1: \documentclass[

2: %preprint%

3:  ,twocolumn%

4:  ,secnumarabic%

5: %,tightenlines%

6: ,amssymb, amsmath,nobibnotes, aps, prb,showpacs]{revtex4}

7: %\usepackage{acrofont}%NOTE: Comment out this line for the release version!

8: %\usepackage{docs}%

9: \usepackage{bm}%

10: \usepackage{amsfonts,amsmath,amsopn}

11: \usepackage{isolatin1}

12: \usepackage{graphicx}

13: \usepackage[german]{varioref}

14: %\usepackage[pdfpagetransition={Box /M /I},pdftex,backref]{hyperref} 

15: \usepackage[dvips]{epsfig}

16: \usepackage{graphics}

17: \usepackage[dvips]{color}

18: \usepackage{color}

19: \usepackage{colordvi}

20: \usepackage{dsfont}

21: \usepackage{psfrag}

22: %\usepackage[colorlinks=true,linkcolor=blue]{hyperref}%

23: %\nofiles

24: \expandafter\ifx\csname package@font\endcsname\relax\else

25:  \expandafter\expandafter

26:  \expandafter\usepackage

27:  \expandafter\expandafter

28:  \expandafter{\csname package@font\endcsname}%

29: \fi

30: 

31: \begin{document}

32: \newcommand{\be}{\begin{equation}}

33: \newcommand{\ee}{\end{equation}}

34: \newcommand{\bea}{\begin{eqnarray}}

35: \newcommand{\eea}{\end{eqnarray}}

36: \def\p#1#2{|#1\rangle \langle #2|}

37: \def\ket#1{|#1\rangle}

38: \def\bra#1{\langle #1|}

39: \def\refeq#1{(\ref{#1})}

40: \def\tb#1{{\overline{{\underline{ #1}}}}}

41: \def\im{\mbox{Im}}

42: \def\re{\mbox{Re}}

43: \def\nn{\nonumber}

44: \def\t{\mbox{tr}}

45: \def\sgn{\mbox{sgn}}

46: \def\Li{\mbox{Li}}

47: \def\P{\mbox{P}}

48: \def\d{\mbox d}

49: %

50: \def\i{\int_{-\infty}^{\infty}}

51: \def\ip{\int_{0}^{\infty}}

52: \def\mi{\int_{-\infty}^{0}}

53: %

54: \def\A{\mathfrak A}

55: \def\AA{{\overline{{\mathfrak{A}}}}}

56: \def\a{\mathfrak a}

57: \def\aa{{\overline{{\mathfrak{a}}}}}

58: \def\B{\mathfrak B}

59: \def\BB{{\overline{{\mathfrak{B}}}}}

60: \def\b{\mathfrak b}

61: \def\bb{{\overline{{\mathfrak{b}}}}}

62: \def\R{\mathcal R}

63: \def\dm{\mathfrak d}

64: \def\dd{{\overline{{\mathfrak{d}}}}}

65: \def\D{\mathfrak D}

66: \def\DD{{\overline{{\mathfrak{D}}}}}

67: \def\c{\mathfrak c}

68: \def\cc{{\overline{{\mathfrak{c}}}}}

69: \def\C{\mathfrak C}

70: \def\CC{{\overline{{\mathfrak{C}}}}}

71: %

72: \def\O{\mathcal O}

73: \def\F{\mathcal F_k}

74: \def\N{\mathcal N}

75: \def\I{\mathcal I}

76: \def\S{\mathcal S}

77: %

78: \def\G{\Gamma}

79: \def\L{\Lambda}

80: \def\la{\lambda}

81: \def\g{\gamma}

82: \def\al{\alpha}

83: \def\s{\sigma}

84: \def\e{\epsilon}

85: \def\te{\text{e}}

86: \def\ti{\text{i}}

87: \def\max{\text{max}}

88: \def\str{\text{str}}

89: \def\tr{\text{tr}}

90: \def\Tr{\text{Tr}}

91: \def\tC{\text C}

92: %

93: \def\Fo{\mathcal{F}_{1,k}}

94: \def\Ft{\mathcal{F}_{2,k}}

95: \def\vs{\varsigma}

96: \def\l{\left}

97: \def\r{\right}

98: \def\up{\uparrow}

99: \def\down{\downarrow}

100: \def\u{\underline}

101: \def\ov{\overline}

102: %

103: \title{$S$=1/2 antiferromagnetic Heisenberg chain with staggered fields:

104:   Copper pyrimidine and copper benzoate using the density matrix

105:   renormalization group for transfer matrices}

106: 

107: \author{S. Glocke$^1$, A. Kl\"umper$^1$, H. Rakoto$^2$, J.M. Broto$^2$, A.U.B. Wolter$^3$, S. S\"ullow$^3$}%

108: \affiliation{$^1$Bergische Universit\"at Wuppertal, Fachbereich Physik, 42097

109:   Wuppertal, Germany}

110: \affiliation{$^2$Laboratoire National des Champs Magn\'etiques Puls\'es, 31432 Toulouse, France}

111: \affiliation{$^3$Institut f\"ur Physik der Kondensierten Materie, TU Braunschweig, 38106 Braunschweig, Germany}

112: \date{\today}%

113: 

114: \begin{abstract}

115: We consider the spin-1/2 antiferromagnetic Heisenberg

116: chain in a staggered magnetic field describing materials such as copper benzoate and copper

117: pyrimidine dinitrate. Using the density-matrix renormalization group for

118: transfer matrices (TMRG) we calculate the magnetization of these

119: materials at finite temperature and arbitrary magnetic field. These results

120: are in excellent agreement with experimental data, allowing for a

121: determination of the inhomogeneity parameter c of copper benzoate ($c = 0.043$) and copper

122: pyrimidine dinitrate ($c= 0.11$). The TMRG approach can be applied to rather

123: low temperatures yielding singular field and temperature dependences of susceptibilities.

124: \end{abstract}

125: \pacs{75.10.Jm, 75.50.Ee, 75.30.Gw, 75.50.Xx}

126: \maketitle

127: 

128: One-dimensional quantum magnets have been of theoretical and

129: experimental interest in recent years because of the rich variety of different

130: magnetic ground states, such as quantum critical behavior or gaps in the spin

131: excitation spectra \cite{Hal83,Den97,Sto03}. 

132: The ideal spin-1/2 antiferromagnetic Heisenberg chain ($S$=1/2 AFHC) with

133: uniform nearest neighbor exchange coupling is of particular interest, since it

134: is exactly solvable using the Bethe ansatz (BA) \cite{Bethe31,FGMSK96,KJ00}. Its

135: ground state is a spin singlet with gapless excitations and quasi-long ranged

136: ground state correlations, hence even small perturbations can change the

137: physical properties fundamentally.

138: 

139: In experimental realizations of spin chains like copper

140: pyrimidine dinitrate

141: $\left[\mbox{PM}\,\mbox{Cu}(\mbox{NO}_3)_2(\mbox{H}_2\mbox{O})_2\right]_n$

142: (CuPM) and copper benzoate $\mbox{Cu}(\mbox{C}_6\mbox{H}_5\mbox{COO})_2\cdot3\mbox{H}_2\mbox{O}$

143: additional terms in the Hamiltonian result from the lack of inversion

144: symmetry. As a consequence of the residual spin-orbit coupling the

145: Dzyaloshinskii-Moriya (DM) interaction and an alternating $g$ tensor have to

146: be taken into account \cite{OA97,AO99}. This gives rise to an effective

147: staggered field $h_s$ perpendicular to the applied magnetic field $H$. The

148: Hamiltonian is written as \cite{OA97,AO99}%\nocite{AO00} 

149: \be

150: \label{Hamiltonian}

151: \hat H= J \sum_i\left[\mathbf{S}_i\mathbf{S}_{i+1} - h_u S_i^z - (-1)^i h_s

152:   S_i^x\right]

153: \ee

154: with the coupling constant $J$, the effective uniform field $h_u = g\mu_B H/J$ 

155: and the induced effective staggered field $h_s = c\, h_u$. For a given field

156: axis the effective $g$ and the inhomogeneity parameter $c$ are determined from

157: the DM interaction $\vec D$ and the alternating $g$ tensor

158: $\overleftrightarrow{g} = \overleftrightarrow{g_u} \pm

159: \overleftrightarrow{g_s}$ in the following way \cite{FAGIMMNS00}

160: %\begin{subequations}

161: %\label{eqn_g_c}

162: %\be

163: %\label{eqn_g}

164: $g = \left|\overleftrightarrow{g_u} \vec H\right|/\left|\vec H\right|$ and % \; ,

165: %\ee

166: \be

167: \label{eqn_g_c}

168: c = \frac{1}{g}\left|\frac{1}{2J} \vec D\times \overleftrightarrow{g_u}\vec H + 

169:   \overleftrightarrow{g_s}\vec H \right|/\left|\vec H\right|\; .

170: \ee

171: %\end{subequations}

172: 

173: In essence, for materials like CuPM and Cu-benzoate, the presence of DM 

174: interaction and staggered $g$ tensor induces new symmetry breakings with 

175: respect to the magnetic properties. In terms of principal axes, 

176: conventionally it is distinguished between the crystallographic unit 

177: cell with the axes $abc$, the coordinate frame $a^\prime b c^\prime$, where

178: the uniform part of the $g$ tensor is diagonal, and the system of the principal

179: magnetic axes 

180: $a^{\prime\prime} b c^{\prime\prime}$ \cite{AO99,FAGIMMNS00}. For both materials the $g$ tensor and the DM vector have 

181: been derived previously, as have the different coordinate frames \cite{AO99,FAGIMMNS00}. 

182: Hence, these quantum spin systems are perfectly suited as model 

183: compounds for a quantitatively exact comparison between theory and 

184: experiment. Correspondingly, in recent years various experimental and 

185: theoretical studies on static or dynamic properties of staggered S=1/2 

186: AFHC have been carried out \cite{ZWXSY03,Kenzel04,Lou05}, even though data analysis 

187: was essentially limited to zero temperature. Therefore, in this combined 

188: theoretical and experimental study we provide a detailed investigation 

189: of the effect of temperature on thermodynamic properties of staggered 

190: S=1/2 AFHCs. For the first time, we derive the induced staggered 

191: magnetization for the staggered S=1/2 AFHC from the low temperature 

192: regime $k_B T \ll J$ up to the paramagnetic range $k_B T > J$ in fields 

193: up to saturation.

194: Since the effective $g$ factor and the $c$ factor depend on the 

195: orientation of the magnetic field $H$ with respect to the crystal axis 

196: \refeq{eqn_g_c}, we have to calculate these physical parameters for each 

197: experimental configuration separately. For this reason, here we 

198: calculate the magnetization for different crystal alignments by use of 

199: the transfer matrix renormalization group (TMRG) method and compare these data with experimental data of CuPM

200: \cite{Wolter03} and Cu-benzoate. The good agreement of numerical and

201: experimental data allows for the determination of $c$ and $g$ factors. 

202: 

203: %%%%%%%%%%%%%%%Method%%%%%%%%%%%%%%%%%%%%%%

204: To study thermodynamical properties at finite temperature the TMRG method

205: provides a powerful numerical tool, because the thermodynamic limit is

206: performed exactly and in contrast to quantum Monte Carlo techniques there is

207: no ``minus sign'' problem. %Therefore finite-size effects, affecting the exact

208: %diagonalization (ED) method, are not important here.

209: The idea of the TMRG method is

210: to express the partition function $Z$ of a one-dimensional quantum system by

211: that of an equivalent two-dimensional classical model by a Trotter-Suzuki

212: mapping \cite{Trotter59,Suzuki85}. For the classical 

213: model a suitable quantum transfer-matrix (QTM) can be defined which allows for the

214: calculation of all thermodynamical quantities in the thermodynamic limit by

215: considering solely the largest eigenvalue of this QTM. Here we 

216: use the Trotter-Suzuki mapping 

217: \cite{JK02} yielding

218: %\be

219: $\label{eqn_part_func}

220: Z = \lim_{M\to\infty}\Tr\left\{\left[T_1(\epsilon)

221:   T_2(\epsilon)\right]^{M/2}\right\}\; ,$

222: %\ee 

223: where $T_{1,2}(\epsilon) = T_{R,L}\exp\left[-\epsilon H + \O

224:   (\epsilon^2)\right]$, $\epsilon = \beta/M$ with $\beta$ being the inverse temperature and

225: $M$ a large integer number. $T_{R,L}$ denote the right- and leftshift

226: operators, respectively. %This mapping allows for the

227: %definition of a QTM with repeat length 1 and

228: The bulk free energy %in the thermodynamic limit 

229: is given by the largest eigenvalue~$\Lambda_0$, 

230: %\be

231: $f = -T \ln\Lambda_0 \; .$

232: %\ee

233: %where $\Lambda_0$ is unique and a real, positive number for all

234: %temperatures. Since a vanishing gap between leading and next-leading

235: %eigenvalue indicates a phase transition, such a degeneracy

236: %is not possible for a one-dimensional quantum system at finite

237: %temperature. 

238: Expectation values of local operators, like the homogeneous and

239: the staggered magnetization, can be expressed in terms of the left and right

240: eigenvectors belonging to the largest eigenvalue.

241: For achieving lower temperatures the length of the transfer matrix is

242: increased in imaginary time direction by an application of the infinite density

243: matrix renormalization group (DMRG) algorithm \cite{DMRG_book}.

244: In all following calculations we have retained

245: between 64 and 100 states and used $\epsilon = 0.05$. There are no finite-size

246: effects as the system size is strictly infinite, however numerical errors due

247: to the discretization of the imaginary time axis and the DMRG truncation are

248: generally within the line width used in our plots as has been checked by

249: comparison with the exact BA data (see Fig. \ref{fig_magn_a16k}). For details

250: of the algorithm the reader is referred to

251: Refs. \cite{JK02,DMRG_book,BXG96,Wang_Xiang97,Shibata97}.

252: 

253: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

254: In the following we compare our numerical results with high-field 

255: magnetization data measured at the Laboratoire National des Champs 

256: Magn\'etiques Puls\'es in Toulouse in pulsed magnetic fields up to 

257: $\mu_0 H = 53$ T (CuPM) and 38 T (Cu-benzoate), respectively (experimental

258: details in Ref. \cite{Wolter03}), and with results obtained by means of

259: $^{13}$C NMR \cite{Wolter05}. 

260: In the presence of a staggered magnetic field $h_s = c\;h_u$ induced by a

261: uniform field $h_u$, the magnetization $m$ is given by the superposition of the uniform

262: $m_u$ and the staggered $m_s$ magnetization components \cite{OA97,AO99}:

263: \be

264: \label{eqn_m_phys}

265:   m = m_u + c\, m_s \; ,

266: \ee

267: where $c$ is the above inhomogeneity parameter. The magnetization $m$ can be

268: measured experimentally. On theory side we can

269: determine the magnetization $m$ by using \refeq{eqn_m_phys} and calculating

270: the uniform $m_u$ and the staggered magnetization $m_s$ separately by TMRG

271: methods.

272: 

273: %%%%%%%%%%%%%%CuPM%%%%%%%%%%%%%%%%

274: Recently, CuPM has been identified as a $S$=1/2 AFHC, with a 

275: magnetic exchange parameter $J/k_B = (36 \pm 0.5)$K \cite{FAGIMMNS00} (in the 

276: following we use $J/k_B = 36.5$K). In the

277: uniform $S$=1/2 AFHC model the (magnetic) saturation field is calculated according to the

278: formula \cite{AO99}

279: \be

280: \label{eqn_hc}

281: H_c = 4 J S / g \mu_B \;.

282: \ee

283: The $g$ tensor and the DM vector of CuPM have been derived from ESR

284: measurements and single-crystal susceptibility studies \cite{FAGIMMNS00}. In

285: the coordinate frame $a^\prime b c^\prime$ the $g$ tensor takes the form 

286: %\bea

287: %\overleftrightarrow{g} &=& 

288: %         \left(\begin{array}{*{3}{c}}

289: %            2.073&0&0\\

290: %            0&2.149&\pm 0.127\\

291: %            0&\pm 0.127&2.287

292: %            \end{array}

293: %            \right)\\

294: %	    &=& \left(\begin{array}{*{3}{c}}

295: %           g_{xx}&0&0\\

296: %            0&g_{yy}&\pm g_s\\

297: %            0&\pm g_s&g_{zz}

298: %            \end{array}

299: %            \right)

300: %	    =\overleftrightarrow{g_u} \pm \overleftrightarrow{g_s} 

301: %\eea

302: \be

303: \overleftrightarrow{g} = 

304:          \left(\begin{array}{*{3}{c}}

305:             2.073&0&0\\

306:             0&2.149&\pm 0.127\\

307:             0&\pm 0.127&2.287

308:             \end{array}

309:             \right)

310: 	    =\overleftrightarrow{g_u} \pm \overleftrightarrow{g_s} 

311: \ee

312: and the DM interaction vector reads 

313: \be

314: \vec D = 0.139 J \left(-0.4115,0,0.9114\right)\; .

315: \ee

316: 

317: If the magnetic field $H$ is directed along the $a^{\prime\prime}$ direction of the CuPM crystal, it will

318: behave like an ideal $S$=1/2 AFHC ($c = 0$). In this case, from \refeq{eqn_g_c} we obtain

319: the effective $g$ factor as $g = 2.14$, with saturation field $\mu_0 H_c =

320: 50.7$ T \refeq{eqn_hc}.

321: \begin{figure}

322: \includegraphics*[width=7.5cm, height=5.5cm]{1_magn_a16k_total.eps}

323: \caption{Plot of theoretical (solid line) and experimental (open circles)

324:   magnetization data for CuPM at $T = 1.6$ K along the $a^{\prime\prime}$ direction. Dashed

325:   line shows TMRG data with a misalignment of $5^\circ$ at $T = 1.6$ K. Note that deviations

326:   are strongest in the vicinity of the saturation field. Lower plot:

327:   difference between TMRG and BA data.} 

328: \label{fig_magn_a16k}

329: \end{figure}

330: In Fig. \ref{fig_magn_a16k} we present the magnetization curve $m$ of CuPM as

331: function of the magnetic field at T = 1.6 K along the $a^{\prime\prime}$ axis,

332: as well as the

333: difference between TMRG and the exact BA results \cite{AK98}. % are shown%

334: Since this difference is always smaller than

335: $2\cdot 10^{-3}$ the TMRG results describe the exact BA results very well. 

336: Comparing the experimental data with the TMRG results we find good agreement

337: with the uniform $S$=1/2 AFHC for magnetic fields up to $\mu_0H < 45$ T. For

338: higher fields the Dzyaloshinskii-Moriya interaction and the alternating $g$

339: tensor may take effect on small crystal misalignments, so that CuPM does not

340: behave like a uniform $S$=1/2 AFHC. Assuming a misalignment of $5^\circ$ of

341: the magnetic field with respect to the $a^{\prime\prime}$ axis \cite{Wolter03} the physical

342: parameters have to be changed to $J/k_B = 36.5$~K, $g = 2.13$ and

343: $c = 0.01$ \refeq{eqn_g_c}.

344: With these slight modifications the calculated

345: magnetization curve of CuPM describes the experimental data well, see

346: Fig. \ref{fig_magn_a16k}.

347: 

348: %%%%%%%%%%%%%%CuPM c'' axis%%%%%%%%%%%%%%%%%

349: \begin{figure}

350: \includegraphics*[width=7.5cm, height=5.5cm]{2_suscep_cupm.eps}

351: \caption{Susceptibility $\chi$ of CuPM as a function of $H$ at \mbox{$T = 1.6$ K}

352:   calculated by TMRG. Solid (dashed) line: susceptibility with magnetic field

353:   along $c^{\prime\prime}$ ($a^{\prime\prime}$) axis. Inset: $\chi$ of CuPM as a function of $T$ at

354:   $h = 0$ calculated by TMRG.} 

355: \label{fig_suszep_CuPM}

356: \end{figure}

357: Along the $c^{\prime\prime}$ axis the effect due to the

358: induced staggered field is largest \cite{FAGIMMNS00}. According to \cite{AO99}

359: a gap is induced $\Delta \propto h^{2/3}$ with multiplicative logarithmic

360: corrections. At low $T$ the magnetization receives strong

361: contributions from the staggered component with singular behavior. For $T=0$ the

362: dependence $m_s \propto h^{1/3}$ was found and for $h=0$ at low $T$ the

363: susceptibility $\chi(T) \propto 1/T$ with multiplicative logarithmic

364: corrections. We stress that this behavior is the result of strong correlations

365: despite vague similarities with paramagnetic impurities. Susceptibility data

366: are shown in Fig. \ref{fig_suszep_CuPM}. Note however, the direct application

367: of DMRG \cite{ZWXSY03} to the Heisenberg chain with full DM-terms does not

368: give evidences of logarithmic corrections.

369: \begin{figure}

370: \includegraphics*[width=7.5cm, height=5.5cm]{3_magn_011.eps}

371: \caption{Theoretical and experimental magnetization curves for CuPM at $T = 1.6$ K

372:   along the $c^{\prime\prime}$ direction. Solid line: magnetization $m$ calculated for $c =

373:   0.11$, dashed line: uniform magnetization $m_u$, dotted line: staggered

374:   magnetization $m_s$, open circles: experimental data. Inset (a) depicts the

375:   magnetization for small fields calculated by TMRG and ED

376:   (dashed-dotted line) in comparison to the experimental data, inset (b)

377:   displays the temperature dependent staggered magnetization of CuPM for a

378:   magnetic field $\mu_0H = 9.3$ T applied along the chain axis (solid line: TMRG

379:   results; open circles: experimental data).} 

380: \label{fig_magn_011}

381: \end{figure}

382: Along this axis we obtain the parameters $c=0.11$ and

383: $g=2.19$ with saturation field $\mu_0 H_c = 49.6$~T according to

384: \refeq{eqn_hc}. \mbox{Figure \ref{fig_magn_011}} shows magnetization curves for CuPM with $H$ along the

385: $c^{\prime\prime}$ axis at temperature $T = 1.6$~K. 

386: For all magnetic fields the magnetization calculated by TMRG is in good agreement with

387: the experimental data. In particular at small 

388: fields, in comparison with the results of exact diagonalization (ED) \cite{Wolter03} the 

389: TMRG describes the experiment much more accurately, because it has no 

390: finite-size effects by construction (inset (a) of Fig. \ref{fig_magn_011}). We stress 

391: that the ED results were calculated for T = 0, 

392: whereas the TMRG results are calculated at the same finite temperatures 

393: chosen in the experiment. 

394: 

395: For magnetic field $H$ parallel to the chain axis \cite{FAGIMMNS00} of a CuPM crystal we find

396: the parameters $c=0.083$ and \mbox{$g = 2.117$}. In inset (b) of

397: Fig. \ref{fig_magn_011} we present the excellent agreement of the calculated

398: and the measured \cite{Wolter05} staggered magnetization curves for magnetic

399: field $\mu_0 H = 9.3$~T. 

400: Since these results are in good agreement with the experimental data, we

401: verified the physical parameters for CuPM, especially $c = 0.11$ along the

402: $c^{\prime\prime}$ axis with accurancy $\pm 0.01$. %Note that we get the same

403: 				%value $c=0.11$ obtained by ED \cite{Wolter03} and that our value differs from the result $c=0.08$ determined in \cite{ZKKF04} from ESR.

404: Note that this is consistent with results based 

405: on ED \cite{Wolter03}, but rules out c=0.08 obtained in [23] from ESR.

406: 

407: %%%%%%%%%%%%%%CuBen%%%%%%%%%%%%%%%%%

408: Cu-benzoate is another example for a staggered $S$=1/2 AFHC, with a behavior very similar to CuPM. Only, for Cu-benzoate the 

409: coupling constant is smaller by a factor $\sim 2$ than for CuPM, that is 

410: $J/k_B = \left(19 \pm 0.5\right)$ K \cite{OA97,AO99}. In consequence, for this

411: material, in a temperature dependent high magnetic field study, the regimes

412: from the fully staggered one $k_B T \ll J$ up to the paramagnetic range $k_B T

413: > J$ in fields up to saturation are accessible and can be compared to the

414: results from TMRG calculations. From ESR measurements the $g$ tensor in the

415: $a^{\prime\prime} b c^{\prime\prime}$ coordinate system takes the form:

416: \be

417: \overleftrightarrow{g} = 

418:        \left(\begin{array}{*{3}{c}}

419:          2.115&\pm0.0190&0.0906\\

420:          \pm 0.0190&2.059&\pm 0.0495\\

421:          0.0906&\pm 0.0495&2.316

422:        \end{array}

423:        \right)  

424: \ee

425: Moreover the DM interaction has been determined \cite{AO99}

426: \be

427:  \vec D = J \left(0.13,0,0.02\right)\; .

428: \ee

429: Cu-benzoate will behave like a homogeneous $S$=1/2 AFHC ($c=0$) with a

430: coupling constant $J/k_B = 19$~K, if the magnetic

431: field is along the $a^{\prime\prime}$ axis. 

432: \begin{figure}[!ht]

433: \includegraphics*[width=7.5cm, height=5.5cm]{4_magn_cu_ben_a.eps} 

434: \caption{Homogeneous magnetization of Cu-benzoate along $a^{\prime\prime}$ axis at

435:   different temperatures (1.7K, 4.2K, 12K and 25K). Solid (dashed) lines show

436:   TMRG (experimental) data. Inset: Magnetization of Cu-benzoate along $c^{\prime\prime}$ axis at different

437:   temperatures (1.7 K, 4.2 K, 12 K and 25 K). %Solid (dashed) lines show TMRG (experimental) data. Moreover

438:   The dotted lines show the

439:   magnetization at $T = 12.8\mbox{K}$ and $T = 22.5\mbox{K}$ calculated by

440:   TMRG.}

441: \label{fig_cuben_a}

442: \end{figure}

443: The saturation field is calculated to

444: $\mu_0 H_c = 26.5$~T by using $g = 2.13$ corresponding to this

445: axis. In Fig. \ref{fig_cuben_a} we show the magnetization \refeq{eqn_m_phys}

446: as a function of the magnetic field at different temperatures. Overall, the

447: TMRG results agree very well with the experimental data obtained in our

448: high-field magnetization study, especially with respect to the saturation  

449: magnetization. 

450: \begin{figure}[b]

451: \includegraphics*[width=7.5cm, height=5.5cm]{5_magn_c_0_048_corrected.eps} 

452: \caption{Staggered magnetization of Cu-benzoate along $c^{\prime\prime}$ axis at different

453:   temperatures (1.7 K, 4.2 K, 12 K and 25 K).} 

454: \label{fig_cuben_c}

455: \end{figure}

456: The field induced gap of Cu-benzoate is largest for magnetic field along

457: the $c^{\prime\prime}$ axis. In this case we

458: use $J/k_B = 18.9$~K and calculate the effective $g$ factor to

459: $g = 2.32$, with a corresponding saturation field

460: $\mu_0 H_c= 24.2$~T. By using \refeq{eqn_g_c} the $c$ factor is determined to $c

461: = 0.043$. The inset of Fig. \ref{fig_cuben_a} shows the magnetization

462: $m$ \refeq{eqn_m_phys}

463: of Cu-benzoate along the $c^{\prime\prime}$ axis at

464: different temperatures. Altogether, within experimental 

465: resolution we find that the magnetization of Cu-benzoate is nicely 

466: described by TMRG results using $c = 0.043\pm 0.01$ (covering the value

467: $c=0.034$ resulting from the DM vector used in Ref. \cite{ZWXSY03}). The somewhat larger 

468: deviations between experiment and theory at higher temperatures ($T = 25$~K) can be ascribed to a larger experimental uncertainty ($\pm 3$~K) in this temperature range.

469: 

470: In Fig. \ref{fig_cuben_c} it can be 

471: clearly seen that in the low temperature staggered regime ($1.7$~K $= 0.09 

472: k_B T/J$) there is a prominent staggered component, with a maximum value 

473: of about 0.6 $\mu_B$/Cu atom at 20 T. This staggered magnetization is 

474: gradually wiped out as temperature is increased up to 25 K $= 1.3 k_B 

475: T/J$, {\it viz.}, the conventional paramagnetic regime.

476: 

477: In summary, a comparative analysis of numerical TMRG and 

478: experimental data for Cu-benzoate and CuPM at low temperatures was 

479: presented. In our comparison we found very good agreement between our 

480: numerical and experimental data. The field dependence of the magnetization was

481: found to show interesting characteristics at high fields corresponding to

482: saturation and for low field due to quantum criticality of the $S$=1/2 AFHC.

483: 

484: We thank H.-H Klauss for providing the NMR data of inset (b) of

485: Fig. \ref{fig_magn_011}. This work has been supported by the DFG

486: under contracts no. SU229/6-1 and KL645/4-2.

487: 

488: \begin{thebibliography}{22}

489: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

490: \expandafter\ifx\csname bibnamefont\endcsname\relax

491:   \def\bibnamefont#1{#1}\fi

492: \expandafter\ifx\csname bibfnamefont\endcsname\relax

493:   \def\bibfnamefont#1{#1}\fi

494: \expandafter\ifx\csname citenamefont\endcsname\relax

495:   \def\citenamefont#1{#1}\fi

496: \expandafter\ifx\csname url\endcsname\relax

497:   \def\url#1{\texttt{#1}}\fi

498: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi

499: \providecommand{\bibinfo}[2]{#2}

500: \providecommand{\eprint}[2][]{\url{#2}}

501: 

502: \bibitem[{\citenamefont{Haldane}(1983)}]{Hal83}

503: \bibinfo{author}{\bibfnamefont{F.~D.~M.}~\bibnamefont{Haldane}},

504:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{50}},

505:   \bibinfo{pages}{1153} (\bibinfo{year}{1983}).

506: 

507: \bibitem[{\citenamefont{Dender et~al.}(1997)\citenamefont{Dender, Hammar,

508:   Reich, Broholm, and Aeppli}}]{Den97}

509: \bibinfo{author}{\bibfnamefont{D.~C.}~\bibnamefont{Dender}},

510:   \bibinfo{author}{\bibfnamefont{P.~R.}~\bibnamefont{Hammar}},

511:   \bibinfo{author}{\bibfnamefont{D.~H.} \bibnamefont{Reich}},

512:   \bibinfo{author}{\bibfnamefont{C.} \bibnamefont{Broholm}}, \bibnamefont{and}

513:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Aeppli}},

514:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{79}},

515:   \bibinfo{pages}{1750} (\bibinfo{year}{1997}).

516: 

517: \bibitem[{\citenamefont{Stone et~al.}(2003)\citenamefont{Stone, Reich,

518:   Broholm, Lefmann, Rischel, Landee, and Turnbull}}]{Sto03}

519:  \bibinfo{author}{\bibfnamefont{M.~B.}~\bibnamefont{Stone}},

520:   \bibinfo{author}{\bibfnamefont{D.~H.} \bibnamefont{Reich}},

521:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Broholm}},

522:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Lefmann}},

523:   \bibinfo{author}{\bibfnamefont{C.} \bibnamefont{Rischel}},

524:   \bibinfo{author}{\bibfnamefont{C.~P.} \bibnamefont{Landee}}, \bibnamefont{and}

525:   \bibinfo{author}{\bibfnamefont{M.~M.}~\bibnamefont{Turnbull}},

526:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{91}},

527:   \bibinfo{pages}{037205} (\bibinfo{year}{2003}).

528: 

529: \bibitem[{\citenamefont{Bethe}(1931)}]{Bethe31}

530: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Bethe}}, \bibinfo{journal}{Z.

531:   Phys.} \textbf{\bibinfo{volume}{71}}, \bibinfo{pages}{205}

532:   (\bibinfo{year}{1931}).

533: 

534: \bibitem[{\citenamefont{Fledderjohann et~al.}(1996)\citenamefont{Fledderjohann,

535:   Gerhardt, M\"utter, Schmitt, and Karbach}}]{FGMSK96}

536: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Fledderjohann}},

537:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Gerhardt}},

538:   \bibinfo{author}{\bibfnamefont{K.~H.} \bibnamefont{M\"utter}},

539:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Schmitt}}, \bibnamefont{and}

540:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Karbach}},

541:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{54}},

542:   \bibinfo{pages}{7168} (\bibinfo{year}{1996}).

543: 

544: \bibitem[{\citenamefont{Kl\"umper and Johnston}(2000)}]{KJ00}

545: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kl\"umper}} \bibnamefont{and}

546:   \bibinfo{author}{\bibfnamefont{D.~C.}~\bibnamefont{Johnston}},

547:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{84}},

548:   \bibinfo{pages}{4701} (\bibinfo{year}{2000}).

549: 

550: \bibitem[{\citenamefont{Oshikawa and Affleck}(1997)}]{OA97}

551: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Oshikawa}} \bibnamefont{and}

552:   \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Affleck}},

553:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{79}},

554:   \bibinfo{pages}{2883} (\bibinfo{year}{1997}).

555: 

556: \bibitem[{\citenamefont{Affleck and Oshikawa}(1999)}]{AO99}

557: \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Affleck}} \bibnamefont{and}

558:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Oshikawa}},

559:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{60}},

560:   \bibinfo{pages}{1038} (\bibinfo{year}{1999}) and \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{62}},

561:   \bibinfo{pages}{9200} (\bibinfo{year}{2000}).

562: 

563: %\bibitem[{\citenamefont{Affleck and Oshikawa}(2000)}]{AO00}

564: %\bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Affleck}} \bibnamefont{and}

565: %  \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Oshikawa}},

566: %  \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{62}},

567: %  \bibinfo{pages}{9200} (\bibinfo{year}{2000}).

568: 

569: \bibitem[{\citenamefont{Feyerherm et~al.}(2000)\citenamefont{Feyerherm, Abens,

570:   G\"unther, Ishida, Mei{\ss}ner, Meschke, Nogami, and Steiner}}]{FAGIMMNS00}

571: \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Feyerherm}},

572:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Abens}},

573:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{G\"unther}},

574:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Ishida}},

575:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Mei{\ss}ner}},

576:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Meschke}},

577:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Nogami}}, \bibnamefont{and}

578:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Steiner}},

579:   \bibinfo{journal}{J. Phys.: Condens. Matter} \textbf{\bibinfo{volume}{12}},

580:   \bibinfo{pages}{8495} (\bibinfo{year}{2000}).

581: 

582: \bibitem[{\citenamefont{Zhao et~al.}(2003)\citenamefont{Zhao, Wang, Xiang, Su,

583:   and Yu}}]{ZWXSY03}

584: \bibinfo{author}{\bibfnamefont{J.~Z.}~\bibnamefont{Zhao}},

585:   \bibinfo{author}{\bibfnamefont{X.~Q.}~\bibnamefont{Wang}},

586:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Xiang}},

587:   \bibinfo{author}{\bibfnamefont{Z.~B.}~\bibnamefont{Su}}, \bibnamefont{and}

588:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Yu}}, \bibinfo{journal}{Phys.

589:   Rev. Lett.} \textbf{\bibinfo{volume}{90}}, \bibinfo{pages}{207204}

590:   (\bibinfo{year}{2003}).

591: 

592: \bibitem[{\citenamefont{Kenzelmann et~al.}(2004)\citenamefont{Kenzelmann,

593:   Chen, Broholm, Reich, and Qiu}}]{Kenzel04}

594: \bibinfo{author}{\bibfnamefont{M.} \bibnamefont{Kenzelmann}},

595:   \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Chen}},

596:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Broholm}},

597:   \bibinfo{author}{\bibfnamefont{D.~H.}~\bibnamefont{Reich}}, \bibnamefont{and}

598:   \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Qiu}},

599:   \bibinfo{journal}{Phys. Rev. Lett.}

600:   \textbf{\bibinfo{volume}{93}}, \bibinfo{pages}{017204}

601:   (\bibinfo{year}{2004}).

602: 

603: \bibitem[{\citenamefont{Lou et~al.}(2005)\citenamefont{Lou, Chen, Zhao, Wang,

604:   Xiang, Su and Yu}}]{Lou05}

605: \bibinfo{author}{\bibfnamefont{J.} \bibnamefont{Lou}},

606:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Chen}},

607:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Zhao}},

608:   \bibinfo{author}{\bibfnamefont{X.}~\bibnamefont{Wang}},

609:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Xiang}},

610:   \bibinfo{author}{\bibfnamefont{Z.}~\bibnamefont{Su}}, \bibnamefont{and}

611:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Yu}},

612:   \bibinfo{journal}{Phys. Rev. Lett.}

613:   \textbf{\bibinfo{volume}{94}}, \bibinfo{pages}{217207}

614:   (\bibinfo{year}{2005}).

615: 

616: \bibitem[{\citenamefont{Wolter et~al.}(2003)\citenamefont{Wolter, Rakoto,

617:   Costes, Honecker, Brenig, Kl\"umper, Klauss, Litterst, Feyerherm, J\'erome,

618:   and S\"ullow}}]{Wolter03}

619: \bibinfo{author}{\bibfnamefont{A.~U.~B.} \bibnamefont{Wolter}},

620:   \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Rakoto}},

621:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Costes}},

622:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Honecker}},

623:   \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Brenig}},

624:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kl\"umper}},

625:   \bibinfo{author}{\bibfnamefont{H.-H.}~\bibnamefont{Klauss}},

626:   \bibinfo{author}{\bibfnamefont{F.~J.}~\bibnamefont{Litterst}},

627:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Feyerherm}},

628:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{J\'erome}}, \bibnamefont{and}

629:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{S\"ullow}},

630:   \bibinfo{journal}{Phys. Rev. B}

631:   \textbf{\bibinfo{volume}{68}}, \bibinfo{pages}{220406(R)}

632:   (\bibinfo{year}{2003}).

633: 

634: \bibitem[{\citenamefont{Trotter}(1959)}]{Trotter59}

635: \bibinfo{author}{\bibfnamefont{H.~F.} \bibnamefont{Trotter}},

636:   \bibinfo{journal}{Proc. Amer. Math. Soc.} \textbf{\bibinfo{volume}{10}},

637:   \bibinfo{pages}{545} (\bibinfo{year}{1959}).

638: 

639: \bibitem[{\citenamefont{Suzuki}(1985)}]{Suzuki85}

640: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Suzuki}},

641:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{31}},

642:   \bibinfo{pages}{2957} (\bibinfo{year}{1985}).

643: 

644: \bibitem[{\citenamefont{Sirker and Kl\"umper}(2002{\natexlab{a}})}]{JK02}

645: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Sirker}} \bibnamefont{and}

646:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kl\"umper}},

647:   \bibinfo{journal}{Europhys. Lett.} \textbf{\bibinfo{volume}{60}},

648:   \bibinfo{pages}{262} (\bibinfo{year}{2002}{\natexlab{a}}) and 

649: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{66}},

650:   \bibinfo{pages}{245102} (\bibinfo{year}{2002}{\natexlab{b}}).

651: 

652: %\bibitem[{\citenamefont{Sirker and Kl\"umper}(2002{\natexlab{b}})}]{JK02_prb}

653: %\bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Sirker}} \bibnamefont{and}

654: %  \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kl\"umper}},

655: %  \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{66}},

656: %  \bibinfo{pages}{245102} (\bibinfo{year}{2002}{\natexlab{b}}).

657: 

658: \bibitem[{\citenamefont{Peschel et~al.}(1999)\citenamefont{Peschel, Wang,

659:   Kaulke, and Hallberg}}]{DMRG_book}

660: \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Peschel}},

661:   \bibinfo{author}{\bibfnamefont{X.}~\bibnamefont{Wang}},

662:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Kaulke}}, \bibnamefont{and}

663:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Hallberg}},

664:   \emph{\bibinfo{title}{Density-Matrix Renormalization, Lecture Notes in

665:   Physics}}, vol. \bibinfo{volume}{528} (\bibinfo{publisher}{Springer, Berlin},

666:   \bibinfo{year}{1999}), \bibinfo{note}{and references therein}.

667: 

668: \bibitem[{\citenamefont{Bursill, Xiang and Gehring}(1996)}]{BXG96}

669: \bibinfo{author}{\bibfnamefont{R.~J.} \bibnamefont{Bursill}},

670:   \bibinfo{author}{\bibfnamefont{T.} \bibnamefont{Xiang}}, \bibnamefont{and}

671:   \bibinfo{author}{\bibfnamefont{G.~A.} \bibnamefont{Gehring}},

672:   \bibinfo{journal}{J. Phys. Conden. Matter} \textbf{\bibinfo{volume}{8}},

673:   \bibinfo{pages}{L583} (\bibinfo{year}{1996}).

674: 

675: \bibitem[{\citenamefont{Wang and Xiang}(1997)}]{Wang_Xiang97}

676: \bibinfo{author}{\bibfnamefont{X.}~\bibnamefont{Wang}} \bibnamefont{and}

677:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Xiang}},

678:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{56}},

679:   \bibinfo{pages}{5061} (\bibinfo{year}{1997}).

680: 

681: \bibitem[{\citenamefont{Shibata}(1997)}]{Shibata97}

682: \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Shibata}}, \bibinfo{journal}{J.

683:   Phys. Soc. Jpn.} \textbf{\bibinfo{volume}{66}}, \bibinfo{pages}{2221}

684:   (\bibinfo{year}{1997}).

685: 

686: \bibitem[{\citenamefont{Wolter et~al.}(2005)\citenamefont{Wolter, Wzietek,

687:   S\"ullow, Litterst, Honecker, Brenig, Feyerhern, and Klauss}}]{Wolter05}

688: \bibinfo{author}{\bibfnamefont{A.~U.~B.} \bibnamefont{Wolter}},

689:   \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Wzietek}},

690:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{S\"ullow}},

691:   \bibinfo{author}{\bibfnamefont{F.~J.}~\bibnamefont{Litterst}},

692:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Honecker}},

693:   \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Brenig}},

694:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Feyerherm}}, \bibnamefont{and}

695:   \bibinfo{author}{\bibfnamefont{H.-H.}~\bibnamefont{Klauss}},

696:   \bibnamefont{et~al.}, \bibinfo{journal}{Phys. Rev. Lett.}

697:   \textbf{\bibinfo{volume}{94}}, \bibinfo{pages}{057204}

698:   (\bibinfo{year}{2005}).

699: 

700: \bibitem[{\citenamefont{Kl\"umper}(1999)}]{AK98}

701: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kl\"umper}},

702:   \bibinfo{journal}{Eur. Phys. J. B} \textbf{\bibinfo{volume}{5}},

703:   \bibinfo{pages}{677} (\bibinfo{year}{1999}).

704: 

705: %\bibitem[{\citenamefont{Zhao et~al.}(2003)\citenamefont{Zhao, Wang, Xiang, Su,

706: %  and Yu}}]{ZWXSY03}

707: %\bibinfo{author}{\bibfnamefont{J.~Z.}~\bibnamefont{Zhao}},

708: %  \bibinfo{author}{\bibfnamefont{X.~Q.}~\bibnamefont{Wang}},

709: %  \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Xiang}},

710: %  \bibinfo{author}{\bibfnamefont{Z.~B.}~\bibnamefont{Su}}, \bibnamefont{and}

711: %  \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Yu}}, \bibinfo{journal}{Phys.

712: %  Rev. Lett.} \textbf{\bibinfo{volume}{90}}, \bibinfo{pages}{207204}

713: %  (\bibinfo{year}{2003}).

714: 

715: \bibitem[{\citenamefont{Zvyagin et~al.}(2004)\citenamefont{Zvyagin, Kolezhuk,

716:   Krzystek, and Feyerherm}}]{ZKKF04}

717: \bibinfo{author}{\bibfnamefont{S.~A.} \bibnamefont{Zvyagin}},

718:   \bibinfo{author}{\bibfnamefont{A.~K.} \bibnamefont{Kolezhuk}},

719:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Krzystek}}, \bibnamefont{and}

720:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Feyerherm}},

721:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{93}},

722:   \bibinfo{pages}{027201} (\bibinfo{year}{2004}).

723: 

724: \end{thebibliography}

725: 

726: \vspace{-0.3cm}

727: \end{document} 

728: 

729: 

730: 

731: 

732: 

733: 

734: 

735: 

736: 

737: 

738: 

739: 

740: 

741: