cond-mat0110315/MT.tex
1: \documentclass{elsart}
2: 
3: \def\ggs{\buildrel\textstyle > \over {\hbox{\raise0.05ex\hbox{$\sim$}}}}
4: \def\lls{\buildrel\textstyle < \over {\hbox{\raise0.05ex\hbox{$\sim$}}}}
5: \def\gsim{\,\lower0.95ex\hbox{$\ggs$}\,}
6: \def\lsim{\,\lower0.95ex\hbox{$\lls$}\,}
7: 
8: \usepackage{graphicx}
9: \usepackage{amssymb}
10: 
11: \topmargin=-1.cm
12: \begin{document}
13: 
14: \begin{frontmatter}
15: 
16: \title{Ising transition in a one-dimensional
17:    quarter-filled electron system with dimerization}
18: 
19:  \author[label1]{M.~Tsuchiizu}\ead{tsuchiiz@yukawa.kyoto-u.ac.jp},
20:  \author[label2]{E.~Orignac}
21:  \address[label1]{Yukawa Institute for Theoretical Physics, 
22:                  Kyoto University,  Kyoto 606-8502, Japan}
23:  \address[label2]{Laboratoire de Physique Th\'eorique, CNRS UMR 8549 
24:      Ecole Normale Sup\'erieure, 24 Rue Lhomond 75005 Paris, France}
25: 
26: 
27: \begin{abstract}
28: We examine critical properties of the 
29:    quarter-filled one-dimensional Hubbard model with dimerization and
30:    with the onsite and nearest-neighbor Coulomb repulsion $U$ and $V$.
31: By utilizing the bosonization method, 
32:    it is shown that the system exhibits an Ising quantum phase 
33:    transition from the Mott insulating state 
34:    to the charge-ordered insulating state.
35: It is also shown that 
36:    the dielectric permittivity exhibits a strong enhancement 
37:    as decreasing temperature with power-law dependence
38:    at the Ising critical point.
39: \end{abstract}
40: 
41: \begin{keyword}
42: A. organic compounds \sep D. electronic structure 
43:       \sep D. phase transitions
44: \end{keyword}
45: \end{frontmatter}
46: 
47: 
48: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
49: 
50: The Bechgaard salts (TMTSF)$_2$X and the Fabre salts 
51:    (TMTTF)$_2$X form a family of organic conductors with 
52:   quasi-one-dimensional band structure, 
53:   with a long history of both theoretical 
54:   and experimental research. 
55: These materials have a 3/4-filled electron (i.e., quarter-filled hole) 
56:    band due to  
57:    a charge transfer from two organic
58:    molecules to one counterion X.
59: Since  the
60:     transfer integral is slightly dimerized along the chain axis,
61:    the system becomes effectively half-filling.  
62: Optical and photoemission experiments on the normal states 
63:    have shown evidences of a
64:    charge gap for both TMTTF salts and TMTSF salts
65:    \cite{Gruner_ICSM,Gruner_science}. This charge gap is believed to
66: be related to the formation of a uniform insulating state. 
67: Recently,  a charge ordered (CO) insulating state
68:    have been discovered by  NMR measurements
69:    in the (TMTTF)$_2$AsF$_6$ and (TMTTF)$_2$PF$_6$ salts \cite{Chow} 
70:    and by  dielectric permittivity measurements
71:    in the (TMTTF)$_2$PF$_6$ salt \cite{Monceau}. 
72: 
73: 
74: 
75: In order to describe theoretically the insulator,
76:    one has to take into account the commensurability effect
77:    of both half-filling and quarter-filling
78:    \cite{Giamarchi_physica}.
79: As a simplified model,
80:    the quarter-filled one-dimensional (1D)
81:    extended Hubbard model with dimerization and 
82:    with the onsite and nearest-neighbor Coulomb repulsion
83:    has been proposed
84:    and it is shown in a mean-field approach that
85:    a transition from the 
86:    state with a uniform charge distribution
87:    to the CO state takes place
88:    by increasing the intersite Coulomb repulsion \cite{Seo}.
89: In order to obtain the characteristics of such a transition,
90:    the effect of quantum fluctuations must be taken into account.
91: For this purpose, the bosonization method is applied to this
92:    system \cite{Bourbonnais,Tsuchiiz}.
93: The model reduces to the 
94:    double-frequency sine-Gordon (DSG) model 
95:    for the charge degrees of freedom    
96:    where two kinds of non-linear terms
97:    appear due to the half-filled and quarter-filled 
98:    umklapp scattering.
99: The DSG model has been analyzed classically \cite{classical_DSG},
100:    and quantum field theoretically 
101:    \cite{Delfino,Fabrizio_PRL,Fabrizio,Tsuchiiz}. 
102: In terms of the renormalization group (RG) method,
103:    it is shown that there are two-kinds of different 
104:    fixed points corresponding to the
105:    Mott insulator and the CO insulator
106:    \cite{Tsuchiiz}.
107: However, in the perturbative RG scheme
108:    the both fixed points are strong coupling
109:    where the perturbative RG treatment breaks down, thus 
110:    the low energy states behavior and
111:    the critical behavior of the transition 
112:    cannot be captured correctly.
113: 
114: Recently, it has been established
115:    that
116:    the critical properties of the DSG model is described as a quantum
117:    Ising transition, i.e.,
118:    $c=1/2$ CFT \cite{Delfino,Fabrizio_PRL,Fabrizio}.
119: By expanding the previous analysis of the DSG model, we examine
120:    the critical properties of the 
121:    quarter-filled 1D extended Hubbard model with dimerization,
122:    and also examine
123:    the physical quantities, such as dielectric response and 
124:    optical conductivity, in the present study. Note that in
125: \cite{Fabrizio_PRL,Fabrizio}, the related problem of the conductivity
126: in a 1D Hubbard model with a {\it staggered} potential has been
127: considered. 
128: We also discuss the experimental results on the
129:    huge enhancement of the dielectric permittivity
130:    obtained in the
131:    (TMTTF)$_2$X salts \cite{Monceau}.
132: 
133: 
134: 
135: We consider a 1D extended Hubbard model with 
136:    quarter-filling, where the corresponding Hamiltonian is given by
137: %==========================================================
138: \begin{eqnarray}
139: H &=& - \sum_{j,\sigma,l}
140:    \left[ t + (-1)^{j} t_d\right] 
141:    \left(  c_{j, \sigma}^{\dagger} 
142:     \, c_{j+1,\sigma} + \mbox{\rm h.c.} \right)
143: \nonumber \\ &&{}
144: + U \sum_{j} \, n_{j, \uparrow} \, n_{j, \downarrow} 
145: + V \sum_{j} \, n_{j} \, n_{j+1} 
146: \label{eq:H}
147: \end{eqnarray}
148: %==========================================================
149:  where 
150:    $n_{j, \sigma} = c_{j, \sigma}^\dagger c_{j, \sigma}$ and
151:    $c_{j, \sigma}$ denotes the annihilation operator of the electron
152:    at the $j$-th site with spin 
153:    $\sigma$($=\uparrow,\downarrow$), and 
154:    $t_d(>0)$ 
155:    denotes dimerization along the chain.
156: The onsite and nearest-neighbor Coulomb repulsions
157:    are denoted by $U$ and $V$, respectively.
158: By utilizing the bosonization method,
159:    the phase Hamiltonian is given by $H=H_c+H_s$ with
160:    \cite{Tsuchiiz}
161: %===================================================================
162: \begin{eqnarray}
163: H_c &=& 
164: \frac{v_\rho}{4\pi} \int {\rm d}x 
165:   \left[\frac{1}{K_\rho} \left(\partial_x \theta_+ \right)^2
166:          + K_\rho \left(\partial_x \theta_- \right)^2
167:   \right]
168: \nonumber \\
169: && {}
170: - \frac{g_{1/2}}{2 \pi^2 \alpha^2} \int {\rm d}x \, \sin 2 \theta_+
171: + \frac{g_{1/4}}{2 \pi^2 \alpha^2} \int {\rm d}x \, \cos 4 \theta_+
172: ,
173: \label{eq:phaseHamiltonian}
174: \\
175: H_s &=& 
176: \frac{v_\sigma}{4\pi} \int {\rm d}x 
177:   \left[\frac{1}{K_\sigma} \left(\partial_x \phi_+ \right)^2
178:          + K_\sigma \left(\partial_x \phi_- \right)^2
179:   \right]
180: \nonumber \\
181: && {}
182: + \frac{g_{s}}{2 \pi^2 \alpha^2} \int {\rm d}x \, \cos 2 \phi_+
183: ,
184: \label{eq:phaseHamiltonian_spin}
185: \end{eqnarray}
186: %===================================================================
187:    where $\theta_\pm$ and $\phi_\pm$ are charge and spin 
188:    phase variables \cite{Tsuchiiz} and
189:    $\alpha$ is a cutoff 
190:    of the order of the lattice constant.
191: In Eq.~(\ref{eq:phaseHamiltonian}),
192:    the velocity and the Tomonaga-Luttinger exponent are
193:    given by
194:    $v_\rho = v_{\rm F} 
195:     [(1+g_{4\rho}/2\pi v_{\rm F})^2-(g_\rho/2\pi v_{\rm F})^2]^{1/2}$,
196:     $v_\sigma = v_{\rm F} 
197:     [(1+g_{4\sigma}/2\pi v_{\rm F})^2
198:          -(g_\sigma/2\pi v_{\rm F})^2]^{1/2}$,
199:     $K_\rho = 
200:     [(2\pi v_{\rm F}+g_{4\rho}-g_\rho)/
201:            (2\pi v_{\rm F}+g_{4\rho}+g_\rho)]^{1/2}$
202:     and
203:     $K_\sigma = 
204:     [(2\pi v_{\rm F}+g_{4\sigma}-g_\sigma)
205:      /(2\pi v_{\rm F}+g_{4\sigma}+g_\sigma)]^{1/2}$.
206:  The constants $g_\rho$, $g_{4\rho}$, $g_\sigma$ and $g_{4\sigma}$
207:     are given by 
208:    $g_{\rho} =g_{4\rho} = (U + 4V)a$ and
209:    $g_{\sigma} =g_{4\sigma} = Ua$
210:    in the lowest order of the interaction, where
211:    $a$ is the lattice constant.
212: The magnitude of half-filled umklapp scattering, $g_{1/2}$,
213:    and that of quarter-filled one, $g_{1/4}$, 
214:    are given by \cite{Tsuchiiz}
215: %===================================================================
216: \begin{eqnarray}
217: g_{1/4} &=& \frac{1}{(2\pi\alpha)^2} \, 
218: \frac{A^4 \, a^5}{v_{\rm F}^2} \, U^2 (U-4V)
219: ,
220: \label{eq:g1/4}
221: \end{eqnarray}
222: %===================================================================
223:    and $g_{1/2} =B  Ua $ in the lowest order where
224:    $A \equiv [1-(t_d/t)^2]/[1+(t_d/t)^2]$ and
225:    $B \equiv [2t_d/t]/[1+(t_d/t)^2]$.
226: We neglect the spin-charge coupling term 
227:    $U^3 \cos 4\theta_+ \cos 2\phi_+$ due to 
228:    its large scaling dimension \cite{Yoshioka}.
229: The order parameters
230:    of the $4k_{\rm F}$ charge density wave (CDW) state and
231:    the $4k_{\rm F}$ bond ordered wave (BOW) state
232:    are given by\cite{Yoshioka}
233: %===================================================================
234: \begin{eqnarray}
235: \mathcal{O}_{4k_{\rm F}\mbox{-}\rm CDW}(x)  
236: &\propto&
237: (-1)^{x/a}
238: \cos  2\theta_+(x)
239: , 
240: \\
241: \mathcal{O}_{4k_{\rm F}\mbox{-}\rm BOW} (x)
242: &\propto&
243: (-1)^{x/a}
244: \sin  2\theta_+(x)
245: , 
246: \end{eqnarray}
247: %===================================================================
248: where $k_{\rm F}=\pi/4a$.
249: 
250: 
251: The spin degrees of freedom becomes gapless due to the 
252:    marginally irrelevance of the $g_{s}$-term. 
253: Therefore we focus hereafter on the charge degrees of freedom only.
254: In the case of $t_d=0$ (i.e., $g_{1/2}=0$),
255:    it has been shown by the RG analysis that 
256:    the system undergoes the
257:    Kosterlitz-Thouless transition by increasing the 
258:    nearest-neighbor repulsion \cite{Yoshioka} where
259:    the gapped phase corresponds to
260:    the CO insulator.
261: In the case of $t_d\neq 0$,
262: the Hamiltonian $H_c$ has two possible ground states with $\langle
263:    \theta_+\rangle = 0,\pi/2$ ($4k_F$-CDW, dominated by $g_{1/4}<0$)
264:     or  $\langle \theta_+\rangle = \pi/4$  ($4k_F$-BOW, dominated by
265: $g_{1/2}$). These two ground states are incompatible, so we expect a
266: phase transition of the Ising type, as discussed in \cite{Fabrizio}. 
267: 
268: 
269: 
270: By following the method of Ref.~\cite{Fabrizio},
271:    we analyze the DSG model
272:    Eq.~(\ref{eq:phaseHamiltonian}).
273: By using the Jordan-Wigner transformation,
274:    the vertex operators
275:    are expressed in terms of the order ($\sigma_1$ and $\sigma_2$) 
276:    and disorder ($\mu_1$ and $\mu_2$) parameters of the 
277:    two Ising models as follows \cite{Gogolin_1d_book,Fabrizio}: 
278: %==========================================================
279: \begin{eqnarray}
280: \sin 2\theta_+ &=& \sigma_1^z \, \sigma_2^z
281: ,
282: \label{eq:sin2theta}
283: \\
284: \cos 2\theta_+ &=& \mu_1^z \, \mu_2^z
285: ,
286: \\
287: \cos 4\theta_+ &=& \frac{1}{2} (\sigma_1^x + \sigma_2^x)
288: .
289: \end{eqnarray}
290: %==========================================================
291: The charge part of the Hamiltonian $H_c$ can be expressed in 
292:    terms of the quantum Ashkin-Teller model which
293:    consists of two quantum Ising  spin chains 
294:    coupled by a four-spin interaction term \cite{Fabrizio}:
295: %===================================================================
296: \begin{eqnarray}
297: H_c &=&
298: H_{\rm QI}[\sigma_1,\sigma_2]
299: + H_{\rm AT} [\sigma_1,\sigma_2]
300: + H_h [\sigma_1,\sigma_2]
301: ,
302: \label{eq:DQAT}
303: \end{eqnarray}
304: %===================================================================
305:    where 
306:    $\sigma_l$ are the Pauli matrices of $l$-th chain $(l=1,2)$.
307: In Eq.~(\ref{eq:DQAT}),
308:    $H_{\rm QI}$ is given by
309: %===================================================================
310: \begin{eqnarray}
311: H_{\rm QI} 
312: &=&
313: -\sum_{l=1,2} \sum_n 
314: \Bigl[
315:   J \,  \sigma_l^z(n) \, \sigma_l^z(n+1) + \Delta \, \sigma_l^x(n)
316: \Bigr]
317: ,
318: \label{eq:Hqi}
319: \end{eqnarray}
320: %===================================================================
321:    where $n$ denotes the lattice site on the respective chain.
322: This Hamiltonian describes two independent
323:    quantum Ising 
324:    (QI) chains in a transverse magnetic field ($\Delta$).
325: In Eq.~(\ref{eq:DQAT}),
326:   $H_{\rm AT}$ is the Ashkin-Teller interaction term given by
327: %===================================================================
328: \begin{eqnarray}
329: H_{\rm AT} 
330: &=&
331: J' \sum_n 
332: \Bigl[
333:   \sigma_1^z(n) \, \sigma_1^z(n+1) \,
334:   \sigma_2^z(n) \, \sigma_2^z(n+1) 
335: \nonumber \\ && \hspace*{1cm} {}
336:  + \sigma_1^x(n)  \, \sigma_2^x(n)
337: \Bigr]
338: ,
339: \end{eqnarray}
340: %===================================================================
341:    where this term describes  a four-spin interaction.
342: It is noted that this interaction term takes a {\it self-dual} form,
343:    thus this term does not make the system non-critical.
344: Finally,
345:    $H_h$ is a ``magnetic field''-type coupling, given by
346: %===================================================================
347: \begin{eqnarray}
348: H_h 
349: &=&
350: -h \sum_n 
351:   \sigma_1^z(n) \, 
352:   \sigma_2^z(n) 
353: .
354: \end{eqnarray}
355: %===================================================================
356: By using the Jordan-Wigner transformation,
357:    the parameters in Eq.~(\ref{eq:phaseHamiltonian}) and
358:    those in Eq.~(\ref{eq:DQAT}) can be related as 
359:   $v_\rho = 2J a [ 1- \left(2J'/\pi J \right)^2 ]^{1/2}$,
360:   $K_\rho = (1/8)
361:     [ \{1+(2J'/\pi J)\}/\{1-(2J'/\pi J)\} ]^{1/2}$,
362:   $g_{1/2}= 2a  h$ and
363:   $g_{1/4} = - 4a (\Delta-J ) $.
364: Note that the sine-Gordon model, Eq.~(\ref{eq:phaseHamiltonian}), 
365:    with $K_\rho=1/8$, $g_{1/2}= 0$ and $g_{1/4} \neq 0$, is equivalent 
366:    to a model of two degenerate massive Majorana fermions,
367:    i.e., two decoupled non-critical Ising models \cite{Gogolin_1d_book},
368:    and,  when the nonlinear term vanishes ($g_{1/4}\to0$), 
369:    two Ising models
370:    become critical simultaneously, i.e., critical properties
371:    are described by $c=1$ CFT.
372: 
373: 
374: The origin of the Ising criticality ($c=1/2$ CFT) in the case
375:    of $g_{1/2}\neq 0$ 
376:    can be captured easily 
377:    by considering the limit of large $h$ \cite{Fabrizio}.
378: In this case, the Ising spins are
379:    locked as $(\sigma_1,\sigma_2)=(1,1)$ or $(-1,-1)$
380:    where $\sigma_i$ is the eigenvalue of $\sigma^z_i$.
381: This fact implies that 
382:    the ``relative'' degrees of freedom 
383:    becomes locked (i.e., this freedom is always non-critical) 
384:    while the ``total'' degrees of freedom 
385:    asymptotically decouples  and 
386:    the system can become
387:    critical where the critical properties are described by $c=1/2$ CFT.
388: In order to analyze the physical quantities on and near 
389:    the criticality, 
390:    we change the basis into ``total''-``relative'' representation,
391:    where the relative Ising mode is described by 
392:    $\tau=\sigma_1\, \sigma_2$, while the total mode 
393:    by  e.g., $\sigma=\sigma_1$. 
394: The operator identities are given by \cite{Fabrizio}
395: %================================================================
396: \begin{eqnarray}
397: \begin{array}{ll}
398:    \sigma^z = \sigma_1^z , 
399: &
400:    \tau^z = \sigma_1^z  \, \sigma_2^z ,
401: \\
402:    \sigma^x=\sigma_1^x \, \sigma_2^x , \hspace*{.2cm}
403: &
404:    \tau^x=\sigma_2^x .
405: \end{array}
406: \end{eqnarray}
407: %===================================================================
408: In terms of these spin variables, 
409:     Eq.~(\ref{eq:DQAT}) is rewritten as
410:    $H_c = H_\sigma + H_\tau + H_{\sigma \tau}$ where
411: %===================================================================
412: \begin{eqnarray}
413: H_\sigma
414: &=& \sum_n
415: \Bigl[
416:   -J \, \sigma^z(n) \, \sigma^z(n+1) 
417:   +J' \, \sigma^x(n)
418: \Bigr]
419: ,
420: \\
421: H_\tau
422: &=& \sum_n
423: \Bigl[
424:    J' \, \tau^z(n) \, \tau^z(n+1) 
425:   -\Delta \, \tau^x(n)
426:   -h \, \tau^z(n)
427: \Bigr]
428: ,
429: \\
430: H_{\sigma \tau}
431: &=& -\sum_n
432: \Bigl[
433:    J \, \sigma^z(n) \, \sigma^z(n+1) \, \tau^z(n) \, \tau^z(n+1) 
434: \nonumber \\ && \hspace*{1.5cm} {}
435:    + \Delta \, \sigma^x(n) \, \tau^x(n)
436: \Bigr]
437: .
438: \end{eqnarray}
439: %===================================================================
440: The Ising transition corresponds the situation 
441:    when the $\sigma$-degrees of freedom go massless
442:    at $\Delta \approx \sqrt{h(2J+J')}$ \cite{Fabrizio}.
443: With no regard to the Ising transition in the $\sigma$-chain,
444:    the $\tau$-chain remains frozen in a 
445:    configuration with both $\langle \tau^z \rangle$ and 
446:    $\langle \tau^x \rangle$ 
447:    nonzero.
448: Here we introduce the spin variables $\mu$ and $\nu$
449:   which are dual to
450:   $\sigma$ and $\tau$, respectively, by 
451:   the Kramers-Wannier transformation:
452:   $\mu^z(n+1/2) \equiv \prod_{m=1}^n \sigma^x(m)$,
453:   $\mu^x(n+1/2) \equiv \sigma^z(n)  \sigma^z(n+1)$,
454:   $\nu^z(n+1/2) \equiv \prod_{m=1}^n \tau^x(m)$ and
455:   $\nu^x(n+1/2) \equiv \tau^z(n)  \tau^z(n+1)$.
456: The spin variables $\mu$ and $\nu$ can be related with
457:     $\mu_1$ and $\mu_2$ by 
458:    $\mu^z=\mu_1^z \, \mu_2^z$,
459:    $\mu^x=\mu_1^x$, 
460:    $\nu^z=\mu_2^z$ and
461:    $\nu^x=\mu_1^x \, \mu_2^x$.
462: The expectation value of $\mu^z$
463:    becomes finite ($\langle \mu^z\rangle\neq 0$) in the case that
464:    $\sigma$-chain is in the disordered phase 
465:    ($\langle \sigma^z\rangle=0$), 
466:    while becomes zero ($\langle\mu^z\rangle= 0$) in the ordered phase
467:    ($\langle \sigma^z\rangle \neq 0$).
468: The order parameter of $4k_{\rm F}$-CDW
469:    is given in terms of the spin variable by
470:    $\mathcal{O}_{4k_{\rm F}\mbox{-}\rm CDW} 
471:     \propto \cos 2\theta_+  =  \mu_1^z \, \mu_2^z=\mu^z$, and then 
472:    one finds $\langle \mathcal{O}_{4k_{\rm F}\mbox{-}\rm CDW} \rangle 
473:     \propto \langle \mu^z \rangle \neq 0$  in the disorder phase only. 
474: Therefore
475:    the transition from the Mott insulator to the
476:    CO insulator corresponds to this order-disorder
477:    Ising quantum phase transition
478:    of the $\sigma$ Ising chain,
479:    where the disordered phase of the Ising system
480:    corresponds to the CO insulating state.
481: Note that the quantity $\langle \mathcal{O}_{4k_{\rm F}\mbox{-}\rm BOW}\rangle$
482:    takes a finite value
483:    and shows no anomaly at the transition point,
484:    since the order parameter of $4k_{\rm F}$-BOW is given in terms
485:    of the noncritical Ising field:
486:    $\mathcal{O}_{4k_{\rm F}\mbox{-}\rm BOW}\propto \sin 2\theta_+
487:     =\sigma_1^z\, \sigma_2^z = \tau^z$.
488: 
489: 
490: 
491: Now we examine  the dielectric permittivity and the
492:    optical conductivity at finite temperature.
493: In the energy scale higher than the 
494:    gap in the $\tau$-chain, the perturbative analysis 
495:    in terms of renormalization group method
496:    works well \cite{Tsuchiiz}.
497: The complex conductivity 
498:    $\sigma(\omega)=\sigma_1(\omega)+i\, \sigma_2(\omega)$
499:    ($\sigma_1\equiv {\rm Re} \, \sigma$ and 
500:     $\sigma_2\equiv {\rm Im} \, \sigma$)
501:    can be calculated from
502:    $\sigma(\omega)=i[\chi(\omega)-\chi(0)]/\omega$,
503:    where $\chi$ is the retarded current-current correlation 
504:    function: $\chi(\omega)=-i\int dx \int_0^\infty dt 
505:    \langle [j(x,t),j(0,0)]\rangle$ $ e^{i\omega t-\delta t}$.
506: In terms of the complex conductivity, 
507:    the dielectric permittivity is given by
508:    $\epsilon(\omega)=\epsilon_1(\omega)+i \, \epsilon_2(\omega)
509:     =1+(4\pi i/\omega)\, \sigma(\omega)$,
510:    where $\epsilon_1\equiv {\rm Re} \, \epsilon$ and
511:    $\epsilon_2\equiv {\rm Im} \, \epsilon$.
512: By utilizing the RG method, 
513:    it can be shown that
514:    the resistivity $\rho(T)$ and the dielectric permittivity
515:    $\epsilon_1(T)$ 
516:    behaves as $\rho(T)=1/\sigma_1(T)\propto T^{16K_\rho-3}$
517:    and $\epsilon_1(T)\propto T^{2-16K_\rho}$ 
518:    in the case that the quarter-filled umklapp scattering
519:    is dominant \cite{Tsuchiiz}.
520: This analysis, however, only holds in the high temperature region.
521: In the energy scale lower than the gap in the $\tau$-chain, 
522:    we analyze these physical quantities in the 
523:    quantum Ising model scheme.
524: By using the spin variable $\mu$,
525:    the current operator are given by
526:    \cite{Fabrizio}
527: %================================================================
528: \begin{eqnarray}
529: j(x) &=& -\frac{1}{\pi} \, \partial_t \theta_+
530:   \, \to -C \, \partial_t \, \mu(x)
531: ,
532: \label{eq:current}
533: \end{eqnarray}
534: %===================================================================
535:    where $\mu(x)$ is a continuum variable of $\mu^z(n)$ and
536:    and $C(\propto \langle \tau^z \rangle)$ is a non-universal
537:    numerical constant.
538: Here we have used the fact that the $\tau$-field is
539:    always non-critical
540:    and has nonzero expectation value. 
541: We introduce the  response function of the 
542:     Ising disorder parameter:
543: %================================================================
544: \begin{eqnarray}
545: D^{(R)}(k,\omega) &=& -i \int dx \, \int_0^\infty dt \,
546:  e^{-ikx+i\omega t} 
547: \left\langle
548:    \left[ \mu(x,t) , \, \mu(0,0)\right]
549: \right\rangle .
550: \label{eq:responseIsing}
551: \end{eqnarray}
552: %===================================================================
553: By using Eq.~(\ref{eq:current}), 
554:    the current-current correlation function is given 
555:    by
556:    $\chi(\omega)=C^2\omega^2 D^{(R)}(0,\omega)$.
557: Then the real and imaginary parts of the conductivity $\sigma_1$ and
558:    $\sigma_2$ are given by
559:   $\sigma_1(\omega) = -C^2 \, \omega \, {\rm Im} 
560:    D^{(R)}(0,\omega)$ and
561:   $\sigma_2(\omega) =  C^2 \, \omega \, {\rm Re} 
562:    D^{(R)}(0,\omega)$.
563: On the Ising criticality, it is known that
564:    the response function 
565:    is given by
566:    $D_{\rm crit}(r) \propto 1/r^{1/4}$
567:    where $r=(x^2+\tau^2)^{1/2}$ 
568:    with the velocity being set to unity.
569: This power-law behavior is still valid away from 
570:    the criticality in the short range region $r\ll 1/\Delta_\sigma$,
571:    where $\Delta_\sigma$ is the gap in the $\sigma$-part of
572:    the Ising chain.
573: By using this asymptotic behavior,
574:    the retarded response functions is given by \cite{Gogolin_1d_book}
575: %================================================================
576: \begin{eqnarray}
577: D_{\rm crit}^{(R)}(0,\omega) 
578:  &=& 
579: -\frac{\sin(2\pi \Delta)}{(2\pi T)^{2-4\Delta}} 
580:   B^2\left(\Delta-i S, 1-2 \Delta \right)
581: ,
582: \end{eqnarray}
583: %===================================================================
584:    where $S=\omega/4\pi T$,
585:    the conformal dimension is $\Delta(=\bar{\Delta})=1/16$ and
586:    $B(x,y)$ is the beta function given by
587:    $B(x,y)=\Gamma(x)\Gamma(y)/\Gamma(x+y)$.
588: The temperature dependence of the dielectric permittivity
589:    $\epsilon_1(\omega,T)$ with 
590:    zero and finite frequency is shown in Fig.~1, where $W$ denotes
591:    the cutoff of the order of the bandwidth and
592:    for simplicity we have set $C=1$.
593: The frequency dependence of $\epsilon_1(\omega,T)$ is also shown
594:    in the inset.
595: The dielectric permittivity shows strong temperature dependence. 
596: In high temperature limit, the dielectric permittivity $\epsilon_1$ 
597:    becomes unity independent of $\omega$,
598:    but as decreasing temperature
599:    the huge enhancement can be obtained and
600:    $\epsilon_1$ takes a maximum 
601:    at $T\approx \omega$.
602: At low temperature ($T  \lsim \omega$), $\epsilon_1$ 
603:    decreases and changes its sign, and finally saturates to a value
604:    that depends on $\omega$.
605: The optical conductivity also increases as decreasing temperature, and
606:    exhibits a maximum at $T\approx \omega$.   
607: The limiting behaviors of $\epsilon_1$ and $\sigma_1$ 
608:    are given by
609:    $\epsilon_1(\omega,T)= 1+ {\rm const.}/T^{7/4}$ and
610:    $\sigma_1(\omega,T)\propto \omega^2/T^{11/4}$
611:    for high temperature ($T \gg \omega$),
612:    and
613:    $\epsilon_1(\omega,T)\propto - 1/\omega^{7/4}$ and
614:    $\sigma_1(\omega,T)\propto 1/\omega^{3/4}$ 
615:    for low temperature ($T \ll \omega$).
616: 
617: 
618: %====================================================================
619: \begin{figure}
620: \begin{center}
621: \includegraphics[width=8cm]{fig1}
622: \end{center}
623: \caption{ 
624: The temperature dependence of the dielectric permittivity
625:    $\varepsilon_1(\omega,T)$ 
626:    for $\omega/W=0$, 0.0005 and 0.001.
627: The inset shows the frequency dependence 
628:    of $\varepsilon_1(\omega,T)$ with fixed temperature
629:    $T/W=0.001$, 0.005 and 0.01.
630: }
631: \end{figure}
632: %======================================================================
633: 
634: 
635: 
636: Finally we discuss the experimental observation 
637:    of the (TMTTF)$_2$X salts.
638: Recent analysis of the dielectric response
639:    exhibits a strong temperature dependence and a huge enhancement 
640:    of the dielectric permittivity as decreasing temperature
641:    \cite{Monceau}.
642: In our analysis of the temperature dependence of the
643:     dielectric permittivity,
644:    we have restricted ourselves to the case on the Ising 
645:    critical point.
646: However, the results would be valid 
647:    in the vicinity of the transition point in the energy region
648:    above $\Delta_\sigma$.
649: Therefore the huge enhancement could be interpreted by 
650:    considering that the system is in the vicinity of the
651:    transition point from the Mott insulator to the CO insulator.
652: 
653: 
654: 
655: 
656: In summary,  we examined the properties of the
657:    dimerized extended Hubbard model at quarter-filling
658:    in one dimension.
659: It was shown that the phase transition
660:    from Mott insulating state to the CO insulating state
661:    in this system
662:    corresponds to the Ising quantum phase transition described by 
663:    $c=1/2$ CFT.
664: We have also show that 
665:    the dielectric permittivity exhibits a strong enhancement 
666:    as decreasing temperature with power-law dependence
667:    at the Ising critical point.
668: 
669: 
670: 
671: One of the authors (M.T.) thanks A.~Furusaki, T.~Hikihara, H.~Seo,
672:    Y.~Suzumura and H.~Yoshioka for helpful discussions.
673:  
674: 
675: %===============================================================
676: \begin{thebibliography}{99}
677: \def\jo #1#2#3#4{#1 {\bf #2} (#3) #4}  
678: %----------------------------
679: \bibitem{Gruner_ICSM}
680: A. Schwartz, 
681: S. Donovan, M. Dressel, L. Degiorgi,  G. Gr\"uner,
682: \jo{Synth. Met.}{86}{1997}{2129}.
683: %----------------------------
684: \bibitem{Gruner_science}
685: V. Vescoli, L. Degiorgi, W. Henderson, G. Gr\"uner,
686:    K. P. Sarkey, L. K. Montgomery, 
687: Science {\bf 281} (1998) 1181.
688: %----------------------------
689: \bibitem{Chow}
690: D.S.~Chow, F.~Zamborszky, B.~Alavi, D.J.~Tanillo, A.~Baur, 
691:    C.A.~Merlic, S.E.~Brown,
692: \jo{Phys. Rev. Lett.}{85}{2000}{1698}.
693: %----------------------------
694: \bibitem{Monceau}
695: F.~Nad, P.~Monceau, C.~Carcel, J.M.~Fabre,
696: \jo{Phys. Rev. B}{62}{2000}{1753}.
697: %----------------------------
698: \bibitem{Giamarchi_physica}
699: T. Giamarchi,
700: \jo{Physica B}{230-232}{1997}{975}.
701: %----------------------------
702: \bibitem{Seo}
703: H.~Seo, H.~Fukuyama,
704: \jo{J. Phys. Soc. Jpn.}{66}{1997}{1249}.
705: %----------------------------
706: \bibitem{Bourbonnais}
707: C.~Bourbonnais, D.~J\'erome,
708: {\it Advances in Synthetic Metals, Twenty years of Progress 
709:    in Science and Technology}, Eds. 
710:    P. Bernier, S. Lefrant, G. Bidan (Elsevier, New York, 1999), 
711:    P.~206-261.
712: %----------------------------
713: \bibitem{Tsuchiiz} 
714: M.~Tsuchiizu, H.~Yoshioka, Y.~Suzumura,
715:  J.~Phys.~Soc.~Jpn. {\bf 70} (2001) 1460.
716: %----------------------------
717: \bibitem{classical_DSG}
718: C.A.~Condat, R.A.~Guyer, M.D.~Miller,
719:    Phys. Rev. B {\bf 27} (1983) 474. 
720: %--------------------------
721: \bibitem{Delfino}
722: G. Delfino, G. Mussardo,
723:    Nucl. Phys. B {\bf 516} (1998) 675.
724: %----------------------------
725: \bibitem{Fabrizio_PRL} 
726: M.~Fabrizio, A.O.~Gogolin, A.A.~Nersesyan,
727:    Phys.~Rev.~Lett. {\bf 83} (1999) 2014.
728: %----------------------------
729: \bibitem{Fabrizio} 
730: M.~Fabrizio, A.O.~Gogolin, A.A.~Nersesyan,
731:    Nucl. Phys. B {\bf 580} (2000) 647.
732: %----------------------------
733: \bibitem{Yoshioka} 
734: H.~Yoshioka, M.Tsuchiizu, Y.~Suzumura,
735:  J.~Phys.~Soc.~Jpn. {\bf 70} (2001) 762, and references therein.
736: %----------------------------
737: \bibitem{Gogolin_1d_book}
738: A.O.~Gogolin, A.A.~Nersesyan, A.M.~Tsvelik,
739: {\it Bosonization and Strongly Correlated Systems}
740: (Cambridge University Press, Cambridge, 1998).
741: %--------------------------
742: 
743: \end{thebibliography}
744: 
745: 
746: \end{document}
747: 
748: