1: \documentclass[twocolumn,letter]{jpsj2}
2: \usepackage{bm,times,mathptmx}
3:
4: \title{
5: Charge-Density-Wave Formation in the Doped Two-Leg Extended Hubbard Ladder
6: }
7:
8: \author{Masahisa {\sc Tsuchiizu} and Yoshikazu {\sc Suzumura}}
9:
10:
11: \inst{
12: Department of Physics, Nagoya University, Nagoya 464-8602, Japan
13: }
14:
15: \abst{
16: We investigate electronic properties
17: of the doped two-leg Hubbard ladder
18: with both the onsite and the nearest-neighbor Coulomb repulsions,
19: by using the the weak-coupling renormalization-group method.
20: It is shown that,
21: for strong nearest-neighbor repulsions,
22: the charge-density-wave state coexisting with the
23: $p$-density-wave state becomes dominant fluctuation
24: where spins form intrachain singlets.
25: By increasing doping rate,
26: we have also shown that the effects of the nearest-neighbor repulsions
27: are reduced and
28: the system exhibits a quantum phase transition into
29: the $d$-wave-like (or rung-singlet) superconducting state.
30: We derive the effective fermion theory which describes the critical properties
31: of the transition point
32: with the gapless excitation of magnon.
33: The phase diagram of the two-leg ladder compound,
34: Sr$_{14-x}$Ca$_x$Cu$_{24}$O$_{41}$, is discussed.
35: }
36:
37: \kword{doped Hubbard ladder, intersite Coulomb repulsion,
38: spin gap, charge density wave, superconductivity}
39:
40:
41: \begin{document}
42: \maketitle
43:
44: Electronic properties on ladder systems have been studied intensively
45: both theoretically and experimentally, since
46: the superconducting (SC) state was discovered in
47: the self-doped two-leg ladder material Sr$_{14-x}$Ca$_x$Cu$_{24}$O$_{41}$
48: with $x \gtrsim 12$ under pressure over 3 GPa.\cite{Uehara,Nagata}
49: The substitution of Ca for Sr changes effectively the
50: hole-doping rate in the ladder Cu sites,
51: where the rate
52: varies monotonically from 0.07 to 0.25 with increasing $x$ from 0 to
53: 12. \cite{Osafune}
54: A characteristic feature is the presence of a gap in
55: magnetic excitations at temperature much higher than
56: the SC transition temperature.
57: \cite{Piskunov,Fujiwara}
58: Besides the SC state,
59: recent experimental studies have focused on
60: the charge dynamics in the slightly doped materials and
61: verified collective modes from
62: the sliding of the charge-density-wave (CDW) developed on ladder sites.
63: \cite{Kitano,Blumberg,Gorshunov,Vuletic}
64: A global phase diagram is obtained on the plane of $x$ and temperature
65: showing that the hole doping suppresses the CDW state
66: followed by the insulating state without the CDW order,
67: and then the high doping leads to the SC state under pressure.
68: \cite{Vuletic}
69: Quite recently the CDW collective modes are also suggested in the
70: highly doped material Sr$_2$Ca$_{12}$Cu$_{24}$O$_{41}$.\cite{Gozar}
71: Therefore it is of particular interest to investigate
72: the competition between the SC state and the CDW state in doped
73: ladder systems.
74:
75: From a theoretical point of view,
76: the origin of the spin gap in the ladder compounds
77: seems to be explained successfully for
78: both the undoped\cite{Dagotto,Schulz} and doped\cite{Schulz} cases,
79: and it is known that
80: the $d$-wave-like SC (SC$d$)
81: state appears in doped ladder systems. \cite{Schulz}
82: In addition, the charge-ordered state is also suggested
83: when intersite interactions are included. \cite{Vojta}
84: Further the competition
85: between the SC$d$ state and the charge-ordered or charge-density-wave (CDW)
86: state has been examined.\cite{Schulz,Vojta,Fradkin2002}
87: However, critical behavior is not yet fully understood.
88: In the present paper,
89: the possible scenario of the instability of the CDW state and the
90: competition between the CDW state and the
91: SC state are proposed
92: in the doped two-leg ladder of
93: the extended Hubbard model (EHM)
94: with nearest-neighbor repulsive interactions
95: by extending the previous analytical calculations
96: \cite{Tsuchiizu2002b,Fradkin2002}
97: to analyze the critical behavior in more detail.
98:
99:
100:
101: We consider the two-leg EHM given by $H=H_0+H_{\mathrm{int}}$.
102: The first term describes the hopping energies along and between legs:
103: %==========================================================
104: \begin{eqnarray}
105: H_0
106: \!\!\!\! &=& \!\!\!\!
107: - t_\parallel \sum_{j,\sigma,l}
108: (c_{j,l,\sigma}^\dagger \, c_{j+1,l,\sigma}^{}+\mathrm{H.c.})
109: \nonumber \\
110: && {} \!\!\!\! - t_\perp \sum_{j,\sigma}
111: (c_{j,1,\sigma}^\dagger \, c_{j,2,\sigma}^{}+\mathrm{H.c.})
112: ,
113: \label{eq:H0}
114: \end{eqnarray}
115: %==========================================================
116: where $c_{j,l,\sigma}$ annihilates an electron of spin
117: $\sigma(=\uparrow,\downarrow)$ on rung $j$ and leg $l(=1,2)$.
118: The Hamiltonian $H_{\mathrm{int}}$ denotes interactions between electrons:
119: %==========================================================
120: \begin{equation}
121: H_{\mathrm{int}}
122: =
123: U \sum_{j,l} n_{j,l,\uparrow} \, n_{j,l,\downarrow}
124: +V_\parallel \sum_{j,l} n_{j,l} \, n_{j+1,l}
125: +V_\perp \sum_{j} n_{j,1} \, n_{j,2},
126: \end{equation}
127: %==========================================================
128: where $U$ represents on-site repulsion and
129: $V_\parallel$ ($V_\perp$) represents intrachain (interchain)
130: nearest-neighbor repulsion with
131: $n_{j,l,\sigma}=c_{j,l,\sigma}^\dagger \, c_{j,l,\sigma}^{}$
132: and $n_{j,l}=n_{j,l,\uparrow}+n_{j,l,\downarrow}$.
133: The $H_0$ term is diagonalized by using the Fourier transform of
134: $c_{\sigma}(\bm{k})$
135: where $\bm{k}=(k_\parallel,k_\perp)$ with $k_\perp = 0$ or $\pi$.
136: The energy dispersion is given by
137: $\varepsilon(\bm{k}) = -2 t_\parallel \cos k_\parallel
138: - t_\perp \cos k_\perp$.
139: Here we consider the case with finite hole doping $\delta$
140: satisfying $t_\perp< 2t_\parallel \cos^2 \frac{\pi}{2}\delta$,
141: in which both the bonding ($k_\perp=0$) and the antibonding
142: ($k_\perp=\pi$) energy bands are partially filled and
143: the Fermi points are located at
144: $ k_{F,0}=\frac{\pi}{2}(1-\delta)+\lambda$ and at
145: $ k_{F,\pi}=\frac{\pi}{2}(1-\delta)-\lambda$ with
146: $\lambda \equiv \sin^{-1}
147: \left[t_\perp/\left(2t_\parallel\cos \frac{\pi}{2}\delta \right)\right]$.
148: We examine the case of small $\delta$ by
149: neglecting the differences in the Fermi velocities of the
150: bonding/antibonding band, i.e., $v_{F,0}=v_{F,\pi} (\equiv v_F)$.
151:
152:
153: Following the standard weak-coupling approach ($g$-ology),
154: the linearized kinetic energy is given by
155: $ H_0 = \sum_{\bm{k},p,\sigma}
156: v_F (pk_\parallel-k_{F,k_\perp}) \,
157: c_{p,\sigma}^\dagger (\bm{k}) \, c^{}_{p,\sigma} (\bm{k})$,
158: where the index $p=+/-$ denotes the right-/left-moving electron.
159: By introducing field operators by
160: $\psi_{p,\sigma,\zeta}(x)=
161: L^{-1/2}\sum_{k_\parallel} e^{ik_\parallel x}
162: c_{p,\sigma}(k_\parallel,k_\perp)$
163: with $\zeta=+(-)$ for $k_\perp=0(\pi)$
164: and $L$ being the system size,
165: the interactions near the Fermi points are rewritten as
166: $H_{\mathrm{int}}= (1/4)
167: \int dx
168: \sum_{p,\sigma}{\sum_{\zeta_i=\pm}}'
169: \mathcal{H}_{\mathrm{int}}$, where $\mathcal{H}_{\mathrm{int}}$ is given by
170: %==========================================================
171: \begin{eqnarray}
172: & & \!\!\!\!
173: g_{1(2)\parallel}^{\epsilon\bar\epsilon} \,
174: \psi_{p,\sigma,\zeta_1}^\dagger \,
175: \psi_{-p,\sigma,\zeta_2}^\dagger \,
176: \psi_{+(-)p,\sigma,\zeta_4}^{} \,
177: \psi_{-(+)p,\sigma,\zeta_3}^{}
178: \nonumber \\
179: &+& \!\!\!\!
180: g_{1(2)\perp}^{\epsilon\bar\epsilon} \,
181: \psi_{p,\sigma,\zeta_1}^\dagger \,
182: \psi_{-p,\bar\sigma,\zeta_2}^\dagger \,
183: \psi_{+(-)p,\bar\sigma,\zeta_4}^{} \,
184: \psi_{-(+)p,\sigma,\zeta_3}^{} ,
185: \quad
186: \label{eq:Hint_g-ology}
187: \end{eqnarray}
188: %==========================================================
189: and $\bar\sigma=\uparrow(\downarrow)$ for
190: $\sigma=\downarrow(\uparrow)$,
191: $\epsilon=\zeta_1\zeta_3$ and $\bar\epsilon=\zeta_1\zeta_2$.
192: The summation of the band index $\zeta_i$ ($i=1,\ldots,4$) is taken
193: under the condition $\zeta_1 \zeta_2 \zeta_3 \zeta_4 = +1$.
194: The coupling constants $g_{i\parallel}^{\epsilon\bar\epsilon}$
195: and $g_{i\perp}^{\epsilon\bar\epsilon}$
196: with $i=1 (2)$ corresponding to the backward (forward) scattering
197: are given by
198: $g_{i\parallel}^{\epsilon\bar\epsilon}=
199: (l_\epsilon V_\perp+ m_{i,\epsilon} V_\parallel)$ and
200: $g_{i\perp}^{\epsilon\bar\epsilon}
201: = (U + l_\epsilon V_\perp
202: + m_{i,\epsilon} V_\parallel)$
203: where $l_\pm = \pm 1$, $m_{1,+}=-2\cos\pi\delta \, \cos 2\lambda$,
204: $m_{1,-} =-2\cos\pi\delta$, $m_{2,+}=+2$, and $m_{2,-}=+2\cos 2\lambda$.
205: We neglect the umklapp scattering processes which become irrelevant for
206: finite doping case and also neglect forward scattering processes which
207: do not yield qualitative changes in the system. \cite{Emery}
208:
209: As possible states,
210: we consider the singlet $d$-wave superconducting (SC$d$) state,
211: the CDW state, and the $p$-density-wave (PDW) state. \cite{Tsuchiizu2002b}
212: The PDW state corresponds to the spin-Peierls state
213: in the limit of $\delta\to 0$.
214: The order parameter of the SC$d$ state is given by
215: $O_{\mathrm{SC}d} = N^{-1} \sum_{j}
216: (c_{j,1,\uparrow} \, c_{j,2,\downarrow}
217: - c_{j,1,\downarrow} \, c_{j,2,\uparrow})$,
218: while those of the density waves are
219: $ O_A = N^{-1}\sum_{\bm{k},\sigma}
220: f_A (\bm{k}) \, c_\sigma^\dagger (\bm{k}) \, c_\sigma^{} (\bm{k+Q})$,
221: with $\bm{Q}=\bigl(\pi(1-\delta),\pi\bigr)$,
222: $f_{\mathrm{CDW}}=1$ and
223: $f_{\mathrm{PDW}}=\sin k_\parallel$.
224: These operators are rewritten in terms of bosonic phase fields by
225: applying the Abelian bosonization method\cite{Emery,Gogolin_book}.
226: The field operators of the right- and left-moving electrons are written as
227: $\psi_{p,\sigma,\zeta}(x) =
228: \eta_{\sigma,\zeta}(2\pi a)^{-1/2} \,
229: \exp [ ipk_{F,k_\perp} x
230: + i p\, \varphi _{p,s,\zeta}(x) ] $
231: where $s=+$ for $\sigma=\uparrow$ and $s=-$ for
232: $\sigma=\downarrow$ and the fields satisfy
233: the commutation relations:
234: $[\varphi_{p,s,\zeta}(x),\varphi_{p,s',\zeta'}(x')]
235: = ip\pi \, \mathrm{sgn}(x-x') \,
236: \delta_{s,s'}\,\delta_{\zeta,\zeta'}$
237: and
238: $[\varphi_{+,s,\zeta},\varphi_{-,s',\zeta'}]
239: = i\pi \,\delta_{s,s'}\,\delta_{\zeta,\zeta'}$.
240: The Klein factors $\eta_{\sigma,\zeta}$
241: are introduced in order to retain the correct anticommutation
242: relations.\cite{Tsuchiizu2002b}
243: For calculating physical quantities, the field
244: $\varphi _{p,s,\zeta} $ is replaced by new bosonic fields:
245: $\phi_{\nu r}=(\phi_{\nu r}^+ + \phi_{\nu r}^-)$ and
246: $\theta_{\nu r}=(\phi_{\nu r}^+ - \phi_{\nu r}^-)$
247: where
248: $\varphi_{p,s,\zeta} =
249: (\phi_{\rho +}^p + \zeta \phi_{\rho -}^p
250: + s \phi_{\sigma +}^p + s \zeta \phi_{\sigma -}^p)$ with
251: $p=\pm$, $s=\pm$, and $\zeta=\pm$.
252: The phase fields $\phi_{\rho\pm}$ and $\phi_{\sigma\pm}$ represent charge
253: and spin fluctuations, respectively and the suffices $\pm$ refers to
254: the even and odd sectors. They satisfy
255: $[ \phi_{\nu r}(x), \theta_{\nu' r'}(x') ] =
256: -i \pi \Theta(-x+x')\delta_{r,r'}$
257: with $\Theta(x)$ being the Heaviside step function.
258: In terms of $\phi_{\nu r}$ and $\theta_{\nu r}$,
259: the order parameters $O=\int dx \mathcal{O}$ are given by
260: %==========================================================
261: \begin{subequations}
262: \begin{eqnarray}
263: &&
264: \mathcal{O}_{\mathrm{SC}d}
265: \propto
266: e^{i \theta_{\rho+}}
267: \cos \theta_{\rho-} \,
268: \cos \phi_{\sigma+} \,
269: \cos \phi_{\sigma-}
270: \nonumber \\ && \qquad\qquad
271: {} - i \,
272: e^{i \theta_{\rho+}}
273: \sin \theta_{\rho-} \,
274: \sin \phi_{\sigma+} \,
275: \sin \phi_{\sigma-}
276: ,
277: \\
278: &&
279: \mathcal{O}_{\mathrm{CDW}}
280: \propto
281: \cos \phi_{\rho+} \,
282: \sin \theta_{\rho-} \,
283: \cos \phi_{\sigma+} \,
284: \cos \theta_{\sigma-}
285: \nonumber , \\
286: && \qquad\qquad
287: {} -
288: \sin \phi_{\rho+} \,
289: \cos \theta_{\rho-} \,
290: \sin \phi_{\sigma+} \,
291: \sin \theta_{\sigma-} ,
292: \label{eq:order_CDW}
293: \\
294: &&
295: \mathcal{O}_{\mathrm{PDW}}
296: \propto
297: \cos \phi_{\rho+} \,
298: \cos \theta_{\rho-} \,
299: \sin \phi_{\sigma+} \,
300: \sin \theta_{\sigma-} ,
301: \nonumber \\
302: && \qquad\qquad
303: {} +
304: \sin \phi_{\rho+} \,
305: \sin \theta_{\rho-} \,
306: \cos \phi_{\sigma+} \,
307: \cos \theta_{\sigma-} .
308: \qquad\quad
309: \end{eqnarray}%
310: \label{order-parameters}%
311: \end{subequations}
312: %==========================================================
313:
314:
315: We can also rewrite
316: the Hamiltonian in terms of the bosonic phase variables.
317: In Eq.\ (\ref{eq:Hint_g-ology}),
318: the phase field $\phi_{\rho-}$ appears
319: in the form $\cos (2\phi_{\rho-}+4\lambda x)$.
320: Since we can safely assume that $t_\perp$ is a relevant perturbation
321: for $t_\perp$ being not so small \cite{Tsuchiizu1999,Tsuchiizu2001},
322: the term with $\cos (2\phi_{\rho-}+4\lambda x)$
323: would become irrelevant, and thus we discard it in the following.
324: We also neglect the $\cos 2\phi_{\sigma-} \, \cos 2\theta_{\sigma-}$ term
325: since its scaling dimension is larger than 2.
326: Then our Hamiltonian reduces to $H=\int dx \mathcal{H}$ with
327: %==========================================================
328: \begin{eqnarray}
329: {\cal H}
330: \!\!\!\! &=& \!\!\!\!
331: \frac{v_F}{\pi} \sum_{r=\pm}
332: \Bigl[
333: \sum_{p=\pm} \left( \partial_x \phi_{\rho r}^p \right)^2
334: + \frac{g_{\rho r}}{\pi v_F}
335: (\partial_x \phi_{\rho r}^+ )
336: (\partial_x \phi_{\rho r}^- )
337: \Bigr]
338: \nonumber \\ && \!\!\!\! {}
339: +
340: \frac{v_F}{\pi}
341: \sum_{r=\pm}
342: \Bigl[
343: \sum_{p=\pm} \left( \partial_x \phi_{\sigma r}^p \right)^2
344: - \frac{g_{\sigma r}}{\pi v_F}
345: \left(\partial_x \phi_{\sigma r}^+ \right)
346: \left(\partial_x \phi_{\sigma r}^- \right)
347: \Bigr]
348: \nonumber \\
349: &&\!\!\!\! {}
350: + \frac{1}{2\pi^2 a^2}
351: (
352: g_{\overline{c-},s+} \,
353: \cos 2 \phi_{\sigma+}
354: + g_{\overline{c-},s-}\,
355: \cos 2 \phi_{\sigma-}
356: \nonumber \\
357: && {} \qquad\qquad
358: + g_{\overline{c-},\overline{s-}}\,
359: \cos 2 \theta_{\sigma-}
360: ) \, \cos 2 \theta_{\rho-}
361: \nonumber \\
362: &&\!\!\!\! {}
363: + \frac{1}{2\pi^2 a^2}
364: (
365: g_{s+, s-}\,
366: \cos 2 \phi_{\sigma+} \,
367: \cos 2 \phi_{\sigma-}
368: \nonumber \\
369: && {} \qquad\qquad
370: + g_{s+,\overline{s-}}\,
371: \cos 2 \phi_{\sigma+}\,
372: \cos 2 \theta_{\sigma-}
373: ),
374: \label{eq:Hboson}
375: \end{eqnarray}
376: %==========================================================
377: where the coupling constants of the harmonic terms are given by
378: $g_{\rho (\sigma) r} = \sum_{\epsilon=\pm} f_{\rho(\sigma)r}^\epsilon
379: ( g_{2\parallel}^{+\epsilon} + (-) g_{2\perp}^{+\epsilon}
380: -g_{1\parallel}^{\epsilon\epsilon})/2 $
381: with $r=\pm$,
382: $f_{\rho+}^\epsilon=1$, $f_{\rho-}^\epsilon=\epsilon$,
383: $f_{\sigma+}^\epsilon=-1$ and $f_{\sigma-}^\epsilon=-\epsilon$.
384: The coupling constants of the nonlinear terms are
385: $g_{\overline{c-},s+}\equiv - g_{1\perp}^{-+}$,
386: $g_{\overline{c-},s-}\equiv - g_{2\perp}^{-+}$,
387: $g_{\overline{c-},\overline{s-}}\equiv
388: (g_{2\parallel}^{-+} - g_{1\parallel}^{-+})$,
389: $g_{s+,s-} \equiv g_{1\perp}^{++}$,
390: $g_{s+,\overline{s-}} \equiv g_{1\perp}^{--}$.
391: These nine coupling constants are not independent
392: since the global spin-rotation SU(2) symmetry leads to
393: \cite{Tsuchiizu2002b}
394: $(g_{\sigma+}+g_{\sigma-} - g_{s+,s-}) =
395: (g_{\sigma+}-g_{\sigma-} - g_{s+,\overline{s-}}) =
396: (g_{\overline{c-},s+} - g_{\overline{c-},s-}
397: -g_{\overline{c-},\overline{s-}}) = 0$.
398: We choose following six coupling constants:
399: %===================
400: \begin{subequations}
401: \begin{eqnarray}
402: && \hspace*{-.5cm}
403: g_{\overline{c-},st}
404: =
405: +U-V_\perp -2 V_\parallel \cos \pi\delta ,
406: , \\
407: && \hspace*{-.5cm}
408: g_{\overline{c-},ss}
409: = {} +U-V_\perp + 2V_\parallel(\cos \pi \delta + 2\cos 2\lambda), \qquad
410: \\
411: && \hspace*{-.5cm}
412: g_{\rho+} =
413: +U+2V_\perp +V_\parallel [4+\cos\pi\delta(1+\cos2\lambda)],
414: \\
415: && \hspace*{-.5cm}
416: g_{\rho-} =
417: -V_\perp - V_\parallel \cos \pi\delta (1- \cos 2\lambda),
418: \\
419: && \hspace*{-.5cm}
420: g_{\sigma+} =
421: +U - V_\parallel \cos\pi\delta (1+\cos 2\lambda ),
422: \\
423: && \hspace*{-.5cm}
424: g_{\sigma-} =
425: +V_\perp + V_\parallel \cos\pi\delta (1-\cos 2\lambda ).
426: \end{eqnarray}%
427: \label{eq:g}%
428: \end{subequations}
429: %===================
430: where
431: $g_{\overline{c-},st}
432: \equiv -g_{\overline{c-},s+}$ and
433: $ g_{\overline{c-},ss}
434: \equiv (-g_{\overline{c-},s-}+g_{\overline{c-},\overline{s-}})$.
435: The present model and the above treatment are quite similar to those
436: in Ref.\ \citen{Fradkin2002}.
437: However, the application of the renormalization-group (RG) method
438: to Eq.\ (\ref{eq:Hboson})
439: is complicated to estimate excitation gaps of spin modes properly.
440: Therefore,
441: we fermionize the spin part of Eq.\ (\ref{eq:Hboson}) \cite{Tsuchiizu2002b}
442: by introducing
443: spinless fermion fields
444: $\psi_{\pm,r}(x) =
445: \eta_r (2\pi a)^{-1/2}
446: \, \exp\left[ \pm i \,2\phi_{\sigma r}^{\pm}(x)\right]$
447: where $r=\pm$ and $\{\eta_r,\eta_{r'}\}=2\delta_{r,r'}$.
448: By using the SU(2) constraints and
449: the Majorana fermions $\xi^n$ ($n=1\sim 4$),
450: the equation (\ref{eq:Hboson}) is rewritten as
451: %===================
452: \begin{eqnarray}
453: \mathcal{H}
454: \hspace*{-.3cm}&=&\hspace*{-.3cm} {}
455: \frac{v_F}{\pi} \sum_{r}
456: \Bigl[
457: \sum_p \left(\partial \phi_{\rho r}^p \right)^2
458: + \frac{g_{\rho r}}{\pi v_F}
459: (\partial_x \phi_{\rho r}^+ )
460: (\partial_x \phi_{\rho r}^- )
461: \Bigr]
462: \nonumber \\ && {}
463: -i\frac{v_F}{2}
464: \left(
465: \bm{\xi}_+ \cdot \partial_x \bm{\xi}_+
466: - \bm{\xi}_- \cdot \partial_x \bm{\xi}_-
467: \right)
468: -\frac{g_{\sigma+}}{2} \,
469: \left(
470: \bm{\xi}_+ \cdot \bm{\xi}_-
471: \right)^2
472: \nonumber \\ && {}
473: -i\frac{v_F}{2}
474: \left(
475: \xi_+^4 \, \partial_x \xi_+^4
476: - \xi_-^4 \, \partial_x \xi_-^4
477: \right)
478: -g_{\sigma-} \,
479: \left( \bm{\xi}_+ \cdot \bm{\xi}_- \right)
480: \, \xi_+^4 \, \xi_-^4
481: \nonumber \\ && {}
482: - \frac{i}{2\pi a}
483: \left(
484: g_{\overline{c-},st} \,
485: \bm{\xi}_+ \cdot \bm{\xi}_-
486: +
487: g_{\overline{c-},ss} \,
488: \xi_+^4 \cdot \xi_-^4
489: \right) \,
490: \cos 2\theta_{\rho-} ,
491: \label{eq:Heff}
492: \end{eqnarray}
493: %===================
494: where
495: $\psi_{p,+}= (\xi_{p}^1+i\xi_{p}^2 )/\sqrt{2}$,
496: $\psi_{p,-} = (\xi_{p}^4+i\xi_{p}^3 )/\sqrt{2}$,
497: and $\bm{\xi}_p=(\xi_p^1,\xi_p^2,\xi_p^3)$.
498: Thus the effective theory for the spin sector becomes
499: O(3)$\times$Z$_2$ symmetric, as seen in the
500: isotropic Heisenberg \cite{Gogolin_book} and half-filled Hubbard ladder.
501: \cite{Tsuchiizu2002b}
502:
503:
504:
505:
506:
507: We investigate the low-energy behavior by using the perturbative
508: RG method with the lattice constant $a\to a e^{dl}$.
509: Following six scaling equations are obtained:
510: %===================
511: \begin{subequations}
512: \begin{eqnarray}
513: && \hspace*{-1.2cm}
514: \frac{d}{dl} G_{\rho-} =
515: - \frac{3}{4} G_{\overline{c-},st}^2
516: - \frac{1}{4} G_{\overline{c-},ss}^2
517: ,
518: \\
519: && \hspace*{-1.2cm}
520: \frac{d}{dl} G_{\sigma+} =
521: - G_{\sigma+}^2
522: - G_{\sigma-}^2
523: -\frac{1}{2} G_{\overline{c-},st}^2
524: ,
525: \\
526: &&\hspace*{-1.2cm}
527: \frac{d}{dl} G_{\sigma-} =
528: - 2 G_{\sigma+} \, G_{\sigma-}
529: - \frac{1}{2} G_{\overline{c-},st} \, G_{\overline{c-}ss}
530: ,
531: \\
532: &&\hspace*{-1.2cm}
533: \frac{d}{dl} G_{\overline{c-},st} =
534: - G_{\rho-} \, G_{\overline{c-},st}
535: - 2 G_{\sigma+} \, G_{\overline{c-},st}
536: - G_{\sigma-} \, G_{\overline{c-},ss} ,
537: \\
538: &&\hspace*{-1.2cm}
539: \frac{d}{dl} G_{\overline{c-},ss} =
540: - G_{\rho-} \, G_{\overline{c-},ss}
541: -3 G_{\sigma-} \, G_{\overline{c-},st}
542: ,
543: \end{eqnarray}
544: \end{subequations}
545: %===================
546: and $dG_{\rho+}/dl =0$ where $G(0)=g/(2\pi v_F)$.
547: We noted that these RG equations can be also derived directly from
548: Eq.\ (\ref{eq:Hboson}).
549: We analyze the RG equations numerically for
550: $ U>0$, $ V_{\parallel}>0 $ and
551: $ V_{\perp} > 0 $.
552: For small $V_{\perp}/U$ and $V_{\parallel}/U$,
553: the limiting behavior of RG equations is given by
554: $(G_{\rho-}^*,G_{\sigma+}^*,G_{\sigma-}^*,
555: G_{\overline{c-},st}^*,G_{\overline{c-},ss}^*)
556: =(-,-,-,+,+)$
557: which corresponds to
558: $(g_{\overline{c-},s+}^*,g_{\overline{c-},s-}^*,
559: g_{\overline{c-},\overline{s-}}^*,g_{s+, s-}^*, g_{s+,\overline{s-}}^*)
560: = (-,-,0,-,0)$ in Eq.\ (\ref{eq:Hboson}).
561: The relevant behavior of coupling constants implies that the phases are
562: locked in order to minimize
563: the cosine potential in Eq.\ (\ref{eq:Hboson}).
564: The positions of phase locking and the corresponding ground states
565: are summarized in Table \ref{table:phase}.
566: Since the $\theta_{\sigma-}$ field is conjugate to
567: $\phi_{\sigma-}$,
568: these two fields cannot be locked at the same time.
569: From Eq.\ (\ref{order-parameters}),
570: the nonvanishing order parameter is $\mathcal{O}_{\mathrm{SC}d}$.
571: Since the correlation function of the operator $e^{i\theta_{\rho+}}$ exhibits
572: power-law behavior,
573: we obtain that the SC$d$ fluctuation becomes quasi-long-range ordered
574: (quasi-LRO) in this case.
575: We note that the SC$d$ state moves to
576: the $D$-Mott or $D^\prime$-Mott state
577: in the limit of $\delta\to 0$.\cite{Tsuchiizu2002b}
578: For large
579: $ V_{\perp}/U$ and $ V_{\parallel}/U$,
580: the limiting behavior of RG equations is now given by
581: $(G_{\rho-}^*,G_{\sigma+}^*,G_{\sigma-}^*,
582: G_{\overline{c-},st}^*,G_{\overline{c-},ss}^*)
583: =(-,-,+,-,+)$, corresponding to
584: $(g_{\overline{c-},s+}^*,g_{\overline{c-},s-}^*,
585: g_{\overline{c-},\overline{s-}}^*,g_{s+, s-}^*, g_{s+,\overline{s-}}^*)$
586: $ = (+,0,+,0,-)$.
587: In this case,
588: the dominant order parameters are given by
589: $\mathcal{O}_{\mathrm{CDW}}$ and
590: $\mathcal{O}_{\mathrm{PDW}}$
591: both of which lead to the quasi-LRO with the same
592: exponent of the correlation functions.
593: We call this coexisting state the CDW+PDW state.
594:
595:
596:
597: In order to analyze the properties near the critical point of the transition
598: between the SC$d$ state and the CDW+PDW state,
599: we restrict ourselves to the case where the mass of the charge mode
600: ($\rho-$) is larger than those of the spin modes ($\sigma\pm$).
601: The $\theta_{\rho-}$ field is locked by the cosine
602: potential below the mass scale of the charge mode $m_{\rho-}$.
603: By replacing $\cos2\theta_{\rho-}$ with its average value
604: $c_{\overline{\rho-}}\equiv
605: \langle \cos 2\theta_{\rho-}\rangle$ in Eq.\ (\ref{eq:Hboson}),
606: the effective low-energy Hamiltonian for
607: the spin degrees of freedom is obtained as
608: \cite{Gogolin_book,Tsuchiizu2002b}
609: %===================
610: \begin{eqnarray}
611: \mathcal{H}_\sigma
612: \!\!\!\!&=&\!\!\!\! {}
613: -i\frac{v_F}{2}
614: \left(
615: \bm{\xi}_+ \cdot \partial_x \bm{\xi}_+
616: - \bm{\xi}_- \cdot \partial_x \bm{\xi}_-
617: \right)
618: - i m_t^0 \,
619: \bm{\xi}_+ \cdot \bm{\xi}_-
620: \nonumber \\ && \!\!\!\! {}
621: -i\frac{v_F}{2}
622: \left(
623: \xi_+^4 \, \partial_x \xi_+^4
624: - \xi_-^4 \, \partial_x \xi_-^4
625: \right)
626: - i m_s^0 \,
627: \xi_+^4 \, \xi_-^4
628: \nonumber \\ && \!\!\!\! {}
629: -\frac{g_{\sigma+}}{2} \,
630: \left(
631: \bm{\xi}_+ \cdot \bm{\xi}_-
632: \right)^2
633: - g_{\sigma-}
634: \left( \bm{\xi}_+ \cdot \bm{\xi}_- \right)
635: \, \xi_+^4 \, \xi_-^4
636: , \qquad
637: \label{eq:Heff_spin}
638: \end{eqnarray}
639: %===================
640: where $m_t^0$ and $m_s^0$ represent bare masses of the Majorana triplet
641: and singlet sector:
642: $m_t^0= (c_{\overline{\rho -}}/2\pi a)
643: ( U-V_\perp -2 V_\parallel \cos \pi\delta )$ and
644: $m_s^0 = (c_{\overline{\rho -}}/2\pi a)
645: [U-V_\perp + 2V_\parallel (\cos \pi \delta + 2\cos 2\lambda)]$.
646: The quantity $m_t^0$ ($m_s^0$) has physical meanings of the gap
647: in the magnon (soliton) excitation in the spin modes of the ladder.
648: \cite{Gogolin_book}
649: Equation (\ref{eq:Heff_spin}) is further analyzed
650: in terms of the following scaling equations for coupling constants:
651: %===================
652: \begin{subequations}
653: \begin{eqnarray}
654: &&
655: \frac{dG_{t}}{dl} =
656: G_{t}-2 G_{t}G_{\sigma+}-G_{s}G_{\sigma-},\\
657: &&
658: \frac{dG_{s}}{dl} =
659: G_{s} -3 G_{t}G_{\sigma-},\\
660: &&
661: \frac{dG_{\sigma+}}{dl} =
662: -G_{\sigma+}^2 - G_{\sigma-}^2 - G_{t}^2,\\
663: &&
664: \frac{dG_{\sigma-}}{dl} =
665: -2G_{\sigma+}G_{\sigma-}-G_{t}G_{s},
666: \end{eqnarray}%
667: \label{eq:dG_spin}%
668: \end{subequations}
669: %===================
670: where $G_{t}=m_t^0/v_F$, $G_{s}=m_s^0/ v_F$, and
671: $G_{\sigma\pm}=g_{\sigma\pm}/2\pi v_F$.
672: The couplings $G_s$ and $G_t$ are relevant,
673: while $G_{\sigma\pm}$ are marginal.
674: In Eq.\ (\ref{eq:dG_spin}), the $G_s$ term as a function
675: of $l$ increases rapidly compared with other $G's$
676: and becomes relevant at $l=l_s$ corresponding to the energy scale
677: of a gap in the Majorana singlet mode $m_s \approx t_\parallel
678: e^{-l_s}$, where we stop the calculation of Eq.\
679: (\ref{eq:dG_spin}).
680: The mode remained below the energy scale of $m_s$
681: is the Majorana triplet sector. The effective theory for this mode is
682: given by
683: $\mathcal{H}_\sigma^{\mathrm{eff}} =
684: -i\frac{1}{2}v_F
685: ( \bm{\xi}_+ \cdot \partial_x \bm{\xi}_+
686: - \bm{\xi}_- \cdot \partial_x \bm{\xi}_-)
687: - i m_t^s \,
688: \bm{\xi}_+ \cdot \bm{\xi}_-
689: -\frac{1}{2}g_{\sigma+}^s \, ( \bm{\xi}_+ \cdot \bm{\xi}_- )^2$
690: where $m_t^s=v_F [G_t(l_s)-G_{\sigma-}(l_s)]$ and $g_{\sigma+}^s=2\pi
691: v_F G_{\sigma+}(l_s)$.
692: Then we solve the RG equations given by
693: $dG_t/dl=G_t-2G_tG_{\sigma+}$ and
694: $dG_{\sigma+}/dl=-2G_{\sigma+}^2-G_t^2$ with the initial conditions
695: $G_t(l_s)=m_t^s/v_F$ and $G_{\sigma+}(l_s)=g_{\sigma+}^s/2\pi v_F$.
696: We easily find that these RG equations have two stable fixed points
697: $(G_t^*,G_{\sigma+}^*)=(+\infty,-\infty)$ and $(-\infty,-\infty)$,
698: corresponding to the SC$d$ state and the CDW+PDW state, respectively,
699: where the magnitude of the gap in the Majorana triplet sector
700: can be estimated from $ m_t \approx t_\parallel e^{-l_t}
701: \mathrm{sgn}(G_t^*)$
702: where $l_t$ is determined by $|G_t(l_t)|=1$
703: (see Table \ref{table:phase}).
704: There are also
705: two unstable fixed points
706: $(G_t^*,G_{\sigma+}^*)=(0,0)$ and $(0,-\infty)$, corresponding
707: to the second-order and first-order phase transitions,
708: \cite{Tsuchiizu2002b}
709: while only the former transition is obtained in the present
710: numerical calculation.
711: %===================
712: \begin{table}[t]
713: \caption{
714: Possible states and the corresponding pattern of phase locking
715: where $I$ is an integer and
716: the symbol $*$ indicates an unlocked bosonic phase field.
717: The signs $+$ and $-$ denote those for
718: renormalized masses $m_t$ and $m_s$ where
719: we have assumed $I$ being even number, i.e., $c_{\overline{\rho-}}>0$.
720: }
721: \label{table:phase}
722: \begin{tabular}{l|cccc|cc}
723: \hline\hline
724: State & $\langle\theta_{\rho-}\rangle$ &
725: $\langle\phi_{\sigma+}\rangle$ &
726: $\langle\phi_{\sigma-}\rangle$ &
727: $\langle\theta_{\sigma-}\rangle$ & $m_t$ & $m_s$ \\ \hline
728: SC$d$ & $\frac{\pi}{2} I$ & $\frac{\pi}{2}I$ & $\frac{\pi}{2}I$ & $*$ &
729: $+$ & $+$ \\
730: CDW + PDW & $\frac{\pi}{2} I$ & $\frac{\pi}{2}(I+1)$ &
731: $*$ & $\frac{\pi}{2}(I+1)$ &
732: $-$ & $+$
733: \\
734: \hline\hline
735: \end{tabular}
736: \end{table}
737: %===================
738:
739: From the numerical integration of the RG equations,
740: we obtain the ground-state phase diagram shown in Fig.\ 1.
741: The SC$d$ state (the CDW+PDW state) is obtained for
742: $ V_{\parallel}/U + V_{\perp}/U \gtrsim 0.4 ( \lesssim 0.4)$.
743: The SC$d$ state, on the one hand, is stabilized by the on-site repulsive
744: interaction, which
745: segregates up-spin from down-spin on the same site
746: and leads to the singlet pairing on a rung.
747: On the other hand, the CDW+PDW state is obtained
748: due to the nearest-neighbor repulsive
749: interactions, which induce density wave leading
750: to the singlet state on the same site or chain.
751: The effect of $V_{\parallel}$ is slightly larger than that of $V_{\perp}$
752: although both the intersite interactions have essentially the same effect of
753: inducing the CDW+PDW state.
754: In Fig.\ 2, the change from the CDW+PDW state to the SC$d$ state is shown
755: with increasing the doping $\delta (> 0.05)$.
756: The novel aspect of the present paper is the competition induced by the
757: doping which reduces the effect of only $V_{\parallel}$
758: as shown in Eq.\ (\ref{eq:g}).
759: In the inset, we show
760: the respective masses estimated
761: from $|m_a| \approx t_\parallel \, \exp(-l_a)$ ($a= t$, $s$, $\rho-$)
762: by noting that the corresponding coupling constant $|G_a|$ becomes
763: of the order of unity at $l = l_a$.
764: Our system exhibits a \textit{second-order} phase transition
765: and
766: the magnon excitation gap vanishes at the quantum critical point (QCP).
767: The critical property for the Majorana triplet sector,
768: which differs from that the conventional Tomonaga-Luttinger liquid,
769: is described by the SU(2)$_2$ Wess-Zumino-Novikov-Witten model with
770: the central charge $c=3/2$.\cite{Gogolin_book}
771: %===================
772: \begin{figure}[tb]
773: \begin{center}
774: \includegraphics[width=6.5cm]{fig1}
775: \vspace{-3mm}
776: \caption[]{
777: The ground-state phase diagram on the plane of $V_\parallel/U$ and
778: $V_\perp/U$, with $U/t_\parallel=2$, $\delta=0.1$, and $t_\perp=t_\parallel$.
779: }
780: \end{center}
781: \label{fig:phase}
782: \end{figure}
783: %===================
784:
785:
786: %===================
787: \begin{figure}[tb]
788: \begin{center}
789: \includegraphics[width=6.5cm]{fig2}
790: \end{center}
791: \caption{
792: The doping dependence of the magnon spin gap $m_t$ with $U/t_\parallel=2$,
793: $V_\parallel/U=V_\perp/U=0.25$, and $t_\perp=t_\parallel=1$.
794: In the inset, the doping dependences of $m_{\rho-}$, $m_s$, and $m_t$ are
795: are shown.
796: }
797: \label{gap}
798: \end{figure}
799: %===================
800:
801:
802:
803: In the present paper, by applying the weak-coupling RG method to the
804: EHM on two-leg ladder, we have shown that
805: the doping $\delta$ suppresses the CDW+PDW quasi-LRO state
806: and yields the system to the QCP, and that the SC$d$
807: quasi-LRO state is stabilized at further doping.
808: Here we discuss the experimental results of the two-leg ladder compound
809: Sr$_{14-x}$Ca$_x$Cu$_{24}$O$_{41}$.
810: The phase diagram of Sr$_{14-x}$Ca$_x$Cu$_{24}$O$_{41}$
811: obtained in Ref.\ \citen{Vuletic} resembles
812: our phase diagram of Fig.\ \ref{gap}
813: if the magnitude of gap $|m_t|$ is regarded as the transition temperature.
814: On closer look,
815: our phase diagram is contrast to the features of
816: Sr$_{14-x}$Ca$_x$Cu$_{24}$O$_{41}$ that
817: the resistivity above the transition
818: temperatures shows an insulating behavior and
819: there is no experimental evidence of the QCP
820: between the CDW state and the SC state.
821: In order to explain the phase diagram of Sr$_{14-x}$Ca$_x$Cu$_{24}$O$_{41}$,
822: the dimensionality effect and/or the disorder effect
823: has been discussed.\cite{Vuletic}
824: The quantum critical behavior would be smeared out
825: by these effects, which
826: are not taken into account in the present paper.
827: However it will be still interesting to examine the competing region
828: in the sense that the magnon gap would become extremely
829: small and anomalous behavior can be expected
830: at temperatures higher than characteristic energies of
831: the disorder and the dimensionality.
832: We note that
833: the origin of the high temperature insulating phase is still unknown and
834: the analysis is left for a future study.
835:
836:
837:
838:
839:
840: \acknowledgements
841: M.T.\ thanks A.\ Furusaki, H.\ Tsunetsugu, N.\ Fujiwara and H.\ Kitano
842: for valuable discussions.
843: This work was supported by a Grant-in-Aid for
844: Scientific Research on Priority Areas of Molecular Conductors
845: (No. 15073213) from
846: the Ministry of Education, Culture, Sports, Science and Technology,
847: Japan.
848:
849:
850:
851: \begin{thebibliography}{}
852:
853: %-------
854: \bibitem{Uehara}
855: M.\ Uehara, T.\ Nagata, J.\ Akimitsu, H.\ Takahashi,
856: N.\ M\^{o}ri, and K.\ Kinoshita:
857: J.\ Phys.\ Soc.\ Jpn.\ \textbf{65} (1996) 2764;
858: %-------
859: \bibitem{Nagata}
860: T.\ Nagata, M.\ Uehara, J.\ Goto, N.\ Komiya, J.\ Akimitsu, N.\ Motoyama,
861: H.\ Eisaki, S.\ Uchida, H.\ Takahashi, T.\ Nakanishi and N.\ M\^{o}ri,
862: Physica C \textbf{282-287} (1997) 153.
863: %-------
864: \bibitem{Osafune}
865: T. Osafune, M. Motoyama, H. Eisaki and S. Uchida:
866: Phys.\ Rev.\ Lett.\ \textbf{78} (1997) 1980.
867: %-------
868: \bibitem{Piskunov}
869: Y.\ Piskunov, D.\ J\'erome, P.\ Auban-Senzier, P.\ Wzietek,
870: C.\ Bourbonnais, U.\ Ammerhal, G.\ Dhalenne, and A.\ Revcolevschi:
871: Eur.\ Phys.\ J.\ B \textbf{24} (2001) 443.
872: %-------
873: \bibitem{Fujiwara}
874: N.\ Fujiwara, N.\ M\^{o}ri, Y.\ Uwatoko, T.\ Matsumoto,
875: N.\ Motoyama, and S.\ Uchida,
876: Phys.\ Rev.\ Lett.\ \textbf{90} (2003) 137001.
877: %-------
878: \bibitem{Kitano}
879: H.\ Kitano, R.\ Inoue, T.\ Hanaguri, A.\ Maeda, N.\ Motoyama,
880: M.\ Takaba, K.\ Kojima, H.\ Eisaki, and S.\ Uchida:
881: Europhys.\ Lett.\ \textbf{56} (2001) 434.
882: %-------
883: \bibitem{Blumberg}
884: G.\ Blumberg, P.\ Littlewood, A.\ Gozar, B.S.\ Dennis,
885: N.\ Motoyama, H.\ Eisaki, and S.\ Uchida:
886: Science \textbf{297} (2002) 584.
887: %-------
888: \bibitem{Gorshunov}
889: B.\ Gorshunov, P.\ Haas, T.\ R\~o\~om, M.\ Dressel,
890: T.\ Vuletic, B.\ Hamzic, S.\ Tomic, J.\ Akimitsu, and T.\ Nagata:
891: Phys.\ Rev.\ B \textbf{66} (2002) 060508.
892: %-------
893: \bibitem{Vuletic}
894: T.\ Vuleti\'c, B.\ Korin-Hamzi\'c, S.\ Tomic,
895: B.\ Gorshunov, P.\ Haas, T.\ R\~o\~om, M.\ Dressel,
896: J.\ Akimitsu, and T.\ Nagata:
897: Phys.\ Rev.\ Lett.\ \textbf{90} (2003) 257002.
898: %-------
899: \bibitem{Gozar}
900: A.\ Gozar, G.\ Blumberg, P.B.\ Littlewood, B.S.\ Dennis,
901: N.\ Motoyama, H.\ Eisaki, and S.\ Uchida,
902: Phys.\ Rev.\ Lett.\ \textbf{91} (2003) 087401.
903: %-------
904: \bibitem{Dagotto}
905: For a review,
906: E.\ Dagotto and T.M.\ Rice:
907: Science \textbf{271} (1996) 618.
908: %-------
909: \bibitem{Schulz}
910: H.J.\ Schulz:
911: \textit{Correlated Fermions and Transport in Mesoscopic Systems},
912: edited by T.\ Martin, G.\ Montambaux, and T.\ Tr\^an Thanh V\^an
913: (Editions Fronti\`eres, Gif-sur-Yvette, France, 1996), p.\ 81,
914: and references therein.
915: %-------
916: \bibitem{Vojta}
917: M.\ Vojta, R.E.\ Hetzel, and R.M.\ Noack,
918: Phys.\ Rev.\ B \textbf{60} (1999) R8417.
919: %-------
920: \bibitem{Tsuchiizu2002b}
921: M.\ Tsuchiizu and A.\ Furusaki:
922: Phys.\ Rev.\ B \textbf{66} (2002) 245106.
923: %-------
924: \bibitem{Fradkin2002}
925: C.\ Wu, W.V.\ Liu, and E.\ Fradkin:
926: Phys.\ Rev.\ B \textbf{68} (2003) 115104.
927: %-------
928: \bibitem{Emery}
929: V.J.\ Emery,
930: \textit{Highly Conducting One-Dimensional Solids},
931: edited by J.\ Devreese, R.\ Evrard, and V.\ van Doren
932: (Plenum, New York, 1979), p. 247.
933: %-------
934: \bibitem{Gogolin_book}
935: A.O.\ Gogolin, A.A.\ Nersesyan, and A.M.\ Tsvelik:
936: \textit{Bosonization and Strongly Correlated Systems}
937: (Cambridge University Press, Cambridge, 1998).
938: %-------
939: \bibitem{Tsuchiizu1999}
940: M.\ Tsuchiizu and Y.\ Suzumura:
941: Phys.\ Rev.\ B \textbf{59} (1999) 12326.
942: %-------
943: \bibitem{Tsuchiizu2001}
944: M.\ Tsuchiizu, P.\ Donohue, Y.\ Suzumura, and T.\ Giamarchi:
945: Eur.\ Phys.\ J.\ B \textbf{19} (2001) 185.
946:
947: \end{thebibliography}
948:
949: \end{document}
950: