Optics Communications 215 (2003) 135–151

www.elsevier.com/locate/optcom

Mapping the nonlinear dynamics of a distributed

feedback semiconductor laser subject to external
optical injection
T.B. Simpson *
Jaycor/Titan, P.O. Box 85154, San Diego, CA 92186-5154, USA
Received 13 August 2002; received in revised form 11 November 2002; accepted 12 November 2002

Abstract

Single-mode semiconductor lasers can exhibit stable, bistable, periodic, quasiperiodic, and chaotic output charac-
teristics when subjected to monochromatic, near-resonant external optical injection. Experimental measurements of the
spectral characteristics of a distributed feedback laser diode under external optical injection identify the various
nonlinear output characteristics. Single-mode operation is maintained in all cases. The output characteristics are
mapped as a function of the strength of the optical injection and the detuning between the injection frequency and the
free-running frequency of the unperturbed laser. A comparison of the observed characteristics with the predictions of a
coupled rate-equation model shows quantitative agreement over a wide range of injection characteristics. The model
shows that the mode hopping previously observed in Fabry–Perot laser diodes subject to external optical injection can
be explained using a simple coupling of the longitudinal modes through the free carriers of the gain medium.
Ó 2002 Elsevier Science B.V. All rights reserved.

PACS: 42.65.Sf; 42.55.Px; 05.45.Pq

Keywords: Semiconductor lasers with optical injection; Nonlinear dynamics in semiconductor lasers

1. Introduction and laser pumping, but also in their usefulness as a

system for understanding the fundamental char-
Lasers made with a semiconductor gain medium acteristics of lasers. The semiconductor laser is a
are among the most important and widely used nonlinear system where quantitative comparisons
today. Their importance lies not only in their many can be made between experiment and theory. It is
applications, from telecommunications and data/ well established that a variety of external pertur-
information processing to environmental sensors bations such as optical injection from an external
source, modulation of the pumping current, ex-
ternal optical feedback from a distant reflector, and
*
Tel.: +1-8587204081; fax: +1-8587204155. electro-optic feedback from a distant photodetec-
E-mail address: [email protected] tor that is added to the bias current, can cause the

0030-4018/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 0 3 0 - 4 0 1 8 ( 0 2 ) 0 2 1 9 2 - 2
136 T.B. Simpson / Optics Communications 215 (2003) 135–151

output of the laser to become unstable and yield perimental observations and model calculations
oscillatory or chaotic output [1]. Of the perturba- has been made where excellent agreement between
tion mechanisms, external optical injection allows experimental data and model calculations was
the greatest degree of control and most direct observed [8]. Previous mappings of the experi-
access to the fundamental mechanisms that lead to mentally observed characteristics have been lim-
the instability. With external feedback, one is lim- ited by the observation of mode hops to new
ited to perturbing the laser with its own output. longitudinal modes [9]. These mode hops tended to
Current modulation and electro-optic feedback can occur in regions where the model predicts a rich
be limited by the fact that semiconductor lasers are array of nonlinear dynamics in the laser output.
circuit elements with inherent capacitance, induc- Further, new theoretical tools have been used re-
tance and resistance. Additional parasitic elec- cently to extract many of the details of the dy-
tronic effects arise when the gain medium is namics of the semiconductor laser under optical
connected into an electronic circuit. The determi- injection [10]. Equally detailed and precise exper-
nation of the mechanisms controlling laser per- imental tools are needed to quantitatively compare
formance can involve a deconvolving of the circuit experiment and theory.
effects with the underlying processes within the la- Here, we present experimental details behind
ser cavity and gain medium. External optical in- the mapping of the nonlinear dynamics of a dis-
jection allows one to perturb the laser with an tributed feedback (DFB) laser diode that main-
independent optical field. The optical injection tains single mode operation under external optical
bypasses the electronic connections and yields di- injection. The mapping is generated through the
rect information about the laser cavity and gain analysis of the optical, regenerative amplification,
medium. In addition external optical injection has and power spectra of the laser subject to external
been proposed and demonstrated as a means to optical injection. This paper emphasizes the ex-
enhance the bandwidth and reduce the noise and perimental techniques used to generate the map-
distortion characteristics of semiconductor lasers ping and compares the results with calculations
[2], generate tunable, microwave modulation of an based on coupled equations with experimentally
optical carrier through double-locked optical out- determined dynamic parameters. We have refined
put [3], and synchronize chaotic dynamics for and augmented past techniques used to generate
possible use in communication systems [4]. the spectra and to determine the key dynamical
Past experimental work on the semiconductor parameters of the coupled equation model. We
laser subject to external optical injection has es- observe very good quantitative agreement between
tablished that a range of optical outputs can be observed and calculated operating characteristics.
produced. For a laser operating at a given bias This paper is organized as follows. Following
current, external optical injection can induce stable this introductory section, the experimental appa-
locked output, multiwave mixing, oscillatory ratus is described. A section describing the ob-
power output due to an undamping of the carrier- served spectra, and the new, more complete
field resonance of the system, and chaotic dy- mapping of the dynamics observed then follows.
namics depending on the amplitude and offset The next section reviews the model of a nearly
frequency of the external optical field [5]. It has single-mode semiconductor laser subject to exter-
been found that a relatively simple model de- nal optical injection, highlighting some new fea-
scribing the coupling between the external and tures. The characterization of the laser using weak
circulating optical fields and the free carriers of the current modulation and external optical injection
gain medium can reproduce the characteristics for determining the model parameters is described.
observed experimentally [6]. Because the key dy- This allows a quantitative comparison between the
namic parameters of the model can be determined model and the experimental data. The results of
experimentally, a quantitative comparison can be model calculations are summarized and show the
made between experimental data and model cal- good agreement with experimentally observed
culations [7]. Recently, a full comparison of ex- data. Also included are numerical results that
 T.B. Simpson / Optics Communications 215 (2003) 135–151 137

point to a mechanism for the mode hop observed was below 1% (experimental uncertainty) of the
in past experiments on Fabry–Perot lasers. We total optical power. Further, under optical injec-
conclude with a brief summary of key results. tion of the principal mode, the optical power in the
side modes never deviated by more than 10%
(again, experimental uncertainty) from its free-
2. Experimental apparatus running value. Therefore, this laser is effectively a
single-mode laser for modeling purposes. For the
Fig. 1 is a schematic of the experimental appa- measurements of the nonlinear dynamics, the laser
ratus used to perform these measurements. Three was operated under dc-current bias. However, a
DFB lasers under independent temperature and high speed current modulation from a frequency
current control are used. All are temperature and synthesizer was added to the dc bias for deter-
current tuned to oscillate at approximately the mining key laser parameters used in the modeling.
same wavelength, 1557 nm, and all are coupled to This will be discussed further below.
single-mode optical fibers. The laser under inves- The apparatus is fiber-coupled, except for an
tigation, the slave laser, is an Alcatel 1915 LMO acousto-optic modulator that is used to generate a
DFB laser package that contains no internal op- replica of the probe output that is frequency shif-
tical isolator. The master laser, used as the optical ted by 80 MHz. It was necessary to use angled
injection source to induce the nonlinear dynamics, physical contact (APC) type connectors for any
and the probe laser, used as a weak optical probe connector not isolated from a laser emitter. The
of the slave laser and/or as a local oscillator for )30 to )40 dB backscatter from the standard
heterodyne measurements of the optical output of physical contact (PC) connectors produced unac-
the slave laser, are both Alcatel 1915 LMI DFB ceptable perturbation of the laser output. The
lasers that are packaged with internal optical iso- master laser is coupled into the slave laser using an
lators. Both probe and master lasers are operated optical circulator after having passed through a
at current levels much farther from threshold than variable attenuator and a polarization rotator. The
the slave laser so that, under the operating con- latter is used to align the optical polarization of the
ditions used here, their noise characteristics can be master with that of the slave. The frequency-shif-
ignored and they can be considered to be essen- ted replica of the probe laser can also be injected
tially monochromatic sources. The optical power into the slave laser after being coupled into the
in all the side modes of the free-running slave laser master laser fiber path with a fiber-optic coupler.

Fig. 1. Schematic of the experimental apparatus. Thin solid lines denote connections via single-mode optical fiber while the thick solid
lines around the acousto-optic (A/O) modulator denote free-space propagation. The dashed line is a microwave coaxial connection.
Arrows denote connections after optical isolation so that propagation is blocked in the opposite direction. The optical circulator
permits coupling only from port 1 to port 2 and from port 2 to port 3. All three laser are under independent bias current and tem-
perature control (not shown). A modulation current could also be added to the dc bias of the slave laser.
138 T.B. Simpson / Optics Communications 215 (2003) 135–151

The output of the slave laser is coupled out of the lating a particular output dynamics to the spectral
optical circulator where it can be combined with features displayed at a given injection level and
the main output of the probe laser. A fast photo- offset frequency [9]. In the optical and regenerative
diode is used to detect the optical signal and its amplification spectra presented here, offset fre-
output is directed to a microwave spectrum ana- quency is relative to the free-running optical fre-
lyzer. quency of the slave laser and signal power as a
Three types of spectra are generated with this function of offset frequency is shown on a loga-
apparatus. If the probe laser is turned off, then the rithmic scale. Most of the spectra show data from
detected signal is a measure of the amplitude two scans overlaid. Because there is some jitter
modulation of the slave laser and the spectrum during a scan, the spectral features do not precisely
analyzer measures the amplitude (power) spec- overlap. For the results presented here, the slave
trum. If the probe laser is on, but the frequency laser was operated at twice the threshold current
shifted output of the acousto-optic modulator is where the relaxation resonance frequency of the
blocked, then the probe laser acts as a local os- free-running laser is approximately 4.7 GHz. More
cillator for a heterodyne measurement of the op- details of the operating parameters in relation to
tical spectrum of the slave laser. To generate a the coupled equation model are given below.
spectrum, we fix the frequency of the microwave Fig. 2 shows the transition from free-running
spectrum analyzer at some convenient frequency, output to limit cycle dynamics as the injected sig-
typically 33 MHz, and then ramp the bias current nal is increased when the master laser is tuned to
to the probe laser. Ramping the bias current to the the free-running optical frequency of the slave la-
probe laser shifts its optical frequency. A low- ser. Optical spectra are shown in the left column
resolution optical spectrum is generated as the and regenerative amplification spectra are shown
probe laser sweeps across the frequency compo- in the right column. Both the free-running optical
nents of the slave laser output. For these mea- and regenerative amplification spectra show the
surements, a spectral resolution of 100 MHz is weak, broad features associated with the relaxa-
sufficient, indeed finer resolution was generally not tion resonance frequency. At relatively low injec-
possible due to fluctuations in the optical fre- tion levels, the laser becomes locked to the master
quencies of the three lasers. Finally, the frequency- signal and the weak probe beam no longer is
shifted output of the probe laser could be injected strongly amplified at the formerly free-running,
simultaneously into the slave laser. If the spectrum now locked, optical frequency. As the injected
analyzer is set to match the 80 MHz frequency optical signal is increased, the damping on this
offset with a very narrow resolution bandwidth, resonance decreases and the sidebands become
typically 10 kHz as compared to 3 MHz with the stronger until the Hopf bifurcation is passed and
other measurements, then a spectrum of the re- the laser output becomes unstable. Note that the
generative amplification of the probe beam in the sideband components grow in magnitude and
slave laser is generated. These three spectra yield sharpen in linewidth, with no pedestal, as the in-
complementary information about the output jection level is increased. Because there were no
characteristics of the slave laser, both free-running abrupt spectral changes to mark the Hopf bifur-
and under external optical injection. cation, criteria had to be developed to define a
change in dynamics. We observed that in the re-
generative amplification spectra a narrow dip and
3. DFB laser nonlinear dynamics peak structure developed at the master laser in-
jection frequency. It is clearly visible at the highest
The fiber-coupled configuration proved to be injection level and is barely showing at the next
robust and stable and we were able to make a lower injection level. The peak and dip reach a
detailed mapping of the nonlinear dynamics as a maximum in between these two injection levels.
function of the detuning and amplitude of the in- Recall that the regenerative amplification spec-
jected optical field. The mapping is made by re- trum measures the small signal optical gain as a
 T.B. Simpson / Optics Communications 215 (2003) 135–151 139

Fig. 2. Spectra with injection at the free-running optical frequency showing the transition from stable to limit cycle dynamics. The left
column contains the optical spectra and the right column consists of the regenerative amplification spectra. The Injection Field is
proportional to the square root of the injected power. The free-running regenerative amplification spectrum has been corrected for
background noise and the variation of local oscillator power during the scan so that it can be compared to the model calculation,
shown as a thin line.
140 T.B. Simpson / Optics Communications 215 (2003) 135–151

function of offset frequency. A dip and peak not the case for an offset frequency of )2 GHz
around an injection feature indicates that the where the enhancement of the relaxation reso-
perturbing probe is having a strong effect on the nance features is considerably less and a transition
spectral feature locked to the strong signal from to stable, locked operation is observed.
the master laser. Because it is occurring as the As the injection level is further increased, the
optical spectrum is showing a change from weakly multiwave mixing features are pulled to the injec-
damped relaxation resonances to simultaneous tion frequency and the features associated with the
oscillation at several optical frequencies, we in- relaxation resonance sharpen and strengthen. In
terpreted the maximum in the peak and dip as an the locked output, shown in the Injection
indicator of the Hopf bifurcation point. Associ- Field ¼ 0.41 spectra, the frequency-pulled multi-
ated with this transition were sidebands approxi- wave mixing components have disappeared and
mately 15–20 dB below the resonant, locked there are narrow, strong spectral components with
feature in the optical spectrum. Therefore, we used pedestals at the relaxation resonance/oscillation
the criterion that a spectral feature exhibited in the frequency. The spectrum shows a laser that is only
optical spectra had to be no more than 20 dB be- marginally stable, or marginally unstable. As the
low the strongest spectral feature for its associated injection level is further increased these pedestals
dynamics to be the characteristic dynamics of the broaden to show weak spectral features around the
system. This choice is made to have a consistent strongest spectral peaks. These new features are
quantitative measure relating the changes in the clearly evident in the Injection Field ¼ 0.52 spectra
observed spectra to the dynamics. and indicate that another incommensurate fre-
This criterion can be used to analyze the dy- quency is present in the dynamics. This is strong
namics associated with more complex spectra. Fig. evidence for the existence of quasiperiodic dy-
3 shows optical and amplitude spectra in the left namics that have been predicted from numerical
and right columns, respectively, at increasing op- calculations [12].
tical injection levels when the master laser is offset As the injection power is further increased, the
by 2 GHz. At the weakest injection levels, the weak features about each strong spectral peak
optical spectrum shows side peaks due to regen- become relatively stronger, broader and shift in
erative amplification of the master laser signal and offset as the injection power is increased Eventu-
four-wave and multiwave mixing between the ally, the broad continuum grows until only a few
master laser and unlocked slave laser frequencies, spectral features are evident. This spectrum indi-
as shown in the Injection Field ¼ 0.14 spectra. The cates that a transition to chaotic dynamics has
multiwave mixing becomes more pronounced as probably occurred. Past calculations have shown
the injection level is increased and mixing peaks that spectra like those at the Injection Field ¼ 0.77
appear as sharp spectral features in the amplitude are associated with chaotic dynamics [13]. There-
modulation spectra. Simultaneously, there is en- fore, the laser has made a transition to chaotic
hancement of the broad features associated with dynamics through a quasiperiodic route, as op-
the relaxation resonance frequency of the free- posed to the period doubling route [13], or abrupt
running laser, particularly in the amplitude spectra saddle-node bifurcation [8], previously observed.
where it quickly becomes the dominant spectral The amplitude spectrum also shows some evidence
component, as shown in the Injection Field ¼ 0.33 for period-three dynamics as well. Other spectra
spectra. The dominance of the multiwave mixing with significant broadening also show different
components in the optical spectra and their rela- periodic motion. The spectra at an Injection
tive weakness in the amplitude spectra show that Field ¼ 1.02 show broadened period-two features.
these features arise primarily from phase modula- At still higher injection levels, the laser makes a
tions of the laser field, consistent with Adler-type transition back to strong, narrow spectral features
frequency pulling towards locking [11]. However, as shown in the Injection Field ¼ 1.3 spectra.
here the frequency pulling is accompanied by a Again, the narrow features have a pedestal, the
destabilization of the relaxation resonance. This is remnants of the higher-order periodic features
 T.B. Simpson / Optics Communications 215 (2003) 135–151 141

Fig. 3. Spectra with injection at +2 GHz relative to the free-running optical frequency showing a range of output characteristics. The
left column contains the optical spectra and the right column consists of the amplitude (power) spectra. The amplitude spectra signal is
relative to the background noise level. The Injection Field is proportional to the square root of the injected power.
142 T.B. Simpson / Optics Communications 215 (2003) 135–151

Fig. 3. (continued)
 T.B. Simpson / Optics Communications 215 (2003) 135–151 143

from the region of more complex dynamics. Fur- stable or periodic dynamics no longer occurs. Re-
ther increase in the injected power eventually sults are given in a companion paper [14].
yields a simple limit cycle spectra, dominated by Fig. 4 is a mapping of the observed dynamics
two strong optical field components, as shown in based on interpretation of spectra like those de-
the Injection Field ¼ 3.01. Note that now the scribed above. The abscissa of the map is propor-
characteristic frequency of the limit cycle has in- tional to the square root of the injected optical
creased well above the free-running relaxation power, the injected field amplitude, and the ordi-
resonance frequency. At these high injection levels nate is the offset frequency of the master laser with
the resonance frequency is a function of the in- respect to the free-running frequency of the slave
jection level and offset frequency of the master laser. Boundaries of the regions of stable and pe-
laser. At still higher injection levels, the laser can riodic dynamics are shown. Diamonds mark the
be expected to reestablish stable, locked operation. saddle-node boundary between stable locked and
However, we feared that we might damage the unlocked dynamics, squares mark the Hopf bifur-
laser and did not attempt to observe this transition cation boundary between stable locked and limit
at this offset frequency. We did observe it at offset cycle dynamics, triangles mark the boundaries of
frequencies between )4 and 1.5 GHz. regions of period two dynamics and circles mark
The clear observation of spectra consistent with the boundary of period four dynamics. The three
quasiperiodic motion in the spectral characteristics
of Fig. 3 confirms theoretical predictions and nu-
merical calculations [12]. This dynamics occurs
under optical injection conditions where the pre-
viously studied Fabry–Perot lasers were induced to
hop to a different longitudinal mode [9]. As we will
show below in the numerical calculations, the
quasiperiodic dynamics occurs with a slight de-
crease in the average power level. This allows one
of the other longitudinal modes of a conventional
edge-emitting, Fabry–Perot laser to become the
dominant mode. With a DFB laser, like the one
used here, the loss for each of the side modes is
sufficiently enhanced so that none achieve signifi-
cant power. Therefore, only single-mode output is
observed.
We have combined spectra such as those in
Figs. 2 and 3 to produce a mapping of the observed
Fig. 4. Mapping of the observed dynamics based on the ex-
dynamics. While spectra that show stable or peri- perimentally measured spectra of the DFB laser under optical
odic dynamics can be interpreted in a straightfor- injection when biased at twice the threshold current. The in-
ward manner, the quasiperiodic and chaotic jection field is proportional to the square root of the injected
spectra are more difficult to distinguish. The pres- optical power. The full diamonds mark the saddle node bifur-
ence of spontaneous emission noise and, poten- cation between stable and unlocked–locked operation while the
open diamonds mark the unlocking–locking transition in a re-
tially, multiple competing attractors means that gion of bistability (see text). Squares mark the Hopf bifurcation
broadening is not enough to distinguish between between stable and limit cycle operation. The triangles bound
periodic motion on multiple attractors, quasiperi- regions of period two operation. Within the period two regions
odic motion with broadening due to noise and true are regions of complex dynamics marked by the shaded lines
and crosses and a region of period four operation that is
chaotic dynamics. Therefore, sophisticated mod-
bounded by the circles. At injection levels below the saddle
eling tools are required to determine the details of node bifurcation line and at low offset frequencies, multiwave
the dynamics. Here, we will not attempt to resolve mixing and Adler-type frequency pulling to locking are ob-
dynamics of the complex spectra that show that served in the lightly shaded regions.
144 T.B. Simpson / Optics Communications 215 (2003) 135–151

regions with shaded lines mark the areas where injection coupling rate, J =ed is the carrier density
more complex dynamics are observed. The two injection rate due to the injection current density J,
positive offset regions were observed in previous cs is the carrier spontaneous decay rate, and e0 n2 is
mappings of the dynamics of a Fabry–Perot laser the non-resonant (real) dielectric constant of the
diode under external optical injection. However, a semiconductor gain medium. When free-running,
significant amount of optical power was transferred the laser oscillates at an optical frequency x0 with
to other longitudinal modes in the past measure- a steady-state field amplitude A0 and a steady-state
ments, and mode hopping obscured the negative carrier density N0 . The complex gain, gðN ; jAjÞ, is
offset region. With the DFB laser, the power, >99%, assumed to vary about the free-running, steady-
remained in the principal mode. There is also a re- state operating point as a linear function of the
gion of bistability between locked and unlocked carrier density and circulating intensity:
dynamics that is observed at negative offsets and gðN ; jAjÞ ¼ ð1  ibÞ

gðN0 ; A0 Þ
relatively high injection levels. This bistable region 
has been observed previously with DFB lasers [15], g 
o

þ ð1  ibÞ  ðN  N0 Þ
and explained as being due to a crossing of the oN 
N0 ;A0
saddle node bifurcation by a torus bifurcation [16]. 
The open diamonds mark the transition from un- 0 o
g  2
þ ð1  ib Þ 2
ðjAj  A20 Þ: ð4Þ
locked to locked output characteristics at the torus ojAj  N0 ;A0
bifurcation while the full diamonds marking the
saddle-node bifurcation show the locking to un- Here, b is the linewidth enhancement factor, the
locking transition. The mapping displays all the ratio of the imaginary to real parts of the complex
regions of stable, bistable, periodic, quasiperiodic gain due to carrier density changes, and b0 is the
and chaotic dynamics. Unlike previous measure- ratio due to saturation effects [17]. These ratios
ments, there was no mode hop or optical power relate the changes in carrier density to refractive
transfer to side modes under optical injection that index changes in the system. The optical resonance
obscured the single-mode dynamics. frequency changes with carrier density. In a simple
two-level system b0 ¼ b, and on resonance b ¼ 0.
The semiconductor laser operates with the char-
4. Coupled equation model acteristics of a detuned oscillator, b 6¼ 0. However,
the two-level model is, at best, an approximation
The basic set of coupled equations for the for the two-band interaction of the semiconductor
complex optical field, A, of a single-mode semi- laser. Theoretical work has suggested that the two
conductor laser that is oscillating at free-running b parameters should be different, and that b0 0
optical frequency x0 , subject to an injected optical for typical values of semiconductor laser operation
field, Ai , at optical frequency x1 , and is coupled to [18]. As we show below, two different parameters,
a carrier density, N, can be written as both non-zero, allow us to match details of our
dA h cc i C experiment with the model, details that cannot be
¼  þ iðx0  xc Þ A þ gðN ; jAjÞA otherwise readily accounted for. Therefore, this is
dt 2 2
a key modification to the model that we have
þ gAi exp½iðx1  x0 Þt; ð1Þ
previously used.
dN J 2e0 n2 2 We make one additional modification. In past
¼  cs N  g
ðN ; jAjÞjAj ; ð2Þ
dt ed
x 0
h work using conventional Fabry–Perot cavity, edge
emitting lasers it has been observed that a mode
g
ðN ; jAjÞ Re½gðN ; jAjÞ: ð3Þ hop to another longitudinal mode is induced under
Here, cc is the photon decay rate from the laser some external optical injection conditions and that
cavity, xc is the cold cavity resonance frequency, C significant power can be emitted by other longi-
is the confinement factor that relates the overlap of tudinal modes under other conditions. Therefore,
the oscillating mode with the gain medium, g is the we add an equation describing a non-resonant
 T.B. Simpson / Optics Communications 215 (2003) 135–151 145

optical mode that is coupled to the others only where cs is the spontaneous carrier relaxation
through the carrier density and saturation effects. rate, cn is the differential carrier relaxation rate,
Eqs. (1) and (2) can then be recast in a convenient cp is the nonlinear carrier relaxation rate, and b is
form, making use of change of variables and in- the linewidth enhancement factor. The parameter
troducing some useful parameters. Because our J~ ¼ ðJ0 =ed  cs N0 Þ=cs N0 , is the differential injec-
primary interest is in how external optical injection tion current density above threshold, normalized
modifies the optical field, we use the steady-state, to the threshold current density. It is often re-
free-running condition of a semiconductor laser ferred to as the pump parameter. Eq. (8) is added
under dc bias above threshold in a single optical to model the side mode amplitude, w, with the
mode as our starting point. The optical injection additional parameter describing the gain defect of
field is characterized by the frequency detuning this secondary mode, l. This formulation as-
f ¼ X=2p ¼ ðxI  x0 Þ=2p and the normalized di- sumes that the secondary mode is weak under
mensionless injection parameter n ¼ gjAi j=cc jA0 j. free-running operation due to a small, relative to
For some of our characterization measurements cc , reduction of the steady-state gain or loss rate.
current modulation with a modulation amplitude Further, the coupling between modes occurs be-
Jm at a modulation frequency fm ¼ Xm =2p is ap- cause of the mutual dependence on the gain me-
plied to the laser. Therefore, the total time- dium, without any significant coupling of the
dependent current density injected to the laser is modal fields. As is shown below, this model re-
J ðtÞ ¼ J0 ð1 þ m cosðXm tÞÞ, where m ¼ Jm =J0 is de- produces key findings for laser diodes that oper-
fined as the modulation index. By expressing the ate in a single spatial and polarization mode
field amplitude and the carrier density of the in- under free-running operation when they are sub-
jection-locked semiconductor laser as A ¼ A0 ject to external optical injection and/or current
ð1 þ aÞeiu and N ¼ N0 ð1 þ n~Þ, respectively, the modulation.
characteristics of its modulation response can be Several important features emerge from a
numerically simulated using the following coupled linearized analysis of the model equations. In
equations:
" # this case, the optical injection and the current
da cc cn cp modulation are weak, causing only small devia-
2 2
¼ n~  ð2a þ a þ w Þ ð1 þ aÞ tions from the steady-state, free-running condi-
dt 2 cs J~ cc
tion. The variables a, n~, and / are assumed to be
þ cc n cos½Xt þ uðtÞ; ð5Þ small and there will be no significant excitation
" # of the weak side mode. The analysis for the case
du bcc cn b0 c p when b ¼ b0 has been published elsewhere [7,19].
¼ n~  ð2a þ a2 þ w2 Þ
dt 2 cs J~ b cc The amplitude modulation spectrum of the free-
cc n running laser due to a weak optical probe does
 sin½Xt þ uðtÞ; ð6Þ not depend on the value of b or b0 and can be
1þa
( written as:
d~
n J~ þ 1
¼  cs J~  m cosðXm tÞ X2 þ ðcr  cp Þ2
dt J~ a2 ¼ c2c n2 ; ð9Þ
" # 2 2
1 cn X þ X2r þ X2 c2r
2 2
þ þ ½ð1 þ aÞ þ w  n~
J~ cs J~ where a ¼ a½max is the normalized Fourier am-
  )
cp 2
plitude of the oscillation at X, X2r ¼ cc cn þ cs cp and
2 2 2
þ 1  ½ð1 þ aÞ þ w  ð2a þ a þ w Þ ; cr ¼ cs þ cn þ cp are the frequency and damping of
cc
the relaxation resonance. For the case when b 6¼ b0
ð7Þ there is a significant prediction for the optical
" #
dw cc cn cp l sidebands generated by a sinusoidal modulation
¼ n~  ð2a þ a2 þ w2 Þ  w; ð8Þ current. The equation for the modulation side-
dt ~
2 cs J cc cc
bands is
146 T.B. Simpson / Optics Communications 215 (2003) 135–151
n 2
o
1 þ ðb  b0 Þðcp =Xm Þ þ b2 c2 c2 ðJ~ þ 1Þ2 measurement is not corrupted by circuit charac-
a2m ¼ c n
m2 : teristics of the laser. With the rates known, Eq.
m r
2
 X 2 þ X 2 þ X 2 c2
m r
4 J~2 (10) can then be used to fit the spectrum of the
ð10Þ ratios of the oppositely shifted sidebands under
weak current modulation to determine b and b0 .
Eq. (10) gives the power ratio of the optical side- Because we are only interested in the ratios of the
band to the central peak under weak injection as sidebands, the measurement also is not sensitive to
would be measured with an optical spectrum an- circuit parasitics. Eq. (11) can be used as a check
alyzer. The effects of b 6¼ b0 are contained in the on the fitting. Therefore, all of the parameters
term in the numerator that disappears when the necessary to compare experimental results with
two parameters are assumed to be equal. Two key model can be determined from spectra of the
modifications to the modulation spectrum due to weakly perturbed system.
this term are an asymmetry in the sidebands, with
the positively offset sidebands enhanced when
b > b0 , and an X2 dependence to the modulation 5. DFB laser characterization
spectra near zero offset frequency. The unphysical
divergence of the latter prediction is a consequence The DFB laser under study exhibited single
of the linearization used in the weak injection mode laser operation under all excitation condi-
analysis. Eq. (10) predicts that the oppositely tions investigated here. Under free-running oper-
shifted sidebands power ratio will peak at a value ation, the optical power emitted in all of the side
that depends only on b with a frequency that de- modes was at least 20 dB below the optical power
pends on b, b0 , and cp . in the principal mode and each side mode power
There will also be modifications to the regen- was more than 30 dB below the principal mode
erative amplification spectrum of a weak optical power. Under external optical injection the total
probe. These spectra will follow the equation power in all of the side modes remained at least 20
88
 2 dB below the principal mode power and increased
jAr j2 << X2r
¼ 2
X þ 0
þ b Xcp from the free-running levels by no more than )30
jA0 j2 :: 2 dB relative to the principal mode power. There-
" #2 9 fore, the principal mode spectra contained the key
Xcp bX2r ðb0  bÞðcr  cp Þcp = information about the characteristics and dynam-
þ Xcr   
2 2 2 ; ics of the laser. For characterization, the free-
, 9 running laser is subjected to weak optical injection
n 2
o= c2 or weak current modulation and the resulting
c 2
 X2 þ X2rr þ X2 c2r n; ð11Þ
; X2 spectra are compared with the weak signal equa-
tions above. A typical fitting comparison is shown
where the equation gives the power ratio of the in the free-running regenerative amplification
optical sideband at the optical injection frequency spectrum in Fig. 2. Fig. 5 plots the relaxation
to the central peak as would be measured by an resonance frequencies, fr ¼ Xr =2p, resonance de-
optical spectrum analyzer. Unlike Eq. (10), Eq. cay frequencies, gr ¼ cr =2p, and saturation decay
(11) has a relatively weak dependence on the pa- frequencies, gp ¼ cp =2p, for the laser at a series of
rameter b0 and predicts spectra similar to the pre- bias currents. The figure also gives the linear least
viously published b ¼ b0 case. squares fits to the data for gr and gp , and the least
Because cn and cp are proportional to the free- squares fit assuming a power law dependence,
running power, the key rate parameters, cc ; cs ; cn , y ¼ axb , for fr . The relaxation resonance fre-
and cp , can be determined by using Eq. (9) to fit quency, fr , should scale with the square root of the
the amplitude modulation spectra of the laser un- current level above threshold while the other two
der weak optical injection at several bias currents. rates should scale linearly because the output
When optical injection is used as a probe, the power was measured to scale linearly.
 T.B. Simpson / Optics Communications 215 (2003) 135–151 147

Table 1
Parameter values determined from experimental measurements
Linewidth enhancement factor (b) 2.4
Ratio factor for nonlinear gain (b0 ) )1.5
Cavity photon decay rate ðcc Þ 5  1011 s1
Carrier spontaneous decay rate ðcs Þ 4  109 s1
Differential gain relaxation rate ðcn Þ 1:7J~  109 s1
Nonlinear gain relaxation rate ðcp Þ 5:4J~  109 s1
Threshold current 5.4 mA

cies. This was determined to be due to the effects of

a weak feedback signal from one of the FC/APC
connectors. Feedback, being an optical injection,
leads to modification to the asymmetry that peaks
Fig. 5. Key frequencies measured from the spectra of the 1915
LMO as a function of the current level above threshold. The near the relaxation resonance frequency. Jitter in
relaxation resonance frequency (diamonds) is plotted along with the laser optical frequency caused the measured
a power-law least squares fit. The resonance decay (squares) and signal to be modified by feedback that was ran-
gain saturation decay (triangles) are plotted along with least domly in and out of phase with the circulating
squares fits. All data are given in frequencies in this chart, as
field, causing the data scatter. However, at the
opposed to the angular rates given elsewhere in the text.
peak of the asymmetry, which is the key point for
the determination of b and b0 , the effects of feed-
Fig. 6 plots the sideband ratio under weak back are relatively small and the data can be rea-
current modulation and the fit to the data using sonably well fit. Note that the data are very poorly
Eq. (11). There is some scatter in the data and a fit by the curve with b0 ¼ 0, at variance with the-
relatively poor fit at higher modulation frequen- oretical predictions and consistent with recently
published data [17].
The parameters determined from the data are
summarized in Table 1. Accuracy for the first two
parameters is estimated to be 15% and 50%,
respectively, for the decay rates 10%, and 0.1
mA for the threshold current.

6. Model calculations of nonlinear dynamics

Computer programs were written to numeri-

cally integrate the four coupled nonlinear differ-
ential equations (5)–(8). To achieve consistent
1
results, the time step had to be ð8cc Þ , or some-
Fig. 6. Asymmetry of the optical sidebands generated under what shorter. Runs using double precision gave
weak current modulation with the 1915 LMO laser biased at
twice the bias current. The ratio of the positive frequency offset
essentially the same results as those using single
sideband power to the negative offset sideband is plotted as a precision though, of course, specific details chan-
function of modulation frequency. The diamonds are experi- ged in the region of chaotic dynamics.
mental data while the solid curve is calculated using Eq. (10), We can compare the observed dynamics with
the relevant rate parameters determined from Fig. 2, and calculated dynamics by relating the injected field to
b ¼ 2:4 and b0 ¼ 1:5, the upper and lower dot-dashed curves
were calculated by changing the value of b to 2.1 and 2.7, re-
the injection parameter, n, that we use in the cal-
spectively, and the dotted curve was calculated with b ¼ 2:4 and culations. Note that the locking–unlocking, sad-
b0 ¼ 0. dle-node bifurcation line shows a nearly linear
148 T.B. Simpson / Optics Communications 215 (2003) 135–151

dependence with injected field, as expected from

both a linearized and full nonlinear analysis. Using
this line of data as a reference, the injection pa-
rameter, n, can be related to the measured injec-
tion signal. This is the only free variable necessary
to make the connection between experimentally
measured and calculated data. In Fig. 7 we show
representative bifurcation diagrams generated
from the calculated time series. The diagram plots
extrema in the field amplitude as a function of the
injection parameter assuming an offset frequency
of 2 GHz for the master laser with respect to the
free-running optical frequency of the slave. A
stable output consists of a single point per injec-
tion parameter value. Limit cycle oscillations are
indicated by two points, period-two oscillations by
four points, etc., and quasiperiodic and chaotic
dynamics show a large set of points. We used the
experimentally measured parameters from Table 1
for the calculations. For single-mode calculations,
a relatively large gain defect, 0:1cc , was used to
suppress any build-up of the optical field in the
side mode. If this value is reduced to 0:01cc , or less,
then there is evidence of power leaking out of the
principal mode and a strong change in the dy-
namics. In Fig. 7, two cases are shown, one with
the gain defect for the side mode set to 0:1cc , and
the other where the gain defect was reduced to a
value of 0:001cc . Fig. 7 also shows the power in the
principal mode, comparing the cases with and
without the participation of the side mode. The
total power from the laser is essentially identical in
both cases. When the gain defect of the side mode
is sufficiently small, power is transferred and the
dynamics are changed. This occurs in and near
regions of complex dynamics and the locking–
unlocking (saddle node) transition, in agreement
with earlier data on Fabry–Perot lasers.
The calculated data from the bifurcation dia-
grams are combined into a mapping of the dy-
Fig. 7. Bifurcation diagrams showing extrema of the optical namics that is shown in Fig. 8(a). At low injection
field as a function of the injection parameter for an offset fre-
quency of 2 GHz. In (a) the gain defect of the side mode is large
levels, the laser is unlocked and undergoing Adler-
so that the laser operates in a single longitudinal mode. In (b) type frequency pulling and, simultaneously, de-
the gain defect is small and significant power leaks into the side stabilization of the relaxation resonance so that
mode when the circulating power is below the free-running the laser locks as it makes a transition to a limit
value. In (c) the relative circulating power level in the injected cycle mode of operation. The agreement between
mode is plotted for single mode, diamonds, and multimode,
open squares, operation.
theory and experiment is excellent, showing that
the model reproduces the experimentally observed
 T.B. Simpson / Optics Communications 215 (2003) 135–151 149

tion levels. Possibly, the inherent noise in the real

semiconductor laser system is sufficient to make
these regions indistinguishable from more complex
dynamics. Alternatively, the noise induces hops
between multiple attractors that may occupy dif-
ferent regions of phase space at these excitation
levels [10]. In Fig. 8(b) the regions where mode
hops alter the bifurcation diagrams are shown.
Comparing this to the previously published map of
a Fabry–Perot laser in [9], there is qualitative
agreement.

7. Discussion and conclusions

The use of a DFB laser diode has permitted us

to map out the nonlinear dynamics of a semicon-
ductor laser under external optical injection and
quantitatively compare the results with the single-
mode, rate-equation model. The agreement, both
qualitative and quantitative, is extraordinary and
makes this system a very good candidate for in-
vestigating nonlinear dynamics in a technologi-
cally relevant physical system. Relative to our
previous work, the work here shows that, in gen-
eral, the saturation effects on the complex gain
have a different proportionality constant than does
the differential gain component, in other words,
Fig. 8. Calculated mappings of the nonlinear dynamics using b0 6¼ b. The influence of non-lasing side modes can
the rate equation model and the experimentally determined
parameters listed in Table 1. The symbols are the same used in
be included, qualitatively at least, by the addition
Fig. 5 with the lightly shaded region covering the range of of an additional field amplitude equation, with this
multiwave mixing and Adler-type frequency pulling and the equation coupled to the original set through the
striped regions the complex dynamics. In (a) single mode op- carrier density equation. When such an equation is
eration is maintained by using a large gain defect for the side included with a gain defect that is on the order of
mode, while in (b) the gain defect is small and the darkly shaded
regions show where significant power has been transferred to
103 of the gain of the principal mode, we observe
the side mode. leakage of a significant fraction of the laser output
to the side mode. The leakage typically occurs
when the injection laser induces output charac-
characteristics. We should note that we were never teristics with an average output power below the
able to observe periodic dynamics of period free-running power of the principal mode. This
greater than six, a weak period three oscillation in often happens in regions of quasiperiodic or cha-
a strong period two feature, and were never able to otic output. This level of gain defect is consistent
observe period doubling beyond period four. The with the observed side mode characteristics of the
DFB laser always broke out of the resonance Fabry–Perot lasers previously investigated. In the
condition back into the more general quasiperi- DFB laser, all side modes are sufficiently sup-
odic, or chaotic, condition. The bifurcation dia- pressed that there is no appreciable leakage out of
grams do show higher periodic dynamics, but the principal mode and we were able to produce a
these regions cover a very narrow range of injec- mapping of the regions of nonlinear dynamics
150 T.B. Simpson / Optics Communications 215 (2003) 135–151

displayed by a semiconductor laser under optical Bernd Krauskopf, Prof. Jia-ming Liu, and Dr.
injection. Sebastian Wieczorek while pursuing this work. Dr.
This map bears a strong similarity to previous Wieczorek first calculated the approximate
mappings of dynamics of a Fabry–Perot cavity, boundary of the region of bistability at negative
edge-emitting laser. The nonlinear dynamics is detunings for this laser. I would also like to thank
more tightly bounded. This is expected, because the Dr. Michel Nizette and Prof. Thomas Erneux for a
laser used here has a smaller linewidth enhance- very careful reading of the manuscript and many
ment factor and larger ratio of the nonlinear gain constructive comments. The work was supported,
factor to the differential gain factor. From this map in part, by the Army Research Office under con-
it is clear that all regions of periodic dynamics with tract DAAG5598C0038.
period two or larger have an associated region of
complex dynamics. The higher period oscillatory
dynamics represent resonances of the competing References
frequencies initially associated with the offset be-
[1] J. Sacher, D. Baums, P. Panknin, W. Els€asser, E.O. G€
obel,
tween the master and slave laser optical frequencies
Phys. Rev. A 45 (1992) 1893–1905.
and the relaxation resonance frequency. Due to the [2] T.B. Simpson, J.M. Liu, IEEE Photon. Technol. Lett.
influence of the carrier density on the refractive 9 (1997) 1322;
index as well as the gain, both the offset and dy- X.J. Meng, D.T.K. Tong, T. Chau, M.C. Wu, IEEE
namic resonance frequencies change as the optical Photon. Technol. Lett. 10 (1998) 1620;
G. Yabre, H. de Waardt, H.P.A. van dem Boom, G.D.
injection is increased. It is the interplay between
Khoe, IEEE J. Quantum Electron. 36 (2000) 385.
these two resonant frequencies that forms the [3] T.B. Simpson, F. Doft, IEEE Photon. Technol. Lett.
overall basis for the global dynamics. 11 (1999) 1476;
Here, we have not rigorously compared exper- T.B. Simpson, Opt. Commun. 170 (1999) 93.
iment and theory in the region of complex dy- [4] H.F. Chen, J.M. Liu, IEEE J. Quantum Electron. 36 (2000)
27.
namics. There is not an exact matching of spectra
[5] T.B. Simpson, J.M. Liu, A. Gavrielides, V. Kovanis, P.M.
but overall, qualitative consistency between model Alsing, Phys. Rev. A 51 (1995) 4181;
and experiment. Indeed, the consistency is suffi- S. Eriksson, A.M. Lindberg, Quantum Semiclassical Opt.
ciently good that the data and model can be 14 (2002) 149.
compared to demonstrate that different transitions [6] S. Wieczorek, B. Krauskopf, D. Lenstra, Opt. Commun.
172 (1999) 279.
to and from chaotic dynamics predicted by theory
[7] J.M. Liu, T.B. Simpson, IEEE J. Quantum Electron.
can be seen in the experimental results. We have 30 (1994) 957.
described the spectral evidence for the quasiperi- [8] S. Wieczorek, T.B. Simpson, B. Krauskopf, D. Lenstra,
odic route to chaos here. More sophisticated the- Phys. Rev. E 65 (2002) 045207(R).
oretical tools than those used here are required to [9] T.B. Simpson, J.M. Liu, K.F. Huang, K. Tai, Quantum
Semiclassical Opt. 9 (1997) 765.
make the connection explicit. These results are
[10] B. Krauskopf, S. Wieczorek, D. Lenstra, Appl. Phys. Lett.
given in a companion paper [14]. Together, these 77 (2000) 1611.
works demonstrate that the agreement between [11] A.F. Siegman, Lasers, University Science Books, Mill
model and experimental data is so good that the Valley, CA, 1986 (Chapter 29).
semiconductor laser subject to external optical [12] B. Krauskopf, N. Tollenaar, D. Lenstra, Opt. Commun.
156 (1998) 158;
injection is an excellent model system for testing
M. Nizette, T. Erneux, A. Gavrielides, V. Kovanis, Proc.
the predictions of nonlinear dynamics theories and SPIE 3625 (1999) 679.
models. [13] V. Kovanis, A. Gavrielides, T.B. Simpson, J.M. Liu, Appl.
Phys. Lett. 67 (1995) 2780.
[14] S. Wieczorek, T.B. Simpson, B. Krauskopf, D. Lenstra,
Opt. Commun. 215 (2003) 125.
Acknowledgements
[15] R. Hui, A. DÕOttavi, A. Mecozzi, P. Spano, IEEE J.
Quantum Electron. 27 (1991) 1688.
The author has benefited from many useful [16] V. Kovanis, T. Erneux, A. Gavrielides, Opt. Commun.
discussions with Dr. Athanaseos Gavrielides, Prof. 159 (1999) 177;
 T.B. Simpson / Optics Communications 215 (2003) 135–151 151

S. Wieczorek, B. Krauskopf, D. Lenstra, Opt. Commun. [18] G.P. Agrawal, IEEE J. Quantum Electron. 26 (1990) 1901;
183 (2000) 215. J. Yao, G.-H. Duan, P. Gallion, IEEE Photon. Technol.
[17] T.B. Simpson, F. Doft, E. Strzelecka, J.J. Liu, W. Lett. 4 (1992) 1103.
Chang, G.J. Simonis, IEEE Photon. Technol. Lett. [19] T.B. Simpson, J.M. Liu, A. Gavrielides, IEEE J. Quantum
13 (2001) 776. Electron. 32 (1996) 1456.