1: \documentclass[aps,prl,twocolumn,superscriptaddress,showpacs,letter]{revtex4}

2: \usepackage{amsfonts,amssymb,amsmath}

3: \usepackage{color}

4: \usepackage{graphicx}

5: \usepackage{dcolumn}% Align table columns on decimal point

6: \usepackage{bm}% bold math

7: 

8: \newcommand{\bra}[1]{\langle #1|}

9: \newcommand{\ket}[1]{|#1 \rangle}

10: \newcommand{\Tr}{\mathop{\mathrm{Tr}}}

11: 

12: \begin{document}

13: \title{Exact results of the quantum phase transition for the topological order}

14: 

15: \author{Jing Yu}

16:   \affiliation{Department of Physics, Beijing Normal University, Beijing 100875, China}

17: 

18: \author{Su-Peng Kou}

19:   \email{spkou@bnu.edu.cn}

20:   \affiliation{Department of Physics, Beijing Normal University, Beijing 100875, China}

21: 

22: \author{Xiao-Gang Wen}

23:   \affiliation{Department of Physics, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139}

24: 

25: \begin{abstract}

26: In this paper{\em \ }a duality between the $d=2$\ Wen-plaquette

27: model in a transverse field and the $d=1$\ Ising model in a

28: transverse field{\em \ }is used to learn the nature of the quantum

29: phase transition (QPT) between a spin-polarized phase and a

30: topological ordered state with the string-net condensation. The QPT

31: is not induced by spontaneous symmetry breaking and there are no

32: conventional Landau-type local order parameters. Instead the

33: string-like non-local order parameters are introduced to describe

34: the QPT. In particular, the duality character between the

35: open-string and closed string for the QPT are explored.

36: 

37: Keywords: topological order, quantum phase transition, the

38: Wen-plaquette model

39: 

40: \end{abstract}

41: 

42: \pacs{75.10.Jm, 05.70.Jk} \maketitle

43: 

44: Landau developed a sysmatical theory for classical statistical

45: systems. Different orders are characterized by different symmetries.

46: The phase transition between different phases are always accompanied

47: with symmetry breaking. However, Landau's theory cannot describe all

48: the continuous phase transition such as the transitions between

49: quantum ordered phase\cite{wen}. Topological ordered state is a

50: special quantum order with full gapped excitations. The first

51: nontrivial example of topological order is the fractional quantum

52: hall (FQH) states which cannot be described by broken symmetries

53: \cite{wen}. Recently, people pay attention to the quantum phase

54: transition (QPT) for the topological orders. It is known that there

55: exists a quantum phase transition between a spin-polarized phase and

56: a topological ordered state. The two phases cannot be characterized

57: by a local order parameter. Thus such a quantum phase transition

58: cannot be described by the Landau theory obviously. To find a new

59: approach to learn the nature of such QPT becomes an interesting

60: issue. The quantum phase transition for the

61: Kitaev toric-code model in a transverse magnetic field was studied in Ref.%

62:  \cite{nayak}. The numerical simulations demonstrate the condensation of

63: `magnetic' excitations and the confinement of `electric' charges of

64: the phase transition out of the topological phase. And then people

65: find that the ``topological entropy'' $S_{\mathrm{top}}$\

66: \cite{Kitaev:06Levin:06,levin} serves as an order

67: parameter\cite{lidar,chamon}. At the QPT the ``topological entropy''

68: jumps abruptly from a finite number to zero out of the topological

69: phase. In addition, in Ref. \cite{xiang} it is revealed that the

70: quantum phase transitions between different topological orders can

71: be exactly characterized by a topological order parameter and shows

72: the one dimensional (1D) properties.

73: 

74: In this paper we focus on the Wen-plaquette model on a square lattice\cite

75: {wen,wen1}. The ground states of the Wen-plaquette model are $Z_2$

76: topological ordered state which is similar to that for the kitaev toric-code

77: model \cite{kitaev}. After adding the transverse field, a quantum phase

78: transition occurs by raising the strength of the transverse field. However,

79: the model cannot be solved exactly. Fortunately, we find\emph{\ }a duality

80: between the $d=2$\ Wen-plaquette model in a transverse field ( transverse

81: Wen-plaquette model) and the $d=1$\ Ising model in a transverse field (

82: transverse\ Ising model). These results indicate a new type of QPT between a

83: topological ordered phase and a non-topological ordered phase, including the

84: non-local order parameters, the 1D properties of the QPT, in particular,

85: \emph{the duality between open-string and closed-string}.

86: 

87: The Hamiltonian for the $d=2$ transverse Wen-plaquette model on square

88: lattice\cite{wen,wen1} is given as:

89: \begin{equation}

90: \mathcal{H}_{\mathrm{wen}}=g\sum_iF_i+h\sum_i\sigma _i^x  \label{wen}

91: \end{equation}

92: with $F_i=\sigma _i^x\sigma _{i+\hat{e}_x}^y\sigma _{i+\hat{e}_x+\hat{e}%

93: _y}^x\sigma _{i+\hat{e}_y}^y$ and $g<0.$ $\sigma _i^x,$ $\sigma _i^y$ are

94: Pauli matrices on sites, $i.$

95: 

96: For the first part of the model, it is an exact solved model\cite{wen1}. The

97: ground state of it is a topological order described by $Z_2$ projective

98: symmetry groups. And there exists three types of quasiparticles: $Z_2$

99: charge, $Z_2$ vortex, and fermions, of which three types of string operators

100: T1, T2, T3, are defined, respectively\cite{wen2}. The fermions can be

101: regarded as bound states of a $Z_2$ charge on even plaquette and a $Z_2$

102: vortex on odd plaquette. The string operator for $Z_2$ charge has a form $%

103: W_c(C)=\prod\limits_{m\in C}\sigma _{i_m}^{l_m}$, where $C$ is a string

104: connecting the even-plaquettes of the neighboring links, and $i_m$ are sites

105: on the string. For $Z_2$ vortex, string operator is $W_V(C)=\prod\limits_{m%

106: \in \tilde{C}}\sigma _{i_m}^{l_m}$ with $\tilde{C}$ as a string connecting

107: the even-plaquettes of the neighboring links. For an open T1 (T2) string,

108: the ends points represent the $Z_2$ charge ( $Z_2$ vortex). It is known that

109: the ground state of the topological order has a condensation of the

110: closed-strings as $\left\langle W_{c,\tilde{c}}(C)\right\rangle \neq 0.$

111: \cite{wen2} On the contrary, the open strings have mass gap without

112: condensation.

113: 

114: It is known that the Wen-plaquette model ($h=0)$ can be mapped onto nearest

115: neighbor Ising chains\cite{ort,hu}. This is because that the Wen-plaquette

116: model ( $h=0$) has an energy spectrum which is identical to that of the one

117: dimensional Ising chain\cite{ort}. Furthermore, we find that the transverse

118: Wen-plaquette model on a square lattice in Eq.(\ref{wen}) corresponds to the

119: one dimensional Ising chain in transverse field. In the following part of

120: this paper, we will show the duality and use it to learn the nature of the

121: QPT between a spin-polarized phase and a topological ordered state.

122: 

123: To obtain the duality we define $\sigma _i^x\sigma _{i+\hat{e}_x}^y\sigma

124: _{i+\hat{e}_x+\hat{e}_y}^x\sigma _{i+\hat{e}_y}^y=A_i,$ $\sigma _i^x=B_i.$

125: The two terms here have the commutation relations as: $\left[ A_i,B_j\right]

126: =2A_iB_j(\delta _{j,i+\hat{e}_x}+\delta _{j,i+\hat{e}_y})$ and $\left[

127: A_i,A_j\right] =0,$ $\left[ B_i,B_j\right] =0.$ Then we denote $A_i$ by $%

128: \tau _{i+\frac 12}^x$. To realize the Pauli algebra for $A_i$ and $B_j$, $%

129: B_j $ is denoted by $\tau _{j-\frac 12}^z\tau _{j+\frac 12}^z$. Here $\tau

130: _{j+\frac 12}^x,$ $\tau _{j+\frac 12}^z$ are Pauli matrices on sites, $%

131: j+\frac 12.$ Then the mapping between the two models is given as

132: \begin{equation}

133: \sigma _i^x\sigma _{i+\hat{e}_x}^y\sigma _{i+\hat{e}_x+\hat{e}_y}^x\sigma

134: _{i+\hat{e}_y}^y\longmapsto \tau _{i+\frac 12}^x,\text{ }\sigma

135: _i^x\longmapsto \tau _{i-\frac 12}^z\tau _{i+\frac 12}^z.  \nonumber

136: \end{equation}

137: 

138: Then we can explicitly denote the original model by the following

139: Hamiltonian, describing the Ising model in a transverse field in $d=1$:

140: \begin{equation}

141: \mathcal{H}_{\mathrm{wen}}\mapsto \mathcal{H}_I=-h\sum_a\sum_i\left( g_I\tau

142: _{a,i+\frac 12}^x+\tau _{a,i-\frac 12}^z\tau _{a,i+\frac 12}^z\right)

143: \label{hamising}

144: \end{equation}

145: where $a$ is the chain-index which is different for the $d=2$ transverse

146: Wen-plaquette model on different lattices with different boundary conditions

147: (see Fig.1). $h$ is an overall energy scale and $g_I=\frac 1{h_I}=\frac gh>0$

148: is a dimensionless coupling constant\cite{commet}. This is a 1D Ising model

149: along diagonal directions. By the mapping above, we find $g$-term in wen's

150: model corresponds to external field term in Ising model, while $h$-term

151: corresponds to Ising-term.

152: 

153: \begin{figure}

154: \includegraphics[width=0.48\textwidth]{fig1.EPS}

155: \caption{The dotted lines are the dual Ising chains with $a$ as the

156: chain-index and $i$ as the site-index. }

157: \label{Fig.1}

158: \end{figure}

159: 

160: In addition, an important feature about the duality is the boundary

161: condition. The $d=2$\ transverse Wen-plaquette model on $N\times N$\ square

162: lattice with the open boundary condition is dual to $2N-1$\ decoupled Ising

163: chains ( in a transverse field). Here the chain-index $a$ is from $1$ to $%

164: M+N-1$. However, on $N\times N$\ square lattice with the periodic boundary

165: condition, the $d=2$\ transverse Wen-plaquette model $1$\ corresponds to $N$%

166: \ decoupled Ising chains.

167: 

168: \textit{Global phase diagram --- }From the duality, one can obtain the

169: global phase diagram for the original model by studying the transverse Ising

170: chain. The model in Eq.(\ref{wen}) has the same phase diagram to that for

171: the $d=1$ transverse Ising model described by Eq.(\ref{hamising}). Then the

172: original model in Eq.(\ref{wen}) has a $T=0$ quantum phase transition at $%

173: g_I=\frac gh=1$ from topological ordered state with ($g_I>1$), to a gapped

174: spin-polarized state with ($g_I<1$). For large $g_I$, $g_I>1$, the ground

175: state is topological order which is dual to the spin-polarized state in the

176: 1D Ising model with $\left\langle \tau _{i+\frac 12}^z\right\rangle =0$ . On

177: the other hand, for small $g_I$, $g_I<1,$ the spin-polarized state is in a

178: superposition of $\sigma ^x$ eigenstates, with different sites uncorrected.

179: This state corresponds to the spin ordered state in the dual 1D Ising model

180: with a spontaneous magnetization $\left\langle \tau _{i+\frac

181: 12}^z\right\rangle \neq 0$.

182: 

183: \label{sec:ising}\textit{The scaling law near the QPT ---} To obtain the

184: scaling law near the QPT, the above spin-$\frac 12$ transverse Ising model

185: in Eq.(\ref{hamising}) can be described by the following Hamiltonian

186: \begin{eqnarray}

187: \mathcal{H}_I &=&h\sum\limits_{a,j}[-(c_{a,j}^{\dagger

188: }-c_{a,j})(c_{a,j+1}^{\dagger }+c_{a,j+1}) \\

189: &&+g_I(c_{a,j}^{\dagger }-c_{a,j})(c_{a,j}^{\dagger }+c_{a,j})]  \nonumber

190: \end{eqnarray}

191: after employing Jordan-Wigner transformation of the spin operators to

192: spinless fermions\cite{lsm,mccoy}, $\tau _{a,j+\frac 12}^x=2c_{a,j}^{\dagger

193: }c_{a,j}-1,$ $\tau _{a,j+\frac 12}^z=(-1)^{j-1}\exp (\pm i\pi

194: \sum\limits_{n=1}^{j-1}c_{a,n}^{\dagger }c_{a,n})(c_{a,j}^{\dagger

195: }+c_{a,j}).$ In the fermionic representation for the transverse Ising chain,

196: the energy spectrum for fermions $E_k=\pm 2h\sqrt{(g_I-\cos (k\tilde{a}%

197: ))^2+\sin ^2(k\tilde{a})}$ has an energy gap $E_{k=0}=2h|g_I-1|$ which tends

198: to zero at $g_I\rightarrow 1$. So the energy gap for the fermions must tend

199: to zero at the transition $m=\frac 1{2h\tilde{a}^2}(g_I-1)~\mbox{and}~~c=2h%

200: \tilde{a},$ where $\tilde{a}$ is the lattice spacing. Here $m$ is the mass

201: for the Majorana fermion which changes sign by tuning the mode across the

202: transition; we have chosen $m>0$ to correspond to the topological ordered

203: side.\cite{ising}.

204: 

205: \textit{The 1D properties for the QPT} \textit{--- }People have guessed that

206: the energy gap for the elementary excitations will always close at the QPT

207: for topological orders. Our results confirm the conjecture. At the critical

208: point, $g_I=1,$ the energy for the fermion excitation becomes $E_k=sim

209: c\cdot k$ where the velocity is $2h\tilde{a}$. Since the result is obtained

210: from the dual Ising chain in Eq.(\ref{hamising}), the energy gap between the

211: ground state and the first excited state is $\delta E_k\sim c\cdot \delta

212: k=c\cdot \frac{2\pi }L$ where $L$ is the length of the Ising chain. The

213: energy gap turns into zero at QPT in the thermodynamic limit. In Ref.(\cite

214: {lidar}) in the thermodynamic limit, it is predicted that the gap between

215: the ground state and the first excited state closes scales as at the

216: critical point $\delta E(\mathcal{N})=E_1-E_0\sim \mathcal{N}^{-\frac 12},$ $%

217: (\mathcal{N}=N^2).$ On a $N\times N$ lattice with periodic boundary

218: condition, the original model is dual to $N $ decoupled Ising chains, then

219: the gap is determined by the length of each Ising chain which indeed scales

220: as $\delta E\sim N^{-1}\sim \mathcal{N}^{-\frac 12}.$ The Fig.2 shows the

221: energy gap $\delta E_k$ at the critical point on different lattices from

222: above results and those from the exact diagonal numerical results. From the

223: results in Fig.2, the energy gap doesn't always scale as $\delta E(\mathcal{N%

224: })\sim \mathcal{N}^{-\frac 12}$\emph{. }

225: 

226: \begin{figure}

227: \includegraphics[width=0.5\textwidth]{fig2.EPS}

228: \caption{The energy gap $\delta E$ (scaled by $4g$ ) between the ground

229: state and the first excited state at the critical point $g=h$ on different

230: lattices from analysis results. The inset shows the energy gap $\delta E$

231: (scaled by $g$ ) at the critical point $g=h$ on different lattices from the

232: exact diagonal numerical results. Here $N*M$ denotes a $N\times M$ lattice.}

233: \label{Fig.2}

234: \end{figure}

235: 

236: In addition, the correlation function at the QPT shows the 1D characters%

237: \textit{. }The long-distance limit of the static correlator for $F_iF_j$ of

238: the 2D transverse Wen-plaquette model can be described by the dual

239: correlator for $\tau _{a,i}^x\tau _{a^{\prime },j^{\prime }}^x$ of the 1D

240: Ising model in the transverse field. For a diagonal case, $F_iF_j=F_iF_{i-n%

241: \hat{e}_x+n\hat{e}_y}$, the correlation function at the QPT is given as\cite

242: {mccoy}\cite{ising}

243: \begin{equation}

244: \left\langle F_iF_{i-n\hat{e}_x+n\hat{e}_y}\right\rangle -\frac 2{\pi

245: ^2}=\frac 4{\pi ^2(4n^2-1)}.

246: \end{equation}

247: Here $m$, $n$ are integer numbers.

248: 

249: \textit{The non-local order parameters ---} Firstly we try to use the

250: expectation value for $\sigma _i^x$ and $F_i$ to be the order parameters to

251: describe the QPT. However, $\langle \sigma _i^x\rangle $ and $\langle

252: F_i\rangle $ are finite in both phases but discontinue at the QPT. So one

253: cannot use the expectation values of $\sigma _i^x$ and $F_i$ to be the order

254: parameters to describe the QPT.

255: 

256: \begin{figure}

257: \includegraphics[clip,width=0.4\textwidth]{fig3.EPS}

258: \caption{The relationship between an open-string and a closed-string. The

259: hexagon black line denotes a closed-string for $F_iF_{i+1}F_{i+2}F_{i+3}.$

260: The black line below it is a open-string for $\sigma _i^x\sigma

261: _{i+1}^x\sigma _{i+2}^x,$ of which two Z2 charges (or Z2 vortexes) live at

262: the end denoted by the black squares.}

263: \label{Fig.3}

264: \end{figure}

265: 

266: Instead, we define two kinds of non-local order parameters : one is $\phi

267: _1=\langle \prod\limits_{i_n=1}^i\sigma _{i_n}^x\rangle $; the other is $%

268: \phi _2=\left\langle \prod\limits_{_{i_n=1}}^iF_{i_n}\right\rangle $. Here $%

269: i_n=i-n\hat{e}_x+n\hat{e}_y$ represents the sites along diagonal direction.

270: Let's explain the physics properties of them. On the one hand, $%

271: \prod\limits_{i_n=1}^i\sigma _{i_n}^x$ is a T1 or T2 open-string operator

272: with two ends, $i_n=1$ and $i_n=i$. On the other hand, $\prod%

273: \limits_{_{i_n=1}}^iF_{i_n}$ is a T1 or T2 close-string operator around the

274: points $i_n=1$ and $i_n=i$. Fig.3 shows the relationship between an

275: open-string and a closed-string. The expectation values of (open or close)

276: string operators\textit{\ }$\prod\limits_{i_n=1}^i\sigma _{i_n}^x$ and $%

277: \prod\limits_{_{i_n=1}}^iF_{i_n}$ are given by the Fig.4. From it one can

278: see that the QPT can be characterized in terms of the two kinds of non-local

279: order parameters.

280: 

281: \textit{The open-string - closed-string duality ---}\emph{\ }Another

282: important feature for the QPT between the topological order and the

283: non-topological order is \emph{the open-string - closed-string duality.}

284: From the duality relationship, $\sigma _i^x\sigma _{i+\hat{e}_x}^y\sigma _{i+%

285: \hat{e}_x+\hat{e}_y}^x\sigma _{i+\hat{e}_y}^y\longmapsto \tau _{i+\frac

286: 12}^x $ and $\sigma _i^x\longmapsto \tau _{i-\frac 12}^z\tau _{i+\frac 12}^z,

287: $ one can define the first order parameter as $\phi _1=\left\langle \tau

288: _{a,1+\frac 12}^z\tau _{a,i+\frac 12}^z\right\rangle $ and the second as $%

289: \phi _2=\left\langle \prod\limits_{_j=i}^i\tau _{a,i+\frac

290: 12}^x\right\rangle $ in the dual model. And $\phi _1$ is just the local

291: order parameter in dual model which is dual to $\phi _2$, a `disorder order

292: parameter' for the transverse Ising chain. This property leads to the

293: duality between open-string and closed string.

294: 

295: From Fig.4, one can see that in the topological ordered phase, $g_I>1$, we

296: have $\left\langle \phi _1\right\rangle =0$ and $\left\langle \phi

297: _2\right\rangle \sim (1-\frac hg)^{\frac 18}\neq 0.$ This result shows that

298: in this phase the T1 and T2 close-strings are both condensed while the

299: open-strings of them are not. It is consistent to the fact that the ground

300: state of the topological order has a condensation of the closed-strings \cite

301: {wen2} $\left\langle W_{c,\tilde{c}}(C)\right\rangle \neq 0$.

302: 

303: \begin{figure}

304: \includegraphics[clip,width=0.4\textwidth]{fig4.EPS}

305: \caption{The expectation values of (open or close) string operators\textit{\

306: }$\prod\limits_{i_n=1}^i\sigma _{i_n}^x$ and $\prod%

307: \limits_{_{i_n=1}}^iF_{i_n}$. }

308: \label{Fig.4}

309: \end{figure}

310: 

311: On the other hand, in the spin-polarized phase, $g_I<1$, a long range order

312: for dual spin correlation function of $\tau _{a,i-\frac 12}^z$ indicates the

313: condensation of the open-string operator near the QPT\cite{xiang}

314: \begin{equation}

315: \left\langle \phi _1\right\rangle =\lim_{i\rightarrow \infty }\langle

316: \prod\limits_{i_n=1}^i\sigma _{i_n}^x\rangle =\lim_{i\rightarrow \infty

317: }\langle \tau _{a,\frac 12}^z\tau _{a,i-\frac 12}^z\rangle \sim (1-\frac

318: gh)^{\frac 14}.

319: \end{equation}

320: It is known that the open-string operator, $\prod\limits_{i_n=1}^i\sigma

321: _{i_n}^x,$ will generate two Z2 vortexes (or Z2 charges) at both ends of the

322: string. The nonzero expectation value of it means that it is \emph{the

323: `condensation' of both Z2 vortex and Z2 charge} that breaks down the

324: topological order. The disappearance of the condensation of closed-string $%

325: \left\langle \phi _2\right\rangle =0$ or $\left\langle W_{c,\tilde{c}%

326: }(C)\right\rangle =0$ means that this phase is not a topological order any

327: more.

328: 

329: The situation here is much difference from that in Ref.\cite{nayak}, of

330: which the topological order is broken down by the condensation of `magnetic'

331: excitations (Z2 vortex ), and the confinement of `electric' charges (Z2

332: charge). Thus our results indicate there may exist different roads to break

333: down the topological order : the condensation for both Z2 vortex and Z2

334: charge or the condensation of Z2 vortex (`magnetic' excitation) together

335: with the confinement of Z2 charge (`electric' charge).

336: 

337: In summary, a duality between the $d=2$\ Wen-plaquette model in a transverse

338: field ( transverse Wen-plaquette model) and the $d=1$\ Ising model in a

339: transverse field ( transverse\ Ising model) are uncovered. By the duality,

340: we find that the $d=2$ transverse Wen-plaquette model on square lattice

341: undergoes a continuous phase transition. The quantum phase transition

342: between a spin-polarized phase and a topological ordered state with the

343: string-net condensation is not induced by spontaneous symmetry breaking and

344: there are no conventional Landau-type local order parameters. Instead, the

345: QPT can be characterized in terms of the two kinds of non-local order

346: parameters, e.g., \textit{the expectation value of (open or close)

347: string-like operators }$\prod\limits_{i_n=1}^i\sigma _{i_n}^x$ and $%

348: \prod\limits_{_{i_n=1}}^iF_{i_n}$. And the topological ordered state are

349: destroyed by the condensation of the open-string operators ($Z_2$ charge and

350: $Z_2$ vortex). These results indicate a new type of QPT between a

351: topological ordered phase and a non-topological ordered phase.

352: 

353: The authors acknowledge stimulating discussions with S. Chen, N.H. Tong.

354: S.P. Kou acknowledges that this research is supported by NFSC Grant no.

355: 10574014.

356: 

357: \begin{thebibliography}{99}

358: 

359: \bibitem{wen}  {X.-G. Wen}, \emph{Quantum Field Theory of Many-Body Systems}%

360: , (Oxford Univ. Press, Oxford, 2004).

361: 

362: \bibitem{nayak}  S. Trebst, P. Werner, M. Troyer, K. Shtengel, C. Nayak,

363: Phys. Rev. Lett. 98, 070602 (2007).

364: 

365: \bibitem{Kitaev:06Levin:06}  A. Kitaev and J. Preskill, Phys. Rev. Lett.

366: \textbf{96}, 110404 (2006).

367: 

368: \bibitem{levin}  M. Levin and X.-G. Wen, Phys. Rev. Lett. \textbf{96},

369: 110405 (2006).

370: 

371: \bibitem{lidar}  A. Hamma, D. A. Lidar, quant-ph/0607145; A. Hamma, W.

372: Zhang, S. Haas, D.A. Lidar, quant-ph/07050026.

373: 

374: \bibitem{chamon}  C. Castelnovo and C. Chamon, cond-mat/07072084.

375: 

376: \bibitem{xiang}  Xiao-Yong Feng, Guang-Ming Zhang, Tao Xiang, Phys. Rev.

377: Lett. 98, 087204 (2007).

378: 

379: \bibitem{wen1}  X. G. Wen, {Phys. Rev. Lett. }\textbf{90} (2), 016803 (2003).

380: 

381: \bibitem{kitaev}  {A. Kitaev}, {Annals Phys.} \textbf{303}, 2 (2003).

382: 

383: \bibitem{wen2}  X. G. Wen, Phys. Rev. D 68, 065003 (2003).

384: 

385: \bibitem{ort}  Z. Nussinov, G. Ortiz, cond-mat/0702377.

386: 

387: \bibitem{hu}  H. D. Chen and J. P. Hu, priprent, cond-mat/0702366.

388: 

389: \bibitem{commet}  A small \textrm{y}-direction transverse field leads to a

390: non-local-interchain interaction for the dual $d=1$ Ising chains. Such

391: relevant interaction couples isolated Ising chains together. And the effects

392: of it will be explored elsewhere.

393: 

394: \bibitem{lsm}  E. Lieb, T. Schultz, and D. Mattis, Ann. of Phys. \textbf{16}%

395: , 406 (1961); P. Pfeuty, Ann. of Phys. \textbf{57}, 79 (1970).

396: 

397: \bibitem{mccoy}  B.M.~McCoy \emph{et. al.}, Phys. Rev. A \textbf{4}, 2331

398: (1971).

399: 

400: \bibitem{ising}  B K Chakrabarti; A Dutta; P Sen, \emph{Quantum Ising phases

401: and transitions in transverse Ising models}\textit{\ }(Lecture Notes in

402: Physics. New Series M, Monographs, M41), (Hardcover - April 1996).

403: 

404: \bibitem{duality1}  E. Fradkin and L. Susskind, Phys. Rev. D \textbf{17},

405: 2637 (1978).

406: 

407: \end{thebibliography}

408: 

409: \end{document}

410: