JP3251615B2 - Semiconductor laser device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates relates to a semiconductor laser device, about the particular semiconductor laser device whose wavelength is changed.

[0002]

2. Description of the Related Art In recent years, research and development of long-distance and large-capacity coherent optical communication systems have been actively conducted. Above all,
A coherent optical communication system using frequency multiplexing is promising because it can dramatically increase the amount of information transmission as compared with the related art. In this frequency multiplexing optical communication system, a semiconductor laser for local oscillation whose wavelength can be changed when a channel is selected on the receiving side is required. At this time, the larger the variable width of the wavelength, the more channels can be received.

[0003] A multi-electrode distributed black reflection (DBR) laser is capable of changing the Bragg wavelength greatly, and currently has the largest continuously variable amount. However, on the other hand, the line width increases to 10 to 20 MHz during wavelength tuning, and it has been difficult to apply the method to coherent optical transmission requiring a narrow line width. This is caused by carrier noise in the wavelength variable region.

On the other hand, a multi-electrode distributed feedback (DFB) laser has a narrow line width and is promising for coherent light transmission.
Conventionally, as discussed in MCWu et al .; CREO'90 P.667, a proposal has been made in which the active region has a uniform quantum well structure and carriers are injected from two electrodes.

The principle of changing the wavelength of the semiconductor laser is described as follows. That is, in the case where the active region has a quantum well structure, as shown in FIG. 28, the carrier density dependence of the gain is non-linear. The gain required for oscillation can be obtained in a state where the differential gain in the region corresponding to the above is different. Therefore, the carrier density of the entire laser can be changed in the oscillation state, and the oscillation wavelength can be changed because the refractive index changes linearly with the carrier density.

However, in this case, since the selection area of the differential gain is limited to one obtained on one carrier density-gain characteristic curve, there is a disadvantage that the variable width of the oscillation wavelength is as small as about 6 nm. Was.

[0007]

As described above, the conventional wavelength tunable semiconductor laser has a problem that when the oscillation wavelength is changed, the spectral line width is widened or the wavelength variation is small.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has as its object to provide a wavelength tunable semiconductor laser device having a wide wavelength tunable width and a spectral line width which does not widen when the wavelength is tunable. Is to do.

[0009]

The gist of the present invention is to obtain a large differential gain difference by changing the carrier density-gain characteristics of the active region in the cavity direction in a distributed feedback semiconductor laser.

That is, the present invention (Claim 1) relates to a plurality of active regions formed of a distributed feedback semiconductor laser and having different carrier density-gain characteristics formed in the same resonator, and corresponding to the plurality of active regions. a plurality of electrodes provided Te, comprises a means for varying the proportion of current flowing to the electrodes
The active regions having different carrier density-gain characteristics.
The quantum well structure, quantum wire structure or quantity
A child box structure is used .

[0011]

[0012]

[0013]

[0014]

[0015]

According to the present invention (claim 2 ), in a distributed feedback semiconductor laser device having a plurality of active regions and a plurality of electrodes corresponding to the plurality of active regions in the resonator direction, the semiconductor layer in contact with the active region has a plurality of doping concentrations. Are characterized by different active regions.

[0017]

[0018]

According to the present invention (claim 1) , by forming a plurality of active regions having different quantum structures in the resonator direction, as shown in the characteristics (1) and (2) of FIG. Each active region has a unique carrier concentration-gain characteristic, and the difference in differential gain can be made larger than in the conventional case. As a result, the oscillation threshold carrier density can be greatly changed, and as a result, the wavelength variable width can be increased.

[0019]

[0020]

According to the present invention (claim 2 ), the carrier density-gain characteristic of the active region is changed by changing the impurity density doped in the active layer for each of a plurality of active layers formed in the resonator direction. Can be changed for each active region, thereby making it possible to increase the variable width of the oscillation wavelength.

[0022]

[0023]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 2 is a sectional view showing a schematic structure of a tunable DFB laser according to the first embodiment of the present invention.
In this embodiment, the number of wells is changed. In the figure, reference numeral 10 denotes an n-type InP substrate, on which a first-order diffraction grating 11 is formed.
An As optical waveguide layer 12 is formed. This InGaAs
On the optical waveguide layer 12, two, four, or two InGaAs wells 13 and an InGaAsP barrier layer 14 are formed in the resonator direction.
Quantum well structure 20 (21, 22a, 22b) made of
Are formed. The quantum well structure 20 has an InP etching stopper layer 15 formed thereon.

InGaAs is formed on the entire surface of the quantum well structure 20.
The sP optical waveguide layer 12 'is formed, and the optical waveguide layer 12'
A p-type InP cladding layer 16 and an InGaAs contact layer 17 are formed on the entire upper surface. The p-side electrode 18 ₁ on the contact layer 17, 18 _2, 18 _3, the substrate 10
An n-side electrode 19 is provided on the lower surface of the.

Next, a method of manufacturing the semiconductor laser having the above-described configuration will be described with reference to the cross-sectional views in FIGS. First, as shown in FIG. 3A, a primary diffraction grating 11 having a period of 240 nm is formed on an n-type InP substrate 10 by a two-beam interference exposure method.

Next, as shown in FIG. 3B, an InGaAs optical waveguide layer 12 (thickness: 50 nm) having a composition of 1.3 μm is formed so as to bury the upper portion of the diffraction grating 11 flat.
On top of this, an InGaAs well layer 13 ₁ (8 nm thick),
The InGaAsP barrier layer 14 ₁ (thickness 10 nm) and the InGaAs well layer 13 ₂ (thickness 8 n) having a composition of 1.3 μm
m), InGaAsP barrier layer 14 ₂ (thickness 3 nm),
InP etching stopper layer 15 (5 nm thick), I
nGaAsP barrier layer 14 ₃ (thickness 3 nm), InGa
As well layer 13 ₃ (thickness 8 nm), InGaAsP barrier layer 14 ₄ (thickness 10 nm), and InGaAs well layer 13 ₄ (thickness 8 nm), metal organic chemical vapor deposition (MO
CVD) (growth conditions: 610 ° C., 200 Torr, total flow rate 10 l / min).

Thereafter, the quantum well structures 22a and 22b in the regions corresponding to both sides of the resonator are partially removed with an etchant of sulfuric acid: hydrogen peroxide: water = 1: 1: 20 (room temperature). At this time, since the InP layer is not etched by the sulfuric acid-based etchant, the quantum well structure portion 21 in the central portion of the resonator has a quantum well structure of four layers of InGaAs well 13, and the quantum well structure portions 22a and 22a on both sides of the resonator.
Reference numeral 22b denotes a quantum well structure having two InGaAs well layers 13.

Next, as shown in FIG.
By the VD method, the entire active region is buried flat with an InGaAsP optical waveguide layer 12 ′ (thickness: 30 nm) having a composition of 1.3 μm, and a p-type InP cladding layer 16 (thickness: 1 μm) is formed.
m), InGaAs contact layer 17 (0.5 μm thick)
m) is grown sequentially.

Finally, electrodes 17 ₁ , 1 are provided at three locations on the InGaAs contact layer 17 at the central portion and on both sides of the resonator, corresponding to the active regions having different differential gains.
The structure shown in FIG. 2 can be obtained by forming the electrodes 18 under the substrates 7 ₂ and 17 ₃ and the substrate 10.

In this way, a total of 1
Injecting a current of 00MA, at equal current supplied to the electrodes 18 _1, 18 _3, 10~90m the current supplied to the electrode 18 ₂
When the value was changed to A, the threshold carrier density was greatly changed, and excellent characteristics compared with the conventional DFB laser having a wavelength variable amount of 10 nm were obtained. On the other hand, the line width is 750 kHz or less even at the time of wavelength tuning.
It showed better characteristics than B laser. Further, since the oscillation threshold current could be reduced, the output for a current of 100 mA doubled from 10 mW to 20 mW.

Although an InGaAsP / InP system is used here for the barrier layer and the etching stopper layer, InGaAlP / GaAs, InGaAsSb / G
Other systems such as aSb or AlGaAsSb / GaSb can also be applied.

FIG. 4 is a sectional view showing a schematic structure of a tunable DFB laser according to a second embodiment of the present invention.
In this embodiment, the amount of distortion is changed. The same parts as those in FIG. 2 are denoted by the same reference numerals, and detailed description thereof will be omitted.

The difference between this embodiment and the first embodiment described above is that the amount of lattice matching distortion of the crystal is changed in the quantum well structure as a means for obtaining a different differential gain. The manufacturing process is almost the same as the first embodiment,
When the active layer is formed, by changing the value of x in (In _x Ga _1-x ) As, a lattice constant larger than InP is grown on the InP substrate, and a compressive strain of 1% is applied. A strained quantum well structure is formed. Then, a region corresponding to the central portion of the resonator is selectively removed by etching using a normal lithography technique, active regions of the strained quantum well structures 24a and 24b are formed at both ends of the resonator, and a region of 4 nm is further removed by etching. A strained quantum well structure 23 is formed by applying a compressive strain of 3% to a thick InGaAs well.

[0035] Also in this case, each electrode 18 _1, 18 ₃ 1
8 where the ratio of current to be ₂ injected into varying similarly to the first embodiment, excellent characteristics that 12nm is obtained as a wavelength variable amount was below 500kHz not observed even spread of the line width.

The present invention is not limited to the embodiment described above. For example, the wavelength can be continuously changed by adding a phase control region. In the above embodiment, an example in which only one means is used as a method for obtaining different differential gains has been described. However, it is also possible to use a plurality of means at the same time, for example, changing the number of wells and changing the amount of distortion. Further, in the above embodiment, the semiconductor substrate is of the n-type, but may be of the p-type. In that case, the conductivity type of each layer may be reversed. Also, for example, in the embodiment shown in FIG. 2, the interface between the quantum well structure and the optical waveguide layer 12 'on the convex side of the quantum well structure 21 is made to have a step-like tapered structure, and the loss due to reflection in the active layer is reduced. Can also be reduced. Further, it is apparent that the present invention can be applied to a quantum wire structure and a quantum box structure in addition to the quantum well structure.

Next, third and fourth embodiments of the present invention will be described.

In this embodiment, when forming a plurality of active regions having different carrier density-gain characteristics in the same resonator,
This is a method for manufacturing a semiconductor laser in which an active region having a multiple quantum well structure in which any of the thickness, composition, and strain amount of a well layer is different using a selective crystal growth technique is formed by a single crystal growth. Specifically, by utilizing the fact that the composition, thickness, and strain amount of a semiconductor thin film grown between a pair of selective growth masks depend on the width of the selective mask, a plurality of different selective growth masks are formed in the resonator direction. By forming, the active region composed of the semiconductor thin film growing between the masks has a plurality of different gain distributions, and the laser oscillation threshold carrier density can be changed corresponding to each gain distribution. As a result, a wide wavelength variable width can be obtained.

In the present embodiment, in the crystal growth provided with the selective crystal growth mask, the composition, thickness and strain amount of the semiconductor thin film formed between the paired masks depend on the width of the mask. The point is that the method was applied to the formation of a plurality of active regions having different differential gains. As shown in FIG. 5, particularly in the InGaAsP / InP system, the composition, thickness and strain of InGaAs used as a well layer in the quantum well structure are InGaAs used as a barrier layer.
It largely depends on the mask width as compared with sP. By taking advantage of this fact and changing only the well layer without changing the barrier layer, the differential gain is greatly different, but multiple quantum wells with the same light coupling and carrier injection efficiency in each well layer are used. The well structure can be formed by one crystal growth.
Further, a phase shift structure can be equivalently provided by partially changing the width between the pair of selection masks.

According to this manufacturing method, not only can an active region having a plurality of carrier density-gain characteristics be formed by one crystal growth, but also the width of the active layer can be accurately controlled before the crystal growth, and the active region can be formed. The layer stripe can be formed on the (111) B plane without striation. Furthermore, since the boundary of the active region composed of a plurality of carrier density-gain characteristics changes smoothly, loss at those boundaries can be reduced.

FIG. 6 is a sectional view showing a schematic structure of a tunable DFB laser according to a third embodiment of the present invention.
In the figure, reference numeral 30 denotes an n-type InP substrate having a (100) plane. A diffraction grating 31 is formed on the n-type InP substrate 30 in the [011] direction. An active region 40 composed of a structure is formed. A plurality of regions are formed in the active region 40 by one selective crystal growth, and the boundary region is smoothly changed. Further, the p-type InP cladding layer 36, InGaAs
The s-contact layer 37 and the ohmic electrode 38 are separately formed corresponding to respective active regions having different structures. 39 is an n-type ohmic electrode, 43 is a separation groove, and 45 is a dielectric anti-reflection coating film.

Next, a method of manufacturing the semiconductor laser having the above configuration will be described below with reference to FIGS. FIG. 7 is a top plan view for explaining the selective crystal growth mask, and is shown corresponding to FIG. FIG. 8 is a cross-sectional view showing a light emitting surface. FIG. 9 is a diagram for explaining a multiple quantum well structure and a transition region, and is an enlarged view of a dashed circle h in FIG. FIG. 10 shows an outline of a semiconductor laser manufacturing process. 10 (a) to 10 (d) are cross-sectional views perpendicular to the axis of the active region.
-J 'section.

First, as shown in FIG. 6, a first-order diffraction grating 31 is placed on an n-type InP substrate 30 having a (100) plane.
Is formed so as to have a period of 240 nm by a two-beam interference exposure method. The diffraction grating 31 is formed such that the phase is shifted by λ / 4 at the center of the laser resonator. Diffraction grating 3
One groove is selected so as to be orthogonal to the [011] direction which is the axis of the active region.

Next, as shown in FIG.
Selective growth mask 47 with 50 nm thick SiO ₂ film on P surface
After being formed by a chemical vapor deposition (CVD) method, two parallel SiO ₂ stripes 47 are formed in the [011] direction by a usual lithography technique.
At this time, the stripe length is set to a period of 300 μm,
The width of the O ₂ mask 47 is 1 μm in the ij section and 4 μm in the i′-j ′ section from the relationship shown in FIG. 5, and the width between the _two parallel SiO ₂ masks 47 is 1 μm.

Next, an InGaAsP light guide layer 32 is grown by MOCVD (metal organic chemical vapor deposition) to a thickness of 70 nm, which is equal to or greater than the thickness of SiO ₂ , and an InGaAs well layer having a thickness of 4 nm in the ij section. Crystal growth is performed alternately with the InGaAsP barrier layer 34 (thickness: 10 nm) so that the number of layers 33 becomes 10 and the InGaAsP light guide layer 32 ′.
Is grown to a thickness of 30 nm, and an InP cap layer is grown to a thickness of 10 nm. InGaAsP up to the thickness of the SiO ₂ mask 47 or more
The crystal growth of the light guide layer 32 is performed because the lateral surface of the selective growth has a flat (111) B surface regardless of the unevenness of the end surface of the SiO ₂ mask 47 when the thickness is equal to or greater than the mask thickness.
This is the same as the 1991 Spring Applied Physics Proceedings 28p-
It is also described in ZK-3.

[0046] In the multiple quantum well structure grown this way, the i-j cross-section, as shown in Figure 5 an In _x Ga
_{X of the 1-x} As quantum well is 0.6, while i'-j '
Although the _x of the In _x Ga _{1 -x} As quantum well in the cross section is 0.7, the composition of the InGaAsP barrier layer hardly changes. Therefore, when x = 0.6, the strain amount of InGaAs is 0.4
% Is 1.2% at x = 0.7, and as a result, these differential gains are greatly different. However, since the barrier layers are almost the same, light coupling and carrier injection in each active region are performed. Efficiency etc. do not change.

Next, a series of steps up to completion of the laser of the embodiment will be described with reference to FIG. FIG. 10A shows a configuration completed through the above steps. In FIG. 10A, 30 is an n-type InP substrate, 47 is a selective growth mask, and 40 is an active region.

Next, as shown in FIG.
The O ₂ selective growth mask 47 is removed by ordinary etching, and a second crystal growth is performed by MOCVD. In this crystal growth, the p-InP cladding layer 36 is
After growth, the p-InGaAs contact layer 37 is grown to 0.5 μm.

Next, as shown in FIG. 10C, a Au / Zn / Au p electrode 38 is vacuum-deposited on the contact layer 37. Further, the p-InGaAs contact layer 37 and the p-electrode 38 are removed by etching corresponding to each active region grown by the first MOCVD method, and are separated to make each active region electrically high in resistance. Thereafter, the back surface is polished to a thickness of 100 μm, and AuGe is vacuum-deposited on the back surface to form an n-electrode 39. After that, a part of the p-InP cladding layer 36 is mesa-etched to a depth of the active region 40.

Next, as shown in FIG. 10D, both sides 40 'of the active region grown simultaneously with the active region are side-etched. Finally, a non-reflective coating film 45 is formed on the laser end face formed by cleavage by p-CVD (plasma C).
By forming by the VD) method, the structure shown in FIG. 6 is obtained.

When the laser oscillation was performed while changing the ratio to the center electrode while the current values applied to the electrodes at both ends of the semiconductor laser manufactured as described above were equal, a large change in the threshold carrier density caused the laser oscillation. Therefore, the wavelength tunable amount is 8 to 12n, which is 10 times or more that of the conventional DFB laser.
m was obtained. Also, the line width is 0 or 8 MHz or less even at the time of wavelength tuning, and has excellent characteristics as compared with a conventional ordinary DFB laser device.

FIG. 11 is a sectional view showing a schematic structure of a tunable DFB laser according to a fourth embodiment of the present invention. In this embodiment, no diffraction grating is formed before the first crystal growth, and a diffraction grating is formed after the crystal growth. That is, at the time of the selective growth, first, n-InP is grown as a buffer layer to have a thickness equal to or greater than the thickness of the SiO ₂ mask 47 so that the entire active region 40 including the light guide layer 32 in the selective growth is removed by the influence of the unevenness of the mask. I have. The diffraction grating 31 is formed after the selective growth and before the removal of the SiO ₂ mask.

The present invention is not limited to the embodiment described above. For example, as shown in FIG. 12, if the active region is formed by changing the mask interval between the central portions 48 of the _two parallel selective growth SiO ₂ masks 47, an equivalent phase shift laser can be obtained.

In addition to the crystal composition, the carrier concentration in the active region changes depending on the presence or absence of the selective crystal growth mask. For example, when Zn and S are added as impurities, the effect is remarkably exhibited. That is, an element having a high atmospheric pressure has a low probability of adhering to a narrow mask, and the carrier concentration is locally reduced. Conversely, when Si is added, it locally increases. By matching this action with the above-mentioned structure action, composition action, strain action and the like of the active region, an improvement in the characteristics can be recognized.

Next, a fifth embodiment of the present invention will be described.

In this embodiment, similarly to the third and fourth embodiments, when forming a plurality of active regions having different carrier density-gain characteristics in the same resonator, a well is formed using a selective crystal growth technique. Active regions having different crystal compositions are formed simultaneously with multiple quantum well structures having different widths and barrier widths.
Specifically, the composition and layer thickness of the crystal grown between a pair of selective growth masks are made different from the composition and layer thickness of the region without the mask, and the laser is locally adapted to the differential gain distribution. The carrier density of the oscillation threshold can be changed, and as a result, the wavelength tunable width can be widened.

In this embodiment, it has been found that in the crystal growth provided with the selective crystal growth mask, the crystal composition and the degree of crystal distortion in the vicinity of the mask can be slightly different from those in the region away from the mask. The point is that the present invention is applied to the formation of the active region. During the crystal growth process, when the reaction gas transported on the mask diffuses into the windowed crystal substrate in an unreacted state and grows the crystal, an element with a low vapor pressure becomes an excess gas state near the mask. This is because the crystal composition is slightly different. The fact that the crystal composition can be varied also makes it possible to impart crystal strain to the quantum well structure by manipulating the crystal growth conditions, and to make the quantum well in the active region grown near the mask a strained quantum well structure. it can. Even when a strained quantum well is formed in the entire region, the degree of strain can be changed between the vicinity of the mask and the other region.

Changes in the crystal growth rate, crystal composition, crystal distortion, etc., depend on the mask interval and the mask coverage, and the smaller the mask interval and the higher the coverage, the larger the amount of change. For example, when a pair of masks having a width of 10 μm are provided, when the mask interval is 50 μm or more, little effect due to crystal composition change or distortion is obtained, and when the mask interval is 1 μm or less, As or P having a high vapor pressure is interposed between the masks. Since the required amount is not supplied, desired crystal growth is not performed with good controllability. In addition, when the mask coverage increases, the growth rate becomes extremely high, and it becomes difficult to form a multiple quantum well structure. When the coverage was represented by the ratio between the mask area and the area of the interval region, an effective composition change was recognized between 1 and 50.

Note that changes in crystal growth rate, crystal composition, crystal distortion, etc. are not only determined by the form of the mask, but also undergo many or complicated changes depending on crystal growth conditions. For example, as a crystal growth condition having crystal distortion in a region without a mask, it is possible to prevent crystal distortion from being applied to a region sandwiched between masks.

According to this manufacturing method, each divided active region is bounded by a smooth transition region by the selective crystal growth technique, the selection of the crystal plane orientation, and the technique of using the contact layer as a blocking layer for lateral etching. Reflection of the laser beam can be prevented, and the mode of the laser beam is not disturbed. Also,
Divided grooves for dividing into complex active regions are excellent in dimensional control of depth and width, and are formed as inclined surfaces or curved surfaces so that the angle between the cavity direction side wall surface and the groove low surface is obtuse. This makes it possible to prevent the disturbance of the mode of the laser beam.

FIG. 13 is a sectional view showing a schematic structure of a tunable DFB laser according to a fifth embodiment of the present invention.
The same parts as those in FIG. 6 are denoted by the same reference numerals, and detailed description thereof will be omitted. The basic configuration is the same as that of the third embodiment. A diffraction grating 31 is formed on an n-type InP substrate 30, and an active region 40 including a light guide layer and a quantum well structure is formed on the diffraction grating 31. Is formed. The active region 40 is divided into two types and three regions by selective crystal growth, and the boundary region is a transition region whose thickness gradually changes. Further, the p-type InP cladding layer 36, In
A GaAs contact layer 37, a p-type ohmic electrode 38,
The n-type ohmic electrode 39, the active region separation groove 43, and the dielectric non-reflective coating film 45 are also formed in the same manner as in the third embodiment.

Next, a method of manufacturing the semiconductor laser having the above configuration will be described below with reference to FIGS.
FIG. 14 is a plan view for explaining a selective crystal growth mask, and is shown corresponding to FIG. FIG. 15 and FIG. 16 show the outline of the manufacturing process of the semiconductor laser. FIG.
5 (a) to 5 (f) are cross-sectional views perpendicular to the axis of the active region, and are ij cross-sections shown in FIG. FIG.
(F) is a cross-sectional view orthogonal to FIGS. 15 (a) to 4 (f), and corresponds to a wavy circle g region in FIG.

First, as shown in FIG. 13, a primary diffraction grating 31 is formed on an n-type InP substrate 30 so as to have a period of 240 nm by a two-beam interference exposure method. The substrate plane orientation is determined so that the grooves of the diffraction grating are orthogonal to the active region axis.
When the surface on which the diffraction grating 31 is formed is the main surface of the substrate and the main surface of the substrate is a (100) crystal plane, the resonator axis direction is the (0-1-1) direction.

Then, as shown in FIG. 14, an amorphous Si film was formed on the main surface of the substrate to a thickness of 500 nm and an SiO film was formed to a thickness of 200 nm.
Is formed by a chemical vapor deposition (CVD) method and then shaped into a pair of masks by a patterning technique. When the mask has a thickness of 500 nm or more, it is preferable that the mask has a laminated structure so as to withstand a crystal growth temperature of about 620 ° C. which will be described later. The thickness of the mask is determined by the required thickness of the active region and the heat resistance. When the thickness of the active region is 100 nm or less, the light guide effect is reduced, and when the thickness is 1200 nm or more, mask cracks occur during crystal growth. FIG. 14 shows a region 52 affected by a mask when forming an active region described later and a region 53 not affected by the mask. Reference numeral 54 denotes a transition area.

Next, an active region is formed through a mask. As shown in FIG. 9, the active region is a buffer In.
The light guide layer 32 and the multiple quantum well layer are composed of a P layer and an InGaAsP layer lattice-matched to InP, and are formed continuously. The multiple quantum well structure is InGaAs
The well layer 33 (thickness 4 to 12 nm) and the InGaAsP barrier layer 34 (thickness 5 to 12 nm) are alternately formed 3 to 10 times. At this time, abnormal crystal growth must be avoided because it will hinder the patterning process described later. Therefore, forming the thickness of the active region to be smaller than the selective growth mask is effective as a measure for preventing abnormal crystal growth on the mask. It is also effective to make the upper layer larger than the lower layer mask in a multilayer mask configuration.

The crystal growth method is CVD using an organic metal.
Each layer was sequentially formed using the MOCVD method (growth conditions: 620 ° C., 150 Torr, total gas flow: 10 l / mi).
n). The active region has a high crystal growth rate near the mask,
The thickness of the active region changes continuously through the transition region, because the delay becomes slower as the mask is removed. At the same time, the crystal composition is also continuously changing. The crystal composition is In
x and y values of _{x Ga 1-x As y P} 1-y well layer varies locally degree 0.02 by the presence or absence of the mask.

When the InGaAs well layer 33 is formed, the crystal growth conditions such as changing the supply amount of the reaction gas, changing the pressure, changing the growth temperature, and the like are manipulated.
The composition of the InGaAs well layer 33 changes, and In _x Ga
The InAs composition x of _1-x As is about 0.6 and 0.3 to 0.
A strained quantum well structure in which 6% compressive strain is strongly applied only to a region sandwiched between masks can be formed.

Next, a series of steps until the laser of the embodiment is completed will be described with reference to FIGS. FIGS. 15A and 16A show the configuration completed through the steps up to this point. FIG. 15 (a) and FIG. 16 (a)
In the figure, 30 is an n-type InP substrate, 51 is a selective growth mask, and 40 is an active region.

Next, as shown in FIGS. 15B and 16B, the active region 40 is processed into a stripe shape by a patterning technique and etching with a mixed acid of hydrogen bromide. The stripe is formed to have a width of 0.8 to 1.2 μm so as to pass between a pair of masks. At this time, by utilizing the fact that the thicknesses of the active layers are partially different, stripes having different active layer widths can be easily formed in a self-aligned manner.
Here, since the crystal part grown in contact with the mask has extremely poor thickness controllability and composition controllability as compared with the crystal part relatively far from the mask, it is necessary to remove the crystal part and prevent the characteristic deterioration. It is.

Next, as shown in FIGS. 15C and 16C, a p-type InP cladding layer 36 (1.5 μm thick) is formed on the main surface of the substrate by a second crystal growth so as to surround the active layer 40. ), An InGaAs contact layer 37 (thickness 0.5 μm) is sequentially formed by the MOCVD method, and then the contact layer 37 is formed by a patterning technique.
Process to a width of 0 μm. At this time, if dummy contact layer stripes (not shown) are formed on both sides of the contact layer 37, lateral etching described later can be completely prevented.

Next, as shown in FIG. 15D and FIG.
The O insulating film 55 is selectively formed, and the AuZn ohmic electrode 56 is formed by lift-off processing. Si
The O insulating film 55 has a thickness of 700 nm by a normal pressure CVD method.
And then patterned. The AuZn ohmic electrode 56 was formed by a vacuum evaporation method.

Next, as shown in FIGS. 15E and 16E, a TiPtAu electrode 57 is selectively formed on the substrate, and the SiOP film 55 is patterned using the TiPtAu electrode 57 as a mask. The width of the TiPtAu electrode 57 is the same as that of the contact layer 37.
Is formed to be about 6 μm wider than the width. By this process,
The p-type InP cladding layer 36 and the InGaAs contact layer 37 are partially exposed.

Next, as shown in FIG.
Using the O film 55 as an etching mask, p-type In
The P-cladding layer 36 is selectively removed by etching to perform so-called element isolation along the resonator axis. At this time, the p-type I
Lateral erosion of the nP cladding layer 36 must not reach the InGaAs contact layer 37. This is because the InGaAs contact layer 37 is formed in a step of forming a dividing groove described later.
This is for preventing the etching of. In addition, as shown in FIG. 16E, the exposed InGaAs contact layer 37 is not etched, and functions as a mask of the InP clad layer 36.

Next, as shown in FIG.
Using the O film 55 as an etching mask, the InGaAs contact layer 37 is removed by etching with a mixed solution of sulfuric acid.
Using the aAs contact layer 37 as a mask, the p-type InP cladding layer 3 is controlled while controlling the division groove to an arbitrary depth.
6 is etched to form the active region dividing groove 43.
To form At this time, the groove 4 formed by etching
The angle formed between the cavity direction side wall surface 43a and the groove bottom surface is an obtuse angle. This is achieved by the above-described designation of the crystal plane orientation. Further, the InGaAs contact layer 3
7, the lateral erosion of the p-type InP cladding layer 36 is prevented, so that the dimensional control of the width of the dividing groove is good.

While the p-type InP cladding layer 36 is etched and the mesas formed by element isolation are further etched in the lateral direction, the etching is stopped by the InGaAs contact layer 37 and the width of the InGaAs contact layer 37 is controlled. A striped mesa can be formed (see FIG. 16F).

An AuGe ohmic electrode 39 provided at the time of forming the TiPtAu electrode 57 is formed on the back surface of the laser wafer substrate thus formed by a vacuum deposition method. 45 is formed by a CVD method to complete a semiconductor laser.

When the laser oscillation was performed while changing the ratio to the center electrode with the current values applied to the electrodes at both ends of the semiconductor laser manufactured as described above being large, the threshold carrier density was largely changed. Therefore, the wavelength tunable amount is 8 to 12n, which is 10 times or more that of the conventional DFB laser.
m was obtained. In addition, the spectral line width was 0,8 MHz or less even at the time of wavelength tuning, and had excellent characteristics as compared with a conventional ordinary DFB laser device.

The present invention is not limited to the embodiment described above. For example, as shown in FIG. 17, after two pairs of selective crystal growth masks 56 and 57 are provided to grow and form an active region, the active region is patterned by using an etching mask to obtain the active layer thickness (quantum well width and barrier). width)
At the same time, the width 58 of the active layer can be changed. In this case, in addition to the above-mentioned effects of the active layer thickness, crystal composition, crystal distortion, etc., the effect of the active layer width can be given simultaneously, and a semiconductor laser device having further excellent characteristics is constituted by a synergistic effect of these. Can be.

Further, the carrier concentration in the active region together with the crystal composition changes depending on the presence or absence of the selective crystal growth mask.
For example, when Zn and S are added as impurities, the effect is remarkably exhibited. That is, an element having a high atmospheric pressure has a low probability of adhering to a space between narrow masks, and the carrier concentration is locally reduced. Conversely, when Si is added, it locally increases. By matching this action with the above-mentioned structure action, composition action, strain action and the like of the active region, an improvement in the characteristics can be recognized.

Next, sixth to eighth embodiments of the present invention will be described.

In this embodiment, for example, a plurality of regions in which the quantum well is a quantum well box and the number of the quantum well boxes per unit volume is different are formed in the same resonator, and a plurality of regions are formed corresponding to these regions. Is a semiconductor laser in which a current can be independently injected into a plurality of regions having different quantum well densities by providing the electrodes.

According to the present embodiment, carrier density-gain characteristics are different in a plurality of regions having different densities of quantum wells formed in the resonator direction. Therefore, the threshold value of the carrier density required for the oscillation can be changed, and the magnitude of the change is determined by the difference in the density of the quantum well, so that it can be changed more than in the conventional case. As a result, the variable width of the oscillation wavelength can be increased.

FIG. 18 is a sectional view showing a schematic structure of a tunable DFB laser according to a sixth embodiment of the present invention. In this embodiment, the number of quantum well thin films per unit thickness is changed. . In the figure, reference numeral 60 denotes an n-type InP substrate, on which a first-order diffraction grating 61 having a period of 240 nm is formed. A 1.3 μm composition of I
An nGaAsP optical waveguide layer (thickness: 50 nm) 62, seven, fifteen, and seven InGaAs quantum well thin films (thickness: 8 nm) in the resonator direction and an InGaAsP barrier layer (thickness: 150 nm, 70 nm, thickness: 1.3 μm) 150 nm) and 1 μm-wide stripe-shaped quantum well active layers 63 a and 63.
3b and 63c are formed.

A p-type InP cladding layer (thickness: 2 μm) 66 is formed so as to bury the active layers 63a, 63b, 63c, and a p-type InGaAsP contact layer (corresponding to the active layers 63a, 63b, 63c) is formed thereon. 0.5μ thickness
m) 67a, 67b, 67c are formed. The p-side electrodes 68a, 68b, 68c are provided on the contact layers 67a, 67b, 67c, and the n-side electrode 69 is provided on the lower surface of the substrate 60.

[0086] Such a semiconductor laser is provided with electrodes 68a,
Currents were injected from 68b and 68c so that the total would be 100 mA, the currents injected from electrodes 68a and 68c were made equal, and the current injected from electrodes 68b was changed from 10 mA to 90 mA. Then, the threshold value of the carrier density required for the oscillation greatly changed, and as a result, the oscillation wavelength changed by 18 nm. This tunable amount is equal to the conventional D
This is an excellent characteristic that the wavelength variable width is three times as large as the wavelength variable amount of the FB laser of 6 nm. In addition, the spectral line width at the time of variable wavelength was 700 kHz or less, which was superior to that of the conventional DFB laser. Furthermore, since the threshold value of the current required for oscillation could be reduced, the optical output for a current injection of 100 mA was 20 mW, which was doubled from 10 mW of the conventional tunable semiconductor laser.

FIG. 19 is a perspective view of a main part of a tunable DFB laser according to a seventh embodiment of the present invention, in which the number of quantum well thin wires per unit cross section is changed. It is. The difference between this embodiment and the previously described sixth embodiment is that the quantum well thin film active layers 63a, 63b, 6
Instead of 3c, as shown in FIG. 19 (a), quantum well thin wire active layers 71a, 71b, 71c are used, and the density of the quantum well thin wire is changed from coarse to fine in the resonator direction. Further, by changing the direction of the quantum well thin wire, FIG.
As shown in the figure, the quantum well fine wire active layers 72a, 72b, 72
c, or as shown in FIG. 19 (c), quantum well fine wire active layers 73a, 73b, 73c.

When the injection current was changed in the semiconductor laser of the seventh embodiment in the same manner as in the sixth embodiment, an excellent characteristic of a wavelength variable amount of 20 nm was obtained, and the spectral line width at this time was also 500 kHz. It was below.

FIG. 20 is a perspective view of a quantum well active layer in which the number of quantum well boxes per unit volume is changed, showing a main part of a tunable DFB laser according to an eighth embodiment of the present invention. It is. The difference between this embodiment and the previously described sixth embodiment is that the quantum well thin film active layers 63a, 63b, 63
Instead of c, the quantum well box active layers 74a, 74b, 74
Using c, the density of the quantum well box was changed from coarse to fine in the resonator direction.

When the injection current was changed in the semiconductor laser according to the eighth embodiment in the same manner as in the sixth embodiment, an excellent characteristic of a wavelength variable amount of 30 nm was obtained, and the spectral line width at this time was also 300 kHz. It was below.

Next, ninth and tenth embodiments of the present invention will be described.

In this embodiment, the doping state of the semiconductor layer in contact with the active layer is changed for each active region, and as a result of impurity diffusion from the semiconductor layer in contact with the active layer to the active layer, the doping state of the active layer is changed. Are different for each active region.

As a specific manufacturing method, a step of forming an active layer on a semiconductor substrate and a step of forming another part in a predetermined region of the semiconductor layer having a wider forbidden band than the active layer provided in contact with the active layer. Forming a different doping region, forming a diffraction grating, diffusing impurities into the active layer, and forming a plurality of electrodes corresponding to the predetermined region and other regions.

Also, a step of forming an active layer on a semiconductor substrate, a step of forming a semiconductor layer having a wider band gap than the active layer, and a step of forming a semiconductor layer having a wider band gap than the active layer The method includes a step of etching a predetermined area, a step of forming a diffraction grating, a step of diffusing impurities into an active layer, and a step of forming a plurality of electrodes corresponding to the predetermined area and other areas.

Here, the order of each step is not limited to the above order. Further, the step of diffusing impurities into the active layer does not need to be an independent step, and includes, for example, a case where diffusion occurs simultaneously in an overgrowth step.

According to the present embodiment, the density of impurities doped into the active layer can be changed for each of a plurality of active regions formed in the resonator direction. In general, the differential gain of the active region increases due to acceptor doping and decreases due to donor doping due to asymmetry between the conduction band and the valence band. That is, according to the present embodiment, the injected carrier-gain characteristics of the active region can be easily changed for each active region. The current injected into each active region, that is, the gain, can be independently controlled by the plurality of electrodes.

As shown by (1) and (2) in FIG. 21, since each active region has a unique injection carrier-gain characteristic, each active region has a characteristic required for obtaining a gain necessary for oscillation. By changing the ratio of the flowing current, the average oscillation threshold carrier density can be largely changed. The oscillation wavelength is proportional to the product of the period of the diffraction grating and the equivalent refractive index. The higher the carrier density, the smaller the equivalent refractive index. Therefore, the wavelength can be largely changed by changing the average carrier density.

FIG. 22 is a sectional view showing a schematic structure of a tunable DFB laser according to a ninth embodiment of the present invention. In the figure, reference numeral 80 denotes an n-type InP substrate. A primary λ / 4 phase shift diffraction grating 81 is formed on the InP substrate 80. A 70 nm-thick InGaAsP optical waveguide layer 82 is lattice-matched to InP and has a composition of 1.55 μm wavelength.
An InGaAsP active layer 83 having a thickness of 100 nm is formed.

[0098] At the center of the predetermined region 91 ₂ on the active layer 83, lattice-matched wavelength 1.3μm composition and the InP etching stop layer 84 having a thickness of 10 nm InP, thickness 50nm
InGaAsP diffusion stop layer 85 is formed. And, in the area 91 _1, 91 ₃ at both ends of the upper and the active layer 83 of the diffusion stop layer 85, p-type InP cladding layer 86, p-type InGaAs contact layer 87 are formed. The p-type dopant is Zn, the Zn concentration of the etching stop layer 84 is 10 ¹⁷ cm ⁻³ or less, and the Zn concentration of the p-type cladding layer 86 is 1 × 10 ¹⁸ cm ⁻³ . On the contact layer 87, p-side electrodes 88 ₁ , 8 corresponding to the respective regions are formed.
8 _2, 88 _3, n-side electrode 89 on the entire lower surface of the substrate 80
Is provided.

Further, a semi-insulating InP layer 92 is formed between the respective regions for insulation separation. The length of each region 91 is about 400 μm, and the laser cavity length is 1.
At 2 mm, the length of the separation region 22 is 5 μm. The phase shift 93 of the diffraction grating 81 is provided at the center. An antireflection coating film 95 having a reflectance of 1% or less is formed on both end surfaces. This semiconductor laser is arranged to supply a current independently to the central electrode 88 ₂ and the electrode 88 ₁ of the two ends, 88 ₃ are mounted on a package (not shown in the Figure).

Next, a method of manufacturing the semiconductor laser having the above configuration will be described with reference to FIG. First, FIG.
As shown in (a), a diffraction grating 81 having a primary period of about 240 nm is formed on an n-type InP substrate 80. At this time, the λ / 4 phase shift region 9 is located
Form 3 Subsequently, the diffraction grating 81 is lattice-matched to InP so as to be buried flat, and has a composition of wavelength 1.15 μm.
70 nm thick undoped InGaAsP optical waveguide layer 82
Is grown and then lattice-matched to InP and has an undoped InGaAsP active layer 83 with a wavelength of 1.55 μm and a thickness of 100 nm, an undoped InP etching stop layer 84 with a thickness of 10 nm, and a lattice-matched with InP and a wavelength and a composition of 1.3 μm A 50 nm undoped InGaAsP diffusion stop layer 85 is grown. In this example, the growth was performed by the reduced pressure metal organic chemical vapor deposition (LP-MOCVD) method, but another growth method may be used.

[0102] Then, as shown in FIG. 23 (b), the SiO ₂ film 96 is formed by CVD on the entire surface, the normal on the region 21 ₂ of length 400μm at the center portion of the laser PEP
To form a resist mask 97. Then, the SiO ₂ film 96 in the portion of the region 91 _1, 91 ₃ formed at both ends was removed with ammonium fluoride solution, followed by diffusion stop layer 8
5 is removed with an etchant of sulfuric acid: hydrogen peroxide: water = 1: 1: 20 (room temperature). Since this etchant has a low etch rate with respect to InP, the etching can be stopped on the etching stop layer 84. Thereafter, the etching stop layer 84 is removed as needed.

[0102] Next, after removing the SiO ₂ film remaining in the central portion 91 _2, as shown in FIG. 23 (c), Z thereon
p-type InP with n concentration of 1 × 10 ¹⁸ cm ⁻³ and thickness of 1.5 μm
Cladding layer 86, Zn concentration 5 × 10 ¹⁸ cm ⁻³ , thickness 0.
The 7 μm p-type InGaAs contact layer 87 is formed by LP-M
It is formed by OCVD. The growth temperature is 620 ° C. At this time, diffusion of Zn occurs from the cladding layer 86 to a layer below. However, since the rate at which Zn is diffused an InGaAsP layer is slower than the rate of diffusion of the InP layer, the diffusion region 91 _2, Zn laser central diffusion stop layer 8
5 and the acceptor concentration of the undoped InP etching stop layer 84 and the active layer 83 remains at 10 ¹⁷ cm ⁻³ or less. On the other hand, at both ends where there is no diffusion stop layer 85, 5 × 10 ¹⁷ c
About m ⁻³ Zn acceptors are diffused and doped.

Thereafter, although not shown, a lateral confinement structure is formed so that the active region has a stripe shape.
Here, a structure in which trenches deeper than the active layer are formed on both sides of the active region and the inside thereof is buried with Fe-doped semi-insulating InP by MOCVD is used. Applicable regardless.

Next, as shown in FIG.
Using an O ₂ film (not shown) as a mask, an isolation groove is formed in a region having a boundary length of 5 μm between each region, and MOCVD is formed therein.
An Fe-doped semi-insulating InP isolation region 92 is selectively grown by the method. Thereafter, p is formed on the contact layer 87 in each region.
After forming the n-side electrode 89 on the entire back surface polished to a thickness of 80 μm, the side electrode 88 is cleaved into a chip having a length of 1.2 mm, and a SiN _x anti-reflection coating film 95 is formed on both end surfaces to form a module. By mounting, bonding, and mounting, a structure as shown in FIG. 22 is obtained.

[0105] Semiconductor lasers fabricated in this way, both end portions 91 ₁ Zn concentration in the active layer of the central portion 91 _2,
It is lower than the Zn concentration of 91 ₃ of the active layer. As a result, both end portions 91 _1, 91 ₃ and the injection carrier vs. gain characteristic of the central portion 91 _2, respectively in FIG. 21 (1), (2)
It becomes as shown in. Now, in the first oscillation state, the operating point is
It is assumed that 1 is at points A1 and A2. Here, while adjusting the total current so that the output power is constant, increasing the current in the both end portions 91 _1, 91 _3, it is assumed to reduce the current in the central portion 91 _2. Both ends 91 ₁ and 91 ₃ are in the center 9
Since the differential gain than 1 ₂ is large, the larger injected carriers reduced in the central portion as compared to the injection carriers increased at both ends. As a result, the operating point moves to B1 and B2, and the overall effective carrier density can be made smaller than the initial state while maintaining the same optical output as the initial state.

Therefore, the overall equivalent refractive index increases,
The oscillation wavelength is longer than in the initial state. At this time, since the black wavelength itself changes together with the mode arrangement, a wide wavelength range of about 100 angstroms can be varied without mode jump or broadening of the oscillation spectrum line width.

[0107] In this example, InP etching stop layer 84 and the InGaAsP diffusion stop layer 85 was undoped, at least one of them to n-doped, n-type active layer of the central portion 91 ₂ in diffusion during overgrowth It is also possible to Since the n-type active layer has a low differential gain, the difference between the differential gains at both ends 91 ₁ and 91 ₃ can be further increased. In this case, the InGaAsP layer 85
Works not as a diffusion stop layer but as a solid phase diffusion layer. Further, the InP etching stop layer 84 and the InGaAs
Perform at least one of the n-type doping of the P layer 85, by not performing the doping during the growth initiation of the cladding layer 86, and only the central portion 91 ₂ in the highly doped n-type Conversely, both end portions 91 _1, 91 ₃ active It is also conceivable to keep the layer undoped or lightly doped. The central portion 91 ₂ and the end portions 9
1 _1, 91 to reverse the ₃ relationship, only the both end portions 91 _1, 9
1 ₃ It is also possible to leave the InGaAsP layer 85.

The InP etching stop layer 84 is not essential, and the diffusion stop layer 85 and the active layer 83 may be in direct contact. InGaAsP diffusion stop layer 85
It is not necessary to completely remove both end portions 91 ₁ and 91 _3. For example, even if the thickness is reduced by etching, the function as a diffusion stop layer can be suppressed. In any case, the doping state of the adjacent region 91 changes,
The differential gain of the active layer 83 changes.

FIG. 24 is a view showing a process of manufacturing a tunable DFB laser according to the tenth embodiment of the present invention.
The same parts as those in FIG. 22 are denoted by the same reference numerals, and detailed description thereof will be omitted. First, as shown in FIG. 24 (a), and growing a thin undoped InP layer 101 on an n-type InP substrate 80, after forming the SiO ₂ film 102, to remove the SiO ₂ film 102 of the laser central portion 91 ₂ , This SiO ₂
Central portion 91 ₂ of the InP layer to the film 102 as a mask 101
Region 103 is doped with n-type dopant Si by vapor phase diffusion in a PH ₃ atmosphere. Then, S
After peeling off the iO ₂ film 102 and slightly etching the surface to expose a flat and less damaged surface, the diffraction grating 8
Form one.

Then, as shown in FIG. 24B, undoped InGa having a wavelength of 1.15 μm and a thickness of 70 nm was lattice-matched to InP so that the diffraction grating 81 was buried flat.
An AsP optical waveguide layer 82 is grown, on which an undoped InGaAsP active layer 83 lattice-matched to InP and having a wavelength of 1.55 μm and a thickness of 100 nm, and undoped InGaAs having a wavelength of 1.3 μm and lattice-matched to InP and having a thickness of 30 nm.
sP optical waveguide layer 82 ′, p-type InP cladding layer 86 having a Zn concentration of 1 × 10 ¹⁸ cm ⁻³ and a thickness of 1.5 μm, and a Zn concentration of 5 ×
A p-type InGaAs contact layer 87 having a thickness of 10 ¹⁸ cm ⁻³ and a thickness of 0.7 μm is formed by an LP-MOCVD method. The growth temperature is 620 ° C.

At this time, Zn is removed from the p-type cladding layer 86.
However, Si diffuses from the n-type InP region 103 into the active layer 83. Whereas only the diffusion of the laser opposite ends 91 _1, 91 _3, Zn is p-type occurs, mutually compensated both central portion 91 _2, Zn and Si, the undoped or low doped region. As a result, both ends 91 ₁ , 9 are relatively formed.
1 ₃ differential gain is increased, the same effect as the ninth embodiment is produced. The electrodes and other manufacturing methods are the same as in the ninth embodiment. Also in this example, as in the ninth embodiment,
Various modified applications are possible.

In the above embodiment, the material is InGaAs.
The P / InP system has been described as an example, but InAs, Ga
As, AlAs, InP, GaP, AlP, InSb,
GaSb, AlSb, GaN, CdTe, HgTe, Z
It can be applied to various materials such as compound semiconductors such as nS, ZnSe, and ZnTe, and ternary mixed crystals and quaternary mixed crystals thereof. Various dopants are also conceivable. The active layer may have a quantum well structure, a strained quantum well structure, a quantum wire structure, or the like, and the optical waveguide layer may have a GRIN structure.

The above-mentioned diffusion control layer may function as a part of the optical waveguide layer or a part of the cladding layer. The diffraction grating may be at the top of the active layer instead of at the bottom. It is clear that the relationship between n-type and p-type may be interchanged for all these examples. In addition, various modifications and applications such as the number and position of regions, the length of each region, end face reflectance, presence / absence of phase shift, method and position, etc. can be made.

Next, eleventh to thirteenth embodiments of the present invention will be described.

In this embodiment, for example, a plurality of regions having different light confinement coefficients are formed in the same resonator by changing the refractive index of the cladding layer, and a plurality of electrodes are provided corresponding to these regions. A semiconductor laser capable of independently injecting current into a plurality of regions having different optical confinement coefficients.

According to this embodiment, the contribution of the gain to the light guided in the resonator differs in a plurality of regions having different light confinement coefficients formed in the resonator direction.
Therefore, the threshold value of the carrier density required for the oscillation can be changed, and the magnitude of the change is determined by the difference in the optical confinement coefficient. As a result, the variable width of the oscillation wavelength can be increased.

FIG. 25 is a sectional view showing a schematic structure of a tunable DFB laser according to an eleventh embodiment of the present invention. In this embodiment, the composition of the cladding layer is changed. The same parts as those in FIG. 2 are denoted by the same reference numerals, and detailed description thereof will be omitted.

On the n-type InP substrate 10, a primary diffraction grating 11 having a period of 240 nm is formed. On the diffraction grating 11, an InGaAsP optical waveguide layer (thickness: 50 μm) having a composition of 1.3 μm is formed.
nm) 12 on which 10 layers of InGaAs are formed.
Quantum well layer (8 nm thick) 13 and 1.3 μm composition In
Width 1 consisting of GaAsP barrier layer (thickness 10 nm) 14
A quantum well active layer 20 having a stripe shape of μm is formed.

In order to embed the active layer 20, InGaAsP, InP,.
P-type cladding layers (2 μm thick) 16a, 16b, 16c made of InGaAsP having a composition of 3 μm, and p-type In
GaAsP contact layer (0.5 μm thickness) 17a, 1
7b and 17c are formed. And contact layer 1
7a, 17b, 17c, p-side electrodes 18a, 18b, 1
8c is provided, and an n-side electrode 19 is provided on the lower surface of the substrate 10.

In this embodiment, in order to obtain a plurality of regions having different light confinement coefficients, the cladding layers 16a, 16b,
16c are semiconductor layers having different compositions, so that the refractive index difference from the active layer is different.

The electrodes 18a, 18a,
Currents were injected from the electrodes 18b and 18c so that the total was 100 mA, the currents injected from the electrodes 18a and 18c were made equal, and the current injected from the electrodes 18b was changed from 10 mA to 90 mA. Then, the threshold value of the carrier density required for the oscillation greatly changed, and as a result, the oscillation wavelength changed by 12 nm. This tunable amount is equal to the conventional D
This is an excellent characteristic that the wavelength variable width is twice as large as the wavelength variable amount of the FB laser of 6 nm. In addition, the spectral line width at the time of variable wavelength was 700 kHz or less, which was superior to that of the conventional DFB laser. Furthermore, since the threshold value of the current required for oscillation could be reduced, the optical output for a current injection of 100 mA was 20 mW, which was doubled from 10 mW of the conventional tunable semiconductor laser.

FIG. 26 shows a schematic structure of a tunable DFB laser according to a twelfth embodiment of the present invention.
(A) is a cross-sectional view in the direction of the resonator, (b) is an arrow A in (a).
It is -A 'sectional drawing. The same parts as those in FIG. 25 are denoted by the same reference numerals, and detailed description thereof will be omitted. This embodiment differs from the eleventh embodiment described above in that the width of the active layer is changed as a means for obtaining a plurality of regions having different light confinement coefficients.

On the optical waveguide layer 12 formed on the substrate 10 on which the diffraction grating 11 is formed, ten InGaAs quantum well layers (8 nm thick) and InGaAs having a composition of 1.3 μm are formed.
Stripe-shaped quantum well active layers 24a, 24b, and 24c each having a width of 0.8 μm, 1.2 μm, and 0.8 μm, each of which is made of an sP barrier layer (thickness: 10 nm), are formed in the resonator direction. A p-type InP cladding layer (2 μm thick) 16 is formed so as to bury the active layers 24a, 24b, and 24c,
Thereon, contact layers 17a, 17b, 17c are formed corresponding to the active layers 24a, 24b, 24c.
The electrodes 18a, 18b, 18c and the electrode 19 are provided in the same manner as above.

The semiconductor laser according to this embodiment has a first
When the injection current was changed in the same manner as in Example 1, excellent characteristics of a wavelength variable amount of 12 nm were obtained, and the spectral line width at this time was 800 kHz or less.

FIG. 27 is a sectional view showing a schematic structure of a tunable DFB laser according to a thirteenth embodiment of the present invention. The same parts as those in FIG. 25 are denoted by the same reference numerals, and detailed description thereof will be omitted. The difference between this embodiment and the eleventh and twelfth embodiments described above is that the composition of the optical waveguide layer is changed as means for obtaining a plurality of regions having different light confinement coefficients.

On the diffraction grating 11 formed on the substrate 10,
In the resonator direction, an InGaAsP optical waveguide layer having a composition of 1.1 μm, a composition of 1.3 μm, and a composition of 1.1 μm (thickness 5
0 nm) 12a, 12b and 12c are formed. Then, an active layer 20 and a cladding layer 16 are formed thereon, and further, contact layers 17a, 17b, and 17c are formed corresponding to the optical waveguide layers 12a, 12b, and 12c. Then, similarly to the previous embodiment, the electrodes 18a, 18
b, 18c and an electrode 19 are provided.

The semiconductor laser according to this embodiment has a first
When the injection current was changed in the same manner as in Example 1, excellent characteristics of a wavelength variable amount of 10 nm were obtained, and the spectral line width at this time was 700 kHz or less.

[0128]

As described above in detail, according to the present invention, a plurality of active regions having different carrier density-gain characteristics are provided in the resonator direction, and a current is independently injected into each active region. Therefore, it is possible to realize a semiconductor laser device in which the wavelength variable amount is large and the line width at the time of wavelength variable is very narrow only by changing the ratio of the injected current.

[Brief description of the drawings]

FIG. 1 is a characteristic diagram illustrating a relationship between a carrier density and a gain for explaining the operation of the present invention;

FIG. 2 is a sectional view showing a schematic structure of a tunable DFB laser according to a first embodiment;

FIG. 3 is a sectional view showing a manufacturing process of the laser according to the first embodiment;

FIG. 4 is a sectional view showing a schematic structure of a tunable DFB laser according to a second embodiment;

FIG. 5 is a characteristic diagram showing a relationship between a mask width and a composition in the third and fourth embodiments;

FIG. 6 is a sectional view showing a schematic structure of a tunable DFB laser according to a third embodiment;

FIG. 7 is a plan view illustrating a selective crystal growth mask according to a third embodiment;

FIG. 8 is a cross-sectional view illustrating a light emitting surface structure according to a third embodiment;

FIG. 9 is an enlarged sectional view showing a multiple quantum well structure in the third embodiment;

FIG. 10 is a sectional view showing a manufacturing process of the laser according to the third embodiment;

FIG. 11 is a sectional view showing a schematic structure of a tunable DFB laser according to a fourth embodiment;

FIG. 12 is a plan view showing an example of a selective crystal growth mask according to a fourth embodiment;

FIG. 13 is a sectional view showing a schematic structure of a tunable DFB laser according to a fifth embodiment;

FIG. 14 is a plan view illustrating a selective crystal growth mask according to a fifth embodiment;

FIG. 15 is a sectional view showing a manufacturing process of the laser according to the fifth embodiment;

FIG. 16 is a sectional view showing a manufacturing process of the fifth embodiment laser.

FIG. 17 is a plan view showing another example of the selective growth mask in the fifth embodiment.

FIG. 18 is a sectional view showing a schematic structure of a tunable DFB laser according to a sixth embodiment;

FIG. 19 is a perspective view showing the main configuration of a wavelength tunable DFB laser according to a seventh embodiment;

FIG. 20 is a perspective view showing a main configuration of a wavelength tunable DFB laser according to an eighth embodiment;

FIG. 21 is a characteristic diagram showing the relationship between carrier density and gain in the ninth and tenth embodiments;

FIG. 22 is a sectional view showing a schematic structure of a tunable DFB laser according to a ninth embodiment;

FIG. 23 is a sectional view showing a manufacturing process of the ninth embodiment laser;

FIG. 24 is a cross-sectional view showing a process of manufacturing the wavelength tunable DFB laser according to the tenth embodiment.

FIG. 25 is a sectional view showing a schematic structure of a tunable DFB laser according to an eleventh embodiment;

FIG. 26 is a sectional view showing a schematic structure of a tunable DFB laser according to a twelfth embodiment;

FIG. 27 is a sectional view showing a schematic structure of a tunable DFB laser according to a thirteenth embodiment;

FIG. 28 is a characteristic diagram showing a relationship between carrier density and gain in a conventional DFB laser.

[Explanation of symbols]

Reference Signs List 10: n-type InP substrate, 11: diffraction grating, 12, 12 ': InGaAsP optical waveguide layer, 13: InGaAs well layer, 14: InGaAsP barrier layer, 15: InP etching stopper layer, 16: p-type InP cladding layer, 17 ... p-type InGaAs contact _{layer, 18 (18 1, 18 2} , 18 3) ... p -side electrode, 19 ... n-side electrode, 20 ... quantum well structure section, 21 ... quantum well structure of well number 2, 22a, 22b ... Quantum well structure with 4 wells, 23% 3% compressive strain quantum well structure, 24a, 24b 1% compressive strain quantum well structure.

──────────────────────────────────────────────────続 き Continuing on the front page (72) Nobuo Suzuki, Inventor No. 1, Komukai Toshiba-cho, Saisaki-ku, Kawasaki-shi, Kanagawa Prefecture Inside Toshiba Research Institute, Inc. No. 1 Toshiba Research Institute, Inc. (72) Inventor Mitsuhiro Kushibe No. 1, Komukai Toshiba-cho, Saiwai-ku, Kawasaki-shi, Kanagawa Prefecture Toshiba Research Institute, Inc. (56) References JP-A-4-76979 (JP, A) JP-A-4-783 (JP, A) JP-A-4-61183 (JP, A) JP-A-64-5088 (JP, A) Appl. Phys. Lett. , 57 [20], p. 2068-2070 Appl. Phys. Lett. , 55 [20], p. 2057-2059 IEICE Technical Report, 91 [75], p. 61-66 J.C. Lightwave. Technology. , LT-5 [4], p. 516-522 (58) Field surveyed (Int. Cl. ⁷ , DB name) H01S 5/00-5/50 G02F 1/35 JICST file (JOIS)

Claims (2)

(57) [Claims] A plurality of active regions formed of a distributed feedback semiconductor laser and having different carrier density-gain characteristics formed in the same resonator; and a plurality of electrodes provided corresponding to the plurality of active regions. , it comprises a means for varying the proportion of current flowing in the electrodes, the carrier density - as different active regions of the gain characteristics,
Quantum well structure, quantum wire structure or quantum box with different strain
A semiconductor laser device having a structure . 2. A distributed feedback semiconductor laser device having a plurality of active regions and a plurality of electrodes corresponding to the active regions in a resonator direction, wherein a doping concentration of a semiconductor layer in contact with the active region is different in the plurality of active regions. A semiconductor laser device characterized by the above-mentioned.