1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %% Manuscript #15103 written using jpsj2.cls
3: %% Date 2005/06/20
4: %% revised 2005/07/19
5: %% for printout 2005/8/3
6: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
7:
8: %\documentclass[twocolumn,letter]{jpsj2} %% for letters
9: \documentclass[preprint, amsmath, amssymb, floatfix, prb]{revtex4} %% for letters
10: \usepackage{graphicx}% Include figure files
11: \usepackage{dcolumn}% Align table columns on decimal point
12: \usepackage{bm}% bold math
13:
14:
15: \begin{document}
16: \title{Dynamics of Macroscopic Wave Packet
17: Passing through \\ Double Slits: Role of Gravity and Nonlinearity}
18:
19: \author{Katsuhiro~Nakamura}
20: \email{nakamura@a-phys.eng.osaka-cu.ac.jp}
21: \author{Naofumi~Nakazono}
22: \affiliation{Department of Applied Physics,
23: Osaka City University, Sumiyoshi-ku,
24: Osaka 558-8585 Japan}
25: \author{Taro~Ando}
26:
27: \affiliation{Hamamatsu Photonics K.K.,
28: Central Research Laboratory, Hirakuchi, Hamamatsu-City, Shizuoka,
29: 434-8601, Japan}
30:
31:
32: \date{June 20, 2005; revised July 19, 2005}
33:
34: \begin{abstract}
35: Using the nonlinear Schr{\"o}dinger equation (Gross-Pitaevskii equation),
36: the dynamics of a macroscopic wave packet for Bose-Einstein condensates
37: falling through double slits is analyzed. This problem is identified with
38: a search for the fate of a soliton showing a head-on collision with a
39: hard-walled obstacle of finite size. We explore the splitting of the wave
40: packet and its reorganization to form an interference pattern. Particular
41: attention is paid to the role of gravity ($g$) and repulsive nonlinearity
42: ($u_0$) in the fringe pattern. The peak-to-peak distance in the fringe
43: pattern and the number of interference peaks are found to be proportional
44: to $g^{-1/2}$ and $u_0^{1/2}g^{1/4}$, respectively. We suggest a way of
45: designing an experiment under controlled gravity and nonlinearity.
46: \end{abstract}
47:
48:
49: \newcommand{\obs}[1]{\ensuremath{\overrightarrow{\boldsymbol{#1}}}}
50: \newcommand{\bol}[1]{\ensuremath{\boldsymbol{#1}}}
51: \newcommand{\suu}{\ensuremath{|\uparrow\uparrow\rangle}}
52: \newcommand{\sud}{\ensuremath{|\uparrow\downarrow\rangle}}
53: \newcommand{\sdu}{\ensuremath{|\downarrow\uparrow\rangle}}
54: \newcommand{\sdd}{\ensuremath{|\downarrow\downarrow\rangle}}
55: \newcommand{\suuu}{\ensuremath{|\uparrow\uparrow\uparrow\rangle}}
56: \newcommand{\suud}{\ensuremath{|\uparrow\uparrow\downarrow\rangle}}
57: \newcommand{\sudu}{\ensuremath{|\uparrow\downarrow\uparrow\rangle}}
58: \newcommand{\sudd}{\ensuremath{|\uparrow\downarrow\downarrow\rangle}}
59: \newcommand{\sduu}{\ensuremath{|\downarrow\uparrow\uparrow\rangle}}
60: \newcommand{\sdud}{\ensuremath{|\downarrow\uparrow\downarrow\rangle}}
61: \newcommand{\sddu}{\ensuremath{|\downarrow\downarrow\uparrow\rangle}}
62: \newcommand{\sddd}{\ensuremath{|\downarrow\downarrow\downarrow\rangle}}
63: \newcommand{\bsuuu}{\ensuremath{\langle\uparrow\uparrow\uparrow|}}
64: \newcommand{\bsuud}{\ensuremath{\langle\uparrow\uparrow\downarrow|}}
65: \newcommand{\bsudu}{\ensuremath{\langle\uparrow\downarrow\uparrow|}}
66: \newcommand{\bsudd}{\ensuremath{\langle\uparrow\downarrow\downarrow|}}
67: \newcommand{\bsduu}{\ensuremath{\langle\downarrow\uparrow\uparrow|}}
68: \newcommand{\bsdud}{\ensuremath{\langle\downarrow\uparrow\downarrow|}}
69: \newcommand{\bsddu}{\ensuremath{\langle\downarrow\downarrow\uparrow|}}
70: \newcommand{\bsddd}{\ensuremath{\langle\downarrow\downarrow\downarrow|}}
71: \newcommand{\balpha}{\boldsymbol{\alpha}}
72: \newcommand{\bbeta}{\boldsymbol{\beta}}
73: \newcommand{\bR}{\mathbf{R}}
74: \newcommand{\br}{\mathbf{r}}
75: \maketitle
76:
77: %\section{Introduction}
78: %
79:
80:
81: Recently a great number of theoretical and experimental effort has been
82: devoted to Bose-Einstein condensates (BECs).
83: The trapping techniques can be used to control the typical size of localized
84: states, such as a Gaussian wave packet, and the technique of Feshbach resonance
85: can allow both the strength and sign of the nonlinearity to be varied.
86:
87: BECs show a variety of interference phenomena that reflect the matter-wave
88: nature.
89: Among them, the BEC analog of Young's interference experiments is the most
90: noteworthy.
91: As demonstrated by Andrews {\it et al.}, a pair of BEC wave packets separately
92: located in double potential wells begins to diffuse after the release of the
93: trap and produces interference fringes\cite{Andre-1, Andre-2, Andre-3}.
94: Since this pioneering experiment, numerous studies on interference have been
95: devoted to creating a stable double-well trap to control the relative phase of
96: split condensates\cite{Steng,Pitaev,Liu,Hind,Hans,Shin,Coll}, except for some
97: novel experiments based on an atom-based Mach-Zehnder interferometer that
98: exploits a superposition of internal states of BEC\cite{Kozu,Hagl,Torii}.
99:
100: On the contrary, however, no work has been devoted to the BEC analog of single
101: electron (photon) dynamics through real double slits (DSs).
102: The release of a double-well trap cannot precisely mimic the passage through DSs.
103: The former treats the second-order interference between two wave packets (WPs)
104: with initially independent phases, while the latter treats the first-order
105: interference between two fractions of the initially identical WP.
106: A single BEC wave packet falling through DSs under uniform gravity will split
107: into distorted pieces.
108: However the fate of these split pieces is not clear because of the complicated
109: effect of diffraction at the slit openings (windows).
110: Furthermore, the effects of gravity and nonlinearity on double-slit interference
111: have not been examined quantitatively up to now.
112:
113: Both trapped and developing BECs are described using the Gross-Pitaevskii
114: equation (GPE) or the nonlinear Schr{\"o}dinger equation
115: (NSE)\cite{Pitaev, Liu, Coll, Holl, Roh}.
116: In this letter, by using NSE (GPE), the dynamics of a macroscopic WP falling
117: through the DSs is analyzed.
118: In contrast to the ordinary quantum mechanics, BEC has advantage in that it shows
119: a continuous time evolution of WP dynamics without being affected by the subtle
120: problem of quantum demolition measurement.
121: We here explore how WP splitting occurs at the slit openings and how the split
122: pieces reorganize to form the interference pattern.
123: We pay special attention to the effect of gravity and nonlinearity on the fringe
124: pattern.
125: We also discuss the designing of the experiment.
126:
127:
128: At low temperatures, $N_{0}$ atoms occupy the single quantum state and the
129: many-body wave function is described by a single macroscopic wave function $\phi$
130: which satisfies NSE, i.e., the time-dependent Schr{\"o}dinger equation with the
131: additional mean-field type nonlinear term.
132: In the presence of a harmonic trap and gravity, GPE (NSE) becomes
133: \begin{equation}
134: i\hbar \frac{\partial \phi}{\partial t}=
135: -\frac{\hbar^{2}}{2m}\nabla^{2}\phi+V_{trap}({\bf r})\phi+
136: U_{gravity}({\bf r})\phi+
137: u_{0}|\phi|^{2}\phi ,
138: \label{eq:2}
139: \end{equation}
140: where $u_{0}$ is the nonlinear interaction defined by
141: $u_{0}=\frac{4 \pi \hbar^{2} a}{ml}N_{0}$ [in 2-dimensional (2D) space] or
142: $\frac{4 \pi \hbar^{2} a}{m}N_{0}$ (in 3D space) with $a$, $m$ and $l$ being
143: scattering length, atomic mass and the characteristic length (to be defined below),
144: respectively.
145: $V_{trap}({\bf r})$ and $U_{gravity}({\bf r})$ are potentials for a harmonic trap
146: and gravity, respectively.
147:
148: Practically, a BEC WP in a harmonic trap exhibits a cigar-like
149: shape.\cite{Andre-1, Andre-2, Andre-3}
150: When the WP lengthens in the $z$ direction to form a nearly cylindrical shape,
151: cross-sectional dynamics of the WP, which is observed on the $x$-$y$ plane
152: bisecting the $z$ axis of the WP, can be approximately described using a 2D GPE
153: without losing any physical meaning.
154: Thus we perform 2D analysis that requires less computational cost than the 3D
155: version.
156: We assume an initial circularly trapped WP that falls toward the positive $y$
157: direction under uniform gravity (gravity constant $g$).
158: Then we set
159: $\nabla^{2}=\frac{\partial ^{2}}{\partial x ^{2}}
160: +\frac{\partial^{2}}{\partial y ^{2}}$
161: and
162: \begin{equation}
163: V_{trap}({\bf r})=\frac{1}{2}m\omega ^{2}(x^{2}+y^{2} ); \qquad
164: U_{gravity}({\bf r})=-mgy.
165: \label{eq:3}
166: \end{equation}
167: Using the characteristic length
168: $l=\sqrt{\hbar/(m\omega)}$, one can make all variables dimensionless
169: as
170: \begin{equation}
171: t^{\prime}=\omega t, \, x^{\prime}=\frac{x}{l}, \, y^{\prime}=\frac{y}{l}, \,
172: \phi^{\prime}=l\phi .
173: \label{eq:4a}
174: \end{equation}
175: Then GPE is reduced to
176: \begin{equation}
177: i\frac{\partial \phi^{\prime}}{\partial t^{\prime}}=
178: -\frac{1}{2} \left( \frac{\partial ^{2}}{\partial x^{\prime 2}}+
179: \frac{\partial^{2}}{\partial y^{\prime 2}} \right) \phi^{\prime}
180: +\frac{1}{2}(x^{\prime 2}+y^{\prime 2})\phi^{\prime}
181: -g^{\prime}y^{\prime}\phi^{\prime}
182: +u_{0}^{\prime }|\phi^{\prime}|^{2}\phi^{\prime},
183: \label{eq:4}
184: \end{equation}
185: % \begin{multline}
186: % i\frac{\partial \phi^{\prime}}{\partial t^{\prime}}=
187: % -\frac{1}{2} \left( \frac{\partial ^{2}}{\partial x^{\prime 2}}+
188: % \frac{\partial^{2}}{\partial y^{\prime 2}} \right) \phi^{\prime}
189: % +\frac{1}{2}(x^{\prime 2}+y^{\prime 2})\phi^{\prime}\\
190: % -g^{\prime}y^{\prime}\phi^{\prime}
191: % +u_{0}^{\prime }|\phi^{\prime}|^{2}\phi^{\prime},
192: % \label{eq:4}
193: % \end{multline}
194: where $u_{0}^{\prime }=4\pi N_{0}a/l$ and
195: $g^{\prime}=g/(l\omega^{2})$ are also dimensionless constants.
196: Initially, we ignore the contribution from gravity and, using the confining
197: potential, prepare a circularly symmetric Gaussian WP with its center of mass
198: at the origin ${\bf r}=(0,0)$,
199: \begin{equation}
200: \phi^{\prime}_{ini}=\frac{1}{ \sqrt{\pi}
201: \left( 1+\frac{u_0^\prime}{2\pi} \right)^{1/4}}
202: \exp \left(-\frac{x^{\prime 2}+y^{\prime 2}}{2 \sqrt{1+\frac{u_{0}^{\prime
203: }}{2\pi}}}\right) ,
204: \label{eq:5}
205: \end{equation}
206: which minimizes the total energy
207: $E^{\prime}=\int d^{2} {\bf r}^{\prime} ( \frac{1}{2}|\nabla
208: \phi^{\prime}|^{2} + V_{trap}^{\prime}|\phi^{\prime}|^{2}
209: + \frac{u_{0}^{\prime}}{2}|\phi^{\prime}|^{4} )$.
210: Here, the de Broglie wavelength of an initially stationary WP can be longer
211: than the size of the WP itself.
212: Thus small variation of the WP shape contributes little to the global phenomena.
213: The Gaussian WP profile is a simple substitute for an exact solution of GPE
214: as far as short-time dynamics of the WP is concerned.
215:
216: Our strategy is as follows:
217: (1) release the confining potential and switch on the uniform gravity;
218: (2) WP falls on DS;
219: (3) WP splits as it passes through the DS;
220: (4) the split pieces reorganize to form a fringe pattern.
221: Dynamics of the WP is obeyed by GPE in eq.(\ref{eq:4}) together with the Dirichlet
222: boundary condition at DS: hard walls forming the DS are taken to be one dimensional
223: along the $x$ direction.
224: We assume the interference phenomena appear in the short-time dynamics shorter than
225: the decoherence time and use GPE even after releasing the trap, closely following
226: the procedures in most of the existing works\cite{Pitaev, Liu, Coll, Holl, Roh}.
227:
228: Given any arbitrary value $u_{0}^{\prime}$, we systematically investigate the effect
229: of gravity on the fringe pattern, starting from a common WP with fixed width.
230: For this purpose, we always choose WP in the case of vanishing nonlinearity
231: ($u_{0}^{\prime} = 0$),
232: \begin{equation}
233: \phi^{\prime }_{ini}=\frac{1}{\sqrt{\pi}}
234: \exp \left[ -\frac{1}{2}(x^{\prime 2}+y^{\prime 2}) \right] ,
235: \label{eq:12}
236: \end{equation}
237: as an initial state for any value of nonlinearity.
238: This means that we have replaced $V_{trap}$ in eq.(\ref{eq:4}) by the
239: $u_{0}^{\prime }$-dependent confining potential,
240: $\frac{1}{2}(1+\frac{u_{0}^{\prime}}{2\pi})(x^{\prime 2}+y^{\prime 2})$, but have
241: taken the scaling in eq.(\ref{eq:4a}) using $l$ proper for $V_{trap}$ with
242: $u_{0}^{\prime}=0$.
243: Since we consider the falling of WP after releasing $V_{trap}$, the formal change
244: of the confining potential has no effect on the dynamics under gravity.
245:
246: We numerically integrate GPE in three ways, i.e., alternating direction implicit
247: method (ADI method), Crank-Nicholson's method, and split-step method, all of which
248: are found to yield identical results.
249: Figure~\ref{linearcase} shows the $x^{\prime}$-$y^{\prime}$ plane of size
250: $40{\times}40$ that we employ.
251: Units for spatial and temporal grids are
252: $\delta x^{\prime}=\delta y^{\prime}=0.1$, $\delta t=0.002$.
253: The center of WP, being located at the origin at $t=0$, begins to fall along the
254: $y^{\prime}$ direction due to the uniform gravity.
255: DS is set at $y^{\prime}=5$. Lengths for the central stopper and a pair of windows
256: are 3 and 1, respectively.
257: The fringe pattern will be observed on a virtual screen located at
258: $y^{\prime}=L_{sc}=18$.
259: The final time of the present computation is the time when the classical particle
260: (center of mass) reaches the screen, and is given by
261: $t=T_{max}=\sqrt{2L_{sc}/g^{\prime}}$.
262: Up to this time, we detect little fraction of WP escaping outside the whole grid
263: area through the boundaries at $x^{\prime}=\pm 20$.
264:
265: \begin{figure}[t]
266: \begin{minipage}{.37\textwidth}
267: \includegraphics[width=8cm]{fig1.eps}
268: \end{minipage}
269: \caption{Wave packet dynamics with $u_{0}^{\prime}=0$, $g^{\prime}=1$.
270: Intensity distributions of WP at $t^{\prime}=0.0$ and $t^{\prime}=$1.0--6.0.
271: Topmost panel illustrates the setting for double-slits
272: experiments.}
273: \label{linearcase}
274: \end{figure}
275: The time evolution of WP in the linear case ($u_{0}^{\prime}=0$) is given in
276: Fig.~\ref{linearcase}.
277: The gravity is fixed at $g^{\prime}=1$.
278: Figure~\ref{linearcase} shows that the falling WP breaks into a pair of pieces
279: during passage through the DS and then begins to form a clear fringe pattern.
280: Above the DS we find a remnant of WP, which has failed to fall through the DS,
281: interferes parallel to DS. Because of the gravity, the remnant will also fall
282: through DS sooner or later in due course.
283: Our additional calculation (to be described elsewhere) indicates that corresponding
284: WP dynamics in the case when either one of the DS windows is closed leads to no
285: fringe pattern and merely to a single humped diffusive WP.
286: To see the fringe pattern, therefore, it is essential to open both windows.
287: The results are consistent with the result of a single electron interference using
288: a DS.
289: \begin{figure}[h]
290: \begin{minipage}{.37\textwidth}
291: \includegraphics[width=8cm]{fig2.eps}
292: \end{minipage}
293: \caption{Intensity distribution of WP.
294: From top to bottom: $g^{\prime}=$1, 3, and 5.
295: From left to right: $u_{0}^{\prime}=-3$, 0, and 20.}
296: \label{nineshots}
297: \end{figure}
298:
299: Taking the result for the linear case and $g^{\prime}=1$ as a standard pattern, let
300: us proceed to the investigation of the role of tunable nonlinearity and gravity.
301: We shall focus on the fringe pattern on the screen at $y^{\prime}=L_{sc}$ at
302: $t=T_{max}$.
303: We choose $u_{0}^{\prime}=-3$, 0, 20 for the nonlinear term and $g^{\prime}=$1, 3, 5
304: for gravity.
305: Figure~\ref{nineshots}, which includes nine fringe patterns, represents the intensity
306: distribution of WP at $t=T_{max}$, and Fig.~\ref{ninesections} shows their cross
307: sections on the screen, with the total probability amplitude for this section being
308: normalized to unity.
309: We find that the nonlinearity is responsible for the localization or delocalization
310: of WP, while gravity governs the velocity of WP arriving at the screen and is
311: responsible for the de Broglie wavelength.
312: For a fixed nonlinearity, the increased gravity will shorten the de Broglie wavelength
313: at time $t=T_{max}$.
314: On the other hand, for a fixed gravity, the increased nonlinearity
315: $u_{0}^{\prime}$ will broaden the fringe pattern.
316: \begin{figure}
317: \begin{minipage}{.37\textwidth}
318: \includegraphics[width=8cm]{fig3.eps}
319: \end{minipage}
320: \caption{Fringe pattern on screen. Probability ($0-0.2$)
321: versus horizontal coordinate. Arrangement of panels is the same as in
322: Fig.~\ref{nineshots}.}
323: \label{ninesections}
324: \end{figure}
325:
326: \begin{figure}[t]
327: \begin{minipage}{.38\textwidth}
328: \begin{center}
329: \includegraphics[width=7.5cm]{fig4.eps}
330: \end{center}
331: \end{minipage}
332: \caption{Main interval versus gravity}
333: \label{peakinterval}
334: \end{figure}
335:
336: Let us develop a theoretical insight in greater detail.
337: Under a fixed nonlinearity, by using Fig.~\ref{ninesections}, one can evaluate the
338: peak-to-peak interval.
339: The gravity dependence of the main interval (between the central and second peaks)
340: is shown in Fig.~\ref{peakinterval}.
341: For each value of $u_{0}^{\prime}$, the interval decreases monotonically as a
342: function of $g^{\prime}$.
343: The peak interval is proportional to the de Broglie wavelength defined by
344: $\lambda ^{\prime} \equiv 2\pi/p^{\prime}$.
345: If we choose $p^{\prime}$ to be the momentum for the classical particle corresponding
346: to the center of WP,
347: $p^{\prime}=V_{max}=g^{\prime}T_{max} = g^{\prime}\sqrt{2L_{sc}/g^{\prime}}=\sqrt
348: {2g^{\prime}L_{sc}}$.
349: Then we obtain
350: \begin{equation}
351: \lambda^{\prime} =\frac{2\pi}{\sqrt{2g^{\prime}L_{sc}}}
352: \end{equation}
353: regardless of the value $u_{0}^{\prime}$, the validity of which is confirmed in
354: Fig.~\ref{peakinterval}.
355:
356:
357: On the other hand, the nonlinearity is responsible for WP broadening, namely,
358: the extension of the area in which a fringe pattern can be formed.
359: We shall make a simple evaluation of WP width at $t=T_{max}$.
360: Let us approximate the falling WP by a cylinder ($\phi = \pi^{-1/2}R^{-1}$) with
361: radius $R$.
362: By suppressing gravity, GPE becomes circularly symmetric and is given by
363: \begin{equation}
364: i\dot{\phi}=\left( -\frac{\partial^{2}}{2\partial
365: R^{2}}-\frac{1}{2R}\frac{\partial}{\partial R}+
366: u_{0}^{\prime}|\phi|^{2} \right) \phi.
367: \end{equation}
368: Using the above cylindrical solution in this equation, we have
369: $\dot{R}=i(u_{0}^{\prime}-\pi/2)\pi^{-1}R^{-1}$, whose solution is
370: $R\approx \sqrt{(u_{0}^{\prime}-\pi/2)t}$ in the large $t$ region.
371: Therefore, at $t=T_{max}$,
372: \begin{equation}
373: R_{sc} \propto \sqrt{\frac{u_{0}^{\prime}-\pi/2}{\sqrt{g^{\prime}}}} .
374: \end{equation}
375:
376: We then proceed to evaluate the peak number.
377: The peak-to-peak interval $\Delta x^{\prime}_{peak}$ is approximately equal to
378: the de Broglie wavelength $\lambda^{\prime}$ at $t=T_{max}$, so
379: $\Delta x^{\prime}_{peak}\propto \lambda^{\prime} \propto 1/\sqrt{g^{\prime}}$.
380: Eventually, the peak number (PN) at the screen can be given by
381: \begin{equation}
382: {\rm PN}=\frac{R_{sc}}{\Delta x^{\prime}_{peak}}\propto
383: \sqrt{\left( u_{0}^{\prime}-\frac{\pi}{2} \right)\sqrt{g^{\prime}}} .
384: \label{eq:19}
385: \end{equation}
386:
387: This approximate analytical expression for PN explains how the peak number
388: depends on the nonlinearity and gravity.
389: Table~\ref{NumberTable} obtained from our numerical fringe pattern on the screen,
390: lists PN for various values of $g^{\prime}$ and $u_{0}^{\prime}$ and is consistent
391: with eq.(\ref{eq:19}).
392: \begin{table}
393: \begin{center}
394: \begin{minipage}{.37\textwidth}
395: \includegraphics[width=6cm]{table1.eps}
396: \end{minipage}
397: \end{center}
398: \caption{Peak number}
399: \label{NumberTable}
400: \end{table}
401:
402: The transmission probability of WP at $t=T_{max}$ is the portion of WP that lies
403: below the DS at that time.
404: This probability is affected by the nonlinearity.
405: Let us define ${\rm Pr}(A)$ and ${\rm Pr}(B)$
406: for each portion lying above and below DS, respectively.
407: They are calculated at $t=T_{max}$ as
408: \begin{equation}
409: {\rm Pr}(A \; {\rm or}\; B)=\int \! \! \int_{A \; {\rm or} \; B}
410: dx^{\prime}dy^{\prime} |\phi^{\prime}|^{2} .
411: \label{eq:20}
412: \end{equation}
413: Noting ${\rm Pr}(A)+{\rm Pr}(B)=1$, ${\rm Pr}(B)$ gives the transmission probability.
414: ${\rm Pr}(B)$ at $t=T_{max}$ is found to depend on $u_{0}^{\prime},g^{\prime}$ and
415: takes values in the range $0.1\leq {\rm Pr}(B)\leq 0.3$.
416: For larger $u_{0}^{\prime}$, WP broadens before reaching the DS and therefore
417: ${\rm Pr}(B)$ diminishes.
418: On the other hand, for increased $g^{\prime}$, WP reaches the DS without broadening
419: very much because of the highly accelerated motion, and a large portion of WP dwells
420: around the central stopper, which suppresses ${\rm Pr}(B)$.
421: Thus ${\rm Pr}(B)$ is sensitive to the values of $u_{0}^{\prime}$ and $g^{\prime}$.
422: However, qualitative features of the interference, such as $g^{\prime}$ and
423: $u_{0}^{\prime}$ dependences of the peak-to-peak distance and peak number, remain
424: unchanged.
425:
426: We have so far investigated WP dynamics using scaled time and coordinates.
427: Here we suggest an experimental situation to which the present result is applicable.
428: Let us introduce the material-specific constant $\alpha=\sqrt{\hbar/m}$.
429: The value $g^{\prime}$ that we have employed determines $\omega$ and $l$.
430: With the use of $l=\sqrt{\hbar/(m\omega)}$ and
431: $g^{\prime}=g/(l\omega^{2})$, one obtains
432: \begin{equation}
433: \omega =\left( \sqrt{\frac{m}{\hbar}}\frac{g}{g^{\prime}} \right)^{2/3}=
434: \left( \frac{1}{\alpha} \right)^{2/3} \left( \frac{g}{g^{\prime}} \right)^{2/3}
435: \label{eq:22}
436: \end{equation}
437: and
438: \begin{equation}
439: l=\alpha^{4/3}\left( \frac{g}{g^{\prime}} \right)^{-1/3}.
440: \label{eq:23}
441: \end{equation}
442: The last value represents the scale of, for example, the WP width, the DS, and the
443: stopper.
444: Let us consider more concretely the case of $^{87}{\rm Rb}$ and $g^{\prime}=1$.
445: Noting that $\alpha=2.726\times 10^{-5}$ and the gravity constant on the earth
446: $g=9.80665$~m/s, we obtain $l=0.383$~$\mu$m.
447: In a laboratory of the space station, one can expect a weak gravity, e.g.,
448: $\hat{g}=g/1000$.
449: Then the experiment corresponding to $g^{\prime}=1$ is realized on a length scale
450: $\hat{l}$ that satisfies
451: \begin{equation}
452: \frac{\hat{l}}{l}= \left( \frac{\hat{g}}{g} \right)^{-1/3}=10,
453: \end{equation}
454: from which we obtain $\hat{l}=3.83$~$\mu$m.
455: The above equality indicates that on decreasing the gravity scale by three orders,
456: the corresponding length scale increases by one order.
457: In other words, the length scale is almost insensitive to the change of gravity and
458: the size of DS can be kept almost constant when designing the present experiment
459: under a weak gravitational field.
460:
461:
462: We have explored the effect of nonlinearity and gravity on the macroscopic fringe
463: pattern generated by BEC passing through DS.
464: Most of the activities so far concerning the BEC analog of Young's interference
465: experiment referred to a double-well trap leading to the 2nd-order interference
466: between two initially separated WPs with uncontrolled independent phases.
467: By contrast we mimic the BEC analog of the 1st-order interference (a single electron
468: interference).
469: A study of the BEC dynamics through the DS is identified with a search for the fate
470: of a 2D soliton showing a head-on collision with a hard-walled obstacle of finite size.
471: In our simulation, we find that split pieces of WP below the DS successfully reorganize
472: to form a clear fringe pattern.
473: The analytic expressions are derived for the peak-to-peak distance and peak number as
474: a function of gravity and nonlinearity.
475: It should be noted that our preliminary simulation for a more realistic 3D version of
476: the present system yielded the same result as found here.
477: A further analysis of wave-function patterns, such as nodal lines and phase rigidity,
478: which is developed in the context of quantum chaos\cite{berg}, are subjects which we
479: intend to study in the future.
480:
481:
482: K. N. is grateful to M. Ueda, M. Hosoda, M. Wilkinson and C. Kosaka for enlightening
483: discussions. T. A. thanks M. Mizobuchi and Y. Ohtake for discussions on the experimental
484: aspects of BECs.
485:
486: \begin{thebibliography}{20}
487: \bibitem{Andre-1}
488: M.~R.~Andrews, C.~G.~Townsend, H.~J.~Miesner, D.~S.~Durfee,
489: D.~M.~Kurn and W.~Ketterle: Science {\bf 275} (1997) 637.
490:
491: \bibitem{Andre-2}
492: M.~R.~Andrews, D.~M.~Kurn, H.~J.~Miesner, D.~S.~Durfee,
493: C.~G.~Townsend, S.~Inouye and W.~Ketterle:
494: Phys. Rev. Lett. {\bf 79} (1997) 553.
495:
496: \bibitem{Andre-3}
497: M.~R.~Andrews, D.~M.~Kurn, H.~J.~Miesner, D.~S.~Durfee,
498: C.~G.~Townsend, S.~Inouye and W.~Ketterle:
499: Phys. Rev. Lett. {\bf 80} (1998) 2967.
500:
501: \bibitem{Steng}
502: J.~Stenger, S.~Inouye, A.~P.~Chikkatur, D.~M.~Stamper-Kurn,
503: D.~E.~Pritchard and W.~Ketterle:
504: Phys. Rev. Lett {\bf 82} (1999) 4569.
505:
506: \bibitem{Pitaev}
507: L.~Pitaevskii and S.~Stringari:
508: Phys. Rev. Lett {\bf 83} (1999) 21.
509:
510: \bibitem{Liu}
511: W.-M.~Liu, B.~Wu and Q.~Niu:
512: Phys. Rev. Lett {\bf 84} (2000) 2294.
513:
514: \bibitem{Hind}
515: E.~A.~Hinds, C.~J.~Vale and M.~G.~Boshier:
516: Phys. Rev. Lett {\bf 86} (2001) 1462.
517:
518: \bibitem{Hans}
519: W.~H{\"a}nsel, J.~Reichel, P.~Hommelhoff and T.~W.~H{\"a}nsch:
520: Phys. Rev. A {\bf 64} (2001) 063607.
521:
522: \bibitem{Shin}
523: Y.~Shin, M.~Saba, T.~A.~Pasquini, W.~Ketterle, D.~E.~Pritchard
524: and A.~E.~Leanhardt: Phys. Rev. Lett {\bf 92} (2004) 050405.
525:
526: \bibitem{Coll}
527: L.~A.~Collins, L.~Pezz{\'e}, A.~Smerzi, G.~P.~Berman and A.~R.~Bishop:
528: Phys. Rev. A {\bf 71} (2005) 033628.
529:
530: \bibitem{Kozu}
531: M.~Kozuma, L.~Deng, E.~W.~Hagley, J.~Wen, R.~Lutwak,
532: K.~Helmerson, S.~L.~Rolston and W.~D.~Phillips:
533: Phys. Rev. Lett {\bf 82} (1999) 871.
534:
535: \bibitem{Hagl}
536: E.~W.~Hagley, L.~Deng, M.~Kozuma, M.~Trippenbach, Y.~B.~Band, M.~Edwards,
537: M.~Doery, P.~S.~Julienne, K.~Helmerson, S.~L.~Rolston and W.~D.~Phillips:
538: Phys. Rev. Lett {\bf 83} (1999) 3112.
539:
540: \bibitem{Torii}
541: Y.~Torii, Y.~Suzuki, M.~Kozuma, T.~Sugiura and T.~Kuga:
542: Phys. Rev. A {\bf 61} (2000) 041602(R).
543:
544: \bibitem{Holl}
545: M.~Holland and J.~Cooper:
546: Phys. Rev. A {\bf 53} (1996) 1954(R).
547:
548: \bibitem{Roh}
549: A.~R{\"o}hrl, M.~Naraschewski, A.~Schenzle and H.~Wallis:
550: Phys. Rev. Lett {\bf 78} (1997) 4143.
551:
552: \bibitem{berg}
553: K. F. Berggren and S. \AA berg (eds.):
554: {\it Quantum Chaos Y2K, Proceedings of Nobel Symposium}
555: (Royal Swedish Academy of Sciences/World Scientific,
556: Singapore, 2001).
557:
558: \end{thebibliography}
559: %\bibliography{BEC-Ando-0527}
560:
561: \end{document}
562: