JP3239809B2 - Optical semiconductor device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an optical semiconductor device having a diffraction grating structure, and more particularly to an optical semiconductor device such as an integrated optical modulator.

[0002]

2. Description of the Related Art In long-distance optical fiber communication over a distance of 40 km to 600 km, low-loss 1.55 .mu.m band light is used. However, commonly laid fibers are:
Since it is a normal fiber having a dispersion of 0 at 3 μm,
The 1.55 μm band is affected by dispersion. In order to eliminate the influence of dispersion, it is necessary to suppress the spread of the wavelength spectrum during modulation, that is, the chirping. Therefore, instead of using a directly modulated semiconductor laser, an external modulation method in which a fixed semiconductor laser beam is externally quenched is used.

The semiconductor laser here is a so-called distributed feedback laser having a built-in diffraction grating. Normal DF
It is called a B (Distributed Feedback) laser, and the oscillation wavelength is determined by the diffraction grating.
The oscillation wavelength spectrum becomes a single mode. Since this oscillation wavelength depends on the refractive index of the active layer of the semiconductor laser, it varies as the carrier density of the active layer varies. Also in the external modulation method, when light returns from the modulator to the semiconductor laser, chirping occurs because the carrier density fluctuates due to the fluctuation of the light density in the semiconductor laser.
In order to prevent this, between the semiconductor laser and the modulator,
It is necessary to insert an isolator. Or JP-A-6-
Means such as “light intensity modulator and light intensity modulator” described in Japanese Patent No. 11670 have been proposed.

In recent years, an integrated optical modulator in which a semiconductor laser and an optical modulator are monolithically integrated with respect to such a single modulator has been developed. The integration not only reduces the size of the device, but also reduces light loss from the semiconductor laser to the modulator. However, since it is monolithic, it is more important to take measures against the return light from the modulator to the semiconductor laser unit. To avoid this, usually, the output end face of the modulator has a window structure and an AR coating film is applied. Take measures. This is described, for example, in Article No. ThA2-1, Y.
Sakata, et al. , "Strained M
QW-BH-LDs and integrated
devices fabricated by sel
active MOVPE ", (8th Int. Co.
nf. on Indium Phosphide an
d Related Materials, paper
ThA2-1 (1996)).
The details will be described below with reference to the drawings.

FIG. 6 is a perspective view of a conventional integrated optical modulator of this type. This manufacturing process is shown in FIG. 7, FIG. 8, and FIG.

First, the plane orientation of the surface is n of the (100) plane.
A diffraction grating 3 in which grooves parallel to the (011) plane are arranged at a pitch of 243 nm is formed in a DFB laser unit 2 having a length of 400 μm on a mold InP substrate 1. Next, the gap is 1.8 μm.
Then, a stripe mask of two pairs of m silicon dioxide films 4 is formed. The width of the silicon dioxide film 4 is 10 μm in the DFB laser section 2 and 4 μm in the modulator section 5. In the section of the modulator section 5 opposite to the end of the DFB laser section 30 μm, two pairs of silicon dioxide The shape of the film 4 is closed. The length of the modulator section 5 is 250 μm.
A transition region portion 6 in which the width of the silicon dioxide film 4 continuously changes is inserted between the DFB laser portion 2 and the DFB laser portion 2 with a length of 50 μm. This silicon dioxide film 4 serves as a growth prevention mask in metal organic chemical vapor deposition (hereinafter, referred to as “MO-VPE”). By changing the mask width between the DFB laser unit 2 and the modulator unit 5, it is possible to change the transition level of the multiple quantum well structure (hereinafter, referred to as "MQW") by MO-VPE.

In this selective growth of MO-VPE, InP
And InGaAsP are epitaxially grown. At this time, the source gas is trimethylindium (hereinafter, “T
MI ". ), Trimethylgallium (hereinafter referred to as “TM
G ". ), Arsine (hereinafter referred to as AsH3.), Phosphine (hereinafter, used as.) "PH _3", organometallic supplies by bubbling of hydrogen.
Regarding doping, disilane (hereinafter referred to as “Si
That ₂ H ₆ ". ), Dimethyl zinc (hereinafter “DMZ”)
n ". ) Is diluted with hydrogen. Also,
The growth pressure is 100 Torr.

After the formation of the silicon dioxide film 4, MO-VPE selective growth is performed with a layer structure as shown in FIG. First, 1.
Starting with the n-InGaAsP guide layer 7 having a composition of 13 μm, the n-InP spacer layer 8 and the n-InP having a composition of 1.2 μm
An nGaAsP SCH layer 9 and a strained MQW layer 10 are sequentially grown. The strained MQW layer 10 has eight layers of compression-strained InGaAsP.
InGaAs having a composition of 1.2 μm between well layers 11
The transition wavelength is 1.56 μm in the DFB laser unit 2 and 1.47 μm in the modulator unit 5. In addition, a 1.2 μm composition In
GaAsP SCH layer 13, InGaAsP intermediate layer 1
4. A p-InGaAsP SCH layer 15 and a p-InP cladding layer 16 are grown. Here, the p-side SCH layer is
By gradually changing the band gap from InGaAsP having a composition of 1.2 μm to InP, hole carriers generated in the modulator can be quickly extracted to the p-side electrode 24. In this way, a selective growth shape as shown in FIG. 7 is obtained. In MO-VPE, since the growth rate of the (111) B plane is slow, (1)
11) Although the B surface is formed, the (111) A surface is formed on the side surface of the window portion 17 of the light emitting portion of the modulator 5.

Next, 1 μm on both sides of the selective growth ridge portion 18
Of silicon dioxide film 4 is removed. The silicon dioxide film 4 is removed so that the entire width of the window portion 17 becomes equal. Then, the p-InP buried layer 1
9, p-InGaAsP contact layer 20, p-InG
An aAs contact layer 21 is grown. Then, as shown in FIG. 8, the epi of the flat growth portion other than the ridge portion is removed.

Next, the p-InGaAsP of the transition region 6 is formed.
Contact layer 20, p-InGaAs contact layer 21
Is removed, and a silicon dioxide film 22 opened only on the selective growth upper surfaces of the DFB laser unit 2 and the modulator unit 5 is formed as shown in FIG.

Next, a polyimide 23 is formed so that the upper surface of the selective growth is exposed, and a p-side electrode 24 is formed on the DFB laser unit 2 and the modulator unit 5. Next, after the back surface is polished to reduce the thickness of the wafer to 120 μm, the n-side electrode 25 is formed.

Finally, the end face is cleaved, and as shown in FIG.
An AR coating film 26 made of a silicon nitride film is applied to the end face on the modulator section 5 side. The end face on the side of the DFB laser section 2 is coated with three layers of silicon dioxide film / amorphous silicon film / silicon dioxide film to obtain an end face reflectivity of 75%.

The reflectivity of the end face of the modulator section 5 can be reduced to 0.04% or less by the structure of the window section 17 + AR coating film 26 formed by the above steps. As described in Aoki et al., "Influence of front-end face light reflection on transmission characteristics in electroabsorption-type optical modulator integrated DFB laser", Paper No. C-311, Proc. Of the 1996 IEICE Electronics Society Conference. , 0.04% or less, 2.48
832 Gb / s of 500 km transmission can be performed.
However, at 10 Gb / s, the influence of chirping becomes larger, so that this level of return light suppression requires 80 km
The transmission distance can be extended only to the extent.

[0014]

In the prior art, 10 Gb / s is limited due to the limitation of the chromatic dispersion of the optical fiber.
Cannot be extended beyond 80 km. The reason is that the return light from the modulator to the DFB laser
Since the light density of the active layer of the DFB laser fluctuates, which changes the carrier density and further changes the refractive index,
This is because the oscillation wavelength of the DFB laser determined by the diffraction grating fluctuates. Since such a fluctuation occurs at a frequency called a relaxation oscillation frequency of the DFB laser, a larger influence is brought when the transmission bit rate approaches the relaxation oscillation frequency. Normally, the relaxation oscillation frequency of a DFB laser is 10
Since the frequency is on the order of GHz, long-distance transmission cannot be performed at 10 Gb / s transmission unless chirping is kept very small.

Also, the Dispersion Compe
There is also a method of compensating for dispersion by inserting a nation fiber (DCF). However, this DCF fiber has a drawback that the device becomes large and the cost is greatly increased.

[0016]

SUMMARY OF THE INVENTION It is an object of the present invention to reduce the occurrence of chirping due to return light inside an integrated optical modulator and realize high-speed, long-distance optical transmission.

[0017]

SUMMARY OF THE INVENTION The present invention provides an optical semiconductor device which integrates a semiconductor laser and an optical modulator, an optical variable
In the part of the modulator, the optical waveguide has its output at its output end face.
It is not perpendicular to the end face but a bent waveguide,
The bend is in a direction perpendicular to the substrate. Also, the optical waveguide of the semiconductor laser is bent and guided.
It may be a wave path.

That is, in the above optical semiconductor device, the modulator and the distributed feedback semiconductor laser are monolithically integrated, and the distributed feedback semiconductor laser is disposed on the side where the optical waveguide and the end face are perpendicular to each other. The modulator is arranged on the side where the optical waveguide and the end face are inclined obliquely.

By bending the optical waveguide in the middle of the chip, the optical waveguide and the end face can be made perpendicular to one end face and oblique to the optical waveguide at the other end face. In the integrated optical modulator, the rear end face of the uniform grating DFB laser needs to have high reflection, and the emission end face on the modulator side needs to have very low reflection. Reflection can be realized.
In the oblique waveguide, light reflected on the end face does not return to the original waveguide, so that the reflection can be substantially reduced.

In the integrated type optical modulator, as the speed becomes longer and the distance becomes longer, suppression of returning light from the modulator to the DFB laser becomes a major problem. The diagonal waveguide at the exit face of the modulator described above is combined with a window structure and an AR coating to return light from the modulator to the DFB laser.
It can be suppressed to 0.01% or less. When the return light is suppressed to 0.01% or less, chirping caused by the return light can be ignored, and 10 Gb / s,
200 km transmission is achieved.

In the case of a curved waveguide, if the radius of curvature is too small, there is a problem that waveguide loss occurs and output power is reduced. This is because the active layer width is 1.5 μm
In this case, if the radius of curvature is 2 mm or more, and if the width of the active layer is 1 μm or more, if the radius of curvature is 5 mm or more, there is no problem.

[0022]

Figure 1 DETAILED DESCRIPTION OF THE INVENTION, reference underlying the present invention
It is a perspective view which shows a technique .

Referring to FIG. 1, the element structure of the integrated optical modulator is roughly divided into a DFB laser section 2, a transition region section 6, and a modulator section 5. In the DFB laser section 2, the n-type In
A diffraction grating 3 is formed on a P substrate 1. Modulator section 5
The active layer has a smaller band gap than the active layer of the DFB laser unit 2, and the DFB
The laser light can be extinguished.
The transition region 6 is a region where the band gap of the active layer changes continuously. The p-side electrode 24 is formed separately from the DFB laser unit 2 and the modulator unit 5.

The feature here is that the optical waveguide of the modulator section 5 is a bent waveguide. The radius of curvature of this bent waveguide is 1 to 3 mm. Modulator section 5
Has a length of 100 μm to 300 μm,
In the emission end face, the angle between the waveguide and the end face is inclined at an angle of 5 ° or more with respect to 90 °. This oblique angle is more desirably 7 ° or more.

The end face of the modulator section 5 is provided with an AR coating film 26, and the inside has a window structure. Thus, AR coating + window structure +
By combining the oblique waveguides, the front end face reflectance is suppressed to 0.01% or less.

As a modification of the present embodiment, the bent waveguide portion can be expanded not only to the modulator section 5 but also to the transition region section 6 and the DFB laser section 2. At this time, the diffraction grating 3 of the DFB laser unit 2
Is a diffraction grating bent as the optical waveguide is bent.
The total length of the chip including the modulator section 5, the transition region section 6 and the DFB laser section 2 is 500 to 900 μm, and the radius of curvature of the whole waveguide is 2 to 6 mm. The output end face of the modulator section 5 is an oblique waveguide in which the optical waveguide and the end face are inclined by 7 ° to 20 ° with respect to 90 °.

Next, an example of the present reference technology will be described in detail with reference to the drawings.

Referring to FIG. 1, in this example , the element is a DF
It comprises a B laser unit 2, a transition region unit 6, and a modulator unit 5, each having a length of 400 μm, 50 μm, 250 μm.
μm. These are all n with a plane orientation of (100).
It is formed monolithically on the mold InP substrate 1. D
In the FB laser unit 2, a diffraction grating 3 having a pitch of 243 nm is formed on the n-type InP substrate 1. The bandgap wavelengths of the active layers of the DFB laser unit 2, the transition region unit 6, and the modulator unit 5 are 1560 nm in the DFB laser unit 2, 1470 nm in the modulator unit 5, and the transition region unit 6 has a continuous gap therebetween. It is something that connects to.

The silicon dioxide film 22 has an opening in the DFB laser unit 2 and the modulator unit 5, and the p-side electrode 2
4 is formed separately in the DFB laser unit 2 and the modulator unit 5. The optical waveguide of the modulator section 5 has a radius of curvature of 2 m.
m, and the output end face is inclined at an angle of 7 °. An AR coating film 26 made of a silicon nitride film is provided on the front end surface.

Next, the manufacturing method of this embodiment will be described in detail with reference to FIGS. 2, 3, 4 and 1.

First, the plane orientation of the surface is n of the (100) plane.
A diffraction grating 3 in which grooves parallel to the (011) plane are arranged at a pitch of 243 nm is formed in a DFB laser unit 2 having a length of 400 μm on a mold InP substrate 1. Next, the gap is 1.8 μm.
Then, a stripe mask of two pairs of m silicon dioxide films 4 is formed. The width of the silicon dioxide film 4 is 10 μm in the DFB laser unit 2 and 4 μm in the modulator unit 5. At the front end portion of the modulator unit 5, the width of the silicon dioxide film 4 is closed at 30 μm. . The length of the modulator section 5 is 2
A transition region section 6 where the width of the silicon dioxide film 4 continuously changes is inserted between the modulator section 5 and the DFB laser section 2 with a length of 50 μm. The stripe of the silicon dioxide film 4 is oriented in the [011] direction in the DFB laser unit 2, but is bent at a curvature radius of 2 mm in the modulator unit 5. The silicon dioxide film 4 serves as a growth blocking mask in MO-VPE growth. By changing the mask width between the DFB laser unit 2 and the modulator unit 5, the transition wavelength of the MQW by the MO-VPE can be changed.

In this selective growth of MO-VPE, InP
And InGaAsP are epitaxially grown. TMI, TMG, AsH ₃ and PH ₃ are used as source gases, and the organic metal is supplied by bubbling hydrogen. For doping, a gas obtained by diluting Si ₂ H ₆ and DMZn with hydrogen is used as appropriate. The growth pressure is 100 Torr.
r.

After the formation of the silicon dioxide film 4, MO-VPE selective growth is performed as shown in FIG. First, starting with the n-InGaAsP guide layer 7 having a composition of 1.13 μm,
nP spacer layer 8, n-InGaAs having a composition of 1.2 μm
A PSCH layer 9 and a strained MQW layer 10 are sequentially grown. Distortion M
The QW layer 10 is composed of eight compression-strained InGaAsP well layers 1
1 and between 1. InGaAsP barrier layer 1 of μm composition
The transition wavelength is 1.56 μm in the DFB laser unit 2 and 1.47 μm in the modulator unit 5. In addition, a 1.2 μm composition InGaAsP SC
H layer 13, InGaAsP intermediate layer 14, p-InGaAs
The sP SCH layer 15 and the p-InP clad layer 16 are grown. Here, the SCH layer on the p-side is made of I
By gradually changing the band gap from nGaAsP to InP, hole carriers generated in the modulator section 5 can be quickly extracted to the p-side electrode 24.
In this way, a selective growth shape as shown in FIG. 2 is obtained. In the MO-VPE, since the growth rate of the (111) B plane is low, the (111) B plane is generally formed on the selective growth side surface. The (111) A plane is formed.

Next, 1 μm on both sides of the selective growth ridge portion 18
Of silicon dioxide film 4 is removed. The silicon dioxide film 4 is removed so that the entire width of the window portion 17 becomes equal. Then, a p-InP buried layer 19,
p-InGaAsP contact layer 20, p-InGaAs
The s-contact layer 21 is grown. Then, as shown in FIG. 3, the epi of the flat growth portion other than the ridge portion is removed.

Next, the p-InGaAsP of the transition region 6
Contact layer 20, p-InGaAs contact layer 21
Is removed, and a silicon dioxide film 22 opened only on the selective growth upper surfaces of the DFB laser unit 2 and the modulator unit 5 is formed as shown in FIG.

Next, a polyimide 23 is formed so that the upper surface of the selective growth is exposed, and a p-side electrode 24 is formed on the DFB laser unit 2 and the modulator unit 5. Next, after the back surface is polished to reduce the thickness of the wafer to 120 μm, the n-side electrode 25 is formed. Finally, the end face is cleaved, and an AR coating film 26 of a silicon nitride film is applied to the end face on the modulator section 5 side as shown in FIG. On the rear end face of the DFB laser unit 2,
An HR coating film having a reflectance of 75% is formed by three-layer coating of a silicon dioxide film / amorphous silicon film / silicon dioxide film.

Through the above manufacturing steps, the output end face of the modulator section 5 becomes a 7 ° oblique waveguide + window structure + AR coating. As a result, the total front end face reflectance is suppressed to 0.01% or less.

In the above example , the InP on the InP substrate
Although the case of strained MQW of GaAsP has been described, it is needless to say that an InGaAlAs-based MQW can be similarly implemented.

FIG. 5 is a perspective view showing an embodiment of the present invention.

Referring to FIG. 5, in this embodiment, the modulator section 5 is a bent waveguide bent in the vertical direction.
In this case, a greater effect can be obtained even with the same radius of curvature. The emission part of the modulator part 5 has a window structure, and the active layer is interrupted at the (111) A plane.
Therefore, when the beam exits from the active layer in the modulator section 5 to the window section, the beam is refracted and bent slightly downward even if it does not form an oblique waveguide. In addition, since the waveguide is bent downward, the properties of the oblique waveguide can be obtained in a more expanded form.

In order to bend the beam downward, the light reflected on the front end face is reflected toward the inside of the n-type InP substrate 1. For this reason, there is also an advantage that there is no concern that the return light is irregularly reflected on the ridge side surface of the selective growth.

When manufacturing this, in the first step, n
A step of etching the type InP substrate 1 is inserted. Due to this etching, the portion of the modulator section 5 has a radius of curvature of 2 mm.
Shape.

This method can be further combined with a horizontally bent waveguide to further enhance the effect.

[0044]

The effect of the present invention described above is as follows. In addition to the window structure and the AR coating, by forming an oblique waveguide at the output end, the return light from the modulator to the DFB laser can be reduced even in an integrated device. That is, it can be suppressed to 40 dB or less. As a result, 10 Gb
/ S optical communication also enables long-distance transmission of 200 km.

The reason is that chirping that limits the transmission distance is induced by the return light from the modulator to the DFB laser. Since the return light is modulated, D
Light density fluctuations occur in the FB laser, which cause carrier density fluctuations of the carriers contributing to laser oscillation, and further cause fluctuations in the refractive index of the active layer. Since the oscillation wavelength of the DFB laser is determined by the pitch of the diffraction grating and the equivalent refractive index of the optical waveguide, the oscillation wavelength changes, that is, chirping occurs due to the change in the refractive index of the active layer. Therefore, if return light can be suppressed, it is possible to suppress chirping.

[Brief description of the drawings]

FIG. 1 is a perspective view showing a reference technique on which the present invention is based .

FIG. 2 is a perspective view in a manufacturing process of a reference technology on which the present invention is based .

FIG. 3 is a perspective view in a manufacturing process of the reference technology on which the present invention is based .

FIG. 4 is a perspective view in a manufacturing process of the reference technology on which the present invention is based .

FIG. 5 is a perspective view showing an embodiment of the present invention.

FIG. 6 is a perspective view showing a conventional technique.

FIG. 7 is a perspective view of a conventional manufacturing process.

FIG. 8 is a perspective view of a conventional manufacturing process.

FIG. 9 is a perspective view of a conventional manufacturing process.

FIG. 10 is a sectional view showing a layer structure of a selective growth portion of the integrated optical modulator.

[Explanation of symbols]

 Reference Signs List 1 n-type InP substrate 2 DFB laser part 3 diffraction grating 4 silicon dioxide film 5 modulator part 6 transition region part 7 n-InGaAsP guide layer 8 n-InP spacer layer 9 n-InGaAsP SCH layer 10 strained MQW layer 11 compressive strained InGaAsP Well layer 12 InGaAsP barrier layer 13 InGaAsP SCH layer 14 InGaAsP intermediate layer 15 p-InGaAsP SCH layer 16 p-InP cladding layer 17 window part 18 selective growth ridge part 19 p-InP buried layer 20 p-InGaAsP contact layer 21 p-InA Contact layer 22 silicon dioxide film 23 polyimide 24 p-side electrode 25 n-side electrode 26 AR coating film

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) H01S 5/00-5/50 G02B 6/42-6/43

Claims (2)

(57) [Claims] In an optical semiconductor device in which a semiconductor laser and an optical modulator are integrated, in an optical modulator portion, an optical waveguide is provided at an output end face thereof.
A bent waveguide that is not perpendicular to the output end face
It has been, an optical semiconductor device in which the bend is perpendicular to the substrate direction, and wherein the. 2. An integrated semiconductor laser and an optical modulator.
In optical semiconductor devices,In the part of the optical modulator, an optical waveguide is provided at its output end face.
A bent waveguide that is not perpendicular to the output end face
The bend is perpendicular to the board, The optical waveguide in the portion of the semiconductor laser becomes a bent waveguide.
ing, An optical semiconductor device, comprising: