quant-ph0206139/pap.tex
1: %\documentstyle[aps,prl,psfig,epsf,epsfig]{revtex}
2: %\setlength{\topmargin}{0pt}
3: \documentstyle[twocolumn,aps,prl,epsf,epsfig]{revtex}
4: \begin{document}
5: \draft
6: \title{Entangling Two Bose-Einstein Condensates by Stimulated Bragg Scattering}
7: \author{B. Deb and G.S. Agarwal}
8: \address{Physical Research Laboratory, Navrangpura, Ahmedabad-380009, India}
9: \date{\today}
10: \maketitle
11: \begin{abstract}
12: We propose an experiment for entangling  two spatially separated Bose-Einstein
13: condensates by  Bragg scattering of light.  When Bragg scattering in two
14: condensates is stimulated by a common  probe, the resulting quasiparticles in
15: the two condensates get entangled due to quantum communication between the
16: condensates via probe beam. The entanglement is shown to be significant and
17: occurs in both number and quadrature phase variables. We present two methods of
18: detecting the generated entanglement.
19: \end{abstract}
20: \pacs{PACS numbers: 03.75.Fi,03.65.Ud,42.50.Dv}
21: Inseparability of quantum states of two or more subsystems is the most
22: significant feature of quantum mechanics. Apparently puzzling, yet most
23: profound, first formulated as a paradox \cite{epr}, this inseparability known
24: as quantum entanglement lies at the very heart of nonclassical physics.
25: Further, as a basic resource for quantum information processing, it has become
26: a focal theme of research in modern physics and many issues in the foundations
27: of quantum mechanics. Generation and manipulation of entanglement is,
28: therefore, of prime  interest.  Bose-Einstein condensates (BEC) \cite{bec} of
29: weakly interacting atomic gases seem to be suitable macroscopic objects for
30: producing many-particle  entanglement  \cite{sorensen}.  A BEC has intrinsic
31: entanglement character  due to reduced quantum fluctuations in momentum 
32: space.  For instance, in the condensate ground state,  a pair of mutually
33: opposite momentum modes is maximally entangled in atomic number variables
34: \cite{deb}. 
35: 
36: Stimuated resonant Bragg scattering of light by a condensate generates
37: quasiparticles \cite{quasi},  predominantly in two momentum side-modes
38: $\mathbf{q}$ and $-\mathbf{q}$, where $\mathbf{q}$ is the momentum transfered
39: from light fields to the atoms. Momentum side-modes are the excited states of a
40: BEC, atoms in such a state collectively behave as quasiparticles. Bragg
41: spectroscopy \cite{bragg} with coherent or classical light produces   coherent
42: states of the quasiparticles in a BEC. When these quasiparticles are projected
43: into particle domain, that is, into the Bogoliubov-transformed momentum modes
44: \cite{bogoliubov}, they form two-mode squeezed as well as entangled state
45: \cite{deb}.  Bragg spectroscopy  with nonclassical light can  generate 
46: tripartite entanlement \cite{deb} in a condensate.  In addition to atom number
47: and phase variables, spin degree-of-freedom of a spinor BEC \cite{spinor} can 
48: be useful in describing entanglement in spin variables. Thus, BECs offer a
49: fertile ground  for studying different aspects of entanglement.  Apart from
50: BECs, multi-atom entanglement in other  macroscopic systems has been realized 
51: \cite{spinexp} on the basis of collective spin squeezing \cite{spin1,spin2}. 
52: Further the  entanglement in collective spin variables of two ensembles  of
53: gaseous Cs atoms  has been experimentally demonstrated \cite{sasha} .
54: Continuous variables like the quadratures of a field  mode (which are 
55: analogous to  position and momentum) have also been employed \cite{kimble} in
56: entanglement studies. 
57: 
58: We here propose a scheme for producing quantum entanglement between two
59: spatially separated BECs of a weakly interacting  atomic gas.  The entanglement
60: we consider is in  quasiparticles  of BECs.    The  proposed experiment is
61: schematically shown in Fig.1. The condensates A and B  are illuminated  by pump
62: lasers L1 and L2, respectively. A single stimulating probe laser L3 passes
63: through both the condensates. All these three lasers are detuned far off the
64: resonance of an electronic excited state of the atoms. The frequencies and the
65: directions of propagation of these lasers  are so chosen  such that Bragg
66: resonance (phase matching) conditions of scattering  in both  the condensates
67: are fulfilled. The Hamiltonian of the system is $H=H_{A}+H_{B}
68: +H_{F}+H_{AF}+H_{BF}$. Retaining the dominant momentum side-modes $\mathbf{q}$
69: and -$\mathbf{q}$ only under Bragg resonance condition, in the Bogoliubov
70: approximation \cite{bogoliubov}, $H_A = \hbar \omega_{q}^B
71: \left(\hat{\alpha}_{{\mathbf{q}}}^{\dagger} \hat{\alpha}_{{\mathbf{q}}}
72: +\hat{\alpha}_{-{\mathbf{q}}}^{\dagger} \hat{\alpha}_{-{\mathbf{q}}}\right)$,
73: where $\hat{\alpha}_{\mathbf{q}}$ represents quasi-particle with mometum
74: $\mathbf{q}$, and  $\omega_{q}^B=\left[(\omega_q + \frac{\mu}{\hbar})^2 -
75: (\frac{\mu}{\hbar})^2\right]^{1/2}$ is the frequency  of Bogoliubov's 
76: quasi-particle \cite{bogoliubov}.  Here $\omega_q = \frac{\hbar^2q^2}{2m}$,
77: and  $\mu = \frac{\hbar^2\xi^{-2}}{2m}$ is the chemical potential with $\xi =
78: (8\pi n_{0}a_{s})^{-1/2}$ being the healing length. Similarly, $H_B = \hbar
79: \omega_{q}^B \left(\hat{\beta}_{{\mathbf{q}}}^{\dagger}
80: \hat{\beta}_{{\mathbf{q}}} +\hat{\beta}_{-{\mathbf{q}}}^{\dagger}
81: \hat{\beta}_{-{\mathbf{q}}}\right)$, 
82: \begin{figure}
83: \psfig{file=dia1.eps,width=3.25in}
84: \caption{ The scheme for creation of entanglement. A and B are two
85: condensates, L1 and L2 are pump lasers, L3 is a common entangling  probe laser.
86: Both the pumps have same wave vector ${\mathbf{k}}_1$, probe's wave vector is
87: ${\mathbf{k}}_2$. The probe is red-detuned from the pumps. The lasers are in
88: Bragg resonance with a particular  momentum mode $\mathbf{q}$ of both the
89: condensates.}
90: \label{fig1}
91: \end{figure}
92: wih $\hat{\beta}_{\mathbf{q}}$ being the quasi-particle operator of the
93: condensate B.  The pumps are treated classically. Let  $\hat{c}$ represents the
94: common probe field mode, then the field Hamiltonian $H_F = -\hbar \delta
95: \hat{c}^{\dagger}\hat{c}$, where  $\delta = \omega_1 - \omega_2$ is the
96: pump-probe detuning.   The quasiparticle operators $\hat{\alpha}(\hat{\beta})$
97: are related to the particle operators  $\hat{a}(\hat{b})$ by Bogoliubov's
98: transformation : $\hat{a}_{{\mathbf{q}}}=
99: u_{q}\hat{\alpha}_{\mathbf{q}}-v_{q}\hat{\alpha}_{-\mathbf{q}}^{\dagger}$,
100: where $v_{q} = (u_{q}^{2}-1)^{1/2} = [\frac{1}{2}(\frac{\omega_q +
101: \mu/\hbar}{\omega_{q}^B}-1)]^{1/2}$.   The atom-field ineteraction Hamiltonian
102: for condensate A is
103: \begin{eqnarray}
104: H_{AF} = \hbar\eta_{A}
105: \hat{c}^{\dagger}
106: (\hat{\alpha}_{\mathbf{q}}^{\dagger} + \hat{\alpha}_{-{\mathbf{q}}})
107: +{\mathrm H.c.}
108: \end{eqnarray}
109: where $\eta_A = \sqrt{N_A}\Omega_A f_q$ is the effecive atom-field coupling
110: constant. Here $N_A$ is the number of atoms in condensate A , $\Omega_A$ is the
111: two-photon Rabi frequency of an atom in A and $f_q = u_q - v_q$. $H_{BF}$ is
112: given by the similar expression as $H_{AF}$ with subscript $A$ replaced by $B$
113: and $\alpha$ replaced by $\beta$. The Heisenberg equations of motion are
114: \begin{eqnarray}
115: \dot{\hat{\alpha}}_{\mathbf{q}} = -i\omega_q\hat{\alpha}_{\mathbf{q}} 
116: -i\eta_A \hat{c}^{\dagger}
117: \end{eqnarray}
118: \begin{equation}
119: \dot{\hat{\alpha}}_{-\mathbf{q}}^{\dagger} =
120: i\omega_q\hat{\alpha}_{-\mathbf{q}}^{\dagger} + i\eta_{A}\hat{c}^{\dagger}
121: \end{equation}
122: \begin{eqnarray}
123: \dot{\hat{c}}^{\dagger} = -i\delta \hat{c}^{\dagger} + i[\eta_A
124: (\hat{\alpha}_{\mathbf{q}} +
125: \hat{\alpha}_{-\mathbf{q}}^{\dagger}) + \eta_B (\hat{\beta}_{\mathbf{q}} +
126: \hat{\beta}_{-\mathbf{q}}^{\dagger})]
127: \end{eqnarray}
128: The Heisenberg equations of $\hat{\beta}_{\mathbf{q}}$ and
129: $\hat{\beta}_{-\mathbf{q}}^{\dagger}$ are similar to  those of $\hat{\alpha}$,
130: but $\hat{\alpha}$ and $\eta_A$ should be replaced by $\hat{\beta}$ and
131: $\eta_B$, respectively.
132: 
133: We next discuss how to quantify entanglement between two  BECs. If the
134: entanglement occurs in number operators of the quasiparticle  modes 1 and 2, 
135: then it can be quantified by the parameter \cite{spin1,deb}
136: \begin{equation}
137: \xi_n(1,2) = \langle[\Delta(\hat{n}_{1}-\hat{n}_{2})]^2\rangle/
138: (\langle\hat{n}_{1}\rangle + \langle\hat{n}_{2}\rangle).
139: \end{equation}
140: If $\xi_n < 1$, then the two modes are entangled.  If the entanglement is
141: described by two noncommuting Gaussian operators $\hat{X}$ and $\hat{P}$ which
142: are analogous to position and momentum variables, then  the entanglement 
143: parameter is defined by \cite{xip}
144: \begin{equation}
145: \xi_p(1,2) = \frac{1}{2}[\langle[\Delta(X_{1}+X_{2})]^2\rangle
146: +\langle[\Delta(P_{1}-P_{2})]^2\rangle]
147: \end{equation}
148: The two modes are entangled in quadrature phase, when  $\xi_p<1$.
149: 
150: For numerical illustration, we consider two homogeneous identical  Na
151: condensates.  We here enlist the important results: 1) If the modes
152: ${\mathbf{q}}_1$ of A and ${\mathbf{q}}_2$ of B are in Bragg-resonance with the
153: the respective Bragg pulses, and if the effective coupling of B ($\eta_B$)  is
154: stronger than that of A, then entanglement arises between ${\mathbf{q}}_1$ of A
155: and $-{\mathbf{q}}_2$ of B only,  other pairs of modes are immune to any
156: entanglement. In fig.2,  we display entanglement parameters between these two
157: chosen modes  as a function of time. We set
158: ${\mathbf{q}}_1={\mathbf{q}}_2=\mathbf{q}$. The effective coupling can be
159: \begin{figure}
160: \psfig{file=fig2.eps,width=3.25in}
161: \caption{ Entanglement parameters $\xi_n$ and $\xi_p$ between 
162: ${\mathbf{q}}$-mode of A and -$\mathbf{q}$-mode of B as a function of time in
163: $\mu$s.  For both the condensates, $\omega_q^B = 0.21$ MHz, $q = 2 \xi^{-1}$
164: and  $\delta=0.17$ MHz. The coupling constants $\eta_A =1.62$ MHz, $\eta_B = 
165: 1.25 \eta_1$. Both the condensates are initially in the ground states  (vacuum
166: of quasi-particle operators), and the common probe field is in coherent state
167: with average number of photons equal to 10.}
168: \end{figure}
169: made different either by using pump lights of different intensities or taking
170: different atom numbers for the two otherwise identical condensates. 2) For
171: equal couplings, there is no entanglement between any pair of modes.  Fig.3 
172: shows the variation of entanglement parameters as a function of the ratio of
173: the two coupling constants at a fixed time.  3) We find entanglement both in
174: quasiparticle (phonon) modes ($\hat{\alpha},\hat{\beta}$), and in the
175: Bogoliubov transformed modes of quasiparticles  which we call particle or 
176: atomic modes ($\hat{a},\hat{b}$). However, in  atomic modes, entanglement is
177: weaker than that in quasiparticle modes.  It is worth mentioning that in a
178: single condensate, as shown in Ref.\cite{deb}, coherent light scattering can
179: generate entanglement only in atomic modes, and not in phonon modes. In
180: contrast, one can generate {\it entanglement in phonon modes in two
181: condensates} by coherent light scattering.    The light scattering events
182: occuring at  A and B are not independent, since a quantum communication has
183: been set between the generated quasiparticles in A and B via the common probe. 
184: Had we treated the common {\it probe classically},  then the Hamiltonian 
185: (Eq.(1))  would have been  linear in the atomic operators. A Hamiltonian linear
186: in Bosonic opeartors  can not generate nonclassical correlation. Therefore, the
187: {\it probe} must be treated {\it quantum mechanically}. The probe carries with
188: it quantum fluctuations  of one condensate and transfers a part of it to the
189: other leading to the entanglement  between the two  condensates.
190: 
191: To explain the results further, we here resort to an approximate analysis. Let
192: us suppose, $\omega_q <\!<\eta_{A(B)}$ and $\delta <\!<\eta_{A(B)}$, then we
193: can neglect the diagonal terms proportional to $\omega_q$ and $\delta$ in the
194: Hamiltonian. For equal 
195: \begin{figure}
196: \psfig{file=fig3.eps,width=3.25in}
197: \caption{  The entanglement parameters $\xi_n$ (solid) and $\xi_p$ (dashed) as
198: a function of $\eta_B/\eta_A $ for a fixed  time $t=0.75$ $\mu$s. The other
199: parameters are the same as in Fig.2.}
200: \end{figure}
201: coupling ($\eta_A = \eta_B$), from  Heisenberg equation of motions,
202: it then follows that  $\hat{\alpha}_{\mathbf{q}}(t) +
203: \hat{\beta}_{-\mathbf{q}}^{\dagger}(t) = \hat{\alpha}_{\mathbf{q}}(0) +
204: \hat{\beta}_{-\mathbf{q}}^{\dagger}(0)$, that is, the superposition operator
205: $\Sigma = \hat{\alpha}_{\mathbf{q}} + \hat{\beta}_{-\mathbf{q}}^{\dagger}$
206: becomes a constant of motion. Let us write the quadratures $X_A =
207: \frac{1}{\sqrt{2}}(\hat{\alpha}_{\mathbf{q}}  +
208: \hat{\alpha}_{\mathbf{q}}^{\dagger})$, $P_{A} = 
209: -i\frac{1}{\sqrt{2}}(\hat{\alpha}_{\mathbf{q}} - 
210: \hat{\alpha}_{\mathbf{q}}^{\dagger})$, and similarly for $X_B$ and $P_B$. Then
211: it can be shown that $\xi_p = \frac{1}{4}[\langle[\Delta(\Sigma +
212: \Sigma^{\dagger})]^2\rangle -
213: \langle[\Delta(\Sigma-\Sigma^{\dagger})]^2\rangle]$. For equal coupling and the
214: initial states being  in vacuum or in coherent states, one obtains $\xi_p =1$,
215: that is, the two modes are  unentangled. Let us then consider the case  of
216: different couplings, for short times characterized by $\eta_At <\!<1$, and 
217: $\eta_Bt <\!<1$, we obtain perturbative solutions of
218: $\hat{\alpha}_{{\mathbf{q}}}(t)$ and $\hat{\beta}_{-{\mathbf{q}}}(t)$  upto the
219: second order in time. Using these solutions, we calculate $\xi_p = 1 -
220: \eta_A\eta_Bt^2(1 - \frac{\eta_A}{\eta_B})$, which is less than unity (the two
221: modes are entangled in quadrature variables) if $\eta_A\eta_Bt^2(1 -
222: \frac{\eta_A}{\eta_B}) > 0$ which is only possible if $\eta_A \ne \eta_B$ and
223: $\eta_A < \eta_B$.  Similarly, we can prove that for -$\mathbf{q}$
224: (off-resonant) of A and $\mathbf{q}$ (resonant) of B, $\xi_p = 1 +
225: (\eta_Bt)^2(1-\frac{\eta_A}{\eta_B})$ which is always greter than unity for
226: $\eta_A < \eta_B$. For the same  resonant $\mathbf{q}$-mode  of A and B, $\xi_p
227: = 1 + \eta_Bt^2/2 + (\eta_A^2 + \eta_B^2)^2t^4/4$, which is always greater than
228: unity.  In the same way, we can show that, for the remaining mode-pair 
229: ($-{\mathbf{q}}, -{\mathbf{q}}$),  $\xi_p$ is also  greater than unity.
230: 
231: Next, we prove that, to generate entanglement in number variables ($\xi_n$),
232: the two coupling parameters should also be different. Substituting $\hat{n}_1 =
233: \hat{\alpha}_{\mathbf{q}}^{\dagger}\hat{\alpha}_{\mathbf{q}}$ and $\hat{n}_2 =
234: \hat{\beta}_{-\mathbf{q}}^{\dagger}\hat{\beta}_{-\mathbf{q}}$ in Eq.(5)  and
235: using the pertubative solutions  we can express $\xi_n= 1-
236: R/(\langle\hat{n}_{1}\rangle + \langle\hat{n}_{2}\rangle)$ where
237: \begin{eqnarray}
238: R = 8\eta_A^2t^4[\eta_B^2 - 2\eta_A^2 + 4 n_p (\eta_B^2 -
239: \eta_A^2)]
240: \end{eqnarray}
241: where $n_p$ is the initial number of photons in the coherent probe beam. Now,
242: $\xi_n < 1$ implies that $R>0$ which amounts to $(\eta_B/\eta_A)^2 > 1+
243: 1/(1+4n_p)$, that is, $\eta_B > \eta_A$. On the other hand, if $\eta_B\le
244: \eta_A$, $\xi_n > 1$.  We also carry out  an aletrnative analysis to check 
245: whether the two resonant modes $\mathbf{q}$ of A and B exhibit any entanglement
246: in other parameter regimes. By neglecting the off-resonant mode $-\mathbf{q}$
247: in both the condensates and keeping only the resonant mode, it can be
248: analytically proved that $\xi_p({\mathbf{q}},{\mathbf{q}}) = 1 + \sinh^2(\eta
249: t)$ and $\xi_n({\mathbf{q}},{\mathbf{q}}) = 1 +
250: (1+\frac{n_p}{n_p+1})\frac{(\eta_A^2 - \eta_B^2)^2}{(\eta_A^2 + \eta_B^2)^2}
251: \sinh^2(\eta t)$, that is, both the  parameters $\xi_p$ and $\xi_n$ are always
252: greater than unity. Here $\eta = \sqrt{\eta_A^2 + \eta_B^2}$. 
253: 
254: We next show how a set up as shown in Fig.4 can be utilized  to verify the
255: generated entanglement.  After the process of generation of entanglement, the
256: duration of which can be typically on the order of 1 to 100 $\mu$s,   is over,
257: the lasers L1, L2 and L3 are switched off. Two different pairs of verifying
258: pump-probe Bragg pulses are applied to the condensates, as described in the
259: caption of Fig.4. The two probes should be derived from a common source. The
260: modes $\mathbf{q}$ of A and $-\mathbf{q}$ of B are in Bragg resonance with the
261: respective Bragg pulses. The effective field-condensate couplings for both the
262: condensates are very small compared to the Bogoliubov frequency $\omega_q^B$.
263: Let $\hat{c}_{probe,A}$ and $\hat{c}_{probe,B}$ denote the verifying probe
264: field modes for the condensates A and B, respectively. By neglecting the
265: off-resonant terms 
266: $\hat{\alpha}_{2{\mathbf{q}}}^{\dagger}\hat{\alpha}_{{\mathbf{q}}}$ and
267: $\hat{\alpha}_{-2{\mathbf{q}}}^{\dagger}\hat{\alpha}_{-{\mathbf{q}}}$ in the
268: Hamiltonian, the time evolution of the output probe modes, in a frame rotating
269: with pump-probe detuning $\delta$, can be written as
270:    \begin{eqnarray*}
271:   \hat{c}_{probe,A}^{(out)} &\simeq&  \hat{c}_{probe,A}^{(in)} +
272:   \frac{\eta_A}{\delta - \omega_q^B}
273:   \left(\exp[i(\delta - \omega_q^B)t]-1\right)\hat{\alpha}_{\mathbf{q}}^{\dagger} \nonumber \\
274:   &+&
275:   \frac{\eta_1}{\delta + \omega_q^B}
276:   \left(\exp[i(\delta + \omega_q^B)t]-1\right)\hat{\alpha}_{-\mathbf{q}}
277:   \end{eqnarray*}
278: \begin{figure}
279: \psfig{file=dia2.eps,width=3.25in}
280: \caption{  The scheme   for  verification of the entanglement.  Apart from pump
281: lasers, two extremely weak verifying probes of the same frequency as that of
282: entangling probe - one each for each condensate - are switched on . The  probes
283: acting on A and B have momentum ${\mathbf{k}}_2$ and  $-{\mathbf{k}}_2$,
284: respectively. The momentum of pump laser for A is  ${\mathbf{k}}_1$, while that
285: for B is $-{\mathbf{k}}_1$. Thus, the mode $\mathbf{q}$ and $-\mathbf{q}$  are
286: in Bragg resonance with A  and B, respectively.  The output from the two phase
287: sensitive detectors (PSD) can be integrated by an integrator followed by
288: another  phase-sensitive detection of the integrated signal.}
289: \label{fig4}
290: \end{figure}
291:  $\hat{c}_{probe,B}^{(out)}$ is given by similar expression as above with
292:  $\alpha$ replaced  by $\beta$; and the subscripts  A and $\mathbf{q}$ replaced
293:  B and $-\mathbf{q}$, respectively. Thus  the output probes have oscillating
294:  parts at frequency $\delta - \omega_q^B$ proportional to the quasiparticle
295:  amplitudes $\alpha_{\mathbf{q}}$ and $\beta_{-\mathbf{q}}$, and at frequency
296:  $\delta + \omega_q^B$ proportional to the amplitues $\alpha_{-\mathbf{q}}$ and
297:  $\beta_{\mathbf{q}}$. Therefore, phase-senitive measurements of the spectral
298:  components of the output probe beams corresponding to these frequencies would
299:  provide  measures of the quasiparticle operators. The output from both the
300:  PSDs can be integrated  by an integrator and the integrated signal can also be
301:  measured by another PSD. By repeating the same measurements under identical
302:  conditions we could  calculate the number variances or correlation functions
303:  of interest, which can be employed to calculate the entanglement parameter in
304:  number operators, i.e., $\xi_n$. For calculating entanglement parameters in
305:  quadrature phase variables, both the output probe beams coming  from A and B,
306:  can be mixed via a beam splitter to form the superposition operators $\Sigma$
307:  which can be measured by a similar phase sensitive detection scheme (not shown
308:  in Fig.4). 
309:   
310:   Following the recent experiment of Ketterle's group \cite{ketter}, we also
311:   suggest that the quasiparticles can be detected by imparting a large momentum
312:   to them with additional Bragg pulses. Alternatively, in the large momentum
313:   regime ($q >\!> \xi^{-1}$), the Bragg-scattered atoms which essentially
314:   behave as free particles ($\omega_q^B \propto q^2$), can be outcoupled by
315:   swithcing off the trap. Since entanglement is between two opposite momentum
316:   states, by proper geometric arrangement, the two moving entangled atomic
317:   ensembles can be made to collide and interfere. From the interference pattern
318:   obtainable via absorption imaging, the atomic number fluctuations can be
319:   deduced using the theoretical model used in Ref.\cite{orzel}, and thus
320:   entanglement parameter in number variables can be calculated. 
321: 
322:    In conclusion, we have theoretically demonstrated how light scattering 
323:    leads to quantum entanglement between two Bose-Einstein condensates. We find
324:    that  the quasiparticle or phonon as well as free-particle momentum modes of
325:    two condensates can be  entangled.  The quasiparticle state can be
326:    sufficiently long-lived due to weak nature of interatomic interaction and
327:    the constraints imposed by  momentum conservations. The generated
328:    entanglement may be  useful in quantum communication using coherent light
329:    \cite{duan}. We have particularly focused on the conditions under which the
330:    entanglement can be obtained. We have also suggested how quasiparticles 
331:    could be studied by using Bragg scattering of far off resonant fields.
332: 
333: \begin{references}
334: 
335: \bibitem{epr} A. Einstein, B. Podolosky, and N. Rosen, Phys.Rev. {\bf 47}, 777
336: (1935).
337: 
338: \bibitem{bec} For a recent
339: review on BEC in atomic gases, see {\it Coherent atomic matter
340: waves}, Les Houches, Session LXXII, edtited by R. Kaiser, C. Westbrook
341: and F. David (Springer, 2000).
342: 
343: \bibitem{sorensen} A. Sorensen {\it et al.},  Nature {\bf 409}, 63 (2001); 
344: L.M. Duan {\it et al.},  Phys. Rev. Lett. {\bf 85}, 3991 (2000); H. Pu. and P.
345: Mestre, Phys.Rev.Lett. {\bf 85}, 3987 (2000); Y. Shi, cond-mat/0107243 (2001); 
346: K. Helmerson, and L. You, Phys.Rev.Lett. {\bf 87}, 17402 (2001).
347: 
348: \bibitem{deb} B. Deb, and G.S. Agarwal, Phys. Rev. A (in press); 
349:  T. Gasenzer et al., Phys.Rev. A {\bf 65}, 021605(2002); 
350: D.C. Roberts et al., quant-ph/0106060
351: 
352: \bibitem{quasi}  D.M. Stamper-Kurn {\it et al.}, Phys.Rev.Lett. {\bf
353: 83}, 2876 (1999).
354: 
355: \bibitem{bragg} J. Stenger  {\it et al.}, Phys.Rev.Lett. {\bf 82}, 4569 (1999);
356:  S. Inouye et al., Phys.Rev.Lett. {\bf 85}, 4225 (2000).
357: 
358: \bibitem{bogoliubov} N.N. Bogoliubov, J. Phys. (USSR), {\bf 11},
359: 23 (1947); A. L. Fetter, cond-mat/9811366 (1998).
360: 
361: \bibitem{spinor} D.M Stemper Kurn {\it at al.}, Phys.Rev.Lett. {\bf 80}, 2027
362: (1998); T.L. Ho, Phys.Rev.Lett. {\bf 81}, 742 (1998); C.K. Law {\it et al.}, 
363: Phys.Rev.Lett. {\bf 81}, 5257 (1998); L.M. Duan {\it et al.}, quant-ph/0107055;
364: O.E. Mustecaplioglu {\it et al.}, quant-ph/0203015.
365: 
366: \bibitem{spinexp} J. Hald {\it et al.}, Phys. Rev. Lett. {\bf 83}, 1319 (1999);
367: A. Kuzmich {\it et al.}, Phys. Rev.Lett. {\bf 85}, 1594 (2000).
368: 
369: \bibitem{spin1} M. Kitagawa and M. Ueda, Phys.Rev. A {\bf 47}, 5138 (1993), D.J.
370: Wineland {\it et al.}, Phys.Rev. A {\bf 50}, 67 (1994).
371: 
372: \bibitem{spin2} G.S. Agarwal and R.R. Puri, Phys.Rev. A {\bf 41}, 3782 (1990);
373: A. Kuzmich {\it et al.}, Phys. Rev.Lett. {\bf 79}, 4782 (1997); E.S. Polzik,
374: Phys.Rev. A {\bf 59}, 4202 (1999); A. Kuzmich  {\it et al.}, Europhys. Lett. A
375: {\bf 42}, 481 (1998).
376: 
377: \bibitem{sasha} B. Julsgaard, A. Kozhekin, and E.S. Polzik, Nature, {\bf 413}, 
378: 400 (2001).
379: 
380: \bibitem{kimble} S.L. Braunstein and H.J. Kimble,  Phys.Rev.Lett. {\bf 80}, 869 (1998); G.J. Milburn and
381: S.L. Braunstein,  Phys.Rev A {\bf 60}, 937 (1999); 
382: T. Opatrny and G. Kurizki, Phys.Rev.Lett. {\bf 86}, 3180 (2001).
383: A. Kitagawa and K. Yamamoto, quant-ph/0202154.
384: 
385: \bibitem{xip} L.M. Duan {\it et al.}, Phys.Rev.Lett. {\bf 84} 2722 (2000); R.
386: Simon, Phys.Rev.Lett. {\it 84}, 2726 (2000).
387: 
388: \bibitem{ketter} J.M. Vogels {\it et al.},  
389: Phys.Rev.Lett. {\bf 88}, 060402 (2002); A. Brunello {\it et al.}, 
390: Phys.Rev.Lett. {\bf 85}, 4422 (2000).
391: 
392: \bibitem{orzel} C. Orzel et al., Science, {\bf 291}, 2386 (2001).
393: 
394: \bibitem{duan} L.M. Duan {\it et al.}, Nature {\bf 414}, 413 (2001).
395: 
396: \end{references}
397: 
398: \end{document}
399: