cond-mat0401478/text.tex
1: % ****** Start of file apssamp.tex ******
2: %
3: % This file is part of the APS files in the REVTeX 3.0 distribution.
4: % Version 3.0 of REVTeX, November 10, 1992.
5: %
6: % Copyright (c) 1992 The American Physical Society.
7: %
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9: %\documentstyle[preprint,eqsecnum,aps]{revtex}
10: %\documentstyle[preprint,aps]{revtex}
11: \documentstyle[multicol,aps,psfig]{revtex}
12: %\topmargin -8mm
13: %\oddsidemargin -8mm \evensidemargin -7mm
14: %-------------------------------------------------------------------
15: \renewcommand{\narrowtext}{\begin{multicols}{2}
16: \global\columnwidth20.5pc\noindent}
17: \renewcommand{\widetext}{\end{multicols}
18: \global\columnwidth42.5pc}
19: \multicolsep = 8pt plus 4pt minus 3pt
20: %-------------------------------------------------------------------
21: \begin{document}
22: \draft
23: \preprint{13 November 2003}
24: \title{Nuclear Magnetic Relaxation in the Haldane-Gap Antiferromagnet
25:        Ni(C$_2$H$_8$N$_2$)$_2$NO$_2$(ClO$_4$)}
26: \author{Shoji Yamamoto and Hiromitsu Hori}
27: \address{Division of Physics, Hokkaido University,
28:          Sapporo 060-0810, Japan}
29: %\date{Received \hspace{4cm}}
30: \date{Received 13 November 2003}
31: \maketitle
32: \begin{abstract}
33: A new theory is proposed to interpret nuclear spin-lattice relaxation-time
34: ($T_1$) measurements on the spin-$1$ quasi-one-dimensional Heisenberg
35: antiferromagnet Ni(C$_2$H$_8$N$_2$)$_2$NO$_2$(ClO$_4$) (NENP).
36: While Sagi and Affleck pioneeringly discussed this subject in terms of
37: field-theoretical languages, there is no theoretical attempt yet to
38: explicitly simulate the novel observations of $T_1^{-1}$ reported
39: by Fujiwara {\it et al.}.
40: By means of modified spin waves, we solve {\it the minimum of $T_1^{-1}$
41: as a function of an applied field}, pending for the past decade.
42: \end{abstract}
43: \pacs{PACS numbers: 76.60.$-$k, 76.50.$+$g, 75.10.Jm}
44: % 75.10.Jm: Quantized spin models
45: % 76.50.$+$g: Ferromagnetic, antiferromagnetic, and ferrimagnetic
46: % 05.30.Jp: Boson systems
47: % 75.40.Mg: Numerical simulation studies
48: %             resonances; spin-wave resonance
49: % 75.50.Xx: Molecular magnets
50: % 76.60.$-$k: Nuclear magnetic resonance and relaxation
51: \narrowtext
52: 
53:    Predicting a striking contrast between integer- and
54: half-odd-integer-spin one-dimensional Heisenberg antiferromagnets, Haldane
55: \cite{H464,H1153} sparked renewed interest in low-dimensional quantum
56: magnetism.
57: The Haldane gap, that is, a magnetic excitation gap immediately above the
58: ground state, was not only calculated by various numerical tools
59: \cite{W3844,G3037,Y3348,S493,W14529,T047203} but indeed observed in
60: spin-$1$ quasi-one-dimensional Heisenberg antiferromagnets such as
61: CsNiCl$_3$ \cite{B371} and Ni(C$_2$H$_8$N$_2$)$_2$NO$_2$(ClO$_4$) (NENP)
62: \cite{R945}.
63: The valence-bond-solid model \cite{A799,A477} significantly contributed
64: toward understanding novel features of the Haldane massive phase such as
65: fractional spins induced on the boundaries \cite{W2863,M913}, a string
66: order hidden in the ground state \cite{N4709,Y9528} and magnon excitations
67: against the hidden order \cite{K627,F8983,Y157,Y1795}.
68: The nonlinear-$\sigma$-model quantum field theory \cite{A397,A409}
69: skillfully visualized the competition between massive and massless phases,
70: while a generalized Lieb-Schultz-Mattis theorem \cite{O1984,T103} gave a
71: criterion for the gap formation in a magnetic field.
72: 
73:    Recent progress in the experimental studies also deserves special
74: mention.
75: Not only the single-magnon dispersion relation \cite{T2313} but also the
76: two-magnon continuum \cite{W3844,Y545} was directly observed by
77: inelastic-neutron-scattering measurements on NENP \cite{M3571} and
78: CsNiCl$_3$ \cite{Z017202}.
79: An applied magnetic field may destroy the Haldane gap and bring back
80: magnetism to the system.
81: Such a field-induced long-range order was indeed realized in a nickel
82: compound Ni(C$_5$H$_{14}$N$_2$)$_2$N$_3$(PF$_6$) \cite{H2566}.
83: Another family of linear-chain nickelates of general formula
84: $R_2$BaNiO$_5$ ($R=\mbox{rare\ earth\ or\ Y}$) \cite{D409} exhibited
85: a novel scenario of one- to three-dimensional crossover \cite{M68,T15189}.
86: When the nonmagnetic Y$^{3+}$ ions are substituted by other magnetic
87: rare-earth ions in Y$_2$BaNiO$_5$ with a disordered ground state, there
88: appears a three-dimensional long-range order, while the one-dimensional
89: gapped excitations persist both above and below the N\'eel temperature
90: \cite{Z6437,Z7210,Y11516}.
91: 
92:    Nuclear spin-lattice relaxation time ($T_1$) has also been measured on
93: Haldane-gap antiferromagnets.
94: The field ($H$) dependence of $T_1^{-1}$ is of particular interest at both
95: high and low temperatures.
96: The high-temperature relaxation rate has been discussed in the context of
97: transport properties.
98: Takigawa {\it et al.} \cite{T2173} measured the high-temperature
99: relaxation rate of $^{31}$P and $^{51}$V nuclei of AgVP$_2$S$_6$, which is
100: also an ideal spin-$1$ Haldane-gap antiferromagnet, and observed the
101: diffusive dynamics $T_1^{-1}\propto H^{-1/2}$ \cite{B2215}.
102: There is a hot argument \cite{S943,F2810,S2712,F2714} whether the spin
103: transport in quantum spin-gapped antiferromagnets should be diffusive or
104: ballistic at finite temperatures.
105: On the other hand, Gaveau {\it et al.} \cite{G647} and Fujiwara
106: {\it et al.} \cite{F7837,F11860} measured the low-temperature relaxation
107: rate of $^1$H nuclei of NENP.
108: As an applied field increases, the first excited state moves down and then
109: crosses the ground-state energy level.
110: Indeed $T_1^{-1}$ reaches a peak near the critical field
111: $H_{\rm c}\equiv{\mit\Delta}_0/g\mu_{\rm B}$ at every temperature, but it
112: is not a monotonically increasing function of $H$, at low temperatures
113: $T\alt{\mit\Delta}_0/k_{\rm B}$ in particular.
114: Sagi and Affleck \cite{S9188} formulated the nuclear magnetic resonance in
115: Haldane-gap antiferromagnets in terms of field-theoretical languages.
116: However, few investigations have followed their pioneering argument and
117: there is no attempt yet to explicitly fit a theory for the above
118: observations.
119: 
120:    In such circumstances, we revisit the low-temperature nuclear magnetic
121: relaxation in Haldane-gap antiferromagnets with particular emphasis on its
122: field dependence.
123: Excluding any phenomenological assumption from our argument, we calculate
124: the nuclear spin-lattice relaxation rate by means of modified spin waves.
125: A new theory claims that {\it as an applied field increases, $T_1^{-1}$
126: should initially decrease logarithmically and then increase
127: exponentially}, well explaining experimental observations.
128: 
129:    We employ the spin-$1$ one-dimensional Heisenberg Hamiltonian
130: \begin{equation}
131:    {\cal H}
132:       =J\sum_{l=1}^L
133:         \mbox{\boldmath$S$}_{l} \cdot \mbox{\boldmath$S$}_{l+1}
134:       -g\mu_{\rm B}H\sum_{l=1}^L S_l^z\,.
135:    \label{E:H}
136: \end{equation}
137: In order to illuminate the essential relaxation mechanism in spin-gapped
138: antiferromagnets as analytically as possible, we do not take any
139: anisotropy into consideration in this letter.
140: Magnetic anisotropy quantitatively affects the gap amplitude but has no
141: qualitative effect on the whole scenario.
142: Scaling temperature and an applied field by the Haldane gap, we present a
143: universal theory.
144: Quantitative refinement of the final product in the presence of single-ion
145: and orthorhombic anisotropy terms will be considered elsewhere.
146: Introducing the Holstein-Primakoff bosonic operators
147: \begin{equation}
148:    \left.
149:    \begin{array}{lll}
150:       S_{2n-1}^+=\sqrt{2S-a_n^\dagger a_n}\ a_n\,,&
151:       S_{2n-1}^z=S-a_n^\dagger a_n\,,\\
152:       S_{2n}^+=b_n^\dagger\sqrt{2S-b_n^\dagger b_n}\ ,&
153:       S_{2n}^z=-S+b_n^\dagger b_n\,,
154:    \end{array}
155:    \right.
156:    \label{E:HP}
157: \end{equation}
158: and retaining only bilinear terms of them, we rewrite the Hamiltonian as
159: \begin{eqnarray}
160:    &&
161:    {\cal H}=-2JS^2N
162:    \nonumber \\
163:    &&\quad\ \ 
164:    +(2JS+g\mu_{\rm B}H)\sum_{n=1}^N a_n^\dagger a_n
165:    +(2JS-g\mu_{\rm B}H)\sum_{n=1}^N b_n^\dagger b_n
166:    \nonumber \\
167:    &&\quad\ \ 
168:    +JS\sum_{n=1}^N
169:     \left(
170:      a_n^\dagger b_n^\dagger+a_n b_n
171:     +b_n^\dagger a_{n+1}^\dagger+b_n a_{n+1}
172:     \right)\,,
173:    \label{E:HHP}
174: \end{eqnarray}
175: where $N\equiv L/2$.
176: In order to preserve the up-down symmetry, or the sublattice symmetry, we
177: optimize the spin-wave distribution functions constraining
178: the total staggered magnetization to be zero
179: \cite{T2494,H4769,T5000,Y769}:
180: \begin{equation}
181:    \sum_n\left(a_n^\dagger a_n+b_n^\dagger b_n\right)=2NS\,.
182:    \label{E:const}
183: \end{equation}
184: The constraint is enforced by introducing a Lagrange multiplier and
185: diagonalizing an effective Hamiltonian
186: \begin{equation}
187:    \widetilde{\cal H}
188:    ={\cal H}
189:    +2J\lambda\sum_n\left(a_n^\dagger a_n+b_n^\dagger b_n\right)\,.
190:    \label{E:effH}
191: \end{equation}
192: Via the Bogoliubov transformation
193: \begin{equation}
194:    \left.
195:    \begin{array}{lll}
196:       {\displaystyle\frac{1}{\sqrt{N}}}
197:       {\displaystyle\sum_n}
198:       {\rm e}^{ {\rm i}k(2n-1/2)}a_n
199:       &=&
200:       \alpha_k{\rm cosh}\theta_k-\beta_k^\dagger {\rm sinh}\theta_k\,,\\
201:       {\displaystyle\frac{1}{\sqrt{N}}}
202:       {\displaystyle\sum_n}
203:       {\rm e}^{-{\rm i}k(2n+1/2)}b_n
204:       &=&
205:       \beta_k {\rm cosh}\theta_k-\alpha_k^\dagger{\rm sinh}\theta_k\,,
206:    \end{array}
207:    \right.
208:    \label{E:FT}
209: \end{equation}
210: with ${\rm tanh}2\theta_k=S{\rm cos}k/(S+\lambda)$, we reach the spin-wave
211: Hamiltonian
212: \begin{eqnarray}
213:    &&
214:    {\cal H}=-2JS^2N-2J(S+\lambda)N+J\sum_k\omega_k
215:    \nonumber \\
216:    &&\qquad
217:    +J\sum_k\left(\omega_k^-\alpha_k^\dagger\alpha_k
218:                 +\omega_k^+\beta_k^\dagger \beta_k\right)\,,
219:    \label{E:SWH}
220: \end{eqnarray}
221: where
222: \begin{equation}
223:    \omega_k^\pm \mp g\mu_{\rm B}H/J
224:    =2\sqrt{(S+\lambda)^2-S^2{\rm cos}^2k}\equiv \omega_k\,.
225: \end{equation}
226: Minimization of the free energy gives the optimum distribution functions
227: as $\bar{n}_k^\pm=({\rm e}^{J\omega_k^\pm/k_{\rm B}T}-1)^{-1}$ and
228: $\lambda$ is then determined through
229: \begin{equation}
230:    \sum_k\left(\bar{n}_k^- +\bar{n}_k^+ +1\right){\rm cosh}2\theta_k
231:    =(2S+1)N\,.
232: \end{equation}
233: 
234:    Now we calculate the relaxation rate in terms of the modified spin
235: waves.
236: Considering the electronic-nuclear energy-conservation requirement,
237: the Raman scattering predominates in spin-gapped antiferromagnets.
238: The Raman relaxation rate is generally given by
239: \begin{eqnarray}
240:    &&
241:    \frac{1}{T_1}
242:     =\frac{4\pi(g\mu_{\rm B}\hbar\gamma_{\rm N})^2}
243:           {\hbar\sum_i{\rm e}^{-E_i/k_{\rm B}T}}
244:      \sum_{i,j}{\rm e}^{-E_i/k_{\rm B}T}
245:    \nonumber \\
246:    &&\qquad\times
247:      \big|
248:       \langle j|{\scriptstyle\sum_l}A_lS_l^z|i\rangle
249:      \big|^2
250:      \,\delta(E_j-E_i-\hbar\omega_{\rm N})\,,
251: \label{E:T1def}
252: \end{eqnarray}
253: where $A_l$ is the dipolar coupling constants between the nuclear and
254: electronic spins in the $l$th site, $\omega_{\rm N}\equiv\gamma_{\rm N}H$
255: is the Larmor frequency of the nuclei with $\gamma_{\rm N}$ being the
256: gyromagnetic ratio, and the summation $\sum_i$ is taken over all the
257: electronic eigenstates $|i\rangle$ with energy $E_i$.
258: By means of the modified spin waves, eq. (\ref{E:T1def}) is rewritten as
259: \begin{eqnarray}
260:    &&
261:    \frac{1}{T_1}
262:     =\frac{4\pi(g\mu_{\rm B}\hbar\gamma_{\rm N})^2}
263:           {\hbar N^2\sum_i{\rm e}^{-E_i/k_{\rm B}T}}
264:      \sum_{i,j}{\rm e}^{-E_i/k_{\rm B}T}
265:      \delta(E_j-E_i-\hbar\omega_{\rm N})
266:    \nonumber \\
267:    &&\qquad\times
268:      \Bigl|
269:       \langle j|
270:       {\scriptstyle\sum_{k,k'}}A_{k'-k}
271:       \bigl[
272:        \big(
273:         \alpha_{k'}^\dagger\alpha_k-\beta_{k'}^\dagger\beta_k
274:        \big){\rm cosh}\theta_{k'}{\rm cosh}\theta_k
275:    \nonumber \\
276:    &&\qquad\quad
277:       -\big(
278:         \alpha_{k'}\alpha_k^\dagger-\beta_{k'}\beta_k^\dagger
279:        \big){\rm sinh}\theta_{k'}{\rm sinh}\theta_k
280:       \bigr]
281:       |i\rangle
282:      \Bigr|^2\,,
283:    \label{E:T1SW}
284: \end{eqnarray}
285: where $A_q=\sum_l{\rm e}^{{\rm i}ql}A_l$.
286: The Fourier components of the hyperfine coupling constant exhibit little
287: momentum dependence when the nuclei take unsymmetrical positions in the
288: crystal, which is the case with the protons in NENP \cite{F11860}.
289: Hence we assume in the following that $A_q\simeq A_{q=0}\equiv A$.
290: Due to the significant difference between the electronic and nuclear
291: energy scales ($\hbar\omega_{\rm N}\alt 10^{-5}J$), eq. (\ref{E:T1SW})
292: ends in
293: \begin{equation}
294:    \frac{1}{T_1}
295:    =\frac{4(g\mu_{\rm B}\hbar\gamma_{\rm N}A)^2}{\pi\hbar}
296:     \int_{-\pi/2}^{\pi/2}
297:     \frac{\sum_{\sigma=\pm}\bar{n}_k^\sigma(\bar{n}_k^\sigma+1)}
298:          {\sqrt{v^2k^2+v\hbar\omega_{\rm N}}}
299:     {\rm d}k\,,
300:    \label{E:T1}
301: \end{equation}
302: where assuming moderate temperatures
303: $k_{\rm B}T\ll\omega_{k=\pi/2}^-=2J(S+\lambda)-g\mu_{\rm B}H$, we have
304: approximated the dispersion relations as
305: \begin{equation}
306:    J\omega_k^\pm\simeq{\mit\Delta}+vk^2\pm g\mu_{\rm B}H\,,
307: \end{equation}
308: with
309: \begin{equation}
310:    {\mit\Delta}=2J\sqrt{\lambda(2S+\lambda)}\,,\ \ 
311:    v=\frac{JS^2}{\sqrt{\lambda(2S+\lambda)}}\,.
312:    \label{E:Deltav}
313: \end{equation}
314: Equation (\ref{E:T1}) claims that an applied field produces two distinct
315: effects on $T_1^{-1}$, one of which originates from the Zeeman energy and
316: appears in $\bar{n}_k^\sigma$, while the other of which appears via the
317: nuclear spins giving the characteristic weight
318: $(v^2k^2+v\hbar\omega_{\rm N})^{-1/2}$ to the electronic transition rate
319: $\bar{n}_k^\sigma(\bar{n}_k^\sigma+1)$.
320: The field effect on the nuclear spins escapes observation in critical spin
321: chains with a linear dispersion at small momenta.
322: The prefactor $(v^2k^2+v\hbar\omega_{\rm N})^{-1/2}$
323: is the consequence of the nature of the delta function,
324: \begin{equation}
325:    \delta[f(x)]=\sum_i\frac{\delta(x-x_i)}{|f'(x_i)|}\,,
326: \end{equation}
327: where $x_i$ is a zero point of an arbitrary regular function $f(x)$, and
328: therefore generally arises from quadratic dispersion relations of the
329: relevant electronic excitations, which are the case with ferromagnets
330: \cite{F433,Y842,Y2324} as well as spin-gapped antiferromagnets.
331: 
332:    Let us fit eq. (\ref{E:T1}) for the proton spin-lattice relaxation-time
333: measurements on NENP \cite{F11860}.
334: Although ${\mit\Delta}$ and $v$, given in eq. (\ref{E:Deltav}), depend
335: on temperature in principle, here we fix
336: $(v^2k^2+v\hbar\omega_{\rm N})^{-1/2}$ to its zero-temperature value in
337: the integration (\ref{E:T1}), which is well justified for
338: $k_{\rm B}T\alt{\mit\Delta}-g\mu_{\rm B}H$ and allows us to inquire
339: further into eq. (\ref{E:T1}) analytically.
340: We compare the calculations with the observations in Fig. \ref{F:T1}.
341: Assuming that $g=2$ and $J/k_{\rm B}=55\,\mbox{K}$ \cite{F11860}, we have
342: set $A$ equal to $0.024\,\mbox{\AA}^{-3}$, which suggests the distance
343: between the interacting proton and electron spins being about
344: $3.5\,\mbox{\AA}$ and is consistent very well with the structural analysis
345: \cite{M1729}.
346: Under the present parametrization, the lowest excitation gap is given by
347: ${\mit\Delta}(T=0)/k_{\rm B}\equiv{\mit\Delta}_0/k_{\rm B}
348:  \simeq 4.0\,\mbox{K}$, which is somewhat smaller than that of NENP,
349: $12.8\,\mbox{K}$ \cite{F11860}.
350: However, the scaled function ${\mit\Delta}(T)/{\mit\Delta}_0$ well
351: reproduces the upward behavior of the Haldane-gap mode as a function of
352: temperature \cite{Y769}.
353: 
354:    With increasing field, the relaxation rate first decreases moderately
355: and then increases much more rapidly, at low temperatures in particular.
356: Although we cannot calculate beyond the critical field
357: $H_{\rm c}\equiv{\mit\Delta}_0/g\mu_{\rm B}\simeq 9.5\,\mbox{T}$ on the
358: present formulation, our theory well reproduces the observations
359: over a wide field range.
360: For $k_{\rm B}T\alt{\mit\Delta}-g\mu_{\rm B}H$,
361: the distribution functions may be approximated as
362: $\bar{n}_k^\pm(\bar{n}_k^\pm+1)\simeq{\rm e}^{-\omega_k^\pm/k_{\rm B}T}$
363: and therefore eq. (\ref{E:T1}) can further be calculated as
364: \begin{equation}
365:    \frac{1}{T_1}\simeq
366:     \frac{8(g\mu_{\rm B}\hbar\gamma_{\rm N}A)^2}{\pi\hbar v}
367:     {\rm e}^{-{\mit\Delta}_0/k_{\rm B}T}
368:     {\rm cosh}\Bigl(\frac{g\mu_{\rm B}H}{k_{\rm B}T}\Bigr)
369:     K_0\Bigl(\frac{\hbar\omega_{\rm N}}{2k_{\rm B}T}\Bigr)\,,
370:    \label{E:T1ap}
371: \end{equation}
372: where $K_0$ is the modified Bessel function of the second kind.
373: Provided
374: $\hbar\omega_{\rm N}\ll k_{\rm B}T\alt{\mit\Delta}-g\mu_{\rm B}H$,
375: we further reach
376: \begin{eqnarray}
377:    &&
378:    \frac{1}{T_1}\simeq
379:     \frac{8(g\mu_{\rm B}\hbar\gamma_{\rm N}A)^2}{\pi\hbar v}
380:     {\rm e}^{-{\mit\Delta}_0/k_{\rm B}T}
381:     {\rm cosh}\Bigl(\frac{g\mu_{\rm B}H}{k_{\rm B}T}\Bigr)
382:    \nonumber \\
383:    &&\qquad\times
384:    \left[
385:     0.80908-{\rm ln}\Bigl(\frac{\hbar\omega_{\rm N}}{k_{\rm B}T}\Bigr)
386:    \right]\,.
387: \end{eqnarray}
388: Thus, {\it as $H$ increases, $T_1^{-1}$ should initially decrease
389: logarithmically and then increase exponentially}.
390: Figure \ref{F:nk} suggests that the initial decrease of $T_1^{-1}$ turns
391: from $T_1^{-1}\propto-{\rm ln}H$ to $T_1^{-1}\propto H^{-1/2}$ with
392: decreasing temperature.
393: The momentum distribution functions $\bar{n}_k^\pm$ are peaked at $k=0$
394: and their peaks are sharpened as $T$ decreases.
395: $\bar{n}_k^\pm$ behaves as $\delta(k)$ in the low-temperature limit.
396: When we replace $\bar{n}_k^\pm$ by $\delta(k)$, eq. (\ref{E:T1}) gives a
397: $H^{-1/2}$-linear field dependence of $T_1^{-1}$.
398: 
399:    The temperature dependence is mainly described by the term
400: ${\rm e}^{-({\mit\Delta}_0-g\mu_{\rm B}H)/k_{\rm B}T}$ but further
401: decorated due to the temperature-dependent energy spectrum.
402: The inelastic-neutron-scattering peak position of the lowest-energy mode
403: exhibits an upward behavior with increasing temperature, for
404: $k_{\rm B}T\agt{\mit\Delta}_0/2$ in particular \cite{R3538,T4677,S3025},
405: where the slope of ${\rm ln}T_1^{-1}$ to $T^{-1}$ correspondingly
406: increases with increasing temperature.
407: 
408:    The nuclear magnetic relaxation in the Haldane-gap antiferromagnet NENP
409: has been interpreted in terms of a modified spin-wave theory.
410: The field dependence of $T_1^{-1}$ was analyzed in detail and {\it the
411: minimum of $T_1^{-1}$ as a function of $H$, pending for the past decade,
412: was solved}.
413: We consider that such a field dependence of $T_1^{-1}$ is qualitatively
414: common to spin-gapped antiferromagnets.
415: We encourage low-temperature $T_1$ measurements on related materials such
416: as the ferromagnetic-antiferromagnetic bond-alternating compound
417: (CH$_3$)$_2$CHNH$_3$CuCl$_3$ \cite{M564} and the two-leg ladder
418: antiferromagnet SrCu$_2$O$_3$ \cite{A3463}.
419: 
420:    The authors are grateful to Dr. N. Fujiwara and Prof. T. Goto for
421: fruitful discussion.
422: This work was supported by the Ministry of Education, Culture, Sports,
423: Science and Technology of Japan, and the Nissan Science Foundation.
424: 
425: \begin{references}
426: \bibitem{H464}
427:    F. D. M. Haldane:
428:       Phys. Lett. {\bf 93A} (1983) 464.
429: 
430: \bibitem{H1153}
431:    F. D. M. Haldane:
432:       Phys. Rev. Lett. {\bf 50} (1983) 1153.
433: 
434: \bibitem{W3844}
435:    S. R. White and D. A. Huse:
436:       Phys. Rev. B {\bf 48} (1993) 3844.
437: 
438: \bibitem{G3037}
439:    O. Golinelli, Th. Jolic{\oe}ur and R. Lacaze:
440:       Phys. Rev. B {\bf 50} (1994) 3037.
441: 
442: \bibitem{Y3348}
443:    S. Yamamoto:
444:       Phys. Rev. Lett. {\bf 75} (1995) 3348.
445: 
446: \bibitem{S493}
447:    U. Schollw\"ock and Th. Jolic{\oe}ur:
448:       Europhys. Lett. {\bf 30} (1995) 493.
449: 
450: \bibitem{W14529}
451:    X. Wang, S. Qin and L. Yu:
452:       Phys. Rev. B {\bf 60} (1999) 14529.
453: 
454: \bibitem{T047203}
455:    S. Todo and K. Kato:
456:       Phys. Rev. Lett. {\bf 87} (2001) 047203.
457: 
458: \bibitem{B371}
459:    W. J. L. Buyers, R. M. Morra, R. L. Armstrong, M. J. Hogan, P. Gerlach
460:    and K. Hirakawa:
461:       Phys. Rev. Lett. {\bf 56} (1986) 371.
462: 
463: \bibitem{R945}
464:    J. P. Renard, M. Verdaguer, L. P. Regnault, W. A. C. Erkelens,
465:    J. Rossat-Mignod and W. G. Stirling:
466:       Europhys. Lett. {\bf 3} (1987) 945.
467: 
468: \bibitem{A799}
469:    I. Affleck, T. Kennedy, E. H. Lieb and H. Tasaki:
470:       Phys. Rev. Lett. {\bf 59} (1987) 799.
471: 
472: \bibitem{A477}
473:    I. Affleck, T. Kennedy, E. H. Lieb and H. Tasaki:
474:       Commun. Math. Phys. {\bf 115} (1988) 477.
475: 
476: \bibitem{W2863}
477:    S. R. White:
478:       Phys. Rev. Lett. {\bf 69} (1992) 2863.
479: 
480: \bibitem{M913}
481:    S. Miyashita and S. Yamamoto:
482:       Phys. Rev. B {\bf 48} (1993) 913.
483: 
484: \bibitem{N4709}
485:    M. den Nijs and K. Rommelse:
486:       Phys. Rev. B {\bf 40} (1989) 4709.
487: 
488: \bibitem{Y9528}
489:    S. Yamamoto and S. Miyashita:
490:       Phys. Rev. B {\bf 48} (1993) 9528.
491: 
492: \bibitem{K627}
493:    S. Knabe:
494:       J. Stat. Phys. {\bf 52} (1988) 627.
495: 
496: \bibitem{F8983}
497:    G. F\'ath and J. S\'olyom:
498:       J. Phys.: Condens. Matter {\bf 5} (1993) 8983.
499: 
500: \bibitem{Y157}
501:    S. Yamamoto:
502:       Phys. Lett. A {\bf 225} (1997) 157.
503: 
504: \bibitem{Y1795}
505:    S. Yamamoto:
506:       Int. J. Mod. Phys. B {\bf 12} (1998) 1795.
507: 
508: \bibitem{A397}
509:    I. Affleck:
510:       Nucl. Phys. B {\bf 257} (1985) 397.
511: 
512: \bibitem{A409}
513:    I. Affleck:
514:       Nucl. Phys. B {\bf 265} (1986) 409.
515: 
516: \bibitem{O1984}
517:    M. Oshikawa, M. Yamanaka and I. Affleck:
518:       Phys. Rev. Lett. {\bf 78} (1997) 1984.
519: 
520: \bibitem{T103}
521:    K. Totsuka:
522:       Phys. Lett. A {\bf 228} (1997) 103.
523: 
524: \bibitem{T2313}
525:    M. Takahashi:
526:       Phys. Rev. Lett. {\bf 62} (1989) 2313.
527: 
528: \bibitem{Y545}
529:    S. Yamamoto and S. Miyashita:
530:       Phys. Lett. A {\bf 235} (1997) 545.
531: 
532: \bibitem{M3571}
533:    S. Ma, C. Broholm, D. H. Reich, B. J. Sternlieb and R. W. Erwin:
534:       Phys. Rev. Lett. {\bf 69} (1992) 3571.
535: 
536: \bibitem{Z017202}
537:    I. A. Zaliznyak, S.-H. Lee and S. V. Petrov:
538:       Phys. Rev. Lett. {\bf 87} (2001) 017202.
539: 
540: \bibitem{H2566}
541:    Z. Honda, H. Asakawa and K. Katsumata:
542:       Phys. Rev. Lett. {\bf 81} (1998) 2566.
543: 
544: \bibitem{D409}
545:    J. Darriet and L. P. Regnault:
546:       Solid State Commun. {\bf 86} (1993) 409.
547: 
548: \bibitem{M68}
549:    S. Maslov and A. Zheludev:
550:       Phys. Rev. B {\bf 57} (1998) 68.
551: 
552: \bibitem{T15189}
553:    Y. Takushima, A. Koga and N. Kawakami:
554:       Phys. Rev. B {\bf 61} (2000) 15189.
555: 
556: \bibitem{Z6437}
557:    A. Zheludev, J. M. Tranquada, T. Vogt and D. J. Buttrey:
558:       Phys. Rev. B {\bf 54} (1996) 6437.
559: 
560: \bibitem{Z7210}
561:    A. Zheludev, J. M. Tranquada, T. Vogt and D. J. Buttrey:
562:       Phys. Rev. B {\bf 54} (1996) 7210.
563: 
564: \bibitem{Y11516}
565:    T. Yokoo, A. Zheludev, M. Nakamura and J. Akimitsu:
566:       Phys. Rev. B {\bf 55} (1997) 11516.
567: 
568: \bibitem{T2173}
569:    M. Takigawa, T. Asano, Y. Ajiro, M. Mekata and Y. J. Uemura:
570:       Phys. Rev. Lett. {\bf 76} (1996) 2173.
571: 
572: \bibitem{B2215}
573:    F. Borsa and M. Mali:
574:       Phys. Rev. B {\bf 9} (1974) 2215.
575: 
576: \bibitem{S943}
577:    S. Sachdev and K. Damle:
578:       Phys. Rev. Lett. {\bf 78} (1997) 943.
579: 
580: \bibitem{F2810}
581:    S. Fujimoto:
582:       J. Phys. Soc. Jpn. {\bf 68} (1999) 2810.
583: 
584: \bibitem{S2712}
585:    S. Sachdev and K. Damle:
586:       J. Phys. Soc. Jpn. {\bf 69} (2000) 2712.
587: 
588: \bibitem{F2714}
589:    S. Fujimoto:
590:       J. Phys. Soc. Jpn. {\bf 69} (2000) 2714.
591: 
592: \bibitem{G647}
593:    P. Gaveau, J. P. Boucher, L. P. Regnault and J. P. Renard:
594:       Europhys. Lett. {\bf 12} (1990) 647.
595: 
596: \bibitem{F7837}
597:    N. Fujiwara, T. Goto, S. Maegawa and T. Kohmoto:
598:       Phys. Rev. B {\bf 45} (1992) 7837.
599: 
600: \bibitem{F11860}
601:    N. Fujiwara, T. Goto, S. Maegawa and T. Kohmoto:
602:       Phys. Rev. B {\bf 47} (1993) 11860.
603: 
604: \bibitem{S9188}
605:    J. Sagi and Affleck:
606:       Phys. Rev. B {\bf 53} (1996) 9188.
607: 
608: \bibitem{T2494}
609:    M. Takahashi:
610:       Phys. Rev. B {\bf 40} (1989) 2494.
611: 
612: \bibitem{H4769}
613:    J. E. Hirsch and S. Tang:
614:       Phys. Rev. B {\bf 40} (1989) 4769.
615: 
616: \bibitem{T5000}
617:    S. Tang, M. E. Lazzouni and J. E. Hirsch:
618:       Phys. Rev. B {\bf 40} (1989) 5000.
619: 
620: \bibitem{Y769}
621:    S. Yamamoto and H. Hori:
622:       J. Phys. Soc. Jpn. {\bf 72} (2003) 769.
623: 
624: \bibitem{F433}
625:    N. Fujiwara and M. Hagiwara:
626:       Solid State Commun. {\bf 113} (2000) 433.
627: 
628: \bibitem{Y842}
629:    S. Yamamoto:
630:       Phys. Rev. B {\bf 61} (2000) R842.
631: 
632: \bibitem{Y2324}
633:    S. Yamamoto:
634:       J. Phys. Soc. Jpn. {\bf 69} (2000) 2324.
635: 
636: \bibitem{M1729}
637:    A. Meyer, A. Gleizes, J. J. Girerd, M. Verdaguer and O. Kahn:
638:       Inorg. Chem. {\bf 21} (1982) 1729.
639: 
640: \bibitem{R3538}
641:    J. P. Renard, M. Verdaguer, L. P. Regnault, W. A. C. Erkelens,
642:    J. Rossat-Mignod, J. Ribas, W. G. Stirling and C. Vettier:
643:       J. Appl. Phys. {\bf 63} (1988) 3538.
644: 
645: \bibitem{T4677}
646:    Z. Tun, W. J. L. Buyers, R. L. Armstrong, K. Kirakawa and
647:    B. Briat:
648:       Phys. Rev. B {\bf 42} (1990) 4677.
649: 
650: \bibitem{S3025}
651:    T. Sakaguchi, K. Kakurai, T. Yokoo and J. Akimitsu:
652:       J. Phys. Soc. Jpn. {\bf 65} (1996) 3025.
653: 
654: \bibitem{M564}
655:    H. Manaka, I. Yamada and K. Yamagushi:
656:       J. Phys. Soc. Jpn. {\bf 66} (1997) 564.
657: 
658: \bibitem{A3463}
659:    M. Azuma, Z. Hiroi, M. Takano, K. Ishida and Y. Kitaoka:
660:       Phys. Rev. Lett. {\bf 73} (1994) 3463.
661: 
662: \end{references}
663: 
664: \begin{figure}
665: \centerline
666: {\mbox{\psfig{figure=Fig1.eps,width=80mm,angle=0}}}
667: \vspace*{1mm}
668: \caption{Dependences of the nuclear spin-lattice relaxation rate on a
669:          field parallel to the chain (a) and temperature (b) in NENP
670:          (symbols) [44] are compared with the modified spin-wave
671:          calculations (lines), where the field and temperature are scaled
672:          by the lowest excitation gap ${\mit\Delta}_0$.}
673:          %[44]=\cite{F11860}
674: \label{F:T1}
675: \end{figure}
676: 
677: \begin{figure}
678: \centerline
679: {\mbox{\psfig{figure=Fig2.eps,width=80mm,angle=0}}}
680: \vspace*{1mm}
681: \caption{The momentum distribution functions $\bar{n}_k^+$ (dotted lines)
682:          and $\bar{n}_k^-$ (solid lines).}
683: \label{F:nk}
684: \end{figure}
685: 
686: \widetext
687: \end{document}
688: