathur98}
44: in the disordered
45: phase close to the magnetic quantum critical point with a temperature dependence
46: of the physical quantities which differs from that of a standard Fermi-liquid.
47: Recent results obtained in $CeCu_{5.9}Au_{0.1}$ by inelastic neutron scattering
48: experiments \cite{schroder98,schroder00}
49: have shown the existence of anomalous $\omega /T$ scaling law for
50: the dynamical spin susceptibility at the antiferromagnetic wavevector
51: which persists over the entire Brillouin zone. This indicates that
52: the spin dynamics are critical not only at large length scales as
53: in the itinerant magnetism picture \cite{hertz76,millis93}
54: but also at atomic length scales.
55: It strongly suggests the presence of locally critical modes beyond
56: the standard spin-fluctuation theory. Alternative theories have then been proposed
57: to describe such a local quantum critical point. In this direction,
58: we will mention recent calculations \cite{si01} based on dynamical mean field
59: theory
60: (DMFT) which seem to lead to encouraging results such as the prediction of a scaling law
61: for the dynamical susceptibility.
62: The other approach which has been
63: put forward to describe the local QCP is based on supersymmetric
64: theory \cite{gca92,pl99,cpt1,cpt2}
65: in which the spin is described in a mixed fermionic-bosonic representation.
66: The interest of the supersymmetric approach is to describe the quasiparticles
67: and the local moments on an equal footing through the fermionic and the bosonic
68: part of the spin respectively. It appears to be specially well-indicated in the
69: case of the locally critical scenario in which the magnetic temperature scale
70: $T_{N}$, and the Fermi scale $T_{K}$ (the Kondo temperature) below which the
71: quasiparticles die, vanish at the same point $\delta _{C}$.
72: %\smallskip
73:
74: Let us now emphasize another aspect in the discussion of the
75: breakdown of the Fermi-liquid theory. It has to do with the question of
76: the instability of the strong coupling (SC) fixed point. A stable SC fixed
77: point is usually associated with a local Fermi-liquid behavior. On the
78: contrary, an instability of the SC fixed point is an indication for the
79: existence of an intermediate coupling fixed point with non-Fermi-liquid
80: behavior. The traditional source of instability of the SC
81: fixed point in the Kondo model is the presence of several channels for the
82: conduction electrons with the existence of two regimes, the underscreened
83: and the overscreened ones with very different associated behavior.
84: In the one-channel antiferromagnetic single-impurity Kondo model,
85: it is well known from
86: renormalization group arguments, that the system flows to a stable
87: SC fixed point \cite{pwa67}
88: with a behavior of the system identified with
89: that of a local Fermi liquid \cite{pn76}. The situation is rather
90: different in the
91: case of the multichannel Kondo
92: model with a number $K$ of channels for the conduction electrons ($K>1$)
93: \cite{nb}.
94: In the underscreened regime when $K<2S$ (where $S$ is the value of the
95: spin in $SU(2)$), the
96: SC fixed point is stable. It becomes unstable in the other regime $K>2S$
97: named as the overscreened regime. The underscreened regime
98: corresponds to the one-stage Kondo effect with the formation of an effective
99: spin $(S-1/2)$ resulting from the screening of the impurity spin by the
100: conduction electrons located on the same site. The system described by the
101: SC fixed point behaves as a local Fermi liquid. The
102: instability of the SC fixed point in the overscreened
103: regime is associated with a multi-stage Kondo effect in which successively
104: the impurity spin is screened by conduction electrons on the same site, and
105: then the dressed impurity is screened by conduction electrons on the
106: neighboring site and so forth.
107: The overscreened regime leads to the existence of an intermediate
108: coupling fixed point with non-Fermi-liquid excitation spectrum and an anomalous residual
109: entropy at zero temperature.
110: It has been recently put forward \cite{cpt1,cpt2} that other sources of instability
111: of the SC
112: fixed point may exist else but the multiplicity of the conduction electron
113: channels. Recent works have shown that the presence of a more general Kondo
114: impurity where the spin symmetry is extended from $SU(2)$ to $SU(N)$,
115: and the impurity has mixed symmetry,
116: %representation is given by a L-shaped Young tableau,
117: may also lead to an
118: unstability of the SC fixed point already in the one-channel case. In the
119: large $N$ limit, Coleman et al. \cite{cpt1} have found that the SC fixed point becomes
120: unstable as soon as the impurity parameter $q$ (defined below)
121: is larger than $N/2$ whatever the value of $2S$ (defined below)
122: is, giving rise to a
123: two-stage Kondo effect. This result opens the route for the existence of an
124: intermediate coupling fixed point with eventually non-Fermi-liquid
125: behavior.
126: %\smallskip
127:
128: It is worth noticing that the supersymmetry theory, or more specifically
129: the taking into account of general Kondo impurities
130: appears to offer valuable insights into the two issues raised by the
131: breakdown of the Fermi liquid theory that we have summarized above, i.e.
132: both the existence of locally critical modes and the question of the
133: instability of the SC fixed point. Somehow it seems that the consideration
134: of general Kondo impurities captures the physics present in real systems
135: with the coexistence of the screening of the spin by conduction electrons
136: responsible for the formation of quasiparticles, and the formation of
137: localized magnetic moment that persists and eventually leads to a phase
138: transition as the coupling to other impurities becomes dominant.
139: %\smallskip
140:
141: The aim of this work is to study the $SU(N)$, extended
142: single-impurity Kondo model in the one-channel case \cite{ljb}. We would like to
143: investigate how the system behaves when not only the values of the
144: parameters $(2S,q)$ of the representation vary, but also the number $n_{d}$
145: of conduction electrons on the neighboring site does. We want to discuss
146: the effect of $n_{d}$ on the stability of the SC fixed point and to further
147: understand the nature of the screening with the possibility of achieving
148: either a one-stage, two-stage or multi-stage Kondo effect depending on the
149: regime considered. Implications for the behavior of physical
150: quantities will be given.
151:
152:
153: \section{Extended $SU(N)$ single-impurity Kondo model}
154:
155: We consider a generalized, single-impurity, Kondo model with one channel of
156: conduction electrons and a spin symmetry group extended from $SU(2)$ to
157: $SU(N)$. An impurity spin, $\mathbf{S}$, is placed at the origin (site $0$).
158: We deal with impurities that can be
159: realized by a combination of $2S$ bosonic ($b^\dagger_\alpha$) and
160: $q-1$ fermionic ($f^\dagger_\alpha$) operators.
161: The hamiltonian describing the model is written as
162: \begin{equation}
163: H=\sum_{\mathbf{k},\alpha }\varepsilon _{\mathbf{k}}c_{\mathbf{k},\alpha
164: }^{\dagger }c_{\mathbf{k},\alpha }+J\sum_{A}\mathbf{S}^{A}\sum_{\alpha
165: ,\beta }c_{\alpha }^{\dagger }(0) {\tau}_{\alpha \beta
166: }^{A}c_{\beta }(0)~, \label{ham1}
167: \end{equation}
168: where $c_{\mathbf{k},\alpha }^{\dagger }$ is the creation operator of a
169: conduction electron with momentum $\mathbf{k}$, and $SU(N)$ spin index
170: $\alpha =a,b,...,r_N$, $c_{\alpha }^{\dagger }(0)=\frac{1}{\sqrt{N_{S}}}
171: \sum_{\mathbf{k}}c_{\mathbf{k},\alpha }^{\dagger }$
172: %is the creation
173: %operator of a conduction electron at the origin,
174: , where $N_{S}$ is the number of sites, and
175: ${\tau }_{\alpha\beta}^{A}$
176: ($A=1,\dots,N^{2}-1$) are the generators of the
177: $SU(N)$ group
178: in the fundamental representation,
179: with $Tr[{\tau}^{A}{\tau}^{B}]=\delta _{AB}/2$. The
180: conduction electrons interact with the impurity spin $\mathbf{S}^{A}$
181: ($A=1,\dots,N^{2}-1$), placed at the origin, via Kondo coupling, $J>0$.
182: When the
183: impurity is in the fundamental representation, we recover the
184: Coqblin-Schrieffer model \cite{cs,hew} describing conduction electrons
185: in interaction with
186: an impurity spin of angular momentum $j$,
187: ($N=2j+1$, $a=j,~b=j-1,\dots,r_N=-j$), resulting from the
188: combined spin and orbit exchange scattering.
189:
190: \subsection{Strong-coupling fixed point}
191:
192: We consider the case where $J\rightarrow \infty$ and we can
193: neglect the kinetic energy in Eq. \ref{ham1}. In this limit the model
194: can be solved exactly in terms of the invariants associated to the
195: spin of the electrons and of the impurities
196: \cite{jaz}. The eigenvalues are of the
197: form
198: \begin{eqnarray}
199: E_{SC} = \frac{J}{2} \left[ {\cal C}_2(R_{SC})-{\cal C}_2(I)-{\cal C}_2(Y)
200: \right]
201: \label{ener}
202: \end{eqnarray}
203: where $I$ denotes the impurity spin, $Y$ the spin of the $n_c$
204: conduction electrons coupled to the impurity
205: at the origin, and $R_{SC}$ the spin of the resulting SC state at the
206: impurity site.
207: The quantities ${\cal C}_2$ are the $SU(N)$ generalization of the
208: $SU(2)$ eigenvalues, $S(S+1)$, and can be readily evaluated
209: (for details, see Ref. \cite{ljb}).
210:
211: The ground state corresponds to having $n_c=(N-q)$ electrons at the
212: origin, partially screening the
213: impurity. It can be written explicitely as the action of $(2S-1)$ bosonic
214: operators on a singlet state. For instance the {\it highest spin} state can
215: be written as
216: \begin{equation}
217: |GS\rangle _{\{a\}aa}^{[2S-1]}=\frac{1}{\scriptstyle\sqrt{(2S-1)!}}%
218: (b_{a}^{\dagger })^{2S-1}
219: \left[ \frac{1}{\gamma }{\mathcal A}(b_{i_{1}}^{\dagger
220: }(\prod_{\alpha =i_{2}}^{i_{q}}f_{\alpha }^{\dagger })(\prod_{\beta
221: =i_{q+1}}^{i_{N}}c_{\beta }^{\dagger }))\right]|0\rangle~
222: \end{equation}%
223: with $\gamma \equiv \sqrt{%
224: (2S+N-1)C_{N-1}^{q-1}}$.
225: Here, $c^\dagger_{\alpha} \equiv c^\dagger_\alpha(0)$.
226: We denote the ground state energy by $E_0$.
227:
228: The effect of the kinetic term in Eq. \ref{ham1} is, to lowest order in
229: perturbation theory, to mix the ground state with excited states where the
230: number of electrons changes by one.
231: There are three such states, which we
232: denote by $|GS+1\rangle^S$, $|GS+1\rangle^A$ and $|GS-1\rangle$.
233: The labels S(A) indicate that the additional electron is coupled
234: symmetrically(antisymmetrically) to the ground state.
235: The states are
236: readily obtained by deriving the relevant SU(N) Clebsch-Gordan coefficients
237: \cite{ljb}.
238: %\begin{equation}
239: %|GS+1\rangle _{\{a\}aaa}^{S}=\frac{1}{\Omega}c_{a}^{\dagger
240: %}|GS\rangle _{\{a\}aa}~,~~~\Omega=\sqrt{\frac{2S+q-1}{2S+N-1}}
241: %\end{equation}
242: %\begin{equation}
243: %|GS+1\rangle _{\{a\}aab}^{A}= \frac{1}{\Lambda}\frac{1}{\sqrt{2S}}
244: %\left[ c_{a}^{\dagger }|GS\rangle _{\{a\}ab}~-\sqrt{2S-1}c_{b}^{\dagger
245: %}|GS\rangle _{\{a\}aa}~\right]~,~~~\Lambda = \sqrt{\frac{q-1}{N-1}}
246: %\end{equation}
247: %\begin{equation}
248: %|GS-1\rangle=\frac{1}{\scriptstyle\sqrt{(2S-1)!}}(b_{a}^{\dagger
249: %})^{2S-1}|\Delta ^{\prime }\rangle ,
250: %\end{equation}
251: %with
252: % \begin{equation}
253: %|\Delta ^{\prime }\rangle \equiv \frac{1}{\gamma ^{\prime }}\mathcal{A}%
254: %(b_{i_{1}}^{\dagger }(\prod_{\alpha =i_{2}}^{i_{q}}f_{\alpha }^{\dagger
255: %})(\prod_{\beta =i_{q+1}}^{i_{N-1}}c_{\beta }^{\dagger }))|0\rangle~,~~~
256: %\gamma ^{\prime }\equiv \sqrt{(2S+N-2)C_{N-1}^{q-1}}~.
257: %\end{equation}
258: %The normalization factors, $\Omega$ and $\Lambda$, appear
259: %because the additional electron has to be antisymmetrized with respect the
260: %$N-q$ electrons already present in the ground state.
261:
262:
263: \subsection{Stability of the strong coupling fixed point}
264:
265: In order to better understand the low-energy physics of the system,
266: we should consider the finite Kondo coupling, allowing virtual hopping from
267: and to the impurity site. These processes generate interactions between the
268: composite at site 0 and the conduction electrons on neighboring sites, that
269: can be treated as perturbations of the SC fixed point.
270: The energy shifts due to the perturbation can be reproduced by introducing
271: an interaction between the spin at the impurity site and the spin
272: of the electrons on the neighboring site, with effective coupling $J_{ef\!f}$.
273: Applying
274: an analysis similar to that of Nozi\`eres and Blandin \cite{nb}
275: to the nature of the
276: excitations, we can argue whether or not the SC fixed point
277: remains stable once virtual hopping is allowed.
278:
279: Thus, if the coupling
280: between the effective spin at the impurity site and that of the electrons
281: on site 1 is ferromagnetic we know, from the scaling analysis at weak
282: coupling, that the perturbation is irrelevant, and the low energy physics
283: is described by a SC fixed point. That is, an underscreened,
284: effective
285: impurity weakly coupled to a gas of free electrons with a phase shift
286: indicating that there are already $(N-q)$ electrons screening the original
287: impurity.
288: %When the impurity is realized just by fermionic operators, the phase shift
289: %corresponds to the unitary limit, $\delta=\pi/2$,
290: %for $SU(2)$, and is a function \cite{pgks} of $q/N$ for $SU(N)$,
291: %reaching the unitary limit
292: %for $q=N/2$.
293:
294:
295: If, on the contrary, the effective coupling is antiferromagnetic, the
296: perturbation
297: is relevant, the SC fixed point is unstable and the
298: low-energy physics of the model corresponds to some intermediate
299: coupling fixed point,
300: to be identified.
301:
302: We explicitely calculate the effects of hopping on the
303: SC fixed point to the lowest order in perturbation theory,
304: that is, second order in $t$. We consider
305: the case with an arbitrary number $n_d$ of conduction electrons in site 1
306: generalizing the case $n_d=1$ considered in ref. \cite{cpt2}.
307:
308: Adding $n_d$ electrons on site 1
309: leads to two different states, that we will call symmetric (S) and
310: antisymmetric (A). These states, degenerate in the SC limit,
311: acquire energy shifts, $\Delta E^A_0$ and $\Delta E^S_0$
312: due to the perturbation given, in the large-N limit, by
313: \begin{eqnarray}
314: \Delta E_0^A = -\left(\frac{2t^2}{J}\right)\left[
315: \left(\frac{N-q}{q}\right)
316: -\frac{n_d}{N}\left(\frac{N(N-2q)}{q(N-q)}\right)\right]~,
317: \label{rdea}
318: \end{eqnarray}
319: \begin{eqnarray}
320: \Delta E_0^S-\Delta E_0^A &=&
321: -\left(\frac{2S+n_d-1}{2S+N-q}\right)\left(\frac{2t^2}{JN}\right)
322: \left(\frac{N(N-2q)}{q(N-q)}\right)~,~
323: \label{resu} \\ &=& \frac{J_{ef\!f}}{2}(2S+n_d-1)~. \nonumber
324: \end{eqnarray}
325: Notice that the behavior of both Eq. \ref{rdea} and Eq. \ref{resu} are
326: controlled by the same factor.
327: This result has the immediate following physical consequence.
328: The change of sign of $J_{ef\!f}$ (Fig. \ref{qpure}, Right)
329: -and hence of the stability of the SC
330: fixed point- is directly connected to the change in the behavior of
331: $\Delta E_0^A \sim \Delta E_0^S$ with $n_d$.
332: In particular, when $\Delta E_0^S=\Delta E_0^A$,
333: $\Delta E_0^A=-(2t^2/J)$ independently \footnote{This is true for
334: arbitrary values of $N$, and $2S$.} of $n_d$, $q$ or $2S$.
335: In the regime where the SC fixed point is stable,
336: $q/N < 1/2$, $J_{ef\!f}<0$, the lowest energy corresponds to $n_d= 1$, whereas
337: for $q/N > 1/2$, $J_{ef\!f}>0$, the energy expressed in Eq. (\ref{rdea}) is
338: minimized for $n_d=(N-1)$ (Fig. \ref{qpure}, Left).
339: This is precisely the mechanism behind
340: the two-stage quenching. The accumulation of electrons on site 1 is not related to $J_{ef\!f}$ which is
341: independent of $n_d$, but results from the dependence of $\Delta E_0^A \sim \Delta E_0^S$ with $n_d$.
342:
343: \begin{figure}[ht]
344: %\begin{figure}[p]
345: \centerline{
346: \begin{tabular}{c@{\hspace{1pc}}c}
347: %(a) & (b) \\
348: \includegraphics[width=6.0cm]{e0d.eps} &
349: \includegraphics[width=6.0cm]{ejc.eps}
350: \end{tabular}}
351: \caption {
352: (Left) Leading term in the energy shift, $\Delta E_0^A \sim \Delta E_0^S$, as a function
353: of $q/N$,
354: for $1<n_d<N-1$ (shaded region), and in
355: the limiting cases $n_d/N \ll 1$ (dashed line), and $n_d/N \approx 1$
356: (solid line). Notice that the
357: value at $q/N=1/2$ is equal to $-2t^2/J$, for any $n_d$.
358: (Right) Energy difference, ($\Delta E_0^S-\Delta E_0^A$), as a function of
359: $q/N$, for different values of $2S$ in the large-N limit.
360: }
361: \label{qpure}
362: \end{figure}
363:
364:
365: We finish by making some remarks on the physical properties of the model in the different
366: regimes. As is common to all models with an antiferromagnetic Kondo coupling, there will be
367: a crossover from weak coupling above a given Kondo scale, $T_K$, to a low-energy regime.
368: When the SC fixed point is stable, we should expect for $T\ll T_K$ a weak
369: coupling of the effective impurity at
370: site-0 with the rest of the electrons. The physical properties at low temperature are
371: controlled by the degeneracy of the effective impurity, $d([2S-1])=C^{N-1}_{N+2S-2}$.
372: Thus, we should expect a residual entropy ${\cal S}^i \sim \ln C^{N-1}_{N+2S-2}$ and a
373: Curie susceptibility, $\chi^i \sim C^{N-1}_{N+2S-2}/T$, with logarithmic corrections \cite{jaz,pgks}.
374: This is the result that we would
375: expect for a purely symmetric impurity.
376: Contrary to the purely bosonic case,
377: only $(N-q)$ electrons are allowed at the origin in the SC limit, instead of
378: $(N-1)$. Thus, we would expect to find different results for quantities that involve the
379: scattering phase shift of electrons off the effective impurity.
380:
381: In the $q>N/2$, we do not have access to the intermediate coupling fixed point that determines the low-energy
382: behavior. Nevertheless, it is reasonable to think that there would be a magnetic contribution
383: to the entropy, and a Curie-like contribution to the susceptibility, since the impurity
384: remains unscreened. This behavior is different from that of the multichannel
385: Kondo model, which is also characterized by an intermediate coupling fixed
386: point, but where the impurity magnetic degrees of freedom are completely
387: quenched. The degeneracy of the true ground state is an open question, but we
388: can assume that the entropy will be smaller than that of the SC fixed point \cite{aflw1}.
389: It is in the scattering properties that we might be able to see the anomalous
390: features of this new fixed point more clearly.
391:
392:
393: \section{Conclusions}
394: We have studied a Kondo model where the spin of the impurity has mixed
395: symmetry, as a way of incorporating the phenomena of local moment screening
396: and
397: magnetic correlation that is observed in some heavy fermion compounds.
398: Such model is naturally realized by extending the spin symmetry to
399: $SU(N)$, and it displays a rich phase diagram. We find that as long as
400: the bosonic component of spin is of order $N$, there is a transition
401: around the point where the fermionic
402: component of the impurity is $q=N/2$. At this particular point, the energy
403: shift
404: is, to lowest order in perturbation theory around the
405: SC fixed point, equal to $-2t^2/J$, independent of the
406: impurity parameters, $q$, $S$ and $N$. When $q<N/2$, the low-energy
407: physics corresponds to the SC fixed point. For $q>N/2$
408: the SC fixed point is unstable and anomalous behavior is
409: expected, in particular, the two-stage quenching effect. This phase
410: diagram is not accidental, but is due to the relation of the
411: effective impurity site in the SC regime to the conduction
412: electrons in neighboring sites, as our study of the dependence of the
413: energy shifts on
414: $n_d$ reveals. If $q<N/2$, the energy is minimized when the dressed
415: impurity repeals the electrons on the next site. That is,
416: when $n_d=1$. At $q=N/2$, the energy shift is also independent of $n_d$.
417: Finally, the lowest energy shift for $q>N/q$ corresponds to a maximal
418: $n_d$, indicating the accumulation of electrons leading to the two-stage
419: Kondo quenching.
420:
421: \vspace{0.2in}
422:
423: \begin{acknowledgements}
424:
425: We would like to thank V. Zlatic and A. Hewson for organizing this
426: very interesting meeting. We are most grateful to N. Andrei for
427: continuous encouragement and discussions. We also thank S. Burdin,
428: P. Coleman, Ph. Nozi\`eres, and C. P\'epin for helpful discussions.
429:
430: \end{acknowledgements}
431:
432: \vspace{0.2in}
433:
434: $^*$ Also at the Centre National de la Recherche Scientifique (CNRS).
435:
436: \vfill\eject
437:
438: \begin{references}
439:
440: \bibitem{aflw1}
441: Affleck, I. and A.~W.~W. Ludwig: 1993,
442: \newblock {\em Phys.\ Rev. B} {\bf 48}, 7297.
443:
444: \bibitem{pwa67}
445: Anderson, P.: 1967.
446: \newblock {\em Phys.\ Rev. B} {\bf 164}, 352.
447:
448: \bibitem{cpt1}
449: Coleman, P., C. P\'epin, and A. M. Tsvelik: 2000a,
450: \newblock {\em Phys.\ Rev. B} {\bf 62}, 3852.
451:
452: \bibitem{cpt2}
453: Coleman, P., C. P\'epin, and A.~M. Tsvelik: 2000b,
454: \newblock {\em Nucl.\ Phys. B} {\bf 586}, 641.
455:
456: \bibitem{cs}
457: Coqblin, B. and J.~R. Schrieffer: 1969,
458: \newblock {\em Phys. Rev} {\bf 185}, 847.
459:
460: \bibitem{gca92}
461: Gan, J., P. Coleman, and N. Andrei: 1992,
462: \newblock {\em Phys. Rev. Lett.} {\bf 68}, 3476.
463:
464: \bibitem{hertz76}
465: Hertz, J.: 1976.
466: \newblock {\em Phys.\ Rev. B} {\bf 14}, 1165.
467:
468: \bibitem{hew}
469: Hewson, A.: 1993, {\em The Kondo problem to heavy fermions}.
470: \newblock Cambridge University Press.
471:
472: \bibitem{jaz}
473: Jerez, A., N. Andrei, and G. Zar\'and: 1998,
474: \newblock {\em Phys.\ Rev. B} {\bf 58}, 3814.
475:
476: \bibitem{ljb}
477: Jerez, A., M. Lavagna, and D. Bensimon:
478: \newblock cond-mat/0212429.
479:
480: \bibitem{vl96}
481: {L\"{o}hneysen}, H. v.: 1996.
482: \newblock {\em J.Phys.: Cond. Mat.} {\bf 8}, 9689.
483:
484: \bibitem{mathur98}
485: Mathur, N. {\it et al.}
486: %, F. Grosche, S. Julian, I. Walker, D. Freye, R. Haselwimmer, and G.
487: % Lonzarich
488: : 1998.
489: \newblock {\em Nature} {\bf 394}, 39.
490:
491: \bibitem{millis93}
492: Millis, A.: 1993.
493: \newblock {\em Phys.\ Rev. B} {\bf 48}, 7183.
494:
495: \bibitem{pn76}
496: Nozi\`eres, P.: 1976.
497: \newblock {\em J. Phys. (Paris)} {\bf 37}, C1--271.
498:
499: \bibitem{nb}
500: Nozi\`eres, P. and A. Blandin: 1980,
501: \newblock {\em J. Phys. (Paris)} {\bf 41}, 193.
502:
503: \bibitem{pgks}
504: Parcollet, O., A. Georges, G. Kotliar, and A. Sengupta: 1998,
505: \newblock {\em Phys.\ Rev. B} {\bf 58}, 3794.
506:
507: \bibitem{pl99}
508: P\'epin, C. and M. Lavagna: 1999,
509: \newblock {\em Phys.\ Rev. B} {\bf 19}, 12180.
510:
511: \bibitem{si01}
512: Si, Q., S. Rabello, K. Ingersent and J.L. Smith: 2001,
513: \newblock {\em Nature} {\bf 413}, 804.
514:
515: \bibitem{schroder98}
516: {Schr\"{o}der}, A., G. Aeppli, E. Bucher, R. Ramazashvili, and P. Coleman:
517: 1998,
518: \newblock {\em Phys. Rev. Lett.} {\bf 80}, 5623.
519:
520: \bibitem{schroder00}
521: {Schr\"oder}, A. {\it et al.}
522: %, G. Aeppli, R. Coldea, M. Adams, O. Stockert, H. v.
523: % {L\"ohneysen}, E. Bucher, R. Ramazashvili, and P. Coleman
524: : 2000,
525: \newblock {\em Nature} {\bf 407}, 351.
526:
527: \bibitem{steglich96}
528: Steglich, F. {\it et al.}
529: %,B. Buschinger, P. Gegenwart, M. Lohmann, R. Helfrich, C.
530: % Langhammer, P. Hellmann, L. Donnevert, S. Thomas, A. Link, C. Geiber, M.
531: % Lang, G. Sparn, and W. Assmus
532: : 1996.
533: \newblock {\em J. Phys.: Cond. Mat.} {\bf 8}, 9909.
534:
535: \end{references}
536:
537: \end{document}
538:
539:
540:
541:
542:
543:
544: