Note: Descriptions are shown in the official language in which they were submitted.
CA 02537358 2006-02-20
V-COUPLED-CAVITY SEMICONDUCTOR LASER
RELATED APPLICATIONS
This application is a continuation-in-part of Canadian Patent Application
No. 2,521,731, filed on October 1 1, 2005, and claims benefit from U.S. Patent
Application Ser. No. 1 1 /306,520, filed on December 30, 2005.
FIELD OF THE INVENTION
This invention relates generally to a semiconductor laser, and more
particularly to a monolithically integrated single-mode wavelength-switchable
semiconductor laser.
BACKGROUND OF THE INVENTION
Widely tunable lasers are of great interest for both long-haul and
metropolitan optical networks. Besides their use for source sparing with the
advantages of reduced inventory and cost, they open the possibility of new
system
architectures with more efficient and more flexible network management. For
example, the combination of tunable lasers with wavelength routers can provide
large format-independent space switches and reconfigurable optical add/drop
functions.
Monolithically integrated semiconductor tunable lasers offer many
advantages over external-cavity tunable lasers assembled from discrete
components. They are compact, low-cost, and more reliable as they contain no
moving parts. A conventional monolithic tunable laser usually comprises a
multi-
electrode structure for continuous tuning. Fig. 1 shows a prior-art example of
a
semiconductor tunable laser consisting of a distributed Bragg reflector (DBR)
CA 02537358 2006-02-20
grating, an active gain section, and a phase shift region. An electrode for
electrical control is disposed on top of each of the three sections. When the
reflection peak wavelength of the DBR grating is tuned by injecting current or
applying an electrical voltage, the phase shift region must be adjusted
simultaneously in order to prevent the laser from hopping from one mode to
another. Besides, the tuning range of such a laser is limited to about 1 Onm
due to
the limitation of commonly achievable refractive index change in semiconductor
materials.
A more sophisticated tunable laser with a wider tuning range and improved
performances was described by V. jarayman, 1. M. Chuang, and L. A. Coldren, in
an article entitled "Theory, design, and performance of extended tuning range
semiconductor lasers with sampled gratings", IEEE ~. Quantum Electron. Vol.
29,
pp. 1824-1834, 1993. It comprises of four electrodes controlling two sampled
grating distributed Bragg reflectors, a phase-shift region and a gain section.
The
wavelength tuning requires complex electronic circuits with multidimensional
current control algorithms and look-up tables. Such complexity reduces the
fabrication yield and increases the cost, and also opens the questions about
the
manufacturability and long term stability of the devices.
A widely tunable or wavelength switchable laser can also be realized by
using two coupled cavities of slightly different lengths. The tuning range is
greatly
increased by using the Vernier effect. The coupled-cavity laser can be
fabricated
either by etching a groove inside a cleaved Fabry-Perot laser, as described in
a
paper entitled "Monolithic two-section GaInAsP/InP active-optical-resonator
devices formed by reactive-ion-etching", by L. A. Coldren et al, Appl. Phys.
Lett.,
2
CA 02537358 2006-02-20
vol. 38, pp. 315317, 1981, or by using a cleaved-coupled-cavity structure, as
described in a paper entitled "The cleaved-coupled-cavity (C3) laser", by W.
T.
Tsang, Semiconductors and Semimetals, vol. 22, p. 257, 1985. However, the
performance of the prior-art coupled-cavity lasers especially in terms of mode
selectivity is not satisfactory, which results in very limited use for
practical
applications.
In co-pending US patent application Ser. No. 10/908,362 and Canadian
patent application No. 2,521,731, both entitled "Wavelength switchable
semiconductor laser", a serially coupled cavity structure with an optimal
coupling
coefficient achieved by using a coupling cavity for improved mode selectivity
is
disclosed for operation as a wavelength switchable laser. The optimal coupling
coefficient between two serially coupled cavities is achieved by using
multiple
reflection trenches producing desired reflectivities and transmission
coefficients.
The present patent application discloses a new parallelly coupled cavity
structure
for achieving a high single-mode selectivity.
Coupled-cavity lasers have also been previously investigated in the form of
a Y-laser, as described in an article entitled "The Y-laser: A Multifunctional
Device
for Optical Communication Systems and Switching Networks", O. Hildebrand, M.
Schilling, D. Baums, W. Idler, K. butting, G. Laube, and K. Wunstel, journal
of
Lightwave Technology, vol. 1 1, no. 2, pp. 2066-2074, 1993, and the references
therein. Fig. 2 shows a schematic diagram of the Y-coupled-cavity laser. The Y-
laser has the advantage of being monolithic without the challenging
fabrication
requirement for deeply and vertically etched trenches. However, the mode
selectivity of the Y-laser is very poor, with a side-mode threshold difference
of
3
CA 02537358 2006-02-20
only about 1 cm-1 for a 450pm laser, compared to over 10 cm-1 for a typical
DFB
laser. This is far from sufficient for stable single-mode operation,
especially when
the laser is directly modulated.
The present invention provides a substantially improved structure using V-
coupled cavities. This design allows an optimized coupling coefficient between
two cavities to be realized to achieve significantly improved single-mode
selectivity, while allowing the lasing wavelength to be tuned over a wide
range.
For many applications, it is not necessary to tune the laser wavelength
continuously. Rather, it is only required that the laser can be set to any
discrete
wavelength channel, e.g. as defined by the ITU (International
Telecommunication
Union). Such applications include linecard sparing, wavelength routing and
optical
add/drop. Key requirements for such wavelength switchable lasers are: 1 ) an
accurate match of the discrete operating wavelengths with the predefined
wavelength channels (e.g. ITU grid); 2) simple and reliable control for the
switching between various wavelength channels; 3) high side-mode suppression
ratio and low crosstalk; 4) fast switching speed; and 5) easy to fabricate and
low
cost.
It is an object of the present invention to provide a monolithically
integrated
single-mode wavelength-switchable semiconductor laser that satisfies all of
the
above requirements.
SUMMARY OF THE INVENTION
In accordance with the invention, there is provided, a semiconductor laser
comprising:
4
CA 02537358 2006-02-20
a first optical cavity comprising a first optical waveguide bounded by two
partially reflecting elements, at least a portion of said first optical
waveguide being
pumped to provide optical gain,
a second optical cavity comprising a second optical waveguide bounded by
two partially reflecting elements, at least a portion of said second optical
waveguide being pumped to provide optical gain,
whereas the first and the second optical waveguides have different optical
path lengths, and are disposed side-by-side on a substrate to form
substantially
V-shaped waveguide branches with substantially no cross-coupling at the open
end and a predetermined cross-coupling at the closed end for achieving a
substantially optimal single-mode selectivity of the laser.
In accordance with another embodiment of the invention, there is provided,
a semiconductor laser comprising:
a Mach-Zehnder interferometer comprising a first and a second waveguide
arms disposed side-by-side on a semiconductor substrate, said first and second
waveguide arms having an uncoupled region between two coupling regions,
a first partially reflecting trench etched through the first waveguide arm in
the uncoupled region,
a second partially reflecting trench etched through the second waveguide
arm in the uncoupled region,
a third partially reflecting trench etched through both the first and the
second waveguide arms in one of the coupling regions to form a first optical
cavity
in the first waveguide arm in conjunction with the first partially reflecting
trench
and a second optical cavity in the second waveguide arm in conjunction with
the
5
CA 02537358 2006-02-20
second partially reflecting trench, said third partially reflecting trench
being
positioned such that the first and the second optical cavities have a
predetermined
amount of cross-coupling for achieving a substantially optimal single-mode
selectivity of the laser.
BRIEF DESCRIPTION OF THE DRAWINGS
Fig. 1 is a prior art semiconductor tunable laser consisting of a DBR grating,
an active gain section, and a phase-shift region.
Fig. 2 is a prior-art Y-coupled-cavity laser.
Fig. 3 is the top view of a monolithic wavelength switchable V-coupled-
cavity laser in accordance with one embodiment of the present invention.
Fig. 4(a) is a typical cross-sectional view perpendicular to the waveguides of
the laser.
Fig. 4(b) is the cross-sectional view along a waveguide showing the shallow
isolation trench and the deeply etched partially-reflecting trench used in
some
embodiments of the present invention.
Fig. 5 is a schematic diagram showing the relationships between the two
sets of resonant peaks of the fixed gain cavity and the channel selector
cavity, and
the material gain spectrum.
Fig. 6 is the effective reflection factors of the wavelength selector cavity
(dotted line) and the fixed gain cavity (solid line) as a function of the
wavelength
when the laser is pumped at the threshold.
Fig. 7 is the small signal gain spectra of the laser near its threshold when
the signal is transmitted through the fixed gain cavity (dotted line) and
through
the wavelength selector cavity (solid line).
6
CA 02537358 2006-02-20
Fig. 8 is the spectra of the effective reflection factor for cross-coupling
coefficient values of 0.1 (solid line) and 0.5 (dotted line).
Fig. 9 is the lasing threshold of the cavity modes for two different cross-
coupling coefficient values of 0.1 (circles) and 0.5 (crosses).
Fig. 10(a) is the threshold gain coefficient of the lowest threshold mode
(solid line) and the next lowest threshold mode (dotted line) as a function of
the
cross-coupling coefficient.
Fig. 10(b) is the threshold difference between the lowest threshold mode
and the next lowest threshold mode.
Fig. 1 1 (a) is the threshold gain coefficient of the lowest threshold mode
(solid line and dashed line) and the next lowest threshold mode (dotted line
and
dash-dotted line) as a function of the cross-coupling coefficient for two
different
pumping conditions corresponding to gL = g'L' and g = g'. The cavity length
difference is doubled compared to the example of Fig. 10(a).
1 5 Fig. 1 1 (b) is the threshold difference between the lowest threshold mode
and the next lowest threshold mode for the cases gL = g'L' (solid line) and g
= g'
(dashed line). The cavity length difference is doubled compared to the example
of
Fig. 10 (b).
Fig. 12 is a schematic diagram of a Y-coupled-cavity laser for comparison
analysis with the V-coupled cavity of the present invnetion.
Fig. 13(a) is the threshold gain coefficient of the lowest threshold mode
(solid line) and the next lowest threshold mode (dotted line) as a function of
the
Y-coupling coefficient of the Y-laser.
7
CA 02537358 2006-02-20
Fig. 13(b) is the threshold difference between the lowest threshold mode
and the next lowest threshold mode of the Y-laser.
Fig. 14 is a preferred embodiment of the present invention.
Fig. 15 is the reflectivity and transmission coefficients of an air trench as
a
function of the trench width at 1550nm wavelength.
Fig. 16 is another embodiment of the present invention where the laser has
both wavelength switching and space switching functionalities.
Fig. 17 is another embodiment of the present invention where the V-
coupled-cavity laser is used as a switchable wavelength converter.
DETAILED DESCRIPTION
Fig. 3 shows the top view of a monolithic wavelength switchable V-coupled-
cavity laser in accordance with one embodiment of the present invention. The
laser consists of two waveguides branches 101 and 102 forming a substantially
V-
shaped geometry, disposed on a semiconductor chip 50. They are closely spaced
or touching on one end (i.e. the closed end) but far apart on the other end
(i.e. the
open end). Bounded by cleaved facets 10 and 20 (or etched trenches in an
alternative embodiment), each of the two waveguide branches forms a Fabry-
Perot
cavity. Near the end facet 20 where the two waveguides are closely spaced or
touching, a small amount of light is coupled from one cavity to the other, due
to
evanescent coupling or overlap of mode fields. An electrode 121 is deposited
on
top of the waveguide branches 101 while two electrodes 122a and 122b are
deposited on top of the waveguide branches 102. In operation, the electrodes
121
and 122a are injected with essentially fixed currents to produce optical gain
for
8
CA 02537358 2006-02-20
the laser, while the electrode 122b is used to apply a variable current or
voltage in
order to change the refractive index of the underlying waveguide and to switch
the
laser wavelength. The variable-index waveguide segment under the electrode
122b is preferably away from the coupling region so that its refractive index
change does not affect the coupling coefficient between the two cavities.
A typical cross-sectional view perpendicular to the waveguides of the laser
is shown in Fig. 4(a). The waveguide structure generally consists of a buffer
layer
1 16, waveguide core layer 1 14 that provides an optical gain when
electrically
pumped (with a current injection), and an upper cladding layer 1 12, deposited
on
a substrate 118. The backside of the substrate is deposited with a metal
electrode layer 120 as a ground plane. Preferably the waveguide core layer
comprises multiple quantum wells and the layers are appropriately doped as in
conventional laser structures. An example material system is InGaAsP/InP. In
the
transverse direction, standard ridge or rib waveguides 101 and 102 are formed
to
1 S laterally confine the optical mode, with electrode 121 and 122 deposited
thereon.
The structure may further include a current blocking layer to improve the
electrical
characteristics.
The waveguide core in the wavelength switching segment under the
electrode 122b preferably has a larger bandgap energy than that of the gain
segments under the electrodes 121 and 122a. This allows a large refractive
index
change to be obtained at the laser wavelength when an electrical current or
voltage is applied on the electrode 122b, without introducing a significant
gain
variation. The different bandgap energies in different sections of the
monolithic
device can be obtained by using a quantum well intermixing technique, or by
9
CA 02537358 2006-02-20
using an etch-and-regrowth method. The electrodes 122a and 122b are separated
by a shallow isolation trench, as shown in Fig. 4(b). Alternatively, a
simplified
embodiment may be envisaged by combining the segments 122a and 122b into a
single gain segment with a variable current injection to provide a gain and a
refractive index change simultaneously.
In accordance with one aspect of the present invention, the length of the
fixed gain cavity 101 is chosen so that its resonance frequency interval
matches
the spacing of the operating frequency grid, an example being the widely used
frequency grid defined by ITU (e.g. spaced at 200GHz,1 OOGHz or 50GHz). The
resonance frequency interval is determined by
2n L
where c is the light velocity in vacuum, n9 the effective group refractive
index of
the waveguide, and L the length of the fixed gain cavity 101.
Similarly, the resonance frequency interval 4f ~ of the second cavity
1 S comprising the waveguide branch 102 (hereafter referred as the channel
selector
cavity) is determined by
,_ c _ c
~ 2n~'L' 2(nGLU +nhLb) (2)
where La and Lb are the lengths of the segments 122a and 122b, respectively,
na
and n6 are the effective group refractive indices of the segments 122a and
122b,
respectively, L' = La + Lb is the total length, and n'9 = (naLa +nbLb )/ L' is
the
average effective group refractive index of the channel selector cavity 102.
CA 02537358 2006-02-20
The resonance frequency interval of ' of the channel selector cavity is chosen
to be slightly different than of so that only one resonant peak coincides with
one
of the resonant peaks of the fixed gain cavity over the material spectral gain
window, as shown in Fig. 5. The distance between two aligned resonant peaks,
which corresponds to the free spectral range (FSR) of the combined cavity, is
determined by
4.f4.~~ ~3)
In order not to have two wavelengths lasing simultaneously, of~ should
generally be larger than the spectral width of the material gain window.
The resonant frequencies of the fixed gain cavity and the channel selector
cavity are determined respectively by
_ me 4a
2nL
,_ m'c 4b
2n' L'
where m and m' are integers, n and n' the averaged effective refractive index
of the
1 S waveguide within the respective cavity, L and L' the cavity lengths. The
frequency
of the channel selector can be tuned by varying the effective index n' or n6
of the
segment 122b. The rate of the tuning is determined by
8.~~ ~rt~ BnbLh
f., _ - n. _ - nbL, ~5)
Since the laser frequency is determined by the resonant peak of the fixed
gain cavity that coincides with a peak of the channel selector cavity, a shift
of ~Of-
Of '~ in the resonant peaks of the channel selector cavity results in a jump
of a
CA 02537358 2006-02-20
channel in the laser frequency. Therefore, the change of the laser frequency
with
the refractive index variation is amplified by a factor of of /oaf-Of y, i.e.,
(6>
The increased tuning range is one of the advantages of the device of the
present invention. Consider an example in which Of =100GHz, and ~f'=90GHz,
the range of the laser frequency variation is increased by a factor of 10 with
respect to what can be achieved by the index variation directly. For this
numerical
example, assuming the effective group refractive index of the waveguide is
3.215,
the lengths of the fixed gain cavity and the channel selector cavity are
L=466.24pm and L'=518.31 um, respectively. The device length is therefore
comparable to conventional DFB or FP lasers.
An important difference between the V-coupled-cavity laser of the present
invention and the prior-art Y-coupled-cavity laser is that the light is
partially
coupled from one branch to the other without going through a common
waveguide section. As we will see later, this allows an optimal amount of
light to
be coupled from one cavity to the other (i. e, cross-coupling), relative to
the
amount of light coupled back to the same cavity (i. e. self-coupling). As a
result, a
much higher single-mode selectivity can be achieved with the V-laser as
compared with the Y-laser.
Referring again to Fig. 3, assume the amplitude reflectivities of the cleaved
facets 10 and 20 are r~ and r2, respectively. The coupling between the
waveguides
occurs near the facet 20. We denote the amplitude coupling coefficients from
waveguide 101 to waveguide 102 (cross-coupling), from waveguide 101 back to
waveguide 101 (self-coupling), from waveguide 102 to waveguide 101 (cross-
12
CA 02537358 2006-02-20
coupling), and from waveguide 102 back to waveguide 102 (self-coupling), as
C~z,
C> >, Cz~ and Czz, respectively. For simplicity, we assume there is no excess
coupling loss. Therefore, we have ~ C> > ~z+~ C~z~z=1 and ~ Cz~ ~z+~ Czzlz=~ .
From
the reciprocity of light wave propagation, we also have C~z=Cz~. Note that the
facet reflectivity is treated separately and is not included in the coupling
coefficients.
In the following section, for simplicity and without losing generality, we
will
treat the waveguide 102 as a uniform waveguide with an average effective
refractive index of n'. To analyze the V-coupled-cavity laser, we consider one
of
the Fabry-Perot cavities as the main laser cavity and include the coupling
effect of
the other cavity in one of the mirrors of the main cavity. First, let us
consider the
fixed gain cavity comprising waveguide 101 as the main cavity. The effective
reflectivity of the facet 20 for this cavity can be written as rze= nrz, where
n is an
effective reflection factor (in amplitude) taking into account the coupling
effect of
the channel selector cavity comprising waveguide 102 and is calculated by
~l =C" +Cz'C'zrirze2~g'+~k~>c~l+Cz2rlrZeZ~g'+~x~>c +C2 r12r2 ea~g'+rx~>c~
+...)
C C r r e2~g'+'k~~~~
=C + z' '2' z
" 1- C r r e2~~'+ix~~L~
z2 ' z
The threshold condition can therefore be written as
r y7~"zez~K+~x>~ -1
In the above equations, k (=2~n/~) and g are, respectively, the propagation
constant and gain coefficient of the waveguide 1 O1, and k' (=2~n'/a) and g'
are,
respectively, the average propagation constant and average gain coefficient of
the
waveguide 102.
13
CA 02537358 2006-02-20
Similarly, we can also consider the channel selector cavity comprising
waveguide 102 as the main laser cavity. The effective reflectivity of the
facet 20
for this cavity can be written as r'ze= n'rz, where n' is the effective
reflection factor
taking into account the coupling effect of the fixed gain cavity and is
calculated by
~7'=C +C C rr ezc~+~k~L(1+C rr ezcg+rk>~ +Czrzrzeacg+~x~c +...)
22 21 12 l 2 ll 1 2 11 1 2
C C r r ez~g+rk>~
= C, + z1 Iz 1 z /9)
zz 1 _ C r r ez~g+rk»
11 1 z
The threshold condition for the laser can then be written as
rl~~rzezcx~+~xoc =1 (10)
After some manipulation, it can be shown that Eqs. (8) and (10) are identical.
They
both lead to the following threshold condition of the V-coupled-cavity laser,
Cllrlrzezc,~+rk~c +Czzrlrzez~K~+rk~>c _(CIICzz -CIZCzI)rlzrz
ezcg+tk~~ez~g~+rx~>c =1 (11)
This complex equation, which can be separated into two equations corresponding
to the real and imaginary parts, determines the wavelengths of the lasing
modes
as well as their corresponding threshold gain coefficients. In the case of
1 S uncoupled cavities, i.e., C~ z=Cz~ = 0 and C> > =Czz = 1, we have n = n' =
1, and
Eqs. (8), (10) and (1 1 ) reduce to the threshold conditions of conventional
Fabry-
Perot cavities.
Now we use a numerical example to illustrate the characteristics of the V-
coupled-cavity laser. Consider the previously mentioned example where n =
n'=3.21 5, L=466.24~m (~f =1 OOGHz), and L'=518.31 ~m (Of'=90GHz). The two
cavities have a common resonance wavelength at 1 550.12nm, corresponding to a
frequency of 193400GHz. Assume C»=Czz=0.95, C~z=Cz~=0.31, and the
reflecting mirrors of the cavities are formed by cleaved facets, i.e., r~ = rz
= (n-
1 )/(n+1 )=0.5255. Also assume that the two cavities are pumped to produce the
14
CA 02537358 2006-02-20
same round trip gain, i.e., gt_ = g'L'. For the lowest threshold mode at the
common
resonance wavelength of 1550.12nm, solving equation (1 1 ) leads to the
threshold
gain coefficients (in intensity) G = 2g = 22.6cm-' and G' = 2g' = 20.4 cm-~.
These
compare to the threshold gain coefficients of uncoupled cavities of Go =
27.6cm-~
and G'o = 24.8cm-', respectively.
The mode selectivity and the wavelength switching function of the V-
coupled-cavity laser can be understood from the effective reflection factors n
and
n', which are wavelength dependent functions with sharp resonant peaks. Fig. 6
plots their squared modulus Inlz (dotted line) and In'12 (solid line), which
are the
effective reflection factors in intensity, as a function of the wavelength
when the
laser is pumped at the threshold. The periodic peaks of the effective
reflection
factor Inl2 occur at the resonant wavelengths of the wavelength selector
cavity
(waveguide 102). The effective reflection factor Inl2 effectively modifies the
reflectivity of one of the mirrors of the fixed gain cavity (waveguide 1 O1 ),
producing a comb of reflectivity peaks similar to those of a prior-art sampled
DBR
grating (but without the fabrication complexity and long device length
associated
with it). Consequently, a resonant mode of the fixed gain cavity that
coincides with
one of the peaks of the effective reflection factor Inl2 is selected as the
lasing
mode. Since the periodic peaks of the effective reflection factor In'IZ
correspond to
the resonant wavelengths of the fixed gain cavity, the lasing wavelength
occurs at
the position where a peak of Inlz overlaps with a peak of In'I2.
Fig. 7 shows the small signal gain spectra of the laser near its threshold
when the signal is transmitted through waveguide 101 (dotted line) and through
the waveguide 102 (solid line). We can see that there are multiple lasing
CA 02537358 2006-02-20
wavelengths separated by about 7.2 nm (900 GHz) as determined by Eq. (3). The
variation of the material gain has not been considered in this calculation. In
practice, the laser is limited to a single mode due to the limited material
gain
window and/or by adding an intracavity filter or a thin-film facet coating
with a
limited passband.
The mode selectivity, which relates to side-mode suppression ratio (SMSR)
or channel crosstalk in the case of wavelength switchable laser, is an
important
consideration in the design of the laser. The mode selectivity can be
optimized by
appropriately choosing the cross-coupling coefficient (i.e., IC~ZIz = ICz~ Iz,
in
intensity). To illustrate the effect of the cross-coupling coefficient, we
calculate
the effective reflection factor Inl2 as a function of wavelength for different
values
of the cross-coupling coefficient. Fig. 8 compares the spectra of the
effective
reflection factor Inlz for cross-coupling coefficient values of 0.1 (solid
line) and
0.5 (dotted line), when the laser is pumped at the corresponding lasing
threshold
of G = 22.6cm-~, and G = 20.1 cm-', respectively. We can see that when the
cross-
coupling coefficient increases, the peaks of the effective reflection factor
Inlz,
which is proportional to the effective reflectivity of one of the mirrors of
the fixed
gain cavity, become narrower while the contrast decreases.
Since the discrimination of side modes is based on the misalignment of
resonant modes between the fixed gain cavity and the channel selector cavity,
the
narrower the effective reflectivity peaks, the better the mode selectivity,
especially
for adjacent modes. Quantitatively, the mode selectivity can be characterized
by
threshold difference between the side modes and the main mode. Fig. 9 shows
the
lasing threshold of the cavity modes for two different cross-coupling
coefficient
16
CA 02537358 2006-02-20
values of 0.1 (circles) and 0.5 (crosses). The threshold difference between
the
lowest threshold mode (the main mode) and the next lowest threshold mode is
about 7.4 cm-' when the cross-coupling coefficient is 0.1, but is only about
1.2
cm-' when the cross-coupling coefficient is 0.5. The threshold (G = 2g) of the
main mode is 22.6cm-' and 20.1 cm-' for the two cases, respectively.
In Fig. 10(a), we plot the threshold gain coefficient G (= 2g) of the lowest
threshold mode (solid line) and the next lowest threshold mode (dotted line)
as a
function of the cross-coupling coefficient (~C~z~z = ~Cz~ ~z). The threshold
difference between the two modes is plotted in Fig. 10(b). We can see that a
cross-
coupling coefficient around 0.1 gives the largest threshold difference,
although a
cross-coupling coefficient around 0.5 produces the lowest threshold for the
main
mode. The threshold difference increases as the cross-coupling coefficient
decreases from 1 to 0.1, because the peak width of the effective reflection
factor
Inlz decreases, resulting in an increased selectivity between the main mode
and its
adjacent modes. As the cross-coupling coefficient further decreases to below
0.1,
the threshold difference decreases. This is because the peak width of the
effective
reflection factor Inlz becomes narrower than the mode spacing and it no longer
affects the threshold difference. Instead, the threshold difference is
determined by
the contrast in the effective reflection factor p2 which decreases with the
decreasing cross-coupling coefficient.
By increasing the length difference between the fixed gain cavity and the
channel selector cavity, the threshold difference between the lowest threshold
mode and the next lowest threshold mode can be increased, at the expense of
reduced free spectral range as determined by Ep. (3). It is also found that
the
17
CA 02537358 2006-02-20
maximum threshold difference is achieved when the simple round trip gains in
the
two cavities (without considering cross-cavity interferences) are equal, i.e.
gL =
g'L' (assuming the cavity mirrors have the same reflectivities) or r~ r2ez9~ =
r~'r2'e2g'~,
Consider the case L=466.24um (0f =1 OOGHz) and L'=570.38~m (~f'=81.7GHz)
with other parameters the same as in the previous example. Fig. 1 1 (a) shows
the
threshold of the lowest threshold mode (solid line and dashed line) and the
next
lowest threshold mode (dotted line and dash-dotted line) as a function of the
cross-coupling coefficient for two different pumping conditions corresponding
to
gL = g'L' and g = g'. Fig. 1 1 (b) gives the threshold difference between the
two
modes for the cases gL = g'L' (solid line) and g = g' (dashed line). Compared
to the
previous example, the cavity length difference is almost doubled, and the free
spectral range of~ is reduced from 900GHz to 446 GHz. The maximum achievable
threshold difference is increased from 7.4 cm-' to 14.5 cm-' (for the case gL
=
g'L'), and the optimum cross-coupling coefficient at which the maximum
threshold
1 5 difference is achieved is increased from 0.1 to 0.26.
As it can also be seen in Fig. 1 1, the maximum achievable threshold
difference decreases as the pumping condition deviates away from the optimal
condition of gL = g'L' (or more generally, equal round trip gains). Therefore,
it is
preferable that gain variations be avoided when the refractive index of the
channel
selector cavity is changed to switch the laser wavelength. This can be
realized by
using a separate tuning section 122b in the channel selector cavity which is
substantially passive (with little gain or loss), as shown in the embodiment
of Fig.
3. This also allows flexible output power control independent of the
wavelength
switching.
18
CA 02537358 2006-02-20
It can also be envisaged to divide the waveguide 101 into two sections, each
with a separate electrode, one for providing a fixed gain to the laser, and
the other
for continuously tuning the resonant wavelength grid of the "fixed" gain
cavity
over a channel spacing. By combining this wavelength tuning with the
wavelength
switching of the wavelength selector cavity, a continuous tuning of the laser
wavelength over a wide wavelength range can be achieved. This tunability can
also
be used to slightly adjust channel wavelengths to compensate for imperfect
match
between the resonant frequencies of the fixed gain cavity and the frequency
grid
of the operating channels.
Now let us examine the performance difference between the V-coupled-
cavity laser of the present invention and the prior-art Y-coupled-cavity
laser.
Consider a Y-coupled-cavity laser as shown in Fig. 12 with two waveguide
branches A, B and a common waveguide section C of length La, Lb and L~,
respectively. The electrodes on the three waveguide sections are separated by
1 5 shallow isolation trenches. The amplitude coupling coefficients from the
common
waveguide section C to waveguide A and B are denoted as C~ and Cz,
respectively.
The amplitude coupling coefficients from waveguide A and B to the common
waveguide section C are denoted as C'~ and C'z, respectively. From the
reciprocity
of light wave propagation, we have C~ = C'~, and Cz = C'z. Assuming there is
no
coupling loss, we have I C~ Iz+I CZIz=1.
Similar to the previous analysis for the V-coupled-cavity laser, we consider
the fixed gain cavity comprising the waveguide sections A and C as the main
laser
cavity whose threshold condition can be written as
TZ'r r gzc~~+rka~L~ ea(g~.+rk ~c~ = 1
19
CA 02537358 2006-02-20
where g and k are the gain coefficient and the propagation constant with the
subscript indicating the corresponding waveguide section, r~ and rz are the
amplitude reflectivities of the cleaved facets (or cavity mirrors), T' = C'~
is the
amplitude transmission coefficient from the waveguide section A to waveguide
section C, and ris the effective amplitude transmission coefficients from the
waveguide section C to waveguide section A. The coupling effect of the channel
selector cavity comprising the waveguide sections B and C is included in the
effective transmission coefficient rwhich is calculated by
, 2(gn+ikh)Lh 2(g~+ik~)L~ 2 '2 2 2 4(gh+ikn)L,, 4(g~+ik'.)L.
z = C~ + C1 CZ C Z r~Y2e a + Ci CZ C 2 r'i r2 a a
__ C1
1-C C' rr e2(gb+ikb)Lbez(g~+ik~.)L~ 13
2 2 1 2
Substituting the expression of rand T' into Eq. (12) and after some
manipulation,
we obtain the following threshold condition of the Y-coupled-cavity laser,
CIC'~ rl~.Ze2(g+ik)L +CZC'2 r~rzez(g'+ik')L~ =1 (14)
where L = La + L~ and L' = Lb + L~ are the total lengths of the fixed gain
cavity and
the wavelength selector cavity, respectively; g = (gala + g~L~)/L and g' _
(gbLb +
g~L~)/L' are the average gain coefficients; and k = (kaLa + k~L~)/L and k' _
(kbLb +
k~L~)/L' are the average propagation constants.
Using the same parameters as those in the example for Fig. 10, i.e.,
L=466.24pm, and L'=518.31 pm. Also assume that the two cavities are pumped to
produce the same round trip gain, i.e., gL = g'L'. Fig. 13(a) shows the
threshold
gain coefficient G (= 2g) of the lowest threshold mode (solid line) and the
next
lowest threshold mode (dotted line) as a function of the Y-coupling
coefficient
~Cz~z. The threshold difference between the two modes is plotted in Fig.
13(b). We
see that the threshold of the main lasing mode is G = 2g = 27.6cm-~ which is
CA 02537358 2006-02-20
independent of the coupling coefficient and is the same as the uncoupled
cavity of
the same length L. The largest threshold difference is only 1.2 cm-' and it
occurs
at ~Cz~z = 0.5, i.e., when the Y-branch is an equal power splitter. These
compare
to a threshold gain coefficient of 22.6cm-~ and a maximum achievable threshold
difference of 7.4 cm-' for the V-coupled-cavity laser of the same cavity
lengths
with a cross-coupling coefficient of 0.1. Therefore, the V-coupled-cavity
laser of
the present invention can produce a much larger side-mode threshold difference
while its lasing threshold for the main mode is lower.
In terms of the threshold condition, the Y-coupled-cavity laser is equivalent
to a special case of the V-coupled-cavity laser when C»Czz - C~zCz~ = 0. The
coupling coefficients between the cavities can be related to the coupling
coefficients of the Y-branch by C» = C~C'~, Czz = CZC'z, C~z = C'~Cz, and Cz~
_
C'zC~. Substituting these equations into Eq. (1 1 ), one can obtain Eq. (14).
One can
also see that, no matter what is the splitting ratio of the Y-branch, the
equation
C»Czz - C~zCz~ = 0 always holds. This limitation is removed for the V-coupled-
cavity of the present invention, which allows the cross-cavity coupling
coefficient
to be optimized, resulting in a much larger side-mode threshold difference,
i.e. a
much higher single-mode selectivity.
The V-coupled-cavity laser of the present invention as depicted in Fig. 3
may be mirrored with respect to the facet 20 to form a substantially X-shaped
geometry with some similar characteristics. It may also be mirrored with
respect to
the facet 10 to form a substantially triangular or diamond shaped geometry
with
similar characteristics.
21
CA 02537358 2006-02-20
A disadvantage of the embodiment of Fig. 3 is that the output power is
emitted from two waveguide branches on both sides of the laser cavity,
reducing
the power utilization efficiency when the light is coupled to a single mode
fiber.
This disadvantage can be overcome by using an external combiner outside of the
laser cavity, either in a monolithic integrated manner or as a separate
discrete
component. Fig. 14 shows a preferred embodiment of the present invention. The
V-coupled-cavity laser as depicted in Fig. 3 is now terminated by the deep-
etched
trenches 10 and 20 which act as partially reflecting mirrors for the laser
cavities.
The cross-section of the deep trench is shown in Fig. 4(b). It has
substantially
vertical sidewalls etched through the waveguide layer 1 14. The Y-branches 130
and 140 combine the light emitted from the two waveguides 101 and 102 into a
single mode waveguide on each side of the laser. The output port, to be
coupled
to a fiber, can be chosen either on the side of the combiner 130 or on the
side of
the combiner 140 (or both). In the case the combiner 130 is chosen as the
output
1 5 port, as depicted in the figure, the facet 30 needs to be anti-reflection
(AR) coated
(or angled) so that no reflection from the facet 30 is fed back to the laser
cavity.
The combiner 130 is made substantially transparent either by electrical
pumping
or by waveguide modification using the quantum-well intermixing or etch-and-
regrowth technique. On the other side of the laser, the combiner 140 may be
reverse-biased to act as an on-chip power monitor. In this case, the light
transmitted through the trench 20 is substantially absorbed by the combiner
140
before reaching the facet 40 so that no reflection from the facet 40 is fed
back to
the laser cavity.
22
CA 02537358 2006-02-20
The reflectivity of the deeply etched trench varies with the trench width.
Fig.
15 shows the reflectivity and transmission coefficient of the air trench as a
function of the trench width at 1550nm wavelength. In order for the deeply
etched
trench to act as a high-reflectivity mirror, the trench width is preferably
designed
to be substantially equal to an odd-integer multiple of quarter-wavelength,
i.e.,
~/4, 3~/4, 5~/4, ...etc., especially for the back reflector (trench 20) for
which a
high reflectivity is desirable. A higher reflectivity of the etched trenches
results in
a lower lasing threshold, without reducing the side-mode threshold difference
of
the V-coupled-cavity laser.
Preferably, the Y-branched combiner outside of the laser cavity has
substantially equal arm lengths so that the light beams emitted from the two
waveguides 101 and 102 of the laser are in phase and constructive for the main
lasing mode at the output port, regardless of the wavelength switching from
channel to channel within the same free spectral range (FSR) as determined by
Eq.
(3). The resonance condition Eq. (4) of the laser cavities ensures that the
optical
path length difference between the waveguide 101 and 102 within the laser
cavity
is an integer multiple of half-wavelength. At the coupling end of the laser
(i.e. at
the deep trench 20 of Fig. 14), the phases of the propagating fields from the
two
waveguide braches 101 and 102 are substantially the same. By setting the
optical
path length difference (electrically adjustable via the electrode 122b) to be
an even
integer multiple of half-wavelength when the resonant frequencies of both
cavities
are aligned for the central channel (fo in Fig. 5), the light beams emitted
from the
two waveguides 101 and 102 of the laser are in phase at the output facet of
the
combiner 130. At the same time, for the next substantially aligned resonant
23
CA 02537358 2006-02-20
frequency (f, in Fig. 5) located at one FSR away, the optical path length
difference
is automatically an odd integer multiple of half-wavelength. As a result, the
beams
emitted from the two waveguides 1 O1 and 102 of the laser are out of phase and
destructive at the output facet of the combiner 130. Therefore, the
interference
effect of the external combiner also helps to filter out the unwanted adjacent
aligned resonant modes.
To avoid interference between the signal modulation and the wavelength
switching mechanism of the laser, the laser is preferably operated in
continuous
wave (CW) mode. An extra-cavity modulator is used instead of direct modulation
to modulate a data signal onto the laser beam. The extra-cavity modulator may
be
monolithically integrated with the laser in the form of electro-absorption
modulator (EAM). One of such embodiments is also represented by Fig. 14 where
the combiner 130 now simultaneously serves as the electro-absorption
modulator.
The bandgap energy of the waveguide material in the EAM section is preferably
slightly larger than that of the laser section so that its absorption
coefficient can
be changed significantly through the electro-absorption effect (or the quantum
confined Stark effect). Again, this bandgap difference can be achieved by
using the
quantum well intermixing, selected-area epitaxy, or etch-and-regrowth
technique.
Wavelength chirp is an important performance parameter for a
semiconductor laser source under high speed modulation. Although the extra-
cavity EAM improves the chirp performance considerably compared to direct
modulated laser, the chirp problem remains due to the fact that the EAM, whose
refractive index intrinsically changes with the modulation of absorption
24
CA 02537358 2006-02-20
coefficient, is placed in the optical path of the laser beam. The EAM also
introduces an extra insertion toss. In a co-pending US patent application Ser.
No.
60/622,326, entitled "Q-modulated semiconductor laser", a modulator comprising
an electro-absorptive waveguide placed inside an anti-resonant Fabry-Perot
cavity
which acts as the rear reflector of a semiconductor laser is described. Such
an
integrated Q-modulation mechanism can be implemented in the V-coupled-cavity
laser as depicted in Fig. 14 by using an electro-absorptive waveguide (with a
reverse-biased voltage or current injection) in the combiner 140 on the back
side
of the laser. In this embodiment the deep trench 20 and the cleaved facet 40
(or
another etched trench) constitute a Fabry-Perot cavity, which is preferably
substantially anti-resonant. The modulation of the absorption coefficient in
the
combiner 140 results in a modulation of the reflectivity of the rear reflector
of the
laser, and hence the Q-factor and threshold, and consequently the output
power.
Such an integrated Q-modulated laser has advantages of low wavelength chirp,
high speed and high extinction ratio, in addition to the high single-mode
selectivity and wavelength switchability offered by the V-coupled-cavity laser
of
the present invention.
Fig. 16 shows another embodiment of the present invention. The Y-
branched combiner of the previous embodiment is replaced by a 2x2 directional
coupler 135 with two output ports. By changing the optical path length
difference
between the waveguide 101 and 102 within the laser cavity (electrically
adjustable
via the electrode 122b) from an even to an odd integer multiple of half-
wavelength for the aligned resonant frequency (i.e. by shifting the resonant
frequency comb of the channel selector cavity by one frequency interval of '),
the
CA 02537358 2006-02-20
laser output is switched from one output port to another. Therefore, this
embodiment allows both wavelength switching and space switching
simultaneously, using the same control signal via the electrode 122b.
Alternatively, a separate electrode can be added in either waveguide branch
inside
or outside the laser cavities to perform the space switching function. It is
also
obvious to one skilled in the art that the directional coupler 135 can be
replaced
by a multimode interference coupler (MMI) with two output ports or other types
of
2x2 optical couplers.
In the embodiment of either Fig. 14 or Fig. 16, the laser can also be seen as
an internally sourced Mach-Zehnder interferometer with two waveguide arms cut
though by two deeply etched trenches which act as partially reflecting mirrors
and
form two coupled optical cavities. While one of the trenches is positioned in
the
uncoupled region of the waveguide arms, the other trench is positioned in the
coupling region of the waveguide interferometer so that the two optical
cavities
have a substantially optimal amount of cross-coupling for obtaining good
single-
mode selectivity of the laser. Obviously, the trench in the uncoupled region
may
be structured as two separate trenches placed at different longitudinal
positions,
each cutting through one of the waveguide arms.
Fig. 17 is yet another embodiment of the invention wherein the V-coupled-
cavity laser is used as a switchable wavelength converter. An input optical
signal is
injected into one arm of the V-coupled-cavity laser after being amplified by
an
on-chip amplifier section 150. The entrance facet of the amplifier is anti-
reflection coated so that it does not influence the laser cavity. The
wavelength-
converted output signal emits from the combiner 140 through its AR-coated
facet
26
CA 02537358 2006-02-20
40. Using the same wavelength switching mechanism as in previous embodiments,
the wavelength of the output signal may be varied over a wide wavelength
range.
The embodiment also has a photodetector 160 integrated on the chip as a power
monitor. Since the combiner 140 may also be designed as an electro-absorption
modulator, the wavelength-switchable laser may be modulated by either an
electrical signal or an input optical signal, adding more versatility to the
device.
The devices of the present invention have many advantages. In particular,
the achievable side-mode threshold difference of the V-coupled-cavity laser is
greatly increased compared to the prior-art Y-coupled-cavity laser, and it
becomes comparable to that of DFB lasers. The complex fabrication process for
DFB grating is not required and the wavelength can be switched over a wide
range.
Therefore, the laser of the present invention is low-cost and has high
functionality, and has a great potential for next-generation reconfigurable
optical
networks.
The present invention also applies to semiconductor lasers based on ring-
resonators or disc-resonators. The two resonators are closely placed or
partially
overlapped so they are coupled between each other. The waveguide lengths of
the
ring cavities (or the perimeters of the disc-resonators) and the optimal
coupling
coefficient are determined using the principle disclosed in the present
invention.
Numerous other embodiments can be envisaged without departing from the
spirit and scope of the invention.
27
