Research Signpost

37/ 661 ( 2) , Fort P.O., Trivandrum - 695 023, Kerala, I ndia

Introduction to Controlling Laser Dynamics, 2008: ISBN: 978-81-308-0262-6

Editor: Alexander N. Pisarchik

Introduction to controlling laser

dynamics
Alexander N. Pisarchik
Centro de Investigaciones en Optica, Loma del Bosque 115, Lomas del
Campestre, 37150 Leon, Guanajuato, Mexico

Abstract
This Introduction is aimed at the general public
who is not well aware of lasers and nonlinear
dynamics and has minimal mathematical background,
to introduce the reader to terminology and
fundamental tenets adopted by the majority of
scientists working in the area of Nonlinear Laser
Dynamics and the theory of control, for better
understanding of the following part of the book.
General principles of laser operation and
fundamentals of nonlinear dynamics are described
and major achievements in laser dynamics are
illustrated with experiments and observations in
different types of lasers. Finally I outline some
perspectives for future research in laser dynamics.
Correspondence/Reprint request: Dr. Alexander N. Pisarchik, Centro de Investigaciones en Optica, Loma del
Bosque 115, Lomas del Campestre, 37150 Leon, Guanajuato, Mexico. E-mail: apisarch@cio.mx
2 Alexander N. Pisarchik

“Without chaos there IS NO ORDER.”

The Scriptures of the Un-Dead a journey into the church of Virtuality/Reality
Abulafia J. Purpleflower - the profit of cynicism (CyberSpace, 1992).

1. Laser as a dynamical system

A laser is an electronic optical device that produces coherent radiation.
The term "laser" has been generally accepted since about 1965. It is an
acronym for Light Amplification by Stimulated Emission of Radiation. At the
beginning it was called an optical quantum generator or an optical maser
(Microwave Amplification by Stimulated Emission of Radiation). The first
masers have been made in 1953-1954 by the Russian physicists Nikolay G.
Basov and Alexander M. Prokhorov and also independently by the American
physicist Charles H. Townes. In 1955 Prokhorov and Basov suggested an
optical pumping of multilevel system as a method for obtaining the population
inversion, which later became one of the main methods of laser pumping.
Townes, Basov, and Prokhorov shared the Nobel Prize in Physics in 1964 “For
fundamental work in the field of quantum electronics, which has led to the
construction of oscillators and amplifiers based on the maser-laser principle.”
Yet in 1974 my teacher, academician Boris Ivanovich Stepanov (1913-1987)
has made an ingenious foresight; in his popular scientific brochure “Dye
Lasers” he wrote: “A laser will enter to every house very soon.” Indeed, the
laser brought about a revolution in optical technology and had a far-reaching
influence in various fields of science and in life.
Unlike ordinary light sources such as lamps, electric bulbs, or discharge
tubes, the laser is an oscillator similar to a radio transmitter. It may be said that
the laser is a nonlinear optical oscillator or an amplifier of light waves.
Therefore, the laser can be described in a similar way as electronic circuits
with the help of circuit theory and rate equations.
Generally, the laser consists of an amplifying (active) medium with
inverted population and a cavity formed by two mirrors M1 and M2, as shown
in Fig. 1. The two mirrors constitute a Fabry-Perot resonator which confines
light at a resonant frequency between the mirrors. Besides plane mirrors,
concave mirrors, diffractive gratings, or Bragg reflectors are also used, while
occasionally an optical resonator with more than two mirrors is used. The
active medium is excited from an external source of energy. There are many
kinds of active media that can be used in the laser to obtain light at different
wavelengths. The active medium plays the essential role in amplifying light
through population inversion and stimulated emission. In fact, a laser is an
optical quantum generator which transforms different kind of energy
(electrical, chemical, optical) to coherent monochromatic radiation.
Introduction to control of laser dynamics 3

Pumping

Active medium

M1 M2

Figure 1. Laser with Fabry-Perot cavity. M1 is a cavity end output mirror and M2 is a
semitransparent mirror.

The more widespread notion of the laser as a system that potentially

exhibits intrinsic dynamics and instabilities resulted in two experiments giving
the first clear evidence of chaotic dynamics of simple lasers: observations on a
loss-modulated CO2 laser [1] and on a multimode infrared He-Ne laser [2].
These results, sharpening the perception of lasers as "unstable" systems,
were then followed by a large number of experimental and theoretical
investigations.
From the beginning of laser industry an important problem was
stabilization of irregular intensity fluctuations in the laser output. However,
instabilities and chaos are intrinsic in the nonlinear nature of laser systems with
more than three degrees of freedom. For instance, in multimode lasers or in
lasers with nonlinear intracavity devices the nonlinear coupling between
modes gives rise to instability and multistability in the laser intensity. The
irregular behavior of such lasers is clearly an undesirable situation for many
practical applications. The scientists clashed with this problem first in a diode-
pumped Nd:YAG (neodymium doped yttrium aluminium garnet) (1.06 µm)
with intracavity second harmonic generation (green light) and thus called this
"green" problem [3]. Several theoretical and experimental works have been
devoted to avoid such a chaotic behavior, for example [4].
The main concepts of Laser Dynamics were taken from the branch of
Applied Mathematics known as Dynamical Systems. Such mathematical terms
as bifurcation, attractor, phase space, deterministic chaos, intermittency, etc.
have been introduced to laser dynamics and now they are commonly used by
physicists for description and analysis of many dynamical phenomena
observed in lasers. Dynamical systems are normally regulated by parameters.
When the parameters change, so do the properties of the system. In particular,
the stability of a system may be investigated by considering the results of small
disturbance. At some points in the space of parameters some of these
properties may change discontinuously as a function of one or more
parameters. These points are called bifurcation points.
4 Alexander N. Pisarchik

“Man muß noch Chaos in sich haben, um einen

tanzenden Stern gebären zu können.”
(F. Nietzsche)

2. Historical overview
The active search for methods for controlling laser dynamics has been
started from the beginning of 90th, just after publications in 1990 and 1992 the
theoretical works of Ott, Grebogi, and Yorke [5] and Pyragas [6] who
proposed the methods for feedback control of chaos in dynamical systems in
general (Fig. 2). Up to now these pioneering works have been cited
respectively 2180 and 1003 times. It is curious, but the Pyragas method was
not recognized in the beginning. His paper was rejected from Physical Review
Letters and not recommended for publication in Physical Review journals.
Now his method is one of the most popular control methods for experimental
realization.
After the discovery of the feedback control of chaos, several methods of
adaptive control have been developed and applied to laser systems. Most of the
methods are based on stabilization of unstable periodic orbits embedded within
a chaotic attractor. This has been realized by applying either feedback (closed-
loop) control to an available system parameter or by a periodic modulation to
one of the system parameters at the appropriate frequency, what is known as
nonfeedback (open-loop) control. The latter implies a nonfeedback technique
which utilizes the effect of external perturbations on evolution of the system.
The open-loop control was introduced by Lima and Pettini [7] and
independently by Braiman and Goldhirsch [8]. Following the terminology
commonly accepted in the control theory, I shall use here the terms “open-
loop” and “closed-loop” control instead of “nonfeedback” and “feedback”
control which also can be often found in the laser literature.

Figure 2. Inventors of feedback control of chaos, theoreticians (from left to right)

Edward Ott, Celso Grebogi, Jaimes A. Yorke, and Kestutis Pyragas.
Introduction to control of laser dynamics 5

The first experimental control of chaotic oscillations in a laser has been

realized by Roy (Fig. 3) and his colleagues in 1992 [4]. By means of a simple
feedback algorithm of Hunt [9] they drove a Nd:YAG laser with a frequency-
doubling crystal into a periodic mode. Then, chaos control has been
experimentally performed in a CO2 laser by the groups of Glorieux [10]
(Fig. 3) at the University of Lille in France and Arecchi [11,12] (Fig. 4) at the
National Institute of Applied Optics (Istituto Nazionale di Ottica Applicata) in
Florence, Italy. Later, successful experiments on controlling laser dynamics
have been carried out with other types of lasers, including multimode solid-
state lasers [13,14], a semiconductor laser [15], a far-infrared laser [16], and a
fiber laser [17].
It should be noted that the control problem is strongly related to the
problem of synchronization, where synchronization can be considered as a
control goal. Synchronization of coupled chaotic lasers has received recently
a great attention due to important practical applications to secure
communications [18,19]. The ability to synchronize coupled chaotic systems
was first demonstrated by Pecora and Carroll yet in 1990 [20]. They have
shown that two identical chaotic systems can be made to synchronize by
linking them with a common signal. Later, in 1994 Roy and Thornburn [21]
have applied this method to synchronize two Nd:YAG lasers.
At the same time, another modification of this method was used in the
group if Shimizu [22] to synchronized two chaotic CO2 lasers, one of which
was driven by the output of another laser. Synchronization of chaotic lasers
were demonstrated experimentally in arrays of three Nd:YAG lasers [23],
Nd:YVO4 microchip lasers [24], erbium-doped fiber lasers [25], and
semiconductor lasers [26]. More recently, the synchronization effects were
studied in a bidirectional Nd:YAG ring laser with modulated pumping [27], in
two pulsing CO2 lasers coupled through a common saturable absorber [28], in

Figure 3. Experimentalists realized control of chaos in lasers: Rajarshi Roy (left) and
Pierre Glorieux (right).
6 Alexander N. Pisarchik

Figure 4. Fortunado Tito Arecchi (left), Riccardo Meucci (right) and Alexander N. Pisarchik
(center) in laser laboratory at National Institute of Applied Optics in Florence during
collaborative research in 1999.

two erbium-doped fiber lasers [29], and in a microchip LiNdP4O12 laser array
with self-mixing feedback modulation [30].
Lasers, among other complex systems, often display multistability or
coexistence of attractors. The coexistence of attractors in a laser system has
been experimentally demonstrated first by Arecchi et al. [1] in a loss-
modulated CO2 laser. Later, multistability has been detected in other lasers,
including a Nd:YAG laser with intracavity second harmonic generation [3] and
a fiber laser [17,31].
Synchronization effects in coupled multistable systems have been recently
demonstrated with Rössler oscillators [32]. Synchronization of multistable
lasers, especially semiconductor lasers, requires special serious consideration.
This topic may be prominent for future research projects.

“Chaos is simply the particular type of self-organizing dynamics that is most easily
studied in terms of simple nonlinear differential and difference equations.”
Ben Goertzel, «The Evolving Mind» (Gordon and Breach, Las Vegas, 1993).
Introduction to control of laser dynamics 7

3. Main achievements in laser dynamics

In the following, I shall consider some important results that, in my
opinion, may guide the reader toward a family of other works, which
demonstrate similar or related phenomena. I am aware that it would be
unrealistic to pretend covering all the body of papers which are offered in the
literature nowadays and it has been necessary to concentrate our treatment only
on few prototypical experiments, duly referring to the other literature on
similar matter. Along this line here I want to illustrate both laser experiments
designed to verify some theoretical aspects, and others that intend, with in the
state of art, to solve some real technological problems. Under this
methodology, I shall attempt to keep the chronological order.
The main stages in development of laser dynamics in the last 20 years can
be characterized by the following achievements:

1. Dynamical classification of lasers.

2. Observation of bifurcations and routes to chaos.
3. Coexistence of attractors (generalized multistability).
4. Spatio-temporal dynamics.
5. Polarization dynamics.
6. Control of bifurcations and chaos.
7. Synchronization of coupled lasers.

3.1. Dynamical classification of lasers

Lasers are usually classified according to the material that provides optical
amplification. This material determines largely the properties of the laser: the
mode of operation (pulsed or continuous), the emission wavelength, the output
power/energy and the coherence properties. Gases, liquids, and solids can
provide optical amplification when properly excited. A new classification
based on time scale considerations introduced by Arecchi et al. [33] opened a
new look at lasers as dynamical systems and can be considered as one of the
most important achievements in laser dynamics. Specifically, lasers operating
in a single emission mode are usually described by three equations. The three
relevant variables: field, population, and polarization, decay on very different
time scales, which are given by the relaxation rates κ, γ⎜⎜, and γ⊥. If one of
these constants is large compared with the others, the corresponding variables
relaxes fast and consequently adiabatically adjusts to the other variables. The
number of equations describing the laser is then reduced. Precisely, single-
mode lasers were called class A, B, and C depending on whether, after suitable
adiabatic elimination of fast variables, the laser dynamics is ruled by one, two,
or three equations, respectively.
Thus, the following classification has been introduced [33].
8 Alexander N. Pisarchik

Class A (e.g., He-Ne, Ar, Kr, dye): γ⊥ ≅ γ⎜⎜ >> κ. One single nonlinear
field equation remains that means stable coherent emission.
Class B (e.g., ruby, Nd:YAG, CO, CO2, semiconductor, fiber): γ⊥ >> κ ≥
γ⎜⎜. Only polarization is adiabatically eliminated and the dynamics is ruled by
two rate equations for field and population that allows for damped oscillations
of the energy between field and inversion (relaxation oscillations).
Class C (e.g., He-Cd, He-Xe, far-infrared gas lasers): γ⊥ ≅ γ⎜⎜ ≅ κ. The
complete set of three equations has to be used, hence Lorenz like chaos is
feasible.
Class B lasers, however, show chaotic dynamics when they are externally
influenced (parametric modulation, injection of external light, feedback,
bidirectional ring cavity, or saturable absorber). Dynamical instabilities and chaos
observed in experiments with a class-A multimode infrared He-Ne laser [2], a
class-B loss-modulated CO2 laser [1], and a class-C ammonia laser [35] have been
properly understood and described mathematically in according with the above
classification. Moreover, this classification scheme allowed one to predict
dynamical behavior of new lasers depending on the relationship of the relaxation
rates of their variables, e.g., later to refer fiber lasers to the class-B lasers.
The onset of deterministic chaos in lasers was studied by referring to low
dimensional systems, in order to isolate the characteristics of chaos from the
random fluctuations due to the coupling with a thermal reservoir. In the case of
a single-mode laser with a homogeneous gain line, this means five coupled
variables or degrees of freedom, namely, complex field amplitude, a complex
polarization, and a population inversion. A corresponding quantum theory,
even for the simplest laser model, does not lead to a closed set of equations,
however the interaction with other degrees of freedom acting as a thermal bath
(atomic collisions, thermal radiation) provides truncation of high order terms
in the atom-field interaction. From the point of view of the associated
information, the standard interferometric or spectroscopic measurements of
classical optics, relying on average field values or on their first order
correlation function, are insufficient. In order to characterize the statistical
features of quantum optics, it was necessary to make extensive use of photon
statistics. From a dynamical point of view, coherence is equivalent of having a
stable steady state and this does not depend on details of the nonlinear
coupling, but on the number of relevant degrees of freedom. Since such a
number depends on the time scales on which the output field is observed
coherence becomes a question of time scales. This is the reason why for some
lasers coherence is a robust quality, persistent even in presence of strong
perturbations, whereas in other cases coherence is easily destroyed by the
manipulations common in the laboratory use of lasers, such as modulation,
feedback or injection from another laser.
Introduction to control of laser dynamics 9

3.2. Observation of bifurcations and routes to chaos

In general, a small variation in one or several laser parameters produces
small (continuous) changes in the laser output, so that the system is said to be
“structurally stable”. However, for some specific parameter values one of the
solutions (or attractors) may suffer a strong qualitative change. Such a
behavior is called a “bifurcation” and the system is said to be “structurally
unstable” for this parameter value. Very often when a control parameter is
varied and a bifurcation appears at some critical value, it is followed by a
sequence of new bifurcations at higher values of the parameter. Each new
attractor appearing in the bifurcation chain is usually more complex than the
previous one and eventually it becomes chaotic. The sequence is called a “road
(or route) to chaos”. The number of types of routes to chaos is unknown but it
has been observed that some of them appear very often, and for this reason
they are called “scenarios”. In the following I shall give a brief description of
the most important scenarios to chaos observed in lasers.

A. Feigenbaum scenario (period-doubling road)

The most common route to chaos for class B lasers is the one through a
sequence of subharmonic bifurcations which appear in the following way

T → µ1 → 2T → µ2 → 4T → µ3 → 8T → µ4 → ... µ∞ → chaos

where 2mT (m = 0,1,2,...) represents the period of a closed orbit attractor and
µn (n = 1,2,3,...) denote the critical values for the control parameter µ. This
sequence known as the Feigenbaum scenario [35] was observed
experimentally first in a loss-modulated CO2 laser [1] and then in many other
lasers of all classes. The typical time series corresponding to the period-
doubling route to chaos are shown in Fig. 5 (left-hand column) for a pump-
modulated fiber laser, as the modulation frequency decreases.
The Feigenbaum scenarios can be also identified by looking at the power
spectra (see right-hand column in Fig. 5). The period-doubling is clearly seen
through the appearance of the first subharmonic of the modulation frequency fm.

B. Ruelle-Takens-Newhouse scenario (quasiperiodicity road)

This scenario consists of a sequence of three Hopf bifurcations1 H1, H2,
and H3 at critical values µ1, µ2, and µ3.

FP→H1(µ1) →T→H2(µ2) →T² →H3(µ3) →T3→…chaos

1
A Hopf bifurcation may be simply thought of as a critical point where a stable steady
state transforms to a periodic orbit.
10 Alexander N. Pisarchik

(dB)
Intensity (V) 0

0.5 -20
-40
0.0 -60
(a)
-80
0.7 0.8 0.9 1.0
0
-20
Intensity (V)

0.5
-40
0.0 -60 (b)
-80
0.7 0.8 0.9 1.0
0
4 -20
Intensity (V)

2 -40
-60 (c)
0 -80
0 1 2 3 4 5 6 7 8 9 10 0 20 40 60 80 100
Time (ms) Frequency (kHz)

Figure 5. Period-doubling route to chaos with time series (left-hand column) and power
spectra (right-hand column) in a pump-modulated fiber laser. (a) Period 1 at fm = 72 kHz,
(b) period 2 at fm = 63 kHz, and (c) chaos at fm = 63 kHz.

As the diagram shows, the attractor is fixed point FP (or stable steady
state) for µ < µ1; at µ = µ1 it transforms into a periodic orbit T; at µ = µ2 it
changes again to a torus T² which entails a quasiperiodic behavior with two
incommensurate frequencies, and at µ = µ3 a new independent frequency
appears, so that in principle a T³ attractor (hypertorus) would be expected, but
in many cases it is unstable towards some kinds of fluctuations and becomes
chaotic. Figure 6 shows the bifurcation diagram of the peak intensity
demonstrated the road to chaos through quasiperiodicity and Fig. 7 represents
the corresponding time series.
The Ruelle-Takens-Newhouse scenario has been identified by observing
the power spectrum of a dynamical variable, which is similar to the schematic
example in Fig. 8. The broad-band spectrum corresponds to the chaotic
dynamical evolution: some peaks are usually still apparent, that indicates that
the previous periodic evolution has not completely disappeared.
The quasiperiodicity as a result of the interaction of three transverse modes
has been observed first by Weiss et al. [2] in a multimode He-Ne laser and by
Biswas and Harrison [36] in a multimode CO2 laser.
Introduction to control of laser dynamics 11

3
T

Intensity (arb. units)

4 H2 2
T

T
3
H1

2 FP H3

2 3 4 5 6
Feedback (ns-1)
µ1 µ2 µ3

Figure 6. Bifurcation diagram of peak intensity of a semiconductor laser with external

cavity with feedback strength as a control parameter.

(a) (b)
2.5 3
Intensity (arb. units)

Intensity (arb. units)

2.0
2

1.5

1
1.0

0 2 4 6 8 10 0 2 4 6 8 10
Time (ns) Time (ns)
5
(c) (d)
Intensity (arb. units)

4
Intensity (arb. units)

3
2
2

1 1

0
0 2 4 6 8 10 0 2 4 6 8 10
Time (ns) Time (ns)

Figure 7. Time series of laser intensity of a semiconductor laser with external cavity for
different feedback strengths demonstrating (a) fixed point (FP), (b) periodic orbit (T),
(c) quasiperiodic orbit (T2), and (d) chaos (T3).
12 Alexander N. Pisarchik

5
10 10
5

10
4
(a) 10
4
(b)
3
10 3
Intensity (dB)

10

Intensity (dB)
2
10 10
2

1
10 10
1

0
10 0
10
-1
10 -1
10
-2
10 -2
0 5 10 15 20 25 10
0 5 10 15 20 25 30 35
Frequency (GHz)
Frequency (GHz)

Figure 8. Power spectra of laser intensity corresponding to (a) quasiperiodic and (b)
chaotic motions. Chaotic attractor has a broad-band spectrum.

C. Intermittency scenario
The intermittency route to chaos is characterized by short, irregular
(turbulent) bursts, interrupting the nearly regular (laminar) motion (Fig. 9).
The duration of the turbulent phases is fairly regular and weakly dependent on
control parameter µ, but the mean duration of the laminar phase decreases as µ
increases beyond its critical value, and eventually they disappear. Hence only
one bifurcation point is associated with the intermittency route to chaos.
Five types of intermittency have been observed in lasers: type-I, type-II,
and type-III of Pomeau-Manneville intermittency [37], on-off [38], and crisis-
induced intermittency [39]. A particular type of intermittency depends on the
type of bifurcation at the critical point. The type-I and on-off intermittency are
associated with saddle-node bifurcations, the type-II and type-III intermittency
with Hopf bifurcation and inverse period-doubling bifurcation, respectively,
and crisis-induced intermittency with crisis of chaotic attractors when two
(or more) chaotic attractors simultaneously collide with a periodic orbit
(or orbits) [40]. Quantitatively, intermittency exhibits characteristic interburst
interval (laminar phase) statistics.
The type-I intermittency road to chaos has been found by Brun et al. [41]
close to periodic windows in the chaotic domain in experiments with a
linewidth-modulated nuclear magnetic resonance (NMR) ruby laser. They also
observed sudden expansions of attractors, giving evidence of “crisis”. The
phenomenon “crisis” has been studied experimentally in more details with a
loss-modulated CO2 laser in the group of Glorieux [42]. The type-II
intermittency route to chaos has been observed in a gain-modulated CO2 laser
with cavity detuning [43] and in an external cavity semiconductor laser [44].
The type-III intermittency road to chaos has been found experimentally in an
optically pumped FIR ring laser [45].
Introduction to control of laser dynamics 13

Intensity (V)
P2 P2
2

0 20 40 60 80 100
Time (ms)

Figure 9. On-off intermittency in a fiber laser with pump modulation. The period-2
regime alternates at random times with the coexisting chaotic regime.

The mechanism for on-off intermittency relies on the time-dependent

forcing (stochastic or periodic) of a bifurcation parameter through a
bifurcation point, while the system switches between two or more unstable
states, which are stable without external forcing. Instead, in Pomeau-
Manneville intermittency and crisis-induced intermittency the parameters are
static. Crisis-induced intermittency has been observed experimentally first in
a NMR laser [46] and on-off intermittency in a semiconductor laser with
external cavity [47].

D. Homoclinic (or Shilnikov) chaos

This phenomenon introduces unavoidable statistical features in nonlinear
laser dynamics. In such case the collective description in terms of a few
dynamical variables breaks down, because of large fluctuations. This was
first observed by Arecchi et al. [49] in a CO2 laser with feedback, and then
similar dynamics has been found out in a laser with a saturable absorber [50]
and in a multimode solid-state laser with intracavity second-harmonic
generation [51].
The experimental data obtained by Arecchi et al. [48] with a CO2 laser
with feedback displayed close agreement with those arising from the theory of
Leonid Shilnikov [49] (Fig. 10). Moreover, they also found out the
phenomenon of fluctuation enhancement in the decay of the unstable state of a
macroscopic system. Normally homoclinic orbits are located within very
narrow parameter ranges between periodic ranges. Recently, it was shown
numerically and experimentally in a CO2 laser with feedback that these narrow
chaotic windows are distributed exponentially as a function of a control
parameter (see Refs. [53,54] and Chapter 1 of this book).
14 Alexander N. Pisarchik

100

I (mV)
50

0
38 40 42 44 46 48
t (ms)

Figure 10. Shilnikov chaos in a CO2 laser with feedback.

3.3. Coexistence of attractors

Nonlinear dissipative systems often exhibit two or more dynamic
equilibrium states for the same values of parameters. Some states may be chaotic
while others are regular (periodic). Such multistability appears to be common for
a variety of nonlinear systems. In a system with coexisting attractors a particular
state is determined by initial conditions. The systematic organization of
coexisting attractors allows one to predict the behavior of lasers, when initial
conditions are allowed to evolve to their final states. Already in the beginning of
the 1980s, the nonlinear behavior of various lasers was experimentally explored
with respect to the emergence of coexisting states.
The first experimental evidence of multistability in lasers has been
demonstrated by Arecchi et al. [1] in a Q-switched CO2 laser. Since then, it has
been shown that lasers exhibit multistability due to pump modulation [55], loss
modulation [56] or optical injection [57,58]. Recently, a rich variety of
bifurcations and coexistence of multiple attractors which appear in the primary
saddle-node bifurcations and their relation to main laser resonances have been
demonstrated in a diode-pumped Erbium-doped fiber laser with pump
modulation [31,59] (Fig. 11).
The dynamics in lasers with an optical feedback due to the reflection from
a mirror often leads to delay differential equations introduced by Lang and
Kobayashi in 1980 [60], where the delayed feedback is a source of the
emergence of complicated behavior including multistability [61,62]. Masoller
[63] has shown that inevitable noise and in some cases a certain amount of
external feedback results in complex hopping dynamics between different
coexisting attractors in the system. For a particular range of noise amplitudes,
this jumping exhibits a resonant behavior due to the interplay of the delayed
feedback and the noise. The residence times probability distribution measuring
Introduction to control of laser dynamics 15

Period 1 Period 4 Period 3

3
Laser intensity (V)
2

0.6 0.7 0.8 0.9 1.0 0.7 0.8 0.9 1.0 0.7 0.8 0.9 1.0
Time (ms) Time (ms) Time (ms)

Figure 11. Coexisting Attractors in a fiber laser with pump modulation.

the time spent by the trajectory in the neighborhood of an attractor exhibits

peaks at multiples of the delay time. The strength of the peaks reaches a certain
maximum for an optimal noise level indicating the resonance.
Laser physics provides also important technological problems which need
to control multistability. One famous example is the so-called “green problem”
[3] present in the operation of intracavity frequency-doubled Nd:YAG lasers.
Usually, these lasers emit infrared light but using a nonlinear optical crystal,
this light can be converted into visible green light. Unfortunately, this
procedure produces irregular fluctuations in the output intensity so that
additional stabilization mechanisms like feedback control need to be applied to
obtain a stable output [64].
The feedback control, however, induces multistability which can be
tackled by the attractor annihilation method [65,66] to make the system
monostable. The method implies a harmonic perturbation leading to attractor
destruction. This method has been realized experimentally first in a loss-
modulated CO2 laser [65] and then in a pump modulated fiber laser [17].

3.4. Spatiotemporal dynamics

So far, when speaking of instabilities we have thought of temporal
behavior, i.e. the onset of periodic or chaotic pulsing. On the other hand, one
can of course also think of spatial instabilities, which means that the spatial
structure of the laser mode changes at, e.g., certain pump strength. One would,
of course, then expect from a combination of spatial and temporal instabilities
the phenomenon of “turbulence” as found in fluids or more generally in
spatially extended systems with time delays in coupling between spatially
separated regions.
The first example indicating that a transverse structure of the laser field
might be important for the dynamics of lasers was given by Lugiato and Milani
16 Alexander N. Pisarchik

[67]. It was shown that, if a fixed Gaussian laser mode transverse intensity
distribution is assumed, in a homogeneously broadened laser there are no “bad-
cavity” or “good-cavity” single-mode laser instabilities. In other words, neither
single-mode instabilities nor instabilities involving multiple longitudinal
modes occur.
In the experiment described by Tamm [68], a laser was forced to oscillate
in modes with zero intensity at the center. This was achieved by a small
absorbing dot placed at or near the center of the laser mode volume. The laser
is then usually emitting in a TEM01 or TEM10 mode, or both. By moving the
absorber dot, the frequency of these modes can be made more or less equal.
Optical bistability was observed between the two modes. This fundamental
laser instability gives the laser a basic capability of “pattern recognition” and a
“content-addressable memory”. Later, spontaneous periodic pulsing was
observed experimentally in a CO2 laser [69].
Chaotic behavior in spatial degree of freedom was recently used for
multichannel communication with a multimode semiconductor laser [70].
Information was encoded in different longitudinal cavity modes, demonstrating
a technique for multiplexing.
The interaction of the cavity modes in a laser results in the formation of
transverse spatial patterns and spatial multistability [71]. Brambilla et al. [72]
suggested this phenomenon could be of importance in the context of optical
information processing like designing associative memories. Two-dimensional
optical patterns can also emerge due to a phase-locking of several wave vectors
with different lengths and orientations [73].

3.5. Polarization dynamics

Devices to specify the state of polarization of light have been
commercially available for many years, but these devices are not designed to
measure light with rapid fluctuations of its sate of polarization because they
require stable state of polarization for time scales from milliseconds to
seconds. An accurate measurement of some optical phenomena, however,
requires tracking a fluctuating state of polarization on much faster time scales.
Van Wiggern and Roy [74] have developed the polarization analyzer capable
of measuring fluctuating state of polarization with time-scales as short as
several nanoseconds. Using this analyzer, they carried out experimental
measurements which provided new insights into the rapid polarization
dynamics of an erbium-doped fiber ring laser.
Polarization dynamics has been studied theoretically by Kulmiskii et al.
[75] who found polarization chaos in an optically pump laser [75]. Later,
Leyva et al. [76] have reported on experimental study of polarization dynamics
of a multimode CO2 laser.
Introduction to control of laser dynamics 17

3.6. Control of bifurcations and chaos

The control usually means achieving safe operations of a laser system onto
some unstable orbits giving the possibility of choosing (and changing) a
controlled, complex, multiperiodic desired behavior without hard changes in
the parameter space of the system. The control of stability in lasers and
nonlinear optical devices is frequently necessary. In lasers where irregular
fluctuations are normally obtained, there is a strong practical interest in
obtaining output intensity or a frequency free of irregular behavior. This is the
main reason why a lot of different control techniques have been used.
Indeed, from the beginning of laser industry a crucial point was to improve
the performance in frequency and intensity, motivated by many problems
inherent to the presence of several longitudinal and transverse modes (pulling
and pushing of laser modes), coupling in nonlinear devices, nonlinearities in
amplifiers and other issues related to the generation of laser radiation or due to
interaction of laser with matter.
Control of oscillations in lasers has some features mostly associated with
the nonlinear nature of the system. Most of the methods are based on the
stabilization of unstable periodic orbits embedded within a chaotic attractor.
This can be realized by applying either "feedback" (closed-loop) control to an
available system parameter or by a periodic modulation to one of the system
parameters at the appropriate frequency, what is known as "nonfeedback"
(open-loop) control. In general, I cannot claim which control method is better
than others for practical experimental implementations. Rather, the
performances of the different proposed techniques depend on the particular
situation under consideration.
As was already mentioned above, Roy and his colleagues [4] carried
out the first experimental control of chaotic oscillations in a laser. By
means of the occasional proportional feedback (OPF) method for selecting
a series of perturbations of limited durations (“kicks”) to the driver, a
diode-pumped Nd:YAG laser with a frequency-doubling crystal was driven
into a periodic mode.
The possibility for controlling bifurcations and chaos in a semiconductor
laser with external cavity was studied in several works [77-81]. A
semiconductor laser is commonly known as a system with extremely well
ordered dynamics. Extending a semiconductor laser by means of an external
resonator providing a weak optical feedback causes high dimensional chaotic
fluctuations of the light intensity. These phenomena, however interesting, have
been known for a long time and are well studied. Risch and Voumand [82]
were the first who described the so-called low-frequency fluctuations (LFFs)
of a semiconductor laser. They occur for very small reflectivities and pump
currents only slightly above the threshold current.
18 Alexander N. Pisarchik

In 1999 Erneux et al. [77] have found theoretically that for a sufficient
feedback rate LFF suppression takes place and chosen the second cavity to
suppress LFF. Later In 2000, Rogister et al. [78] confirmed these results
experimentally. Recently, it was demonstrated theoretically that additional
resonators with different round trip times may turn these fluctuations into more
ordered oscillations or even lead back to stable steady state operation [81].

3.7. Synchronization of coupled lasers

The control of chaotic behavior has another important application, namely
the synchronization of chaotic systems. If we consider two identical chaotic
systems starting from different initial conditions, then the critical sensitivity to
initial conditions implies that their difference grows exponentially in time, and
that they will evolve in an unsynchronized manner. The feeding of the right
signal from one system to another can, however, reduce to zero such
difference, and push the two systems into a synchronized manifold, wherein
the chaotic motion is now developed so as the system are in step during the
course of time. This proposal was intensively pushed forward at the last decade
of the 20th century.
During recent years, an interest in studying synchronization phenomena in
lasers has received much attention. The main reasons of such interest are
fundamental aspects for deeper understanding of coherent dynamical behavior
of coupled systems and due to the important practical application to secure
communications. By “synchronization” is commonly meant that the states of
ensembles of two and more coupled oscillators with different individual
frequencies are properly adjusted. Depending on the coupling strength and
individual frequencies of coupled systems, synchronization can be either
“complete” so that the states of the interacting systems coincide (master-slave
synchronization) or have opposite phases (antiphase synchronization) or
“partial” so that the states are slightly different. Particular cases of partial
synchronization are “phase synchronization”, when the phases of coupled
oscillators are equal while their amplitudes are noncorrelated and “lag
synchronization”, when the amplitudes are correlated but one system lags in
time to the other. In general, synchronization can be defined as the presence of
a functional relation between the states of master and slave systems
(“generalized synchronization”). The typical example of phase synchronization
in a dual-cavity CO2 laser is shown in Fig. 12 [83].
Lasers are among the most convenient dynamical systems as for theoretical
and experimental study of synchronization phenomena as for secure
communications. Recently, different stages of synchronization, from complete
synchronization to partial synchronization including phase and lag
synchronizations, have been studied in a class-B laser with modulated losses in
Introduction to control of laser dynamics 19

(a)
25

u1 (arb. units)
20

15

10

5
u2 (arb. units) 0 t
-5

(b) 2.04
(c)
u2 (arb. units)

10

5 T2 (ms) 2.02

0 2.00
0 10 20 30 2.00 2.02 2.04
u1 (arb. units) T1 (ms)

Figure 12. (a) Time series, (b) phase space of laser intensities, and (c) times of laser
spikes. The amplitudes are noncorrelated while the phases are completely synchronized.

one of the coupled channels [83]. Such a dual-wavelength optical source has
attracted much attention for a potential application in wavelength division
multiplexes transmission systems and optical signal processing. In particular, a
dual-wavelength CO2 laser is promising in remote sounding of the atmosphere,
isotope separation, and metrological applications. In molecular gas lasers many
wavelengths can be emitted as a result of several specified vibro-rotational
transitions. The coupling between lasing lines was realized through the rotational
relaxation, i.e. collision processes of gas molecules.
For stronger coupling or time delay in the coupling an intriguing effect
gives rise where the oscillators pull each other off their limit cycles and
collapse to a state of zero amplitude. This effect known as "oscillation death"
has been theoretically predicted by several authors and recently demonstrated
experimentally by my colleagues from Universitat Autonoma de Barcelona
[84] with a pare of thermo-optical oscillators linearly coupled by heat transfer.
Recently, the oscillation death has been found theoretically in a dual-cavity
CO2 laser with loss modulation [83] and experimentally verified in coupled
chaotic Nd:YVO4 lasers [85].
The majority of works on synchronization in lasers is devoted to
semiconductor lasers because these lasers are among the fastest and cheapest lasers
20 Alexander N. Pisarchik

in the market and hence they have potential application in communications. Here I
will not concentrate too much on this topic because two chapters in this book are
devoted directly to synchronization of semiconductor lasers.

“Life is chaos. Chaos is life. Control is an illusion.”

(Gene Roddenberry)

4. Control methods
In general, all existing methods of the dynamical control in lasers can be
schematically classified, depending on the control technique or on the type of the
control, into three main classes, namely: closed-loop or feedback methods, open-
loop or nonfeedback methods, and open-plus-closed-loop control (Fig. 13).
In the first class of the methods some information from the chaotic system
is extracted and used to design a weak correction signal applied to suitable
control parameter or variable. The possibility that chaos and instabilities can be
controlled efficiently using closed-loop schemes to stabilize unstable periodic
orbits (UPOs) was described by Ott, Grebogi, and Yorke (OGY) in 1990 [5].
The majority of the feedback methods are based on the OGY method which
allows stabilization of a given periodic orbit embedded in a chaotic attractor in
the form of small parameter perturbations, which are proportional to the
deviation of the system from the unstable fixed point. Following this work,
several algorithms have been developed to control chaos in various systems

METHODS OF DYNAMICAL CONTROL IN LASERS

CLOSED-LOOP CONTROL OPEN-LOOP CONTROL COMBINED CONTROL

OGY TDF Harmonic modulation Short pulses TDF+Slow modulation

OPF MED

Resonant Near-resonant Non-resonant

Tracking

Autonomous Nonautonomous Fast modulation Slow modulation

Figure 13. Schematic classification of control methods.

Introduction to control of laser dynamics 21

including lasers, for example, the “occasional proportional feedback” (OPF)

method proposed by Hunt [9] and the “minimal expected deviation” (MED)
method proposed by Reyl and his colleagues [86]. The OPF method has been
realized in the Nd:YAG [4,13] laser and lead-salt diode lasers [87] and the
MED method in a nuclear magnetic resonance laser [86].
Another approach to feedback control has been introduced by Pyragas [88]
who proposed to modify state variables instead of control parameters. The
“time delayed feedback” (TDF) Pyragas method implies a correction signal to
be proportional to the difference between values of a given variable at different
times; the delay time is selected equal to the period of the unstable orbit to be
stabilized. The last method has been successfully applied to a CO2 laser [10]
and a semiconductor laser [89].
The second class of the control methods includes the strategies in which a
small periodic perturbation is used to alter drastically the dynamics of the
system. This approach is known as nonfeedback control. These methods do not
require a prior knowledge of the system behavior. Therefore, they are
particularly appealing for lasers whose state is impossible or difficult to
measure in real time and where feedback control is difficult to realize. The
nonfeedback control can be realized in the forms of the resonant, near-
resonant, or nonresonant modulation applied either to an available system
parameter (cavity losses, pump parameter, or cavity detuning) or to a system
variable (usually as an injected radiation from another laser).
Finally, the third class of the control methods implies the combination of
the former two classes, the feedback and nonfeedback schemes [90,91], to
achieve a certain control goal. This type of control has been used in a chaotic
laser with periodically modulated feedback strength [90]. The delayed
feedback stabilizes simultaneously several periodic orbits, making the system
periodic but multistable, while the external modulation annihilates undesirable
periodic attractors.
The three classes of the control methods are of completely different
natures, but they both share the concept of controlling dynamics through the
use of small perturbations.
Usually the control in its general sense is thought of as the application of a
small signal to a system to achieve some control goals. Therefore, it also
makes sense to introduce another classification of the control methods, which
is based on the control goals. The traditional goals of the control are
stabilization of unstable fixed points and UPOs, tracking of UPOs, and
modification of the asymptotic behavior of the system (e.g., to change the type
of equilibrium, to change the type of limit cycle, to change of the bifurcation
point and/or the type of the bifurcation point in the parameter space, or to
create oscillations with desired properties). Thus, in the majority of works on
nonlinear laser dynamics the control is traditionally taken to mean stabilization
22 Alexander N. Pisarchik

of UPOs or modification of asymptotic behavior of a laser with a vanishing

control signal. There are also a large number of works where the authors
explore a generalization of the control, which includes situations when the
control signal is not small. Such a control, referred to as generalized control
allows one to reach some additional goals among the traditional ones, e.g.
dynamical tracking of unstable periodic orbits [92-94], targeting trajectories of
the dynamical system (targeting unstable or coexisting stable periodic orbits)
[95], the change of the type of attractor (from a fixed point into a limit cycle or
chaos), the change in organization of the basins of attraction [96], the creation
of additional attractors (attractor splitting) [97], and the annihilation of
undesirable coexisting attractors [66]. A particular control goal can be
achieved by different ways, i.e. with the use of open-loop or closed-loop
techniques, or their combination.
The number of theoretical and experimental works on controlling laser
dynamics and synchronization is very large and increases drastically every
year. The reader should understand that it is very difficult or even impossible
to discuss all of them. That is why in the following I will try to summarize the
most important, in my opinion, results on controlling laser dynamics and
synchronization with the emphasis mainly on experimental works.

4.1. Closed-loop control

The method of the closed-loop (or feedback) control of chaos was
enunciated first by Ott, Grebogi, and Yorke [5] (OGY method) who suggested
to stabilize an unstable periodic orbit embedded into a chaotic attractor. The
main idea consisted in waiting for a natural passage of the chaotic orbit close
to the desired periodic behavior, and then applying a small properly chosen
perturbation, in order to stabilize the unstable periodic orbit. This method is
based on extremely high sensitivity of a chaotic system to perturbations of its
initial conditions that may be, in fact, very desirable in practical situations.
Indeed, since a small perturbation can gives rise to a very large system
response, a proper choice of such a perturbation can direct the trajectory to the
neighborhood of the desired periodic orbit that can now be stabilized.
Many alternative approaches to the OGY method have been proposed to
stabilize unstable periodic orbits in a chaotic system. Among the feedback
methods, we should mention two the most popular methods. These are the so-
called occasional proportional feedback method introduced by Hunt in 1991
[9] and the method invented by Pyragas [6], who applied a delayed feedback to
one of the system variable. Until now the Pyragas method remains the most
prevalent method for controlling chaotic dynamics, because in many cases the
control parameters are strongly influenced by the environmental conditions,
and a variation of them is not easy to be performed.
Introduction to control of laser dynamics 23

A. Occasional proportional feedback control

The basic building scheme of a generic feedback blocks for stabilizing
unstable periodic orbits in a laser system is shown schematically in Fig. 14. It
consists of the chaotic laser that is to be controlled, a device to sense the
dynamical state (photodetector), a processor to generate the feedback signal,
and an actuator that adjusts the accessible laser parameter (modulator).
In their original conceptualization of the control scheme, OGY suggested
the use of discrete proportional feedback because of its simplicity and because
the control parameters can be determined straightforwardly from experimental
observations. In this particular form of feedback control, the state of the system
is sensed and adjustments are made to the accessible system parameter as the
system passes through a surface of the Poincaré section. Figure 15 illustrates a
portion of a trajectory in a three-dimensional phase space and one possible
surface of section that is oriented so that all trajectories pass through it. The
dots on the plane indicate the location where the trajectory pierces the surface.

Chaotic laser

Actuator Measure state of

the laser

Generate feedback
signal

Figure 14. Feedback scheme for controlling laser dynamics.

X3

X2

X1

Figure 15. A segment of a trajectory in a three-dimensional phase space and Poincaré

section.
24 Alexander N. Pisarchik

In the OGY control algorithm, the size of the adjustments is proportional

to the difference between the current and desired states of the system.
Specifically, consider a system whose dynamics on a surface of section is
governed by the m-dimensional map

zi+1 = F(zi, pi), (1)

where zi is its location on the ith piercing of the surface and pi is the value of
an externally accessible control parameter that can be adjusted about a nominal
value p0. The map F is a nonlinear vector function that transforms a point on
the plane with position vector zi to a new point with position vector zi+1.
Feedback control of the desired UPO, characterized by the location z*( p0) of
its piercing through the section, is achieved by adjusting the accessible
parameter by an amount

δpi = pi – p0 = - γn ⋅ [zi – z*( p0)] (2)

on each piercing of the section when zi is small neighborhood of z*(p0), where

γ is the feedback gain and n is an m-dimensional unit vector that is directed
along the measurement direction. The location of the unstable fixed point
z*(p0) must be determined before control is initiated. It can be determined from
experimental observations of zi in the absence of control. The feedback gain γ
and the measurement direction n necessary to obtain control is determined
from the local linear dynamics of the system about z*(p0) using the standard
techniques of modern control engineering and chosen so that the adjustments
δpi force the system onto the local stable manifold of the fixed point on the
next piercing of the system. Successive iterations of the map in the presence of
control direct the system to z*(p0). It is important to note that δpi vanishes
when the system is stabilized; the control only has to counteract the
destabilizing effects of noise.
In 1991, Hunt developed a modification of the OGY method called
occasional proportional feedback (OPF) to find, stabilize, and switch between
different orbits of a diode resonator [9]. In contrast to the original OGY
method, the Hunt's algorithm allows fairly large perturbations which can
change the chaotic system slightly, permitting high period orbits. With
completely analog technique, the OPF method allows the system to find stable
orbits itself. This is the most important advantage of the OPF method, because
detailed knowledge of the dynamics is not necessary, nor is it required to know
initially where the fixed points of the system lie. The basic algorithm of OPF is
to sample an accessible dynamical system variable, and if the value of the
variable falls within a prescribed range or "window", a system parameter is
Introduction to control of laser dynamics 25

modulated with amplitude proportional to the difference between the value of

the sampled variable and the centre of the window. The OPF method has been
particularly effective for high-dimensional systems where control perturbations
induce transients off the unstable manifold that hinder the effectiveness of the
OGY method [4,13].
However, the OPF method also requires the careful tuning of additional
experimental adjustment of control perturbation pulse width. An extension of
this method has been made by Carr and Schwartz [98,99], who proposed to use
the duration of the feedback control (pulse width) as an additional control
parameter to control steady states in systems with high complexity. Using this
additional control parameter, it is possible to compensate transients that occur
on the stable manifold. The authors consider the general system

•
z = F ( z , P ), (3)

where z is an n-dimensional state variable, P is a scalar parameter of the

system that is used as the control variable, and F is a nonlinear vector function
of the state and control variable. The existence of a steady state solution given
by (z(P),P) was assumed. The purpose of the control is to stabilize a particular
steady state when P = P and z (P ) = z. To achieve this control goal, the
authors induce the following approximation about this point

•
x = A • x + Bp, (4)

dF ( z , P )
A = D z F ( z , P ), B= , (5)
dP

where x = z − z << 1 and p = P − P << 1 are small deviations from the

steady states z and P , respectively. The authors also assume that there is a
single complex-conjugate pair of unstable eigenvalues σu(p) ± iω(p), where
σu > 0 and dσu/dp ≠ 0. This implis the existence of a Hopf bifurcation for some
lower value of P. The restriction on the additional eigenvalues is that they have
a negative real part. For model-independent control, the parameters A and B
can be determined by embedding the flow of the physical system into a virtual
phase space. If A is non-singular, i.e., P is not a bifurcation point, then the
general solution of Eq. (4) is

x(t) = eAt ⋅ [x(0) + A-1 ⋅ Bp] – A-1 ⋅ Bp. (6)

26 Alexander N. Pisarchik

The n × n matrix A can be block diagonalized as A = S ⋅ Λ ⋅ S -1, where S

is composed of the right eigenvectors ei, i = 1,...,n, and S -1 is composed of the
left eigenvectors fi. The right eigenvectors e1 and e2 are associated with the
unstable complex-conjugate pair of eigenvalues, and similarly the left
eigenvectors f1 and f2 with the corresponding left eigenvalues.
The first successful experiment on controlling chaotic laser dynamics has
been carried out by Roy et al. in 1992 [4] in a diode-pumped Nd:YAG laser
with an intracavity frequency-doubling crystal (KTP) (potassium titanyl
phosphate). The multimode, autonomous chaotic laser has been controlled by
the technique of occasional proportional feedback, related to the OGY control
scheme [5]. The basic technique for achieving dynamical control was as
follows. The authors first prepared the laser as to operate in a weak chaotic
regime, by adjusting the rotational orientation of the YAG and KTP crystals
such that only small amount (less than a few µW) of green light was generated
in the doubling crystal. Then, a system variable (the total laser output intensity)
was sampled within a window of selected offset and width. The sampling
frequency was related to the relaxation oscillation frequency of the laser. A
signal proportional to the deviation of the sampled intensity from the centre of
the window was generated and applied to perturb a system parameter (the drive
current) from its ambient value. This control signal repeatedly attempts to
bring the system closer to a periodic unstable orbit that is embedded in the
chaotic attractor, resulting in its stabilization. The control voltage fluctuations
become very small once the steady state is achieved. The authors reported that
adjustments of the control signal width and gain, as well as the wave-form
offset and window width, also were necessary to optimize the stability of the
wave forms obtained. The control of period-2, 4, and 9 orbits was obtained
with small amplitude perturbations (a few percent of the ambient-bias current).
The main advantage of this method is that a detailed model of the system
is not necessary.

B. Tracking unstable steady states

The OPF technique has been extended by application of the tracking
procedure [100,101] to stabilize an unstable steady state in a multimode solid-
state laser over a wide range of parameter values for the system, including both
chaotic and nonchaotic regions. In the experiments of Roy et al. [4], stable
steady state output has been obtained only for a very small range of pump
powers, for the given set of laser operating parameters. The application of both
tracking and control technique, the range of pump excitation over which
stabilization can be maintained has been extended by more than an order of
magnitude as the pump excitation is slowly varied.
Introduction to control of laser dynamics 27

The unstable chaotic fluctuations of the output intensity with no control

signal are stabilized to the steady state when the control is applied. The
stabilized steady-state intensity has the same average value as that of the
chaotic output. The fluctuations of the laser intensity are a few percent of the
steady-state dc level, reduced from the nearly 100% fluctuations in the
chaotic state. These results demonstrate that it is possible to successfully
stabilize the unstable steady state for a given pump power and control
parameter setting.
In order to achieve these results the authors steadily increased the pump
power in small steps from a value close to threshold, where stable steady state
was obtained without control. When unstable oscillations of the output
intensity were detected, control was switched on, and the reference intensity
level was optimized to obtain an essentially zero dc component of the control
signal. They found that the laser could be stabilized over the entire range of
pump power by tracking the steady state through adjustment of the intensity as
the pump power was increased, and through minor adjustment of the period 1
at which the output is sampled and the gating period over which the correction
signal is applied to the ambient value of the diode drive current. The tracking
technique allowed the author to obtain about 15 times more output power in a
stable steady state for a given set of laser operating parameters. The stabilized
steady-state values were very close to the average values of the fluctuating
unstable laser output for the same pump power. Thus, by the tracking
procedure allowed the extension of the stability regime for a pump power was
extended to more than 3 times above the threshold.

C. Time-delayed feedback control

A large amount of works on controlling laser dynamics is constituted by
different implementations of the method originally introduced by Pyragas
[6]. This method implies the application of a continuous time-delayed
feedback.
The main contribution of Pyragas was to introduce a correction based on
delayed variable as stated later, rather than action on a variable. The correction
signal is proportional to the difference between values of a given variable at
different times; the delay time is selected equal to the period of the unstable
periodic orbit (UPO) to be stabilized. Both approaches imply the action on the
state variable, whether this is directly or indirectly (through a system
parameter). Moreover, many methods, including variants of the OGY method,
do not require detailed knowledge of the system, otherwise they would not
have been implemented experimentally so rapidly in lasers; the knowledge on
the system necessary to select the perturbation can be obtained by simply
observing the system for a suitable learning time. The advantage of the
Pyragas method is that it does not require complete information about the
28 Alexander N. Pisarchik

target UPO; but it only uses a constant delay in time in the feedback loop. In
practice, the Pyragas method transforms a system of ordinary differential
equations into a delayed dynamical system. This increases its dimensionality
so as the desired unstable periodic orbit, which was unstable in the ordinary
differential equation system, becomes now stable in the new delayed
dynamical system.
Thus, the Pyragas method is based on the construction of a special form of
time-continuous perturbation, which does not change the form of the desired
UPO, but under certain conditions, can stabilize it. Pyragas considered a
system governed by the ordinary differential equations

•
x = Q(x, y ), y = P(x, y ) + F(t ), (7)

where x ∈ ℜn-1 describes the variables of the system that are not available or
not of interest and only y ∈ ℜ can be measured as a system output, Q: ℜn-1 ×
ℜ → ℜn-1 and P: ℜn-1 × ℜ → ℜ, and F(t) is the control signal. He proposed
two types of feedback control, delayed feedback control [88]

F(t ) = η[ y(t − τ ) − y(t )] (8)

and external feedback control

F(t ) = δ[ y(t ) − y(t )], (9)

to stabilize an UPO embedded in the chaotic attractor. Here y(t ) is the UPO to
be stabilized and τ is its period. If the period of external modulation T = 1/f or
time delay T0 is equal to τ, then it is possible to find feedback strengths η and δ
which allow stabilization of the UPO.
The delayed feedback control technique has been successfully used for
stabilization of UPOs in a Nd-doped fiber laser [102], a CO2 laser [10,103], a
semiconductor laser [104] and an optically pumped far-infrared gas laser
[105]. The efficiency of the delayed feedback method has been improved by
the adaptive control technique [106] that was implemented experimentally
with a CO2 laser [107,108].

4.2. Open-loop control

The open-loop control methods imply nonfeedback techniques which
consider the effect of external perturbations on evolution of the system. The
general scheme of this type of control is shown in Fig. 16.
Introduction to control of laser dynamics 29

Chaotic laser Measure state of the laser

Actuator

Figure 16. Schematic of open-loop control.

Usually, a harmonic modulation is applied to an available control parameter a

in the form

a = a0 [1 - m sin (2πfct)], (10)

where a0 is the initial value of the control parameter and m and fc are the
amplitude and frequency of control modulation.
In nonfeedback schemes, an orbit similar to the desired unstable state is
entrained by adjusting an accessible system parameter about its nominal
value by a weak periodic signal, usually in the form of a continuous
sinusoidal modulation [7,8]. Such a modulation is shown drastically alter the
dynamics of chaotic systems, leading eventually to stabilization of some
periodic orbits. This is somewhat simpler than feedback schemes because it
does not require real-time measurement of the state of the system and
processing of a feedback signal. The main advantage of these methods is that
they are independent on the knowledge of the actual dynamical state.
However, these approaches are limited, in general, by the facts that the
control perturbations are not sufficiently small and the final state cannot be
decided in advance. In the other words, periodic modulation fails in many
cases to entrain the UPO and its success or failure is highly dependent on the
specific form of the dynamical system.

A. Resonant parameter modulation

Generally, a resonant control means that the ratio between the control
frequency fc and a characteristic frequency of the system fx, is a rational
number, i.e., fc = (1/n) fx (n = 1,2,...). The efficiency of the nonfeedback
control depends strongly on the frequency of the control modulation. The
modifications of the method are different in autonomous and nonautonomous
systems. For the autonomous systems the characteristic frequency means the
relaxation oscillation frequency or the fundamental frequency of a system,
fx = fr, whereas for the nonautonomous systems the characteristic frequency
is the driving frequency, i.e. fx = fd. Similarly to the OGY method,
30 Alexander N. Pisarchik

autonomous chaotic systems can be locked to one of the unstable periodic

orbits by a parameter modulation with a frequency corresponding to the
frequency of an unstable periodic orbit imbedded into chaotic attractor. The
locking of a passive Q-switched chaotic CO2 laser to a small modulation of
the losses has been realized by Tsukamoto and his colleagues [103]. The
control modulation changes the width of periodic windows by shifting
bifurcation points. Recently, the locking of unstable periodic orbits has been
realized in a chaotic CO2 laser with feedback [104]. The main requirement
for this method is that the control frequency should get in resonance with the
selected orbit.
A principal feature of the resonant control systems is to make a system
bistable by splitting the original attractor into two new ones [90]. The
distance between these two attractors depends on the phase of the control
modulation. Thus, the resonant control is nothing neither more nor less than
the attraction of a system to the attractor with lower complexity, but not a
stabilization of an unstable periodic orbit. The resonant control has been
used by Chizhevsky and co-workers [105] to induce fast switching between
coexisting attractors.

B. Near-resonant control
A near-resonant control means a parameter modulation with the frequency
close to a subharmonic of the driving frequency, i.e. fc = (1/n)fd + δ (δ being a
small frequency detuning) whereas a nonresonant control (at fc ≠ nfd) requires a
relatively fast or a relatively slow parameter modulation. The terms “fast” and
“slow” mean that the frequency of the control modulation is much higher or
smaller than the driving frequency in nonautonomous systems. The near-
resonant control was used by Chizhevsky and Corbalán [106] to control
positions of bifurcation points in a CO2 laser.

C. Slow modulation control

The control by slow modulation can change drastically the global structure
of coexisting attractors, change the type of attractor, shift critical points [107],
and even destroy one or more undesired attractors by involving them to crisis
[31,65,66,89]. Moreover, slow modulation induces parametric resonances in a
period-doubled system. The last effect has been found numerically and
experimentally in a loss-driven CO2 laser [108].

4.3. Open-plus-closed-loop control

Recently a new type of control has been introduced [83]. It combines both the
open and close-loop control methods. The scheme of this type of control is shown
in Fig. 17.
Introduction to control of laser dynamics 31

Time delay

Chaotic laser Measure state of the laser

Actuator

Figure 17. Schematic of open-plus-closed-loop control with a time-delayed feedback laser.

It was shown [83] that in a simplest chaotic system dx/dt = Q(x,y), a time
delay in feedback F(t) = h[y(t-τ) – y(t)] can stabilize simultaneously several
UPOs which may coexist with the original chaotic attractor. Then, the external
control in the form of a harmonic modulation Eq. (10) applied to the feedback
strength locks one of these periodic orbits. Thus, it was demonstrated that
chaos can be controlled to a single periodic state.

“The reason for chaos is unmethodological

attempts to introduce an order.”

5. Recent trends and perspectives in laser dynamics

A huge growth of papers devoted to laser dynamics indicates on a growing
interest on this topic in scientific community. It is impossible to predict all
future trends in this highly developing area of optics. Here I shall distinguish
only the most important directions in the future development of this area of
science, based mainly on my own experience. Thus, I shall consider only the
following problems.

1. Control of bifurcations and chaos.

2. Control of multistability (green problem).
3. Application to communications.
4. Synchronization of array of lasers and controlling synchronization.
5. Control of intermittency.
6. Synchronization of multistable lasers.

5.1. Control of bifurcations and chaos

Nonlinear dynamics of many kinds of lasers requires further fundamental
investigations. Up to now, only few types of bifurcations were controlled in
laser systems, namely, period-doubling, Hopf, saddle-node, inverse period-
doubling, homoclinic bifurcations, and crisis points. The control of a large
number of other bifurcations has not yet been demonstrated in lasers. In some
32 Alexander N. Pisarchik

kind of lasers many dynamical effects has been studied only theoretically, e.g.
complex dynamics in a fiber laser with modulated losses [109]. Control of
polarization and spatial temporal dynamics as well as the control of Lorenz
chaos still remain theoretical problems [110]. Controlling chaos has not been
yet reported in a fiber laser. Complex dynamics, including generalized
multistability, in a fiber laser with modulated losses has been studied only
theoretically [109].

5.2. Control of multistability (green problem)

A crucial issue is connected with the important irregular intensity
fluctuations which appear in the generation of second harmonic with an
intracavity nonlinear device. The nonlinear couplings between modes that are
produced in nonlinear crystals give rise to irregular fluctuations in the optical
cavity. These fluctuations are amplified from the beginning by the quality factor
Q of the laser cavity and by the presence of the laser amplifier media. Strong
fluctuations arise then in the laser intensity. This is clearly a undesirable situation
for practical applications. To give an example of what discussed above, the
second harmonic generation (green light) in a diode-pumped intracavity doubled
Nd:YAG laser (1.06 µm) [3], is normally accompanied with strong intensity
fluctuations. The irregular behavior in this kind of systems was largely
investigated and attributed to the destabilization of relaxation oscillations, always
present in this kind of lasers, due to the nonlinear coupling of longitudinal
modes. One of the reasons of such a behavior is the coexistence of multiple
stable states (attractors) which appear normally in a system with many degrees of
freedom. The irregularity in the laser intensity results from involuntary switches
between the coexisting attractors.
One of the methods which allows one to avoid this kind of instabilities is
the method of "attractor annihilation" [65,66]. Attractor annihilation has been
realized theoretically in a CO2 laser [65] and in a semiconductor laser with
modulated injection current [83]. Recently, control of multistability has been
successfully realized experimentally in a diode-pumped fiber laser [17]. The
attractor annihilation method may be also applied for controlling multistability
in other types of lasers and in coupled lasers.
Multistability can be also controlled by noise. In this case, noise converts
the system with coexisting attractors to a system with a single, but intermittent
attractor when the trajectory visits randomly coexisting unstable states which
were stable without noise. Such noise-induced hoping between coexisting
states has been observed in lasers only theoretically [63].
I believe in the nearest future the control of multistability in a solid-state
laser with an intracavity frequency doubled crystal will help to solve the green
problem.
Introduction to control of laser dynamics 33

5.3. Application to communications

There is a simple connection between chaos and communication theory.
Chaotic systems can be viewed, indeed, as information sources that naturally
produce digital communication signals because they are characterized by
having positive entropies and thus they are information sources. By assigning a
discrete alphabet to the system state space using the formalism of symbolic
dynamics, the chaotic system becomes a symbol source, and because it is a
continuous-time wave form source, it is also a digital signal source. A chaotic
laser is, therefore, a natural source of digital communication signals.
Among the practical unlimited possible applications of the control of chaotic
behavior in lasers, herewith I concentrate on two applications, which have attracted
considerable attention in the scientific community over the past few years; namely
the control of chaotic behavior for communicating with chaos and for the
synchronization of chaotic systems for various communication schemes. In the
former case, a chaotic system is conveniently perturbed, in order to give rise to a
particular chaotic trajectory carrying a given message. In the latter case, the process
of chaos synchronization is applied to a communication line between a message
sender and a message receiver, allowing synchronization between them.

A. Digital communication
As early as 1994 Colet and Roy [109] have demonstrated possibilities for
application of synchronization of coupled lasers to digital communication with
chaotic single-mode Nd:YAG laser. In secure communication with chaotic lasers
the transmitted message should be encoded within the noise-like output of a
chaotic transmitter. Extraction of the message requires a receiver in which the
same chaos is generated as in the transmitter and hence the two lasers should be
synchronized. Experiments on communication with chaos have been carried out
in the same group, showing the possibility of encoding and decoding messages
with chaotic lasers [110], and the possibility of transmitting a desired message in
a very fast way using high-dimensional chaotic waveforms [111].
Recently, secure communication systems based on chaos in erbium-doped
fiber lasers were proposed and studied with message masking and chaos shift
keying [112]. Useful progress was made towards the experimental encryption
and decryption by use of a form of wavelength chaos [113]. A successful
demonstration of chaotic transmission of a message was reported recently using
a fiber laser [111] and external-cavity diode lasers [114,115]. It should noted that
among all lasers, diode lasers are most convenient for the purpose of
communication, because of the very high frequency of relaxation oscillations (in
the order of tens GHz). Therefore, chaotic optical encryption and decryption
demonstrated recently in synchronized external-cavity diode lasers [116] may
have very useful application for secure communication.
34 Alexander N. Pisarchik

High-speed modulation of Q-switched lasers may be of interest for developing

large capacity information transmission and ultrafast optical processing systems.
Such regimes can be useful for laser communications because they are
characterized by shorter pulse duration and higher amplitude in compare with
regular regimes. Moreover, the coexistence of different chaotic regimes may be
prominent for high secure communication by a chaotic on-off keying method [117]
if a proper switch between the chaotic attractors is organized.
Recent theoretical studies of anticipating synchronization of two chaotic
laser diodes by incoherent optical coupling [118] may be also interesting for
practical application in secure communication.

B. Oscillation death
Early it was thought that oscillation death (or quenching) has no analogy
in the case of periodic forcing and direct coupling. However, the oscillation
death phenomenon has been found numerically in nonautonomous coupled
lasers with modulated losses. The simulations have been carried out for a dual-
cavity CO2 laser [83]. The losses in one of the coupled cavities were
periodically modulated. The boundary conditions for the death state have been
theoretically evaluated [119]. These conditions are of fundamental importance
in communications with parametrically modulated coupled systems because
they indicate the dead region for information transmission by such systems.
Although the amplitude death has been recently demonstrated experimentally
in coupled lasers [85], the observation of a similar phenomenon in driven
lasers is still awaited.

C. Spatiotemporal communication
In most optical realizations of communications with chaos, the message to
be encoded drives the nonlinear transmitter, so that the message and carrier
become mixed in a nontrivial way. The resulting output is injected into a
receiver, which, upon synchronization to the transmitter, allows for recovery of
the original signal. The optical schemes developed so far have used purely
temporal chaotic signals as information carriers. Recently, a new nonlinear
optical device exhibiting spatiotemporal chaos as the basis of a communication
system has been proposed [120]. This device is capable of transmitting
information “in space and time”. Further development and experimental
realization of this potentially useful concept have to be awaited.

5.4. Synchronization of array of lasers and controlling

synchronization
The synchronization phenomenon, which occurs in systems without
control (self-synchronization), is well known and quite well studied. However,
Introduction to control of laser dynamics 35

far less attention was paid to synchronization as a control goal. Investigations

of the controlled synchronization have started only recently. Alternatively, the
slave laser can be performed as an array of lasers coupled with the master laser
[23]. Experimental synchronization of an array of many coupled lasers still
remains an open problem. This problem is very important for the array of
pulsed diode lasers in order to increase their combined power.
Many types of synchronization, e.g. complete synchronization, phase
synchronization of chaos, generalized synchronization, and oscillation death in
coupled laser systems, will be studied because all of them have significant
potential for application in optical communications. I believe that in the nearest
future the control of synchronized dynamical states will be realized in different
laser systems, including those with intracavity doubled crystal.

5.5. Control of intermittency

Closed-loop and open-loop control of on-off intermittency have been
demonstrated only theoretically by Nagai, et al. [121] and our group [122].
However, it was not still realized experimentally. Specifically, the objective
was to control the average length of the laminar phase ("off" state). For this
purpose, in closed-loop control a feedback signal, and in open-loop control an
external modulation, are applied to a control parameter near onset of
intermittency. The advantage of the open-loop control is that in experiments
one knows the driving signal well and hence the appropriate modulation
parameters can be computed and applied to the system in order to control the
mean laminar phase even without knowledge of an adequate theoretical model.
The control of intermittency is still waiting for realization in lasers.

5.6. Synchronization of multistable lasers

Synchronization of two systems which being isolated has coexisting attractors
has been demonstrated recently with electronic Rössler-like circuits [32,123].
Some types of lasers, e.g. semiconductor lasers with external cavity produce
complex dynamics which allows multistability, i.e. coexistence of steady-state,
periodic or chaotic attractors [124]. This makes these lasers good candidates for
studying synchronization phenomena in coupled multistable lasers. Moreover,
synchronization of semiconductor multistable lasers might be of interest for
communication if information could be encrypted in switching between coexisting
attractors. I guess special efforts will be directed to find out general relationships in
control methods and their applications in secure communications.

6. Summary
This short review introduced the reader to controlling laser dynamics and
synchronization. Only few achievements of laser dynamics were considered
36 Alexander N. Pisarchik

among a huge number of experimental and theoretical works. I concentrated

mainly on most important experimental works which demonstrated new
dynamical phenomena for the first time in laser systems. Finally I considered
only few trends in future development of this topic. Actually, a number of
works on controlling laser dynamics increases drastically. New achievements
in laser technology stimulate new research in laser dynamics that can help
engineers to solve some important technological problems. I hope this book
will bridge a gap in the literature and stimulate new discussions and
fundamental issues to deeper level of understanding the laser dynamics. The
results of this exercise might be also useful in definition of scientific and
technological programs related to controlling laser dynamics and
synchronization.

Acknowledgment
The author acknowledges support from CONACYT project No. 46973.

References
1. Arecchi, F. T., Meucci, R., Puccioni, G., and Tredicce, J. 1982, Phys. Rev. Lett.,
49, 1217.
2. Weiss, C. O., Godone, A., and Olafsson, A. 1983, Phys. Rev. A, 28, 369.
3. Baer, T. 1986, J. Opt. Soc. Am. B, 3, 1175.
4. Roy, R., Murphy, Jr., T. W., Maier, T. D., Gills, Z., and Hunt, E. R. 1992, Phys.
Rev. Lett., 68, 1259.
5. Ott E., Grebogi C., Yorke J. 1990, Phys. Rev. Lett., 64, 1196.
6. Pyragas, K. 1992, Phys. Lett. A, 170, 421.
7. Lima, R., and Pettini, M. 1990, Phys. Rev. A, 41, 726.
8. Braiman, Y., and Goldhirsch, I. 1991, Phys. Rev. Lett., 66, 2545.
9. Hunt, E. R. 1991, Phys. Rev. Lett., 67, 1953.
10. Bielawski, S., Derozier, D., and Glorieux, P. 1994, Phys. Rev. E, 49, R971.
11. Meucci, R., Gadomski, W., Ciofini, M., and Arecchi, F. T. 1994, Phys. Rev. E, 49,
R2528.
12. Ciofini, M., Meucci, R., and Arecchi, F. T. 1995, Phys. Rev. E, 52, 94.
13. Gills, Z., Iwata, C., Roy, R., Schwartz, I., and Triandaf, I. 1992, Phys. Rev. Lett.,
69, 3169.
14. Otsuka, K., Chern, J.-L., and Lih, J.-S. 1997, Optics Letters, 22, 292.
15. Hohl, A., and Gavrielides, A. 1998, Opt. Lett., 23, 1606.
16. Dykstra, R., Rayner, A., Tang D. Y., and Heckenberg, N. R. 1998, Phys. Rev. E,
57, 397.
17. Pisarchik, A. N., Barmenkov, Y. O., and Kir’yanov, A. V. 2003, Phys. Rev. E, 68,
066211.
18. Cuomo, M., and Oppenheim, A. V. 1993, Phys. Rev. Lett., 71, 65, 1993.
19. Luo, L. G., and Chu, P. L. 1998, J. Opt. Soc. Am. B, 15, 2524.
20. Pecora, L. M., and Carroll, T. L., 1990, Phys. Rev. Lett., 64, 821.
Introduction to control of laser dynamics 37

21. Roy, R., and Thornburg, K. S. 1994, Phys. Rev. Lett., 72, 2009.
22. Sugawara, T., Tachikawa, M., Tsukamoto, T., and Shimizu, T. 1994, Phys. Rev.
Lett., 72, 3502.
23. Terry, J. R., Thornburg, Jr., K. S., DcShazer, D. J., Van-Waggeren, G. D., Zhu, S.,
Ashwin, P., and Roy, R. 1999, Phys. Rev. E, 59, 4036.
24. Möller, M., Forsmann, B., and Lange, W. 1998, Quantum Semiclass. Opt., 10, 839.
25. Uchida, A, Shinozuka, M., Ogawa, T., and Kannari, F. 1999, Opt. Lett., 24, 890.
26. Mirasso, C. R., Colet, P., and Fernandez, P. G. 1996, IEEE Photon. Technol. Lett.,
8, 299.
27. Klimenko, D. N., Kravtsov, N. V., Lariontsev, E. G., and Firsov, V. V. 1997, Sov.
J. Quantum Electron., 24, 649.
28. Barsella, A., Lepers, C., Dangoisse, D., Glorieux, P., and Erneux, T. 1999, Opt.
Commun., 165, 251.
29. Luo, L., Chu, P. L., Whitbread, T., and Peng, R. F. 2000, Opt. Commun., 175, 213.
30. Otsuka, K., Kawai, R., Hwong, S.-L., Ko, J.-Y., and Chern, J.-L. 2000, Phys. Rev.
Lett., 84, 3049.
31. Pisarchik, A. N., Barmenkov, Y. O., and Kir’yanov, A.V. 2003, IEEE J. Quantum.
Electron., 39, 1567.
32. Pisarchik, A. N., Jaimes-Reátegui, R., Villalobos-Salazar, J. R., Garcia-Lopez, J.
H., and Boccaletti, S. 2006, Phys. Rev. Lett., 96, 244102.
33. Arecchi, F. T., Lippi, G. L., Puccioni, and G. P., Tredicce, J. R. 1984, Opt.
Commun., 51, 308.
34. Weiss, C. O., Klische, W., Ering, P. S., and Cooper, M. 1985, Opt. Comm. 52, 405.
35. Eckmann, J.-P. 1981, Rev. Mod. Phys., 53, 643.
36. Biswas, D. J., and Harrison, R. G. 1985, Phys. Rev. A, 32, 3835.
37. Pomeau, Y., and Manneville, P. 1980, Comm. Math. Phys., 74, 189.
38. Spiegel, E. A. 1981, Ann. N.Y. Acad. Sci., 617, 305.
39. Fujisaka, H., Kamifukumoto, H., and Inoue M. 1982, Prog. Theor. Phys., 69, 333.
40. Ott, E. 1993, Chaos in Dynamical Systems, Cambridge University Press.
41. Brun, E., Derighetti, B., Meier, D., Holzner, R., and Ravani, M. 1985, J. Opt. Soc.
Am. B, 2, 156.
42. Dangoisse, D., Glorieux, P., and Hennequin, D. 1986, Phys. Rev. Lett., 57, 2657.
43. Biswas, D. J., Dev, V., and Chatterjee, U. K. 1987, Phys. Rev. A, 35, 456.
44. Sacher, J., Elsässer, W., and Göbel, E. O. 1989, Phys. Rev. Lett., 63, 2224.
45. Tang, D. Y., Pujol, J., and Weiss, C. O. 1991, Phys. Rev. A, 44, R35.
46. Ditto, W. L., Rauseo, S., Cawley, R., Grebogi, G. H., Hsu, C., Kostelich, E., Ott,
E., Savage, H. T., Segnam, R., Spano, M. L., and Yorke, J. A. 1989, 63, 923.
47. Pisarchik, A. N., and Pinto-Robledo, V. J. 2002, Phys. Rev. E, 66, 027203.
48. Arecchi, F. T., Meucci, R., and Gadomski, W. 1987, Phys. Rev. Lett., 58, 2205.
49. Shilnikov, L. P. 1965, Dokl. Akad. Nauk SSSR, 160, 558.
50. Arecchi, F. T., Degiorgio, V., and Querzola, B. 1967, Phys. Rev. Lett., 19, 1168.
51. Hennequin, D., Tomasi, F. de, Zambon, B., and Arimondo, E. 1988, Phys. Rev. A,
37, 2243.
52. Viktorov, E. A., Klimer, D. R., Karim, M. A. 1995, Opt. Commun., 113, 441.
53. Pisarchik, A. N., Meucci, R., and Arecchi, F. T. 2000, Phys. Rev. E, 62, 8823.
54. Pisarchik, A. N., Meucci, R., and Arecchi, F. T. 2001, Eur. Phys. J. D, 13, 385.
38 Alexander N. Pisarchik

55. Lauterborn, W., and Steinhoff, R. 1988, J. Opt. Soc Am. B, 5, 1097.
56. Saucedo-Solorio, J. M., Pisarchik, A. N., Kir'yanov, A. V., and Aboites, V. 2003,
J. Opt. Soc. Am. B, 20, 490.
57. Hohl, van der Linden, H., Roy, R., Goldsztein, G., Broner, F., and Strogatz, S.
1996, Phys. Rev. Lett., 74, 2220.
58. Wieczorek, S., Simpson, T., Krauskopf, B., and Lenstra, D. 2002, Phys. Rev. E,
65, 045207R.
59. Reategui, R. J., Kir’yanov, A. V., Pisarchik, A. N., Barmenkov, Y. O., and
Il’ichev, N. N. 2004, Laser Physics, 14 (10), 1277.
60. Lang, R., and Kobayashi, K. 1980, IEEE J. Quantum Electron. 16, 347--355.
61. Masoller, C., and Abraham, N. 1998, Phys. Rev. A, 57, 1313.
62. Krauskopf, B. 2006, in Trends in Dynamical Systems, F. Dumortier, D. Roose, and
A. Vanderbauwhede (Eds.), Handelingen van de Koninklijke Vlaamse Academie
van Belgie, pp. 33-45.
63. Masoller, C. 2002, Phys. Rev. Lett., 88, 034102.
64. Pyragas, K., Lange, F., Letz, T., Parisi, J., and Kittel, A. 2000, Phys. Rev. E, 61,
3721.
65. Pisarchik, A. N., and Goswami, B. K. 2000, Phys. Rev. Lett., 84, 14236.
66. Pisarchik, A. N. 2001, Phys. Rev. E, 64, 046203.
67. Lugiato, L. A., and Milani, H. 1983, Opt. Commun., 46, 57.
68. Tamm, C. 1988, Phys. Rev. A., 38, 5960.
69. Lippi, G. L, Abraham, N. B., Puccioni, G. D., Arecchi, F. T., and Tredicce, J. R.
1987, Phys. Rev. A, 35, 3978.
70. White, J. K, and Moloney, J. V. 1999, Phys. Rev. A, 59, 2422.
71. Brambilla, M., Lugiato, L. A., Penna, V., Prati, F., Tamm, C., and Weiss, C. 1991,
Phys. Rev. A, 43, 5114.
72. Brambilla, M., Lugiato, L. A., Pinna, M., Prati, F., Pagani, P., Vanotti, V., Li, M.,
and Weiss, C. 1992, Opt. Commun., 92, 145.
73. Pampaloni, E., Residor, S., Soria, S., and Arecchi, F. 1997, Phys. Rev. Lett., 78,
1042.
74. Van Wiggeren, G. D., and Roy, R. 1999, Opt. Commun., 164, 107.
75. Kulminskii, A., Vilaseca, R., and Corbalan, R. 1995, Optics Letters, 20 (23), 2390.
76. Leyva, I., Allaria, E., and Meucci, R. 2003, Phys. Rev. A 68, 053806.
77. Erneux, T., Rogister, F., M´egret, P., Deparis, O., and Blondel, M. 1999, Opt. Lett.,
24, 1218.
78. Rogister, F., Sukow, D. W., Gavrielides, A., M´egret, P., Deparis, O., and Blondel,
M. 2000, Opt. Lett., 25, 808.
79. Erneux, T., Rogister, F., Gavrielides, A., and Kovanis, V. 2000, Opt. Commun.,
183, 467.
80. Sukow, D. W., Hegg, M. C., and Wright, J. L. 2002, Opt. Lett., 27, 827.
81. Ruiz-Oliveras, F. R., and Pisarchik, A. N. 2006, Optics Express, 14 (26) 12859.
82. Risch, C., and Voumand, C. 1977, J. App. Phys., 48, 2083 (1977).
83. Kuntsevich, B. F., and Pisarchik, A. N. 2001, Phys. Rev. E, 64, 046221.
84. Herrero, R., Figueras, M., Rius, J., Pi, F., and Orriols, G. 2000, Phys. Rev. Lett.,
84, 5312.
85. Wei, M.-D., and Lun, J.-C. 2007, Appl. Phys. Lett., 91, 061121.
Introduction to control of laser dynamics 39

86. Reyl, C., Flepp, L., Badii, R., and Brun, E. 1993, Phys. Rev. E, 47, 267.
87. Chin, G., Senesac, L. R., Blass, W. E., and Hillman, J. J. 1996, Science, 274, 1498.
88. Pyragas, K. 1995, Phys. Lett. A, 206, 323.
89. Simmendinger, C., and Hess, O. 1996, Phys. Lett. A, 216, 97.
90. Martínez-Zérega, B. E., Pisarchik, A. N., and Tsimring, L. 2003, Phys. Lett. A,
318, 102.
91. Martínez-Zérega, B. E., and Pisarchik, A. N. 2005, Phys. Lett. A, 340 (1-4), 212.
92. Pisarchik, A. N., Corbalán, R., Chizhevsky, V. N., Vilaseca, R., and Kuntsevich,
B. F. 1998, Int. J. Bifurcation and Chaos, 8, 1783.
93. Pisarchik, A. N. 1998, Phys. Lett. A, 242, 152.
94. Pisarchik, A. N., Kuntsevich, B. F., and Corbalán, R. 1998, Phys. Rev. E, 57, 4046.
95. Chizhevsky, V. N., and Glorieux, P. 1995, Phys. Rev. E, 51, R2701.
96. Pisarchik, A.N., and Corbalán, R. 2001, Physica D, 150, 14.
97. Chizhevsky, V. N., Corbalán, R., and Pisarchik, A. N. 1997, Phys. Rev. E 56,
1580.
98. Carr, T. W, and Schwartz, I. B. 1994, Phys. Rev. E, 50, 3410.
99. Carr, T. W, and Schwartz, I. B. 1995, Rev. E , 51, 5109.
100. Schwartz, I. B., and Triandaf, I. 1992, Phys. Rev. A, 46, 7439.
101. Carroll, T., Triandaf, I., Schwartz, I.B., and Pecora, L. 1992 Phys. Rev. A 46,
6189.
102. Bielawski, S., Bonazacui, M., Derozier, D., and Glorieux, P. 1993, Phys. Rev. A,
47, 3276.
103. Erneux, T., Bielawski, S., Derozier, D., and Glorieux, P. 1995, Quantum
Semiclass. Opt., 7, 951.
104. Simmendinger, C., and Hess, O. 1996, Phys. Lett. A, 216, 97.
105. Dykstra, R., Tang, D. Y., Hechenberg, N. R. 1998, Phys. Rev. E, 57, 6596.
106. Boccaletti, S., and Arecchi, F. T. 1995, Europhys. Lett., 31, 127.
107. Ciofini, M., Labate, A., Meucci, R., and Arecchi, F. T. 1999, Eur. Phys. J. D, 7, 9.
108. Arecchi, F. T., Meucci, R., Allaria, E., Di Garbo, A., and Tsimring, L. S. 2002,
Phys. Rev. E, 65, 046237.
109. Colet, P., and Roy, R. 1994, Opt. Lett., 19, 2056.
110. Alsing, P. M., Gavrielides, A., Kovanis, V., Roy, R., and Thornburg, K. S. 1997,
Phys. Rev. E, 56, 6302.
111. Van Wiggeren, G. D., and Roy, R. 1998, Phys. Rev. Lett., 81, 3547.
112. Luo, L. G., and Chu, P. L. 1998, J. Opt. Soc. Am. B, 15, 2524.
113. Goedgebuer, J. P., Larger, L., and Porte, H. 1998, Phys. Rev. Lett., 80, 2249.
114. Sivaprakasam, S., and Shore, K. A. 1999, Opt. Lett., 24, 1200.
115. Sivaprakasam, S., and Shore, K. A. 2000, IEEE J. Quantum Electron., 36, 35.
116. Sivaprakasam, S., and Shore, K.A. 2000, Phys. Rev. E, 61, 5997.
117. Uchida, A., Yoshimori, S., Shinozuka, M., Ogawa, T., Kannari, F. 2001, Optics
Letters, 26, 866.
118. Rogister, F., Pieroux, D., Sciamanna, M., Mégret, P., and Blondel, M. 2002, Opt.
Commun., 207, 295.
119. Pisarchik, A. N. 2003, Phys. Lett. A, 318, 65.
120. García-Ojalvo, J., and Roy, R. 2002, Phys. Rev. Lett., 66, 017206.
121. Nagai, Y., Hua, X.-D., and Lai Y.-C. 1996, Phys. Rev. E, 54, 1190.
40 Alexander N. Pisarchik

122. Reategui, R. J., and Pisarchik, A. N. 2004, Phys. Rev. E, 69, 067203.
123. Pisarchik, A. N., Jaimes-Reátegui, R., and García-Lopez, J. H. 2008, Phil. Trans.
Roy. Soc., Ser. A, 366, 459.
124. Ruiz-Oliveras, F. R., and Pisarchik, A. N. 2006, 14 (26), 12859.