JP2012059963A - Method of manufacturing semiconductor optical integrated element - Google Patents

Abstract

A method of manufacturing a semiconductor optical integrated device capable of widening the control range of light emission wavelength is provided.
A plurality of semiconductor layers constituting a quantum well structure including a well layer and a barrier layer on a semiconductor substrate containing sulfur, wherein the semiconductor layers have different thicknesses from each other. A semiconductor layer forming step to be formed; an optical element forming step of forming a plurality of optical elements each including the plurality of semiconductor layers; and a heat treatment step of performing a heat treatment for shifting emission wavelengths of the plurality of semiconductor layers. Including. Preferably, the heat treatment step is performed in a range of 600 ° C to 800 ° C.
[Selection] Figure 1

Description

  The present invention relates to a method for manufacturing a semiconductor optical integrated device including a plurality of optical devices having a semiconductor layer on a semiconductor substrate.

  With increasing demand for information communication, the capacity of high-density wavelength division multiplexing (DWDM) communication systems is increasing. As a light source used in a large-capacity communication system, an optical integrated device in which a plurality of semiconductor lasers having different laser oscillation wavelengths are formed on the same substrate has been studied for the purpose of reducing cost and size. As a technique for realizing semiconductor lasers with different laser oscillation wavelengths on the same substrate, there is a method using selective growth (for example, see Non-Patent Document 1). In this method, dielectric crystal masks having different widths are formed on both sides of a region where a semiconductor laser is to be formed, and semiconductor crystal growth is performed, whereby each region has a different thickness depending on the width of the mask. This is a method of growing active layers having different thicknesses and making the emission wavelengths of the respective active layers different. Here, it is assumed that the active layer has a single or multiple quantum well structure including a well layer and a barrier layer. According to this selective growth, the thickness of the well layer can be increased by increasing the mask width. When the well layer is thickened, the quantum level formed in the well layer shifts to the low energy side in the conduction band and shifts to the high energy side in the valence band, so the emission wavelength of the active layer is shifted to the long wavelength side. Can be shifted. Similarly, the emission wavelength of the active layer can be shifted to the short wavelength side by narrowing the mask width.

  In the case of an active layer having a quantum well structure, a phenomenon is known in which element mutual diffusion occurs at the interface between the well layer and the barrier layer due to heat treatment, and the emission wavelength shifts to the short wavelength side (for example, non-patent literature). 2).

T. Sasaki, M. Yamaguchi, K. Komatsu, and I. Mito, IEICE Trans. Electron., E80-C, pp.654-663 (1997) S. D. Perrion, et al., J. Elect. Mat. 23 (1994) 81.

  When the selective growth described above is used, the emission wavelength can be controlled by adjusting the thickness of the active layer by adjusting the mask width. However, in the case of selective growth, there is a problem that the control range of the emission wavelength is limited by the adjustment range of the mask width and the adjustment range of the thickness of the well layer. In particular, from the standpoint of crystal quality and emission intensity, the lower limit of the preferred well layer thickness is limited, so that there is a problem that the shift amount of the emission wavelength to the short wavelength side is limited.

  The present invention has been made in view of the above, and an object of the present invention is to provide a method of manufacturing a semiconductor optical integrated device capable of widening the control range of the emission wavelength.

  In order to solve the above-described problems and achieve the object, a method of manufacturing a semiconductor optical integrated device according to the present invention includes a plurality of quantum well structures including a well layer and a barrier layer on a semiconductor substrate containing sulfur. A semiconductor layer forming step of forming a plurality of semiconductor layers having different thicknesses of the well layer or the barrier layer, and an optical element forming step of forming a plurality of optical elements each including the plurality of semiconductor layers And a heat treatment step for performing a heat treatment for shifting the emission wavelengths of the plurality of semiconductor layers.

  In the method for manufacturing a semiconductor optical integrated device according to the present invention as set forth in the invention described above, the heat treatment step is performed in a range of 600 ° C. to 800 ° C.

  In the method of manufacturing a semiconductor optical integrated device according to the present invention as set forth in the invention described above, the semiconductor layer forming step forms the well layer or the barrier layer in a thickness of 1 nm to 10 nm.

  The semiconductor optical integrated device manufacturing method according to the present invention is characterized in that, in the above invention, the semiconductor layer forming step grows the plurality of semiconductor layers using a selective growth method.

In the method of manufacturing a semiconductor optical integrated device according to the present invention, the semiconductor substrate includes sulfur having a carrier concentration of 1 × 10 ¹⁸ cm ⁻³ to 1 × 10 ¹⁹ cm ^{−3 in} the above invention. And

  The semiconductor optical integrated device manufacturing method according to the present invention is characterized in that, in the above invention, the heat treatment step is performed by forming a temperature distribution in the surface of the semiconductor substrate.

  According to the present invention, it is possible to realize a method for manufacturing a semiconductor optical integrated device capable of widening the control range of the emission wavelength.

FIG. 1 is a schematic configuration diagram of an integrated semiconductor laser device according to an embodiment. FIG. 2 is a partial cross-sectional view taken along line AA of the integrated semiconductor laser device shown in FIG. FIG. 3 is a cross-sectional view showing the structure of the active layer shown in FIG. FIG. 4 is a cross-sectional view showing the structure of the active layer shown in FIG. 5 is a partial cross-sectional view of the integrated semiconductor laser device shown in FIG. 6 is a partial cross-sectional view taken along line CC of the integrated semiconductor laser device shown in FIG. FIG. 7 is a view (a cross section taken along the line DD shown in FIG. 1) for explaining a step of growing a buffer layer in the method of manufacturing the integrated semiconductor laser device shown in FIG. FIG. 8 is a diagram illustrating a process of forming a mask in the method of manufacturing the integrated semiconductor laser device shown in FIG. FIG. 9 is an enlarged view showing the shape of the mask shown in FIG. FIG. 10 is a diagram illustrating a process of growing a plurality of semiconductor layers including an active layer in the method of manufacturing the integrated semiconductor laser device shown in FIG. FIG. 11 is a cross-sectional view showing the structure of the active layer grown by the process of FIG. FIG. 12 is a cross-sectional view showing the structure of the active layer grown by the process of FIG. FIG. 13 is a diagram illustrating an example of the relationship between the mask width and the thickness of the well layer. FIG. 14 is a diagram illustrating an example of the relationship between the mask width, the emission wavelength, and the energy gap. FIG. 15 is a diagram for explaining a step of forming a diffraction grating in the method of manufacturing the integrated semiconductor laser device shown in FIG. FIG. 16 is a diagram illustrating a process of growing a core layer and an upper clad layer in the method of manufacturing the integrated semiconductor laser device shown in FIG. FIG. 17 is a plan view showing a state after performing the mesa structure forming step in the method of manufacturing the integrated semiconductor laser device shown in FIG. FIG. 18 is a view showing a cross section in a region where a semiconductor optical amplifier is formed in the manufacturing method of the integrated semiconductor laser device shown in FIG. FIG. 19 is a diagram showing an example of the relationship between the thickness of the well layer and the shift amount of energy corresponding to the shift of the emission wavelength. FIG. 20 is a diagram illustrating an example of the relationship between the heat treatment temperature and the shift amount of the emission wavelength. FIG. 21 is a diagram illustrating an example of the relationship between the heat treatment time and the shift amount of the emission wavelength. FIG. 22 is a diagram illustrating an example of the relationship between the sulfur carrier concentration of the substrate and the shift amount of the emission wavelength. FIG. 23 is a partial cross-sectional view illustrating an example of a method for forming a temperature distribution on a substrate. FIG. 24 is a partial cross-sectional view illustrating an example of a method for forming a temperature distribution on a substrate.

  According to the present invention, the light emission wavelength of a semiconductor layer having a quantum well structure grown on a substrate containing sulfur is shifted by the heat treatment according to the thickness of the semiconductor layer and the sulfur carrier concentration of the substrate. It has been completed by finding that it shifts only and conceiving that this is applied to control of the emission wavelength of the semiconductor layer.

  Embodiments of a method for manufacturing a semiconductor optical integrated device according to the present invention will be described below in detail with reference to the drawings. Note that the present invention is not limited to the embodiments. Moreover, in each drawing, the same code | symbol is attached | subjected suitably to the same or corresponding component. Further, the drawings are schematic, and it should be noted that the relationship between the thickness and width of each layer, the ratio of each layer, and the like may differ from the actual situation. Also in the drawings, there are included portions having different dimensional relationships and ratios.

(Embodiment)
An integrated semiconductor laser element that is a semiconductor optical integrated element according to an embodiment of the present invention will be described. FIG. 1 is a schematic configuration diagram of an integrated semiconductor laser device according to the present embodiment.

  As shown in FIG. 1, the integrated semiconductor laser device 100 according to the first embodiment includes a plurality of DFB (Distributed Feed-Back) laser stripes 11-1 to 11-n (n is an integer of 2 or more), It has a configuration in which a plurality of optical waveguides 12-1 to 12-n that are passive optical waveguides, an optical combiner 13 that is a passive optical waveguide, and a semiconductor optical amplifier 14 are integrated on one semiconductor substrate. Note that the periphery of these optical elements is embedded by the embedded portion 15. Further, trench grooves 17-1 to 17-m (m = n-1) are formed in the buried portion 15 between the DFB laser stripes 11-1 to 11-n.

  The DFB laser stripes 11-1 to 11-n are edge-emitting lasers having a stripe-shaped embedded structure, and are formed at one end in the length direction of the integrated semiconductor laser element 100 at a predetermined pitch in the width direction. Yes. The DFB laser stripes 11-1 to 11-n are configured such that the laser oscillation wavelengths of the output laser beams are different from each other within a predetermined wavelength range, for example, a range of 1158 nm to 1326 nm. The laser oscillation wavelengths of the DFB laser stripes 11-1 to 11-n can be finely adjusted by changing the set temperature of the integrated semiconductor laser element 100.

  The optical combiner 13 is of an MMI (Multi Mode Interferometer) type, for example, and is formed near the center of the integrated semiconductor laser device 100. The optical waveguides 12-1 to 12-n are formed between the DFB laser stripes 11-1 to 11-n and the optical combiner 13, and the DFB laser stripes 11-1 to 11-n and the optical combiner 13 13 is connected. The semiconductor optical amplifier 14 is formed to be connected to the optical output port 13a of the optical combiner 13, and has an optical output end 14a.

  Next, a cross-sectional structure of the integrated semiconductor laser device 100 will be described. Below, a cross-sectional structure will be described in the order of the DFB laser stripes 11-1 to 11-n, the optical combiner 13, and the semiconductor optical amplifier 14.

  First, the cross-sectional structure of the DFB laser stripes 11-1 to 11-n will be described. 2 is a partial cross-sectional view taken along line AA of the integrated semiconductor laser device 100 shown in FIG. As shown in FIG. 2, the DFB laser stripe 11-1 has a lower cladding on a substrate 22 made of n-type InP containing sulfur (S), which is an n-type dopant, with an n-side electrode 21 formed on the back surface. Also serving as an n-type buffer layer 23 made of InP, a lower SCH (Separate Confinement Heterostructure) layer 24 made of InGaAsP whose composition is continuously changed, and an active layer 25-1 having a multi-quantum well (MQW) structure The upper SCH layer 26 made of InGaAsP with a continuously changing composition, the spacer layer 27 made of InP, the grating layer 28-1 made of InGaAsP, the upper cladding layers 29 and 30 made of p-type InP, and the p-type InGaAs The contact layers 31 are sequentially stacked. In the grating layer 28-1, a diffraction grating set to a period corresponding to the wavelength of the laser beam to be output from the DFB laser stripe 11-1 is formed.

  The upper part of the buffer layer 23 to the upper cladding layer 29 has a mesa structure, and both sides of the mesa structure are a lower current blocking layer 32a made of p-type InP and an upper current blocking layer 32b made of n-type InP. Embedded in a current blocking layer 32 composed of

  The surface of the semiconductor layer on which the DFB laser stripe 11-1 is stacked is covered with a protective film 33 made of SiN, and the p-side electrode is in ohmic contact with the contact layer 31 through the opening 33a of the protective film 33. 34 is formed.

  The DFB laser stripe 11-2 electrically separated by the DFB laser stripe 11-1 and the trench groove 17-1 has the same cross-sectional structure as the DFB laser stripe 11-1, but the active layer 25-1 and The grating layer 28-1 is replaced with an active layer 25-2 and a grating layer 28-2 different from these. The grating layer 28-2 is formed with a diffraction grating set at a period corresponding to the wavelength of the laser beam to be output from the DFB laser stripe 11-2. The other DFB laser stripes 11-3 to 11-n have the same cross-sectional structure as the DFB laser stripe 11-1, but the active layer 25-1 and the grating layer 28-1 are different from each other in the active layer and the grating layer. It has the structure replaced by.

  Next, the structure of the active layer will be described. FIG. 3 is a sectional view showing the structure of the active layer 25-1 shown in FIG. FIG. 4 is a cross-sectional view showing the structure of the active layer 25-2 shown in FIG. As shown in FIG. 3, the active layer 25-1 has a configuration in which well layers 25a-1 and barrier layers 25b-1 are alternately stacked. As shown in FIG. 4, the active layer 25-2 has a configuration in which well layers 25a-2 and barrier layers 25b-2 are alternately stacked. The number of well layers is not limited to 3 as shown in FIGS. 3 and 4 and can be 2 or more.

  The well layers 25a-1 and 25a-2 have the same composition but different thicknesses. Further, the barrier layers 25b-1 and 25b-2 have the same composition but different thicknesses. Specifically, the well layer 25a-2 and the barrier layer 25b-2 are thicker. Here, as described above, the emission wavelength of the active layer shifts to a shorter wavelength as the well layer becomes thinner, but the emission wavelengths of the active layers 25-1 and 25-2 are the well layers 25a-1 and 25a. −2 with respect to the wavelength shifted depending on the thickness of −2 by a shift amount corresponding to the thickness of the well layers 25a-1 and 25a-2 and the sulfur carrier concentration of the substrate 22, It has become. This will be described in detail later when a method for manufacturing the integrated semiconductor laser device 100 according to the present embodiment is described.

  Similarly, each active layer of the DFB laser stripes 11-1 to 11-n has the same composition of the well layer and the composition of the barrier layer, but the thickness of the well layer and the thickness of the barrier layer are as follows. The thickness increases from the DFB laser stripe 11-1 toward the DFB laser stripe 11-n. Therefore, the emission wavelengths of the active layers of the DFB laser stripes 11-1 to 11-n are also longer from the DFB laser stripe 11-1 toward the DFB laser stripe 11-n.

  The emission wavelength of each active layer of the DFB laser stripes 11-1 to 11-n is close to the selected wavelength (that is, the desired laser oscillation wavelength) of the diffraction grating formed on the grating layer located above each active layer. Preferably, they are set to coincide. As a result, the DFB laser stripes 11-1 to 11-n can output laser light of each desired wavelength with a low threshold current and high power efficiency.

  Next, a cross-sectional structure of the optical combiner 13 will be described. FIG. 5 is a partial cross-sectional view taken along line BB of the integrated semiconductor laser device 100 shown in FIG. As shown in FIG. 5, the optical combiner 13 is composed of a buffer layer 23 and InGaAsP on a substrate 22 having an n-side electrode 21 formed on the back surface, and laser light output from the DFB laser stripes 11-1 to 11-n. A transparent core layer 41, an upper clad layer 42 made of i-type InP, an upper clad layer 30, and a contact layer 31 are sequentially laminated. The upper part of the buffer layer 23 to the upper cladding layer 42 has a mesa structure, and both sides of the mesa structure are embedded by a current blocking layer 32 including a lower current blocking layer 32a and an upper current blocking layer 32b. Yes. The surface of the contact layer 31 is covered with a protective film 33.

  The optical waveguides 12-1 to 12-n have the same cross-sectional structure as the optical combiner 13, but the core layer width (mesa width) is output by the DFB laser stripes 11-1 to 11-n. The difference is that the laser beam is set to be guided in a single mode.

  Next, a cross-sectional structure of the semiconductor optical amplifier 14 will be described. FIG. 6 is a partial cross-sectional view of the integrated semiconductor laser device 100 shown in FIG. As shown in FIG. 6, the semiconductor optical amplifier 14 includes a buffer layer 23, a lower SCH layer 24, an active layer 43 having an MQW structure, an upper SCH layer 26, a spacer layer on a substrate 22 having an n-side electrode 21 formed on the back surface. 27, upper cladding layers 44 and 30 made of p-type InP and a contact layer 31 are sequentially stacked. The upper part of the buffer layer 23 to the upper cladding layer 44 has a mesa structure, and both sides of the mesa structure are embedded by a current blocking layer 32 composed of a lower current blocking layer 32a and an upper current blocking layer 32b. Yes. The surface of the contact layer 31 is covered with a protective film 33, and a p-side electrode 45 is formed so as to make ohmic contact with the contact layer 31 through the opening 33 b of the protective film 33.

  The structure of the active layer 43 of the semiconductor optical amplifier 14 is the same as that of the active layer 25-1, but the emission wavelength takes into account the shift amount according to the thickness of the well layer and the sulfur carrier concentration of the substrate 22. Thus, it is preferable to set the composition and thickness of the well layer and the barrier layer so as to be near the center of the wavelength range of the laser light output from the DFB laser stripes 11-1 to 11-n.

(Operation)
The operation of the integrated semiconductor laser device 100 will be described. First, one DFB laser stripe selected from the DFB laser stripes 11-1 to 11-n is driven. Since the trench grooves 17-1 to 17-m electrically isolate the DFB laser stripes 11-1 to 11-n, the isolation resistance between the DFB laser stripes increases, and the DFB laser stripes 11-1 to 11-n One of them can be easily selected and driven.

  Next, the optical waveguide optically connected to the driven DFB laser stripe among the plurality of optical waveguides 12-1 to 12-n guides the laser light from the driven DFB laser stripe. The optical combiner 13 passes the light guided through the optical waveguide and outputs it from the optical output port 13a. The semiconductor optical amplifier 14 amplifies the laser beam output from the optical output port 13a and outputs it from the optical output end 14a. The semiconductor optical amplifier 14 is used to compensate for the loss of light by the optical combiner 13 of the laser light from the driving DFB laser stripe and obtain a light output with a desired intensity from the output end.

  As described above, the laser oscillation wavelengths of the DFB laser stripes 11-1 to 11-n can be finely adjusted by changing the temperature of the integrated semiconductor laser device 100. This integrated semiconductor laser device 100 is set so that the gap between the laser oscillation wavelengths of the DFB laser stripes 11-1 to 11-n can be covered by fine adjustment by temperature setting. A continuous and wide wavelength tunable range can be realized by switching and temperature control.

(Production method)
Next, an example of a method for manufacturing the integrated semiconductor laser device 100 according to the present embodiment will be described. In addition, the following FIG. 7, FIG. 10, FIG. 15, FIG. 16 is a figure which shows the cross section corresponding to the DD line | wire cross section shown in FIG. Further, regions E1 to E4 in the figure are regions for forming DFB laser stripes 11-1 to 11-n, regions for forming optical waveguides 12-1 to 12-n, regions for forming the optical combiner 13, respectively. A region for forming the semiconductor optical amplifier 14 is shown.

  First, as shown in FIG. 7, a buffer layer 23 is grown on a substrate 22 to which sulfur has been added by using a metal organic chemical vapor deposition (MOCVD) method.

  Next, as shown in FIG. 8, a mask M 1 for selective growth is formed on the buffer layer 23. The mask M1 is made of a dielectric material such as SiN, and the width direction of the regions E1-1 to E1-n and E4-1 where the DFB laser stripes 11-1 to 11-n and the semiconductor optical amplifier 14 are to be formed. It is formed so as to sandwich both sides.

  FIG. 9 is an enlarged view showing the shape of the mask M1 shown in FIG. 8 in the region E1. As shown in FIG. 9, when the widths (mask widths) of the portions M1-1 to 1-n sandwiching both sides of the regions E1-1 to E1-n of the mask M1 are W1 to Wn, respectively, W1 <. Set to be Wn. W1 to Wn are set to a range of 0 μm to 40 μm, for example. The mask width sandwiching both sides in the width direction of the region E4-1 where the active layer of the semiconductor optical amplifier 14 is to be formed is, for example, an average mask width of W1 to Wn.

  Next, as shown in FIG. 10, using the MOCVD method, the lower SCH layer 24, the DFB laser stripes 11-1 to 11-n and the active layers of the semiconductor optical amplifier 14, the upper SCH layer 26, the spacer layer 27, A grating layer 28 and an upper cladding layer 44 are grown sequentially. 10 corresponds to the cross section taken along the line D-D in FIG. 1, so that the active layer is the active layer 25-3 of the DFB laser stripe 11-3 in the region E1 and the activity of the semiconductor optical amplifier 14 in the region E4. Layer 43 is shown. Regions E2 and E3 indicate active layers to be removed later.

  FIG. 11 is a diagram showing a cross-sectional structure along the line AA of the active layer 25-1 grown by the process of FIG. FIG. 12 is a diagram showing a cross-sectional structure along the line AA of the active layer 25-2 grown by the process of FIG. As shown in FIG. 11, portions M1-1 of a mask M1 having a mask width W1 are formed on both sides of a region E1-1 where the active layer 25-1 is formed. Therefore, more material is supplied to the region E1-1 depending on the mask width W1 than when the portion M1-1 is not present. As a result, the well layer 25a-1 and the barrier layer 25b-1 are formed slightly thicker than the case where there is no portion M1-1 depending on the width W1.

  On the other hand, as shown in FIG. 12, portions M1-2 of a mask M1 having a mask width W2 wider than W1 are formed on both sides of the region E1-2 in which the active layer 25-2 is formed. Therefore, more material is supplied to the region E1-2 than the region E1-1 according to the mask width W2. As a result, the well layer 25a-2 is formed thicker than the well layer 25a-1. The barrier layer 25b-2 is also formed thicker than the barrier layer 25b-1. As for the active layers formed in the other regions E1-3 to E1-n, the well layer and the barrier layer are formed thicker as the mask width becomes wider.

FIG. 13 is a diagram illustrating an example of the relationship between the mask width and the thickness of the well layer. The width between the masks is 15 μm. As shown in FIG. 13, the thickness of the well layer increases as the mask width increases. FIG. 14 is a diagram illustrating an example of the relationship between the mask width, the emission wavelength, and the energy gap corresponding to the mask width. The unit of the energy gap is an electron volt (eV, 1 eV is about 1.60 × 10 ⁻¹⁹ joules). As shown in FIG. 14, since the thickness of the well layer increases as the mask width increases, the energy gap between the quantum levels formed in the well layer decreases, so that the emission wavelength is longer. Wavelength. In the example shown in FIG. 14, by setting the mask width to 80 μm, the energy gap can be made smaller by about 100 meV than when there is no mask, and the emission wavelength can be made longer by about 0.14 μm. In the present manufacturing method, an active layer having well layers having different thicknesses is formed in each of the regions E1-1 to E1-n using the relationship shown in FIG.

  Next, after removing the mask M1, a SiN film is deposited again on the entire surface, and the DFB laser stripes 11-1 to 11-n in the region E1 are formed at the positions corresponding to the respective laser oscillation wavelengths. Patterning is performed so as to obtain a diffraction grating pattern. Further, patterning is also applied to the areas E2 to E4. Then, etching is performed using the SiN film as a mask to form a grating groove 46 serving as a diffraction grating in the grating layer 28 so as to penetrate the upper cladding layer 44 in the region E1, and thereafter, all the grating layers 28 in the regions E2 to E4 are formed. remove. Next, after removing the mask of the SiN film, upper cladding layers 29 and 44 are grown by MOCVD (see FIG. 15).

  Next, after a SiN film is deposited on the entire surface, patterning is performed on each of the regions E1 and E4 so that the pattern has a slightly wider shape than the DFB laser stripe and the semiconductor optical amplifier. Then, using this SiN film as a mask, the upper cladding layers 29 and 44, the grating layer 28, the spacer layer 27, the upper SCH layer 26, the active layer, and the lower SCH layer 24 other than the portion covered with the mask are removed. Then, using the SiN film as it is as a selective growth mask, a core layer 41 and an upper clad layer 42 are sequentially grown by MOCVD as shown in FIG.

  Next, after removing the mask of the SiN film, a new SiN film is deposited, and the DFB laser stripes 11-1 to 11-n, the optical waveguides 12-1 to 12-n shown in FIG. Then, patterning is performed so that a pattern corresponding to the semiconductor optical amplifier 14 is obtained. Then, this SiN film is used as a mask to form a mesa structure corresponding to the DFB laser stripes 11-1 to 11-n, the optical waveguides 12-1 to 12-n, the optical combiner 13, and the semiconductor optical amplifier 14. At the same time, the buffer layer 23 is exposed.

  FIG. 17 is a plan view showing a state after this step is performed. In the regions E1 to E4, mesa structures MESA1 to MESA4 having shapes corresponding to the DFB laser stripes 11-1 to 11-n, the optical waveguides 12-1 to 12-n, the optical combiner 13, and the semiconductor optical amplifier 14, respectively. Is formed.

  Next, the lower current blocking layer 32a and the upper current blocking layer 32b are sequentially grown on the exposed buffer layer 23 using the MOCVD method using the SiN film mask used in the previous step as a selective growth mask. A current blocking layer 32 is formed. Next, after removing the mask of the SiN film, the upper clad layer 30 and the contact layer 31 are sequentially grown on the entire surface of the regions E1 to E4 using the MOCVD method. FIG. 18 is a view showing a cross section taken along line CC in FIG. 1, that is, a cross section in the region E4 where the semiconductor optical amplifier 14 is formed.

  Next, heat treatment is performed to shift the emission wavelength of each active layer included in the DFB laser stripes 11-1 to 11-n and the semiconductor optical amplifier 14 to the short wavelength side. The shift amount of the emission wavelength is defined by the following parameters (1) to (4): (1) heat treatment temperature, (2) heat treatment time, (3) thickness of the well layer included in the active layer, (4 ) Since it depends on the sulfur carrier concentration of the substrate, the emission wavelength of each active layer is shifted to a desired wavelength by appropriately combining these parameters. The emission wavelengths of the active layers of the DFB laser stripes 11-1 to 11-n are preferably shifted so as to be close to the selected wavelength of the diffraction grating formed in each grating layer.

Here, the relationship between each parameter and the shift amount of the emission wavelength by heat treatment is illustrated. First, FIG. 19 is a diagram showing an example of the relationship between the thickness of the well layer and the shift amount of the emission wavelength. FIG. 19 shows the case where the heat treatment is performed for 60 minutes with the heat treatment temperature set to 800 ° C. and the sulfur carrier concentration of the substrate set to 8 × 10 ¹⁸ cm ⁻³ . As shown in FIG. 19, the shift amount of the emission wavelength toward the short wavelength side can be increased as the thickness of the well layer is reduced. The thickness of the well layer is preferably 10 nm or less because the shift amount can be sufficiently increased. Moreover, if it is 1 nm or more, a well layer with good crystallinity can be grown, and if it is 3 nm or more, it is more preferable because it can be easily controlled to a desired thickness. The thickness of the barrier layer is not particularly limited, but is about 9 to 20 nm, for example.

Next, FIG. 20 is a diagram illustrating an example of the relationship between the heat treatment temperature and the shift amount of the emission wavelength. FIG. 20 shows a case where the heat treatment is performed for 60 minutes with the thickness of the well layer being 3 nm and the sulfur carrier concentration of the substrate being 8 × 10 ¹⁸ cm ⁻³ . As shown in FIG. 20, the higher the heat treatment temperature, the larger the shift amount of the emission wavelength to the short wavelength side. The heat treatment temperature is preferably 600 ° C. or higher, for example, because the shift amount can be sufficiently increased. Further, it is preferable that the temperature is such that the quality of the semiconductor crystal constituting each optical element is not deteriorated by a high temperature, for example, 800 ° C. or less.

Next, FIG. 21 is a diagram showing an example of the relationship between the heat treatment time and the shift amount of the emission wavelength. FIG. 21 shows a case where the thickness of the well layer is 3.9 nm, the heat treatment temperature is 800 ° C., and the sulfur carrier concentration of the substrate is 8 × 10 ¹⁸ cm ⁻³ . As shown in FIG. 21, the longer the heat treatment time, the larger the shift amount of the emission wavelength to the short wavelength side. The heat treatment time is preferably, for example, 10 minutes or longer because the shift amount can be sufficiently increased. In addition, when the heat treatment time is high, it is preferable that the heat treatment time is such that it does not deteriorate the quality of the semiconductor crystals constituting each optical element, for example, 60 minutes or less.

Next, FIG. 22 is a diagram illustrating an example of the relationship between sulfur carrier concentration and emission wavelength shift amount. FIG. 22 shows a case where the thickness of the well layer is 3 nm, the heat treatment temperature is 800 ° C., and the heat treatment is performed for a sufficiently long time of about 30 minutes. As shown in FIG. 22, as the sulfur carrier concentration is increased, the shift amount of the emission wavelength to the short wavelength side can be increased. For example, a carrier concentration of 1 × 10 ¹⁸ cm ⁻³ or more is preferable because a shift toward the short wavelength side can be generated. Moreover, since there exists a possibility that crystal quality may deteriorate, it is preferable to set it as 1 * 10 < ¹⁹ > cm <^-3> or less.

19 to 22, for example, the sulfur carrier concentration of the substrate 22 is set to 8 × 10 ¹⁸ cm ^−3, and in each active layer of the DFB laser stripes 11-1 to 11-n, the well layer When the thickness is in the range of 3 nm to 9 nm and heat treatment is performed at a temperature of 800 ° C. for 60 minutes, the maximum shift amount to the short wavelength side is about 37 meV (when the thickness of the well layer is 3 nm). The minimum value is about 10 meV (when the thickness of the well layer is 9 nm). Therefore, the wavelength range that can be controlled by 27 meV in terms of energy can be further widened than the wavelength range that can be controlled by the mask width in the conventional selective growth method.

  For example, in an integrated semiconductor laser device manufactured using conventional selective growth, the wavelength range of the output laser light is 1.20 μm to 1.34 μm, and a control wavelength range of 140 nm (about 100 meV) is realized. It had been. On the other hand, if the relationship shown in FIGS. 19 to 22 is used, a wider control wavelength range of 1168 nm to 1326 nm of 168 nm (127 meV) can be realized.

  Note that the SiN film mask is removed after the heat treatment. Thereafter, after depositing a SiN film on the entire surface, patterning is performed so as to form a pattern corresponding to the trench groove, and etching is performed using this SiN film as a mask to form the trench grooves 17-1 to 17-m shown in FIG. Form. The trench grooves 17-1 to 17-m are formed to a depth reaching the buffer layer 23. However, the trench grooves 17-1 to 17-m are not limited to this, and may be formed to a depth at which the DFB laser stripes 11-1 to 11-n can be electrically separated. . After the etching process, the SiN film mask is removed. Then, openings 33a and 33b for the DFB laser stripes 11-1 to 11-n and the semiconductor optical amplifier 14 are formed in the SiN film again deposited on the entire surface to form the SiN protective film 33. Next, after depositing a three-layer conductive film made of Ti / Pt / Au on the entire surface, the p-side is patterned by patterning into shapes corresponding to the DFB laser stripes 11-1 to 11-n and the semiconductor optical amplifier 14. Electrodes 34 and 45 are formed. On the other hand, an n-side electrode 21 having a two-layer structure made of AuGeNi / Au is formed on the back surface of the substrate 22. Through the above manufacturing process, the semiconductor optical amplifier 14, which is an optical element, the DFB laser stripes 11-1 to 11-n, the optical waveguides 12-1 to 12-n, and the optical combiner 13 are formed.

  Finally, the substrate 22 is cleaved into a bar shape in which a plurality of integrated semiconductor laser elements 100 are arranged, and antireflection films are coated on both end surfaces on which the DFB laser stripes 11-1 to 11-n and the semiconductor optical amplifier 14 are formed. Thereafter, the integrated semiconductor laser device 100 is completed by separating each device.

  As described above, according to this manufacturing method, the control range of the emission wavelength can be widened, so that the integrated semiconductor laser device 100 having a wider band can be manufactured.

  By the way, in this manufacturing method, the shift amount of the emission wavelength of each active layer depends on the heat treatment temperature. This heat treatment temperature can be adjusted by adjusting the temperature of the substrate 22. The temperature of the substrate 22 is usually uniform within the substrate surface, but if the temperature distribution is formed within the substrate surface, the emission wavelength of the active layer can be controlled more freely.

  FIG. 23 is a partial cross-sectional view illustrating an example of a method for forming a temperature distribution on a substrate. In FIG. 23, a substrate 47 on which an active layer and a core layer of a passive optical waveguide are formed is mounted on a mounting table 50 including a susceptor 51 made of carbon and a tray 52 made of quartz glass mounted on the susceptor 51. ing. In addition, temperature distribution forming films 53 and 54 having different thermal conductivities are formed on the back surface of the substrate 47 facing the tray 52. The temperature distribution forming films 53 and 54 are made of, for example, a dielectric film such as SiNx or a metal film, and can be formed using a CVD method or a vapor deposition method.

  When the susceptor 51 is heated by high frequency induction using an RF coil or the like while rotating the mounting table 50 around the axis, the susceptor 51 heats the substrate 47 through the tray 52 by heat conduction. Here, since the temperature distribution forming films 53 and 54 exist between the substrate 47 and the tray 52, the amount of heat transferred to the substrate 47 on the temperature distribution forming film 53 and the temperature distribution forming film 54. Therefore, a temperature distribution is formed in the plane of the substrate 47.

  If heat treatment is performed in this manner, the heat treatment can be performed at different temperatures depending on the position in the plane of the substrate 47. Therefore, the emission wavelength of the active layer formed on the substrate 47 can be shifted by different shift amounts depending on the position of the active layer. Therefore, for example, an active layer with a large shift amount and an active layer with a suppressed shift amount can be formed simultaneously.

  FIG. 24 is a partial cross-sectional view illustrating another example of a method for forming a temperature distribution on a substrate. In FIG. 24, the substrate 47 is mounted on a mounting table 60 including a susceptor 51 and a tray 62 made of quartz glass mounted on the susceptor 51. Here, a recess 62 a is formed in the tray 62. Since the substrate 47 receives heat conduction through the air on the concave portion 62a, the amount of heat transferred is smaller than that in other regions, so that a temperature distribution can be generated in the plane of the substrate 47. Therefore, the emission wavelength of the active layer formed on the substrate 47 can be shifted by different shift amounts depending on the position of the active layer. In the example shown in FIG. 24, the recess 62 a is formed in the tray 62, but the same effect can be obtained by forming a recess on the back surface of the substrate 47.

  Further, the temperature distribution forming films 53 and 54 and the recesses formed on the back surface of the substrate 47, and the recesses 62a formed on the tray 62 are formed in a fine pattern having the same size as that of the optical element by using, for example, a photolithography technique. Therefore, a temperature distribution corresponding to the size of the individual optical elements can be formed in the plane of the substrate 47. Thereby, the shift amount of the emission wavelength of the active layer of each optical element can be controlled to a different shift amount for each optical element.

  As described above, if the temperature distribution is formed in the substrate surface, the emission wavelength of the active layer can be controlled more freely. Further, by forming a temperature distribution in the substrate plane, the active layer of the DFB laser stripes 11-1 to 11-n suppresses the shift amount of the emission wavelength, and the band of the core layer of the passive optical waveguide. The gap energy can be greatly shifted to the short wavelength side. As a result, the transmittance of the core layer of the optical waveguides 12-1 to 12-n at the wavelength of the laser light output from the DFB laser stripes 11-1 to 11-n can be increased, and the transmission loss can be reduced. The core layers of the optical waveguides 12-1 to 12-n may have a multilayer structure including a quantum well structure. In this case, by shifting the band gap energy of each core layer by an amount corresponding to the laser oscillation wavelength of each DFB laser stripe 11-1 to 11-n connected thereto, each DFB laser stripe 11-1 Optical waveguides 12-1 to 12-n optimized for 11-n can be formed.

  In the above embodiment, the emission wavelength is shifted by heat treatment by a shift amount corresponding to the thickness of the well layer included in the active layer and the sulfur carrier concentration of the substrate, but the barrier layer included in the active layer The amount of shift may be shifted according to the thickness of the substrate and the carrier concentration of sulfur in the substrate.

  In the integrated semiconductor laser element according to the above-described embodiment, the material, size, etc., of the compound semiconductor and electrodes constituting the substrate and each semiconductor layer are set according to the wavelength used. However, each material, size, and the like can be appropriately set according to the wavelength used, and are not particularly limited. Moreover, in the said embodiment, although an active layer is a multiple quantum well structure, it is good also as a single quantum well structure.

  Further, although the above embodiment is an integrated semiconductor laser device, the present invention is not limited to this, and can be applied to an optical integrated device in which optical devices having an active layer having a quantum well structure are integrated. For example, as an optical element having an active layer having a quantum well structure, for example, an electroabsorption optical modulator or a photodiode can be used. The electroabsorption optical modulator and the photodiode can shift the light absorption wavelength in accordance with the shift of the emission wavelength of the active layer, so that the wavelength dependence of the optical characteristics can be controlled. Therefore, the present invention can be applied to various types of optical integrated devices in which optical devices such as semiconductor laser devices, semiconductor optical amplifiers, electroabsorption optical modulators, and photodiodes are integrated on the same substrate. It is.

11-1 to 11-n DFB laser stripes 12-1 to 12-n Optical waveguide 13 Optical combiner 13a Optical output port 14 Semiconductor optical amplifier 14a Optical output end 15 Buried portion 17-1 to 17-m Trench groove 21 n side Electrodes 22, 47 Substrate 23 Buffer layer 24 Lower SCH layer 25-1 to 25-3, 43 Active layer 25a-1, 25a-2 Well layer 25b-1, 25b-2 Barrier layer 26 Upper SCH layer 27 Spacer layer 28, 28-1, 28-2 Grating layer 29, 30, 42, 44 Upper cladding layer 31 Contact layer 32 Current blocking layer 32a Lower current blocking layer 32b Upper current blocking layer 33 Protective film 33a, 33b Opening 34, 45 p-side electrode 41 Core layer 46 Lattice groove 50, 60 Mounting table 51 Susceptor 52, 62 Tray 53, 54 Temperature distribution film 62a Concave portion 100 Integrated semiconductor laser element E1 to E4, E1-1 to E1-n region M1 mask M1-1 to M1-n part MESA1 to MESA4 mesa structure

Claims (6)

A plurality of semiconductor layers constituting a quantum well structure composed of a well layer and a barrier layer on a semiconductor substrate containing sulfur, wherein the semiconductor layers form a plurality of semiconductor layers having different thicknesses from each other. Forming process;
An optical element forming step of forming a plurality of optical elements each including the plurality of semiconductor layers;
A heat treatment step for performing a heat treatment for shifting emission wavelengths of the plurality of semiconductor layers;
A method for manufacturing a semiconductor optical integrated device, comprising:   The method of manufacturing a semiconductor optical integrated device according to claim 1, wherein the heat treatment step is performed in a range of 600 ° C. to 800 ° C.   3. The method of manufacturing a semiconductor optical integrated device according to claim 1, wherein in the semiconductor layer forming step, the well layer or the barrier layer is formed to a thickness of 1 nm to 10 nm.   The method of manufacturing a semiconductor optical integrated device according to claim 1, wherein the semiconductor layer forming step grows the plurality of semiconductor layers using a selective growth method. 5. The semiconductor optical integrated device according to claim 1, wherein the semiconductor substrate contains sulfur having a carrier concentration of 1 × 10 ¹⁸ cm ⁻³ to 1 × 10 ¹⁹ cm ⁻³ . Production method.   6. The method of manufacturing a semiconductor optical integrated device according to claim 1, wherein the heat treatment step is performed by forming a temperature distribution in the surface of the semiconductor substrate.

Images