1: %\documentclass[prl,floatfix,twocolumn,showpacs,amsmath,amssymb,letter]{revtex4}
2: %\documentclass[aps,onecolumn,showpacs,superscriptaddress]{revtex4}
3: \documentclass[twocolumn,floatfix,showpacs,amsmath,amssymb,aps,prl]{revtex4}
4: %\documentclass[preprint,floatfix,showpacs,amsmath,amssymb,aps]{revtex4}
5: \usepackage{graphicx}% Include figure files
6: \usepackage{dcolumn}% Align table columns on decimal point
7: \usepackage{bm}% bold math
8: \usepackage{epsfig}
9: \def\avg#1{\langle#1\rangle}
10: \def\Re{\rm{Re}}
11: \def\Im{\rm{Im}}
12: \def\be{\begin{equation}} \def\ee{\end{equation}}
13: \def\bea{\begin{eqnarray}} \def\eea{\end{eqnarray}}
14: \def\pp{\parallel}
15:
16:
17:
18: \begin{document}
19: \def \cH{{\cal H }}
20: \def\nd{{^{\vphantom{\dagger}}}}
21: \def\yd{^\dagger}
22: \def \bea{\begin{eqnarray}}
23: \def \eea{\end{eqnarray}}
24:
25: \title{Current Carrying Ground State in a Bi-layer Model}
26:
27: \author{Sylvain Capponi}
28: \affiliation{Laboratoire de Physique Th\'eorique UMR 5152,
29: Universit\'e Paul Sabatier, 118 route de Narbonne, 31062 Toulouse,
30: France}
31: \affiliation{Department of Physics, McCullough Building, Stanford
32: University, Stanford CA~~94305-4045}
33: \author{Congjun Wu}
34: \author{Shou-Cheng Zhang}
35: \affiliation{Department of Physics, McCullough Building, Stanford
36: University, Stanford CA~~94305-4045}
37: \date{\today}
38:
39: \begin{abstract}
40: %PACS numbers: 71.27.+a, 71.10.-w, 71.10.Fd \\ \\
41: Strongly interacting systems have been conjectured to
42: spontaneously develop current carrying ground states under certain
43: conditions. We conclusively demonstrate the existence of a
44: commensurate staggered interlayer current phase in a bi-layer
45: model by using the recently discovered quantum Monte-Carlo
46: algorithm without the sign problem. A pseudospin SU(2) algebra and
47: the corresponding anisotropic spin-1 Heisenberg model are
48: constructed to show the competition among the staggered interlayer
49: current, rung singlet and charge density wave phases.
50: \end{abstract}
51: \pacs{71.10.Fd,71.10.Hf, 71.30.+h, 74.20.Mn } \maketitle
52:
53: Strongly correlated systems can spontaneously break symmetries of
54: the microscopic Hamiltonian. A particularly interesting class of
55: ground states spontaneously break the time reversal symmetry
56: and carry a persistent current in the ground state. Such states
57: are known by different synonyms, {\it e.g.} the orbital
58: antiferromagnetic phase (OAF), the staggered flux (SF) or the
59: D-density wave (DDW) phase. In the context of high Tc
60: superconductivity, these current carrying ground states have been
61: proposed as competing states for the pseudogap
62: phase\cite{AFFLECK1988,HSU1991,WEN1996,VARMA1999,CHAKRAVARTY2001A,schroeter2002}.
63: The SF or the DDW phase has the attractive feature that the nodal
64: quasi-particles have an energy spectrum similar to that of the
65: $d-$wave superconducting state.
66:
67: Whenever new ground states are proposed, it is important to
68: establish for which microscopic Hamiltonian such states are
69: realized. Because of their relative simplicity and availability of
70: reliable analytical and numerical methods, the ladder system has
71: been used as a theoretical laboratory to investigate the DDW
72: phase. Weak coupling bosonization methods combined with the
73: renormalization group (RG) analysis on extended two-leg Hubbard
74: ladders show the existence of commensurate DDW phase at
75: half-filling \cite{FJAERESTAD2002,WU2003A,TSUCHIIZU2002} and
76: incommensurate power law fluctuating DDW order away from
77: half-filling\cite{SCHULZ1996,ORIGNAC1997,WU2003A}. While the DDW
78: state does not appear to be the ground state of the t-J ladder~\cite{SCALAPINO2001,TSUTSUI2002},
79: numerical works using the density matrix renormalization
80: group(DMRG) found commensurate DDW order at
81: half-filling\cite{MARSTON2002} and incommensurate DDW order at low
82: doping\cite{SCHOLLWOCK2003} in a ladder model first proposed by
83: Scalapino, Zhang and Hanke\cite{SCALAPINO1998}. The work of
84: Schollw\"ock {\it et al} has generated significant interest in
85: connection with the DDW proposal for the
86: cuprates\cite{CHAKRAVARTY2001A}.
