1: \documentclass[aps,twocolumn]{revtex4}

2: %\documentclass[aps]{revtex4}

3: \usepackage[dvips]{graphics,graphicx}

4: \begin{document}

5: \title{Dipole and Bloch oscillations of cold atoms in a parabolic lattice}

6: \author{A.~V.~Ponomarev and A.~R.~Kolovsky} 

7: \affiliation{Kirensky Institute of Physics, Ru-660036 Krasnoyarsk, Russia}

8: \date{\today}

9: 

10: \begin{abstract}

11: The paper studies the dynamics of a Bose-Einstein condensate

12: loaded into a 1D parabolic optical lattice, and excited by a sudden

13: shift of the lattice center. Depending on the magnitude of the initial shift, 

14: the condensate undergoes either dipole or Bloch oscillations. 

15: The effects of dephasing and of atom-atom interactions on

16: these oscillations are discussed.

17: \end{abstract}

18: 

19: %\pacs{ }

20: \maketitle

21: 

22: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

23: 1. Bloch oscillations (BO) of a quantum particle in a periodic

24: potential are one of the most fascinating phenomena of quantum

25: physics \cite{Zene34}. Since the pioneering experiment \cite{Daha96}

26: in 1996, this phenomenon has been intensively studied

27: for cold atoms in optical lattices \cite{63}, with recent

28: emphasis on quantum statistical (Fermi or Bose)

29: and atom-atom interaction effects. In particular, the dynamics of 

30: degenerate Bose gases, on which we will focus here, was studied

31: experimentally in \cite{Mors01,Scot04,Modu04}. It should

32: be stressed from the very beginning that, when addressing this

33: problem theoretically, one has to distinguish between quasi

34: one-dimensional lattices (created by two counter-propagating laser

35: beams) and truly 1D lattices (or so-called modulated quantum

36: tubes). Indeed, in the former case the number of atoms per well

37: of the optical lattice can be as large as $10^3-10^4$, and a

38: mean field approach (based on the Gross-Pitaevskii or nonlinear

39: Schr\"odinger equation) is generally justified. This is not the

40: case of the truly 1D lattices, where only few atoms occupy a single

41: well, and, hence, a microscopic analysis is required. For a

42: tilted infinite lattice such analysis, based on the

43: Bose-Hubbard model, was presented in 

44: \cite{57,61,60}, where two regimes of BO -- quasiperiodic and

45: irreversible decaying -- were identified.

46: 

47: When referring to the typical laboratory experiments, an additional

48: complication stems from the harmonic confinement along the lattice.

49: Clearly, harmonic confinement should modify BO of bosonic atoms, and

50: the aim of this work is to estimate its effect. At the same

51: time, parabolic lattices have their own interest, because

52: they allow to study dipole oscillations of BECs. Recent

53: experiments \cite{Fert05} have shown that there is a fundamental

54: difference between dipole oscillations in quasi-

55: and truly 1D lattices. While in the former case the main effect of

56: the periodic potential can be taken into account by simply

57: substituting the atomic mass by its effective mass in the ground

58: Bloch band \cite{Cata03}, one observes a rapid decay of

59: oscillations in the latter case. In the present paper we also

60: briefly discuss dipole oscillations of a BEC in truly 1D

61: lattices, partially overlapping in this part with recent

62: theoretical work \cite{Rey05}.

63: 

64: 

65: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

66: 2. We consider atoms in a parabolic lattice potential

67: $V(x)=M\omega^2x^2/2-V\cos^2(2\pi x/d)$, where the atoms are set

68: into motion by a sudden shift of the trap origin. The relevant

69: parameters of the system are the hopping matrix element $J$

70: (defined by the amplitude of the periodic potential $V$),

71: the `parabolicity' $\nu=M\omega^2 d^2$, and the initial shift $l_0=\Delta x/d$.

72: Using the single band approximation and neglecting atom-atom interactions, the

73: dynamics of the system is described by the pendulum model \cite{63,preprint},

74: %********************************************************

75: \begin{equation}

76: \label{1}

77: i\hbar\dot{a}_l=\frac{\nu}{2}l^2a_l

78: -\frac{J}{2}\left(a_{l+1}+a_{l-1}\right) \;,

79: \end{equation}

80: %

81: where $a_l(t)$ is the complex amplitude of the atoms in the $l$-th

82: well of the optical lattice. The separatrix of the pendulum corresponds to the

83: shift 

84: %***************************************************

85: \begin{equation}

86: \label{2}

87: l^*=2(J/\nu)^{1/2} \;.

88: \end{equation}

89: %

90: If the initial shift $l_0<l^*$, the pendulum shows oscillations around the

91: equilibrium point and, referring to the original system, this regime is

92: regarded as that of dipole oscillations of the atoms. If $l_0>l^*$,

93: the pendulum is in the rotational regime, and the dynamics of the

94: atoms can be regarded as BO in a local static field $F=\nu l_0/d$.

95: 

96: 

97: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

98: 3. We begin with analyzing the regime of BO for an ideal bosonic

99: system (i.e., no atom-atom interactions).

100: Because the local static force $F$ is not homogeneous, one has an

101: additional process of dephasing of BO in a parabolic lattice, as

102: compared to the paradigmatic case of a homogeneously tilted lattice.

103: When discussing the mean atomic momentum, we can estimate this

104: effect of dephasing by evaluating the sum

105: %***************************************************

106: \begin{equation}

107: \label{3}

108: p(t)\sim\sum_m \exp(-m^2/2\gamma^2) \sin[(\omega_B+\nu m/\hbar)t] \;,

109: \end{equation}

110: %

111: where $\omega_B=\nu l_0/\hbar$, $m=l-l_0$, and $\gamma$ is the width of the atomic wave packet (measured in units of the lattice period). Substituting the sum by an integral, we obtain

112: %***************************************************

113: \begin{equation}

114: \label{4}

115: p(t)\sim\exp(-t^2/2\tau_\gamma^2)\sin(\omega_B t) \;,

116: \end{equation}

117: %

118: where

119: %***************************************************

120: \begin{equation}

121: \label{5}

122: \tau_\gamma=\hbar/\gamma\nu \;.

123: \end{equation}

124: %

125: It is seen in Eq.~(\ref{5}),  that the dephasing time $\tau_\gamma$

126: is defined by both the wave packet width and by the trap frequency

127: (see Fig.~\ref{fig1} below). On the basis of this result one might

128: conclude that a narrow wave packet is preferable for studying BO

129: in parabolic lattices. This is, however, not exactly true

130: because a narrow wave packet implies a lower contrast of the

131: interference pattern measured in laboratory experiments. Thus,

132: one has to keep a compromise between the contrast and dephasing,

133: when preparing the initial wave packet.

134: 

135: The irreversible decay of BO according to Eq.~(\ref{4}) is a

136: consequence of our approximation of the sum by an integral.

137: Without this approximation, the decay of oscillations is followed

138: by periodic revivals with a period $T_\nu=2\pi\hbar/\nu$

139: \cite{remark1}. One of these revivals is illustrated in the upper

140: panel of Fig.~\ref{fig1}, which shows the dynamics of the mean

141: momentum of the non-interacting atoms in the parabolic lattice

142: with parabolicity $\nu=0.04 J$.  As initial state of the

143: system we choose here the ground state of the atoms in a

144: parabolic lattice with a slightly tighter confinement $\nu'=4\nu$,

145: which was then shifted by a distance $l_0=8l^*=80$. Note that

146: by changing $\nu'$ we change only the dephasing time [through the

147: change of the wave packet width $\gamma=\gamma(\nu')$], while the

148: revival time is defined exclusively by the parameter $\nu$.

149: %############################################################

150: \begin{figure}[t!]

151: \center

152: \includegraphics[width=8.5cm, clip]{fig1.eps}

153: \caption{Bloch oscillations of $N=5$ atoms in a parabolic

154: lattice with parabolicity $\nu=0.04J$: (a) mean momentum of

155: non-interacting atoms; (b) mean momentum of interacting

156: ($W=0.2J$) atoms; (c) macroscopic coherence of the 

157: interacting atoms. Initial shift $l_0=8l^*=80$.} \label{fig1}

158: \end{figure}

159: 

160: 

161: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

162: 4. Next we address the effect of atom-atom interactions. In the

163: case of quasi one-dimensional parabolic lattices, Bloch and dipole

164: oscillations of the interacting atoms were studied in a number of

165: papers, using the mean-field approach

166: \cite{Kram02,Chio01,Smer02,Adhi03,Nesi04,preprint}. As known, the

167: mean-field approach is justified in the limit of large occupation

168: number $\bar{n}\rightarrow\infty$ and vanishing microscopic interaction

169: constant $W\rightarrow 0$,  and leads (in the simplest case) to the

170: discrete nonlinear Schr\"odinger equation,

171: %********************************************************

172: \begin{equation}

173: \label{6}

174: i\hbar\dot{a}_l=\frac{\nu}{2}l^2 a_l

175: -\frac{J}{2}\left(a_{l+1} +a_{l-1}\right)+g|a_l|^2a_l \;,

176: \end{equation}

177: %

178: where $g=WN$ is the macroscopic interaction constant. In our

179: present work, we focus on the case of truly one-dimensional

180: lattices, where the mean occupation number $\bar{n}\sim 1$.

181: Clearly, the mean field approach is not applicable here and one

182: has to treat the system microscopically by using, for example, the

183: Bose-Hubbard model,

184: %********************************************************

185: \begin{equation}

186: \label{7}

187: H=\sum_l \frac{\nu}{2} l^2 \hat{n}_l

188: -\frac{J}{2}\left(\sum_l \hat{a}^\dag_{l+1}\hat{a}_l+h.c.\right)

189:   +\frac{W}{2}\sum_l \hat{n}_l(\hat{n_l}-1) \;.

190: \end{equation}

191: %

192: The main question we address below is the effect of atom-atom

193: interactions on the Bloch dynamics depicted in the upper panel of

194: Fig.~\ref{fig1}.

195: 

196: First we shall discuss the initial conditions in some more detail.

197: Throughout the paper we shall consider the ground many-body state

198: of the atoms in a parabolic lattice as the initial wave packet

199: (which is shifted then by the distance $l_0$). Clearly, along with

200: the ratio $J/\nu$ this state is also defined by the ratio of the

201: interaction constant to the hopping matrix element. Namely, it is

202: essentially given by the symmetrized product of the

203: single-particle atomic state for $W<J$ , while it may resemble

204: the Mott-insulator state for $W\gg J$ \cite{Rey05}. In what follows we restrict

205: ourselves by a relatively weak interaction. Then the ground state

206: of the system can be well approximated by the many-body wave function

207: %*****************************************************

208: \begin{equation}

209: \label{9}

210: |\tilde{\Psi}_0\rangle=\sum_{\bf n} c_{\bf n}|\bf{n}\rangle \;,\quad

211: c_{\bf n}=\sqrt{N!}\prod_l\frac{a_l^{n_l}}{\sqrt{n_l!}} \;,

212: \end{equation}

213: %

214: where $|{\bf n}\rangle=|\ldots,n_{-1},n_0,n_1,\ldots\rangle$ is

215: the Fock basis and the $a_l$ satisfy the stationary nonlinear

216: Schr\"odinger equation

217: %*****************************************************

218: \begin{equation}

219: \label{10}

220: \frac{\nu}{2}l^2 a_l-\frac{J}{2}\left(a_{l+1}+a_{l-1}\right)

221: +g|a_l|^2a_l=E_0a_l \;.

222: \end{equation}

223: %

224: For example, for $N=5$, $\nu=0.04J$ and $W=0.2J$, the overlap of

225: the state (\ref{9}) with the exact ground state $|\Psi_0\rangle$

226: is $|\langle\Psi_0|\tilde{\Psi}_0\rangle|^2=0.97$. We note that

227: the state (\ref{9}) is completely coherent and is analogous to the

228: super-fluid state in a homogeneous lattice. We shall characterize

229: the macroscopic coherence of the given many-body state

230: $|\Psi\rangle$ by the maximal eigenvalue $\lambda$ of the

231: single-particle density matrix

232: %*****************************************************

233: \begin{equation}

234: \label{8}

235: \rho_{l,m}=N^{-1}\langle\Psi|\hat{a}_l^+\hat{a}_m|\Psi\rangle \;.

236: \end{equation}

237: %

238: Then the macroscopic coherence of the state (\ref{9}) is $\lambda=1$.

239: 

240: We proceed with the dynamics. The middle panel in Fig.~\ref{fig1}

241: shows the mean momentum of $N=5$ interacting atoms ($W=0.2J$). In

242: comparison with the noninteracting case (upper panel), a qualitative

243: change is noticed. This change can be understood by analyzing the

244: macroscopic coherence of the system, shown in the lower panel. It

245: is seen that the macroscopic coherence oscillates with some

246: characteristic period $T_W$. In the case of a homogeneously tilted

247: lattice these oscillations were studied in Ref.~\cite{57}. The origin

248: of the oscillations was shown to be the Stark localization of the

249: single-particle wave functions which, together with the discreetness

250: of the atom number, leads to the following expression for the

251: macroscopic coherence,

252: %*****************************************************

253: \begin{equation}

254: \label{11} \lambda=\exp(-2\bar{n}[1-\cos(Wt/\hbar)]) \;.

255: \end{equation}

256: %

257: In Eq.~(\ref{11}) $\bar{n}$ is the mean number of atoms per

258: lattice site \cite{remark2} and the limit $Fd\gg J$ is implicitly

259: assumed. Since for the considered local static force $Fd=\nu

260: l_0=3.2J$ Stark localization is not complete, the oscillations

261: of the macroscopic coherence decay in time.

262: Nevertheless, if this irreversible decay of coherence is slow on

263: the time scale of the dephasing time, one can observe the revival

264: of BO of the interacting atoms -- an effect which attracts much

265: attention because it provides an independent and accurate method

266: for measuring the microscopic interaction constant $W$.

267: 

268: 

269: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

270: 5. Let us now turn to the case $l_0<l^*$. Here we meet dipole

271: oscillations of a BEC with a characteristic frequency given by the

272: frequency of small pendulum oscillations $\omega_0=(\nu

273: J)^{1/2}/\hbar$. (We recall in passing that the frequency of BO

274: was given by $\omega_B=\nu l_0/\hbar\approx 2\omega_0 l_0/l^*$,

275: $l_0\gg l^*$.) For vanishing atom-atom interactions these dipole

276: oscillations are shown in the upper panel of Fig.~\ref{fig5},

277: where $l_0=l^*/2=5$, and the other parameters are the same as in

278: Fig.~\ref{fig1}. The dephasing time $\tau_\gamma$ is again given

279: by Eq.~(\ref{5}) but with the parameter $\nu$ substituted by the

280: nonlinearity parameter $\tilde{\nu}=\nu/8$ \cite{Lich83}. (The

281: latter parameter also defines the revival time.) The middle and

282: lower panels in Fig.~\ref{fig5} refer to interacting atoms.

283: An exponential decay of the macroscopic coherence is noticed. The

284: other point to which we want to draw the attention of the reader is that a

285: moderate interaction stabilizes the dipole oscillations against

286: dephasing. Within the mean-field approach (which reduces the

287: Bose-Hubbard model to the discrete nonlinear Schr\"odinger

288: equation), this phenomenon is discussed in Ref.~\cite{preprint}.

289: %############################################################

290: \begin{figure}[t!]

291: \center

292: \includegraphics[width=8.5cm, clip]{fig2.eps}

293: \caption{Dipole oscillations of $N=4$ atoms in the parabolic

294: lattice with parabolicity $\nu=0.04J$: (a) mean momentum of

295: non-interacting atoms; (b) mean momentum of interacting

296: ($W=0.2J$) atoms; (c) macroscopic coherence of interacting

297: atoms. Initial shift $l_0=l^*/2=5$.} \label{fig5}

298: \end{figure}

299: 

300: 

301: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

302: 6. In conclusion, we have shown that the dynamics of cold atoms in

303: parabolic lattices is governed by the relation between two

304: characteristic times -- the dephasing time $\tau_\gamma$ and the

305: decoherence time $\tau_W$.

306: 

307: The dephasing time is inversely proportional to the width $\gamma$

308: of the initial wave packet and the nonlinearity $\tilde{\nu}$,

309: which, in turn, is defined by the initial shift $l_0$ of the wave

310: packet relative to the separatrix $l^*=2(J/\nu)^{1/2}$. Namely,

311: $\tilde{\nu}=\nu/8$ for $l_0\ll l^*$, and $\tilde{\nu}=\nu$ for

312: $l_0\gg l^*$. It is interesting to estimate the dephasing

313: time in a typical laboratory experiment. Taking, as an example,

314: the recent experiment \cite{Fert05} with rubidium atoms in an

315: array of axially modulated quantum tubes, we have $\nu=0.0014E_R$

316: and $J=0.38E_R$ for the modulation amplitude (depth of the optical

317: lattice) of one recoil energy. This gives a separatrix position $l^*=33$,

318: and a period $T_0=12.1\ \rm ms$ of small dipole oscillations. Assuming

319: a dilute gas (which, in fact, is not the case realized in the

320: cited experiment) the width of the initial wave packet is

321: $\gamma\approx(J/4\nu)^{1/4}=\sqrt{l^*}/2\approx3$ and, hence, the

322: dephasing time $\tau_\gamma=85\ \rm ms$ for dipole oscillations, and

323: $\tau_\gamma=10.6\ \rm ms$ for BO. Note that these are upper

324: estimates for the dephasing times, and for the initial shift $l_0$

325: closer to the separtrix the dephasing times are essentially

326: smaller. It is also worth noting that there is a maximal shift

327: $l_0$ above which the single band approximation (used throughout

328: the paper) is not valid. The crucial parameter here is the

329: energy gap between the Bloch bands ($\Delta=0.5E_R$ for the

330: specified parameters). The analysis of BO in a parabolic lattice

331: beyond the single-band approximation will be subject of a

332: separate paper.

333: 

334: The decoherence time $\tau_W$ is defined by the characteristic

335: density of the atomic gas $\bar{n}$ and by the value of the

336: microscopic interaction constant $W$. The latter, in turn, is

337: defined by the $s$-wave scattering length and by the degree of

338: confinement of the atoms in the wells of the optical potential. In

339: particular, in the experiment \cite{Fert05}, the quantum tubes

340: were created by two crossing quasi 1D optical lattices with an

341: amplitude $V=30E_R$. For the axial modulation with $V=E_R$ this

342: gives $W=0.73E_R$. For this relatively high value of the

343: interaction constant few atoms per one tube are enough to destroy

344: the dipole/Bloch oscillations on a very short time scale. This

345: qualitatively explains the results of the experiment

346: \cite{Fert05}, where the number of atoms per one quantum tube was

347: around 20. To observe the effects discussed in this paper one has

348: to decrease either the atomic density or the interaction constant

349: (e.g. by use of a Fishbach resonance), as compared to those of

350: Ref.~\cite{Fert05}.

351: %

352: %We also would like to stress that in this work we restrict

353: %ourselves by considering the weak interaction limit where

354: %the initial state of the system is a Bose-Einstein condensate

355: %(i.e., the macroscopic coherence $\lambda\approx 1$ at $t=0$). This is

356: %no way a general case and for the different initial states (realized,

357: %for example, for deeper optical lattices) the atomic dynamics

358: %may fundamentally differ from the discussed Bloch and dipole

359: %oscillations.

360: 

361: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

362: \begin{thebibliography}{10}

363: 

364: \bibitem{Zene34}

365: C.~Zener, Proc. R. Soc. Lond. {\bf A145}, 523  (1934).

366: 

367: \bibitem{Daha96}

368: M.~Ben Dahan, E.~Peik, J.~Reichel, Y.Castin, and C.~Salomon,

369: Phys. Rev. Lett. \textbf{76}, 4508 (1996).

370: 

371: \bibitem{63}

372: For the introductory review see A.R.Kolovsky and H.J.Korsch,

373: %{\em Bloch oscillations of cold atoms in optical lattices},

374: International J. of Mod. Physics {\bf 18}, 1235 (2004).

375: 

376: \bibitem{Mors01}

377: O.~Morsch, J.~H.~M\"uller, M.~Cristani, D.~Ciampini, and E.~Arimondo,

378: %{\em Bloch oscillations and mean-field effects of

379: %Bose-Einstein condensates in 1D optical lattices},

380: Phys. Rev. Lett. \textbf{87}, 140402 (2001).

381: 

382: \bibitem{Scot04}

383: R.~G.~Scott, A.~M.~Martin, S.~Bujkiewicz, T.~M.~Fromhold,

384: N.~Malossi, O.~Morsch, M.~Cristiani, and E.~Arimondo,

385: %{\em Transport and disruption of Bose-Einstein condensates

386: %in optical lattices},

387: Phys. Rev. A {\bf 69}, 033605 (2004).

388: 

389: \bibitem{Modu04}

390: M.~Modugno, E.~de~Mirandes, F.~Ferlaino, H.~Ott, G.~Roati, and M.~Inguscio,

391: %{\em Atom interferometry in a vertical optical lattice},

392: Fortschr. Phys. \textbf{52}, 1173 (2004).

393: 

394: \bibitem{57}

395: A.~R.~Kolovsky,

396: %{\em New Bloch period for interacting cold atoms in 1D optical lattices},

397: Phys. Rev. Lett. \textbf{90}, 213002 (2003).

398: 

399: \bibitem{61}

400: A.~Buchleitner and A.~R.~Kolovsky,

401: %{\em Interaction-induced decoherence of atomic Bloch oscillations},

402: Phys. Rev. Lett. \textbf{91}, 253002 (2003).

403: 

404: \bibitem{60}

405: A.~R.~Kolovsky and A.~Buchleitner,

406: %{\em Floquet-Bloch operator for the Bose-Hubbard model with static field},

407: Phys. Rev. E {\bf 68}, 056213 (2003).

408: 

409: \bibitem{Fert05}

410: C.~D.~Fertig, K.~M.~O'Hara, J.~H.~Huckans, S.~L.~Rolston, W.~D.~Phillips,

411: and J.~V.~Porto,

412: %{\em Strongly inhibited transport of a degenerate 1D Bose gas in a lattice},

413: Phys. Rev. Lett. {\bf 94}, 120403 (2005).

414: 

415: \bibitem{Cata03}

416: F.~S.~Cataliotti, L.~Fallani, F.~Ferlaino, C.~Fort, P.~Maddaloni, and M.~Inguscio,

417: %{\em Superfluid current disruption in a chain of weakly coupled BEC},

418: New J. of Phys. \textbf{5}, 71.1 (2003).

419: 

420: \bibitem{Rey05}

421: A.~M.~Rey, G.~Pupillo, C.~W.~Clark, and C.~J.~Williams,

422: %{\em Ultracold atoms confined in an optical lattice plus parabolic

423: %potential: a closed-form approach},

424: e-print: cond-mat/0503477.

425: 

426: \bibitem{preprint}

427: J.~Brand and A.~R.~Kolovsky,

428: %{\em Pendulum model approach to the dynamics of a Bose-Einstein

429: %condensate in a parabolic lattice},

430: e-print: cond-mat/0412549.

431: 

432: \bibitem{remark1}

433: Indeed, at times $t$ multiple to $T_\nu$ the phases $\nu mt/\hbar$ in

434: the sum (\ref{3})  are multiple of $2\pi$ and, hence, the momentum takes

435: its initial value. (More rigourosly, the revivals of ocsillations follow

436: from the quadratic dependence for the eigenenergies of the quantum pendulum

437: in the asymptotic region $l\gg l^*$.)

438: 

439: \bibitem{Kram02}

440: M.~Kr\"amer, L.~Pitaevskii, and S.~Stringari,

441: %{\em Macroscopic Dynamics of a trapped BEC in the presence of 1D and

442: %2D optical lattices}

443: Phys. Rev. Lett. \textbf{88}, 180404 (2002).

444: 

445: \bibitem{Chio01}

446: M.~L.~Chiofalo and M.~P.~Tosi,

447: %{\em Coherent and dissipative transport of BEC inside an optical lattice}

448: J. Phys. B: At. Mol. Opt. Phys. {\bf 34}, 4551 (2001).

449: 

450: \bibitem{Smer02}

451: A.~Smerzi, A.~Trombettoni, P.~G.~Kevrekidis, and A.~R.~Bishop,

452: %{\em Dynamical superfluid-insulator transition in a chain of

453: %weakly coupled BECs},

454: Phys. Rev. Lett. \textbf{89}, 170402 (2002).

455: 

456: \bibitem{Adhi03}

457: S.~K.~Adhikari,

458: %{\em Mean-field model for Josephson oscillation in BEC on an 1D optical trap},

459: Eur. Phys. J. D {\bf 25}, 161 (2003).

460: 

461: \bibitem{Nesi04}

462: F.~Nesi and M.~Modugno,

463: %{\em Loss and revival of phase coherence in a BEC moving through an optical lattice},

464: J.~Phys.~B: At. Mol. Opt. Phys. {\bf 37}, S101 (2004).

465: 

466: \bibitem{remark2}

467: In the considered case of a finite wave packet it looks

468: reasonable to associate $\bar{n}$ with the number of atoms $N$

469: devided by the wave packet width $\gamma$.

470: 

471: \bibitem{Lich83}

472: A.~J.~Lichtenberg and M.~A.~Libermann, {\em Regular and chaotic

473: dynamics} (Springer, Berlin, 1983).

474: 

475: %\bibitem{remark3}

476: %According to our preliminary results, the decoherence time scales

477: %as $\tau_W\sim\hbar/W\bar{n}$.

478: 

479: \end{thebibliography}

480: \end{document}

481: 

482: