1: \documentclass[12pt]{iopart}

2: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3: %\usepackage{iopams}

4: \usepackage[dvips]{graphicx}

5: 

6: \begin{document}

7: 

8: \letter{Entanglement and spin

9: squeezing in the two-atom Dicke model}

10: 

11: \author{A Messikh\dag,\ Z Ficek\dag\ddag\ and M R B Wahiddin\dag}

12: 

13: \address{\dag\ Centre for Computational and Theoretical Sciences,

14: Kulliyyah of Science, International Islamic University Malaysia,

15: 53100 Kuala Lumpur, Malaysia}

16: 

17: \address{\ddag\ Department of Physics, The University of Queensland,

18: Brisbane, QLD 4072, Australia}

19: 

20: \eads{\mailto{ficek@physics.uq.edu.au}, \mailto{mridza@iiu.edu.my}}

21: 

22: 

23: \begin{abstract}

24:    We analyze the relation between the entanglement and spin-squeezing

25:    parameter in the two-atom Dicke model and identify the source of the

26:    discrepancy recently reported by Banerjee and Zhou \etal that one can

27:    observe entanglement without spin squeezing. Our calculations

28:    demonstrate that

29:    there are two criteria for entanglement, one associated with the

30:    two-photon coherences that create two-photon entangled states, and

31:    the other associated with populations of the collective states.

32:    We find that the spin-squeezing parameter correctly predicts

33:    entanglement in the two-atom Dicke system only if it is associated

34:    with two-photon entangled states, but fails to predict entanglement

35:    when it is associated with the entangled symmetric state.

36:    This explicitly identifies the source of the discrepancy and explains

37:    why the system can be entangled without spin-squeezing. We

38:    illustrate these findings in three examples of the interaction of the

39:    system with thermal, classical squeezed vacuum and quantum squeezed

40:    vacuum fields.

41: \end{abstract}

42: 

43: 

44: \pacs{03.67.Mn, 42.50.Dv, 42.50.Fx}

45: 

46: \submitto{\JOB}

47: 

48: %\maketitle

49: \nosections

50: %\section{Introduction}\label{int}

51: 

52: Entanglement, the most intriguing property of multiparticle systems

53: (or qubits), is one of the key problems in quantum physics and has

54: been the subject of active research in recent years~\cite{si}.

55: It describes a multiparticle system which has the astonishing property

56: that the results of a measurement on one particle cannot be specified

57: independently of the results of measurements on the other particles.

58: Therefore, the generation of entanglement between atoms is fundamental

59: not only to demonstrate quantum nonlocality but also would constitute

60: a valuable resource in the fields of quantum information processing,

61: cryptography and quantum computation~\cite{ben}.

62: In this context, it is not surprising that a tremendous number of

63: theoretical proposals have been made to produce entanglement between

64: separate particles~\cite{ft}. Several different criteria have

65: been proposed to identify entanglement in two-particle systems, but no

66: definite measure of entanglement exists for a number of particles larger

67: than two. Entanglement between two particles can be identified by

68: calculating, for example, the Wootters entanglement measure

69: (concurrence)~\cite{woo}, or a measure proposed by

70: Peres~\cite{per} and Horodecki~\cite{horo} given in terms of the

71: negative eigenvalues

72: of the partial transposition of the density matrix of the two-particle

73: system. Recently, S\o rensen \etal~\cite{sdcz} have proposed a measure

74: of multiparticle entanglement in terms of the spin-squeezing

75: parameter~\cite{ku,ibbg,wbi,sor}

76: \begin{eqnarray}

77:      \xi_{\bi{n}_{i}} =\frac{N_{a}\langle \left(\Delta

78:      S_{\bi{n}_{i}}\right)^{2}\rangle}{ \langle S_{\bi{n}_{j}}\rangle^{2}

79:      +\langle S_{\bi{n}_{k}}\rangle^{2}} \ ,\label{eq1}

80: \end{eqnarray}

81: where $N_{a}$ is the number of particles, $\bi{n}_{i}, \bi{n}_{j}$

82: and $\bi{n}_{k}$ are

83: three mutually orthogonal unit vectors oriented such that the mean

84: value of one of the spin components, say  $\langle S_{\bi{n}_{k}}\rangle$,

85: is different from zero, while the other components $S_{\bi{n}_{i}}$ and

86: $S_{\bi{n}_{j}}$ have zero mean values. The variance

87: $\langle \left(\Delta S_{\bi{n}_{i}}\right)^{2}\rangle$ should be calculated

88: in the plane orthogonal to the mean spin direction.

89: A multiatom system in a coherent state has variances normal to the mean

90: spin direction equal to the standard quantum limit of $N_{a}/4$. In this

91: case, $\xi_{\bi{n}_{i}}=\xi_{\bi{n}_{j}}=1$. A system with the variance

92: reduced below

93: the standard quantum limit in one direction normal to the mean spin

94: direction is characterized by $\xi_{\bi{n}_{i}}<1$, that is spin squeezed

95: in the direction $\bi{n}_{i}$.

96: S\o rensen \etal~\cite{sdcz} have shown that multiparticle spin

97: squeezed systems also exhibit entanglement.

98: 

99: However, in recent studies of entanglement in the two-atom Dicke

100: system~\cite{ban,zsl} it has been discovered that the spin-squeezing

101: parameter $\xi_{\bi{n}_{i}}$ is not sufficient for predicting entanglement

102: in a multiparticle system. Banerjee~\cite{ban} and Zhou~\etal~\cite{zsl}

103: have shown that the two-atom Dicke system driven by a single mode thermal

104: field, can exhibit an entanglement and at the same time

105: $\xi_{\bi{n}_{i}}>1$. They have found that in the thermal field the time

106: evolution of the system is represented by a diagonal density matrix

107: \begin{eqnarray}

108:      \hat{\rho}(t) = \rho_{gg}(t)|g\rangle \langle g|

109:      +\rho_{ee}(t)|e\rangle \langle e|

110:      +\rho_{ss}(t)|s\rangle \langle s| \ ,\label{eq2}

111: \end{eqnarray}

112: where

113: \begin{eqnarray}

114:      |g\rangle &=& |g_{1}\rangle |g_{2}\rangle \ ,\nonumber \\

115:      |e\rangle &=& |e_{1}\rangle |e_{2}\rangle \ ,\nonumber \\

116:      |s\rangle &=& \frac{1}{\sqrt{2}}\left(|e_{1}\rangle

117:      |g_{2}\rangle +|g_{1}\rangle |e_{2}\rangle \right)  \ ,\nonumber

118:      \\

119:      |a\rangle &=& \frac{1}{\sqrt{2}}\left(|e_{1}\rangle

120:      |g_{2}\rangle -|g_{1}\rangle |e_{2}\rangle \right)

121:      \label{eq3}

122: \end{eqnarray}

123: are the collective states of the two-atom system~\cite{dic}, and

124: $|g_{i}\rangle, |e_{i}\rangle$ are the ground and excited states of

125: the $i$th atom, respectively. In the Dicke system the antisymmetric

126: state $|a\rangle$ is completely decoupled from the remaining states,

127: and then the simple three-state representation of the two-atom Dicke

128: system can be applied with the ground product

129: state $|g\rangle$, the excited product state $|e\rangle$ and the

130: maximally entangled symmetric state $|s\rangle$.

131: Since the density matrix of the system is

132: diagonal and the symmetric state $|s\rangle$ is a maximally entangled

133: state, an entanglement can be produced in the Dicke system

134: by a suitable population of the state $|s\rangle$. This

135: is exactly the situation considered by Banerjee~\cite{ban} and

136: Zhou~\etal~\cite{zsl}.

137: 

138: In this letter, we clarify the discrepancy between entanglement and

139: the spin-squeezing parameter. The parameter $\xi_{\bi{n}_{i}}$ has been

140: proposed as a simple and robust method to identify entanglement of a

141: large number of atoms, so we believe that a detailed analysis of the

142: discrepancy is of general interest.

143: We show that in the two-atom Dicke model,

144: the parameter $\xi_{\bi{n}_{i}}$ correctly predicts entanglement only

145: if the system is in the two-photon entangled states

146: which are linear superpositions of the

147: collective ground state $|g\rangle$ and the upper state $|e\rangle$,

148: but fails to predict entanglement if the system is in the entangled

149: symmetric state $|s\rangle$.

150: 

151: In order to show this more quantitatively, we start from the

152: definition of the parameter $\xi_{\bi{n}_{i}}$, which we can write

153: in terms of the density matrix elements of the system as

154: \begin{eqnarray}

155:      \xi_{\bi{n}_{i}} =2\langle \left(\Delta

156:      S_{\bi{n}_{i}}\right)^{2}\rangle =1 +\rho_{ss} -2|\rho_{eg}|

157:      \cos \theta \ ,\label{eq4}

158: \end{eqnarray}

159: where $\theta$ is the angle between $\bi{n}_{i}$ and the direction

160: of maximum squeezing. In the derivation of \eref{eq4},

161: we have used the Kitagawa and Ueda's~\cite{ku} definition

162: of $\xi_{\bi{n}_{1}}$ in which the variance $\langle \left(\Delta

163: S_{\bi{n}_{i}}\right)^{2}\rangle$, calculated in the $\bi{n}_{i}$

164: direction, is compared to the maximum spin

165: $\langle S_{\bi{n}_{k}}\rangle =N_{a}/2$ in the normal $\bi{n}_{k}$

166: direction.

167: For simplicity, we have assumed that the mean spin direction coincides

168: with the $z$ axis and calculated the variance in the $\bi{n}_{i}$ direction

169: which coincides with the $x$ axis. This is not an essential feature if

170: the system is driven by a thermal or squeezed vacuum field, since in

171: this case the mean values $\langle S_{x}\rangle$ and $\langle

172: S_{y}\rangle$ are zero for all values of the parameters

173: involved~\cite{ft}. In a more general case of a coherently driven

174: atoms, where $\langle S_{x}\rangle$ and $\langle

175: S_{y}\rangle$ are different from zero, one can adjust the angle

176: $\theta$ such that the maximum squeezing will coincide with the

177: direction of the rotated nonzero spin components.

178: 

179: We see from \eref{eq4} that the parameter $\xi_{\bi{n}_{i}}$

180: depends on the population $\rho_{ss}$ of the entangled symmetric state

181: and the two-photon coherence $\rho_{eg}$. Hence,

182: spin squeezing will be produced in the direction $\theta$ when

183: $|\rho_{eg}|>\rho_{ss}/2$. Note that the spin-squeezing parameter

184: involves the two-photon coherences with no dependence on

185: one-photon coherences. This indicates that the spin

186: squeezing can only be generated by two-photon processes.

187: Thus, the spin squeezing is inherent multi-atom

188: effect arising from the collective evolution of the Dicke system.

189: 

190: We now determine general conditions for entanglement in the two-atom

191: Dicke model using the Peres-Horodecki measure of entanglement given

192: by the quantity~\cite{per,horo}

193: \begin{eqnarray}

194:      E = {\rm max}\left(0, -2\sum_{i}\mu_{i-}\right) \ ,\label{eq5}

195: \end{eqnarray}

196: where the sum is taken over the negative eigenvalues $\mu_{i-}$

197: of the partial transposition of the density matrix $\hat{\rho}$ of the

198: system. The value $E=1$ corresponds to maximum entanglement between

199: the atoms whilst $E=0$ describes completely separated atoms.

200: 

201: Since the generation

202: of the spin squeezing is independent of the one-photon coherences, we

203: will look into conditions for entanglement which are determined by the

204: population of the collective states and the two-photon coherences.

205: Note, that in the Dicke model, $\rho_{aa}=0$. In

206: this case, the density matrix of the system in the basis $\{

207: |e_{1},e_{2}\rangle, |e_{1},g_{2}\rangle, |g_{1},e_{2}\rangle,

208: |g_{1},g_{2}\rangle \}$ can be written as

209: \begin{eqnarray}

210: \hat{\rho} &=& \left(

211: \begin{array}{cccc}

212: \rho_{ee} & 0 & 0 & \rho_{eg} \\

213: 0 & \frac{1}{2}\rho_{ss} & \frac{1}{2}\rho_{ss} & 0 \\

214: 0 & \frac{1}{2}\rho_{ss} & \frac{1}{2}\rho_{ss} & 0 \\

215: \rho_{ge} & 0 & 0 & \rho_{gg}

216: \end{array}

217: \right) \ . \label{eq6}

218: \end{eqnarray}

219: Following the Peres-Horodecki criterion for entanglement, we find that

220: the eigenvalues of the partial transposition of $\hat{\rho}$ are

221: \begin{eqnarray}

222:      \mu_{1\pm} &=& \frac{1}{2}\rho_{ss} \pm |\rho_{eg}| \ ,\nonumber

223:      \\

224:      \mu_{2\pm} &=& \frac{1}{2}\{\left(\rho_{ee}+\rho_{gg}\right) \pm

225:      \left[\left(\rho_{ee}-\rho_{gg}\right)^{2}

226:      +\rho_{ss}^{2}\right]^{\frac{1}{2}}\} \ .\label{eq7}

227: \end{eqnarray}

228: It is obvious that $\mu_{1+}$ and $\mu_{2+}$ are always positive. The

229: eigenvalues $\mu_{1-}$ and $\mu_{2-}$ become negative if and only if

230: \begin{eqnarray}

231:      |\rho_{eg}| >\frac{1}{2}\rho_{ss} \ ,\label{eq8}

232: \end{eqnarray}

233: or

234: \begin{eqnarray}

235:      \rho_{ss} > 2\sqrt{\rho_{ee}\rho_{gg}} \ .\label{eq9}

236: \end{eqnarray}

237: We are now in a position to understand quantitatively the discrepancy

238: between entanglement and the spin squeezing parameter.

239: It is seen that there are {\it two} criteria for entanglement in the

240: two-atom Dicke model. The first criterion, \Eref{eq8}, is

241: associated with the two-photon coherence and population of the

242: symmetric state. The second criterion, \Eref{eq9}, is associated

243: only with the populations of the collective states. It is evident that

244: the criterion \eref{eq8} overlaps with the criterion for spin squeezing,

245: see \Eref{eq4}. Therefore, in the absence of the two-photon

246: coherences, the two-atom system can still be entangled, in accordance

247: with the criterion \eref{eq9}, but cannot exhibit spin-squeezing,

248: which is associated with the criterion \eref{eq8}. This explicitly

249: identifies the source of the discrepancy found by Banerjee~\cite{ban}

250: and Zhou~\etal~\cite{zsl} and explains why the two-atom Dicke system

251: can be entangled without spin-squeezing.

252: 

253: In the situations where the criterion \eref{eq8} is satisfied, there

254: are entangled states generated which can be found by the

255: diagonalization of the density matrix \eref{eq6}. We find that the

256: diagonalization leads to eigenstates

257: \begin{eqnarray}

258: |\Psi_{+}\rangle &=& \left[\left(\Pi_{+}-\rho_{ee}\right)|g\rangle

259: +\rho_{eg}|e\rangle

260: \right]/\left[\left(\Pi_{+}-\rho_{ee}\right)^{2}

261: +\left|\rho_{eg}\right|^{2}\right]^{\frac{1}{2}} \ ,\nonumber \\

262: |\Psi_{-}\rangle &=& \left[\rho_{ge}|g\rangle +

263: \left(\Pi_{-}-\rho_{gg}\right)|e\rangle

264: \right]/\left[\left(\Pi_{-}-\rho_{gg}\right)^{2}

265: +\left|\rho_{eg}\right|^{2}\right]^{\frac{1}{2}} \ ,\nonumber \\

266: |\Psi_{s}\rangle &=& |s\rangle \ ,\nonumber \\

267: |\Psi_{a}\rangle &=& |a\rangle \ ,\label{eq10}

268: \end{eqnarray}

269: with the diagonal probabilities

270: \begin{eqnarray}

271: \Pi_{+} &=& \frac{1}{2}\left(\rho_{gg}+\rho_{ee}\right)

272: +\frac{1}{2}\left[\left(\rho_{gg}-\rho_{ee}\right)^{2}

273: +4\left|\rho_{eg}\right|^{2}\right]^{\frac{1}{2}} \ ,\nonumber \\

274: \Pi_{-} &=& \frac{1}{2}\left(\rho_{gg}+\rho_{ee}\right)

275: -\frac{1}{2}\left[\left(\rho_{gg}-\rho_{ee}\right)^{2}

276: +4\left|\rho_{eg}\right|^{2}\right]^{\frac{1}{2}} \ ,\nonumber \\

277: \Pi_{s} &=& \rho_{ss} \ ,\nonumber \\

278: \Pi_{a} &=& 0 \ .\label{eq11}

279: \end{eqnarray}

280: It is evident from \eref{eq10}, that in the presence of the

281: two-photon coherence, the system evolves

282: into entangled states which are linear superpositions of the

283: collective ground state $|g\rangle$ and the upper state $|e\rangle$.

284: The entangled symmetric state remains unchanged in the presence of

285: two-photon processes. Thus, spin squeezing and entanglement created

286: by the two-photon coherences are both associated with the two-photon

287: entangled states $|\Psi_{\pm}\rangle$.

288: 

289: As an example to illustrate our findings, consider the two-atom Dicke

290: system driven by a broadband squeezed vacuum field. In the steady-state,

291: nonzero matrix elements are~\cite{ft,pk90}

292: \begin{eqnarray}

293: \rho_{ee} &=& \frac{N^{2}\left(2N+1\right)

294: -\left(2N-1\right)|M|^{2}}{\left(2N+1\right)

295: \left(3N^{2} +3N +1 -3|M|^{2}\right)} \ ,\nonumber \\

296: \rho_{ss} &=& \frac{N\left(N+1\right)

297: -|M|^{2}}{3N^{2} +3N +1 -3|M|^{2}} \ ,\nonumber \\

298: |\rho_{eg}| &=& \frac{|M|}{\left(2N+1\right)

299: \left(3N^{2} +3N +1 -3|M|^{2}\right)} \ ,\label{eq12}

300: \end{eqnarray}

301: where $N$ is the intensity of the squeezed field and $M$ is the

302: two-photon correlation function~\cite{dfs}.

303: 

304: For the interaction of the system with a thermal field, $M=0$, and

305: then using \eref{eq12} it is straightforward to prove that

306: both criteria \eref{eq8} and \eref{eq9} are not satisfied for all

307: values of $N$. Moreover, the inequality $\xi_{\bi{n}_{x}}>1$ always

308: holds indicating that both entanglement and spin-squeezing are not

309: present in the steady-state two-atom Dicke system driven by a thermal

310: field.

311: 

312: The situation is different when the system is driven by

313: a classical squeezed field with the maximal two-photon

314: correlations $M=N$. In this case the inequality

315: $\rho_{ss}<2\sqrt{\rho_{ee}\rho_{gg}}$ always holds in contradiction

316: to \eref{eq9}. However, we find that the inequality \eref{eq8}

317: can be satisfied as $|\rho_{eg}|\neq 0$. According to \eref{eq4}

318: and \eref{eq8}, the criterion for both entanglement and spin

319: squeezing can be determined by positive values of a parameter

320: \begin{eqnarray}

321:      2|\rho_{eg}|-\rho_{ss} = \frac{N\left(1-2N\right)}

322:      {\left(2N+1\right)\left(3N +1\right)} \ .\label{eq13}

323: \end{eqnarray}

324: \Eref{eq13} shows that the system driven by the

325: classical squeezed field will exhibit entanglement and

326: spin-squeezing when $N<1/2$. We illustrate these features in

327: \Fref{fig1}(a), where we plot the entanglement measure $E$ and

328: the squeezing parameter $\xi_{\bi{n}_{x}}$

329: as a function of the intensity $N$. The figure clearly demonstrates

330: that with the condition \eref{eq8}, the squeezing parameter

331: correctly predicts entanglement induced by the two-photon coherences.

332: We should note here that the steady-state of the system driven by a

333: classical squeezed field is a mixed state. Thus, the squeezing

334: parameter correctly predicts entanglement in a mixed state if the

335: entanglement is generated by the two-photon coherences.

336: %

337: \begin{figure}[h]

338: \begin{center}

339: \includegraphics[width=10cm]{job.eps}

340: %\epsfbox{rid.eps}

341: \end{center}

342: \caption{\label{fig1} Entanglement measure $E$ (solid line) and the

343: squeezing parameter $\xi_{\bi{n}_{x}}$ (dashed line) as a function of

344: the intensity $N$ for (a) classical squeezed field with $|M|=N$, and

345: (b) quantum squeezed field with $|M|=\sqrt{N(N+1)}$. }

346: \end{figure}

347: 

348: When the system is driven by a quantum squeezed field with perfect

349: correlations $|M|^{2}=N(N+1)$, the populations of the diagonal

350: states~(\ref{eq10}) are profoundly affected by the presence of the

351: strong two-photon correlations $M$ such that the steady-state of the

352: system is a pure state~\cite{ft,pk90,pa}.

353: Since $\rho_{ss}=0$, the criterion \eref{eq9} is not

354: satisfied, and therefore entanglement is determined solely by

355: the criterion \eref{eq8} which is always satisfied as $|\rho_{eg}|>0$.

356: Since the inequality $|\rho_{eg}|>0$ always holds, the system

357: exhibits entanglement and spin-squeezing for all $N$.

358: This feature is seen in \Fref{fig1}(b), where we show the

359: entanglement measure $E$ and squeezing parameter $\xi_{n_{x}}$ for the

360: quantum squeezed field. We see that the entanglement and spin

361: squeezing are present for all $N$. The entanglement and spin squeezing

362: increase with increasing $N$ and attain their maximal values, $E=1$

363: and $\xi_{\bi{n}_{x}}=0$, for large $N$.

364: 

365: As we have mentioned above, the steady-state of the system driven by

366: the quantum squeezed field is a pure state.

367: We find from \eref{eq10} that the pure state is the entangled

368: state given by~\cite{pk90,pa}

369: \begin{equation}

370: |\Psi_{+}\rangle = \frac{1}{\sqrt{2N+1}}

371: \left[\sqrt{N+1}|g\rangle

372: +\sqrt{N}|e\rangle\right] \ .\label{eq14}

373: \end{equation}

374: The pure state is a non-maximally entangled state, and reduces to a

375: maximally entangled state for $N\gg 1$.

376: 

377: In summary, we have clarified the discrepancy between the entanglement

378: and spin-squeezing parameter recently reported by Banerjee~\cite{ban}

379: and Zhou~\etal~\cite{zsl}. We have found that there are two criteria

380: for entanglement in the two-atom Dicke system, one associated with

381: the two-photon coherences and population of the symmetric state, and

382: the other associated with populations of all the collective states.

383: We have shown that the criterion for spin squeezing overlaps with

384: only one of the two criteria for entanglement, that involving the

385: two-photon coherences. Therefore, if entanglement is produced by the

386: other criterion, one obtains entanglement without spin squeezing.

387: Thus, our calculations demonstrate that the spin-squeezing parameter

388: correctly predicts entanglement in the two-atom Dicke system if the

389: entanglement is created by the two-photon coherences.

390: Moreover, the current study provides a clear physical picture of

391: different processes which can create entanglement in the two-atom

392: Dicke system.

393: 

394: 

395: \ack

396: This research has been supported in

397: part by Malaysia IRPA research grant 09-02-08-0203-EA002.

398: ZF would like to thank the International Islamic University

399: Malaysia for hospitality and financial support.

400: 

401: \section*{References}

402: \begin{thebibliography}{99}

403: 

404: \bibitem{si} Nielsen M A and Chuang I J 2000 {\it Quantum Computation

405: and Quantum Information} (Cambridge: Cambridge University Press)

406: 

407: \bibitem{ben} Bennett C H and DiVincenzo 2000 {\it Nature} {\bf 404} 247

408: 

409: \bibitem{ft} Ficek Z and Tana\'s R 2002 {\it Phys. Rep.} {\bf 372} 369

410: 

411: \bibitem{woo} Wootters W K 1998 \PRL {\bf 80} 2245;

412: Hill S and Wootters W K 1997 \PRL {\bf 78} 5022

413: 

414: \bibitem{per} Peres A 1996 \PRL {\bf 77} 1413

415: 

416: \bibitem{horo} Horodecki P 1997 \PL A {\bf 232} 333

417: 

418: \bibitem{sdcz} S\o rensen A S, Duan L M, Cirac J I and Zoller P 2001

419: {\it Nature} {\bf 409} 63

420: 

421: \bibitem{ku} Kitagawa M and Ueda M 1993 \PR A {\bf 47} 5138

422: 

423: \bibitem{ibbg} Itano W M, Bergquist J C, Bollinger J J, Gilligan J M,

424: Heinzen D J, More F L, Raizen M G and Wineland D J 1993 \PR A

425: {\bf 47} 3554

426: 

427: \bibitem{wbi} Wineland D J, Bollinger J J, Itano W M and Heinzen D J

428: 1994 \PR A {\bf 50} 67

429: 

430: \bibitem{sor}  Hald J, S\o rensen J L, Schori C and Polzik E S 1999

431: \PRL {\bf 83} 1319

432: 

433: \bibitem{ban} Banerjee A 2001 {\it Preprint} quant-ph/0110032

434: 

435: \bibitem{zsl} Zhou L, Song H S and Li C 2002 \JOB {\bf 4} 425

436: 

437: \bibitem{dic} Dicke R H 1954 \PR {\bf 93} 99

438: 

439: \bibitem{pk90} Palma G M and Knight P L 1989 \PR A {\bf 39}

440: 1962

441: 

442: \bibitem{dfs} Dalton B J, Ficek Z and Swain S 1999 {\it J. Mod. Opt.}

443: {\bf 46} 379

444: 

445: \bibitem{pa} Agarwal G S and Puri R R 1990 \PR A {\bf 41} 3782

446: 

447: 

448: \end{thebibliography}

449: 

450: \end{document}

451: