JP2002344068A - Optical semiconductor device and manufacturing method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce optical absorption in a waveguide and oscillate a stable laser of low threshold, in an optical semiconductor device which is used as a light source for optical communication, optical disk, optical interconnection, etc. SOLUTION: This optical semiconductor device is provided with a striped quantum well structure layer 6, and an upper clad layer 8 that is formed on the quantum well structure layer 6; the quantum well structure layer 6 is formed starting from a gain region A to a mode conversion region B on a semiconductor substrate 1, whose film is formed thinner in the mode conversion region B as it becomes far away from the gain region A, and where the energy band gap due to resulted quantum effect becomes larger apart from the gain region A. Its film thickness in the boundary between the mode conversion region B and gain region A is two times or more that at the light outgoing end of the mode conversion region B.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical semiconductor device and a method of manufacturing the same, and more particularly, to an optical semiconductor device used for a light source such as an optical communication device, an optical disk device, and an optical interconnection, and a method of manufacturing the same.

[0002] With the advancement of semiconductor communication technology, the fabrication technology of semiconductor lasers has also been improved. In recent years, research on integrating semiconductor lasers and other optical semiconductor elements has been actively conducted. For example,
There are devices in which a DFB laser and an optical modulator are integrated, and devices in which a DBR laser and a mode converter (beam size converter) are integrated.

A mode converter is a mechanism for narrowing an output beam of a semiconductor laser which has an emission angle as wide as 30 to 40 degrees, and facilitates optical coupling when a semiconductor laser and an optical fiber are modularized. It is.

[0004]

2. Description of the Related Art In a semiconductor laser, as the light confinement of an optical waveguide becomes stronger, that is, as the light spot diameter becomes smaller, the oscillation threshold value becomes smaller and the light emission efficiency improves. Becomes difficult.

Therefore, a mode converter integrated Fabry-Perot semiconductor laser (hereinafter, referred to as an FP-LD) having a waveguide for converting a light beam diameter made of a semiconductor has been proposed in, for example, the following document.

[1] TL KOCH et al., IEEE PHOTONIC
S TECHNOLOGY LETTERS, VOL.2, NO.2, 1990 The FP-LD, as shown in FIG.
A first InP cladding layer 102, a waveguide layer 103, a multiple quantum well (MQW) active layer 104, and a second InP cladding layer
105 are formed in order. MQW active layer 104
Is composed of an InGaAs well layer and an InGaAsP barrier layer, and is formed only in the gain region 110. In addition, the waveguide layer 103
Are formed in both the gain region 110 and the mode conversion region 111. Reference numeral 106 denotes the second InP cladding layer 10.
The contact layer 107 formed on 5 indicates an etching stop layer formed in the second cladding layer.

The waveguide layer in the mode conversion region 111
103 has a tapered shape in the thickness direction,
It fades away from 110. Waveguide layer 10
3 has a structure in which InGaAsP layers 103a and InP layers 103b are alternately stacked. Then, using the InP layer 103b as an etching stop layer and changing the etching resistance mask a plurality of times, the InGaAsP layer 103a and the InP layer 103b are patterned in a stepwise manner, whereby the waveguide layer 103 in the mode conversion region is formed.
Has a tapered shape.

The light excited by the MQW active layer 104 and guided through the waveguide layer 103 has a weaker light confinement in the mode conversion region 111 than in the gain region 110, so that the near-field image at the tapered tip portion expands. As a result, the far-field image, which is the diffraction pattern of the near-field image, becomes narrow. Therefore, the exit angle of the beam emitted from the tapered tip becomes narrow, and the coupling with the optical fiber becomes easy.

[0009]

By the way, the waveguide layer 103 constituting the resonator has a structure in which the number of layers decreases as approaching the output end. The tapered shape of the waveguide layer 103 is such that the etching-resistant mask is used a plurality of times. Since the structure is obtained by replacement, the structure tends to include crystal defects, and the characteristics of the light-emitting element are likely to deteriorate.

When integrating a tapered waveguide layer into a semiconductor laser, it is important to prevent the waveguide in the mode conversion region from becoming an absorption medium for oscillating light.

In the structure described in the above document [1], the active layer 104 in the gain region 110 and the tapered waveguide layer 103 in the mode conversion region 111 are different layers and have substantially the same composition. Part of the laser light is applied to the tapered waveguide layer
It is easy to be absorbed in the portion 103 near the gain region 110.

Therefore, compared with a laser device in which the mode converter is not integrated, the light output and the slope efficiency (the rising differential value of the current-light intensity characteristic line) are inevitably reduced. In particular, in the FP-LD described in the document [1], the tapered waveguide layer 103 in the resonator affects the basic characteristics of the semiconductor laser. It has risen to 70 mA.

In addition, a device at both ends is cleaved at room temperature continuously (C
W) There is no report that oscillation has occurred.

Further, a device in which a semiconductor laser and a tapered waveguide are integrated is described in [2] JP-A-63-233584 and [3] JP-A-64-53487. Also in these structures, the laser active layer and the tapered waveguide have the same composition, and the tapered region has a large absorption loss. To cancel this absorption loss, a large current is injected into the entire tapered waveguide. There is a need.

However, in the current-light intensity characteristics of the publication [2], as is apparent from the figure, an oscillation threshold of 42 mA and a differential quantum efficiency of only 0.15 mW / mA are obtained. Further, since the tapered shape of the waveguide layer can be obtained by devising an etching method, it is difficult to always form the tapered shape uniformly, so that the controllability of the beam spot shape is not good and the yield is reduced. There is fear. In addition, a crystal defect easily enters the waveguide layer constituting the resonator, and the characteristic of the light emitting element is likely to deteriorate. Further, the waveguide layer is made of a single material and has a tapered MQ shape as described in Reference [1].
It does not have a W layer.

In the publication [3], an optical semiconductor device having a tapered shape in the lateral direction is required. Sub-micron-order processing is required at the tip of the tapered waveguide. It is difficult to produce a shape with good reproducibility.

The present invention has been made in view of the above problems, and has been made in view of the above circumstances. An optical semiconductor device having a mode converter capable of reducing a light absorption in a waveguide and performing stable laser oscillation at a low threshold value, and manufacturing the same. The aim is to provide a method.

[0018]

Means for Solving the Problems The above-mentioned problems are shown in FIG.
As illustrated in FIGS. 10 and 11, a semiconductor substrate 1 containing impurities is formed on the semiconductor substrate 1 from the gain region A to the mode conversion region B, and the gain region A is formed in the mode conversion region B. The energy band gap increases as the distance from the gain region increases due to the quantum effect due to the film thickness being reduced as the distance from the mode conversion region increases. A quantum well structure layer 6 having a stripe shape that is at least twice as thick as the film thickness of the light emitting end of B, and an upper cladding layer 8 formed on the quantum well structure layer 6 are provided. It is solved by the optical semiconductor device which does.

In the above-described optical semiconductor device, the quantum well structure layer 6 has a multiple quantum well structure 6 having a plurality of well layers 6a and a plurality of barrier layers 6b sandwiching the well layers 6a. It may be constituted by a single quantum well structure having a barrier layer 6b sandwiching the well layer 6a.

In the optical semiconductor device described above, a lower cladding layer 2 may be formed between the quantum well structure layer 6 and the semiconductor substrate 1.

In the above optical semiconductor device, SCH layers 3 and 7 may be formed above or below the quantum well structure layer 6. In this case, the SCH layers 3 and 7 may have a structure in which the film thickness gradually decreases toward the light emitting end.

In the optical semiconductor device described above, the first electrode 18 formed below the semiconductor substrate 1 and the second electrode 17 formed on the gain region A in the upper cladding layer 8 May be further provided.

As shown in FIGS. 1 and 6 (a), the above-mentioned problem is solved by forming a semiconductor substrate 1 containing an impurity, and forming the semiconductor substrate 1 from the gain region A to the mode conversion region B on the semiconductor substrate 1, and In the mode conversion region B, a stripe-shaped quantum well structure layer 6 whose energy band gap is widened away from the gain region due to a quantum effect because the film thickness is reduced as the distance from the gain region A is increased; An upper cladding layer 8 formed on the well structure layer 6, a first electrode 18 formed on the lower surface of the semiconductor substrate 1, and a range from the gain region A to a part of the mode conversion region B. The optical semiconductor device has a second electrode 17 formed on the upper clad layer 8.

In the above-mentioned optical semiconductor device, the second
Of the mode conversion region B and the gain region A
May be separated in the vicinity of the boundary.

The above-mentioned problem is caused by the problem that the semiconductor substrate 1 containing impurities is formed on the semiconductor substrate 1 from the gain region A to the mode conversion region B, and in the mode conversion region B, as the distance from the gain region A increases. A stripe-shaped quantum well structure layer 6 whose energy band gap widens away from the gain region due to a quantum effect due to the thin film thickness, and an upper cladding layer formed on the quantum well structure layer 6 8, a first electrode 18 formed on the lower surface of the semiconductor substrate 1, a second electrode 17 formed on the upper cladding layer 8 in the gain region A, and a second electrode 17 A contact layer 14 formed between the upper cladding layer 8 and extending to the mode conversion region B side with respect to the second electrode 17; It is solved by an optical semiconductor device according to claim to.

In the optical semiconductor device, the contact layer 14 extends from the gain region A to the mode conversion region B
May be extended halfway. In the optical semiconductor device, it is preferable that the thickness of the quantum well structure layer 6 in the mode conversion region B is twice or more at the light incident end as compared with the light emitting end.

In each of the above-described optical semiconductor devices, the ratio of the width of both ends of the quantum well structure layer 6 is 0.8 to 0.8.
It is preferably 1.2.

In each of the above optical semiconductor devices, the well layer 6a in the quantum well structure layer 6 is made of InGaAsP.
Alternatively, the upper cladding layer 8 may be formed of a compound semiconductor layer containing GaAs, and the upper cladding layer 8 may be formed of InP. The well layer 6a in the quantum well structure layer 6 is made of Ga
The upper cladding layer 8 may be made of AlGaAs. Further, the well layer 6 in the quantum well structure layer 6 may be formed of GaInP, and the upper cladding layer 8 may be formed of AlGaInP.

In each of the above optical semiconductor devices, a structure in which a diffraction grating constituting a distributed feedback mirror is formed along the quantum well structure layer 6 in the gain region A may be employed. Further, a structure in which a diffraction grating forming a distributed reflection type mirror is formed in front of or behind the gain region A along the waveguide direction may be employed.

As shown in FIGS. 2 to 4, the above-described problem is caused by forming a first opening 5 a in the form of a stripe on the semiconductor substrate 1 containing impurities and connecting the first opening 5 a to the first opening 5 a. Forming a mask 4 having a second opening 5b wider than the first opening 5a, and growing a semiconductor layer in the first and second openings 5a and 5b, The maximum thickness of the first opening 5a is equal to the second opening 5b.
An optical semiconductor device comprising: a step of forming a quantum well structure layer 6 having a thickness at least twice the minimum film thickness of the above; and a step of forming an upper cladding layer 8 on the quantum well structure layer 6. Is solved.

In the method for manufacturing an optical semiconductor device, the quantum well structure layer 6 includes a plurality of well layers 6a and
This may be either a multiple quantum well structure having a plurality of barrier layers 6b sandwiching a, or a single quantum well structure having a well layer 6a and a barrier layer sandwiching the well layer 6b.

It should be noted that the above-mentioned figure numbers and reference numerals are cited for facilitating the understanding of the present invention, and the present invention is not limited thereto.

[0033]

According to the present invention, the active layer in the gain region and the waveguide for mode conversion are composed of one stripe-shaped quantum well structure layer, and the number of the quantum well structure layers of the waveguide is changed. Instead, the light beam is made thinner as it goes away from the gain region, and the diameter of the light beam in the film thickness direction is converted.

As for the thickness of the quantum well structure layer in the mode conversion region, the film thickness at the incident end in the mode conversion region B is more than twice as large as the film thickness at the emission end. According to this, the beam spot emitted from the quantum well structure layer has a circular shape or a shape very close to a circular shape, and the beam emission angle is in the range of 20 to 10 degrees. The coupling efficiency with the above is greatly improved.

In this structure, the quantum well layer forming the mode conversion waveguide is thinner than the thickness of the quantum well layer forming the active layer, and the ground level of the quantum well is more active in the mode conversion waveguide. Since the height is higher than that of the layer, the light absorption wavelength end of the mode conversion waveguide at the emission end is shorter than the oscillation wavelength of the active layer.

Thus, light guided through the quantum well structure layer at the emission end is not absorbed in the mode conversion region.

The thickness of the quantum well structure layer is controlled by using the property that the growth rate changes by changing the area of the opening of the mask used for crystal growth. However, in the vicinity of the gain region, the thickness of the mode conversion waveguide is not sufficiently reduced, resulting in a large loss.

According to the present invention, the current is caused to flow locally by providing an electrode in a portion close to the gain region of the mode conversion waveguide or extending the contact layer. Are transparent and light absorption is eliminated. An electrode formed in a portion near the gain region of the mode conversion waveguide may extend the electrode in the gain region or may be separated from the electrode in the gain region. It becomes easy to control.

According to another aspect of the present invention, the ratio of the minimum value of both end widths of the striped quantum well structure layer is set to 0.8 to 1.
By setting to 2, a circular beam spot can be obtained, and optical coupling becomes easy.

The materials of the well layers and the cladding layers of the quantum well structure layer are not particularly limited. For example, if the well layer is InGa
The AsP cladding layer may be formed from InP, the well layer may be formed from GaAs, the cladding layer may be formed from AlGaAs, the well layer may be formed from GaInP, and the cladding layer may be formed from AlGaInP. The laser formed in such an optical semiconductor device may be of a Fabry-Perot type, a distributed feedback type having a diffraction grating or a distributed reflection type.

[0041]

FIG. 1 is a perspective sectional view showing a Fabry-Perot semiconductor laser according to a first embodiment of the present invention.

This semiconductor laser has a gain peak of 1.3.
The structure appears in the μm band, and the structure will be described along with the manufacturing process.

First, as shown in FIG. 2A, an n-type (n-)
A first mask 4 having at least a gain region A and a mode conversion region B is formed on an InP substrate 1.

The mask 4 is obtained by patterning an insulating film such as a silicon nitride film by photolithography. The opening 5 of the mask 4 has a rectangular shape of 400 μm in length extending in the light traveling direction in the gain region A. Part 5a
And a 350 μm long sector 5 whose width is continuously increased at a constant angle as the distance from the end of the rectangular portion 5a increases
b and an extended portion 5c that is further widened from the end of the sector 5b.

Next, as shown in FIG. 2 (b) and FIG. 3 (a) showing a cross section taken along the line XX of FIG.
An n-InP cladding layer 2 and an In _x Ga _1-x
The first S consisting of As _y P _1-y (0 <x <1, 0 <y <1)
CH layer 3, MQW layer _{6, In x Ga 1-x} As y P 1-y consisting of the second SCH layer 7, a p-InP cladding layer 8 continuously epitaxially grown. For the thickness of each layer,
The n-InP cladding layer 2 is 100 nm, the first and second SCH
Layers 3 and 7 are 100 nm, and p-InP cladding layer 8 is 200 nm.
And

The crystals of these layers do not grow on the mask 4 without the growth nuclei, and the n-type
Grow selectively on InP substrate 1. In addition, these layers are formed in a striped gain region A in the opening 5 of the mask 4.
In the mode conversion region B extending therefrom, the film thickness becomes thinner as the distance from the gain region A increases. Such a film growth method is called a selective growth method.

The thickness of the well layer 6a of the MQW layer 6 at the tip of the mode conversion region B is 1 to the thickness in the gain region A.
/ 3. The MQW layer 6 functions as an active layer in the gain region A, and functions as a waveguide in the mode conversion region B. Since the MQW layer 6 is formed in a tapered shape in the mode conversion region B, that region is hereinafter also referred to as a tapered region.

The MQW layer 6 has three In _{x '} Ga _{1 -x'} As _{y '}
Well layer 6a composed of P _{1-y ′} (0 <x ′ <1, 0 <y ′ <1) and In _{x ″} Ga _{1-x ″} As _{y ″} P sandwiched therebetween
_{1-y ''} (0 <x ''<1, 0 <y ''<1), which is composed of a barrier layer 6b, and a well layer 6a in the gain region A.
Is 7 nm, and the thickness of the barrier layer 6b is 15 nm.

After removing the first mask 4 described above,
A second mask 9 is formed on the p-InP cladding layer 8 along the light traveling direction as shown in FIG. 4 (a) showing a cross section taken along line YY in FIG. 2 (b). The mask 9 is formed by patterning a silicon nitride film into a rectangular stripe shape having a width of 1.5 μm. At this time, the ratio of the minimum value of the stripe width in the gain region A and the mode conversion region B is set to 0.8 to 1.2.
To make the beam spot circular or substantially circular.

Thereafter, as shown in FIG. 4B, the portion from the p-InP cladding layer 8 to the upper portion of the n-InP cladding layer 2 is etched in a substantially vertical direction using the second mask 9 and shaped. After that, without removing the second mask 9, the first p-InP is buried on the n-InP cladding layer 2 on both sides of the second mask 9 as shown in FIG. Layer 10 and n-InP buried layer 11
Are formed in order.

Next, after removing the second mask 9, p-
By stacking p-InP on the InP clad layer 8 and the n-InP buried layer 11, the thickness of the second clad layer 8 is increased, and the second p-InP buried layer 1
Then, a current block layer 13 having a thyristor structure as shown in FIG. 1 is formed on both sides of the MQW layer 6. Further, FIG.
As shown in (b), p ⁺ -type InGaAsP is laminated on p-InP and used as the contact layer 14. The contact layer 14 is prevented from being present in the tapered region B by either using a mask for preventing crystal growth or patterning the semiconductor layer.

Thereafter, an insulating film 15 is formed on the contact layer 14, the second cladding layer 8, and the second p-InP buried layer 8, and is patterned by photolithography to form the MQW of the gain region A. Opening 16 only above layer 6
To form

After completion of the lamination of the compound semiconductor and the insulator in this manner, as shown in FIG.
4 is formed on the lower surface of the n-InP substrate 1.
An electrode 18 is formed. Further, of the InP substrate 1 and the respective semiconductor layers thereon, the end surfaces of the gain region A and the mode conversion region B which are not connected to each other are cleaved. Note that a reflection film may be formed on the cleavage plane of the gain region A.

FIG. 1 is a partially cutaway perspective view of the Fabry-Perot type semiconductor laser formed by the above steps. MQW of this semiconductor laser
The tapered shape of the layer in the mode conversion region B can be obtained by devising the shape of the mask during film growth without changing the etching method.
Crystal defects are less likely to occur in the QW layer. Therefore, there is no deterioration in characteristics due to the tapered crystal defect. Moreover,
The tapered shape is formed with a small error with a high yield, and a beam spot with a uniform shape can be obtained.

In the MQW 6 of this semiconductor laser, the light oscillated in the gain region A resonates by reflection between the end surface of the gain region A and the end surface of the mode conversion region B, and is output from the end surface on the side of the mode conversion region B. Become. In this case, the ground level potential of the quantum well increases as the well layer 6a becomes thinner due to the quantum size effect, and the band gap between the conduction band and the valence band of the well also increases. Since the band gap Eg ₁ in the gain region A shown in FIG. 5A is smaller than the band gap Eg _{2 in the} mode conversion region B shown in FIG. 5B, the MQW layer 6 in the mode conversion region B The light oscillated at A is transmitted.

However, the mode conversion region (taper region) B
Since the change in the film thickness of the MQW layer 6 is gentle, the portion of the tapered MQW layer 6 close to the gain region A becomes a saturable absorption layer. If a saturable absorption layer is present, light is absorbed until the carriers are saturated in the saturable absorption layer. Therefore, the threshold value becomes high, and the current-light output characteristic (IL characteristic) curve is kinked, and Light intensity cannot be obtained.

Therefore, as shown in FIG. 6A, the contact layer 14 formed in the gain region A of the semiconductor laser and the
It is preferable that the electrode 17 extends to the saturable absorption region C near the tapered region B, and the current is injected also into the MQW layer 6 in the saturable absorption region C.

According to this, the current is also supplied to the MQW layer 6 in the saturable absorption region C through the electrode 17, and the absorption of light in that region is canceled by the injection of carriers, and the MQW layer 6 becomes a saturable absorption layer. It becomes transparent and transmits light. As a result, kink does not occur in the threshold rise characteristic curve, and the characteristic shown in FIG. 7 is obtained.

In the IL characteristic curve shown in FIG. 7, the oscillation threshold current of 19 mA is obtained for both cleaved devices, which is lower than the structure shown in the literatures [1] and [2], and the differential quantum efficiency is 0.25.
It was as good as mW / mA. In the mode conversion area B,
Since the end of the light absorption wavelength band is on the shorter wavelength side than the oscillation wavelength and almost no saturable absorption region is generated, no hysteresis or kink is observed in the IL curve, and the gain region A and the tapered region in the MQW layer 6 are not observed. It was confirmed that the reflection and emission of light at the junction of B were also sufficiently suppressed.

As a result of observing the far-field image, the full width at half maximum (FWHM) of the emitted beam was 11.8 ° in the film thickness (vertical) direction as shown in FIGS. It was 8.0 ° in the (horizontal) direction, and a narrower beam emission angle than the conventional one was obtained.

When the relationship between the vertical and horizontal directions of the beam spot diameter in the near-field image and the length (taper length) of the tapered region B was examined, the relationship shown in FIG. 9 was obtained. In this case, the thickness of the well layer 6a of the MQW layer 6 in the tapered region B is one of the thickness of the well layer 6a in the gain region A.
/ 5.

In FIG. 9, if the length at which the film thickness changes abruptly, that is, the taper length is 50 μm or less, the light mode cannot follow the change in the film thickness, and a sufficient function as a mode converter cannot be obtained. On the other hand, when the taper length is 50 μm or more, the change in spot diameter is gradual, and the change in film thickness is sufficient for a mode converter. Although not particularly shown, when the taper length is 500 μm or more, there is no change in the spot diameter, and a longer length is not particularly necessary.

Therefore, it can be said that the optimum taper length is 50 μm or more, preferably 50 to 500 μm.

Next, the optimum change in the thickness of the MQW layer 6 in the mode conversion region B will be described.

FIG. 10 shows an MQW layer 6 serving as a tapered waveguide.
When the film thickness ratio (incident end film thickness / outgoing end film thickness) between the incident end and the outgoing end of the mode conversion region B is changed from 1 to 5,
It is the result of calculating how the beam radiation angle from the exit end changes.

The calculation was performed using the ordinary three-dimensional beam propagation method, and the calculation was performed on the MQW waveguide in the 1.3 μm band. As shown in FIG. 11, the thickness of the MQW layer 6 at the incident end is 0.1 μm, the SCH layers 3 and 7 are 0.1 μm, the length l of the waveguide is 200 μm, and the change in the film thickness is uniform. A lossless waveguide is assumed. The width of the tapered waveguide was constant at 1 μm over the entire length of the waveguide.

In FIG. 10, square dots indicate the radiation angle of the beam in the horizontal direction with respect to the substrate surface, and rhombic dots indicate the radiation angle of the beam in the direction perpendicular to the substrate surface. According to FIG. 10, when the film thickness ratio is set to 1, only a radiation angle close to 30 degrees can be obtained similarly to a general semiconductor laser.

The radiation angle becomes narrower as the film thickness ratio increases, and becomes 15 degrees when the film thickness ratio is 3 and 12 degrees when the film thickness ratio is 4. In addition, the fact that the values of the radiation angles in the horizontal direction and the vertical direction are close indicates that the beam shape approaches a circular shape, and the closer to the circular shape, the better the result for coupling with the optical fiber.

Considering the coupling with the optical fiber, the coupling efficiency is greatly improved when the beam radiation angle in both the horizontal and vertical directions is less than half the conventional value, that is, 15 degrees or less.

This corresponds to the case where the film thickness ratio is 3 or more.
Therefore, it is clear from FIG. 10 that the thickness ratio of the MQW layer 6 is desirably 3 or more when used in a mode converter (spot size converter). However, the radiation angle is 20
Even when the degree, that is, the film thickness ratio is 2, the effect of improving the coupling efficiency can be expected because the beam approaches a circle more than the conventional laser and the mode shape of the laser approaches the mode shape of the optical fiber.

Incidentally, an optical modulator integrated light source in which the film thickness of the MQW layer is changed is disclosed in Japanese Patent Application Laid-Open No. 1-31998.
No. 6, JP-A-3-284891.

In the former publication, the width of the mesa stripe shape of the substrate is changed to make the MQW at the boundary between the gain region and the modulation region.
It describes changing the layer thickness. Also,
The latter publication describes changing the thickness of the MQW layer at the boundary between the gain region and the modulation region by changing the stripe width of the mask. These are for forming the film thickness of the MQW layer in the modulation region to be constant, and there is no description or suggestion about forming the MQW layer in a tapered shape.

In the light source in which the optical modulator is integrated, the MQW layer in the light modulation region has a shorter wavelength composition than the composition of the MQW layer in the gain region. For example, when the composition of the gain region is 1.3 μm, 1.20 μm in the light modulation region.
m to 1.22 μm.

When the composition of the gain region is 1.55 μm, the light modulation region has a composition of 1.40 μm to 1.45 μm.
m.

When this is converted into the film thickness ratio of the MQW layer, the film thickness ratio exists in the range of 1.5 to 2. In the case of a film thickness ratio larger than this, since the composition wavelength in the light modulation region is too large, sufficient light absorption cannot be obtained unless the voltage for light modulation is increased. As a result, the operating voltage of the device must be increased. Actually, the thickness of the MQW layer in the gain region of the optical modulator integrated light source disclosed in these patent publications is twice or less the thickness of the MQW layer in the light modulation region.

Therefore, in a light source in which an optical modulator is integrated,
The change in the film thickness of the MQW layer is desirably small, which is different from the effective change in the film thickness of the semiconductor laser in which the mode converter of the present invention is integrated.

As described above, the present inventors have found that the effective range of the change in film thickness between the light source integrated with the light modulation and the light source integrated with the optical mode converter is different. In the light source integrated with the above, the thickness of the waveguide layer of the mode converter needs to be twice or more the incidence end with respect to the emission end, and by making it five times, a circular beam can be obtained.

As described above, by setting the thickness ratio between the start point and the end point of the tapered region of the MQW layer 6 to be more than twice, a function as a mode converter can be obtained. Further, the change in the film thickness does not need to be uniform, but may change abruptly in the vicinity of the gain region A and change gradually as the distance increases, and conversely, change gradually in the vicinity of the gain region A. Alternatively, the rate of change of the film thickness may be gradually changed a plurality of times. Further, the film thickness may be changed in a gentle step shape.

When the current injected into the saturable absorption region C may be smaller than that in the gain region A, as shown in FIG.
The two p-electrodes 17 and 21 may be formed separately. According to this, the amount of current injection in the saturable absorption region C is reduced, and power consumption is reduced.

Further, since the amount of light absorption in the saturable absorption region C of the tapered MQW layer 6 gradually decreases as the distance from the gain region A increases, the current value for eliminating the light absorption decreases as the distance from the gain region A decreases. I'm done. Accordingly, only the contact layer 14 may be extended from the gain region A to the saturable region C as shown in FIG. According to this, the contact layer 1 made of the impurity-containing semiconductor
4 has a higher resistance than the p-electrode 17 made of metal, and the current injected into the MQW layer 6 via the contact layer 14 decreases as the distance from the gain region A increases, so that the supply of an excessive current can be suppressed.

(Other Embodiments) Although the Fabry-Perot structure is used in the above embodiment, the MQW layer 6 in the mode conversion region B
A DBR laser structure may be provided by providing a diffraction grating below or above. Further, a DFB laser structure may be provided by providing a diffraction grating along the MQW layer in the gain region.

Further, in the above embodiment, the MQW is set to In.
Although it is made of GaAsP / InP-based material,
A material using a GaAs material, an AlGaInP / GaInP material, or another compound semiconductor material may be used.

The growth of the compound semiconductor layer described above
MOCVD, MBE, and other epitaxial growth methods.

[0084]

According to the present invention, the active layer in the gain region and the waveguide for mode conversion are composed of one stripe-shaped quantum well structure layer, and the number of the quantum well structure layers of the waveguide is changed. Instead, the light beam is made thinner as it moves away from the gain region, and the diameter of the light beam in the film thickness direction is converted.

In the structure, the thickness of the quantum well structure layer in the mode conversion region is twice or more the thickness of the incident end in the incident mode conversion region B as compared with the thickness of the exit end. The beam spot shape emitted from the well structure layer has a circular shape or a shape very close to a circular shape, and the beam emission angle exists in a range of 20 degrees to 10 degrees, so that the coupling efficiency with an optical fiber or the like is increased. Can be greatly improved.

In this structure, the quantum well layer forming the mode conversion waveguide is thinner than the thickness of the quantum well layer forming the active layer, and the ground level of the quantum well is more active in the mode conversion waveguide. Since the light absorption wavelength end of the mode conversion waveguide at the emission end is shorter than the oscillation wavelength of the active layer, light guided through the quantum well structure layer is subjected to mode conversion. In the region, the amount of absorption can be reduced.

Further, since an electrode is provided on the cladding layer in a portion close to the gain region of the mode conversion waveguide, or the contact layer is extended, current is caused to flow locally. It is possible to make the portion transparent and eliminate light absorption.

The electrodes may extend from the electrodes in the gain region or may be separated from the electrodes in the gain region. By separating the electrodes, the amount of light absorption can be easily controlled.

According to another aspect of the present invention, the ratio of the minimum value of both end widths of the striped quantum well structure layer is set to 0.8 to 1.
By setting to 2, a circular beam spot can be obtained, and optical coupling is facilitated.

[Brief description of the drawings]

FIG. 1 is a perspective sectional view of an optical semiconductor device showing a first embodiment of the present invention.

FIGS. 2A and 2B are perspective views showing a part of a manufacturing process of the optical semiconductor device according to the first embodiment of the present invention.

FIGS. 3A to 3C are side sectional views showing a part of the manufacturing process of the optical semiconductor device according to the first embodiment of the present invention.

FIGS. 4A to 4C are front sectional views showing a part of a manufacturing process of the optical semiconductor device according to the first embodiment of the present invention.

FIGS. 5A and 5B are diagrams showing energy bands of the quantum well structure according to the first embodiment of the present invention.

FIGS. 6 (a) to 6 (c) are cross-sectional views showing a plurality of structures of an electrode according to the optical semiconductor device of one embodiment of the present invention.

FIG. 7 is a characteristic diagram showing a relationship between current and light output in the optical semiconductor device according to the first embodiment of the present invention.

FIGS. 8A and 8B are vertical and horizontal light intensity distribution diagrams of a far-field image of the optical semiconductor device of the first embodiment.

FIG. 9 is a characteristic diagram showing a relationship between a taper length of a waveguide and a spot diameter of a near-field image of the optical semiconductor device according to the first embodiment of the present invention.

FIG. 10 is a characteristic diagram illustrating a relationship between a film thickness ratio and a beam radiation angle of the optical semiconductor device according to the first embodiment of the present invention.

FIG. 11 is a cross-sectional view showing a model of the waveguide when calculating the characteristics shown in FIG. 10;

FIG. 12 is a cross-sectional view showing an example of a semiconductor laser in which a conventional mode conversion waveguide is integrated.

[Explanation of symbols]

 A Gain region B Mode conversion region 1 Substrate 2, 8 Cladding layer 3, 7 SCH layer 6 Multiple quantum well structure 6a Quantum well layer 6b Barrier layer 13 Current block layer 14 Contact layer 15 Insulating film 16 Opening 17 p electrode 18 n electrode

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Mitsuru Egawa 1015 Uedanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Prefecture Inside Fujitsu Limited (72) Inventor Jiro Okazaki 1015 Kamikodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Fujitsu Limited ( 72) Inventor Shoichi Ogita 1015 Kamiodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Inside Fujitsu Limited (72) Inventor Takuya Fujii 1015 Kamikodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Fujitsu Limited F-term (reference) 5F073 AA21 AA45 AA61 AA67 AA73 AA74 AA86 AA87 AA89 BA02 BA04 BA09 CA04 CA06 CA12 CA14 CB11 DA05 DA06 DA35 EA20 EA29

Claims (19)

[Claims] A semiconductor substrate containing an impurity, wherein the semiconductor substrate is formed on the semiconductor substrate from a gain region to a mode conversion region, and the film thickness of the mode conversion region decreases as the distance from the gain region increases. Due to the quantum effect, the energy band gap increases as the distance from the gain region increases, and the film thickness at the boundary between the mode conversion region and the gain region is twice or more as large as the film thickness at the light emitting end of the mode conversion region. An optical semiconductor device, comprising: a striped quantum well structure layer; and an upper cladding layer formed on the quantum well structure layer. 2. The quantum well structure layer, wherein the quantum well structure layer has a multiple quantum well structure having a plurality of well layers and a plurality of barrier layers sandwiching the well layer, or a single quantum well structure having a well layer and a barrier layer sandwiching the well layer. 2. The optical semiconductor device according to claim 1, wherein the optical semiconductor device has a single quantum well structure. 3. The optical semiconductor device according to claim 1, wherein a lower cladding layer is formed between said quantum well structure layer and said semiconductor substrate. 4. An SCH above or below the quantum well structure layer.
The optical semiconductor device according to claim 1, wherein a layer is formed. 5. The SCH layer according to claim 4, wherein the thickness of the SCH layer gradually decreases toward the light emitting end.
An optical semiconductor device according to item 1. 6. The semiconductor device according to claim 1, further comprising a first electrode formed below said semiconductor substrate, and a second electrode formed on said gain region in said upper cladding layer. 2. The optical semiconductor device according to 1. 7. A semiconductor substrate containing an impurity, wherein the semiconductor substrate is formed on the semiconductor substrate from a gain region to a mode conversion region, and the film thickness of the mode conversion region decreases as the distance from the gain region increases. A stripe-shaped quantum well structure layer in which an energy band gap is spread away from the gain region due to a quantum effect; an upper cladding layer formed on the quantum well structure layer; and a lower surface of the semiconductor substrate. An optical semiconductor device comprising: a first electrode; and a second electrode formed on the upper cladding layer in a range from the gain region to a part of the mode conversion region. 8. The optical semiconductor device according to claim 7, wherein said second electrode is separated near a boundary between said mode conversion region and said gain region. 9. A semiconductor substrate containing impurities, wherein the semiconductor substrate is formed on the semiconductor substrate from the gain region to the mode conversion region, and the film thickness of the mode conversion region is reduced as the distance from the gain region increases. A stripe-shaped quantum well structure layer in which an energy band gap is spread away from the gain region due to a quantum effect; an upper cladding layer formed on the quantum well structure layer; and a lower surface of the semiconductor substrate. A first electrode, a second electrode formed on the upper cladding layer in the gain region, and a second electrode formed between the second electrode and the upper cladding layer. An optical semiconductor device comprising: a contact layer extending to the mode conversion region side. 10. The optical semiconductor device according to claim 9, wherein said contact layer extends from said gain region to a part of said mode conversion region. 11. The quantum well structure layer according to claim 7, wherein the thickness of the quantum well structure layer in the mode conversion region is twice or more at the light incident end as compared with the light emitting end. Optical semiconductor device. 12. The optical semiconductor device according to claim 1, wherein a ratio of widths at both ends of the quantum well structure layer is 0.8 to 1.2. 13. The quantum well structure layer according to claim 1, wherein said well layer is made of InGa.
10. The optical semiconductor device according to claim 1, wherein the optical semiconductor device is formed of a compound semiconductor layer containing AsP or GaAs, and the upper cladding layer is formed of InP. 14. The well layer in the quantum well structure layer is formed of GaAs.
10. The optical semiconductor device according to claim 1, wherein the upper cladding layer is made of AlGaAs. 15. The quantum well structure layer according to claim 15, wherein the well layer is GaInP.
10. The optical semiconductor device according to claim 1, wherein the upper cladding layer is formed of AlGaInP. 16. The gain region according to claim 1, wherein a diffraction grating constituting a distributed feedback mirror is formed along the quantum well structure layer. Optical semiconductor device. 17. A diffraction grating constituting a distributed reflection type mirror is formed in front of or behind said gain region along a waveguide direction. An optical semiconductor device according to item 1. 18. A stripe-shaped first opening and a second opening connected to the first opening and wider than the first opening on a semiconductor substrate containing impurities. Forming a mask having the same, and growing a semiconductor layer in the first and second openings, so that the maximum thickness of the first opening is twice the minimum thickness of the second opening. A method for manufacturing an optical semiconductor device, comprising: forming a quantum well structure layer as described above; and forming an upper cladding layer on the quantum well structure layer. 19. The quantum well structure layer has a multiple quantum well structure having a plurality of well layers and a plurality of barrier layers sandwiching the well layer, or a single quantum well structure having a well layer and a barrier layer sandwiching the well layer. 19. The method according to claim 18, wherein the optical semiconductor device has a single quantum well structure.