1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %
3: % Format of paper cond-mat
4: % REVTeX.4 version
5: %
6: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
7: %
8: % document class definition
9: %
10: %\documentclass[11pt,showpacs,preprintnumbers]{revtex4}
11: %\documentclass[aps,prb,twocolumn,superscriptaddress,showpacs]{revtex4}
12: \documentclass[11pt,aps,prb,superscriptaddress,showpacs]{revtex4}
13: %
14: \usepackage[dvips]{graphicx}
15: \usepackage{epsfig}
16: \usepackage{amsmath,amssymb,amscd}
17: %
18: \begin{document}
19: %
20: % title
21: %
22: \title{
23: Phase diagram of S=1/2 two-leg XXZ spin ladder systems
24: }
25: %
26: % author
27: %
28: \author{Keigo Hijii}
29: \affiliation{Department of Physics,Kyushu University, Fukuoka 812-8581, Japan}
30: \author{Atsuhiro Kitazawa}
31: \affiliation{Research Institute for Applied Mechanics, Kyushu University,
32: Kasuga, Fukuoka 816-858, Japan}
33: \author{Kiyohide Nomura}
34: \affiliation{Department of Physics,Kyushu University, Fukuoka 812-8581, Japan}
35: %
36: %
37: \date{\today}
38: %
39: %
40: \large
41: %
42: \begin{abstract}
43: We investigate the ground state phase diagram
44: of the S=1/2 two-leg $XXZ$ spin ladder system
45: with an isotropic interchain coupling.
46: In this model, there is
47: the Berezinskii-Kosterlitz-Thouless transition
48: which occurs at the $XY$-Haldane and the $XY$-rung singlet phase boundaries.
49: It was difficult to determine the transition line
50: using traditional methods.
51: We overcome this difficulty
52: using the level spectroscopy method combined with the twisted boundary
53: condition method, and we check the consistency.
54: We find out that the phase boundary between $XY$ phase and Haldane phase
55: lies on the $\Delta=0$ line.
56: And we show that there exist two different $XY$ phases,
57: which we can distinguish investigating
58: a $XX$ correlation function.
59: \end{abstract}
60:
61: \pacs{75.10.Jm, 75.40.Mg, 75.40.Cx}
62: %\pacs{75.10.Jm}
63:
64: \maketitle
65:
66: \section{INTRODUCTION}
67: %\hspace*{\parindent}
68: Low dimensional quantum spin systems have attracted
69: some attention.
70: Especially, ground state properties are interesting
71: since the quantum fluctuation often plays the dominant role.
72: For example, $S=1/2$ quantum spin ladder systems have been studied
73: from both theoretical and experimental points of view \cite{D-R},
74: related to Haldane's conjecture and high-T$_c$ superconductors.
75: Haldane has predicted that the properties of integer spin chains are
76: different from those of half-odd-integer spin chains \cite{Haldane}.
77: Half odd integer spin chains do not have excitation energy gaps
78: \cite{L-S-M}.
79: By contrast, for integer spin chains there is an energy gap
80: between the ground state and the lowest excited state\cite{Haldane}.
81: We can understand this fact, using the valence bond solid picture
82: \cite{VBS}.
83: From the other point of view,
84: Haldane's conjecture is related to spin ladder systems,
85: since we can describe arbitrary spin-$S$ chains as $S=1/2$ $2S$-leg
86: ladder systems approximately \cite{Schulz}.
87:
88: The model which we study in this paper
89: consists of two coupled $S=1/2$ $XXZ$ chains.
90: We consider the following Hamiltonian,
91: \begin{align}
92: {\cal H} &= {\cal H}_{\rm leg} + {\cal H}_{\rm rung}, \nonumber \\
93: {\cal H}_{\rm leg} &=
94: \sum_{i=1}^L\sum_{\alpha=1,2} \left(
95: S_{\alpha,i}^x S_{\alpha,i+1}^x + S_{\alpha,i}^y S_{\alpha,i+1}^y
96: + \Delta S_{\alpha,i}^z S_{\alpha,i+1}^z
97: \right), \nonumber \\
98: {\cal H }_{\rm rung} &= J_{\rm rung} \sum_{i=1}^L {\bf S}_{1,i}
99: \cdot {\bf S}_{2,i},
100: \label{eq:Hamiltonian}
101: \end{align}
102: where $\alpha=1,2$ are indices of chains (see Fig. \ref{fig:ladder}),
103: $\Delta$ is an anisotropy of the leg coupling, and $J_{\rm rung}$ is
104: the rung coupling. $L$ is the system size ($L=N/2$, $N$ is the number of
105: sites). We shall discuss the boundary condition later.
106:
107: So far, S=1/2 two-leg quantum spin ladder systems have been studied
108: by many authors.
109: Strong and Mills have presented the phase diagram using the bosonization
110: approach \cite{S-M}.
111: Watanabe {\it et al.} have discussed the same model
112: using the bosonization with the aid of Wilson's renormalization
113: group method, and presented the phase diagram partially \cite{W-N-T}.
114: Narushima {\it et al.} have investigated
115: Haldane-N\'eel transition \cite{N-N-T},
116: using the density matrix renormalization group (DMRG) method
117: \cite{DMRG}.
118: These theoretical studies have been done mainly with antiferromagnetic legs,
119: since real materials have antiferromagnetic interaction,
120: and those are investigated in detail experimentally
121: (e.g., Sr$_{14}$Cu$_{24}$O$_{41}$).
122: Sr$_{14-x}$Ca${_x}$Cu$_{24}$O$_{41+\delta}$ is known as
123: the superconductor under high pressure\cite{U-N-A-T-M-K}.
124: Recently, ferromagnetic legs cases attract much interest,
125: for example, as Kolezhuk and Mikeska have discussed \cite{K-M}.
126: There exists a linear chain with ferromagnetic interaction,
127: and we can expect the discovery of the ladder material
128: with ferromagnetic interaction\cite{K-T-J1,K-T-J2}.
129: Vekua {\it et al.} have studied ferromagnetic cases using
130: the bosonization approach in the weak-coupling limit \cite{V-J-M}.
131: They found out that an extended gapless phase appears
132: not only in the ferromagnetic rung coupling region but also
133: in the antiferromagnetic rung coupling region.
134: This phase was also referred by Strong and Mills \cite{S-M} too.
135: The stability of this gapless phase for the $\Delta <0$ region
136: was discussed by Legeza and Solyom. \cite{L-S}.
137: $\Delta= 1$ cases have been well studied using the bosonization
138: \cite{S-N-T,L-O}.
139: In this case, the system is massive.
140: Lecheminant and Orignac showed that
141: the edge state does not exist in this model with $J_{\rm rung}>0$
142: %there does not exist the edge state in this model with $J_{\rm rung}>0$
143: under the open boundary condition (OBC) \cite{L-O}.
144: This means that there is no Haldane phase in $J_{\rm rung}>0$ region,
145: since the edge state appears in Haldane phase under the OBC.
146:
147: There are several suggestions for the phase diagram of this model.
148: However there is no unified view.
149: There remain controversies in three points.
150: The first point is the extent of the XY phase.
151: The second point is the possibility of existence of two different $XY$ phases,
152: and the third point is the discussion about the multicritical point
153: near the Ferromagnetic phase.
154:
155: For the first point, there are some previous research.
156: For example, Strong and Mills \cite {S-M} and Vekua {\it et al}.
157: \cite{V-J-M} have insisted that the $XY$ phase extend to the $\Delta >0$ region.
158: Watanabe {\it et al}, who used the bosonization approach,
159: have insisted the different extension in the $\Delta >0$ region
160: \cite{W-N-T}.
161: Narushima {\it et al}. \cite{N-N-T} and Legeza {\it et al}.
162: \cite{L-S} have insisted that the XY phase does not extended
163: in the $\Delta >0$ region.
164: They calculated the energy gap and the correlation function,using DMRG.
165: But Narushima {\it et al}. and Legeza {\it et al}. did not
166: determine the Berezinskii-Kosterlitz-Thouless (BKT) transition line,
167: since the correlation length becomes large compared to the system
168: size, especially near the $XY$ phase.
169: Also there are peculiar difficulties in determining the BKT transition
170: point \cite{LS}.
171: %Since there are controversies, we study this system
172: %in order to solve this problem.
173:
174: The $XY$ phase belongs to the universality class
175: of Tomonaga-Luttinger (TL) liquid \cite{T-L}, which is characterized by
176: a gapless excitation and a power-law decay of the correlation function.
177: In general, one component TL liquid is described by the $c=1$
178: conformal field theory (CFT) with the U(1) symmetry.
179: One of the instabilities of the TL liquid is the BKT transition \cite{BKT}.
180: In the BKT transition, the traditional finite size scaling method
181: (the phenomenological renormalization group \cite{F-B})
182: can not be applied \cite{fsBKT,I-N}.
183: Moreover, since there exists the logarithmic correction,
184: it was difficult to determine the BKT transition point.
185: Recently, combining the renormalization group
186: with the symmetry consideration,
187: the "level spectroscopy" method has been developed
188: in order to overcome these difficulties \cite{LS}.
189: Combining this method and twisted boundary condition \cite{Kitazawa,N-K},
190: we will numerically determine the phase boundary between the $XY$ phase
191: and other massive phases in this paper.
192:
193:
194: For the second point, the possibility of
195: two different $XY$ phases in ladder systems was pointed out
196: by Vekua {\it et al.}\cite{V-J-M}
197: in a little different model.
198: However, in their following paper, they did not
199: comment on two different $XY$ phases\cite{V-J-M2}.
200: We discuss this point using a different approach in this paper.
201: According to the Marshall-Lieb-Mattis's theorem,
202: the $XY1$ phase is in $J_{\rm rung}<0$ region and
203: the $XY2$ phase is in the $J_{\rm rung}>0$ region %are different
204: (see section.\ref{sec:unitary}).
205:
206:
207: The third point is the discussion about the multicritical point
208: near the Ferromagnetic phase.
209: There are different suggestions\cite{S-M,K-M,V-J-M}.
210: We clarified the slope toward the multicritical point at
211: $\left(\Delta ,J_{\rm rung} \right)=\left(-1,0\right)$
212: concerning the contradicting results\cite{K-M,V-J-M}.
213:
214:
215: We discuss the phase diagram of S=1/2 two-leg $XXZ$ spin ladder systems
216: in this paper (see Fig. \ref{fig:pdm})
217: We organize this paper as follows.
218: In the next section, we consider the phase diagram
219: under several limits.
220: In section 3, we discuss the phase boundary,
221: using the variational approach.
222: In section 4, we discuss unitary transformations and correlation
223: functions.
224: In section 5, we determine some phase boundaries from exact results,
225: considering the number of degeneracies of the ground state,
226: and we consider the weak coupling region.
227: In section 6, we numerically determine the phase boundary between the
228: $XY$ phase
229: and the rung-singlet phase and between the $XY$ phase and the
230: Haldane phase, using the level spectroscopy method,
231: and we discuss the multicriticality.
232: The last section is the conclusion.
233:
234:
235: \section{Several limits}
236: %\hspace*{\parindent}
237: In this section, we consider the ground state in some limits.
238:
239: In the $J_{\rm rung}=0$ case,
240: the system consists of two independent $S=1/2$ $XXZ$ chains,
241: and it was solved exactly \cite{Bethe,C-G,B-I-K}.
242: In the $\Delta<-1$ region, there is the ferromagnetic phase.
243: In the $-1< \Delta <1$ region, the XY (spin fluid) phase appears.
244: In the $\Delta>1$ region, the N\'eel phase appears.
245:
246: In the $J_{\rm rung} \rightarrow -\infty$ case,
247: this model can be mapped onto an $S=1$ $XXZ$ spin chain.
248: The $S=1$ $XXZ$ spin chain has been studied by many authors
249: \cite{dN-R,C-H-S}.
250: In this system, there are four phases, i.e., the ferromagnetic phase,
251: the $XY$ phase, the Haldane phase, and the N\'eel phase.
252: This Haldane phase is the gapped phase with the hidden
253: $Z_2 \times Z_2$ symmetry breaking.
254:
255: In the $\Delta\rightarrow -\infty$ case, we can expect that
256: ferromagnetic ordered phases appear in each leg directions.
257: And we can classify them as follows.
258: If $J_{\rm rung}$ is positive, the ground state is called as the stripe
259: ferromagnetic phase \cite{V-J-M}.
260: The staggered magnetization appears in the rung direction.
261: This phase has the following correlation function:
262: \begin{equation}
263: \left< S^z_{\alpha,i}S^z_{\beta,j} \right> \propto
264: \left(-1\right)^{\alpha+\beta},
265: \end{equation}
266: (see Fig. \ref{fig:Stripe}(a)).
267: If $J_{\rm rung}$ is negative, the ground state is the fully ferromagnetic
268: phase.
269:
270: In the $\Delta \rightarrow + \infty$ case, we can expect that
271: antiferromagnetic (N\'eel) ordered phases appear in each leg direction.
272: If $J_{\rm rung}$ is positive, the ground state is
273: the N\'eel phase in the rung direction.
274: (see Fig. \ref{fig:Stripe}(c)).
275: If $J_{\rm rung}$ is negative, the ground state is
276: the ferromagnetic correlation in the rung direction.
277: This phase is called as the stripe N\'eel phase.
278: The staggered magnetization appears in the leg direction.
279: This phase has the following correlation function,
280: \begin{equation}
281: \left< S^z_{\alpha,i}S^z_{\beta,j} \right> \propto
282: \left(-1\right)^{i-j},
283: \end{equation}
284: (see Fig. \ref{fig:Stripe}(b)).
285:
286: In the $\Delta \sim 0$ and $J_{\rm rung} \gg 1$ case, we can expect that
287: the rung-singlet phase appears.
288: This state is a nondegenerate, massive phase.
289: We can understand this state as the direct
290: product of singlets in the rung direction.
291:
292:
293:
294:
295: \section{Variational approach and Phase boundaries}
296: %\hspace{\parindent}
297: In this section, we roughly estimate phase boundaries using the
298: variational method in some limits under the periodic boundary condition
299: (PBC).
300:
301:
302:
303: \subsection{Rung Singlet-Stripe Ferromagnetic phase transition}
304: %\hspace{\parindent}
305: At first, we consider the phase transition between
306: the rung singlet phase and the stripe ferromagnetic phase.
307: The variational energy of the pure rung-singlet state is given as follows:
308: \begin{equation}
309: \frac{E_{\rm RS}}{L}=
310: \frac{\left< \mbox{RS} \left| {\cal H} \right| \mbox{RS} \right>}{L}
311: =-\frac{3}{4}J_{\rm rung},
312: \end{equation}
313: where $L$ is the system size ($L=N/2$, $N$ is the number of sites).
314: On the other hand,
315: the variational energy of the pure stripe-ferromagnetic state is given
316: as follows:
317: \begin{equation}
318: \frac{E_{\rm SF}}{L}=
319: \frac{\left< \mbox{SF} \left| {\cal H} \right| \mbox{SF} \right>}{L}
320: =\frac{1}{2}\Delta-\frac{1}{4}J_{\rm rung}.
321: \end{equation}
322: So we can roughly estimate that the phase boundary between the rung
323: singlet phase and the stripe ferromagnetic phase is $\Delta=-J_{\rm rung}$
324: in the $\Delta \rightarrow -\infty$ and
325: $J_{\rm rung} \rightarrow \infty$ limit.
326:
327: This phase transition is related to an
328: spontaneously ${\bf Z}_2$ symmetry breaking.
329: Therefore, this is the second order phase
330: transition and its universality class may be the
331: two dimensional (2D) Ising type.
332:
333:
334: \subsection{N\'eel-Rung Singlet phase transition}
335: %\hspace{\parindent}
336: Next, we consider the phase transition between
337: the N\'eel phase and the rung singlet phase.
338:
339: The variational energy of the pure N\'eel state is given as follows:
340: \begin{equation}
341: \frac{E_{\rm Neel}}{L}=
342: \frac{\left< \mbox{N\'eel} \left| {\cal H} \right| \mbox{N\'eel} \right>}{L}
343: =-\frac{1}{2}\Delta -\frac{1}{4} J_{\rm rung} .
344: \end{equation}
345: So we can roughly estimate the phase boundary between the rung singlet
346: phase and the N\'eel phase is $\Delta=J_{\rm rung}$,
347: in $\Delta \rightarrow \infty$ and
348: $J_{\rm rung} \rightarrow \infty$ limit.
349: This phase transition is related to the ${\bf Z}_2$ symmetry breaking.
350: Therefore, this phase transition is the second order phase
351: transition and its universality class may be the 2D-Ising type.
352:
353:
354:
355: \section{Unitary transformation and Phase transition}
356: \label{sec:unitary}
357: %\hspace*{\parindent}
358: In this section, we treat the bipartite systems.
359: That is,
360: the system size ($L=N/2$) is even under the periodic boundary condition,
361: or the system size ($L$) is arbitrary under the open boundary
362: condition, where $N$ is the number of spins.
363:
364: \subsection{$XY$ phase - Ferromagnetic phase transition}
365: %\hspace*{\parindent}
366: In the ferromagnetic coupling case ($\Delta<-1,J_{\rm rung}<0$),
367: we found that the ferromagnetic phase appears.
368: Vekua {\it et al.} have obtained that the phase boundary
369: between $XY$ phase and the ferromagnetic phase
370: locates at $\Delta=-1$ using the instability analysis of the
371: spin wave theory for the ferromagnetic phase \cite{V-J-M}.
372: Here we try to use another approach for this phase transition.
373: We transform the original Hamiltonian (\ref{eq:Hamiltonian})
374: using a following unitary operator:
375: \begin{equation}
376: U=\exp \left[ i\pi \sum_{j,\alpha}S^z_{\alpha,2j+1} \right]
377: \label{eq:unitary1}
378: \end{equation}
379: (see Fig \ref{fig:nrung}).
380: This unitary operator transform
381: $S^{\pm}_{\alpha,2j+1} \rightarrow -S^{\pm}_{\alpha,2j+1}$.
382: Especially, in the $\Delta=-1$ case,
383: the transformed Hamiltonian is described as the pure Heisenberg model
384: with a negative coupling constant.
385: Now we can choose a set of basis vectors as eigenvectors which
386: diagonalize $S^z$.
387: Then all off-diagonal elements of the transformed Hamiltonian
388: have a negative sign.
389: From Perron-Frobenius's theorem,
390: the ground state is unique in the fixed
391: $S^z_T=\sum_{j,\alpha}S^z_{\alpha,j}$ space in the finite system,
392: and we can choose that all the coefficients are positive.
393: Since the transformed system has an $SU(2)$ symmetry
394: in the $\Delta=-1$ case,
395: states with $S^z_T=0,\pm1,\pm2,\cdots$ are degenerate (SU(2) Ferro).
396: This means that $J_{\rm rung}<0, \Delta=-1$ line is the phase boundary,
397: between the fully ferromagnetic phase ($\Delta<-1$)
398: and the $XY$ phase ($\Delta>-1$).
399:
400: Especially, in the $(J_{\rm rung}=0,\Delta=-1)$ case
401: and the system size $L(=N/2))$, where $L$ is even,
402: there are the number of $(L+1)^2$ degenerated ground states,
403: since this system has an $SU(2) \times SU(2)$ symmetry.
404: We can consider that this point is the multicritical point,
405: where the BKT transition line meets the 2D Ising type phase transition
406: line.
407:
408:
409: \subsection{Off critical case.1 ($J_{\rm rung}<0$)}
410: %\hspace*{\parindent}
411: In this case, we consider the unitary transformation (\ref{eq:unitary1}),
412: (see Fig \ref{fig:nrung}).
413: After this unitary transformation, all off-diagonal elements of
414: Hamiltonian become negative.
415: From the discussion of the previous subsection,
416: signs of correlation functions are
417: represented in the original Hamiltonian (\ref{eq:Hamiltonian}) as follows,
418: \begin{equation}
419: \left(-1\right)^{i-j}
420: \left<S^x_{\alpha,i}S^x_{\beta,j}\right> >0.
421: \end{equation}
422: This corresponds to the Ferromagnetic phase, the $XY1$ phase,
423: Haldane phase and the Stripe N\'eel phase.
424:
425:
426: \subsection{Off critical case.2 ($J_{\rm rung}>0$)}
427: %\hspace*{\parindent}
428: In this case, we consider the following unitary transformation:
429: \begin{equation}
430: U=\exp \left[ i\pi \sum_{j} \left( S^z_{1,2j+1} + S^z_{2,2j} \right)
431: \right]
432: \label{eq:unitary2}
433: \end{equation}
434: (see Fig \ref{fig:prung}).
435: This unitary operator transform
436: $S^{\pm}_{1,2j+1}\rightarrow -S^{\pm}_{1,2j+1}$ and
437: $S^{\pm}_{2,2j}\rightarrow -S^{\pm}_{2,2j}$.
438: After this unitary transformation, all off-diagonal elements of
439: Hamiltonian become negative.
440: From the Marshall-Lieb-Mattis's theorem\cite{M-L-M},
441: this case has following signs of the correlation function
442: in the original Hamiltonian (\ref{eq:Hamiltonian}):
443: \begin{equation}
444: \left(-1\right)^{\alpha+\beta}\left(-1\right)^{i-j}
445: \left<S^x_{\alpha,i}S^x_{\beta,j}\right>>0
446: \end{equation}
447: This corresponds to the Stripe Ferromagnetic phase, the $XY2$ phase,
448: the rung singlet phase and N\'eel phase.
449:
450: \subsection{Off critical case.3 ($\Delta=-1,J_{\rm rung}>0$)}
451: %\hspace*{\parindent}
452: In this subsection, we consider ($\Delta=-1,J_{\rm rung}>0)$ case.
453: Using the unitary transformation (\ref{eq:unitary1}),
454: we can see that the system has an $SU(2)$ symmetry.
455: From Mermin-Wagner's theorem, since there is no long range order
456: with the spontaneously continuous symmetry breaking,
457: the ground state is not the Stripe Ferromagnetic phase.
458: Since there is no soft mode with the wave number $q=\pi$
459: for the transformed Hamiltonian,
460: there is no possibility that it is described as
461: a $k=1,2$ $SU(2)$ Wess-Zumino-Witten model, which is massless.
462:
463: From the Marshall-Lieb-Mattis's theorem \cite{M-L-M},
464: the sign of the correlation function
465: transformed by unitary operator (\ref{eq:unitary1}),
466: \begin{equation}
467: \left< S^{x}_{1,i}S^{x}_{2,j}\right><0.
468: \end{equation}
469: Thus this ground state is not an $SU(2)$ ferromagnetic phase.
470:
471: Therefore we can conclude that the ground state is the rung singlet
472: phase in the ($\Delta=-1,J_{\rm rung}>0$) case.
473:
474:
475:
476:
477: \section{Approach from decoupled chains}
478: %\hspace*{\parindent}
479: In this section, we discuss the phase diagram from the exact solutions.
480: This system with $J_{\rm rung}=0$ consists of two independent chains.
481: From the exact solution we know that $S=1/2$ $XXZ$ chain with anisotropy
482: $\left|\Delta\right|>1$ is massive, and the ground state is twofold
483: degenerate.
484: Now the system is two decoupled chains, therefore, the ground state
485: is fourfold degenerate.
486: In following subsections, we discuss the neighborhood of
487: $J_{\rm rung}=0$.
488:
489:
490:
491: \subsection{Ferromagnetic phase - Stripe Ferromagnetic phase transition}
492: %\hspace{\parindent}
493: In $\Delta<-1$ region, we can find two different two ferromagnetic
494: ordered phases, as suggested by Vekua {\it et al} \cite{V-J-M}.
495: The usual ferromagnetic ordered phase is two fold degenerate.
496: However on the $\Delta<-1,J_{\rm rung}=0$ line, the ground state
497: is four fold degenerate.
498: In the first case all spins of both chains are up;
499: in the second case all spins of both chains are down;
500: in the third case all spins of 1-chain is up and all spins of 2-chain
501: is down;
502: and in the forth case all spins of 1-chain is down and
503: all spins of 2-chain
504: is down (see Fig.(\ref{fig:Stripe})(a)).
505: These four states have the same energy exactly on the $J_{\rm rung}=0$.
506: A nonzero $J_{\rm rung}$ term breaks this degeneracy.
507: This means that $\Delta<-1,J_{\rm rung}=0$ line is the phase boundary.
508: In addition,
509: the system, which consists of two independent gapped system is gapped.
510: So this phase transition is the first order transition.
511:
512:
513: \subsection{N\'eel phase - Stripe N\'eel phase transition}
514: %\hspace{\parindent}
515: The discussion in the previous subsection can be applied to
516: the antiferromagnetic ordered phase similarly.
517:
518: Firstly, we consider the single $S=1/2$ $XXZ$ chain.
519: From the exact solution, in $\Delta > 1$ region,
520: the ground state is the antiferromagnetic
521: (N\'eel) ordered state,
522: which is two fold degenerate.
523: Secondly, we consider independent double $S=1/2$ $XXZ$ chains.
524: The ground state is fourfold degenerate
525: in $\Delta > 1$ region,
526: since this total system is described as the direct product of
527: the single chain system.
528: These ground states are shown in Fig.\ref{fig:Stripe} (b),(c).
529: Lastly, we consider the coupled $S=1/2$ $XXZ$ chains
530: described by the Hamiltonian (\ref{eq:Hamiltonian}).
531: The rung-coupling term breaks the four fold degeneracy in the above case.
532: In $J_{\rm rung}<0$ case,
533: the ferromagnetic ordered state in the rung direction
534: (see, Fig.\ref{fig:Stripe} (b)),
535: which is called as the stripe N\'eel phase.
536: In $J_{\rm rung}>0$ case,
537: the antiferromagnetic ordered state in the rung direction
538: (see, Fig.\ref{fig:Stripe} (c)),
539: which is called as the N\'eel phase.
540: Both states are antiferromagnetic ordered in the leg direction.
541:
542: Since the system is gapped in the $J_{\rm rung}=0$ case,
543: this phase transition between the N\'eel phase and the stripe N\'eel phase
544: is the first order transition.
545:
546:
547: \subsection{Weak coupling region}
548: %\hspace{\parindent}
549: In this subsection, we consider the weak coupling S=1/2 XXZ chain,
550: $-1 \le \Delta \le 1$ and $\left|J_{\rm rung}\right| \ll 1$.
551:
552: At first, we consider two independent chains, $J_{\rm rung}=0$ case.
553: This case is exactly solved using Bethe ansatz\cite{Bethe,C-G,B-I-K}.
554: One chain in the $-1<\Delta<1$ region is described by TL liquid
555: \cite{T-L}.
556: Two independent chains are described by the direct product of
557: two TL liquids.
558: This model has critical properties in extended regions in the
559: parameter space.
560: Here we consider the case that there are some perturbations for this
561: system.
562:
563: These systems can be analyzed using the bosonization method
564: \cite{S-M,V-J-M}.
565: Then two decoupled chains are described in two bosonic fields,
566: $(\phi_1,\theta_1)$ and $(\phi_2,\theta_2)$ of each other.
567: We obey the notation used in Ref.\cite{V-J-M} and introduce
568: the symmetric and antisymmetric combination of the bosonic field,
569: \begin{equation}
570: \left\{
571: \begin{array}{c}
572: \phi_\pm \propto \phi_1 \pm \phi_2, \\
573: \theta_\pm \propto \theta_1 \pm \theta_2. \\
574: \end{array}
575: \right.
576: \end{equation}
577: We consider the case with $J_{\rm rung}\neq 0$.
578: Behavior of the symmetric field
579: is
580: governed by the effective sine-Goldon model.
581: $\cos\sqrt{\frac{2\pi}{K_-}}\theta_-$ is always relevant.
582: $\cos\sqrt{2\pi K_+}\phi_+$ is relevant for $\Delta>0$,
583: irrelevant for $\Delta<0$,
584: where $K_{\pm}$ are constant \cite{V-J-M}.
585:
586:
587: The phase transition between the $XY$ phase and
588: the nondegenerate massive phase is the $K=4$ BKT type transition
589: when the system has a simple $U(1)$ symmetry,
590: where $K$ is the TL parameter.
591: In this case, a $(S^z_T=0,q=\pi)$ mode is always massive,
592: and a $(S^z_T=0,q=0)$ mode becomes massless.
593:
594:
595: We can distinguish two different $XY$ phases.
596: One $XY$ phase ($XY1$) is in $J_{\rm rung}<0$,
597: another $XY$ phase ($XY2$) is in $J_{\rm rung}>0$ region.
598: In the $J_{\rm rung}>0$ region,
599: from the Marshall-Lieb-Mattis's theorem,
600: $\left<S^x_{1,j}S^x_{2,j}\right> < 0$.
601: On the other hand,
602: in the $J_{\rm rung}<0$ region, $\left<S^x_{1,j}S^x_{2,j}\right> > 0$.
603: Therefore these two $XY$ phases have different symmetry.
604: A $J_{\rm rung}=0$ line is the second order phase transition line.
605: The system on the $J_{\rm rung}=0$ line is described as two component
606: TL liquid.
607: This second order phase transition is explained as
608: that one component of the TL liquid remains massless,
609: but another component become massive.
610:
611:
612: Here we summarize this subsection in the CFT language.
613: In the $\left|\Delta\right|<1$ region,
614: the system is described by the central charge
615: $c=2$ CFT on the $J_{\rm rung}=0$ line.
616: And there exists a region which is described as $c=1$ CFT
617: in the $0<\left|J_{\rm rung}\right| \ll 1$ and $-1 \leq\Delta\leq 0$.
618:
619:
620:
621:
622: \section{Phase Transitions and Numerical Results}
623: %\hspace*{\parindent}
624: In this section, we determine the critical points and the universality class.
625: We show the numerical results of $N=12,16,20,24$ systems using
626: the exact diagonalization, and the level spectroscopy method with the twisted
627: boundary condition \cite{LS,Kitazawa,N-K}.
628:
629: \subsection{XY-Rung singlet transition}
630: %\hspace*{\parindent}
631: We can expect that this phase transition is the BKT type.
632: Using the level spectroscopy method,
633: we determine critical points of this phase transition.
634: Considering symmetries, we can identify critical points as cross points
635: of two low-lying excitations.
636: One has quantum numbers $S^z_T=\pm 2, P=1, q=0$
637: under the periodic boundary condition (PBC),
638: another has quantum numbers $S^z_T=0, P^*=1$ under the
639: twisted boundary condition (TBC)
640: $(S^{x,y}_{i,j} = - S^{x,y}_{i,j+L})$.
641: Here $S^z_T$ is total magnetizations of the system
642: ($S^z_T\equiv\sum_{i,\alpha}S^z_{\alpha,i}$),
643: and $P$ is the parity defined under the following transformation
644: $(S^{\alpha}_{i,j} \leftrightarrow S^{\alpha}_{i,L-j+1})$,
645: $P^*$ is the parity under the TBC.
646: $q$ is the wave number defined under the periodic boundary condition,
647: $L$ is the system size which is a half of the number of sites.
648: We show crossing points in Fig. \ref{fig:cp-xy-rs}.
649:
650: However values of crossing points have size dependence.
651: We can remove logarithmic corrections, which is proportional to
652: $1/\ln L$, using the level spectroscopy method.
653: However, there remain other corrections
654: which come from the lattice structure.
655: For example, there are correction terms
656: from the $x=4$ irrelevant field ($L_{-2}\overline{L}_{-2}{\bf 1}$)
657: \cite{Cardy}.
658: Therefore we extrapolate crossing points as follows,
659: \begin{equation}
660: J_{\rm rung}^{\rm cross} \left( L \right)=J_{\rm rung}^{\rm cross}
661: \left( \infty \right)
662: + a \frac{1}{L^2} + o \left( \frac{1}{L^4} \right)
663: \end{equation}
664: or
665: \begin{equation}
666: \Delta^{\rm cross} \left( L \right)=\Delta ^{\rm cross}
667: \left( \infty \right) + a
668: \frac{1}{L^2} + o \left( \frac{1}{L^4} \right)
669: \end{equation}
670: where $a$ is a fitting parameter. We neglect higher order terms
671: $o\left(\frac{1}{L^4} \right)$.
672: In Fig. \ref{fig:sizeeff}, we show the size dependence of the crossing
673: points and its extrapolation.
674: Then we can find the phase boundary under the infinite system limit
675: in Fig.(\ref{fig:pd-xy-rs}).
676:
677: In order to confirm the consistency of our results,
678: we need check the universality class using CFT \cite{CFT}.
679: In this case, the critical theory which describes the transition line
680: is the $c=1$ CFT.
681: The central charge appears as an universal finite size correction
682: for the ground state energy under the PBC
683: \cite{B-C-N,Affleck,Cardy1},
684: \begin{equation}
685: E_{\rm g} \left( L \right)=\epsilon L -\frac{\pi v }{6L}
686: \left( c +o \left( \frac{1}{ \left( \ln L \right)^3 }\right)\right)
687: \end{equation}
688: where $L$ is the system size, $\epsilon$ is the ground state energy per
689: site in the infinite system, $v$ is the spin wave velocity.
690: Now $o\left( \frac{1}{ \left( \ln L \right)^3 }\right)$ correction
691: terms are small enough for numerical data, so we neglect it.
692: In order to numerically determine the central charge,
693: we should obtain $v$ in addition to the ground state energy.
694: We can obtain the spin wave velocity as follows:
695: \begin{equation}
696: v \left( L \right) = \frac{L}{2\pi} \left(
697: E \left( q=\frac{2\pi}{L} \right) - E_g
698: \right).
699: \end{equation}
700: where $q$ is a wave number.
701: Then we extrapolate $v \left( L \right)$ as
702: \begin{equation}
703: v\left( L \right) = v \left( \infty \right) + a \frac{1}{L^2}
704: +b \frac{1}{L^4} + \mbox{higher order},
705: \end{equation}
706: where $a$ and $b$ are fitting parameters. We neglect higher order terms.
707:
708: Here we define the effective central charge in numerical calculations
709: as follows,
710: \begin{equation}
711: E_{\rm g} \left( L \right)=\epsilon L -\frac{\pi v \tilde{c}}{6L},
712: \end{equation}
713: where $\tilde{c}$ is the effective central charge.
714: The effective central charge is equivalent to the central charge
715: on the critical point or in the TL liquid phase.
716: And this changes rapidly from $\tilde{c}=1$ (the TL phase)
717: to $\tilde{c}\rightarrow0$ (the massive phase) \cite{I-N}.
718: In Fig.(\ref{fig:cc-xy-rs}), we show the effective central charge
719: obtained on the transition points.
720: In this figure, we find that the central charge
721: decreases from $\tilde{c}=2$ ($J_{\rm rung}=0$)
722: to $\tilde{c}=1$ (far from $J_{\rm rung}=0$).
723: This reflects Zamolodchikov's c-theorem \cite{c-theo}.
724: And we consider the phase transition between the rung-singlet phase
725: (massive )and the $XY$ phase (massless).
726: We see that the effective central charge $\tilde{c}$
727: is $1$ in normal $XY$ phases,
728: and $\tilde{c}$ is $2$ in $XY$ phase in two independent chains
729: $J_{\rm rung}=0$,
730: and $\tilde{c}$ becomes zero increasing the system size
731: ($L\rightarrow\infty$) in the rung-singlet phase
732: (see Fig.(\ref{fig:ecc})).
733:
734:
735: And then we should calculate scaling dimensions removing logarithmic
736: corrections.
737: From CFT, scaling dimensions in the finite size system under the PBC
738: are related to excitation energies as follows \cite{Cardy2}:
739: \begin{equation}
740: E_i \left( L \right) - E_g \left( L \right)= \frac{2\pi v x_i}{L},
741: \end{equation}
742: where $E_{i}$ is an excitation energy, $v$ is the spin wave velocity,
743: $L$ is the system size, $x_i$ is the scaling dimension.
744: In fact, there exists additional logarithmic corrections.
745: \begin{table}
746: \caption{
747: $P=1^*$ is even parity under TBC,
748: $P=-1^*$ is odd parity under TBC,
749: , $q$ is the wave number, $x$ is the scaling dimension,
750: $y_l \propto 1/\ln L$
751: }
752: \label{tb:sdim}
753: \begin{center}
754: \begin{tabular}{|c|c|c|c|c|c|}
755: \hline
756: $S_z^T$ & P & q & BC & x & abbr \\
757: \hline
758: $\pm2$ & 1 & 0 & PBC & $\frac{1}{2}-\frac{y_1}{4}$ & $x_{\pm2,0}$ \\
759: \hline
760: 0 & -1$^*$ & & TBC & $\frac{1}{2}+\frac{y_1}{4}-\frac{y_2}{2}$ & $x_{0,sin}^{TBC}$ \\
761: \hline
762: 0 & 1$^*$ & & TBC & $\frac{1}{2}+\frac{y_1}{4}+\frac{y_2}{2}$ & $x_{0,cos}^{TBC}$ \\
763: \hline
764: $\pm4$ & 1 & 0 & PBC & $2-y_1$ & $x_{\pm4,0}$ \\
765: \hline
766: 0 & 1 & 0 & PBC & $2-y_1 \left( 1+\frac{4t}{3} \right)$ & $x_{marg}$ \\
767: \hline
768: 0 & -1 & 0 & PBC & $2+y_1 $ & $x_{0,sin}$ \\
769: \hline
770: 0 & 1 & 0 & PBC & $2+2y_1 \left( 1+ \frac{2t}{3} \right) $ & $x_{0,cos}$ \\
771: \hline
772: \end{tabular}
773: \end{center}
774: \end{table}
775: From table (\ref{tb:sdim}), we obtain the following relation:
776: \begin{equation}
777: 2 x_{\pm2,0} + x_{0,{\rm sin}}^{\rm TBC} + x_{0,{\rm cos}}^{\rm TBC} =2,
778: \end{equation}
779: removing leading logarithmic corrections.
780: In Fig. \ref{fig:sd-xy-rs}, we show its size dependence.
781:
782: Further more, we check the consistency of BKT phase transition.
783: We calculate the ratio of two excitation energies, $S^z_T=1$ and $S^z_T=2$.
784: \begin{eqnarray}
785: \Delta E_{1}&=&E(S^z_T=1) - E_{\rm g}, \nonumber \\
786: \Delta E_{2}&=&E(S^z_T=2) - E_{\rm g},\nonumber \\
787: f&=&\frac{\Delta E_2}{\Delta E_1},
788: \label{eq:ratio}
789: \end{eqnarray}
790: where $E_{\rm g}$ is the ground state energy, $E(S^z_T=1,2)$ is the lowest
791: energy with the quantum number $S^z_T=1,2$.
792: This $f$ is equivalent to the ratio of scaling dimensions
793: on the TL liquid.
794: For Gaussian model \cite{K-B}, scaling dimensions are described as
795: \begin{equation}
796: x_{m,n}=\frac{1}{2}\left( \frac{m^2}{K} + n^2K \right),
797: \label{eq:scalingdimension}
798: \end{equation}
799: under the PBC, where $m$ is related to the magnetization of the system,
800: $n$ is not related to any conserved quantities,
801: $K$ is the TL parameter. $m$ and $n$ are integers.
802: From eq.(\ref{eq:ratio}) and eq.(\ref{eq:scalingdimension})
803: $f$ should be $4$ in one component TL liquid phase.
804:
805: However, when the system consists of two independent chain,
806: this $f$ should be 2 in TL liquid phase for the following reason.
807: In $S^z=1$ case, one chain has the ground state energy $E_{\rm g}$,
808: another chain has the excitation energy $E(S^z=1)$,
809: thus the system has the excitation energy $\Delta E_1$.
810: In $S^z=2$ case, each chain has the excitation energy $E(S^z=1)$,
811: so the system has the excitation energy $2\Delta E_1$
812: (see Appendix \ref{sec:ind-chain}).
813: In this case, $f$ should be 2.
814:
815: On the other hand, in a massive phase, the energy of two independent
816: magnons is twice the energy of one magnon in $L\rightarrow\infty$ limit.
817: If the interaction of magnons is repulsive,
818: then the ratio is more than 2 in the finite system.
819: If the interaction of magnons is attractive then the ratio is less than 2
820: in the finite system.
821: We show the ratio $f$ in Fig. \ref{fig:ratio}.
822: This figure support the second order phase transition between
823: the $XY1$ phase and $XY2$ phase, together with Fig. \ref{fig:ecc}.
824:
825:
826:
827:
828: \subsection{$XY$-Haldane transition}
829: %\hspace*{\parindent}
830: The procedure in the previous subsection can be applied to the
831: $XY$-Haldane transition, since we can expect that $XY$-Haldane transition is
832: BKT type.
833: In this case, we can identify critical points as cross points of two
834: low-lying excitations.
835: One has quantum number $S^z_T=\pm 2, P=1, q=0$ under the PBC,
836: another has quantum number $S^z_T=0, P^*=-1$ under the TBC.
837: We show crossing points in Fig.(\ref{fig:cp-xy-haldane}).
838: We can find the transition line is on $\Delta=0$.
839: As before, we calculate the effective central charge
840: and the scaling dimensions after removing logarithmic correction.
841: We show results in Fig. \ref{fig:cc-xy-haldane} for the central charge,
842: and in Fig. \ref{fig:sd-xy-haldane} for the scaling dimensions.
843: As a result, we obtain the phase boundary is on a $\Delta =0$ line.
844: This is analogous to the phase boundary
845: between XY phase and Haldane phase in $S=1$ $XXZ$ spin chain
846: \cite{dN-R,C-H-S}.
847:
848: We have found analytically that S=1 $XY$ spin chain has an additional
849: $SU(2)$ symmetry under the OBC and an artificial boundary
850: condition(ABC) in our previous paper\cite{K-H-N}.
851: In this paper, we show that there is an additional $SU(2)$
852: symmetry for $S=1/2$ two-leg $XXZ$ spin ladder system with $\Delta=0$
853: under the OBC and ABC (see Appendix \ref{sec:addsu2} and Ref.\cite{K-H-N}).
854: This supports our numerical calculations,
855: since an $SU(2)$ symmetry is related to the BKT transition
856: \cite{Ginsparg}.
857:
858:
859:
860: \subsection{Multicritical point $(\Delta,J_{\rm rung})=(0,0)$}
861: %\hspace*{\parindent}
862: In this subsection, we investigate excitations near the multi-critical
863: point $J_{\rm rung}=0$ and $\Delta=0$.
864: On this point, BKT transition line meets Gaussian transition line.
865: BKT transition line is a phase boundary between $XY$ phase and Haldane
866: phase, and between $XY$ phase and rung-singlet phase.
867: Gaussian transition line is a phase boundary between rung-singlet phase
868: and Haldane phase.
869: BKT transition line consists of two parts,
870: one is the crossing line ($S^z_T=\pm 2, P=1, q=0$ under the PBC)
871: and ($S^z_T=0, P^*=1$ under the TBC),
872: another is the crossing line ($S^z_T=\pm 2, P=1, q=0$ under the PBC)
873: and ($S^z_T=0, P^*=-1$ under the TBC).
874:
875: We show excitation energies with the following quantum numbers and
876: boundary conditions:
877: \begin{itemize}
878: \item $S^z_T=\pm 2, P=1, q=0$ under the PBC
879: \item $S^z_T=0, P^*=1$ under the TBC
880: \item $S^z_T=0, P^*=-1$ under the TBC
881: \end{itemize}
882: in Fig. (\ref{fig:mcp-xy-rs-o}) along the $J_{\rm rung}=\Delta$ line
883: and in Fig. (\ref{fig:mcp-xy-h-o}) along the $J_{\rm rung}=-4\Delta$
884: line in the parameter space.
885: We show these crossing points in Fig.(\ref{fig:cp-parity}).
886: This show that $(\Delta,J_{\rm rung})=(0,0)$ is a multicritical point.
887: $XY1$-Haldane transition line and $XY2$-rung singlet transition line
888: are continuously connected but these are not smoothly connected.
889:
890: From bosonization studies \cite{S-M,V-J-M}, $XY$-rung singlet transition
891: line seems to smoothly connect $XY$-Haldane transition line.
892: However, this is not consistent with CFT analysis.
893: Our numerical calculations support CFT analysis.
894: We can consider that this reflects an additional SU(2) symmetry
895: on $\Delta=0$ line \cite{K-H-N}.
896:
897:
898:
899:
900: \subsection{Multicritical point $(\Delta,J_{\rm rung})=(-1,0)$}
901: %\hspace*{\parindent}
902: In this subsection, we discuss the other multicritical point
903: $(\Delta,J_{\rm rung})=(-1,0)$.
904: This multicritical point is among the $XY$ phase, the Ferromagnetic
905: phase, the Stripe Ferromagnetic phase, and the rung-single phase.
906: This multicritical point has been studied by Kolezhuk and Mikeska \cite{K-M}
907: and Vekua {\it et al}.\cite{V-J-M}.
908: Kolezhuk and Mikeska discussed this,
909: mapping onto the anisotropic non-linear $\sigma$ model.
910: The renormalization group analysis of the anisotropic non-linear
911: $\sigma$ model was developed by Nelson and Pelcovits\cite{N-P}.
912: Vekua {\it et al}. also discussed the ferromagnetic leg case,
913: using the bosonization \cite{V-J-M}.
914: Outlines of the phase diagram are the same,
915: however there is different point.
916: Kolezhuk and Mikeska suggested that the shape of the phase boundary
917: between $XY$ phase and rung-singlet phase is the exponential function
918: $J_{\rm rung} \propto \exp \left(-a/\left|\Delta+1\right|\right)$,
919: where $a$ is a positive constant.
920: On the other hand, Vekua {\it et al}. suggested
921: that the shape of phase boundary between the $XY$ phase and rung-singlet phase
922: is the linear line on $\Delta-J_{\rm rung}$ plane.
923: Our numerical result strongly supports Kolezhuk's suggestion
924: (See Fig.\ref{fig:pd-xy-rs}).
925: This shows that non-linear $\sigma$ model with anisotropy
926: is more appropriate to describe the region near the multicritical point.
927:
928:
929:
930:
931: \section{CONCLUSIONS}
932: %\hspace{\parindent}
933: In this paper, we have studied the ground state phase diagram of S=1/2
934: two-leg spin ladder system.
935: We have accurately numerically determined the phase boundary
936: between the $XY$ phase and the Haldane phase,
937: and between the XY phase and the rung-singlet phase,
938: analyzing the exact diagonalization data using the level spectroscopy
939: method, TBC, and CFT.
940: And we have checked the universality class.
941: As a result, in $-1< \Delta < 0$ region, the $XY$ phase extends over the
942: $J_{\rm rung}=0$ line from the $J_{\rm rung}<0$ region to
943: the $J_{\rm rung}>0$ region.
944: And there does not exist the $XY$ phase in $\Delta>0$ region
945: except on the $J_{\rm rung}=0$ line,
946: since the BKT transition line is the $\Delta=0$ line.
947: We can understand this result, considering an additional $SU(2)$
948: symmetry \cite{K-H-N}.
949:
950: We have roughly discussed phase boundaries between the rung-singlet
951: phase and N\'eel phase, and between the rung-singlet phase and
952: the stripe ferromagnetic phase, using the variational method.
953: We think that these phase transitions are the second order phase
954: transitions, considering the symmetry,
955: and that the universality class is the Ising type.
956:
957: We have determined phase boundaries between the stripe ferromagnetic
958: phase and the fully ferromagnetic phase, and between the N\'eel phase
959: and the stripe N\'eel phase, considering degeneracies of ground states.
960: These phase transitions are of the first order type.
961:
962: We find that there are two different $XY$ phase.
963: We can distinguish these $XY$ phases, considering $XX$ (or $YY$)correlation
964: function.
965: This is based on the Marshall-Lieb-Mattis's theorem.
966: One $XY$ phase ($XY2$) is the $J_{\rm rung}>0$ region, other $XY$ phase ($XY1$) is
967: the $J_{\rm rung}<0$ region,
968: and it is understood that the second order phase transition occurs
969: numerically between the $XY1$ phase and the XY2 phase
970: from Fig. \ref{fig:ecc} and Fig. \ref{fig:ratio}.
971:
972: $J_{\rm rung}=0,\Delta=-1$ point is a multicritical point, among
973: the rung-singlet phase, the stripe ferromagnetic phase,
974: the fully ferromagnetic phase, and the $XY$ phase.
975: Seeing the $J_{\rm rung}>0$ region, the phase diagram is quite characteristic.
976: This phase diagram is similar to one of the anisotropic
977: nonlinear $\sigma$ model without the topological term
978: whose phase diagram and the renormalization group flow
979: have been discussed by Nelson and Pelcovits \cite{N-P}.
980: The same analysis can be applied to this problem.
981:
982:
983:
984:
985:
986: \section*{ACKNOWLEDGEMENTS}
987:
988: %\hspace*{\parindent}
989: Numerical calculations in this paper was based on T.I.T.pack ver.2
990: coded by Prof. H. Nishimori.
991:
992:
993:
994:
995: \appendix
996:
997: \section{Excitations of the two independent chains system}
998: \label{sec:ind-chain}
999: %\hspace*{\parindent}
1000: In $S=1/2$ two-leg ladder system,
1001: the $J_{\rm rung}=0$ case is described as two independent $S=1/2$ XXZ chains.
1002: An $S=1/2$ XXZ chain is exactly solved \cite{C-G},
1003: using the Bethe ansatz \cite{Bethe}.
1004: This model is described by the Hamiltonian
1005: \begin{equation}
1006: {\cal H}=J \sum_{i} \left( S^x_{i} S^x_{i+1} + S^y_{i} S^y_{i+1}
1007: + \Delta S^z_{i} S^z_{i+1} \right) .
1008: \end{equation}
1009: For $\Delta >1$, the system becomes the N\'eel (Anti Ferromagnetic)
1010: phase, whereas for $\Delta <-1$, the system becomes the Ferromagnetic
1011: phase. In the intermediate case $-1 \leq \Delta \leq 1$, the system
1012: becomes XY (spin fluid) phase.
1013:
1014: In this appendix, we discuss the excitation structure of two independent
1015: S=1/2 XY spin chain.
1016:
1017: In XY phase, the system is described as $c=1$ gaussian model.
1018: Now we have two independent chains.
1019: We describe the state of all systems as $\Psi$.
1020: We think that $\Psi_1$ is the state of chain 1,
1021: and $\Psi_2$ is that of chain 2.
1022: \begin{itemize}
1023: \item $\Psi = \Psi_1 \otimes \Psi_2$
1024: \item $q=q_1+q_2$ : wave number
1025: \item $S^T_z=S^{T1}_z+S^{T2}_z$ : magnetization
1026: \end{itemize}
1027:
1028: Excitations of the single chain are, for example,
1029: \begin{itemize}
1030: \item $q=0$,$S^T_z=0$,$x=0$ : the ground state
1031: \item $q=\pi$,$S^T_z=0$,$x=1$
1032: \item $q=0$,$S^T_z=0$,$x=2$
1033: \item $q=\pi$,$S^T_z=\pm1$,$x=\frac{1}{4}$.
1034: \end{itemize}
1035:
1036:
1037:
1038: Excitation of the system are as follows,
1039: \begin{itemize}
1040:
1041: \item $
1042: \left( q=\pi, S^{T1}_z=0, x=1 \right)
1043: \otimes
1044: \left( q=\pi, S^{T2}_z=0, x=1 \right)
1045: =
1046: \left( q=0(2\pi), S^{T}_z=0, x=2 \right)
1047: $
1048:
1049: \item $
1050: \left( q=0, S^{T1}_z=0, x=2 \right)
1051: \otimes
1052: \left( q=0, S^{T2}_z=0, x=0 \right)
1053: =
1054: \left( q=0(2\pi), S^{T}_z=0, x=2 \right)
1055: $
1056:
1057: \item $
1058: \left( q=\pi, S^{T1}_z=+1, x=\frac{1}{4} \right)
1059: \otimes
1060: \left( q=\pi, S^{T2}_z=-1, x=\frac{1}{4} \right)
1061: =
1062: \left( q=0(2\pi), S^{T}_z=0, x=\frac{1}{2} \right)
1063: $
1064:
1065: \end{itemize}
1066:
1067: The last one transforms as follows,
1068: under the spin reversion $S_z \leftrightarrow -S_z$,
1069:
1070: \begin{center}
1071:
1072: $\left( q=\pi, S^{T}_z=+1, x=\frac{1}{4} \right)
1073: \otimes
1074: \left( q=\pi, S^{T}_z=-1, x=\frac{1}{4} \right)$ \\
1075: $\Downarrow$ \\
1076: $ \left( q=\pi, S^{T}_z=-1, x=\frac{1}{4} \right)
1077: \otimes
1078: \left( q=\pi, S^{T}_z=+1, x=\frac{1}{4} \right) $
1079: \end{center}
1080:
1081: This has the odd rung-parity.
1082: This excitation is relevant, it does not appear
1083: in the space where the rung-parity is restricted to even.
1084:
1085:
1086: \section{An additional $SU(2)$ symmetry of the
1087: $S=1/2$ two-leg spin ladder system\cite{K-H-N}}
1088: \label{sec:addsu2}
1089: %\hspace*{\parindent}
1090: We present a brief review of an additional $SU(2)$ symmetry
1091: of the $S=1/2$ two-leg ladder system.
1092: In the $S=1$ XY chain and the $S=1/2$ two-leg ladder system,
1093: there exists an additional $SU(2)$ symmetry
1094: besides the usual spin $SU(2)$ symmetry.
1095: In our previous paper \cite{K-H-N},
1096: we show that the one dimensional spin-$1$ XY model has
1097: an additional $SU(2)$ symmetry.
1098: And we also show the case of $S=1/2$ two-leg ladder system.
1099:
1100: We consider the Hamiltonian (\ref{eq:Hamiltonian}).
1101: At first, we introduce the following local operators:
1102: \begin{equation}
1103: \tilde{s}^{\pm}_j=S^{\pm}_{1,j}S^{\pm}_{2,j}
1104: , \hspace{1cm}
1105: \tilde{s}^z_j=\frac{1}{2}\left(S^z_{1,j}+S^z_{2,j}\right).
1106: \end{equation}
1107: After the simple calculation,
1108: we obtain the commutation relation of $SU(2)$,
1109: \begin{equation}
1110: \left[\tilde{s}^z_j, \tilde{s}^{\pm}_k \right]=\pm\delta_{jk}
1111: \tilde{s}^{\pm}_j , \hspace{1cm}
1112: \left[\tilde{s}^+_j, \tilde{s}^-_k\right]=2\delta_{jk}\tilde{s}^z_j
1113: \end{equation}
1114: The operator $\sum_j \tilde{s}^z_j$ commutes with the Hamiltonian
1115: (\ref{eq:Hamiltonian}), but operators $\sum_j\tilde{s}^{\pm}_j$ do not.
1116:
1117: Thus we introduce new non-local operators,
1118: \begin{equation}
1119: s^{\pm}_j=S^{\pm}_{1,j}S^{\pm}_{2,j}\prod^{j-1}_{l=1}
1120: \left(-4S^z_{1,l}S^z_{2,l}\right) \ \ \ \
1121: s^z_j=\frac{1}{2}\left(S^z_{1,j}+S^z_{2,j}\right).
1122: \end{equation}
1123: New operators satisfy the commutation relation of $SU(2)$,
1124: \begin{equation}
1125: \left[s^z_j,s^{\pm}_k\right]=\pm\delta_{jk}s^{\pm}_j \ \ \ \
1126: \left[s^+_j,s^-_k\right]=2\delta_{jk}s^z_j
1127: \end{equation}
1128: and
1129: \begin{equation}
1130: \left[s^z_T,s^{\pm}_T\right]=\pm s^{\pm}_T \ \ \ \
1131: \left[s^+_T,s^-_T\right]=2s^z_T.
1132: \end{equation}
1133: where $s^{\pm}_T=\sum_j s^{\pm}_j$ and
1134: $s^z_T\sum_j s^z_j$ are total operators.
1135:
1136: After the some calculation,
1137: we can show that these total operators
1138: commute with the Hamiltonian (\ref{eq:Hamiltonian})
1139: with $\Delta=0$ under the open boundary condition
1140: and an artificial boundary condition.
1141: For a discussion in detail, please see our previous paper\cite{K-H-N}.
1142:
1143:
1144: \begin{thebibliography}{99.}
1145:
1146: \addcontentsline{toc}{section}{References}
1147:
1148:
1149: \bibitem{D-R} E. Dagotto and T. M. Rice, Science. {\bf 271}, 618 (1996);
1150: E. Dagotto, Rep. Prog. Phys. {\bf 62}, 1525 (1999)
1151:
1152: \bibitem{Haldane} F. D. M. Haldane, Phys. Rev. Lett. {\bf 50}, 1153 (1983)
1153: ; Phys. Lett. {\bf 93A}, 464 (1983)
1154:
1155: \bibitem{L-S-M} E. H. Lieb, T. Schulz and D. Mattis, Ann. Phys. {\bf 16},
1156: 407 (1961)
1157:
1158: I. Affleck and E. H. Lieb, Lett. Math. Phys. {\bf 12}, 57 (1986)
1159:
1160: \bibitem{VBS} I. Affleck, T. Kennedy, E. H. Lieb and H. Tasaki,
1161: Phys. Rev. Lett. {\bf 59}, 799 (1987); Commun. Math. Phys.
1162: {\bf 115}, 477 (1988)
1163:
1164: \bibitem{Schulz} H. J. Schulz, Phys. Rev. B. {\bf 34}, 6372 (1986)
1165:
1166: \bibitem{S-M} S. P. Strong and A. J. Mills, Phys. Rev. Lett.{\bf 69},
1167: 2419 (1992) ; Phys. Rev. B. {\bf 50}, 9911 (1994)
1168:
1169: \bibitem{W-N-T} H. Watanabe, K. Nomura and S. Takada,
1170: J. Phys. Soc. Jpn. {\bf 62}, 2845 (1993)
1171:
1172: \bibitem{N-N-T} T. Narushima, T. Nakamura and S. Takada,
1173: J. Phys. Soc. Jpn. {\bf 64}, 4322 (1995)
1174:
1175: \bibitem{DMRG} S. R. White, Phys. Rev. Lett, {\bf 69}, 2863 (1992)
1176: ; Phys. Rev. B, {\bf48}, 10345 (1993)
1177:
1178: \bibitem{U-N-A-T-M-K}M. Ueda, T. Nagata, J. Akimitsu, H. Takahashi,
1179: N. M\^{o}ri, and K. Kinoshita,
1180: J. Phys. Soc. Jpn, {\bf 65}, 2764, (1996)
1181:
1182: \bibitem{K-M} A. K. Kolezhuk and H.-J. Mikeska, Phys. Rev. B.
1183: {\bf 53}, R8848 (1996)
1184:
1185: \bibitem{K-T-J1}K. Kopinga, A. M. C. Tinus, and W. J. M. de Jonge,
1186: Phys. Rev. B. {\bf 25}, 4685, (1982)
1187:
1188: \bibitem{K-T-J2}K. Kopinga, A. M. C. Tinus, and W. J. M. de Jonge,
1189: Phys. Rev. B. {\bf 29}, 2868, (1984)
1190:
1191: \bibitem{V-J-M} T. Vekua, G. I. Japaridze and H.-J. Mikeska,
1192: Phys. Rev. B. {\bf 67}, 064419 (2003)
1193:
1194: \bibitem{L-S} O. Legeza and J. Solyom,
1195: Phys. Rev. B. {\bf 56}, 14449, (1997)
1196:
1197: \bibitem{S-N-T} D. G. Shelton, A. A. Nersesyan and A. M. Tsvelik,
1198: Phys. Rev. B. {\bf 53}, 8521 (1996)
1199:
1200: \bibitem{L-O} P. Lecheminant and E. Orignac, Phys. Rev. B. {\bf 65},
1201: 174406, (2002)
1202:
1203: \bibitem{LS} K. Nomura and K. Okamoto, J. Phys. A: Math. Gen.
1204: {\bf 27}, 5773 (1994); K. Nomura, J. Phys. A: Math. Gen.
1205: {\bf 28}, 5451 (1995)
1206:
1207: \bibitem{T-L} S. Tomonaga Prog. Theo. Phys, {\bf 5}, 544 (1950)
1208: ;J. M. Luttinger J. Math. Phys, {\bf 4},1154 (1963)
1209: ;D. C. Mattis, and E. H. Lieb J. Math. Phys, {\bf 6}, 304, (1965)
1210:
1211:
1212: \bibitem{BKT}Z. L. Berezinskii, Zh. Eksp. Teor. Fiz, {\bf 59}, 907
1213: (1970) (Sov. Phys. JETP, {\bf 32}, 493 (1971))
1214: ; Zh. Eksp. Teor. Fiz, {\bf 61}, 1144, (1971)
1215: (Sov. Phys. JETP, {\bf 34}, 610 (1972));
1216: J. M. Kosterlitz and D. J. Thouless, J. Phys. C, {\bf 6}, 1181 (1973);
1217: J. M. Kosterlitz, J. Phys. C, {\bf 7}, 1046 (1974)
1218:
1219: \bibitem{F-B}M. E. Fisher and M. N. Barber, Phys. Rev. Lett. {\bf 28},
1220: 1516 (1972)
1221:
1222: \bibitem{fsBKT} J. C. Bonner and G. M\"uller, Phys. Rev. B. {\bf 29},
1223: 5216 (1984); J. S\'olyom and T. A. L. Ziman, Phys. Rev. B.
1224: {\bf 30}, 3980 (1984)
1225:
1226: \bibitem{I-N}H. Inoue and K. Nomura, Phys. Lett. A, {\bf 262}, 96 (1999)
1227:
1228: \bibitem{Kitazawa} A. Kitazawa, J. Phys. A: Math. Gen. {\bf 30}, L285 (1997)
1229: \bibitem{N-K} K. Nomura and A. Kitazawa, J. Phys. A.: Math. Gen. {\bf 31},
1230: 7341 (1998)
1231:
1232: \bibitem{V-J-M2} T. Vekua, G. I. Japaridze and H.-J. Mikeska,
1233: Phys. Rev. B. {\bf 70}, 014425 (2004)
1234:
1235: \bibitem{Bethe} H. A. Bethe, Z. Phys. {\bf 71}, 205 (1931)
1236:
1237: \bibitem{C-G} J. des Cloizeaux and M. Gaudin, J. Math. Phys.
1238: {\bf 7}, 1384 (1966)
1239:
1240: \bibitem{B-I-K} N. M. Bogoliubov, A. G. Izergin and V. E. Korepin,
1241: Nucl. Phys. {\bf B 275}, 687 (1986)
1242:
1243: \bibitem{dN-R} M. den Nijs and K. Rommelse, Phys. Rev. B {\bf 40}, 4709
1244: (1989)
1245: \bibitem{C-H-S} W. Chen, K. Hida and B. C. Sanctuary, Phys. Rev. B {\bf 67},
1246: 104401 (2003)
1247:
1248: \bibitem{M-L-M} W. Marshall, Proc. Roy. Soc. {\bf A 232}, 48, (1955)
1249:
1250: E. Lieb and D. Mattis, J. Math. Phys. {\bf 3}, 749, (1962)
1251:
1252: \bibitem{Cardy}J. L. Cardy, Nucl. Phys. B. {\bf 270}, 186, (1986)
1253:
1254: \bibitem{CFT} P. Di Francesco, P. Mathieu, and D. S\'en\'echal,
1255: {\it Conformal Field Theory} (Springer-Verlag, New York,1997)
1256:
1257: \bibitem{B-C-N} H. W. J. Bl\"ote, J. L. Cardy, and M. P. Nightingale,
1258: Phys. Rev. Lett. {\bf 56}, 742 (1986)
1259: \bibitem{Affleck} I. Affleck, Phys. Rev. Lett. {\bf 56}, 746 (1986)
1260: \bibitem{Cardy1} J. L. Cardy, J. Phys. A. Math. Gen {\bf 19}, L1093
1261: (1986)
1262:
1263: \bibitem{c-theo} A. B. Zamolodchikov, Pis'ma Zh. Eksp. Theor. Fiz.
1264: {\bf 43}, 565 (1986) [JETP Lett. {\bf 43}, 730 (1986)]
1265:
1266: \bibitem{Cardy2} J. L.Cardy, J.Phys.A.Math.Gen {\bf 17} L385 (1984)
1267:
1268: \bibitem{K-B} L. P. Kadanoff and A. C. Brown, Ann. Phys. {\bf 121}
1269: 318, (1979)
1270:
1271: \bibitem{K-H-N} A. Kitazawa, K. Hijii and K. Nomura,
1272: J. Phys. A: Math. Gen. {\bf 36}, L351-L357 (2003)
1273:
1274: \bibitem{Ginsparg}P. Ginsparg, Nucl.Phys. {\bf B 295}, 153, (1988)
1275:
1276: \bibitem{N-P} D. R. Nelson and R. A. Pelcovits, Phys. Rev. B. {\bf 16}
1277: 2191 (1977)
1278:
1279: \end{thebibliography}
1280:
1281:
1282: \newpage
1283:
1284:
1285: \begin{figure}
1286: \begin{center}
1287: \epsfig{file=./ladder.eps,width=10cm}
1288: \caption{Schematic structure of a S=1/2 two-leg spin ladder of Eq (1)}
1289: \label{fig:ladder}
1290: \end{center}
1291: \end{figure}
1292:
1293:
1294: \begin{figure}
1295: \begin{center}
1296: \epsfig{file=./pdm.eps,width=10cm}
1297: \caption{
1298: Schematic phase diagram in the S=1/2 XXZ ladder system
1299: on $\Delta-J_{\rm rung}$ plane. Dotted lines are phase boundaries.
1300: }
1301: \label{fig:pdm}
1302: \end{center}
1303: \end{figure}
1304:
1305:
1306:
1307: \begin{figure}
1308: \begin{center}
1309: \epsfig{file=./state.eps,width=10cm}
1310: \caption{
1311: (a) Schematic picture of the Stripe Ferromagnetic state.
1312: This phase has ferromagnetic order in the leg direction,
1313: and antiferromagnetic order in the rung direction.
1314: (b) Schematic picture of Stripe N\'eel state.
1315: This phase has antiferromagnetic order in the leg direction,
1316: and ferromagnetic order in the rung direction.
1317: (c) Schematic picture of N\'eel state.
1318: This phase has antiferromagnetic order in the leg direction,
1319: and antiferromagnetic order in the rung direction.
1320: }
1321: \label{fig:Stripe}
1322: \end{center}
1323: \end{figure}
1324:
1325:
1326: \begin{figure}
1327: \begin{center}
1328: \epsfig{file=./negative.eps,width=10cm}
1329: \caption{
1330: Schematic figure for the unitary transformation in the $J_{\rm rung}<0$
1331: case.
1332: Empty circles are uncchanged.
1333: Full circles are $S^{\pm}_{\alpha,j} \rightarrow -S^{\pm}_{\alpha,j}$
1334: }
1335: \label{fig:nrung}
1336: \end{center}
1337: \end{figure}
1338:
1339:
1340: \begin{figure}
1341: \begin{center}
1342: \epsfig{file=./positive.eps,width=10cm}
1343: \caption{
1344: Schematic figure for the unitary transformation in the $J_{\rm rung}>0$
1345: case.
1346: Empty circles are uncchanged.
1347: Full circles are $S^{\pm}_{\alpha,j} \rightarrow -S^{\pm}_{\alpha,j}$
1348: }
1349: \label{fig:prung}
1350: \end{center}
1351: \end{figure}
1352:
1353:
1354:
1355: \begin{figure}
1356: \begin{center}
1357: \epsfig{file=./cp24site-xy-rs.eps,width=10cm}
1358: \caption{
1359: Excitation energies of $L=12(=N/2)$, $\Delta=-0.5$ near the XY-Rung
1360: Singlet transition point. $\times$'s are $S^z_T=\pm2$,$q=0$,$P=1$ under
1361: the periodic boundary condition. $+$'s are $S^z_T=0$,$P^*=1$ under
1362: the twisted boundary condition.$*$'s are $S^z_T=0$,$P^*=-1$ under
1363: the twisted boundary condition.
1364: }
1365: \label{fig:cp-xy-rs}
1366: \end{center}
1367: \end{figure}
1368:
1369:
1370: \begin{figure}
1371: \begin{center}
1372: \epsfig{file=./deltam050.eps,width=10cm}
1373: \caption{Size dependence of the crossing points. The extrapolated
1374: value is 0.3727 at $\Delta=-0.5$.
1375: Dotted line is determined using a linear least method.
1376: }
1377: \label{fig:sizeeff}
1378: \end{center}
1379: \end{figure}
1380:
1381: \begin{figure}
1382: \begin{center}
1383: \epsfig{file=./pd-xy-rs.eps,width=10cm}
1384: \caption{
1385: crossing points and their size dependence for XY-Rung Singlet transition,
1386: for N=12,16,20,24,and extrapolated values.
1387: }
1388: \label{fig:pd-xy-rs}
1389: \end{center}
1390: \end{figure}
1391:
1392:
1393: \begin{figure}
1394: \begin{center}
1395: \epsfig{file=./cc-xy-rs.eps,width=10cm}
1396: \caption{The effective central charge on XY-Rung Singlet transition
1397: line. $\times$'s are extrapolated using $L(=N/2)=6,8,10$.
1398: $+$'s are extrapolated using $L(=N/2)=8,10,12$.
1399: }
1400: \label{fig:cc-xy-rs}
1401: \end{center}
1402: \end{figure}
1403:
1404: \begin{figure}
1405: \begin{center}
1406: \epsfig{file=./cc.eps,width=10cm}
1407: \caption{
1408: The effective central charge $\tilde{c}$
1409: as a function of $J_{\rm rung}$ at $\Delta=-0.5$.
1410: $\times$ are extrapolated using $N=12,16,20$.
1411: $+$ are extrapolated using $N=16,20,24$.
1412: }
1413: \label{fig:ecc}
1414: \end{center}
1415: \end{figure}
1416:
1417:
1418:
1419: \begin{figure}
1420: \begin{center}
1421: \epsfig{file=./scal-xy-rs.eps,width=10cm}
1422: \caption{Size dependence of scaling dimensions removed logarithmic
1423: correction on the XY-Rung Singlet transition line
1424: for the system size $N=(2L)=12,16,20,24$.
1425: }
1426: \label{fig:sd-xy-rs}
1427: \end{center}
1428: \end{figure}
1429:
1430:
1431: \begin{figure}
1432: \begin{center}
1433: \epsfig{file=./ratio.eps,width=10cm}
1434: \caption{
1435: The ratio of energy gaps as a function of $J_{\rm rung}$
1436: at $\Delta=-0.5$ and its size dependence.
1437: }
1438: \label{fig:ratio}
1439: \end{center}
1440: \end{figure}
1441:
1442:
1443:
1444: \begin{figure}
1445: \begin{center}
1446: \epsfig{file=./cp24site-xy-h.eps,width=10cm}
1447: \caption{
1448: Excitation energies of $L=12(=N/2)$, $J_{\rm rung}=-0.5$ near the XY-Haldane
1449: transition point. $\times$'s are $S^z_T=\pm2$,$q=0$,$P=1$ under
1450: the periodic boundary condition. $+$'s are $S^z_T=0$,$P^*=1$ under
1451: the twisted boundary condition.$*$'s are $S^z_T=0$,$P^*=-1$ under
1452: the twisted boundary condition.
1453: }
1454: \label{fig:cp-xy-haldane}
1455: \end{center}
1456: \end{figure}
1457:
1458:
1459:
1460: \begin{figure}
1461: \begin{center}
1462: \epsfig{file=./cc-xy-h.eps,width=10cm}
1463: \caption{The effective central charge on XY-Haldane transition line.
1464: $\times$'s are extrapolated using $L(=N/2)=6,8,10$.
1465: $+$'s are extrapolated using $L(=N/2)=8,10,12$.
1466: }
1467: \label{fig:cc-xy-haldane}
1468: \end{center}
1469: \end{figure}
1470:
1471:
1472: \begin{figure}
1473: \begin{center}
1474: \epsfig{file=./scal-xy-h.eps,width=10cm}
1475: \caption{Size dependence of scaling dimensions removed logarithmic
1476: correction on the XY-Haldane transition line ($\Delta=0$)
1477: for the system size $N(=2L)=12,16,20,24$.
1478: }
1479: \label{fig:sd-xy-haldane}
1480: \end{center}
1481: \end{figure}
1482:
1483:
1484:
1485: \begin{figure}
1486: \begin{center}
1487: \epsfig{file=./24xy-rs-o.eps,width=10cm}
1488: \caption{
1489: Excitation energies of $L=12(=N/2)$, on $\Delta=J_{\rm rung}$ line
1490: near the multicritical point.
1491: $\times$'s are $S^z_T=\pm2$,$q=0$,$P=1$ under
1492: the periodic boundary condition. $+$'s are $S^z_T=0$,$P^*=1$ under
1493: the twisted boundary condition. $*$'s are $S^z_T=0$,$P^*=-1$ under
1494: the twisted boundary condition.
1495: }
1496: \label{fig:mcp-xy-rs-o}
1497: \end{center}
1498: \end{figure}
1499:
1500:
1501:
1502: \begin{figure}
1503: \begin{center}
1504: \epsfig{file=./24xy-h-o.eps,width=10cm}
1505: \caption{
1506: Excitation energies of $L=12(=N/2)$, on $\Delta=-4J_{\rm rung}$ line
1507: near the multicritical point.
1508: $\times$'s are $S^z_T=\pm2$,$q=0$,$P=1$ under
1509: the periodic boundary condition. $+$'s are $S^z_T=0$,$P^*=1$ under
1510: the twisted boundary condition. $*$'s are $S^z_T=0$,$P^*=-1$ under
1511: the twisted boundary condition.
1512: }
1513: \label{fig:mcp-xy-h-o}
1514: \end{center}
1515: \end{figure}
1516:
1517:
1518:
1519:
1520:
1521: \begin{figure}
1522: \begin{center}
1523: \epsfig{file=./parity.eps,width=10cm}
1524: \caption{
1525: Extrapolated cross points of excitation energies.
1526: $\times$'s are cross points of the excitation with
1527: $S^z_T=\pm2$,$q=0$,$P=1$ under the periodic boundary condition and
1528: the excitation with $S^z_T=0$,$P^*=1$ under the twisted boundary condition.
1529: $+$'s are cross points of the excitation with $S^z_T=\pm2$,$q=0$,$P=1$ under
1530: the periodic boundary condition and the excitation $S^z_T=0$,$P^*=-1$
1531: under the twisted boundary condition.
1532: }
1533: \label{fig:cp-parity}
1534: \end{center}
1535: \end{figure}
1536:
1537:
1538:
1539:
1540:
1541:
1542:
1543:
1544:
1545: \end{document}
1546: