1: %%\hoffset -1.4truecm \voffset -2truecm \textheight 24truecm
2: %%\textwidth 17truecm \pagenumbering {arabic}
3:
4:
5: \documentclass[showpacs,pre,onecolumn]{revtex4}
6: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
7: \usepackage{amsmath}
8: \usepackage{caption2}
9: \usepackage{graphicx}
10: \usepackage{subfigure}
11:
12: \setcounter{MaxMatrixCols}{10}
13: %TCIDATA{OutputFilter=Latex.dll}
14: %TCIDATA{Version=5.50.0.2890}
15: %TCIDATA{<META NAME="SaveForMode" CONTENT="1">}
16: %TCIDATA{BibliographyScheme=Manual}
17: %TCIDATA{LastRevised=Saturday, February 25, 2006 15:24:15}
18: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
19:
20: \renewcommand{\baselinestretch}{1.5}
21: \input{tcilatex}
22: \begin{document}
23:
24: \title{Two-dimensional solitons in saturable media with a
25: quasi-one-dimensional lattice potential}
26: \author{Thawatchai Mayteevarunyoo}
27: \affiliation{Department of Telecommunication Engineering, Mahanakorn University of
28: Technology, Bangkok 10530, Thailand}
29: \author{Boris A. Malomed}
30: \affiliation{Department of Interdisciplinary Studies, School of Electrical Engineering,
31: Tel Aviv University, Tel Aviv 69978, Israel}
32:
33: \begin{abstract}
34: We study families of solitons in a two-dimensional (2D) model of the light
35: transmission through a photorefractive medium equipped with a
36: (quasi-)one-dimensional photonic lattice. The soliton families are bounded
37: from below by finite minimum values of the peak and total power. Narrow
38: solitons have a single maximum, while broader ones feature side lobes.
39: Stability of the solitons is checked by direct simulations. The solitons can
40: be set in motion across the lattice (actually, made tilted in the spatial
41: domain), provided that the respective boost parameter does not exceed a
42: critical value. Collisions between moving solitons are studied too.
43: Collisions destroy the solitons, unless their velocities are sufficiently
44: small. In the latter case, the colliding solitons merge into a single stable
45: pulse.
46: \end{abstract}
47:
48: \maketitle
49:
50: \section{Introduction}
51:
52: As was first predicted in Refs. \cite{DemetriMoti}, a periodic lattice
53: potential induced in a photorefractive medium, which is characterized by the
54: saturable nonlinearity, may be an efficient tool for creation and
55: stabilization of two-dimensional (2D) spatial solitons of various types. The
56: prediction was followed by the experimental observation of ordinary solitons
57: \cite{Nature}, localized vortices with the topological charge $1$ belonging
58: to the first (lowest) \cite{vortex} or second \cite{second-band} bandgap in
59: the respective linear spectrum (stable higher-order vortices and \textit{%
60: supervortices} in such systems were recently predicted too \cite{Hidetsugu}%
61: ), dipole- and quadrupole-mode solitons \cite{Jianke}, steady patterns in
62: the form of soliton necklaces \cite{necklace}, and some others. A review of
63: the field was recently given in Ref. \cite{ExpressReview}.
64:
65: In the experiment, the periodic potential is induced by the photonic
66: lattice, which is created as a superposition of counterpropagating laser
67: beams illuminating the photorefractive crystal in the ordinary polarization,
68: in which the light propagation is nearly linear; then, the spatial solitons
69: are built in a probe beam launched through the lattice in the extraordinary
70: polarization (i.e., polarized along the crystalline $c$ axis), which is
71: subject to strong nonlinearity induced by the dc bias electric field applied
72: to the crystal \cite{DemetriMoti,ExpressReview}. Using this technique, the
73: square photonic lattice can be created by two pairs of counterpropagating
74: beams illuminating the crystal in directions orthogonal to each other and to
75: the probe beam that gives rise to the soliton(s).
76:
77: On the other hand, 2D solitons can also be supported by a low-dimensional,
78: i.e., quasi-one-dimensional (Q1D) periodic potential, that may be readily
79: induced by a single pair of counterpropagating beams illuminating the bulk
80: crystal in the ordinary polarization. Recently, it has been predicted that
81: the Q1D lattice may efficiently stabilize 2D solitons in the model with the
82: cubic (rather than saturable) nonlinearity \cite{BBB}. This result directly
83: applies to Bose-Einstein condensates with attractive interactions between
84: atoms, trapped in a photonic lattice; moreover, it has also been
85: demonstrated \cite{BBB,Barcelona} that a quasi-2D lattice can stabilize 3D
86: solitons against strong collapse that the cubic nonlinearity gives rise to
87: in the latter case. 2D solitons supported by the Q1D lattice naturally
88: demonstrate strong anisotropy, which makes them essentially different from
89: the usual 2D solitons. A significant advantage offered by the use of the Q1D
90: lattice is the fact that the remaining free direction allows the solitons to
91: move (in the spatial domain, ``motion" means a tilt of the soliton beam),
92: which opens a way to study collisions between them, formation of bound
93: states, etc. \cite{BBB}. The mobility of 2D solitons may also be strongly
94: anisotropic in some 2D lattices \cite{preprint}.
95:
96: The objective of the present work is to introduce 2D solitons in the model
97: of the photorefractive media with the saturable nonlinearity and Q1D
98: periodic potential (Q1D solitons -- which are, effectively, one-dimensional
99: objects -- in the model of the 2D photorefractive medium with the Q1D
100: lattice were introduced in Ref. \cite{Lena}). The soliton solutions are
101: constructed in Section 2. We demonstrate that the Q1D solitons are
102: characterized by minimum peak and total intensities necessary for their
103: existence, and they are stable in the entire existence region. Moving
104: solitons and collisions between them are studied in Section 3. We find that
105: collisions destroy the solitons, unless the collision ``velocity" (in fact,
106: the relative tilt of the two spatial solitons) is small enough; in the
107: latter case, the colliding solitons merge into a single one, irrespective of
108: the orientation of the velocity vector relative to the Q1D lattice.
109:
110: The model outlined above is based on an equation for the spatial evolution
111: of the probe field (the slowly varying amplitude $U$ of the electromagnetic
112: wave in the extraordinary polarization), which follows the standard
113: description of photorefractive media \cite{DemetriMoti}. In normalized
114: units, the equation takes the form:
115:
116: \begin{equation}
117: i\frac{\partial U}{\partial z}+\frac{\partial ^{2}U}{\partial x^{2}}+\frac{%
118: \partial ^{2}U}{\partial y^{2}}-\frac{U}{1+I_{0}\cos ^{2}\left( \pi
119: x/d\right) +\left\vert U\right\vert ^{2}}=0, \label{PhR}
120: \end{equation}%
121: where $I_{0}$ and $d$ are the peak intensity and period of the photonic
122: lattice induced by the superposition of counterpropagating waves launched
123: along the $x$ axis in the ordinary polarization. The normalized propagation
124: and transverse coordinates, $z$ and $\left( x,y\right) $, are proportional
125: to their counterparts measured in physical units, $Z$ and $\left( X,Y\right)
126: $, so that $z=Z/\left( k_{0}\Delta n_{0}\right) $ and $\left( x,y\right)
127: =\left( X,Y\right) /\sqrt{2k_{0}n_{0}\Delta n_{0}}$, where $k_{0}$ is the
128: propagation constant of the probe wave, and $\Delta n_{0}=n_{0}^{3}r_{33}E/2$
129: is the change of the refractive index $n_{0}$ (accounted for by the
130: electro-optic coefficient $r_{33}$) caused by the dc electric field $E$.
131:
132: \section{Two-dimensional solitons and their stability}
133:
134: Soliton solutions to Eq. (\ref{PhR}) are searched for as
135: \begin{equation}
136: U=u\left( x,y\right) e^{-i\mu z}, \label{mu}
137: \end{equation}%
138: where $-\mu $ is the shift of the propagation constant in the soliton, and
139: the real function $u$ satisfies the equation%
140: \begin{equation}
141: \frac{\partial ^{2}u}{\partial x^{2}}+\frac{\partial ^{2}u}{\partial y^{2}}+%
142: \left[ \mu -\frac{1}{1+I_{0}\cos ^{2}\left( \pi x/d\right) +u^{2}}\right]
143: u=0. \label{NE}
144: \end{equation}%
145: Solutions $u(x,y)$ of this equation were found by means of an iteration
146: procedure in the Fourier space. To this end, following a numerical method
147: elaborated in Ref. \cite{Jianke}, Eq. (\ref{NE}) was rewritten for the
148: Fourier transform $\widehat{u}(k_{x},k_{y})$ in the form%
149: \begin{equation}
150: \widehat{u}=\frac{1}{k_{x}^{2}+k_{y}^{2}-\mu +1}\left\{ \mathcal{F}\left(
151: \frac{I_{0}\cos ^{2}\left( \pi x/d\right) \cdot u}{1+I_{0}\cos ^{2}\left(
152: \pi x/d\right) +\left\vert u\right\vert ^{2}}\right) +\mathcal{F}\left(
153: \frac{\left\vert u\right\vert ^{2}u}{1+I_{0}\cos ^{2}\left( \pi x/d\right)
154: +\left\vert u\right\vert ^{2}}\right) \right\} , \label{Fu}
155: \end{equation}%
156: $\mathcal{F}\left( ...\right) $ standing for the 2D Fourier transform. A
157: direct iteration procedure applied to Eq. (\ref{Fu}) does not converge, in
158: the general case. Therefore, the equation was modified as follows: defining
159: integral factors,%
160: \begin{eqnarray}
161: \alpha &=&\int \int \left\{ \left( k_{x}^{2}+k_{y}^{2}-\mu +1\right)
162: \widehat{u}-\mathcal{F}\left( \frac{I_{l}}{1+I_{0}\cos ^{2}\left( \pi
163: x/d\right) +\left\vert u\right\vert ^{2}}u\right) \right\} \widehat{u}^{\ast
164: }dk_{x}dk_{y} \label{alpha} \\
165: \beta &=&\int \int \mathcal{F}\left( \frac{\left\vert u\right\vert ^{2}u}{%
166: 1+I_{0}\cos ^{2}\left( \pi x/d\right) +\left\vert u\right\vert ^{2}}\right)
167: \widehat{u}^{\ast }dk_{x}dk_{y}, \label{beta}
168: \end{eqnarray}%
169: where $\ast $\ stands for the complex conjugation, the following iterative
170: equation was introduced,%
171: \begin{eqnarray}
172: \widehat{u}_{n+1} &=&\frac{1}{k_{x}^{2}+k_{y}^{2}-\mu +1}\left\{ \left(
173: \frac{\alpha _{n}}{\beta _{n}}\right) ^{1/2}\mathcal{F}\left( \frac{I_{l}}{%
174: 1+I_{0}\cos ^{2}\left( \pi x/d\right) +\left\vert u_{n}\right\vert ^{2}}%
175: u_{n}\right) \right. \notag \\
176: &&\left. +\left( \frac{\alpha _{n}}{\beta _{n}}\right) ^{3/2}\mathcal{F}%
177: \left( \frac{\left\vert u_{n}\right\vert ^{2}u_{n}}{1+I_{0}\cos ^{2}\left(
178: \pi x/d\right) +\left\vert u_{n}\right\vert ^{2}}\right) \right\} ,
179: \label{iter}
180: \end{eqnarray}%
181: where $\alpha _{n}$ and $\beta _{n}$ are the factors (\ref{alpha}) and (\ref%
182: {beta}) corresponding to the function $\widehat{u}_{n}$. Fixed points of Eq.
183: (\ref{iter}), corresponding to $\underset{n\longrightarrow \infty }{\lim }%
184: \left( \alpha _{n}/\beta _{n}\right) =1$, yield solutions to Eq. (\ref{Fu})
185: as well. The iterative procedure based on Eq. (\ref{iter}) provides for fast
186: convergence, and produces solutions displayed below.
187:
188: Typical examples of the 2D solitons are shown in Fig. \ref{fig1}. In
189: particular, the picture observed in the left column of the figure is typical
190: to cases when the lattice potential is weak, and/or the soliton's peak
191: intensity essentially exceeds the lattice's amplitude $I_{0}$: the soliton
192: is practically isotropic, without a conspicuous effect of the lattice. The
193: picture in the right column is typical for a relatively strong lattice with
194: a large period: then, the soliton is almost entirely trapped in one
195: potential trough, assuming an elliptic shape. In either case, the soliton's
196: shape features a single lobe, either circular or elliptic one.
197:
198: The most interesting typical case is represented by the central column in
199: Fig. \ref{fig1}, which demonstrates a soliton with well-pronounced side
200: lobes in a moderately strong lattice (in physical units, this case
201: corresponds to typical values of parameters available in the experiment).
202: Naturally, the multi-humped structure is observed only along the axis $x$,
203: while in the free direction, $y$, the soliton always features a simple
204: single-hump form. Results reported below are given for the same period, $%
205: d=2\pi $, which gives rise to the central column in Fig. \ref{fig1}. For
206: other values of $d$ which are neither very small nor very large, the results
207: are quite similar.
208:
209: The solitons are characterized by the peak intensity (power), $I_{p}\equiv
210: \left\vert u(x=0,y=0)\right\vert ^{2}$, and integral intensity,%
211: \begin{equation}
212: P\equiv \int_{-\infty }^{+\infty }\int_{-\infty }^{+\infty }\left\vert
213: u\left( x,y\right) \right\vert ^{2}dxdy. \label{P}
214: \end{equation}%
215: Accordingly, soliton families for given values of $I_{0}$ and $d$ are
216: represented by the dependences $P(\mu )$ and $I_{p}(\mu )$, see Fig. \ref%
217: {fig2} [recall $-\mu $ is the soliton propagation-constant shift, defined in
218: Eq. (\ref{mu})]. An important feature observed in this figure is that the 2D
219: solitons exist, with a given lattice strength $I_{0}$, only for $P$ and $%
220: I_{p}$ exceeding certain finite minimum (threshold) values, $P_{\min }$ and $%
221: I_{p,\min }$; for instance, $I_{p,\min }=0.0411$ for $I_{0}=5$ and $d=2\pi $%
222: . Both minimum values are shown, as functions of the strength $I_{0}$ of the
223: photonic lattice, in Fig. \ref{fig3}. It should be noticed that a lower
224: intensity threshold necessary for the existence of 2D solitons in models
225: combining lattice potentials with various nonlinearities was found in
226: earlier works \cite{DemetriMoti,BBBfirst,BBB}.
227:
228: It is easy to check that the soliton families shown in Figs. \ref{fig1} and %
229: \ref{fig2} belong to the semi-infinite bandgap in the spectrum of the
230: linearized equation (\ref{PhR}). It is known that, on top of the
231: semi-infinite gap, a sufficiently strong lattice may give rise to solitons
232: in finite bandgaps of models combining a periodic potential and saturable
233: nonlinearity (see, e.g., Ref. \cite{Merhasin}). In this work, we do not aim
234: to look for gap solitons of such types.
235:
236: Proceeding to the investigation of stability of the 2D solitons, we first of
237: all note that the negative slope of the dependence $P(\mu )$, obvious in
238: Fig. \ref{fig2}, suggests possible stability of the soliton families
239: pursuant to the Vakhitov-Kolokolov (VK)\ criterion \cite{VK}, which may
240: guarantee the absence of unstable modes of small perturbations corresponding
241: to real eigenvalues of the instability growth rate. The full stability is
242: not provided by the VK criterion, and, moreover, the applicability of the
243: criterion even to perturbations with real eigenvalues was not proven in the
244: present context. In fact, a counter-example is known: soliton subfamilies
245: which should be VK-unstable in a 1D model combining a periodic potential of
246: the Kronig-Penney type and cubic-quintic nonlinearity (another variety of
247: the saturation) were found to be \emph{completely stable} \cite{Merhasin}.
248:
249: We have tested stability of the 2D solitons in direct simulations of Eq. (%
250: \ref{PhR}), making use of the fast Fourier transform in directions $x$ and $%
251: y $ and Runge-Kutta method to advance in $z$. In each simulation, some
252: amount of random noise was added to the soliton as a perturbation. It has
253: been concluded that \emph{all the solitons} are stable, as (formally)
254: predicted by the VK criterion. An example illustrating the stability of a
255: large-amplitude 2D soliton with $I_{0}=5$ and $I_{p}=56.76$ is displayed in
256: Fig. \ref{fig4}(a). Broad solitons, with the peak intensity close to the
257: threshold value $I_{p,\min }$, are stable too, although their peak intensity
258: may slowly grow with $z$, as shown in Fig. \ref{fig4}(b) for the same
259: lattice strength, $I_{0}=5$, as in Fig. \ref{fig4}(a), but $I_{p}=1.1$. A
260: plausible explanation to the latter effect is that the combination of the
261: lattice potential and self-focusing nonlinearity (for relatively small $%
262: I_{p} $, the saturation does not suppress the self-focusing) leads to
263: sucking the perturbation wave field into the spot where the soliton's
264: maximum is located, and the build-up of the additional wave field around the
265: soliton's peak may be conspicuous, against the backdrop of the relatively
266: small value of $I_{p}$.
267:
268: \subsection{Moving two-dimensional solitons and their collisions}
269:
270: Solutions $\tilde{U}(x,y,z)$ for solitons ``moving" (actually, tilted) along
271: the free direction $y$ can be generated from the ``quiescent" solitons
272: reported in the previous section, $U(x,y,z)$, by means of the Galilean
273: transformation,%
274: \begin{equation}
275: \tilde{U}(x,y,z)=U(x,y-2Q_{y}z,z)\exp \left( iQ_{y}x-iQ_{y}^{2}z\right) ,
276: \label{Galileo}
277: \end{equation}%
278: where $Q_{y}$ is an arbitrary \textit{boost parameter}, that determines the
279: soliton's ``velocity" (tilt) $2Q_{y}$. On the other hand, generation of
280: solitons tilted along the $x$ axis is a nontrivial problem. In the 1D
281: version of Eq. (\ref{PhR}), without the term $\partial ^{2}U/\partial y^{2}$%
282: , a family of tilted solitons was introduced in Ref. \cite{Lena}.
283:
284: We looked for solitons ``moving" along the $x$ axis by simulating the
285: evolution of initial states of the form%
286: \begin{equation}
287: U_{0}(x,y)=U(x,y)\exp \left( iQ_{x}x\right) , \label{U0}
288: \end{equation}%
289: cf. Eq. (\ref{Galileo}), where $U(x,y)$ corresponds to a zero-velocity
290: stationary soliton solution. The boost parameter in Eq. (\ref{U0}), $Q_{x}$,
291: was gradually increased from a run to a run, until reaching a critical
292: (maximum) value $Q_{\max }$, at which the initial configuration (\ref{U0})
293: does not generate any soliton, but rather gets destroyed into
294: small-amplitude waves. Results of the simulations are summarized in Fig. \ref%
295: {fig5}, in the form of plots showing $Q_{\max }$ as a function of the
296: soliton's peak intensity $I_{p}$, for different values of the lattice
297: strength $I_{0}$. Measurement of the established value of the average
298: ``velocity" (tilt) of the solitons generated by the initial condition (\ref%
299: {U0}) with $Q_{x}<Q_{\max }$ produces values which are quite close to ones ($%
300: 2Q_{x}$) corresponding to the Galilean transform (\ref{Galileo}) in the free
301: space. Finally, applying the Galilean transformation (\ref{Galileo}) to the
302: soliton already moving along the $x$ direction, one can generate a pulse
303: moving in any direction (in particular, along the diagonal, $x=y$, see
304: below).
305:
306: Once the stability limits for the moving solitons are available, the next
307: step is to consider collisions between them. The outcome of the collision
308: may depend on the magnitude and direction of the velocities, and the aiming
309: mismatch (its zero value corresponds to the head-on collision).
310:
311: Numerous simulations demonstrate that the collision completely destroys both
312: solitons, unless their velocities are sufficiently small. An example of the
313: destructive collision is displayed in Fig. \ref{fig6}. In particular, in the
314: case of $I_{0}=5$, the solitons must not be set in motion by the boost with $%
315: \left\vert Q_{x}\right\vert $ or $\left\vert Q_{y}\right\vert $ in excess of
316: $0.05,$ otherwise the collision will destroy them. This critical value is
317: the same for the motion in the $x$ and $y$ directions, up to the accuracy of
318: the simulations. Moreover, if the solitons are boosted in the diagonal
319: direction and then collide, the same limit, $\left\vert Q\right\vert =0.05$,
320: was found for the absolute value of the corresponding vectorial boost, $%
321: \left\vert Q\right\vert =\sqrt{Q_{x}^{2}+Q_{y}^{2}}$, i.e., the critical
322: boost is virtually isotropic.
323:
324: Collisions between solitons boosted by $|Q|$ with values smaller than the
325: critical one result in their \emph{merger} into a single pulse. An example
326: of the merger resulting from the collision with a finite aiming mismatch is
327: displayed in Fig. \ref{fig7}. Further, the mergers caused by head-on
328: collisions in the diagonal or vertical ($y$) direction are presented in Fig. %
329: \ref{fig8}. As seen from the figures, the pulse generated by the merger
330: performs intrinsic vibrations, but remains stable.
331:
332: The above examples displayed collisions of single-lobe solitons. Collisions
333: between their broader counterparts, which feature side lobes in the $x$%
334: -direction, are quite similar. Figure \ref{fig9} presents an example of the
335: head-on collision between sufficiently slowly moving solitons of the latter
336: type. The collision again leads to the merger of the pulses into a single
337: one, which then performs conspicuous intrinsic vibrations, but remains a
338: stable object. Additional simulations show that the soliton produced by the
339: merger of two multi-lobe ones can also easily move across the lattice, if
340: given a push.
341:
342: Thus we conclude that collisions between the 2D solitons are always \emph{%
343: strongly inelastic} (both the destruction and merger of the colliding
344: solitons are inelastic outcomes), attesting to the fact that the present
345: model is far from any integrable limit, where collisions between solitons
346: would be elastic. For comparison, we note that the collisions may be less
347: inelastic in the 2D model with the Q1D periodic potential if the
348: nonlinearity is cubic \cite{BBB}. In that model, elastic collisions are
349: possible (passage of the solitons in the case of a finite aiming mismatch,
350: or their mutual bounce after the head-on collision, if the phase shift
351: between them is $\pi $). An additional inelastic outcome of the head-on
352: collisions between in-phase 2D solitons, which was observed in the model
353: with the cubic nonlinearity, but cannot occur if the nonlinearity is
354: saturable, is collapse of the single pulse formed after the merger of two
355: solitons (i.e., formation of a singularity after a finite propagation
356: distance).
357:
358: \section{Conclusion}
359:
360: In this paper, we have proposed a model of the two-dimensional (2D) medium
361: with the saturable nonlinearity and quasi-1D lattice potential, that can be
362: realized in the spatial domain in a bulk photorefractive crystal. The
363: subject of the analysis were 2D solitons (ones belonging to the
364: semi-infinite bandgap in the linear spectrum). It was demonstrated that they
365: form families bounded from below by finite minimum values of the peak and
366: total intensities. Narrow pulses feature a single maximum, while broad
367: solitons contain side lobes. Direct simulations confirm that the solitons
368: are stable. They can be set in motion (actually, tilted in the spatial
369: domain) in an obvious way along the quasi-1D lattice, and also across the
370: lattice, provided that, in the latter case, the boost parameter does not
371: exceed a critical value, beyond which the soliton is destroyed.
372:
373: Collisions between stable moving solitons were studied in detail, with the
374: conclusion that the collisions destroy the solitons, unless their velocities
375: are sufficiently small. In the latter case, the colliding solitons merge
376: into a single stable pulse, that performs intrinsic vibrations. The
377: predictions reported in this paper can be readily implemented experimentally
378: in photorefractive crystals.
379:
380: \begin{thebibliography}{99}
381: \bibitem{DemetriMoti} N. K. Efremidis, S. Sears, D. N. Christodoulides, J.
382: W. Fleischer, and M. Segev, Phys. Rev. E \textbf{66}, 046602 (2002); N. K.
383: Efremidis, J. Hudock, D. N. Christodoulides, J. W. Fleischer, O. Cohen, and
384: M. Segev, Phys. Rev. Lett. \textbf{91}, 213906 (2003).
385:
386: \bibitem{Nature} J.~W. Fleischer, M. Segev, N.~K. Efremidis, and D.~N.
387: Christodoulides, Nature \textbf{422}, 147 (2003); J. W. Fleischer, G.
388: Bartal, O. Cohen, O. Manela, M. Segev, J. Hudock, and D. N. Christodoulides,
389: Phys. Rev. Lett. \textbf{92}, 123904 (2004).
390:
391: \bibitem{vortex} D N. Neshev, J. Alexander, E. A. Ostrovskaya, Y. S.
392: Kivshar, H. Martin, I. Makasyuk, and Z. Chen, Phys. Rev. Lett. \textbf{92},
393: 123903 (2004); J. W. Fleischer, G. Bartal, O. Cohen, O. Manela, M. Segev, J.
394: Hudock, and D. N. Christodoulides, Phys. Rev. Lett. \textbf{92}, 123904
395: (2004).
396:
397: \bibitem{second-band} O. Manela, O. Cohen, G. Bartal, J. W. Fleischer, and
398: M. Segev, Opt. Lett. \textbf{29}, 2049 (2004).
399:
400: \bibitem{Hidetsugu} H. Sakaguchi and B. A. Malomed, \textit{Higher-order
401: vortex solitons, multipoles, and supervortices on a square optical lattice},
402: Europhys. Lett., in press.
403:
404: \bibitem{Jianke} J. Yang, I. Makasyuk, A. Bezryadina, and Z. Chen, Stud.
405: Appl. Math. \textbf{113}, 389 (2004).
406:
407: \bibitem{necklace} J. Yang, I. Makasyuk, P. G. Kevrekidis, H. Martin, B. A.
408: Malomed, D. J. Frantzeskakis, and Z. Chen, Phys. Rev. Lett. \textbf{94},
409: 113902 (2005).
410:
411: \bibitem{ExpressReview} J. W. Fleischer, G. Bartal, O. Cohen, T. Schwartz,
412: O. Manela, B. Freedman, M. Segev, H. Buljan, N. K. Efremidis, Opt. Exp.
413: \textbf{13,} 1780 (2005).
414:
415: \bibitem{BBB} B. B. Baizakov, B. A. Malomed and M. Salerno, Phys. Rev. A
416: \textbf{70}, 053613 (2004); in: \textit{Nonlinear Waves: Classical and
417: Quantum Aspects}, ed. by F. Kh. Abdullaev and V. V. Konotop, p. 61 (Kluwer
418: Academic Publishers: Dordrecht, 2004; also available at
419: http://rsphy2.anu.edu.au/\symbol{126}asd124/Baizakov\_2004\_61%
420: \_NonlinearWaves.pdf)
421:
422: \bibitem{Barcelona} D. Mihalache, D. Mazilu, F. Lederer, Y. V. Kartashov,
423: L.-C. Crasovan, and L. Torner, Phys. Rev. E 70, 055603(R) (2004).\newline
424:
425: \bibitem{preprint} R. Fischer, D. Traeger, D. N. Neshev, A. A. Sukhorukov,
426: W. Krolikowski, C. Denz, and Y. S. Kivshar, e-print physics/0509255 (2005).
427:
428: \bibitem{Lena} B. A. Malomed, T. Mayteevarunyoo, E. A. Ostrovskaya, and Y.
429: S. Kivshar, Phys. Rev. E \textbf{71}, 056616 (2005).
430:
431: \bibitem{BBBfirst} B. B. Baizakov, B. A. Malomed, and M. Salerno, Europhys.
432: Lett. \textbf{63}, 642 (2003).
433:
434: \bibitem{Merhasin} I. M. Merhasin, B. V. Gisin, R. Driben, and B. A.
435: Malomed, Phys. Rev. E \textbf{71}, 016613 (2005).
436:
437: \bibitem{VK} M. G. Vakhitov and A. A. Kolokolov, Izv. Vuz. Radiofiz. \textbf{%
438: 16,} 1020 (1973) [in Russian; English translation: Sov. J. Radiophys.
439: Quantum Electr. \textbf{16}, 783 (1973)]; see also L. Berg{\'{e}}, Phys.
440: Rep. \textbf{303}, 260 (1998).
441:
442: \begin{figure}[tbp]
443: \renewcommand{\captionfont}{\small \sffamily} \renewcommand{%
444: \captionlabelfont}{} \centering\includegraphics[width=4in]{Fig1.eps} %
445: \renewcommand{\figurename}{Fig.}
446: \caption{Typical examples of two-dimensional solitons found from Eq. (%
447: \protect\ref{NE}) with a moderately strong quasi-one-dimensional photonic
448: lattice, corresponding to $I_{0}=5$. Panels (a), (b), and (c) show the
449: intensity distribution in the lattice, $I_{\mathrm{phl}}(x)=I_{0}\cos
450: ^{2}\left( \protect\pi x/d\right) $, for $d=\protect\pi /5$, $2\protect\pi $%
451: , and $10\protect\pi $, respectively. The two-dimensional intensity field in
452: the respective solitons is shown in panels (d), (e), and (f), and the
453: corresponding intensity profiles in two cross sections of the solitons,
454: along $y=0$ and $x=0$, are displayed in panels (g), (h), and (i).}
455: \label{fig1}
456: \end{figure}
457: \end{thebibliography}
458:
459: \begin{figure}[tbp]
460: \renewcommand{\captionfont}{\small \sffamily} \renewcommand{%
461: \captionlabelfont}{} \centering%
462: \subfigure[]{\label{fig:subfig:a}
463: \includegraphics[width=4in]{Fig2a.eps}}
464: \subfigure[]{\label{fig:subfig:b}
465: \includegraphics[width=4in]{Fig2b.eps}} \renewcommand{\figurename}{Fig.}
466: \caption{The total intensity (power) (a) and peak intensity (b) of families
467: of two-dimensional solitons, vs. the absolute value of the
468: propagation-constant shift, $\protect\mu $, for $d=2\protect\pi $, and
469: different values of the photonic-lattice strength $I_{0}$. The curves
470: terminate at points where the solitons cease to exist.}
471: \label{fig2}
472: \end{figure}
473:
474: \begin{figure}[tbp]
475: \renewcommand{\captionfont}{\small \sffamily} \renewcommand{%
476: \captionlabelfont}{} \centering\includegraphics[width=4in]{Fig3.eps} %
477: \renewcommand{\figurename}{Fig.}
478: \caption{{}The minimum peak intensity, $I_{p,\min }$, and total power, $%
479: P_{\min }$, necessary for the existence of the two-dimensional solitons, vs.
480: the strength, $I_{0}$, of the underlying quasi-one-dimensional lattice.}
481: \label{fig3}
482: \end{figure}
483:
484: \begin{figure}[t]
485: \renewcommand{\captionfont}{\small } \renewcommand{\captionlabelfont}{} %
486: \centering%
487: \subfigure[]{\label{fig:subfig:b}
488: \includegraphics[width=4in]{Fig4a.eps}}
489: \subfigure[]{\label{fig:subfig:b}
490: \includegraphics[width=4in]{Fig4b.eps}} \renewcommand{\figurename}{Fig.}
491: \caption{{}(a) Stable evolution of a high-intensity soliton, with $I_{0}=5$
492: and $I_{p}=56.76$, under random perturbations, is illustrated by the $z$%
493: -dependence of the soliton's peak intensity. (b) The same for a
494: low-intensity soliton, with $I_{p}=1.1$.}
495: \label{fig4}
496: \end{figure}
497:
498: \begin{figure}[tbp]
499: \renewcommand{\captionfont}{\small \sffamily} \renewcommand{%
500: \captionlabelfont}{} \centering\includegraphics[width=4in]{Fig5.eps} %
501: \renewcommand{\figurename}{Fig.}
502: \caption{The maximum value of the boost parameter, $Q_{x}$, which generates
503: solitons moving (actually, tilted in the spatial domain) along the $x$ axis,
504: vs. the soliton's peak intensity, $I_{p}$, for different value of the
505: strength $I_{0}$ of the underlying lattice. For $Q_{x}>Q_{\max }$, the
506: application of the boost destroys the soliton.}
507: \label{fig5}
508: \end{figure}
509: \
510:
511: \begin{figure}[tbp]
512: \renewcommand{\captionfont}{\small \sffamily} \renewcommand{%
513: \captionlabelfont}{} \centering\includegraphics[width=4in]{Fig6.eps} %
514: \renewcommand{\figurename}{Fig.}
515: \caption{A typical example of destructive collisions between moving
516: solitons, for $I_{0}=5$ and $I_{p}=1.78$. The velocity vectors of the
517: colliding solitons (shown by arrows in the first panel, in this figure and
518: below) have components $Q_{x,y}=\pm 0.12$.}
519: \label{fig6}
520: \end{figure}
521:
522: \begin{figure}[tbp]
523: \renewcommand{\captionfont}{\small \sffamily} \renewcommand{%
524: \captionlabelfont}{} \centering
525: %%\subfigure[$Q_{y}=0.01$]{\label{fig:subfig:a}
526: \includegraphics[width=4in]{Fig7.eps} %%\figurename}{Fig.}
527: \caption{Merger of colliding solitons which were set in motion by the boost $%
528: Q_{x}=\pm 0.01$, with a finite aiming mismatch in the $y$ direction. In this
529: case, $I_{0}=25$ and $I_{p}=83.97.$}
530: \label{fig7}
531: \end{figure}
532:
533: \begin{figure}[tbp]
534: \renewcommand{\captionfont}{\small \sffamily} \renewcommand{%
535: \captionlabelfont}{} \centering%
536: \subfigure[]{\label{fig:subfig:a}
537: \includegraphics[width=4in]{Fig8a.eps}} %%\renewcommand{\subfigure
538: \subfigure[]{\label{fig:subfig:b}
539: \includegraphics[width=4in]{Fig8b.eps}} %%\renewcommand{\figurename}{Fig.}
540: \caption{{}Merger as a result of head-on collisions between slowly moving
541: solitons that were set in motion (a) along the diagonal direction, by the
542: application of the boost with $Q_{x}=Q_{y}=\pm 0.01$, and (b) in the
543: vertical direction, by the boost $Q_{y}=\pm 0.01$. In case (a), $I_{0}=25$
544: and $I_{p}=83.97$; in case (b), $I_{0}=15$ and $I_{p}=27.92$.}
545: \label{fig8}
546: \end{figure}
547:
548: \begin{figure}[tbp]
549: \renewcommand{\captionfont}{\small \sffamily} \renewcommand{%
550: \captionlabelfont}{} \centering\includegraphics[width=4in]{Fig9.eps}
551: %%\renewcommand{\figurename}{Fig.}
552: \caption{{}Merger of broad (multi-lobed) solitons caused by the head-on
553: collision. The solitons were set in motion by the vertical boost with $%
554: Q_{y}=\pm 0.01$. In this case, $I_{0}=5$ and $I_{p}=1.786$.}
555: \label{fig9}
556: \end{figure}
557:
558: \end{document}
559: