1: %\documentclass[preprint,prl,aps]{revtex4}
2: \documentclass[twocolumn,prl,aps,showpacs]{revtex4}
3: \usepackage{graphicx}
4:
5:
6: %\documentclass[twocolumn,prl,aps,showpacs]{revtex4}
7: %\usepackage{graphicx}
8: %\documentstyle[epsfig,twocolumn,prl,aps]{revtex}
9:
10: \begin{document}
11:
12: \title{\bf Low-energy Effective Theory for Spin Dynamics of Fluctuating Stripes}
13:
14: \author{Chi-Ho Cheng}
15: \email{phcch@phys.sinica.edu.tw} \affiliation{Institute of
16: Physics, Academia Sinica, Taipei, Taiwan} \affiliation{Institute
17: of Theoretical Physics, Chinese Academy of Sciences, Beijing}
18:
19: \date{\today}
20:
21:
22: \begin{abstract}
23: We derive an effective Hamiltonian for spin dynamics of
24: fluctuating smectic stripes from the {\it t-J} model in the weak
25: coupling limit $t\gg J$. Besides the modulation of spin magnitude,
26: the high energy hopping term would induce a low-energy
27: anti-ferromagnetic interaction between two neighboring ``blocks of
28: spins". Based on the effective Hamiltonian, we applied the linear
29: spin-wave theory and found that the spin-wave velocity is almost
30: isotropic for $\rm La_{2-x}Sr_x CuO_4$ unless the structural
31: effect is considered. The intensity of the second harmonic mode is
32: found to be about 10\% to that of the fundamental mode.
33: \end{abstract}
34:
35: \pacs{75.30.Fv, 75.30.Ds, 75.10.Jm, 75.50.Ee}
36:
37: \maketitle
38:
39: \vspace{-0.10cm}
40:
41: There's still a lot of interest on stripe physics
42: \cite{review,review2}. The stripe modulation was found in cuprate
43: superconductor \cite{tranquada} of the sample $\rm La_{2-x}Sr_x
44: CuO_4$ (LSCO) in the low-temperature orthorhombic (LTO) structure.
45: It is also consistent with the incommensurate magnetic peaks
46: observed in the inelastic neutron scattering experiments
47: \cite{incomm-expt}, in which the peak shift from $(\pi,\pi)$
48: towards $(\pi,0)$ with the derivation very near to $2\pi\delta$ at
49: doping concentration $\delta$. By partial substitution of Nd for
50: La, the LTO lattice structure is distorted to the low-temperature
51: tetragonal (LTT) structure \cite{crawford}, in which the
52: horizontal stripe is enhanced and the fluctuating (dynamic) stripe
53: becomes more ordered (static).
54:
55: Theoretically, by the mean-field analysis of the single-band
56: Hubbard model \cite{hubbard}, there is a possibility that the
57: stripe phase is formed. On the other hand, by employing the
58: Schwinger-boson mean-field theory \cite{auerbach} to the {\it t-J}
59: model, it is found that the spiral spin state \cite{incomm-th} can
60: also give the deviation from $(\pi,\pi)$ upon doping. However, it
61: was studied by many approaches that the uniform phase is unstable
62: towards phase separation in the range of interest ratio {\it t/J}
63: \cite{phase-separation}. One of the consequence of the phase
64: separation is to form stripes, in order to be consistent with the
65: neutron scattering experiments. Because of the stripe fluctuation
66: around the hole domain, the spins across the hole domain should be
67: anti-parallel. By considering just a small transverse fluctuation
68: of the stripe, it was found that two neighboring spins across the
69: hole domain feels an anti-ferromagnetic interaction of coupling
70: $J$ \cite{zachar02}.
71:
72: %Furthermore, the idea of this kind of phase separation is one of
73: %the candidate to explain the normal state behavior of cuprate
74: %superconductors \cite{emery} in which the Coulomb interaction
75: %plays an important role.
76:
77: Although there is still controversy about the stability of stripes
78: \cite{stability}, in this paper, the fact that the fluctuating
79: smectic stripe phase is stable is our assumption.
80: %(rigorous treatment is to include Coulomb interaction to stablize it)
81: By including the Coulomb repulsion (which is neglected in the {\it
82: t-J} model), the holes can only phase separate at microscopic
83: scale instead of full phase separation. In order to balance the
84: hole kinetic energy and the Coulomb repulsion, stripes is the
85: simplest solution among those inhomogeneous states at microscopic
86: scale.
87:
88: In the following, we are going to derive the low-energy spin
89: dynamics from the 2D {\it t-J} model of the site-centered stripe
90: along $y$-direction. The period of hole domain is $2R$ where $R\gg
91: 1$ in our model.
92:
93: The hopping term $H_t$ in the {\it t-J} model is written as
94: \begin{eqnarray}
95: H_t &=& H_t^{\perp} + H_t^{\parallel} \nonumber \\
96: &=& \sum_\alpha \left( H_t^{\perp,\alpha} +
97: H_t^{\parallel,\alpha} \right)
98: \end{eqnarray}
99: where $H_t^{\perp}$ is the transverse fluctuation of vertical
100: stripes, and $H_t^{\parallel}$ is the 1D kinetic motion along the
101: vertical stripes. Now
102: \begin{eqnarray} \label{htalpha}
103: H_t^{\perp,\alpha} &=& -t \sum_{\sigma} {\sum_{i}}'
104: \left(
105: c_{i+\hat x,\alpha,\sigma}^\dagger c_i + c_{i+2\hat x,\alpha,\sigma}^\dagger c_{i+\hat x,\alpha,\sigma} +
106: \ldots \right. \nonumber \\
107: && \left. + c_{i+R\hat x,\alpha,\sigma}^\dagger c_{i+(R-1)\hat x,\alpha,\sigma}
108: \right) + {\rm h.c.}
109: \end{eqnarray}
110: where $\alpha$ label the smectic stripes and $i_x$ is summed over
111: the super-cell of period $2R$. And also
112: \begin{eqnarray}
113: H_t^{\parallel,\alpha} = -t \sum_{\sigma} {\sum_{i}}'
114: c_{i+\hat y,\alpha,\sigma}^\dagger c_{i,\alpha,\sigma}
115: + {\rm h.c.}
116: \end{eqnarray}
117:
118: In the weak coupling limit, $t \gg J$, the fluctuating stripe
119: induces a charge-density wave along $x$-direction. There's two
120: almost degenerate ground states for $H_t^{\perp,\alpha}$,
121: $|\psi_{\rm S}\rangle$ and $|\psi_{\rm T}\rangle$, which are
122: expressed as
123: \begin{eqnarray} \label{psi_st}
124: |\psi_{\rm S}\rangle &=& \frac{1}{\sqrt{2}} \left(
125: |\psi_1\rangle - |\psi_2\rangle \right) \\
126: |\psi_{\rm T}\rangle &=& \frac{1}{\sqrt{2}} \left(
127: |\psi_1\rangle + |\psi_2\rangle \right)
128: \end{eqnarray}
129: where
130: \begin{eqnarray}
131: |\psi_1\rangle &=&
132: a_0 |\cdot\cdot \downarrow \uparrow \circ \downarrow \uparrow \cdot\cdot\rangle
133: + a_1 |\cdot\cdot \downarrow \uparrow \downarrow \circ \uparrow
134: \cdot\cdot \rangle
135: \nonumber \\
136: && + a_{-1} |\cdot\cdot \downarrow \circ \uparrow \downarrow \uparrow \cdot\cdot \rangle + \ldots \\
137: |\psi_2\rangle &=&
138: a_0 |\cdot\cdot \uparrow \downarrow \circ \uparrow \downarrow
139: \cdot\cdot\rangle
140: + a_1 |\cdot\cdot \uparrow \downarrow \uparrow \circ \downarrow
141: \cdot\cdot \rangle
142: \nonumber \\
143: && + a_{-1} |\cdot\cdot \uparrow \circ \downarrow \uparrow \downarrow \cdot\cdot \rangle
144: + \ldots
145: \end{eqnarray}
146: The basis of $|\psi_1\rangle$ and $|\psi_2\rangle$ are different
147: by the up and down spins. The corresponding coefficients are
148: equal. Two subspaces $\{|\psi_1\rangle\}$ and $\{|\psi_2\rangle\}$
149: are orthogonal to each other. By Eq.(\ref{psi_st}), which is an
150: orthogonal transformation from $|\psi_1\rangle$ and
151: $|\psi_2\rangle$ to $|\psi_{\rm S}\rangle$ and $|\psi_{\rm
152: T}\rangle$, two subspaces ${\cal H}_S$ and ${\cal H}_T$ spanned by
153: $\{|\psi_{\rm S}\rangle\}$ and $\{|\psi_{\rm T}\rangle\}$
154: respectively are also orthogonal to each other. Eq.(\ref{htalpha})
155: can be written in the following matrix form under the above basis,
156: {\it i.e.},
157: \begin{eqnarray}
158: H_t^{\perp,\alpha} = (-t) \left(\begin{array}{ccccc}
159: 0 & 1 & 1 & 0 & \ldots \\
160: 1 & 0 & 0 & 1 & \ldots \\
161: 1 & 0 & 0 & 0 & \ldots \\
162: 0 & 1 & 0 & 0 & \ldots \\
163: \vdots & \vdots & \vdots & \vdots & \ddots
164: \end{array}\right) \bigotimes (-t)
165: \left(\begin{array}{ccccc}
166: 0 & 1 & 1 & 0 & \ldots \\
167: 1 & 0 & 0 & 1 & \ldots \\
168: 1 & 0 & 0 & 0 & \ldots \\
169: 0 & 1 & 0 & 0 & \ldots \\
170: \vdots & \vdots & \vdots & \vdots & \ddots
171: \end{array}\right) \nonumber \\
172: \end{eqnarray}
173: where the two identical $(2R+1)$-dimensional matrices correspond
174: to two orthogonal subspaces spanned by $\{|\psi_{\rm S}\rangle\}$
175: and $\{|\psi_{\rm T}\rangle\}$. Simultaneously, the Heisenberg
176: term in the {\it t-J} model is written in the matrix form
177: \begin{eqnarray}
178: H_J^{\perp,\alpha} = (-\frac{3J}{4}) \left(\begin{array}{ccccc}
179: 0 & 0 & 0 & 0 & \ldots \\
180: 0 & 1 & 0 & 0 & \ldots \\
181: 0 & 0 & 1 & 0 & \ldots \\
182: 0 & 0 & 0 & 1 & \ldots \\
183: \vdots & \vdots & \vdots & \vdots & \ddots
184: \end{array}\right) \bigotimes (\frac{J}{4}) \left(\begin{array}{ccccc}
185: 0 & 0 & 0 & 0 & \ldots \\
186: 0 & 1 & 0 & 0 & \ldots \\
187: 0 & 0 & 1 & 0 & \ldots \\
188: 0 & 0 & 0 & 1 & \ldots \\
189: \vdots & \vdots & \vdots & \vdots & \ddots
190: \end{array}\right) \nonumber \\
191: \end{eqnarray}
192:
193:
194:
195: Now because two subspaces ${\cal H}_{\rm S}$ and ${\cal H}_{\rm
196: T}$ are orthogonal to each other, the first-order correction to
197: the ground state energy due to perturbed term $H_J^{\perp,\alpha}$
198: is
199: \begin{eqnarray} \label{ergs}
200: E_{\rm S}^{(1)} &=& \langle \psi_{\rm
201: S}|H_J^{\perp,\alpha}|\psi_{\rm S}\rangle + O(J^2/t) \\
202: \label{ergt} E_{\rm T}^{(1)} &=& \langle \psi_{\rm
203: T}|H_J^{\perp,\alpha}|\psi_{\rm T}\rangle + O(J^2/t)
204: \end{eqnarray}
205: and no correction to the wavefunction since two subspaces are
206: orthogonal. Solving the eigenproblem for $H_t^{\perp,\alpha}$
207: gives the transverse density $\rho(x)$ in $(-R,R)$ to the leading
208: order can be obtained from the
209: \begin{eqnarray} \label{rho}
210: \rho(x) &=& |\langle x|\psi_{\rm S}\rangle|^2 = |\langle
211: x|\psi_{\rm T}\rangle|^2 \nonumber \\ &=& \frac{\delta}{1+2/\pi}
212: \left( 1+ \cos(\frac{\pi x}{2R}) \right)
213: \end{eqnarray}
214: and repeat for a period $2R$. It would be convenient to define
215: $\vec q_0 =(q_0,0)$, where $q_0 = \pi/(2R)$. The charge-density
216: wave induced by stripe fluctuation is
217: \begin{eqnarray}
218: \rho(\vec r) = \delta + \rho_1 \cos(2 \vec q_0 \cdot \vec r)
219: + \rho_2 \cos(4 \vec q_0 \cdot \vec r) +
220: \ldots
221: \end{eqnarray}
222: where $\rho_1=\frac{4\delta}{3(2+\pi)}$ and
223: $\rho_2=-\frac{4\delta}{15(2+\pi)}$.
224:
225: Eqs.(\ref{ergs})-(\ref{ergt}) tells that two almost degenerate
226: states are split by an energy difference $E_{\rm T}^{(1)}-E_{\rm
227: S}^{(1)}$, and hence the energy difference between two ``block of
228: spins" is
229: \begin{eqnarray}
230: J' = 2R \delta(E_{\rm T}^{(1)}-E_{\rm S}^{(1)})
231: = \frac{2J \delta R^2 }{R+1}
232: \end{eqnarray}
233: Substitution of $q_0=2\pi\delta$ gives $J'=J/(2(1+4\delta))$.
234: Notice also that the energy difference per site between anti-phase
235: and in-phase becomes $J'/(2R)=J \delta /(1+4\delta)$, which is
236: almost linear in $\delta$ for $\delta\ll 1$. For $\delta = 1/8$,
237: energy difference is $J/12 \simeq 130K$ (take $J=135meV$), which
238: is consistent with the observation of the incommensurate peak up
239: to around $100K$ \cite{incomm-expt}.
240:
241: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
242: % spin2.eps
243:
244:
245: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
246:
247: There is then an effective anti-ferromagnetic coupling term
248: between two neighboring blocks of spins. The anti-phase described
249: by $|\psi_{\rm S}\rangle$ is of lower energy $J'$ than the
250: in-phase by $|\psi_{\rm T}\rangle$. Notice that it is different
251: from the case that two neighboring spins across the hole domain
252: interacts with an anti-ferromagnetic coupling
253: \cite{zachar02,otherHam}. The illustration of how to flip a block
254: of spins is shown in Fig.\ref{spin4.eps}. Several mean-field type
255: studies neglecting this low-energy interaction \cite{ogata} cannot
256: distinguish anti-phase and in-phase.
257:
258: Because of no-double occupancy at every site, the spin magnitude
259: also form a period of $2R$, in which
260: \begin{eqnarray}
261: S(\vec r) &=& \frac{1}{2}(1-\rho(\vec r)) \nonumber \\ &=& S_0 +
262: S_1 \cos(2 \vec q_0 \cdot \vec r) + S_2 \cos(4 \vec q_0 \cdot \vec
263: r ) + \ldots
264: \end{eqnarray}
265: where $S_0 = \frac{1}{2}(1-\delta)$, $S_1 =
266: -\frac{2\delta}{3(2+\pi)}$, and $S_2=\frac{2\delta}{15(2+\pi)}$.
267:
268: Because of the anti-ferromagnetic interaction between two blocks
269: of spins across the hole stripes, we can determine {\it
270: classically} the $z$-component of the spin $S^z(\vec r)$ which
271: gives
272: \begin{eqnarray} \label{sz}
273: S^z(\vec r) {\rm e}^{i\vec Q \cdot \vec r} = S^z_1 \sin(\vec q_0
274: \cdot \vec r ) + S^z_3 \sin(3 \vec q_0 \cdot \vec r) + \ldots
275: \end{eqnarray}
276: where $\vec Q = (\pi,\pi)$,
277: $S^z_1=\frac{2}{\pi}-\frac{3\delta}{2+\pi}$, and
278: $S^z_3=\frac{2}{3\pi}-\frac{5\delta}{3(2+\pi)}$. Notice that the
279: components $S^z_2$, $S^z_4$, and etc vanish because of the
280: anti-phase symmetry. Including the effect from lattice distortion
281: gives the same result. The non-vanishing higher harmonic should
282: have wavevector $3q_0$, $5q_0$, and etc. The schematic diagram of
283: the classical spin state is shown in Fig.\ref{spin2.eps}.
284:
285: %Eq.(\ref{sz}) has an important implication to neutron scattering
286: %experiment. The neutron scattering peak occurs at $\vec k = \vec Q
287: %\pm \vec q_0$.
288:
289: Estimated up to the order of magnitude, the neutron scattering
290: intensity is more or less proportional to
291: \begin{eqnarray}
292: && \int d{\vec r} {\rm e}^{-\vec k \cdot \vec r}
293: \langle S^z(\vec r) S^z(0) \rangle \nonumber \\
294: &\propto& (S^z_1)^2 \left[\delta(\vec k-\vec Q+\vec q_0)+\delta(\vec k-\vec Q-\vec
295: q_0)\right] \nonumber \\
296: && + (S^z_3)^2 \left[\delta(\vec k-\vec Q+ 3 \vec q_0)+\delta(\vec k-\vec Q- 3\vec
297: q_0)\right] \nonumber \\ && + \ldots
298: \end{eqnarray}
299: For $\delta\ll 1$, $(S^z_3/S^z_1)^2 \simeq 1/9$. Substitute
300: $\delta = 1/8$, the ratio is 0.093. In general, the ratio of the
301: intensities of the second harmonic ($\vec Q \pm 3\vec q_0$) to
302: that of the fundamental mode ($\vec Q \pm \vec q_0$) is about
303: 10\%. However, the signal-to-noise ratio in current neutron
304: scattering experiments \cite{tranquada} is not high enough to
305: observe the second harmonic peak. We expect further experiments on
306: a larger pure single crystal measurement can verify our
307: prediction. The first harmonic occuring at $\vec Q\pm 2\vec q_0$
308: vanishes. For LSCO, structural effect at doping concentration
309: $\delta=1/8$ enhances the fundamental mode so that the second
310: harmonic is even harder to be observed.
311:
312: In order to obtain the low-energy properties, we apply the linear
313: spin-wave expansion \cite{hp}, {\it i.e.},
314: \begin{eqnarray}
315: \hat{S}^z(\vec r_i) &=& \pm (S(\vec r_i)-b_i^\dagger b_i) \\
316: \hat{S}^\pm(\vec r_i) &=& \sqrt{S(\vec r_i)} b_i \\
317: \hat{S}^\mp(\vec r_i) &=& \sqrt{S(\vec r_i)} b_i^\dagger
318: \end{eqnarray}
319: in which $b_i^\dagger$'s and $b_i$'s describe the quantum
320: fluctuation from the classical ground state. The effective
321: Hamiltonian for spin dynamics is
322: \begin{eqnarray}
323: H_J &=& E_0 + 4J\sum_i S(\vec r_i) b_i^\dagger b_i \nonumber \\ &&
324: + 2J\sum_i \left( \sqrt{S(\vec r_i)S(\vec r_{i+1})}b_ib_{i+1} +
325: {\rm h.c.} \right)
326: \end{eqnarray}
327: In the long-wavelength limit, we replace the geometric mean
328: $\sqrt{S(\vec r_i)S(\vec r_{i+1})}$ by the arithmetic mean
329: $(S(\vec r_i)+S(\vec r_{i+1}))/2$, and then perform the Fourier
330: transform to the variables $b_i^\dagger$ and $b_i$. Notice that
331: the sites of classical spin pointing up and down are merged
332: together in their Fourier modes. Then
333: \begin{eqnarray}
334: H_J &=& E_0 + 2S_0 J \sum_{\vec k} \left(
335: 2 b_{\vec k}^\dagger b_{\vec k} + \gamma_{\vec k} b_{\vec
336: k}b_{-\vec k} + \gamma_{\vec k} b_{-\vec k}^\dagger b_{\vec
337: k}^\dagger \right) \nonumber \\
338: && + S_1 J \sum_{\vec k} \left( 2 b_{\vec k+2 \vec q_0}^\dagger b_{\vec k}
339: + \gamma_{\vec k -2\vec q_0} b_{\vec k} b_{-\vec k + 2\vec q_0}
340: \right.
341: \nonumber \\
342: &&\left. +\gamma_{\vec k +2\vec q_0} b_{\vec k} b_{-\vec k - 2\vec q_0} +
343: {\rm h.c.}
344: \right)
345: \end{eqnarray}
346: where $\gamma_{\vec k}=(\cos(k_x)+\cos(k_y))/2$, and the terms
347: involving higher harmonics are neglected. The Hamiltonian is
348: quadratic and in principle can be straightforwardly diagonalized.
349: Since we are only interested in the excitation spectrum around
350: $\vec k = \pm 2\vec q_0$, write
351: \begin{eqnarray}
352: H_J|_{\vec k \simeq \pm 2\vec q_0} =
353: \sum_{\vec k \simeq 2\vec q_0} \eta_+^\dagger M_+ \eta_+
354: + \sum_{\vec k \simeq -2\vec q_0} \eta_-^\dagger M_- \eta_-
355: \end{eqnarray}
356: where $\eta_\pm=(b_{\vec k} \ b_{\vec k \mp \vec q} \ b_{-\vec k
357: }^\dagger \ b_{-\vec k \pm \vec q}^\dagger)^T$ and
358: \begin{eqnarray}
359: M_\pm = J \left(\begin{array}{cccc}
360: 4S_0 & 2S_1 & 4S_0 \gamma_{\vec k} & 2S_1 \gamma_{\vec k \mp 2 \vec q} \\
361: 2S_1 & 4S_0 & 2S_1 \gamma_{\vec k \mp 2 \vec q} & 4S_0 \gamma_{\vec k \mp 2 \vec q} \\
362: 4S_0 \gamma_{\vec k} & 2S_1 \gamma_{\vec k \mp 2 \vec q} & 4S_0 & 2S_1 \\
363: 2S_1 \gamma_{\vec k \mp 2 \vec q} & 4S_0 \gamma_{\vec k \mp 2 \vec
364: q} & 2S_1 & 4S_0
365: \end{array} \right) \nonumber \\
366: \end{eqnarray}
367: Note that it is the bosonic version of the spin-density wave
368: induced by charge stripes instead of the fermionic one due to
369: Fermi surface instability \cite{gruner}.
370:
371: Diagonalize the Hamiltonian, we get the gapless excitation at
372: $\vec k = \pm 2\vec q_0$, with the anisotropic spin-wave
373: velocities, in which
374: \begin{eqnarray} \label{ekx}
375: \epsilon_{\vec k=(k_x,0)} = 4S_0 J \left[1 -
376: \frac{S_1^2}{4S_0^2}\sec^2(q_0)\right]^\frac{1}{2} |k_x \pm 2 q_0|
377: \end{eqnarray}
378: and
379: \begin{eqnarray} \label{eky}
380: \epsilon_{\vec k=(\pm 2 q_0,k_y)} = 4S_0 J |k_y|
381: \end{eqnarray}
382: One can estimate the ratio of the anisotropic spin-wave velocities
383: by the above Eqs.(\ref{ekx})-(\ref{eky}),
384: \begin{eqnarray} \label{vyvx}
385: \frac{v_x}{v_y} = \left[1 -
386: \frac{S_1^2}{4S_0^2}\sec^2(q_0)\right]^\frac{1}{2}
387: \end{eqnarray}
388: For LSCO, $q_0= 2\pi\delta$. $v_x/v_y=1$ at $\delta=0$. At $\delta
389: =1/8$, $v_x/v_y=0.9997$. Without the structural effect, one can
390: safely ignore the anisotropy of spin-wave velocities. However,
391: perturbed structural effect can be straightforwardly considered by
392: including a term
393: \begin{eqnarray}
394: - V \sum_i \cos(2 \vec q_0 \cdot \vec r_i)b_i^\dagger b_i
395: \end{eqnarray}
396: where $V > 0$ enhances the spin ordering of magnitude of period
397: $2R$. The term behaves as a perturbed term $\delta M$ to enhance
398: the first harmonic component $S_1$ of the spin magnitude.
399: \begin{eqnarray}
400: \delta M = \frac{-V}{2} \left(\begin{array}{cc}
401: \begin{array}{cc}
402: 0 & 1 \\
403: 1 & 0
404: \end{array} & 0 \\
405: 0 & \begin{array}{cc}
406: 0 & 1 \\
407: 1 & 0
408: \end{array}
409: \end{array}\right)
410: \end{eqnarray}
411:
412: For perturbed $V \ll t$ where the weak coupling limit is still
413: valid, the density profile $\rho(x)$ in Eq.(\ref{rho}) does not
414: vary. Eqs.(\ref{ekx}) and (\ref{eky}) are just modified by
415: replacing $S_1$ by $S_1 + V/(4J)$. At $\delta \ll 1$, we can
416: estimate
417: \begin{eqnarray} \label{vyvx2}
418: \frac{v_x}{v_y} \simeq \sqrt{1-\frac{V^2}{4J^2}}
419: \end{eqnarray}
420: The anisotropy in the neutron scattering measurement is around
421: 0.75 \cite{aeppli}. Eq.(\ref{vyvx2}) gives an estimate of $V
422: \simeq 1.3J$.
423:
424: If the structural effect is strong enough ($V/t \gtrsim 1$), for
425: example, in the sample of the ordered stripe phase of $\rm
426: La_2NiO_{4+\delta}$, we should go beyond the weak coupling limit
427: such that the density profile found in Eq.(\ref{rho}) should also
428: depend on $V$. The case of the ordered stripe phase will be
429: reported elsewhere.
430:
431:
432: % ... related to the NL$\sigma$M and Landau theory of stripes
433:
434:
435:
436:
437: % conclusion
438:
439: %In conclusion, the low-energy effective Hamiltonian for spin
440: %dynamics of the fluctuating stripe is derived from the {\it t-J}
441: %model in the weak coupling limit $t\gg J$. Besides the modulation
442: %of the spin magnitude, the high energy hopping term would induce a
443: %low-energy anti-ferromagnetic interaction between two neighboring
444: %``blocks of spins". Based on the effective Hamiltonian, we applied
445: %the linear spin-wave theory and found that the spin wave velocity
446: %is almost isotropic for the sample of LSCO unless the lattice
447: %distortion effect is considered.
448:
449:
450:
451: % acknowledgement
452:
453: The author would like to thank T.K. Lee for introducing him to the
454: stripe problem, and acknowledges numerous discussion with the
455: colleagues at the Institute of Theoretical Physics, Chinese
456: Academy of Sciences (ITP, CAS). The work was supported by the
457: National Science Council of Taiwan under Grant No.
458: NSC89-2816-M-001-0012-6, and the Visiting Scholar Program of ITP,
459: CAS under 20C905.
460:
461:
462: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
463: % spin4.eps
464:
465:
466: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
467:
468: \begin{thebibliography}{10}
469: \bibitem{review}
470: E.W. Carlson {\it et al.}, in {\it The Physics of Conventional and
471: Unconventional Superconductors}, edited by K.H. Bennemann and J.B.
472: Ketterson (Springer-Verlag) and references therein.
473: \bibitem{review2}
474: S.A. Kivelson {\it et al.}, \rmp {\bf 75}, 1201 (2003) and
475: references therein.
476: \bibitem{tranquada}
477: J.M. Tranquada {\it et al.}, Nature {\bf 375}, 561 (1995).
478: % expt (incommensurate peaks)
479: \bibitem{incomm-expt}
480: S.W. Cheong {\it et al.}, \prl {\bf 67}, 1791 (1991); T.E. Mason,
481: G. Aeppli, and H.A. Mook, \prl {\bf 68}, 1414 (1992); T.R.
482: Thurston {\it et al.}, \prb {\bf 46}, 9128 (1992).
483: \bibitem{crawford}
484: M.K. Crawford {\it et al.}, \prb {\bf 44}, R7749 (1991).
485: % stripe from Hubbard model
486: \bibitem{hubbard}
487: D. Poilblanc and T.M. Rice, \prb {\bf 39}, R9749 (1989); J. Zaanen
488: and O. Gunnarsson, \prb {\bf 40}, R7391 (1989).
489: % slave fermion mean-field (incommenurate peaks)
490: \bibitem{auerbach}
491: A. Auerbach and D.P. Arovas, \prl {\bf 61}, 617 (1988).
492: % theories (incommursurate peaks)
493: \bibitem{incomm-th}
494: C. Jayaprakash, H.R. Krishnamurthy, and S. Sarker, \prb {\bf 40},
495: R2610 (1989); C.L. Kane, P.A. Lee, T.K. Ng, B. Chakraborty, and N.
496: Read, \prb {\bf 41}, R2653 (1990); B. Normand and P.A. Lee, \prb
497: {\bf 51}, 15519 (1995); C.D. Batista {\it et al.}, Europhys. Lett.
498: {\bf 38}, 147 (1997).
499: % phase separation
500: \bibitem{phase-separation}
501: V.J. Emery, S.A. Kivelson, and H.Q. Lin, \prl {\bf 64}, 475
502: (1990); M. Marder, N. Papanicolaou, and G.C. Psaltakis, \prb {\bf
503: 41}, 6920 (1990); T.I. Ivanov, \prb {\bf 44}, R12077 (1991); C.T.
504: Shih, Y.C. Chen, and T.K. Lee, \prb {\bf 57}, 627 (1998); C.H.
505: Cheng and T.K. Ng, Europhys. Lett. {\bf 52}, 87 (2000).
506: \bibitem{zachar02}
507: O. Zachar, \prb {\bf 65}, 174411 (2002).
508:
509: \bibitem{stability}
510: S.A. Kivelson and V.J. Emery, Synthetic Metals {\bf 80}, 151
511: (1996); S.R. White and D.J. Scalapino, \prl {\bf 80}, 1272 (1998);
512: C.S. Hellberg and E. Manousakis, \prl {\bf 83}, 132 (1999); N.G.
513: Zhang and C.L. Henley, \prb {\bf 68}, 014506 (2003).
514:
515: \bibitem{otherHam}
516: %If the effective term is the antiferromagnetic coupling between
517: %two neighboring spins across the hole stripes, the excitation
518: %spectrum is totally different from our theory. For example, the
519: %spin excitation is in general gapped.
520: Ref.[10] considered the hole stripes fluctuation inside the
521: three-leg ladder, and it is somehow equivalent to introducing the
522: structural effect implicitly.
523: % theories on stripes
524: \bibitem{ogata}
525: For example, A. Himeda, T. Kato, and M. Ogata, \prl {\bf 88},
526: 117001 (2002). The energy difference between the anti-phase and
527: in-phase spin domains at $\delta=1/8$ in their study is about
528: $0.002t=8K$ (take $t=0.4eV$), while the incommensurate peaks can
529: still be observed up to around $100K$.
530: %, which is easily destroyed by thermal fluctuation.
531: \bibitem{hp}
532: T. Holstein and H. Primakoff, Phys. Rev. {\bf 58}, 1098 (1940).
533: \bibitem{gruner}
534: G. Gr\"{u}ner, \rmp {\bf 66}, 1 (1994).
535: \bibitem{aeppli}
536: G. Aeppli {\it et al.}, \prl {\bf 62}, 2052 (1989).
537:
538:
539: %\bibitem{emery1}
540: %V.J. Emery, S.A. Kivelson, and J.M. Tranquada, Proc. Natl. Acad.
541: %Sci. {\bf 96}, 8814 (1999).
542:
543: %\bibitem{kivelson}
544: %S.A. Kivelson {\it et al.}, Nature {\bf 393}, 550 (1998).
545:
546:
547:
548:
549:
550:
551:
552: %\bibitem{pryadko}
553: %L.P. Pryadko {\it et al.}, \prb {\bf 60}, 7541 (1999).
554:
555: %\bibitem{zachar2}
556: %O. Zachar, S.A. Kivelson, and V.J. Emery, \prb {\bf 57}, 1422
557: %(1998).
558: %\bibitem{neto}
559: %A.H. Castro Neto and D. Hone, \prl {\bf 76}, 2165 (1996).
560:
561: % numerical (incommensurate peaks)
562: %\bibitem{leung}
563: %R.J. Gooding, K.J.E. Vos, and P.W. Leung, \prb {\bf 49}, 4119
564: %(1994).
565: % expt (supporting stripes)
566: %\bibitem{Hellberg}
567: %C.S. Hellberg and E. Manousakis, \prl {\bf 83}, 132 (1999).
568: %\bibitem{dhlee}
569: %D.H. Lee, Int. Jour. Mod. Phys. B {\bf 15}, 1117 (2001).
570:
571: %\bibitem{emery}
572: %V.J. Emery, \prl {\bf 58}, 2794 (1987).
573:
574:
575:
576: \end{thebibliography}
577: \newpage
578:
579: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
580: \vspace{15pt}
581: \begin{figure}[tbh]
582: \begin{center}
583: \includegraphics[width=3in]{spin2.eps}
584: \end{center}
585: \vspace{-5pt}
586: \caption{Schematic diagram of spin state at doping concentration $\delta=1/8$
587: of LSCO ($q_0 = 2\pi\delta = \pi/4$).
588: The circles represent the zero average of spin direction. The dashed line shows the
589: envelope of the spin component in $z$-direction. Any two neighboring ``blocks of spins" feel
590: an anti-ferromagnetic interaction of coupling $J'$.}
591: \label{spin2.eps}
592: %\vspace{-5pt}
593: \end{figure}
594:
595: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
596:
597: %\newpage
598:
599: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
600: \vspace{15pt}
601: \begin{figure}[tbh]
602: \begin{center}
603: \includegraphics[width=3in]{spin4.eps}
604: \end{center}
605: \vspace{-5pt}
606: \caption{Illustration of the flip of a ``block of spins" from (a) to (b).}
607: \label{spin4.eps}
608: \vspace{-5pt}
609: \end{figure}
610:
611: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
612:
613: \end{document}
614: