1: % ****** Start of file apssamp.tex ******
2: %
3: % This file is part of the APS files in the REVTeX 4 distribution.
4: % Version 4.0 of REVTeX, August 2001
5: %
6: % Copyright (c) 2001 The American Physical Society.
7: %
8: % See the REVTeX 4 README file for restrictions and more information.
9: %
10: % TeX'ing this file requires that you have AMS-LaTeX 2.0 installed
11: % as well as the rest of the prerequisites for REVTeX 4.0
12: %
13: % See the REVTeX 4 README file
14: % It also requires running BibTeX. The commands are as follows:
15: %
16: % 1) latex apssamp.tex
17: % 2) bibtex apssamp
18: % 3) latex apssamp.tex
19: % 4) latex apssamp.tex
20: %
21: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
22: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
23:
24: % Some other (several out of many) possibilities
25: %\documentclass[preprint,aps]{revtex4}
26: %\documentclass[preprint,aps,draft]{revtex4}
27: %\documentclass[prl]{revtex4}% Physical Review A
28:
29: \usepackage{graphicx}% Include figure files
30: \usepackage{dcolumn}% Align table columns on decimal point
31: \usepackage{bm}% bold math
32:
33: %\nofiles
34:
35: \begin{document}
36:
37: %\preprint{APS/123-MW}
38:
39: \title{Dynamical instability and domain formation in a spin-1
40: Bose condensate}
41:
42: \author{Wenxian Zhang, D. L. Zhou, M. -S. Chang, M. S.
43: Chapman, and L. You}
44:
45: \affiliation{School of Physics, Georgia Institute of Technology,
46: Atlanta, Georgia 30332-0430}
47:
48: %\affiliation{$^2$Institute of Theoretical Physics, The Chinese Academy of
49: %Sciences, Beijing 100080, China}
50:
51: %\author{Wenxian Zhang, D. L. Zhou, M. -S. Chang, M. S. Chapman, and L. You}
52:
53: %\affiliation{School of Physics, Georgia Institute of Technology, 837 State
54: %Street, Atlanta, Georgia 30332-0430}
55:
56: \date{\today}
57:
58: \begin{abstract}
59: We interpret the recently observed spatial domain formation in spin-1 atomic
60: condensates as a result of dynamical instability. Within the mean field theory,
61: a homogeneous condensate is dynamically unstable (stable) for ferromagnetic
62: (antiferromagnetic) atomic interactions. We find this dynamical instability
63: naturally leads to spontaneous domain formation as observed in several recent
64: experiments for condensates with rather small numbers of atoms. For trapped
65: condensates, our numerical simulations compare quantitatively to the
66: experimental results, thus largely confirming the physical insight from our
67: analysis of the homogeneous case.
68: \end{abstract}
69:
70: \pacs{03.75.Mn, 03.75.Kk, 47.54.+r, 89.75.Kd}
71: % PACS, the Physics and Astronomy Classification Scheme.
72: % 03.75.Kk Dynamic properties of condensates; collective and hydrodynamic
73: % excitations, superfluid flow
74: % 03.75.Mn Multicomponent condensates; spinor condensates
75: % 05.30.Jp Boson systems (for static and dynamic properties of Bose-Einstein
76: % condensates, see 03.75.Hh and 03.75.Kk)
77: % 45.70.Qj Pattern formation of granular systems
78: % 47.54.+r Pattern selection; pattern formation of rarefied gas dynamics
79: % 42.60.Jf Beam characteristics: profile, intensity, and power;
80: % spatial pattern formation of random laser
81: % 89.75.Kd Patterns of complex systems
82: %\keywords{Suggested keywords}%Use showkeys class option if keyword
83: %display desired
84: \maketitle
85:
86:
87: %\section{Introduction}
88:
89: Spatial domains or pattern formation is a common feature of nonlinear dynamics
90: in extended systems. It has been actively researched in nonlinear optics
91: \cite{AgrawalBook89}, classical fluids \cite{Gollub99}, granular materials
92: \cite{Pooley04}, and recently in atomic Bose-Einstein condensates
93: (BECs) \cite{Strecker02, Lewandowski02, Stenger98, Sarlo04, Kasamatsu04,
94: Kourakis05, Robins01}. It is generally understood that the unstable modes
95: of a dynamically unstable system can grow exponentially and eventually lead to
96: the appearance of spatial domain structures that last for a long time.
97:
98: Many earlier studies have suggested interesting mechanisms for
99: spontaneous domain formation in atomic condensates \cite{Strecker02,
100: Lewandowski02, Stenger98, Sarlo04}. Most focus on the single- or
101: two-component condensates, where the number(s) of atoms for each component is
102: conserved. The dynamical instability due to
103: attractive atomic interactions is the most prominent among
104: all proposed scenarios for domain formation \cite{Strecker02, Lewandowski02,
105: Kasamatsu04}. The attractive interaction in a single-component condensate is
106: also believed to be responsible for the formation of a train of solitons,
107: consistent with the fact that it is dynamically unstable \cite{Strecker02}. For a
108: two-component condensate, again it is found that an effective attractive
109: interaction is responsible for the dynamical instability and domain formation
110: \cite{Lewandowski02, Kasamatsu04, Kourakis05}.
111:
112: Several groups have also studied three-component, or spin-1 condensates
113: ($F=1$), which are distinct as the spin mixing interaction \cite{Stenger98,
114: Robins01,Saito05} allows for exchanging atoms among spin components
115: $2|m_F=0\rangle \leftrightarrow |m_F=+1\rangle + |m_F=-1\rangle$ (hereafter as
116: $|0\rangle$, $|+\rangle$, and $|-\rangle$). The number of atoms for each
117: component therefore can change, but the total number of atoms and the system
118: magnetization are conserved. Significant interest now exists for spin-1
119: condensates because of the recent progress from several experimental groups
120: \cite{Barrett01}, in particular, the observation of spontaneous domain
121: formation in $^{87}$Rb condensates \cite{Chapman05}. Robins {\it et al.} were
122: among the first to study dynamical instability in a spin-1 condensate
123: \cite{Robins01}. They discovered a particular type of stationary state
124: dynamically unstable for ferromagnetic interactions, evidenced by the sudden
125: collapse when propagated with Gross-Pitaevskii (GP) equations, presumably
126: resulting from the amplification of numerical discretization errors. Through
127: extensive numerical simulations, Saito and Ueda also investigated very recently
128: the spontaneous multi-domain formation induced by the dynamical instability in
129: a spin-1 condensate with ferromagnetic interactions \cite{Saito05}. A clear
130: picture, however, is still lacking as indicated by the general lack of
131: comparisons with experimental reports. Our work aims at providing a complete
132: understanding for domain formation in a spin-1 condensate.
133:
134: To begin with, we consider a homogeneous condensate at an off-equilibrium state
135: initially. For example, a spin-1 condensate in the ground state at a certain
136: nonzero magnetic (B-) field for time $t<0$ will become off-equilibrium when the
137: external B-field is changed for $t\ge 0$. This causes the spin-1
138: condensate to collectively oscillate analogous to a nonrigid pendulum as we
139: recently showed \cite{Zhang05a}. In addition, we assume the condensate size is
140: much larger than the spin healing length at least in one direction so that
141: domains may be formed. Within our mean field description, the evolution of a
142: spin-1 condensate is described by the coupled GP equations \cite{Ho98}
143: \begin{eqnarray}
144: i\hbar {\partial \over \partial t}\Phi_\pm &=&
145: \left[{\cal H}
146: + c_2(n_\pm+n_0-n_\mp)\right]\Phi_\pm+c_2\Phi_0^2\Phi_\mp^*,\nonumber\\
147: i\hbar {\partial \over \partial t}\Phi_0 &=&
148: \left[{\cal H}
149: + c_2(n_++n_-)\right]\Phi_0+2c_2\Phi_+\Phi_-\Phi_0^*,\label{eq:gpe}
150: \end{eqnarray}
151: where ${\cal H}=-{(\hbar^2/ 2{\sf m})}\nabla^2+V_{\rm ext}+c_0n$,
152: $\Phi_j$ is the j-th spin component condensate wave function, and $n_j=|\Phi_j|^2$.
153: $c_0=4\pi\hbar^2(a_0+2a_2)/3{\sf m}$ and $c_2=4\pi\hbar^2(a_2-a_0)/3{\sf m}$
154: with $a_0$ and $a_2$ the scattering lengths for the two colliding atoms
155: in the symmetric channels of total spin $0$ and $2$,
156: respectively. The interaction is ferromagnetic (antiferromagnetic) if $c_2<0$ ($>0$).
157:
158: Let $\Phi_j=\sqrt {n_j}e^{i\theta_j}$ and define a relative phase $\theta =
159: \theta_++\theta_--2\theta_0$, Eqs. (\ref{eq:gpe})
160: simplifies to the following \cite{Zhang05a}
161: \begin{eqnarray}
162: \dot n_0 &=& {2c_2\over \hbar}n_0 \sqrt{(n-n_0)^2-m^2} \sin \theta,
163: \label{eqn0}\\
164: \dot \theta &=& {2c_2\over \hbar}\left[(n-2n_0)
165: +{(n-n_0)(n-2n_0)-m^2 \over \sqrt{(n-n_0)^2-m^2}} \cos\theta\right], \nonumber
166: \end{eqnarray}
167: due to the conservation of
168: atomic density ($n=n_++n_0+n_-$) and the magnetization ($m=n_+-n_-$).
169: Equations (\ref{eqn0}) defines an energy conserving dynamics,
170: with the effective energy per unit of volume given by
171: \begin{eqnarray}
172: {\cal E}={E\over V} &=&{1\over 2}c_0n^2 + {1\over 2}c_2
173: \left[m^2+2n_0(n-n_0)\right. \nonumber \\
174: && \left.+ 2n_0\sqrt{(n-n_0)^2-m^2}
175: \cos\theta \right]. \label{eq:he} % Homogeneous Energy
176: \end{eqnarray}
177: We note Eq. (\ref{eqn0}) for a homogeneous condensate differs from a trapped
178: one even under single spatial mode approximation despite sharing the same
179: dynamical Eq. (\ref{eqn0}).
180:
181: %\begin{figure}
182: %\includegraphics[width=1.5in]{SpinDomain_Fig1.eps}
183: %\caption{Iso-energy contour plot of spin-1 $^{87}$Rb condensate with $m=0$
184: %(left) and the total spin $f$.} \label{fig1}
185: %\end{figure}
186: %\begin{figure}
187: %\includegraphics[width=1.5in]{./SpinSlice.eps}
188: %\caption{(Color) The dependence of the total spin $f$ of a homogeneous spin-1
189: %$^{87}$Rb condensate on the magnetization $m$, the relative phase $\theta$, and
190: %the $|0\rangle$ component fraction $\rho_0$. $f$ ranges from 0 to 1 as denoted
191: %by the legend to the right.} \label{fig:ss}
192: %\end{figure}
193:
194:
195: %It is easy to find that the off-equilibrium is periodic and the period is
196: %generally \cite{Pu99, Zhang05}
197: %\begin{eqnarray}
198: %T=.
199: %\end{eqnarray}
200: Within the mean field approximation, the average spin of a condensate
201: $\vec f = f_x \hat x + f_y \hat y + m \hat z$, where
202: $f_j=\langle F_j\rangle$ with $F_{x,y,z}$ being the
203: spin-1 matrices,
204: is also conserved in addition to the conservations of $n$ and $m$ \cite{Pu99}.
205: The energy functional Eq. (\ref{eq:he}) thus becomes
206: $
207: {\cal E} = {1\over 2}c_0n^2 + {1\over 2}c_2 f^2,
208: $
209: if $f^2\equiv f_x^2+f_y^2+m^2$. We note that the mean field theory
210: model cannot be applied to extreme cases
211: such as $N_0=0 $ and $N_0=N$, where quantum effects are important.
212:
213: We adopt three approaches to study dynamical stability of the off-equilibrium
214: collective oscillations of a condensate: the effective potential method; the
215: Bogoliubov method; and direct numerical simulations. By going into a rotating
216: frame, an entire orbit reduces to a stationary point in the phase space
217: \cite{Zhang05a}. The effective potential then becomes ${\cal F} = (c_0/2)n^2 +
218: (c_2/2) \left(m^2+f_x^2+f_y^2\right)-\mu n-\eta m-\delta_xf_x-\delta_yf_y$,
219: where parameters $\{\mu, \eta, \delta_x, \delta_y\}=\{c_0n, c_2m, c_2f_x,
220: c_2f_y\}$ defines the rotating frame and are obtained through
221: \begin{eqnarray}
222: {\partial {\cal F} \over \partial n} = 0,\;\; {\partial {\cal F} \over
223: \partial m} = 0,\;\; {\partial {\cal F} \over \partial f_x} &=& 0,\;\;
224: {\partial {\cal F} \over \partial f_y} = 0. \nonumber
225: \end{eqnarray}
226:
227: Our system is dynamically stable if its Hessian matrix of ${\cal F}$ with
228: respect to $\{n, m, f_x, f_y\}$ is positive definite and dynamically unstable
229: if the Hessian matrix has any negative eigenvalue. It is easy to check that the
230: eigenvalues of the Hessian matrix are $\{c_0, c_2, c_2, c_2\}$. Thus
231: an antiferromagnetically
232: interacting spin-1 condensate is dynamically stable,
233: while a ferromagnetically interacting one is
234: dynamically unstable since $c_2<0$.
235:
236: We next employ the Bogoliubov transformation to find out the corresponding
237: unstable modes. Starting from the stationary point in the rotating frame as found
238: above, the equation of motion for collective excitations
239: can be cast in a matrix form \cite{Maldonado93} as $ {\cal M}\cdot
240: \vec x = \hbar \omega \vec x$, with a vector $\vec x =
241: (\delta\Psi_+, \delta\Psi_0, \delta\Psi_-, \delta\Psi_+^*, \delta\Psi_0^*,
242: \delta\Psi_-^*)^T$. $\delta\Psi_j$ and $\delta\Psi_j^*$ denote the
243: deviations from the stationary point, and the
244: associated matrix is
245: \begin{eqnarray}
246: {\cal M} &=& \left(\begin{array}{cc}{\cal A}&{\cal B}\\-{\cal B}^*&-{\cal
247: A}^*\end{array}\right), \nonumber
248: \end{eqnarray}
249: with
250: \begin{widetext}
251: \begin{eqnarray}
252: {\cal A} &=& \left(\begin{array}{ccc}\varepsilon_k + (c_0+c_2)n_+ + c_2n_0
253: &c_0\Phi_0^* \Phi_+ + c_2\Phi_0\Phi_-^*
254: &(c_0-c_2) \Phi_-^* \Phi_+ \\
255: c_0\Phi_0 \Phi_+^* + c_2\Phi_0^*\Phi_- &\varepsilon_k + c_0n_0+ c_2(n_++n_-)
256: &c_0\Phi_0 \Phi_-^* + c_2\Phi_0^* \Phi_+ \\
257: (c_0-c_2) \Phi_- \Phi_+^* &c_0\Phi_0^* \Phi_- + c_2\Phi_0 \Phi_+^*
258: &\varepsilon_k + (c_0+c_2)n_- + c_2n_0\end{array}\right), \nonumber
259: \end{eqnarray}
260: and
261: \begin{eqnarray}
262: {\cal B} &=& \left(\begin{array}{ccc}(c_0 + c_2) \Phi_+^2 &(c_0 + c_2) \Phi_0
263: \Phi_+
264: &c_2\Phi_0^2 + (c_0 - c_2)\Phi_-\Phi_+\\
265: (c_0 + c_2) \Phi_0 \Phi_+ &c_0\Phi_0^2 + 2c_2\Phi_-\Phi_+
266: &(c_0 + c_2) \Phi_0\Phi_-\\
267: c_2\Phi_0^2 + (c_0 - c_2)\Phi_-\Phi_+ &(c_0 + c_2) \Phi_0\Phi_- &(c_0
268: +c_2)\Phi_-^2\end{array}\right). \nonumber
269: \end{eqnarray}
270: \end{widetext}
271: $\varepsilon_k = \hbar^2 k^2/2{\sf m}$ is kinetic energy of the collective
272: excitation mode with wave vector $k$.
273:
274: The eigen-frequencies of the Bogoliubov excitations are obtained
275: from the characteristic equation $\det({\cal M}-\hbar \omega I)=0$,
276: explicitly given by
277: \begin{eqnarray}
278: \left[2c_s\varepsilon_k + \varepsilon_k^2+c_s^2{\sf
279: f}^2-(\hbar\omega)^2\right]&&
280: \nonumber \\
281: \times \left[\left(\varepsilon_k^2-(\hbar\omega)^2\right)
282: \left(2c_s\varepsilon_k+\varepsilon_k^2-(\hbar\omega)^2\right)\right. &&\nonumber \\
283: \left.+2c_n\varepsilon_k\left(\varepsilon_k^2+2c_s\varepsilon_k(1-{\sf
284: f}^2)-(\hbar\omega)^2\right)\right] &=& 0,
285: \end{eqnarray}
286: with $c_n=c_0n$, $c_s=c_2n$, and ${\sf f}=f/n$.
287: The frequencies are then given by
288: \begin{eqnarray}
289: (\hbar\omega)^2_{1,2} &=&
290: \varepsilon_k\left[(c_n+c_s+\varepsilon_k)+\sqrt{(c_n-c_s)^2+4c_nc_s{\sf
291: f}^2}\,\right],
292: \nonumber \\
293: (\hbar\omega)^2_{3,4} &=&
294: \varepsilon_k\left[(c_n+c_s+\varepsilon_k)-\sqrt{(c_n-c_s)^2+4c_nc_s{\sf
295: f}^2}\,\right],
296: \nonumber \\
297: (\hbar\omega)^2_{5,6} &=& (\varepsilon_k+c_s)^2-c_s^2 (1-{\sf f}^2).
298: \label{eq:ee}
299: \end{eqnarray}
300: The corresponding modes are termed as density modes (solid lines),
301: spin modes (dotted lines), and quadrupolar spin modes (dashed lines as in Fig. \ref{fig:ee})
302: by Ho \cite{Ho98}. Figure \ref{fig:ee}(a) and
303: \ref{fig:ee}(b) show the real and imaginary parts of the typical
304: dispersion relation for a $^{87}$Rb spin-1 condensate \cite{para2},
305: respectively. All frequencies are real for
306: antiferromagnetically interacting (e.g. $^{23}$Na) condensate.
307:
308: \begin{figure}
309: \begin{center}
310: \includegraphics[width=3.25in]{spindomain_fig1.eps}
311: \caption{(Color online) Real part (a) and imaginary part (b) of a typical
312: Bogoliubov spectrum for a spin-1 condensate with ferromagnetic interaction.
313: Panel (c) shows the smallest wavelength of a homogeneous spin-1 $^{87}$Rb
314: condensate at $n=1.9\times 10^{14}$ cm$^{-3}$.} \label{fig:ee}
315: \end{center}
316: \end{figure}
317:
318: Our analysis here parallels that of Refs.
319: \cite{Kasamatsu04,Kourakis05} for a two-component condensate.
320: We find two interesting features in Fig. \ref{fig:ee}(b).
321: One of them at $k_-$ is the most unstable mode at
322: the maximum imaginary frequency,
323: it determines short
324: time behavior such as the time scale for domains to emerge. The
325: other is the largest wave vector $k_m$ with an (infinitesimal)
326: imaginary frequency which determines the long time behavior such
327: as the final domain size. From Eq. (\ref{eq:ee}), we find
328: the time scale for the emergence of a domain is $\sim
329: 1/|\omega(k_-)| = h/(|c_s|\sqrt{1-{\sf f}^2})$ and the domain
330: width is about $\lambda_m = 2\pi/k_m = h /
331: \sqrt{2m|c_s|(1+\sqrt{1-{\sf f}^2})}$.
332: It is typically of the order of spin healing length.
333: Figure \ref{fig:ee}(c)
334: displays the ${\sf f}$-dependence of $\lambda_m$ at $n=1.9\times
335: 10^{14}$ cm$^{-3}$ for a homogeneous $^{87}$Rb spin-1 condensate.
336: For the domains to form, a
337: condensate has to be larger than $\lambda_m$, at least in one direction.
338:
339: We note the spin domain formation as discussed here is different
340: from striation patterns as observed in (antiferromagnetic)
341: $^{23}$Na condensates. The stripe patterns arise from interplay of
342: an external B-field, a field gradient, and immiscibility
343: among different spin components \cite{Stenger98}. No domains were
344: observed in Stenger {\it et al.}'s experiment \cite{Stenger98} at
345: negligible B-fields where the $|+\rangle$ and $|-\rangle$
346: components coexist. At finite values of B-fields, phase
347: separation between the $|0\rangle$ and the $|+$/$-\rangle$
348: components occurs \cite{Zhang03}. In Miesner {\it et al.} and
349: Stamper-Kurn {\it et al.}'s experiments \cite{Stenger98},
350: only two spin components were involved due to the relatively large
351: bias B-field ($\sim 15$ G).
352:
353: \begin{figure}
354: \includegraphics[width=3.25in]{spindomain_fig2.eps}
355: \caption{(Color online) Surfaces of $d{\cal E}/dm=0$ (top) and the
356: cross-section at $\theta=0$ (bottom). The plus signs denote $d{\cal E}/dm>0$
357: and the minus signs denote $d{\cal E}/dm<0$.}
358: \label{fig:dfdi} % Domain Formation from Dynamical Instability
359: \end{figure}
360:
361: Our analysis shows the formation of spin domains is
362: a direct consequence of dynamic instability for a condensate with
363: ferromagnetic interactions. To provide a clearer physical picture
364: for domain formation,
365: we now work in the lab frame. We focus on $d{\cal E}/dm$,
366: which in fact calibrates the formation of the spin domain. We find
367: \begin{eqnarray}
368: {d{\cal E} \over dm} &=& c_2m\left[ 1 -
369: {n_0\cos\theta\over \sqrt{(n-n_0)^2-m^2}} \right].
370: \end{eqnarray}
371: Figure \ref{fig:dfdi} shows the surfaces where the above first order derivative
372: is zero. The region below the saddle surface in Fig. \ref{fig:dfdi} of an orbit
373: is unstable if $c_2<0$. Here the meaning of ``unstable" is generalized,
374: referring to the dynamical property where the local magnetization tends to
375: deviate further from $m=0$. For example, in the lower right allowed region of
376: $d{\cal E}/dm < 0$, $\Delta m>0$ is required to lower the local energy ${\cal
377: E} \approx {\cal E}(m) + (d{\cal E}/dm) \Delta m$. Thus $m$ tends to increase.
378: Similarly the lower left allowed region would make $m$ decrease. The combined
379: effect is dynamically unstable orbits, separation of $|+\rangle$ and
380: $|-\rangle$ component, and the eventual formation of spin domains.
381:
382: For antiferromagnetic interactions ($c_2>0$), $\theta$
383: usually oscillates around $\pi$. Thus $d{\cal E}/dm>0$ for $m>0$ and $d{\cal
384: E}/dm<0$ for $m<0$. So the magnetization always oscillates around zero and
385: no domain forms. This coincides with the findings of the dynamical stability
386: analysis for an antiferromagnetically interacting condensate.
387:
388: \begin{figure}
389: \includegraphics[width=3.25in]{./spindomain_fig3.eps}
390: \caption{(Color online) Typical evolutions for spin domain formation in a
391: $^{87}$Rb condensate. The initial state is the
392: ground state at $B=0.3$ G. The B-field is then set to zero and a small
393: white noises are added throughout the evolution. The left contour plot is
394: for the $|+\rangle$ component. The right column shows the density
395: distribution of all three components at times $t=0,160,320$ ($1/\omega_z$).
396: Solid, dash-dotted, and dashed lines denote respectively the $|+\rangle$,
397: $|0\rangle$, and $|-\rangle$ component. Dotted lines are for the total density.
398: The axial density is $n_z\equiv\int \sum_j|\Phi_j|^2 2\pi r dr$.
399: $a_z=\sqrt{\hbar/{\sf m}\omega_z}\simeq 2.2\mu$m and the average condensate
400: density is $\sim 1.9\times 10^{14}$ cm$^{-3}$.}
401: \label{fig:sd} % Spin Domains
402: \end{figure}
403:
404: Finally we perform numerical simulations of Eq. (\ref{eq:gpe}) to confirm the
405: mechanism of dynamical instability-induced spin domain formation. The initial
406: conditions are as in the experiment \cite{Chapman05}, with $^{87}$Rb
407: condensates ($N_0(0)/N=0.744$, $\theta(0)=0$, for the ground state of
408: $N=2.0\times 10^5$ at $B=0.3$ G and $M=0$), in a trap $V_{\rm ext}(\vec r) =
409: ({\sf m}/2)(\omega_x^2x^2 +\omega_y^2y^2+\omega_z^2z^2)$ with
410: $\omega_x=\omega_y= (2\pi) 240$ Hz and $\omega_z=(2\pi) 24$ Hz. In one of the
411: simulations, we intentionally include additive small white noise ($\sim
412: 1.0\times 10^{-5}$), although still much larger than numerical errors
413: \cite{Kramer04} during the propagation. We find that it takes a shorter time
414: for the $|+\rangle$ and $|-\rangle$ components to separate when white-noise is
415: included. Figure \ref{fig:sd} shows the evolution of axial density
416: distributions. Phase separation between the $|+\rangle$ and $|-\rangle$
417: components is seen accompanied by the formation of domains. This proves again
418: that dynamical instability causes the formation of domains. The domain width
419: (an upper limit) as estimated from $\lambda_m = {h/\sqrt{2{\sf m}|c_2|\langle
420: n\rangle(1+\sqrt{1-{\sf f}^2})}} \approx 15 \mu$m, is consistent with both
421: simulations and experimental observations \cite{Chapman05}.
422:
423: In a spinor condensate, spin wave excitations normally refer to
424: dynamically stable (or relatively more stable) collective modes.
425: Once excited, they lead to coherent cyclic dynamics in both
426: spatial and temporal dimensions. Spin domains, on the other hand,
427: refers to unstable modes, with a fixed pattern in the long time limit.
428:
429: In conclusion we have presented a systematic study of dynamical stability and
430: the accompanied mechanism for domain formation in a spin-1 condensate. Our
431: results affirm that a ferromagnetically interacting condensate is dynamically
432: unstable and evolves spontaneously into multi-domain structures contrary to
433: dynamically stable antiferromagnetic condensates. Our work provides
434: a clear physical picture for recently observed spontaneous domain formations in
435: spin-1 condensates \cite{Chapman05}.
436:
437: This work is supported by NASA and NSF.
438:
439: \begin{thebibliography}{widest-label}
440:
441: \bibitem{AgrawalBook89}
442: G. P. Agrawal, {\it Nonlinear fiber optics}, (Academic Press, Boston, 1989).
443:
444: \bibitem{Gollub99} %Pattern formation in nonequilibium fluid
445: J. P. Gollub and J. S. Langer, Rev. Mod. Phys. {\bf 71}, S397 (1999); M. C.
446: Cross and P. C. Hohenberg, Rev. Mod. Phys. {\bf 65}, 851 (1993).
447:
448: \bibitem{Pooley04} % Stripe pattern formation in granular
449: C. M. Pooley and J. M Yeomans, Phys. Rev. Lett. {\bf 93}, 118001 (2004); T.
450: Shinbrot and F. J. Muzzio, Nature (London) {\bf 410}, 251 (2001).
451:
452: \bibitem{Strecker02} % One-component soliton and modulational instability
453: K. E. Strecker {\it et al.}, Nature (London) {\bf 417}, 150 (2002);
454: %K. E. Strecker, G. B. Partridge, A. G. Truscott, and R. G. Hulet, Nature (London)
455: %{\bf 417}, 150 (2002).
456: L. Salasnich, A. Parola, and L. Reatto, Phys. Rev. Lett. {\bf 91},
457: 080405 (2003);
458: L. D. Carr and J. Brand, Phys. Rev. Lett. {\bf 92}, 040401 (2004).
459:
460: \bibitem{Lewandowski02} % Domain in spin-1/2 BEC
461: H. J. Lewandowski {\it et al.}, Phys. Rev. Lett. {\bf 88}, 070403 (2002);
462: %H. J. Lewandowski, D. M. Harber, D. L. Whitaker, and E. A. Cornell,
463: %Phys. Rev. Lett. {\bf 88}, 070403 (2002).
464: P. Ao and S. T. Chui, Phys. Rev. A {\bf 58}, 4836 (1998);
465: P. Ao and S. T. Chui, J. Phys. B {\bf 33}, 535 (2000).
466:
467: \bibitem{Stenger98} % Domain formation in spin-1 BEC
468: J. Stenger {\it et al.}, {Nature} (London) {\bf 396}, 345 (1998); H.-J. Miesner
469: {\it et al.}, Phys. Rev. Lett. {\bf 82}, 2228 (1999);
470: %J. Stenger, S. Inouye, D. M. Stamper-Kurn, H. -J. Miesner, A. P. Chikkatur, and
471: %W. Ketterle, {Nature} {\bf 396}, 345 (1998).
472: D. M. Stamper-Kurn {\it et al.}, Phys. Rev. Lett. {\bf 83}, 661 (1999); D.M.
473: Stamper-Kurn and W. Ketterle, in {\it Coherent Atomic Matter Waves}, Les
474: Houches Summer School Session LXXII in 1999, edited by R. Kaiser, C. Westbrook,
475: and F. David (Springer, NewYork, 2001), pp. 137-217; T. Isoshima, K. Machida,
476: and T. Ohmi, Phys. Rev. A {\bf 60}, 4857 (1999).
477:
478: \bibitem{Sarlo04}
479: L. De Sarlo {\it et al.}, e-print cond-mat/0412279 v1; C. Tozzo, M. Kr\"amer,
480: and F. Dalfovo, e-print cond-mat/0505625 v1.
481:
482: \bibitem{Kasamatsu04} % 2-component Na23 domains
483: K. Kasamatsu and M. Tsubota, Phys. Rev. Lett. {\bf 93}, 100402 (2004).
484:
485: \bibitem{Kourakis05}
486: I. Kourakis {\it et al.}, e-print cond-mat/0505324 v1; H. Pu {\it et al.},
487: Phys. Rev. Lett. {\bf 88}, 070408 (2002).
488:
489: \bibitem{Robins01} % Modulational instability of Rb87 spin-1 condensate
490: N. P. Robins {\it et al.}, Phys. Rev. A {\bf 64}, 021601(R) (2001).
491: %N. P. Robins, W. Zhang, E. A. Ostrovskaya, and Y. S. Kivshar,
492: %Phys. Rev. A {\bf 64}, 021601R (2001).
493:
494: \bibitem{Saito05} % Spin domain in Rb87, numerical simulation
495: H. Saito and M. Ueda, e-print cond-mat/0504398 v1.
496:
497: \bibitem{Barrett01} % Spin-1 experiemnts
498: M. D. Barrett, J. A. Sauer, and M. S. Chapman, Phys. Rev. Lett. {\bf 87}, 010404
499: (2001);
500: M. -S. Chang {\it et al.}, Phys. Rev. Lett. {\bf 92}, 140403 (2004);
501: %M. -S. Chang, C. D. Hamley, M. D. Barrett, J. A. Sauer, K. M. Fortier,
502: %W. Zhang, L. You, and M. S. Chapman, Phys. Rev. Lett. {\bf 92}, 140403 (2004).
503: H. Schmaljohann {\it et al.}, Phys. Rev. Lett. {\bf 92}, 040402 (2004);
504: %H. Schmaljohann, M. Erhard, J. Kronj\"{a}ger, M. Kottke, S. van Staa, L.
505: %Cacciapuoti, J. J. Arlt, K. Bongs, and K. Sengstock, Phys. Rev. Lett. {\bf 92},
506: %040402 (2004).
507: T. Kuwamoto {\it et al.}, Phys. Rev. A {\bf 69}, 063604 (2004);
508: %T. Kuwamoto, K. Araki, T. Eno, and T. Hirano, Phys. Rev. A {\bf
509: %69}, 063604 (2004).
510: J. M. Higbie {\it et al.}, e-print cond-mat/0502517.
511: %J. M. Higbie, L. E. Sadler, S. Inouye, A. P. Chikkatur, S.
512: %R. Leslie, K. L. Moore, V. Savalli, and D. M. Stamper-Kurn, e-print
513: %cond-mat/0502517.
514:
515: \bibitem{Chapman05}
516: We have observed the multidomain structure in a spin-1 $^{87}$Rb condensate and
517: submitted recently a publication.
518:
519: %\bibitem{Erhard04}
520: %M. Erhard {\it et al.}, e-print cond-mat/0402003.
521: %M. Erhard, H. Schmaljohann, J. Kronj\"{a}ger, K. Bongs,
522: %and K. Sengstock, cond-mat/0402003.
523:
524: %\bibitem{Szirmai03}
525: %G. Szirmai, P. Sz\'{e}pfalusy and K. Kis-Szab\'{o},
526: %Phys. Rev. A {\bf 68}, 023612 (2003).
527:
528: \bibitem{Zhang05a} % spin mixing in B field under SMA
529: W. Zhang {\it et al.}, Phys. Rev. A {\bf 72}, 013602 (2005);
530: %W. Zhang, D. L. Zhou, M. -S. Chang, M. S. Chapman, and L. You, e-print
531: %cond-mat/0502061.
532: D. R. Romano and E. J. V. de Passos, Phys. Rev. A {\bf 70},
533: 043614 (2004).
534:
535: \bibitem{Ho98}
536: T. -L. Ho, {Phys. Rev. Lett.} {\bf 81}, 742 (1998); T. Ohmi and K. Machida, {J.
537: Phys. Soc. Jpn.} {\bf 67}, 1822 (1998); C. K. Law, H. Pu, and N. P. Bigelow,
538: {Phys. Rev. Lett.} {\bf 81}, 5257 (1998).
539:
540: %\bibitem{Pu00} % Spin mixing in B fields
541: %H. Pu, S. Raghavan, and N. P. Bigelow, Phys. Rev. A {\bf 61},
542: %023602 (2000).
543:
544: %\bibitem{Yi02}
545: %S. Yi, \"{O}. E. M\"{u}stecapl{\i}o\u{g}lu, C. P. Sun
546: %and L. You, {Phys. Rev. A} {\bf 66}, 011601 (2002).
547:
548: \bibitem{Pu99} % Spin mixing
549: H. Pu {\it et al.}, Phys. Rev. A {\bf 60}, 1463 (1999).
550: %H. Pu, C. K. Law, S. Raghavan, J. H. Eberly, and N. P. Bigelow,
551: %Phys. Rev. A {\bf 60}, 1463 (1999).
552:
553: \bibitem{Maldonado93}
554: O. Maldonado, J. Math. Phys. {\bf 34}, 5016 (1993).
555:
556: \bibitem{para2} % Rb87 scattering length
557: E. G. M. van Kempen {\it et al.}, {Phys. Rev. Lett.} {\bf 88}, 093201 (2002).
558: %E. G. M. van Kempen, S. J. J. M. F. Kokkelmans, D. J. Heinzen,
559: %and B. J. Verhaar, {Phys. Rev. Lett.} {\bf 88}, 093201 (2002).
560:
561: %\bibitem{para3} % Na23 scattering length
562: %A. Crubellier, O. Dulieu, F. Masnou-Seeuws, M. Elbs,
563: %H. Kn\"{o}ckel, and E. Tiemann, {Eur. Phys. J. D} {\bf 6}, 211 (1999).
564:
565: \bibitem{Zhang03}
566: W. Zhang, S. Yi, and L. You, {New J. Phys.} {\bf 5}, 77 (2003).
567:
568: \bibitem{Kramer04} % Parametric instability of BEC in optical lattice
569: M. Kr\"amer, C. Tozzo, and F. Dalfovo, e-print cond-mat/0410122 v1.
570:
571: \end{thebibliography}
572:
573: \end{document}
574: