1: \documentclass[prl,epsfig,twocolumn,amssymb]{revtex4}
2: %\documentclass[prl,epsfig,amssymb]{revtex4}
3: %\renewcommand{\baselinestretch}{2.0}
4:
5: \usepackage{graphicx}
6:
7:
8: \begin{document}
9: \title{Dynamics of a quantum phase transition in a ferromagnetic Bose-Einstein
10: condensate}
11: \begin{abstract}
12: We discuss dynamics of a slow quantum phase transition in a spin-1
13: Bose-Einstein condensate. We determine analytically the scaling properties of the
14: system magnetization and verify them with numerical simulations in a one
15: dimensional model.
16: \end{abstract}
17: \author{Bogdan Damski and Wojciech H. Zurek}
18: \affiliation{
19: Theory Division, Los Alamos National Laboratory, MS-B213, Los Alamos, NM 87545, USA
20: }
21: \maketitle
22:
23: Studies of phase transitions have traditionally focused
24: on {\it equilibrium} scalings of various properties near the critical point.
25: Dynamics of the phase transition presents new challenges and there is a strong motivation
26: for analyzing it. Nonequilibrium phase
27: transitions may play a role in the evolution of the early Universe \cite{kibble}.
28: Their analogues can be studied in the condensed matter
29: experiments. The latter observation led to
30: development of the theory based on the universality of critical
31: behavior \cite{zurek}, which in turn resulted
32: in a series of beautiful experiments \cite{eksperymenty}.
33: The recent progress in the cold atom experiments allows for
34: time dependent realizations of different models undergoing a quantum phase
35: transition (QPT) \cite{lewenstein,nature_kurn}.
36: These experimental developments are only a proverbial tip of the iceberg, but they call
37: for an in-depth theoretical understanding of the QPT dynamics.
38:
39:
40: A QPT is a fundamental change in {\it ground state} (GS) of the system as a result of
41: small variations of an external parameter, e.g.,
42: a magnetic field \cite{sachdev}.
43: It takes place ideally at zero absolute temperature, which is in striking
44: contrast to thermodynamical phase transitions.
45: The most complete description of the QPT dynamics has been obtained
46: so far in spin models
47: \cite{dorner,spiny} that are exactly solvable.
48: In these systems the gap in the
49: excitation spectrum goes to zero at the critical point, which precludes the
50: adiabatic evolution across the phase boundary. It leads to creation of
51: excitations whose density and scaling with a quench rate
52: follow from a quantum version \cite{dorner,bodzio} of the Kibble-Zurek (KZ) theory
53: \cite{kibble,zurek}.
54:
55: We study dynamics of a ferromagnetic condensate of
56: spin-1 particles \cite{fer}.
57: For simplicity, we consider
58: 1D homogeneous (untrapped) clouds: atoms in a box as in the
59: experiment \cite{raizen} with spinless bosons.
60: We adopt the parameters for our 1D model
61: such that the length and time scales are comparable to experimental
62: ones \cite{units}. Assuming that the system is placed in a
63: magnetic field $B$ aligned in the $z$ direction, one gets the following
64: dimensionless mean-field energy functional \cite{units}
65: \begin{eqnarray}
66: \label{E}
67: E[\Psi]= \int&dz&\frac{1}{2}\frac{d\Psi^\dag}{dz}\frac{d\Psi}{dz}
68: + \frac{c_0}{2}\left(\Psi^\dag\Psi\right)^2
69: + Q\langle\Psi|F_z^2|\Psi\rangle\nonumber\\&+&
70: \frac{c_1}{2}\sum_\alpha\langle\Psi|F_\alpha|\Psi\rangle^2
71: \end{eqnarray}
72: where
73: $\Psi^T=(\psi_1,\psi_0,\psi_{-1})$ describes the $m=0,\pm1$ condensate components,
74: $\int dz \sum_m|\psi_m|^2=1$, and
75: $F_{x,y,z}$ are spin-1 matrices \cite{ueda_broken}.
76: The first term in (\ref{E}) is the kinetic energy,
77: the second and the fourth term describe spin-independent
78: and spin-dependent atom interactions respectively, the third term is a
79: quadratic Zeeman shift coming from atom interactions with a magnetic field.
80: For $^{87}$Rb atoms considered
81: here $c_1<0$, which results in an interesting phase diagram due to
82: the competition between the last two terms in (\ref{E}).
83: Restricting analysis to zero longitudinal magnetization case,
84: and introducing
85: $$
86: q=Q/(n|c_1|),\ \ n=\Psi^\dag\Psi
87: $$
88: one finds a polar phase for $q>2$, described by $\Psi_P^T\sim(0,1,0)$,
89: and the broken-symmetry phase where
90: $$
91: \Psi_B^T\sim(\sqrt{4-2q}e^{i\chi_1},
92: \sqrt{8+4q}e^{i(\chi_1+\chi_{-1})/2},\sqrt{4-2q}e^{i\chi_{-1}})
93: $$
94: for $0\le q<2$.
95: The freedom of choosing the $\chi_{\pm1}$ results in rotational symmetry
96: of the transverse magnetization on the $(x,y)$ plane.
97: The transition between these phases can be
98: driven by the change of the magnetic field $B$ imposed on the
99: atom cloud, $q\sim Q\sim B^2$ \cite{ueda_pra2007}, which was experimentally
100: done in \cite{nature_kurn}.
101:
102: The dynamics of a QPT depends on the rate of quench
103: driving the system across the phase boundary.
104: For very fast ``impulse'' transition,
105: the system has no time to adjust to the changes of the Hamiltonian and
106: arrives in a region where a new phase is expected with
107: the ``old'' wave-function untouched during the evolution.
108: Slow transitions are different: the system has time to ``probe''
109: various broken symmetry ``vacua'' in the neighborhood of the critical
110: point where it gets excited.
111: We are interested in evolutions that are fast enough
112: to produce macroscopic excitations of the system, but slow enough to
113: reflect scalings of the critical region. By comparing analytical
114: findings to numerical simulations for experimentally relevant parameters
115: we provide the first complete description of QPT dynamics in a ferromagnetic
116: condensate.
117:
118: Fast transitions were realized in the Berkeley experiment
119: \cite{nature_kurn}. The 3D numerical simulations closely following this
120: experiment were reported in \cite{ueda_pra2007}. Analytical studies
121: of the evolution after ``impulse'' quench were
122: presented in \cite{austen,uwe}. The paper of Lamacraft
123: \cite{austen} also discusses dynamics of non instantaneous transitions
124: in 2D spinor condensates focusing on analytical predictions on
125: the growth of the transverse magnetization correlation functions.
126:
127: We start with a qualitative discussion.
128: Considering small perturbations around the GS of the broken-symmetry phase
129: one finds three Bogolubov modes as in \cite{ueda_broken}
130: where quantum fluctuations are studied.
131: In the long wavelength limit (important for slow transitions)
132: there is only one nonzero eigenvalue mode:
133: the gapped mode having eigenenergy $\Delta\sim\sqrt{4-q^2}$.
134: Suppose now that we drive the system from polar to broken-symmetry phase.
135: The system reaction time
136: to Hamiltonian changes in the broken-symmetry phase is given by the inverse
137: of the gap: $\tau\sim\frac{1}{\Delta}$ \cite{dorner,bodzio}. For example,
138: when $\tau$ is
139: small enough the evolution becomes adiabatic so the system adjusts
140: fast to parameter changes. Right after entering the broken-symmetry
141: phase, the reaction time is
142: large with respect to the transition time, $\Delta/\frac{d\Delta}{dt}$, and so
143: the system undergoes the ``impulse'' evolution where its state is ``frozen''.
144: The gapped mode starts to be populated around the instant $\hat t$ after
145: entering the broken-symmetry phase: the system leaves the ``impulse'' regime to
146: catch up with instantaneous GS solution. This happens when the two time scales
147: become comparable: $1/\Delta(\hat t)\sim\Delta/\frac{d\Delta}{dt}|_{t=\hat t}$.
148: We consider here transitions driven by
149: \begin{equation}
150: \label{q_od_t}
151: q(t)=2-t/\tau_Q,
152: \end{equation}
153: where $\tau_Q$ is the quench time inversely proportional to the speed of
154: driving the system through the phase transition.
155: For slow transitions of interest here, $\tau_Q\gg1$, we obtain
156: \begin{equation}
157: \label{t_hat}
158: \hat t\sim \tau_Q^{1/3}.
159: \end{equation}
160:
161: In the following we analyze dynamics induced by a linear decrease of $q(t)$ (\ref{q_od_t}).
162: The evolution starts from $t<0$, i.e., in the polar phase, and ends
163: at $t=2\tau_Q$ ($q=0$). Such $q(t)$ dependence is achieved
164: by ramping down the magnetic field as $\sim\sqrt{2-t/\tau_Q}$.
165: The initial state is chosen as a slightly
166: perturbed GS in the polar phase,
167: $
168: \Psi^T\sim(\delta\psi_1, 1/\sqrt{L}+\delta\psi_0, \delta\psi_{-1}),
169: $
170: where $|\delta\psi_m|\ll 1/\sqrt{L}$ are random.
171: We generate the real and imaginary
172: part of $\delta\Psi_m$ at different grid points with the
173: probability distribution $p(x)=\exp(-x^2/2\sigma^2)/\sqrt{2\pi}\sigma$.
174: We take $\sigma= 10^{-4}$ to start evolution
175: closely to the polar phase GS.
176:
177: To find the full numerical solution within the mean-field approximation,
178: we integrate three coupled nonlinear Schr\"odinger
179: equations for the $\psi_m$ condensates that can be easily obtained by the
180: variation of (\ref{E}).
181: During evolution we look at the magnetization of the
182: sample
183: $$
184: f_\alpha=\langle\Psi|F_\alpha|\Psi\rangle, \ \ \alpha=x,y,z.$$
185: {\it The transverse magnetization.} A total transverse (to the magnetic field
186: in the $z$ direction) magnetization reads
187: \begin{equation}
188: \label{MT}
189: M_T(t)=\int dz [f_x^2(z,t)+f_y^2(z,t)]=\int dz\,m_T,
190: \end{equation}
191: and is experimentally measurable.
192: It disappears in the GS of the polar phase and
193: equals $(1-q^2/4)/L$ in the broken-symmetry GS.
194: Its typical
195: evolution is depicted in Fig. \ref{fig1}. We see there that nothing happens
196: in the polar phase. The system starts nontrivial evolution in the
197: broken-symmetry phase at a distance
198: $\hat t/\tau_Q$ after the critical point was passed.
199: The magnetization grows fast
200: from that point until it exceeds the static
201: prediction and starts oscillations with the amplitude decreasing
202: in time. We consider slow transitions. Therefore,
203: by the end of time evolution, when $q=0$, the system is in the slightly
204: perturbed ferromagnetic GS: globally $M_TL\approx1$ (Fig. \ref{fig1})
205: and locally $L^2m_T(z)\approx1$ (Fig. \ref{fig3}).
206: We can now ask: Does the scaling (\ref{t_hat}) hold?
207: To find out we define arbitrarily $\hat t$ as the instant when $M_TL$ intersects $1\%$.
208: A fit to numerics for $\tau_Q\ge10$ yields
209: $\ln\hat t=(0.056\pm0.01)+(0.332\pm0.002)\ln\tau_Q$ which confirms prediction
210: (\ref{t_hat}). This fit is presented in Fig. \ref{fig1}a, where the gradual departure of the
211: numerical data for $\tau_Q<10$ from
212: $\hat t\sim\tau_Q^{1/3}$ indicates that
213: $\tau_Q\gg1$ or $37ms$ has to be taken for the observation of $1/3$
214: exponent: quench has to be slow enough to reflect the critical
215: dynamics.
216: \begin{figure}[t]
217: \includegraphics[width=\columnwidth,clip=true]{fig1.eps}
218: \caption{(color) Main plot: numerical solution (black solid line) vs.
219: static prediction (red dashed line).
220: The arrow depicts direction of evolution.
221: Inset (a): the same
222: as in the main plot plus a numerically obtained
223: solution of the linearized problem (green divergent line). Inset (b):
224: numerical data vs. fit to $\tau_Q\ge10$ data only (see text for details).
225: In the main plot $\tau_Q= 10$ (see \cite{units} for
226: units).
227: }
228: \label{fig1}
229: \end{figure}
230: \begin{figure}[t]
231: \includegraphics[width=\columnwidth,clip=true]{fig2a.eps}\\
232: \includegraphics[width=\columnwidth,clip=true]{fig2b.eps}
233: \caption{The vectors represent $(f_x(z),f_y(z))\times10^3$. Plot (a):
234: snapshot at $q(t=2.81)= 1.72$, i.e., at the first peak in
235: $M_TL$ (see Fig. \ref{fig1}). Plot (b): snapshot by the
236: end of time evolution: $q(t=20)=0$. The results come from the same numerical simulation
237: as in Fig. \ref{fig1} (see \cite{units} for units).
238: }
239: \label{fig2}
240: \end{figure}
241:
242: In the GS configuration of the
243: broken-symmetry phase the vector $(f_x,f_y)$ can have arbitrary orientation,
244: so in the dynamical problem considered here it is interesting to find
245: out how is this rotational symmetry broken.
246: When unstable evolution starts, spatial correlations
247: in magnetization appear (Fig. \ref{fig2}a).
248: In the subsequent evolution these correlations evolve such that the
249: correlation length increases: see Fig. \ref{fig2}b obtained by the end of
250: time evolution. This is a generic picture though the details
251: depend on the quench time $\tau_Q$ and initial state of the system.
252: This behavior suggests creation of spin textures \cite{stephens,vilenkin}.
253: In our case,
254: topological textures are spin configurations where
255: the magnetization direction varies in space so that the kinetic energy term in
256: (\ref{E}) is not minimized, but magnetization magnitude
257: follows closely a GS result.
258: Such structures appear in 1D when the first homotopy group of the
259: vacuum manifold $\cal M$ is nontrivial, which happens here:
260: $\pi_1({\cal M})={\mathbb Z}$ \cite{makela}. These textures are characterized
261: by the winding number,
262: $
263: \frac{1}{2\pi}\int dz \frac{d}{dz}{\rm Arg}(f_x+if_y),
264: $
265: which is not conserved. Indeed,
266: it reads $+1$ in Fig. \ref{fig2}a, while
267: by the end of that evolution (Fig. \ref{fig2}b) it equals $0$.
268:
269: \begin{figure}[t]
270: \includegraphics[width=\columnwidth,clip=true]{fig3.eps}
271: \caption{(color) Magnetization of the system at $t=2\tau_Q$
272: ($q=0$). The dashed lines facilitate observation of
273: extrema coincidences. Results come from the same
274: simulation as in Figs. \ref{fig1}, \ref{fig2}; see \cite{units} for units.
275: }
276: \label{fig3}
277: \end{figure}
278:
279: Are different stages of this evolution
280: experimentally observable? Let's look at $\tau_Q=10$ case
281: presented in Figs. \ref{fig1}-\ref{fig3}. The evolution from the phase
282: boundary to
283: the first peak in magnetization $M_T$ (the $q=0$ point)
284: takes $2.81\times37ms\cong104ms$ ($2\tau_Q=740ms$).
285: Both these time scales are
286: well within the reach of the experiment \cite{nature_kurn}.
287:
288: {\it The longitudinal magnetization.}
289: Initially, $f_z(z)\approx0$ so that $\int dz f_z\approx0$.
290: The conservation of the latter allows only
291: for creation of a network
292: of magnetic domains (nontopological structures with fixed $f_z$ sign)
293: having opposite polarizations.
294: The domains
295: appear by the time when the system enters unstable evolution and the
296: maxima of $|f_z|$ tend to move towards the minima of $m_T$
297: (Fig. \ref{fig3}).
298: More quantitatively,
299: we performed $N_r$ evolutions starting from different initial conditions, but
300: fixed $\sigma$. As in the experiment \cite{nature_kurn}, we average over these
301: runs to wash out shot-to-shot fluctuations.
302: In Fig. \ref{fig4} we plot the mean domain size:
303: $\xi=\sum_i\xi_z(i)/N_r$, where $i=1,...,N_r$ and
304: $\xi_z(i)$ is the mean domain size in the $i$-th run.
305: As shown in Fig. \ref{fig4}a, for $t\lesssim\hat t$ we observe
306: $\xi\approx f(t/\tau_Q^{1/3})$ as for $M_T(t)$.
307: The domains are formed on a time scale of $\sim\hat t$.
308: A simple analysis based on KZ theory \cite{kibble,zurek} suggests that
309: their characteristic {\it post-transition} size, $\hat \xi$, should be
310: roughly given by $\int_0^{\sim\hat t} dt v_s(t)$,
311: where $v_s(t)$ is a sound velocity.
312: There are two sound modes in the broken-symmetry phase that propagate both spin and density
313: fluctuations \cite{ueda_broken}: the faster (slower) one
314: has velocity $\sim\sqrt{c_0}$ ($\sim\sqrt{q|c_1|}$).
315: Putting any of these as $v_s$ into the integral, and
316: assuming $\tau_Q\gg1$ for the slower mode, we get
317: \begin{equation}
318: \label{XI}
319: \hat\xi\sim\tau_Q^{1/3}.
320: \end{equation}
321: This result correctly predicts the scaling
322: property of the size of {\it post-transition} ``defects'' as is evident from
323: the overlap of different curves in Fig. \ref{fig4}, which
324: shows up for $\tau_Q\ge25$ or $0.9s$. Quantitatively,
325: we define $\hat\xi$ as the value of $\xi$ averaged over $q\in[1/2,1]$
326: to wash out post-transition fluctuations.
327: A fit got us $\ln\hat\xi= (-0.38\pm0.03)+ (0.30\pm0.01)\ln\tau_Q$,
328: in good agreement with (\ref{XI}). The fit was done to
329: $\tau_Q\ge30$ data and is presented in Fig. \ref{fig4}b which illustrates
330: that smaller $\tau_Q$ data gradually departs from $1/3$ scaling law.
331:
332:
333: Now we focus on the analytical calculations providing predictions
334: about early stages of time-evolution. We assume
335: that the wave-function stays close to the polar phase GS,
336: $
337: \Psi^T=(\delta\psi_1(t), 1/\sqrt{L}+\delta\psi_0(t), \delta\psi_{-1}(t))
338: \exp(-i\mu t)
339: $,
340: where the chemical potential is $\mu=c_0/L$, $|\delta\psi_m|\ll1/\sqrt{L}$,
341: and $\int dz(\delta\Psi_0+\delta\Psi_0^*)\equiv0$ to keep $\int dz
342: \Psi^\dag\Psi=1+O(\delta\Psi^2)$.
343: Linearizing
344: the coupled nonlinear-Schr\"odinger equations that describe the system
345: we get $f_\chi={\rm Re} G_\chi$, where
346: $\chi=x,y$, $G_x= \sqrt{2}(\delta\Psi_1+\delta\Psi_{-1})/\sqrt{L}$,
347: $G_y= i\sqrt{2}(\delta\Psi_1-\delta\Psi_{-1})/\sqrt{L}$, and
348: $$
349: i\frac{d}{dt}G_\chi=
350: -\frac{1}{2}\frac{d^2}{dz^2}G_\chi+\frac{\alpha}{2}qG_\chi-\frac{\alpha}{2}(G_\chi+G_\chi^*),
351: $$
352: where $\alpha=2|c_1|/L$.
353: To solve this equation we go to momentum space, $a_\chi(k)=\int dz
354: f_\chi\exp(ikz)$ and $b_\chi(k)=\int dz {\rm Im}G_\chi\exp(ikz)$,
355: getting
356: \begin{equation}
357: \label{ewol}
358: \frac{d}{dt}
359: \left[\begin{array}{c}
360: a_\chi \\ b_\chi
361: \end{array}
362: \right]=\frac{1}{2}
363: \left(
364: \begin{array}{cc}
365: 0 & k^2+\alpha q \\
366: 2\alpha-k^2-\alpha q & 0
367: \end{array}
368: \right)
369: \left[\begin{array}{c}
370: a_\chi \\ b_\chi
371: \end{array}
372: \right].
373: \end{equation}
374: Diagonalizing the matrix from Eq. (\ref{ewol}) we see that there is
375: instability for $k^2/\alpha<2-q$ as in the
376: Bogolubov spectrum of this model \cite{ueda_broken}.
377: Thus, the system is stable in the polar
378: phase, and so small initial perturbations do not grow during
379: the evolution towards broken-symmetry phase.
380: The instability for $q<2$ is responsible for the magnetization jump
381: in Fig. \ref{fig1} and the subsequent breakdown of the linear approach (Fig.
382: \ref{fig1}b).
383:
384: \begin{figure}[t]
385: \includegraphics[width=\columnwidth,clip=true]{fig4.eps}
386: \caption{(color) Dynamics of magnetic domains in $f_z$. Black line ($\tau_Q=30$),
387: red line ($\tau_Q=50$), green line ($\tau_Q=80$). The arrow show direction
388: of evolution on the main plot.
389: Inset (a): early stages of $\xi(t)$ evolution.
390: Inset (b): dependence of the typical post-transition
391: domain size, $\hat\xi$, on quench time. The fit was done to $\tau_Q\ge30$ data
392: (see text for details).
393: The figure shows results averaged over $N_r=44$ runs;
394: see \cite{units} for units.
395: }
396: \label{fig4}
397: \end{figure}
398:
399: To solve Eq. (\ref{ewol}) with $q(t)$ given by (\ref{q_od_t}) we derive the
400: equation for $d^2a_\chi(t)/dt^2$, keep leading order terms
401: in the slow transition ($\tau_Q\gg1$)
402: and long-wavelength \protect{($k^2/\alpha\ll2$)} limits,
403: and get that
404: \begin{equation}
405: \label{aki}
406: a_\chi(k,t)= \alpha_{k\chi}{\rm Ai}(s)+ \beta_{k\chi}{\rm Bi}(s), \
407: \frac{s}{\kappa}= \frac{t}{\tau_Q^{1/3}}-\frac{k^2\tau_Q^{2/3}}{\alpha},
408: \end{equation}
409: where $\kappa=(\alpha^2/2)^{1/3}$,
410: $\alpha_{k\chi}$ and $\beta_{k\chi}$ are constants given by initial
411: conditions, while Ai and Bi are Airy functions.
412: From (\ref{aki}) we see that the instability
413: arises from unbounded increase of the Bi$(s)$ function happening for
414: $s>0$, i.e., $k^2/\alpha<2-q(t)$, which is a
415: dynamical manifestation of the static result for unstable modes.
416: This solution works till $t\sim\hat t\sim \tau_Q^{1/3}$ when
417: a significant increase of $f_\chi$
418: invalidates the linearized theory: this calculation
419: rigorously derives scaling (\ref{t_hat}). Additionally, the solution (\ref{aki})
420: can be reliably used as long as $\tau_Q\gg1$ or $37ms$, which is also supported
421: by numerics (Fig. \ref{fig1}a). The quench time scale in the experiment \cite{nature_kurn}
422: is much smaller than this bound. Finally, these results hold for any initial state
423: spread over the $k$ modes.
424:
425: The (re)scalings $t/\tau_Q^{1/3}$ and
426: $\hat\xi\sim\hat t\sim\tau_Q^{1/3}$ derived above in a 1D system
427: were also found by different means in a 2D spinor condensate \cite{austen}.
428: A trivial extension of our mean-field analytical calculations to 2D and 3D systems
429: shows that they hold for any number of spatial dimensions.
430:
431: To summarize, we have developed a theory of the dynamics of symmetry-breaking
432: in the quantum phase transition inspired by the experiment \cite{nature_kurn},
433: but for the range of quench rates that are sufficiently slow so that the
434: critical scalings can determine phase transition dynamics. This regime should
435: be accessible by a ``slower'' version of the quench \cite{nature_kurn}.
436: Our analysis points to a
437: Kibble-Zurek-like scenario, where the state of the system departs from the
438: old symmetric vacuum with a delay $\sim\hat t$ after the critical point was
439: crossed. This sets up an initial post-transition state with a characteristic
440: length scale $\hat\xi$ (\ref{XI}). This scale should determine the initial density of
441: topological features. In our 1D simulations textures appear, but we predict
442: that in real 3D experiments other topological defects are created
443: (as they were in \cite{nature_kurn}),
444: and the distance between them should be initially $\sim\hat\xi$. Such
445: topological defects are more stable than textures so measurement of their
446: density should be possible and would be a good test of the theory we have
447: presented.
448:
449: \begin{thebibliography}{99}
450:
451: \bibitem{kibble} T.W.B. Kibble, J. Phys. A {\bf 9}, 1387 (1976);
452: Phys. Rep. {\bf 67}, 183 (1980).
453:
454: \bibitem{zurek} W.H. Zurek, Nature (London) {\bf 317}, 505 (1985);
455: Acta Phys. Pol. B {\bf 24}, 1301 (1993);
456: Phys. Rep. {\bf 276}, 177 (1996).
457:
458: \bibitem{eksperymenty}
459: I. Chuang {\it et al.}, Science {\bf 251}, 1336 (1991);
460: M.J. Bowick {\it et al.}, {\it ibid.} {\bf 263}, 943
461: (1994);
462: C. Bauerle {\it et al.}, Nature (London) {\bf 382}, 332 (1996);
463: V.M.H. Ruutu {\it et al.}, {\it ibid.} {\bf 382}, 334 (1996);
464: S. Ducci {\it et al.}, {\it ibid.} {\bf 83}, 5210 (1999);
465: A. Maniv, E. Polturak, and G. Koren, Phys. Rev. Lett. {\bf 91}, 197001 (2003);
466: R. Monaco {\it et al.}, {\it ibid.} {\bf 96}, 180604 (2006).
467:
468:
469: \bibitem{lewenstein} M. Lewenstein {\it et al.}, Adv. Phys. {\bf 56}, 243 (2007).
470:
471: \bibitem{nature_kurn} L.E. Sadler, J.M. Higbie, S.R. Leslie, M.
472: Vengalattore, and D. M. Stamper-Kurn,
473: {\it et al.}, Nature (London) {\bf 443}, 312 (2006).
474:
475: \bibitem{sachdev} S. Sachdev, Quantum Phase Transitions (Cambridge University
476: Press, Cambridge UK, 2001).
477:
478: \bibitem{dorner}
479: W.H. Zurek, U. Dorner, and P. Zoller, Phys. Rev. Lett. {\bf 95}, 105701 (2005).
480:
481: \bibitem{spiny}
482: J. Dziarmaga, Phys. Rev. Lett. {\bf 95}, 245701 (2005); R.W. Cherng and L.S. Levitov,
483: Phys. Rev. A {\bf 73}, 043614 (2006).
484:
485: \bibitem{bodzio} B. Damski, Phys. Rev. Lett. {\bf 95}, 035701 (2005).
486: B. Damski and W.H. Zurek, Phys. Rev. A {\bf 73}, 063405 (2006).
487:
488: \bibitem{fer} T.-L. Ho, Phys. Rev. Lett. {\bf 81}, 742 (1998);
489: T. Ohmi and K. Machida, J. Phys. Soc. Jpn. {\bf 67}, 1822 (1998).
490:
491: \bibitem{raizen} T.P. Meyrath {\it et al.},
492: %F. Schreck, J.L. Hanssen,
493: %C.-S. Chuu and M.G. Raizen,
494: Phys. Rev. A {\bf 71}, 041604(R) (2005).
495:
496: \bibitem{units}
497: The experiment \cite{nature_kurn} is done in the harmonic potential
498: $m\omega_z^2(\lambda_x^2x^2+\lambda_y^2y^2+z^2)/2$, where $m$ is the
499: $^{87}$Rb mass,
500: $\omega_z=2\pi\times4.3{\rm Hz}$, $\lambda_x=13$, and $\lambda_y=81.4$.
501: We use the oscillator units through the paper for
502: time ($1/\omega_z=37ms$), length ($\sqrt{\hbar/m\omega_z}=5.1\mu m$), and
503: energy ($\hbar\omega_z$), and
504: assume that the system stays in the harmonic oscillator ground states
505: in $x$ and $y$ directions. Then, after skipping constant terms,
506: the 3D energy functional \cite{fer} reduces to dimensionless (\ref{E})
507: with the additional $\Psi^\dag\Psi z^2/2$ term, and
508: $c_i=2N\alpha_i\sqrt{m\omega_z/\hbar}\sqrt{\lambda_x\lambda_y}/3$
509: ($N=2\times10^6$ is the atom number,
510: $\alpha_0= 16nm$, and $\alpha_1=-\alpha_0/216.1$).
511: To approximate such a system by a box we assume that
512: the total density of atoms in the harmonic trap center is the same as in the
513: box of size $L$. Neglecting in the Thomas-Fermi limit
514: the first and the last term of (\ref{E}) we get
515: $L=2(\sqrt{c_0}2/3)^{2/3}\approx78$, i.e., $78\times5.1\mu m\approx0.4 mm$.
516:
517: \bibitem{ueda_broken} K. Murata, H. Saito, and M. Ueda, Phys. Rev. A {\bf 75},
518: 013607 (2007).
519:
520: \bibitem{ueda_pra2007} H. Saito, Y. Kawaguchi, and M. Ueda, Phys. Rev. A {\bf 75},
521: 013621 (2007).
522:
523: \bibitem{austen} A. Lamacraft, Phys. Rev. Lett. {\bf 98}, 160404 (2007).
524:
525: \bibitem{uwe} M. Uhlmann, R. Sch\"utzhold, and U.R. Fischer, cond-mat/0612664.
526:
527: \bibitem{stephens} G.J. Stephens, Phys. Rev. D {\bf 61}, 085002 (2000).
528:
529: \bibitem{vilenkin} A. Vilenkin and E.P.S. Shellard, {\it Cosmic strings
530: and other topological defects} (Cambridge University Press, Cambridge, 1994).
531:
532: \bibitem{makela} H. M\"akel\"a, Y. Zhang, and K.-A. Suominen, J. Phys. A {\bf
533: 36}, 8555 (2003).
534:
535: \end{thebibliography}
536:
537: \end{document}
538: