cond-mat0506108/LP.tex
1: \documentstyle[aps,preprint,pra,epsfig]{revtex}
2: %\documentstyle[aps,multicol,pra,epsfig]{revtex}
3: %\documentclass{article}                  
4: 
5: \begin{document}
6: 
7: \draft
8: 
9: \title{On Defect-Mediated Transitions in Bosonic Planar Lattices}
10: 
11: \author{A. Trombettoni$^{1}$, A. Smerzi$^{2,3}$, and P. Sodano$^{4}$}
12: \address{$^1$ Istituto Nazionale per la Fisica della Materia and
13: Dipartimento di Fisica, Universita' di Parma,
14: parco Area delle Scienze 7A, I-43100 Parma, Italy\\
15: $^2$ Istituto Nazionale per la Fisica della Materia BEC-CRS and
16: Dipartimento di Fisica, Universita' di Trento, I-38050 Povo, Italy\\
17: $^3$ Theoretical Division, Los Alamos
18: National Laboratory, Los Alamos, NM 87545, USA\\
19: $^4$ Dipartimento di Fisica and Sezione I.N.F.N., Universit\`a di 
20: Perugia, Via A. Pascoli, I-06123 Perugia, Italy\\
21: }
22: \date{\today}
23: \maketitle
24: 
25: \begin{abstract}
26: We discuss the finite-temperature properties of Bose-Einstein condensates 
27: loaded on a $2D$ optical lattice. In an experimentally attainable range 
28: of parameters the system is described by the $XY$ model, which undergoes 
29: a Berezinskii-Kosterlitz-Thouless (BKT) transition driven by the 
30: vortex pair unbinding. 
31: The interference pattern of the expanding condensates provides 
32: the experimental signature of the BKT transition: 
33: near the critical temperature, the $\vec{k}=0$ component 
34: of the momentum distribution sharply decreases.
35: \end{abstract}
36: %\pacs{PACS: 63.20.Pw, 05.45.-a}
37: 
38: %begin{multicols}{2}
39: 
40: \newpage
41: 
42: Bose-Einstein condensates (BECs) are nowadays routinely 
43: stored in $1D$ 
44: \cite{anderson98,cataliotti01,morsch01,hensinger01,eiermann03},  
45: $2D$ \cite{greiner01} and 
46: $3D$ \cite{greiner02} optical lattices. 
47: In this paper we review and further discuss  
48: the finite-temperature properties of Bose-Einstein condensates 
49: loaded on a $2D$ optical lattice \cite{cm}: there we showed 
50: that at a critical temperature 
51: lower than the temperature $T_{BEC}$ at which condensation 
52: in each well occurs, $2D$ lattices of BECs may undergo a
53: phase transition to a superfluid regime 
54: where the phases of the single-well condensates are coherently aligned
55: allowing for the observation of a 
56: Berezinskii-Kosterlitz-Thouless (BKT) transition. 
57: 
58: The BKT transition is the paradigmatic example of a 
59: defect-mediated transition \cite{nelson83} 
60: which is exhibited by the two-dimensional 
61: $XY$ model \cite{minnaghen87,kadanoff00}, 
62: describing $2$-components spins on a two-dimensional lattice. 
63: In the low-temperature phase, characterized by the presence of 
64: bound vortex-antivortex pairs, the spatial correlations exhibit a power-law 
65: decay; above a critical temperature $T_{BKT}$, the decay is exponential 
66: and there is a proliferation of free vortices. 
67: The BKT transition has been observed in superconducting Josephson arrays 
68: \cite{resnick81} and its predictions are well verified in 
69: the measurements of the superfluid density 
70: in $^4He$ films \cite{bishop78}. In finite magnetic systems 
71: with planar symmetry \cite{bramwell93}, the BKT transition point 
72: is signaled by the drop of a suitably defined
73: magnetization \cite{berezinskii73}. We shall discuss in the sequel 
74: the analogous of this magnetization for atomic systems.  
75: 
76: For $2D$ optical lattices, when the polarization vectors of the two 
77: standing wave laser fields are orthogonal, the resulting periodic 
78: potential for the atoms is
79: $V_{opt}=V_0 [ \sin^2{(kx)} + \sin^2{(ky)} ]$
80: where $k=2 \pi/\lambda$ is the wavevector of the laser beams. 
81: The potential maximum of a single standing wave 
82: $V_0=sE_R$ may be controlled by changing the intensity of the optical 
83: lattice beams, and it is conveniently measured in units of the recoil energy
84: $E_R=\hbar^2 k^2/2m$ ($m$ is the atomic mass), 
85: while, typically, $s$ can vary from $0$ up to $30$. Gravity 
86: is assumed to act along the $z$-axis. 
87: Usually, superimposed to the optical potential 
88: there is an harmonic magnetic potential  
89: $V_{m}=(m/2) [ \omega_r^2 (x^2+y^2) + \omega_z^2 z^2 ]$,
90: where $\omega_z$ ($\omega_r$) is the axial (radial) trapping frequency. 
91: The minima of the lattice
92: are located at the points $\vec{j}=(j_1,j_2) \cdot \frac{\lambda}{2}$ 
93: with $j_1$ and $j_2$ integers, and the potential around the minima is
94: $V_{opt} \approx 
95: \frac{m}{2} \tilde{\omega}_r^2 [ (x-\lambda j_1/2)^2 +  
96: (y-\lambda j_2/2)^2 ]$ with $\tilde{\omega}_r=\sqrt{2 V_0 k^2 / m}$. 
97: When $\tilde{\omega}_r \gg \omega_r,\omega_z$, 
98: the system realizes a square array of tubes, i.e. an array of harmonic traps 
99: elongated along the $z$-axis \cite{greiner01}. 
100: In the following we analyze only 
101: the situation in which the axial degrees of freedom are frozen out, 
102: which is realizable if $\omega_z$ is sufficiently large. We also
103: assume that the harmonic oscillator length 
104: of the magnetic potential $\sqrt{\hbar/m \omega_r}$ is 
105: enough larger than the size $L$ (in lattice units) of the optical 
106: lattice in order to reduce density inhomogeneity effects and to allow for 
107: a safe control of finite size effects.
108: 
109: When all the relevant energy scales 
110: are small compared to the excitation 
111: energies one can expand the field operator \cite{jaksch98} as 
112: $\hat{\Psi}(\vec{r},t)= \sum_{\vec{j}} \hat{\psi}_{\vec{j}}(t) 
113: \Phi_{\vec{j}}(\vec{r})$ with $\Phi_{\vec{j}}(\vec{r})$ 
114: the Wannier wavefunction localized in the $\vec{j}$-th well 
115: (normalized to $1$) and $\hat{N}_{\vec{j}}=\hat{\psi}^{\dag}_{\vec{j}} 
116: \hat{\psi}_{\vec{j}}$ the bosonic number operator.
117: Substituting the expansion in the full quantum Hamiltonian 
118: describing the bosonic system, leads to 
119: the Bose-Hubbard model (BHM) \cite{fisher89,jaksch98} 
120: \begin{equation} 
121: \hat{H}=-K \sum_{<\vec{i}, \vec{j}>} 
122: (\hat{\psi}^{\dag}_{\vec{i}} \hat{\psi}_{\vec{j}}+ h.c.)
123: + \sum_{\vec{j}} \Bigg[ {U \over 2} \hat{N}_{\vec{j}} 
124: (\hat{N}_{\vec{j}}-1) \Bigg]
125: \label{B-H}
126: \end{equation}
127: where $\sum_{<\vec{i}, \vec{j}>}$ denotes a sum over 
128: nearest neighbours, $ U = (4 \pi \hbar^2 a / m) \int d\vec{r} \, 
129: \Phi_{\vec{j}}^4$ ($a$ is the $s$-wave scattering length) and 
130: $K \simeq - \int d\vec{r} \, \big[ \frac{\hbar^2}{2m} 
131: \vec{\nabla} \Phi_{\vec{i}} \cdot \vec{\nabla} \Phi_{{\vec{j}}} + 
132: \Phi_{\vec{i}}  V_{ext} \Phi_{\vec{j}} \big]$. 
133:  
134: Upon defining $J \approx 2 K N_0$ (where $N_0$ is the 
135: average number of atoms per site),   
136: when $N_0 \gg 1 $ and $J/N_0^2 \ll U$, 
137: the BHM reduces to 
138: $\hat{H}=H_{XY}- {U \over 2}\sum_{\vec{j}} 
139: \frac{\partial^2}{\partial \theta_{\vec{j}}^2}$, 
140: which describes the so-called quantum phase model 
141: (see e.g. \cite{simanek94,fazio01}): 
142: $\theta_{\vec{j}}$ is the phase of the $j$-th condensate and 
143: \begin{equation}
144: H_{XY}=-J \sum_{<\vec{i},\vec{j}>} \cos{(\theta_{\vec{i}}-
145: \theta_{\vec{j}})}
146: \label{X-Y}
147: \end{equation}
148: stands for the Hamiltonian of the classical $XY$ model. 
149: When $J \gg U$ and at temperatures $T \gg U/k_B$, 
150: the quantum phase Hamiltonian may be 
151: well approximated by Eq.(\ref{X-Y}); 
152: therefore, the pertinent partition function 
153: describing the thermodynamic behaviour 
154: of the BECs stored in an optical lattice can be computed 
155: in these limits with the classical $XY$ model.
156: 
157: We remark that the BHM - for $U=0$ -  
158: describes harmonic oscillators and 
159: thus cannot sustain any BKT transition. However, 
160: in the limits above specified ($N_0 \gg 1 $ and $J/N_0^2 \ll U \ll J$) 
161: the Bose-Hubbard Hamiltonian reduces in the $XY$ model 
162: (\ref{X-Y}), which displays the BKT transition. We observe that in 
163: the chosen range of parameters, the condition 
164: $U N_0^2/J=U N_0/2K \gg 1$, satisfies the 
165: finite-temperature stability criterion 
166: recently derived in \cite{tsuchiya03}.  
167: In superconducting networks 
168: $N_0$ is the average number of excess Cooper pairs and 
169: thus, the requirements $N_0 \gg 1 $ and $J/N_0^2 \ll U$ are 
170: always satisfied; at variance, in bosonic lattices 
171: $N_0$ varies usually between $\sim 1$ and $\sim 10^3$ and the validity 
172: of the mapping in the quantum phase Hamiltonian is not always guaranteed.      
173:   
174: A simple estimate of the coefficients $J$ and 
175: $U$ may be obtained by 
176: approximating the Wannier functions with gaussians, 
177: whose widths are determined variationally \cite{cm}. 
178: For a $2D$ lattice with $V_0$ between $20$ and $25E_R$ (and 
179: $N_0 \approx 170$ as in \cite{greiner01}), 
180: the conditions $J \gg U \gg J/N_0^2$ are rather well 
181: satisfied and the BKT critical temperature, $T_{BKT} \sim J/k_B$, 
182: is between $10$ and $30nK$. 
183: 
184: We observe that, 
185: using a coarse-graining approach to determine the finite-temperature  
186: phase-boundary line of the Bose-Hubbard model (\ref{B-H}) 
187: \cite{kampf93}, 
188: one gets - in $2D$ for $N_0 \gg 1$ and $J / U \gg 1 $ - 
189: a critical temperature $T_{BKT} \approx J / k_B$. 
190: Indeed, within a coarse-graining approach, the 
191: phase-boundary line of the Bose-Hubbard model at a temperature 
192: $T$ is given by \cite{kampf93}
193: \begin{equation}
194: 1=K d \int_{0}^{\beta} d \tau G(\tau)
195: \label{PBL}
196: \end{equation}
197: where $d$ is the lattice dimension and 
198: $\beta=1/k_B T$. In Eq.(\ref{PBL}) $G(\tau)$ is the correlation function, 
199: which, for $N_0 \gg 1$, is given by
200: \begin{equation}
201: G(\tau) \approx N_0 \frac{\vartheta_3 \bigg( 
202: \pi(N_0+\tau/\beta),q \bigg)}{\vartheta_3 \bigg( 
203: \pi N_0,q \bigg)} e^{-\tau U(1-\tau/\beta)/2}
204: \label{correlatore}
205: \end{equation}
206: with $\vartheta_3(z,q)=1+2 \sum_{n=1}^{\infty} \cos{(2nz)} 
207: \cdot q^{n^2}$. Since $d=2$, and $J=2KN_0$, 
208: from Eq.(\ref{PBL}) it follows $1=J G(\omega=0)/N_0$ where 
209: $G(\omega=0)=\int_0^\beta d \tau G(\tau)$; one has then  
210: \begin{equation}
211: \frac{G(\omega=0)}{N_0}=
212: \frac{\sum_{m} \exp{(-\beta U \phi_n^2 / 2)} \cdot 
213: \frac{1-\exp{(-\beta U (\phi_n + 1 / 2)}}{U(\phi_m+1/2)}} 
214: {\sum_m \exp{(-\beta U \phi_n^2 / 2)} }:
215: \label{trasf_corr}
216: \end{equation}
217: where $\phi_m=N_0-m$. Eq.(\ref{trasf_corr}) yields, for $U \ll k_B T$,  
218: $G(\omega=0) / N_0 \approx \beta$, thus implying 
219: $\beta J \approx 1$, i.e. $k_B T_{BKT} \approx J$, 
220: in fair agreement with $XY$ estimates.
221: 
222: We notice that the $XY$ Hamiltonian, 
223: in the limits $N_0 \gg 1$ and $J \gg U \gg J/N_0^2$, can be retrieved  
224: extending the Gross-Pitaevskii (GP) Hamiltonian to 
225: finite temperature $T \gg U/k_B$. 
226: In fact, replacing the tight-binding ansatz 
227: $\Psi(\vec{r},t)= \sum_{\vec{j}} \psi_{\vec{j}}(t) 
228: \Phi_{\vec{j}}(\vec{r})$ (where $\psi_{\vec{j}}$ are classical fields 
229: and not operators) in the GP Hamiltonian 
230: one gets the lattice Hamiltonian 
231: $H=-K \sum_{<\vec{i}, \vec{j}>} 
232: (\psi^{\ast}_{\vec{i}} \psi_{\vec{j}}+ c.c.)
233: + (U/2) \sum_{\vec{j}} N_{\vec{j}} (N_{\vec{j}}-1)$, 
234: which is the classical version of the Bose-Hubbard model 
235: ($N_{\vec{j}} \equiv \mid \psi_{\vec{j}}\mid^2$). 
236: Writing $N_{\vec{j}}=N_0 + \delta N_{\vec{j}}$, one may 
237: neglect 
238: the quadratic terms [i.e. $(\delta N_{\vec{j}})^2$] 
239: in the hopping part of the Hamiltonian 
240: $-K \sum_{<\vec{i}, \vec{j}>} 
241: (\psi^{\ast}_{\vec{i}} \psi_{\vec{j}}+ c.c.)$, which then 
242: reduces to $H_{XY}$ (\ref{X-Y}). 
243: 
244: Accurate Monte Carlo simulations yield - for the 
245: $XY$ model - the BKT critical temperature 
246: $T_{BKT}=0.898J/k_B$ \cite{gupta88}.
247: When $U \ll J$, 
248: a BKT transition still occurs at a slightly 
249: lower critical temperature $T_{BKT}(U)$. 
250: To evaluate more accurately the effects of the interaction 
251: term on the BKT critical temperature one may use a
252: semiclassical approximation to evaluate a renormalized Josephson energy
253: \cite{jose84,rojas96}: 
254: the renormalization group equations yields then  
255: the following equation for ${\cal K} \equiv \beta T_{BKT}(U)$
256: \begin{equation}
257: 2-\pi {\cal K}  \Bigr[1-\frac{1}{4 {\cal K}}-\frac{1}{X_u 
258: {\cal K}^3} \, g({\cal K},X_u) \Bigr]=0
259: \label{radice}
260: \end{equation}
261: where $X_u=U/\pi J$ and $g({\cal K},X_u) \equiv \sum_{n=1}^{\infty} 
262: [X_u {\cal K}^2 / 2-(n^2/2) 
263: ln (1+ X_u {\cal K}^2/n^2)]$. One may see \cite{smerzi04} that 
264: for $U/J<36/\pi$, Eq.(\ref{radice}) has an {\it unique} solution; 
265: at variance, for $U/J>36/\pi$ Eq.(\ref{radice}) 
266: does not have any solution. 
267: The value $U/J=36/\pi$ is in reasonable agreement with the known 
268: mean-field prediction $2 z J / U \approx 1$ 
269: for the $T=0$ phase transition 
270: (see, e.g., Eq. (30) of Ref. \cite{vanoosten01} for $N_0 \gg 1$). 
271: In the inset of Fig.1 we plot the value of $T_{BKT}(U)/T_{BKT}$ 
272: (where $T_{BKT}$ is the BKT critical temperature of the XY model 
273: with $U=0$) as a function of $U/J$ from the numerical solution 
274: of Eq.(\ref{radice}). 
275: 
276: The emerging physical picture is the following: 
277: There are two relevant temperatures for the system, 
278: the temperature $T_{BEC}$, at which in each well there is a condensate, 
279: and the temperature $T_{BKT}$ at which the condensates phases 
280: start to coherently point in the same direction. Of course, 
281: in order to have well defined condensates phases one should have 
282: $T_{BKT} < T_{BEC}$.
283: The critical temperature $T_{BEC} \approx 0.94 \hbar N_0^{1/3} 
284: \bar{\omega}/k_B$ where $\bar{\omega}= (\tilde{\omega}_r^2 \omega_z)^{1/3}$.
285: With the numerical values given in \cite{greiner01}, 
286: $T_{BEC}$ turns out to be  
287: $ \gtrsim \ 500 nK$ for $s \gtrsim 20$. When $T < T_{BEC}$, 
288: the atoms in the well $\vec{j}$ 
289: of the $2D$ optical lattice may be described by a macroscopic 
290: wavefunction $\psi_{\vec{j}}$. 
291: Furthermore, when the fluctuations around 
292: the average number of atoms per site $N_0 \gg 1$ are strongly suppressed, 
293: one may put, apart from the factor $\sqrt{N_0}$ constant across the array,  
294: $\psi_{\vec{j}}=e^{i \theta_{\vec{j}}}$. 
295: The temperature $T_{BKT}$ is of order 
296: of $J/k_B$: with the experimental parameters of 
297: \cite{greiner01} and $V_0$ between $20$ and $25E_R$, 
298: one has that $T_{BKT}$ is between $10$ and $30nK$, which is 
299: sensibly smaller than the condensation temperature 
300: $T_{BEC}$ of the single well. 
301: A similar picture describes  
302: also $2D$ arrays of superconducting Josephson junctions 
303: \cite{simanek94,fazio01}: they 
304: exhibit a temperature $T_{BCS}$ at which the metallic grains 
305: placed on the sites of the array become (separately) superconducting 
306: and the Cooper pairs may be described by macroscopic wavefunctions. 
307: At a temperature $T_{BKT}<T_{BCS}$, 
308: the array undergo a BKT transition and the system - as a whole - 
309: becomes superconducting.
310: 
311: The experimental signature of the BKT transition in bosonic planar lattices 
312: is obtained by measuring, as a function of the temperature, 
313: the central peak of the interference pattern 
314: obtained after turning off the confining potentials \cite{cm}. 
315: In fact, the peak of the momentum distribution at 
316: $\vec{k}=0$ is the direct analog of the magnetization of a 
317: finite size $2D$ $XY$ magnet. For the $XY$ magnets, the spins can be written 
318: as $\vec{S}_{\vec{j}}=(\cos{\theta_{\vec{j}}},\sin{\theta_{\vec{j}}})$ and 
319: the magnetization is defined as $M=(1/N) \cdot  
320: \langle \, \mid \sum_{\vec{j}} \vec{S}_{\vec{j}} \mid \, \rangle$ 
321: where $\langle \cdots \rangle$ stands for the thermal average. 
322: A spin-wave analysis at low temperatures yields 
323: $M=(2N)^{-k_B T/8 \pi J}$ \cite{bramwell93,berezinskii73}. 
324: With discrete BECs at $T=0$, all the phases 
325: $\theta_{\vec{j}}$ are equal at the equilibrium and 
326: the lattice Fourier transform of $\psi_j$, 
327: $\tilde{\psi}_{\vec{k}}=\frac{1}{N} 
328: \sum_{\vec{j}} \psi_{\vec{j}} \, \, \, e^{-i \vec{k} \cdot {\vec{j}}}$, 
329: exhibits a peak at $\vec{k}=0$ ($\vec{k}$ 
330: is in the first Brillouin zone of the $2D$ square lattice) and
331: the magnetization is: 
332: \begin{equation}
333: M=\langle \, \mid \tilde{\psi}_0 \mid \, \rangle.
334: \label{M}
335: \end{equation}  
336: 
337: The intuitive picture of the BKT transition is then the following: 
338: at $T=0$, all the spins point in the same direction.  
339: Increasing the temperature, bound vortex pairs are thermally induced.  
340: In Fig.2, left inset, 
341: a single free vortex is depicted: we plot a configuration 
342: of the phases such that $\theta_{\vec{j}}$ equals the polar angle of the 
343: site $\vec{j}$ with respect to the core vortex 
344: (notice that with $\theta_{\vec{j}}$ equals 
345: the polar angle plus $\pi / 2 $ one would have 
346: the usual plot of a vortex). 
347: As one can see, a single vortex modifies the phase 
348: distribution also much far from its core and the square modulus of 
349: its lattice Fourier transform $\mid \tilde{\psi}_{\vec{k}} \mid^2$
350: has a minimum at $\vec{k}=0$. At variance, a vortex-antivortex  
351: pair [see Fig.2, right inset] 
352: modifies the phase distribution only near the center of 
353: the pair (in this sense is a {\em defect}) and its lattice Fourier transform 
354: has a maximum at $\vec{k}=0$. 
355: Increasing the temperature, vortices are thermally induced. 
356: For $T < T_{BKT}$ only bound vortex pairs are present, and in
357: average the spins continue to point in the same direction. 
358: When the condensates expand, a large peak (i.e., a magnetization) 
359: is observed in the central $\vec{k}=0$ momentum component. 
360: Rising further the temperature, 
361: due to the increasing number of vortex pairs, the central peak density 
362: decreases from the $T=0$ value. For $T \approx T_{BKT}$ 
363: the pairs starts to unbind and free vortices begin to appear, determining 
364: a sharp decrease around $T_{BKT}$ of the magnetization. 
365: At high temperatures, only free vortices are present,
366: leading to a randomization 
367: of the phases and to a vanishing magnetization. 
368: In Fig.1 we plot the intensity of the central peak of the 
369: momentum distribution (normalized to the value at $T=0$) in a 2D 
370: lattice as a function of the reduced temperature $k_B T/J$, 
371: evidencing the sharp decrease 
372: of the magnetization around the critical temperature \cite{cm}. 
373: 
374: %Let us now turn our attention to the the interference patterns 
375: %of the expanding condensates: after the bosonic system reaches the 
376: %equilibrium, one may switch off both the harmonic trap and the optical
377: %lattice. At this time ($t=0$) the momentum distribution 
378: %$\tilde{\psi}(\vec{p},t=0)$ is given by the Fourier transform of 
379: %$\psi(\vec{r},t=0)=\sum_{\vec{j}} \psi_{\vec{j}} \Phi_{\vec{j}}(\vec{r})$:
380: %$\tilde{\psi}(\vec{p},t=0)= \tilde{\Phi}(\vec{p}) \tilde{\psi}_{\vec{k}}$
381: %where $\tilde{\Phi}(\vec{p})$ is the Fourier transform 
382: %of the $3D$ Wannier functions and $\hbar \vec{k}=(p_x,p_y)$ is the 
383: %momentum projection on the 
384: %first Brillouin zone of the $2D$ optical lattice. 
385: %Using a gaussian for the Wannier functions, 
386: %$\Phi(\vec{r}) \propto e^{-x^2/2\sigma_x^2} 
387: %e^{-y^2/2\sigma_y^2}e^{-z^2/2 \sigma_z^2}$ 
388: %(with $\sigma_z \gg \sigma \equiv \sigma_x=\sigma_y$, 
389: %since $\tilde{\omega}_r \gg \omega_z$), 
390: %one has $\tilde{\Phi}(\vec{p})=\chi(p_x) \chi(p_y) \chi(p_z) \propto 
391: %e^{-\sigma^2 (p_x^2+p_y^2)/2\hbar^2}$, where $\chi(p_x) \propto 
392: %e^{-\sigma_x^2 p_x^2/ 2 \hbar^2}$ and similarly for $\chi(p_y)$ and 
393: %$\chi(p_z)$. 
394: %Since for $s \gg1$ $\sigma / d =1/ \pi s^{1/4}$, 
395: %the momentum distribution exhibits well 
396: %pronounced peaks in the centers of each Brillouin zone. 
397: %The peaks have different heights and the central $\vec{k}=0$ peak 
398: %is the largest (see Figs.2 and 3 of \cite{greiner01}).
399:  
400: At $t=0$ the amplitude of the $\vec{k}=0$ peak of the 
401: momentum distribution is simply given by the thermal average 
402: of $\tilde{\psi}_{0}$. By measuring the $\vec{k}=0$ 
403: peak (i.e., $\langle \, \mid \tilde{\psi}_0  \mid^2 \,    \rangle$) 
404: at different temperatures, one obtains the results plotted in Fig.1. 
405: The figure has been obtained using a Monte Carlo simulation of the 
406: $XY$ magnet for a $40 \times 40$ array: we find $k_B T_C \approx 1.07 J$. 
407: In Fig.1 we also plot the low-temperature 
408: spin wave prediction \cite{berezinskii73} (solid line). 
409: At times different from  
410: $t=0$, the density profiles 
411: are well reproduced by the free expansion of the ideal gas. 
412: One obtains $\tilde{\psi}(\vec{p},t)=\chi(p_z) 
413: e^{-i[(p_z+mgt)^3-p_z^3]/6m^2g\hbar} \tilde{\varphi}(p_x,p_y,t)$ 
414: [where $\chi(p_x) \propto  
415: e^{-\sigma_x^2 p_x^2/ 2 \hbar^2}$ and similarly for $\chi(p_y)$ and  
416: $\chi(p_z)$], i.e. 
417: a uniformly accelerating motion along $z$ and a free motion in the 
418: plan $x-y$, with $\tilde{\varphi}(p_x,p_y,t)= 
419: \chi(p_x) \chi(p_y) \tilde{\psi}_{\vec{k}}e^{-i (p_x^2+p_y^2) t/2\hbar m}$
420: giving the central and lateral peaks of 
421: the momentum distribution as a function of time for different 
422: temperatures. 
423: 
424: An intense experimental work is now focusing on the 
425: Bose-Einstein condensation in two dimensions: 
426: at present a crossover to two-dimensional behaviour 
427: has been observed for $Na$ \cite{gorlitz01} and $Cs$ atoms \cite{hammes03}. 
428: Our analysis relies on two basic assumptions; namely, 
429: the validity of the tight-binding approximation 
430: for the Bose-Hubbard Hamiltonian and the requirement 
431: that the condensate in the optical lattice may be regarded as planar 
432: \cite{petrov00}. It is easy to see that the first assumption 
433: is satisfied if, at $T=0$, $V_0 \gg \mu$ 
434: (where $\mu$ is the chemical potential), and, at finite temperature, 
435: $\hbar \tilde{\omega}_r \gtrsim k_B T$. The second assumption is much more 
436: restrictive since it requires freezing the transverse excitations; 
437: for this to happen one should require a condition on the transverse 
438: trapping frequency $\omega_z$. Namely, one should have that, at $T=0$, 
439: $\hbar \omega_z \gtrsim 8 K$ and that, at finite temperature, 
440: $\hbar \omega_z \gtrsim k_B T$ (since $\omega_z \ll \tilde{\omega}_r$, 
441: the latter condition also implies that $\hbar \tilde{\omega}_r 
442: \gtrsim k_B T$). In \cite{greiner01} it is  
443: $V_0=s \cdot k_B \cdot 0.15 \mu K$ and for $s \gtrsim 20$ 
444: the tight-binding conditions are satisfied since 
445: $10 Hz \gtrsim 8K/2\pi \hbar$; furthermore, if $\omega_z=2 \pi \cdot 1kHz$, 
446: one may safely regard our finite temperature analysis to be valid at least 
447: up to $T \sim 50 nK$. We notice that the experimental signature for the  
448: BKT transition for a continuous 
449: (i.e., without optical lattice) weakly
450: interacting 2D Bose gas \cite{prokofev01} is also given 
451: by the central peak of the atomic density of the 
452: expanding condensates.
453: 
454: Our study evidences the possibility that Bose-Einstein
455: condensates loaded on a $2D$ optical lattice may exhibit 
456: - at finite temperature - a new coherent 
457: behaviour in which all the phases of the condensates located in each well 
458: of the lattice point in
459: the same direction. The finite temperature transition, which is due 
460: to the thermal atoms in each well, is mediated by vortex defects and may 
461: be experimentally detectable by looking at the
462: interference patterns of the expanding condensates. Our analysis
463: strengthens - and extends at finite temperature - the 
464: striking and deep analogy of bosonic planar systems 
465: with $2D$ Josephson junction arrays \cite{anderson98}. 
466: 
467: %We stress out that the transition we propose to observe 
468: %is different from the quantum phase transition reported in \cite{greiner02}. 
469: %There the system is at $T=0$ and the insulator phase
470: %(signaled by a reversible
471: %destruction of the phase coherence across the lattice)
472: %is reached varying the optical lattice 
473: %parameters. Since, in a mean-field approach, 
474: %the Mott insulator-superfluid transition 
475: %occurs at $T=0$ approximately at $2 z J/U \sim 1$ \cite{fisher89,fazio01} 
476: %($z$ is the number of nearest neighbours), from Fig.1 
477: %one sees that it is much easier to detect a quantum phase transition  
478: %for the three-dimensional array (and indeed it has been recently detected  
479: %\cite{greiner02}), while no quantum phase transition has yet been 
480: %reported in literature for $2D$ and $1D$ optical lattices, for whom 
481: %a larger laser power $V_0$ is required. At variance, here
482: %we propose to fix the system parameters in the superfluid phase and 
483: %increase the temperature $T$ until the thermally induced vortex 
484: %proliferation determines the BKT transition.
485: 
486: {\em Acknowledgements} 
487: This work has been supported by MIUR through grant No. 2001028294 
488: and by the DOE.
489: 
490: \begin{thebibliography}{10}
491: 
492: \bibitem{anderson98} B. P. Anderson and M. A. Kasevich, 
493: Science {\bf 282}, 1686 (1998). 
494:  
495: \bibitem{cataliotti01} F. S. Cataliotti, S. Burger, 
496: C. Fort, P. Maddaloni, F. Minardi, A. Trombettoni, 
497: A. Smerzi, and M. Inguscio, Science {\bf 293}, 843 (2001).  
498:  
499: \bibitem{morsch01} O. Morsch, J. H. M\"uller, 
500: M. Cristiani, D. Ciampini, and E. Arimondo   
501: Phys. Rev. Lett. {\bf 87}, 140402 (2001).  
502:  
503: \bibitem{hensinger01} W. K. Hensinger, H. Haffer, A.  Browaeys, 
504: N. R. Heckenberg, K. Helmerson, C. McKenzie, G. J. Milburn, 
505: W. D. Phillips, S. L. Rolston, H. Rubinsztein-Dunlop, and B. Upcroft, 
506: Nature {\bf 412}, 52 (2001).  
507:  
508: \bibitem{eiermann03} 
509: B. Eiermann, P. Treutlein, T. Anker, 
510: M. Albiez, M. Taglieber, K.-P. Marzlin, and M. K. Oberthaler 
511: Phys. Rev. Lett. {\bf 91}, 060402 (2003). 
512: 
513: \bibitem{greiner01}  
514: M. Greiner, I. Bloch, O. Mandel, T. W. H\"ansch, and 
515: T. Esslinger,  Phys. Rev. Lett. {\bf 87}, 160405 (2001); 
516: Appl. Phys. B {\bf 73}, 769 (2001). 
517:  
518: \bibitem{greiner02} M. Greiner, O. Mandel, T. Esslinger, 
519: T. W. H\"ansch, and I. Bloch, 
520: Nature {\bf 415}, 39 (2002). 
521: 
522: \bibitem{cm} A. Trombettoni, A. Smerzi, and P. Sodano, 
523: New J. Phys. {\bf 7}, 57 (2005).
524: 
525: \bibitem{nelson83} D. R. Nelson, in {\em Phase Transitions and 
526: Critical Phenomena}, vol. 7, eds. C. Domb and J. L. Lebowitz 
527: (New York, Academic Press, 1983) and references therein.
528: 
529: \bibitem{minnaghen87} P. Minnaghen, Rev. Mod. Phys. {\bf 59}, 
530: 1001 (1987).
531: 
532: \bibitem{kadanoff00} L. P. Kadanoff, {\em Statistical Physics: Statics, 
533: Dynamics and Renormalization}, (Singapore, World Scientific, 2000), 
534: Chapt.s 16-17 and reprints therein.
535: 
536: \bibitem{resnick81} D. J. Resnick, J. C. Garland, 
537: J. T. Boyd, S. Shoemaker, and R. S. Newrock, 
538: Phys. Rev. Lett. {\bf 47}, 1542 (1981).
539: 
540: \bibitem{bishop78}  D. J. Bishop and J. D. Reppy, 
541: Phys. Rev. Lett. {\bf 40}, 1727 (1978).
542: 
543: \bibitem{bramwell93} S. T. Bramwell and P. C. W. Holdsworth, 
544: J. Phys.: Condensed Matter {\bf 5}, L53 (1993); 
545: Phys. Rev. B {\bf 49}, 8811 (1994).
546: 
547: \bibitem{berezinskii73} V.L. Berezinskii and A.Ya. Blank, 
548: Sov. Phys. JETP {\bf 37}, 369 (1973).
549: 
550: \bibitem{jaksch98} D. Jaksch, C. Bruder, J. I. Cirac, 
551: C. W. Gardiner, and P. Zoller,
552: Phys. Rev. Lett. {\bf 81}, 3108 (1998). 
553:  
554: \bibitem{fisher89} M. P. A. Fisher, P. B. Weichman, G. Grinstein, 
555: and D. S. Fisher, Phys. Rev. B {\bf 40}, 546 (1989).  
556: 
557: %\bibitem{anglin01} J.R. Anglin{\em et al.}, 
558: %Phys. Rev. A {\bf 64}, 063605 (2001).
559: 
560: %\bibitem{batrouni02} G.G. Batrouni {\em et al.}, 
561: %Phys. Rev. Lett. {\bf 89}, 117203 (2002).
562: 
563: %\bibitem{dalfovo99} F. Dalfovo {\em et al.}, 
564: %Rev. Mod. Phys. {\bf 71}, 463 (1999).
565: 
566: \bibitem{simanek94} E. Sim\`anek, {\em Inhomogeneous Superconductors},
567: Oxford University Press, New York, 1994.
568: 
569: \bibitem{fazio01} R. Fazio and H. van der Zant, 
570: Phys. Rep. {\bf 355}, 235 (2001).
571: 
572: %\bibitem{kampf93} A.P. Kampf and G.T. Zimanyi, 
573: %Phys. Rev. B {\bf 47}, 279 (1993).
574: 
575: \bibitem{tsuchiya03} S. Tsuchiya and A. Griffin, 
576: Phys. Rev. A {\bf 70}, 023611 (2004). 
577: 
578: \bibitem{kampf93} A. P. Kampf and G. T. Zimanyi,  
579: Phys. Rev. B {\bf 47}, 279 (1993). 
580: 
581: \bibitem{gupta88} R. Gupta, J. DeLapp, G. G. Batrouni, 
582: G. C. Fox, C. F. Baillie, and J. Apostolakis, 
583: Phys. Rev. Lett. {\bf 61}, 1996 (1988).
584: 
585: \bibitem{jose84} J.V. Jos\'e, 
586: Phys. Rev. B {\bf 29}, 2836 (1984).
587: 
588: \bibitem{rojas96} C. Rojas and J.V. Jos\'e, 
589: Phys. Rev. B {\bf 54}, 12361 (1996).
590: 
591: \bibitem{smerzi04} A. Smerzi, P. Sodano, and A. Trombettoni,
592: J. Phys. B {\bf 37}, S265 (2004).
593: 
594: %\bibitem{sheshadri93} K. Sheshadri {\em et al.}, Europhys. Lett. 
595: %{\bf 22}, 257 (1993).
596: 
597: %\bibitem{freericks95} J.K. Freericks and H. Monien, Europhys. Lett. 
598: %{\bf 26}, 545 (1993).
599: 
600: \bibitem{vanoosten01} 
601: D. van Oosten, P. van der Straten, and H. T. C. Stoof,   
602: Phys. Rev. A {\bf 63}, 053601 (2001). 
603: 
604: %\bibitem{pedri01} P. Pedri {\em et al.}, 
605: %Phys. Rev. Lett. {\bf 87}, 220401 (2001).
606: 
607: \bibitem{gorlitz01} A. G\"orlitz, J. M. Vogels, 
608: A. E. Leanhardt, C. Raman, T. L. Gustavson, 
609: J. R. Abo-Shaeer, A. P. Chikkatur, S. Gupta, 
610: S. Inouye, T. Rosenband, and W. Ketterle,  
611: Phys. Rev. Lett. {\bf 87}, 130402 (2001).
612: 
613: \bibitem{hammes03} M. Hammes, 
614: D. Rychtarik, B. Engeser, H.-C. N\"agerl, and R. Grimm, 
615: Phys. Rev. Lett. {\bf 90}, 173001 (2001).
616: 
617: \bibitem{petrov00} D. S. Petrov, M. Holzmann, 
618: and G. V. Shlyapnikov, Phys. Rev. Lett. {\bf 84}, 2551 (2000).
619: 
620: %\bibitem{popov83} V. N. Popov, {\em Functional Integrals in Quantum 
621: %Field Theory and Statistical Physics} (Reidel, Dordrecht, 1983). 
622: 
623: \bibitem{prokofev01} N. Prokof\'ev, O. Ruebenacker, and B. Svistunov,
624: Phys. Rev. Lett. {\bf 87}, 270402 (2001).
625: 
626: \end{thebibliography}
627: 
628: \begin{figure}
629: \centerline{\psfig{figure=fig1.ps,width=60mm,angle=270}}
630: \caption{Intensity of the central peak of the momentum distribution 
631: (normalized to the value at $T=0$) as a function of the reduced temperature 
632: $t=k_B T/J$; empty circles: Monte Carlo simulations; 
633: solid line: low-temperature spin wave prediction 
634: \protect\cite{berezinskii73}. 
635: Inset: the BKT critical temperature $T_{BKT}(U)$ 
636: (in units of the $XY$ critical temperature) as a function  
637: of $U/J$ from Eq.(\ref{radice}).}
638: \label{fig1}
639: \end{figure}
640: 
641: \begin{figure}
642: \centerline{\psfig{figure=fig2.ps,width=60mm,angle=0}}
643: \caption{Left: a lattice vortex; right: a vortex-antivortex 
644: pair.}
645: \label{fig2}
646: \end{figure}
647: 
648: %end{multicols}
649: 
650: \end{document}
651: 
652: 
653: 
654: 
655: 
656: 
657: 
658: 
659: 
660: 
661: 
662: 
663: