JP2780829B2 - Method for manufacturing optical semiconductor integrated device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] The present invention relates to a method of manufacturing a direct coupling (bat joint) type optical semiconductor integrated device in which waveguides of two or more optical semiconductor elements are abutted and matched with each other, and relates to a direct coupling type optical semiconductor integrated device. In a method of manufacturing a device, there is provided a method of manufacturing an optical semiconductor device capable of reducing a change in thickness at a coupling portion while securing a desired thickness of an optical waveguide layer of a plurality of optical semiconductor elements. A method of manufacturing an optical semiconductor integrated device in which first and second optical semiconductor elements are integrated on a same semiconductor substrate by directly arranging waveguide portions thereof, wherein the first waveguide is A first semiconductor layer to be formed is selectively formed, and a first semiconductor layer is formed thereon.
And a second mask formed along the region where the second waveguide is to be formed, separated from the region where the first waveguide is to be formed by a predetermined distance or more, and the region where the first waveguide is to be formed. And selectively growing a second semiconductor layer on which a second waveguide is to be formed using the regions of the first and second masks.

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for manufacturing an optical semiconductor integrated device, and more particularly to a direct coupling (butt joint) type optical semiconductor integrated device in which waveguides of two or more optical semiconductor elements are aligned by abutting each other. It relates to a manufacturing method.

[Prior Art] Conventionally, when two or more optical semiconductor elements are directly coupled with each other by abutting respective waveguides, one optical semiconductor element is formed to some extent, covered with a mask, and the other optical semiconductor element is covered with a mask. Unnecessary substances on a region where a semiconductor element is to be formed are removed by etching, and the other optical semiconductor element is formed adjacent to one optical semiconductor element. This pattern will be described with reference to FIG.

In FIG. 2A, after forming a first waveguide layer 54 on which one optical semiconductor element is to be formed and a protective layer 55a thereon on a base crystal 50, a mask 5 made of, for example, SiO ₂ or the like is formed.
1 is formed, and the protective layer 55 in the exposed portion is
a, The first waveguide layer 54 was removed by etching, and the second waveguide layer 56 for forming the second optical semiconductor element and the protective layer 55b were formed on the exposed base crystal 51 this time.

However, when the second optical semiconductor device is formed adjacent to the first optical semiconductor device, the growth thickness of the second waveguide layer 56 becomes large near the mask 51, and the thickness of the growth layer becomes uneven. A phenomenon occurs. In other words, as shown in the figure, the growth film thickness suddenly increases in the immediate vicinity of the joint portion, gradually decreases and forms a bulge portion, and becomes almost uniform at a position separated by a certain degree or more. In the protective layer 55b on the second waveguide layer 56, the non-uniform thickness distribution tends to relax, but the change in thickness does not immediately disappear. For this reason, after removing the mask 51, the cladding layer is grown again,
I try to get a flat plane.

In FIG. 2A, the growth was performed so that the waveguide layer 54 of the first optical semiconductor element and the uniform portion of the second waveguide layer 56 of the second optical semiconductor element had substantially the same thickness. Although a case has been shown, in this case, a large step occurs at the coupling portion as illustrated. Such a step not only hinders the manufacturing process but also causes light scattering and the like, which is not preferable in terms of the characteristics of the optical semiconductor element.

At the coupling portion, it is possible to control so that the waveguide layers of the two optical semiconductor elements have substantially the same thickness. This is shown in FIG. 2 (B). Second waveguide layer of second optical semiconductor device
When the thickness of the second waveguide layer 56 is controlled to be small to have substantially the same thickness at the joint with the first waveguide layer 54, the overall thickness of the second waveguide layer 56 is reduced. For this reason, the thickness of the second waveguide layer 56 becomes thinner as the distance from the coupling portion increases, and it becomes too thin in a uniform portion, so that desired characteristics cannot be exhibited. For this reason, a phenomenon in which light escapes in the thin uniform portion occurs, and the desired characteristics of the optical semiconductor element cannot be realized.

As described above, it is possible to eliminate the difference in thickness at the joint, but in this case, the characteristics of the optical semiconductor element itself cannot be guaranteed.

The thickness of the thick part close to the mask and the uniform part having a uniform thickness away from the mask are, for example, 3 to 4
Times different. Therefore, when trying to obtain a desired thickness in a uniform portion, a portion 3 to 4 times thicker than the desired thickness is generated from the joining portion to about 20 to 30 μm in the vicinity thereof, and the joining portion has A large step occurs.

[Problems to be Solved by the Invention] As described above, when an optical semiconductor integrated device of a direct coupling type is to be manufactured by the conventional technique, the thickness of the optical waveguide layer is reduced at the coupling portion between two optical semiconductor elements. A large step occurs.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a method for manufacturing a direct-coupling type optical semiconductor integrated device, in which a desired thickness of an optical waveguide layer of a plurality of optical semiconductor elements is secured while reducing a change in thickness at a coupling portion. It is an object of the present invention to provide a method for manufacturing an optical semiconductor device which can be performed.

[Means for Solving the Problems] FIG. 1 is an explanatory view of the principle of the present invention.

FIG. 1A schematically shows the structure of an optical semiconductor integrated device to be manufactured. The first waveguide 2a and the second waveguide 3a abut on the semiconductor substrate 1 in a straight line to form a direct coupling. Light in the first waveguide 2a travels along the first waveguide and enters the second waveguide 3a. These waveguides 2
a and 3a can be formed by removing unnecessary portions other than the striped region from the first semiconductor layer 2 formed on the half plane and the second semiconductor layer formed on the remaining half plane.

A method for manufacturing an optical semiconductor integrated device as shown in FIG. 1 (A) will be described with reference to FIGS. 1 (B) to 1 (E).

In FIG. 1B, a first semiconductor layer 2 on which a first waveguide 2a is to be formed is selectively formed on a semiconductor substrate 1, and a first semiconductor layer 2 having a function of suppressing crystal growth thereon is formed. A mask 4 is prepared. Further, a second mask 5 is prepared along the region where the second waveguide 3a is to be formed. First
The second mask 5 is cut off in the vicinity of the region to be the waveguide 2a.

Using the first and second masks 4 and 5, crystal growth is performed on the semiconductor substrate 1 to grow the second semiconductor layer 3.

FIG. 1C is a plan view for explaining the arrangement of the first and second masks 4 and 5. Under the first mask 4, the first semiconductor layer 2 on which the first waveguide 2a is to be formed has already been formed.
Are formed, and the first mask 4 is formed thereon. It is planned to create a second waveguide 3a on the extension of the first waveguide. The second mask 5 is formed along the region where the second waveguide 3a is to be formed, and the region where the second waveguide 3a (the first waveguide 2a) is to be formed near the first mask 4. From a predetermined distance or more.

By using such masks 4 and 5, crystal growth of the second waveguide layer 3 is performed on the semiconductor substrate 1 as shown in FIG. The thickness of the second semiconductor layer 3 is substantially equal to the thickness of the first semiconductor layer 2.

[Operation] In the conventional technology, crystal growth is performed using only a mask corresponding to the first mask 4. At that time, the thickness of the semiconductor layer grown near the joint increased, and the thickness became uneven as a whole.

When the crystal is grown by the above method, the second semiconductor layer 3 having a substantially uniform thickness can be grown as shown in FIG. 1 (D). The cause is considered as follows.

FIG. 1E conceptually shows the operation of the first and second masks. Since the first mask 4 and the second mask 5 have a function of suppressing crystal growth, no crystal growth occurs on the source material even when the source material reaches the mask surface. On the other hand, in the portion 6 where the semiconductor substrate 1 is exposed, the source material reaching the surface is crystallized and solidified. Therefore, the source material is continuously transported to the exposed portion 6 of the semiconductor substrate 1 from above. On the other hand, the source material directed to the first mask 4 and the second mask 5 cannot be deposited on the surface and diffuses in the lateral direction. This is indicated by the arrow in FIG. That is, the source material diffused from above the masks 4 and 5 is
When it reaches the top, crystal growth occurs there. Therefore, crystal growth becomes active on the exposed surface of the semiconductor substrate 1 near the ends of the masks 4 and 5.

Conventionally, only the first mask 4 exists, so the mask 4
The growth layer became thick only in the vicinity. At a location away from the first mask 4, the growth layer becomes relatively thin. Therefore, in this thinned region, by forming the second mask 5 along the region where the second waveguide 3a is to be formed, both masks 4 are formed over the entire length of the region where the second waveguide 3a is to be formed. If the source material is supplied uniformly from 5 and above, the crystal growth on the region where the second waveguide 3a is to be formed becomes uniform.

In the figure, in a portion further away from the second mask 5, the influence of the second mask 5 disappears, and a film thickness distribution appears. However, since such a portion is not used as an optical semiconductor integrated device, the influence is small. Absent.

After the second semiconductor layer 3 is formed using the first mask 4 and the second mask 5 in this manner, a mask is formed on the first waveguide 2a and the second waveguide 3a. And
By performing etching in a stripe shape, it is possible to obtain an optical semiconductor integrated device structure as shown in FIG.

Example FIGS. 3A and 3B show an example of an optical semiconductor integrated device to be manufactured according to the present invention. FIG. 3 (A) shows a longitudinal sectional view, and FIG. 3 (B) shows a perspective view.

In the figure, a DFB laser 20 and an optical modulator 21 are integrated and manufactured on an n-type InP substrate 11. A diffraction grating 12 is formed on the left surface of the n-type InP substrate 11. A waveguide layer (guide layer) 13 made of, for example, InGaAsP is formed on a substrate 11, and an etching stop layer 17 made of, for example, InP for stopping selective etching is formed thereon. On this etching stop layer 17, an active layer 14 and an absorption layer 16 made of, for example, InGaAsP having different compositions are formed. For these active layers and absorption layers, first, after growing the active layer over the entire surface, unnecessary portions are removed by etching,
It is created by selectively growing an absorbing layer on the rest. In the case of the figure, the active layer 14 of the DFB laser 20 having a relatively long photoluminescence (PL) emission wavelength is formed first, and after the right portion is removed by etching, the PL is removed.
The absorption layer 16 of the optical modulator 21 having a relatively short wavelength is formed. On the active layer and the absorption layer, a cladding layer 15 made of, for example, p-type InP is formed, and a part thereof is converted into, for example, a high-resistance separation layer 18 doped with Fe. A contact layer 22 made of p ⁺ -type InGaAsP is formed on the cladding layer 15, and p-side electrodes 23a and 23b are formed thereon. Note that an n-side substrate 24 is formed on the InP substrate 11. On the end face of the optical modulator 21, an antireflection (AR) film 25 made of, for example, silicon nitride (SiN) is formed.

In FIG. 3 (A), if the thicknesses of the first active layer 14 and the second absorption layer 16 are different, a step will occur at the joint. The steps not only hinder the subsequent manufacturing process, but also have undesirable effects such as optically causing light scattering. Therefore, it is better that the step is as small as possible.

4 (A) to 4 (H) are views for explaining a method of manufacturing an optical semiconductor integrated device as shown in FIGS. 3 (A) and 3 (B).

In FIG. 4 (A), a diffraction grating 12 is selectively formed on the surface of an n-type InP substrate 11 by an optical interference method.
A guide layer 13 made of InGaAsP having a PL wavelength of 1.1 μm and having a thickness of about 0.15 μm is grown by a liquid phase growth method. Thickness on this
An etching stop layer 17 composed of an InP layer of 200 to 300 ° is grown, and a non-doped InGaAs having a PL wavelength of 1.55 μm is formed thereon.
First active layer 1 to be a light emitting layer of about 0.15 μm in thickness composed of P
4. A protective layer 15a of P-type InP is formed. Protective layer 15a
Plays a role of protecting the first active layer 14 and constitutes a part of the cladding layer 15 later. On the protective layer 15a, SiO
_A mask 31 made of _{2 is formed} to a thickness of about 2500 mm by CVD. The mask 31 has a function of covering the DFB laser portion and preventing it from being etched. Further, the mask 31 extends so as to sandwich the optical waveguide portion of the optical modulator as described later.

FIG. 4B shows a selective etching step. mask
31 and the underlying protective layer 15a and the first active layer 14
Is selectively etched. Etching stops on the underlying etching stop layer 17.

Thereafter, as shown in FIG. 4C, using this mask 31 as it is as a crystal growth mask, crystal growth is selectively performed on the exposed InP etching stop layer 17.
Crystal growth is performed by, for example, liquid crystal growth. On the etching stop layer 17, an absorption layer 16 serving as a light absorption layer made of non-doped InGaAsP having a PL wavelength of 1.40 μm is grown to a thickness equivalent to that of the first active layer 14, and a p-type InP protective layer 15b is formed thereon. Is grown to a thickness of 0.3 μm or more.

The reason why the absorption layer 16 and the protective layer 15b thereon can be grown with a uniform thickness is due to the shape of the mask 31.

FIG. 4D shows an example of a planar shape of the SiO ₂ mask 31. In the figure, the mask 31 extends not only on the first mask 4 on the first waveguide 3a side but also on both sides of the portion of the second waveguide 3b to form the second mask 5. That is, the second waveguide 3b is formed so as to form a rice scoop-shaped opening having a flattened tip.
It has a shape sandwiching. In addition, the mask is removed in a portion 32 that is more than a predetermined distance from the second waveguide 3b.

For example, in the case of liquid crystal growth, the source material that has moved downward in the melt grows there when it reaches the crystal surface, but cannot grow when it reaches the mask. It moves to the surroundings while maintaining the degree of supersaturation. Then, the degree of supersaturation moves from above the first and second masks 4 and 5 to the periphery thereof. If this supersaturated flow is made uniform on the second waveguide 3b, the finally obtained second waveguide 3b
Can have a uniform thickness. Therefore, the second mask 5 may be disposed along the second waveguide 3b at a portion separated from the first waveguide 3a by a certain distance or more. Since there is a degree of supersaturation supplied from the first mask 4 in a portion near the first waveguide 3a, if the degree of supersaturation is further multiplied, the degree of supersaturation becomes too high, which also causes nonuniform film thickness. . Therefore, in the vicinity of the first mask 4 on the first waveguide 3a, the second mask 5 on the second waveguide 3b side is arranged so as to be separated from the waveguide portion by a certain degree or more. For example, the flow of supersaturation supplied from above the mask extends up to a width of about 20 μm, and particularly has an effect up to a width of about 10 μm. There is almost no effect if the distance is 30 μm or more.
Thus, for example, the width of the region where the second waveguide 3b is to be formed is set to, for example, about 10 μm, and the second mask 5 having a width of 20 μm or more is disposed on both sides thereof. Also, the second mask 5 is kept away from the first mask 4 on the first waveguide 3a up to about 20 μm in width, for example, the width of the widest part in the drawing is about 100 μm. I do. The width between the masks in the portion of the second waveguide 3b is set to about 10 μm in order to finally obtain a waveguide having a uniform thickness of about 1.5 μm, for example. If the width is too small, the shape may be raised at the end.

FIG. 4E shows an example of the film thickness distribution when the crystal growth with the degree of supersaturation supplied from the mask portion being adjusted as described above. First semiconductor layer forming first waveguide
14 and the second semiconductor layer 16 forming the second waveguide have a substantially uniform thickness distribution.

Thereafter, the mask 31 is removed, and as shown in FIG. 4 (F), p-type InP is grown thereon to form the cladding layer 15. On top of this, a P ⁺ type InGaAsP
A cap layer 22 made of is grown. After stacking is performed in this manner, as shown in FIG. 4 (G), a mask 35 composed of a SiO ₂ layer having a width of about 3 μm and a thickness of about 2500 ° is formed in a vertical direction, and mesa etching is performed. Do. The mesa etching is performed until the substrate 11 is reached.

Next, as shown in FIG. 4H, a high-resistance InP (Fe-doped) layer 19 is grown around the mesa portion by MOCVD. Further, by forming a p-side electrode made of a Ti / Pt stack and an n-side electrode made of an Au-Ge alloy, and forming an AR film made of SiN on the emission surface, the structure shown in FIGS. An optical semiconductor integrated device as shown is prepared.

The method for manufacturing an optical semiconductor integrated device has been described in accordance with the above embodiments. Here, a mask used for crystal growth will be further described.

FIGS. 5A to 5E show various mask shapes. FIG. 5A shows a mask 31a having a substantially T-shaped opening. The upper part of the T-shape constitutes the first mask 4, and the part on the right side in the figure constitutes the second mask 5. The portion along the T-shaped vertical bar is arranged along the waveguide to be created. In the vicinity of the first mask 4, the second masks 5 are separated from each other by a sufficient distance until the supply of the degree of supersaturation from the second mask 5 side becomes substantially zero. T
At the connection between the upper bar and the vertical bar, the second mask 5
Is designed so that the sum of the supersaturation supplied by the first mask 4 and the supersaturation supplied by the first mask 4 becomes substantially constant.

FIG. 5B shows another shape 31b of the mask. First
The first mask 4 disposed on the waveguide has a substantially semi-planar shape, and the second mask 5 formed across the second waveguide is separated from the first mask 4 by a predetermined distance. Released, second
Are arranged on both sides in parallel with the waveguide. The reason why the second mask 5 is formed to have a width in consideration of the fact that it is difficult to control the crystal growth when the area for crystal growth is too small may be considered. For example, if the area of the base crystal is 100 and the portion 99 is covered with the mask, it may be difficult to precisely control the crystal growth performed on the remaining portion.

FIG. 5 (C) shows still another embodiment. In FIG. 5C, the width of the first mask 4 is also limited. Since only the first waveguide portion having a short width is ultimately used, it is meaningless if crystals outside the first waveguide portion are present or not. On the other hand, if the area for crystal growth is too small, it may be difficult to control the crystal growth. Therefore, the width of the first mask 4 is also limited, and crystal growth is performed on both sides. Things.

FIG. 5 (D) shows a further modification of the shape of the second mask at a portion where the effects of the first mask 4 and the second mask 5 are added. The degree of supersaturation supplied from the first mask 4 decreases as the distance from the first mask 4 increases. In accordance with this, the distance between the pair of first masks 5 is gradually reduced so that the supersaturation degree supplied from the second mask 5 gradually increases, and the first masks 5 are arranged so as to extend in parallel.

The supersaturated flow, which suppresses crystal growth, goes from above the mask to the exposed surface without the mask, but is not very sensitive to the fine features of the mask. As shown in FIG. 5 (E), even if a notch-shaped portion 41 exists in the second mask 5, a substantially uniform degree of supersaturation is supplied in a region sandwiched between the pair of second masks 5. It is possible. It is also possible to control the degree of supersaturation supplied by controlling the width of the mask itself, like the tip 42 of the second mask 5.

Although the present invention has been described with reference to the embodiments, the present invention is not limited thereto. For example, it will be apparent to those skilled in the art that various modifications, improvements, combinations, and the like can be made.

[Effects of the Invention] As described above, according to the present invention, after a first semiconductor layer is selectively formed, a mask is prepared thereon, and a second semiconductor layer is selectively formed on the remaining portion. In the case of the growth, the step at the joint portion can be reduced.

The step at the coupling portion can be reduced, and the thickness distribution of the second waveguide formed from the second semiconductor layer can be made substantially uniform.

Therefore, an optical semiconductor integrated device having excellent optical coupling efficiency can be manufactured.

[Brief description of the drawings]

1 (A) to 1 (E) are explanatory views of the principle of the present invention. FIG. 1 (A) is a schematic perspective view of an optical semiconductor integrated device, and FIG. 1 (B) is a perspective view of a structure in which a mask is formed. Fig. 1 (C)
Is a plan view of the mask, FIG. 1 (D) is a schematic cross-sectional view of the structure in which the crystal has been grown, FIG. 1 (E) is a conceptual diagram for explaining the operation of the mask, FIG. 2B shows a conventional technique, FIG. 2A is a sectional view showing a case where the growth amount of the second semiconductor layer is large, and FIG. 2B is a sectional view showing a case where the growth amount of the second semiconductor layer is small. 3A and 3B are a sectional view and a perspective view showing an example of an optical semiconductor integrated device manufactured according to the embodiment of the present invention, and FIGS. 4A to 4H are embodiments of the present invention. FIG. 4A is a view for explaining a method of manufacturing an optical semiconductor integrated device according to FIG.
4C are cross-sectional views for explaining various manufacturing steps, FIG. 4D is a plan view of a mask, and FIG. 4E is a schematic diagram showing a film thickness distribution of a growth layer that has been subjected to selective growth. Sectional view shown in FIG.
FIG. 4 (F) is a cross-sectional view of the substrate after the epitaxial growth, FIG. 4 (G) is a cross-sectional view showing a mesa etching step, and FIG.
FIG. 5H is a sectional view showing a burying growth step, and FIGS. 5A to 5E are schematic plan views showing various mask shapes. In the figure, 1 ... semiconductor substrate 2 ... first semiconductor layer 2a ... first waveguide 3 ... second semiconductor layer 3a ... second waveguide 4, 5 ... mask 6 ... semiconductor Exposed portion of substrate 11 n-type InP substrate 12 diffraction grating 13 waveguide layer 14 active layer 15 clad layer 16 absorption layer (light absorption layer) 17 etching stop layer 18 … High-resistance separation layer 19… High-resistance InP buried layer 20… DFB laser 21… Optical modulator 23… Plus electrode 24… Negative electrode 25… AR film 31… SiO ₂ mask 32… Notch 35 ... Stripe-shaped etch mask 41 ... Notch 42 ... Tip

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int.Cl. ⁶ , DB name) H01S 3/18 G02F 1/025

Claims (3)

(57) [Claims] A first and a second semiconductor substrate on the same semiconductor substrate;
A method for manufacturing an optical semiconductor integrated device in which the optical semiconductor elements are integrated by directly arranging their waveguide portions (2a, 3a), wherein a first waveguide (2a) to be formed is formed. Semiconductor layer (2)
Is selectively formed, and is separated from the region of the first mask (4) covering the first waveguide (2a) by a predetermined distance or more from the region where the first waveguide (2a) is to be formed. And a region of a second mask (5) formed along a region where a second waveguide (3a) is to be formed, and a second mask is formed by using the regions of the first and second masks. A method for manufacturing an optical semiconductor integrated device, comprising a step of selectively growing a second semiconductor layer (3) on which a waveguide (3a) is to be formed. 2. The method according to claim 1, wherein the second mask is formed on both sides of a region where a second waveguide is to be formed. 3. A mask region having an opening width larger than the opening width of the region of the second mask (5) is provided between the region of the first mask (4) and the region of the second mask (5). 3. The method for manufacturing an optical semiconductor integrated device according to claim 2, wherein the method is provided.