1: \documentclass[eqsecnum,aps,showpacs,twocolumn]{revtex4} % PH. REV. FINAL FORMAT STYLE
2: \usepackage{graphicx}
3:
4: \begin{document}
5:
6: \title{Effect of an impulsive force on vortices in a rotating Bose-Einstein
7: condensate}
8:
9: \author{Sadhan K. Adhikari}
10:
11: \author{Paulsamy Muruganandam}
12:
13: \affiliation{Instituto de F\'{\i}sica Te\'orica, Universidade Estadual
14: Paulista, 01.405-900 S\~ao Paulo, S\~ao Paulo, Brazil}
15:
16: \date{\today}
17:
18: \begin{abstract}
19:
20: The effects of a sudden increase and decrease of the interatomic interaction
21: and harmonic-oscillator trapping potential on vortices in a quasi
22: two-dimensional rotating Bose-Einstein condensate are investigated using the
23: mean-field Gross-Pitaevskii equation. Upon increasing the strength of
24: interaction suddenly the condensate enters a nonstationary oscillating phase
25: which starts to develop more vortices. The opposite happens if the strength is
26: reduced suddenly. Eventually, the number of vortices attains a final value at
27: large times. Similarly, the number of vortices increases (decreases) upon a
28: sudden reduction (augmentation) in the trapping potential. We also study the
29: decay of vortices when the rotation of the condensate is suddenly stopped.
30: Upon a free expansion of a rotating BEC with vortices the radius of the vortex
31: core increases more rapidly than the radius of the condensate. This makes the
32: counting and detection of multiple vortex easier after a free expansion.
33:
34: \end{abstract}
35:
36: \pacs{03.75.Fi}
37:
38: \maketitle
39:
40: \section{Introduction}
41:
42: Since the successful detection \cite{1} of Bose-Einstein condensates (BEC)
43: in dilute trapped bosonic atoms at ultra-low temperature, one problem of
44: extreme interest is the formation of vortices in a rotating condensate
45: in an axially symmetric
46: trap.
47: The experimental observation
48: \cite{exp,exp1,expm,jila,hal,hal1,mit,mit1} of vortices in
49: the condensate has intensified theoretical investigations on vortices
50: using the mean-field Gross-Pitaevskii (GP) equation \cite{gp}. For large
51: frequencies of rotation, a very large number of vortices arranged in a
52: triangular lattice, frequently referred to as a vortex lattice, have been
53: observed. The
54: vortex lattices are highly excited collective states of BECs with an
55: angular momentum of up to $60\hbar$ per particle \cite{mit}.
56: There have
57: been many theoretical studies on different aspects of BEC \cite{2,3,11}
58: and specially, on the vortices in a rotating BEC in
59: axially symmetric traps
60: \cite{2a,2c,2d,2e,sin,2f,2ff,2f1,2g,2h,fc,ad,ueda,dis}.
61:
62:
63: When an ordinary fluid is placed in a rotating container the fluid rotates
64: with the container because of viscous force. For superfluid $^4$He(II) in
65: a rotating container no motion of the fluid is observed below a critical
66: rotational frequency. Above the critical frequency, lines of singularity
67: appear in its velocity field. These lines of singularity correspond to
68: vortices and appear in the form of a lattice specially for large
69: rotational frequencies. Superfluidity exhibits many peculiar features:
70: absence of viscosity, reduction of moment of inertia, occurrence of
71: persistent current, appearance of quantized vortices among others
72: \cite{11}. The formation of vortex lattice in a rotating $^4$He(II) is
73: intimately related to its superfluidity and is a manifestation of
74: quantized vortices \cite{exp,exp1,expm,jila,mit,mit1}. However, because of
75: the
76: strong interaction between $^4$He atoms, the theoretical description of
77: this system is not an easy task.
78:
79: Similar vortices can be generated in theoretical mean-field models of
80: trapped BEC based on the GP equations
81: \cite{2a,2c,2d,2e,sin,2f,2ff,2f1,2g,2h,fc,ad,ueda} and have been observed
82: experimentally \cite{exp,exp1,expm,jila,hal,hal1,mit,mit1} in BEC with
83: repulsive interaction. In contrast to liquid $^4$He(II), a trapped BEC is
84: a very dilute and weakly interacting system, which makes a mean-field
85: analysis appropriate. This analogy of a trapped BEC with liquid $^4$He(II)
86: suggests the presence of superfluidity in the BEC. Many different ways
87: for generating quantized vortices in a BEC have been suggested \cite{2d} ,
88: e.g., via spontaneous vortex formation in evaporative cooling \cite{2f},
89: via controlled excitation to an excited state of atoms \cite{2ff}, by
90: stirring a BEC using an external laser with an angular frequency above a
91: critical value for generating a quantized vortex with a single quantum
92: \cite{2c}, or by the rotation of an axially symmetric trap with an angular
93: frequency above a similar critical value \cite{2g}.
94:
95: Experimentally, quantized vortices have been generated by different groups
96: in a rotating BEC using a variety of techniques. Vortices have been
97: detected in a $^{87}$Rb condensate in a cylindrical trap by Madison et al
98: at ENS \cite{exp}, where angular momentum is generated by a stirring laser
99: beam. At MIT Raman et al \cite{expm} also studied the nucleation of
100: vortices in a Bose-Einstein condensate stirred by a laser beam. They have
101: been observed by Matthews et al \cite{jila} at JILA in coupled BEC's
102: comprised of two spin states of $^{87}$Rb in a spherical trap, where
103: angular momentum is generated by a controlled excitation of the atoms
104: between the two states. The tilting of the vortex axis in a nonspherical
105: trapping potential has been studied at JILA by Haljan et al \cite{hal}.
106: Vortices have also been generated by evaporatively spinning up a normal
107: gas of $^{87}$Rb and then cooling below quantum degeneracy at JILA by
108: Haljan et al \cite{hal1}. A large number of vortices in the form of a
109: lattice have been observed at MIT by Abo-Shaeer et al \cite{expm,mit} as
110: well as at ENS by Madison et al \cite{exp} for sufficiently large
111: rotational frequency. The interesting dynamics of the formation and decay
112: of vortices has also been studied \cite{mit1}. Madison et al \cite{exp}
113: also confirmed that before a vortex pattern is formed the condensate
114: passes through dynamically unstable distorted configurations. The
115: dynamical instability transforms the distorted condensate to an
116: appropriately symmetric condensate with vortices \cite{sin,ueda}.
117:
118: There have been theoretical studies of how vortices are generated in a BEC
119: via these dynamically unstable configurations using the mean-field GP
120: equations in two \cite{ueda} and three \cite{fc} dimensions. The
121: theoretical study of the dynamical instability using mean-field approach
122: in the generation of vortex is of interest as one can then compare the
123: results of the models with experiments. This will provide a more stringent
124: test of the mean-field theories than the comparison of stable
125: configurations of BEC.
126:
127: However, there is another type of instability which has been studied in
128: relation to a nonrotating BEC under the application of an impulsive
129: force. One can now
130: suddenly change the trapping potential on a stable BEC. This can be done
131: in experiment by varying the current in the coil responsible for
132: confinement. One can also
133: change the strength of interatomic interaction suddenly by changing the
134: magnetic field near a Feshbach resonance \cite{yy1}. These set in
135: dynamical instabilities in the BEC, which can be studied both
136: experimentally \cite{ex} and theoretically \cite{th}.
137:
138: Similar impulsive force can be applied on a rotating BEC with vortices and the
139: resultant oscillation studied. One can study how the number of vortices change
140: after the application of the impulsive force on a rotating BEC. In this paper
141: we perform such mean-field analyses of the dynamics of a rotating BEC subject
142: to three types of impulsive forces. We consider a sudden change in (i) the
143: strength of interaction near a Feshbach resonance \cite{yy1} induced by a
144: variation of an external magnetic field, (ii) the trapping potential, and
145: (iii) the rotation of the BEC. Specifically, we study the dynamics when the
146: strength of interaction or the trapping potential is suddenly increased or
147: reduced. We also study the decay of vortices once the rotation of the
148: condensate is suddenly stopped. There have been certain experimental
149: measurements on such decay of vortices \cite{exp,mit1}. Finally, the trapping
150: potential can be removed suddenly in a rotating BEC with vortices and the
151: resultant expansion studied.
152:
153:
154:
155: We base our study on the numerical solution \cite{3,2a,2h} of the nonlinear
156: time-dependent mean-field GP equation \cite{gp}, which should provide a
157: faithful description of the formation and evolution of vortices. The
158: experiments on the vortices used a highly anisotropic trap which may be
159: simulated by a quasi two-dimensional model. Hence as in the recent studies on
160: the subject in Refs. \cite{ad,ueda} we employ the GP equation in two
161: dimensions. We use a split-step method where in the time evolution of the GP
162: equation the kinetic energy and angular momentum (derivative) terms in the two
163: directions are dealt with in two independent steps by the Crank-Nicholson
164: rule complimented by the known boundary conditions \cite{2h,koo}. All the
165: harmonic oscillator and the nonlinear terms are treated in a third separate
166: step. We find that this time-dependent split-step approach leads to good
167: convergence.
168:
169: In Sec. II we describe briefly the time-dependent GP equation for a rotating
170: BEC and a numerical method for its solution. In Sec. III we report the
171: numerical results of the present investigation and finally, in Sec. IV we give
172: a summary of our study.
173:
174: \section{Nonlinear Gross-Pitaevskii Equation}
175:
176: At zero temperature, the time-dependent Bose-Einstein condensate wave
177: function $\Psi({\bf r};\tau)$ in two dimensions at position ${\bf r}$ and
178: time $\tau $ may
179: be described by the following mean-field nonlinear GP equation
180: \cite{gp,11}
181: \begin{eqnarray}\label{a} \biggr[ -\frac{\hbar^2}{2m}\nabla^2
182: &+& V({\bf r})
183: + gN|\Psi({\bf
184: r};\tau)|^2- \hat \Omega L_{\hat z} \nonumber \\
185: &-&i\hbar\frac{\partial
186: }{\partial \tau} \biggr]\Psi({\bf r};\tau)=0. \end{eqnarray} Here $m$
187: is
188: the mass and $N$ the number of atoms in the
189: condensate,
190: $g$ the strength of interatomic interaction.
191: The trapping potential in two dimensions may be written as $ V({\bf
192: r}) =\frac{1}{2}m \omega ^2(\hat x^2+ \hat y^2)$ where
193: $\omega$ is the angular frequency. The rotational term
194: $-\hat \Omega L_{\hat z}
195: = i\hbar \hat \Omega (\hat x\partial _{\hat y} -\hat y\partial_{\hat x})$
196: where $\hat \Omega $ is the
197: rotational frequency along the $\hat z$ direction with $L_{\hat z}$ the
198: $\hat z$ component
199: of angular momentum.
200: The normalization condition of the wave
201: function is
202: $ \int d{\bf r} |\Psi({\bf r};\tau)|^2 = 1. $
203: The GP
204: equation (\ref{a}) can
205: accommodate
206: quantized vortex
207: states with rotational motion of the condensate around the $\hat z$ axis.
208:
209:
210: It is convenient to use the dimensionless variables
211: defined by $x =\hat x /l$, $y=\hat y/l$, $t=\tau \omega/2, $
212: $l\equiv \sqrt {\hbar/(m\omega)}$, $\Omega=2\hat \Omega/\omega $,
213: and
214: $\varphi(x,y;t)=l
215: { \Psi({\bf r};t)}$. In these units Eq. (\ref{a}) becomes
216: \begin{eqnarray}\label{d}
217: \biggr[ -\frac{\partial^2}{\partial
218: x^2}&-&\frac{\partial^2}{\partial
219: y^2}
220: +\left(x^2+y^2\right)+i\Omega\left( x\frac{\partial }{\partial y}
221: - y\frac{\partial }{\partial x}\right)
222: \nonumber \\
223: &+& {\cal N}\left| {\varphi({x,y};t)}\right|^2
224: -i\frac{\partial
225: }{\partial t} \biggr]\varphi({ x,y};t)=0,
226: \end{eqnarray}
227: where
228: $ {\cal N} = 2g Nm /\hbar^2.$
229: The normalization condition of the wave
230: function is \begin{equation}\label{5} \int_{-\infty} ^\infty
231: dx \int _{-\infty}^\infty dy|\varphi(x,y;t)|
232: ^2 =1. \end{equation}
233:
234:
235: The explicit inclusion of the rotational energy in the GP hamiltonian
236: above simulates the time evolution of a rotational BEC which permits
237: multiple-vortex formation. The initial input solution to the GP equation
238: is a circularly symmetric state obtained with $\Omega =0$. The method of
239: solution with $\Omega \ne 0$ is time iteration using an appropriate
240: algorithm. If one iterates the GP equation above in time using the full
241: angular momentum term the circular symmetry is maintained always and a
242: solution with multiple vortices is not generated. It is essential that the
243: circular symmetry of the solution is broken at some stage before multiple
244: vortices are generated during time iteration.
245:
246:
247:
248: The circular symmetry of the
249: solution can be broken in the process of time evolution of
250: the GP equation above by
251: introducing \cite{fc} a phase in the initial wave function for $\Omega
252: =0$
253: of the form
254: \begin{equation}\label{ph} \delta= \sum_{i} \arctan [(y-y_i)/(x-x_i)],
255: \end{equation} where $x_i$ and $y_i$ are a set of parameters such that the
256: positions $(x_i,y_i)$ fall inside the boundary of the BEC. This
257: initial phase preserves the norm of the wave function but accelerates the
258: formation of vortex nucleation after time evolution of the GP equation.
259: However, there is no direct relation between the parameters $(x_i,y_i)$
260: and the positions or the numbers of the vortices in the final BEC. Another
261: way of facilitating vortex formation is to include a small absorptive term
262: in the GP equation as has been advocated in Ref. \cite{2f} and employed in
263: Ref. \cite{ueda}.
264:
265:
266: We solve the GP equation (\ref{d}) with $\Omega = 0 $
267: using a
268: time-iteration method elaborated in Refs. \cite{2h,koo} starting
269: from the analytic two-dimensional harmonic-oscillator solution for ${\cal N}=0$. The full
270: GP Hamiltonian is conveniently broken into three parts $H_x$, $H_y$ and
271: $H_{\mbox{res}}$ $-$ the first containing the $x$-dependent derivative
272: terms, the second containing the $y$-dependent derivative terms with
273: the harmonic-oscillator and nonlinear interaction terms contributing to
274: the third residual part. The GP equations along the $x$ and $y$ directions
275: are discretized
276: on a two-dimensional set of grid points $N_x \times N_y$ using the
277: Crank-Nicholson discretization method \cite{koo}.
278: The second (first) derivatives are discretized using a three-point (two-point)
279: finite difference rule.
280: Effectively, each time iteration of the GP equation is broken up into
281: three parts $-$ using $H_x$, $H_y$ and $H_{\mbox{res}}$.
282: The first two parts lead to two sets of tridiagonal equations
283: along $x$ and $y$ directions which are solved alternately by the Gaussian
284: elimination method \cite{koo}. The third part does not contain any
285: derivative and is solved essentially exactly. For a small time
286: step $\Delta$ the error involved in this break-up procedure along $x$ and
287: $y$ directions is quadratic in $\Delta$ and hence can be neglected.
288: Using the above procedure the solution of the GP equation corresponding to
289: $\Omega
290: =0$ and ${\cal N}\ne 0$
291: is obtained. This solution serves as the initial input for the
292: simulation of multiple vortex generation using the time iteration of the
293: GP equation.
294:
295:
296: The initial solution with $\Omega =0$ is circularly symmetric. For the
297: generation of multiple vortex the circular symmetry is broken by
298: introducing the phase $\delta$ of Eq. (\ref{ph}) to the solution with
299: $\Omega =0$ and iterating the GP equation in time with full $\Omega$. We
300: also included a small dissipative term in the GP equation. As the
301: formation and decay of vortices is intrinsically dissipative \cite{mit1},
302: it is more realistic to include a small dissipative term \cite{ueda,dis}.
303: However, from a mathematical point of view, vortices and vortex lattice
304: can
305: be generated in numerical simulation without including any dissipative
306: term \cite{fc}. The time evolution of the GP equation is performed as
307: described
308: above using three independent steps. The system passes through dynamically
309: unstable deformed configurations and during time iteration the multiple
310: vortex centers appear for $\Omega$ greater than a critical value.
311:
312:
313:
314:
315:
316:
317: \section{Numerical Results}
318:
319: In the present numerical study on vortices we use an equilibrated
320: condensate with ${\cal N}=500$ trapped in the stationary potential as the
321: initial state. Figure 1 shows the dynamics of the condensate as it starts
322: to rotate suddenly with an angular frequency $\Omega =1.5$. In the
323: simulation we use as input the wave function with phase $\delta$ given by
324: Eq. (\ref{ph}). In the beginning the system passes through deformed
325: configurations with unstable oscillating boundaries. The vortices are
326: generated in the peripheral region via the oscillation of the boundary.
327: More and more vortices are generated which move towards the center and
328: which arrange themselves in a regular form in the condensate. For a
329: sufficiently large number of vortices they are arranged as in a lattice
330: and maintain that pattern during rotation. When such a regular
331: equilibrated final pattern is formed the condensate attains the symmetry
332: of the problem $-$ circular in the present case.
333: \begin{figure}[!ht]
334: \begin{center}
335: \includegraphics[width=\linewidth]{fig1}
336: \end{center}
337: \caption{Formation of the vortices for ${\cal N} = 500$ and $\Omega =1.5$ at
338: different times; size of each square is $10\times 10$.}
339: \end{figure}
340:
341: In actual experiment the size of the condensate reduces with time due to
342: inelastic atomic collisions \cite{mit}. However, throughout the present study
343: we renormalize the number of particles during time iteration so that the
344: size is conserved. This should not affect the generation of vortices.
345:
346: For $\Omega =1.5$ and ${\cal N}=500$ one has a hexagonal lattice structure with
347: seven vortices at large times. In recent theoretical studies \cite{fc,ad,ueda}
348: some general features of vortex formation have been established which are
349: confirmed in the present study. For a fixed large ${\cal N}$ there is a
350: critical $\Omega $ for the appearance of a certain number of vortices
351: \cite{exp,mit}. The critical value of $\Omega $ increases as the number of
352: vortices increases. For a fixed $\Omega $ the number of vortices increases as
353: ${\cal N}$ increases. It is interesting to recall that for ${\cal N}=0$ only
354: states with a single vortex at the center can be formed.
355:
356: An interesting set of experiments can be performed where on a stable rotating
357: BEC with vortices the strength of atomic interaction or the harmonic
358: oscillator trapping potential is suddenly changed. Once either of these
359: parameters are changed suddenly on a vortex lattice with hexagonal structure,
360: the system enters again a dynamically unstable asymmetric configuration with
361: highly oscillating deformed boundary. Through the oscillating deformed boundary
362: new vortices can enter the condensate or some of the existing vortices can get
363: out. Eventually, a new equilibrated BEC with vortices is formed in the
364: condensate with circular symmetry. However, as the condensate oscillates
365: after the application of the impulsive force it is more difficult to obtain
366: perfect symmetry numerically in the final state. Small imperfections remain
367: near the boundary and in the arrangement of the vortices in the final state.
368:
369: \begin{figure}[!ht]
370:
371: \begin{center}
372: \includegraphics[width=\linewidth]{fig2}
373: \end{center}
374:
375: \caption{ Dynamics of the condensate of Fig. 1 at different times $t$ upon
376: sudden increase of $\cal N$ from 500 to 1200 at $t=0$; size of each square is
377: $10\times 10$.}
378:
379: \end{figure}
380:
381: \begin{figure}[!ht]
382:
383: \begin{center}
384: \includegraphics[width=\linewidth]{fig3}
385: \end{center}
386: \caption{Dynamics of the condensate of Fig. 1 at different times $t$ upon
387: sudden decrease of $\cal N$ from 500 to 100 at $t=0$; size of each square is
388: $10\times 10$. }
389: \end{figure}
390:
391: In the simulation presented in Fig. 2 the nonlinear parameter $\cal N$ on
392: the hexagonal vortex lattice of Fig. 1 is suddenly changed to 1200 from
393: 500 at $t=0$ and the evolution of the resultant condensate studied. This
394: corresponds to an increase of the strength of interatomic interaction by a
395: factor 2.4. The size of the condensate increases and it accommodates more
396: vortices. The system seems to keep on oscillating with $(12\pm 1)$
397: vortices. This oscillation seems to be a natural consequence of the
398: sudden change of the strength of interaction. Next we consider in Fig. 3
399: the sudden change in the nonlinear parameter $\cal N$ from 500 to 100 at
400: $t=0$ on the final configuration of Fig. 1. With the reduction of the
401: strength of interaction by a factor 0.2, the condensate shrinks in size by
402: passing through dynamically unstable and asymmetric configurations and
403: starts to oscillate. The condensate tries to reach a equilibrated stage
404: by getting rid of vortices. However, at large times the system continues
405: to oscillate with ($6\pm 1$) vortices.
406: \begin{figure}[!ht]
407: \begin{center}
408: \includegraphics[width=\linewidth]{fig4}
409: \end{center}
410: \caption{ Dynamics of the condensate of Fig. 1 at different times upon sudden
411: increase of the harmonic oscillator trapping potential by a factor of 1.5 at
412: $t=0$; size of each square is $10\times 10$.}
413: \end{figure}
414: \begin{figure}[!ht]
415: \begin{center}
416: \includegraphics[width=\linewidth]{fig5}
417: \end{center}
418: \caption{ Dynamics of the condensate of Fig. 1 at different times upon sudden
419: decrease of the harmonic oscillator trapping potential by a factor of 0.75 at
420: $t=0$; size of each square is $10\times 10$. }
421: \end{figure}
422: \begin{figure}[!ht]
423: \begin{center}
424: \includegraphics[width=\linewidth]{fig6}
425: \end{center}
426: \caption{ Decay of the condensate of Fig. 1 at different times after the
427: rotation is suddenly stopped at $t=0$; size of each square is $10\times 10$.}
428: \end{figure}
429:
430: Next we consider a sudden change in the harmonic oscillator potential.
431: First, the harmonic oscillator potential is increased by a factor of 1.5
432: on the rotating BEC of Fig. 1 at $t=0$. The time evolution of the
433: resultant condensate is shown in Fig. 4. With the sudden increase in the
434: trapping potential the condensate starts to shrink and oscillate. It
435: passes through dynamically unstable asymmetric configurations and reaches
436: an equilibrium stage by getting rid of a couple of vortices. Eventually, a
437: condensate emerges with four vortices. The system executes small
438: oscillations at large times with a fixed number of vortices. In the next
439: simulation the harmonic oscillator potential is reduced by a factor of
440: 0.75 on the configuration of Fig. 1 at $t=0$. Because of the sudden
441: decrease in the trapping potential the condensate starts to swell and
442: oscillate. More vortices are generated near the boundary which move
443: inwards. Eventually, one has a larger condensate with 11 vortices at large
444: times.
445:
446: We study the decay of the vortices of Fig. 1 once the rotation of the BEC is
447: suddenly stopped at $t=0$. The dynamics of the decay of vortices has been
448: studied experimentally in Refs. \cite{exp,mit1}. The numerical simulation of
449: the decay of vortices is shown in Fig. 6. The vortices do not decay
450: immediately after stopping the rotation. Rather they stay for a long time after
451: the rotation stops. As time passes vortices get out of the condensate until one
452: with a single vortex results. This BEC with a single vortex survives for a
453: long time. If we consider $\omega =2\pi\times 219 $ s$^{-1}$ as in the
454: experiment of Madison et al \cite{exp}, one unit of present dimensionless
455: time is 0.000726 s. The system gets rid of six vortices in about 600 units of
456: time or in 0.4 s. However, the state with one vortex stays up to $t\approx
457: 3000$, which gives a lifetime of about 2400 units of time or about 1.6 s.
458: Madison et al \cite{exp} measured the half life of a single vortex state and
459: found it to be about 1 s, which is of the same order of magnitude as in the
460: present simulation. However, a quantitative comparison between the two is not
461: to the point because of approximate nature of the present simulation compared
462: to experiment. The present calculation used a quasi two-dimensional model
463: opposed to a full three-dimensional one. Also, the nonlinearity of the present
464: model is not the same as in the experiment. However, in a rotating BEC with
465: many vortices, most of the vortices decay within one second whereas a small
466: number of vortices survive for a long time. This was also experimentally
467: confirmed by Abo-Shaeer et al \cite{mit1} who used $\omega \approx 2\pi \times
468: 85$ s$^{-1}$.
469: \begin{figure}[!ht]
470: \begin{center}
471: \includegraphics[width=\linewidth]{fig7}
472: \end{center}
473: \caption{ Free expansion of the condensate Fig. 1 at different times after
474: the removal of the trap at $t=0$; size of each square is $20\times 20$. }
475: \end{figure}
476:
477: Finally, we consider the free expansion of the vortices in Fig. 1 when the
478: harmonic oscillator trap is suddenly removed. The dynamics of free expansion is
479: shown in Fig. 7. It is interesting that the vortex pattern is not destroyed
480: immediately after the removal of the trap. On the other hand, upon expansion
481: the radius of the vortex core is found to increase more rapidly than the radius
482: of the condensate. Due to the very small radius of the vortex core it is
483: difficult to observe the vortices and count them properly in an experiment
484: \cite{mit1}. However, because of the increase in the radius of the vortex core
485: after a free expansion the detection and counting of vortices will be easier
486: from a picture of the condensate after expansion \cite{mit1}. In Fig. 7 the
487: vortices are more prominently visible at $t=1.6$ than at $t=0.2$. Using the
488: radial trapping frequency $\omega = 2\pi \times 85$ s$^{-1}$ as in the
489: experiment of Abo-Shaeer et al \cite{mit1}, $t=1.6$ of Fig. 7 corresponds to
490: about 3 ms. In the actual experiment \cite{mit1} after 42.5 ms of expansion the
491: vortex cores were magnified 20 times their size in the trap. The present
492: simulation has again led to the right order of magnitude.
493:
494: \section{Summary}
495:
496: We have investigated the effect of an impulsive force on a rotating BEC with
497: vortices using the mean-field GP equation. We have considered the effect of
498: suddenly increasing or decreasing the strength of atomic interaction and the
499: harmonic oscillator trapping potential. Upon application of such an impulsive
500: force the condensate passes through highly unstable distorted configurations
501: with oscillating boundary. New vortices can enter the condensate via the
502: oscillating boundary. Also the condensate can get rid of some of the vortices
503: in a similar fashion. Eventually, a (more) symmetric condensate with a
504: different number of vortices result. When the atomic interaction is suddenly
505: increased (decreased) the final condensate grows (shrinks) in size and
506: accommodates more (less) vortices. When the trapping potential is suddenly
507: increased (decreased) the final condensate shrinks (grows) in size and
508: accommodates less (more) vortices. Experiments can be performed when similar
509: impulsive force in applied on a rotating BEC with vortices and compared with
510: the mean-field prediction.
511:
512: We also study the decay of vortices in a rotating BEC when the rotation is
513: suddenly stopped. Most of the vortices decay quickly (in a fraction of a
514: second), but a few vortices survive during a relatively long time before they
515: decay. In a typical experimental situation \cite{exp,mit1} the single vortex at
516: the end survives for a unusually long interval of time (a couple of seconds).
517:
518:
519: In another simulation of a free expansion of a rotating BEC we find that the
520: vortices are not immediately lost during expansion. Rather, upon expansion the
521: radius of the vortex core increases at a faster rate than that of the
522: condensate. Hence the vortices are more prominently visible in a picture of the
523: condensate during the free expansion of the condensate. Experimentally, it is
524: difficult to observe and count the vortices correctly as the vortex cores have
525: usually small radii \cite{mit1}. The present simulation suggests that the free
526: expansion could aid in the observation and counting of vortices in a rotating
527: BEC, specially, when there is a large number of vortices. This was noted in the
528: experiment of Abo-Shaeer et al \cite{mit1}. In a typical experimental
529: situation, on free expansion the vortices may be observed during few tens of
530: milliseconds.
531:
532: \acknowledgments
533:
534: The work is supported in part by the Conselho Nacional de Desenvolvimento
535: Cient\'\i fico e Tecnol\'ogico and Funda\c c\~ao de Amparo \`a Pesquisa do
536: Estado de S\~ao Paulo of Brazil.
537:
538:
539: \begin{thebibliography}{11}
540: \bibitem{1} J. R.
541: Ensher, D. S. Jin, M. R. Matthews, C. E. Wieman, and E. A. Cornell, Phys.
542: Rev. Lett. {\bf 77}, 4984 (1996); K. B. Davis, M. O. Mewes, M. R.
543: Andrews, N. J. van Druten, D. S. Durfee, D. M. Kurn, and W. Ketterle,
544: {\it ibid.} {\bf 75}, 3969 (1995); D. G. Fried, T. C. Killian, L.
545: Willmann, D. Landhuis, S. C. Moss, D. Kleppner, T. J. Greytak, {\it
546: ibid.} {\bf 81}, 3811 (1998);
547: F. Pereira Dos Santos, J. L\'eonard, Junmin Wang, C. J. Barrelet,
548: F. Perales, E. Rasel, C. S. Unnikrishnan, M. Leduc, and C.
549: Cohen-Tannoudji, {\it ibid.} {\bf 86}, 3459 (2001);
550: C. C. Bradley, C. A. Sackett,
551: J. J. Tollett, and R. G. Hulet,
552: {\it ibid.} {\bf 75}, 1687 (1995);
553: A. Robert, O. Sirjean, A. Browaeys, J. Poupard, S. Nowak,
554: D. Boiron, C. I. Westbrook, and A. Aspect, Science
555: {\bf 292}, 461 (2001); G. Modugno, G. Ferrari, G. Roati, R. J. Brecha,
556: A. Simoni, and
557: M. Inguscio, {\it ibid.} {\bf 294}, 1320 (2001).
558:
559:
560: \bibitem{exp} K. W. Madison, F. Chevy, W. Wohlleben, and J. Dalibard,
561: Phys. Rev. Lett. {\bf 84}, 806 (2000).
562: \bibitem{exp1} K. W. Madison, F. Chevy, V.
563: Bretin, and J. Dalibard, Phys. Rev. Lett. {\bf 86}, 4443 (2001).
564:
565: \bibitem{expm} C. Raman, J. R. Abo-Shaeer, J. M. Vogels, K. Xu, and
566: W. Ketterle, Phys. Rev. Lett. {\bf 87}, 210402 (2001).
567:
568:
569: \bibitem{jila} M. R.
570: Matthews, B. P. Anderson, P. C. Haljan, D. S. Hall, C. E. Wieman, E. A.
571: Cornell, Phys. Rev. Lett. {\bf 83}, 2498 (1999).
572:
573:
574: \bibitem{hal} P. C. Haljan, B. P. Anderson, I. Coddington, and
575: E. A. Cornell, Phys. Rev. Lett. {\bf 86}, 2922 (2001).
576: \bibitem{hal1} P. C. Haljan, I. Coddington, P. Engels and
577: E. A. Cornell, Phys. Rev. Lett. {\bf 87}, 210403 (2001).
578:
579: \bibitem{mit}J. R. Abo-Shaeer, C. Raman, J. M. Vogels, and W. Ketterle,
580: Science {\bf 292}, 476 (2001).
581:
582:
583: \bibitem{mit1}J. R. Abo-Shaeer, C. Raman, and W. Ketterle,
584: Phys. Rev. Lett. {\bf 88}, 070409 (2002).
585:
586:
587: \bibitem{gp}E. P. Gross, Nuovo Cimento {\bf 20}, 454 (1961); L. P.
588: Pitaevskii, Zh. Eksp. Teor. Fiz. {\bf 40}, 646 (1961)[Sov. Phys. JETP
589: {\bf 13}, 451 (1961)].
590:
591:
592: \bibitem{2} S. K. Adhikari, { Phys.
593: Lett. A} {\bf 281}, 265 (2001);
594: {\bf 265} 91 (2000);
595: Phys. Rev. A {\bf
596: 63}, 043611 (2001);
597: J. Phys. B {\bf 34}, 4231 (2001).
598:
599: \bibitem{3} P. A. Ruprecht, M. J. Holland, K. Burnett, and M. Edwards,
600: {Phys. Rev. A} { \bf 51}, 4704 (1995);
601: S. K. Adhikari, Phys. Rev. E {\bf 63}, 056704 (2001); {\bf 62}, 2937
602: (2000);
603: M. J. Holland,
604: D. S. Jin, M. L. Chiofalo, and J. Cooper,
605: Phys. Rev. Lett. {\bf 78}, 3801 (1997);
606: M. Holland and J. Cooper, Phys. Rev. A {\bf 53},
607: R1954 (1996).
608:
609:
610:
611:
612: \bibitem{11} F. Dalfovo, S. Giorgini, L. P. Pitaevskii, and S. Stringari,
613: Rev. Mod. Phys. {\bf 71}, 463 (1999).
614:
615:
616:
617: \bibitem{2a}F. Dalfovo and S. Stringari, Phys. Rev. A {\bf 53},
618: 2477 (1996);
619: F. Dalfovo and M. Modugno, {Phys. Rev. A} {\bf
620: 61},
621: 023605 (2000); S. K. Adhikari,
622: {\it ibid.}{\bf 65}, 033616 (2002);
623: E. Lundh, C. J. Pethick, and H. Smith, {\it ibid.} {\bf 58}, 4816
624: (1998); Y. Castin and R. Dum, Eur. Phys. J. D
625: {\it 7}, 399 (1999).
626:
627: \bibitem{2c}B. Jackson, J. F. McCann, and C. S. Adams, Phys. Rev. A {\bf
628: 61},
629: 013604 (1999).
630:
631:
632:
633: \bibitem{2d}
634: D. L. Feder, C. W. Clark, and B. I. Schneider,
635: Phys. Rev. Lett. {\bf
636: 82},
637: 4956 (1999).
638:
639: \bibitem{2e}
640: J. J. Garc\'ia-Ripoll and
641: V. M. P\'erez-Garc\'ia; Phys. Rev. Lett. {\bf
642: 84},
643: 4264 (2000);
644: V. M. P\'erez-Garc\'ia and J. J. Garc\'ia-Ripoll, Phys. Rev. A {\bf
645: 62},
646: 033601 (2000);
647: G. M. Kavoulakis, B. Mottelson, and C. J. Pethick, {\it ibid.}
648: {\bf
649: 62},
650: 063605 (2000).
651:
652: \bibitem{sin} S. Sinha and Y. Castin, Phys. Rev. Lett. {\bf 87}, 190402
653: (2001).
654:
655: \bibitem{2f}R. J. Marshall, G. H. C. New, K. Burnett, and S. Choi, Phys. Rev. A
656: {\bf 59}, 2085 (1999).
657:
658: \bibitem{2ff}R. Dum, J. I. Cirac, M. Lewenstein, and P. Zoller,
659: Phys. Rev. Lett. {\bf 80}, 2972 (1998).
660:
661: \bibitem{2f1}L. Dobrek, M. Gajda, M. Lewenstein,
662: K. Sengstock, G. Birkl, and W. Ertmer, Phys. Rev. A
663: {\bf 60}, 3381 (1999).
664:
665: \bibitem{2g} A. A. Svidzinsky and A. L. Fetter, Phys. Rev. A {\bf
666: 62},
667: 063617 (2000).
668:
669: \bibitem{2h}S. K. Adhikari, Phys. Rev. E {\bf 65}, 016703 (2002).
670:
671:
672:
673: \bibitem{fc}D. L. Feder and C. W. Clark, Phys. Rev. Lett. {\bf 87},
674: 190401 (2001).
675: \bibitem{ad}A. Aftalion and Q. Du, Phys. Rev. A {\bf 64}, 063603 (2001).
676:
677:
678: \bibitem{ueda}
679: M. Tsubota, K. Kasamatsu, and M. Ueda, Phys. Rev. A {\bf 65},
680: 023603 (2002).
681:
682:
683: \bibitem{dis}O. N. Zhuravlev, A. E. Muryshev, and P. O. Fedichev,
684: Phys. Rev. A {\bf 64}, 053601 (2001); P. O. Fedichev and
685: G. V. Shlyapnikov,
686: {\it ibid. } {\bf 60}, R1779 (1999).
687:
688:
689:
690: \bibitem{yy1}S. Inouye, M. R. Andrews, J. Stenger, H. J. Miesner,
691: D. M. Stamper-Kurn, and W. Ketterle, Nature {\bf 392}, 151
692: (1998); J. Stenger, S. Inouye, M. R. Andrews, H. J. Miesner,
693: D. M. Stamper-Kurn, and W. Ketterle, Phys. Rev. Lett. {\bf 82}, 2422
694: (1999);
695: J. L. Roberts, N. R. Claussen,
696: S. L. Cornish, E. A. Donley, E. A. Cornell,
697: C. E. Wieman, {\it ibid.} {\bf 86}, 4211 (2001).
698:
699:
700: \bibitem{ex}E. A. Donley, N. R. Claussen, S. L. Cornish, J. L. Roberts,
701: E. A. Cornell, and C. E. Wieman, Nature {\bf 412}, 295 (2001);
702: N. R. Claussen, E. A. Donley, S. T. Thompson, and C. E. Wieman,
703: LANL preprint
704: cond-mat/0201400.
705:
706: \bibitem{th}H. Saito and M. Ueda, Phys. Rev. A {\bf 65}, 033624 (2002); P.
707: Muruganandam and S. K. Adhikari, {\it ibid.} {\bf 65}, 043608 (2002);
708: S. K. Adhikari, LANL preprint cond-mat/0203071.
709:
710:
711:
712:
713: \bibitem{koo} S. E. Koonin and D. C. Meredith, Computational Physics
714: Fortran Version, Addison-Wesley Pub. Co., Reading, 1990, pp 169-180.
715:
716:
717:
718:
719:
720:
721:
722:
723: \end{thebibliography}
724:
725:
726:
727: \end{document}
728:
729: