1:
2: \documentstyle[prl,aps,epsfig]{revtex} % 2 col mode
3: \begin{document}
4: \draft % preprint mode
5: % 2 col mode:
6: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname @twocolumnfalse\endcsname
7:
8:
9: \title{Spatial Resonator Solitons}
10:
11: \author{V. B. Taranenko, G. Slekys, C.O. Weiss}
12:
13: \address{Physikalisch-Technische Bundesanstalt \\
14: 38116 Braunschweig, Germany}
15:
16: %\date{\today{}}
17: \maketitle
18:
19: \begin{abstract}
20: Spatial solitons can exist in various kinds of nonlinear optical
21: resonators with and without amplification. In the past years
22: different types of these localized structures such as vortices,
23: bright, dark solitons and phase solitons have been experimentally
24: shown to exist. Many links appear to exist to fields different
25: from optics, such as fluids, phase transitions or particle
26: physics. These spatial resonator solitons are bistable and due to
27: their mobility suggest schemes of information processing not
28: possible with the fixed bistable elements forming the basic
29: ingredient of traditional electronic processing. The recent
30: demonstration of existence and manipulation of spatial solitons
31: in semiconductor microresonators represents a step in the
32: direction of such optical parallel processing applications. We
33: review pattern formation and solitons in a general context, show
34: some proof of principle soliton experiments on slow systems, and
35: describe in more detail the experiments on semiconductor
36: resonator solitons which are aimed at applications.
37: \end{abstract}
38: \pacs{PACS 42.65.Tg, 47.54.+r, 42.65.Sf} \vskip1pc ]
39: % BEGIN TEXT HERE
40: \section{Introduction: a multidisciplinary view at pattern formation and solitons}
41: %
42: In nonequilibrium nonlinear systems of all kinds patterns or
43: "structures" are known to form. Their properties seem to be
44: determined by a maximum dissipation principle. A system without
45: structure is only able to dissipate as much energy as given by
46: the microscopic dissipation processes. If then the system with
47: structure (compatible with the system equations, parameters,
48: boundary conditions and possibly initial conditions) has a higher
49: dissipation, the structure will appear spontaneously e.g. through
50: a modulational instability. This has been discussed at length in
51: relation with the Bénard convection, but is easily found to be
52: true in other fields.
53:
54: In optics, when pumping a medium enough, then the microscopic
55: relaxation processes do not provide the highest possible
56: dissipation. As is well-known a coherent and structured and well
57: ordered optical laser field can build which gives the system a
58: much higher dissipation. If the pumping is increased, the laser
59: emission may convert to a pulsed instead of a continuous emission
60: in the form of regular or chaotic pulsing as described by the
61: Lorenz Model. This pulsed laser emission is again increasing the
62: system's capability to dissipate over the continuous emission.
63: Such pulsing means a modulational instability in the temporal
64: domain. Equally modulations or patterns appear in space, such as
65: a hexagonal structure of the field in the emission cross section
66: \cite{tag:1}.
67:
68: If noise sources act, the "coherent" structures permitting the
69: high dissipation can appear stochastically. Examples are
70: appearance of vortices in wave turbulence \cite{tag:2},
71: vortex-antivortex generation in a noise driven Ginzburg Landau
72: equation \cite{tag:3}, and, evidently, phenomena such as
73: earthquakes and stock market crashes \cite{tag:4}. The mentioned
74: vortex-antivortex generation would seem to us to be a good
75: picture for the spontaneous appearance of particle-antiparticle
76: pairs out of vacuum in quantum physics and we would think that it
77: is also the proper model for the enigmatic phenomenon of 1/f
78: noise \cite{tag:5}, particularly since this description contains
79: the two best accepted explanations "self induced criticality"
80: \cite{tag:6} and the often mentioned coupled oscillator model as
81: special cases.
82:
83: "Coherence", "structure" or "order" in space or time occurs thus
84: in nonequilibrium systems to maximize the energy dissipation of
85: the system. If one accepts such a principle (to which to our
86: knowledge no counter examples are known) one can see this world
87: as driven by a "pressure to dissipate" (where the pump evidently
88: is the sun). In the way in which water pressed through a porous
89: material or through rocks with cracks will form highly complex
90: flow paths through the material, one can visualize by analogy the
91: complex pattern formation in nature, culminating in biological
92: structures, as forming under the pressure to dissipate and in the
93: cracks and pores given by the boundary conditions in the world.
94:
95: In this picture there is no need for a "plan" or "intention" to
96: generate the complex structures which we find developed under the
97: irradiation of the sun (absence of "teleological" elements).
98: Everything is dictated by a simple extremum principle, which is
99: "blind" in the sense that it can only find a nearest potential
100: minimum. (And the system will directly head there even if "death"
101: lurks in this minimum. There are good examples for such
102: "suicidal" systems in laser physics \cite{tag:7}.) A
103: "teleological" element or "planning" and an "intention" appears
104: only with the appearance of a brain, which can simulate the world
105: and find out that there are several minima of which the nearest
106: may not be the deepest (or that at the nearest minimum "death"
107: may lurk). Even then it appears that the teleological element
108: seems to be the exception. Even among the most complex structure,
109: the human species, most decisions are done unconsciously.
110:
111: It is evidently one of the goals of nonlinear physics to clarify
112: and understand the origin of life and of biological structures.
113: Although the path from simple periodic patterns, appearing as a
114: modulational instability, and spatial solitons, to structures as
115: complex as biological ones would seem exceedingly long at first
116: glance, this may not be so at second glance.
117:
118: A simple periodic pattern such as a hexagonal field structure in
119: the cross section of a nonlinear resonator can disintegrate into
120: spatial resonator solitons \cite{tag:8}. A closer examination of
121: these structures shows many common traits with biological
122: structures.
123:
124: Spatial solitons in nonlinear dissipative resonators\\
125: 1) are mobile (allowing them to move to the point in space where
126: they are most stable i.e. find the best "living conditions" (i.e.
127: they can search around for food));\\
128: 2) they possess a "metabolism": they have internal energy flows
129: and spatially varying dissipation, which stabilize them; \\
130: 3) they can "die": when the field sustaining them ("food") is
131: reduced, they will extinguish (die). And even a re-increase of
132: the field to a strength which allowed the solitons to exist
133: before will not reignite them, because of their bistability
134: (death is irreversible);\\
135: 4) they can multiply ("self-replicate") by repeated splitting
136: \cite{tag:9}.\\
137: There are further common traits of dissipative resonator solitons
138: and biological structures, among them the fact that both exist
139: only at medium nonlinearities and not at large or small
140: nonlinearities (see below).
141:
142: Consequently one could very well conjecture that the simplest
143: biological ("live") structures are not viruses (as commonly
144: taught) but dissipative solitons. This, by the way, shedding a
145: completely new light on the question of what biological life is:
146: life is usually ascribed to chemistry-based structures (with
147: carbon the central compound). Whereas if one takes the analogies
148: above seriously, "life" is not bound to chemistry but constitutes
149: itself as organizational "forms" of matter or fields. Thus the
150: probability of finding "live" structures is much larger than
151: finding "life" only in its form bound to carbon or chemistry.
152:
153: These analogies of dissipative solitons with biological
154: structures are one of the reasons for which one might want to
155: study spatial solitons. Furthermore these solitons are all by
156: definition bistable i.e. they constitute natural carriers of
157: information. As opposed to the information carrying elements in
158: electronic information processing, they are mobile and would
159: therefore perhaps permit processing functions beyond the reach of
160: electronic computing. We have recently studied the "aggregation"
161: of spatial solitons and find again common traits with biology.
162: Considering that these "aggregations" of information carriers
163: constitute something like a "brain" (a collection of
164: information-carrying elements, bound loosely to one another,
165: capable of exchanging information among the elements) there is
166: another motivation for the study: One might hope, in the way
167: there are analogies between solitons and simple biological
168: ("live") structures, to find operation principles of the brain by
169: studying the structure and properties of the "aggregations" of
170: spatial solitons.
171:
172: Optics and particularly resonator optics suggest themselves for
173: the investigations, because of possible photonic applications
174: such as parallel processing, urgently awaited already e.g. in
175: telecom. For this reason we have conducted in the past a series
176: of "proof of existence"-experiments on resonator solitons as well
177: as of their manipulation and properties \cite{tag:10}. Presently
178: spatial resonator solitons in semiconductor micro-resonators are
179: investigated, which are most likely best suited for
180: applications.\\
181:
182: \section{Proof of existence experiments of resonator solitons with slow materials}
183:
184: Spatial resonator solitons are self-formed localized structures,
185: which are free to move around. We have investigated several types:
186:
187: \begin{figure}[htbf] \epsfxsize=60mm
188: \centerline{\epsfbox{Fig01.eps}} \vspace{0.7cm} \caption{In a
189: laser with internal nonlinear absorber (bacterio-rhodopsin) a
190: soliton can be written in arbitrary places in the resonator cross
191: section. The resonator used is degenerate for all transverse modes
192: to allow any arbitrary field configuration to be resonant.}
193: \end{figure}
194:
195: Vortices occur in resonators with a gain medium without phase
196: preference, such as laser resonators. They are helical defects in
197: the wavefront and possess therefore, strictly speaking,
198: "tristability": right- as well as left-handed vortices coexist
199: with the wave without defect. We have given an overview of laser
200: vortices in \cite{tag:10} including examples of their dynamics in
201: Internet films \cite{tag:11}. Vortices were until now not studied
202: with regard to their usefulness as binary information carriers
203: (except for some basic experiments relating to pattern
204: recognition with lasers \cite{tag:12}).
205:
206: \begin{figure}[htbf] \epsfxsize=40mm
207: \centerline{\epsfbox{Fig02.eps}} \vspace{0.7cm} \caption{A soliton
208: can be made to move across the resonator cross section by a phase
209: gradient in the resonator.}
210: \end{figure}
211:
212: \begin{figure}[htbf]
213: \epsfxsize=40mm \centerline{\epsfbox{Fig03.eps}} \vspace{0.7cm}
214: \caption{In a resonator with a "phase trough" at the center a
215: soliton is drawn from all sides to the resonator axis and is
216: trapped there.}
217: \end{figure}
218:
219: Bright solitons are more reminiscent of the original idea of
220: optical information processing namely the use of optically
221: bistable resonators. They have therefore been considered more
222: closely for such tasks. In order to convince oneself that such
223: bright solitons in bistable resonators exist generally, we
224: conducted experiments with nonlinear media with slow response -
225: for the purpose of being able to observe the phenomena (which
226: entail 2D space-time dynamics) on a convenient time scale. As
227: opposed to semiconductors which have characteristic times of ps
228: to ns, e.g. photorefractive media and nonlinear absorbers like
229: bacterio-rhodopsine \cite{tag:13} have time constants of 10 ms to
230: 1 s, quite suitable for recording with ordinary video equipment.
231: In general a resonator containing only such a non-linear absorber
232: is too lossy for bistability. Therefore we used resonators
233: containing additionally gain elements to compensate for the
234: losses. A system like a laser with nonlinear absorber results
235: which is well suited to study these transverse effects such as
236: spatial modulational instabilities, patterns and solitons.
237:
238: \begin{figure}[htbf]
239: \epsfxsize=75mm \centerline{\epsfbox{Fig04.eps}} \vspace{0.4cm}
240: \caption{Larger number of stationary bright solitons coexisting.}
241: \end{figure}
242:
243: The first of such experiments yielded evidence of the existence of
244: spatial solitons \cite{tag:14}. In refined experiments we showed
245: the existence, the bistability and the manipulations of bright
246: soliton in this system \cite{tag:15}. FIG.~1 shows that the
247: solitons can be "ignited" anywhere in the cross section of the
248: resonator. FIG.~2 shows the motion of a bright soliton in a phase
249: gradient, and FIG.~3 shows the "trapping" of a bright soliton in a
250: "phase-trough". These are functions as they are required for
251: applications. FIG.~4 shows that large numbers of such bright
252: solitons can exist at the same time \cite{tag:9}, equally as
253: necessary for technical applications.
254:
255: \begin{figure}[htbf]
256: \epsfxsize=45mm \centerline{\epsfbox{Fig05.eps}} \vspace{0.7cm}
257: \caption{Phase solitons coexisting with non stationary phase
258: fronts. At the fronts and surrounding the soliton the phase of the
259: field changes by 180 degrees producing by interference the dark
260: ring surrounding the solitons and the dark fronts at the
261: phase-domain boundaries.}
262: \end{figure}
263:
264: As an interesting feature of these solitons it was found
265: \cite{tag:16} that these solitons have quantized velocity
266: magnitudes (including velocity magnitude zero). Although the
267: velocity magnitudes of such solitons are fixed, their direction
268: of motion is completely free, and varies e.g. under the action of
269: noise.
270:
271: It may be mentioned that in resonators with phase-sensitive gain,
272: such as gain supplied by 4-wave-mixing or degenerate parametric
273: mixing, solitons which concern only the phase structure of the
274: field (phase-solitons) were predicted \cite{tag:17} and
275: experimentally observed \cite{tag:18}. See FIG.~5.\\
276:
277: \section{Semiconductor Resonator Solitons }
278:
279: To make bright solitons suitable for technical applications, fast
280: materials have to be used. We found that the optimum system
281: concerning speed of response and nonlinearity is the
282: semiconductor micro-resonator. This is a resonator of $\sim$ 1
283: wavelength length which provides the shortest conceivable
284: resonator response time of 100 fs - 1 ps. The finesse of these
285: resonators has to be around 100 - 200 to provide bistability
286: together with a nonlinear medium. For the latter a semiconductor
287: slice of the length of the resonator ($\sim$ 1$\lambda$) is
288: suitable, and we stress that the response time of semiconductor
289: material is well matched to that of the resonator. Importantly:
290: it makes no sense to use "faster" materials than semiconductors
291: in such resonators. Since the small length of the resonator
292: provides a response time which is at the conceivable limit for
293: optical resonators, a faster nonlinear material does not decrease
294: the system response time. However, since higher speed comes at the
295: cost of a smaller nonlinearity, a faster material would require
296: unnecessarily high light intensities. Thus the $\sim$ 1$\lambda$
297: microresonator combined with a semiconductor material as the
298: nonlinear ingredient represents the optimum in terms of response
299: time and required light intensity. This structure is incidentally
300: the structure of VCSELs (vertical cavity surface emitting laser)
301: which during the last years have already been developed to a
302: certain degree of perfection. For the existence of spatial
303: solitons in such resonators absorbing ("passive") as well as
304: population-inverted ("active") material is suitable. Thus a
305: direct connection with VCSEL technology exists.
306:
307: FIG.~6 shows the structure of the nonlinear resonator. Its length
308: is $\sim$ 1$\lambda$ while the transverse size is typically 5 cm.
309: The short length and wide area of this microresonator permit only
310: one longitudinal mode (FIG.~7~(a)) while allowing an enormous
311: number of transverse modes so that a very large number of spatial
312: solitons can coexist. The resonator is obviously of the plane
313: mirror type, implying frequency degeneracy of all transverse
314: modes and thus allowing arbitrary field patterns to be resonant
315: inside the resonator. This is another prerequirement for
316: existence and manipulability of spatial solitons.
317:
318: \begin{figure}[htbf]
319: \epsfxsize=80mm \centerline{\epsfbox{Fig06.eps}} \vspace{0.7cm}
320: \caption{Schematic of semiconductor microresonator consisting of
321: two plane distributed Bragg reflectors (DBR) and multiple quantum
322: wells (MQW). Observation is in reflection.}
323: \end{figure}
324:
325: \begin{figure}[htbf]
326: \epsfxsize=85mm \centerline{\epsfbox{Fig07.eps}} \vspace{0.7cm}
327: \caption{Semiconductor microresonator reflectance spectrum (a),
328: typical bistability loop in reflection (b) and dynamics of
329: switching (c). Arrows mark the driving field that is detuned from
330: the resonator resonance and the pump field that is tuned to be
331: coupled into the resonator through one of the short-wavelength
332: interference notches of the resonator reflectance spectrum. (c)
333: shows that by short positive (negative) pulses the resonator can
334: be switched from one branch of bistability loop to the other.}
335: \end{figure}
336:
337: The resonator soliton existence is closely linked with the plane
338: wave resonator bistability (FIG.~7~(b)) caused by longitudinal
339: nonlinear effects: the nonlinear changes of the resonator length
340: (due to nonlinear refraction changes) and finesse (due to
341: nonlinear absorption changes) \cite{tag:19}. The longitudinal
342: nonlinear effects combined with transverse nonlinear effects
343: (such as self-focusing) can balance diffraction and form resonator
344: solitons. Generally these nonlinear effects can cooperate or act
345: oppositely, with the consequence of reduced soliton stability in
346: the latter case.
347:
348: \subsection{Model and numerical analysis}
349: %
350:
351: As a guide for the experiments we use a phenomenological model of
352: a driven wide area multiquantum well (MQW)-semiconductor
353: microresonator similar to \cite{tag:20,tag:21}. The optical field
354: $A$ inside the resonator is described in the mean-field
355: approximation \cite{tag:22}. The driving incident field {$A_{\rm
356: in}$} is assumed to be a stationary plane wave. Nonlinear
357: absorption and refractive index changes induced by the
358: intracavity field in the vicinity of the MQW-structure band edge
359: are assumed to be proportional to the carrier density $N$
360: (normalized to the saturation carrier density). The equation of
361: motion for $N$ includes optical pumping $P$, carrier recombination
362: and diffusion. The resulting coupled equations describing the
363: spatio-temporal dynamics of $A$ and $N$ have the form:
364:
365: \begin{eqnarray}
366: \cr {\partial A}/{\partial t}=A_{\rm in}-\sqrt{T}A\{[1+C{\rm
367: Im}(\alpha)(1-N)]+\cr{+i(\theta-C{\rm
368: Re}(\alpha)N-\nabla^{2}_{\bot})\}}\,, \cr\cr{\partial N}/{\partial
369: t}=P-\gamma[N-|A|^2(1-N)-d\nabla^{2}_{\bot}N]\,\,,
370: \end{eqnarray}
371:
372: where $C$ is the saturable absorption scaled to the resonator
373: transmission $T$ ($T$ is assumed to be small since the mirror
374: reflectivity is typically $\geq$ 0.995). ${\rm Im}(\alpha)(1-N)$
375: and ${\rm Re}(\alpha)N$ describe the absorptive and refractive
376: nonlinearities, respectively. $\theta$ is the detuning of the
377: driving field from the resonator resonance. $\gamma$ is the photon
378: lifetime in the resonator normalized to the carrier recombination
379: time. $d$ is the diffusion coefficient scaled to the diffraction
380: coefficient, and $\nabla^{2}_{\bot}$ is the transverse Laplacian.
381:
382: \begin{figure}[htbf]
383: \epsfxsize=70mm \centerline{\epsfbox{Fig08.eps}} \vspace{0.7cm}
384: \caption{Numerical solutions of Eq.(1) for unpumped and mixed
385: absorptive/self-focusing case. Area limited by dashed lines is
386: optical bistability domain for plane waves. Shaded areas are
387: domains of stability for bright/dark solitons and patterns. Insets
388: are bright and dark solitons in 3D representation.}
389: \end{figure}
390:
391: Linear effects in the resonator are spreading of light by
392: diffraction, and carrier diffusion (terms with
393: $\nabla^{2}_{\bot}$ in (1)). The material nonlinearity that can
394: balance this linear spreading can do this in various ways. It has
395: a real (refractive) and imaginary (dissipative) part and can act
396: longitudinally and transversely. The nonlinear changes of the
397: resonator finesse (due to nonlinear absorption change) and length
398: (due to nonlinear refractive index change) constitute
399: longitudinal nonlinear effects, also known under the name
400: \emph{nonlinear resonance} \cite{tag:17}. The transverse effects
401: of the nonlinear refractive index can be self-focusing (favorable
402: for bright and unfavorable for dark solitons) and self-defocusing
403: (favorable for dark and unfavorable for bright solitons).
404: Absorption (or gain) saturation (bleaching) leading to
405: \emph{nonlinear gain guiding} (in laser parlance), is a transverse
406: effect. Longitudinal and transverse effects can work oppositely,
407: or cooperate.
408:
409: \begin{figure}[htbf]
410: \epsfxsize=70mm \centerline{\epsfbox{Fig09.eps}} \vspace{0.7cm}
411: \caption{Maxwellian switched domain in reflection (b) numerically
412: calculated for parameters corresponding to the plane-wave OB area
413: in FIG.~8 and for Gaussian profile (a) of the driving beam.}
414: \end{figure}
415:
416: \begin{figure}[htbf]
417: \epsfxsize=70mm \centerline{\epsfbox{Fig10.eps}} \vspace{0.7cm}
418: \caption{Bright- (a) and dark- (b) spot-hexagons numerically
419: calculated for parameters corresponding to the Pattern existence
420: domain of FIG.~8. Driving light intensity increases from (a) to
421: (b).}
422: \end{figure}
423:
424: There are two principal external control parameters: the driving
425: field intensity $|A_{\rm in}|^2$ and the resonator detuning
426: $\theta$. FIG. 8 shows typical existence domains (in coordinates
427: $\theta$, $|A_{\rm in}|^2$) for all possible structures
428: (patterns, bright and dark spatial solitons) and plane-wave
429: bistability as calculated from (1) for the case of a mixed
430: absorptive/dispersive nonlinearity. At large resonator detuning
431: the intracavity field is transversely homogeneous and stable. If
432: in this case the driving beam has a Gaussian profile (FIG.~9~(a))
433: and its amplitude in the maximum exceeds the switching-on
434: threshold then the part of the beam cross section which is
435: limited by so called Maxwellian intensity \cite{tag:23} is
436: switched, thus forming a dark switched domain in reflection
437: (FIG.~9~(b)).
438:
439: For small resonator detuning and for driving intensities not
440: quite sufficient for reaching the resonance condition for the
441: whole resonator area, the system "chooses" to distribute the
442: light intensity in the resonator in isolated spots where the
443: intensity is then high/low enough to reach the resonance
444: condition, thus forming bright/dark patterns (FIG.~10). Instead
445: of saying "the system chooses" one would more mathematically
446: express this by describing it as a modulational instability. The
447: detuned plane wave field without spatial structure with intensity
448: insufficient to reach the resonance condition is unstable against
449: structured solutions. According to our numerical solutions of (1)
450: a large number of such structured solutions coexist and are
451: stable (see e.g. patterns in FIG.~18).
452:
453: The bright/dark soliton structures (inset in FIG.~8) can be
454: interpreted as small circular switching fronts, connecting two
455: stable states: the low transmission and the high transmission
456: state. Such a front can in 2D surround a domain of one state.
457: When this domain is comparable in diameter to the "thickness" of
458: the front, then each piece of the front interacts with the piece
459: on the opposite side of the circular small domain, which can
460: lead, particularly if the system is not far from a modulational
461: instability (see FIG.~8, Patterns), to a stabilisation of the
462: diameter of the small domain. In this case the small domain is an
463: isolated self-trapped structure or a dissipative resonator
464: soliton.
465:
466: \subsection{Experimental arrangement}
467: %
468:
469: FIG.~11 shows the optical arrangement for the semiconductor
470: soliton experiments, and in particular for their switching on or
471: off. The semiconductor microresonator consists of a MQW
472: (GaAs/AlGaAs or GaInAs/GaPAs) structure sandwiched between
473: high-reflectivity ($\geq$ 0.995) DBR-mirrors (FIG.~6) used at room
474: temperature. The microresonator structures were grown on GaAs
475: substrates by a molecular beam epitaxy technique that allows
476: growing MQW structures with small radial layer thickness
477: variation. The best sample used in our experiments has only
478: $\sim$ 0.3 nm/mm variation of the resonator resonance wavelength
479: over the sample cross section of 5 cm diameter.
480:
481: The driving light beam was generated by either a tunable (in the
482: range 750-950 nm) Ti:Sa laser or a single-mode laser diode
483: ($\sim$ 854 nm), both emitting continuously. For experimental
484: convenience and to limit thermal effects, the experiments are
485: performed within a few microseconds, by admitting the light
486: through an acusto-optical modulator. A laser beam of suitable
487: wavelength is focused onto the microresonator surface in a spot
488: of $\sim$ 50 $\mu$m diameter, thus providing quite large Fresnel
489: number ($\geq$ 100).
490:
491: \begin{figure}[htbf]
492: \epsfxsize=85mm \centerline{\epsfbox{Fig11.eps}} \vspace{0.7cm}
493: \caption{Experimental setup. Laser: Ti:Sa (or diode) laser, AOM:
494: acousto-optic modulator, $\lambda$/2: halfwave plate, PBS:
495: polarization beam spliters, EOM: electro-optical amplitude
496: modulators, BE: beam expander, PZT: piezo-electric transducer, P:
497: polarizer, L: lenses, BS: beam splitters, PD: photodiode. In some
498: cases an optical pump is used ("Pump"), see text.}
499: \end{figure}
500:
501: Part of the laser light is split away from the driving beam and
502: is superimposed with the main beam in a Mach-Zehnder
503: interferometer arrangement, to serve as a writing/erasing
504: (address) beam. This beam is tightly focused and directed to some
505: particular location in the illuminated area to create or destroy
506: a spatial soliton. The switching light is opened for a few
507: nanoseconds using an electro-optic modulator. For the case of
508: incoherent switching the polarization of the address beam is
509: perpendicular to that of the main beam to avoid interference. For
510: the case of coherent switching the polarizations are parallel and
511: a phase control of the switching field is always needed: for
512: switching on as well as switching off a soliton. One of the
513: interferometer mirrors can be moved by a piezo-electric element
514: to control the phase difference between the background light and
515: the address light.
516:
517: Optical pumping of the MQW-structures was done by a multi-mode
518: laser diode or a single-mode Ti:Sa laser. To couple the pump
519: light into the microresonator the pump laser wavelength was tuned
520: into the short wavelength reflection minimum as shown in
521: FIG.~7~(a).
522:
523: The observations are done in reflection (because the GaAs
524: substrate is opaque) by a CCD camera combined with a fast shutter
525: (another electro-optic modulator), which permits to take
526: nanosecond snapshots at a given time, of the illuminated area on
527: the resonator sample. Recording movies on the nanosecond time
528: scale is also possible. To follow the intensity in time in
529: certain points (e.g. at the location of a soliton) a fast
530: photodiode can be imaged onto arbitrary locations within the
531: illuminated area.
532:
533:
534: \subsection{Results and discussions}
535: %
536:
537: Switched (dark) domains exist for small negative resonator
538: detuning in the vicinity of the absorption band edge. The
539: switching front surrounding these domains is defined by the
540: spatial profile of the driving beam. The switching front is
541: located where the driving intensity equals the Maxwellian
542: intensity. FIG.~12 shows snapshots of various switched domains
543: demonstrating that they have the same shapes as the driving
544: beams. When changing the intensity of the driving beam in time
545: the fronts around the switched domains follow obviously an
546: equiintensity contour of the incident light (FIG.~13).
547:
548: \begin{figure}[htbf]
549: \epsfxsize=70mm \centerline{\epsfbox{Fig12.eps}} \vspace{0.7cm}
550: \caption{Dark switched domains observed at small (negative)
551: resonator detuning for round (a) and oval (b) driving beam
552: shapes.}
553: \end{figure}
554:
555: \begin{figure}[htbf]
556: \epsfxsize=60mm \centerline{\epsfbox{Fig13.eps}} \vspace{0.7cm}
557: \caption{Incident driving light intensity, center of Gaussian beam
558: (a) and reflectivity of sample (on diameter of circular driving
559: beam cross section) along with equiintensity contours of incident
560: light (b) demonstrating that borders of dark switched area follow
561: one of the equiintensity contours (heavy line).}
562: \end{figure}
563:
564: In contrast to these dark Maxwellian switched domains, spatial
565: patterns and solitons are self-sustained objects independent of
566: boundary conditions or a beam profile and can be both bright and
567: dark. FIG.~14 shows snapshots of bright and dark small ($\sim$ 10
568: $\mu$m) round spots (solitons), at large (negative) resonator
569: detuning, whose shape/size is independent of the shape/intensity
570: of the driving beam \cite{tag:24}. This independence of boundary
571: conditions allows to distinguish between patterns and solitons on
572: the one side and Maxwellian switched domains on the other side.
573:
574: \begin{figure}[htbf]
575: \epsfxsize=80mm \centerline{\epsfbox{Fig14.eps}} \vspace{0.7cm}
576: \caption{Bright- and dark-spot switched structures observed
577: experimentally at large (negative) resonator detuning. The
578: independence of the structures of the driving beam shape shows
579: that the structures are self-localized (patterns or spatial
580: solitons).}
581: \end{figure}
582:
583: \begin{figure}[htbf]
584: \epsfxsize=60mm \centerline{\epsfbox{Fig15.eps}} \vspace{0.7cm}
585: \caption{Recording of coherent switching-on (a) and switching-off
586: (b) of a bright soliton. Heavy arrows mark the application of
587: switching pulses. Dotted traces: incident intensity; solid traces:
588: reflected intensity at center of soliton. The insets show
589: intensity snapshots, namely unswitched state (left) and soliton
590: (right). Details see text.}
591: \end{figure}
592:
593: Resonator solitons can be written and erased by focused optical
594: (coherent) pulses independent of other bright (dark) spots
595: (solitons).
596:
597: FIG.~15 shows how solitons can be switched by light coherent with
598: the background light \cite{tag:25}. FIG.~15~(a) shows switching a
599: bright soliton on. The driving light intensity is chosen slightly
600: below the spontaneous switching threshold. At ${t}$ $\approx$ 1.2
601: $\mu$s the writing pulse is applied. It is in phase with the
602: driving light, as visible from the constructive interference. A
603: bright soliton results, showing up in the reflected intensity
604: time trace as a strong reduction of the intensity. FIG.~15~(b)
605: shows switching a soliton off. The driving light is increased to
606: a level where a soliton is formed spontaneously. The address
607: pulse is then applied in counterphase to the driving light, as
608: visible from the destructive interference. The soliton then
609: disappears, showing up in the reflected intensity time trace as
610: reversion to the incident intensity value. The FIG.~15 insets
611: show 2D snapshots before and after the switching pulses for
612: clarity.
613:
614: Thus depending on conditions we find switched domains, patterns
615: and bright and dark resonator solitons. Periodic patterns can be
616: distinguished from collections of solitons by the mutual
617: independence of the latter.
618:
619: To find the most stable resonator solitons for applications one
620: can play with the nonlinear (absorptive/dispersive) material
621: response by choice of the driving field wavelength and intensity,
622: with the resonator detuning, and finally with the carrier
623: population inversion (when using pumping). We recall that all
624: nonlinearities change their sign at transparency i.e. at the
625: point where the valence- and the conduction band populations are
626: equal. Going from below transparency (absorption) to above
627: transparency (population inversion, producing light
628: amplification), nonlinear absorption changes to nonlinear gain,
629: self-focusing changes to self-defocusing and vice versa, and
630: decrease of optical resonator length with intensity changes to
631: increase (and vice versa). The population of the bands can be
632: controlled by pumping ($P$ in (1)) i.e. transferring electrons
633: from the valence band to the conduction band. We do this by
634: optical excitation \cite{tag:26}, with radiation of a wavelength
635: shorter than the band edge wavelength. If the structures were
636: suited to support electrical currents (i.e. if it were a real
637: VCSEL-structure) pumping could evidently be effected by
638: electrical excitation.
639:
640:
641: \subsubsection{Illumination below bandgap (defocusing nonlinearity)}
642: %
643:
644: Working well below the bandgap (in the dispersive/defocusing
645: limit) with the driving field wavelength $\sim$ 30 nm longer than
646: the band edge wavelength we observe spontaneous formation of
647: hexagonal patterns (FIG.~16). The hexagon period scales linearly
648: with $\theta^{-1/2}$ \cite{tag:27} indicating that the hexagons
649: are formed by the tilted-wave mechanism \cite{tag:28}, which is
650: the basic mechanism for resonator pattern formation \cite{tag:29}.
651: Dark-spot hexagons (FIG.~16~(a)) convert to bright-spot hexagons
652: (FIG.~16~(b)) when the driving intensity increases. This is in
653: qualitative agreement with numerical simulations (FIG.~10).
654:
655: \begin{figure}[htbf]
656: \epsfxsize=70mm \centerline{\epsfbox{Fig16.eps}} \vspace{0.7cm}
657: \caption{(a) Bright (dark in reflection) and (b) dark (bright in
658: reflection) hexagonal patterns for the dispersive/defocusing case.
659: Driving light intensity increases from (a) to (b).}
660: \end{figure}
661:
662: \begin{figure}[htbf]
663: \epsfxsize=80mm \centerline{\epsfbox{Fig17.eps}} \vspace{0.7cm}
664: \caption{Switching-off of individual spots of a hexagonal
665: structure with address pulses aimed at different bright spots
666: (marked 1 and 2) of the pattern.}
667: \end{figure}
668:
669: \begin{figure}[htbf]
670: \epsfxsize=75mm \centerline{\epsfbox{Fig18.eps}} \vspace{0.7cm}
671: \caption{Numerical solutions of Eq.(1) for intracavity light
672: intensity as function of incident intensity: homogeneous solution
673: (dashed line marks modulationally unstable part of the curve) and
674: patterns (a-d). Shaded area marks existence range for dark-spot
675: hexagons.}
676: \end{figure}
677:
678: At high driving intensity we find that the bright spots in such
679: hexagonal patterns can be switched independently from one another
680: by focused optical (incoherent) pulses \cite{tag:27}. FIG.~17
681: shows the experimental results. FIG.~17~(a) shows the hexagonal
682: pattern formed. The focused light pulse can be aimed at
683: individual bright spots such as the ones marked "1" or "2".
684: FIG.~17~(b) shows that after the switching pulse was aimed at
685: "1", spot "1" was switched off. FIG.~17~(c) shows the same for
686: spot "2". We note that in these experiments we speak of true
687: logic switching: the spots remain switched off after the
688: switching pulse, (if the energy of the pulse is sufficient,
689: otherwise the bright spot reappears after the switching pulse).
690: These observations of local switching indicate that these
691: hexagonal patterns are not coherent patterns. The individual
692: spots are spatial solitons: they are independent, even at this
693: dense packing where the spot distance is about the spot size.
694:
695: \begin{figure}[htbf]
696: \epsfxsize=80mm \centerline{\epsfbox{Fig19.eps}} \vspace{0.7cm}
697: \caption{Stable hexagonal arrangements of dark spatial solitons:
698: (a) without defects, (b) with single-soliton defect, (c) with
699: triple-soliton defect. Parameters as for FIG.~18 (d). Compare with
700: the experiment FIG.~17 and [27].}
701: \end{figure}
702:
703: These experimental findings can be understood in the frame of the
704: model (1) \cite{tag:8}. FIG.~18 shows the bistable plane wave
705: characteristic of the semiconductor resonator for conditions
706: roughly corresponding to the experimental conditions. At the
707: intensities marked (a) to (d) patterned solutions exist.
708:
709: The pattern period in FIG.~18~(a) scales again linearly with
710: $\theta^{-1/2}$ due to the tilted wave mechanism: When the driving
711: field is detuned, the resonance condition of the resonator cannot
712: be fulfilled by plane waves travelling exactly perpendicularly to
713: the mirror plane. However, the resonance condition can be
714: fulfilled if the wave plane is somewhat inclined with respect to
715: the mirror plane (the tilted wave mechanism \cite{tag:28}). The
716: system chooses therefore to support resonant, tilted waves.
717: FIG.~18~(a) is precisely the superposition of six tilted waves
718: that support each other by (nonlinear) 4-wave-mixing. The pattern
719: period corresponds to the resonator detuning as in the experiment
720: for structures FIG.~16~(a). In this pattern (FIG.~18~(a)) the
721: bright spots are not independent. Individual spots cannot be
722: switched as in the experiment FIG.~17.
723:
724: On the high intensity pattern FIG.~18~(d) the pattern period is
725: remarkably different from FIG.~18~(a) even though the (external)
726: detuning is the same. This is indication that the internal
727: detuning is smaller and means that the resonator length is
728: nonlinearly changed by the intensity-dependent refractive index
729: (nonlinear resonance). From the ratio of the pattern periods of
730: FIG.~18~(a) and (d) one sees that the nonlinear change of
731: detuning is about half of the external detuning. That means the
732: nonlinear detuning is by no means a small effect. This in turn
733: indicates that by spatial variation of the resonator field
734: intensity the detuning can vary substantially in the resonator
735: cross section. In other words, the resonator has at the higher
736: intensity a rather wide freedom to (self-consistently) arrange
737: its field structure. One can expect that this would allow a large
738: number of possible stable patterns between which the system can
739: choose - or which are chosen by initial conditions.
740:
741: FIG.~19 shows that at the high intensity corresponding to
742: FIG.~18~(d) the model (1) allows to reproduce the experimental
743: findings on switching individual bright spots. FIG.~19~(a) is the
744: regular hexagonal pattern, at high intensity. FIG.~19~(b) shows
745: the field with one bright spot switched off as a stable solution
746: and FIG.~19~(c) shows a triple of bright spots switched off as a
747: stable solution, just as observed in the experiments \cite{tag:8}.
748:
749: Thus while FIG.~18~(a) is a completely coherent space filling
750: pattern, FIG.~18~(d) is really a cluster of (densest packed)
751: individual dark solitons. The increase of intensity from (a) to
752: (d) allows the transition from extended "coherent" patterns to
753: localized structures, by the increased nonlinearity, which gives
754: the system an additional internal degree of freedom. We note that
755: the transition from the coherent low intensity pattern to the
756: incoherent higher intensity structure proceeds through stripe
757: patterns as shown in FIG.~18~(b) \cite{tag:8}. For the intensity
758: of FIG.~18~(c) the individual spots are still not independent,
759: corresponding to the experiment FIG.~16~(b).
760:
761: \subsubsection{Illumination near bandgap (absorptive/defocusing nonlinearity)}
762: %
763:
764: Working at wavelengths close to the band edge we find bright and
765: dark solitons (FIG.~20), as well as collections of several spots
766: (FIG.~14~(b),(d),(e)) (all independent on the illumination beam
767: profile).
768:
769: The nonlinearity of the MQW structure near the band edge is
770: predominantly absorptive. For comparing with calculations we can
771: therefore in the first approximation neglect the refractive part
772: of the complex nonlinearity in the model equations (1) and
773: describe the nonlinear medium as a saturable absorber. Numerical
774: simulations for this case (FIG.~21) confirm existence of both
775: bright and dark resonator solitons as they are observed in the
776: experiment (Figs~14,~20). We can contrast these purely dissipative
777: resonator spatial solitons with propagating spatial solitons (in a
778: bulk nonlinear material) \cite{tag:30}: the latter can not be
779: supported by saturable absorption.
780:
781: \begin{figure}[htbf]
782: \epsfxsize=70mm \centerline{\epsfbox{Fig20.eps}} \vspace{0.7cm}
783: \caption{Dark and bright solitons experimentally observed near
784: bandgap.}
785: \end{figure}
786:
787: FIG.~22 shows details of the spontaneous formation of such bright
788: solitons as in FIG.~14~(c) and FIG.~20~(a). As discussed in
789: \cite{tag:24} material heating leads in this case to a slow
790: spontaneous formation of solitons, associated with the shift of
791: the semiconductor band edge by temperature \cite{tag:31}. In
792: FIG.~22 at ${t}$ $\approx$ 1.3 $\mu$s (arrow) the resonator
793: switches to high transmission (small reflection). The switched
794: area then contracts slowly to the stable structure FIG.~22~(b),
795: which is existing after ${t}$ $\approx$ 3.0 $\mu$s.
796:
797: \begin{figure}[htbf]
798: \epsfxsize=70mm \centerline{\epsfbox{Fig21.eps}} \vspace{0.7cm}
799: \caption{Numerical solutions of Eq.(1) for unpumped ($P$=0 ),
800: absorptive (${\rm Re}(\alpha)=0$ ) case. Area limited by dashed
801: lines is optical bistability domain for plane waves. Shaded areas
802: are domains of existence of bright and dark solitons. }
803: \end{figure}
804:
805: After the resonator has switched to low reflection its internal
806: field and with it the dissipation is high. A rising temperature
807: $\Delta$$T$ decreases the band gap energy \cite{tag:31} ($E_{\rm
808: g}$ $\approx$ $E_{\rm go}$ - $\alpha$$\Delta$$T$) and therefore
809: shifts the bistable resonator characteristic towards higher
810: intensity. Thus the basin of attraction for solitons which is
811: located near the plane wave switch-off intensity (see locations of
812: the existence domains for the bright solitons and the plane wave
813: bistability in FIG.~21) is shifted to the incident intensity,
814: whereupon a soliton can form. Evidently for different parameters
815: the shift can be substantially larger or smaller than the width of
816: bistability loop, in which case no stable soliton can appear. In
817: consequence, we note that in absence of thermal effects (good
818: heat-sinking of sample) solitons would not appear spontaneously,
819: but would have to be switched on by local pulsed light injection.
820:
821: \begin{figure}[htbf]
822: \epsfxsize=60mm \centerline{\epsfbox{Fig22.eps}} \vspace{0.7cm}
823: \caption{Bright soliton formation below bandgap. Reflectivity on a
824: diameter of the illuminated area as a function of time (b).
825: Intensity of incident (dotted) and reflected (solid) light, at the
826: center of the soliton as a function of time (a). Arrows mark the
827: switch-on and -off. Details see text.}
828: \end{figure}
829:
830: \begin{figure}[htbf]
831: \epsfxsize=60mm \centerline{\epsfbox{Fig23.eps}} \vspace{0.7cm}
832: \caption{Recording of incoherent switching-on (a) and
833: switching-off (b) of a soliton. Snapshot pictures show unswitched
834: state (left), circular switched domain (center) and a soliton
835: (right). Dotted trace: incident intensity, solid trace: reflected
836: intensity at center of soliton.}
837: \end{figure}
838:
839: Besides the coherent switching (FIG.~15) bright solitons can also
840: be switched by light incoherent with the background field. Under
841: these conditions the switching occurs due to change of carrier
842: density alone. FIG.~23~(a) shows incoherent switch-on of a bright
843: soliton, where a perpendicularly polarized switching pulse of
844: $\approx$ 10 ns duration is applied at ${t}$ $\approx$ 4.0
845: $\mu$s. As apparent, a soliton forms after this incoherent light
846: pulse. The slow formation of the soliton is apparent in
847: FIG.~23~(a) (using roughly the time from ${t}$ $\approx$ 4.0
848: $\mu$s to ${t}$ $\approx$ 4.5 $\mu$s), indicating again the
849: influence of material heating.
850:
851: It should be emphasized that this heating is not instrumental for
852: switching a soliton on. However, it allows switching a soliton
853: off incoherently \cite{tag:32}. This is shown in FIG.~23~(b)
854: where the driving light is initially raised to a level at which a
855: soliton forms spontaneously. The slow soliton formation due to the
856: heating is again apparent. The incoherent switching pulse is then
857: applied which leads to disappearance of the soliton.
858:
859: \begin{figure}[htbf]
860: \epsfxsize=65mm \centerline{\epsfbox{Fig24.eps}} \vspace{0.7cm}
861: \caption{Reflected light intensity measured at the center of a
862: soliton showing the regenerative pulsing (repeated switching on
863: and off of a soliton) resulting from combined thermal and
864: electronic effects.}
865: \end{figure}
866:
867: \begin{figure}[htbf]
868: \epsfxsize=65mm \centerline{\epsfbox{Fig25.eps}} \vspace{0.7cm}
869: \caption{Formation of a dark soliton (compare FIG.~22 and caption)
870: and motion. The motion visible in the streak picture (bottom) is
871: shown in 2D snapshots (top) for clarity. }
872: \end{figure}
873:
874: The soliton can thus be switched on and also off by an incoherent
875: pulse. The reason for the latter is thermal effect: Initially the
876: material is "cold". A switching pulse leads then to the creation
877: of a soliton. Dissipation in the material at the location of the
878: soliton raises the temperature and the soliton is slowly formed.
879: At the raised temperature the band edge (and with it the
880: bistability characteristic and the existence range of solitons)
881: is shifted so that a new pulse brings the system out of the range
882: of existence of solitons. Consequently the soliton is switched
883: off.
884:
885: Thus switching on a soliton is possible incoherently with the
886: "cold" material and switching off incoherently with the "heated"
887: material. When the driving intensity is chosen to be slightly
888: below the spontaneous switching threshold the nonlinear resonator
889: is cold. An incoherent pulse increases the carrier density
890: locally and can switch the soliton on, which causes local
891: heating. Another incoherent pulse aimed into the heated area can
892: then switch the soliton off and thereby return the resonator to
893: its initial temperature, so that the soliton could be switched
894: on/off again.
895:
896: This thermal effect combined with electronic nonlinearity can
897: cause spatial and temporal instabilities. A bright soliton
898: switching on spontaneously in the cold material will then heat
899: locally and can thereby destroy the condition for its existence,
900: so that it switches off. After the material has cooled the
901: soliton switches on again etc. Regenerative pulsing of the
902: soliton results, an example of which is shown in the observation
903: FIG.~24.
904:
905: As opposed to the bright soliton that heats the material locally
906: a dark soliton cools the material locally. This results in a
907: shift of the band edge and with it the switching characteristic
908: opposite to the bright soliton case. The consequence is that the
909: dark soliton moves laterally to places of uncooled material.
910: Here, however, it cools the material again so that a continuous
911: motion results, analogously to what we have described as the
912: "restless vortex" in \cite{tag:33}. Dark solitons in connection
913: with material heating effects tend therefore to be non
914: stationary. Such motion of a dark soliton was observed e.g. in
915: \cite{tag:24} and is shown in FIG.~25.
916:
917: \subsubsection{Illumination above bandgap (absorptive/focusing case)}
918: %
919:
920: At excitation above bandgap bright solitons form \cite{tag:34}
921: which have the same appearance as the bright solitons below/near
922: band gap (Figs~14~(c) and 20~(a)).
923:
924: FIG.~26 shows the dynamics of the bright soliton formation for
925: excitation above bandgap. The difference between solitons
926: existing above and below bandgap can be understood from the
927: model. From (1) we obtain the reflected intensity as a function
928: of incident intensity for wavelengths above the bandgap(${\rm
929: Re}(\alpha)>0$), as well as below the bandgap (${\rm
930: Re}(\alpha)<0$), for plane waves (FIG.~27). One sees that the
931: bistability range is large below and small above the bandgap.
932: Solving (1) numerically, the typical bright soliton (top of
933: FIG.~27) is found coexisting with the homogeneous intensity
934: solutions in the shaded regions of FIG.~27~(a),~(b).
935:
936: \begin{figure}[htbf]
937: \epsfxsize=60mm \centerline{\epsfbox{Fig26.eps}} \vspace{0.7cm}
938: \caption{Bright soliton formation above bandgap. Reflectivity on a
939: diameter of the illuminated area as a function of time (b).
940: Intensity of incident (dotted) and reflected (solid) light, at the
941: center of the soliton as a function of time (a). Arrows mark the
942: switch-on and -off. The soliton form fast (without mediation by a
943: thermal effect as in FIG.~22).}
944: \end{figure}
945:
946: \begin{figure}[htbf]
947: \epsfxsize=85mm \centerline{\epsfbox{Fig27.eps}} \vspace{0.7cm}
948: \caption{Steady-state plane wave solution of Eq.(1) above bandgap
949: (a) and below bandgap (b). The soliton solution shown exists for
950: incident intensities corresponding to the shaded areas, in
951: coexistence with the homogeneous solution. For a temperature
952: increase the characteristics shift together with soliton existence
953: ranges to higher incident intensities. Reflected and incident
954: intensities normalized to the same value. Note the different
955: widths of the bistability ranges above and below bandgap.}
956: \end{figure}
957:
958: Above the band gap FIG.~26 shows that the soliton is switched on
959: "immediately" without the slow thermal process described above.
960: FIG.~27~(a) shows why. The plane wave characteristic of the
961: resonator above band gap is either bistable in a very narrow
962: range, or even monostable (due to the contribution of the
963: self-focusing reactive nonlinearity \cite{tag:35}) but still with
964: bistability between the soliton state (not plane wave) and the
965: unswitched state. In this case the electronic switching leads
966: directly into the basis of attraction for solitons and the
967: switch-on of the soliton is purely electronic and fast. The widths
968: of the bistability characteristics observed experimentally
969: \cite{tag:34} correspond to the calculated ones FIG.~27.
970: Conversely the wide bistability range below bandgap requires the
971: (slow) thermal shift of the characteristic to reach the soliton
972: existence range.
973:
974: Nonetheless, also above bandgap there is strong dissipation after
975: the switch-on. The associated temperature rise influences and can
976: even destabilize a soliton. The destabilization effect can be seen
977: in FIG.~26~(a). Over a time of a few µs after the soliton
978: switch-on the soliton weakens (reflectivity increases slowly)
979: presumably by the rise of temperature and the associated shift of
980: the band gap. At 6.5 $\mu$s the soliton switches off, although the
981: illumination has not yet dropped.
982:
983: Thus, while the dissipation does not hinder the fast switch-on of
984: the soliton, it can finally destabilize the soliton. After the
985: soliton is switched off, the material cools and the band gap
986: shifts back so that the soliton can switch on again.
987:
988: \subsubsection{Optical pumping}
989: %
990:
991: The thermal effects discussed above result from the local heating
992: caused by the high intracavity intensity within the bright
993: soliton. They limit the switching speed of solitons and they will
994: also limit the speed at which solitons could be moved around,
995: thus limiting applications. The picture is that a soliton carries
996: with it a temperature profile, so that the temperature becomes a
997: dynamic and spatial variable influencing the soliton stability.
998:
999: As opposed, a spatially uniform heating will not cause such
1000: problems, as it shifts parameters but does not constitute a
1001: variable in the system. The unwanted heating effects are directly
1002: proportional to the light intensity sustaining a soliton. For
1003: this reason and quite generally it is desirable to reduce the
1004: light intensities required for sustaining solitons.
1005:
1006: Conceptually this can be expected if part of the power sustaining
1007: a resonator soliton could be provided incoherently with the
1008: driving field, e.g. by means of optical or electrical pumping.
1009: Pumping of the MQW structure generates carriers and allows
1010: conversion from absorption to gain. If pumping is strong enough
1011: the semiconductor microresonator can emit light as a laser
1012: \cite{tag:36}.
1013:
1014: FIG.~28 shows the variation of the plane-wave bistability domain
1015: with the pump as calculated from (1). The increase of the pump
1016: intensity leads to a shrinking of the bistability domain for
1017: plane waves (FIG.~28, pump intensities from $P_{\rm 1}$ to
1018: $P_{\rm 3}$) and the resonator solitons' existence domain while
1019: reducing the light intensity necessary to sustain the solitons.
1020: This reduction of the sustaining light intensity was observed
1021: experimentally (FIG.~29) \cite{tag:26}. When pumping below the
1022: transparency point of the material and with the driving laser
1023: wavelength near the semiconductor MQW structure band edge, bright
1024: and dark solitons form similarly to the unpumped case (FIG.~20).
1025: As the pump reduces the sustaining light intensity of the
1026: solitons, the heating effects are weak and the solitons switch on
1027: fast and unmediated by heating \cite{tag:26}.
1028:
1029: \begin{figure}[htbf]
1030: \epsfxsize=70mm \centerline{\epsfbox{Fig28.eps}} \vspace{0.7cm}
1031: \caption{Calculated plane-wave bistability domains as a function
1032: of the pump intensity. Dark shaded areas show bright soliton
1033: existence domains for two extreme cases: without pump ($P_{\rm
1034: 1}$=0) and with pump near lasing threshold ($P_{\rm 6}$).}
1035: \end{figure}
1036:
1037: \begin{figure}[htbf]
1038: \epsfxsize=87mm \centerline{\epsfbox{Fig29.eps}} \vspace{0.7cm}
1039: \caption{Measured (a) and calculated (b) switch-on and switch-off
1040: intensities for the driving light as a function of the pump
1041: intensity. The unphysical crossing of the on and off curves is an
1042: artifact from material heating.}
1043: \end{figure}
1044:
1045: When the pump intensity approaches the transparency point of the
1046: semiconductor material, the resonator solitons' domain of
1047: existence disappears (FIG.~28, pump intensities between $P_{\rm
1048: 3}$ and $P_{\rm 4}$). It reappears above the transparency point
1049: (FIG.~28, pump intensities from $P_{\rm 4}$ to $P_{\rm 6}$) and
1050: expands with the pump intensity (FIG.~30).
1051:
1052: Switched structures observed below the lasing threshold are shown
1053: in FIG.~31. FIG.~32 shows structures observed slightly above
1054: lasing threshold when a driving field is used simultaneously with
1055: the laser emission. The structures of FIG.~32 are reminiscent of
1056: the solitons in electrically pumped resonators \cite{tag:37}. It
1057: becomes clear that these structures (FIG.~32) must be
1058: soliton-collections or patterns, and not (linear) mode patterns,
1059: when looking at FIG.~32~(c). Here three bright spots are visibly
1060: separated by darker lines. In a "mode-pattern" the phase change
1061: from one bright spot to the next is $\pi$ (resulting in black
1062: lines due to destructive interference separating the bright
1063: spots). Therefore a "flower" mode pattern must have (and has)
1064: always an even number of "petals", whereas here we observe an odd
1065: number, inconsistent with phases in mode patterns. Thus we can
1066: conclude that there is no phase change between the bright spots
1067: and the latter are formed by self-localization, i.e. they are
1068: solitons or patterns in a resonator above transparency of the
1069: nonlinear material.
1070:
1071: \begin{figure}[htbf]
1072: \epsfxsize=70mm \centerline{\epsfbox{Fig30.eps}} \vspace{0.7cm}
1073: \caption{Results of numerical simulations of below bandgap
1074: solitons using the model (1) for a microresonator pumped close to
1075: the lasing threshold. Shaded areas are domains of existence of
1076: bright resonator solitons and patterns. Area limited by dashed
1077: lines is optical bistability domain for plane waves. }
1078: \end{figure}
1079:
1080: We note that optically pumped resonators allow more homogeneous
1081: pumping conditions than electrical pumping \cite{tag:38}. This
1082: suggests that optical pumping lends itself more readily for
1083: localization and motion control of solitons than electrical
1084: pumping.
1085:
1086: \begin{figure}[htbf]
1087: \epsfxsize=80mm \centerline{\epsfbox{Fig31.eps}} \vspace{0.7cm}
1088: \caption{Snapshots of typical optical structures at optical pump
1089: intensities slightly below lasing threshold (driving light
1090: intensity increases from (a) to (c)). With driving intensity dark
1091: spot patterns change to bright spot patterns.}
1092: \end{figure}
1093:
1094: There is a difference between resonator solitons below bandgap in
1095: pumped and unpumped material. The nonlinear resonance mechanism of
1096: soliton formation \cite{tag:17} uses a defocusing nonlinearity
1097: below transparency and a focusing nonlinearity above
1098: transparency. Defocusing nonlinearity stabilizes dark solitons
1099: and focusing nonlinearity stabilizes bright solitons. It follows
1100: that dark solitons should prevail for unpumped material and
1101: bright solitons for pumped material. FIG.~30 shows typical
1102: examples of calculated resonator solitons for a pumped
1103: semiconductor microresonator. Bright solitons have a large
1104: existence range in the pumped case (FIG.~28, at $P_{\rm 6}$),
1105: dark solitons exist, though with smaller range of stability, in
1106: the unpumped case (FIG.~28, at $P_{\rm 1}$).
1107:
1108: \begin{figure}[htbf]
1109: \epsfxsize=80mm \centerline{\epsfbox{Fig32.eps}} \vspace{0.7cm}
1110: \caption{Snapshots of typical optical structures at optical pump
1111: intensities slightly above lasing threshold (driving light
1112: intensity increases from (a) to (c)). Structures are always bright
1113: spot patterns.}
1114: \end{figure}
1115:
1116: Thus pumped semiconductor resonators are well suited for
1117: sustaining solitons below bandgap: (i) the background light
1118: intensity necessary to sustain and switch resonator solitons is
1119: substantially reduced by the pumping and therefore destabilizing
1120: thermal effects are minimized, (ii) the nonlinear resonance
1121: effect and the transverse nonlinear effect (self-focusing)
1122: cooperate to stabilize bright solitons, therefore the domain of
1123: existence of below bandgap (purely dispersive) bright solitons
1124: can be quite large.
1125:
1126: \section{Conclusion}
1127: %
1128:
1129: We have shown in these experiments that optical solitons
1130: generally exist in nonlinear optical resonators. They can be
1131: vortices, phase solitons, or bright and dark solitons. For
1132: technical applications the experiments on semiconductor
1133: microresonators have shown the existence of bright and dark
1134: solitons. One can experimentally distinguish switched areas,
1135: patterns and solitons from each other and from mode-fields (i.e.
1136: fields whose structure is standing transverse waves resulting
1137: from boundaries). Thermal effects can lead to spatial and
1138: temporal instabilities. They can be controlled in various ways
1139: e.g. by working above bandgap or by pumping. Pumping allows
1140: furthermore to control the magnitude and the sign of the material
1141: nonlinearities, and thus to maximize stability ranges for bright
1142: (dark) solitons. Coherent periodic patterns can disintegrate with
1143: increasing driving field into independent spatial solitons.
1144:
1145: Optical resonator solitons show properties reminiscent of simple
1146: biological structures and one may speculate that aggregations of
1147: resonator solitons could allow information processing reminiscent
1148: of brain functions.
1149: \\
1150:
1151: Acknowledgment\\ This work was supported by Deutsche
1152: Forschungsgemeinschaft under grant We743/12-1.
1153:
1154: %\bibliographystyle{empty}
1155: \begin{thebibliography}{99}
1156: \vspace*{-0.7cm}
1157: \bibitem{tag:1} W.J. Firth, A.J. Scroggie: Europhys. Lett. \textbf{26}, 521 (1994)
1158:
1159: \bibitem{tag:2} U. Frisch: \emph{Turbulence}. The Legacy of A.N.Kolmogorov, Cambridge University Press
1160: (1995)
1161:
1162: \bibitem{tag:3} K. Staliunas: Int. Journ. of Bifurcation and Chaos \textbf{11}, 2845 (2001)
1163:
1164: \bibitem{tag:4} K. Staliunas: "Anticorrelation and Subdiffusion in Financial Systems", xxx.lanl.gov, cond.mat/0203591 (2002); subm. Phys.
1165: Rev. E (2002)
1166:
1167: \bibitem{tag:5} For a review see e.g. P. Dutta, P. M. Horn: Rev. Mod. Phys. \textbf{53}, 497 (1981)
1168:
1169: \bibitem{tag:6} P. Bak: Phys. Rev. Lett. \textbf{59}, 381 (1987)
1170:
1171: \bibitem{tag:7} See as examples: A. Jaques, P. Glorieux: Opt. Comm. \textbf{40}, 455 (1982) and S. Rushin, S.M. Bauer: Appl. Phys. \textbf{24}, 45 (1981)
1172:
1173: \bibitem{tag:8} V.B. Taranenko, C.O. Weiss, B. Sch\"apers: Phys. Rev. A \textbf{65}, 013812 (2002)
1174:
1175: \bibitem{tag:9} G. Slekys, K. Staliunas, C.O. Weiss: Opt. Comm. \textbf{149}, 113 (1998)
1176:
1177: \bibitem{tag:10} C.O. Weiss, M. Vaupel, K. Staliunas, G. Slekys, V.B. Taranenko: Appl. Phys. B \textbf{68}, 151 (1999)
1178:
1179: \bibitem{tag:11} See the supplementary electronic material to \cite{tag:10}: http://link.springer.de/jounals/apb
1180:
1181: \bibitem{tag:12} C.P. Smith, Y. Dihardja, C.O. Weiss, L.A. Lugiato, F. Prati, P. Vanotti: Opt. Comm. \textbf{102}, 105 (1999)
1182:
1183: \bibitem{tag:13} D. Oesterhelt, W. Stoekenius: Nature \textbf{233}, 149 (1971)
1184:
1185: \bibitem{tag:14} V.Yu Bazhenov, V.B. Taranenko, M.V. Vasnetsov: \emph{Proc. SPIE}, \textbf{1840}, 183 (1992)
1186:
1187: \bibitem{tag:15} V.B. Taranenko, K. Staliunas, C.O. Weiss: Phys. Rev. A \textbf{56}, 1582 (1997)
1188:
1189: \bibitem{tag:16} K. Staliunas, V.B. Taranenko, G. Slekys, R. Viselga, C.O. Weiss: Phys. Rev. A \textbf{57}, 599 (1998)
1190:
1191: \bibitem{tag:17} K. Staliunas, V.J. Sanchez-Morcillo: Opt. Comm. \textbf{139}, 306 (1997)
1192:
1193: \bibitem{tag:18} V.B. Taranenko, K. Staliunas, C.O. Weiss: Phys. Rev. Lett. \textbf{81}, 2226 (1998)
1194:
1195: \bibitem{tag:19} H.M. Gibbs: \emph{Optical Bistability - Controlling Light with Light}, Academic Press, (1985)
1196:
1197: \bibitem{tag:20} L. Spinelli, G. Tissoni, M. Brambilla, F. Prati, L.A. Lugiato: Phys. Rev. A \textbf{58}, 2542 (1998)
1198:
1199: \bibitem{tag:21} D. Michaelis, U. Peschel, F. Lederer: Phys. Rev. A \textbf{56}, R3366 (1997)
1200:
1201: \bibitem{tag:22} L.A. Lugiato, R. Lefever: Phys. Rev. Lett. \textbf{58}, 2209 (1987)
1202:
1203: \bibitem{tag:23} N.N. Rosanov:, Prog. Opt. \textbf{35}, 1 (1996)
1204:
1205: \bibitem{tag:24} V.B. Taranenko, I. Ganne, R. Kuszelewicz, C. O. Weiss: Phys. Rev. A \textbf{61}, 063818 (2000)
1206:
1207: \bibitem{tag:25} V.B. Taranenko, F.-J. Ahlers, K. Pierz: Appl. Phys. B \textbf{75}, 75 (2002)
1208:
1209: \bibitem{tag:26} V.B. Taranenko, C.O. Weiss, W. Stolz: Opt. Lett. \textbf{26}, 1574 (2001)
1210:
1211: \bibitem{tag:27} V.B. Taranenko, I. Ganne, R. Kuszelewicz, C. O. Weiss: Phys. Rev.A \textbf{61}, 063818 (2000)
1212:
1213: \bibitem{tag:28} P.K. Jacobsen, J.V. Moloney, A.C. Newell, R. Indik: Phys. Rev. A \textbf{45}, 8129 (1992)
1214:
1215: \bibitem{tag:29} W.J. Firth, A.J. Scroggie: Europhys. Lett. \textbf{26}, 521 (1994)
1216:
1217: \bibitem{tag:30} G.I. Stegeman, M. Segev, Science \textbf{286}, 1518 (1999)
1218:
1219: \bibitem{tag:31} T. Rossler, R.A. Indik, G.K. Harkness, J.V. Moloney, C.Z. Ning: Phys. Rev. A \textbf{58}, 3279 (1998)
1220:
1221: \bibitem{tag:32} V.B. Taranenko, C.O. Weiss: Appl. Phys. B \textbf{72}, 893 (2001)
1222:
1223: \bibitem{tag:33} C.O. Weiss, H.R. Telle, K. Staliunas, M. Brambilla: Phys. Rev. A \textbf{47}, R1616 (1993)
1224:
1225: \bibitem{tag:34} V.B. Taranenko, C.O. Weiss, W. Stolz: J. Opt. Soc. Am. B \textbf{19}, 8129 (2002)
1226:
1227: \bibitem{tag:35} S.H. Park, J.F. Morhange, A.D. Jeffery, R.A. Morgan, A. Chavez-Pirson, H.M. Gibbs, S.W. Koch, N. Peyghambarian, M.
1228: Derstine, A.C. Gossard, J.H. English, W. Weidmann: Appl. Phys.
1229: Lett. \textbf{52}, 1201 (1988)
1230:
1231: \bibitem{tag:36} V.B. Taranenko, C.O. Weiss: "Spatial solitons in an optically pumped semiconductor microresonator", xxx.lanl.gov, nlin.PS/0204048 (2002)
1232:
1233: \bibitem{tag:37} Report as given in www.pianos-int.org
1234:
1235: \bibitem{tag:38} W.J. Alford, T.D. Raymond, A.A. Allerman: J. Opt. Soc. Am. B \textbf{19}, 663 (2002)
1236:
1237: \end{thebibliography}
1238:
1239: \end{document}
1240: