0% found this document useful (0 votes)
5 views9 pages

Influence of A Nonuniform Coupling Coefficient On The Static and Large Signal Dynamic Behavior of Bragg-Detuned DFB Lasers

This paper investigates the dynamics of Bragg-detuned Distributed Feedback (DFB) lasers, specifically comparing conventional corrugation-pitch-modulated DFB lasers (CPM-DFB) with those that have a longitudinally dependent coupling coefficient (CPM-DCC-DFB). The study demonstrates that while CPM-DCC-DFB lasers can achieve better steady-state singlemode operation and reduced modal instabilities, they may experience side mode excitation during transient intervals. The findings suggest that optimizing the coupling coefficient profile can enhance the performance of DFB lasers in terms of spectral stability and modulation depth capabilities.

Uploaded by

fan ping
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
5 views9 pages

Influence of A Nonuniform Coupling Coefficient On The Static and Large Signal Dynamic Behavior of Bragg-Detuned DFB Lasers

This paper investigates the dynamics of Bragg-detuned Distributed Feedback (DFB) lasers, specifically comparing conventional corrugation-pitch-modulated DFB lasers (CPM-DFB) with those that have a longitudinally dependent coupling coefficient (CPM-DCC-DFB). The study demonstrates that while CPM-DCC-DFB lasers can achieve better steady-state singlemode operation and reduced modal instabilities, they may experience side mode excitation during transient intervals. The findings suggest that optimizing the coupling coefficient profile can enhance the performance of DFB lasers in terms of spectral stability and modulation depth capabilities.

Uploaded by

fan ping
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 16, NO.

3, MARCH 1998 419

Influence of a Nonuniform Coupling Coefficient


on the Static and Large Signal Dynamic
Behavior of Bragg-Detuned DFB Lasers
Thierry Fessant

Abstract—A time domain model is used to study the large signal structures with multiple discrete phase shifts located along
dynamics of general Bragg-detuned cavities. Such structures, the cavity [8], [9] or phase-adjusted DFB lasers with a step-
in which the grating pitch is spatially nonuniform so that a like nonuniform stripe width structure [10], [11]. Alternatively,
distributed phase shift is produced along the cavity, are often
called corrugation-pitch-modulated distributed feedback lasers DFB lasers with modulation of their grating pitch (CPM–DFB
(CPM–DFB). Furthermore, by introducing a longitudinally de- for corrugation-pitch-modulated) provide a longitudinal Bragg
pendent coupling coefficient , a corrugation-pitch-modulated detuning so that a distributed and more gradual phase shift
distributed-coupling-coefficient DFB laser (CPM-DCC-DFB) with is produced [12]–[14]. The discontinuity in the corrugation of
quite different spectral properties can be realized. We present in QWS–DFB cavities is then smoothed and, consequently, the
this paper a comparison between these two kinds of components.
It is shown that an adequate profile of the coupling coefficient photon density exhibits a less pronounced maximum. Although
(stronger coupling at the center of the cavity) in a CPM-DCC- good results are obtained at threshold, static above-threshold
DFB structure can help to achieve better steady-state singlemode studies revealed that these structures could be subject to modal
operation with respect to conventional CPM-DFB lasers (uniform instabilities.
coupling coefficient). However, in CPM-DCC-DFB lasers, side We propose in this paper a multisection DFB laser with
modes are liable to be excited during the turn-on transient
interval, thus imposing a limitation to the modulation depth longitudinal variation in both its grating pitch and coupling
requirement. coefficient profile (CPM-DCC-DFB for corrugation-pitch-
modulated distributed-coupling-coefficient DFB). Threshold
Index Terms— Coupled mode analysis, distributed feedback
lasers, optical modulation, semiconductor laser, time domain studies performed on a three-section CPM-DCC-DFB cavity
analysis. [14] have shown that this particular grating configuration
help to improve simultaneously the gain margin and the
intracavity field flatness. To date, results about large signal
I. INTRODUCTION dynamics of CPM-DCC-DFB cavities have not been presented

C ONVENTIONAL quarter-wave shift DFB lasers (QWS-


DFB) are often problematic because, even if singlemode
operation is ensured at threshold thanks to the built-in phase
and the paper is based on a comparison between dynamics
of conventional CPM-DFB ( -dependent , uniform ) and
CPM-DCC-DFB lasers ( and both -dependent). Transient
shift, longitudinal spatial hole burning (LSHB) occurs at higher responses to a step current, dynamic (during the turn-on
injection currents, contributing to degrade the laser side mode transient) or steady-state Fourier spectra, SMSR versus biasing
suppression ratio (SMSR) [1]–[3]. Indeed, in QWS-DFB lasers current and longitudinal profiles of intrinsic parameters are
with strong coupling, the photon density tends to accumulate analyzed in order to illustrate the fundamental differences
around the phase shift and LSHB is just a consequence of between these two kinds of structures, due to the -dependent
spatial nonuniformities in the intracavity field (because of coupling coefficient.
induced nonuniformities of the effective index). So, further
optimizations are required to improve the laser performance
and multisection structures is a possibility. II. OUTLINE OF THE TIME DOMAIN MODEL
In order to reduce the sensitivity of the structure to LSHB, Starting from the multimode time domain model of
two criterions are generally chosen. The first one consists in Marcenac et al. at Cambridge University [15], [16], we have
enlarging the threshold gain margin. For example, attention developed and implemented a numerical simulator which has
have been paid to longitudinal variations of the coupling been used to predict the large signal dynamics of multisection
coefficient profile in a QWS-DFB laser [4], [5] or to gain DFB structures. For example, quarter-wave shift distributed
coupling mechanisms in Bragg gratings [6], [7]. The second coupling coefficient DFB lasers (QWS-DCC-DFB) with a
criterion consists in flattening the intracavity field, in order stronger center coupling have been found to exhibit better
to eliminate the main cause of LSHB. Some works report spectral stability than conventional QWS-DFB cavities, due
to an enlarged threshold gain margin [17]. Moreover, if the
Manuscript received April 21, 1997; revised October 15, 1997. dynamic model is used to study the steady-state properties
The author is with École Nationale d’Ingénieurs de Brest, laboratoire RESO,
Technopôle Brest-Iroise, CP 15, 29608, Brest Cedex, France. of DFB cavities, comparisons made with our static above-
Publisher Item Identifier S 0733-8724(98)01901-X. threshold model [18] show a good adequation between these
0733–8724/98$10.00  1998 IEEE
420 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 16, NO. 3, MARCH 1998

two approaches. So, the static and dynamic softwares may


provide common features (steady-state SMSR, carrier density
or effective index longitudinal profiles) and, principally,
enable a complementary analysis of complex structures to
be performed.
The model is based on the time-dependent coupled wave
equations for the forward and reverse fields inside the cavity,
respectively and , which can be expressed as

(1)

(2)

is the longitudinal coordinate of the laser while is the


time. is the group velocity, is the material gain,
are the internal loss and the detuning from the
Bragg frequency. The material gain and frequency detuning
are expressed as

Fig. 1. Schematic diagram of a corrugation-pitch-modulated


(3) distributed-coupling-coefficient, three-section DFB laser (CPM-DCC-DFB).
The coupling coefficient  and the grating pitch 3 are z -dependent (s and
s in the side section, c and 3c in the center section). In our analysis,
(4) c > s and 3c < 3s .

In (3) and (4), is the optical confinement factor and is the effective index at transparency and is the
is the differential gain. stands for the carrier phase-amplitude coupling factor.
density defined further while is the carrier density at The longitudinally dependent coupling coefficient en-
transparency. is the photon density and the nonlinear ables distributed-coupling-coefficient structures to be taken
gain coefficient. is an approximate emission wavelength into account. describes the spontaneous emission and
and the effective index. is the longitudinally- is generated from a Gaussian distributed random number
dependent (structural) grating pitch, allowing Bragg-detuned generator that satisfies the correlation
DFB structures to be considered. In the model, a flat mate-
rial gain (with no frequency-dependence) is assumed.
However, the frequency dependence can be taken into account (8)
by using a digital filter [19], [20].
The carrier density is described by its rate equation is the spontaneous coupling factor, the Petermann factor
and the total length of the laser.
All equations involved in the calculation are solved nu-
merically. LSHB is described by discretizing the intrinsic
parameters all along the cavity. All parameters are constant
(5) in a subsection but can vary from subsection to subsection.
The time-domain model is principally developed for studying
is the uniform biasing current density, the electric charge large signal dynamics of various structures but, by allowing
and the thickness of the active layer. is the linear the transient response to settle for a given injection current,
recombination lifetime, the radiative spontaneous emission steady-state characteristics can also be obtained.
coefficient and the Auger recombination coefficient.
The total photon density is the sum of the photon III. NUMERICAL RESULTS
densities of the two traveling wave
The proposed CPM-DCC-DFB structure (Fig. 1) is a sym-
(6) metric three-section laser of total length m, with
no facet reflection. The length, grating pitch and normalized
where the star represents the complex conjugate. coupling coefficient of center section are denoted by ,
The effective index is given by and , respectively while those of side sections are called
and . The CPM-DFB structure is characterized by
a -dependent grating pitch and a uniform coupling coefficient
(7)
. In the CPM-DCC-DFB cavity, both and
FESSANT: DYNAMIC BEHAVIOR OF BRAGG-DETUNED DFB LASERS 421

are -dependent. An optimization of CPM-DCC-DFB lasers


at threshold previously revealed that an adequate value of
the grating pitch difference and a stronger coupling in the
center section allows to enhance simultaneously the threshold
gain margin and field flatness with respect to conventional
CPM-DFB structures [14]. Furthermore, the static singlemode
stability of Bragg-detuned cavities with respect to the onset of
a side mode is strongly dependent upon the sign of the relative
grating pitch difference [defined in (9)]. So, cavities with
positive or negative Bragg-detuning are not equivalent above
threshold [13]. In the present case, a negative is assumed,
\cdots, . The following values of relative section
lengths, coupling coefficient and grating pitch difference are (a)
used in our dynamic simulations of CPM-DFB and CPM-
DCC-DFB lasers

(9)

are common to CPM-DFB and CPM-DCC-DFB lasers. More-


over, for the CPM-DCC-DFB cavity, the coupling coefficient
ratio is

(10)

The 120 subsections are used for numerical calculations. The (b)
lasers are biased at transparency and switched to a high current
Fig. 2. Transient response to a heavy-side step, from transparency to twice
density according to a Heaviside step, i.e., the threshold, of the output photon density. (a) Conventional CPM-DFB laser
(uniform coupling) and (b) CPM-DCC-DFB laser (z -dependent ).
with if and if

corresponds to the biasing current (or modulation depth).


The output photon density dynamics corresponding to var- that the turn-on-transient appears to be different in the case of
ious biasing currents are shown in Figs. 2–4. When the mod- CPM-DCC-DFB cavities and oscillations of small amplitude
ulation depth is twice the threshold (Fig. 2), no difference indicating a possible mode beating during the first pulse can
between the conventional CPM-DFB and CPM-DCC-DFB be seen [small figure at the top right of Fig. 4(b)].
cavities is observed. The turn-on-transient after the lasers have A fast Fourier transform (FFT) on the (sampled) time-
switched on are similar and typical relaxation oscillations are dependent complex output fields has been carried out on both
obtained in both cases. For higher modulation depth, ( CPM-DFB and CPM-DCC DFB lasers. A Hanning window
at Fig. 3 and at Fig. 4), the conventional CPM-DFB covering either the turn-on-transient interval or the steady-
structure is subject to mode beating and its output photon state regime is used to characterize the spectral (singlemode
density starts to oscillate in strong amplitude as the result of the or multimode) regime of each structure during all the dynamic
interference between the involved modes. On the other hand, process. In fact, a rigorous steady-state regime occurs only for
the CPM-DCC-DFB laser maintains a steady-state regime in the CPM-DCC-DFB and the term “steady-state” is incorrect
which the output power becomes stabilized (no mode beating), as far as the CPM-DFB laser is considered but we will use it
whatever the modulation biasing current. Note also the higher by opposition to the turn-on-transient interval corresponding
output power in the case of the CPM-DCC-DFB structure to the laser switch-on.
with respect to the conventional CPM-DFB laser for a given Let us first consider the conventional CPM-DFB structure.
normalized injection current. Reasons of this phenomenon will During the turn-on-transient, a broadened but singlemode
be given later. spectrum is obtained, showing that the LSHB effects, i.e., the
The mode beating comes from the onset of a side mode transition to a multimode regime, do not take place during the
in the cavity as the result of spatial hole burning. This phe- first overshoot but rather in the steady-state regime (Fig. 5).
nomenon being power-dependent, the amplitude of the beating Moreover, because the dynamic photon density increases with
increases. The conventional Bragg-detuned cavity, with uni- modulation depth, larger dynamic variations of the effective
form coupling coefficient, is rapidly subject to LSHB while index are induced, especially during the turn-on-transient
the CPM-DCC-DFB structure seems to be widely prevented (Figs. 2–4). So, the dynamic spectrum is more and more
against it. So, a -dependent coupling coefficient appears to be broadened. This broadening indicates a gradually chirped
a good means to fight against the effects of LSHB in the static pulse. In the steady-state regime, the laser is singlemode
regime. However, a careful study of Figs. 3(b) and 4(b) shows for a biasing current of twice the threshold [Fig. 6(a)] but
422 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 16, NO. 3, MARCH 1998

(a) (a)

(b) (b)
2
Fig. 3. Transient response to a heaviside step, from transparency to 4 2
Fig. 4. Transient response to a heaviside step, from transparency to 8
threshold, of the output photon density. (a) Conventional CPM-DFB laser threshold, of the output photon density. (a) Conventional CPM-DFB laser
(uniform coupling) and (b) CPM-DCC-DFB laser (z -dependent ). (uniform coupling) and (b) CPM-DCC-DFB laser (z -dependent ).

becomes rapidly multimode with increasing modulation depth properties. That is, a higher value of leads to
[Fig. 6(b) and (c)]. So, the spectral stability with respect to the same results as far as the spectral stability is concerned: the
the excitation of a second mode is very poor and would cause Fourier spectra are almost unchanged (excepted that the noise
detrimental pulse-spreading in communication systems. The floor is higher) while the dynamic and steady-state SMSR keep
onset of multiple side modes is believed to be the result of the same values.
modeling a flat material gain (so that the gain is uniformly The same calculations have been repeated for a range of
distributed upon all the modes) but a frequency-dependent biasing currents going from up to in order to
gain with a gain peak located around the main mode would estimate both steady-state and dynamic SMSR (during the
probably reduce the side modes influence [19]. turn-on-transient) versus normalized biasing current. Follow-
The case of the CPM-DCC-DFB cavity is entirely different ing [22], we assume that singlemode operation is ensured if
(Fig. 7). Indeed, the turn-on-transient appears to be multimode the drop between the starting and final values of the SMSR
and all the involved modes broadened, despite the gain margin do not exceed a fixed value (say 40 dB). It is shown in
is large at threshold . However, for the range of Fig. 9 that the steady-state SMSR of the conventional CPM-
operating biasing currents, the main mode remains preponder- DFB structure falls down rapidly with while that of the
ant even at high injection current (SMSR drops from a value of CPM-DCC-DFB cavity is clamped to a high value.
65 dB at to a value of 30 dB at ). On the other side, Better static spectral stability in CPM-DCC-DFB lasers
in the steady-state regime, the laser remains singlemode with with respect to conventional CPM-DFB structures comes from
a good SMSR, whatever the modulation depth (Fig. 8). This several physical phenomena.
time, a transition from a multimode regime to a singlemode 1) First, higher threshold gain margins are achieved in such
one is obtained. structures when , so that the possible onset of
It must be pointed out that a relatively low value of the side modes induced by LSHB is delayed (with respect
spontaneous emission factor has been used in the simulations to increasing bias current).
and higher values can be found in the literature. 2) Second, modulating spatially the coupling coefficient
On the other side, it has been shown that the actual value of induces a nonlasing side mode competition. The modes
(as well as that of the nonlinear gain coefficient) was a source closer to the stop-band do not necessarily have the
of damping in the relaxation oscillations [21]. In this work, lowest net gains and this phenomenon may help to
we have not really focused on the relaxation frequency and enhance the lasing mode contribution in the above-
damping phenomena but rather on the spectral and dynamic threshold regime.
FESSANT: DYNAMIC BEHAVIOR OF BRAGG-DETUNED DFB LASERS 423

(a) (a)

(b)
(b)

(c)
Fig. 5. Fourier transform spectra of time-domain signals for the conventional (c)
CPM-DFB laser. The Hanning window covers the turn-on-transient interval
and the modulation depth is increased from twice the threshold to 8 2 Fig. 6. Steady-state Fourier spectra of time-domain signals for the conven-
threshold. tional CPM-DFB laser. The modulation depth is increased from twice the
2
threshold to 8 threshold.

3) Salzman et al. [13] have shown that CPM-DFB lasers


with positive or negative Bragg detuning behaved dif- better mode selectivity are obtained. Anyway, in the
ferently above threshold (although they are equiva- two cases, the effective index spatial profile is well
lent at threshold as far as the gain margin and field modified with increasing bias current as the result of
flatness are concerned) and this analysis can be ap- LSHB. On the other side, an above-threshold static study
plied to CPM-DCC-DFB lasers. Spatial hole burning of CPM-DCC-DFB structures shows that the steady-
leads to a -dependent Bragg wavelength through a state longitudinal profile of the effective index is not
-dependent effective index along the cavity, which dramatically modified with increasing bias currents [14].
may be responsible for the onset of a side mode. Thanks to the cusps located at the coupling coefficient
So, depending on the sign of , the effective discontinuities, which increase the photon density near
index in the center and side sections may either depart the facets, the effective index is averaged along the
from each other, or come close to each other above cavity and remains relatively uniform so that quite less
threshold. In the first case, the overall filter response sensitivity to LSHB is achieved.
is broadened and the selectivity becomes poor. In the The dynamic SMSR corresponding to the first overshoot
second case, a narrower overall filter response and a remains high for the conventional CPM-DFB but decreases
424 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 16, NO. 3, MARCH 1998

(a)
(a)

(b)
(b)

(c)
Fig. 7. Fourier transform spectra of time-domain signals for the
CPM-DCC-DFB laser. The Hanning window covers the turn-on-transient (c)
interval and the modulation depth is increased from twice the threshold
2
to 8 threshold.
Fig. 8. Steady-state Fourier spectra of time-domain signals for the
CPM-DCC-DFB laser. The modulation depth is increased from twice the
2
threshold to 8 threshold.
slightly for the CPM-DCC-DFB laser with biasing current
(Fig. 10). We believe that the onset of side modes at high modulation depth requirement but this limit is high so that
bias currents during the turn-on transient interval in the CPM- a -dependent is a good mean to achieve very good static
DCC-DFB laser may result from the interaction between and dynamic singlemode stability in Bragg-detuned cavities
the gain distribution along the cavity and the spatial profile for the range of usual bias currents.
of the side modes, i.e., these side modes may be favored Another kind of mode instability has been reported in index-
temporarily before the gain is clamped to its steady-state coupled or complex-coupled DFB lasers [23], [24]. Indeed,
value. That is, the CPM-DCC-DFB cavity tolerates rather high depending on the bias current and for particular sets of
bias currents and will become problematic only for very high physical parameters, DFB structures may be subject to mode
ones. The maximum bias current for single mode operation hoppings as the result of spatial hole burning. The lasing mode,
is found to be around . Our criterion of dynamic stable and dominant at a given injection current, may become
singlemode operation is nevertheless kept in the range of unstable at a different bias current and then hops to another
operating currents. laser mode. More spectacular with this modal instability is the
So, in CPM-DCC-DFB structures, the turn-on-transient can possible induced asymmetry in the carrier density and effective
be worrying because side modes are liable to be excited index longitudinal profile of this new mode as well as that of
during this interval. This behavior imposes a limitation to the the output power. Such instabilities have not been predicted in
FESSANT: DYNAMIC BEHAVIOR OF BRAGG-DETUNED DFB LASERS 425

Fig. 9. Steady-state side mode suppression ratio versus biasing current for Fig. 11. Evolution of the carrier density longitudinal profile in the
the conventional CPM-DFB and CPM-DCC-DFB structures. steady-state regime for the CPM-DFB cavity with increasing biasing current.
The typical “break” of QWS-DFB lasers located at the cavity center is
removed thanks to the distributed phase shift.

Although this time domain model is principally thought


to study large signal dynamics, steady-state characteristics
can also be obtained by allowing the transient response to
settle for a given injection current. For example, it can be
used to visualize the combined effects of growing stimulated
emission and LSHB on the carrier density longitudinal profile
. for the CPM-DFB laser is plotted in Fig. 11 for
increasing biasing currents. A gradual distortion of the curve
with respect to its uniform profile at threshold is observed.
However, thanks to the distributed phase shift generated by
Fig. 10. Dynamic (during the turn-on-transient) side-mode suppression ratio the Bragg-detuned section, the photon density accumulates
versus biasing current for the conventional CPM-DFB and CPM-DCC-DFB quite less in the cavity center with respect to QWS-DFB
structures. lasers, so that the carrier density does not exhibit the typical
“break” of QWS-DFB’s at the phase shift location. The CPM-
TABLE I DCC-DFB structure exhibits also such distortions (Fig. 12) but
PARAMETERS USED IN THE MODEL
the important feature concerns the two cusps located at the
coupling coefficient discontinuities and .
The carrier profile is no longer smooth and the carriers are
rather depleted near the facets. This means that the photon
density is large at the ends of the cavity so that the output
power is enlarged. That’s why the output photon density is
found to be larger for CPM-DCC-DFB lasers with respect to
conventional CPM-DFB structures (Figs. 2–4).
Longitudinal profiles of carrier density can also indicate
the occurrence of a multimode operation in DFB structures.
Indeed, the time domain model is basically multimode and the
whole carrier density is the result of the interplay with all the
involved modes. So, when the laser is singlemode, the carrier
density is depleted only through the main mode but, when the
laser is multimode, this depletion is due to a superposition of
several modes [25]. An illustration is given in Fig. 13. Let
us consider two distinct instants chosen in the steady-state
regime. For the conventional CPM-DFB laser, the first one
is taken before the mode beating and the second one during
the structures studied in this paper with the physical parameters this mode beating. For the CPM-DCC-DFB structure, these
given in Table I. However, for a similar CPM-DCC-DFB particular instants are chosen no matter when after the laser
cavity in which the grating configuration is such that the center has stabilized.
pitch is longer than the side pitch , we have found Fig. 13(a) shows that the carrier density profiles of the
an asymmetry in the carrier density and output power [14]. conventional CPM-DFB structure depart from each other even
426 JOURNAL OF LIGHTWAVE TECHNOLOGY, VOL. 16, NO. 3, MARCH 1998

Fig. 14. Longitudinal profiles at threshold of the photon density in the


Fig. 12. Evolution of the carrier density longitudinal profile in the CPM-DFB structure. Main mode, first side one and normalized superposition
steady-state regime for the CPM-DCC-DFB cavity with increasing biasing of both, showing the flattening of the whole intracavity field.
current. The two cusps located at the coupling coefficient discontinuities are
responsible for the enlargement of the output photon density.
the carrier density but during the beating, two modes are
mainly involved in this depletion and is flattened. This
phenomenon can be understood by a precise observation of the
photon density profiles at threshold. The main mode, the first
side one and a normalized superposition of both are plotted in
Fig. 14. As far as the photon density profile of the side mode
has a minimum where the main mode is maximum (center of
the cavity), the superposition of both becomes more uniform.
So, during the beating, the carrier density is depleted by the
(more uniform) total photon density and thus, tends to flatten.
On the other side, the carrier density of the CPM-DCC-DFB
structure is no longer modified whatever the instants chosen
in the steady-state regime, i.e., that no side mode is excited
and the laser remains singlemode.

(a)
IV. CONCLUSION
Large signal dynamics of general Bragg-detuned cavities
have been analyzed. Conventional CPM-DFB structures (uni-
form coupling coefficient ) are found to exhibit poor steady-
state singlemode operation because of spatial hole burning
effects which allows side modes to be excited. A solution to
this limitation is to modulate spatially the coupling coefficient
so that a CPM-DCC-DFB cavity is produced. A stronger
at the cavity center tends to enhance the static singlemode
properties because these particular structures are less sensi-
tive to LSHB. Unfortunately, in CPM-DCC-DFB lasers, side
modes are liable to be excited during the turn-on transient
interval, thus imposing a limitation to the modulation depth
(b) requirement. However, in the case of the optimized CPM-
Fig. 13. Variations induced on the steady-state carrier density profile, due to DCC-DFB cavity studied in this paper, this limit is high
the mode beating in the cavity. (a) In the conventional CPM-DFB, the carrier (around ) so that a -dependent is a good means to
profile before and during the beating departs from each other, due to the onset
of a side mode. (b) The CPM-DCC-DFB cavity remains singlemode and the achieve very good static and dynamic singlemode stability in
carrier profile is clamped, whatever the instants chosen in the steady-state the range of operating bias currents.
regime. The simultaneous longitudinal and time dependence of this
multimode model can be used to visualize LSHB as well as the
occurrence of a side mode in the cavity. A characterization of
if the biasing current is clamped after the switch-on (the multimode operation in complicated DFB structures is carried
biasing current is kept constant). The carrier profile out through transient responses (mode beating), Fourier spectra
is thus modified because of the occurrence of a second mode and longitudinal carrier density profiles (gradual departure of
in the cavity: before the beating, only the main mode depletes the carrier density at two instants of the steady-state regime).
FESSANT: DYNAMIC BEHAVIOR OF BRAGG-DETUNED DFB LASERS 427

ACKNOWLEDGMENT [14] T. Fessant, “Optimization des paramètres structurels de lasers à


semi-conducteurs DFB multi-sections. Influence sur leurs propriétés
The author wishes to thank J. Le Bihan and F. Ropars from modales, statiques et dynamiques,” Ph.D. dissertation, Ecole Nationale
École Nationale d’Ingénieurs de Brest for their encouragement. d’Ingénieurs de Brest, Brest, 1996.
[15] C. F. Tsang, D. D. Marcenac, J. E. Carroll, and L. M. Zhang, “Compar-
ison between “Power Matrix Model (PMM)” and “Time Model Domain
REFERENCES (TDM)” in modeling large signal responses of DFB lasers,” in Inst. Elec.
Eng. Proc.—Part J., Apr. 1994, vol. 141, no. 2, pp. 89–96.
[1] H. Soda, Y. Kotaki, H. Sudo, H. Ishikawa, S. Yamakoshi, and H. Imai, [16] L. M. Zhang, S. F. Yu, M. C. Nowell, D. D. Marcenac, J. E. Carroll, and
“Stability in single longitudinal mode operation in GaInAsP/InP phase- R. G. S. Plumb, “Dynamic analysis of radiation side-mode suppression
adjusted DFB lasers,” IEEE J. Quantum Electron., vol. QE-23, pp. in a second-order DFB laser using time-domain large-signal traveling
804–814, June 1987. wave model,” IEEE J. Quantum Electron., vol. 30, pp. 1389–1395, June
[2] J. E. Whiteaway, G. H. B. Thompson, A. J. Collar, and C. J. Armistead, 1994.
“The design and assessment of =4 shifted DFB structures,” IEEE J. [17] T. Fessant, “Enhanced dynamics of QWS-DFB lasers by longitudinal
Quantum Electron., vol. 25, pp. 1261–1279, June 1989. variation of their coupling coefficient,” IEEE Photon. Technol. Lett.,
[3] P. Correc, “Stability of phase-shifted DFB lasers against hole burning,” vol. 9, pp. 1075–1077, Aug. 1997.
IEEE J. Quantum Electron., vol. 30, pp. 2467–2476, Nov. 1994. [18] , “Threshold and above-threshold analysis of corrugation-pitch-
[4] T. Fessant and J. Le Bihan, “Analysis of the spectral properties of multi- modulated DFB lasers with inhomogeneous coupling coefficient,” to be
section shifted DFB lasers with inhomogeneous coupling coefficient,” in published in Inst. Elec. Eng. Proc.—Optoelectron., vol. 144, no. 6, Dec.
Proc. Semiconductor Integrated Opto-Electron. Conf., SIOE’95, Cardiff, 1997.
Wales, Apr. 1995. [19] A. J. Lowery, “Modeling the static and dynamic behavior of quarter-
[5] B. S. K. Lo and H. Ghafouri-Shiraz, “Spectral characteristics of DFB wave-shifted DFB lasers,” IEEE J. Quantum Electron., vol. 28, pp.
laser diodes with distributed coupling coefficient,” J. Lightwave Tech- 1874–1883, Sept. 1992.
nol., vol. 13, pp. 200–212, Feb. 1995. [20] D. J. Jones, L. M. Zhang, J. E. Carroll, and D. D. Marcenac, “Dynamics
[6] E. Kapon, A. Hardy, and A. Katzir, “The effect of complex coupling of monolithic passively mode-locked semiconductor lasers,” IEEE J.
coefficient on distributed feedback lasers,” IEEE J. Quantum Electron., Quantum Electron., vol. 31, pp. 1051–1058, June 1995.
vol. 18, pp. 66–71, Jan. 1982. [21] A. J. Lowery, “A new time-domain model for spontaneous emission in
[7] K. David, J. Buus, and R. G. Baets, “Basic analysis of AR-coated, partly semiconductor lasers and its use in predicting their transient response,”
gain-coupled DFB lasers: The standing wave effect,” IEEE J. Quantum Int. J. Numerical Modeling: Electron. Networks, Devices and Fields, vol.
Electron., vol. 28, pp. 427–433, Feb. 1992. 1, pp. 153–164, 1988.
[8] J. I. Kinoshita and K. Matsumoto, “Yield analysis of SLM DFB lasers [22] S. F. Yu and H. Li, “Proposed enhancement of side-mode suppression
with axially-flattened internal field,” IEEE J. Quantum Electron., vol. ratio in =4 shifted distributed feedback lasers with nonuniform diffused
25, pp. 1324–1332, June 1989. quantum wells,” IEEE Photon. Technol. Lett., vol. 8, pp. 482–484, Apr.
[9] J. E. A. Whiteaway, B. Garret, G. H. B. Thompson, A. J. Collar, C. 1996.
J. Armistead, and M. J. Fice, “The static and dynamic characteristics [23] A. J. Lowery and H. Olesen, “Dynamics of mode instabilities in quarter-
of single and multiple phase-shifted DFB laser structures,” IEEE J. wave-shifted DFB semiconductor lasers,” Electron. Lett., vol. 30, no.
Quantum Electron., vol. 28, npp. 1277–1293, May 1992. 12, pp. 965–967, June 1994.
[10] H. Soda, K. Wakao, H. Sudo, T. Tanahashi, and H. Imai, “GaInAsP/InP [24] B. Jonsson, A. J. Lowery, H. Olesen, and B. Tromborg, “Instabilities
phase-ajusted distributed feedback lasers with a step-like nonuniform and nonlinear L-I characteristics in complex-coupled DFB lasers with
stripe width structure,” Electron. Lett., vol. 20, no. 24, pp. 1016–1018, antiphase gain and index gratings,” IEEE J. Quantum Electron., vol. 32,
1984. pp. 839–850, May 1996.
[11] H. Soda, H. Ishikawa, and H. Imai, “Design of DFB lasers for high- [25] S. F. Yu, R. G. S. Plumb, L. M. Zhang, M. C. Nowell, and J. E.
power single mode operation,” Electron. Lett., vol. 22, no. 20, pp. Carroll, “Large-signal dynamic behavior of distributed-feedback lasers
1047–1049, 1986. including lateral effects,” IEEE J. Quantum Electron., vol. 30, no. 8, pp.
[12] M. Okai, “Spectral characteristics of distributed feedback semicon- 1740–1750, Aug. 1994.
ductor lasers and their improvements by corrugation-pitch modulated
structure,” J. Appl. Phys., vol. 75, no. 1, pp. 1–29, 1994.
[13] J. Salzman, H. Olesen, A. Moller-Larsen, O. Albrektsen, J. Hanberg, J.
Norregaard, B. Jonsson, and B. Tromborg, “Distributed feedback lasers
with an S-bent waveguide for high-power single-mode operation,” IEEE Thierry Fessant, photograph and biography not available at the time of
J. Select. Topics Quantum Electron., vol. 1, npp. 346–355, Feb. 1995. publication.

You might also like