Surface emitting laser element and method for manufacturing the same

Abstract

The present invention comprises: a first guide layer including a photonic-crystal layer that is formed on a c surface of a group III nitride semiconductor and that has holes which are arranged with two-dimensional periodicity in a plane parallel to the layer, and an embedded layer that is formed on the photonic-crystal layer and that closes the holes; an active layer that is formed on the first guide layer; and a second guide layer that is formed on the active layer. A hole set, which includes at least a main hole and a sub hole that is smaller in size than the main hole, is disposed at each square lattice point in the plane parallel to the photonic-crystal layer. The main holes have a regular hexagonal prism shape, an elongated hexagonal prism shape, or an elliptical prism shape in which the longer axis is parallel to a <11-20> axis.

Description

[Technical field]

[0001] The present invention relates to a surface-emitting laser element and a manufacturing method thereof, and more particularly to a photonic crystal surface-emitting laser element and a manufacturing method thereof. [Background technology]

[0002] 2. Description of the Related Art In recent years, the development of photonic crystal surface-emitting lasers using photonic crystals (PC) has been progressing.

[0003] For example, Patent Document 1 describes a two-dimensional photonic crystal surface-emitting laser light source having a two-dimensional photonic crystal in which a plate-shaped base material is provided with a large number of periodically arranged modified refractive index area assemblies, each of which is composed of a plurality of regions having a refractive index different from that of the base material and at least two of the regions having thicknesses different from each other. It is disclosed that this configuration makes it possible to reduce symmetry in a plane parallel to the base material compared to a two-dimensional photonic crystal in which cylindrical modified refractive index areas are periodically arranged, and to suppress a decrease in the efficiency of laser light extraction due to cancellation of antisymmetric modes caused by interference.

[0004] When a burying layer is grown on a two-dimensional photonic crystal in which a large number of such voids are periodically arranged within the crystal plane, there is a problem that the surface of the burying layer becomes rough, deteriorating the quality of the active layer grown on the burying layer. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent No. 4294023 Summary of the Invention [Problem to be solved by the invention]

[0006] As described above, in photonic crystal surface emitting lasers, there has been a problem in that the surface of the burying layer in which the two-dimensional photonic crystal layer is buried becomes rough, deteriorating the quality of the active layer formed on the burying layer.

[0007] In addition, the inventors of the present application have found that when a two-dimensional photonic crystal in which sets of air holes of different sizes are periodically arranged is embedded with an embedding layer, the surface of the embedding layer becomes rougher than in the case of a two-dimensional photonic crystal consisting of air holes of a single size.

[0008] In particular, the inventors found that in the process of forming the buried layer, the shape of the voids changes due to mass transport, and the shape changes of adjacent voids interfere with each other due to the change in the shape of the voids, causing unevenness on the surface of the buried layer.

[0009] The present invention has been made in consideration of the above-mentioned points, and has an object to provide a photonic crystal surface-emitting laser and a manufacturing method thereof, which have a two-dimensional photonic crystal (hereinafter also referred to as a multi-lattice photonic crystal) configured by arranging a plurality of air holes of different sizes at lattice points, in which sets of the plurality of air holes are periodically arranged, and in which the surface flatness of the embedding layer in which the multi-lattice photonic crystal is embedded is greatly improved.

[0010] Another object of the present invention is to provide a photonic crystal surface-emitting laser in which the active layer formed on the multilattice photonic crystal layer has high flatness and crystallinity, has high light extraction efficiency, and is capable of oscillating at a low threshold current density and high quantum efficiency, and a method for manufacturing the same. [Means for solving the problem]

[0011] A surface-emitting laser element according to one embodiment of the present invention is a surface-emitting laser element made of a group III nitride semiconductor, a first guide layer including: a photonic crystal layer formed on a c-plane of a Group III nitride semiconductor and having holes arranged with two-dimensional periodicity in a plane parallel to the layer; and a burying layer formed on the photonic crystal layer and closing the holes; an active layer formed on the first guide layer; a second guide layer formed on the active layer; a hole set including at least a main hole and a subhole having a size smaller than that of the main hole is disposed at each of square lattice points in a plane parallel to the photonic crystal layer, The main holes have a regular hexagonal prism shape, a long hexagonal prism shape, or a long cylindrical shape with their major axes parallel to the <11-20> axis.

[0012] A method for manufacturing a surface emitting laser element according to another embodiment of the present invention includes the steps of forming a guide layer on a c-plane of a group III nitride semiconductor; forming an etching mask on the guide layer, the etching mask having an opening set including at least a main opening and a sub-opening having a size smaller than that of the main opening at each of the square lattice points; Etching the guide layer using the etching mask to form main holes and sub-holes; forming a buried layer that closes the openings of the main holes and the sub-holes by performing crystal growth including mass transport, thereby forming a multi-lattice photonic crystal layer in which a hole set including a main hole and a sub-hole having a size smaller than that of the main hole is arranged at each of the square lattice points; forming a semiconductor layer including an active layer on the multilattice photonic crystal layer; The main holes are characterized in that they have a regular hexagonal prism shape, an elongated hexagonal prism shape, or an elongated cylindrical shape with their major axes parallel to the <11-20> axis. [Brief description of the drawings]

[0013] [Figure 1A] FIG. 1 is a cross-sectional view showing a schematic example of the structure of a photonic crystal laser (PCSEL) 10. [Figure 1B] 1B is an enlarged cross-sectional view showing a schematic diagram of photonic crystal layer 14P and air hole pairs 14K arranged in photonic crystal layer 14P in FIG. 1A. FIG. [Figure 2A] 1 is a plan view showing a schematic view of the upper surface of photonic crystal laser 10. FIG. [Figure 2B] 1B is a cross-sectional view that illustrates a schematic cross section of photonic crystal layer 14P in a plane parallel to n-guide layer 14 (cross section AA in FIG. 1B). [Figure 2C] 2 is a plan view showing a schematic view of the bottom surface of photonic crystal laser 10. FIG. [Diagram 3] FIG. 13 is a top view that typically shows a resist in which opening pairs, each of which is an oval-shaped main opening K1 and sub-opening K2, are two-dimensionally arranged in a square lattice pattern within the plane. [Figure 4] 1 is an SEM image of a GaN surface on which holes 14H1 and 14H2 are formed. [Diagram 5] 1 is a SEM image of the upper surface of a guide layer, showing the shapes of main holes 14K1 and sub-holes 14K2 in this embodiment. [Figure 6A] 6 is an SEM image showing a cross section taken along line AA in FIG. 5. [Figure 6B] 6 is an SEM image showing a cross section taken along line BB in FIG. 5. [Figure 7] 1 is an SEM image showing cylindrical holes CH formed in a GaN surface portion in Comparative Example 1. [Figure 8A] FIG. 2 is a graph showing the IL characteristics of the PCSEL device 10 of Example 1 and the PCSEL device of Comparative Example 1. [Figure 8B] FIG. 2 shows emission spectra near the threshold current of the PCSEL device 10 of Example 1 and the PCSEL device of Comparative Example 1. [Figure 9A] 13 is a top SEM image showing a main hole 14H1 and a sub-hole 14H2 of a structure (structure A) in which main holes and sub-holes are formed similarly to Example 1. [Figure 9B] 1 is a top view SEM image showing a main hole CH1 and a sub-hole CH2 of Comparative Example 2 (structure B). [Figure 10A] 1 is an AFM image showing the surface morphology of the buried layer after filling the main voids and the sub-voids in the case of the structure A (Example 2). [Figure 10B] 11 is an AFM image showing the surface morphology of the buried layer after the main voids and the sub-voids are filled in the case of Structure B (Comparative Example 2). [Figure 11A]FIG. 11 is a top view diagrammatically illustrating a change in shape of a hole in the process of forming a buried layer in the case of Structure A (Example 2). [Figure 11B] FIG. 13 is a top view illustrating a schematic change in shape of a hole in the process of forming a buried layer in the case of Structure B (Comparative Example 2). [Figure 12A] 13 is a graph showing the vertical cavity loss αn when the relative positions Δx, Δy (Δx=Δy) of the subhole 14K2 with respect to the main hole 14K1 are changed from 0.0PC to 0.5PC in the structure of Example 1. [Figure 12B] 13 is a graph showing the horizontal cavity loss αp when the relative position Δx, Δy (Δx=Δy) of the subhole 14K2 with respect to the main hole 14K1 is changed from 0.0PC to 0.5PC in the structure of Example 1. [Figure 12C] 1 is a graph showing the ratio Rn of the vertical cavity loss αn to the total cavity loss (αp+αn). [Figure 13] FIG. 2 is a diagram showing a photonic band structure near the Γ point of a square lattice photonic crystal. [Figure 14] 1 is a graph showing the threshold gain (resonator loss) of each mode obtained from the coupled wave theory when Δx and Δy are changed in Example 1. [Figure 15] 3 is a diagram showing a schematic diagram of the refractive index, carrier density, and photon density of a current injection region. FIG. [Figure 16A] 1 is a diagram showing a schematic diagram of the frequency of the photonic band edge in the vicinity of the current injection region in the cases of band edge modes A and B. FIG. [Figure 16B] 10 is a diagram showing a schematic diagram of the frequency of the photonic band edge in the vicinity of the current injection region in the cases of band edge modes C and D. FIG. [Figure 17A] FIG. 1 is a diagram showing the refractive index, carrier density, and photon density in the current injection region when lasing in band edge modes A and B. The solid and dashed lines show the cases when the photonic band effect is taken into account and when it is not taken into account, respectively. [Figure 17B]The graphs show the refractive index, carrier density, and photon density in the current injection region when lasing in band edge modes C and D. The solid and dashed lines show the cases when the photonic band effect is taken into account and when it is not, respectively. [Figure 18] 1 is an SEM image of a GaN surface on which holes 14H1 and 14H2 are formed. [Figure 19A] 13 is a SEM image of the upper surface of a guide layer, showing the hole shapes of a main hole 14K1 and a sub-hole 14K2 in Example 3. [Figure 19B] 19B is an SEM image showing a cross section taken along line AA in FIG. 19A. [Figure 19C] 19B is an SEM image showing a cross section taken along line BB in FIG. 19A. [Figure 20A] FIG. 1 shows the IL characteristics of the PCSEL device 10 of Example 3 and the PCSEL device of Comparative Example 1. [Figure 20B] FIG. 1 shows emission spectra near the threshold current of the PCSEL device 10 of Example 3 and the PCSEL device of Comparative Example 1. [Figure 21A] FIG. 13 is a graph showing the vertical cavity loss αn versus d (the relative position of the subhole to the main hole). [Figure 21B] FIG. 13 is a diagram showing the horizontal cavity loss αp versus the relative position d. [Figure 21C] FIG. 13 shows the dependence of Rn on the relative position d. [Figure 21D] 13 is a graph showing the threshold gain (resonator loss) of each mode obtained from the coupled wave theory when Δx and Δy (relative position d) are changed in Example 3. [Figure 22A] FIG. 13 is a graph showing Rn versus hole filling rate ratio FF1 / FF2. [Figure 22B] FIG. 13 is a graph showing the sum of the vertical and horizontal cavity losses (αn+αp) versus the hole filling factor ratio FF1 / FF2. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] In the following, preferred embodiments of the present invention will be described, which may be modified and combined as appropriate. In the following description and accompanying drawings, substantially the same or equivalent parts are designated by the same reference numerals. [Photonic crystal surface-emitting laser structure] A photonic crystal surface-emitting laser (hereinafter also referred to as PCSEL) is an element that has a resonator layer in a parallel direction to the semiconductor light-emitting structure layers (n-guide layer, light-emitting layer, p-guide layer) that constitute the light-emitting element, and emits coherent light in a direction perpendicular to the resonator layer.

[0015] Meanwhile, distributed Bragg reflector (DBR) lasers are known that have a pair of resonator mirrors (Bragg reflectors) that sandwich a semiconductor light-emitting structure layer, but photonic crystal surface-emitting lasers (PCSELs) differ from DBR lasers in the following respects: In photonic crystal surface-emitting lasers (PCSELs), light waves propagating within a plane parallel to the photonic crystal layer are diffracted by the diffraction effect of the photonic crystal to form a two-dimensional resonance mode, and are also diffracted in a direction perpendicular to the parallel plane. In other words, the light extraction direction is perpendicular to the resonance direction (within the plane parallel to the PC layer).

[0016] Fig. 1A is a cross-sectional view showing a schematic example of a structure (first PCSEL structure) of a photonic crystal laser element (PCSEL element) 10. As shown in Fig. 1A, a semiconductor structure layer 11 is formed on a substrate 12. The semiconductor structure layer 11 is made of a hexagonal nitride semiconductor. The semiconductor structure layer 11 is made of, for example, a GaN-based semiconductor.

[0017] More specifically, a semiconductor structure layer 11, that is, an n-clad layer (first clad layer of a first conductivity type) 13, an n-guide layer (first guide layer) 14, an active layer 15, a p-guide layer (second guide layer) 16, an electron barrier layer (EBL: Electron Blocking Layer) 17, a p-clad layer (second clad layer of a second conductivity type) 18, and a p-contact layer 19 are formed in this order on a substrate 12. Note that, although a case will be described in which the first conductivity type is n-type and the second conductivity type, which is the opposite conductivity type to the first conductivity type, is p-type, the first conductivity type and the second conductivity type may be p-type and n-type, respectively.

[0018] The n-guide layer 14 is made up of a lower guide layer 14A, a photonic crystal layer (air hole layer or PC layer) 14P, and an burying layer 14B. The burying layer 14B is made up of a first burying layer 14B1 and a second burying layer 14B2.

[0019] In this specification, "n-" and "p-" mean "n-side" and "p-side", respectively, and do not necessarily mean that they have an n-type or p-type. For example, an n-guide layer means a guide layer provided on the n-side of the active layer, and may be an undoped layer (or an i-layer).

[0020] In addition, the n-cladding layer 13 may be composed of multiple layers instead of a single layer, and in that case, all layers do not need to be n layers (n-doped layers) and may include an undoped layer (i layer). The same applies to the guide layer 16 and the p-cladding layer 18.

[0021] Although the specific and detailed configuration of the semiconductor layers of photonic crystal laser element 10 has been described above, this is merely one example of the element structure. In short, it is sufficient that the element has a first semiconductor layer (or guide layer) having photonic crystal layer 14P, a second semiconductor layer (or guide layer), and an active layer (light emitting layer) sandwiched between these layers, and is configured to emit light by injecting a current into the active layer.

[0022] For example, the photonic crystal laser element does not need to have all of the above-mentioned semiconductor layers, or may have various semiconductor layers (e.g., a hole barrier layer, a light confinement layer, a current confinement layer, a tunnel junction layer, etc.) for improving element characteristics.

[0023] In addition, an n-electrode (cathode) 20A is formed on the back surface of the substrate 12, and a p-electrode (anode) 20B is formed on (the upper surface of) the p-contact layer 19. The side surfaces of the semiconductor structure layer 11 and the side surfaces of the upper part of the substrate 12 are covered with an insulating film 21 such as SiO2. In addition, the side surfaces of the p-electrode 20B and the surface of the p-contact layer 19 are covered with the insulating film 21 so as to cover the edge portion of the upper surface of the p-electrode 20B.

[0024] Light directly emitted from photonic crystal layer (PC layer) 14P (directly emitted light Ld) and light emitted from photonic crystal layer 14P and reflected by p-electrode 20B (reflected emitted light Lr) are emitted to the outside from light emission region 20L on the back surface of substrate 12.

[0025] Fig. 1B is an enlarged cross-sectional view showing photonic crystal layer 14P and air hole pairs 14K arranged in photonic crystal layer 14P in Fig. 1A. Air hole pairs 14K (main air holes 14K1 and sub-air holes 14K2) are formed in a crystal growth plane (semiconductor layer growth plane), i.e., in a plane (AA cross section in the figure) parallel to n-guide layer 14, with a period PC in the shape of, for example, a square lattice, and air hole pairs 14K are two-dimensionally arranged at square lattice point positions and embedded in n-guide layer 14.

[0026] FIG. 2A is a plan view diagrammatically showing the top surface of photonic crystal laser (PCSEL) 10, FIG. 2B is a cross-sectional view diagrammatically showing a cross-section of photonic crystal layer (PC layer) 14P in a plane parallel to n-guide layer 14 (cross-section AA in FIG. 1B), and FIG. 2C is a plan view diagrammatically showing the bottom surface of photonic crystal laser (PCSEL) 10.

[0027] As shown in Fig. 2B, in photonic crystal layer 14P, air hole pairs 14K are arranged periodically within, for example, a rectangular air hole formation region 14R. As shown in Fig. 2C, n-electrode (cathode) 20A is arranged in an annular shape on the outside of air hole formation region 14R so as not to overlap air hole formation region 14R when viewed from a direction perpendicular to photonic crystal layer 14P. The region inside n-electrode 20A is light emitting region 20L. Also, it is electrically connected to n-electrode 20A, and is provided with bonding pad 20C for connecting a wire for power supply from outside. EXAMPLES

[0028] 1. Fabrication process of photonic crystal laser (PCSEL) 10 The following is a detailed description of the steps for fabricating the PCSEL device 10. Metalorganic Vapor Phase Epitaxy (MOVPE) was used as the crystal growth method, and the semiconductor structure layer 11 was grown on the growth substrate 12 by normal pressure (atmospheric pressure) growth. In the steps described below, Sn means step n.

[0029] Moreover, the layer thicknesses, carrier concentrations, Group 3 (Group III) and Group 5 (Group V) materials, temperatures, and the like shown below are merely examples unless otherwise specified. [S1: Board preparation process] We prepared a GaN single crystal with a "+c" plane as the main surface, which is the (0001) plane with Ga atoms arranged on the outermost surface. The main surface may be just the same, or it may be a substrate offset by up to about 1° in the m-axis direction. For example, a substrate offset by up to about 1° in the m-axis direction can obtain mirror-finish growth under a wide range of growth conditions.

[0030] The substrate surface (rear surface) on which the light emission region 20L is provided and which faces the main surface is the “-c” surface, which is the (000-1) surface on which N atoms are arranged. The -c surface is resistant to oxidation and the like, and is therefore suitable as a light extraction surface.

[0031] In this embodiment, an n-type GaN single crystal was used as the GaN substrate 12. The n-type GaN substrate 12n functions as a contact layer with the electrode. [S2: n-clad layer formation process] On the +c-plane GaN substrate 12, an n-clad layer 13 having an Al composition of 4% is formed. 0.04 Ga 0.96 The N layer was grown to a thickness of 2 μm. The AlGaN layer was grown by supplying trimethylgallium (TMG) and trimethylaluminum (TMA) as group III atom sources to a GaN substrate heated to 1100°C.

[0032] Carrier doping was performed by simultaneously supplying silane (SiH4) and the above-mentioned raw materials (Si doping). In this case, the carrier concentration at room temperature was approximately 4 × 10 17 cm -3 It was. [S3a: Lower guide layer + hole preparation layer formation process] Next, TMG was supplied to grow n-type GaN to a thickness of 250 nm as a preparatory layer for the n-guide layer 14. Silane (SiH4) was supplied at the same time to dope the carriers, as in the case of the AlGaN layer. The carrier concentration at this time was approximately 4×10 17 cm -3 It was.

[0033] This growth layer is a preparation layer for forming a layer consisting of lower guide layer 14A and photonic crystal layer 14P.

[0034] In the following, for ease of explanation and understanding, the substrate 12 on which such a growth layer is formed (substrate with growth layer) may be simply referred to as a substrate. [S3b: Hole and void formation process] After the preparation layer was formed, the substrate was removed from the chamber of the MOVPE apparatus, and fine holes were formed on the surface of the growth layer. More specifically, the substrate was washed to obtain a clean surface, and then a silicon nitride film (Si x N y) was formed using plasma CVD. A resist for electron beam lithography was applied on top of this by spin coating, and the resist was placed in an electron beam (EB) lithography device to pattern a two-dimensional periodic structure.

[0035] As shown in Fig. 3, a pair of openings, each consisting of an oval main opening K1 and a sub-opening K2 smaller than the main opening K1, was two-dimensionally arranged in the resist plane in a square lattice pattern with a period PC = 164 nm. For clarity of the drawing, the openings are shown with hatching.

[0036] More specifically, the main apertures K1 are arranged in a square lattice pattern with their centers of gravity CD1 in two mutually orthogonal directions (x direction and y direction) at a period PC=164 nm, and the sub-apertures K2 are also arranged in a square lattice pattern with their centers of gravity CD2 in the x direction and y direction at a period PC=164 nm.

[0037] The major axes of the main aperture K1 and the sub aperture K2 are parallel to the <11-20> direction of the crystal orientation, and the minor axes of the main aperture K1 and the sub aperture K2 are parallel to the <1-100> direction.

[0038] Also, the center of gravity CD2 of the sub-aperture K2 is spaced apart from the center of gravity CD1 of the main aperture K1 by Δx and Δy. Here, Δx=Δy. That is, the center of gravity CD2 of the sub-aperture K2 is spaced apart from the center of gravity CD1 of the main aperture K1 in the <1-100> direction. Specifically, the center-of-gravity distance Δx in the x direction and the center-of-gravity distance Δy in the y direction were 65.6 nm (=PC×0.4).

[0039] The main opening K1 had a major axis of 125 nm and a minor axis of 50 nm, with a ratio of the major axis to the minor axis (major axis / minor axis) of 2.50, and the sub-opening K2 had a major axis of 57.5 nm and a minor axis of 50 nm, with a ratio of the major axis to the minor axis (major axis / minor axis) of 1.15.

[0040] After developing the patterned resist, the Si x N yThe film was selectively dry etched. As a result, main openings K1 and sub-openings K2 arranged in a square lattice pattern with a period of 164 nm were formed in the Si x N y formed to penetrate the membrane.

[0041] The period (hole spacing) Pc was calculated as Pc=λ / n=164 nm, assuming an oscillation wavelength (λ) of 410 nm and a refractive index (n) of GaN of 2.5.

[0042] Next, the resist is removed and the patterned Si x N y The film was used as a hard mask to form holes in the GaN surface. The GaN was dry-etched in the depth direction using a chlorine-based gas and argon gas in an ICP-RIE apparatus to form holes 14H1 and 14H2, which are oblong cylindrical holes dug vertically in the GaN surface. In this process, the holes dug in the GaN surface by this etching are hereinafter referred to simply as holes to distinguish them from air holes in the photonic crystal layer 14P. [S3c: Cleaning process] The substrate in which the holes 14H1 and 14H2 are formed is degreased and cleaned, and then the Si x N y The film was then removed. A SEM (Scanning Electron Microscope) image of the GaN surface at this point is shown in Figure 4.

[0043] As shown in the surface SEM image in Figure 4, a plurality of hole pairs 14H (main holes 14H1 and sub-holes 14H2) were formed in a square lattice pattern, i.e., two-dimensionally on the square lattice points, with a period PC of 164 nm. The holes 14H1 and 14H2 are approximately oval cylindrical holes or apertures that open to the top surface (GaN surface).

[0044] More specifically, the major axis LH1 of the opening of the main hole 14H1 on the GaN surface was 124 nm and the minor axis WH1 was 50 nm (RH1 = major axis / minor axis = 2.49), and the major axis LH2 of the opening of the sub-hole 14H2 was 57.1 nm and the minor axis WH1 was 50.0 nm (RH2 = major axis / minor axis = 1.14).

[0045] In this case, the distance between the center of gravity CD1 of the main hole 14H1 and the center of gravity CD2 of the sub-hole 14H2 in the x and y directions was 65.3 nm (=0.4 × PC), respectively. The major axes of the main hole 14H1 and the sub-hole 14H2 were arranged parallel to the <11-20> axis (i.e., the a-axis).

[0046] 4, the x and y directions are inclined at 45° with respect to the long axis direction (<11-20> direction) and the short axis direction (<1-100> direction) of the opening of the main hole 14H1 and the opening of the sub-hole 14H2, respectively. In this specification, the xy coordinates are also referred to as hole coordinates. [S3d: Buried layer formation process] This substrate was again introduced into the reactor of the MOVPE apparatus, and ammonia (NH3) was supplied to raise the temperature to 950°C (first embedding temperature). Trimethylgallium (TMG) and NH3 were then supplied to block the opening of the main hole 14H1 and the sub-hole 14H2, thereby forming the first embedded layer 14B1.

[0047] First, in the process of reaching the first temperature region (800°C or higher and 1100°C or lower) (heat-up process), mass transport of Ga atoms on the growth substrate surface occurs in the supplied NH3 atmosphere, forming an overhanging portion composed of {1-101} and {1-100} planes that closes the opening of the hole formed in the n-cladding layer.

[0048] Next, the above-mentioned eaves grow toward the center of the hole by TMG supplied after the temperature reaches the first temperature region, and then coalesce to close and fill the main hole 14H1 and the sub-hole 14H2, thereby forming a first buried layer 14B1.

[0049] Next, after the main hole 14H1 and the sub-hole 14H2 were blocked, a second buried layer 14B2 having a thickness of 50 nm was grown. The second buried layer 14B2 was grown by lowering the substrate temperature to 820°C (second buried temperature), and then supplying triethylgallium (TEG) and trimethylindium (TMI) as a group 3 atom source and NH3 as a nitrogen source. The second buried temperature is lower than the first buried temperature, and is 700°C or higher and lower than 900°C.

[0050] In addition, the In composition of the second buried layer 14B2 in this embodiment is 2% (i.e., Ga 0.98 In 0.02 Second embedded layer 14B2 functions as a light distribution adjusting layer for adjusting the coupling efficiency (light field) between light and photonic crystal layer 14P.

[0051] By the above embedding process, a photonic crystal layer 14P having a double lattice structure in which a pair of holes 14K, each consisting of a main hole 14K1 and a sub-hole 14K2, is arranged at each of the square lattice points is formed. [S4: Light-emitting layer formation process] Next, a multiple quantum well (MQW) layer was grown as the active layer 15, which is a light-emitting layer. The barrier layer and well layer of the MQW were GaN and InGaN, respectively. The barrier layer was grown by lowering the temperature of the substrate to 820°C, and then supplying triethylgallium (TEG) as a group 3 atom source and NH3 as a nitrogen source. The well layer was grown at the same temperature as the barrier layer, and by supplying TEG and trimethylindium (TMI) as group 3 atom sources and NH3 as a nitrogen source. The center wavelength of the PL (Photoluminescence) emission from the active layer in this example was 412 nm. [S5: p-guide layer formation process] After the active layer was grown, the substrate was heated to 1050° C., and GaN was grown to a thickness of 120 nm as the p-guide layer 16. The p-guide layer 16 was grown by supplying TMG and NH3 without doping with a dopant. [S6: Electron barrier layer formation process] After the growth of the p-guide layer 16, the electron barrier layer (EBL) 17 was grown while maintaining the substrate temperature at 1050°C. The EBL 17 was grown by supplying TMG and TMA as the group 3 atom source and NH3 as the nitrogen source. Cp2Mg was also supplied as the p-dopant. As a result, the EBL 17 with an Al composition of 18% and a layer thickness of 15 nm was formed. [S7: p-cladding layer formation process] After the growth of the electron barrier layer (EBL) 17, the p-cladding layer 18 was grown while maintaining the substrate temperature at 1050°C. The p-cladding layer 18 was grown by supplying TMG and TMA as the group 3 atom source and NH3 as the nitrogen source. Cp2Mg was also supplied as the p-dopant. As a result, a p-cladding layer 18 with an Al composition of 6% and a layer thickness of 600 nm was formed. Note that the carrier concentration of the p-cladding layer (p-AlGaN) 18 was 2×10 17 cm -3 It was. [S8: p-contact layer formation process] After growing the p-cladding layer 18, the p-contact layer 19 was grown while maintaining the substrate temperature at 1050° C. The growth of the p-contact layer 19 was performed by supplying TMG as a group 3 atom source and NH3 as a nitrogen source. Cp2Mg was also supplied as a dopant. [S9: Electrode formation process] The surface of the p-contact layer 19 of the substrate with the epitaxially grown layer is attached to a support substrate, and the substrate 12 is thinned to a predetermined thickness by a polishing device.

[0052] Thereafter, a mask covering the p-contact layer 19 side except for the element isolation groove was formed, and etching was performed until the n-clad layer 13 or the substrate 12 was exposed. Thereafter, the mask was removed, the support substrate was removed, and the element isolation groove was formed. [S10: Electrode formation process] (Anode electrode formation) A palladium (Pd) film and a gold (Au) film were formed in this order as a p-electrode 20B on the surface of the epitaxial growth substrate 12 by electron beam deposition. The formed electrode metal film was then photolithographically cut to a size of 200×200 μm. 2 Then, the resulting film was patterned to form a p-electrode 20B. (Cathode electrode formation) Next, Ti and Au were deposited in this order on the back surface of the substrate 12 by electron beam deposition to form an n-electrode 20A. [S11: Protective film formation process] The bottom surface of the substrate after the electrode formation was attached to a support substrate, and a mask was formed to cover the anode electrode. After that, a SiO2 film was formed as a protective film on the top and side surfaces of the element by sputtering. [S12: Singulation process] Finally, laser scribing was performed along the center lines of the substrate separation grooves to obtain individualized PCSEL elements (hereinafter referred to as PCSEL elements or simply PCSELs) 10. 2. Main and secondary holes In order to confirm the shape of the filled voids in this embodiment, the laminated structure was processed from the surface by a focused ion beam (FIB) until the voids in the void layer were exposed, and then SEM observation was performed. The void shapes of the main void 14K1 and the sub-void 14K2, which is smaller in size than the main void 14K1, are shown in the top SEM image of Figure 5. Note that the sub-void 14K2 only needs to be smaller than the main void 14K1 in at least one of the void diameter and depth.

[0053] 5. FIG. 6A is an SEM image showing a cross section taken along line AA in FIG. 5, and FIG. 6B is an SEM image showing a cross section taken along line BB in FIG.

[0054] As shown in FIG. 5, the major axis LK1 of the main hole 14K1 was 75.2 nm, the minor axis WK1 was 45 nm, and the ratio of the major axis to the minor axis (major axis / minor axis) RK1=1.67. The major axis LK2 of the subhole 14K2 was 45.8 nm, the minor axis WK2 was 39.2 nm, and the ratio of the major axis to the minor axis (major axis / minor axis) RK2=1.17.

[0055] As shown in FIGS. 6A and 6B, the depth DK1 of the main void 14K1 was 102 nm, and the depth DK2 of the subvoid 14K2 was 78.5 nm.

[0056] In this way, it was confirmed that a photonic crystal layer 14P was formed in which hole pairs, each consisting of a main hole 14K1 and a sub-hole 14K2, were arranged at each square lattice point with a period PC=164 nm. That is, the main hole 14K1 is arranged at a square lattice point with a period PC, and the sub-hole 14K2 is arranged at a square lattice point with the same period PC. The hole pairs are arranged such that the distance between the center of gravity D1 of the main hole 14K1 and the center of gravity D2 of the sub-hole 14K2 (center-to-center distance) is Δx in the x direction and Δy (constant) in the y direction.

[0057] In this specification, the "distance between the centers of gravity" of a main hole and a subhole refers to the distance between the central axis of the main hole and the central axis of the subhole, and is expressed as the distance in the x and y directions that the subhole is spaced from the main hole.

[0058] The main hole 14K1 and the subhole 14K2 had a long hexagonal column shape with the long axis parallel to the <11-20> axis. The main hole 14K1 and the subhole 14K2 had a {1-102} facet at their bottoms (on the substrate 12 side), but the rest of the holes had a long hexagonal column shape.

[0059] Specifically, the distances Δx and Δy between the center of gravity D1 of the main hole 14K1 and the center of gravity D2 of the subhole 14K2 were both 65.4 nm (Δx = Δy = 0.4 × PC), and did not change before filling. In addition, the major axes of the main hole 14K1 and the subhole 14K2 were arranged to be parallel to the <11-20> axis (i.e., the a-axis).

[0060] In this specification, the "long axis or short axis" of a hole (or a void) refers to the long axis or short axis of the cross section (opening surface) of the hole in a plane parallel to the photonic crystal layer.

[0061] The percentage of the hole area as viewed from a direction perpendicular to the plane of photonic crystal layer 14P divided by the square of the hole period PC is called the hole filling factor FF. The hole filling factors FF1 and FF2 of main hole 14K1 and subhole 14K2 were calculated to be FF1=10.5% and FF2=5.1%. 3. Device characteristics and evaluation 3.1 Double lattice structure and single lattice structure (1) Comparative Example 1 As Comparative Example 1 for the PCSEL device 10 according to this embodiment, a PCSEL device in which a single hole (single lattice photonic crystal) was formed at a lattice point was fabricated. The only difference from the above-mentioned manufacturing process was S3b (hole formation process), and this point will be explained.

[0062] 7 is an SEM image showing cylindrical holes CH formed on the GaN surface portion in Comparative Example 1. Specifically, the holes CH have a perfect circular opening with a diameter of 80 nm, and are arranged at square lattice points with a period Pc = 164 nm in the x and y directions.

[0063] More specifically, a resist pattern was formed so that the x-axis, which is the hole arrangement direction, was parallel to the <1-100> axis (i.e., the m-axis) and the y-axis was parallel to the <11-20> axis (i.e., the a-axis). Then, dry etching was performed using an ICP-RIE device to form cylindrical holes 14C arranged in a square lattice in the x and y directions. The cylindrical holes 14C had a diameter of 79 nm and a period PC of 164 nm. (2) Threshold Gain (Coupled Wave Theory) From the electric field intensity distribution of the fundamental mode in the structures of Example 1 and Comparative Example 1, the optical confinement coefficient (Γact) of the active layer and the optical confinement coefficient (Γmg) of the Mg-doped layers (p-contact layer, p-cladding layer, and electron barrier layer) were estimated. In addition, the cavity loss (α p ), the cavity loss in the direction perpendicular to the photonic crystal layer 14P (α n These are shown in Table 1.

[0064] In both Example 1 and Comparative Example 1, the absorption coefficient αmg of the Mg-doped layer was 160 cm -1 From Γmg and αmg, the absorption loss αi in each structure that satisfies the following (Equation 1) was calculated as the value shown in Table 1.

[0065] αi = Γmg × αmg (Equation 1) The threshold gain Gth of the laser is calculated by (Equation 2), and as shown in Table 1, in Example 1, it is 1119 cm -1 , and 724 cm in Comparative Example 1. -1 It was estimated that:

[0066] Γact Gth = α p +α n + αi (Equation 2)

[0067] [Table 1]

[0068] In Table 1, α of Example 1 n is α of Comparative Example 1 n This is because the double lattice structure breaks the rotational symmetry of the lattice point structure. In Comparative Example 1, the lattice point structure has two-fold rotational symmetry, so that light that is diffracted in the vertical direction among the light propagating through the hole layer is cancelled out by evanescent interference. In Example 1, the rotational symmetry is reduced, so this evanescent interference is weakened and more light is diffracted in the vertical direction. That is, α n In other words, α n In order to increase the lattice point structure, it is desirable to make the lattice point structure rotationally symmetrical once, and the vacancy set of the main vacancy 14K1 and the sub-vacancy 14K2 may be shaped or arranged so that they coincide with each other after a 360° rotation. (3) Light output characteristics, emission spectrum The IL characteristics (current-light output characteristics) of the PCSEL device 10 of Example 1 and the PCSEL device of Comparative Example 1 are shown in Fig. 8A, and the emission spectra near the threshold current are shown in Fig. 8B. The measurements were performed using a pulse current drive with a pulse width of 100 ns and a pulse period of 1 kHz.

[0069] The PCSEL device 10 of Example 1 had a threshold current of 1.24 A (threshold current density: 3.9 kA / cm 2 ), and a strongly unimodal laser oscillation was observed at a threshold current of 0.71 A (threshold current density: 2.2 kA / cm2). On the other hand, the PCSEL element of Comparative Example 1 exhibited a strongly unimodal laser oscillation at a threshold current of 0.71 A (threshold current density: 2.2 kA / cm2). The reason why the threshold current of Example 1 is larger than that of Comparative Example 1 is due to an increase in the threshold gain Gth in Table 1. In Example 1 and Comparative Example 1, the rate of increase in the threshold gain Gth and the rate of increase in the threshold current are approximately the same, and it is believed that the increase in the threshold current is due to an increase in the resonator loss and the absorption loss, and that even when a double lattice structure is introduced as in Example 1, there is no impact on the quality of the active layer.

[0070] On the other hand, the slope efficiency of the PCSEL device 10 of Example 1 is 0.23 W / A, which is significantly higher than 0.10 W / A of Comparative Example 1. This is because, as described above, in Example 1, α n This is because the vertical leakage of light (i.e. the light component that contributes to output) is large, so a large amount of light can be extracted as output. In other words, by making the lattice points in the hole layer into a double lattice structure, a large output can be obtained with a smaller current. 3.2 Arrangement of main and sub-holes in a double lattice structure EXAMPLES

[0071] (1) Example 2 and Comparative Example 2 In order to evaluate the surface shape after filling by introducing a double lattice structure, photonic crystal layers with two structures were fabricated, and the steps up to S3d were carried out in the same manner as in the above example to fill the holes.

[0072] More specifically, two structures were investigated in which the main voids and the sub-voids have the same size and shape, but the major axis directions of the main voids and the sub-voids differ by 90°.

[0073] 9A is a top SEM image showing main holes 14H1 and sub-holes 14H2 formed in a GaN surface portion in a structure (structure A) in which main holes and sub-holes are formed similarly to Example 1. Specifically, in structure A (Example 2), hole pairs each consisting of an elliptical cylindrical main hole 14H1 with a major axis / minor axis ratio of 2.75 (=79.8 nm / 29 nm) and an elliptical cylindrical sub-hole 14H2 with a major axis / minor axis ratio of 1.14 (=42.8 nm / 37.6 nm) are formed in a square lattice pattern with a period PC=164 nm.

[0074] 9B is a top-view SEM image showing main holes CH1 and sub-holes CH2 formed in a GaN surface portion in Comparative Example 2. In Comparative Example 2 (structure B), hole pairs each consisting of an elliptical cylindrical main hole CH1 with a major axis / minor axis ratio of 2.76 (=79.4 nm / 28.8 nm) and an elliptical cylindrical sub-hole CH2 with a major axis / minor axis ratio of 1.10 (=41.8 nm / 37.9 nm) are formed in a square lattice pattern with a period PC=164 nm.

[0075] In structure A (FIG. 9A), the major axes of the main hole 14H1 and the sub-hole 14H2 are arranged parallel to the <11-20> axis (i.e., the a-axis), whereas in structure B (FIG. 9B), the major axes of the main hole CH1 and the sub-hole CH2 are arranged perpendicular to the <11-20> axis (i.e., the a-axis).

[0076] That is, Structure A (Example 2) and Structure B (Comparative Example 2) have main holes and sub-holes of approximately the same size and shape, but the long axis direction of the holes in Structure B (Comparative Example 2) is 90° different from the long axis direction of the holes in Structure A (Example 2).

[0077] Therefore, by filling the main holes 14H1 and the sub-holes 14H2 of the structure A, a photonic crystal layer 14P is obtained in which the major axes of the main holes 14K1 and the sub-holes 14K2 are arranged parallel to the <11-20> axis (see FIG. 11A). By filling the main holes CH1 and the sub-holes CH2 of the structure B, a photonic crystal layer is obtained in which the major axes of the main holes CK1 and the sub-holes CK2 are arranged perpendicular to the <11-20> axis (see FIG. 11B). (2) Surface roughness of the buried layer 10A and 10B are atomic force microscope (AFM) images showing the surface morphology of the buried layers after filling the main voids and the sub-voids in the case of Structure A (Example 2) and Structure B (Comparative Example 2), respectively. Note that the horizontal axis of the AFM image is the m-axis, and the vertical axis is the a-axis.

[0078] As shown in FIG. 10A, when the long axis of each pore was parallel to the a-axis (<11-20> axis) (Structure A: Example 2), a flat surface with a surface roughness (RMS) of 0.594 nm was obtained.

[0079] On the other hand, as shown in FIG. 10B, when the long axis of each hole was perpendicular to the a-axis (Structure B: Comparative Example 2), undulations with a height of about 7 nm appeared on the surface, and the surface roughness (RMS 0.836 nm) was larger than that of Structure A (Example).

[0080] If an active layer is grown on a buried layer with such large irregularities on its surface, the In composition becomes non-uniform in the uneven areas, degrading the quality of the active layer. For this reason, the threshold current increases, so it is preferable that the long axis of the holes in the photonic crystal layer is parallel to the a-axis.

[0081] 11A and 11B are top views showing the change in the shape of the hole during the formation of the buried layer. With reference to these figures, the following will discuss the difference in surface roughness of the buried layer depending on whether the long axis of the hole is parallel to the a-axis of the crystal orientation or perpendicular to it.

[0082] When filling holes in group III nitrides, mass transport occurs, causing the hole shape to change to a shape composed of thermally stable planes, forming a vacancy. That is, in a +c-plane substrate, the side of the vacancy changes shape to a {1-100} plane (i.e., m-plane). That is, the shape changes from an oblong cylindrical shape to a long hexagonal prism whose side is composed of m-planes.

[0083] Figure 11A shows a schematic diagram of the shape change of the vacancy when the long axis of the long cylindrical hole is parallel to the a-axis (<11-20> axis) (structure A). The size of the holes 14H1 and 14H2 is reduced after deformation, but the Group III nitride atoms that make up the difference are supplied by mass transport from the surrounding crystal parts (arrows in the figure). Then, the shape changes to long hexagonal columnar vacancies 14K1 and 14K2 that are inscribed in the original shape of the holes 14H1 and 14H2.

[0084] On the other hand, Figure 11B shows a schematic diagram of the shape change of the voids when the long axis of the oval cylindrical hole is perpendicular to the a-axis (structure B). The size of the holes CH1 and CH2 shrinks due to mass transport (arrows in the figure). This shape change is larger than that of the hole whose long axis is parallel to the a-axis (Figure 11A).

[0085] In a multiple lattice photonic crystal, the distance between the holes is shorter than that in a single lattice photonic crystal, so adjacent holes (i.e., main and sub holes) are arranged at a distance of about the size of the holes. Therefore, when the shape of the holes changes due to mass transport, the shape changes of the adjacent holes interfere with each other. Large irregularities (height differences) are formed on the embedded surface between the parts where the shape changes interfere with each other and the parts where they do not. Therefore, as shown in Figure 11B, when the long axis of the hole and the a-axis are perpendicular to each other, a large shape change is accompanied, resulting in surface roughness. On the other hand, when the long axis of the hole and the a-axis are parallel, it is believed that the shape change can be minimized and a flat surface can be obtained after embedding. (3) Positional relationship between main and sub-holes and cavity loss When the air holes are arranged in a square lattice pattern in the air hole layer (photonic crystal layer), it is preferable that the air holes have the same feedback effect (diffraction efficiency) in the x-axis and y-axis directions in order to prevent direction-dependent uneven distribution of light within the two-dimensional photonic crystal and to stably obtain laser oscillation at a single wavelength.

[0086] That is, it is preferable that the two-dimensional photonic crystal forming the hole layer is symmetrical with respect to the axes inclined at 45° from the x-axis and y-axis of the square lattice. When the main hole 14K1 and the sub-hole 14K2 have a symmetrical structure with respect to the axes inclined at 45° from the x-axis and y-axis, it is preferable that the distances Δx and Δy between the centers of gravity of the main hole 14K1 and the sub-hole 14K2 in the x-axis and y-axis directions are equal (Δx=Δy).

[0087] In the structure of Example 1, the distances (spacing) Δx and Δy between the centers of gravity of main holes 14K1 and sub-holes 14K2 that form photonic crystal layer 14P are changed from 0.0PC to 0.5PC (PC is the period of a square lattice) with Δx=Δy, and the resonator losses α in the vertical and horizontal directions with respect to photonic crystal layer 14P are n ,α p 12A and 12B, in which the horizontal axis indicates the ratio d to the period PC (i.e., d=Δx / PC=Δy / PC).

[0088] FIG. 12C shows the R calculated from the following (Equation 3): n That is, R n is the total cavity loss (α p +α n ) versus the vertical cavity loss α n is the percentage.

[0089] R n = α n / (α p +α n ) (Formula 3) In a photonic crystal surface emitting laser, the slope efficiency ηSE is calculated by the fraction of the total loss, including the absorption loss αi, as shown in Equation 4. n (vertical cavity loss).

[0090]

number

[0091] Therefore, if the absorption loss αi due to the constituent materials could be made zero, ηSE would be R n In other words, in order to increase the emission efficiency (i.e., ηSE) of a photonic crystal surface emitting laser and obtain a photonic crystal surface emitting laser capable of achieving high output, R n It is desirable to increase

[0092] R in a conventional single-lattice photonic crystal surface-emitting laser (Comparative Example 1) n is about 0.18 from Table 1. In the double lattice photonic crystal surface emitting laser (Example 1), R n Conventional laser (R n =0.18), the relative position d (=Δx / PC=Δy / PC) of the subhole to the main hole is preferably 0.06 or more or 0.47 or less, as shown in FIG. 12C. (4) Oscillation mode Referring to FIG. 12C, the distance (spacing) Δx, Δy (Δx=Δy) between the centers of gravity of the main hole 14K1 and the sub-hole 14K2 is 0.28×PC and 0.40×PC. n changes discontinuously. This is because the oscillation mode changes at the boundary between the centers of gravity.

[0093] Figure 13 shows the photonic band structure near the Γ point of a square lattice photonic crystal. In a photonic crystal surface emitting laser, the standing wave state of light at the band edge of the Γ point of the photonic band structure of the hole layer is used as the resonance effect. At the band edge of the Γ point, there are four band edge modes (indicated by black circles), A, B, C, and D, from the lowest frequency side. Laser oscillation is obtained in the mode with the lowest threshold gain (smallest resonator loss) among the four band edge modes.

[0094] In Example 1, the threshold gain (resonator loss) of each mode obtained from the coupled-wave theory when the distances Δx and Δy between the centers of gravity are changed is shown in FIG. 14. It can be seen that for 0.28×PC or less (d≦0.28), mode A; for 0.28PC to 0.40PC (0.28<d<0.40), mode D; and for 0.40PC or more (0.40≦d), mode B becomes the mode with the smallest threshold gain.

[0095] FIG. 15 schematically shows the refractive index, carrier density, and photon density in the current injection region during laser oscillation of a surface-emitting laser. In a group III nitride semiconductor, generally, due to the carrier plasma effect, the refractive index of the current injection region becomes low. On the other hand, in a semiconductor laser, since the stimulated emission rate is high and the carrier density is low at the center of the resonator where the photon density is high, and also because the temperature of the central part rises due to heat generation, the refractive index of the central part of the current injection region becomes higher than that of its surroundings.

[0096] FIG. 16A schematically shows the frequencies of the photonic band edges at the Γ point in the vicinity of the current injection region during current injection for the band-edge modes A and B, and FIG. 16B shows the same for the band-edge modes C and D.

[0097] The frequency ω is expressed as in (Equation 5) using the refractive index n, the speed of light c, and the wave number k.

[0098] ω = c / n × k (Equation 5) From (Equation 5), when the refractive index distribution changes as shown in FIG. 15 due to current injection (laser oscillation), the frequency profiles of the band-edge modes A, B, C, and D also change with the change in the refractive index (from a trapezoidal shape to a trapezoidal shape with a concave center), as shown in FIGS. 16A and 16B.

[0099] As shown in FIG. 16A, in the case of the band-edge modes A and B (hereinafter simply referred to as modes A and B), the oscillation mode frequency exists within the photonic band gap at the center of the current injection region. Therefore, the presence of photons at the center of the injection region is suppressed, and photons are generated in the region outside the center of the injection region.

[0100] 17A shows a schematic diagram of the refractive index, carrier density, and photon density in the current injection region when lasing in the band edge modes A and B. The solid lines show the case where the photonic band effect is taken into account, and the dashed lines show the case where the photonic band effect is not taken into account.

[0101] Taking the photonic band effect into consideration, as described above, in modes A and B, the photons are distributed over a wider region, and the photon density is flattened over the entire current region as the injection current increases.

[0102] Correspondingly, the carrier density and refractive index are also uniformed over the entire current injection region. Therefore, when oscillating in modes A and B, even when a large current is injected for high-output operation, the photon density distribution is flattened over the entire injection region as the current increases, and the emitted beam has a stable beam pattern.

[0103] On the other hand, as shown in Fig. 16B, in the case of the band-edge modes C and D, the oscillation mode frequency exists within the band gap in the region outside the center of the current injection region, and therefore the photons are localized in the center of the injection region.

[0104] 17B shows a schematic diagram of the refractive index, carrier density, and photon density in the current injection region when lasing in the band edge modes C and D. The solid lines show the case where the photonic band effect is taken into account, and the dashed lines show the case where the photonic band effect is not taken into account.

[0105] When the photonic band effect is taken into consideration, photons are concentrated and localized in the center of the current injection region, and therefore the carrier density and refractive index also vary significantly locally in the center of the current injection region, resulting in a non-uniform distribution in the current injection region.

[0106] For this reason, when oscillating in band edge modes C and D, if a large current is injected for high output operation, the photon density distribution becomes localized and concentrated as the current increases, and the emitted beam has an unstable beam pattern. In addition, when a large current is injected, the carrier density and refractive index distributions become significantly non-uniform, so the oscillating operation tends to become multi-mode, making it impossible to obtain stable oscillating operation.

[0107] As described above, in order to obtain stable oscillation when a high current is injected, it is preferable to oscillate in modes A and B, and it is preferable that the relative position d (=Δx / PC=Δy / PC) of the subhole 14K2 to the main hole 14K1 is 0.06PC to 0.28PC (0.06≦d≦0.28) or 0.40PC to 0.47PC (0.40≦d≦0.47). EXAMPLES

[0108] In Example 3, a PCSEL device 11 was examined that has a hole layer (photonic crystal layer) with a dual lattice structure consisting of two regular hexagonal prism-shaped main holes 14K1 and sub-holes 14K2. 1. Fabrication Process The process for fabricating the PCSEL device 10 of Example 3 will be described in detail below. Differences from the fabrication process of the above-mentioned Examples and major points will be described. The structure of the PCSEL device 10 and the layer configuration of the semiconductor structure layer 11 were the same as those shown in FIG. 1A. [S3b: Hole and air hole formation process] A clean surface was obtained by cleaning the substrate with a growth layer on which an n-type GaN layer was grown as a preparatory layer for n-guide layer 14. This n-type GaN growth layer was a preparatory layer for forming a layer consisting of lower guide layer 14A and photonic crystal layer 14P.

[0109] The substrate with the growth layer was washed to obtain a clean surface, and then a silicon nitride film (Si x N y) was formed using plasma CVD. A resist for electron beam lithography was applied on top of this by spin coating, and the resist was placed in an electron beam (EB) lithography device to pattern a two-dimensional periodic structure.

[0110] This resulted in patterning in which aperture pairs, each consisting of a substantially circular main aperture K1 and a sub-aperture K2 smaller than the main aperture K1, were two-dimensionally arranged on square lattice points with a period PC=164 nm within the resist plane (see FIG. 3).

[0111] More specifically, the main apertures K1 are arranged at square lattice points with their centers of gravity CD1 at a period PC=164 nm in two mutually orthogonal directions (x-direction and y-direction), and the sub-apertures K2 are arranged at square lattice points with their centers of gravity CD2 at a period PC=164 nm in the x-direction and y-direction.

[0112] The main opening K1 had a major axis of 76 nm and a minor axis of 66 nm (major axis / minor axis = 1.15), and the sub-opening K2 had a major axis of 59 nm and a minor axis of 51 nm (major axis / minor axis = 1.12). The major axis of each opening was patterned to be parallel to the <11-20> axis (i.e., the a-axis).

[0113] The patterning was performed so that the distance Δx=Δy between the centers of gravity of the main opening K1 and the sub-opening K2 in the x and y directions was 65.6 nm (=0.4×164 nm).

[0114] After developing the patterned resist, the Si x N y The film was selectively dry etched. As a result, main openings K1 and sub-openings K2 arranged at square lattice points with a period of 164 nm were formed in the Si x N y formed to penetrate the membrane.

[0115] Although the main opening K1 and the sub-opening K2 have an elliptical shape that is slightly deviated from a perfect circle, they may be formed to have a perfect circle shape.

[0116] Next, the resist is removed and the patterned Si x Ny Using the film as a hard mask, holes were formed on the GaN surface. A plurality of hole pairs 14H (main holes 14H1 and sub-holes 14H2) perpendicular to the GaN surface were formed by dry etching the GaN with a chlorine-based gas using an ICP-RIE apparatus. [S3c: Cleaning process] The substrate in which the holes 14H1 and 14H2 are formed is degreased and cleaned, and then the Si x N y The film was then removed. The SEM image of the GaN surface at this stage is shown in Figure 18.

[0117] 18, a plurality of hole pairs 14H (main holes 14H1 and sub-holes 14H2) were formed in a square lattice pattern, i.e., two-dimensionally on the square lattice points, with a period PC of 164 nm. The holes 14H1 and 14H2 are approximately cylindrical holes that open to the upper surface (GaN surface).

[0118] More specifically, the major axis LH1 of the opening of the main hole 14H1 on the GaN surface was 76 nm and the minor axis WH1 was 68 nm, i.e., RH1 (=major axis / minor axis) = 1.12, and the major axis LH2 of the opening of the sub-hole 14H2 was 59 nm and the minor axis WH2 was 53 nm, i.e., RH2 (=major axis / minor axis) = 1.11.

[0119] In this case, the distances in the x and y directions between the center of gravity CD1 of the main hole 14H1 and the center of gravity CD2 of the sub-hole 14H2 were 65.3 nm (=0.4 × PC), respectively. The major axes of the main hole 14H1 and the sub-hole 14H2 were arranged parallel to the <11-20> axis (i.e., the a-axis).

[0120] Although the main hole 14H1 and the sub-hole 14H2 have an elliptical cylindrical shape that is slightly deviated from a perfect cylindrical shape, they may be formed to have a perfect cylindrical shape. [S3d: Buried layer formation process] This substrate was again introduced into the reactor of the MOVPE apparatus, and ammonia (NH3) was supplied to raise the temperature to 950°C (first embedding temperature). Trimethylgallium (TMG) and NH3 were then supplied to block the opening of the main hole 14H1 and the sub-hole 14H2, thereby forming the first embedded layer 14B1.

[0121] In this temperature range, N atoms are attached to the top surface of the growth substrate, so N-polar faces are grown preferentially. Therefore, {1-101} facets are grown preferentially on the surface. When the opposing {1-101} facets collide with each other, the holes are blocked and buried in the GaN layer.

[0122] Next, after the main hole 14H1 and the sub-hole 14H2 were blocked, the second buried layer 14B2 was grown to a thickness of 50 nm. The second buried layer 14B2 was grown by lowering the substrate temperature to 820°C (second burying temperature) and then supplying triethylgallium (TEG) and trimethylindium (TMI) as a group 3 atom source and NH3 as a nitrogen source.

[0123] The In composition of the second buried layer 14B2 is 2% (i.e., Ga 0.98 In 0.02 Second embedded layer 14B2 functions as a light distribution adjusting layer for adjusting the coupling efficiency (light field) between light and photonic crystal layer 14P. [S4: Light-emitting layer formation process] Next, a multiple quantum well (MQW) layer was grown as the active layer 15, which is the light emitting layer. More specifically, as in the above-mentioned embodiment, the barrier layer and well layer of the MQW were GaN and InGaN, respectively. The barrier layer was grown by lowering the temperature of the substrate to 820°C, and then supplying triethylgallium (TEG) as a group 3 atom source and NH3 as a nitrogen source. The well layer was grown at the same temperature as the barrier layer, and by supplying TEG and trimethylindium (TMI) as group 3 atom sources and NH3 as a nitrogen source. The center wavelength of the PL emission from the active layer in this embodiment was 412 nm.

[0124] The steps from S5 (p-side guide layer formation step) onwards were the same as the fabrication steps described above, and the PCSEL device 10 of Example 3 was formed. 2. Main and secondary holes In order to confirm the shape of the air holes formed by filling in the examples, the laminated structure was processed from the surface by FIB until the air holes in the air hole layer were exposed, and then SEM observation was performed. The air holes of the main air hole 14K1 and the sub-air hole 14K2, which is smaller in size than the main air hole 14K1, are shown in the top SEM image of FIG. 19A. It is sufficient that the sub-air hole 14K2 is smaller than the main air hole 14K1 in at least one of the air hole diameter and the depth.

[0125] Moreover, FIG. 19B is an SEM image showing a cross section taken along line AA in FIG. 19A, and FIG. 19C is an SEM image showing a cross section taken along line BB in FIG. 19A.

[0126] 19A, it was confirmed that the main void 14K1 and the sub-void 14K2 had a hexagonal shape. In addition, SEM observation revealed that the major axis LK1 of the main void 14K1 was 57.2 nm, the minor axis WK1 was 49.5 nm, and the ratio of the major axis to the minor axis (major axis / minor axis) RK1 was 1.15. In addition, the depth DK1 of the main void 14K1 was 91.3 nm.

[0127] The major axis LK2 of the sub-hole 14K2 was 43.5 nm, the minor axis WK2 was 37.7 nm, and the ratio of the major axis to the minor axis (major axis / minor axis) RK2 was 1.15. The depth DK2 of the sub-hole 14K2 was 79.4 nm.

[0128] The ratio of the major axis to the minor axis of a regular hexagon is RR=2 / 3. 1 / 2 =1.15, it was confirmed from the ratio RK1 of the main hole 14K1 and the ratio RK2 of the subhole 14K2 that the main hole 14K1 and the subhole 14K2 have a regular hexagonal column shape.

[0129] Moreover, the distances (center-to-center distances) Δx and Δy between the center of gravity D1 of the main hole 14K1 and the center of gravity D2 of the subhole 14K2 were 65.4 nm (=0.4×PC), respectively, and did not change from before the filling.

[0130] The major axes of the main hole 14K1 and the subhole 14K2 were arranged parallel to the <11-20> axis (that is, the a-axis). 3. Device characteristics and evaluation (1) Comparative Example 1 and Comparative Example 2 In evaluating the PCSEL device 10 of Example 3, it was compared with the PCSEL device of Comparative Example 1 described above, which has a single lattice structure in which a single void is formed at a lattice point. It was also compared with the PCSEL device of Comparative Example 2 described above, in which the long axis directions of the main voids and sub-voids differ by 90° from those of Example 3. (2) Threshold Gain (Coupled Wave Theory) From the electric field intensity distribution of the fundamental mode in the structures of Example 3 and Comparative Example 1, the optical confinement coefficient (Γact) of the active layer and the optical confinement coefficient (Γmg) of the Mg-doped layers (p-contact layer, p-cladding layer, and electron barrier layer) were estimated. In addition, the cavity loss (α p ), the cavity loss in the direction perpendicular to the photonic crystal layer 14P (α n ) was estimated.

[0131] In both Example 3 and Comparative Example 1, the absorption coefficient α mg is 160cm -1 The optical confinement factor (Γmg) and absorption coefficient α mg From the above (Equation 1), the absorption loss α i The threshold gain Gth of the laser was calculated by the above formula (2), and as shown in Table 2, in Example 3, it was 1068 cm -1 , and 724 cm in Comparative Example 1. -1 These are summarized in Table 2.

[0132] [Table 2]

[0133] In Table 2, α of Example 3 n is α of Comparative Example 1 nThis is because the double lattice structure breaks the 90° rotational symmetry of the lattice point structure. In Comparative Example 1, since the lattice point structure has 90° rotational symmetry, the light propagating through the hole layer that is diffracted in the vertical direction is cancelled out by evanescent interference.

[0134] In the third embodiment, the 90° rotational symmetry is reduced, so this evanescent interference is weakened and more light is diffracted in the vertical direction. n is increasing.

[0135] In this way, it was confirmed that, also in Example 3 in which the main holes 14K1 and the subholes 14K2 have a regular hexagonal column shape, the optical component propagating in the vertical direction was significantly increased compared to a photonic crystal layer having a single lattice structure. (3) Light output characteristics, emission spectrum The IL characteristics (current-light output characteristics) of the PCSEL device 10 of Example 3 and the PCSEL device of Comparative Example 1 are shown in Fig. 20A, and the emission spectra near the threshold current are shown in Fig. 20B. The measurements were performed using a pulse current drive with a pulse width of 100 ns and a pulse period of 1 kHz.

[0136] The PCSEL device 10 of Example 3 had a threshold current of 1.21 A (threshold current density: 3.8 kA / cm 2 ), which exhibited strong single-peaked laser oscillation. On the other hand, the PCSEL element of Comparative Example 1 exhibited strong single-peaked laser oscillation at a threshold current of 0.71 A (threshold current density: 2.2 kA / cm2). The reason why the threshold current of Example 3 is larger than that of Comparative Example 1 is that the threshold gain Gth shown in Table 2 is increased.

[0137] In Example 3 and Comparative Example 1, the rate of increase in threshold gain Gth and the rate of increase in threshold current are approximately the same, and it is considered that the increase in threshold current is due to increases in resonator loss and absorption loss. Therefore, even when a double lattice structure is introduced as in Example 3, it is considered that there is no effect on the quality of the active layer.

[0138] On the other hand, the slope efficiency of the PCSEL device 10 of Example 3 is 0.35 W / A, which is significantly higher than 0.10 W / A of Comparative Example 1. This is because, as described above, in Example 3, the α n This is because the increase in the light propagating in the vertical direction is large, and therefore the amount of light extracted as output is large. In other words, by making the lattice structure in the hole layer a double lattice structure, it is possible to obtain a large output with a smaller current. (4) Positional relationship between main and sub-holes and resonator loss In Example 3, a PCSEL element was fabricated having a hole layer with a double lattice structure consisting of main holes 14K1 and sub-holes 14K2 each having a regular hexagonal prism shape with two lattice points.

[0139] In the structure of Example 3, the distances Δx, Δy (Δx=Δy) between the centers of gravity of the main hole 14K1 and the subhole 14K2 are changed in the range from 0.0PC to 0.5PC (PC is the period of the square lattice), and the resonator loss α in the vertical and horizontal directions with respect to the photonic crystal layer 14P is n ,α p 21A and 21B, where the horizontal axis indicates the ratio d of the center-of-gravity distances (the distance between the main hole and the subhole) Δx, Δy to the square lattice period PC (i.e., d=Δx / PC=Δy / PC).

[0140] FIG. 21C shows the R calculated from the above (Equation 3). n This shows the dependence of the distance between the centers of gravity on R n is the total cavity loss (α p +α n ) versus the vertical cavity loss α n is the percentage.

[0141] In addition, in a photonic crystal surface emitting laser, the slope efficiency ηSE is calculated by dividing the total loss, including the absorption loss αi, by the n (vertical cavity loss).

[0142] Therefore, if the absorption loss αi due to the constituent materials could be made zero, ηSE would be R nIn other words, in order to increase the emission efficiency (i.e., ηSE) of a PCSEL device and obtain a PCSEL device capable of achieving high output, R n It is desirable to increase

[0143] R in a conventional PCSEL device with a single grating structure (Comparative Example 1) n is about 0.18 according to Table 2. In the PCSEL device of Example 3 having a double lattice structure consisting of main and sub-holes in the shape of a regular hexagonal prism, R n The single-lattice structure PCSEL element (R n =0.18), the distance between the centers of gravity of the main hole 14K1 and the sub-hole 14K2 (relative position d) (=Δx / PC=Δy / PC) is preferably 0.06 or more or 0.47 or less, as shown in FIG. 21C.

[0144] It is considered possible to break the four-fold (90°) rotational symmetry of the spatial refractive index distribution within the lattice points by making the air hole shape a long hexagonal prism with a larger long axis / short axis ratio. However, even with a regular hexagonal prism-shaped air hole with a small long axis / short axis ratio of 1.15, a high R n It was found that it is possible to obtain

[0145] However, in PCSEL devices using GaN-based materials, it is preferable to change the shape of the side surface of the voids during growth to a thermally stable m-plane in order to stably fill the voids. In other words, it is preferable that the major axis / minor axis ratio of the voids forming the void layer is at least 1.15. (5) Oscillation mode FIG. 21D shows the threshold gain (resonator loss) of each mode obtained from the coupled wave theory when the relative position d (=Δx / PC=Δy / PC) of the subhole 14K2 with respect to the main hole 14K1 is changed in the third embodiment.

[0146] When the relative position d is 0.24×PC or less (d≦0.24), it is mode A; when it is 0.24PC to 0.28PC (0.24<d≦0.28), it is mode B; when it is 0.28PC to 0.34PC (0.28<d<0.34), it is mode C; when it is 0.34PC to 0.40PC (0.34≦d<0.40), it is mode D; when it is 0.40PC or more (0.40≦d), it can be seen that mode B becomes the mode with the smallest threshold gain.

[0147] As described with reference to FIGS. 16A and 16B, in order to obtain stable oscillation when a high current is injected, it is preferable to oscillate in modes A and B. (6) Hole filling ratio (FF: filling factor) In the structure of Example 3, with the hole filling ratio FF2 of the secondary holes 14K2 fixed at 4.5%, the resonator loss when the hole filling ratio FF1 of the primary holes 14K1 was varied in the range from 4.5% to 40% was obtained by two-dimensional coupled-wave theory. R with respect to the hole filling ratio ratio FF1 / FF2 n , the sum of the resonator losses in the vertical and horizontal directions (α n +α p ) are shown in FIGS. 22A and 22B, respectively.

[0148] As described above, in this specification, the hole filling ratio is the ratio of the area occupied by each hole per unit area of a two-dimensional periodic array. Specifically, assuming that the areas of the primary holes 14K1 and secondary holes 14K2 in the hole layer are S1 and S2, respectively, in the case of a square lattice (period PC), the hole filling ratio FF1 of the primary holes 14K1 = S1 / PC 2 , and the hole filling ratio FF2 of the secondary holes 14K2 = S2 / PC 2 .

[0149] Generally, the photonic band structure near the Γ point of a square lattice photonic crystal is as shown in FIG. 13 described above, and there are four band-edge modes. As shown in FIGS. 22A and 22B, as the hole filling ratio ratio FF1 / FF2 changes, the band-edge mode with the smallest loss changes.

[0150] As described in detail in Examples 1 and 2, in order to obtain stable oscillation when a high current is injected, it is preferable to oscillate in modes A and B. That is, in order to oscillate in modes A and B, it is preferable that the hole-filling ratio RF=FF1 / FF2 is in the range of 1.7 to 7.5, that is, 1.7≦(FF1 / FF2)≦7.5, as shown in Figures 22A and 22B.

[0151] In the above-described embodiments, the main holes 14K1 and the sub-holes 14K2 have a long hexagonal prism shape and a regular hexagonal prism shape. However, the sub-holes 14K2 may have a cylindrical shape with a perfectly circular cross section, a regular hexagonal prism shape, or the like.

[0152] Although the case where the main hole 14K1 after embedding has a long hexagonal column shape and a regular hexagonal column shape has been described, the present invention is not limited thereto. The main hole 14K1 may have an oval column shape, or may have an intermediate shape in the process in which the hole changes from a cylindrical shape or an oval column shape to a regular hexagonal column shape or a long hexagonal column shape due to embedding growth. Therefore, in this specification, "regular hexagonal column shape, long hexagonal column shape, or long hexagonal column shape" includes an intermediate shape in the process in which the hole changes from a cylindrical shape or an oval column shape to a regular hexagonal column shape or a long hexagonal column shape.

[0153] Furthermore, when the primary void 14K1 has a regular hexagonal prism shape, a long hexagonal prism shape, or an oval cylinder shape whose major axis is parallel to the <11-20> axis, the secondary void 14K2 may have a regular hexagonal prism shape, a long hexagonal prism shape, or an oval cylinder shape whose major axis is parallel to the <11-20> axis.

[0154] In the above embodiment, the photonic crystal layer 14P is described in which a hole pair consisting of a main hole 14K1 and a subhole 14K2 is arranged at each square lattice point. However, a multi-lattice photonic crystal layer may be formed in which a hole set consisting of a main hole 14K1, a subhole 14K2 and at least one hole is arranged at each square lattice point.

[0155] The numerical values ​​in the above-mentioned embodiment are merely examples and can be appropriately modified and applied.

[0156] As described above in detail, according to the above-described embodiment, it is possible to provide a photonic crystal surface-emitting laser (PCSEL) element and a manufacturing method thereof in which the flatness of the surface of the embedding layer in which the multilattice photonic crystal is embedded is greatly improved.

[0157] In addition, it is possible to provide a photonic crystal surface-emitting laser having an active layer grown on a multilattice photonic crystal layer of high quality and crystallinity, with high light extraction efficiency, and capable of oscillating at a low threshold current density and high quantum efficiency, and a method for manufacturing the same. [Explanation of symbols]

[0158] 10: PCSEL element, 11: semiconductor structure layer, 12: substrate, 13: first cladding layer, 14: first guide layer, 14A: lower guide layer, 14P: photonic crystal layer (PC layer), 14B: buried layer, 15: active layer, 16: second guide layer, 17: electron barrier layer, 18: second cladding layer, 19: contact layer, 20A: first electrode, 20B: second electrode, 20L: light emission region, 25: second buried layer, CD1, CD2: center of gravity, K1 / K2: main / secondary opening, 14H1 / 14H2: main / secondary hole, 14K1 / 14K2: main / secondary hole

Claims

1. A surface emitting laser element made of a group III nitride semiconductor, a first guide layer including: a photonic crystal layer formed on a c-plane of a group III nitride semiconductor and having holes arranged with two-dimensional periodicity in a plane parallel to the layer; and a burying layer formed on the photonic crystal layer and closing the holes; an active layer formed on the first guide layer; a second guide layer formed on the active layer; a hole set including at least a main hole and a subhole having a hole diameter and a depth smaller than those of the main hole is disposed at each of square lattice points in a plane parallel to the photonic crystal layer, The main holes have a regular hexagonal prism shape, an elongated hexagonal prism shape, or an elongated cylindrical shape, with the major axis parallel to the <11-20> axis.

2. 2. The surface-emitting laser element according to claim 1, wherein the sub-hole has a regular hexagonal prism shape, a long hexagonal prism shape, or an oval cylinder shape, the long axis of which is parallel to the <11-20> axis.

3. 2. The surface-emitting laser element according to claim 1, wherein the main hole and the sub-hole have a regular hexagonal prism shape with a major axis parallel to a <11-20> axis.

4. 4. The surface-emitting laser element according to claim 1, wherein the sub-hole is disposed at a position such that the center of gravity of the sub-hole is spaced apart in the <1-100> direction from the center of gravity of the main hole.

5. 4. The surface-emitting laser element according to claim 1, wherein the hole sets are arranged in x and y directions perpendicular to each other and at angles of 45° with respect to a <11-20> axis and a <1-100> axis, and the relative positions Δx and Δy of the subholes with respect to the main hole satisfy Δx = Δy, and when a period of the square lattice is PC and Δx = Δy = d × PC, 0.06≦d≦0.28 or 0.40≦d≦0.47 is satisfied.

6. 6. The surface-emitting laser element according to claim 1, wherein the main hole and the sub-hole have a hexagonal prism shape with side surfaces being m-planes.

7. 7. The surface-emitting laser element according to claim 1, wherein when the hole filling rates of the main holes and the subholes in the photonic crystal layer are FF1 and FF2, respectively, a hole filling rate ratio RF=FF1 / FF2 satisfies 1.7≦RF≦7.

5.

8. the guide layer is a GaN layer, 8. The surface-emitting laser element according to claim 1, wherein the burying layer comprises a first burying layer which is a GaN layer, and a second burying layer formed on the first burying layer and containing In (indium) in its composition.

9. A method for manufacturing a surface emitting laser element made of a group III nitride semiconductor, comprising the steps of: forming a guide layer on a c-plane of a Group III nitride semiconductor; forming an etching mask on the guide layer, the etching mask having an opening set including at least a main opening and a sub-opening having a size smaller than that of the main opening at each of the square lattice points; Etching the guide layer using the etching mask to form main holes and sub-holes; forming a buried layer that closes the openings of the main holes and the sub-holes by performing crystal growth including mass transport, and forming a multi-lattice photonic crystal layer in which a hole set including a main hole and a sub-hole having a hole diameter and depth smaller than those of the main hole is arranged at each of the square lattice points; forming a semiconductor layer including an active layer on the multilattice photonic crystal layer; The main holes have a regular hexagonal prism shape, a long hexagonal prism shape, or a long cylindrical shape with their major axes parallel to the <11-20> axis.

10. 10. The method for manufacturing a surface-emitting laser element according to claim 9, wherein the sub-hole has a regular hexagonal prism shape, a long hexagonal prism shape, or an oval cylinder shape, the long axis of which is parallel to the <11-20> axis.

11. A surface emitting laser element made of a group III nitride semiconductor, a first guide layer including: a photonic crystal layer formed on a c-plane of a group III nitride semiconductor and having holes arranged with two-dimensional periodicity in a plane parallel to the layer; and a burying layer formed on the photonic crystal layer and closing the holes; an active layer formed on the first guide layer; a second guide layer formed on the active layer; a hole set including at least a main hole and a subhole having a size smaller than that of the main hole is disposed at each of square lattice points in a plane parallel to the photonic crystal layer, The main hole and the sub-hole have a regular hexagonal column shape with the major axis parallel to the <11-20> axis.

Images