JP5892534B2 - Semiconductor laser element - Google Patents

Description

  The present invention relates to a semiconductor laser device provided with a diffraction grating layer.

  Patent Document 1 discloses a distributed feedback semiconductor laser element (DFB-LD) including a diffraction grating. This document discloses that the saturation output of the laser beam can be increased by controlling the transverse mode along the width direction of the waveguide in the DFB-LD.

  Non-Patent Document 1 describes a surface emitting semiconductor laser device using a two-dimensional diffraction grating. The two-dimensional diffraction grating is called a photonic crystal (PC), and has a structure with periodic irregularities at intervals of about the wavelength of light. It has various functions such as localization of light and control of propagation direction and speed. Application becomes possible.

  A photonic crystal surface emitting laser (PCSEL) having a two-dimensional diffraction grating is a laser light source to which a PC is applied, and a standing wave generated by the PC formed in the plane is caused by a diffraction effect. This device is taken out in the direction perpendicular to the plane. In PCSEL, in principle, it is possible to increase the area of a PC while maintaining a single mode, and it is possible to achieve high output while maintaining high beam quality.

JP 2002-324948 A

K. Sakai et al, “Lasing Band-Edge Identification for a Surface-Emitting Photonic Crystal Laser” IEEE Journal On Selected Area In Communications, Vol. 23, No. 7, (2005) pp. 1335 S.Iwahashi et al, "Air-holedesign in vertical direction for high-power two-dimensional photonic-crystalsurface-emiting lasers" J.Opt., Soc. Am. B / Vol.27, No.6 / (2010) pp .1204

  However, from the viewpoint of the intensity and stability of the laser beam, a semiconductor laser element capable of outputting a more efficient laser beam is expected. The present invention has been made in view of such problems, and an object of the present invention is to provide a semiconductor laser element capable of outputting a highly efficient laser beam.

In order to solve the above-described problems, a semiconductor laser device according to the present invention is sandwiched between a lower clad layer made of a semiconductor, an upper clad layer made of a semiconductor, the lower clad layer, and the upper clad layer. A core layer comprising a plurality of stacked semiconductor layers and having an average refractive index higher than any of the lower cladding layer and the upper cladding layer, the core layer being adjacent to the guide layer, and a quantum well layer And a single diffraction grating layer separated from the active layer via a carrier block layer having an energy band gap larger than that of the guide layer , and in the core layer during operation. The electric field intensity distribution in the thickness direction is distributed in the first-order mode of TE-polarized light having two peaks, and the positions of the valleys between the peaks are the active layer and the Is set in the region between the diffraction grating layer, said core layer is between said active layer and said upper cladding layer, said upper cladding layer and is doped with impurities of the same conductivity type, the impurity concentration thereof is 1 A doped layer of × 10 ¹⁷ / cm ³ or more is further provided, and a part or all of the doped layer is located within a range of ± 50 nm from the position of the valley that is in the vicinity of the position of the valley between the two peaks. It is characterized by being set to .

In this case, although the electric field intensity distribution is in the core layer, the position of the valley is not the position of either the active layer or the diffraction grating layer. The laser light generation function can be sufficiently functioned, and the laser light can be generated with high efficiency.
In addition, when the positions of the peaks are set in the active layer and the diffraction grating layer, respectively, the electric field intensity at these positions can be increased, and high-intensity laser light can be generated. It becomes possible.

  Further, in the core layer, the optical confinement coefficient Γqw (1) in the active layer in the first-order mode of TE polarized wave that gives the two peaks, and the optical confinement coefficient Γqw (1 in the active layer in the zero-order mode) 0) preferably satisfies the relational expression Γqw (1)> Γqw (0). The light confinement coefficient Γ is a ratio (Γ = S ′ / S) of the intensity (S ′) in the target layer to the total intensity (S) of light. That is, the optical confinement factor can be increased by setting the two peak positions of the electric field intensity in the active layer and the diffraction grating layer as described above. In this case, it is possible to increase the electric field intensity at these positions, and it is possible to generate high-intensity laser light, and it is possible to obtain high-intensity laser light compared to the case where the 0th-order mode is mainly used. It becomes.

The core layer is a doped layer having an impurity concentration of 1 × 10 ¹⁷ / cm ³ or more added between the active layer and the upper clad layer with an impurity having the same conductivity type as the upper clad layer. Furthermore, it is preferable that a part or all of the doped layer is set in the vicinity of a valley position between the two peaks. “Nearby” means a region within ± 50 nm from the valley position. By providing the doped layer, the Fermi level changes to form a barrier, and carriers flowing from the lower cladding layer can be prevented from flowing toward the upper cladding layer, and the carrier concentration in the active layer is increased. be able to.

  Further, when the impurity concentration is high, energy of light and carriers is absorbed and loss occurs. The positions of the valleys of the two peaks in the first-order mode correspond to the peak positions in the zero-order mode. Accordingly, the provision of the doped layer causes optical loss at the peak position in the 0th-order mode, suppresses the occurrence of the 0th-order mode, stabilizes light emission by the first-order mode, and uses the first-order mode more effectively. be able to.

  The diffraction grating layer is made of a group III-V compound semiconductor, the group III element is selected from the group consisting of Ga, Al, and In, and the group V element is selected from the group consisting of As, P, N, and Sb. It is selected. That is, since a compound semiconductor including these elements can be a direct transition semiconductor, light can be easily emitted by carrier recombination.

  According to the semiconductor laser device of the present invention, it is possible to output a highly efficient laser beam.

It is a figure which shows the longitudinal cross-sectional structure of the semiconductor laser element which cut the semiconductor laser element shown in FIG. 3 along the II arrow. It is a graph which shows the structure and material of each layer of a semiconductor laser element. It is a top view of a semiconductor laser element. It is a figure which shows the horizontal cross section in a diffraction grating layer. It is a perspective view of a semiconductor laser element. It is a figure which shows the horizontal cross section in another diffraction grating layer. It is a figure for demonstrating the manufacturing method of a semiconductor laser element. It is a figure for demonstrating the manufacturing method of a semiconductor laser element. It is a figure for demonstrating the manufacturing method of a semiconductor laser element. It is a figure which shows the longitudinal cross-sectional structure of the semiconductor laser element which cut the semiconductor laser element shown in FIG. 11 along the XX arrow. It is a top view of a semiconductor laser element. It is a figure which shows the horizontal cross section in a diffraction grating layer. It is a perspective view of a semiconductor laser element. It is a figure which shows the horizontal cross section in another diffraction grating layer. It is a figure for demonstrating the manufacturing method of a semiconductor laser element. It is a figure for demonstrating the manufacturing method of a semiconductor laser element. It is a figure for demonstrating the manufacturing method of a semiconductor laser element. It is a figure which shows the longitudinal cross-section structure of an edge-emitting type semiconductor laser element. It is a graph which shows the structure and material of each layer of a semiconductor laser element. It is a graph which shows the detailed parameter of each layer of a semiconductor laser element. It is a perspective view of a semiconductor laser element. It is a figure which shows the electric field strength distribution along the thickness direction of a semiconductor laser element. The structure (FIG. 23 (a)), the energy band gap (FIG. 23 (b)), and the electric field intensity distribution (FIG. 23 (c): zero-order mode) along the thickness direction of the main part of the edge-emitting semiconductor laser element. FIG. 23 (d): primary mode). It is a figure for demonstrating the laminated structure in a core layer. It is a graph which shows the relationship between the normalized frequency V (V parameter) and the normalized propagation constant b. It is a graph which shows the relationship between the normalization frequency V with respect to a mode order, and the _cutoff film thickness _dcutoff (nm) of a core layer. 5 is a graph showing the relationship between the refractive index n and the electric field strength E along the thickness direction of the semiconductor laser element. Structure (FIG. 28 (a)), energy band gap (FIG. 28 (b)), and electric field intensity distribution (FIG. 28 (c): zero-order mode) along the thickness direction of the main part of the semiconductor laser device having a diffraction grating layer FIG. 28 (d): primary mode). It is a figure which shows the laminated structure in a core layer. It is a graph which shows the detailed parameter of each layer of a semiconductor laser element. It is a graph which shows the relationship between the equivalent refractive index and band edge wavelength (nm) in fundamental mode (0th-order mode) and 1st-order mode. It is a graph which shows the relationship between the wavelength in which resonance occurs, the emission wavelength, and the intensity. It is a graph which shows the optical confinement factor in each layer. 5 is a graph showing the relationship between the refractive index n and the electric field strength E along the thickness direction of the semiconductor laser element. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and an equivalent refractive index neff. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and optical confinement coefficient (GAMMA) qw. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and the optical confinement coefficient (GAMMA) g. It is a graph which shows minimum value Min (nm) and maximum value Max (nm) of thickness D of the diffraction grating layer which satisfy | fill various conditions. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and the optical confinement coefficient (GAMMA) dope. It is a graph which shows the detailed parameter of each layer of a semiconductor laser element. It is a graph which shows the relationship between the equivalent refractive index and band edge wavelength (nm) in fundamental mode (0th-order mode) and 1st-order mode. It is a graph which shows the optical confinement factor in each layer. 5 is a graph showing the relationship between the refractive index n and the electric field strength E along the thickness direction of the semiconductor laser element. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and an equivalent refractive index neff. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and optical confinement coefficient (GAMMA) qw. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and the optical confinement coefficient (GAMMA) g. It is a graph which shows minimum value Min (nm) and maximum value Max (nm) of thickness D of the diffraction grating layer which satisfy | fill various conditions. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and the optical confinement coefficient (GAMMA) dope. It is a graph which shows the detailed parameter of each layer of a semiconductor laser element. It is a graph which shows the relationship between the equivalent refractive index and band edge wavelength (nm) in fundamental mode (0th-order mode) and 1st-order mode. It is a graph which shows the optical confinement factor in each layer. 5 is a graph showing the relationship between the refractive index n and the electric field strength E along the thickness direction of the semiconductor laser element. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and an equivalent refractive index neff. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and optical confinement coefficient (GAMMA) qw. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and the optical confinement coefficient (GAMMA) g. It is a graph which shows minimum value Min (nm) and maximum value Max (nm) of thickness D of the diffraction grating layer which satisfy | fill various conditions. It is a graph which shows the relationship between the thickness D of a diffraction grating layer, and the optical confinement coefficient (GAMMA) dope. It is a graph which shows the minimum value Min (nm) and the maximum value Max (nm) of the thickness d of the core layer in which the secondary mode does not occur and the primary mode occurs. It is a graph which shows numerical formula. 5 is a chart showing a correlation among a lattice constant a, a refractive index n, an emission wavelength λ, a composition ratio (1-X) of an element having a large atomic radius, a composition ratio X of an element having a small atomic radius, and an energy band gap Eg.

  Hereinafter, the semiconductor laser device according to the embodiment will be described. In addition, the same code | symbol shall be used for the same element and the overlapping description is abbreviate | omitted.

  FIG. 1 is a diagram showing a vertical cross-sectional configuration of a semiconductor laser element obtained by cutting the semiconductor laser element shown in FIG. 3 along arrows II, and FIG. 2 is a chart and diagram showing the structure and material of each layer of the semiconductor laser element. 3 is a plan view of the semiconductor laser element.

  The semiconductor laser element shown in FIG. 1 is a photonic crystal surface emitting laser element (PCSEL) 10, and on a semiconductor substrate 1 made of a compound semiconductor, a lower clad layer 2, a guide layer 3A each made of a compound semiconductor, and an active layer. The layer 3B, the spacer layer 3C, the carrier block layer 3D, the diffraction grating layer 4, the buffer layer 4 ′, the upper cladding layer 5, and the contact layer 6 are sequentially stacked. The lower electrode E1 is provided in contact with the back surface of the semiconductor substrate 1, and the upper electrode E2 is provided in contact with the contact layer 6.

  The compound semiconductor layer is preferably made of a direct transition type III-V compound semiconductor. In such a III-V group compound semiconductor, the group III element can be selected from the group consisting of Ga, Al and In, and the group V element can be selected from the group consisting of As, P, N and Sb. Since a compound semiconductor including these elements can be a direct transition semiconductor, light can be easily emitted by carrier recombination inside the semiconductor.

  Carriers composed of electrons and holes can be injected into the semiconductor by applying a bias between the lower electrode E1 and the upper electrode E2. These injected carriers can be recombined in the active layer 3B.

  The active layer 3B is composed of a quantum well layer. The quantum well layer includes a single quantum well structure (SQW) having one well layer and a multiple quantum well structure (MQW) having two or more well layers. The quantum well structure uses two or more kinds of materials having different energy band gaps, and a thin film (well layer) having a small band gap is sandwiched between thin films (barrier layers) having a large material.

  The thicknesses of the well layer and the barrier layer are on the order of nm. Thereby, electrons or holes as carriers are confined in a layer (well) of a material having a small energy band gap. Carriers confined in the well layer are reduced in the degree of freedom in the direction perpendicular to the well layer, exhibit two-dimensionality, and enable stable light emission. Note that, when the number of well layers is increased, the optical confinement factor tends to increase, and from this viewpoint, MQW is preferable to SQW.

  Referring to FIG. 2, the basic structure of the semiconductor laser device 10 is obtained by laminating a lower clad layer B, a core layer A, an upper clad layer C, and a contact layer on a semiconductor substrate. The semiconductor substrate 1 is made of N-type GaAs, the lower clad layer B (lower clad layer 2) is made of N-type AlGaAs, and the upper clad layer C (upper clad layer 5) is made of P-type AlGaAs. The core layer A located between the clad layers B and C includes a guide layer 3A, an active layer 3B, a spacer layer 3C, a carrier block layer 3D, a diffraction grating layer 4, and a buffer layer 4 '.

  The material of the active layer 3B is made of a III-V compound semiconductor. In this example, it has a structure in which a plurality of InGaAs layers and AlGaAs layers are stacked. An InGaAs layer having a narrow energy band gap is used as the well layer, and AlGaAs having a wide energy band gap is used as the barrier layer.

  Both the guide layer 3A and the spacer layer 3C sandwiching the active layer 3B are made of AlGaAs. The guide layer 3A has a function of suppressing carriers overflowing from the active layer 3B and confining light in the active layer 3B, and has an energy band gap equal to or greater than that of the barrier layer of the active layer 3B. The spacer layer 3C is a layer for adjusting the distance from the carrier block layer 3D as necessary, and has the same energy band gap as that of the active layer 3B.

The lower cladding layer 2 is made of N-type AlGaAs, has an energy band gap larger than that of the barrier layer of the active layer 3B, and has a low refractive index. The upper cladding layer 5 is made of P-type AlGaAs, has an energy band gap larger than that of the barrier layer of the active layer 3B, and has a low refractive index. Examples of N-type impurities include Se and Si, and examples of P-type impurities include Zn, Mg, and C. The impurity concentration of the lower cladding layer 2 is 1 × 10 ¹⁷ / cm ³ to 1 × 10 ¹⁹ / cm ³ / cm ³ , and the impurity concentration of the upper cladding layer 5 is 1 × 10 ¹⁷ / cm ³ to 1 × 10 ¹⁹ / It is cm ³ / cm ³ , and the core layer A between them is basically added with no impurities, and even when added, it is set to less than 1 × 10 ¹⁷ / cm ³ except for a doped layer described later. can do. Actually, the conductivity type in the region between the P-type cladding layer 5 and the active layer 3B is a low-concentration P-type, and the conductivity in the region between the N-type cladding layer 2 and the active layer 3B. The mold is assumed to be a low concentration N type.

  In addition to the spacer layer 3C, a carrier block layer 3D, a diffraction grating layer 4, and a buffer layer 4 'are disposed between the active layer 3B and the upper cladding layer 5. The carrier block layer 3D and the buffer layer 4 ′ are made of AlGaAs, and the diffraction grating layer 4 is originally made of the basic layer 4A made of GaAs, and the different refractive index portion 4B having a different refractive index is formed inside the basic layer 4A. Are periodically embedded two-dimensionally to form a two-dimensional diffraction grating. The different refractive index portion 4B is made of AlGaAs, and is embedded in a plurality of holes formed in the basic layer 4A and formed as a buried layer in the initial stage when the buffer layer 4 'is formed. If the formation of the buffer layer 4 'is completed when the hole is filled, the buffer layer 4' formed closer to the upper electrode than the upper surface on the basic layer 4A does not exist. The buffer layer 4 ′ located above the diffraction grating layer 4 may be provided as necessary and can be omitted.

  The carrier block layer 3D has an energy band gap higher than that of the guide layer adjacent to the carrier block layer 3D, and thus the refractive index is set low. This can increase the tendency to confine carriers in the active layer 3B to the inside of the carrier block layer 3D, but the thickness is relatively small and does not have a function of greatly preventing light emission in the active layer 3B. . The carrier block layer 3D has a larger energy band gap and refractive index than the active layer 3B.

  Since the active layer 3B and the diffraction grating layer 4 have a refractive index distribution in the two-dimensional plane (YZ plane), in contrast to other layers, the in-plane average refractive index is the refractive index, The in-plane average energy band gap is defined as the energy band gap. In addition, since the plurality of layers constituting the core layer A are laminated in the thickness direction, the refractive index of the core layer A is distributed two-dimensionally as described above, with the average refractive index in the thickness direction as its refractive index. If there is, the average refractive index in the surface is obtained, and this is set as the average refractive index in the thickness direction. That is, the average refractive index of the core layer A is a value obtained by multiplying the thickness of each layer by the in-plane average refractive index for all layers, integrating the values, and dividing the integrated value by the total thickness. Given in. Also in the case of a clad layer, when there is a distribution in the thickness direction or in-plane direction, the average value is calculated by this method, and this is used as the refractive index of the clad layer, which is the equivalent refractive index. Also, in the case of the energy band gap, the energy band gap of each layer is obtained with the average value by the same method as the refractive index.

  In designing the energy band gap and the refractive index, these can be adjusted by changing the composition ratio of elements in each compound semiconductor layer. FIG. 60 shows the correlation among the lattice constant a, the refractive index n, the emission wavelength λ, the composition ratio (1-X) of the element having a large atomic radius, the composition ratio X of the element having a small atomic radius, and the energy band gap Eg. It is a chart which shows. The intersection value between the parameter indicated by the row and the parameter indicated by the column indicates the correlation between these parameters. When there is a positive correlation, “+” is indicated. When there is a negative correlation, “−” is indicated. Is described.

As a general property of a compound semiconductor, as the composition ratio of an element having a small atomic radius increases, for example, for Al _X Ga _1-X As, the lattice constant a tends to decrease as the Al composition ratio X increases. It is in. Further, when the lattice constant a is decreased, the refractive index n and the emission wavelength λ are decreased (positive correlation), while the energy band gap Eg is increased (negative correlation). That is, the lattice constant a, the refractive index n, the emission wavelength λ, and the composition ratio (1-X) of the elements having a large atomic radius have a positive correlation with each other, and they are in an energy band gap Eg or small atoms. The composition ratio X of the element of the radius tends to have a negative correlation.

  Therefore, in order to increase the energy band gap Eg, in AlGaAs, the composition ratio X of Al having a small atomic radius may be increased, and in order to decrease the energy band gap Eg, Al having a small atomic radius may be used. The composition ratio X may be reduced. Therefore, in the AlGaAs system, the energy band gap Eg becomes zero when the Al composition ratio X is decreased, that is, it becomes minimum in GaAs, but when the energy band gap Eg is further reduced, each of GaAs InGaAs having an atomic radius larger than that of the constituent elements may be mixed to form InGaAs. Since the energy band gap Eg and the refractive index n are generally negatively correlated, the refractive index adjustment may be designed according to the adjustment of the energy band gap Eg.

  For example, in the diffraction grating layer 4, a plurality of different refractive index portions 4B are embedded in the base basic layer 4A, but if the basic layer 4A is made of GaAs and the different refractive index portion 4B is made of AlGaAs, then Al The higher the composition ratio X is in the different refractive index portion 4B, the lower the refractive index of the different refractive index portion 4B. The larger this refractive index difference, the stronger the diffraction. The refractive index basically means the refractive index for light generated inside, but the refractive index at a wavelength of 0.9 μm at room temperature is n = 3.59 in the case of GaAs and in the case of AlAs. N = 2.97, and when GaAs and AlGaAs are used, a maximum refractive index difference of 0.62 can be obtained, and a diffraction grating having a sufficient diffraction effect can be manufactured. Of course, since there is a difference in the composition ratio X, it can function as a different refractive index portion. Therefore, both the basic layer 4A and the different refractive index portion 4B are made of AlGaAs, and the Al composition ratio X of the different refractive index portion 4B is changed to the basic layer 4A. By changing, it can function effectively as a diffraction grating also in this case.

  This semiconductor laser element is a surface emitting type, and a laser beam LB is emitted in the thickness direction of the active layer 3B. The horizontal cross section (YZ cross section) of the laser beam LB can take various shapes. In this example, the outer edge of the horizontal cross section of the laser beam LB is an ellipse, a circle, or a rectangle, as shown in FIG. Furthermore, in a plan view (a shape in the YZ plane as viewed from the X-axis direction), the region extends to an area outside the upper electrode E2. In plan view, the upper electrode E2 is set in the diffraction grating formation region RPC, and the area of the upper electrode E2 is smaller than the area of the diffraction grating formation region RPC.

  FIG. 4 is a diagram showing a horizontal section in the diffraction grating layer 4.

  A plurality of different refractive index portions 4B are two-dimensionally embedded in the diffraction grating formation region RPC in the basic layer 4A. In this example, the diffraction grating formation region RPC is set to a rectangle. The plurality of different refractive index portions 4B are aligned along the Y-axis direction and aligned along the Z-axis direction, and the gravity centers of the individual different refractive index portions 4B are regarded as lattice points and pass through the lattice points. When a lattice is set, this constitutes a rectangular lattice or a square lattice. As the type of the lattice, a triangular lattice, an orthorhombic lattice, a hexagonal lattice, or the like is applicable. Each of the different refractive index portions 4B has a polygonal column shape extending along the X axis that is the thickness direction of the basic layer 4A. The polygon here includes not only a triangle, a quadrangle, a pentagon, and a hexagon, but also an elliptical shape having infinite corners, and the elliptical shape includes a circular shape. In addition, although the shape of the different refractive index portion 4B of this example is circular in a plan view (YZ cross-sectional shape viewed from the X-axis direction), there is a slight manufacturing error in the actual manufacturing process, and the complete circular shape is It doesn't mean.

  FIG. 5 is a perspective view of the semiconductor laser device.

  The three-dimensional shape of the semiconductor laser element 10 is a rectangular parallelepiped, and emits laser light LB along the X-axis direction. When an orthorhombic semiconductor constituent material is used, the three-dimensional shape of the semiconductor laser element 10 may be a parallelepiped. As described above, the semiconductor laser device 10 includes the lower cladding layer 2, the light emitting layer 3, the diffraction grating layer 4, the buffer layer 4 ′, the upper cladding layer 5, and the contact layer 6 on the semiconductor substrate 1. The light emitting layer 3 is a semiconductor layer located between the lower cladding layer 2 and the diffraction grating layer 4. In this example, the light emitting layer 3 includes the guide layer 3A, the active layer 3B, the spacer layer 3C, It consists of a carrier block layer 3D. When a driving voltage is applied between the lower electrode E1 and the upper electrode E2 and a driving current is passed between them, carriers are concentrated in the active layer 3B in the light emitting layer 3, and electrons injected in the region Holes recombine and light emission occurs. This light emission resonates in the region between the clad layers, that is, in the core layer A including the light emitting layer 3 and the diffraction grating layer 4, and is output to the outside as the laser light LB. The diffraction grating layer 4 determines the wavelength of the resonating laser light.

In principle, light emission occurs in a path through which a drive current flows, that is, a region immediately below the upper electrode E2 in which the electric resistance is reduced. Since the upper electrode E2 is disposed on the upper surface of the semiconductor laser element 10, the laser beam LB is somewhat disturbed by the upper electrode E2. As a material for the upper electrode E2, a metal such as Au, Ag, Cu, Ni or Al, a mixture of these with a semiconductor such as Ge, or an impurity was added at a high concentration of 1 × 10 ¹⁹ / cm ³ or more. A compound semiconductor can be used. The material of the lower electrode E1 is the same, but a material that transmits the laser beam LB, that is, a transparent electrode such as ITO, can be used for the upper electrode E2.

  FIG. 6 is a diagram showing a horizontal cross section in another diffraction grating layer having a stripe pattern.

  The diffraction grating layer of FIG. 4 in the semiconductor laser element 10 can be replaced with that shown in FIG. A plurality of different refractive index portions 4B are embedded in the diffraction grating formation region RPC in the basic layer 4A. The diffraction grating formation region RPC is set to a rectangle. Each of the different refractive index portions 4B extends linearly along the Z-axis direction in plan view and is aligned along the Y-axis direction. Each of the different refractive index portions 4B has an aspect of the maximum dimension in the longitudinal direction with respect to the maximum dimension in the width direction, with the X axis being the thickness direction of the basic layer 4A as the depth direction, the Z axis as the longitudinal direction, and the Y axis as the width direction. It has a rectangular parallelepiped shape with a high ratio (2 or more). The different refractive index portion 4B is rectangular in a plan view, but may be other polygons such as a triangle, a pentagon, and a hexagon with a high aspect ratio. Polygon includes the meaning of an ellipse with infinite corners.

  7 to 9 are views for explaining a method of manufacturing the semiconductor laser element shown in FIG.

  First, an N-type (first conductivity type) semiconductor substrate (GaAs) 1 is prepared (FIG. 7A). Next, on the semiconductor substrate 1, an N-type lower cladding layer (AlGaAs) 2, a guide layer (AlGaAs) 3A, an active layer (MQW: InGaAs / AlGaAs) 3B, an optical guide layer (GaAs / AaGaAs) or a spacer layer ( AlGaAs) 3C, carrier block layer (AlGaAs) 3D, and basic layer (GaAs or AlGaAs) 4A serving as a diffraction grating layer are epitaxially grown sequentially using MOCVD (metal organic chemical vapor deposition) (FIG. 7B). . In the MOCVD method, the Al source can be supplied from TMA (trimethylaluminum), the Ga source can be supplied from TMG (trimethylgallium), and the In source can be supplied from TMI (trimethylindium).

Next, a mask layer FL1 made of SiN is formed on the basic layer 4A by plasma CVD (FIG. 7C). As a raw material gas for plasma CVD in forming SiN, SiH ₄ is used as a Si raw material, N ₂ , NH ₃ or the like is used as an N raw material, and the gas is diluted with H _{2 as} necessary. Further, a resist RG1 is applied on the mask layer FL1 (FIG. 7 (d)), a two-dimensional fine pattern is drawn with an electron beam drawing apparatus, and developed to form a two-dimensional (or one-dimensional) fine pattern on the resist RG1. (Corresponding to the position of the different refractive index portion) is formed (FIG. 7E). As a result, a plurality of holes H1 having a fine pattern are formed in the resist RG1, and each hole H1 reaches the surface of the mask layer FL1.

Next, the mask layer FL1 is etched using the resist RG1 as a mask, and the fine pattern of the resist is transferred to the mask layer FL1 (FIG. 7F). For this etching, reactive ion etching (RIE) can be used. As the SiN etching gas, a fluorine-based gas (CF ₄ , CHF ₃ , C ₂ F ₆ ) can generally be used. By this etching, holes H2 are formed in the mask layer FL1, and each hole H2 reaches the surface of the basic layer 4A.

  Next, the resist RG1 is removed by applying a stripping solution to the resist RG1 and further ashing the resist RG (FIG. 8G). As the ashing, photoexcitation ashing or plasma ashing can be used. As a result, only the mask layer FL1 having a fine pattern having a plurality of holes H3 remains on the basic layer 4A.

Thereafter, using the mask layer FL1 as a mask, the basic layer 4A is etched, and the fine pattern of the mask layer FL1 is transferred to the basic layer 4A (FIG. 8H). For this etching, dry etching is used. In dry etching, a chlorine-based or fluorine-based gas can be used as an etching gas. For example, Cl ₂ , SiCl _4, SF ₆ or the like can be used as a main etching gas, and Ar gas or the like mixed therein can be used. In addition to normal plasma etching, etching using inductively coupled plasma (ICP) can also be employed. By this etching, the depth of the hole H4 formed in the basic layer 4A is about 200 nm, and the depth of the hole H4 is smaller than the thickness of the basic layer 4A. The hole H4 may reach the surface of the semiconductor layer that is the base of the basic layer 4A.

Next, only the mask layer FL1 made of SiN is removed by reactive ion etching (RIE), and the opening end face of the hole H5 continuous to the hole H4 is exposed, that is, the surface of the basic layer 4A is exposed (FIG. 8 ( i)). As the SiN etching gas, a fluorine-based gas (CF ₄ , CHF ₃ , C ₂ F ₆ ) can be used as described above. Thereafter, surface treatment such as surface cleaning including thermal cleaning of the basic layer 4A is performed.

  Next, a different refractive index portion (buried layer) 4B is formed (regrown) in the hole H5 of the basic layer 4A by using the MOCVD method. In this regrowth process, AlGaAs is supplied to the surface of the basic layer 4A. The supplied AlGaAs has a different composition ratio between the basic layer 4A and Al. In the initial stage of regrowth, AlGaAs fills the hole H5 and becomes the different refractive index portion 4B. When the hole H5 is filled, AlGaAs supplied thereafter is laminated on the basic layer 4A as the buffer layer 4 '. Thereafter, a P-type (second conductivity type) clad layer (AlGaAs) 5 and a P-type contact layer (GaAs) 6 are sequentially grown on the buffer layer 4 ′ by MOCVD (FIG. 8J). )). The Al composition ratio X in the P-type cladding layer (AlGaAs) 5 is equal to or greater than the Al composition ratio X in the buffer layer 4 ′. In this example, these are equal, but X = 0.4 is adopted. Can do. Of course, the Al composition ratio X in the different refractive index portion 4B and the buffer layer 4 ′ is set to, for example, X = 0.35, and the Al composition ratio X is gradually increased with the growth, so that the Al composition in the upper cladding layer 5 is increased. The ratio X can also be set to X = 0.4. The crystal growth described above is all epitaxial growth, and the crystal axes of the respective semiconductor layers are coincident.

  Next, a resist RG2 is applied on the contact layer 6 (FIG. 8 (k)). The resist RG2 is exposed in a pattern so that a square opening is formed after the development process. That is, when a positive resist is used, the exposure area is square, and when a negative resist is used, the non-exposure area is square and the peripheral area is exposed. Thereafter, the resist RG2 is developed to form a square opening pattern at the center of the resist RG2. Electrode material E2 'is deposited on the exposed surfaces of resist RG2 and contact layer 6 using resist RG2 having the opening pattern as a mask (FIG. 8L). For the formation of this electrode material, an evaporation method or a sputtering method can be used. In this example, the evaporation method is used.

  Further, the resist RG2 is removed by lift-off, and a square electrode material is left on the contact layer 6 to form the upper electrode E2 (FIG. 9 (m)). Thereafter, the back surface of the N-type semiconductor substrate 1 is polished, and then the lower electrode E1 is formed on the entire polished back surface (FIG. 9 (n)). The electrode can be formed by vapor deposition or sputtering.

  In the case where a plurality of semiconductor laser elements are manufactured from a single wafer, it is necessary to form a groove for element isolation. This is done by applying a resist on the contact layer 6 and exposing and developing the resist. As a result, an opening pattern for element isolation is formed, and the contact layer 6 is wet etched using this resist to form a groove. The etching depth is about 10 μm. Thereafter, the resist may be removed using an organic solvent.

  In the embodiment, the electron beam exposure method has been described as a method for forming the hole H1, but other fine processing such as nanoimprint, interference exposure, focused ion beam (FIB), and optical exposure using a stepper is also used. Techniques can also be used. Furthermore, in this example, the pattern drawn on the resist is once transferred to the SiN, and then transferred to the substrate by dry etching. However, the pattern drawn on the resist is omitted as it is on the substrate by dry etching. It may be transferred. In the above description, the example in which one diffraction grating layer 4 is provided on the upper side of the active layer 3B has been described. However, this may be provided with a diffraction grating layer on the lower side of the active layer 3B. It can also be set as the structure provided with a diffraction grating layer on each of upper and lower sides.

  Next, a semiconductor laser device in which the structures of the lower electrode E1 and the upper electrode E2 are different from the above will be described.

  10 is a diagram showing a vertical cross-sectional configuration of the semiconductor laser element obtained by cutting the semiconductor laser element shown in FIG. 11 along arrows XX, FIG. 11 is a plan view of the semiconductor laser element, and FIG. 12 is a diffraction grating layer. FIG. 13 is a perspective view of the semiconductor laser device.

  The semiconductor laser device shown in FIG. 10 is a photonic crystal surface emitting laser device (PCSEL) 10, and on a semiconductor substrate 1, a lower cladding layer 2, a guide layer 3A, an active layer 3B, a spacer each made of a compound semiconductor. The layer 3C, the carrier block layer 3D, the diffraction grating layer 4, the buffer layer 4 ′, the upper clad layer 5, and the contact layer 6 are sequentially stacked. These structures and materials are the same as those of the semiconductor laser element described in FIG.

  There are only two differences from the one shown in FIG.

  The first difference is that a lower insulating layer F1 is formed on the back surface of the semiconductor substrate 1, and a lower electrode E1 is formed in an opening provided in the lower insulating layer F1. The formation region of the lower electrode E1 is not an entire region of the semiconductor substrate 1, but an annular region, and is disposed so as to surround the upper electrode E2 when viewed from the X-axis direction.

  The second difference is that an upper insulating layer F2 is formed on the contact layer 6, and an upper electrode E2 is formed in an opening provided in the upper insulating layer F2. The other configuration is the same as that shown in FIG.

  Note that the lower electrode E1 is in contact with the back surface of the semiconductor substrate 1 and the upper electrode E2 is in contact with the contact layer 6 as in FIG.

  In this semiconductor laser device, when a driving voltage is applied between the lower electrode E1 and the upper electrode E2 and a driving current is passed between them, carriers are concentrated in the active layer 3B in the light emitting layer 3, and the region is in this region. The electrons and holes injected in step recombine to generate light. This light emission resonates in the region between the clad layers, that is, in the core layer A including the light emitting layer 3 and the diffraction grating layer 4, and is output to the outside as the laser light LB. The diffraction grating layer 4 determines the wavelength of the resonating laser light.

  In principle, light emission occurs in a path through which a drive current flows, that is, in a region where electric resistance is reduced. Since the upper electrode E2 is disposed on the upper surface of the semiconductor laser element 10 and the annular lower electrode E1 is disposed on the rear surface, the light emitting region is a region between them. When a metal having high reflectance such as Al is used for the upper electrode E2, the laser beam LB traveling in the thickness direction of the semiconductor layer is reflected on the back surface of the upper electrode E2, and the laser beam LB is emitted from the opening of the lower electrode E1. (See FIGS. 10 and 13). The shape of the opening may be circular or rectangular. In order to enhance the effect of reflection of the light diffracted in the direction of the upper electrode E2 in the direction of the lower electrode E1, AlGaAs layers with different thicknesses of λ / 4 wavelength having different Al compositions are alternately stacked as the P clad layer material. A semiconductor multilayer film (referred to as DBR) can be introduced at an appropriate position.

  Also in the semiconductor laser element shown in FIGS. 10 to 13, the diffraction grating layer 4 having a striped planar pattern can be employed.

  FIG. 14 is a diagram showing a horizontal section in a diffraction grating layer having a stripe pattern.

  The diffraction grating layer of FIG. 12 in the semiconductor laser element 10 can be replaced with that shown in FIG. A plurality of different refractive index portions 4B are embedded in the diffraction grating formation region RPC in the basic layer 4A. The diffraction grating formation region RPC is set to a rectangle. Each of the different refractive index portions 4B extends linearly along the Z-axis direction in plan view and is aligned along the Y-axis direction. Each of the different refractive index portions 4B has an aspect of the maximum dimension in the longitudinal direction with respect to the maximum dimension in the width direction, with the X axis being the thickness direction of the basic layer 4A as the depth direction, the Z axis as the longitudinal direction, and the Y axis as the width direction. It has a rectangular parallelepiped shape with a high ratio (2 or more). The different refractive index portion 4B is rectangular in a plan view, but may be other polygons such as a triangle, a pentagon, and a hexagon with a high aspect ratio. Polygon includes the meaning of an ellipse with infinite corners.

  15 to 17 are views for explaining a method of manufacturing the semiconductor laser element shown in FIG.

  First, an N-type (first conductivity type) semiconductor substrate (GaAs) 1 is prepared (FIG. 15A). Next, on the semiconductor substrate 1, an N-type lower cladding layer (AlGaAs) 2, a guide layer (AlGaAs) 3A, an active layer (MQW: InGaAs / AlGaAs) 3B, an optical guide layer (GaAs / AaGaAs) or a spacer layer ( AlGaAs) 3C, carrier block layer (AlGaAs) 3D, and basic layer (GaAs or AlGaAs) 4A serving as a diffraction grating layer are epitaxially grown sequentially using MOCVD (metal organic chemical vapor deposition) (FIG. 15B). . In the MOCVD method, the Al source can be supplied from TMA (trimethylaluminum), the Ga source can be supplied from TMG (trimethylgallium), and the In source can be supplied from TMI (trimethylindium).

Next, a mask layer FL1 made of SiN is formed on the basic layer 4A by plasma CVD (FIG. 15C). As a raw material gas for plasma CVD in forming SiN, SiH ₄ is used as a Si raw material, N ₂ , NH ₃ or the like is used as an N raw material, and the gas is diluted with H _{2 as} necessary. Further, a resist RG1 is applied on the mask layer FL1 (FIG. 15 (d)), a two-dimensional fine pattern is drawn by an electron beam drawing apparatus, and developed to form a two-dimensional (or one-dimensional) fine pattern on the resist RG1. (Corresponding to the position of the different refractive index portion) is formed (FIG. 15E). As a result, a plurality of holes H1 having a fine pattern are formed in the resist RG1, and each hole H1 reaches the surface of the mask layer FL1.

Next, the mask layer FL1 is etched using the resist RG1 as a mask, and the fine pattern of the resist is transferred to the mask layer FL1 (FIG. 15F). For this etching, reactive ion etching (RIE) can be used. As the SiN etching gas, a fluorine-based gas (CF ₄ , CHF ₃ , C ₂ F ₆ ) can generally be used. By this etching, holes H2 are formed in the mask layer FL1, and each hole H2 reaches the surface of the basic layer 4A.

  Next, the resist RG1 is removed by applying a stripping solution to the resist RG1 and further ashing the resist RG (FIG. 16G). As the ashing, photoexcitation ashing or plasma ashing can be used. As a result, only the mask layer FL1 having a fine pattern having a plurality of holes H3 remains on the basic layer 4A.

Thereafter, using the mask layer FL1 as a mask, the basic layer 4A is etched, and the fine pattern of the mask layer FL1 is transferred to the basic layer 4A (FIG. 16H). For this etching, dry etching is used. In dry etching, a chlorine-based or fluorine-based gas can be used as an etching gas. For example, Cl ₂ , SiCl _4, SF ₆ or the like can be used as a main etching gas, and Ar gas or the like mixed therein can be used. In addition to normal plasma etching, etching using inductively coupled plasma (ICP) can also be employed. By this etching, the depth of the hole H4 formed in the basic layer 4A is about 200 nm, and the depth of the hole H4 is smaller than the thickness of the basic layer 4A. The hole H4 may reach the surface of the semiconductor layer that is the base of the basic layer 4A.

Next, only the mask layer FL1 made of SiN is removed by reactive ion etching (RIE), and the opening end face of the hole H5 continuous to the hole H4 is exposed, that is, the surface of the basic layer 4A is exposed (FIG. 16 ( i)). As the SiN etching gas, a fluorine-based gas (CF ₄ , CHF ₃ , C ₂ F ₆ ) can be used as described above. Thereafter, surface treatment such as surface cleaning including thermal cleaning of the basic layer 4A is performed.

  Next, a different refractive index portion (buried layer) 4B is formed (regrown) in the hole H5 of the basic layer 4A by using the MOCVD method. In this regrowth process, AlGaAs is supplied to the surface of the basic layer 4A. The supplied AlGaAs has a different composition ratio between the basic layer 4A and Al. In the initial stage of regrowth, AlGaAs fills the hole H5 and becomes the different refractive index portion 4B. When the hole H5 is filled, AlGaAs supplied thereafter is laminated on the basic layer 4A as the buffer layer 4 '. Thereafter, a P-type cladding layer (AlGaAs) 5 and a P-type contact layer (GaAs) 6 are sequentially grown on the buffer layer 4 'by MOCVD (FIG. 16J). The Al composition ratio X in the P-type cladding layer (AlGaAs) 5 is equal to or greater than the Al composition ratio X in the buffer layer 4 ′. In this example, these are equal, but X = 0.4 is adopted. Can do. Of course, the Al composition ratio X in the different refractive index portion 4B and the buffer layer 4 ′ is set to, for example, X = 0.35, and the Al composition ratio X is gradually increased with the growth, so that the Al composition in the upper cladding layer 5 is increased. The ratio X can also be set to X = 0.4. The crystal growth described above is all epitaxial growth, and the crystal axes of the respective semiconductor layers are coincident.

  In the case where a plurality of semiconductor laser elements are manufactured from a single wafer, it is necessary to form a groove for element isolation. This is done by applying a resist on the contact layer 6 and exposing and developing the resist. As a result, an opening pattern for element isolation is formed, and the contact layer 6 is wet etched using this resist to form a groove. The etching depth is about 10 μm. Thereafter, the resist is removed using an organic solvent.

Next, an upper insulating layer F2 made of SiN is formed on the contact layer 6 (FIG. 16 (k)). A plasma CVD method can be used as a method of forming the upper insulating layer F2. SiH ₄ is used as the Si raw material for plasma CVD, N ₂ , NH ₃ or the like is used as the N raw material, and the gas is diluted with H _{2 as} necessary.

Next, a resist is applied on the upper insulating layer F2, and this is exposed and developed to form a square opening pattern on the resist. Thereafter, the upper insulating layer F2 is formed using the opening pattern as a mask. Etching is performed to form an upper insulating layer F2 having an opening pattern (FIG. 16L). For the etching at this time, reactive ion etching (RIE) can be used. As the SiN etching gas, a fluorine-based gas (CF ₄ , CHF ₃ , C ₂ F ₆ ) can be used as described above.

  Next, a resist RG3 is applied on the exposed surfaces of the upper insulating layer F2 and the contact layer 6 (FIG. 17M), and a pattern exposure is performed so that a square opening is formed after the development process. When a positive resist is used, the exposure area is square, and when a negative resist is used, the non-exposure area is square and the peripheral area is exposed. Thereafter, the resist RG3 is developed to form a square opening pattern at the center of the resist RG3 (FIG. 17 (n)). The opening pattern in the upper insulating layer F2 is set inside the opening pattern of the resist RG3, and the area is smaller than the opening pattern of the resist RG3. Note that a negative type is used as the resist RG3 in this example.

  Using the resist RG3 having an opening pattern as a mask, an electrode material is deposited on the exposed surfaces of the resist RG3, the upper insulating layer F2, and the contact layer 6, and then the resist RG3 is lifted off so that the upper electrode E2 is brought into the opening pattern. It forms (FIG. 17 (o)). For the formation of this electrode material, an evaporation method or a sputtering method can be used. In this example, the evaporation method is used.

Thereafter, the back surface of the N-type semiconductor substrate 1 is mirror-polished by mechanical polishing and chemical mechanical polishing (CMP). Next, a lower insulating layer F1 made of SiN is deposited on the back surface of the semiconductor substrate 1 using a plasma CVD method. SiH ₄ is used as the Si raw material for plasma CVD, N ₂ , NH ₃ or the like is used as the N raw material, and the gas is diluted with H _{2 as} necessary. The optical film thickness of the lower insulating layer F1 is set to λ / 4 when the emission wavelength of the semiconductor laser element is λ, and the reflected light from the lower insulating film F1 has an opposite phase to the incident light and attenuates.

Next, after applying a resist on the lower insulating layer F1, exposure and development are performed so that an annular opening pattern is formed in the resist, and the lower insulating layer made of SiN is used with the annular opening pattern of the resist as a mask. F1 is etched, and then the resist is removed (FIG. 17 (p)). Thereby, the lower insulating layer F1 is processed to have an annular opening pattern. SiN can be etched using dry etching or RIE. As described above, chlorine-based (Cl ₂ , SiCl ₄ ) or fluorine-based (SF ₆ ) gas can be used as an etching gas for dry etching of SiN, and fluorine-based gas can be used as an etching gas for RIE. (CF ₄ , CHF ₃ , C ₂ F ₆ ) can be employed.

  Finally, after applying a resist on the lower insulating layer F1 having the annular opening pattern and the exposed back surface of the semiconductor substrate 1, an annular opening pattern having a larger width than the opening pattern of the lower insulating layer F1 is formed on the resist. Then, using the annular opening pattern of the resist as a mask, an electrode material is deposited on the resist and in the opening of the annular opening pattern, and then the resist is removed by lift-off (FIG. 17 (q)). . The electrode material can be formed by vapor deposition or sputtering.

  In addition, although the manufacturing method by the electron beam exposure method was described in the embodiment as the manufacturing method of the above-described hole (hole) H1, nanoimprint, interference exposure, focused ion beam (FIB), optical exposure using a stepper, etc. Other microfabrication techniques can also be used. Furthermore, in this example, the pattern drawn on the resist is once transferred to the SiN, and then transferred to the substrate by dry etching. However, the pattern drawn on the resist is omitted as it is on the substrate by dry etching. It may be transferred. In the above description, the example in which one diffraction grating layer 4 is provided on the upper side of the active layer 3B has been described. However, this may be provided with a diffraction grating layer on the lower side of the active layer 3B. It can also be set as the structure provided with a diffraction grating layer on each of upper and lower sides.

  Next, a method for designing the electric field intensity distribution of light propagating in the core layer located between the clad layers will be described by taking an edge-emitting normal semiconductor laser as an example.

  18 is a diagram showing a longitudinal sectional configuration of an edge-emitting semiconductor laser device, FIG. 19 is a chart showing the structure and material of each layer of the semiconductor laser device, and FIG. 20 is a detailed parameter of each layer of the semiconductor laser device. FIG. 21 and FIG. 21 are perspective views of the semiconductor laser device.

  A semiconductor laser element 10 shown in FIG. 18 is an edge-emitting laser element that does not include a diffraction grating layer. On the semiconductor substrate 1, a lower cladding layer 2, a guide layer 3A, an active layer 3B, each made of a compound semiconductor, The carrier block layer 3D, the guide layer 4G, the upper cladding layer 5, and the contact layer 6 are sequentially stacked.

  Referring to FIG. 19, the basic structure of the semiconductor laser device 10 is obtained by laminating a lower clad layer B, a core layer A, an upper clad layer C, and a contact layer on a semiconductor substrate. The semiconductor substrate 1 is made of N-type GaAs, the lower clad layer B (lower clad layer 2) is made of N-type AlGaAs, and the upper clad layer C (upper clad layer 5) is made of P-type AlGaAs. The core layer A located between the clad layers B and C includes a guide layer 3A, an active layer 3B, a carrier block layer 3D, and a guide layer 4G. This semiconductor laser device is different from the semiconductor laser device shown in FIG. 1 in that it does not have a spacer layer, a diffraction grating layer, and a buffer layer, and instead has a guide layer 4G, and the other points are the same.

Referring to FIG. 20, the structure of the core layer A and the cladding layers B and C that contribute to light propagation will be described in detail. The lower cladding layer 2 is made of N-type Al _X Ga _1-X As (X = 0.7). will guide layer 3A is made of _{Al X Ga 1-X As (} X = 0.1), the active layer 3B is _{InGaAs / Al X Ga 1-X} As multi-quantum well structure (X = 0.1) (MQW The carrier block layer 3D is made of Al _X Ga _1-X As (X = 0.4), the guide layer 4G is made of Al _X Ga _1-X As (X = 0.1), and the upper cladding layer 5 is made of _{Al X Ga 1-X As (} X = 0.4). The number of wells in MQW can be set to 1 or 2 or more.

  Regarding the thickness, preferably, for example, the thickness of the lower cladding layer 2 is 2000 nm, the thickness of the guide layer 3A is 50 nm, the thickness of the active layer 3B is 40 nm, the thickness of the carrier block layer 3D is 30 nm, and the thickness of the guide layer 4G is 420 nm. The thickness of ˜430 nm and the upper cladding layer 5 can be 2000 nm.

  As shown in FIG. 21, the edge-emitting semiconductor laser element 10 has a resonance length in the Z-axis direction, and is located at the light emitting end face (at the positive end of the Z-axis in the semiconductor laser element). The laser light reciprocates between the XY plane) and the light reflection end face (XY plane located at the negative end of the Z axis), and the laser light LB is emitted in the Z axis direction. The planar shape of the upper electrode E <b> 2 is a rectangle extending in parallel with the Z axis, and the lower electrode E <b> 1 on the back surface side is formed on the entire back surface of the semiconductor substrate 1. The upper electrode E2 extends from the light emitting end face to the light reflecting end face, and is configured such that carriers are supplied within the resonance length range in the Z-axis direction.

  FIG. 22 is a diagram showing an electric field intensity distribution along the thickness direction (X direction) of the semiconductor laser element.

  When the drive current is supplied between the lower electrode E1 and the upper electrode E2, a laser beam is generated inside the semiconductor laser device 10 and output to the outside. The intensity peak of this laser beam (the peak of the electric field strength) ) Exists in the core layer A.

  The center position in the thickness direction of the core layer A is the origin O, and the thickness of the core layer A is d1. In this case, the electric field intensity distribution in the core layer A (the electric field intensity Ey in the Y-axis direction in the TE mode) has a unimodal intensity peak in the vicinity of the origin O in the fundamental mode (0th order mode) (FIG. 22 (a)).

  On the other hand, when the center position in the thickness direction of the core layer A is the origin O and the thickness of the core layer A is d2 (d2> d1), the electric field strength distribution in the core layer changes to the primary mode. In this case, the electric field intensity distribution in the core layer A (the electric field intensity Ey in the Y-axis direction in the TE mode) has an intensity peak valley in the vicinity of the origin O in the primary mode, and two intensities on both sides thereof. It has a peak (FIG. 22 (b)). In the figure, the peak positions in the primary mode are set to positions of approximately + (d2) / 4 and − (d2) / 4.

  FIG. 23 shows the structure (FIG. 23 (a)), the energy band gap (FIG. 23 (b)), and the electric field strength distribution (FIG. 23 (c)) along the thickness direction of the main part of the edge-emitting semiconductor laser device. It is a figure which shows 0th mode, FIG.23 (d): primary mode). The electric field strength Ey indicates the electric field strength in the Y-axis direction in the TE mode.

  The average energy band gap Eg of the core layer A is smaller than the average energy band gap Eg of the cladding layers B and C, and the energy band gap Eg of the carrier block layer 3D is the maximum value of the energy band gap Eg in the active layer 3B. Bigger than. As shown in FIG. 60, since the energy band gap Eg and the refractive index n have a negative correlation, the average refractive index n of the core layer A is greater than the average refractive index n of the cladding layers B and C. The refractive index n of the carrier block layer 3D is smaller than the maximum value of the refractive index n in the active layer 3B. Since light propagates through a high refractive index, light propagates through the core layer A in the three-layer slab structure.

  In the case of the 0th-order mode (FIG. 23C), the electric field intensity Ey at the position of the active layer 3B is smaller than the electric field intensity Ey at the peak position, and the intensity of the laser beam is not high. On the other hand, in the case of the primary mode (FIG. 23 (d)), the active layer 3B is located at one peak position of the electric field intensity Ey, and the intensity of the laser light is increased.

  Since the primary mode is a higher-order mode of the fundamental mode, the electric field intensity distribution of the fundamental mode (zero-order mode) exists even when light propagation in the primary mode is performed.

  FIG. 24 is a view for explaining a laminated structure in the core layer A. FIG.

  When the core layer A includes a plurality of stacked layers (Layer 1 to Layer (i + 1)) (where i is a natural number), each of these layers has a dielectric constant (ε (1) to ε (i + 1)). And a film thickness (d (1) to d (i + 1)). In order to manufacture a waveguide structure in which a primary transverse mode occurs in the core layer A, it is necessary to satisfy several conditions. That is, the film thickness of the core layer A needs to have a refractive index (n) and a film thickness (d) such that the first-order higher mode exists as an eigenstate.

  Such a range of refractive index and film thickness is calculated by the following method.

The electromagnetic field distribution existing in the waveguide structure can be calculated by numerically solving Maxwell's wave equation under boundary conditions. The layer structure shown in FIG. 20 can be equivalently regarded as a two-layer asymmetric slab waveguide structure as shown in FIG. At this time, the dielectric constant and film thickness of each layer in the core layer A are defined as shown in FIG. In this case, an equivalent dielectric constant (average dielectric constant) ε _Eq is calculated as shown in Expression (1) shown in FIG.

Also, equation (2) shown in FIG. 59 represents the normalized frequency V (V parameter). In Equation (2), k ₀ = 2π / λ (λ is the laser beam wavelength), which indicates the propagation constant (wave number) of a plane wave in vacuum, and n1 is the average refractive index of the core layer A (the core layer A ( _Determined from the average dielectric constant ε _Eq ). The refractive index (n1) has a positive correlation with the square root of the dielectric constant (ε _Eq ). In equation (2), Δ represents the relative refractive index difference between the core layer A and the clad layer B, and d represents the film thickness of the core layer.

Incidentally, it is given by the relative refractive index difference _{^{Δ = (n 1 2 -n 2}} 2) / 2n 1 2. Here, the average refractive index of the core layer A and n _1, the average refractive index of the cladding layer B and n _2. Even when the average refractive indexes of the cladding layers B and C are different, the relative refractive index Δ is given by the above equation.

Note that the propagation mode is a set of light that propagates without being reflected while being reflected at a specific angle. From the one having a small incident angle, the zero-order mode, the first-order mode, the second-order mode, and the like. Said. When the angle (propagation angle) of light propagation with respect to the longitudinal direction (Z-axis) of the core layer A is θ, the propagation constant β = k ₀ n ₁ cos θ is given. Equation (3) shows the normalized propagation constant b, which is given by the refractive index n and the propagation constant β.

  Further, the V parameter can be expressed using the normalized propagation constant b, which is shown in the equation (4) in FIG. In Equation (4), a ′ is an asymmetric parameter, which is given by Equation (5), χi is given by Equation (6), N corresponds to the mode order, and π represents the circular ratio. . Further, ni (i = 2, 3) in the formula (6) indicates the average refractive index in the cladding layers B and C, respectively.

  As described above, there is a correlation between the V parameter and the normalized propagation constant b, and the normalized propagation constant b is given by the propagation constant β in the specific mode propagating at the propagation angle θ. Therefore, when the propagation mode changes, the normalized propagation constant b changes, and the V parameter that satisfies this changes. The number of modes in the three-layer asymmetric slab waveguide is characterized by the V parameter and the normalized propagation constant b.

FIG. 25 is a graph showing the relationship between the normalized frequency V (V parameter) and the normalized propagation constant b. This graph is obtained in the TE mode in the structure shown in FIG. The wavelength λ is 980 nm, the refractive index n ₁ of the core layer A is 3.45, the refractive index n ₂ of the lower cladding layer B is 3.26, and the refractive index n ₃ of the upper cladding layer C is 3.11.

  When the value of the normalized frequency V shown in FIG. 25 is in the range indicated by oblique lines (3.756 ≦ V ≦ 6.898), the normalized propagation constant b is N = 0, N in the equation (4) For = 1, we have two solutions. In other words, when the normalized frequency V is a value within the hatched range, it is possible to propagate the 0th-order mode light and the first-order mode light.

The normalized frequency V is proportional to the film thickness d of the core layer A. However, when the film thickness d is smaller than a predetermined value, the N-order mode is cut off and cannot exist. The film thickness d at this time is defined as a cut-off film thickness d _cutoff (nm).

FIG. 26 is a chart showing the relationship between the normalized frequency V with respect to the mode order (N = 0, 1, 2, 3) and the _cutoff film thickness d _cutoff (nm) of the core layer. The cut-off film thickness d _cutoff of the first-order mode calculated by this method is 515 (nm). That is, when the film thickness d is decreased from the state where the primary mode (N = 1) exists, the normalized frequency V decreases, and d _cutoff = 515 (nm) and V = 3.756. At that time, the normalized propagation constant b of the first order mode becomes 0, and when the film thickness d is smaller than this, only the 0th order mode (N = 0) exists. Similarly, the cut-off film thickness d _cutoff of the zeroth-order mode is 84 (nm), the cut-off film thickness of the second-order mode is d _cutoff = 946 (nm), and the cut-off film thickness of the third-order mode is d _cutoff = 1377 (nm). It is.

  In other words, the thickness d of the core layer A needs to be 515 nm or more and smaller than 946 nm in order to oscillate the primary mode laser beam and not to oscillate the secondary mode. .

The thickness d _cutoff of the core layer in which the primary mode is blocked was calculated by the step matrix method, which is the numerical calculation method of the Maxwell equation. As a result, the thickness d _{cutoff of the} primary mode was 545 nm. The error due to these calculation methods is as small as 6%. Therefore, even if this method is used, the ranges of the refractive index n ₁ and the film thickness d of the core layer A can be determined, and these parameters can also be set using the step matrix method.

  FIG. 27 is a graph showing the relationship between the refractive index n along the thickness direction of the semiconductor laser element and the electric field intensity E (electric field intensity Ey in the Y-axis direction). This graph is obtained by the staircase matrix method. In the figure, regions of the core layer A, the lower clad layer B, and the upper clad layer C are shown. The alternate long and short dash line indicates the refractive index n, the solid line indicates the electric field intensity distribution in the first-order mode, and the dotted line indicates the electric field intensity distribution in the zero-order mode. Regarding the electric field strength E, the calculation data curve in the first-order mode is shown slightly shifted upward from the calculation data curve in the zero-order mode.

  FIG. 27A shows data when the thickness d of the core layer A is 540 nm, FIG. 27B shows data when the thickness d of the core layer A is 545 nm, and FIG. The data when the thickness d of the core layer A is 550 nm is shown.

  The refractive index of the core layer A is higher than the refractive indexes of the cladding layers B and C, and the refractive index of the well layer in the active layer 3B is higher than the refractive index of the barrier layer adjacent thereto, and the carrier block layer 3D The refractive index is lower than the refractive index of the guide layer 4G adjacent thereto. As shown in FIG. 60, the relationship of the energy band gap Eg in each layer is opposite to the relationship of the refractive index n.

  As shown in the figure, when the thickness d of the core layer A is 545 nm or more ((b), (c)), the primary mode is present, and the position of the active layer 3B is aligned in the vicinity of one intensity peak position. Thus, the laser light intensity can be increased.

  This principle can be applied to the semiconductor laser device including the diffraction grating layer shown in FIGS. 1 to 17 as follows. The following semiconductor laser elements show modifications of the core layer in the semiconductor laser elements shown in FIG. 1 to FIG. 9 or FIG.

  FIG. 28 shows a structure (FIG. 28 (a)), an energy band gap (FIG. 28 (b)), and an electric field intensity distribution (FIG. 28 (c)) along the thickness direction of the main part of the semiconductor laser device having a diffraction grating layer. : 0th-order mode, FIG. 28 (d): first-order mode). The electric field strength Ey indicates the electric field strength in the Y-axis direction in the TE mode.

  The average energy band gap Eg of the core layer A is smaller than the average energy band gap Eg of the cladding layers B and C, and the energy band gap Eg of the carrier block layer 3D is the maximum value of the energy band gap Eg in the active layer 3B. Bigger than. As shown in FIG. 60, since the energy band gap Eg and the refractive index n have a negative correlation, the average refractive index n of the core layer A is greater than the average refractive index n of the cladding layers B and C. The refractive index n of the carrier block layer 3D is smaller than the maximum value of the refractive index n in the active layer 3B. Since light propagates through a high refractive index, light propagates through the core layer A in the three-layer slab structure.

  In the case of the 0th order mode (FIG. 28C), the electric field intensity Ey at the position of the active layer 3B is smaller than the electric field intensity Ey at the peak position, and the intensity of the laser beam is not high. On the other hand, in the case of the primary mode (FIG. 28 (d)), the active layer 3B is located at one peak position of the electric field strength Ey, and the diffraction grating layer 4 (4A, 4B) It is located at another peak position of the intensity Ey, and the intensity of the laser light is increased.

  Next, the laminated structure in the core layer A will be described.

  FIG. 29 is a diagram showing several stacked structures in the core layer A. FIG. Each of them shows a modification of the core layer A in the semiconductor laser element shown in FIGS. 1 to 9 or the semiconductor laser element shown in FIGS.

  FIG. 29A shows a core layer A of the type shown in FIGS. 1 and 28, and includes a guide layer 3A, an active layer 3B, a spacer layer 3C, a carrier block layer 3D, and a diffraction grating layer 4 (4A, 4B). They are sequentially stacked. Note that the buffer layer 4 ′ shown in FIG. 1 may be positioned on the diffraction grating layer 4.

  FIG. 29B shows a core layer A of a type having a doped layer, which is a guide layer 3A, an active layer 3B, a spacer layer 3C, a carrier block layer 3D, a guide layer 3E, a doped layer 3F, and a diffraction grating layer 4 (4A). , 4B) are sequentially laminated. Note that the buffer layer 4 ′ shown in FIG. 1 may be positioned on the diffraction grating layer 4. The structure of the spacer layer 3C is as shown in FIGS.

  FIG. 29 (c) shows a core layer A of the type obtained by removing the spacer layer 3C from the core layer A shown in FIG. 29 (b). The guide layer 3A, the active layer 3B, the carrier block layer 3D, the guide layer 3E, A doped layer 3F and a diffraction grating layer 4 (4A, 4B) are sequentially laminated. Note that the buffer layer 4 ′ shown in FIG. 1 may be positioned on the diffraction grating layer 4. Details of the core layer A having this structure will be described later with reference to FIG. 30 (structure (1)), FIG. 40 (structure (2)), and FIG. 49 (structure (3)).

  FIG. 29D shows a core layer A of a type in which a guide layer 3G is newly added to the core layer A shown in FIG. 29C. The guide layer 3A, the active layer 3B, the carrier block layer 3D, and the guide layer 3E, a doped layer 3F, a guide layer 3G, and a diffraction grating layer 4 (4A, 4B) are sequentially laminated. Note that the buffer layer 4 ′ shown in FIG. 1 may be positioned on the diffraction grating layer 4. The configuration of the guide layer 3G can be set similarly to the configuration of the guide layer 3E.

  FIG. 30 is a chart showing detailed parameters of each layer of the semiconductor laser device having the core layer A shown in FIG. 29C (referred to as structure (1)).

Referring to FIG. 30, the structure of the core layer A and the clad layers B and C that contribute to light propagation will be described in detail. The lower clad layer 2 is made of N-type Al _X Ga _1-X As (X = 0.7). becomes, the guide layer 3A is made of _{Al X Ga 1-X As (} X = 0.1), the active layer 3B is _{InGaAs / Al X Ga 1-X} As multi-quantum well structure (X = 0.1) (MQW The carrier block layer 3D is made of Al _X Ga _1-X As (X = 0.4), the guide layer 3E is made of Al _X Ga _1-X As (X = 0.1), and the doped layer 3F Is made of P-type Al _X Ga _1-X As (X = 0.4), and the diffraction grating layer 4 has a basic layer 4A made of Al _X Ga _1-X As (X = 0.1). rate unit 4B is a _{Al X Ga 1-X As (} X = 0.4), upper-cladding layer 5 l _X Ga _1-X As consists (X = 0.4). The number of wells in MQW can be set to 1 or 2 or more.

  Regarding the thickness, for example, the thickness of the lower cladding layer 2 is 2000 nm, the thickness of the guide layer 3A is 50 nm, the thickness of the active layer 3B is 40 nm, the thickness of the carrier block layer 3D is 30 nm, and the thickness of the guide layer 3E is 260 nm. The thickness of the doped layer 3F can be 40 nm, the thickness of the diffraction grating layer 4 can be 300 nm, and the thickness of the upper cladding layer 5 can be 2000 nm.

Further, the occupation area ratio (filling factor: FF) of the different refractive index portion 4B in the diffraction grating formation region RPC (see FIG. 4) of the diffraction grating layer 4 can be set to 20%. The occupied area ratio FF (%) is a ratio of the total area of the plurality of different refractive index portions 4B to the area of the diffraction grating formation region RPC in the horizontal cross section (YZ plane). In other words, the occupied area ratio FF (%) is a ratio of the area occupied by the different refractive index portion 4B per unit area. The shape of one different refractive index portion 4B is circular in this example, and the area of the circle is, for example, 18000 nm ² for a square lattice having a lattice constant of 300 nm, and the depth in the X-axis direction embedded in the basic layer 4A The thickness is assumed to be 300 nm. That is, there are 0.2 cm ² of different refractive index portions 4B per 1 cm ² .

  31 shows an equivalent refractive index neff of the TE mode in the fundamental mode (0th order mode) and the first order mode, and the band edge of the photonic band gap in the diffraction grating layer 4 in the semiconductor laser device having the structure shown in FIG. It is a graph which shows the relationship of a wavelength (nm). The oscillation wavelength λ is 980 nm. The thickness D of the diffraction grating layer 4 is 300 nm (0.3 μm), and the thickness d of the core layer A is 720 nm. The number of quantum wells in the active layer 3B is 2.

The refractive index n ₁ of the core layer A = 3.42, the refractive index n ₂ of the lower cladding layer B = 3.10, and the refractive index n ₃ of the upper cladding layer C = 3.26. As for these refractive indexes, the value of the refractive index at a wavelength of 980 nm was obtained by MSEO (Modified Single Effective Oscillator) method and Equation (1). In this case, the cut-off thickness _{d cutoff} the primary mode to the propagation mode occurs _(1), the cut-off thickness _{d cutoff} the second-order mode occurs ₍₂₎ is 573 and (nm) 1047 and (nm). That is, if the thickness d of the core layer A is not less than 573 (nm) and less than 1047 (nm), the condition that the primary mode is generated and the secondary mode is not generated is satisfied. The range of the region taking the first-order higher-order mode is calculated as 600 (nm) to 1040 (nm) from the electromagnetic field calculation by the staircase matrix method, and these ranges are almost the same.

  Since the total film thickness d of the core layer A is 720 nm as described above, the primary mode generation condition is satisfied, and the TE mode electromagnetic field mode includes the fundamental 0th mode and the primary mode.

  From the difference in the equivalent refractive index neff shown in FIG. 31, when a triangular lattice photonic crystal having a lattice constant of 344 nm is used for the diffraction grating layer 4, the band edge wavelengths (1009.03 nm, 980. 43 nm) difference is about 30 nm.

  FIG. 32 is a graph showing the relationship (a) between the wavelength and intensity at which resonance occurs, and the relationship (b) between the emission wavelength and intensity.

  As described above, there is a difference of about 30 nm between the band edge wavelength λp0 in the 0th order mode of the photonic bandgap of the diffraction grating layer 4 and the band edge wavelength λp1 in the 1st order mode. Resonance of light emission for generating laser light can be generated at each wavelength. On the other hand, the peak wavelength λp of the gain spectrum of light emission in the active layer 3B made of a quantum well is the composition and energy band of the active layer 3B. It can be moved by changing the gap or the like. Therefore, if the peak wavelength λp of the gain spectrum and the band edge wavelength λp1 (980 nm) of the first-order mode are matched, the first-order mode laser oscillation becomes dominant and the zero-order mode laser oscillation is suppressed. Become. In the above case, since the band edge wavelength difference is about 30 nm, if the band edge wavelength λp1 of the primary mode exists within the range of the peak wavelength λp ± 5 nm of the gain spectrum, the primary mode is selective and efficient. Will occur.

  FIG. 33 is a chart showing optical confinement factors in each layer.

  In the case of the setting shown in FIG. 30, the optical confinement coefficient Γqw in the active layer (quantum well layer) 3B is 1.62% in the case of the fundamental mode (0th order mode) and 2 in the case of the primary mode. In the case of the primary mode, more light can be confined in the active layer 3B. The optical confinement coefficient Γg in the diffraction grating door layer 4 is 30.1 (%) in the case of the fundamental mode (0th order mode) and 39.3 (%) in the case of the first order mode. In the mode, more light can be confined in the diffraction grating layer 4. Further, the optical confinement coefficient Γdope in the doped layer 3F is 6.7 (%) in the case of the fundamental mode (0th-order mode) and 0.2 (%) in the case of the first-order mode. In this case, the influence of the light in the doped layer 3F is further suppressed, and the 0th-order mode light is lost by the impurities in the doped layer 3F.

  As described above, in the first-order mode, since the optical confinement coefficient Γqw in the active layer 3B is high, there is an advantage that laser oscillation with a low threshold is possible.

  Further, the optical confinement coefficient Γg in the diffraction grating layer 4 in the first order mode is higher than that in the fundamental 0th order mode. The optical confinement coefficient Γg is proportional to the coupling coefficient κ representing the magnitude of the diffraction effect in the direction perpendicular to the plane, and the diffraction grating layer 4 functions sufficiently. Note that the value of the optical confinement factor Γg = 39.3% is sufficiently high even when compared with the optical confinement factor Γg = 18.1% of the semiconductor laser element of Non-Patent Document 1.

Further, P-type impurities of 1 × 10 ¹⁷ / cm ³ or more are added to the doped layer 3F. In this example, the P-type impurity concentration of the doped layer 3F is 1 × 10 ¹⁸ / cm ³ . As a result, the Fermi level of the doped layer 3F can be shifted to the valence band side, and electrons injected from the N-type lower cladding layer 2 side can be blocked efficiently, and holes can be formed without resistance in the active layer 3B. Can be injected into the inside. In addition, the optical confinement coefficient Γdope of the doped layer 3F is a very small value of 0.2 (%) in the first-order mode as compared with 6.7 (%) in the basic 0th-order mode. Due to this difference, the optical loss is selectively increased in the fundamental 0th order mode, so that the primary mode can oscillate in a single mode more stably. In addition, it is preferable that the P-type impurity concentration of the doped layer 3F is 1 × 10 ²⁰ / cm ³ or less from the viewpoint of increase in light absorption due to diffusion of the dopant and crystallinity.

  Moreover, in the diffraction grating layer 4, since the hole for embedding the different refractive index part 3B is formed in advance by dry etching, processing damage remains on the bottom surface of the hole. In other words, a damaged region due to processing is formed in a region of about several tens of nanometers from the bottom of the basic layer 4A constituting the diffraction grating layer 4. In the damaged region, lattice defects, impurity levels, and the like are formed, which can cause light absorption. Therefore, if the electric field strength is reduced in the vicinity of the bottom of the diffraction grating layer 4 (or in the vicinity of the bottom of the hole), the influence of the light absorption can be reduced.

  In the 0th-order mode, it is difficult to reduce the electric field strength near the bottom or bottom. This is because the zero-order mode is basically unimodal, and in order to increase both the optical confinement coefficient Γqw of the active layer 3B and the optical confinement coefficient Γg of the diffraction grating layer 4, the bottom of the diffraction grating in the middle thereof This is because the electric field strength cannot be reduced.

  On the other hand, when the primary mode is used, as shown in FIG. 34, the position of the valley between the two intensity peaks in the primary mode is set in the vicinity of the doped layer 3F, so that the diffraction grating positioned above this is set. Also at the bottom of the layer 4, the electric field strength can be reduced and light absorption in the damaged region can be reduced. FIG. 34 is a graph showing the relationship between the refractive index n and the electric field strength E along the thickness direction of the semiconductor laser element, and the primary mode data is shown slightly shifted from the fundamental mode data. It is.

  It should be noted that the electric field strength can be zero at the position of the valley (node) of the electric field strength. For this reason, the node of the electric field strength can be arranged in the damaged region at the bottom of the diffraction grating layer, and the influence of the damaged region can be suppressed. Further, since the active layer 3B layer can be disposed away from the bottom of the diffraction grating layer 4 as compared with the 0th mode, the influence of direct processing damage to the active layer 3B can be suppressed.

  From this point of view, when considering the positional relationship along the X axis, the end surface (bottom surface) of the diffraction grating layer 4 on the active layer 3B side is within a range of ± 50 nm from the position of the valley of the electric field strength peak in the primary mode. The center position in the thickness direction of the doped layer 3F is preferably located within a range of ± 50 nm from the position of the valley of the electric field intensity peak in the primary mode. In order to efficiently use the two intensity peaks of the first mode, the end face (bottom face) of the diffraction grating layer 4 on the active layer 3B side is from the center position of the active layer 3B with respect to the oscillation wavelength λ. , Λ / (4 × neff) nm is desirable, but in order to exert the influence of the diffraction grating layer 4 on the active layer 3B, the distance between these center positions is λ / (2 × neff) nm. It is preferable that it is about.

In addition, since the electric field is more widely distributed in the primary mode than in the 0th mode, the optical confinement coefficient of the upper cladding layer 5 is increased, and the internal loss is increased in the wavelength region of about λ = 1 μm due to the influence of free electron absorption. There are also factors. Actually, since the optical confinement coefficient of the entire upper cladding layer 5 is 25.0%, the absorption loss can be increased. However, in the first-order mode, most of the light is confined in the vicinity of the diffraction grating layer 4, and 75% of the entire upper cladding layer is concentrated in a region 200 nm from the end surface of the diffraction grating layer 4 on the upper cladding layer 5 side. Therefore, a laser structure with a small internal loss can be realized by reducing the P-type impurity concentration in the region within the range of 1 × 10 ^{16 to} 3 × 10 ¹⁷ cm ⁻³ .

  FIG. 35 is a graph showing the relationship between the thickness D of the diffraction grating layer and the equivalent refractive index neff. As shown in FIGS. 30 and 31, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the equivalent refractive index neff is 3.3870 in the 0th-order mode and 3 in the first-order mode. 2910. When the occupied area ratio FF is 10%, 20%, or 30%, the equivalent refractive index neff is smaller in the first-order mode than in the zero-order mode, and the thickness of the diffraction grating layer D It can be seen that the equivalent refractive index neff increases as increases.

  FIG. 36 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γqw in the active layer 3B. As shown in FIGS. 30 and 33, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γqw in the active layer 3B is 1.62 (% ) 2.28 (%) in the primary mode. The optical confinement coefficient Γqw in the first-order mode is larger than the optical confinement coefficient Γqw in the 0th-order mode (condition A) when the thickness D is a predetermined value or more, and at least the thickness D is up to 600 nm. This magnitude relationship is maintained. The minimum value (predetermined value) and maximum value of the thickness D that satisfy the condition A are as shown in FIG.

  FIG. 37 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γg. As shown in FIGS. 30 and 33, when the occupied area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γg in the diffraction grating layer 4 is 30.1 ( %) 39.3 (%) in the primary mode. The optical confinement factor Γg in the first-order mode is larger than the optical confinement factor Γg in the zero-order mode (referred to as condition B) when the thickness D is within a predetermined range. A predetermined range (minimum value and maximum value) of the thickness D satisfying the condition B is as shown in FIG.

  FIG. 38 is a chart showing the minimum value Min (nm) and the maximum value Max (nm) of the thickness D of the diffraction grating layer that satisfies the conditions A and B in the structure (1).

  When the occupied area ratio FF is 10%, in order to satisfy the condition A, the thickness D needs to be 180 nm or more, and this condition is satisfied if it is 600 nm or less. In order to satisfy the condition B, the thickness D needs to be 200 nm or more and 300 nm or less. That is, in order to satisfy both conditions, the thickness D needs to be 200 nm or more and 300 nm or less.

  When the occupied area ratio FF is 20%, in order to satisfy the condition A, the thickness D needs to be 200 nm or more, and if it is 600 nm or less, this condition is satisfied. In order to satisfy the condition B, the thickness D needs to be 200 nm or more and 450 nm or less. That is, in order to satisfy both conditions, the thickness D needs to be 200 nm or more and 450 nm or less.

  When the occupation area ratio FF is 30%, in order to satisfy the condition A, the thickness D needs to be 270 nm or more, and if it is 600 nm or less, this condition is satisfied. In order to satisfy the condition B, the thickness D needs to be 200 nm or more and 600 nm or less. That is, in order to satisfy both conditions, the thickness D needs to be 270 nm or more and 600 nm or less.

  When the occupation area ratio FF is 10% to 30%, when the thickness D is 200 nm or more and 300 nm or less, both conditions are always satisfied.

  FIG. 39 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γdope of the doped layer. As shown in FIGS. 30 and 33, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γdope in the doped layer 3F is 6.7 (%) in the 0th-order mode. ) 0.2 (%) in the primary mode. The optical confinement coefficient Γdope in the first-order mode becomes smaller than the optical confinement coefficient Γdope in the 0th-order mode (condition C) at least in the thickness D of 150 nm to 600 nm.

  FIG. 40 is a chart showing detailed parameters of each layer of the semiconductor laser device having another structure having the core layer A shown in FIG. 29C (referred to as structure (2)).

  The difference between the structure shown in FIG. 40 and that shown in FIG. 30 is only the composition ratio X and the film thickness of aluminum, and the other basic structures are the same. The details will be described below.

Referring to FIG. 40, the structure of the core layer A and the clad layers B and C that contribute to the light propagation will be described in detail. The lower clad layer 2 is made of N-type Al _X Ga _1-X As (X = 0.7). becomes, the guide layer 3A is made of _{Al X Ga 1-X As (} X = 0.25), the active layer 3B is _{InGaAs / Al X Ga 1-X} As multi-quantum well structure (X = 0.25) (MQW The carrier block layer 3D is made of Al _X Ga _1-X As (X = 0.7), the guide layer 3E is made of Al _X Ga _1-X As (X = 0.25), and the doped layer 3F Is made of P-type Al _X Ga _1-X As (X = 0.7), and the diffraction grating layer 4 has a basic layer 4A made of Al _X Ga _1-X As (X = 0.25). rate portion 4B consists _{Al X Ga 1-X As (} X = 0.6), the upper cladding Layer 5 is made of _{Al X Ga 1-X As (} X = 0.6). The number of wells in MQW can be set to 1 or 2 or more.

  Regarding the thickness, for example, the thickness of the lower cladding layer 2 is 2000 nm, the thickness of the guide layer 3A is 50 nm, the thickness of the active layer 3B is 40 nm, the thickness of the carrier block layer 3D is 30 nm, and the thickness of the guide layer 3E is 100 nm. The thickness of the doped layer 3F can be 30 nm, the thickness of the diffraction grating layer 4 can be 300 nm, and the thickness of the upper cladding layer 5 can be 2000 nm.

Further, the occupation area ratio (filling factor: FF) of the different refractive index portion 4B in the diffraction grating formation region RPC (see FIG. 4) of the diffraction grating layer 4 can be set to 20%. The shape of one of the different refractive index portions 4B is also circular in this example, and the area of the circle is, for example, 18000 nm ² for a square lattice with a lattice constant of 300 nm, and the depth in the X-axis direction embedded in the basic layer 4A The thickness is assumed to be 300 nm. That is, there are 0.2 cm ² of different refractive index portions 4B per 1 cm ² .

  41 shows the equivalent refractive index neff of the TE mode in the fundamental mode (0th order mode) and the first order mode, and the band edge of the photonic band gap in the diffraction grating layer 4 in the semiconductor laser device having the structure shown in FIG. It is a graph which shows the relationship of a wavelength (nm). The oscillation wavelength λ is 800 nm. The thickness D of the diffraction grating layer 4 is 300 nm (0.3 μm), the thickness d of the core layer A is 550 nm, and the thickness excluding the diffraction grating layer 4 is 250 nm. The number of quantum wells in the active layer 3B is 2.

The refractive index n ₁ of the core layer A = 3.42, the refractive index n ₂ of the lower cladding layer B = 3.19, and the refractive index n ₃ of the upper cladding layer C = 3.25. For these refractive indexes, the value of the refractive index at a wavelength of 980 nm was determined by the MSEO method and the formula (1). In this case, the cut-off thickness _{d cutoff} the first-order mode occurs in propagation mode _(1), second order mode occurs cutoff thickness _{d cutoff (2)} becomes 428 and (nm) 800 and (nm). That is, if the thickness d of the core layer A is not less than 428 (nm) and less than 800 (nm), the condition that the primary mode is generated and the secondary mode is not generated is satisfied. The region range where the first-order higher-order mode is taken is calculated as 430 (nm) to 800 (nm) from the electromagnetic field calculation by the staircase matrix method.

  As described above, since the total film thickness d of the core layer A is 550 nm, the conditions for generating the first-order mode are satisfied, and the fundamental zero-order mode and the first-order mode are formed in the electromagnetic field mode of the TE mode.

  From the difference in the equivalent refractive index neff shown in FIG. 41, when a triangular lattice photonic crystal having a lattice constant of 281 nm is used for the diffraction grating layer 4, the band edge wavelengths (821.51 nm, 799. 04 nm) difference is about 20 nm. Therefore, as in the case of the structure (1) (see FIG. 32), if the peak wavelength λp of the gain spectrum is matched with the band edge wavelength λp1 (799 nm) of the primary mode, the laser oscillation of the primary mode is dominant. Thus, laser oscillation in the 0th order mode is suppressed. In the above case, since the band edge wavelength difference is about 20 nm, if the band edge wavelength λp1 of the primary mode exists within the range of the peak wavelength λp ± 5 nm of the gain spectrum, the primary mode is selective and efficient. Will occur.

  FIG. 42 is a chart showing optical confinement factors in each layer.

  40, the optical confinement coefficient Γqw in the active layer (quantum well layer) 3B is 2.67 (%) in the fundamental mode (0th order mode), and is 4. in the primary mode. 15 (%), and more light can be confined in the active layer 3B in the primary mode. The optical confinement coefficient Γg in the diffraction grating door layer 4 is 50.1 (%) in the fundamental mode (0th order mode) and 33.8 (%) in the 1st order mode. In the mode, more light can be confined in the diffraction grating layer 4. Furthermore, the optical confinement coefficient Γdope in the doped layer 3F is 6.3 (%) in the case of the fundamental mode (0th order mode) and 0.5 (%) in the case of the first order mode. In this case, the influence of the light in the doped layer 3F is further suppressed, and the 0th-order mode light is lost by the impurities in the doped layer 3F.

  As described above, in the first-order mode, since the optical confinement coefficient Γqw in the active layer 3B is high, there is an advantage that laser oscillation with a low threshold is possible.

  As in the case of the structure (1), the optical confinement coefficient Γg in the diffraction grating layer 4 in the first-order mode is not so low as compared with the fundamental zero-order mode, and the diffraction grating layer 4 functions sufficiently. It will be.

Further, P-type impurities of 1 × 10 ¹⁷ / cm ³ or more are added to the doped layer 3F. In this example, the P-type impurity concentration of the doped layer 3F is 1 × 10 ¹⁸ / cm ³ . Similar to the structure (1), holes can be injected into the active layer 3B without resistance while efficiently blocking electrons injected from the N-type lower cladding layer 2 side. Since the light loss selectively increases in the next mode, the primary mode can oscillate in a single mode more stably. In addition, it is preferable that the P-type impurity concentration of the doped layer 3F is 1 × 10 ²⁰ / cm ³ or less from the viewpoint of increase in light absorption due to diffusion of the dopant and crystallinity.

  Similarly to the case of the structure (1), when the primary mode is used, as shown in FIG. 43, the position of the valley between the two intensity peaks in the primary mode is set in the vicinity of the doped layer 3F. Thus, even at the bottom of the diffraction grating layer 4 positioned above, it is possible to reduce the electric field strength and reduce light absorption in the damaged region at the time of hole formation in the diffraction grating layer 4. FIG. 43 is a graph showing the relationship between the refractive index n and the electric field strength E along the thickness direction of the semiconductor laser element, and the primary mode data is shown slightly shifted from the fundamental mode data. It is.

  It should be noted that the electric field strength can be zero at the position of the valley (node) of the electric field strength. For this reason, the node of the electric field strength can be arranged in the damaged region at the bottom of the diffraction grating layer, and the influence of the damaged region can be suppressed. Further, since the active layer 3B layer can be disposed away from the bottom of the diffraction grating layer 4 as compared with the 0th mode, the influence of direct processing damage to the active layer 3B can be suppressed.

  From this point of view, when considering the positional relationship along the X axis, the end surface (bottom surface) of the diffraction grating layer 4 on the active layer 3B side is within a range of ± 50 nm from the position of the valley of the electric field strength peak in the primary mode. The center position in the thickness direction of the doped layer 3F is preferably located within a range of ± 50 nm from the position of the valley of the electric field intensity peak in the primary mode. In order to efficiently use the two intensity peaks of the first mode, the end face (bottom face) of the diffraction grating layer 4 on the active layer 3B side is λ / (4 × neff) from the center position of the active layer 3B. Although it is desirable that the distance is about nm, in order to influence the diffraction grating layer 4 on the active layer 3B, the distance between the center positions is preferably about λ / (2 × neff) nm.

In order to reduce the factor of increasing the internal loss due to the influence of free electron absorption in the wavelength range of about λ = 1 μm, a P-type impurity in a region of 200 nm or less from the end surface on the upper cladding layer 5 side of the diffraction grating layer 4 is used. By reducing the concentration within the range of 1 × 10 ¹⁶ / cm ^{3 to} 3 × 10 ¹⁷ / cm ^3, a laser structure with small internal loss can be realized.

  FIG. 44 is a graph showing the relationship between the thickness D of the diffraction grating layer and the equivalent refractive index neff. As shown in FIGS. 40 and 41, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the equivalent refractive index neff is 3.3758 in the 0th-order mode and 3 in the first-order mode. 2835. When the occupied area ratio FF is 10%, 20%, or 30%, the equivalent refractive index neff is smaller in the first-order mode than in the zero-order mode, and the thickness of the diffraction grating layer D It can be seen that the equivalent refractive index neff increases as increases.

  FIG. 45 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γqw in the active layer 3B. As shown in FIGS. 40 and 42, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γqw in the active layer 3B is 2.67 (% ) 4.15 (%) in the primary mode. The optical confinement coefficient Γqw in the first-order mode is larger than the optical confinement coefficient Γqw in the 0th-order mode (condition A) when the thickness D is a predetermined value or more, and at least the thickness D is up to 550 nm. This magnitude relationship is maintained. The minimum value (predetermined value) and maximum value of the thickness D that satisfy the condition A are as shown in FIG.

  FIG. 46 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γg. As shown in FIGS. 40 and 42, when the occupied area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γg in the diffraction grating layer 4 is 50.1 ( %) 33.8 (%) in the primary mode. The optical confinement factor Γg in the first-order mode does not become larger than the optical confinement factor Γg in the zero-order mode (condition B).

  FIG. 47 is a chart showing the minimum value Min (nm) and the maximum value Max (nm) of the thickness D of the diffraction grating layer that satisfies the condition A in the structure (2).

  When the occupied area ratio FF is 10%, in order to satisfy the condition A, the thickness D needs to be 200 nm or more, and this condition is satisfied if it is 550 nm or less.

  When the occupied area ratio FF is 20%, in order to satisfy the condition A, the thickness D needs to be 230 nm or more, and this condition is satisfied if it is 550 nm or less.

  When the occupied area ratio FF is 30%, in order to satisfy the condition A, the thickness D needs to be 270 nm or more, and this condition is satisfied if it is 550 nm or less.

  When the occupation area ratio FF is 10% to 30%, the condition A is always satisfied when the thickness D is 270 nm or more and 550 nm or less.

FIG. 48 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γdope. As shown in FIGS. 40 and 42, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γdope in the doped layer 3F is 6.3 (% in the 0th-order mode. ) 0.5 (%) in the primary mode. Optical confinement factor Γdope in the primary mode, 0-order mode is smaller than the light confinement coefficient Γdope the case of (a condition C) is at least the thickness D is less than 150nm or more _{D C.} When the occupied area ratio FF is 10%, the thickness Dc is 480 nm, when the occupied area ratio FF is 20%, the thickness Dc is 540 nm, and when the occupied area ratio FF is 30%, the thickness Dc is 540 nm.

  FIG. 49 is a chart showing detailed parameters of each layer of the semiconductor laser device having still another structure having the core layer A shown in FIG. 29C (referred to as structure (3)).

  The difference between the structure shown in FIG. 49 and that shown in FIG. 30 is only the composition ratio X of aluminum in the doped layer 3F and the film thickness of the carrier block layer 4D, and the other structures are the same. The details will be described below. That is, the aluminum composition ratio X in the doped layer 3F is set to X = 0.1, and the thickness of the carrier block layer 3D is changed to 40 nm. Since other structures are the same, description thereof is omitted.

  FIG. 50 shows the equivalent refractive index neff of the TE mode in the fundamental mode (0th order mode) and the first order mode, and the band edge of the photonic band gap in the diffraction grating layer 4 in the semiconductor laser device having the structure shown in FIG. It is a graph which shows the relationship of a wavelength (nm). The oscillation wavelength λ is 980 nm. The thickness D of the diffraction grating layer 4 is 300 nm (0.3 μm), the thickness d of the core layer A is 730 nm, and the thickness excluding the diffraction grating layer 4 is 430 nm. The number of quantum wells in the active layer 3B is 2.

The refractive index n ₁ of the core layer A = 3.44, the refractive index n ₂ of the lower cladding layer B = 3.10, and the refractive index n ₃ of the upper cladding layer C = 3.26. For these refractive indexes, the value of the refractive index at a wavelength of 980 nm was determined by the MSEO method and the formula (1). In this case, the cut-off thickness _{d cutoff} the primary mode to the propagation mode occurs _(1), the cut-off thickness _{d cutoff} the second-order mode occurs ₍₂₎ is 541 (nm) 992 and (nm). That is, if the thickness d of the core layer A is not less than 541 (nm) and less than 992 (nm), the condition that the primary mode is generated and the secondary mode is not generated is satisfied. The region range where the first-order higher-order mode is taken is calculated as 570 to 1040 nm from the electromagnetic field calculation by the staircase matrix method, and there is no significant deviation.

  Since the total film thickness d of the core layer A is 730 nm as described above, the primary mode generation condition is satisfied, and the TE mode electromagnetic field mode includes a fundamental 0th order mode and a primary mode.

  From the difference in the equivalent refractive index neff shown in FIG. 50, when a triangular lattice photonic crystal having a lattice constant of 344 nm is used for the diffraction grating layer 4, the band edge wavelengths (1013.34 nm, 980. 70 nm) difference is about 30 nm. Therefore, as in the case of the structure (1) (see FIG. 32), if the peak wavelength λp of the gain spectrum is matched with the band edge wavelength λp1 (980 nm) of the first mode, the laser oscillation of the first mode is dominant. Thus, laser oscillation in the 0th order mode is suppressed. In the above case, since the band edge wavelength difference is about 30 nm, if the band edge wavelength λp1 of the primary mode exists within the range of the peak wavelength λp ± 5 nm of the gain spectrum, the primary mode is selective and efficient. Will occur.

  FIG. 51 is a chart showing optical confinement factors in each layer.

  49, the optical confinement coefficient Γqw in the active layer (quantum well layer) 3B is 1.2 (%) in the case of the fundamental mode (0th order mode) and 2 in the case of the primary mode. 39 (%), and more light can be confined in the active layer 3B in the primary mode. The optical confinement coefficient Γg in the diffraction grating door layer 4 is 32.0 (%) in the case of the fundamental mode (0th order mode) and 37.7 (%) in the case of the first order mode. In the mode, more light can be confined in the diffraction grating layer 4. Further, the optical confinement coefficient Γdope in the doped layer 3F is 8.2 (%) in the case of the fundamental mode (0th order mode) and 0.2 (%) in the case of the first order mode. In this case, the influence of the light in the doped layer 3F is further suppressed, and the 0th-order mode light is lost by the impurities in the doped layer 3F.

  As described above, in the first-order mode, since the optical confinement coefficient Γqw in the active layer 3B is high, there is an advantage that laser oscillation with a low threshold is possible.

  As in the case of the structure (1), the optical confinement coefficient Γg in the diffraction grating layer 4 in the first-order mode is higher than that in the fundamental zero-order mode, and the diffraction grating layer 4 functions sufficiently.

Further, P-type impurities of 1 × 10 ¹⁷ / cm ³ or more are added to the doped layer 3F (1 × 10 ¹⁸ / cm ^{3 in} this example). Similar to the structure (1), holes can be injected into the active layer 3B without resistance while efficiently blocking electrons injected from the N-type lower cladding layer 2 side. Since the light loss selectively increases in the next mode, the primary mode can oscillate in a single mode more stably. In addition, it is preferable that the P-type impurity concentration of the doped layer 3F is 1 × 10 ²⁰ / cm ³ or less from the viewpoint of increase in light absorption due to diffusion of the dopant and crystallinity.

  Similarly to the case of the structure (1), when the primary mode is used, as shown in FIG. 52, the position of the valley between the two intensity peaks in the primary mode is set in the vicinity of the doped layer 3F. Thus, even at the bottom of the diffraction grating layer 4 positioned above, it is possible to reduce the electric field strength and reduce light absorption in the damaged region at the time of hole formation in the diffraction grating layer 4. FIG. 52 is a graph showing the relationship between the refractive index n and the electric field intensity E along the thickness direction of the semiconductor laser element, and the primary mode data is shown slightly shifted from the fundamental mode data. It is.

  It should be noted that the electric field strength can be zero at the position of the valley (node) of the electric field strength. For this reason, the node of the electric field strength can be arranged in the damaged region at the bottom of the diffraction grating layer, and the influence of the damaged region can be suppressed. Further, since the active layer 3B layer can be disposed away from the bottom of the diffraction grating layer 4 as compared with the 0th mode, the influence of direct processing damage to the active layer 3B can be suppressed.

  From this point of view, when considering the positional relationship along the X axis, the end surface (bottom surface) of the diffraction grating layer 4 on the active layer 3B side is within a range of ± 50 nm from the position of the valley of the electric field strength peak in the primary mode. The center position in the thickness direction of the doped layer 3F is preferably located within a range of ± 50 nm from the position of the valley of the electric field intensity peak in the primary mode. In order to efficiently use the two intensity peaks of the first mode, the end face (bottom face) of the diffraction grating layer 4 on the active layer 3B side is λ / (4 × neff) from the center position of the active layer 3B. Although it is desirable that the distance is about nm, in order to influence the diffraction grating layer 4 on the active layer 3B, the distance between the center positions is preferably about λ / (2 × neff) nm.

In order to reduce the factor of increasing the internal loss due to the influence of free electron absorption in the wavelength range of about λ = 1 μm, a P-type impurity in a region of 200 nm or less from the end surface on the upper cladding layer 5 side of the diffraction grating layer 4 is used. By reducing the concentration within the range of 1 × 10 ¹⁶ / cm ^{3 to} 3 × 10 ¹⁷ / cm ^3, a laser structure with small internal loss can be realized.

  FIG. 53 is a graph showing the relationship between the thickness D of the diffraction grating layer and the equivalent refractive index neff. As shown in FIGS. 49 and 50, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the equivalent refractive index neff is 3.4015 in the 0th-order mode and 3 in the first-order mode. 2919. When the occupied area ratio FF is 10%, 20%, or 30%, the equivalent refractive index neff is smaller in the first-order mode than in the zero-order mode, and the thickness of the diffraction grating layer D It can be seen that the equivalent refractive index neff increases as increases.

  FIG. 54 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γqw in the active layer 3B. As shown in FIGS. 49 and 51, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γqw in the active layer 3B is 1.20 (% ) 2.39 (%) in the primary mode. The optical confinement coefficient Γqw in the first-order mode is larger than the optical confinement coefficient Γqw in the 0th-order mode (condition A) when the thickness D is equal to or greater than a predetermined value and the occupation area ratio FF is 30. In the case of% or less, this magnitude relationship is maintained at least until the thickness D is 600 nm. The minimum value (predetermined value) and maximum value of the thickness D that satisfy the condition A are as shown in FIG.

  FIG. 55 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γg. As shown in FIGS. 49 and 51, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γg in the diffraction grating layer 4 is 32.0 ( %) 37.7 (%) in the primary mode. The optical confinement coefficient Γg in the first-order mode is larger than the optical confinement coefficient Γg in the zero-order mode (condition B) when the thickness D of the diffraction grating layer is 200 nm or more. As shown.

  FIG. 56 is a chart showing the minimum value Min (nm) and the maximum value Max (nm) of the thickness D of the diffraction grating layer that satisfies the conditions A and B in the structure (3).

  When the occupation area ratio FF is 10%, in order to satisfy the condition A, the thickness D needs to be 150 nm or more, and if it is 600 nm or less, this condition is satisfied. In order to satisfy the condition B, the thickness D needs to be 200 nm or more, and this condition is satisfied if it is 300 nm or less. Both conditions are satisfied when the thickness is 200 nm or more and 300 nm or less.

  When the occupied area ratio FF is 20%, in order to satisfy the condition A, the thickness D needs to be 150 nm or more, and if it is 600 nm or less, this condition is satisfied. In order to satisfy the condition B, the thickness D needs to be 200 nm or more, and this condition is satisfied if it is 600 nm or less. Both conditions are satisfied when the thickness is 200 nm or more and 600 nm or less.

  When the occupied area ratio FF is 30%, in order to satisfy the condition A, the thickness D needs to be 180 nm or more, and if it is 600 nm or less, this condition is satisfied. In order to satisfy the condition B, the thickness D needs to be 200 nm or more, and this condition is satisfied if it is 600 nm or less. Both conditions are satisfied when the thickness is 200 nm or more and 600 nm or less.

  When the occupied area ratio FF is 40%, in order to satisfy the condition A, the thickness D needs to be 210 nm or more, and if it is 450 nm or less, this condition is satisfied. In order to satisfy the condition B, the thickness D needs to be 210 nm or more, and this condition is satisfied if it is 600 nm or less. Both conditions are satisfied when the thickness is 210 nm or more and 450 nm or less.

  When the occupation area ratio FF is 10% to 40%, the condition A and the condition B are always satisfied when the thickness D is 210 nm or more and 450 nm or less.

  FIG. 57 is a graph showing the relationship between the thickness D of the diffraction grating layer and the optical confinement coefficient Γdope. As shown in FIGS. 49 and 51, when the occupation area ratio FF is 20% and the thickness D = 0.3 μm, the optical confinement coefficient Γdope in the doped layer 3F is 8.2 (%) in the 0th-order mode. ) 0.2 (%) in the primary mode. The optical confinement coefficient Γdope in the first-order mode is smaller than the optical confinement coefficient Γdope in the 0th-order mode (condition C) at least in the thickness D of 150 nm to 600 nm.

  58 shows the minimum value Min (nm) and the maximum value Max (nm) of the thickness d of the core layer A in which the secondary mode does not occur and the primary mode occurs in the structures (1), (2), and (3) described above. It is a chart showing. In the structure (1), a primary mode occurs at 573 (nm) or more and less than 1047 (nm), and in the structure (2), a primary mode occurs at 428 (nm) or more and less than 800 (nm), In the structure (3), a primary mode occurs at 541 (nm) or more and less than 992 (nm).

  Considering such a relationship in relation to the oscillation wavelength λ in the semiconductor laser element, the thickness d of the core layer A is 0.585 × λ ≦ d <1.068 × λ in the structure (1). In the structure (2), the relationship of 0.535 × λ ≦ d <1 × λ is satisfied, and in the structure (3), the relationship of 0.552 × λ ≦ d <1.012 × λ is satisfied. Yes. That is, in any structure, the relationship of 0.585 × λ ≦ d <1 × λ is satisfied, and in this case, the primary mode can be selectively generated in the core layer A. .

  In the above-described structures (1) to (3), when the occupied area ratio FF is greatly increased, the inversion phenomenon that the Γqw in the 0th-order mode becomes larger than the Γqw in the higher-order 1st-order mode occurs. . Further, when the occupation area ratio FF is further increased to 80% or more, the first-order higher-order mode is not confined in the core layer. This is because the optical confinement effect is weakened because the refractive index of the diffraction grating layer is lowered. Accordingly, in any of the structures (1) to (3), it is desirable that the occupied area ratio FF <80%.

  In the structure (1), the inversion phenomenon occurs when the occupation area ratio FF is 40% or more. Therefore, it is desirable that the occupied area ratio FF <40%.

  Also in the structure (2), the reverse phenomenon occurs when the occupied area ratio FF is 40% or more. Therefore, it is desirable that the occupied area ratio FF <40%.

  In the structure (3), the inversion phenomenon occurs when the occupied area ratio FF is 50% or more. Therefore, it is desirable that the occupied area ratio FF <50%.

  In addition, the shape of the different refractive index portion in the diffraction grating layer can be selected from an arbitrary shape such as a circle or a rectangle. The principle described above is not limited to a two-dimensional DFB (distributed feedback type) surface light emitting element, and can also be applied to an end surface light emitting element having a similar structure. Further, if the diffraction grating layer 4 has a one-dimensional structure, it can be applied to a one-dimensional DFB structure. By making the pattern of the diffraction grating layer 4 and the upper electrode structure longer as shown in FIG. It is also possible to apply to a type of one-dimensional DFB laser element. Of course, the shape of the upper electrode takes an arbitrary aspect ratio, and the upper electrode may not be in contact with the end face. In the end face emission type structure, various end face treatments can be performed.

  As described above, it has been found that the semiconductor laser device described above has the following advantages.

(1) Mode selectivity Since the equivalent refractive index difference between the fundamental 0th-order mode and the first-order mode is large, the difference between the band edge wavelengths formed by each mode can be increased by 20 nm or more. For this reason, only the primary mode can be oscillated by adjusting the peak wavelength of the gain spectrum to the band edge wavelength of the primary mode. On the other hand, in the second or higher order mode, the equivalent refractive index difference from the first order mode becomes small, and it becomes difficult to select the mode using the relative position of the peak wavelength of the gain spectrum.
(2) Low threshold, high efficiency, mode stabilization Generally, in the first-order mode, the electric field strength has spatially peaks at two positions. By arranging a diffraction grating layer and an active layer at the peak positions, The optical confinement coefficients of the diffraction grating layer and the active layer can be increased, and as a result, a photonic crystal surface emitting laser element (PCSEL) having a low threshold, high efficiency, and mode stability can be realized.
(3) High concentration doping to carrier block layer From the viewpoint of carrier confinement and electrical conduction, it is desirable to perform high concentration doping to the carrier block layer inside the guide layer. However, since it is difficult to reduce the optical confinement coefficient of the carrier block layer in the 0th-order mode, light absorption increases. On the other hand, the primary mode can be designed so that the node of the electric field strength is located in the carrier block layer, and a structure suitable for high output with low loss and excellent carrier confinement can be realized.
(4) Suppressing the effect of processing damage When processing the diffraction grating layer by dry etching, a damaged region is formed in the region of several tens of nanometers from the bottom of the diffraction grating. In the damaged region, lattice defects, impurity levels, and the like are formed, which can cause light absorption. Near the bottom of the diffraction grating layer, it is difficult to reduce the electric field strength in the 0th-order mode, but the node of the electric field strength can be located in the first-order mode, and the influence of light absorption in the damaged region can be suppressed. it can. Further, since the electric field distribution is wider in the primary mode than in the basic 0th order mode, the active layer can be arranged away from the damaged region, and the influence of direct processing damage to the active layer can be suppressed.
(5) Beam pattern When this structure is applied to a normal one-dimensional edge-emitting DFB laser, there are two electric field intensity peaks spatially in the primary mode. The far field pattern (FFP) has two peaks, but even in such a beam, it is low threshold, high efficiency and electric in applications such as solid excitation light sources for high power lasers such as processing applications. The advantages of this structure with excellent characteristics can be utilized. In addition, in a laser beam emitting in the direction perpendicular to the plane using second-order diffraction, a uniform beam can be obtained in principle.

  From the above viewpoint, the two eigenmodes of the electric field in one polarization mode are the fundamental 0th-order mode and the first-order higher-order mode, and the node A of the electric field strength of the first-order mode is the diffraction grating layer and the quantum well layer. In the PCSEL structure between the two, a PCSEL structure having a low threshold, high efficiency, and mode stability can be realized as compared with the conventional structure. In addition, a PCSEL structure with a low threshold, high efficiency, and mode stability can be easily manufactured by changing only the composition and film thickness of each semiconductor layer.

  As described above, the semiconductor laser device 10 described above is sandwiched between the lower cladding layer B made of semiconductor, the upper cladding layer C made of semiconductor, the lower cladding layer B, and the upper cladding layer C in the semiconductor laser device. A core layer A comprising a plurality of stacked semiconductor layers and having a higher average refractive index than any of the lower clad layer B and the upper clad layer C, the core layer A being an active layer comprising a quantum well layer 3B and the diffraction grating layer 4, and the electric field strength distribution Ey in the thickness direction in the core layer A during operation has at least two peaks, and the position of the valley between the peaks is the active layer 3B. And in the region between the diffraction grating layer 4.

  In this case, although the electric field intensity distribution is in the core layer A, the position of the valley is not in any position of the active layer 3B and the diffraction grating layer 4, so that these two are suppressed while suppressing a decrease in laser light generation gain. The laser light generating action in the two layers can sufficiently function, and the laser light can be generated with high efficiency.

  Further, when the positions of the peaks are set in the active layer 3B and the diffraction grating layer 4, respectively, the electric field strength at these positions can be increased, and high-intensity laser light can be generated. It becomes possible.

  Further, in the core layer A, the optical confinement coefficient Γqw (1) in the active layer 3B in the first-order mode of TE polarized wave that gives the two peaks, and the optical confinement coefficient Γqw (in the active layer 3B in the zero-order mode) 0) satisfies the relational expression Γqw (1)> Γqw (0). By setting the two peak positions of the electric field intensity in the active layer 3B and the diffraction grating layer 4 as described above, the optical confinement factor can be increased. In this case, it is possible to increase the electric field intensity at these positions, and it is possible to generate high-intensity laser light, and it is possible to obtain high-intensity laser light compared to the case where the 0th-order mode is mainly used. It becomes. Note that Γ is a numerator obtained by integrating the electric field E (x) from −infinity to + infinity along the X-axis direction, and integrating the electric field E (x) in a specific section x1 to x2. (See equation (7) in FIG. 59).

The core layer A is doped with an impurity having the same conductivity type as that of the upper clad layer C between the active layer 3B and the upper clad layer C, and the impurity concentration thereof is 1 × 10 ¹⁷ / cm ³ or more. And part or all of the doped layer is set in the vicinity of the position of the valley between the two peaks. “Nearby” means a region within ± 50 nm from the valley position. By providing the doped layer 3F, the Fermi level changes to form a barrier, and carriers flowing from the lower cladding layer can be prevented from flowing toward the upper cladding layer, and the carrier concentration in the active layer 3B can be suppressed. Can be increased.

  Further, when the impurity concentration is high, energy of light and carriers is absorbed and loss occurs. The positions of the valleys of the two peaks in the first-order mode correspond to the peak positions in the zero-order mode. Therefore, by providing the doped layer 3F, light loss is caused at the peak position in the 0th-order mode, the occurrence of the 0th-order mode is suppressed, the light emission by the first-order mode is stabilized, and the primary mode is used more effectively. can do.

  The diffraction grating layer 4 is made of a group III-V compound semiconductor, the group III element is selected from the group consisting of Ga, Al, and In, and the group V element is selected from the group consisting of As, P, N, and Sb. Since the selected compound semiconductor composed of these elements can be a direct transition semiconductor, it can easily emit light by carrier recombination.

  In the case of emission in the direction perpendicular to the surface, as shown in Non-Patent Document 2, the depth of the diffraction grating has a depth at which the radiation coefficient in the vertical direction becomes small due to the influence of vanishing interference in the depth direction. Also in the present application, it is desirable to select the diffraction grating layer thickness so that the vertical emission coefficient does not become small. It should be noted that the above effect is not affected by the emission in the end face direction.

  Further, the shape of one different refractive index portion 4B of the diffraction grating layer 4 may not be uniform with respect to the X-axis direction along the epilayer thickness, and may be inclined with respect to the X-axis direction. good.

A ... core layer, B, C ... cladding layer, 3B ... active layer, 3F ... doped layer, 4 ... diffraction grating layer.

Claims (4)

In a semiconductor laser device,
A lower cladding layer made of semiconductor;
An upper cladding layer made of semiconductor;
A core layer sandwiched between the lower clad layer and the upper clad layer and composed of a plurality of stacked semiconductor layers, having a higher average refractive index than any of the lower clad layer and the upper clad layer,
With
The core layer is
An active layer adjacent to the guide layer and comprising a quantum well layer;
Only one diffraction grating layer separated from the active layer via a carrier block layer having an energy band gap larger than that of the guide layer ;
Including
The electric field strength distribution in the thickness direction in the core layer during operation is distributed in the first-order mode of the TE polarized wave having two peaks,
The position of the valley between the peaks is set in a region between the active layer and the diffraction grating layer ,
The core layer further includes a doped layer in which an impurity having the same conductivity type as that of the upper clad layer is added between the active layer and the upper clad layer, and the impurity concentration thereof is 1 × 10 ¹⁷ / cm ³ or more. ,
Part or all of the doped layer is set at a position within a range of ± 50 nm from the position of the valley between the two peaks.
A semiconductor laser device.   2. The semiconductor laser device according to claim 1, wherein the positions of the peaks are set in the active layer and the diffraction grating layer, respectively. In the core layer, the optical confinement factor Γqw (1) in the active layer in the first-order mode of TE polarized wave giving the two peaks, and the optical confinement factor Γqw (0) in the active layer in the zero-order mode Satisfies the relational expression Γqw (1)> Γqw (0).
The semiconductor laser device according to claim 1, wherein the semiconductor laser device is a semiconductor laser device. The diffraction grating layer is made of a III-V group compound semiconductor,
The group III element is selected from the group consisting of Ga, Al and In;
The group V element is selected from the group consisting of As, P, N and Sb;
The semiconductor laser device according to any one of claims 1 to 3, characterized in that.