1: \documentclass[prl,twocolumn,aps,showpacs]{revtex4}
2: \usepackage{graphicx}
3: \begin{document}
4:
5: \title{Spin-3 Chromium Bose-Einstein Condensates}
6: \author{L. Santos$^{(1)}$ and T. Pfau$^{(2)}$}
7: \affiliation{
8: (1) Institut f\"ur Theoretische Physik III,
9: Universit\"at Stuttgart, Pfaffenwaldring 57 V, D-70550 Stuttgart \\
10: (2) 5. Physikalisches Institut,
11: Universit\"at Stuttgart, Pfaffenwaldring 57 V, D-70550 Stuttgart \\
12: }
13:
14: \begin{abstract}
15: We analyze the physics of spin-3 Bose-Einstein condensates, and in particular the
16: new physics expected in on-going experiments with condensates
17: of Chromium atoms. We first discuss the ground-state properties, which, depending
18: on still unknown Chromium parameters, and for low magnetic fields
19: can present various types of phases.
20: We also discuss the spinor-dynamics in Chromium spinor condensates, which
21: present significant qualitative differences when compared to other
22: spinor condensates. In particular, dipole-induced spin relaxation may lead
23: for low magnetic fields to transfer of spin into angular momentum similar to the
24: well-known Einstein-de Haas effect. Additionally, a rapid large transference
25: of population between distant magnetic states becomes also possible.
26: \end{abstract}
27: \maketitle
28:
29: % Introduction. Why spinor BEC is interesting?
30:
31: Within the very active field of ultra cold atomic gases,
32: multicomponent (spinor) Bose-Einstein condensates (BECs) have recently
33: attracted a rapidly growing attention. Numerous works have addressed
34: the rich variety of phenomena revealed by spinor BEC, in particular
35: in what concerns ground-state and spin dynamics
36: \cite{Ho98,Ohmi98,Law98,Koashi00,Ciobanu00,Koashi00b,Pu01,You04,Ho00b,Hall98,
37: Stenger98,Barret01,Schmaljohann04,Chang04,Kuwamoto04,Wheeler04,Higbie05,Widera05}.
38: The first experiments on spinor BEC were performed at JILA using mixtures of
39: $^{87}$Rb BEC in two magnetically confined internal states (spin-$1/2$ BEC)
40: \cite{Hall98}. Optical trapping of spinor condensates was first realized
41: in spin-$1$ Sodium BEC at MIT \cite{Stenger98} constituting a major
42: breakthrough since, whereas magnetic trapping confines the BEC to weak-seeking
43: magnetic states, an optical trap enables confinement of all magnetic
44: substates. In addition, the atoms in a magnetic substate can be converted into
45: atoms in other substates through interatomic interactions. Hence, these
46: experiments paved the way towards the above mentioned fascinating
47: phenomenology
48: originating from the spin degree of freedom. Various experiments have been
49: realized since then in spin-$1$ BEC, in particular in $^{87}$Rb in the $F=1$
50: manifold. It has been predicted that spin-$1$ BEC can present just two
51: different ground states phases, either ferromagnetic or polar \cite{Ho98}.
52: In the case of $F=1$ $^{87}$Rb it has been shown that the ground-state
53: presents a ferromagnetic behavior \cite{Schmaljohann04,Chang04,Klausen01}.
54: These analyses have
55: been recently extended to spin-$2$ BEC, which presents an even richer
56: variety of possible ground-states, including in addition the
57: so-called cyclic phases \cite{Ciobanu00,Koashi00b}. Recent experiments have
58: shown a behavior compatible with a polar ground-state, although in the very
59: vicinity of the cyclic phase \cite{Schmaljohann04,Klausen01}. Recently, spin
60: dynamics has attracted a major interest, revealing the
61: fascinating physics of the coherent oscillations between the different
62: components of the manifold \cite{Barret01,Schmaljohann04,Chang04,Kuwamoto04,
63: Wheeler04,Higbie05}.
64:
65: % Chromium as a spin 3 and dipolar gas
66:
67: Very recently, a Chromium BEC (Cr-BEC) has been achieved at Stuttgart University
68: \cite{Griesmaier05}. Cr-BEC presents fascinating new features when
69: compared to other experiments in BEC. On one hand, since the ground state of
70: $^{52}$Cr is $^7$S$_3$, Cr-BEC constitutes the first accessible example
71: of a spin-3 BEC. We show below that this fact may have very important
72: consequences for both the ground-state and the dynamics of spinor Cr-BEC.
73: On the other hand, when aligned into the state with magnetic quantum number $m=\pm 3$,
74: $^{52}$Cr presents a magnetic moment $\mu=6\mu_B$, where $\mu_B$ is
75: the Bohr magneton. This dipolar moment
76: should be compared to alkali atoms, which have a
77: maximum magnetic moment of $1\mu_B$, and hence $36$ times smaller
78: dipole-dipole interactions. Ultra cold dipolar gases have attracted a
79: rapidly growing attention, in particular in what
80: concerns its stability and excitations \cite{Dipoles}. The interplay of the
81: dipole-dipole interaction and spinor-BEC physics has been also considered
82: \cite{Pu01,You04}. Recently, the dipolar
83: effects have been observed for the first time in
84: the expansion of a Cr-BEC \cite{Stuhler05}.
85:
86: % This Letter
87:
88: This Letter analyzes spin-3 BEC, and in particular
89: the new physics expected in on-going experiments in Cr-BEC. After deriving the
90: equations that describe this system, we focus on the ground-state, using single-mode approximation (SMA), showing
91: that various phases are possible, depending on the applied magnetic field,
92: and the (still unknown) value of the $s$-wave scattering length for the channel of total spin zero.
93: This phase diagram presents certain differences with respect to the diagram first worked out recently by Diener and Ho \cite{Diener05}.
94: In the second part of this Letter, we discuss the spinor dynamics, departing from the SMA.
95: The double nature of Cr-BEC as a spin-3 BEC and a dipolar BEC is shown to lead to significant qualitative
96: differences when compared to other spinor BECs. The larger spin
97: can allow for fast population transfer from $m=0$ to $m=\pm 3$ without a
98: sequential dynamics as in $F=2$ $^{87}$Rb \cite{Schmaljohann04}. In addition, dipolar
99: relaxation violates spin conservation, leading to rotation of the
100: different components, resembling the well-known Einstein-de Haas (EH)
101: effect~\cite{Einstein15}.
102:
103: % Model
104:
105: In the following we consider an optically trapped Cr-BEC with
106: $N$ particles. The Hamiltonian regulating Cr-BEC is of the form
107: $\hat H=\hat H_0+\hat V_{sr}+\hat V_{dd}$.
108: The single-particle part of the Hamiltonian, $\hat H_0$, includes the
109: trapping energy and the linear Zeeman effect (quadratic Zeeman
110: effect is absent in Cr-BEC), being of the form
111: \begin{equation}
112: \hat H_0=\int d{\bf r} \sum_m \hat\psi_m^\dag({\bf r})
113: \left [\frac{-\hbar^2}{2M}\nabla^2+U_{trap}({\bf r})+pm \right ]
114: \hat\psi_m({\bf r}),
115: \end{equation}
116: where $\hat \psi_m^\dag$ ($\hat\psi_m$) is the creation (annihilation)
117: operator in the $m$ state, $M$ is the atomic mass, and $p=g\mu_B B$, with $g=2$ for $^{52}$Cr,
118: and $B$ is the applied magnetic field.
119:
120: The short-range (van der Waals) interactions are given by $\hat V_{sr}$.
121: For any interacting pair with spins ${\bf S}_{1,2}$, $\hat V_{sr}$
122: conserves the total spin, $S$, and
123: is thus described in terms of the projector operators on different
124: total spins $\hat{\cal P}_S$, where $S=0$, $2$, $4$, and $6$
125: (only even $S$ is allowed)\cite{Ho98} :
126: \begin{equation}
127: \hat V_{sr}=\frac{1}{2}\int d{\bf r} \sum_{S=0}^6 g_S \hat{\cal P}_S({\bf r}),
128: \end{equation}
129: where $g_S=4\pi\hbar^2a_S/M$, and $a_S$ is the $s$-wave scattering length for a
130: total spin $S$. Since ${\bf S}_1\cdot{\bf S}_2=(S^2-S_1^2-S_2^2)/2$, then
131: $\sum_S \hat{\cal P}_S({\bf r}) = :\hat n^2({\bf r}):$,
132: $\sum_S \lambda_S \hat{\cal P}_S({\bf r}) = :\hat F^2({\bf r}):$,
133: and
134: $\sum_S \lambda_S^2 \hat{\cal P}_S({\bf r}) = :\hat O^2({\bf r}):$,
135: where $::$ denotes normal order,
136: $\lambda_S=[S(S+1)-24]/2$,
137: $\hat n({\bf r})=\sum_m \hat\psi_m^\dag({\bf r})\hat\psi_m({\bf r})$,
138: $\hat F^2=\sum_{i=x,y,z} \hat F_i^2$, with
139: $\hat F_{i}({\bf r})=\sum_{mn} \hat\psi_m^\dag({\bf r})S_{mn}^i\hat\psi_n({\bf r})$,
140: and $\hat O^2=\sum_{i,j} \hat O_{ij}^2$, with
141: $\hat O_{ij}=\sum_{m,n} \hat\psi_m^\dag({\bf r})(S^iS^j)_{mn}\hat\psi_n(\vec
142: r)$. $S^{x,y,z}$ are the spin matrices.
143: Hence,
144: \begin{equation}
145: \hat V_{sr}=\frac{1}{2}\int d{\bf r} [c_0 :\hat n^2({\bf r}) +
146: c_1 \hat F^2({\bf r}) +c_2 \hat{\cal P}_0({\bf r})+c_3 \hat O^2({\bf r}):],
147: \label{Vsr}
148: \end{equation}
149: where
150: $\hat {\cal P}_0({\bf r})=\frac{1}{7}\sum_{m,n}(-1)^{m+n}
151: \hat\psi_m^\dag\hat\psi_{-m}^\dag\hat\psi_n\hat\psi_{-n}$,
152: and
153: $c_0=(-11g_2+81g_4+7g_6)/77$,
154: $c_1=(g_6-g_2)/18$,
155: $c_2=g_0+(-55g_2+27g_4-5g_6)/33$,
156: $c_3=g_2/126-g_4/77+g_6/198$.
157: For the case of $^{52}$Cr \cite{Werner05}, $a_6=112a_B$, where $a_B$
158: is the Bohr radius, and
159: $c_0=0.65g_6$, $c_1=0.059g_6$, $c_2=g0+0.374g_6$, and $c_3=-0.002g_6$.
160: The value of $a_0$ is unknown, and
161: hence, in the following, we consider different scenarios depending on the
162: value of $g_0/g_6$. Note that Eq.~(\ref{Vsr}) is similar to that obtained
163: for spin-2 BEC \cite{Ciobanu00,Koashi00b},
164: the main new feature being the $c_3$ term, which
165: introduce qualitatively new physics as discussed below.
166:
167: The dipole-dipole interaction $\hat V_{dd}$ is of the form
168: \begin{eqnarray}
169: \hat V_{dd}&=&\frac{c_d}{2}\int d{\bf r}\int d{\bf r}' \frac{1}{|{\bf r}-{\bf r}'|^3}
170: \hat\psi_m^\dag ({\bf r})\hat\psi_m'^\dag ({\bf r}') \nonumber \\
171: &&\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\! \!\!\!\!\left [
172: {\bf S}_{mn}\cdot{\bf S}_{m'n'}-3({\bf S}_{mn}\cdot{\bf e})
173: ({\bf S}_{m'n'}\cdot{\bf e})
174: \right ]
175: \hat\psi_n ({\bf r})\hat\psi_n'({\bf r}'),
176: \end{eqnarray}
177: where $c_d=\mu_0\mu_B^2g_S^2/4\pi$, with $\mu_0$ the magnetic permeability of
178: vacuum, and ${\bf e}=({\bf r}-{\bf r}')/|{\bf r}-{\bf r}'|$. For $^{52}$Cr $c_d=0.004g_6$.
179: $\vec V_{dd}$ may be re-written as:
180: \begin{eqnarray}
181: \hat V_{dd}&=&\sqrt{\frac{3\pi}{10}}c_d\int \int \frac{d{\bf r} d{\bf r}'}{|{\bf r}-{\bf r}'|^3}
182: :\left \{
183: {\cal F}_{zz}({\bf r},{\bf r}') Y_{20}
184: \right\delimiter 0 \nonumber \\
185: &&\left\delimiter 0 +{\cal F}_{z,-}({\bf r},{\bf r}') Y_{21}
186: + {\cal F}_{z,+}({\bf r},{\bf r}') Y_{2-1} \right\delimiter 0
187: \nonumber \\
188: &&\left\delimiter 0
189: + {\cal F}_{-,-}({\bf r},{\bf r}')Y_{22}
190: + {\cal F}_{+,+} ({\bf r},{\bf r}') Y_{2-2}
191: \right\}:,
192: \end{eqnarray}
193: where ${\cal F}_{zz}({\bf r},{\bf r}')=\sqrt{2/3} [3\hat F_z({\bf r})\hat F_z({\bf r}')-
194: \hat {\bf F}({\bf r})\cdot \hat {\bf F}({\bf r}')]$,
195: ${\cal F}_{z,\pm}({\bf r},{\bf r}')=\pm[\hat F_\pm({\bf r})\hat F_z({\bf r}')+
196: \hat F_z({\bf r})\hat F_\pm({\bf r}')]$,
197: ${\cal F}_{\pm,\pm}({\bf r},{\bf r}')=\hat F_\pm({\bf r})\hat F_\pm({\bf r}')$,
198: $\hat F_\pm=\hat F_x\pm i\hat F_y$, and $Y_{2m}({\bf r}-{\bf r}')$ are the spherical
199: harmonics. Note that contrary to the short-range interactions, $\hat V_{dd}$ does
200: not conserve the total spin, and
201: may induce a transference of angular momentum into the center of mass (CM) degrees of freedom.
202:
203: We first discuss the ground-state of the spin-3 BEC for
204: different values of $g_0$, and the
205: magnetic field, $p$. We consider mean-field (MF)
206: approximation $\hat\psi_m({\bf r})\simeq\sqrt{N}\psi_m({\bf r})$.
207: In order to simplify the analysis of the possible
208: ground-state solutions we perform
209: SMA: $\psi_m({\bf r})=\Phi({\bf r})\psi_m$,
210: with $\int d{\bf r} |\Phi({\bf r})|^2=1$.
211: Apart from spin-independent terms the energy per particle is of the form:
212: \begin{eqnarray}
213: E&=& pf_z+\frac{Nc_1\beta}{2}(f_z^2+f_+f_-) +\frac{2c_2N\beta}{7}|s_-|^2
214: \nonumber \\
215: &+& \frac{Nc_3\beta}{2}\sum_{ij}O_{ij}^2 +
216: \sqrt{\frac{3\pi}{10}}Nc_d
217: \left [
218: \sqrt{2/3}
219: \Gamma_0 (2f_z^2-f_+f_-) \right\delimiter 0 \nonumber \\
220: &-&\left\delimiter 0 2\Gamma_{+1}f_zf_-+2\Gamma_{-1}f_zf_+
221: +\Gamma_{+2}f_-^2+\Gamma_{-2}f_+^2
222: \right ]
223: ,
224: \end{eqnarray}
225: where $\beta=\int d{\bf r} |\Phi({\bf r})|^4$,
226: $s_-=\frac{1}{2}\sum_m (-1)^m \psi_m\psi_{-m}$,
227: $f_i=\sum_{m,n} \psi_m^* S^i_{mn} \psi_n$,
228: $O_{ij}=\sum_{m,n} \psi_m^* (S^iS^j)_{mn} \psi_n$
229: and $\Gamma_{m}=\int d{\bf r}\int d{\bf r}'
230: |\Phi({\bf r}')|^2|\Phi({\bf r})|^2 Y_{2m}({\bf r}-{\bf r}')/|{\bf r}-{\bf r}'|^3$.
231: Note that $\Gamma_{\pm 1}=0$ for any symmetric density $|\Phi({\bf r})|^2$.
232: %% FIGURE 1
233: \begin{figure}%[ht]
234: \begin{center}\
235: \vspace*{-0.5cm}
236: \includegraphics[width=6.2cm,angle=0]{forfig1_new2.eps}
237: \end{center}
238: \caption{Phase diagram function of $g_0/g_6$ and
239: $\tilde p/g_6$, where $\tilde p=2g_S\mu_BB/N\beta$.}
240: \label{fig:1}
241: \end{figure}
242:
243: Let us consider a magnetic field in the $z$-direction. Hence,
244: the ground-state magnetization must be aligned with the external field, and
245: $f_+=f_-=0$. Then, we obtain that
246: $\sum_{ij}O_{ij}^2=77-12\gamma_z+3\gamma_z^2/2-f_z^2/2+|\eta|^2/2+2|\sigma|^2$,
247: where $\gamma_z =\sum_m m^2 |\psi_m|^2$,
248: $\eta=\sum_{m=-3}^1 \sqrt{[12-(m+2)(m+1)][12-m(m+1)]} \psi_{m+2}^*\psi_m$, and
249: $\sigma=\sum_{m=-3}^2 m\sqrt{[12-m(m+1)]} \psi_{m+1}^*\psi_m$.
250: Removing spin-independent terms, the energy becomes $E=N\beta\epsilon/2$, with
251: \begin{equation}
252: \epsilon=\tilde p f_z +\tilde c_1 f_z^2+\frac{4c_2}{7}|s_-|^2+
253: c_3 \left ( \frac{3\gamma_z^2}{2}-12\gamma_z +\frac{|\eta|^2}{2}+2|\sigma|^2 \right),
254: \label{eps}
255: \end{equation}
256: where $\tilde p=2p/N\beta$, and $\tilde c_1=c_1-c_3/2+\sqrt{16\pi/5}\Gamma_0
257: c_d/\beta$. Since $c_d\ll c_1$, dipolar effects are not relevant
258: for the equilibrium discussion. We will hence set $\Gamma_0=0$.
259:
260: %For $c_3=0$ Eq.~(\ref{eps}) resembles that of spin-2 BEC, and one may work out
261: %that the phase diagram would present a ferromagnetic phase for $\tilde p\geq 6\tilde c_1\simeq 0.35g_6$ and
262: %$\tilde p>6\tilde c_1-2c_2/21$, a polar phase for $c_2<0$ ($g_0/g_6<-0.374$) and $\tilde p<6\tilde c_1-2c_2/21$, and
263: %a set of degenerated cyclic phases for $c_2>0$ and $\tilde p<6\tilde c_1$. The presence of $c_3$ changes this picture.
264:
265: We have minimized Eq.~(\ref{eps}) with respect
266: to $\psi_m$, under the constraints $\sum_m|\psi_m|^2=1$ and $f_+=0$ \cite{footnote1}.
267: Figs.~\ref{fig:1} shows the corresponding phase diagram, which although in basic agreement with that worked out recently in Ref.~\cite{Diener05}
268: presents certain differences in its final form. For the phases discussed below $\sigma=0$ \cite{footnoteZ}. For sufficiently negative $g_0$ the system is in a polar phase
269: $P=({\bf c}\theta,0,0,0,0,0,{\bf s}\theta)$, with ${\bf c}\equiv\cos$, and ${\bf s}\equiv\sin$. This phase
270: is characterized by $f_z\simeq -3\tilde p/\tilde p_c$ ($p_c\simeq 6\tilde c_1\simeq 0.36$), $4|s_-|^2=1-(\tilde p/\tilde p_c)^2$,
271: $\gamma_z=9$, and $|\eta|=0$. The polar phase extends up to $g_0/g_6=0.01$ for $p=0$.
272: For sufficiently large $g_0$ and $\tilde p$, $CY_{-3,2}=({\bf c}\theta,0,0,0,0,{\bf s}\theta,0)$
273: occurs. This phase is cyclic ($s_-=0$) and it is characterized
274: by $f_z=-3\tilde p/\tilde p_c$, $\gamma_z\simeq 6+3\tilde p/\tilde p_c$, and $|\eta|=0$. Both P and $CY_{-3,2}$ continuously
275: transform at $\tilde p=\tilde p_c$ into a ferromagnetic phase $F=(1,0,0,0,0,0,0)$. For $0.04<\tilde p\leq 2\tilde c_1+9c_3$
276: $CY_{-3,2}$ becomes degenerated with the cyclic phase
277: $CY_{-1,2}=(0,0,{\bf c}\theta,0,0,0,{\bf s}\theta,0)$. The latter state differs
278: from $CY_{-3,2}$, since $\gamma_z=2-3\tilde p/\tilde p_c$. For sufficiently large $g_0$ and $\tilde p<0.04$,
279: a cyclic phase ($CY_{-3,-1,1,3}$) of the form $(\psi_{-3},0,\psi_{-1},0,\psi_{1},0,\psi_{-3})$ occurs. Contrary to the
280: other cyclic phases this phase is characterized by $|\eta|>0$ \cite{FootnoteBiaxial}. Finally for a region around $g_0=0$
281: an additional phase with $|\eta|>0$, $|s_-|>0$ is found.
282: In this phase two possible ground states are degenerated \cite{footnoteG},
283: namely $S_{-2,0,2}=(0,\psi_{-2},0,\psi_{0},0,\psi_{+2},0)$, and $S_{-3,-1,1,3}$, which has a similar form as $CY_{-3,-1,1,3}$.
284:
285: For a Cr-BEC in a spherical trap of frequency $\omega$
286: in the Thomas-Fermi regime, $\tilde p=\tilde p_c$ for a magnetic field (in mG)
287: $B\simeq 1.25
288: \left (
289: N/10^5
290: \right )^{2/5}
291: \left (
292: \omega/10^3\times 2\pi
293: \right )^{6/5}
294: $. Hence, most probably, magnetic shielding seems necessary for the observation of
295: non-ferromagnetic ground-state phases. We would like to stress as well, that there are
296: still uncertainties in the exact values of $a_{2,4,6}$. In this sense, we have checked for $a_6=98a_B$, $a_4=64a_B$, and
297: $a_2=-27a_B$, that apart from a shift of $0.2$ towards larger values of $g_0/g_6$, the phase diagram remains
298: qualitatively the same.
299:
300: In the last part of this Letter, we consider the spinor dynamics
301: within the MF approximation, but abandoning the
302: SMA. The equations for the dynamics are obtained by deriving
303: the MF Hamiltonian
304: with respect to $\psi_m^*({\bf r})$:
305: \begin{eqnarray}
306: i\hbar\frac{\partial}{\partial t}\psi_m({\bf r})&=&
307: \left [
308: \frac{-\hbar^2\nabla^2}{2M}+U_{trap}+pm\right] \psi_m
309: \nonumber \\
310: %
311: &+& N
312: \left [
313: c_0 n+ m(c_1 f_z+c_d {\cal A}_0) \right ] \psi_m \nonumber \\
314: %
315: &+&\frac{N}{2}
316: \left[c_1 f_-+ 2c_d {\cal A}_{-} \right ] S^+_{m,m-1} \psi_{m-1}\nonumber \\
317: %%
318: &+& \frac{N}{2}
319: \left[c_1f_+ + 2c_d {\cal A}_{+}\right ] S^-_{m,m+1}\psi_{m+1}\nonumber \\
320: %%
321: &+& (-1)^m \frac{2N c_2}{7}s_-\psi_{-m}^* \nonumber \\
322: &+& N c_3 \sum_{n} \sum_{i,j}O_{ij}(S^i S^j)_{mn}\psi_n,
323: \label{dynamic}
324: \end{eqnarray}
325: where
326: ${\cal A}_0=\sqrt{6\pi/5} [\sqrt{8/3}\Gamma_{0,z}+\Gamma_{1,-}+\Gamma_{-1,+}]$,
327: ${\cal A}_\pm=\sqrt{6\pi/5} [-\Gamma_{0,\pm}/\sqrt{6}\mp\Gamma_{\pm,z}+\Gamma_{\pm 2,\mp}]$,
328: $\Gamma_{m,i}=\int d{\bf r}' f_i({\bf r}')
329: Y_{2m}({\bf r}-{\bf r}')/|{\bf r}-{\bf r}'|^3$, and
330: $S^\pm_{m,m\mp 1}=\sqrt{12-m(m\mp 1)}$.
331: Note that all $\psi_m$, $n$, $f_i$, $O_{ij}$, and ${\cal A}_i$ have now a spatial dependence.
332: We first consider $p=0$, and discuss on $p\neq 0$ below.
333: There are two main features in the spinor dynamics in Cr-BEC
334: which are absent (or negligible) in other spinor BECs.
335: On one side the $c_3$ term allows for jumps in the spin manifold
336: larger than one, and hence for a rapid dynamics from e.g. $m=0$
337: to $m=\pm 3$. On the other side,
338: the dipolar terms induce a EH-like transfer of spin into CM
339: angular momentum.
340: %% FIGURE 2
341: \begin{figure}[ht]
342: \begin{center}\
343: \vspace*{-0.6cm}
344: \hspace*{-1.2cm}
345: \includegraphics[width=10.5cm,angle=0]{fig2_new.eps}
346: \end{center}
347: \caption{$|\psi_{m=-2}({\bf r})|^2$ at $\omega t=40$ (a) and $120$ (b) for
348: $p=0$, $g_0=0$, $\omega_z=1$kHz, $N=10^4$ atoms, and $\psi(t=0)=\psi_{m=-3}$.
349: The $x$ and $y$ axes are in $\sqrt{\hbar/m\omega}$ units.}
350: \label{fig:2}
351: \end{figure}
352:
353: We consider for simplicity a quasi-2D BEC, i.e.
354: a strong confinement in the $z$-direction by a harmonic potential
355: of frequency $\omega_z$. Hence
356: $\psi_m({\bf r})=\phi_0(z)\psi_m({\bf \rho})$, where $
357: \phi_0(z)=\exp[-z^2/2l_z^2]/\pi^{1/4}\sqrt{l_z}$,
358: with $l_z=\sqrt{\hbar/m/\omega_z}$. We have then
359: solved the 2D equations using Crank-Nicholson method, considering a harmonic
360: confinement of frequency $\omega$ in the $xy$-plane.
361: In 2D, $\Gamma_{\pm 1,z}=0$, but these terms
362: vanish also in 3D due to symmetry, and hence the 2D physics
363: is representative of the 3D one. The vanishing of $\Gamma_{\pm 1,z}$ is rather important,
364: since if
365: $\psi(t=0)=\psi_{m=\pm 3}$, $\Gamma_{\pm 1,z}$
366: is responsible for a fast dipolar relaxation (for $\omega t\sim 1$).
367: Hence, a BEC with $\psi(t=0)=\psi_{\pm 3}$ does not
368: present any significant fast spin dynamics.
369: However significant spin relaxation appears in the
370: long time scale, due to the $\Gamma_{2,-}({\bf r})$ terms.
371: This is the case of Fig.~\ref{fig:2}, where we $\psi(t=0)=\psi_{m=-3}$ \cite{footnote2}.
372:
373: One of the most striking effects related with spin relaxation
374: is the transference of spin into CM angular
375: momentum, which resembles the famous EH effect \cite{Einstein15}.
376: The analysis of this effect motivated us to avoid the SMA in the
377: study of the spinor dynamics. In Figs.~\ref{fig:2} we
378: show snapshots of the spatial distribution of
379: the $m=-2$ component.
380: Observe that the wavefunction clearly loose its polar symmetry, since spin is converted
381: into orbital angular momentum. The spatial patterns become progressively more complicated in time.
382:
383: The other special feature of Cr-BEC, namely the appearance of the $c_3$ term
384: can have significant qualitative effects in the dynamics both for short and
385: for long time scales. The evolution at long time scale may present interesting
386: features, as large revivals, and it will be considered in future work. Here,
387: we would like to focus on the short-time scales, where the $c_3$ term
388: may produce fast transference from $\psi_{m=0}$
389: to the extremes $\psi_{m=\pm 3}$. The latter is illustrated in Fig.~\ref{fig:3}, where we consider $g_0=g_6$.
390: The population is initially
391: all in the $m=0$ \cite{footnote2}. Contrary to the case of $F=2$ in $^{87}$Rb
392: \cite{Schmaljohann04}, there is at short time scales a jump to the extremes (the
393: population of $\pm m$ is the same due to symmetry since $p=0$).
394: This large jump is absent if $c_3=0$, and depends on the value of $g_0$.
395: In particular, if $g_0=0$
396: one obtains at short time scales a sequential population as for $F=2$ $^{87}$Rb
397: \cite{Schmaljohann04}.
398: %% FIGURE 3
399: \begin{figure}
400: \begin{center}\
401: \vspace*{-0.6cm}
402: \includegraphics[width=5.8cm,angle=0]{fig3_new.eps}
403: \end{center}
404: \caption{Population of $\psi_{0,1,2,3}$ versus $\omega t$ for $g_0=g_6$, $p=0$, and
405: $\psi (0)=\psi_{0}$. Note the rapid growth of $m=\pm 3$ (arrow).}
406: \label{fig:3}
407: \end{figure}
408:
409: We finally comment on the dynamics if $p\neq 0$. In the case of
410: $F=1$ or $F=2$ $^{87}$Rb, the dynamics is independent of $p$ since the
411: linear Zeeman effect may be gauged out by transforming
412: $\psi_m\rightarrow\psi_m e^{ipmt/\hbar}$, due to the conservation of the
413: total spin. In Cr-BEC the situation is very different, since
414: the $\Gamma_{\pm 1,z}$ and $\Gamma_{\pm 2,\mp}$ do not conserve the total spin,
415: and hence oscillate with the Larmor frequency $\omega_L=p/\hbar$ and
416: $2\omega_L$, respectively. If $\omega_L\gg \omega$ one may perform rotating-wave
417: approximation and eliminate these terms. Hence
418: the coherent EH-like effect disappears for sufficiently large applied magnetic fields.
419:
420: In conclusion, spin-3 Cr-BEC is predicted to show different types of spin phases
421: depending on $a_0$ and the magnetic field.
422: The spinor dynamics also presents novel features, as a fast transference
423: between $\psi_0\rightarrow\psi_{\pm 3}$, and the Einstein-de Haas-like transformation of spin into rotation of
424: the different components due to the dipole interaction.
425:
426: We would like to thank M. Fattori for enlightening discussions,
427: and the German Science Foundation (DFG) (SPP1116 and
428: SFB/TR 21) for support. We thank H. M\"akel\"a and K.-A. Suominen for pointing us a mistake
429: in previous calculations, and T.-L. Ho for enlightening e-mail exchanges.
430: During the elaboration of the final version of this paper,
431: the EH-effect has been also discussed in Ref.~\cite{Kawaguchi05}.
432:
433: \begin{thebibliography}{99}
434:
435: \bibitem{Ho98} T.-L. Ho, Phys. Rev. Lett. {\bf 81}, 742 (1998).
436:
437: \bibitem{Ohmi98} T. Ohmi and K. Machida,
438: J. Phys. Soc. Jpn. {\bf 67}, 1822 (1998).
439:
440: \bibitem{Law98} C. K. Law, H. Pu, and N. P. Bigelow, Phys. Rev. Lett.
441: {\bf 81}, 5257 (1998); H. Pu et al., Phys. Rev. A {\bf 60}, 1463 (1999).
442:
443: \bibitem{Koashi00} M. Koashi and M. Ueda, Phys. Rev. Lett. {\bf 84},
444: 1066 (2000).
445:
446: \bibitem{Ciobanu00} C. V. Ciobanu, S.-K. Yip, and T.-L. Ho, Phys. Rev. A,
447: {\bf 61}, 033607 (2000).
448:
449: \bibitem{Koashi00b} M. Koashi and M. Ueda, Phys. Rev. Lett. {\bf 84},
450: 1066 (2000); M. Ueda and M. Koashi, Phys. Rev. A {\bf 65}, 063602 (2002).
451:
452: \bibitem{Pu01} H. Pu, W. Zhang, and P. Meystre, Phys. Rev. Lett.
453: {\bf 87}, 140405 (2001).
454:
455: \bibitem{You04} S. Yi, L. You, and H. Pu, Phys. Rev. Lett. {\bf 93},
456: 040403 (2004).
457:
458:
459: \bibitem{Ho00b} T.-L. Ho and S. K. Yip, Phys. Rev. Lett. {\bf 84},
460: 4031 (2000); \"O. E. M\"ustercapl\i o\u{g}lu {\it et al}, Phys. Rev. A {\bf 68},
461: 063616 (2003).
462:
463: \bibitem{Hall98} D. S. Hall {\it et al.}, Phys. Rev. Lett. {\bf 81},
464: 1539 (1998).
465:
466: \bibitem{Stenger98} J. Stenger {\it et al.}, Nature {\bf 396}, 345 (1998).
467:
468: \bibitem{Barret01} M. D. Barret, J. A. Sauer, and M. S. Chapman,
469: Phys. Rev. Lett. {\bf 87}, 010404 (2001).
470:
471: \bibitem{Schmaljohann04} H. Schmaljohann {\it et al.}, Phys. Rev. Lett.
472: {\bf 92}, 040402 (2004).
473:
474: \bibitem{Chang04} M.-S. Chang {\it et al.}, Phys. Rev. Lett.
475: {\bf 92}, 140403 (2004); W. Zhang {\it et al.}, Phys. Rev. A {\bf 72},
476: 013602 (2005).
477:
478: \bibitem{Kuwamoto04}
479: T. Kuwamoto {\it et al.}, Phys. Rev. A {\bf 69}, 063604 (2004).
480:
481: \bibitem{Wheeler04} M. H. Wheeler {\it et al.}, Phys. Rev. Lett. {\bf 93}, 170402 (2004).
482:
483: \bibitem{Higbie05} J. M. Higbie {\it et al.} Phys. Rev. Lett. {\bf 95},
484: 050401 (2005).
485:
486: \bibitem{Widera05} A. Widera {\it et al.}, cond-mat/0505492.
487:
488: \bibitem{Klausen01} N. N. Klausen, J. L. Bohn, and C. H. Green, Phys. Rev. A
489: {\bf 64}, 053602 (2001).
490:
491: \bibitem{Griesmaier05} A. Griesmaier {\it et al.}, Phys. Rev. Lett. {\bf 94},
492: 160401 (2005).
493:
494: \bibitem{Dipoles} S. Yi and L. You, Phys. Rev. A {\bf 61}, 041604 (2000);
495: K. G\'oral, K. Rz\c a\.zewski, and T. Pfau, Phys. Rev. A
496: {\bf 61}, 10501601(R) (2000); L. Santos {\it et al.},
497: Phys. Rev. Lett. {\bf 85}, 1791 (2000);
498: L. Santos, G. V. Shlyapnikov, and M. Lewenstein,
499: Phys. Rev. Lett. {\bf 90}, 250401 (2003).
500:
501: \bibitem{Stuhler05} J. Stuhler {\it et al.} Phys. Rev. Lett. {\bf 95}, 150406 (2005).
502:
503: \bibitem{Diener05} R. Diener, and T.-L. Ho, cond-mat/0511751.
504:
505: \bibitem{Einstein15} A. Einstein and W. J. de Haas, Verhandl. Deut. Physik Ges. {\bf 17}, 152 (1915).
506:
507: \bibitem{Werner05} J. Werner {\it et al.}, Phys. Rev. Lett. {\bf 94}, 183201
508: (2005).
509:
510: \bibitem{footnote1} For every $g_0$ and $\tilde p$
511: we performed up to $2000$ different runs of a simulated annealing
512: method to avoid the numerous local minima.
513:
514: \bibitem{footnoteZ} We also found numerically states (similar to the Z states of Ref.~\cite{Diener05})
515: with tiny $\sigma$, which are however very slight variations of neighboring phases,
516: with which they are in practice degenerated.
517:
518: \bibitem{FootnoteBiaxial} $|\eta|\neq 0$ leads to the interesting possibility of biaxial nematics, as recently
519: pointed out for the first time in Ref.~\cite{Diener05}.
520:
521: \bibitem{footnoteG} These phases are degenerated for any practical purposes, with relative energy differences $<0.01\%$.
522:
523: \bibitem{footnote2} A small seed in the other components triggers the mean-field evolution.
524:
525: \bibitem{Saito05} H. Saito, and M. Ueda, cond-mat/0506520.
526:
527: \bibitem{Kawaguchi05} Y. Kawaguchi, H. Saito, and M. Ueda, cond-mat/0511052.
528:
529: \end{thebibliography}
530:
531: \end{document}