1:
2: \documentstyle[aps,floats,psfig,epsf,twocolumn]{revtex}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: \draft
5:
6: \begin{document}
7: \title{Antiferromagnetic fluctuations, symmetry and shape of the gap function
8: \\in the electron-doped superconductors: the functional renormalization-group analysis}
9: \author{A. A. Katanin$^{a,b}$ }
10: \address{$^a$Max-Planck-Institut f\"ur Festk\"orperforschung, D-70569 Stuttgart,
11: Germany\\
12: $^b$Institute of Metal Physics, 620219 Ekaterinburg, Russia}
13:
14: \address{~\\
15: \parbox{14cm}{\rm \medskip \vskip0.2cm
16: %%\abstract{
17: The problem of the symmetry of the superconducting
18: pairing and the form of the gap function in the electron-doped
19: superconductors is reconsidered within the temperature-cutoff functional
20: renormalization group approach combined with the Bethe-Salpeter equations.
21: The momentum dependence of the order parameter for antiferromagnetic and superconducting
22: instabilities in these compounds is analyzed. The gap function in the
23: antiferromagnetic (particle-hole) channel has its maxima at the hot-spots,
24: or at the diagonal of the Brilloin zone in their absence. The wavefunction in the singlet
25: superconducting channel is non-monotonic in the vicinity of the ($\pi ,0$)
26: and ($0,\pi $) points, {deviating therefore from the conventional d-wave
27: form} in striking similarity with recent experimental data. An
28: instability in the triplet superconducting channel is much weaker than the
29: singlet one and has an f-wave like form of the gap function.
30: \vskip0.05cm\medskip PACS Numbers: 71.10.Fd;
31: 71.27.+a;74.25.Dw
32: }}
33:
34: \maketitle
35:
36: \tighten
37:
38: It is by now well established that the superconducting order parameter in
39: high-T$_c$ compounds is described by a nearly ($\cos k_x-\cos k_y$)-momentum
40: dependence: its absolute value is largest at the Fermi surface (FS) points close to $(\pi ,0)
41: $ and $(0,\pi )$ and vanishes at the FS crossings on the Brillouin zone (BZ)
42: diagonals.
43:
44: Accurate measurements of the gap function in the hole-doped
45: cuprates found, however, a slight deviation from this d$_{x^2-y^2}$-wave
46: momentum dependence \cite{Mesot}, with a flatter angular dependence near the
47: nodal points. On the other hand, recent experiments on the electron-doped cuprates
48: revealed non-monotonicity of the gap function as a function of an angle
49: around the FS \cite{Bluemberg,Matsui}. In particular, the observed gap function has its maxima away from the points
50: of the Fermi surface which are closest to the $(\pi,0)$ and $(0,\pi)$ points of the
51: Brillouin zone. Since the symmetry and the details of the momentum dependence of the
52: superconducting order parameter are closely related to the structure of the
53: effective pairing interaction between the electrons, the momentum dependence of the gap function contains
54: important information about the pairing mechanism in high-$T_c$ compounds.
55:
56:
57: One of the proposed candidates for the pairing mechanism in cuprates is the antiferromagnetic (AFM) spin fluctuations scenario \cite{Scalapino,BSW,Pines,Chubukov}. While the theoretical
58: studies of the hole-doped compounds found that antiferromagnetic
59: fluctuations lead to a d-wave symmetry of the gap function \cite{SLH,SWhte},
60: there is currently a theoretical controversy about the symmetry of the
61: superconducting pairing in the electron-doped cuprates. As it was argued earlier \cite{Bluemberg,Matsui},
62: with increasing electron doping the hot-spots of the Fermi surface (i.e. the points connected by the
63: antiferromagnetic wave vector $\bf{Q}=(\pi ,\pi )$) move towards each other (see Fig.1a). Since the sign of
64: the d-wave gap is opposite at the hot-spots, this leads to the suppression of the tendency towards the d-wave pairing.
65:
66: In this situation, different
67: types of the pairing symmetry were proposed. An s-wave pairing at large electron dopings
68: was proposed\cite{Abrikosov}, as arising mainly due to the effect of disorder in the model with both short- and
69: long-range Coulomb parts of the interaction. On the other hand, for short-range interactions in the
70: absence of disorder, p-wave superconductivity was considered as a possible candidate
71: in the presence of strong charge fluctuations \cite{Yakovenko}. With these theoretical proposals,
72: it remains an important problem which pairing symmetry is favored at intermediate and large electron
73: dopings and what is the theoretically expected
74: ground state superconducting order parameter for the electron-doped superconductors.
75:
76: To study this problem, we apply in the present paper the functional renormalization-group (fRG) technique \cite
77: {Zanchi,Metzner,SalmHon,SalmHon1,KK}, which proved successful in describing the interplay of antiferromagnetism and
78: d-wave superconductivity in the $t$-$t^{\prime }$ Hubbard model. Although a previous application of this technique
79: to the microscopic analysis of electron-doped superconductors within the Hubbard model found d-wave pairing even at relatively
80: high dopings \cite{Carsten}, the results of this study will be reconsidered in the present paper in two points.
81:
82: First, the momentum-cutoff renormalization group approach of Ref. \cite{Carsten} does not
83: allow to search for the charge and spin instabilities in the forward scattering
84: (zero-momentum transfer) channel. This approach,
85: therefore, may miss the possibility of the triplet pairing, which is often closely related to such instabilities.
86: This drawback is overcome in the
87: temperature-cutoff fRG approach \cite{SalmHon1}, which uses temperature as a natural cutoff parameter
88: and proved successful in describing both AFM and ferromagnetic instabilities together with
89: singlet- and triplet superconducting pairing \cite{SalmHon1,KK}.
90:
91: Second, a recent extension of the temperature cutoff renormalization group approach - its combination with the
92: Bethe-Salpeter (BS) approach - was proposed as a very
93: convenient tool of studing symmetry and the form of the wavefunctions of different order parameters \cite{BS}.
94: Contrary to previous fRG approaches, this method is not based on the knowledge of the wavefunctions at high temperatures.
95: %%, and therefore unbiased with respect to the original choice of the wavefunction symmetry.
96: For the hole-doped superconductors
97: it predicts correctly the flattening of the gap function near the nodal points \cite{BS} in agreement with
98: the experimental observation \cite{Mesot}.
99:
100: In the present paper we use the temperature cutoff functional renormalization group approach in combination
101: with the Bethe-Salpeter equations to extract eigenfunctions and eigenvalues of the effective interaction in the
102: particle-particle (pp) or the particle-hole (ph) channel in the
103: electron-doped superconductors.
104:
105: \begin{figure}[t!]
106: \psfig{file=Fig1.eps,width=90mm,silent=} \vspace{2mm}
107: \caption{(a) The shape of the Fermi-surface (solid line) in the electron-doped
108: superconductors, the dashed line shows the umklapp surface $k_x+k_y=\pi $, black
109: circles mark hot-spots, arrows indicate the direction of movement of the Fermi-surface
110: and hot-spots with increasing doping. (b) The actual non-interacting Fermi surfaces
111: of the model (\ref{H}) at $t^{\prime }=0.3t,$ and $\mu =0.8t$ ($\delta=n-1=0.09$)
112: and $\mu =1.4t$ ($\delta=0.26$). The dots mark the positions of the centers of 12 patches,
113: located in the upper-right quarter of the Brilloin zone; $\alpha $ is an angle
114: corrdinate of the patch centers.}
115: \label{fig:Fig1}
116: \end{figure}
117:
118:
119: {\it The model}. As a convenient model which describes electron-doped superconductors, we
120: consider the 2D $t$-$t^{\prime }$ Hubbard model $H_\mu =H-(\mu -4t^{\prime
121: })N$ with
122: \begin{equation}
123: H=-\sum_{ij\sigma }t_{ij}c_{i\sigma }^{\dagger }c_{j\sigma
124: }+U\sum_in_{i\uparrow }n_{i\downarrow } \label{H}
125: \end{equation}
126: where $t_{ij}=t$ for nearest neighbor (nn) sites $i$ and $j$ and $%
127: t_{ij}=-t^{\prime }$ for next-nn sites ($t,t^{\prime }>0$) on a square
128: lattice. The chemical potential $\mu$ is determined by the filling
129: $n>1$; for convenience we have shifted the chemical potential $\mu $ by $%
130: 4t^{\prime }$.
131:
132: {\it The method}. We follow the many-patch temperature cutoff fRG for one-particle
133: irreducible Green functions as proposed in Ref. \cite{SalmHon1}. This scheme accounts for
134: excitations with momenta far from and close to the FS, which is necessary
135: for the description of instabilities arising from zero-momentum transfer ph
136: scattering. The momentum dependence of the effective interaction
137: between electrons at a given temperature, $V_{T}({\bf k}_1,{\bf k}_2,{\bf k}_3,{\bf k}_4)$ is parametrized
138: by the position of incoming ${\bf k}_1,{\bf k}_2$ and outgoing electron momentum
139: ${\bf k}_3$ at the Fermi surface. The fourth momentum, ${\bf k}_4$ is determined by the
140: momentum conservation law. Neglecting the frequency dependence of the
141: vertices, which is expected to have minor relevance in the weak-coupling
142: regime, the RG differential equation for the interaction vertex has the form
143: \cite{SalmHon1}
144: \begin{eqnarray}
145: \frac{{\rm d}V_T}{{\rm d}T}=-V_T\circ \frac{{\rm d}L_{{\rm pp}}}{{\rm d}T}%
146: \circ V_T+V_T\circ \frac{{\rm d}L_{{\rm ph}}}{{\rm d}T}\circ V_T\,,
147: \label{dV}
148: \end{eqnarray}
149: where $\circ $ is a short notation for summations over intermediate momenta
150: and spin,
151: \begin{equation}
152: L_{\text{ph,pp}}({\bf k},{\bf k}^{\prime })=\frac{f_T(\varepsilon _{{\bf k}%
153: })-f_T(\pm \varepsilon _{{\bf k}^{\prime }})}{\varepsilon _{{\bf k}}\mp
154: \varepsilon _{{\bf k}^{\prime }}}, \label{Lphpp}
155: \end{equation}
156: and $f_T(\varepsilon )$ is the Fermi function. The upper sign in Eq.(\ref
157: {Lphpp}) is for $L_{\text{ph}}$ and the lower sign for $L_{\text{pp}}$, respectively. Eq.(%
158: \ref{dV}) has to be solved with the initial condition $V_{T_0}({\bf k}_1,%
159: {\bf k}_2,{\bf k}_3,{\bf k}_4)=U$; the initial temperature is chosen as
160: large as $T_0=400t$.
161:
162: We discretize the momentum space in $N_p=48$ patches using the same patching
163: scheme as in Ref. \cite{SalmHon1}. This reduces the integro-differential
164: equations (\ref{dV}) to a set of 5824 differential equations,
165: which were solved numerically. The position of the centers of the patches at the
166: Fermi surface for two different filling is shown in Fig. 1b.
167:
168: To perform an analysis of possible instabilities within the fRG+BS approach, we
169: consider the solution of the Bethe-Salpeter equations \cite{BS,BetheSal}
170: \begin{eqnarray}
171: \sum_{{\bf p}}\Gamma _{\text{ph}}^T({\bf k},{\bf p})L_{\text{ph}}(%
172: {\bf p},{\bf p}+{\bf Q})\phi _{{\bf p}}^{\text{ph}} &=\displaystyle{\frac{\lambda _{\text{ph}}\phi _{%
173: {\bf k}}^{\text{ph}}}{1-\lambda _{\text{ph}}}}
174: \,, \nonumber \\
175: -\sum_{{\bf p}}\Gamma _{\text{pp}}^T({\bf k},{\bf p}%
176: )L_{\text{pp}}({\bf p},-{\bf p})\phi _{{\bf p}}^{\text{pp}} &=\displaystyle{\frac{\lambda _{\text{pp}}\phi _{%
177: {\bf k}}^{\text{pp}}}{1-\lambda _{\text{pp}}}}\,, \label{BSirr}
178: \end{eqnarray}
179: The 2-particle reducible vertices $\Gamma _{\text{ph,pp}}^T({\bf k},{\bf k}^{\prime })$ can be directly extracted
180: from the fRG flow according to
181:
182: \begin{equation}
183: \Gamma _{\text{ph,pp}}^T ({\bf k},{\bf k}^{\prime })=\left\{
184: \begin{array}{cl}
185: V_T({\bf k},{\bf k}^{\prime },{\bf k}^{\prime }+{\bf Q}) \text{,\ \ \ ph (AFM)} \\
186: V_T({\bf k},-{\bf k},{\bf k}^{\prime })\pm V_T({\bf k},-{\bf k},-{\bf k}^{\prime }), \\
187: \text{\ \ \ \ \ \ \ \ \ \ \ \ pp (singlet, triplet SC)}
188: \end{array}
189: \right. \label{Gamma}
190: \end{equation}
191:
192: The value $\lambda _{\text{ph,pp}}=1$ corresponds to an ordering instability with the
193: symmetry of the eigenfunction $\phi _{{\bf k}}^{\text{ph,pp}}$. Therefore, tracing
194: the temperature dependence of eigenvalues and -functions allows to identify
195: both, the leading instabilities {\it and} their concomitant order parameter
196: structure. We stop the fRG flow at the temperature $T_X=0.001t$ (the maximum
197: interaction vertex $V_{\max}\equiv\max\{V({\bf k}_1,{\bf k}_2;{\bf k}_3,
198: {\bf k}_4)\}$ at this temperature remains smaller than the bandwidth).
199: We have verified that the results for the eigenfunctions are only weakly dependent
200: on the choice of $T_X$.
201:
202: \begin{figure}[t!]
203: \psfig{file=Fig2.eps,width=90mm,silent=} \vspace{2mm}
204: \caption{Eigenvalues (a) and angular-dependence on the
205: FS of the eigenfunctions $\phi _{{\bf p}}$ of the
206: Bethe-Salpeter equation at $T=T_X$ (b,c,d) for
207: $t^{\prime }=0.3t,$ $U=3.5t,$ $\mu=0.8t$ ($\delta=0.09$).
208: $T_X$ is the lowest temperature reached in the fRG flow.
209: The dots are the value of the eigenfunctions at the centers of corresponding
210: patches, cf. Fig. 1}
211: \label{fig:Fig2}
212: \end{figure}
213:
214: {\it Results.} Below we discuss the results of the numerical solution of the
215: Bethe-Salpeter equations in the electron-doping regime. We choose $t^{\prime
216: }=0.3t$ which is close to typical values considered for the electron-
217: doped cuprates. To remain in the weak-coupling regime, where the
218: considered technique is applicable, we put $U=3.5t.$ Although this value is
219: substantially smaller than that which is expected for the cuprate materials, we show
220: that it allows us to reproduce the main features which are observed experimentally.
221: We also do not expect qualitative change of the results even for higher values of $U.$
222:
223: %At the values of chemical potential $0.55<\mu<0.7t,$ which corresponds to
224: %very low dopings $\delta<0.05$, the leading
225:
226: We start in Fig. 2 with the results for the chemical potential $\mu
227: =0.8t,$ which corresponds to a filling $n=$ $1.09,$ i.e. $9\%$ of the
228: electron doping. At this filling the eigenvalue corresponding to the AFM
229: instability saturates at a value $\lambda<1$ at the smallest temperature $T_X$
230: which one can reach in the fRG flow. Therefore, the AFM is
231: not expected to be the leading ground-state instability
232: (more generally, the AFM is found not to dominate at $\mu>0.7t$, i.e. $\delta>0.05$). The
233: corresponding eigenfunction in the particle-hole channel (Fig. 2b) has its maxima at the hot-spots of the
234: Fermi surface.
235:
236: The eigenvalue $\lambda_{pp}$ in the singlet pairing channel is monotonically increasing with decreasing
237: temperature, and it is
238: biggest close to $T_X$. Therefore, the singlet pairing
239: is expected to be the leading instability at $T\rightarrow 0$. The corresponding wave function
240: is non-monotonic (Fig. 2c) and has a shape which is strikingly
241: similar to the recent experimental data \cite{Bluemberg,Matsui}.
242: The maximum of the wavefunction lies between the 3-rd and 4-th patch of the Fermi surface,
243: i.e. it is shifted from the position of the hot-spot towards
244: the point of the FS closest to the $(\pi ,0)$. The eigenvalue corresponding
245: to the triplet pairing instability is much smaller than for the singlet one,
246: although it increases with decreasing temperature.
247: The corresponding wave function (Fig. 2d) has nodes at the diagonals, and it is maximal near hot spots
248: having therefore f-wave like symmetry.
249:
250: \begin{figure}[t!]
251: \psfig{file=Fig3.eps,width=90mm,silent=} \vspace{2mm}
252: \caption{Same as Fig. 2 for $\mu=1.4t$ ($\delta=0.26$).}
253: \label{fig:Fig3}
254: \end{figure}
255:
256: Now we consider the case of stronger doping $\mu =1.40t,$ i.e. $%
257: n=1.26$ (26\% of the electron doping), see Fig. 3. In this case the hot-spots of the Fermi surface
258: are absent. The eigenvalue corresponding to the AFM instability is decreasing
259: with decreasing temperature at low $T$, so that this instability is again not favored in the ground state.
260: The corresponding eigenfunction has its maximum at the diagonal of the Brillouin zone.
261: Both eigenvalues corresponding to the singlet- and triplet
262: superconducting pairing are smaller than in the case of $\mu =0.80t,$ although the
263: singlet superconductivity remains the leading instability in the $T\rightarrow 0$ limit.
264: The shapes of the wavefunctions
265: for singlet and triplet pairing are similar to those at $\mu =0.80t$.
266: The maximum of the wavefunction in the singlet superconducting channel is further shifted
267: towards the $(\pi ,0)$ point of the Brillouin zone (i.e. in the direction opposite
268: to the hot-spots), while maxima of the eigenfunction in the triplet pairing
269: channel are shifted towards the diagonal.
270:
271: Now we discuss the physical origin of the shape of the gap functions found in the numerical
272: analysis. To this end, we plot the effective pairing interaction $\Gamma _{pp}^T ({\bf k},{\bf k}^{\prime })$
273: in the singlet- and triplet channel as a function of the momenta ${\bf k},{\bf k}^{\prime }$ on
274: the Fermi surface (Fig.4). One can see that at low electron doping ($\mu=0.8t$) the maxima
275: of the attractive interaction in both, the singlet and the triplet channels are located near the hot-spots (Fig. 4a,b), which leads
276: to the maximum of the pairing gap near these points. The difference in 1 patch between the position of the hot-spots
277: and the maxima of the gaps can be explained in this case by small incommensurability of spin fluctuations and the contribution of
278: the other channels of electronic scattering. With increasing doping the maximum of the interaction in the singlet channel spreads in a broader momentum range
279: (Fig. 4c), which, however, leads to almost the same position of the maxima of the gap, as for small electron doping.
280: The maximum of the attraction in the triplet channel at the points with
281: %%${\bf k}-{\bf k}^{\prime }\approx {\bf Q}$
282: ${\bf k}_{x,y}=-{\bf k}_{y,x}^{\prime }$ shifts
283: towards the diagonal of the Brillouin zone, where it is compensated, however, by the strong repulsion, which arises at ${\bf k}={\bf k^{\prime}}$ (Fig. 4d).
284: As a ``compromise", the maxima of the gap are located in the parts of the momentum space where neither repulsive nor attractive
285: interaction is strong, i.e. again remain almost unchanged with respect to small doping.
286:
287: \begin{figure}[t!]
288: \psfig{file=Fig4.eps,width=90mm,silent=} \vspace{2mm}
289: \caption{Contour plots of the pairing interaction $\Gamma _{pp}^T ({\bf k},{\bf k}^{\prime })$
290: in singlet (a,c) and
291: triplet (b,d) channels at $t^{\prime }=0.3t,$ $U=3.5t,$ $\mu=0.8t$ (a,b)
292: and $\mu=1.4t$ (c,d) as a function of the position of the momenta ${\bf k},{\bf k}^{\prime }$ at the Fermi surface
293: (labeled by the number of patch). Black color corresponds to the attractive (negative),
294: white - to the repulsive (positive) interaction. White contours show regions
295: with the strongest attraction, dashed lines mark the points with ${\bf k}_{x,y}=-{\bf k}_{y,x}^{\prime }$.
296: }
297: \label{fig:Fig4}
298: \end{figure}
299:
300: Therefore, while at small doping the shape of the gap functions are determined by the spin fluctuations
301: with the wavevector close to $\bf Q$, at larger dopings the spin fluctuations with broader range of momenta
302: start to play an important role. Note that the maximum of the singlet gap is
303: located away from the Fermi surface points closest to $(\pi ,0)$ and $(0, \pi )$ only for the fillings $n>1$;
304: on the hole doped side (i.e. at fillings $n<1$) the nonmonotonicity of the singlet gap function quickly disappears \cite{BS}.
305:
306: In conclusion, we have investigated the symmetry of the leading
307: instabilities and the shape of the corresponding wavefunctions of the
308: 2D electron doped superconductors within the 2D $t$-$t^{\prime }$
309: Hubbard model using as a novel tool the combination of the Bethe-Salpeter
310: equation and the fRG approach. At $U=3.5t$ we have found the $d$-wave pairing
311: instability to be the strongest instability for dopings $\delta > 0.05$
312: (one can expect that the range of dopings where the AFM instability is the leading one,
313: increases further with increasing interaction strength). The angular dependence of the singlet pairing
314: gap function is found to be non-monotonic and its shape is strikingly similar to the
315: recent experimental data. The position of the maximum of the gap is close to the position
316: of the hot spots at low doping level and deviates from it at higher dopings, where
317: the antiferromagnetic fluctuations with wavevectors with broad momenta range become
318: essential. The triplet instability has subleading eigenvalues and f-wave like
319: form of the wavefunction. The maxima of the wavefunction are close to the hot-spots
320: at low electron dopings and are close to the diagonal for higher dopings.
321:
322: I am grateful to V. M. Yakovenko for stimulating discussions and pointing my attention
323: to the problem of electron-doped cuprates and to the ICCMP (Brasilia) where
324: this work was initiated.
325:
326: \begin{references}
327: \bibitem{Mesot} J. Mesot et al., Phys. Rev. Lett. {\bf 83}, 840 (1999).
328:
329: \bibitem{Bluemberg} G. Bluemberg et al., Phys. Rev. Lett. {\bf 88}, 107002
330: (2002).
331:
332: \bibitem{Matsui} H. Matsui et al., Phys. Rev. Lett. {\bf 95}, 017003 (2005).
333:
334: \bibitem{Scalapino} D. J. Scalapino, J. Low Temp. Phys. {\bf 117,} 179
335: (1999).
336:
337: \bibitem{BSW} N. E. Bickers, et al., Phys. Rev. Lett. {\bf 62,} 961 (1989).
338:
339: %\bibitem{Zhang} S. C. Zhang, Science {\bf 275,} 1089 (1997); E. Demler and
340: %S. C. Zhang, Nature (London) {\bf 396,} 733 (1998).
341:
342: \bibitem{Pines} J. Schmalian, et al., Phys. Rev. Lett. {\bf 80,} 3839 (1998).
343:
344: \bibitem{Chubukov} A. Chubukov and D. Morr, Phys. Rep. {\bf 288}, 355
345: (1997); A. Abanov et al., Phys. Rev. Lett. {\bf 84}, 5608 (2000).
346:
347: \bibitem{SLH} D. J. Scalapino, et al., Phys. Rev. B {\bf 34}, 8190 (1986);
348: ibid. {\bf 35}, 6694 (1987).
349:
350: \bibitem{SWhte} S. R. White et al., Phys. Rev. B {\bf 39}, 839 (1989);
351: N. Bulut et al., Phys. Rev. B {\bf 47}, 2742 (1993)
352:
353: \bibitem{Abrikosov} A. A. Abrikosov, Phys. Rev. B {\bf 53}, R8910 (1996);
354: ibid. {\bf 52}, R15738 (1995)
355:
356: \bibitem{Yakovenko} V. A. Khodel, et. al., Phys. Rev. B {\bf 69}, 144501
357: (2004).
358:
359: %\bibitem{ARPES} A. Ino, C. Kim, M. Nakamura, T. Yoshida, T. Mizokawa, A.
360: %Fujimori, Z.-X. Shen, T. Kakeshita, H. Eisaki, and S. Uchida, Phys. Rev. B
361: %{\bf 65}, 094504 (2002).
362:
363: %\bibitem{ARPES1} P.V. Bogdanov, A. Lanzara, X.J. Zhou, S.A. Kellar, D.L.
364: %Feng, E.D. Lu, H. Eisaki, J.-I. Shimoyama, K. Kishio, Z. Hussain, and Z. X.
365: %Shen, Phys. Rev. B {\bf 64}, 180505 (2001).
366:
367: %\bibitem{ARPES2} D.L. Feng, C. Kim, H. Eisaki, D.H. Lu, A. Damascelli, K.M.
368: %Shen, F. Ronning, N.P. Armitage, N. Kaneko, M. Greven, J. Shimoyama, K.
369: %Kishio, R. Yoshizaki, G.D. Gu, and Z.-X. Shen, Phys. Rev. B {\bf 65},
370: %220501(R) (2002).
371:
372: %\bibitem{Damascelli} A. Damascelli, D. H. Lu, K. M. Shen, N. P. Armitage,
373: %F. Ronning, D. L. Feng, C. Kim, Z.-X. Shen, T. Kimura, Y. Tokura, T.
374: %Tsukuba, Q. Mao, and Y. Maeno, Phys. Rev. Lett. {\bf 85}, 5194 (2000).
375:
376: %\bibitem{SrBand} T. Oguchi, Phys. Rev. B {\bf 51}, 1385 (1995); D. J.
377: %Singh, Phys. Rev. B {\bf 52}, 1358 (1995).
378:
379: %\bibitem{KL} W. Kohn and J. M. Luttinger, Phys. Rev. Lett. {\bf 15}, 524
380: %(1965).
381:
382: %%\bibitem{Santos} R. R. Santos, Phys. Rev. B {\bf 39}, 7259 (1989).
383:
384: \bibitem{Zanchi} D. Zanchi and H.J. Schulz, Phys. Rev. B {\bf 54}, 9509
385: (1996); ibid. {\bf 61}, 13609 (2000).
386:
387: \bibitem{Metzner} C. J. Halboth and W. Metzner, Phys. Rev. B {\bf 61}, 7364
388: (2000); Phys. Rev. Lett. {\bf 85}, 5162 (2000).
389:
390: \bibitem{SalmHon} C. Honerkamp, et al., Phys. Rev. B {\bf 63}, 035109 (2001).
391:
392: \bibitem{SalmHon1} C. Honerkamp and M. Salmhofer, Phys. Rev. Lett. {\bf 87}%
393: , 187004 (2001); Phys. Rev. B{\bf \ 64}, 184516 (2001).
394:
395: \bibitem{KK} A. A. Katanin and A. P. Kampf, Phys. Rev. B {\bf 68}, 195101
396: (2003); ibid. {\bf 67}, 125104 (2003).
397:
398: %\bibitem{Neumayr} See also the renormalized perturbation theory analysis of
399: %A. Neumayr and W. Metzner, Phys. Rev. B {\bf 67}, 035112 (2003).
400:
401: \bibitem{Carsten} C. Honerkamp, Eur. Phys. J. B {\bf 21}, 81 (2001).
402:
403: \bibitem{BS} A. A. Katanin and A. P. Kampf, cond-mat/0408246 (unpublished);
404: Phys. Rev. B. in press.
405:
406: \bibitem{BetheSal} N. Bulut, D. J. Scalapino, and S. R. White, Phys. Rev. B
407: {\bf 47}, 6157 (1993); ibid. {\bf 47}, 14599 (1993).
408:
409: %%\bibitem{SalmHon02} C. Honerkamp, et al., Eur. Phys. J. B {\bf 27}, 127 (2002).
410:
411: \end{references}
412:
413: \end{document}