87:
88:
89: \begin{figure}
90: \includegraphics[width=0.4\textwidth]{Staggered.eps}
91: \caption{\label{fig:current} (a) Sketch of a staggered interlayer
92: current (SIC) phase. For clarity, we do not show the bottom layer
93: current. By conservation, each site acts as a source or drain for
94: the current within the bi-layers. (b) Top view of the the
95: bi-layer. (c) Sketch of the SF or the DDW current pattern for
96: comparison.}
97: \end{figure}
98:
99: To the best of our knowledge, the existence of a current carrying
100: ground state has not been conclusively demonstrated in any higher
101: dimensional models. Following the insights we learned from the 1D
102: systems, we investigate the current carrying ground state in a
103: bi-layer version of the model constructed in Ref.
104: \cite{SCALAPINO1998}. This model was originally constructed and
105: extensively investigated because of the exact $SO(5)$ symmetry
106: when coupling constants satisfy a simple relation, and is commonly
107: referred to as the SZH
108: model\cite{BOUWKNEGT1999,DUFFY1998,LIN1998,FRAHM2001}. Here we
109: show that the recently discovered fermionic quantum Monte Carlo
110: (QMC) algorithm without the sign problem\cite{WU2003} can also be
111: applied to this model at and away from half-filling, for a large set of
112: parameters, including purely repulsive interactions. Using this
113: highly accurate numerical method, we can conclusively demonstrate
114: the existence of a current carrying ground state in this model.
115: The current carrying ground state is illustrated in Fig.
116: \ref{fig:current}, with staggered interlayer currents (SIC)
117: between the bi-layers and alternating source to drain currents
118: within the bi-layers. Viewed from the top of the bi-layers, this
119: current pattern is different from the SF or the DDW current
120: pattern, since it has a $s$-wave symmetry. While the SF or the DDW
121: currents are divergence free within the layer, the SIC current is
122: curl free within the layer. These two flow patterns can be
123: considered as dual to each other in two dimensions. In this paper,
124: we shall first discuss the physics of the SIC phase by mapping
125: onto an effective spin one Heisenberg model, and then proceed with
126: the QMC results.
127:
128: The Hamiltonian for the SZH model~\cite{SCALAPINO1998} generalized
129: straightforwardly to the bi-layer system reads
130: \begin{widetext}
131: \begin{eqnarray}\label{SZHham} H&=&-t_\parallel \sum_{\langle ij\rangle}
132: \big\{c^\dagger_{i\sigma} c_{j\sigma} +d^\dagger_{i\sigma}
133: d_{j,\sigma}+h.c.\big\}-t_\perp \sum_i \Big \{c^\dagger_{i,\sigma}
134: d_{i,\sigma}+h.c.\Big \} -\mu\sum_i \Big \{c^\dagger_{i,\sigma}
135: c_{i,\sigma}+ d^\dagger_{i,\sigma} d_{i,\sigma} \Big \}\nonumber
136: +J\sum_{i} \vec{S}_{i,c}\cdot
137: \vec{S}_{i,d}\\
138: &+& U\sum_{i} (n_{i,\uparrow,c}-1/2)( n_{i,\downarrow,c} -1/2) +
139: (n_{i,\uparrow,d}-1/2) (n_{i,\downarrow,d}-1/2) +V\sum_{i}
140: (n_{i,c}-1)(n_{i,d}-1) ,
141: \end{eqnarray}
142: \end{widetext}
143: where $c$ and $d$ denotes fermionic operators in the upper and the
144: lower layers, respectively, $\sigma$ corresponds to up and down
145: spins. At half-filling, $\mu=0$, and the model is particle-hole
146: symmetric. $t_\parallel=1$ sets the unit of energy. Before
147: discussing the SIC phase, we first discuss some general properties
148: of the SZH model which were not known before. The SZH model was
149: known to have a $SO(5)$ symmetry when $J=4(U+V)$ and $\mu=0$, which
150: unifies antiferromagnetism with
151: superconductivity\cite{SCALAPINO1998}. Remarkably, it also has
152: another $SO(5)$ symmetry when
153: \begin{eqnarray}\label{SO5ph} J=4(U-V),\ \ \ t_\perp=0,
154: \end{eqnarray}
155: valid for all filling factors. In order to distinguish between the
156: two different $SO(5)$ symmetries, we call the former the
157: particle-particle $SO(5)$ symmetry, denoted by $SO(5)_{pp}$, and
158: call the latter the particle-hole $SO(5)$ symmetry, denoted by
159: $SO(5)_{ph}$. The mathematical structure associated with the
160: $SO(5)_{ph}$ algebra, not necessarily the symmetry itself, plays a
161: crucial role in constructing the fermionic QMC algorithm without
162: minus sign problem.
163:
164: We construct a four component fermion field $\Psi=\{c_\sigma,
165: d_\sigma\}$. Using the five Dirac $\Gamma_a$ matrices given in
166: Ref.\cite{WU2003}, we construct the fermion bi-linears \bea
167: n_a=\Psi^\dagger \frac{\Gamma_a}{2}\Psi \hspace{5mm}
168: L_{ab}=\Psi^\dagger \frac{\Gamma_{ab}}{2} \Psi \eea It is
169: straightforward to check that $[H,L_{ab}]=0$ when Eq.
170: (\ref{SO5ph}) is satisfied, thus demonstrating the exact
171: $SO(5)_{ph}$ symmetry. The SZH model can be mapped exactly to the
172: spin $3/2$ Hubbard model\cite{WU2003}, by the identification $
173: c_\uparrow = c_{3/2},\ c_\downarrow = c_{1/2},\ d_\uparrow =
174: c_{-1/2},\ d_\downarrow = c_{-3/2}$, and the $SO(5)_{ph}$ symmetry
175: maps exactly onto the $SO(5)$ symmetry of the spin $3/2$ Hubbard
176: model. Because of the exact mapping from the SZH model to the spin
177: $3/2$ Hubbard model, we are able to use the QMC algorithm
178: discovered in Ref. \cite{WU2003}, which works without the minus
179: sign problem in a large parameter regime.
180:
181: In studying the strong coupling phase diagram, SZH identified the
182: $E_0$ phase where the rung singlet state, depicted in Fig.
183: \ref{fig:rung}b, is the lowest energy state, and the $E_3$ phase,
184: where the CDW states, depicted in Fig. \ref{fig:rung}a and
185: \ref{fig:rung}c are the lowest energy states. The new insight
186: gained from Ref. \cite{MARSTON2002,SCHOLLWOCK2003} reveals that
187: the competition between these two phases could result in the DDW
188: phase. In view of this insight, let us consider the following
189: operators
190: \begin{eqnarray*}
191: n_1(i)&=& i/2\sum_\sigma \big\{c^\dagger_\sigma(i) d_\sigma(i)-
192: d^\dagger_\sigma(i) c_\sigma(i) \big \}
193: ,\nonumber \\
194: n_5(i)&=& 1/2 \sum_\sigma \big \{ c^\dagger_\sigma(i) d_\sigma(i)
195: +d^\dagger_\sigma(i) c_\sigma(i) \big \}, \nonumber \\
196: Q(i) &=& L_{15}=1/2 \sum_\sigma \big \{ c^\dagger_\sigma(i)
197: c_\sigma(i) -d^\dagger_\sigma(i) d_\sigma(i) \big \},
198: \end{eqnarray*}
199: where $\vec{\sigma}$ are the Pauli matrices. These operators
200: describe rung current ($n_1$), rung kinetic energy ($n_5$) and the
201: CDW order parameter ($Q$). These three operators form a
202: pseudo-spin $SU(2)$ algebra which are important for our discussion
203: of the SIC phase.
204:
205: \begin{figure}
206: \centering\epsfig{file=Rung.eps,clip=1,width=6cm,angle=0}
207: \caption{ The double occupancy state a) and c) and the rung
208: singlet (b). a), b), c) are spin SU(2) singlets and form the
209: triplet representation of the pseudospin SU(2) group. }
210: \label{fig:rung}
211: \end{figure}
212:
213:
214: There are 16 states on each rung, including 8 bosonic states with
215: particle number 0, 2 or 4 and eight fermionic states with particle
216: number 1 or 3. We are interested in the three rung states shown in
217: Fig.~\ref{fig:rung}a,b,c, which form a spin-1 representation of
218: the pseudospin SU(2) algebra defined above, with eigenvalues
219: $Q=1,0,-1$. $n_1\pm i n_5$ act as pseudospin raising and lowering
220: operators which connect these three states to each other. At
221: half-filling and under the condition that $\max(U, V-3/4 J)<
222: \min(V+J/4, U+2V, U/2+V)$, these are the three lowest energy
223: states. Furthermore, at $U=V-3/4 J$, these three states are
224: degenerate. In the strong coupling limit, we can construct an
225: effective theory to describe the low energy physics by using a
226: pseudospin-1 antiferromagnetic Heisenberg model
227: \bea\label{exchange} H_{ex}= J_p \sum_{\langle i,j \rangle} \Big
228: \{n_5(i) n_5(j) +n_1(i) n_1(j) +Q(i) Q(j) \Big \}, \eea with $J_p=
229: 2 t_\pp^2 / (V+\frac{3}{4} J) $. Several terms break the
230: pseudospin SU(2) symmetry. The intra-rung hopping $t_\perp$ term
231: acts as an uniform external magnetic field which couples to $n_5$.
232: Also, the deviation of $U$ from $V -3/4 J$ removes the degeneracy
233: between a), c) and b) states. These symmetry breaking terms are
234: described by the on-site part as \bea\label{break} H_{on}&=&
235: \sum_{i} \{-2 t_\perp n_5(i) + \Delta U (Q^2(i)-1/2) \} \eea where
236: $\Delta U= U-(V-3/4 J)$. The nonzero value of $\Delta U$ also
237: gives different corrections to the three exchange terms at the
238: order of $J_p \Delta U /U$. We will neglect these corrections
239: below because the more important symmetry breaking effect from
240: $\Delta U$ has already been taking into account in the on-site
241: part. $H=H_{ex}+H_{on}$ describes a 2D antiferromagnetic spin one
242: Heisenberg model in an uniform magnetic field $t_\perp$, with
243: either easy axis ($\Delta U<0$) or easy plane ($\Delta U>0$)
244: anisotropy.
245:
246:
247: \begin{figure}
248: \centering\epsfig{file=PhaseDiagram.eps,clip=1,width=7cm,angle=0}
249: \caption{\label{PhaseDiagram} Phase diagram in the strong coupling
250: limit. Both SO(5) lines are shown as well as QMC region with no
251: minus-sign problem for any filling (hatched area): $g>0$, $g'>0$
252: and $U_c>0$. There is also another region with $V<0$ (not shown).
253: In the yellow region, the low-energy bosonic states are $a$, $b$
254: and $c$ shown in Fig.~\ref{fig:rung}. This is where we expect the
255: competition between the SIC and the rung-singlet phase. Black dots
256: correspond to models for which we have performed QMC simulations.}
257: \end{figure}
258:
259: For the easy axis case with $\Delta U<0$, the effective
260: Hamiltonian reduces to an Ising model with $Q=\pm 1$ states, in a
261: transverse magnetic field\cite{SCALAPINO1998,TSUCHIIZU2002}. For
262: $t_\perp =0$, and $\Delta U > 0$, the rung singlet state (b) has
263: the lowest energy. However, in this case, there is a competition
264: between the $\Delta U > 0$ term and the Heisenberg exchange term
265: $J_p$. For $\Delta U
266: > zJ_p$, where $z=4$ is the coordination number, the ground
267: state is a featureless Mott insulating state which can be
268: described as a product of the rung singlet state on each site. On
269: the other hand, for $\Delta U < zJ_p$, it is more favorable to
270: form linear combinations between the (a), (b) and (c) states, such
271: that a staggered ground state expectation value of $\langle n_1
272: \rangle$ and $\langle n_5 \rangle$ is spontaneously developed,
273: thus lowering the Heisenberg exchange energy $H_{ex}$. In this
274: case, and for $t_\perp =0$, the pseudo-spin vector can lie along
275: in any direction in the $(n_1,n_5)$ plane. At $\Delta U=0$, a
276: finite value of $t_\perp > 0$ corresponds to a pseudo-spin
277: magnetic field along the $n_5$ direction, which creates an easy
278: plane in the $(n_1, Q)$ space. The antiferromagnetic component of
279: the pseudo-spin moment lies in the $(n_1, Q)$ plane, but the
280: uniform component of the pseudo-spin moment points along the $n_5$
281: direction. The pseudo-spin moment becomes fully polarized when
282: $t_\perp > \frac{z}{2} J_p$, and the antiferromagnetic component
283: vanishes beyond this point. We see that $t_\perp > 0$ favors the
284: $(n_1, Q)$ easy plane while $\Delta U < zJ_p$ favors the $(n_1,
285: n_5)$ easy plane, therefore, when both conditions are satisfied,
286: the intersection between the two easy planes, namely the $n_1$
287: easy axis, is selected. This is exactly the staggered inter-layer
288: current (SIC) order. Combining all these considerations, we can
289: summarize the subtle criteria for the SIC phase as \bea
290: &&V-\frac{3}{4} J < U < \min(V+\frac{J}{4}, 2V),~~ V>0 \nonumber \\
291: && t_\perp < \frac{1}{2} z J_p \sqrt{1-(\Delta U/z
292: J_p)^2},~~\Delta U< z J_p, \label{sicMF} \eea The first two robust
293: conditions ensure that the (a), (b) and (c) states are the lowest
294: and next lowest energy states among the 16 states on the rung,
295: while the last two conditions are the rough mean field estimates
296: discussed above.
297:
298: On Fig.~\ref{PhaseDiagram}, we show some specific regions on the
299: phase diagram, obtained in the strong coupling limit. There are two
300: additional axis for $t_\parallel$ and $t_\perp$. If $t_\parallel$ and/or $t_\perp$
301: gets larger, we can expect some
302: phases to have larger or smaller extension. In the case of
303: ladders, a similar phase diagram has been
304: proposed~\cite{TSUCHIIZU2002,SCHOLLWOCK2003}. In order to obtain
305: significant current correlations, one should be close enough from
306: the line $V=U+3/4 J$ shown on Fig.~\ref{PhaseDiagram} where states
307: a, b and c become degenerate.
308:
309: Now we proceed to discuss the QMC calculation of the SIC phase. We
310: first express the interaction terms of the SZH model as
311: \begin{equation}\label{NLsigma}
312: H_{int}=-g (n_1^2 + n_5^2) - g' (n_2^2+n_3^2+n_4^2) - U_c
313: (n-2)^2,
314: \end{equation}
315: up to a constant term. Here $4U_c=-U-3V+3J/4$, $4g=V-U+3J/4$ and
316: $4g'=U-V+J/4$. The $SO(5)_{ph}$ symmetry is clearly recovered when
317: $g=g'$, i.e., when $U=V+J/4$. We now introduce auxiliary
318: Hubbard-Stratonovich fields to decouple each of the three terms
319: above. Wu, Hu and Zhang\cite{WU2003} have shown that the QMC
320: algorithm is free of the minus sign problem provided all three
321: coefficients, $g$, $g'$ and $U_c$ are positive. It corresponds to
322: a wedge in the phase diagram shown on Fig.~\ref{PhaseDiagram}, and
323: most remarkably, it includes a region with purely repulsive
324: interactions, where $U$, $V$ and $J$ are all positive. A simpler
325: case containing only $n_4^2$ interaction, which explicitly breaks
326: the SU(2) spin rotation invariance, has been studied in another
327: context~\cite{ASSAAD2003}. The ground-state (GS) properties of our
328: model are conveniently studied with the projector auxiliary field
329: QMC algorithm. The basic idea is to apply the operator $\exp
330: (-\theta H)$ to a trial state. When $\theta$ becomes large enough
331: and with a proper normalization, this state converges
332: exponentially to the GS. Details of the algorithm may be found
333: in~\cite{ASSAAD1997}. The Trotter discretization was chosen to be
334: $\Delta \tau=0.1$ but we checked that it does not change the
335: results.
336:
337: \begin{figure}
338: \includegraphics[width=0.8\linewidth]{model8.eps}
339: \caption{\label{model8} Parameters are $t_\perp=0.1$, $U=0$,
340: $V=0.5$, $J=2.0$ and correspond to $g=0.25$, $g'=0$ and $U_c=0$.
341: Scaling of ${\cal J}(\vec{Q})/N$ and $J(L/2,L/2)$ vs $1/L$ showing
342: almost no finite-size effects and proving long-range order in the
343: thermodynamic limit. (Inset shows the convergence of ${\cal
344: J}(\vec{Q})/N$ with the projection parameter $\theta$. Typically
345: the GS value is obtained for $\theta=20$).}
346: \end{figure}
347:
348:
349: %{\it If we have some room left, we can add a few algorithmic
350: %details} Practically, it is more efficient in sampling with
351: %discrete HS transformation using two Ising-like fields $\eta, s$
352: %for each quartic fermion term as in Ref. \cite{ASSAAD1997} \bea
353: %&&e^{\Delta \tau V (n(i,\tau)-2)^2}=\sum_{l,s=\pm1}
354: %{\gamma_l\over4} e^{s\eta_l\sqrt{\Delta\tau V}
355: %(n(i,\tau)-2)}+O(\Delta\tau^4), \nonumber\\
356: %&&e^{\Delta \tau W n_a^2(i,\tau)}=\sum_{l,s=\pm1} {\gamma_l\over4}
357: %e^{s\eta_l\sqrt{\Delta\tau W} n_a(i,\tau)}+O(\Delta\tau^4)
358: %\nonumber, \eea where $\gamma_l=1+{\sqrt{6}\over 3} l,
359: %\eta_l=\sqrt{2(3-\sqrt{6}l)}$.
360:
361: We compute correlations between rung currents $ n_1(\vec{r})$ and
362: perform its Fourier transform
363: \begin{equation}\label{correlator}
364: {\cal J}(\vec{q})=\frac{1}{N}\sum_{\vec{r}} e^{i \vec{q}\cdot
365: \vec{r}} \sum_i \langle n_1(i) n_1(i+\vec{r}) \rangle.
366: \end{equation} The strongest signal
367: in the Fourier transform is found for $\vec{Q}=(\pi,\pi)$,
368: suggesting a staggered current pattern as shown in Fig.
369: \ref{fig:current}. This quantity converges to its GS value as the
370: projector parameter $\theta$ increases as shown in the inset of
371: Fig.~\ref{model8}. In order to obtain information in the
372: thermodynamic limit, one has to make an extrapolation of these GS
373: values with a $1/L$ finite-size scaling, where $L$ is the linear
374: size ($L=4$, 6 and 8 in our simulations).
375:
376: \begin{figure}
377: \includegraphics[width=0.8\linewidth]{noLRO.eps}
378: \caption{\label{noLRO} Finite-size scaling of the current correlations ${\cal
379: J}(\vec{Q})/N$ showing no long-range order in the thermodynamic limit. Parameters are~:
380: (i) same as Fig.~\ref{model8} except for $t_\perp=0.5$ (ii) $U=V=0.3$, $J=1.6$ and
381: $t_\perp=0.5$ at half-filling; (iii) $U=0.75$, $V=0$, $J=1$ and $t_\perp=0$ at 1/8-doping.
382: Typically
383: the GS value is obtained for $\theta=20$.
384: }
385: \end{figure}
386:
387: Following our previous mean-field arguments, in order to prefer a
388: phase with staggered current, we
389: choose $g > g'$ and $U_c=0$, with a small $t_\perp$. As shown on
390: Fig.~\ref{model8} for $U=0$, $V=0.5$ and $J=2$ when $t_\perp$ is
391: small at 0.1, our values are rather constant with size, as expected
392: in an Ising-like phase. Both the largest distance real-space
393: correlations $J(L/2,L/2)$ and the Fourier transform ${\cal
394: J}(\vec{Q})/N$ converge to the same finite value (within our error
395: bars), meaning long-range order in the thermodynamic limit.
396:
397: As expected from our analytical estimates in (\ref{sicMF}), if
398: $\Delta U$ or $t_\perp$ gets too large, long-range order disappear
399: as shown on Fig.~\ref{noLRO}. Since we can also perform the QMC
400: simulation at finite-doping without the sign problem, we have
401: chosen to work at 1/8-doping for some parameters shown on
402: Fig.~\ref{noLRO}. Again, rung-current correlations vanish in the
403: thermodynamic limit since the Fermi surface is not nested anymore.
404: From the analytical estimates based on the mapping to the spin one
405: antiferromagnetic Heisenberg model and the detailed QMC
406: calculations shown above, we can conclusively demonstrate the
407: existence of the SIC phase at half-filling, and also note that
408: this is a rather subtle phase which can be easily destabilized by
409: large $U$ and doping.
410:
411: %\begin{figure}
412: %\includegraphics[width=0.8\linewidth,angle=-90]{model5_tp05.ps}
413: %\caption{\label{model5} Rung current correlations at $(\pi,\pi)$
414: %as a function of inverse effective temperature $\beta$ for various
415: %linear sizes $L$. Parameters are $U=V=0.3$, $J=1.6$, $t_\perp=0.5$
416: %It corresponds to $g=0.15$, $g'=0.05$ and $U_c=0$. No long range
417: %order.}
418: %\end{figure}
419:
420:
421: %\begin{figure}
422: %\includegraphics[width=0.7\linewidth,angle=-90]{model5_tp01.ps}
423: %\caption{\label{model5_tp01} Rung current correlations at
424: %$(\pi,\pi)$ as a function of inverse effective temperature $\beta$
425: %for various linear sizes $L$. Parameters are $U=V=0.3$, $J=1.6$,
426: %$t_\perp=0.1$ It corresponds to $g=0.15$, $g'=0.05$ and $U_c=0$.
427: %No long range order.}
428: %\end{figure}
429:
430: \acknowledgements This work is supported by the NSF under grant
431: numbers DMR-0342832 and the US Department of Energy, Office of
432: Basic Energy Sciences under contract DE-AC03-76SF00515. S.C.
433: thanks IDRIS (Orsay) and SLAC (Stanford) for allocation of
434: CPU-time.
435:
436: %\bibliography{sic}
437: \begin{thebibliography}{21}
438: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
439: \expandafter\ifx\csname bibnamefont\endcsname\relax
440: \def\bibnamefont#1{#1}\fi
441: \expandafter\ifx\csname bibfnamefont\endcsname\relax
442: \def\bibfnamefont#1{#1}\fi
443: \expandafter\ifx\csname citenamefont\endcsname\relax
444: \def\citenamefont#1{#1}\fi
445: \expandafter\ifx\csname url\endcsname\relax
446: \def\url#1{\texttt{#1}}\fi
447: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
448: \providecommand{\bibinfo}[2]{#2}
449: \providecommand{\eprint}[2][]{\url{#2}}
450:
451: \bibitem[{\citenamefont{Affleck and Marston}(1988)}]{AFFLECK1988}
452: \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Affleck}} \bibnamefont{and}
453: \bibinfo{author}{\bibfnamefont{J.~B.} \bibnamefont{Marston}},
454: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{37}},
455: \bibinfo{pages}{3774} (\bibinfo{year}{1988}).
456:
457: \bibitem[{\citenamefont{Hsu et~al.}(1991)\citenamefont{Hsu, Marston, and
458: Affleck}}]{HSU1991}
459: \bibinfo{author}{\bibfnamefont{T.~C.} \bibnamefont{Hsu}},
460: \bibinfo{author}{\bibfnamefont{J.~B.} \bibnamefont{Marston}},
461: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Affleck}},
462: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{43}},
463: \bibinfo{pages}{2866} (\bibinfo{year}{1991}).
464:
465: \bibitem[{\citenamefont{Wen and Lee}(1996)}]{WEN1996}
466: \bibinfo{author}{\bibfnamefont{X.~G.} \bibnamefont{Wen}} \bibnamefont{and}
467: \bibinfo{author}{\bibfnamefont{P.~A.} \bibnamefont{Lee}},
468: \bibinfo{journal}{Physical Review Letters} \textbf{\bibinfo{volume}{76}},
469: \bibinfo{pages}{503} (\bibinfo{year}{1996}).
470:
471: \bibitem[{\citenamefont{Varma}(1999)}]{VARMA1999}
472: \bibinfo{author}{\bibfnamefont{C.~M.} \bibnamefont{Varma}},
473: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{83}},
474: \bibinfo{pages}{3538} (\bibinfo{year}{1999}).
475:
476: \bibitem[{\citenamefont{Chakravarty et~al.}(2001)\citenamefont{Chakravarty,
477: Laughlin, Morr, and Nayak}}]{CHAKRAVARTY2001A}
478: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Chakravarty}},
479: \bibinfo{author}{\bibfnamefont{R.~B.} \bibnamefont{Laughlin}},
480: \bibinfo{author}{\bibfnamefont{D.~K.} \bibnamefont{Morr}}, \bibnamefont{and}
481: \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Nayak}},
482: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{63}},
483: \bibinfo{pages}{94503} (\bibinfo{year}{2001}).
484: %\bibitem{CHAKRAVARTY2001A}{S.~Chakravarty {\it et al.}, Phys. Rev. B {\bf 63}, 94503 (2001).}
485:
486: \bibitem[{\citenamefont{Schroeter and Doniach}(2002)}]{schroeter2002}
487: \bibinfo{author}{\bibfnamefont{D.~F.} \bibnamefont{Schroeter}}
488: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Doniach}},
489: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{66}},
490: \bibinfo{pages}{75120} (\bibinfo{year}{2002}).
491:
492: \bibitem[{\citenamefont{Fj{\ae}restad and Marston}(2002)}]{FJAERESTAD2002}
493: \bibinfo{author}{\bibfnamefont{J.~O.} \bibnamefont{Fj{\ae}restad}}
494: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{J.~B.}
495: \bibnamefont{Marston}}, \bibinfo{journal}{Phys. Rev. B}
496: \textbf{\bibinfo{volume}{65}}, \bibinfo{pages}{125106}
497: (\bibinfo{year}{2002}).
498:
499: \bibitem[{\citenamefont{Wu et~al.}(2003{\natexlab{a}})\citenamefont{Wu, Liu,
500: and Fradkin}}]{WU2003A}
501: \bibinfo{author}{\bibfnamefont{C.~J.} \bibnamefont{Wu}},
502: \bibinfo{author}{\bibfnamefont{W.~V.} \bibnamefont{Liu}}, \bibnamefont{and}
503: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Fradkin}},
504: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{68}},
505: \bibinfo{pages}{115104} (\bibinfo{year}{2003}{\natexlab{a}}).
506:
507: \bibitem[{\citenamefont{Tsuchiizu and Furusaki}(2002)}]{TSUCHIIZU2002}
508: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Tsuchiizu}} \bibnamefont{and}
509: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Furusaki}},
510: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{66}},
511: \bibinfo{pages}{245106} (\bibinfo{year}{2002}).
512:
513: \bibitem[{\citenamefont{Schulz}(1996)}]{SCHULZ1996}
514: \bibinfo{author}{\bibfnamefont{H.~J.} \bibnamefont{Schulz}},
515: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{53}},
516: \bibinfo{pages}{R2959} (\bibinfo{year}{1996}).
517:
518: \bibitem[{\citenamefont{Orignac and Giamarchi}(1997)}]{ORIGNAC1997}
519: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Orignac}} \bibnamefont{and}
520: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Giamarchi}},
521: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{56}},
522: \bibinfo{pages}{7167} (\bibinfo{year}{1997}).
523:
524: \bibitem[{\citenamefont{Scalapino et~al.}(2001)\citenamefont{Scalapino, White, and Affleck}}]{SCALAPINO2001}
525: \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Scalapino}},
526: \bibinfo{author}{\bibfnamefont{S.~R.} \bibnamefont{White}},
527: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Affleck}},
528: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{64}},
529: \bibinfo{pages}{100506} (\bibinfo{year}{2001}).
530:
531: \bibitem[{\citenamefont{Tsutsui et~al.}(2002)\citenamefont{Tsutsui, Poilblanc and Capponi}}]{TSUTSUI2002}
532: \bibinfo{author}{\bibfnamefont{K.} \bibnamefont{Tsutsui}},
533: \bibinfo{author}{\bibfnamefont{D.} \bibnamefont{Poilblanc}},
534: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Capponi}},
535: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{65}},
536: \bibinfo{pages}{020406(R)} (\bibinfo{year}{2002}).
537:
538: \bibitem[{\citenamefont{Marston et~al.}(2002)\citenamefont{Marston,
539: Fj{\ae}restad, and Sudb{\o}}}]{MARSTON2002}
540: \bibinfo{author}{\bibfnamefont{J.~B.} \bibnamefont{Marston}},
541: \bibinfo{author}{\bibfnamefont{J.~O.} \bibnamefont{Fj{\ae}restad}},
542: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Sudb{\o}}},
543: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{89}},
544: \bibinfo{pages}{56404} (\bibinfo{year}{2002}).
545:
546: %\bibitem[{\citenamefont{Schollw\"ock et~al.}(2003)\citenamefont{Schollw\"ock,
547: % Chakravarty, Fj{\ae}restad, Marston, and Troyer}}]{SCHOLLWOCK2003}
548: %\bibinfo{author}{\bibfnamefont{U.}~\bibnamefont{Schollw\"ock}},
549: % \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Chakravarty}},
550: % \bibinfo{author}{\bibfnamefont{J.~O.} \bibnamefont{Fj{\ae}restad}},
551: % \bibinfo{author}{\bibfnamefont{J.~B.} \bibnamefont{Marston}},
552: % \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Troyer}},
553: % \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{90}},
554: % \bibinfo{pages}{186401} (\bibinfo{year}{2003}).
555: \bibitem{SCHOLLWOCK2003}{U.~Schollw\"ock {\it et al.}, Phys. Rev. Lett. {\bf 90}, 186401 (2003).}
556:
557: \bibitem[{\citenamefont{Scalapino et~al.}(1998)\citenamefont{Scalapino, Zhang,
558: and Hanke}}]{SCALAPINO1998}
559: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Scalapino}},
560: \bibinfo{author}{\bibfnamefont{S.~C.} \bibnamefont{Zhang}}, \bibnamefont{and}
561: \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Hanke}},
562: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{58}},
563: \bibinfo{pages}{443} (\bibinfo{year}{1998}).
564:
565: \bibitem[{\citenamefont{Bouwknegt and Schoutens}(1999)}]{BOUWKNEGT1999}
566: \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Bouwknegt}} \bibnamefont{and}
567: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Schoutens}},
568: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{82}},
569: \bibinfo{pages}{2757} (\bibinfo{year}{1999}).
570:
571: \bibitem[{\citenamefont{Duffy et~al.}(1998)\citenamefont{Duffy, Haas, and
572: Kim}}]{DUFFY1998}
573: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Duffy}},
574: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Haas}}, \bibnamefont{and}
575: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Kim}},
576: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{58}},
577: \bibinfo{pages}{R5932} (\bibinfo{year}{1998}).
578:
579: \bibitem[{\citenamefont{Lin et~al.}(1998)\citenamefont{Lin, Balents, and
580: Fisher}}]{LIN1998}
581: \bibinfo{author}{\bibfnamefont{H.~H.} \bibnamefont{Lin}},
582: \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Balents}}, \bibnamefont{and}
583: \bibinfo{author}{\bibfnamefont{M.~P.~A.} \bibnamefont{Fisher}},
584: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{58}},
585: \bibinfo{pages}{1794} (\bibinfo{year}{1998}).
586:
587: \bibitem[{\citenamefont{Frahm and Stahlsmeier}(2001)}]{FRAHM2001}
588: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Frahm}} \bibnamefont{and}
589: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Stahlsmeier}},
590: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{63}},
591: \bibinfo{pages}{125109} (\bibinfo{year}{2001}).
592:
593: \bibitem[{\citenamefont{Wu et~al.}(2003{\natexlab{b}})\citenamefont{Wu, Hu, and
594: Zhang}}]{WU2003}
595: \bibinfo{author}{\bibfnamefont{C.~J.} \bibnamefont{Wu}},
596: \bibinfo{author}{\bibfnamefont{J.~P.} \bibnamefont{Hu}}, \bibnamefont{and}
597: \bibinfo{author}{\bibfnamefont{S.~C.} \bibnamefont{Zhang}},
598: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{91}},
599: \bibinfo{pages}{186402} (\bibinfo{year}{2003}{\natexlab{b}}).
600:
601: %\bibitem[{\citenamefont{Assaad et~al.}(2003)\citenamefont{Assaad, Rousseau,
602: % Hebert, Feldbacher, and Batrouni}}]{ASSAAD2003}
603: %\bibinfo{author}{\bibfnamefont{F.~F.} \bibnamefont{Assaad}},
604: % \bibinfo{author}{\bibfnamefont{V.}~\bibnamefont{Rousseau}},
605: % \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Hebert}},
606: % \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Feldbacher}},
607: % \bibnamefont{and} \bibinfo{author}{\bibfnamefont{G.~G.}
608: % \bibnamefont{Batrouni}}, \bibinfo{journal}{Europhys. Lett.}
609: % \textbf{\bibinfo{volume}{63}}, \bibinfo{pages}{569} (\bibinfo{year}{2003}).
610: \bibitem{ASSAAD2003}{F.~F.~Assaad {\it et al.}, Europhys. Lett. {\bf 63}, 569 (2003).}
611:
612: \bibitem[{\citenamefont{Assaad et~al.}(1997)\citenamefont{Assaad, Imada, and
613: Scalapino}}]{ASSAAD1997}
614: \bibinfo{author}{\bibfnamefont{F.~F.} \bibnamefont{Assaad}},
615: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Imada}}, \bibnamefont{and}
616: \bibinfo{author}{\bibfnamefont{D.~J.} \bibnamefont{Scalapino}},
617: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{56}},
618: \bibinfo{pages}{15001} (\bibinfo{year}{1997}).
619:
620: \end{thebibliography}
621:
622:
623: \end{document}
624: