JP2011181638A - Method of manufacturing distributed feedback semiconductor laser - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method of manufacturing a DFB semiconductor laser that can suppress an increase in manufacturing cost when the DFB semiconductor laser having a current constriction structure is manufactured. <P>SOLUTION: The DFB semiconductor laser 100 is composed of a compound semiconductor, and has a lower clad layer 2, a lower light guide layer 3, a semi-insulating active layer 4, an upper light guide layer 5, an upper first clad layer 6, an optical block layer (etch stop layer) 7, an upper second clad layer 8, a diffraction grating (layer) 9, an upper third clad layer 10, and a contact layer (cap layer) 11 formed in order on a semiconductor substrate 1. Any of the upper second clad layer 8, diffraction grating 9, upper third clad layer 10, and contact layer 11 contains Ga and As, and etching is carried out using the same etchant. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a method for manufacturing a distributed feedback semiconductor laser (DFB semiconductor laser).

  A Fabry-Perot laser (FP laser) using a Fabry-Perot resonator has a structure in which light is confined between both end faces of the resonator to oscillate, and laser light is emitted from one end face of the resonator. However, the FP laser has an unstable oscillation wavelength, the oscillation wavelength changes according to the drive current and temperature in a steady state, and operates in multimode during high-speed modulation. On the other hand, since the DFB semiconductor laser has a diffraction grating inside the resonator and selectively amplifies laser light having a specific wavelength, it can oscillate in a single mode. A conventional DFB semiconductor laser is described in Patent Document 1, for example. In Patent Document 1, a current confinement layer is provided above the active layer in order to increase the current density flowing into the active layer.

Japanese Patent Laid-Open No. 2003-069145

  However, when the structure is complicated and the number of semiconductor layers is increased, it is necessary to perform dry etching with different ionic species as etching before forming the current confinement layer, resulting in an increase in manufacturing cost.

  The present invention has been made in view of such problems, and provides a method for manufacturing a DFB semiconductor laser capable of suppressing an increase in manufacturing cost when manufacturing a DFB semiconductor laser having a current confinement structure. Objective.

  In order to solve the above-described problem, a method for manufacturing a DFB semiconductor laser according to the present invention includes a step of sequentially laminating an optical guide layer, an upper first cladding layer, an etch stop layer, and an upper second cladding layer on an active layer. Forming a diffraction grating on the upper second cladding layer; forming an upper third cladding layer on the diffraction grating; and forming a contact layer on the third cladding layer. A pattern extending in the resonance length direction is formed on the contact layer, and the contact layer, the upper third cladding layer, the diffraction grating, and the upper second cladding layer are etched using the pattern as a mask, Exposing the side surfaces of the contact layer, the upper third cladding layer, the diffraction grating, and the upper second cladding layer and the surface of the etch stop layer; A current confinement layer is formed on the exposed region on the top layer, and the current confinement layer covers the exposed side surfaces of the contact layer, the upper third cladding layer, the diffraction grating, and the upper second cladding layer. The upper second cladding layer, the diffraction grating, the upper third cladding layer, and the contact layer all contain Ga and As, and the etching is performed using the same etching solution. It is characterized by being performed by.

  In the case of this structure, since it is possible to perform wet etching using a single etching solution, manufacturing cost can be reduced.

In particular, the upper second cladding layer is made of Al _X Ga _1-X As (X = 0.43 ± 0.01), and the diffraction grating is Al _X Ga _1-X As (X = 0.30 ± 0.0. 01), the upper third cladding layer is made of Al _X Ga _1-X As (X = 0.43 ± 0.01), the contact layer is made of GaAs, and the etch stop layer is made of Al _X Ga _{1. In the case of -XAs} (X = 0.70 ± 0.01), etching can be performed using a citric acid-based etchant as an etchant.

  The manufacturing method further includes a step of forming the light guide layer as an upper light guide layer and forming the active layer on the lower light guide layer, and the total thickness of the lower light guide layer and the upper light guide layer is The lower and upper light guide layers are grown so as to have a thickness of 50 nm to 200 nm. In the case of such a complicated structure, the performance of the DFB semiconductor laser can be remarkably improved. Since such a complicated structure can also form each layer using the same kind of material, the manufacturing process is simplified. The manufacturing cost can be further reduced by using the above-described manufacturing method.

  According to the method for manufacturing a DFB semiconductor laser of the present invention, manufacturing costs can be reduced.

1 is a perspective view of a DFB semiconductor laser 100. FIG. FIG. 2 is a cross-sectional view taken along the line II-II of the DFB semiconductor laser 100 shown in FIG. It is a front view of a DFB semiconductor laser. It is a graph which shows the physical property of a semiconductor layer. It is sectional drawing of the diffraction grating vicinity. It is a figure which shows a beam shape. It is a graph which shows the relationship between thickness _TG of a light guide layer, radiation angle (theta), and optical confinement coefficient (GAMMA) QW. It is a graph which shows the thickness _TG of a light guide layer, the radiation angle (theta), the optical confinement coefficient (GAMMA) QW, and the figure of merit FOM. It is a graph which shows optical output Po (mW) with respect to drive current If (mA), optical output differential value SE (W / A), voltage Vf (V) between electrodes, and resistance Rd (ohm) between electrodes. It is a graph which shows the relationship of the laser beam intensity | strength I (au) (normalization) with respect to radiation angle (theta) (deg). It is a chart which shows the data of the graph of FIG. It is a graph which shows the relationship of laser beam intensity | strength I (au) (normalization) with respect to the radiation angle (theta) (deg) (solid line) of a perpendicular direction, and the radiation angle (theta) (deg) (dotted line) of a horizontal direction. The relationship between the laser beam intensity I (au) (normalized) with respect to the vertical radiation angle θ (deg) (solid line) and the horizontal radiation angle θ (deg) (dotted line) of the DFB laser according to the comparative example. It is a graph to show. It is a figure for demonstrating the manufacturing method of a DFB semiconductor laser. It is a figure for demonstrating the manufacturing method of a DFB semiconductor laser. It is a figure for demonstrating the manufacturing method of a DFB semiconductor laser. It is a figure for demonstrating the manufacturing method of a DFB semiconductor laser. It is a figure for demonstrating the manufacturing method of a DFB semiconductor laser. It is a figure for demonstrating the manufacturing method of a DFB semiconductor laser. It is a figure for demonstrating the manufacturing method of a DFB semiconductor laser. It is a graph which shows the temperature characteristic of a DFB semiconductor laser. It is a graph which shows the relationship between the drive current If (mA) of DFB semiconductor laser, and optical output Po (mW). The wavelength dependence of the laser which concerns on an Example is shown, The horizontal axis shows wavelength (lambda) (nm) and the vertical axis | shaft has shown optical output Po (mW) (logarithm). Example _(Y G = 90 nm) of the laser beam intensity to the emission angle theta (deg) in the FFP where I (A.U.) Is a graph showing the relationship between the (normalized). It is a graph which shows optical output Po (mW) with respect to dynamic current If (mA) in the case of the comparative example of B, and optical output differential value SE (W / A). 14 is a chart showing data of the graphs of FIGS. 12 and 13. 14 is a chart showing data of the graphs of FIGS. 12 and 13.

  The distributed feedback semiconductor laser (DFB semiconductor laser) according to the embodiment will be described below. In the description, the same reference numerals are used for the same elements, and duplicate descriptions are omitted.

1 is a perspective view of the DFB semiconductor laser 100, FIG. 2 is a sectional view taken along the line II-II of the DFB semiconductor laser 100 shown in FIG. 1, and FIG. 3 is a front view of the DFB semiconductor laser. Further, specific materials, Al composition ratio X, thickness, refractive index, energy band gap, conductivity type, and impurity concentration of each element are as shown in the table of FIG. In addition, although the conductivity type I means an intrinsic semiconductor, it actually means a P-type semiconductor having an extremely low impurity concentration of 1 × 10 ¹⁵ cm ⁻³ or less.

  The DFB semiconductor laser 100 is made of a compound semiconductor, and has a lower cladding layer 2, a lower light guide layer 3, a semi-insulating active layer 4, an upper light guide layer 5, and an upper first cladding layer on a semiconductor substrate 1. 6, an optical block layer (etch stop layer) 7, an upper second cladding layer 8, a diffraction grating (layer) 9, an upper third cladding layer 10, and a contact layer (cap layer) 11 are sequentially formed. The refractive index of the cladding layer is lower than the average refractive index of the active layer 4 and functions as an optical confinement layer. All the cladding layers 2, 6, 8, and 10 have the same Al composition ratio and the same energy band gap, but there may be some difference. Further, the refractive index of the cladding layer is lower than the refractive indexes of the adjacent light guide layers 3 and 5. The light guide layers 3 and 5 have an energy band gap larger than that of the well layer constituting the active layer 4 having the quantum well structure, and have a refractive index lower than that of the well layer.

  The upper electrode E1 is in contact with the contact layer 11 and extends along the Z-axis direction. Further, the lower electrode E2 is in contact with the entire back surface of the semiconductor substrate 1. Note that the impurity concentration of the semiconductor at the interface can be increased at the interface between the electrode and the semiconductor so that these good ohmic contacts can be obtained. Gold (Au) can be used as the electrode material. The Z-axis direction coincides with the direction of the laser resonance length (L: FIG. 2). In the figure, an orthogonal coordinate system including an X axis, a Y axis, and a Z axis is shown. Assuming that the direction of the resonance length is the Z axis, the stacking direction of the semiconductor layers is the Y axis direction, and the direction (width direction) perpendicular to both of these axes is the X axis direction. In this DFB semiconductor laser, a laser beam LB is emitted from an end surface perpendicular to the Z axis of the active layer 4. An end face film R such as an antireflection (AR) film or a high reflection (HR) film is formed on the end face opposite to the light emitting face of the active layer 4.

  The upper third cladding layer 10, the diffraction grating 9, and the upper second cladding layer 8 in the outer region along the X-axis direction of the contact layer 11 are etched and processed into a mesa shape. The current passing region is limited to the region immediately below 11, and a current confinement structure is formed. In the etched region, a current confinement layer (buried layer) 12 (12A, 12B) is formed on the etch stop layer 7, and the refractive index of the current confinement layer 12 is lower than that of the clad layer. Is suppressed. The opening width W in the X-axis direction between the current confinement layers 12A and 12B is set to 5 μm in this example. In addition, the resonance length L of this example is 1.0 mm.

  Etching of the cladding layer or the like is stopped by the optical block layer (etch stop layer) 7 and the upper first cladding layer 6 remains without being etched. The etch stop layer 7 has improved resistance to the etching solution by increasing the Al composition ratio more than the cladding layers 8 and 10. Note that the higher the Al content, the higher the resistance and the lower the refractive index, but the larger the energy band gap.

  Since the etch stop layer 7 has a low refractive index, it functions as an optical block layer that suppresses light from the active layer 4 to the diffraction grating 9. That is, the refractive index of the diffraction grating 9 is higher than that of the cladding layers 6 and 8, and light tends to be absorbed into the semiconductor layer having a high refractive index. Therefore, in such a case, part of the optical power that is originally concentrated on the active layer 4 is largely moved to the diffraction grating 9, and the laser light intensity is reduced. The etch stop layer 7 functions as an optical block layer with a low refractive index, thereby suppressing the optical power from entering the diffraction grating 9 more than necessary and preventing the laser light intensity from decreasing.

  The active layer 4 has a quantum well structure composed of AlInGaAs and AlGaAs, and has a single quantum well structure when there is one well layer and a double quantum well structure when there are two well layers. The number of well layers may be 3 or more. The energy band gap of the barrier layer is larger than the energy band gap of the well layer. When a multi-quantum well structure is used, a high-power laser beam can be obtained, but COD (catalytic optical damage) may increase as the number of wells increases.

  The Al composition ratio X of each semiconductor layer can take at least the following ranges in consideration of errors and the like.

Composition ratio X in the lower clad layer 2: 0.43 ± 0.01
Composition ratio X in the lower light guide layer 3: 0.36 ± 0.01
Composition ratio X in the semi-insulating active layer 4: 0.15 ± 0.01
Composition ratio X in the upper light guide layer 5: 0.36 ± 0.01
Composition ratio X in upper first cladding layer 6: 0.43 ± 0.01
Composition ratio X in the light blocking layer 7: 0.70 ± 0.01
Composition ratio X in upper second cladding layer 8: 0.43 ± 0.01
Composition ratio X in the diffraction grating (layer) 9: 0.30 ± 0.01
Composition ratio X in upper third cladding layer 10: 0.43 ± 0.01

Various shapes of the diffraction grating 4 are conceivable. The high-refractive-index semiconductor layer periodically embedded in the low-refractive-index clad layer is composed of stripe-shaped semiconductor regions extending along the X-axis direction, and an interval Zp between these semiconductor regions (of the diffraction grating) The wavelength of the laser light is determined by the period: see FIG. That is, assuming that the effective refractive index of the waveguide including the diffraction grating 4 is n and the order of the diffraction grating 4 is m, the Bragg wavelength λ _B is approximately given by 2nZp / m, and this wavelength is selectively amplified, and the laser It is emitted as light.

  Of course, the surface shape in the YZ section may be a triangular wave shape as shown in FIG. That is, the surface of the cladding layer 8 is processed into a triangular wave shape in the YX cross section by etching, and a diffraction grating (layer) 9 having a high refractive index is formed thereon, and the diffraction grating layer 9 is further formed by the cladding layer 10. The structure can be embedded. In the case of a film thickness of 50 nm, it is difficult to accurately control the film thickness by simple unevenness, so it is preferable to form the diffraction grating in a stripe shape.

  The reason why the diffraction grating layer 9 is embedded in the third cladding layer 10 is that the distance from the active layer to the diffraction grating layer changes when the oscillation wavelength is changed, but even in such a case, the diffraction grating layer This is because it is possible to cope with the problem by simply changing the embedding position, so that there is an advantage of high versatility when changing the design.

  FIG. 6 shows the shape of the laser beam LB in the XY plane. (A) shows the far-field pattern (FFP), and (B) shows the near-field pattern (NFP). The boundary line of each pattern is shown at a position giving a half width. That is, the light intensity peak is located at the center position (origin), but the distance to a position that is half the light intensity at this position is the half width, and this position is shown as the outer edge of the beam pattern. .

In NFP, the vertical distance (Y _NFP ) is the half width σ _YN, and this position is shown as the outer edge of the beam pattern. The distance 2σ _{YN, which} is twice the half width σ _YN , is the full width at half maximum (FWHM). is there. Similarly, the distance 2σ _{XN which} is twice the half width σ _{XN in} the horizontal direction of the intensity at the origin is the full width at half maximum (FWHM). Also in the FFP, the vertical distance (Y _FFP ) is the half width σ _YF, and this position is shown as the outer edge of the beam pattern. The distance 2σ _{YF, which} is twice the half width σ _YF , is the full width at half maximum (FWHM). It is. Similarly, the distance 2σ _{XF which} is twice the horizontal half-value width σ _XF of the intensity at the origin is the full width at half maximum (FWHM).

The radiation angle of the laser light is given by a vector extending from the outer edge position of the NFP to the corresponding outer edge position of the FFP and the angle formed with the Z axis, and corresponds to the FWHM of the light intensity distribution. In the vertical direction, a vector extending from the position of + Y _{NFP to} the position of + Y _FFP has twice the angle formed with the Z axis as the radiation angle θ. Since the intensity distribution of the laser beam is a Gaussian distribution and is symmetric with respect to the center position, in other words, the radiation angle θ is a vector extending from the position of + Y _{NFP to} the position of + Y _{FFP and} the position of −Y _NFP . Is equal to the angle formed by the vector extending from to -Y _FFP .

Similarly, in the horizontal direction, a vector extending from the position of + X _{NFP to} the position of + X _FFP has twice the angle formed with the Z axis as the horizontal radiation angle. Note that the radiation angle is particularly problematic in the radiation angle having the larger angle, that is, the radiation angle θ in the vertical direction.

FIG. 7 is a graph showing the relationship between the total thickness Y _G (nm) of the light guide layers 3 and 5, the vertical radiation angle θ (deg), and the optical confinement coefficient ΓWQ (%). This graph is obtained by simulation.

In the graph, 760-SQW-FFP shows data of the vertical radiation angle θ in the FFP when the center wavelength of the laser light is 760 nm and the structure of the active layer is SQW (single quantum well structure). Similarly, 760-DQW-FFP shows data of the vertical radiation angle θ in the FFP when the center wavelength of the laser beam is 760 nm and the structure of the active layer is DQW (double quantum well structure). 830-SQW-FFP shows data of the radiation angle θ in the vertical direction in the FFP when the center wavelength of the laser beam is 830 nm and the structure of the active layer is SQW. 830-DQW-FFP shows data of the vertical emission angle θ in the FFP when the center wavelength of the laser beam is 830 nm and the structure of the active layer is DQW. 980-SQW-FFP shows data of the vertical emission angle θ in the FFP when the center wavelength of the laser beam is 980 nm and the structure of the active layer is SQW. The values shown in the table of FIG. 4 were used at wavelengths of 760 nm and 830 nm, but in the 980 nm band, Al _0.22 Ga _0.78 As as the light guide layer and Al _0.35 Ga _0.65 As as the cladding layer. Only the point using is different.

  In the graph, 760-SQW-Γ indicates data of the optical confinement coefficient ΓQW when the center wavelength of the laser beam is 760 nm and the structure of the active layer is SQW. Similarly, 760-DQW-Γ shows data on the optical confinement coefficient ΓQW when the center wavelength of the laser beam is 760 nm and the structure of the active layer is DQW. 830-SQW-Γ indicates data of the optical confinement coefficient ΓQW when the center wavelength of the laser beam is 830 nm and the structure of the active layer is SQW. 830-DQW-Γ shows data of the optical confinement coefficient ΓQW when the center wavelength of the laser beam is 830 nm and the structure of the active layer is DQW. 980-SQW-Γ shows data of the optical confinement factor ΓQW when the center wavelength of the laser beam is 980 nm and the structure of the active layer is SQW.

In the region where the thickness of the light guide layer ranges from about 1000 nm (1 μm) to 2000 nm, it has been found that the emission angle in the vertical direction is reduced by increasing the thickness of the light guide layer as in the prior art. When the light guide layer becomes extremely thin, that is, at least 200 nm or less, there is a critical thickness range (R _PEAK ) in which the radiation angle decreases and the oscillation threshold decreases. I found it. Of course, when the thickness of the light guide layer is extremely reduced, the oscillation threshold value rises. However, when the thickness is at least 50 nm or more, there is a thickness range in which the oscillation threshold value can be sufficiently lowered. did.

The vertical radiation angle θ in the far-field image is preferably small and the oscillation threshold (current) is preferably small, but the oscillation threshold is generally proportional to the inverse of the optical confinement factor ΓQW, and thus the optical confinement factor is large. This is preferable. That is, if the figure of merit FOM of this DFB semiconductor laser is defined as FOM = (ΓQW / θ) ² , the FOM has a peak when the thickness of the light guide layer is 50 nm to 200 nm.

FIG. 8 shows the total thickness Y _G (nm) of the light guide layers 3, 5, normalized FOM (au), vertical radiation angle θ (deg), and optical confinement factor ΓWQ × It is a graph which shows the relationship of 10 (%). It can be seen that the FOM has a peak between 50 nm and 200 nm. FIG. 11 is a chart showing the graph data.

In this graph, the square of the value obtained by dividing the radiation angle θ of the 830 nm-DQW structure by the optical confinement coefficient ΓQW is shown as FOM. Since ΓQW includes an error, a total of four data in the 760 nm band (SQW and DQW) and 830 nm band (SQW and DQW) were averaged and used for the calculation. There is a peak between 50 nm and 200 nm, and the FOM peak can be similarly set to Y _G = 50 nm to 200 nm with respect to the radiation angle θ in other wavelength bands. Of course, Y _G = 80 nm to 140 nm is more preferable, and Y _G = 100 nm to 120 nm is more preferable. In addition, the AlGaAs material forming each layer of the crystal structure has an error of about Al of Al composition ± 1% with the accuracy of the growth apparatus, and a refractive index change accompanying this. Since the value of ΓQW is affected by this refractive index change, there is an error of about ± 0.5% in ΓQW, but the superiority of the range R _{PEAK in} the conclusion does not change.

If optical confinement is not realized well, ΓQW is less than 0.5%. This means that the amount of light in the gain region is insufficient, and oscillation characteristics cannot be obtained. When the guide width _{T G} was changed in 20nm increments to 80nm~140nm at 830nm band, the vertical radiation angle becomes 15 degrees to 20 degrees. If the number of quantum wells is 3 or more, the width of about 50 nm is occupied by itself, so that the degree of freedom in design is reduced. Therefore, the number of wells is preferably 1 or 2. If the guide width _TG is 50 nm or less, ΓQW will be too low, and good oscillation characteristics cannot be obtained. Moreover, when it becomes 200 nm or more, the expansion of the vertical radiation angle is induced.

FIG. 9 shows the optical output Po (mW), the optical output differential value SE (W / A), the interelectrode voltage Vf (V), and the interelectrode resistance Rd with respect to the drive current If (mA) when Y _G = 80 nm to 140 nm. It is a graph which shows (ohm) (experimental value). Threshold of oscillation with increasing Y _G is seen to decrease.

FIG. 10 is a graph showing the relationship between the laser beam intensity I (au) (normalized) and the radiation angle θ (deg) in the FFP when Y _G = 80 nm to 140 nm (experimental value). In the graph, FFP-H represents the light intensity distribution in the horizontal direction, and FFP-V represents the light intensity distribution in the vertical direction. The radiation angle θ giving the FWHM of the light intensity distribution in the vertical direction is 16.1 (deg) when Y _G = 80 nm, 18.3 (deg) when Y _G = 100 nm, and Y _G = 120 nm. In this case, it is 20.2 (deg), and in the case of Y _G = 140 nm, it is 21.8 (deg).

FIG. 12 shows the laser light intensity I (au) with respect to the vertical radiation angle θ (deg) (solid line) and the horizontal radiation angle θ (deg) (dotted line) in the FFP when Y _G = 80 nm to 140 nm. ) (Normalization) is a graph showing the relationship (calculated value). FIG. 13 is a graph showing data of the DFB laser according to the comparative example. Laser light with respect to the vertical radiation angle θ (deg) (solid line) and the horizontal radiation angle θ (deg) (dotted line) in the FFP. The relationship of intensity I (au) (normalized) is shown (calculated value). 26 and 27 are tables showing original data for obtaining the data of FIGS. 12 and 13, respectively, and show data before normalization and data after normalization, respectively. Note that FFP-H in the table indicates data in the horizontal direction, and FFP-V indicates data in the vertical direction.

  According to the graph, in the numerical values of the examples (A) to (D), the radiation angle θ in the FWHM is narrow. In addition, the aspect ratio (ratio of the radiation angle in the horizontal direction and the radiation angle in the vertical direction) of the embodiment is small, and the beam shape is close to a perfect circle, so that application to a device is easy.

  In the DFB semiconductor laser having the above-described structure, a diode having a P-type upper cladding layer as an anode and an N-type lower cladding layer as a cathode is configured. The drive current flows in the forward direction of the diode, that is, from the upper electrode E1 to the lower electrode E2, and at this time, a part of the light generated by carrier recombination in the active layer 4 exceeds the light blocking layer 7, The wavelength of the laser beam LB that reaches the diffraction grating 9 and is emitted from the end face of the active layer 4 is fixed to a single value. The distance between the diffraction grating 9 and the active layer 4 is preferably as close as possible in order to reduce the manufacturing time of the cladding layers 6 and 8 and the loss in light propagation in the vertical direction. If the light blocking layer 7 is not provided, a large amount of light is absorbed by the diffraction grating 9 and the laser light output is reduced.

  That is, since the DFB semiconductor laser includes the optical block layer 7, even when the diffraction grating 9 is brought close to the active layer 4, light can be emitted with high efficiency. In other words, this structure is particularly effective when the distance (t5 + t6 + t7 + t8) from the active layer 4 to the diffraction grating 9 is set to 300 nm to 700 nm. For the same reason, the lower end position of the diffraction grating layer 9 is preferably located within 1/3 from the upper end of the etch stop layer with respect to the total thickness (t8 + t9 + t10 + t11) of the ridge portion (protruded portion). In addition, the optical block layer 7 also functions as an etch stop layer when forming a buried type current confinement structure, and thus is useful from the viewpoint of the manufacturing process. Since the light blocking layer 7 also suppresses light from the active layer 4 to the contact layer 11, the contact layer 11 can be suppressed from deterioration due to light absorption.

  14 to 20 are views for explaining the above-described method for manufacturing the DFB semiconductor laser, in which (A) is a front view and (B) is a side view.

First, as shown in FIG. 14, the layers 2, 3, 4, 5, 6, 7, 8, and 9 are sequentially stacked on the substrate 1. That is, the light guide layer 5, the upper first clad layer 6, the etch stop layer 7, and the upper second clad layer 8 are formed on the semiconductor substrate 1 on the lower clad layer 2, the light guide layer 3, and the active layer 4. A compound semiconductor layer (9) having a high refractive index to be a diffraction grating 9 is further formed thereon, and a high refractive index compound semiconductor layer (9) is formed thereon. Note that an MOCVD (metal organic chemical vapor deposition) apparatus is used for crystal growth in this stacking step. TMA (trimethylaluminum) can be used as the Al raw material during production, TMG (trimethylgallium) can be used as the Ga raw material, and arsine (AsH ₃ ) can be used as the As raw material, and the growth temperature is 600 ° C. to 750 ° C. Can do. After the growth of each layer 1-9, the substrate is removed from the crystal growth furnace.

  Next, as shown in FIG. 15, a photoresist PR1 is applied to P-type AlGaAs (Al composition ratio X = 0.30) on the outermost surface, and the period Zp of the diffraction grating is set to about 250 nm by interference exposure. To form a diffraction grating made of photoresist PR1.

  Next, as shown in FIG. 16, the diffraction grating layer 9 is etched using the diffraction grating formed of the photoresist PR1 as a mask to remove the photoresist PR1. Thereby, the diffraction grating 9 is formed on the upper second cladding layer 8. For this etching, wet etching can be used. As an etchant in the case of wet etching, a Br-based etchant can be used.

  Next, as shown in FIG. 17, the substrate is introduced into an MOCVD crystal growth furnace, and a P-type AlGaAs (Al composition ratio X = 0.43) clad layer 10 and a P-type GaAs are formed on the diffraction grating 9. The contact layer 11 is grown sequentially. The raw materials and temperature during growth are as described above. That is, this manufacturing method includes a step of forming the upper third cladding layer 10 on the diffraction grating 9 and a step of forming the contact layer 11 on the upper third cladding layer 10.

Next, as shown in FIG. 18, a silicon nitride (SiN _x ) layer 13 is vapor-deposited on the eye of the contact layer 11, and subsequently, a photoresist PR2 is applied thereon, and the photoresist PR2 is subjected to lithography by exposure and development. The resist pattern having a width W (see FIG. 3) of about 5 mm is formed. Using the patterned photoresist PR2, the exposed region of the silicon nitride layer 13 is removed using a RIE (Reactive Ion Etching) apparatus.

  Next, as shown in FIG. 19, the photoresist PR2 is removed, and using the silicon nitride layer 13 as a mask, a citric acid-based etching solution (citric acid aqueous solution) is used, and the optical block layer (etch stop layer) 7 is clad. The layer and the diffraction grating layer are etched to form a ridge structure.

  More specifically, in this manufacturing method, a pattern (silicon nitride layer 13) extending in the resonance length direction is formed on the contact layer 11, and using this pattern as a mask, the contact layer 11, the upper third cladding layer 10, the diffraction The lattice 9 and the upper second cladding layer 8 are etched. As a result, the side surfaces of the contact layer 11, the upper third cladding layer 10, the diffraction grating 9, and the upper second cladding layer 8 are exposed, and the surface of the etch stop layer 7 is also exposed. In the case of this structure, since it is possible to perform wet etching using a single etching solution, manufacturing cost can be reduced.

  Next, as shown in FIG. 20, the substrate on which the ridge structure is formed is introduced into an MOCVD crystal growth furnace, and selective burying growth is performed to form a current confinement layer 12. The material of the current confinement layer 12 is n-type AlGaAs (Al composition ratio X = 0.50). That is, the current confinement layer 12 is formed on the exposed region on the etch stop layer 7, and the current confinement layer 12 forms the contact layer 11, the upper third cladding layer 10, the diffraction grating 9, and the upper second cladding layer 8. Cover the exposed sides. The upper second cladding layer 8, the diffraction grating 9, the upper third cladding layer 10, and the contact layer 11 all contain Ga and As, and the etching is performed with the same etching solution. This etching can be suitably performed when the Al composition ratio X is in the above range.

  Finally, the electrodes E1 and E1 are formed by vapor deposition, and the end face film R is formed on the back side of the laser, whereby the DFB semiconductor laser 100 shown in FIGS. 1 to 3 is completed.

  Note that in the formation of the above-described compound semiconductor layer, Si can be used as an N-type impurity and Zn can be used as a P-type impurity.

  FIG. 21 is a graph showing temperature characteristics of the DFB semiconductor laser described above. The horizontal axis indicates the temperature T (° C.), and the vertical axis indicates the optical output Po (mW) and the wavelength λ (nm) of the laser beam. The solid line indicates the optical output data, the dotted line indicates the wavelength data, A indicates the example shown in FIG. 4 (where the thickness of the lower cladding layer is 3 μm), and B indicates the data of the comparative example. The driving current of the example is 250 mA, and the driving current of the comparative example is 220 mA.

  In this comparative example, the thickness of the light guide layer is increased. The lower light guide layer has a thickness of 500 nm, and the upper light guide layer has a thickness of 150 nm. In this comparative example, the thickness of the lower cladding layer is 1.5 μm, the composition X is 0.43, the refractive index n is 3.3270, the composition ratio X of the lower and upper light guide layers is 0.387, and the refractive index is 3 3533, a light guide layer (composition X = 0.387, thickness 350 nm) is also inserted on the etch stop layer instead of the cladding layer 8, and the upper third cladding layer has a thickness of 1.05 μm and a composition ratio X = 0.43 cladding layer is introduced. The radiation angle θ in the vertical direction was 23.5 (deg) according to the calculation and 23.7 ± 0.2 (deg) according to the experiment.

  The temperature characteristic of the example has the advantage that the change rate of the light output is smaller than that of the comparative example. In addition, the vertical radiation angle θ of the example was 18 (deg), and the wavelength change rate in the example and the comparative example was +0.07 nm (common to both) on the average of the measurement range per 1 ° C., respectively. . Moreover, the temperature change rate of the optical output in an Example and a comparative example was 0.6 mW and 2.8 mW per 1 degreeC on the average of a measurement range, respectively. In the examples, the temperature from 20 ° C. to 70 ° C. is changed, and in the comparative example, the temperature from 0 ° C. to 50 ° C. is changed.

  In the comparative example of B, when the Al composition ratio of the light guide layer is 38.2%, the radiation angle θ in the vertical direction is 24.4 (deg) (calculation), 24.7 ± 0.2 ( deg) (experiment).

  FIG. 22 is a graph showing the relationship between the drive current If (mA) and the optical output Po (mW) of the DFB semiconductor laser according to the above-described embodiment. The differential value SE of the light output is 0.8 to 1.0 (W / A). It can be seen that the threshold current increases as the temperature increases.

  FIG. 23 shows the wavelength dependence of the laser according to the example, where the horizontal axis indicates the wavelength λ (nm) and the vertical axis indicates the optical output Po (mW) (logarithm). When the optical output is 50 mW, the central wavelength λ is 835.52 nm, FWHM is 0.114 (nm), and when the optical output is 100 mW, the central wavelength λ is 855.68 nm, FWHM is 0.11 (nm), and the optical output At 150 mW, the center wavelength λ is 835.84 nm, FWHM is 0.148 (nm), and when the optical output is 200 mW, the central wavelength λ is 836.32 nm, FWHM is 0.104 (nm), and the optical output is 250 mW. When the center wavelength λ is 836.48 nm, FWHM is 0.162 (nm), and the optical output is 300 mW, the center wavelength λ is 836.88 nm and FWHM is 0.144 (nm), and the wavelength variation is small. I understand that.

FIG. 24 is a graph showing the relationship between the laser beam intensity I (au) (normalized) and the radiation angle θ (deg) in the FFP in the case of the above-described embodiment (Y _G = 90 nm) (experimental value). . In the graph, FFP-H represents the light intensity distribution in the horizontal direction, and FFP-V represents the light intensity distribution in the vertical direction. The radiation angle θ giving the FWHM of the light intensity distribution in the vertical direction is 18.2 (deg), the radiation angle θ in the horizontal direction is 5.1 (deg), and the light output Po is 100 mW.

  FIG. 25 is a graph showing the light output Po (mW) and the light output differential value SE (W / A) with respect to the drive current If (mA) in the comparative example B (experimental value). The threshold value is higher than that of the present invention shown in FIG. That is, the present invention has an advantage that the oscillation threshold is lower (lower than 45 mA) than the comparative example. The dotted line indicates the characteristics when the element is placed in a cooling type package, and it can be seen that the linearity is improved. Note that. The data of the comparative example of the solid line is measured in a state where the element is housed in a metal package having a diameter of 9 mm.

DESCRIPTION OF SYMBOLS 1 ... Semiconductor substrate, 2 ... Lower clad layer, 3 ... Lower light guide layer, 4 ... Active layer, 5 ... Upper light guide layer, 6 ... Upper 1st clad layer, 7 ... optical block layer (etch stop layer), 8 ... upper second cladding layer, 9 ... diffraction grating, 10 ... upper third cladding layer, 11 ... contact layer (cap layer) .

Claims (3)

In the manufacturing method of the distributed feedback semiconductor laser,
A step of sequentially laminating a light guide layer, an upper first cladding layer, an etch stop layer, and an upper second cladding layer on the active layer;
Forming a diffraction grating on the upper second cladding layer;
Forming an upper third cladding layer on the diffraction grating;
Forming a contact layer on the upper third cladding layer;
A pattern extending in the resonance length direction is formed on the contact layer, and the contact layer, the upper third cladding layer, the diffraction grating, and the upper second cladding layer are etched using the pattern as a mask, and the contact Exposing a side surface of the layer, the upper third cladding layer, the diffraction grating and the upper second cladding layer and the surface of the etch stop layer;
A current confinement layer is formed on the exposed region on the etch stop layer, and the current confinement layer forms exposed side surfaces of the contact layer, the upper third clad layer, the diffraction grating, and the upper second clad layer. A covering step;
With
The upper second cladding layer, the diffraction grating, the upper third cladding layer, and the contact layer all contain Ga and As, and the etching is performed with the same etching solution. A method of manufacturing a distributed feedback semiconductor laser. The upper second cladding layer is made of Al _X Ga _1-X As (X = 0.43 ± 0.01),
The diffraction grating is made of Al _X Ga _1-X As (X = 0.30 ± 0.01),
The upper third cladding layer is made of Al _X Ga _1-X As (X = 0.43 ± 0.01),
The contact layer is made of GaAs;
The etch stop layer is made of Al _X Ga _1-X As (X = 0.70 ± 0.01).
The method of manufacturing a distributed feedback semiconductor laser according to claim 1. The light guide layer is an upper light guide layer,
Further comprising the step of forming the active layer on the lower light guide layer,
3. The distributed feedback type according to claim 1, wherein the lower and upper light guide layers are grown so that a total thickness of the lower light guide layer and the upper light guide layer is not less than 50 nm and not more than 200 nm. Semiconductor laser manufacturing method.

Images