1: \documentclass[prl,aps,twocolumn,floats,showpacspsfig]{revtex4}
2: %\documentclass[preprint,floats,aps,showpacs]{revtex4}
3: \usepackage{amssymb}
4:
5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
6: \usepackage{epsfig}
7:
8: %\documentclass[preprint,floats,aps,showpacs,epsf]{revtex4}
9:
10:
11: \newcommand{\be}{\begin{equation}}
12: \newcommand{\ee}{\end{equation}}
13: \newcommand{\bea}{\begin{eqnarray}}
14: \newcommand{\eea}{\end{eqnarray}}
15: \newcommand{\rD}{\mbox{D}}
16: \newcommand{\reff}{\mbox{eff}}
17: \newcommand{\rR}{{\rm R}}
18: \newcommand{\rL}{{\rm L}}
19: \newcommand{\p}{\partial}
20: \newcommand{\s}{\sigma}
21: \newcommand{\rF}{{\rm F}}
22: \newcommand{\rf}{{\rm f}}
23: \newcommand{\la}{\langle}
24: \newcommand{\ra}{\rangle}
25: \newcommand{\rd}{\mbox{d}}
26: \newcommand{\ri}{\mbox{i}}
27: \newcommand{\re}{\mbox{e}}
28: \newcommand{\rc}{{\rm c}}
29: \newcommand{\rs}{{\rm s}}
30: \newcommand{\rt}{{\rm t}}
31:
32:
33:
34: \begin{document}
35: %\draft
36: \title{Doped Spin Liquid: Luttinger Sum Rule and Low Temperature Order}
37:
38: \author{ R. M. Konik, T. M. Rice$^{*}$ and A. M. Tsvelik}
39: \affiliation{ Department of Physics, Brookhaven National Laboratory, Upton, NY 11973-5000, USA\\
40: $^{*}$ Institute f\"ur Theoretische Physik, ETH-H\"onggerberg, CH-8093 Z\"urich, Switzerland}
41: \date{\today}
42:
43:
44: \begin{abstract}
45: We analyze a model of two-leg Hubbard ladders weakly coupled by interladder
46: tunneling. At half filling a semimetallic state with small Fermi pockets is induced
47: beyond a threshold tunneling strength. The sign changes in the single electron Green's
48: function relevant for the Luttinger Sum Rule now take place at surfaces with both zeroes
49: and infinities with important consequences for the interpretation of ARPES experiments.
50: Residual interactions between electron and hole-like quasi-particles cause
51: a transition to long range order at low temperatures. The theory can be extended to small
52: doping leading to superconducting order.
53:
54: \end{abstract}
55:
56: \pacs{PACS numbers: 71.10.Pm, 72.80.Sk}
57: \maketitle
58: %\narrowtext
59:
60: \sloppy
61: %\section{Introduction}
62:
63: While the properties of doped spin liquids in more than one dimension are notoriously difficult
64: to analyze, they are nonetheless highly relevant. In one dimension the two leg Hubbard ladder at half
65: filling is the spin liquid epitome and as such ladder systems have attracted strong interest
66: (for an early review see \cite{Dagotto}). Powerful analytic techniques such as bosonization and
67: Bethe ansatz have been applied to single ladders with weak interactions (see \cite{ConTs} and
68: references therein) and have led to comprehensive understanding of both doped and undoped ladders.
69: In this letter we report the extension of these results to higher dimensions through the introduction of a
70: small long range interladder tunneling in an ensemble of uncoupled half-filled Hubbard ladders.
71: Increasing the tunneling amplitude leads to the formation of closed electron and hole Fermi pockets.
72: The Luttinger Sum Rule (LSR) now takes on a novel form with the sign changes in the one electron Green's
73: function appearing both as zeroes and infinities. This result has strong implications for the interpretation
74: of ARPES results on underdoped cuprates. In the pseudogap phase the experiments interpret infinities as a
75: set of disconnected Fermi arcs \cite{marnorm}, but do not (and cannot) observe the zeroes. Lastly we analyze possible
76: instabilities of the carriers in the Fermi pockets using interactions derived from the low energy effective
77: field theory for the ladders.
78:
79: The dynamics of the component half-filled ladders in our ensemble are governed at low energies by an
80: effective field theory. As demonstrated in \cite{so8}, half-filled ladders with generically repulsive interactions
81: experience under renormalization an enhancement in the symmetry of the bare Hubbard lattice
82: Hamiltonian. With this enhancement,
83: the effective field theory for the ladder is the SO(8) Gross-Neveu model, $H^{SO(8)}$ \cite{so8}.
84: We couple the half-filled ladders together by long range single-particle tunneling.
85: The complete Hamiltonian is then
86: \bea
87: \sum_i H^{SO(8)}_i + \sum_{i,i' \atop l,l',n}
88: \!\!t^\perp_{ii'll'}\!\!\int dx (c^\dagger_{nli\s}(x)c_{nl'i'\s}(x) + \mbox{h.c.}),
89: \eea
90: where $c^\dagger_{nli\sigma}$ creates an electron at the $n$-th site
91: on the $l$-th leg ($l=1,2$)
92: of the i-th ladder.
93: By making the hopping amplitude long range, we acquire a small parameter, as was done for a similar
94: model of coupled Hubbard chains in \cite{EsTs}.
95: We assume the following hierarchy of energy scales: $W ~{\rm (bandwidth)} \gg
96: \Delta ~{\rm (spectral~gap)} \gg t^\perp$. The first inequality guarantees that the
97: ladders can be described using the effective field theory.
98:
99: The Gross-Neveu model is exactly solvable for all
100: semi-simple symmetry groups and a great deal is known about its thermodynamics and correlation functions.
101: In the SO(8) case the correlation functions were studied in \cite{KonLud,EsKon}.
102: The spectrum of this model consists of three octets of particles of mass $\Delta$ and a
103: multiplet of 29 excitons with mass $\sqrt 3\Delta$.
104: Two octets consist of quasi-particles of different chirality transforming according to the
105: two irreducible spinor representations of SO(8), while the third octet consists of vector particles.
106: The latter include magnetic excitations as well as
107: the Cooperon (a particle with charge $\pm 2e$).
108: The 16 kink fields, carrying charge, spin, orbit, and parity indices, are direct descendants
109: of the original electron lattice operators on the ladders.
110: According to \cite{KonLud,EsKon},
111: the corresponding single electron Green's function is given by
112: \bea
113: G_{a}^{(0)}(\omega, k_x) = \left\{Z_a\frac{[\omega + \epsilon_a(k_x)]}
114: {\omega^2 - \epsilon_a^2(k_x) - \Delta^2} + G_{a,reg}\right\}, \label{G0}
115: \eea
116: where $a = \s,j$ with $\s$ spin and $j =\pm$ indexing the bonding/antibonding bands.
117: $\epsilon_a(k_x)$ is the bare dispersion in the corresponding band.
118: There are no off diagonal Green's function involving electrons from opposite Fermi points,
119: a result of right and left moving quasi-particles belonging
120: to different irreducible representations of the SO(8) group.
121: As was demonstrated in \cite{EsKon},
122: the incoherent part of the Green's function $G_{a,reg}$ yields a negligible
123: contribution to the spectral weight. Thus the quasi-particle weight satisfies $Z_a \approx 1$.
124:
125: \noindent{\bf RPA Analysis:} We will study the properties of our coupled ladders
126: close to the Mott-Hubbard transition. Our approach follows closely the
127: one developed in \cite{EsTs}.
128: The interladder hopping is diagonal in the bonding/antibonding indices because of the bands' differing
129: Fermi wavevectors.
130: A Random Phase approximation (RPA) (diagrammatically pictured in Fig. 1)
131: yields the following expression for the full 2D Green's function:
132: \bea
133: G^{RPA}_a(k_x,{\bf k_\perp}) = \left\{(G_a^{(0)}(k_x))^{-1} - t_a({\bf k}_{\perp})\right\}^{-1} .\label{G}
134: \eea
135: In our model the interchain tunneling amplitude has strong peaks at ${\bf k}_{\perp} =0, {\bf G}/2$,
136: where ${\bf G}$ is the inverse lattice vector in the direction perpendicular to the chains.
137: (The peak at ${\bf G}/2$ follows from particle-hole symmetry, i.e. $t^\perp(k) = -t^\perp(k+{\bf G}/2)$.)
138: Near these points the following expansion is valid:
139: \begin{equation}
140: t_a({\bf k}_{\perp} + (1 \mp 1){\bf G}/4) = \mp t_{a0}[1 - ({\bf k}_{\perp})^2/\kappa_0^2 + ...],\label{t}
141: \end{equation}
142: where the dots stand for terms of higher order in $|{\bf k}_{\perp}|/\kappa_0$ and $\kappa_0 << G$
143: is the small parameter of the theory.
144: We note that $t_{+0} >0$ (bonding) while $t_{-0} < 0$ (anti-bonding).
145:
146: The quasi-particle spectrum is given by
147: \bea
148: \omega - \epsilon_a(k_x) - \Delta^2(\omega + \epsilon_a(k_x))^{-1} - t_a
149: ({\bf k}_{\perp}) = 0 .\label{secular}
150: \eea
151: At this point we note that the RPA Green's function (\ref{G}) together with (\ref{secular})
152: bears a remarkable resemblance to the
153: single electron Green's function of underdoped cuprates conjectured in \cite{norman}
154: on phenomenological grounds.
155: In both cases the numerator of the self energy is modified. In contrast, for a conventional superconductor
156: $t_a({\bf k_\perp})$ would be expected to modify $\epsilon_a(k_x)$.
157:
158:
159: \begin{figure}
160: \begin{center}
161: \epsfxsize=0.45\textwidth
162: \epsfbox{rpa1.eps}
163: \end{center}
164: \caption{The RPA equation for single particle Green's function (thick line), $G_a^{RPA}$.
165: The double line is the bare
166: hopping amplitude, $t_a$ while the thin line is the bare Green's function $G_0$.}
167: \label{rpa}
168: \end{figure}
169:
170:
171: The dispersion relationships, $E_{a\pm}({\bf k})$, of the quasi-particles are given from Eq. \ref{secular} by
172: \bea
173: E_{a\pm}({\bf k}) = t_a({\bf k_\perp})/2 \pm \sqrt{(\epsilon_a(k_x) + t_a({\bf k_\perp})/2)^2 + \Delta^2}.
174: \eea
175: The FS are determined by solving $E_{a\pm}=0$.
176: Doing so to leading order for ${\bf k_\perp}$ near $0$ and ${\bf G}/2$ yields
177: \bea
178: [2(k_x - k_{Fa})v_F \mp t_{a0}]^2 + 2t_{a0}^2{\bf k_{\perp}}^2/\kappa_0^2 = t^2_{a0} - 4\Delta^2.
179: \eea
180: Gapless excitations, i.e. Fermi surfaces (FS),
181: then only emerge when max$|t_a({\bf k_\perp})| > 2\Delta$.
182: For $\pm t_{a0} < 0$, the FS pockets are electron-like, while for $\pm t_{a0} > 0$, they are hole-like.
183: Now $G^a_{RPA}= (\omega+\epsilon_a({\bf k}))/(\omega-E_{a+}({\bf k}))(\omega-E_{a-}({\bf k}))$.
184: Near the FS of the pockets,
185: $\epsilon_a(k_x) \sim \pm\Delta$ and $E_{a+}({\bf k}) \sim \pm t_{a0}$. The effective quasi-particle weight
186: can then be read off to be
187: $ Z_{RPA} \sim 1/2 $.
188: Thus RPA yields well defined quasi-particles.
189: \begin{figure}
190: %[ht]
191: \begin{center}
192: \epsfxsize=0.35\textwidth
193: \epsfbox{FS.eps}
194: \end{center}
195: \caption{Electron (red and magenta) and hole (green and blue) pockets. The difference in size
196: between pockets formed from bonding and antibonding orbitals originates from possible difference between
197: the corresponding tunneling amplitudes. The dashed lines represent the Luttinger surfaces $\epsilon_{\pm}(k)=0$.
198: The thick dashed red lines are loci of the gap minima, $\epsilon_a(k) = - t_a({\bf k})/2$.}
199: \label{FS}
200: \end{figure}
201:
202: \noindent{\bf Luttinger Sum Rule:}
203: The Green's functions (\ref{G0}) and (\ref{G}) satisfy the Luttinger Sum Rule (LSR) in the form which,
204: though being well known theoretically, has had limited applications.
205: The LSR relates the electron density to the volume in momentum space in which $G(\omega=0,{\bf k}) >0$.
206: This volume is bounded by the surface where $G(\omega=0,{\bf k})$ changes sign \cite{AGD}.
207: The sign change can occur either at an infinity of $G(\omega=0,{\bf k})$ (Fermi surface)
208: or a zero (Luttinger surface).
209: The first possibility is denied for a Mott insulator. For example,
210: the Green's function (\ref{G0}) at $\omega = 0$
211: vanishes at $k_{Fa}$ where $\epsilon_a(k_{Fa}) = 0$, i.e. at momenta where the noninteracting Green's
212: function had infinities. In this way the LSR is satisfied for this nonperturbative ladder ground state.
213: In the presence of interladder tunneling, the Green's function (\ref{G}) continues to have zeroes at $k_{Fa}$
214: independent of the ${\bf k}_{\perp}$ component (see Fig. 2). However, when Fermi surface
215: pockets appear, the Green's
216: function additionally changes sign through the newly formed infinities.
217: Electron like pockets add and hole like pockets subtract from the total electron density,
218: but the LSR remains valid.
219: This example demonstrates, as does the case described in \cite{EsTs},
220: that in doped spin liquids it is generally
221: necessary to determine both the Fermi and Luttinger surfaces in order to obtain the
222: electron density from the LSR.
223: It is important to point out that the Luttinger surface, determined by the zeroes of
224: the Green's function, differs dramatically
225: from the surface of minimum gap (see Fig. 2). The latter is often used in ARPES
226: experiments to extrapolate to an underlying
227: Fermi surface. This however leads to difficulties in the pseudogap phase of
228: underdoped cuprates where the enclosed
229: area manifestly exceeds one electron per unit cell, inconsistent with hole doping (e.g. see \cite{shen}).
230:
231:
232: \begin{figure}
233: \begin{center}
234: \epsfxsize=0.35\textwidth
235: \epsfbox{vertex.eps}
236: \end{center}
237: \caption{Approximation of the four-quasi-particle interaction by the emission of an intermediate vector boson. }
238: \label{43}
239: \end{figure}
240:
241: \noindent{\bf Instabilities and Doping Dependence:}
242: As was demonstrated in \cite{EsTs}, the RPA solution may become unstable at low temperatures.
243: The instability is
244: driven by the residual interactions between Fermi quasi-particles and collective modes of the
245: spin liquid. This interaction
246: may receive added strength from nesting of the Fermi surfaces of particles and holes.
247: To describe the instability one
248: needs to move beyond RPA. We follow here the paper \cite{EsTs} and write down an effective
249: action for quasi-particles
250: interacting with collective excitations. The interaction comes from the diagram depicted
251: on the l.h.s. of Fig.3. This four-point
252: function can be approximated as shown on the r.h.s. of this same figure, leading to
253: the following effective action $(p= (\omega,{\bf k}))$:
254: \bea
255: &&S = \frac{1}{2}\sum_{p,j}A_j(-p)[\omega^2 - (v_F k_x)^2 - \Delta^2]A_j(p) + \nonumber\\
256: && \sum_{p,\alpha}\psi_{\pm}^{\bar\alpha}(-p)G^{-1}_{RPA}(p)\psi_{\pm}^{\alpha}(p) + \\\label{H}
257: && \sum_{q,k} \Gamma(\frac{q}{\Delta},\frac{k}{\Delta})A_j(q)\psi_+^{\alpha}(k)
258: (C\gamma^j)_{\alpha\beta}\psi_-^{\beta}(-k -q) ,\nonumber
259: \eea
260: where all fields are real and $\bar\alpha$ is a charge conjugate of $\alpha$. $A_j$ is the
261: bosonic field transforming according to the vector representation of the SO(8) group, $\psi^{\alpha}_{\pm}$
262: are spinor fields of right and left chirality, and $\gamma^j$ are gamma matrices of the SO(8) group.
263: In principle, there is an interaction within each particle multiplet,
264: but we neglect it by accepting the approximation of Fig. 3.
265: Such interactions lead to the creation of bound states
266: with spectral gap $\sqrt 3\Delta$. We, however, treat these as high energy processes.
267: From general considerations, supported
268: by the calculation that follows, we conclude that $\Gamma \sim \sqrt{v_F}\Delta$.
269:
270: In coupling the ladders together, the SO(8) symmetry is reduced to SO(6)
271: (provided $|t_{+0}|=t_{-0}|$. The quasi-particles
272: that transformed in an 8 dim. representation under SO(8) now are arranged into 4 dim.
273: spinors. These spinors are precisely the same that appear in the SO(5) theory of
274: superconductivity \cite{zhang}. We thus expect the same phenomenology present in SO(5) models
275: to be present here. At half-filling the coupling of the ladders will
276: lead to a spontaneous breaking of the SO(6) symmetry. Possible ordered states
277: include superconductivity (SC), antiferromagnetism (AFM), and a staggered flux phase (SFP). The physics
278: of explicit SO(6) breaking terms will be studied in later work.
279:
280: Moving away from half-filling introduces a nonzero
281: chemical potential $\mu$, and the SO(6) symmetry is reduced down to the SU(2)$\times$U(1). The chemical
282: potential acts on the vector bosons as a ``magnetic field'' moving the Cooperon down in energy.
283: At the same time it partially removes the nesting such that the electron and hole pockets become unequal in
284: size and pockets of one type may even disappear. In this case SC becomes the leading instability.
285:
286: We are able to provide an estimate of the superconducting ordering temperature, $T_c$.
287: The RPA result remains valid at $\mu < \Delta/2$
288: so that the Cooperon has not yet condensed and the ground state of the single ladder remains unchanged.
289: (For $\Delta \approx 2\mu$ where Cooperons do condense, a more sophisticated
290: approach similar to \cite{chubukov} is required.)
291: The dispersion relations
292: are modified to
293: \bea
294: && E_{a}({\bf k}) \approx
295: \frac{(k_x - p_0)^2}{2m_{\parallel}} + \frac{{\bf k}^2_{\perp}}{2m_{\perp}} - \epsilon_F - \mu ;\\
296: && E_{a}({\bf k} + {\bf G}/2)
297: \approx -\frac{(k_x - p_{{\bf G}/2})^2}{2m_{\parallel}} -\frac{{\bf k}^2_{\perp}}{2m_{\perp}} + \epsilon_F - \mu , \nonumber
298: \eea
299: where $\epsilon_F = t_{a0}/4 - \Delta^2/t_{a0}$,
300: $p_{(1\mp 1){\bf G}/4} = k_{Fa} \pm t_{a0}/2v_F$, $m_{\perp} = \kappa_0^2/t_{a0}$, and $m_{\parallel} = t_{a0}/2v_F^2$.
301: In two dimensions the density of states on the Fermi surface
302: is $\rho_F = \frac{\sqrt{m_{\parallel}m_{\perp}}}{\pi} = \frac{\kappa_0}{\sqrt{2}\pi v_F}$.
303: The pairing susceptibility is
304: \bea
305: \chi^{-1} = [(\omega + 2\mu)^2 - (v_Fq_x)^2 - \Delta^2] + \Gamma^2\Pi(\omega,{\bf q}) \label{chi}.
306: \eea
307: Since the interaction decays at high energies at the scale $\Delta$, we can take it as the
308: high energy cut-off
309: in the polarization operator. At $\omega, q =0$ we have
310: $\chi^{-1}(0,0) \approx -\Delta^2 + 4\mu^2 + (v_F\rho_F\Gamma^2/\Delta)\ln(\sqrt{(\epsilon_F + \mu)\Delta}/T)$
311: which determines the mean field temperature of the transition to a superconductor with a stiffness
312: determined by the dopant density:
313: \bea
314: T_c \approx \sqrt{(\epsilon_F(t_{a0}) + \mu)\Delta}\exp\left[-\frac{1 - (2\mu/\Delta)^2}{\mbox{const}\times(\kappa_0 v_F/\Delta)}\right] \label{Tc}.
315: \eea
316: This expression is valid only when $t_{a0}$ exceeds $2\Delta$.
317: As $\mu$ increases so does $T_c$. The $T_c$ of other possible instabilities
318: (AFM and SFP) is found approximately (ignoring the consequences of the destruction of a $(\pi,\pi)$ nesting)
319: by setting $\mu=0$ in Eqn. \ref{Tc}. These instabilities are thus exponential disfavoured.
320:
321: While the expression for $T_c$ is not valid if $\mu > \Delta/2$, we can still ask what
322: happens to the Fermi surface for temperatures above any putative $T_c$.
323: Upon increasing $\mu > \Delta/2$, the single particle gap {\it on the ladder} decreases but never
324: vanishes \cite{KonLud} and so
325: the ladder Greens function retains its zeros. It is these zeros that prevent the RPA
326: electron pockets from merging together (see Fig. 2). We thus do not expect, within the validity of the model,
327: a transition to a large Fermi surface for some $\mu_c$.
328:
329: To provide an estimate of $\Gamma$ entering the expression for $T_c$,
330: we calculate the three-point correlation function
331: in the SO(8)
332: Gross-Neveu model using the formfactor approach.
333: To evaluate this correlator, we insert a resolution of the identity
334: between fields, reducing the correlation function
335: to a sum over matrix elements. Keeping only matrix elements
336: involving single particle states we find upon Fourier transformation,
337: \begin{widetext}
338: \bea
339: && \frac{1}{v_F^2}D_{\alpha\beta}^a(p_1,p_2) = \frac{\la a,-p_1|A^a|0\ra}{E_1(E_1 + \omega_1)}\left[\frac{\la 0|\psi_+^{\alpha}|\bar{\alpha},-p_2\ra \la\bar{\alpha},-p_2|\psi_-^{\beta}|a,-p_1\ra}{E_2(\omega_2 - E_2)} +
340: \frac{\la 0|\psi_-^{\beta}|\bar{\beta},-p_1-p_2\ra\la\bar{\beta},-p_1-p_2|\psi_+^{\alpha}|a,-p_1\ra}{E_{12}(\omega_1+\omega_2 + E_{12})}\right] + \nonumber\\
341: && -\frac{\la 0|A^a|a,p_1\ra}{E_1(\omega_1-E_1)}\left[
342: - \frac{\la a,p_1|\psi_-^{\beta}|\bar{\beta},-p_2\ra \la\bar{\beta},-p_2|\psi_+^{\alpha}|0\ra}{E_2(\omega_2 + E_2)} +
343: \frac{\la a,p_1|\psi_+^{\alpha}|\beta,p_1+p_2\ra\la\beta,p_1+p_2|\psi_-^{\beta}|0\ra}{E_{12})(\omega_1+\omega_2-E_{12})}\right] \label{3}\\
344: && - \frac{\la 0|\psi_+^{\alpha}|\bar{\alpha},p_2\ra\la\bar{\alpha},p_2|A^a|\beta,p_1+p_2\ra\la \beta,p_1+p_2|\psi_-^{\beta}|0\ra}
345: {E_{12}E_2(\omega_2 - E_2)(\omega_1+\omega_2 - E_{12})}
346: + \frac{\la 0|\psi_-^{\beta}|\bar{\beta},-p_1-p_2\ra\la \bar{\beta},-p_1-p_2|A^a|\bar{\alpha},-p_2\ra\la\bar{\alpha},-p_2|\psi_+^{\alpha}|0\ra}
347: {E_{12}E_2(\omega_2+E_2)(\omega_1+\omega_2+E_{12})}.\nonumber
348: \eea
349: \end{widetext}
350: Each state is labeled by its isotopic index and momentum, $p$. Momentum and energy are parameterized
351: in terms of rapidities, $\theta_i$ via $p_i = \Delta\sinh(\theta_i)/v_F$, $E_i=\Delta\cosh(\theta_i)$,
352: and $E_{12}=\sqrt{v_F^2(p_1+p_2)^2+\Delta^2}$.
353: The matrix elements of the Fermi operators are given by \cite{KonLud}:
354: \bea
355: && \la 0|\psi_{\pm}^{\alpha}|\rho, \theta\ra = A\re^{\pm\ri\pi/4}C_{\alpha\rho}\re^{\pm\theta/2}; \\
356: && \la\rho, \theta_1|\psi_{\pm}^{\beta}|a,\theta_2\ra = (C\gamma^aC)_{\alpha\rho}\re^{\pm(\theta_1+\theta_2)/4}g(\theta_1-\theta_2) ;\nonumber \\
357: && g(\theta) = \frac{B}{1/2 - \cosh\theta} \times\nonumber\\
358: &&\exp\left\{\int_0^{\infty}\frac{\rd x\sin^2(x\theta/2\pi)}{x\sinh x\cosh(x/2)}\left[2\cosh(x/6) + \re^{-7x/6}\right]\right\},\nonumber
359: \eea
360: where $C$ is the charge conjugation matrix.
361: $A$ and $B$ are related constants on the order of ${\sim \sqrt{\Delta/v_F}}$.
362: For the Bose operators we have $\la 0|A^a|b,\theta\ra = \la b,\theta|A^a|0\ra = A_B \delta_{ab}, \la\rho|A^a|\eta\ra =0$,
363: where $A_B$ is ${\cal O} (v_F^{-1/2})$.
364: The vertex is then given in terms of the three point function via
365: $\Gamma^a_{\alpha\beta} = 2\pi G_a^{-1}(p_1)G_\alpha^{-1}(p_2)G_\beta^{-1}(-p_1-p_2)D^a_{\alpha\beta}(p_1,p_2)$ where
366: \bea
367: &&D_{\alpha\beta}^a(p_1,p_2) =
368: \frac{A_B A (C\gamma^a)_{\alpha\beta}\re^{-\theta_{12}/4}g(\theta_{12})}{E_1E_2}\times\\&&
369: \hskip -.2in \left[\frac{\re^{-\ri\pi/4}}{(\omega_1 \!-\! E_1)(\omega_2 \!+\! E_2)} \!-\! \frac{\re^{\ri\pi/4}}{(\omega_1 \!+\! E_1)(\omega_2 \!-\! E_2)}
370: \right]\bigg|_{{{\frac{p_1}{\Delta}}=\frac{\sinh(\theta_1)}{v_F}}\atop{-\frac{p_2}{\Delta}=\frac{\sinh(\theta_2)}{v_F}}} \nonumber \\
371: && - ((p_1,p_2,\omega_1,\omega_2) \rightarrow (-p_1,p_1+p_2,-\omega_1,-\omega_1-\omega_2)), \nonumber
372: \eea
373: and the $G's$ are the corresponding non-interacting propagators.
374: %The vertex is antisymmetric with respect to permutation of fermions and the change of their chirality, as it must be.
375: We also note that $g(x) \sim -B\re^{-|x|/4}$ for $|x| >> 1$.
376: As we see, the vertex is a smooth function of momenta and frequencies changing
377: with a characteristic scale $\Delta$ as written in (\ref{H}).
378: This derivation justifies Eqs. (\ref{Tc}).
379:
380: In conclusion, we have constructed a toy model of a doped spin liquid. This model possesses a number of
381: interesting features. Vis-a-vis ARPES measurements, it offers an alternative framework in which to understand the
382: observed arcs in underdoped cuprates \cite{shen,marnorm}: such arcs may be Fermi pockets unresolved
383: by ARPES due to disorder and limited experimental accuracy. It further suggests using an observed line of minimal
384: gap will lead to overestimate of the number of electrons present in a band. Beyond implications for ARPES,
385: the model (at half-filling) possesses an
386: SO(6) symmetry and so encompasses the same phenomenology as SO(5) models of superconductivity including
387: a $\pi$-resonance at energy $2\mu$. Away from half-filling
388: superconductivity is preferred and the model is under sufficient control to
389: provide an estimate for the superconducting $T_c$. Finally we point out above $T_c$ the model predicts
390: the existence of a low lying Cooperon excitation. This excitation's existence and its concomitant
391: near coherent
392: propagation provides a means to understand
393: the giant proximity effect seen in LSCO/LCO/LSCO thin films
394: where the Josephson current was measured through LCO in its normal state \cite{bozovic}.
395:
396: RMK and AMT acknowledge the support
397: from US DOE under contract number DE-AC02 -98 CH 10886. TMR
398: acknowledges hospitality from the Institute for Strongly
399: Correlated and Complex Systems at BNL.
400: \begin{thebibliography}{99}
401: \bibitem{Dagotto} E. Dagotto and T. M. Rice, Science {\bf 271}, 618 (1996).
402: \bibitem{ConTs} D. Controzzi and A. M. Tsvelik, Phys. Rev B {\bf 72}, 035110 (2005).
403: \bibitem{marnorm} D. S. Marshall et al., Phys. Rev. Lett. 76, 4841 (1996);
404: M. R. Norman et al., Nature {\bf 392}, 157 (1998).
405: \bibitem{so8} H. L. Lin, L. Balents and M. P. A. Fisher, Phys. Rev. B{\bf 58}, 1794 (1998).
406: \bibitem{EsTs} F. H. L. Essler and A. M. Tsvelik, Phys. Rev B. {\bf 65}, 115117 (2002); {\it ibid.} {\bf 71}, 195116 (2005).
407: \bibitem{KonLud}R. Konik and A. Ludwig, Phys. Rev. B{\bf 64}, 155112 (2001);
408: R. Konik et al., Phys. Rev. B {\bf 61}, R4983 (2000).
409: \bibitem{EsKon} F. H. L. Essler and R. M. Konik in ``From Fields to Strings: Circumnavigating Theoretical Physics'', ed. by M. Shifman, A. Vainshtein and J. Wheather, World Scientific, Singapore (2005);
410: cond-mat/0412421.
411: \bibitem{norman} M.R. Norman, M. Randeria, H. Ding and J.C. Campuzano, Phys. Rev B{\bf 57}, R11093 (1998).
412: \bibitem{AGD} A. A. Abrikosov, L. P. Gorkov and I. E. Dzyaloshinskii,
413: {\it Methods of Quantum Field Theory in Statistical Physics''},
414: ed. by R. A. Silverman, revised edn. Dover, New York.; see also A. M. Tsvelik,
415: {\it Quantum Field Theory in Condensed Matter Physics''}, CUP, 2003.
416: \bibitem{shen} K. M. Shen et al., Science {\bf 307}, 901 (2005).
417: \bibitem{zhang} S. Rabello et al., Phys. Rev. Let. {\bf 80}, 3586 (1998); S. Zhang, Science {\bf 275}, 1069 (1997).
418: \bibitem{chubukov} A. V. Chubukov and J. Schmalian, cond-mat/0507562.
419: \bibitem{bozovic} I. Bozovic et al., Phys. Rev. Lett. {\bf 93}, 157002 (2004).
420: \end{thebibliography}
421: \end{document}
422:
423:
424:
425:
426: