1: %------- from here, half-1c.tex ------------
2: %------- FINAL VERSION !! ------------------
3: %------- JPSJ style ------
4: \documentclass[twocolumn]{jpsj2}
5: %\documentclass{jpsj2}
6: %%\documentclass[letter]{jpsj2} %% for letters
7: %%\documentclass[shortnote]{jpsj2} %% for short notes
8: %%\documentclass[comment]{jpsj2} %% for comments
9: %%\documentclass[addenda]{jpsj2} %% for addenda
10: %%\documentclass[errata]{jpsj2} %% for errata
11: %%\documentclass[twocolumn,letter]{jpsj2} %% two-column layout
12: %%\documentclass[seceq]{jpsj2} %% It makes equation numbers included within the section number (for regular paper only).
13: %%% The following is the list of packages loaded automatically into this class file.
14: %% amsmath.sty
15: %% amssymb.sty
16: %% graphicx.sty
17: %% overcite.sty
18: %
19:
20:
21: %
22:
23: %
24: %\hyphenation{
25: %salts unit Since eqs e-qua-tions small fact leads
26: %photo-emission dif-fer-ent model Hubbard
27: %}
28: %
29: %\renewcommand\figureheight[1]{\vspace{24pt}\mbox{\rule{0cm}{#1}}}
30: \def\runtitle{
31: Knight Shift and Nuclear Spin Relaxation Rate
32: in a Charge-Ordered State of the One-Dimensional Extended Hubbard Model
33: at Quarter Filling
34: }
35: \def\runauthor{Hideo {\sc Yoshioka}}
36: %%---------- equation --------------------------------
37: %\makeatletter
38: % \@addtoreset{equation}{section}
39: %\makeatother
40: %\def\theequation{\thesection.\arabic{equation}}
41: %%------------------------------------------------------
42: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
43: %\documentstyle[preprint,epsf,eqsecnum,aps]{revtex}
44: %\documentstyle[preprint,eqsecnum,aps]{revtex}
45: %\documentstyle[preprint,aps,epsf]{revtex}
46: %\documentstyle[preprint,aps]{revtex}
47: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
48: \def\d{{\rm d}}
49: \def\lan{\left\langle}
50: \def\ran{\right\rangle}
51: \def\dl{{d \over{dl}} \,}
52: \def\dx{dx\,}
53: \def\e{{\rm e}}
54: \def\virg{\;\;,}
55: \def\point{\;\,.}
56: \def\vf{v_{\rm F}}
57: \def\kf{k_{\rm F}}
58: \def\ef{\epsilon_{\rm F}}
59: \def\ek{\epsilon_{k}}
60: \def\ekq{\epsilon_{k+q}}
61: \def\Ek{E_{k}}
62: \def\Ekq{E_{k+q}}
63: \def\vrho{v_\rho}
64: \def\ggs{\buildrel\textstyle > \over {\hbox{\raise0.2ex\hbox{$\sim$}}}}
65: \def\lls{\buildrel\textstyle < \over {\hbox{\raise0.2ex\hbox{$\sim$}}}}
66: \def\gsim{\,\lower0.75ex\hbox{$\ggs$}\,}
67: \def\lsim{\,\lower0.75ex\hbox{$\lls$}\,}
68: \def\N{\hat{N}}
69: \def\tt{\tilde{t}}
70: \def\im{{\rm i}}
71: \def\ie{{\it i.e.}, }
72: \def\eg{{\it e.g.}, }
73: \def\delx{\partial_x}
74: \def\deltau{\partial_\tau}
75: \def\on{\omega_n}
76: \def\en{\epsilon_n}
77: %----- related to Journal by Hideo Yoshioka 1995/11/3 -----------
78: %-- 1:name of Journal, 2:volume, 3:year, 4:page ---
79: \def\jo #1#2#3#4{#1 {\bf #2} (#3) #4} %For J. Phys. Soc. Jpn.
80: %\def\jo #1#2#3#4{#1 {\bf #2}, #4 (#3)} %For Phys. Rev. B
81: %----- related to Journal by Masahisa Tsuchiizu 1997/10/13 ------
82: %\def\jo #1#2#3#4{#1 {\bf #2} (#3), #4} %For Prog. Theor. Phys.
83: %-------------------------------------------------------
84: \def\PR{Phys.\ Rev.}
85: \def\PRB{Phys.\ Rev.\ B}
86: \def\PRL{Phys.\ Rev.\ Lett.}
87: \def\PRS{Proc.\ Roy.\ Soc.}
88: \def\SSC{Solid State Commun.}
89: \def\APL{Appl.\ Phys.\ Lett.}
90: \def\IJQC{Int.\ J.\ Quantum\ Chem.}
91: \def\JAP{J.\ Appl.\ Phys.}
92: \def\JCP{J.\ Chem.\ Phys.}
93: \def\JPC{J.\ Phys.\ C}
94: \def\JPF{J.\ Phys.\ France}
95: \def\JPIF{J.\ Phys.\ I\ France}
96: \def\JPIVF{J.\ Phys.\ IV\ France}
97: \def\JDP{J.\ de\ Phys.}
98: \def\JPCON{J.\ Phys.\ Condens. Matter}
99: \def\JJAP{Jpn.\ J.\ Appl.\ Phys.}
100: \def\JPSJ{J.\ Phys.\ Soc.\ Jpn.}
101: \def\PSCA{Physica}
102: \def\RMP{Rev.\ Mod.\ Phys.}
103: \def\PTP{Prog.\ Theor.\ Phys.}
104: \def\ZP{Z.\ Phys.}
105: \def\ZPB{Z.\ Phys.\ B}
106: \def\ADV{Adv.\ Phys.}
107: \def\ANP{Ann.\ Phys.}
108: \def\SOV{Sov.\ Phys.\ JETP}
109: \def\SL{JETP\ Lett.}
110: \def\PHB{Physica B}
111: \def\PHC{Physica C}
112: \def\EPL{Europhys.\ Lett}
113: \def\PLA{Phys.\ Lett.\ A}
114: \def\PL{Phys.\ Lett.}
115: \def\JLP{J.\ Low Temp.\ Phys.}
116: \def\IJMPB{Int.\ J.\ Mod.\ Phys.\ B}
117: \def\IJMP{Int.\ J.\ Mod.\ Phys.}
118: \def\SM{Synth.\ Met.}
119: \def\JMP{J.\ Math.\ Phys.}
120: \def\MCLC{Mol.\ Cryst.\ Liq.\ Cryst.}
121: \def\PS{Physica\ Scripta}
122: \def\BCSJ{Bull.\ Chem.\ Soc.\ Jpn.}
123: \def\JPCS{J.\ Phys.\ Chem.\ Solids}
124: \def\CR{Chem.\ Rev.}
125: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
126: %
127: \title
128: {
129: Knight Shift and Nuclear Spin Relaxation Rate
130: in a Charge-Ordered State of the One-Dimensional Extended Hubbard Model
131: at Quarter Filling
132: }
133:
134: \author{
135: Hideo {\sc Yoshioka}\thanks{E-mail: h-yoshi@cc.nara-wu.ac.jp}
136: }
137:
138: \inst{
139: Department of Physics, Nara Women's University, Nara 630-8506
140: }
141:
142:
143: \recdate{\today}
144:
145: \abst{
146: We investigate Knight shift and nuclear spin relaxation rate
147: in a charge ordered state of the one-dimensional extended Hubbard
148: model with a quarter filled band by using RPA around the mean-field
149: solution.
150: It is shown that both quantities show splitting below the critical
151: temperature of the charge order, as is experimentally observed.
152: The relationship between the mount of the splitting in the both quantities and
153: the charge disproportionation rate is discussed.
154: }
155:
156: \kword{extended Hubbard model, quarter filling, charge order,
157: Knight shift, nuclear spin relaxation rate
158: }
159:
160: \begin{document}
161: %\sloppy
162: \maketitle
163:
164:
165: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
166: \section{Introduction}
167: Organic conductors have been known to exhibit a variety of states
168: depending on the temperature, pressure and so on.
169: Such a property is indeed due to the low dimensional electronic structure and
170: the mutual interaction.
171: Thus, the materials are one of the most suitable systems for
172: studying exotic states realized by the interaction between electrons
173: in low dimensional systems.\cite{review}
174:
175: Recently, the remarkable state with inhomogeneous but regular charge
176: arrangement called as the charge ordering (CO) has been found
177: in the organic materials with 2:1 composition ratio $A_2 B$,
178: such as (DI-DCNQI)$_2$Ag,\cite{Hiraki1998PRL,Kanoda1999JPhys,Nogami1999JPhys} (TMTTF)$_2 X$\cite{Chow2000PRL,Monceau2001PRL,Zamborszky2002PRB,Nakamura2003JPSJ} and (BEDT-TTF)$_2X$\cite{Miyagawa2000PRB,Tajima2000PRB,Chiba2001JPCS,Takano2001JPCS,Yamamoto2002PRB,W2003PRB}
179: with $X$ being a monovalent anion.
180: Since the molecule $B$ is ionized as $B^+$ or $B^-$ in order to form a closed shell,
181: the average valence of $A$ becomes $-1/2$ or $+1/2$.
182: As a result, the energy band composed of LUMO or HOMO of the molecule $A$
183: becomes quarter filled in terms of electrons or holes, respectively.
184: Thus, the CO phenomenon is seen in several molecular conductors with the
185: quarter-filled band.
186: Theoretically, the CO phenomenon has been analyzed based on the extended
187: Hubbard model with the parameters reflecting the actual band and crystal
188: structure, and it has been found that the phenomenon is due to the strong mutual
189: interaction between electrons, especially the repulsion between the
190: nearest neighbor molecules.\cite{Seo97JPSJ,Seo00JPSJ,Seo04CR}
191:
192: The nuclear magnetic resonance (NMR) is the powerful experimental technique which
193: can explore the local electronic state.
194: It has been considered to be one of the experimental evidence of the CO
195: that the shift of the resonance frequency shows splitting and/or
196: the relaxation rate of the nuclear spin becomes multi-components.
197: It is because the equivalent molecules in the absence of the CO becomes
198: inequivalent under the CO.
199: However, to the best of our knowledge, no one show theoretically
200: that the Knight shift exhibit separation and the NMR relaxation rate
201: has several components in the presence of CO.
202: It has been found that CO does not always lead to the splitting of
203: the Knight shift and the multi component relaxation rate.
204: In the case of the extended Hubbard model with half filling,
205: the Knight shift show no splitting and
206: the relation rate is only one component even in the presence of CO.\cite{Yoshioka02JPSJ}
207: This result is due to the fact that the particle-hole transformation,
208: under which the Hamiltonian is invariant,
209: exchanges the charge rich site with the poor site, vise versa.
210: In the state with CO at the quarter filling,
211: there is no such a symmetry.
212:
213: In the present paper,
214: we investigate the Knight shift and the nuclear spin relaxation rate of the charge
215: ordered state in the system with the quarter filled band.
216: We use the one-dimensional extended Hubbard model and apply the random
217: phase approximation (RPA) around the mean-filed solution to the
218: Hamiltonian.
219: It has been shown that the splitting of the both quantities occurs
220: below the critical temperature of the charge ordering, $T_{\rm CO}$.
221: The result is different from the case of the half-filling.
222: In addition, the relationship between the amount of the splitting of the
223: both quantities and the charge disproportionation is discussed.
224:
225: The paper is organized as follows.
226: The model we consider and it's solution by the mean-field
227: treatment are shown in \S.2.
228: The formulation for calculating the Knight shift and the nuclear spin
229: relaxation rate based on the path integral method is also given in
230: \S.2.
231: The obtained results are shown in \S.3.
232: \S.4 is devoted to Summary.
233:
234: \section{Model and Formulation}
235: We consider the one-dimensional extended Hubbard model at
236: the quarter filling written by the following Hamiltonian,
237: ${\cal H} = {\cal H}_{\rm k} + {\cal H}_{\rm int}$,
238: %=========== (1) (2) =================================
239: \begin{align}
240: {\cal H}_{\rm k}
241: =& - t \sum_{j,s}
242: \left( c_{j,s}^\dagger c_{j+1,s} + {\rm h.c.}\right),
243: \label{eqn:Hk} \\
244: %
245: {\cal H}_{\rm int} =& \frac{U}{2} \sum_{j,s}
246: n_{j,s} n_{j, -s} +
247: V \sum_{j,s,s'} n_{j,s} n_{j+1,s'},
248: \label{eqn:Hint}
249: \end{align}
250: %================================================================
251: where $t$, $U$ and $V$ are the transfer energy between the
252: nearest-neighbor site, the on-site repulsion and the interaction between
253: the nearest-neighbor site, respectively.
254: Here,
255: $c_{j,s}^\dagger$ denotes the creation
256: operator of the electron at the $j$-th site with spin $s = \pm$,
257: $n_{j,s}=c_{j,s}^\dagger c_{j,s}$ and $n_j = \sum_s n_{j,s}$.
258: The possible ordered states of the model obtained by the standard mean-filed
259: approximation
260: are $4\kf$-CDW (CO),
261: $2\kf$-SDW and the coexistence of both states.
262: The phase diagram at the finite temperature
263: on the $V/t$-$T/t$ plane for $U/t=4.0$ is shown in
264: Fig.\ref{fig:phase},\cite{Tomio01JPCS} where (I), (II) and (III) denote the normal state,
265: $4\kf$-CDW and the coexistent state of $4\kf$-CDW and $2\kf$-SDW,
266: respectively.
267: Note that the pure $2\kf$-SDW state appears for the small $V$ region,
268: which is not shown in Fig.\ref{fig:phase}.
269: With decreasing temperature, the transition from the normal state to the
270: coexistent state occurs for $V<V_c$, whereas the pure $4\kf$-CDW is
271: realized between them for $V>V_c$.
272: Figure \ref{fig:rho} shows the charge disproportionation $\rho$ at $U/t=4.0$
273: as a function of
274: $T/t$ for the several choices of $V$, where the quantity $\rho$ is
275: defined as $\langle n_j \rangle = 1/2 - (-1)^j \rho$.
276: At the transition temperature of the coexistent state,
277: the charge disproportionation shows sudden increase.
278: The transition from the normal state to the pure CO is the second-order.
279: For $V \gg V_c$, the transition to the coexistent state is continuous,
280: whereas the transition become the first order for $V \sim V_c$.\cite{Tomio01JPCS}
281: %----------- Fig.1 -----------
282: \begin{figure}
283: \centerline{\includegraphics[width=7.0truecm]{co-sdw-phase-U4-paper.eps}}
284: \caption{
285: The phase diagram on the plane of $V/t$ and $T/t$ at $U/t=4.0$.
286: Here, (I), (II) and (III) denote the normal state,
287: $4\kf$-CDW and the coexistent state of $4\kf$-CDW and $2\kf$-SDW,
288: respectively.
289: }
290: \label{fig:phase}
291: \end{figure}
292: %----------------------
293: %----------- Fig.2 -----------
294: \begin{figure}
295: \centerline{\includegraphics[width=7.0truecm]{r-T-U4-paper.eps}}
296: \caption{
297: The charge disproportionation $\rho$ as a function of temperature
298: for the several choices of $V/t$ at $U/t = 4.0$.
299: }
300: \label{fig:rho}
301: \end{figure}
302: %----------------------
303: In the following, we assume that
304: the pure $4\kf$-CDW state (II) corresponds to the charge ordered state
305: without magnetic order
306: realized from 5.5 K to 220 K in (DI-DCNQI)$_2$Ag and
307: investigate Knight shift and the nuclear spin
308: relaxation rate in the state (II)
309: by using RPA around the mean-field solution.
310:
311: The fluctuation around the mean-filed solution can be taken into account
312: by the path integral method.
313: The effective action up to the second order of the charge and spin
314: fluctuation, which correspond to the RPA approximation where the
315: coupling between the different modes is neglected,
316: is divided into the spin and
317: the charge part ( see Appendix ).
318: The spin part, $S_{\sigma}$ is written as follows,
319: %=========== (3) ================================================
320: \begin{align}
321: S_\sigma
322: =& \sum_{0 \leq q < Q_0} \sum_{\omega_n}
323: (\delta \sigma (-q, -\im \omega_n ), \delta \sigma (-q + Q_0, -\im \omega_n))
324: \nonumber \\
325: &\times \left(
326: \begin{array}{cc}
327: A_\sigma(q, \im \omega_n)& B_\sigma(q, \im \omega_n) \\
328: B_\sigma(q, \im \omega_n)& A_\sigma(q-Q_0, \im \omega_n)
329: \end{array}
330: \right) \nonumber \\
331: &\times \left(
332: \begin{array}{c}
333: \delta \sigma (q, \im \omega_n ) \\
334: \delta \sigma (q-Q_0, \im \omega_n )
335: \end{array}
336: \right),
337: \label{eqn:s-action}
338: \end{align}
339: %================================================================
340: where $Q_0 = \pi/a$ with $a$ being the lattice spacing and
341: \begin{align}
342: A_\sigma (q, \im \omega_n) &= - \frac{v_s(q)}{N_L} \left\{ 1 + 4 v_s(q) K(q, \im
343: \omega_n) \right\},
344: \label{eqn:AS}
345: \\
346: B_\sigma (q, \im \omega_n) &= - \frac{4 v_s(q) v_s(q-Q_0)}{N_L} H(q, \im
347: \omega_n),
348: \label{eqn:BS}
349: \end{align}
350: with $v_s (q) = - U/4$ and $N_L$ is the total number of the lattice.
351: Here, the quantities $K(q,\im \omega_n)$ and $H(q,\im \omega_n)$
352: are defined as follows,
353: \begin{align}
354: & K(q,\im \omega_n ) \nonumber \\
355: =& - \frac{1}{N_L} \sum_k
356: \Bigg\{
357: \frac{1}{4} \left( 1 +
358: \frac{\epsilon_k}{E_k}\frac{\epsilon_{k+q}}{E_{k+q}} +
359: \frac{\Delta}{E_k}\frac{\Delta}{E_{k+q}} \right) \nonumber \\
360: &\times
361: \Big(
362: \frac{f(E_k - \mu) - f(E_{k+q} - \mu)}{\im \omega_n + E_k - E_{k+q} }
363: \nonumber \\
364: &\hspace{4em}+
365: \frac{f(-E_k - \mu) - f(-E_{k+q} - \mu)}{\im \omega_n - E_k + E_{k+q} }
366: \Big) \nonumber \\
367: &+ \frac{1}{4} \left( 1 -
368: \frac{\epsilon_k}{E_k}\frac{\epsilon_{k+q}}{E_{k+q}} -
369: \frac{\Delta}{E_k}\frac{\Delta}{E_{k+q}} \right) \nonumber \\
370: &\times
371: \Big(
372: \frac{f(E_k - \mu) - f(-E_{k+q} - \mu)}{\im \omega_n + E_k + E_{k+q} }
373: \nonumber \\
374: &\hspace{4em}+
375: \frac{f(-E_k - \mu) - f(E_{k+q} - \mu)}{\im \omega_n - E_k - E_{k+q} }
376: \Big)
377: \Bigg\}, \label{eqn:K}
378: \end{align}
379: and
380: \begin{align}
381: & H(q,\im \omega_n ) \nonumber \\
382: =& - \frac{1}{N_L} \sum_k
383: \Bigg\{
384: \frac{1}{4} \left(
385: \frac{\Delta}{E_k} + \frac{\Delta}{E_{k+q}} \right) \nonumber \\
386: &\times
387: \Big(
388: \frac{f(E_k - \mu) - f(E_{k+q} - \mu)}{\im \omega_n + E_k - E_{k+q} }
389: \nonumber \\
390: &\hspace{4em}-
391: \frac{f(-E_k - \mu) - f(-E_{k+q} - \mu)}{\im \omega_n - E_k + E_{k+q} }
392: \Big) \nonumber \\
393: &+ \frac{1}{4} \left(
394: \frac{\Delta}{E_k}- \frac{\Delta}{E_{k+q}} \right) \nonumber \\
395: &\times
396: \Big(
397: \frac{f(E_k - \mu) - f(-E_{k+q} - \mu)}{\im \omega_n + E_k + E_{k+q} }
398: \nonumber \\
399: &\hspace{4em}-
400: \frac{f(-E_k - \mu) - f(E_{k+q} - \mu)}{\im \omega_n - E_k - E_{k+q} }
401: \Big)
402: \Bigg\} \nonumber \\
403: =& H(q-Q_0, \im \omega_n).
404: \label{eqn:H}
405: \end{align}
406: Here, $\mu$ is the chemical,
407: $\epsilon_k = - 2 t \cos ka$ and $E_k = \sqrt{\epsilon_k^2 +
408: \Delta^2}$.
409: The energy gap $\Delta \geq 0$ which opens at $\pm \pi/(2a)$ is related to the charge
410: disproportionation $\rho$ as
411: \begin{align}
412: \Delta = - 2 v_c (Q_0) \rho.
413: \label{eqn:DvsRho}
414: \end{align}
415: where $v_c(q) = U/4 + V \cos qa$.
416: By using the action,
417: the correlations of the spin fluctuation are obtained as follows,
418: \begin{align}
419: &\langle \delta \sigma (q, \im \omega_n) \delta \sigma (-q,-\im
420: \omega_n) \rangle \nonumber \\
421: =&
422: \frac{1}{2} \frac{A_\sigma (q-Q_0, \im \omega_n)}
423: {A_\sigma (q, \im \omega_n) A_\sigma (q-Q_0, \im \omega_n) -
424: B_\sigma^2(q, \im \omega_n)}, \label{eqn:COL0}\\
425: &\langle \delta \sigma (q, \im \omega_n) \delta \sigma (-q+Q_0,-\im
426: \omega_n) \rangle \nonumber \\
427: =&
428: - \frac{1}{2} \frac{B_\sigma (q, \im \omega_n)}
429: {A_\sigma (q, \im \omega_n) A_\sigma (q-Q_0, \im \omega_n) -
430: B_\sigma^2(q, \im \omega_n)},
431: \label{eqn:COL1}
432: \end{align}
433: for $-\pi/a \leq q < \pi/a$
434: and the others are zero.
435: Note that
436: the CO leads to the coupling between $q$-component and
437: $-q+Q_0$-component in the fluctuation because the spacing between the
438: equivalent site becomes twice.
439: In the case of the half filling, \ie $\mu = 0$,
440: such a coupling diminishes because
441: the quantity $H(q,\on)$ in eq. (\ref{eqn:H}) vanishes.
442:
443: The spin susceptibility in the site representation, $\tilde{\chi}_\sigma (x_i, x_j ; \im \omega_n)$
444: is obtained as follows,
445: \begin{align}
446: &\tilde{\chi}_\sigma (x_i, x_j ; \im \omega_n) \nonumber \\
447: =& \frac{1}{N_L^2} \sum_{q,q'} \e^{-\im (q'x_i + qx_j)}
448: \int_0^\beta {\rm d} \tau \e^{\im \omega_n \tau}
449: \langle T_\tau m(q',\tau) m(q,0)
450: \rangle \nonumber \\
451: %
452: =& \frac{1}{N_L^2} \sum_{q,q'}
453: \e^{-\im (q'x_i + qx_j)} \nonumber \\
454: &\times \left\{
455: \langle \delta \sigma (q, \im \omega_n) \delta \sigma (q', -\im
456: \omega_n) \rangle + \delta_{q',-q} \frac{1}{2} \frac{N_L}{v_s(q)}
457: \right\} \nonumber \\
458: %
459: %=& \frac{1}{N_L^2} \sum_{q}
460: % \e^{\im q(x_i - x_j)} \nonumber \\
461: %&\times \Bigg[
462: %\left\{
463: %\langle \delta \sigma (q, \im \omega_n) \delta \sigma (-q, -\im
464: % \omega_n) \rangle + \frac{1}{2} \frac{N_L}{v_s(q)}
465: %\right\} \nonumber \\
466: %&+ (-1)^i \langle \delta \sigma (q, \im \omega_n) \delta \sigma (-q+Q_0, -\im
467: % \omega_n) \rangle
468: %\Bigg]
469: %\nonumber \\
470: %
471: \equiv& \frac{1}{N_L} \sum_q \e^{\im q (x_i - x_j)} \nonumber \\
472: \times&\left\{
473: \chi_\sigma (q,-q;\im \omega_n) + (-1)^i \chi_\sigma (q,-q+Q_0;\im \omega_n)
474: \right\},
475: \end{align}
476: where
477: \begin{align}
478: & \chi_\sigma (q,-q;\im \omega_n) \nonumber \\ =&
479: \frac{\langle \delta \sigma (q, \im \omega_n) \delta \sigma (-q, -\im
480: \omega_n) \rangle}{N_L} + \frac{1}{2v_s(q)}, \label{eqn:CHI-0}\\
481: & \chi_\sigma (q,-q+Q_0;\im \omega_n) \nonumber \\ =&
482: \frac{\langle \delta \sigma (q, \im \omega_n) \delta \sigma (-q+Q_0, -\im
483: \omega_n) \rangle}{N_L}.
484: \label{eqn:CHI-1}
485: \end{align}
486: The Knight shift $S_i$ at the location $x_i$ is proportional to the local
487: magnetic moment at the $i$-th site under
488: the uniform and static magnetic magnetic field.
489: Therefore, $S_i$ is written by using eqs.(\ref{eqn:CHI-0}) and (\ref{eqn:CHI-1}) as follows,
490: \begin{align}
491: S_i &\propto \sum_{j} \tilde{\chi}_\sigma (x_i,x_j;0) \nonumber \\
492: &\propto \chi_\sigma (0,0;0) + (-1)^i \chi_\sigma (0,Q_0;0).
493: \end{align}
494: Thus, the average between the Knight shift at the charge rich site and
495: that at the poor site is proportional to the magnetic
496: susceptibility\cite{Hiraki1998PRL}.
497: When the hyperfine coupling constant is assumed to be independent of the
498: wave number, the nuclear spin relaxation rate $R_i = (T_1 T)_i^{-1}$ is
499: written as follows,
500: \begin{align}
501: R_i \propto& \lim_{\omega \to 0}
502: \frac{{\rm Im} \tilde{\chi}_\sigma (x_i,x_i;\omega + \im
503: \delta)}{\omega} \nonumber \\
504: =& \frac{1}{N_L} \sum_q \lim_{\omega \to 0}
505: \Bigg\{
506: \frac{{\rm Im} \chi_\sigma (q,-q;\omega + \im
507: \delta)}{\omega} \nonumber \\
508: & + (-1)^i
509: \frac{{\rm Im} \chi_\sigma (q,-q+Q_0;\omega + \im
510: \delta)}{\omega}
511: \Bigg\},
512: \end{align}
513: where $\delta = +0$.
514: Note that the transverse susceptibility is the same as the longitudinal
515: one because the order of the spin does not appear in the region (II).
516: The charge rich and poor sites are the odd site ($i$=odd) and the even
517: site ($i$=even), respectively.
518: In the case of the half-filling, \ie $\mu = 0$,
519: the quantity, $H (q, \im \omega_n)$ in eq. (\ref{eqn:H}) vanishes, as is
520: already noted.
521: Therefore,
522: the Knight shift and the relaxation rate do not show splitting.
523: However, in the present quarter-filling case,
524: the splitting of both quantities appear, as will be shown in the next
525: section concretely.
526:
527: \section{Results}
528:
529: In this section, we discuss the Knight shift and the nuclear spin
530: relaxation rate in the pure CO state (II) in the phase diagram of
531: Fig. \ref{fig:phase}.
532:
533: Figure \ref{fig:shift1} shows the Knight shift at the charge rich site $S_+$ and that at
534: the poor site $S_-$ as a function of temperature for $U/t=4.0$ and
535: $V/t=3.0$.
536: These quantities are normalized by the shift in the non-interacting case at $T=0$,
537: $S_0$.
538: %----------- Fig.3 -----------
539: \begin{figure}[hob]
540: \centerline{\includegraphics[width=7.0true cm]{knight-U4-V3-normalize-paper.eps}}
541: \caption{
542: The normalized Knight shift at the charge-rich site $S_+/S_0$ and that at the poor site
543: $S_-/S_0$ as a function of the temperature for the pure charge ordered
544: state in the case of $U/t=4.0$ and $V/t=3.0$
545: Here $S_0$ is the shift of the non-interaction case at $T=0$.
546: The dashed curve expresses the
547: the average of $S_+$ and $S_-$,
548: \ie magnetic susceptibility.
549: The shift without taking account of the charge order is denoted by
550: the dotted curve.
551: }
552: \label{fig:shift1}
553: \end{figure}
554: %----------------------
555: The shift splits into the two components below $T_{\rm CO}$
556: and the quantity at the rich site is larger than that at the poor site.
557: The fact can be easily understood by considering the case where the
558: charge order is complete,
559: \ie $\rho = 1/2$.
560: In the charge poor site where there are no carriers,
561: the spin excitation is impossible.
562: Therefore the shift is equal to zero.
563: On the other hand, in the charge rich site with one electron per site,
564: the spin excitation is possible and the Knight shift is finite.
565: The dashed curve and dotted one express the average of $S_+$ and $S_-$,
566: \ie the uniform and static magnetic susceptibility and the quantity
567: in the absence of the charge order, respectively.
568: The magnetic susceptibility is enhanced by the appearance of the charge
569: order.
570: It is due to the enhancement of the density of state at the Fermi energy
571: by opening the gap at $\pm 2 \kf$.
572: It should be noticed that
573: the magnetic susceptibility does not show any singularity at the
574: critical temperature of the charge order,\cite{Tomio2002DT} which has been observed experimentally.\cite{Hiraki1998PRL}
575:
576: The nuclear spin relaxation rate for $U/t=4.0$ and $V/t=3.0$
577: normalized by the quantity in the
578: absence of the interaction at $T=0$, $R_0$
579: is shown in Fig. \ref{fig:rate1},
580: where $R_+$ and $R_-$ are $(T_1 T)^{-1}$ at the charge rich site and
581: that of poor site, respectively.
582: %----------- Fig.4 -----------
583: \begin{figure}[htb]
584: \centerline{\includegraphics[width=7.0truecm]{rate-U4-V3-normalize-paper.eps}}
585: \caption{
586: The normalized nuclear spin relaxation rate at the charge-rich site $R_+/R_0$ and that at the poor site
587: $R_-/R_0$ as a function of the temperature for the pure charge ordered
588: state in the case of $U/t=4.0$ and $V/t=3.0$
589: Here $R_0$ is the rate of the non-interaction case at $T=0$.
590: The dashed curve expresses the
591: the average of $R_+$ and $R_-$.
592: The rate without taking account of the charge order is denoted by
593: the dotted curve.
594: }
595: \label{fig:rate1}
596: \end{figure}
597: %---------------------
598: Here the dashed curve expresses the average of $R_+$ and $R_-$ and
599: the rate in the absence of CO is shown by the dotted curve.
600: As is seen in the Knight shift,
601: the rate shows splitting and
602: the quantity $R_+$ becomes larger $R_-$
603: because the spin fluctuation at the charge rich site is larger than that
604: at the poor site.
605: Note that the average seems to have a singularity at $T = T_{\rm CO}$,
606: which is different from the magnetic susceptibility.
607:
608: Next, the relationship between the mount of the splitting in the these
609: quantities and the charge disproportionation $\rho$ is discussed.
610: The experiments have been analyzed by assuming
611: that the relative shift, $S_+/(S_+ + S_-)$ or $S_-/(S_+ + S_-)$
612: measures the local charge\cite{Hiraki1998PRL} and that
613: the relaxation rate is proportional to the local charge
614: $(1/2 \pm \rho)^2$,\cite{Zamborszky2002PRB} respectively.
615: If the assumptions are valid,
616: the following relationship should hold,
617: \begin{align}
618: \frac{S_+ - S_-}{S_+ + S_+} &= 2\rho, \label{eqn:aS}\\
619: \frac{R_+ - R_-}{R_+ + R_+} &= \frac{2\rho}{1/2 + 2 \rho^2}.
620: \label{eqn:aR}
621: \end{align}
622: Figures \ref{fig:shift2} and \ref{fig:rate2} show
623: the l.h.s. of eq.(\ref{eqn:aS}) and that of eq.({\ref{eqn:aR})
624: as a function of the r.h.s. of the respective equations.
625: %----------- Fig.5 -----------
626: \begin{figure}[htb]
627: \centerline{\includegraphics[width=7.0truecm]{knight-U4-normalize-withR-paper-III.eps}}
628: \caption{
629: The mount of splitting of the Knight shift $S_+ - S_-$ normalized by
630: $S_+ + S_-$ as a function of $2\rho$ for several
631: choices of $V/t$ at $U/t=4.0$.
632: }
633: \label{fig:shift2}
634: \end{figure}
635: %---------------------
636: %----------- Fig.6 -----------
637: \begin{figure}[htb]
638: \centerline{\includegraphics[width=7.0truecm]{rate-U4-normalize-withR-paper-III.eps}}
639: \caption{
640: The mount of splitting of the nuclear spin relaxation rate $R_+ - R_-$ normalized by
641: $R_+ + R_-$ as a function of $2\rho/(1/2+2\rho^2)$ for several
642: choices of $V/t$ at $U/t=4.0$.
643: }
644: \label{fig:rate2}
645: \end{figure}
646: %---------------------
647: In Fig. \ref{fig:shift2},
648: the relative shift is well proportional to $2 \rho$, but
649: the coefficient becomes larger than unity.
650: The coefficient approaches to unity when $V$ increases.
651: Similarly, the quantity $(R_+ - R_-)/(R_+ + R_-)$ approaches to
652: $2 \rho / (1/2 + 2 \rho^2)$ for large $V$.
653: Thus, the assumptions, eqs. (\ref{eqn:aS}) and (\ref{eqn:aR}), seems to hold
654: approximately well for the large $V$ in which the CO state becomes
655: robust,
656: but with decreasing $V$ such relationship become worse.
657: It should be noted that
658: eqs.(\ref{eqn:aS}) and (\ref{eqn:aR}) gives the smaller value of the
659: charge disproportionation that the actual one.\cite{Zamborszky2002PRB}
660:
661: \section{Summary}
662: In the present paper,
663: we investigated the Knight shift and the nuclear spin relaxation rate
664: in the state with charge ordering based on the one-dimensional extended Hubbard
665: model with a quarter filled band with taking account of the fluctuation
666: around the mean-filed solution by RPA.
667: It is shown that both quantities shows splitting below the critical
668: temperature of the charge order.
669: The result is different from the case of the half-filling where
670: the splitting does not occur even in the CO state due to the
671: particle-hole symmetry.
672: The quantities at the charge rich site are larger than those at the poor
673: site due to the larger spin fluctuation.
674: Thus, in the materials with the quarter filling,
675: the splitting of these quantities can be considered to be the
676: experimental evidence of appearance of the charge order as has been
677: already experimentally insisted.
678: The relationship between the amount of the splitting normalized by the
679: average is discussed.
680: It is shown that
681: the assumptions used in the analysis of the experiments
682: hold well when $V$ becomes larger,
683: \ie the CO is robust,
684: though the estimated mount of the charge disproportionation becomes
685: generally smaller compared with the actual value.
686:
687: In the present study, we used the method in which
688: we apply the mean-field treatment to the present model and take into
689: account of the fluctuation around the solution by using RPA.
690: Though this method can succeed to show the splitting of the Knight shift
691: and the relaxation rate,
692: the transport properties of the actual materials such as
693: (DI-DCNQI)$_2$Ag cannot be described.
694: The materials show the insulating behavior even at the room temperature,
695: \ie $T > T_{\rm CO}$,\cite{Hiraki1998PRL}.
696: In addition, the derivative of the resistivity in terms of the temperature has a
697: peak at $T = T_{\rm CO}$.\cite{Ito04PRL} where the Metal-Insulator
698: transition occurs.
699: On the other hand, the present mean-field treatment leads to
700: the metallic state both $T > T_{\rm CO}$ and $T < T_{\rm CO}$.
701: That is, the metal insulator transition do not happen even at $T =
702: T_{\rm CO}$ because the energy gap opens at not $\pm \kf$ but $\pm 2
703: \kf$.
704: Therefore the further studies are necessary to describe the CO state
705: consistently in both the transport and magnetic properties.
706:
707:
708:
709:
710:
711: %------------Acknowledgement--------
712: \section*{Acknowledgments}
713: The author would like to H. Fukuyama, Y. Suzumura, M. Ogata, K. Kanoda,
714: S.E. Brown,
715: %J. Kishine,
716: H. Seo, M. Tsuchiizu and Y. Fuseya
717: for useful discussion.
718: This work was supported by Grant-in-Aid for Scientific Research on
719: Priority Area of Molecular Conductors (No.16038218) and
720: Grant-in-Aid for Scientific Research (C) (No.15540343)
721: from MEXT.
722:
723: \appendix
724: \section{Effective action of the fluctuation around the pure $4\kf$-CDW state}
725:
726: The partition function corresponding to the Hamiltonian, (\ref{eqn:Hk}) and
727: (\ref{eqn:Hint}), is written by utilizing the Stratonovich-Hubbard
728: transformation as follows,
729: %--- A1-A3 -----
730: \begin{align}
731: Z =& \int {\cal D}(c^*,c) {\cal D}\rho {\cal D} \sigma \e^{-(S_{\rm k} +
732: S_{\rm int})}, \\
733: S_{\rm k} =& \int_0^\beta \d \tau
734: \sum_{k,s} c^*_{k,s}(\tau) (\partial_\tau - \mu + \epsilon_k)
735: c_{ks}(\tau) , \\
736: S_{\rm int} =& \int_0^\beta \d \tau
737: \sum_q \nonumber \\
738: \times& \Bigg\{ \frac{v_c(q)}{N_L}\left[-\rho(q,\tau)\rho(-q,\tau) + 2
739: \rho(q,\tau) n(-q,\tau)
740: \right] \nonumber \\
741: +& \frac{v_s(q)}{N_L}\left[-\sigma(q,\tau)\sigma(-q,\tau) + 2
742: \sigma(q,\tau) m(-q,\tau)
743: \right]
744: \Bigg\},
745: \end{align}
746: %-------------
747: where $\mu$ is the chemical potential,
748: $N_L$ is the total number of sites,
749: $\beta = 1/T$,
750: $v_c(q) = U/4 + V \cos q a$, $v_c(q) = - U/4$,
751: $n(q,\tau) = \sum_{k,s} c^*_{k+q,s}(\tau) c_{k,s}(\tau)$ and
752: $m(q,\tau) = \sum_{k,s} s c^*_{k+q,s}(\tau) c_{k,s}(\tau)$.
753: Here we assume that the pure $4\kf$-CDW is realized.
754: Then the Stratonovich-Hubbard fields, $\rho(q,\tau)$ and
755: $\sigma(q,\tau)$, are rewritten as follows,
756: %---- A4-A5 ---
757: \begin{align}
758: \rho(q,\tau) &=
759: \begin{cases}
760: \rho(0) + \delta \rho(0,\tau), & \text{for $q=0$} \\
761: \rho(Q_0) + \delta \rho(Q_0,\tau), & \text{for $q=Q_0$} \\
762: \delta \rho(q,\tau), & \text{for the others}
763: \end{cases} \\
764: \sigma(q,\tau) &= \delta \sigma (q,\tau).
765: \end{align}
766: %------------
767: When eqs. (A$\cdot$4) and (A$\cdot$5) are substituted into (A$\cdot$3),
768: the action of the interaction part is divided as $S_{\rm int} =
769: S_{{\rm int},0} + S_{{\rm int},0}^{e} + S_{{\rm int},1} + S_{{\rm
770: int},2}$, where
771: %------ A6-A9 ----
772: \begin{align}
773: S_{{\rm int},0} =& - \beta \frac{v_c(0)}{N_L} \rho(0)^2 - \beta
774: \frac{v_c(Q_0)}{N_L} \rho(Q_0)^2, \\
775: S_{{\rm int},0}^e =& \int_0^\beta \d \tau \nonumber \\
776: \times& \left\{
777: 2 \frac{v_c(0)}{N_L} \rho(0) n(0,\tau) + 2 \frac{v_c(Q_0)}{N_L} \rho(Q_0) n(Q_0,\tau)
778: \right\}, \\
779: S_{{\rm int},1} =& \int_0^\beta \d \tau \bigg\{
780: \sum_{q \neq 0, Q_0} 2 \frac{v_c(q)}{N_L} \delta \rho (q,\tau) n(-q,\tau) \nonumber \\
781: &+ 2 \frac{v_c(0)}{N_L} ( n(0,\tau) - \rho(0) ) \delta \rho(0,\tau)
782: \nonumber \\
783: &+ 2 \frac{v_c(Q_0)}{N_L} ( n(Q_0,\tau) - \rho(Q_0) ) \delta
784: \rho(Q_0,\tau) \nonumber \\
785: &+ \sum_{q} 2 \frac{v_s(q)}{N_L} \delta \sigma (q,\tau) m(-q,\tau)
786: \bigg\}, \\
787: S_{{\rm int},2} =& - \int_0^\beta \d \tau \sum_q \bigg\{
788: \frac{v_c(q)}{N_L} \delta \rho (q,\tau) \delta \rho (-q,\tau) \nonumber
789: \\
790: &+
791: \frac{v_s(q)}{N_L} \delta \sigma (q,\tau) \delta \sigma (-q,\tau)
792: \bigg\}.
793: \end{align}
794: %---------------------
795: By integrated out the electron degree of freedom, the partition function
796: is calculated as,
797: %---- A10 ----
798: \begin{align}
799: Z = \e^{-S_{{\rm int},0}} Z_e^0 \int {\cal D}\delta \rho {\cal D}\delta
800: \sigma
801: \e^{-S_{{\rm int},2}} \langle \e^{-S_{{\rm int},1}} \rangle,
802: \end{align}
803: %------------
804: where $Z_e^0 = \int {\cal D}(c^*,c) \e^{-(S_{\rm k} + S_{{\rm
805: int},0}^e)}$ and $\langle \cdots \rangle$ denotes
806: the average in terms of $S_{\rm k} + S_{{\rm int},0}^e$ given as
807: follows,
808: %------- A11 -----
809: \begin{align}
810: & S_{\rm k} + S_{{\rm int},0}^e \nonumber \\
811: =& \int_0^\beta \d \tau \sum_{0<k\leq
812: Q_0,s}
813: (c_{k,s}^*(\tau), c_{k-Q_0,s}^*(\tau)) \nonumber \\
814: &\times
815: \left(
816: \begin{array}{cc}
817: \partial_\tau - \mu + \epsilon_k & \Delta \\
818: \Delta & \partial_\tau - \mu - \epsilon_k
819: \end{array}
820: \right)
821: \left(
822: \begin{array}{c}
823: c_{k,s}(\tau) \\
824: c_{k-Q_0,s}(\tau)
825: \end{array}
826: \right) \nonumber \\
827: =& \int_0^\beta \d \tau \sum_{0<k\leq
828: Q_0,s}
829: (c_{k,s}^*(\tau), c_{k-Q_0,s}^*(\tau)) \nonumber \\
830: &\times \left(- \overleftrightarrow{G}(k,\tau) \right)^{-1}
831: \left(
832: \begin{array}{c}
833: c_{k,s}(\tau) \\
834: c_{k-Q_0,s}(\tau)
835: \end{array}
836: \right),
837: \end{align}
838: %-------------------
839: with $\Delta = 2 v_c (Q_0) \rho(Q_0)/N_L \equiv -2 v_c(Q_0) \rho$.
840: Here
841: %-------- A12 ------
842: \begin{align}
843: \overleftrightarrow{G}(k,\tau) =&
844: \begin{pmatrix}
845: G(k,\tau) & F(k,\tau) \\
846: F(k-Q_0,\tau) & G(k-Q_0,\tau)
847: \end{pmatrix}
848: \nonumber \\
849: =&
850: \frac{1}{\beta} \sum_{\epsilon_n} \e^{-\im \epsilon_n \tau}
851: \nonumber \\
852: & \times
853: \left(
854: \begin{array}{cc}
855: G(k, \im \epsilon_n) & F(k, \im \epsilon_n)\\
856: F(k-Q_0, \im \epsilon_n) & G(k-Q_0, \im \epsilon_n)
857: \end{array}
858: \right),
859: \end{align}
860: %---------------------
861: where
862: %----- A13-A14 -------
863: \begin{align}
864: G(k,\im \epsilon_n) &= \frac{u_k^2}{\im \epsilon_n + \mu - E_k}
865: + \frac{v_k^2}{\im \epsilon_n + \mu + E_k}, \\
866: F(k,\im \epsilon_n) &= \frac{u_k v_k}{\im \epsilon_n + \mu - E_k}
867: - \frac{u_k v_k}{\im \epsilon_n + \mu + E_k},
868: \end{align}
869: for $- Q_0 < k \leq Q_0$
870: with $E_k = \sqrt{ \epsilon_k^2 + \Delta^2}$ ($\Delta > 0$) and
871: %---- A15-A16 ------
872: \begin{align}
873: u_k = \sqrt{\frac{1}{2}\left(1 + \frac{\epsilon_k}{E_k}\right)}, \\
874: v_k = \sqrt{\frac{1}{2}\left(1 - \frac{\epsilon_k}{E_k}\right)}.
875: \end{align}
876: %------------------
877: We expand $\langle \e^{-S_{{\rm int},1}} \rangle$ up to the second
878: order.
879: The condition $\langle S_{{\rm int},1} \rangle = 0$ gives rise to
880: the following equations,
881: %---- A17-A18 ----
882: \begin{align}
883: \rho (0) =& \frac{N_L}{2} = 2 \sum_{0<k\leq Q_0} \left\{f(-E_k - \mu) +
884: f(E_k - \mu) \right\}, \label{eqn:Ne}\\
885: \Delta =& 4v_c(Q_0) \frac{1}{N_L} \sum_{0<k\leq Q_0}
886: \frac{\Delta}{E_k}
887: \left\{f(E_k - \mu) -
888: f(-E_k - \mu) \right\}.
889: \label{eqn:gapII}
890: \end{align}
891: %-------------
892: Note that the chemical potential $\mu$ is determined by
893: eq.(\ref{eqn:Ne}) and eq.(\ref{eqn:gapII}) is the self-consistent
894: equation of $\Delta$.
895: On the other hand,
896: the quantity $\langle S_{{\rm int} 1}^2 \rangle / 2$ is calculated as
897: follows,
898: %------- A19 -----
899: \begin{align}
900: & \frac{\langle S_{{\rm int} 1}^2 \rangle}{2} \nonumber \\ =&
901: \frac{1}{2} \int_0^\beta \d \tau \d \tau' \sum_{q,q'} \nonumber \\
902: &\times\bigg\{
903: \frac{2 v_c(q)}{N_L}\frac{2 v_c(q')}{N_L} \delta \rho(q,\tau) \delta \rho(q',\tau')
904: \langle : n(-q,\tau): :n(-q',\tau'): \rangle \nonumber \\
905: &+ \frac{2 v_s(q)}{N_L}\frac{2 v_s(q')}{N_L} \delta \sigma(q,\tau) \sigma \rho(q',\tau')
906: \langle : m(-q,\tau): :m(-q',\tau'): \rangle
907: \bigg\} \nonumber \\
908: =& 4 \int_0^\beta \d \tau \d \tau' \frac{1}{N_L} \sum_{q,q'} \nonumber
909: \\
910: &\times \left\{
911: v_c(q) v_c(q') \delta \rho (q,\tau) \delta \rho (q',\tau')
912: + v_s(q) v_s(q') \delta \sigma (q,\tau) \delta \sigma (q',\tau')
913: \right\} \nonumber \\
914: &\times \left\{
915: \delta_{q+q',0} K (q, \tau-\tau') + \delta_{q+q',Q_0} H (q, \tau-\tau')
916: \right\},
917: \end{align}
918: %-----------
919: where $:A: = A - \langle A \rangle$ and
920: %---- A20-A21 ---
921: \begin{align}
922: & K(q,\tau) \nonumber \\
923: =& - \frac{1}{N_L} \sum_k \bigg\{
924: G(k-q,-\tau) G(k,\tau) + F(k-q,-\tau) F(k,\tau)
925: \bigg\} \nonumber \\
926: =& \frac{1}{\beta} \sum_{\omega_n} \e^{-\im \omega_n \tau} K(q,\im
927: \omega_n), \\
928: & H(q,\tau) \nonumber \\
929: =& - \frac{1}{N_L} \sum_k \bigg\{
930: G(k-q,-\tau) F(k,\tau) + F(k-q,-\tau) G(k,\tau)
931: \bigg\} \nonumber \\
932: =& \frac{1}{\beta} \sum_{\omega_n} \e^{-\im \omega_n \tau} H(q,\im
933: \omega_n).
934: \end{align}
935: %-------------------
936: The quantities, $K(q, \im \omega_n)$ and $H(q, \im \omega_n)$ are given
937: as follows,
938: \begin{align}
939: K(q,\im \on) &= - \frac{1}{\beta N_L} \sum_{k, \en} \big\{
940: G(k-q, \im \en) G(k, \im \en + \im \on) \nonumber \\
941: &+ F(k-q, \im \en) F(k, \im \en + \im \on)
942: \big\}, \\
943: H(q,\im \on) &= - \frac{1}{\beta N_L} \sum_{k, \en} \big\{
944: G(k-q, \im \en) F(k, \im \en + \im \on) \nonumber \\
945: &+ F(k-q, \im \en) G(k, \im \en + \im \on)
946: \big\},
947: \end{align}
948: and obtained in eq.(\ref{eqn:K}) and (\ref{eqn:H}), respectively.
949: When the Fourier transformation, $\delta \nu (q,\im \omega_n) =
950: 1/\sqrt{\beta} \int_0^\beta \d \tau \exp(-\im \omega_n \tau) \delta \nu
951: (q,\tau)$ ($\nu = \rho, \sigma$)
952: is introduced, eq.(A$\cdot$19) is written as follows,
953: %----- A22 -----
954: \begin{align}
955: &\frac{\langle S_{{\rm int}1}^2 \rangle}{2} \nonumber \\
956: =& \sum_{0<q\leq Q_0}
957: \sum_{\omega_n} \left(\delta \rho(-q,-\im \omega_n), \delta
958: \rho(-q+Q_0,-\im \omega_n) \right) \nonumber \\
959: &\times
960: \begin{pmatrix}
961: \dfrac{4 v_c^2(q)}{N_L} K (q, \im \omega_n) & \dfrac{4 v_c(q)v_c(q-Q_0)}{N_L} H (q, \im \omega_n) \\
962: \dfrac{4 v_c(q)v_c(q-Q_0)}{N_L} H (q, \im \omega_n) & \dfrac{4 v_c^2(q-Q_0)}{N_L} K (q-Q_0, \im \omega_n)
963: \end{pmatrix} \nonumber \\
964: &\times
965: \begin{pmatrix}
966: \delta \rho (q,\im \omega_n) \\ \delta \rho (q-Q_0,\im \omega_n)
967: \end{pmatrix} \nonumber \\
968: +&
969: \sum_{0<q\leq Q_0}
970: \sum_{\omega_n} \left(\delta \sigma(-q,-\im \omega_n), \delta
971: \sigma(-q+Q_0,-\im \omega_n) \right) \nonumber \\
972: &\times
973: \begin{pmatrix}
974: \dfrac{4 v_s^2(q)}{N_L} K (q, \im \omega_n) & \dfrac{4 v_s(q)v_s(q-Q_0)}{N_L} H (q, \im \omega_n) \\
975: \dfrac{4 v_s(q)v_s(q-Q_0)}{N_L} H (q, \im \omega_n) & \dfrac{4 v_s^2(q-Q_0)}{N_L} K (q-Q_0, \im \omega_n)
976: \end{pmatrix} \nonumber \\
977: &\times
978: \begin{pmatrix}
979: \delta \sigma (q,\im \omega_n) \\ \delta \sigma (q-Q_0,\im \omega_n)
980: \end{pmatrix}.
981: \end{align}
982: %-------------------
983: On the other hand, $S_{{\rm int}2}$ is written as follows,
984: %------- A23 -------
985: \begin{align}
986: & S_{{\rm int}2} \nonumber \\
987: =& - \sum_{0<q\leq Q_0}
988: \sum_{\omega_n} \left(\delta \rho(-q,-\im \omega_n), \delta
989: \rho(-q+Q_0,-\im \omega_n) \right) \nonumber \\
990: &\times
991: \begin{pmatrix}
992: \dfrac{v_c(q)}{N_L} & 0 \\
993: 0 & \dfrac{v_c(q-Q_0)}{N_L}
994: \end{pmatrix}
995: \begin{pmatrix}
996: \delta \rho (q,\im \omega_n) \\ \delta \rho (q-Q_0,\im \omega_n)
997: \end{pmatrix} \nonumber \\
998: &-
999: \sum_{0<q\leq Q_0}
1000: \sum_{\omega_n} \left(\delta \sigma(-q,-\im \omega_n), \delta
1001: \sigma(-q+Q_0,-\im \omega_n) \right) \nonumber \\
1002: &\times
1003: \begin{pmatrix}
1004: \dfrac{v_s(q)}{N_L} & 0 \\
1005: 0 & \dfrac{v_s(q-Q_0)}{N_L}
1006: \end{pmatrix}
1007: \begin{pmatrix}
1008: \delta \sigma (q,\im \omega_n) \\ \delta \sigma (q-Q_0,\im \omega_n)
1009: \end{pmatrix}.
1010: \end{align}
1011: %-------------
1012: Thus the effective action $S_{\rm eff} = S_{{\rm int}2} - \langle
1013: S_{{\rm int}1}^2 \rangle/2$ is given by the sum of the charge and spin
1014: part, $ \sum_{\nu = \rho, \sigma} S_\nu$,
1015: where
1016: %---- A23 ----
1017: \begin{align}
1018: {\cal S}_\nu
1019: =& \sum_{0 < q \leq Q_0} \sum_{\omega_n}
1020: (\delta \nu (-q, -\im \omega_n ), \delta \nu (-q + Q_0, -\im \omega_n))
1021: \nonumber \\
1022: &\times \left(
1023: \begin{array}{cc}
1024: A_\nu(q, \omega_n)& B_\nu(q, \omega_n) \\
1025: B_\nu(q, \omega_n)& A_\nu(q-Q_0, \omega_n)
1026: \end{array}
1027: \right) \nonumber \\
1028: &\times \left(
1029: \begin{array}{c}
1030: \delta \nu (q, \im \omega_n ) \\
1031: \delta \nu (q-Q_0, \im \omega_n )
1032: \end{array}
1033: \right),
1034: \end{align}
1035: %-------------------------
1036: with
1037: %-------- A24-A27 ------
1038: \begin{align}
1039: A_\rho (q, \omega_n) &= - \frac{v_c(q)}{N_L} \left\{ 1 + 4 v_c(q) K(q, \im
1040: \omega_n) \right\}, \\
1041: B_\rho (q, \omega_n) &= - \frac{4 v_c(q) v_c(q-Q_0)}{N_L} H(q, \im
1042: \omega_n), \\
1043: A_\sigma (q, \omega_n) &= - \frac{v_s(q)}{N_L} \left\{ 1 + 4 v_s(q) K(q, \im
1044: \omega_n) \right\}, \\
1045: B_\sigma (q, \omega_n) &= - \frac{4 v_s(q) v_s(q-Q_0)}{N_L} H(q, \im
1046: \omega_n).
1047: \end{align}
1048: %--------------------
1049:
1050:
1051: %===============================================================
1052: \begin{thebibliography}{99}
1053:
1054: %---------------------
1055: \bibitem{review}
1056: For review, see,
1057: T. Ishiguro, K. Yamaji and G. Saito:
1058: Organic Superconductors (Springer-Verlag, Berlin, 1998),
1059: Chem. Rev. {\bf 104} (2004).
1060: %---------------
1061: \bibitem{Hiraki1998PRL}
1062: K. Hiraki and K. Kanoda:
1063: \jo{\PRL}{80}{1998}{4737}.
1064: %---------------
1065: \bibitem{Kanoda1999JPhys}
1066: K. Kanoda, K. Miyagawa, A. Kawamoto and K. Hiraki:
1067: \jo{\JPIVF}{9}{1999}{Pr10-353}.
1068: %--------------
1069: \bibitem{Nogami1999JPhys}
1070: Y. Nogami, K. Oshima, K. Hiraki and K. Kanoda:
1071: \jo{\JPIVF}{9}{1999}{Pr10-357}.
1072: %---------------
1073: \bibitem{Chow2000PRL}
1074: D.S. Chow, F. Zamborszky, B. Alavi, D.J. Tantillo, A. Baur, C.A. Nerlic
1075: and S.E. Brown:
1076: \jo{\PRL}{85}{2000}{1698}.
1077: %---------------
1078: \bibitem{Monceau2001PRL}
1079: P. Monceau, F. Ya. Nad and S. Brazovskii:
1080: \jo{\PRL}{86}{2001}{4080}.
1081: %--------------
1082: \bibitem{Zamborszky2002PRB}
1083: F. Zamborszky, W. Yu, W. Raas, S.E. Brown, B. Alavi, C.A. Merlic and
1084: A. Baur:
1085: \jo{\PRB}{66}{2002}{081103}.
1086: %-------------
1087: \bibitem{Nakamura2003JPSJ}
1088: T. Nakamura:
1089: \jo{\JPSJ}{72}{2003}{213}.
1090: %-------------
1091: \bibitem{Miyagawa2000PRB}
1092: K. Miyagawa, A. Kawamoto and K. Kanoda:
1093: \jo{\PRB}{62}{2000}{7679}.
1094: %-------------
1095: \bibitem{Tajima2000PRB}
1096: H. Tajima, S. Kyoden, H. Mori,
1097: S. Tanaka, A. Kawamoto:
1098: \jo{\PRB}{62}{2000}{9378}.
1099: %-------------
1100: \bibitem{Chiba2001JPCS}
1101: R. Chiba, H. M. Yamamoto, T. Nakamura, and T. Takahashi:
1102: \jo{\JPCS}{62}{2001}{389}.
1103: %-------------
1104: \bibitem{Takano2001JPCS}
1105: Y. Takano, K. Hiraki, H. M. Yamamoto, T. Nakamura, and T. Takahashi:
1106: \jo{\JPCS}{62}{2001}{393}.
1107: %-------------
1108: \bibitem{Yamamoto2002PRB}
1109: K. Yamamoto, K. Yakushi, K. Miyagawa, K. Kanoda and A. Kawamoto:
1110: \jo{\PRB}{65}{2002}{085110}.
1111: %-------------
1112: \bibitem{W2003PRB}
1113: R. Wojciechowski, K. Yamamoto, K. Yakushi, M. Inokuchi and A. Kawamoto:
1114: \jo{\PRB}{67}{2003}{224105}.
1115: %----------------
1116: \bibitem{Seo97JPSJ}
1117: H. Seo and H. Fukuyama:
1118: \jo{\JPSJ}{66}{1997}{1249}.
1119: %---------------
1120: \bibitem{Seo00JPSJ}
1121: H. Seo:
1122: \jo{\JPSJ}{69}{2000}{805}
1123: %---------------
1124: \bibitem{Seo04CR}
1125: For recent theoretical review on $A_2 B$ compounds,
1126: H. Seo, C. Hotta, and H. Fukuyama:
1127: \jo{\CR}{104}{2004}{5005}.
1128: %--------------
1129: \bibitem{Yoshioka02JPSJ}
1130: H. Yoshioka:
1131: \jo{\JPSJ}{71}{2002}{1812}.
1132: %-------------
1133: \bibitem{Tomio01JPCS}
1134: The similar phase diagram has been obtained in the presence of
1135: dimerization,
1136: Y. Tomio and Y. Suzumura:
1137: \jo{\JPCS}{62}{2001}{431}.
1138: %--------------
1139: \bibitem{Tomio2002DT}
1140: Y. Tomio: Ph. D. Thesis (Nagoya University, Nagoya, 2002).
1141: %--------------
1142: \bibitem{Ito04PRL}
1143: T. Itou, K. Kanoda, K. Murata, T. Matsumoto, K. Hiraki and T. Takahashi:
1144: \jo{\PRL}{93}{2004}{216408}.
1145: \end{thebibliography}
1146:
1147:
1148:
1149: \end{document}
1150:
1151:
1152: