1: \documentclass[prl,aps,onecolumn,superscriptaddress]{revtex4}
2: \usepackage{graphicx}
3:
4: \newcommand{\beq}{\begin{equation}}
5: \newcommand{\eeq}{\end{equation}}
6: \newcommand{\ba}{\begin{array}}
7: \newcommand{\ea}{\end{array}}
8: \newcommand{\bean}{\begin{eqnarray*}}
9: \newcommand{\eean}{\end{eqnarray}}
10: \newcommand{\bea}{\begin{eqnarray}}
11: \newcommand{\eea}{\end{eqnarray}}
12: \newcommand{\bc}{\begin{center}}
13: \newcommand{\ec}{\end{center}}
14: \newcommand{\bt}{\begin{table}}
15: \newcommand{\et}{\end{table}}
16: \newcommand{\eps}{\epsilon}
17: \newcommand{\bpsi}{{\bf \psi}}
18: \newcommand{\bphi}{{\bf \phi}}
19: \newcommand{\la}[1]{\label{#1}}
20: \newcommand{\p}{\partial}
21: \renewcommand{\d}{{\rm d}}
22: \newcommand{\de}{\delta}
23: \newcommand{\vp}{\varphi}
24: \newcommand{\ds}{\displaystyle}
25: \newcommand{\no}{\noindent}
26: \newcommand{\pp}[2]{{\partial #1 \over \partial #2}}
27: \newcommand{\ppn}[3]{{\partial^{#1} #2 \over \partial #3^{#1}}}
28: \newcommand{\Pain}{Painlev\'{e} }
29: \newcommand{\mbf}[1]{\mbox{\boldmath {$#1$}}}
30: \newcommand{\binomial}[2]{\left(\ba{c}#1 \\ #2 \ea\right)}
31: \newcommand{\rf}[1]{(\ref{#1})}
32: \newcommand{\beqno}{\begin{displaymath}}
33: \newcommand{\eeqno}{\end{displaymath}}
34: \newcommand{\dd}[2]{{\delta #1 \over \delta #2}}
35: \newcommand{\ddn}[3]{{\delta^{#1} #2 \over \delta #3^{#1}}}
36: \newcommand{\ie}{{\em i.e.}}
37: \newcommand{\been}{\begin{enumerate}}
38: \newcommand{\een}{\end{enumerate}}
39: \renewcommand{\aa}[2]{\alpha_{#1}^{(#2)}(r)}
40: \newcommand{\al}[1]{\alpha_{#1}(r)}
41: \renewcommand{\iff}{\Leftrightarrow}
42: \newcommand{\since}[1]{\stackrel{\mbox {using \rf{#1}}}{=}}
43: \newcommand{\used}[2]{#1 \dotfill #2\\}
44: \newcommand{\call}{{\cal L}}
45: \newcommand{\sn}{{\rm sn}}
46: \newcommand{\cn}{{\rm cn}}
47: \newcommand{\dn}{{\rm dn}}
48:
49: \begin{document}
50:
51:
52: \title{High-Order-Mode Soliton Structures in Two-Dimensional Lattices with Defocusing Nonlinearity}
53: \author{P. G.\ Kevrekidis}
54: \affiliation{Department of Mathematics and Statistics, University of
55: Massachusetts, Amherst MA 01003-4515}
56: \author{H.\ Susanto}
57: \affiliation{Department of Mathematics and Statistics, University of
58: Massachusetts, Amherst MA 01003-4515}
59: \author{Z.\ Chen}
60: \affiliation{Department of Physics and Astronomy, San Francisco State University,
61: San Francisco, CA 94132}
62: \affiliation{TEDA Applied Physics School, Nankai University, Tianjin, China}
63: \
64: \begin{abstract}
65: While fundamental-mode discrete solitons have been demonstrated with
66: both self-focusing and defocusing nonlinearity, high-order-mode localized
67: states in waveguide lattices have been studied thus far only for the
68: self-focusing case.
69: In this paper, the existence and stability regimes
70: of dipole, quadrupole and vortex soliton
71: structures in two-dimensional lattices induced with a defocusing nonlinearity
72: are examined by the theoretical and numerical analysis of a generic envelope
73: nonlinear lattice model. In particular,
74: we find that the stability of such high-order-mode solitons is quite
75: different from that with self-focusing nonlinearity. As a simple example, a
76: dipole (``twisted'') mode soliton which may be
77: stable in the focusing case becomes
78: unstable in the defocusing regime. Our results
79: may be relevant to other two-dimensional defocusing periodic nonlinear
80: systems such as Bose-Einstein condensates with a positive scattering length
81: trapped in optical lattices.
82: \end{abstract}
83:
84: \maketitle
85:
86: %{\it Introduction.}
87:
88: \section{Introduction}
89:
90: Ever since the suggestion of optically induced
91: lattices in photorefractive media such as Strontium Barium Niobate
92: (SBN) in \cite{efrem}, and its experimental realization in
93: \cite{moti1,neshevol03,martinprl04}, there has been an explosive growth in the
94: area of nonlinear waves and solitons in periodic lattices.
95: A stunning array of structures has been predicted and
96: experimentally obtained in lattices induced with a self-focusing
97: nonlinearity, including
98: (but not limited to) discrete dipole \cite{dip},
99: quadrupole \cite{quad}, necklace \cite{neck} and other
100: multi-pulse patterns (such as e.g. soliton stripes \cite{multi}),
101: %impurity
102: %solitons \cite{fedele},
103: discrete vortices
104: \cite{vortex}, and
105: rotary solitons \cite{rings}.
106: %, and others \cite{us_yuri}.
107: Such structures have a potential to be used as carriers and conduits for
108: data transmission and processing, in the context of all-optical
109: schemes. A recent review of this direction can be found in \cite{moti3}
110: (see also \cite{zc4}).
111:
112: %Another similarly growing in recent years direction of
113: %intriguing phenomenology arising due to the interplay of
114: %nonlinearity and discrete diffraction can be found
115: Many of these studies in induced lattices were
116: also triggered by the pioneering work done in fabricated AlGaAs
117: waveguide arrays \cite{7}. In the latter setting a multiplicity
118: of phenomena such as discrete diffraction, Peierls barriers,
119: diffraction management \cite{7a} and gap solitons \cite{7b} among
120: others \cite{eis3} were experimentally obtained. These phenomena, in turn,
121: triggered a tremendous increase also on the theoretical side of
122: the number of studies addressing such effectively discrete media;
123: see e.g. \cite{review_opt,general_review} for a number of relevant
124: reviews.
125:
126: Finally, yet another area where such considerations and structures
127: are relevant is that of soft-condensed matter physics, where droplets
128: of Bose-Einstein condensates (BECs) may be trapped in an (egg-carton)
129: two-dimensional optical lattice potential \cite{bloch}. The latter
130: field has also experienced a huge growth over the past few years,
131: including the prediction and manifestation of modulational instabilities
132: \cite{pgk}, the observation of gap solitons \cite{markus} and Landau-Zener
133: tunneling \cite{arimondo} among many other salient features; reviews of
134: the theoretical and experimental findings in this area have also been
135: recently appeared in \cite{konotop,markus2}.
136:
137: In light of all the above activity, it is interesting to note that
138: the only structure that has been experimentally observed in
139: two-dimensional (2d) lattices in ``defocusing'' media
140: consists of self-trapped ``bright'' wave packets (so-called ``staggered''
141: or gap solitons) excited in the vicinity of the edge of the first
142: Brillouin zone
143: \cite{moti1}.
144: However more complex
145: coherent structures have not yet been explored in lattices with
146: defocusing nonlinearity and their stability properties
147: have not yet been examined, to
148: the best of our knowledge. It
149: should be mentioned that the defocusing context is accessible
150: in the aforementioned settings. E.g., in the photorefractive
151: lattices, this can be done by appropriate reversal of the applied voltage
152: to the relevant crystal,
153: %in the waveguide arrays, it can be achieved
154: %by appropriate angle of the incident beam \cite{eis3};
155: while in
156: BECs, the defocusing nonlinearity corresponds to the most typical
157: case arising in dilute gases of $^{87}$Rb or $^{23}$Na.
158:
159: It is the aim of the present work to examine the non-fundamental soliton
160: structures (e.g., dipoles, multipoles, and vortices) in lattices with a
161: defocusing nonlinearity, and to illustrate the similarities and differences
162: in comparison to their counterparts in the focusing case. In particular,
163: we study dipole structures (consisting of two peaks) and quadrupole
164: structures (featuring four peaks), as well as vortices of topological
165: charge $S=1$ (cf. \cite{vortex}) in a 2D induced lattice with a defocusing
166: nonlinearity. These structures will be analyzed in detail for both cases,
167: namely, the ``on-site'' excitation (where the center of the structure is
168: on an empty lattice site between the excited ones) and
169: the ``inter-site'' excitation (where their
170: center is between two lattice sites and no empty lattice site
171: exists between the excited ones).
172:
173: Our study of these structures will be conducted
174: analytically and numerically (in the next two sections) in
175: the context of the most prototypical generic envelope lattice
176: model, the so-called discrete nonlinear Schr{\"o}dinger
177: (DNLS)
178: equation with a defocusing nonlinearity
179: \cite{dnls} which is related to all of the above
180: contexts \cite{review_opt,konotop}.
181: %This will, in turn, motivate
182: %us to initialize our defocusing photorefractive SBN crystals
183: %with the corresponding configurations and observe their dynamics
184: %experimentally (in the fourth section).
185: When we find the relevant structures to be unstable, we will
186: also briefly address the dynamical evolution of the instability,
187: through appropriately crafted numerical experiments.
188: Finally, in the last section, we will summarize
189: our findings and present our conclusions, and the interesting
190: experimental manifestations that they suggest.
191:
192: \section{Model and Theoretical Setup}
193:
194: As our generic envelope model
195: encompassing the main features of discrete diffraction and
196: defocusing nonlinearity we use the two-dimensional (2D) DNLS
197: equation:
198: \begin{equation}
199: i\dot{u}_{\mathbf{n}}=-C \left( \Delta _{2}u\right)
200: _{\mathbf{n}} + |u_{\mathbf{n}}|^{2}u_{\mathbf{n}},
201: \label{DNLS}
202: \end{equation}
203: where $u_{\mathbf{n}}$ is a complex amplitude of the electromagnetic wave in
204: nonlinear optics
205: \cite{review_opt}, or the BEC wave function at the nodes of
206: a deep 2D optical lattice \cite{konotop}; $\mathbf{n}$
207: is the (two-dimensional in the present study)
208: vector lattice index, and $\Delta_{2} $ the
209: standard discrete Laplacian. Furthermore, $C $ is the constant
210: of the intersite coupling (associated with the
211: interwell ``tunnelling rate'' \cite{konotop}),
212: and the overdot stands for the
213: derivative with respect to the evolution variable, which can be $z$ in optical
214: waveguide arrays, or the time
215: $t$ in the BEC model. We focus on standing-wave solutions
216: of the form $u_{\mathbf{n}}=\exp (-i\Lambda t)\phi _{\mathbf{n}}$, with
217: $\phi _{\mathbf{n}}$ satisfying the
218: %stationary
219: equation,
220: %
221: \begin{equation}
222: f(\phi _{\mathbf{n}},C) \equiv -\Lambda \phi _{\mathbf{n}}-C \Delta
223: _{2}\phi _{\mathbf{n}}+|\phi _{\mathbf{n}}|^{2}\phi _{\mathbf{n}}=0.
224: \label{steady}
225: \end{equation}
226: Perturbing around the solutions of Eq. (\ref{steady}) gives rise
227: to the linearization operator
228: \begin{eqnarray}
229: \mathcal{H}_{\mathbf{n}}^{(C)} = \left(
230: \begin{array}{cc}
231: -\Lambda +2|\phi _{\mathbf{n}}|^{2} & \phi _{\mathbf{n}}^{2} \\
232: \bar{\phi}_{\mathbf{n}}^{2} & -\Lambda +2|\phi _{\mathbf{n}}|^{2}\end{array}\right)
233: %\nonumber \\
234: -C \Delta _{2}\left(
235: \begin{array}{cc}
236: 1 & 0 \\
237: 0 & 1\end{array}\right),
238: \label{oper}
239: \end{eqnarray}
240: with the overbar denoting complex conjugation. Through an appropriate rescaling
241: of the equation, we can fix $\Lambda \equiv 1$. Our analysis uses as
242: a starting point the so-called anti-continuum limit, i.e., the case of
243: $C=0$, where
244: for the uncoupled sites,
245: \begin{eqnarray}
246: \phi_{\mathbf{n}}=r_{\mathbf{n}} e^{i\theta _{\mathbf{n}}},
247: \label{AC}
248: \end{eqnarray}
249: with the amplitude $r_{\mathbf{n}}$
250: being $0$ or $\sqrt{\Lambda}$, and the phase $\theta _{\mathbf{n}}$ being
251: an arbitrary constant. Continuation of such a solution to
252: nontrivial couplings necessitates that a certain, so-called
253: Lyapunov-Schmidt condition be satisfied \cite{dep}. The latter
254: imposes for the projection of
255: eigenvectors of the kernel of $\mathcal{H}_{\mathbf{n}}^{(0)}$
256: onto the system of stationary
257: equations to be vanishing. This solvability condition provides
258: a nontrivial constraint at every ``excited'' (i.e., $r_{\mathbf{n}}\neq0$)
259: site of the AC limit, namely:
260: \begin{equation}
261: -2ig_{\mathbf{n}}(\theta ,C ) \equiv -C e^{-i\theta
262: _{\mathbf{n}}}\Delta _{2}\phi _{\mathbf{n}}+C e^{i\theta
263: _{\mathbf{n}}} \Delta _{2}\bar{\phi}_{\mathbf{n}}=0.
264: \label{inter}
265: \end{equation}
266: It is interesting (and crucial for stability purposes) to note
267: that this equation has an extra $(-)$ sign in comparison to its
268: focusing counterpart.
269: The derivation of these solvability conditions is especially important
270: because the corresponding Jacobian
271: \begin{eqnarray}
272: \mathcal{M}_{ij}=
273: \partial g_{i}/\partial \theta_{j}
274: \label{jacobian}
275: \end{eqnarray}
276: has eigenvalues $\gamma$ that are
277: directly related to the ``regular'' eigenvalues of the linearization problem
278: $\lambda $, through the equation
279: \begin{eqnarray}
280: \lambda =\pm \sqrt{2\gamma }.
281: \label{eig}
282: \end{eqnarray}
283: Hence, the method that we use to derive the eigenvalues $\lambda$
284: (which fully determine the crucial issue of stability of the solution
285: for small $C$) consists of a perturbative expansion of the solution
286: from the AC limit
287: \begin{eqnarray}
288: \phi _{\mathbf{n}}=\phi _{\mathbf{n}}^{(0)}+C \phi
289: _{\mathbf{n}}^{(1)}+\dots,
290: \label{expansion}
291: \end{eqnarray}
292: which allows us to derive the principal
293: bifurcation conditions for a specific configuration and therefore
294: infer its linear stability properties through the eigenvalues of
295: $\mathcal{M}$ and their connection to the linearization eigenvalues
296: $\lambda$. Recall that a nonzero real part of {\it any}
297: eigenvalue is a necessary
298: and sufficient condition for an exponential instability in Hamiltonian
299: systems, such as the one considered herein.
300:
301: \section{Comparison of Analytical and Numerical Results}
302:
303: \subsection{General Terminology}
304: We start with
305: some general terminology
306: that we will use in this section. The designation in-phase
307: (IP) will be used for two sites such that their relative phase difference
308: is $0$, while out-of-phase (OP) will signify that it is $\pi$. Furthermore,
309: on-site (OS) will mean that the center of the configuration is on an
310: empty lattice
311: site (between the excited ones), while inter-site (IS) will signify that
312: the center is located between the excited lattice
313: sites (and no empty site exists between them).
314: For all modes, in the figures below, we show their power $P=\sum
315: |u_{\mathbf{n}}|^2$ as a function of the coupling strength $C$, as
316: well as the real and imaginary parts of the key eigenvalues (the
317: ones determining the stability of the configuration).
318: %In figures
319: %\ref{fig2} and \ref{fig3},
320: We start with the dipole configuration (consisting primarily of two
321: lattice sites; see Figs. \ref{fig1}-\ref{fig1c}). We also examine
322: the more complex quadrupole (see Figs. \ref{fig2}-\ref{fig2c})
323: and vortex (see Figs. \ref{fig3}-\ref{fig3a})
324: configurations. In all the cases, we
325: offer typical examples of the mode profiles and
326: stability for select values of $C$. When the configurations are
327: found to be unstable, we also give a typical example of the
328: instability evolution, for a relevant value of the coupling strength.
329: Another general feature that applies
330: to all modes is a continuous spectrum band extending for
331: $\lambda_i \in [\Lambda-8C, \Lambda]$. This latter trait significantly
332: affects the stability intervals of the structures in
333: comparison with their focusing counterparts as we will see also below
334: (since configurations may be stable for small $C$, but not for larger
335: $C$).
336:
337: The presentation of the figures will be uniform throughout the
338: manuscript in that in each pair of figures, we examine two
339: types of configurations (one in the left column and one in the
340: right column). The first figure of each pair will have five panels showing
341: $P$ as a function of $C$, the principal real eigenvalues
342: (second panel) and imaginary eigenvalues (third panel). In
343: these plots, the numerical results are shown by the solid (blue)
344: line, while the analytical results by the dashed (red) line.
345: The fourth and fifth panels show typical examples of the
346: relevant configuration (obtained through a fixed point iteration
347: of the Newton type) and its stability eigenvalues
348: (shown through the spectral plane $(\lambda_r,\lambda_i)$
349: for the eigenvalues $\lambda=\lambda_r + i \lambda_i$).
350: The accompanying second figure will show the result of a typical
351: evolution of an unstable mode, perturbed by a random perturbation
352: of amplitude $10^{-4}$, in order to accelerate the instability evolution.
353: The four contour plot panels (one set on the left and one on the
354: right) will display the solution's squared absolute value for four
355: different values of the evolution variable; the bottom panel will
356: show the dynamical evolution of the sites chiefly ``participating''
357: in the solution. A fourth-order Runge-Kutta scheme has been used for
358: the numerical integration results presented herein.
359:
360: To facilitate the reader, a summary of the results, encompassing
361: our main findings reported below is offered in Table 1. The
362: table summarizes the configurations considered, their
363: linear stability and the outcome of their dynamical evolution
364: for appropriate initial conditions in the instability regime.
365: Note that if the solutions are unstable for all $C$, they
366: are denoted as such, while if they are partially stable for
367: a range of coupling strengths, their interval of stability is
368: explicitly mentioned. Details of our analytical results and
369: their connection/comparison with the numerical findings
370: are offered in the rest of this section.
371:
372: \begin{table*}[t]
373: \begin{center}
374: \begin{tabular}{|l|c|c|c|c|}
375: \hline
376: & \multicolumn{2}{|c|}{On-site} & \multicolumn{2}{|c|}{Inter-site} \\
377: \cline{2-5}
378: Type & Stability & Instability Outcome & Stability & Instability Outcome \\ \hline
379: In-phase Dipole & Unstable & 1-Site Pulse & $C<0.064$ & 1-Site Pulse \\ \hline
380: Out-of-phase Dipole & $C<0.092$ & Decay & Unstable & 1-Site Pulse \\ \hline
381: In-phase Quadrupole & Unstable & Breathing Behavior & $C<0.047$ & 1-Site Pulse
382: \\ \hline
383: Out-of-phase Quadrupole & $C<0.08$ & 1-Site Pulse & Unstable & 2-Site Mode \\ \hline
384: Vortex & $C<0.095$ & 1-Site Pulse & $C<0.095$ & 1-Site Pulse \\ \hline
385: \end{tabular} \end{center}
386: \caption{Summary of the stability results for all the configurations
387: presented below. For partially stable (near the anti-continuum limit)
388: solutions their interval of stability (for $\Lambda=1$) is given.
389: In each case, the outcome of the instability evolution for the parameters
390: and initial conditions considered below is also mentioned.} \label{table1}
391: \end{table*}
392:
393:
394: \subsection{Dipole Configurations}
395:
396: \begin{figure}[tbp]
397: \begin{center}
398: \hskip-0.15cm
399: \begin{tabular}{cc}
400: \includegraphics[height=6cm,width=6cm]{offS0pp1.eps}
401: \includegraphics[height=6cm,width=6cm]{onS0pp1.eps} & \\
402: \includegraphics[height=6cm,width=6cm]{offS0pp2.eps}
403: \includegraphics[height=6cm,width=6cm]{onS0pp2.eps} &
404: \end{tabular}\end{center}
405: \par
406: \vskip-0.7cm
407: \caption{(Color Online)
408: The first line of panels shows the power $P$ vs. coupling $C$
409: for the inter-site (IS), in-phase (IP) mode (left) and on-site (OS),
410: IP mode (right). The second lines show their maximal real eigenvalues
411: and the third their first few imaginary eigenvalues.
412: The solid (blue) lines illustrate the numerical results, while
413: the dashed (red) lines the analytical ones.
414: The fourth and fifth panels show the contour plot of
415: the mode profile (fourth panel) and the corresponding spectral plane
416: of eigenvalues $\lambda=\lambda_r + i \lambda_i$ (fifth panel);
417: The left two panels are for the IS-IP mode for $C=0.08$ and $C=0.116$
418: respectively. The right panel shows the OS-IP mode for $C=0.08$.}
419: \label{fig1}
420: \end{figure}
421:
422: \begin{figure}[tbp]
423: \begin{center}
424: \hskip-0.15cm
425: \begin{tabular}{cc}
426: \includegraphics[height=6cm,width=6cm]{ISIP1.eps}
427: \includegraphics[height=6cm,width=6cm]{OSIP1.eps} & \\
428: \includegraphics[height=6cm,width=6cm]{ISIP2.eps}
429: \includegraphics[height=6cm,width=6cm]{OSIP2.eps} &
430: \end{tabular}
431: \end{center}
432: \par
433: \vskip-0.7cm
434: \caption{(Color Online) The four panels at the top left corner show the contour
435: plots of the dynamical evolution of the unstable inter-site,
436: in phase (IS-IP) solution for $C=0.08$. The respective times are
437: $t=50$ and $t=150$ in the top and $t=250$ and $t=350$ in
438: the second row. The panel at the bottom left shows the dynamical
439: evolution of the square modulus of the
440: principal two sites participating in the IS-IP
441: solution as a function of time. From both of the above,
442: it is clear that the configuration relaxes into a single site soliton.
443: The right panels show the same features for the on-site, in-phase
444: (OS-IP) solution, which also relaxes (but more slowly)
445: into a single-site configuration.}
446: \label{fig1a}
447: \end{figure}
448:
449: \subsubsection{Inter-site, In-Phase Mode}
450: Figures \ref{fig1}-\ref{fig1a} encompass our results for the
451: two types of IP dipole solutions
452: (i.e., initialized at the AC limit with two in-phase excited sites).
453: The
454: IS-IP mode of the left panels
455: is theoretically found to possess 1 imaginary eigenvalue
456: pair (and, hence, is stable for small $C$)
457: \begin{eqnarray}
458: \lambda \approx \pm 2 \sqrt{C} i.
459: \label{ISIP}
460: \end{eqnarray}
461: The collision with the continuous
462: spectrum described above
463: causes the mode to become unstable for sufficiently large
464: $C$; the theoretically predicted instability
465: threshold (obtained by equating the eigenvalue of
466: Eq. (\ref{ISIP}) with the lower edge of the phonon
467: band located at $\Lambda-8 C$) is $C=0.0625$, the
468: numerically found one is $C \approx 0.064$. Additional
469: instability may ensue when the monotonicity of the $P$ vs. $C$ curve
470: changes (we have found this to be a general feature of the defocusing
471: branches). The fourth and fifth panels show the mode and its
472: linearization eigenvalues for $C=0.08$ and $C=0.116$.
473: In fact, the dynamical evolution of the mode is demonstrated
474: for the case of $C=0.08$, illustrating that only one of the two
475: sites eventually persists, after the demonstrably oscillatory
476: instability destroys the configuration for $t>100$.
477:
478: \subsubsection{On-site, In-Phase Mode}
479: The OS-IP mode of the right panels of Figs. \ref{fig1}-\ref{fig1a}
480: is always unstable due to a real pair, theoretically found to be
481: \begin{eqnarray}
482: \lambda \approx \pm 2 C,
483: \label{OSIP}
484: \end{eqnarray}
485: for small $C$. Notice once again the remarkable accuracy
486: of this theoretical prediction, in comparison with the numerically
487: obtained eigenvalue.
488: The fourth and fifth right panels of Fig. \ref{fig1} show the
489: mode and its stability for $C=0.08$. Its dynamical evolution in the
490: right column of Fig. \ref{fig1a} shows its slow disintegration
491: into a single-site solitary wave.
492:
493: \subsubsection{Inter-site, Out-of-phase Mode}
494:
495: \begin{figure}[tbp]
496: \begin{center}
497: \hskip-0.15cm
498: \begin{tabular}{cc}
499: \includegraphics[height=6cm,width=6cm]{offS0pm1.eps}
500: \includegraphics[height=6cm,width=6cm]{onS0pm1.eps} & \\
501: \includegraphics[height=6cm,width=6cm]{offS0pm2.eps}
502: \includegraphics[height=6cm,width=6cm]{onS0pm2.eps} &
503: \end{tabular}
504: \end{center}
505: \par
506: \vskip-0.7cm
507: \caption{(Color Online) Similar to Fig. \ref{fig1}, but now
508: for the inter-site, out-of-phase (IS-OP) mode (left panels)
509: and for the on-site, out-of-phase mode (OS-OP). The fourth
510: and fifth rows of panels are for $C=0.08$ and for
511: $C=0.116$ in both cases.}
512: \label{fig1b}
513: \end{figure}
514:
515:
516: \begin{figure}[tbp]
517: \begin{center}
518: \hskip-0.15cm
519: \begin{tabular}{cc}
520: \includegraphics[height=6cm,width=6cm]{ISOP1.eps}
521: \includegraphics[height=6cm,width=6cm]{OSOP1.eps} & \\
522: \includegraphics[height=6cm,width=6cm]{ISOP2.eps}
523: \includegraphics[height=6cm,width=6cm]{OSOP2.eps} &
524: \end{tabular}
525: \end{center}
526: \par
527: \vskip-0.7cm
528: \caption{(Color Online) Same as in Fig. \ref{fig1a}, but now for the IS-OP mode
529: (left panels) and the OS-OP mode (right panels). The former shown
530: at times $t=25$, $t=50$ (top row), $t=150$ and $t=250$ (second row),
531: in the case of
532: $C=0.08$ finally results into a single-site configuration (as is
533: also illustrated by the bottom panel showing the two principal
534: sites participating in the mode). The latter, shown for $C=0.1$,
535: at times $t=50$, $t=100$ (top row), $t=150$ and $t=200$ (second row)
536: is entirely destroyed by
537: the instability resulting into small amplitude excitations.}
538: \label{fig1c}
539: \end{figure}
540:
541: Figures \ref{fig1b}-\ref{fig1c} illustrate the two dipole, out-of-phase
542: modes.
543: The left panels of the figures correspond to the IS-OP mode;
544: this one is also immediately unstable (as one departs from the
545: anti-continuum limit),
546: due to a real pair which is
547: \begin{eqnarray}
548: \lambda \approx 2 \sqrt{C},
549: \label{ISOP}
550: \end{eqnarray}
551: for small $C$. The fourth and fifth panels of Fig. \ref{fig1b}
552: show the relevant mode for $C=0.08$ and $C=0.116$, showing its
553: 1 and 2 unstable real eigenvalue pairs respectively.
554: The numerical experiment highlighting the evolution of the
555: mode for the case of $C=0.08$ is shown in the left panel
556: of Fig. \ref{fig1c}. Clearly, in this case as well, the
557: positive real eigenvalue leads to the growth of one of the
558: two sites constituting the dipole, and the eventual formation
559: of a single-site solitary pulse.
560:
561: \subsubsection{On-site, Out-of-phase Mode}
562: The right panels of Fig. \ref{fig1b}-\ref{fig1c} show the
563: OS-OP mode. The stability analysis of this waveform shows
564: that it
565: possesses an imaginary
566: eigenvalue
567: \begin{eqnarray}
568: \lambda \approx 2 C i.
569: \label{OSOP}
570: \end{eqnarray}
571: This leads to an instability upon
572: collision (occurring
573: theoretically for $C= 0.1$, numerically for $C\approx 0.092$)
574: with the lower edge (located at $\Lambda-8 C$) of the
575: continuous band of phonon modes.
576: The mode is shown for $C=0.08$ and $C=0.116$ in the right panels
577: of Fig. \ref{fig1b}.
578: The direct integration
579: of the unstable solution with $C=0.1$ is shown in the right panels
580: of Fig. \ref{fig1c}, indicating that in this case
581: the mode completely
582: disappears (because of the oscillatory instability) into
583: extended wave, small amplitude radiation.
584:
585: \subsection{Quadrupole Confirgurations}
586:
587: \subsubsection{Inter-site, In-phase Mode}
588:
589: \begin{figure}[tbp]
590: \begin{center}
591: \hskip-0.15cm
592: \begin{tabular}{cc}
593: \includegraphics[height=6cm,width=6cm]{offS0qpp1.eps}
594: \includegraphics[height=6cm,width=6cm]{onS0qpp1.eps} & \\
595: \includegraphics[height=6cm,width=6cm]{offS0qpp2.eps}
596: \includegraphics[height=6cm,width=6cm]{onS0qpp2.eps} &
597: \end{tabular}\end{center}
598: \par
599: \vskip-0.7cm
600: \caption{(Color Online) The top three lines of panels show the same features
601: as the corresponding ones of figure \ref{fig1} but now for the quadrupole
602: IS-IP mode (left) and the quadrupole OS-IP mode (right).
603: The contour
604: plot of the real part of the modes and the spectral plane
605: of their linearization eigenvalues are shown in
606: the fourth and fifth rows for $C=0.05$ and $C=0.1$ in the case of
607: the former mode, while the latter is only shown for $C=0.05$.}
608: \label{fig2}
609: \end{figure}
610:
611: \begin{figure}[tbp]
612: \begin{center}
613: \hskip-0.15cm
614: \begin{tabular}{cc}
615: \includegraphics[height=6cm,width=6cm]{QISIP1.eps}
616: \includegraphics[height=6cm,width=6cm]{QOSIP1.eps} & \\
617: \includegraphics[height=6cm,width=6cm]{QISIP2.eps}
618: \includegraphics[height=6cm,width=6cm]{QOSIP2.eps} &
619: \end{tabular}
620: \end{center}
621: \par
622: \vskip-0.7cm
623: \caption{(Color Online) Same as Fig. \ref{fig1a}, but now for the quadrupole
624: IS-IP mode with $C=0.08$ (left panels) and the quadrupole OS-IP mode
625: with $C=0.05$ (right
626: panels). The left panels are for $t=200$, $t=300$ (top),
627: $t=400$ and $t=500$ (middle) and illustrate alongside the
628: bottom panel (containing the evolution of the principal
629: four sites participating in the structure)
630: how the configuration eventually degenerates to a single-site
631: soliton. The right panels are for $t=50$, $t=150$, (top row) $t=250$
632: and $t=350$ (second row) and show together with the bottom panel the complex
633: oscillation (breathing) involved in the behavior of the
634: quadrupole OS-IP mode for $C=0.05$.}
635: \label{fig2a}
636: \end{figure}
637:
638: Figures \ref{fig2}-\ref{fig2a} show the quadrupolar mode
639: with four in-phase participating sites when centered
640: between lattice sites in the left panels of the figures.
641: This mode is theoretically predicted to have two
642: imaginary (for small $C$) eigenvalue pairs with
643: \begin{eqnarray}
644: \lambda \approx 2 \sqrt{C} i
645: \label{QISIP}
646: \end{eqnarray}
647: and one imaginary pair with
648: \begin{eqnarray}
649: \lambda \approx \sqrt{8 C} i.
650: \label{QISIP2}
651: \end{eqnarray}
652: As a result, this mode (shown in the fourth row
653: panels of Fig. \ref{fig2} for $C=0.05$ and $C=0.1$)
654: is unstable due to the collision of the above eigenvalues
655: with the continuous spectrum occurring theoretically for
656: $C \approx 0.0477$, while in the numerical computations it happens
657: for $C \approx 0.047$. The outcome of the instability shown
658: in Fig. \ref{fig2a} for $C=0.08$ is the degeneration of
659: the quadrupolar mode into a single-site excitation.
660:
661: \subsubsection{On-site, In-phase Mode}
662:
663: The right panels of Figs. \ref{fig2}-\ref{fig2a} show the
664: case of the on-site, in-phase mode. The latter is found to
665: always be unstable due to a real eigenvalue pair of
666: \begin{eqnarray}
667: \lambda \approx \pm 4 C
668: \label{QOSIP}
669: \end{eqnarray}
670: and a double, real eigenvalue pair of
671: \begin{eqnarray}
672: \lambda \pm \sqrt{12} C.
673: \label{QOSIP2}
674: \end{eqnarray}
675: This can also be clearly observed in the fourth and fifth
676: panels of Fig. \ref{fig2}, showing the mode and its stability for $C=0.05$.
677: The dynamical evolution of the unstable mode for $C=0.05$
678: is shown in the panels of Fig. \ref{fig2a}.
679: Both from the contour plots at the different times and from
680: the dynamical evolution of the main sites participating in
681: the structure, it can be inferred that the mode embarks in
682: an oscillatory breathing, without being ultimately destroyed
683: in this case.
684:
685: \subsubsection{Inter-site, Out-of-Phase Mode}
686:
687: \begin{figure}[tbp]
688: \begin{center}
689: \hskip-0.15cm
690: \begin{tabular}{cc}
691: \includegraphics[height=6cm,width=6cm]{offS0qpm1.eps}
692: \includegraphics[height=6cm,width=6cm]{onS0qpm1.eps} & \\
693: \includegraphics[height=6cm,width=6cm]{offS0qpm2.eps}
694: \includegraphics[height=6cm,width=6cm]{onS0qpm2.eps} & \\
695: \end{tabular}
696: \end{center}
697: \par
698: \vskip-0.7cm
699: \caption{(Color Online) Similar to Fig. \ref{fig1} but for the
700: quadrupole IS-OP mode (left panels) and the quadrupole OS-OP mode
701: (right panels). The fourth and fifth rows show the modes and their
702: stability for $C=0.08$ and $C=0.116$ in each case.}
703: \label{fig2b}
704: \end{figure}
705:
706: \begin{figure}[tbp]
707: \begin{center}
708: \hskip-0.15cm
709: \begin{tabular}{cc}
710: \includegraphics[height=6cm,width=6cm]{QISOP1.eps}
711: \includegraphics[height=6cm,width=6cm]{QOSOP1.eps} & \\
712: \includegraphics[height=6cm,width=6cm]{QISOP2.eps}
713: \includegraphics[height=6cm,width=6cm]{QOSOP2.eps} &
714: \end{tabular}
715: \end{center}
716: \par
717: \vskip-0.7cm
718: \caption{(Color Online) Same as Fig. \ref{fig1a}, but for
719: the evolution of the quadrupole IS-OP mode with $C=0.08$ (left panels)
720: the quadrupole OS-OP mode with $C=0.1$ (right panels). The former
721: is shown for $t=25$, $t=50$ (top row), $t=150$ and $t=250$ (second row)
722: substantiating
723: (together with the bottom panel showing the principal four sites
724: of the branch) its resulting into a long-lived, breathing two-site waveform.
725: The latter is shown for $t=50$, $t=150$ (top row), $t=250$ and $t=350$
726: (second row)
727: indicating its degeneration into a single-site configuration.}
728: \label{fig2c}
729: \end{figure}
730:
731: We next consider the case of the IS-OP mode in Figs. \ref{fig2b}-\ref{fig2c}.
732: Our analytical results for this mode show that for small values of
733: $C$, we should expect to find it to be immediately unstable due to
734: three real pairs of eigenvalues, namely a single one with
735: \begin{eqnarray}
736: \lambda \approx \pm \sqrt{8 C}
737: \label{QISOP1}
738: \end{eqnarray}
739: and a double one with
740: \begin{eqnarray}
741: \lambda \approx \pm 2
742: \sqrt{C}.
743: \label{QISOP2}
744: \end{eqnarray}
745: This expectation is once again confirmed by the numerical results
746: of the left panel of Fig. \ref{fig2b}. The fourth and fifth panels
747: show the mode and the spectral plane of its linearization
748: for the cases of $C=0.08$ and $C=0.116$. The dynamical evolution of this
749: mode also gives an interesting result, in that it produces, upon
750: manifestation of the instability, a long-lived, two-site oscillatory
751: mode, as is illustrated in the left panels of Fig. \ref{fig2c}
752: for $C=0.08$.
753:
754: \subsubsection{On-site, Out-of-phase Mode}
755:
756: Finally, the last one among the quadrupolar modes is the
757: OS-OP mode, examined in the right panels of Figs. \ref{fig2b}-\ref{fig2c}.
758: Our theoretical analysis predicts that this mode should have
759: a double imaginary eigenvalue pair of
760: \begin{eqnarray}
761: \lambda\approx \pm 2 C i
762: \label{QOSOP1}
763: \end{eqnarray}
764: and a single imaginary pair of
765: \begin{eqnarray}
766: \lambda \approx 4 C i.
767: \label{QOSOP2}
768: \end{eqnarray}
769: These, in turn, imply that the mode is stable for small $C$, but
770: becomes destabilized upon collision of the larger one among these
771: eigenvalues with the continuous band of phonons. This is numerically
772: found to occur for $C \approx 0.08$, while it is theoretically
773: predicted, based on the above eigenvalue estimates, to take place
774: for $C =0.083$. The mode's stability analysis is
775: shown in the fourth and fifth panel
776: of Fig. \ref{fig2b} for $C=0.08$ and $C=0.116$; for $C=0.1$, and
777: its dynamical evolution is examined in the right panels of Fig.
778: \ref{fig2c}. In this case, we do find that the mode essentially
779: degenerates to a single-site solitary wave.
780:
781: \subsection{Vortex Configuration}
782:
783: \begin{figure}[tbp]
784: \begin{center}
785: \hskip-0.15cm
786: \begin{tabular}{cc}
787: \includegraphics[height=6cm,width=6cm]{offS1p1.eps}
788: \includegraphics[height=6cm,width=6cm]{onS1p1.eps} & \\
789: \includegraphics[height=6cm,width=6cm]{offS1p2.eps}
790: \includegraphics[height=6cm,width=6cm]{onS1p2.eps} &
791: \end{tabular}\end{center}
792: \par
793: \vskip-0.7cm
794: \caption{(Color Online) The same features as in Figure 1
795: are shown here for the IS vortex of topological charge $S=1$
796: (left) and the OS vortex of $S=1$ (right). In this case both
797: the real (fourth line) and imaginary (fifth line) parts of the
798: solution are shown (and their stability in the sixth line) for
799: $C=0.08$ and $C=0.116$.}
800: \label{fig3}
801: \end{figure}
802:
803: \begin{figure}[tbp]
804: \begin{center}
805: \hskip-0.15cm
806: \begin{tabular}{cc}
807: \includegraphics[height=6cm,width=6cm]{VIS1.eps}
808: \includegraphics[height=6cm,width=6cm]{VOS1.eps} & \\
809: \includegraphics[height=6cm,width=6cm]{VIS2.eps}
810: \includegraphics[height=6cm,width=6cm]{VOS2.eps} &
811: \end{tabular}
812: \end{center}
813: \par
814: \vskip-0.7cm
815: \caption{(Color Online) Same as in Fig. \ref{fig1a} but
816: for the dynamical evolution of the IS (left panels) and
817: the OS (right panels) vortex of topological charge
818: $S=1$. Both cases are shown for $C=0.1$. Both the left
819: and right panels of evolution are for $t=50$, $t=150$ (top row),
820: $t=250$ and $t=350$ (second row) and alongside the bottom panels showing
821: the principal four sites of the vortex they show how
822: the vortices transform themselves into fundamental
823: solitary waves centered on a single site (in the latter case,
824: interestingly, the site of the largest excitation of the ensuing
825: wave is not one among the four principal excitation sites of
826: the original OS vortex).}
827: \label{fig3a}
828: \end{figure}
829:
830: \subsubsection{Inter-site Vortices}
831:
832: Finally, Figs. \ref{fig3}-\ref{fig3a} show similar features,
833: but now for the IS (left panels) and OS (right panels) vortex
834: solutions \cite{dnls,dep}.
835: The former has a theoretically predicted double pair of eigenvalues
836: \begin{eqnarray}
837: \lambda \approx \pm 2 C i.
838: \label{VIS}
839: \end{eqnarray}
840: leading to
841: an instability upon collision with the continuum band for
842: $C\geq 0.095$ ($C\geq 0.1$ theoretically). In this
843: case, there is also an eigenvalue of higher
844: order
845: \begin{eqnarray}
846: \lambda \approx \pm 4 C^2 i.
847: \label{VIS2}
848: \end{eqnarray}
849: which obviously depends more weakly on $C$.
850: The fourth and fifth panels of Fig. \ref{fig3} show the
851: real and the imaginary parts of the vortex configuration
852: for $C=0.08$ and $C=0.116$ and the sixth panel the
853: corresponding spectral planes. The dynamical evolution
854: of the vortex of topological charge $S=1$, for $C=0.1$ is shown
855: in the left panels of Fig. \ref{fig3a}, indicating that the
856: vortex also, upon the occurrence of the oscillatory instability,
857: becomes a single-site solitary wave.
858:
859: \subsubsection{On-site Vortices}
860:
861: On the other hand, the OS vortices are shown in the right panels
862: of Figs. \ref{fig3}-\ref{fig3a}. In this case, we theoretically
863: find that the vortex, for small $C$, should have a double pair
864: of eigenvalues
865: \begin{eqnarray}
866: \lambda \approx 2 C i
867: \label{VOS}
868: \end{eqnarray}
869: and a single, higher order pair of eigenvalues
870: \begin{eqnarray}
871: \lambda \approx \pm \sqrt{32} C^3 i.
872: \label{VOS2}
873: \end{eqnarray}
874: The former eigenvalue pairs, upon collision with the continuous
875: spectrum, lead to an instability, theoretically predicted to
876: occur at $C=0.1$ and numerically found to happen for $C \approx 0.095$.
877: The on-site mode (and its stability) is shown in the fourth-sixth right
878: panels of Fig. \ref{fig3} for $C=0.08$ and $C=0.116$. Its evolution
879: (for $C=0.1$) is shown in the right panel of Fig. \ref{fig3a}, where
880: it is again seen that the mode degenerates from an $S=1$ to an $S=0$
881: structure, i.e., a single-site solitary wave with no topological charge.
882:
883:
884: \subsection{General Principles Derived From Stability Considerations}
885:
886: It is interesting to note as an over-arching conclusion that the
887: stability intervals of the defocusing structures are different from
888: those of their focusing counterparts (especially when they are stable
889: close to the AC limit) because of the collisions with the continuous
890: spectrum
891: band edge; the latter is at $\lambda=\Lambda$ in the focusing case, while it
892: is at $\lambda=\Lambda-8C$ in the defocusing setting. Another similarly general
893: note is an immediate inference on whether the structures are stable or
894: not; this can be made based on the knowledge of whether their focusing
895: counterparts are stable or not and the transformation from the former
896: to the latter through the staggering transform: $u_{n,m}=
897: (-1)^{n+m} v_{n,m}$. For instance, IP two-site configurations (both OS and
898: IS) are known to be generically unstable in the focusing regime \cite{dep};
899: through the staggering transformation,
900: OS-IP of the focusing case
901: remains OS-IP in the defocusing, while IS-IP of the focusing
902: becomes IS-OP in the defocusing.
903: Hence, these two should be expected to be always
904: unstable, while the remaining two (OS-OP in both focusing and
905: defocusing and IS-OP of the focusing, which becomes IS-IP in the
906: defocusing) should similarly be expected
907: to be linearly stable close to the AC-limit, as is indeed observed.
908: Notice that, interestingly enough, for the vortex states the staggering
909: transformation indicates that the stability is not modified between
910: the focusing and defocusing cases. This is because for an IS vortex, it
911: transforms an $S=1$ state into an $S=-1$ state (which is equivalent
912: to the former, in terms of stability properties), while the OS vortex
913: remains unchanged by the transformation.
914: However, as mentioned above, these considerations are not sufficient to
915: compute the instability thresholds for initially stable modes, among other
916: things. They do, nonetheless, provide a guiding principle for inferring
917: the near-AC limit stability of the defocusing staggered states, based
918: on their focusing counterparts.
919:
920: \section{Conclusions and Future Challenges}
921:
922: In this paper, we have studied in detail some of the principal
923: multi-site solitary wave structures that emerge in the context
924: of defocusing nonlinearities, examining, in particular,
925: dipole, quadrupole and vortex configurations. We have illustrated
926: which ones among these states can potentially be stable
927: (e.g. IS-IP and OS-OP for both dipoles and quadrupoles, as
928: well as the vortices) and those that will {\it always} be unstable
929: (e.g. IS-OP and OS-IP modes for both dipoles and quadrupoles).
930: We have also provided detailed analytical estimates of the
931: stability eigenvalues associated with these modes, in
932: very good agreement with the observed numerical results.
933: The analytical calculations also empower us to identify,
934: even for the stable (close to the AC-limit)
935: modes, the relevant intervals of stability
936: of those waveforms. We have corroborated our analytical
937: calculations with detailed computations that identify the
938: corresponding modes and numerically analyze their linear
939: stability. In addition, for each of the modes, we have shown some
940: typical examples of their dynamical evolution, when they become
941: unstable (either directly, or subsequently due to eigenvalue
942: collisions).
943:
944: These results offer immediate suggestions for experiments
945: in arrays of optical waveguides, Bose-Einstein condensates
946: (e.g. of $^{87}$Rb or $^{23}$Na, which feature repulsive
947: interactions amounting, at the mean-field level,
948: to a defocusing nonlinearity) mounted on a deep optical
949: lattice. In the latter case, the nodes of the lattice
950: considered herein
951: would correspond to BEC droplets in the respective
952: wells of the optical potential. Finally, they are also
953: suggestive of similar experiments in the recently and
954: rapidly growing theme of photorefractive crystal lattices
955: (where, however, the nonlinearity is slightly different,
956: featuring a saturable form).
957:
958: We close by suggesting that these results also indicate
959: that higher charge configurations \cite{dep,zhigang_pre}
960: may similarly be possible and could potentially also
961: be stable in a defocusing setting, similarly to the
962: $S=1$ states discussed above. It would certrainly be
963: of interest to examine such states in the near future,
964: as well as to study the effect of additional components \cite{dep1}
965: (i.e., multi-component states, relevant to the above
966: optical settings when multiple
967: polarizations are present, or to BECs when multiple hyperfine states
968: are studied), or that of higher-dimensional
969: structures \cite{ricardo}.
970:
971: \vspace{5mm}
972: {\bf Acknowledgements.} PGK gratefully acknowledges the support
973: of NSF through the grants DMS-0204585,
974: DMS-CAREER, DMS-0505663 and DMS-0619492. ZC was supported by
975: AFOSR, NSF, PRF and NSFC.
976:
977:
978: \begin{thebibliography}{99}
979:
980: \bibitem{efrem} N.K. Efremidis,
981: S. Sears, D. N. Christodoulides, J. W. Fleischer, and M. Segev
982: Phys. Rev. E {\bf 66}, 46602 (2002).
983:
984: \bibitem{moti1} J.W. Fleischer, M. Segev, N.K. Efremidis and
985: D.N. Christodoulides,
986: Nature {\bf 422}, 147 (2003); J.W. Fleischer, T. Carmon,
987: M. Segev, N.K. Efremidis and D.N. Christodoulides,
988: Phys. Rev. Lett. {\bf 90}, 23902 (2003).
989:
990: \bibitem{neshevol03} D. Neshev, E. Ostrovskaya, Yu.S. Kivshar
991: and W. Krolikowski,
992: Opt. Lett. {\bf 28}, 710 (2003).
993:
994: \bibitem{martinprl04} H. Martin, E.D. Eugenieva,
995: Z. Chen and D.N. Christodoulides,
996: Phys. Rev. Lett. {\bf 92}, 123902 (2004).
997:
998: \bibitem{dip} J. Yang, I. Makasyuk, A. Bezryadina and Z. Chen,
999: Opt. Lett. \textbf{29}, 1662 (2004).
1000:
1001: \bibitem{quad} J. Yang, I. Makasyuk, A. Bezryadina and Z. Chen,
1002: Stud. Appl. Math. \textbf{113}, 389
1003: (2004).
1004:
1005: \bibitem{neck} J. Yang,
1006: I. Makasyuk, P. G. Kevrekidis, H. Martin, B. A. Malomed,
1007: D. J. Frantzeskakis, and Zhigang Chen,
1008: Phys. Rev. Lett. \textbf{94}, 113902
1009: (2005).
1010:
1011: \bibitem{multi} D. Neshev, Yu. S. Kivshar, H. Martin, and
1012: Z. Chen,
1013: %``Soliton stripes in two-dimensional nonlinear photonic lattices,''
1014: Opt. Lett. {\bf 29}, 486-488 (2004).
1015: %Z. Chen,
1016: %H. Martin, E.D. Eugenieva, J. Xu, and A. Bezryadina
1017: %Phys. Rev. Lett. \textbf{92},
1018: %143902 (2004).
1019:
1020: %\bibitem{fedele} F. Fedele, J. Yang and Z. Chen, Opt. Lett.
1021: %{\bf 30}, 1506 (2005).
1022:
1023:
1024: \bibitem{vortex} D. N. Neshev,
1025: T.J. Alexander, E.A. Ostrovskaya, Yu.S. Kivshar,
1026: H. Martin, I. Makasyuk and Z. Chen,
1027: Phys. Rev. Lett. 92, 123903
1028: (2004); J. W. Fleischer,
1029: G. Bartal, O. Cohen, O. Manela, M. Segev,
1030: J. Hudock, and D.N. Christodoulides
1031: Phys. Rev. Lett. \textbf{92},
1032: 123904 (2004).
1033:
1034: \bibitem{rings} X. Wang, Z. Chen, and P. G. Kevrekidis, Phys. Rev. Lett.
1035: \textbf{96}, 083904 (2006).
1036:
1037:
1038: \bibitem{moti3} J.W. Fleischer,
1039: G. Bartal, O. Cohen, T. Schwartz, O. Manela, B. Freedman, M. Segev,
1040: H. Buljan and N.K. Efremidis,
1041: Opt. Express {\bf 13}, 1780
1042: (2005).
1043:
1044: \bibitem{zc4} Z. Chen, H. Martin, A. Bezryadina, D. N. Neshev,
1045: Yu.S. Kivshar, and D. N. Christodoulides,
1046: %``Experiments on Gaussian beams and vortices in optically-induced
1047: %photonic lattices,''
1048: J. Opt. Soc. Am. B {\bf 22}, 1395-1405 (2005).
1049:
1050: \bibitem{7} H.S. Eisenberg, Y. Silberberg,
1051: R. Morandotti, A.R. Boyd and J.S. Aitchison,
1052: Phys. Rev. Lett. {\bf 81}, 3383 (1998).
1053:
1054: \bibitem{7a} R. Morandotti, U. Peschel,
1055: J.S. Aitchison, H.S. Eisenberg and Y. Silberberg,
1056: Phys. Rev. Lett. {\bf 83}, 2726 (1999);
1057: H.S. Eisenberg, Y. Silberberg, R. Morandotti and
1058: J.S. Aitchison,
1059: Phys. Rev. Lett. {\bf 85}, 1863 (2000).
1060:
1061: \bibitem{7b} D. Mandelik, R. Morandotti,
1062: J.S. Aitchison, and Y. Silberberg
1063: Phys. Rev. Lett.
1064: {\bf 92}, 93904 (2004).
1065:
1066: \bibitem{eis3} R. Morandotti, H.S. Eisenberg, and Y. Silberberg,
1067: M. Sorel and J. S. Aitchison,
1068: Phys. Rev. Lett. {\bf 86}, 3296 (1999).
1069:
1070:
1071: \bibitem{review_opt} D. N.\ Christodoulides,
1072: F.\ Lederer, and Y.\ Silberberg,
1073: Nature \textbf{424}, 817 (2003); A. A.\ Sukhorukov,
1074: Y. S.\ Kivshar, H. S.\ Eisenberg, and Y.\ Silberberg,
1075: IEEE J. Quant. Elect. \textbf{39}, 31 (2003).
1076:
1077: \bibitem{general_review} S.\ Aubry, \newblock Physica \textbf{103D}, 201
1078: (1997); S.\ Flach and C. R.\ Willis, \newblock Phys.\ Rep.\ \textbf{295},
1079: 181 (1998); D. K.\ Campbell, S. Flach, and Y. S. Kivshar, Phys. Today,
1080: January 2004, p. 43.
1081:
1082: \bibitem{bloch} S. Burger, F. S. Cataliotti, C. Fort, P. Maddaloni,
1083: F. Minardi and M. Inguscio,
1084: Europhys. Lett. {\bf 57}, 1 (2002).
1085: %M. Greiner {\it et al.}, Nature {\bf 415}, 39 (2002).
1086:
1087: \bibitem{pgk} A. Smerzi, A. Trombettoni, P. G. Kevrekidis, and A. R. Bishop,
1088: Phys. Rev. Lett.
1089: {\bf 89}, 170402 (2002); F.S. Cataliotti,
1090: L. Fallani, F. Ferlaino, C. Fort, P. Maddaloni and M. Inguscio,
1091: New J. Phys. {\bf 5}, 71 (2003).
1092:
1093: \bibitem{markus} B. Eiermann, Th. Anker,
1094: M. Albiez, M. Taglieber, P. Treutlein, K.-P. Marzlin, and M.K. Oberthaler
1095: Phys. Rev. Lett. {\bf 92}, 230401 (2004)
1096:
1097: \bibitem{arimondo} M. Jona-Lasinio,
1098: O. Morsch, M. Cristiani, N. Malossi, J. H. M{\"u}ller,
1099: E. Courtade, M. Anderlini, and E. Arimondo
1100: Phys. Rev. Lett. {\bf 91}, 230406 (2003)
1101:
1102: \bibitem{konotop} V. A. Brazhnyi and V. V. Konotop, Mod. Phys. Lett. B
1103: \textbf{18}, 627 (2004); P. G. Kevrekidis and D. J. Frantzeskakis, Mod.
1104: Phys. Lett. B \textbf{18}, 173 (2004).
1105:
1106: \bibitem{markus2} O. Morsch and M. Oberthaler,
1107: Rev. Mod. Phys. {\bf 78}, 179 (2006).
1108:
1109:
1110: \bibitem{dnls} P.G. Kevrekidis, K.{\"O}. Rasmussen and A.R. Bishop,
1111: Int. J. Mod. Phys. B {\bf 15}, 2833 (2001).
1112:
1113: \bibitem{dep} D.E. Pelinovsky, P.G. Kevrekidis and D.J. Frantzeskakis
1114: Physica D \textbf{212}, 1 (2005); \textit{ibid.} \textbf{212}, 20 (2005).
1115:
1116: \bibitem{zhigang_pre} P.G. Kevrekidis, B.A. Malomed,
1117: Z. Chen and D.J. Frantzeskakis,
1118: Phys. Rev. E {\bf 70}, 056612 (2004).
1119:
1120: \bibitem{dep1} See e.g. P.G. Kevrekidis and D.E. Pelinovsky,
1121: Proc. Roy. Soc. London A {\bf 462}, 2073 (2006)
1122: %http://www.journals.royalsoc.ac.uk/link.asp?id=d55367845kj74308
1123: and references therein.
1124:
1125: \bibitem{ricardo} R. Carretero-Gonz{\'{a}}lez, P. G. Kevrekidis,
1126: B. A. Malomed,and D. J. Frantzeskakis, Phys. Rev.
1127: Lett. \textbf{94}, 203901 (2005); T. J. Alexander, E.A. Ostrovskaya,
1128: A.A. Sukhorukov and Yu.S. Kivshar, Phys.
1129: Rev. A \textbf{72}, 043603 (2005).
1130:
1131:
1132:
1133: \end{thebibliography}
1134:
1135: \end{document}
1136:
1137:
1138: