1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: % INSTITUTE OF PHYSICS PUBLISHING %
3: % %
4: % `Preparing an article for publication in an Institute of Physics %
5: % Publishing journal using LaTeX' %
6: % %
7: % LaTeX source code `ioplau2e.tex' used to generate `author %
8: % guidelines', the documentation explaining and demonstrating use %
9: % of the Institute of Physics Publishing LaTeX preprint files %
10: % `iopart.cls, iopart12.clo and iopart10.clo'. %
11: % %
12: % `ioplau2e.tex' itself uses LaTeX with `iopart.cls' %
13: % %
14: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15: %
16: %
17: % First we have a character check
18: %
19: % ! exclamation mark " double quote
20: % # hash ` opening quote (grave)
21: % & ampersand ' closing quote (acute)
22: % $ dollar % percent
23: % ( open parenthesis ) close paren.
24: % - hyphen = equals sign
25: % | vertical bar ~ tilde
26: % @ at sign _ underscore
27: % { open curly brace } close curly
28: % [ open square ] close square bracket
29: % + plus sign ; semi-colon
30: % * asterisk : colon
31: % < open angle bracket > close angle
32: % , comma . full stop
33: % ? question mark / forward slash
34: % \ backslash ^ circumflex
35: %
36: % ABCDEFGHIJKLMNOPQRSTUVWXYZ
37: % abcdefghijklmnopqrstuvwxyz
38: % 1234567890
39: %
40: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
41: %
42: \documentclass[12pt]{iopart}
43: % Uncomment next line if AMS fonts required
44: \usepackage{iopams}
45: \usepackage{graphicx}% Include figure files
46: \newcommand{\bra}{\langle}
47: \newcommand{\ket}{\rangle}
48: \newcommand{\nn}{\nonumber}
49: \renewcommand{\i}{{\rm{i}}}
50: \renewcommand{\e}{{\rm{e}}}
51: %
52: %
53: \begin{document}
54: \newcommand{\bv}[1]{\mbox{\boldmath$#1$}}
55:
56: \title[Creation
57: of Vortices in BEC]{Continuous Creation
58: of a Vortex in a Bose-Einstein Condensate with Hyperfine Spin $F=2$}
59:
60: \author{Mikko M\"ott\"onen$^1$, Naoki Matsumoto$^2$, Mikio Nakahara$^{1,2}$
61: \footnote[3]{To
62: whom correspondence should be addressed (nakahara@math.kindai.ac.jp)}
63: and Tetsuo Ohmi$^3$
64: }
65: \address{
66: Materials Physics Laboratory, Helsinki University of Technology,
67: P.O. Box 2200 FIN-02015 HUT, Finalnd$^1$}
68: \address{
69: Department of Physics, Kinki University, Higashi-Osaka 577-8502, Japan$^2$}
70: \address{Department of Physics, Kyoto University, Kyoto 606-8502, Japan$^3$}
71: \begin{abstract}
72: It is shown that a vortex can be
73: continuously created in a Bose-Einstein condensate
74: with hyperfine spin $F=2$ in a Ioffe-Pritchard trap
75: by reversing the axial magnetic field
76: adiabatically. It may be speculated that the condensate cannot
77: be confined in the trap since the weak-field seeking state
78: makes transitions to the neutral and the strong-field seeking states
79: due to the degeneracy of these states along
80: the vortex axis when the axial field vanishes. We have solved the
81: Gross-Pitaevskii equation numerically with given external
82: magnetic fields to show that this is not the case.
83: It is shown that a considerable fraction of the condensate remains in the
84: trap even when the axial field is reversed rather slowly.
85: This scenario is also analysed in the presence of an optical plug
86: along the vortex axis. Then the condensate remains within
87: the $F_z=2$ manifold, with respect to the local magnetic field,
88: throughout the formation of a vortex
89: and hence the loss of atoms does not take place.
90: \end{abstract}
91:
92: %Uncomment for PACS numbers title message
93: %\pacs{00.00, 20.00, 42.10}
94:
95: % Uncomment for Submitted to journal title message
96: %\submitto{\JPA}
97:
98: % Comment out if separate title page not required
99: \maketitle
100:
101: \section{Introduction}
102:
103: It has been observed that the Bose-Einstein condensate (BEC)
104: of alkali atom gas becomes superfluid \cite{rf:1, rf:2}. Superfluidity
105: of this system is different from the previously known
106: superfluid $^4$He in many aspects. For example, the BEC is
107: a weak-coupling gas for which the Gross-Pitaevskii equation
108: is applicable while superfluid $^4$He is a strong-coupling system.
109: One of the most remarkable differences is that
110: the BEC has spin degree of freedom
111: originating from the hyperfine spin of the atom, and that this degree of
112: freedom couples to external magnetic fields.
113: Accordingly the order parameter of the condensate is also
114: controlled at will by external magnetic fields.
115: Superfluid $^3$He also has similar internal degrees of freedom,
116: which, however, are rather difficult to control by external fields \cite{rf:3}.
117:
118: Taking advantage of this observation, we proposed a simple method
119: to create a vortex in a BEC with the hyperfine spin $F=1$ \cite{tomoya, ogawa};
120: a vortex-free BEC is intertwined topologically by manipulating the magnetic
121: fields in the Ioffe-Pritchard trap to form a vortex with the winding
122: number 2. This is achieved by reversing the axial magnetic field adiabatically
123: while the planar quadrupole field is kept fixed.
124:
125: In the present paper, a similar scenario is analysed for
126: a BEC with $F=2$, taking $^{87}$Rb as an example.
127: The difference between the present case and that for $F=1$
128: will be emphasised in our analysis.
129: In the next section, we briefly review the order parameter of
130: BEC with hyperfine spin $F=2$ and the Gross-Pitaevskii (GP) equation which
131: describes the time-evolution of the order parameter.
132: In section 3,
133: the ground state order parameter of the BEC in the weak-field seeking state
134: confined in a harmonic potential is obtained.
135: Then the time-evolution of the condensate, as the axial field
136: is adiabatically reversed, is studied by solving the GP equation
137: numerically. Cases with different reverse time are analysed
138: to find the best possible reversing time for which
139: the the fraction of the remaining condensate in the vortex state is maximised.
140: It is shown that the condensate in the end of this scenario has
141: the winding number 4. In section 4,
142: the GP equation is solved in the presence of an optical plug along
143: the vortex axis. The BEC remains in the
144: $F_z=2$ weak-field seeking state, with respect to the local
145: magnetic field, throughout the development,
146: and hence no atoms will be lost during the formation of a vortex.
147: Section 5 is devoted to conclusions and discussions.
148:
149: \section{Order Parameter of $F=2$ BEC}
150:
151: \subsection{General $F=2$ condensate and Gross-Pitaevskii equation}
152:
153: Suppose a uniform magnetic field $\bv{B}$ parallel to the $z$-axis is applied
154: to a BEC of alkali atoms with the hyperfine spin $F=2$.
155: Then the hyperfine spin state of the atom is quantised along
156: this axis; the eigenvalue $m$ of $F_z$ takes a value $-2 \leq m \leq +2$,
157: where $F_z | m \ket = m |m \ket$.
158: Let us introduce the following conventions
159: $$
160: |2 \ket = \left(
161: \begin{array}{c}
162: 1\\ 0\\ 0\\ 0\\ 0
163: \end{array} \right),
164: \ |1 \ket = \left(
165: \begin{array}{c}
166: 0\\ 1\\ 0\\ 0\\ 0
167: \end{array} \right),
168: \ |0 \ket = \left(
169: \begin{array}{c}
170: 0\\ 0\\ 1\\ 0\\ 0
171: \end{array} \right),
172: \ |-1 \ket = \left(
173: \begin{array}{c}
174: 0\\ 0\\ 0\\ 1\\ 0
175: \end{array} \right),
176: \ |-2 \ket =
177: \left( \begin{array}{c}
178: 0\\ 0\\ 0\\ 0\\ 1
179: \end{array} \right).
180: $$
181: The order parameter $|\Psi \ket$ is expanded in terms of $|m \ket$ as
182: \begin{equation}
183: |\Psi \ket = \sum_{m=-2}^2 \Psi_m |m \ket =
184: (\Psi_2,\Psi_1,\Psi_0,\Psi_{-1},\Psi_{-2})^{\rm T},
185: \end{equation}
186: where T denotes the transpose.
187:
188: The representation of the angular momentum operators $F_k\ (k=x, y, z)$ for $F=2$
189: is easily obtained from the well-known formulae
190: \begin{eqnarray*}
191: \bra 2, m|F_+|2, m' \ket &=& \sqrt{(2-m)(3+m)} \delta_{m, m'+1},\\
192: \bra 2, m|F_-|2, m' \ket &=& \sqrt{(2+m)(3-m)} \delta_{m, m'-1},\\
193: \bra 2, m|F_z|2, m' \ket &=& m \delta_{m, m'},
194: \end{eqnarray*}
195: where $F_{\pm} = F_x \pm i F_y$.
196:
197: The dynamics of the condensate in the limit of zero temperature is given,
198: within the mean field approximation, by the time-dependent
199: Gross-Pitaevskii (GP) equation with spin degrees of freedom.
200: This equation, obtained by Ciobanu, Yip and Ho \cite{Ciobanu}
201: (see also \cite{ueda}),
202: is written in components $\Psi_m$ as
203: \begin{eqnarray}\label{eq:gptime}
204: i\hbar\frac{\partial}{\partial t}\Psi_{m} &=& \left[-\frac{\hbar^2}{2M}
205: \nabla^2 +V(r) \right] \Psi_m \nn\\
206: & & +g_1|\Psi_n|^2 \Psi_m
207: + g_2\left[\Psi_{n}^\dagger (F_{k})_{np} \Psi_{p}\right]
208: (F_{k})_{mq}\Psi_q \nn\\
209: & &+ 5g_3\Psi_n^\dagger \bra 2m2n|00\ket\bra 00|2p2q\ket\Psi_p\Psi_q
210: +\frac{1}{2}\hbar\omega_{Lk}(F_k)_{mn}\Psi_n,
211: \end{eqnarray}
212: where summations over $k=x,y,z$ and $-2 \leq n,p,q \leq 2$ are understood.
213: Here, $M$ is the mass of the atom and $V(r)$ is
214: the possible external potential.
215: The Larmor frequency is defined as $\hbar\omega_{Lk}=\gamma_\mu B_k$, where
216: $\gamma_\mu\quad \lsimeq \mu_B$ is the gyromagnetic ratio of the atom and
217: $\mu_B$ the Bohr magneton. The interaction parameters are
218: expressed in terms of the $s$-wave
219: scattering length $a_F$, $F$ being the total
220: hyperfine spin of the two-body scattering state, and are given by
221: \cite{Ciobanu}
222: \begin{eqnarray}
223: g_1 = \frac{4\pi\hbar^2}{M}\frac{4a_2+3a_4}{7} \nn \\
224: g_2 = \frac{4\pi\hbar^2}{M}\frac{a_2-a_4}{7} \nn \\
225: g_3 = \frac{4\pi\hbar^2}{M}\left(\frac{a_0-a_4}{5}-\frac{2a_2-2a_4}{7}\right),
226: \end{eqnarray}
227: where $a_0=4.73$nm, $a_2=5.00$nm and $a_4=5.61$nm for $^{87}$Rb atoms.
228: This should be compared with $F=1$ BEC where there are only two types
229: of scattering state and hence two interaction terms in the GP equation.
230:
231: \subsection{Weak-field seeking state}
232:
233: Suppose a strong magnetic field $\bv{B}$ is applied along the $z$-axis.
234: Then the components with $F_z = 1$ and $2$ are in the weak-field seeking state
235: (WFSS)
236: while those with $F_z = -1$ and $-2$ are in the strong-field seeking
237: state (SFSS). The presence of two WFSSs leads to an interesting
238: two-component vortex that is not observed in $F=1$ BEC as we see in
239: the next section.
240: The energy of the state with $F_z=0$ is independent of the magnetic field
241: and will be called the neutral state (NS) hereafter.
242: Suppose a uniform condensate is in the state with $F_z=2$.
243: The order parameter of the condensate takes the form
244: \begin{equation}
245: |\Psi_0 \ket = f_0 \left( 1, 0, 0, 0, 0
246: %\begin{array}{c} 1\\ 0\\ 0\\ 0\\ 0 \end{array}
247: \right)^{\rm T}
248: \label{eq:b0}
249: \end{equation}
250: where $|f_0|^2$ is the number density of the condensate. Now let us consider
251: a state which is quantised along an arbitrary local magnetic field
252: \begin{equation}
253: \bv{B}(\mathbf{r})= B \left( \begin{array}{c}
254: \sin \beta \cos \alpha\\
255: \sin \beta \sin \alpha\\
256: \cos \beta
257: \end{array} \right).
258: \label{eq:mag}
259: \end{equation}
260: Let $F_B \equiv \bv{B} \cdot \bv{F}/B$
261: be the projection of the hyperfine
262: spin vector along the local magnetic field.
263: The WFSS $|\Psi \ket$ which satisfies $F_B |\Psi \ket
264: = +2 |\Psi \ket$ is obtained by rotating $|\Psi_0 \ket$ by
265: Euler angles $\alpha, \beta$ and $\gamma$ and is given by
266: \begin{eqnarray}
267: |\Psi (\mathbf{r}) \ket&=&
268: \exp(-i \alpha F_z) \exp(-\beta F_y) \exp(-i\gamma F_z) |\Psi_0 \ket
269: \nonumber\\
270: &= &f_0 e^{-2 i\gamma} \left( \begin{array}{c}
271: e^{-2i \alpha} \cos^4 \frac{\beta}{2}\\
272: 2 e^{-i \alpha} \cos^3 \frac{\beta}{2}\sin \frac{\beta}{2}\\
273: \sqrt{6} \cos^2 \frac{\beta}{2}\sin^2 \frac{\beta}{2}\\
274: 2 e^{i \alpha} \cos \frac{\beta}{2}\sin^3 \frac{\beta}{2}\\
275: e^{2i \alpha} \sin^4 \frac{\beta}{2}
276: \end{array}
277: \right) \equiv f_0 |v \ket.
278: \label{eq:wfss}
279: \end{eqnarray}
280:
281: The GP equation restricted within the WFSS is obtained by substituting
282: Eq.~(\ref{eq:wfss}) into Eq.~(\ref{eq:gptime}), see the next section.
283:
284: \section{Vortex formation without optical plug}
285:
286: The formation of a vortex in the $F=2$ condensate is analysed in this and the
287: next sections. In the present section, we study the scenario without an
288: optical plug along the vortex axis. Although some fraction of the condensate
289: is lost from the trap in this scenario, the experimental setup will be much easier
290: without introducing an optical plug. In fact, it will be shown below that
291: a considerable amount of the condensate remains in the trap by
292: properly choosing the time-dependence of the magnetic field.
293:
294: \subsection{Magnetic fields}
295:
296: Suppose a condensate is confined in a Ioffe-Pritchard trap.
297: It is assumed that the trap is translationally invariant along
298: the $z$ direction and rotationally invariant around the $z$-axis.
299: The quadrupole magnetic field of the trap
300: takes the form
301: \begin{equation}
302: \bv{B}_{\perp} (\bv{r}) = B_{\perp}(r) \left( \begin{array}{c}
303: \cos (-\phi)\\
304: \sin (-\phi)\\
305: 0
306: \end{array}
307: \right),
308: \end{equation}
309: where $\phi$ is the polar angle.
310: The magnitude $B_{\perp}(r)$ is proportional to the radial distance
311: $r$ near the origin; $B_{\perp}(r) \sim B_{\perp}' r$.
312: The uniform time-dependent field
313: \begin{equation}
314: \bv{B}_z(t) = \left( \begin{array}{c}
315: 0\\
316: 0\\
317: B_z(t)
318: \end{array}
319: \right)
320: \end{equation}
321: is also applied along the $z$-axis to prevent
322: Majorana flips from taking place at $r \sim 0$
323: where $\bv{B}_{\perp}$ vanishes.
324: %Note that the field $\bv{B}_z$ is assumed to be time-dependent.
325: Now the total magnetic field is given by
326: \begin{equation}
327: \bv{B}(\bv{r}, t) = \bv{B}_{\perp}(\bv{r}) + \bv{B}_z(t)
328: = \left( \begin{array}{c}
329: B_{\perp}(r) \cos (-\phi)\\
330: B_{\perp}(r) \sin (-\phi)\\
331: B_z(t)
332: \end{array}
333: \right).
334: \label{eq:magx}
335: \end{equation}
336: Comparing this equation with Eq. (\ref{eq:mag}), it is found that
337: \begin{equation}
338: \alpha = -\phi \qquad \beta = \tan^{-1}\left[\frac{B_{\perp}(r)}{B_z(t)}
339: \right].
340: \label{eq:phi}
341: \end{equation}
342:
343: It has been shown in the previous works for $F=1$ BEC that a
344: vortex-free condensate in the beginning will end up with a
345: condensate with a vortex of the winding number 2 if $\bv{B}_z$
346: reverses its direction while $\bv{B}_{\perp}$ is kept unchanged \cite{tomoya,
347: ogawa}.
348: We expect that the same magnetic field manipulation to lead
349: to the vortex formation in a BEC with $F=2$.
350: The uniform axial field $\bv{B}_z(t)$ must reverse its direction
351: as
352: \begin{equation}
353: B_z(t) = \left\{
354: \begin{array}{cl}
355: B_z(0) \left(1-\frac{2t}{T}\right)& 0 \leq t \leq T\\
356: -B_z(0)& T < t.
357: \end{array} \right.
358: \end{equation}
359: to create a vortex along the $z$-axis. This gives a ``twist'' to
360: the condensate leading to the formation of a vortex with winding
361: number 4, see below.
362:
363: Before we start the detailed analysis, it will be useful
364: to outline the idea underlying our scenario.
365: Suppose one has WFSS with $F_B = 2$
366: in the trap at $t=0$. The magnetic field at $r \sim 0$
367: points $+z$ direction (i.e., $\beta \sim 0$)
368: and hence the WFSS takes the form $\Psi_0$
369: of Eq.~(\ref{eq:b0}). Then the angle $\gamma$ must satisfy
370: $\gamma = -\alpha$ for the BEC to be vortex-free, see Eq.~(\ref{eq:wfss}).
371: The field $B_z(t)$ vanishes at $t =T/2$, for which $\beta = \pi/2$,
372: and the hyperfine spin is parallel to the quadrupole field $\bv{B}_{\perp}$.
373: Accordingly one must choose $\alpha = -\phi$ for this condition to
374: be satisfied, see Eqs.~(\ref{eq:magx}) and (\ref{eq:phi}).
375: This also implies $\gamma = + \phi$.
376: When the field $B_z$ is completely reversed
377: at $t=T$, the magnetic field at $r \sim 0$ points down and hence
378: $\beta \sim \pi$ there. Substituting $\alpha= -\gamma =\phi$ and
379: $\beta = \pi$ into Eq.~(\ref{eq:wfss}), one finds the
380: order parameter at $t=T$;
381: \begin{equation}
382: |\Psi \ket = f_0 \e^{-2\i \phi} \left(
383: %\begin{array}{c} 0\\ 0\\ 0\\ 0\\
384: 0, 0,0,0,\e^{-2\i \phi}
385: %\end{array}
386: \right)^{\rm T},
387: \end{equation}
388: which shows that a vortex with the winding number 4 has been created.
389:
390: \subsection{Initial state}
391:
392: Suppose a vortex-free BEC is confined in a Ioff-Pritchard trap,
393: whose magnetic field takes the form (\ref{eq:magx}) and that
394: the condensate is in the eigenstate $F_B =2$ with respect
395: to the local magnetic field $\bv{B}$ with $B_z=B_z(t=0)$.
396: The condensate wave function is then obtained by solving the
397: stationary state Gross-Pitaevskii equation. Substitution of
398: Eq.~(\ref{eq:wfss}) with $\alpha = -\gamma = -\phi$ into
399: Eq.~(\ref{eq:gptime}) yields
400: \begin{eqnarray*}
401: -\frac{\hbar^2}{2M} \nabla^2(f_0 v_m) +(g_1+ 4 g_2) f_0^3 v_m
402: + \hbar \omega_L f_0 v_m = \mu f_0 v_m,
403: \end{eqnarray*}
404: where we have put $\Psi_m \equiv f_0 v_m$.
405: Note that the $g_3$ term vanishes identically for the present state.
406: The condensate wave amplitude
407: $f_0(r)$ is taken to be a real function without loss of generality.
408: The eigenvalue $\mu$ is identified with the chemical potential.
409: If one multiplies the above equation by $\{v_m\}^{\dagger}$ from the left and uses the
410: identity $\sum_m|v_m|^2 =1$ and other identities derived from this,
411: one obtains the reduced GP equation for $f_0(r)$;
412: \begin{equation}
413: -\frac{\hbar^2}{2M} \left[f_0'' +\frac{f_0'}{r} + (v_m^* \nabla^2 v_m) f_0
414: \right]
415: + (g_1 + 4 g_2) f_0^3 + \hbar \omega_L f_0 = \mu f_0,
416: \end{equation}
417: where
418: \begin{equation}
419: v_m^* \nabla^2 v_m = \left[\frac{2}{r^2}(3 \cos^2 \beta - 5) \sin^2
420: \frac{\beta}{2} - \beta'^2 \right]
421: \end{equation}
422: comes from the rotation of the five-dimensional local orthonormal frame
423: that defines the order parameter.
424: The reduced GP equation looks similar to the ordinary scalar GP
425: equation except that there is an extra term $\beta'^2$
426: in $v_m^* \nabla^2 v_m $.
427: %Interesting consequencies of the
428: %existence of this term is under study and will be published elsewhere.
429:
430: It is convenient to introduce the energy scale $\hbar \omega$
431: and the length scale $a_{\rm HO}$ define by
432: \begin{equation}
433: \omega = \sqrt{\frac{\gamma_{\mu}}{M B_z(0)} }B_{\perp}'
434: \qquad
435: a_{\rm HO} = \sqrt{\frac{\hbar}{M \omega}}.
436: \end{equation}
437: For a typical values $B_z(0) = 1{\rm G}, B_{\perp}' = 200{\rm G/cm}$
438: for $^{87}$Rb, one obtains $\hbar \omega \simeq 1.69\cdot10^{-24} \rm{erg}$
439: and $a_{\rm HO} \simeq
440: 0.68 \mu\rm{m}$. After scaling all the physical quantities by these units, one
441: obtains the dimensionless form of the reduced GP equation;
442: \begin{eqnarray}
443: \lefteqn{
444: -\frac{1}{2} \left[\tilde{f}_0'' +\frac{\tilde{f}_0'}{\tilde{r}} +
445: \left[\frac{2}{\tilde{r}^2}(3 \cos^2 \beta - 5) \sin^2
446: \frac{\beta}{2} - \beta'^2 \right]
447: \tilde{f}_0\right]}
448: \nn\\
449: & & + (\tilde{g}_1 + 4 \tilde{g}_2)
450: \tilde{f}_0^3 + \tilde{\omega}_L \tilde{f}_0 =
451: \tilde{\mu} \tilde{f}_0,
452: \end{eqnarray}
453: where the dimensionless quantities are denoted with tilde. For example,
454: $\tilde{r} = r/a_{\rm HO}, \tilde{f}_0 = f_0 a_{\rm HO}^{3/2}$ and
455: $\tilde{g}_k = g_k/(\hbar\omega a_{\rm HO}^3)$. The tilde will be dropped hereafter
456: unless otherwise stated explicitly.
457: \begin{figure}
458: \begin{center}
459: \includegraphics[width=7cm]{initial.eps}
460: \end{center}
461: \caption{\label{fig1}The initial
462: condensate wave function $f_0(r)$ in dimensionless
463: form. The radial coordinate $r$ is also dimensionless.}
464: \end{figure}
465:
466: Figure 1 shows the ground state condensate wave function
467: obtained by solving Eq.~(16) numerically.
468: We have chosen $f_0(r=0) = 6$
469: which yields the central density $n_0 \sim 1.17 \cdot 10^{14}{\rm cm}^{-3}$.
470: This is roughly of the same order as that realised experimentally.
471: The difference between the
472: chemical potential and the Larmor energy at the origin is
473: $\delta \mu = \mu - \omega_L = 3.95$, which amounts to
474: $\delta\mu = 6.68 \cdot 10^{-24} {\rm erg}$
475: in dimensionful form.
476:
477: \subsection{Time development}
478:
479: Now the time-dependent GP equation (\ref{eq:gptime}) is solved numerically with
480: the initial condition $\Psi_m = f_0 (r) v_m$ with $f_0$
481: been obtained in the previous subsection.
482: We have introduced a tanh-shaped cutoff to mimic the loss of atoms
483: from the trap;
484: particles reach at $L \gg a_{\rm HO}$ vanish from the system.
485: We have made several choices of the reversing time $T$ and
486: maximised
487: the fraction of the condensate left in the trap in the final
488: equilibrium state. The details of the algorithm are given in
489: \cite{ogawa} and will not be repeated here.
490:
491: \begin{figure}
492: \begin{center}
493: \includegraphics[width=11cm]{t1000wav.eps}
494: \end{center}
495: \caption{Time dependence of the order parameter
496: $|\Psi_m|$ for the reversing time $T/\tau = 1000$.}
497: \end{figure}
498: Figure 2 shows the wave functions $|\Psi_m|$ for the choice
499: $T/\tau = 1000$, where $\tau = 2\pi/\omega_L$ is the time scale
500: set by the Larmor frequency at $t=0$ and $r=0$.
501: One obtains $\tau \sim 7.14 \cdot 10^{-7} {\rm sec}$ for the
502: parameters given in the previous subsection.
503: The parameter $\tau$
504: is expected to be the measure of the adiabaticity.
505: There are two weak-field seeking states possible for $F=2$, those
506: with $F_B = 2$ and $F_B=1$. It turns out that the final
507: vortex state is a mixture of these two states.
508: %
509: When the axial field $B_z(t)$ vanishes at $t=T/2$, the
510: gaps among WFSSs, SFSSs and NS disappear at $r=0$ and the
511: level crossing takes place there. Then the adiabatic
512: assumption breaks down and some fraction of the condensate
513: transforms into SFSSs and NS as well as $F_B=1$ WFSS.
514: Those components in SFSSs and NS eventually leave the trap
515: and the final condensate is made of $F_B=2$ and $F_B=1$ components.
516: It is a remarkable feature of the $F=2$ BEC, compared to
517: its $F=1$ counterpart, that the vortex state thus created is
518: mixture of these two WFSSs.
519: %
520: The composite nature of the final vortex state is best revealed
521: by projecting $|\Psi(\bv{r}) \ket$ to $F_B = 1$ and $2$ states.
522: Let $|v \ket$ be the vector defined in Eq.~(\ref{eq:wfss})
523: and $|u \ket = \exp(-i \alpha F_z) \exp(-\beta F_y) \exp(-i\gamma F_z)|1 \ket$.
524: Then $\Pi_2(\bv{r}) \equiv \bra v(\bv{r})|\Psi(\bv{r}) \ket$ and
525: $\Pi_1(\bv{r}) \equiv \bra u(\bv{r})|\Psi(\bv{r}) \ket$
526: depict
527: the projected amplitudes
528: of $|\Psi(\bv{r}) \ket$ to the local $F_B = 2$ and $F_B=1$ state,
529: respectively. These amplitudes are shown in Fig.~3 for $|\Psi(\bv{r})
530: \ket$ at $t= 10 T$. It is interesting to note that the $F_B=2$ component has
531: a winding number 4 while $F_B=1$ has 3.
532: \begin{figure}
533: \begin{center}
534: \includegraphics[width=10cm]{proj.eps}
535: \end{center}
536: \caption{The projected amplitudes $|\Pi_2(r)|$ and $|\Pi_1(r)|$
537: obtained from the order parameter $|\Psi(r)\ket$ at $t=10T$.}
538: \end{figure}
539:
540: The fraction of the condensate
541: left in the trap at time $t$ has been plotted in Fig.~4
542: for $T/\tau = 1000$. It should be noted that $\sim 2/5$
543: of the condensate is left in the trap when the system reaches
544: an equilibrium at $t \gg T$.
545: \begin{figure}
546: \begin{center}
547: \includegraphics[width=7cm]{t1000n.eps}
548: \end{center}
549: \caption{The fraction of the condensate left in the
550: trap, as a function of the dimensionless time $t/\tau$,
551: for the reversing time $T/\tau = 1000$.}
552: \end{figure}
553:
554: Figure 5 shows the fraction of the condensate left in the trap
555: in the equilibrium state at $t \gg T$ for various $T$.
556: \begin{figure}
557: \begin{center}
558: \includegraphics[width=7cm]{nt.eps}
559: \end{center}
560: \caption{\label{fig1}The fraction of the condensate left in the
561: trap, as a function of $T/\tau$, when the BEC reaches equilibrium
562: at $t \gg T$. The curve is shown for guide.}
563: \end{figure}
564: It can be seen from this figure that a considerable amount of the condensate
565: is left in the trap for a wide variety of the reversing time $T$.
566:
567: In the next section, we analyse the creation of a vortex in
568: the presence of an optical plug along the centre of the system.
569: It will be shown that the vortex thus created is purely made
570: of $F_B = 2$ WFSS.
571:
572: \section{Vortex formation with optical plug}
573:
574: The loss of the condensate in the previous section takes place
575: since the energy gaps among WFSSs, NS and SFSSs disappear
576: at $r=0$ when $B_z$ vanishes at $t=T/2$. One may introduce
577: an optical plug along the vortex axis to prevent the condensate
578: from entering this ``dangerous'' region. An optical plug
579: may be simulated by a repulsive potential
580: \begin{equation}
581: V(r) = V_0 \exp\left(-\frac{r^2}{r_0^2}\right)
582: \end{equation}
583: where $V_0$ is determined by the power of the blue-detuned laser
584: while $r_0$ by its waste size. We take $V_0 =
585: 9.27 \cdot 10^{-21}\ {\rm erg}$ and $r_0 = 5 {\rm \mu m}$
586: in our computation below.
587:
588: Now the time-independent GP equation is given by
589: \begin{eqnarray}
590: \lefteqn{
591: -\frac{1}{2} \left[\tilde{f}_0'' +\frac{\tilde{f}_0'}{\tilde{r}} +
592: \left[\frac{2}{\tilde{r}^2}(3 \cos^2 \beta - 5) \sin^2
593: \frac{\beta}{2} - \beta'^2 +\tilde{V}(r)\right]
594: \tilde{f}_0\right]
595: }
596: \nn\\
597: & & \qquad+ (\tilde{g}_1 + 4 \tilde{g}_2)
598: \tilde{f}_0^3 + \tilde{\omega}_L \tilde{f}_0 =
599: \tilde{\mu} \tilde{f}_0
600: \end{eqnarray}
601: in dimensionless form,
602: where $\tilde{V}(r) = V(r)/\hbar \omega$.
603: The angle $\beta$ is given by $\beta(r) = \tan^{-1}[B_{\perp}(r)/B_z(0)]$.
604: We will drop tilde from
605: dimensionless quantities hereafter unless otherwise stated.
606: The ground state
607: condensate wave function is obtained by solving this equation
608: numerically. We find the relative chemical potential $\delta \mu =
609: \mu - \omega_L = 173$, which amounts to $\delta \mu = 2.92\cdot 10^{-22} {\rm
610: erg}$ in dimensionful form, and the condensate wave function $f_0$
611: shown in Fig. 5.
612:
613: \begin{figure}
614: \begin{center}
615: \includegraphics[width=8cm]{pwav10000.eps}
616: \end{center}
617: \caption{Time dependence of the condensate amplitude
618: $f_0$ in the presence of the
619: optical plug. The reversing time is $T/\tau = 10000$.
620: The condensate amplitudes at $t=0$ and $t=T$ are almost
621: degenerate.
622: }
623: \end{figure}
624: The time-dependent GP equation
625: \begin{eqnarray}
626: \lefteqn{i\frac{\partial f_0}{\partial t}=
627: -\frac{1}{2} \left[{f}_0'' +\frac{{f}_0'}{{r}} +
628: \left[\frac{2}{{r}^2}(3 \cos^2 \beta - 5) \sin^2
629: \frac{\beta}{2} - \beta'^2 +{V}(r)\right]
630: {f}_0\right]
631: }
632: \nn\\
633: & & \qquad+ ({g}_1 + 4 {g}_2) {f}_0^3 + {\omega}_L {f}_0
634: \end{eqnarray}
635: is solved with the initial wave
636: function obtained above.
637: Here $\beta = \beta(r, t) \equiv \tan^{-1}[B_{\perp}(r)/B_z(r)]$.
638: It should be stressed again that the
639: condensate remains within the $F_B=+2$ WFSS throughout the
640: temporal evolution. Figure 7 shows the time-dependence of
641: the components $\Psi_m$
642: for the choice $T/\tau = 10000$, namely $T = 7.14 {\rm ms}$
643: in dimensionful form.
644: In contrast with the case without optical plug, the time-dependence
645: of the order parameter is independent of the choice of $T$
646: so far as $T/\tau$ is large enough so that the adiabaticity is
647: observed.
648:
649: The vortex thus obtained has a region near the origin ($r \sim 0$)
650: where the condensate cannot approach due to the presence of the
651: optical plug. The vortex current flows around a multiply connected
652: region. This situation is analogous to the superconducting
653: current flowing around a ring.
654: It is natural to expect that a vortex without the
655: optical plug may be obtained if one withdraws the optical plug
656: after the persistent current is established at $t=T$.
657: (Note that the optical plug has been introduced to prevent
658: Majorana flips at $r \sim 0$ at $t \sim T/2$. Accordingly the optical
659: plug is not required anymore for $t \geq T$.)
660: Let us suppose that the optical plug is slowly turned off
661: after $t=T$ with the time constant $t_0$;
662: \begin{equation}
663: V(r, t) = \left\{ \begin{array}{cl}
664: V_0 \exp (-r^2/r_0^2)& 0< t <T\\
665: V_0 \exp (-r^2/r_0^2) \exp [-(t-T)/t_0]& T < t.
666: \end{array} \right.
667: \end{equation}
668: It is found that the condensate oscillates back and forth
669: for small $t_0$. For sufficiently large $t_0$, however, the
670: condensate smoothly rearranges itself to a vortex state without
671: the optical plug.
672: Figure 7 shows our numerical result for $T/\tau = t_0/\tau = 10000$,
673: for which one still observes such oscillations.
674: \begin{figure}
675: \begin{center}
676: \includegraphics[width=8cm]{pn10000.eps}
677: \end{center}
678: \caption{The time dependence of the particle numbers in unit length
679: $N_m(t) = 2\pi\int |\Psi_m(r, t)|^2 r dr$ for $T/\tau = 10000$.}
680: \end{figure}
681:
682: \begin{figure}
683: \begin{center}
684: \includegraphics[width=8cm]{pwav10000po.eps}
685: \end{center}
686: \caption{Time dependence of the condensate amplitude
687: $f_0$ for the reversing time $T/\tau = 10000$. The potential decays
688: exponentially with the time constant $t_0 = T$ for $t > T$ , see Eq. (20).}
689: \end{figure}
690:
691: A vortex with a winding number 4 is thus created without
692: losing any atoms from the trap. It should be noted, however,
693: that it is technically difficult, albeit not impossible
694: \cite{mit}, to
695: introduce an optical plug with a few microns of radius along
696: the centre of the BEC whose radial dimension without optical plug
697: is of the order of a few microns.
698:
699: \section{Conclusions and Discussions}
700:
701: The formation of a vortex in a BEC with $F=2$ in a Ioffe-Pritchard trap
702: has been considered by fully utilising the spinor degrees of
703: freedom. It was shown that a vortex with winding number 4
704: is created continuously from a condensate without a vortex,
705: by simply reversing the axial magnetic field $B_z(t)$.
706: This scenario has been studied with and without
707: an optical plug at the centre of the vortex.
708: Some amount of the BEC is lost from the trap in the absence of
709: an optical plug while no atoms are lost in the presence of it.
710: Our numerical analysis shows that there remains a considerable
711: fraction of BEC even without the optical plug. The introduction
712: of an optical plug in a trapped BEC is difficult, albeit not
713: impossible.
714:
715: Our vortex has a large winding number 4 and is expected to be
716: unstable against decay into four singly quantised vortices
717: in the absence of an optical plug.
718: Whether a vortex with such a large winding number may be observable
719: depends on how large the lifetime of the metastable state is
720: compared to the trapping time of the BEC. Our prelimiary analysis
721: of the Bogoliubov equation suggests that the lifetime is of
722: the order of 100ms and these highly-quantised vortices exist for
723: a considerable duration of time.
724:
725: We would like to thank Kazushige Machida and Takeshi Mizushima
726: for discussions.
727: One of the authors (MN)
728: thanks Takuya Hirano, Ed Hinds and Malcolm Boshier
729: for discussions. He also thanks Martti M. Salomaa for support
730: and warm hospitality in the Materials Physics Laboratory at
731: Helsinki University of Technology, Finland. MN's work
732: is partially supported by Grand-in-Aid from Ministry of Education,
733: Culture, Sports, Science and Technology, Japan (Project Nos.
734: 11640361 and 13135215).
735: \vspace{1cm}\\
736: {\it Note Added} --- After we submitted our manuscript,
737: the MIT group reported the formation of vortices
738: according to the present scenario \cite{mit2}. They employed
739: hyperfine spin states $F=1$ and $F=2$ of $^{85}$Rb and
740: found that the vortex thus created had the winding number two in the
741: former case while four in the latter case, in consistent
742: with our prediction. The vortex state has considerably long lifetime,
743: at least 30ms after its formation, in spite of higher winding number,
744: which suggests that these vortices are rather stable.
745: The stability analysis
746: of highly-quantised vortices is outside the scope of the present work and
747: will be published elsewhere.
748:
749: We are grateful to Aaron Leanhardt for informing us of their result.
750:
751: \section*{References}
752: \begin{thebibliography}{99}
753: \bibitem{rf:1} Inguscio~M, Stringari~S, and Wieman~C E (eds.),
754: {\it Bose-Einstein Condensation in Atomic Gases}, (IOS Press, Amsterdam,
755: 1999).
756: \bibitem{rf:2} Martellucci~S,
757: Chester~A~N, Aspect~A, and Inguscio~M (eds.),
758: {\it Bose-Einstein Condensates and Atomic Lasers}, (Kluwer Academic/
759: Plenum Publishers, New York, 2000).
760: \bibitem{rf:3} Vollhardt~D and P. W\"olfle,
761: {\it The Superfluid Phases of Helium 3}, (Taylor and Francis, London, 1990).
762: \bibitem{tomoya} Isoshima~T, Nakahara~M, Ohmi~T and Machida~K 2000
763: {\it Phys. Rev. A} {\bf 61} 10000
764: \bibitem{ogawa} Ogawa~S-I, M\"ott\"onen~M, Nakahara~M,
765: Ohmi~T and Shimada~H 2002 Phys. Rev. A {\bf 66} 013617
766: \bibitem{Ciobanu} Ciobanu~C~V, Yip~S-K and Ho~T-L 2000 {\it Phys. Rev. A} {\bf 61}
767: 033607
768: \bibitem{ueda} Koashi~M and Ueda~M 2000 {\it Phys. Rev. Lett.} {\bf 84}
769: 1066
770: \bibitem{mit} Davis~K B {\it et al} 1995 {\it Phys. rev. Lett.} {\bf 75}
771: 3969
772: \bibitem{mit2} A.E. Leanhardt {\it et al} 2002 cond-mat/0206303
773: \end{thebibliography}
774:
775: \end{document}
776:
777:
778:
779: