JP2014195069A - Semiconductor light-emitting element, manufacturing method of the same and optical base material - Google Patents

Abstract

A semiconductor light emitting device with improved internal quantum efficiency and light extraction efficiency and an optical substrate for manufacturing the same are provided.
A semiconductor light emitting device (100) includes a first semiconductor layer (30) and a light emitting semiconductor layer (30) on an optical substrate (10) having a fine concavo-convex structure (20) formed on a part or the entire surface thereof. 40) and the second semiconductor layer (50) on the surface on which the average pitch Pav of the fine relief structure (20) and the fine relief structure (20) of the optical substrate (10) are formed. The ratio (Pav / λ) of the emitted light emitted from the light emitting semiconductor layer (40) in the first semiconductor layer (30) to the wavelength λ is 2.5 or more and 8 or less. 20) The ratio of the in-plane average height Hav to the wavelength λ (Hav / λ) of the emitted light is 0.1 or more and 1 or less, and the aspect ratio of the fine concavo-convex structure (20) is 0.2 or more and 1 It is as follows.
[Selection] Figure 1

Description

  The present invention relates to a semiconductor light emitting device, a manufacturing method thereof, and an optical substrate.

  In recent years, various studies have been made in order to improve the light emission efficiency of semiconductor light emitting devices such as organic electroluminescence (OLED), phosphor, and light emitting diode (LED). Such a semiconductor light emitting element has a configuration in which a high refractive index region including a light emitting portion is sandwiched between low refractive index regions. For this reason, the emitted light emitted from the light emitting portion of the semiconductor light emitting element becomes a waveguide mode that guides the inside of the high refractive index region, is confined inside the high refractive index region, is absorbed in the waveguide process, and is attenuated as heat. As described above, in the semiconductor light emitting device, the emitted light cannot be extracted outside the semiconductor light emitting device, and there is a problem that the light emission efficiency is greatly reduced.

  In the case of the LED element, as described below, a highly efficient LED element can be manufactured by improving the light extraction efficiency LEE and the internal quantum efficiency IQE or the light extraction efficiency LEE and the electron injection efficiency EIE at the same time.

A GaN-based semiconductor element typified by a blue LED is manufactured by laminating an n layer, a light emitting layer, and a p layer on a single crystal substrate by epitaxial growth. As the single crystal substrate, a sapphire single crystal substrate or a SiC single crystal substrate is generally used. However, since there is a lattice mismatch between the sapphire crystal and the GaN-based semiconductor crystal, dislocations are generated inside the GaN-based semiconductor crystal (see, for example, Non-Patent Document 1). This dislocation density reaches 1 × 10 ⁹ pieces / cm ² . By this dislocation, the internal quantum efficiency of the LED, that is, the efficiency of light emission of the semiconductor decreases, and as a result, the external quantum efficiency decreases.

  The refractive index of the GaN-based semiconductor layer is larger than that of the sapphire substrate. For this reason, the light generated in the semiconductor light emitting layer is not emitted at an angle greater than the critical angle from the interface between the sapphire substrate and the GaN-based semiconductor layer. That is, a waveguide mode is formed, and heat is attenuated in the waveguide process. For this reason, the light extraction efficiency is lowered, and as a result, the external quantum efficiency is lowered. In addition, when a SiC substrate having a very large refractive index is used as the single crystal substrate, light emission at an angle greater than the critical angle from the interface between the SiC substrate and the air layer does not occur. In addition, a waveguide mode is generated, and the light extraction efficiency LEE decreases.

  That is, the internal quantum efficiency is reduced due to dislocation defects inside the semiconductor crystal, and the light extraction efficiency is reduced due to the formation of the waveguide mode, so that the external quantum efficiency of the LED is greatly reduced.

  In view of this, a technique has been proposed in which a concavo-convex structure is provided on a single crystal substrate and the light guide direction of the semiconductor crystal layer is changed to increase the light extraction efficiency (see, for example, Patent Document 1).

  In addition, a technique has been proposed in which the size of the concavo-convex structure provided on the single crystal substrate is nano-sized, and the pattern of the concavo-convex structure is randomly arranged (see, for example, Patent Document 2). Note that it has been reported that when the pattern size provided on the single crystal substrate is nano-sized, the light emission efficiency of the LED is improved as compared to a micro-sized pattern base material (for example, see Non-Patent Document 2).

JP 2003-318441 A JP 2007-294972 A

IEEE photo. Tech. Lett. , 20, 13, 1193 (2008) J. et al. Appl. Phys. , 103, 014314 (2008)

  In general, factors that determine external quantum efficiency EQE (External Quantum Efficiency) indicating the light emission efficiency of an LED include electron injection efficiency EIE (Electron Injection Efficiency), internal quantum efficiency IQE (Internal Quantum Efficiency E light efficiency) Extraction Efficiency). Among these, the internal quantum efficiency IQE depends on the dislocation density caused by the crystal mismatch of the GaN-based semiconductor crystal. The light extraction efficiency LEE is improved by breaking the waveguide mode inside the GaN-based semiconductor crystal layer by light scattering or light diffraction by the concavo-convex structure provided on the single crystal substrate.

  In the prior art, the light emission efficiency of the semiconductor light emitting element is not maximized. In the technique described in Patent Document 1, the light extraction efficiency LEE is improved by the effect (2), but the effect of improving the internal quantum efficiency IQE by the effect (1) is small. This is because dislocation defects are reduced by unevenness on the single crystal substrate because the chemical vapor deposition (CVD) growth mode of the GaN-based semiconductor layer is disturbed by the unevenness, and dislocation defects accompanying layer growth collide and disappear. It is. For this reason, it is effective to improve the internal quantum efficiency IQE because the size of the unevenness has a nano-order pitch, which has a great effect on the reduction of the dislocation density.

  However, in the conventional technology, the concavo-convex structure is not an optimal structure for the internal quantum efficiency IQE and the light extraction efficiency LEE necessary for maximizing the light emission efficiency of the LED.

  The present invention has been made in view of the above-described problems. A semiconductor light-emitting device having improved both internal quantum efficiency and light extraction efficiency, a method for manufacturing the same, and an optical for manufacturing the semiconductor light-emitting device An object is to provide a substrate.

  As a result of intensive studies to solve the above problems, the present inventor has found that both the internal quantum efficiency IQE and the light extraction efficiency LEE are obtained when the fine uneven structure of the nanopattern provided on the optical substrate satisfies a specific condition. It has been found that the luminous efficiency of the semiconductor light emitting device is maximized. That is, the present invention is as follows.

  The semiconductor light-emitting device of the present invention has a structure in which a first semiconductor layer, a light-emitting semiconductor layer, and a second semiconductor layer are stacked on an optical substrate having a fine relief structure formed on a part or the entire surface. The average pitch Pav of the fine concavo-convex structure and the wavelength of the emitted light emitted by the light-emitting semiconductor layer in the optical medium layer provided on the surface of the optical substrate on which the fine concavo-convex structure is formed The ratio (Pav / λ) to λ is 2.5 or more and 8 or less, and the ratio (Hav / λ) of the in-plane average height Hav of the fine concavo-convex structure to the wavelength λ of the emitted light is 0.1 or more. The aspect ratio of the fine concavo-convex structure is 0.2 or more and 1 or less.

  With this configuration, both the internal quantum efficiency IQE and the light extraction efficiency LEE are improved, and the light emission efficiency of the semiconductor light emitting device is maximized. Further, the wave nature of light is emphasized, and the emitted light has directivity. This increases the degree of freedom in designing the electrode area and material.

In the semiconductor light emitting device of the present invention, it is preferable that the wavelength λ of the emitted light is obtained by dividing the vacuum wavelength λ ₀ of the emitted light by the refractive index n of the first semiconductor layer.

  In the semiconductor light emitting device of the present invention, it is preferable that the optical medium layer is the first semiconductor layer.

  In the semiconductor light emitting device of the present invention, it is preferable that the bottom of the concave portion constituting the fine concavo-convex structure is a flat surface.

  In the semiconductor light emitting device of the present invention, it is preferable that the convex part constituting the fine concavo-convex structure has a smaller diameter at the top part than at the bottom part.

  In the semiconductor light emitting device of the present invention, the optical base material is preferably a sapphire substrate, a Si substrate, a SiC substrate, a GaN substrate, a GaAs substrate, or a spinel substrate.

  In the semiconductor light emitting device of the present invention, it is preferable that the optical base material is a c-plane sapphire substrate, and an offset angle is 0.05 degree or more and 0.8 degree or less.

  In the semiconductor light emitting device of the present invention, it is preferable that the first semiconductor layer, the light emitting semiconductor layer, and the second semiconductor layer are III-V group semiconductors.

  In the semiconductor light emitting device of the present invention, it is preferable that the first semiconductor layer, the light emitting semiconductor layer, and the second semiconductor layer are GaN-based semiconductors.

  A method for manufacturing a semiconductor light emitting device according to the present invention is a method for manufacturing a semiconductor light emitting device as described above, wherein the step of preparing the optical substrate and the surface of the optical substrate on which the fine concavo-convex structure is formed are provided. And sequentially stacking at least the first semiconductor layer, the light emitting semiconductor layer, and the second semiconductor layer.

  The optical substrate of the present invention is an optical substrate having a fine concavo-convex structure formed on a part or the entire surface thereof, which is used in a semiconductor light emitting device having a structure in which a first semiconductor layer, a light emitting semiconductor layer and a second semiconductor layer are laminated. A ratio between an average pitch Pav of the fine concavo-convex structure and a wavelength λ of emitted light emitted by the light-emitting semiconductor layer in an optical medium layer provided on a surface on which the fine concavo-convex structure is formed. (Pav / λ) is 2.5 or more and 8 or less, and the ratio (Hav / λ) of the in-plane average height Hav of the fine concavo-convex structure to the wavelength λ of the emitted light is 0.1 or more and 1 or less. The aspect ratio of the fine concavo-convex structure is 0.2 or more and 1 or less.

In the optical substrate of the present invention, the wavelength λ of the emitted light is preferably obtained by dividing the vacuum wavelength λ ₀ of the emitted light by the refractive index n of the first semiconductor layer.

  In the optical base material of the present invention, it is preferable that the optical medium layer is the first semiconductor layer.

  In the optical base material of the present invention, it is preferable that the bottom of the concave portion constituting the fine concavo-convex structure is a flat surface.

  In the optical base material of the present invention, it is preferable that the convex part constituting the fine concavo-convex structure has a smaller diameter at the top than at the bottom.

  In the optical base material of the present invention, a sapphire substrate, Si substrate, SiC substrate, GaN substrate, GaAs substrate or spinel substrate is preferable.

  In the optical base material of the present invention, it is a c-plane sapphire substrate, and the offset angle is preferably 0.05 degrees or more and 0.8 degrees or less.

  The method for producing a semiconductor light-emitting device of the present invention includes a step of preparing the above-described optical base material, and at least the first semiconductor layer and the light-emitting semiconductor layer on the surface of the optical base material on which the fine concavo-convex structure is formed. And sequentially stacking the second semiconductor layers.

  According to the present invention, there are provided a semiconductor light emitting device that improves both internal quantum efficiency and light extraction efficiency, and further has a high degree of freedom in designing an electrode portion, a method for manufacturing the same, and an optical substrate for manufacturing the semiconductor light emitting device. Can be provided.

1 is a schematic cross-sectional view showing a semiconductor light emitting element according to an embodiment. It is a schematic diagram for demonstrating the diffraction azimuth angle of the diffracted light in the semiconductor light-emitting device concerning this Embodiment. It is a schematic diagram for demonstrating the relationship between the diffraction azimuth angle and electrode part which concerns on this Embodiment and in the conventional semiconductor light-emitting device. It is a schematic diagram which shows the example of the arrangement | sequence of the continuous convex part in the semiconductor light-emitting device concerning this Embodiment. It is a schematic diagram which shows the example of the arrangement | sequence of the continuous convex part in the semiconductor light-emitting device concerning this Embodiment. It is the top view which looked at the 1st semiconductor layer of the semiconductor light-emitting device concerning this Embodiment from the surface side in which the fine concavo-convex structure was formed. It is the top view which looked at the 1st semiconductor layer of the semiconductor light-emitting device concerning this Embodiment from the surface side in which the fine concavo-convex structure was formed. It is a top view which shows the case where the fine grooving | roughness structure of the semiconductor light-emitting device concerning this Embodiment is a dot structure. It is a cross-sectional schematic diagram which shows the fine concavo-convex structure in the line segment position corresponded to the pitch P shown in FIG. It is a top view which shows the case where the fine concavo-convex structure of the semiconductor light emitting element concerning this Embodiment is a hole structure. It is a cross-sectional schematic diagram which shows the fine concavo-convex structure in the line segment position corresponded to the pitch P shown in FIG. It is a cross-sectional schematic diagram which shows the sapphire substrate which has an offset angle which concerns on this Embodiment. It is explanatory drawing explaining the sapphire substrate which has an offset angle and a fine uneven structure. It is explanatory drawing explaining the sapphire substrate which has an offset angle and a fine uneven structure. It is a cross-sectional schematic diagram which shows the other example of the semiconductor light-emitting device concerning this Embodiment.

  Hereinafter, an embodiment of the present invention (hereinafter abbreviated as “embodiment”) will be described in detail. In addition, this invention is not limited to the following embodiment, It can implement by changing variously within the range of the summary.

  First, the structure of the semiconductor light emitting device according to the present embodiment will be described with reference to the drawings. FIG. 1 is a schematic cross-sectional view showing a semiconductor light emitting device according to the present embodiment. As shown in FIG. 1, in the semiconductor light emitting device 100, the optical substrate 10 has a fine concavo-convex structure 20 on the surface thereof. A first semiconductor layer 30, a light emitting semiconductor layer 40, and a second semiconductor layer 50 are sequentially stacked on the surface of the optical substrate 10 including the fine concavo-convex structure 20. Here, the emitted light generated in the light emitting semiconductor layer 40 is extracted from the second semiconductor layer 50 side or the optical substrate 10. Further, the first semiconductor layer 30 and the second semiconductor layer 50 are different semiconductor layers. Here, it is preferable that the first semiconductor layer 30 planarizes the fine concavo-convex structure 20. By providing the first semiconductor layer 30 so as to planarize the fine concavo-convex structure 20, the performance of the first semiconductor layer 30 as a semiconductor can be reflected in the light emitting semiconductor layer 40 and the second semiconductor layer 50. Therefore, the internal quantum efficiency IQE is improved.

  The first semiconductor layer 30 may be composed of an undoped first semiconductor layer 31 and a doped first semiconductor layer 32, as shown in FIG. Here, the undoped first semiconductor layer 31 is provided so as to planarize the fine concavo-convex structure 20, so that the performance of the undoped first semiconductor layer 31 as a semiconductor is the same as that of the doped first semiconductor layer 32 and the light emitting semiconductor layer. 40 and the second semiconductor layer 50 can be reflected, so that the internal quantum efficiency IQE is improved.

  Furthermore, the undoped first semiconductor layer 31 preferably includes a buffer layer 33 as shown in FIG. In the semiconductor light emitting device 100, the buffer layer 33 is provided on the fine concavo-convex structure 20, and then the undoped first semiconductor layer 31 and the doped first semiconductor layer 32 are sequentially stacked, thereby crystallizing the first semiconductor layer 30. Nucleation and nucleus growth, which are initial conditions for growth, are improved, and the performance of the first semiconductor layer 30 as a semiconductor is improved, so that the degree of improvement in internal quantum efficiency IQE is improved. Here, the buffer layer 33 may be arranged so as to planarize the fine concavo-convex structure 20, but since the growth rate of the buffer layer 33 is slow, from the viewpoint of shortening the manufacturing time of the semiconductor light emitting device 100, the buffer layer 33. It is preferable to planarize the fine concavo-convex structure 20 with the undoped first semiconductor layer 31 provided thereon. By providing the undoped first semiconductor layer 31 so as to planarize the fine concavo-convex structure 20, the performance of the undoped first semiconductor layer 31 as a semiconductor is improved by the doped first semiconductor layer 32, the light emitting semiconductor layer 40, and the first semiconductor layer 31. 2 can be reflected in the semiconductor layer 50, so that the internal quantum efficiency IQE is improved. In FIG. 1, the buffer layer 33 is disposed so as to cover the surface of the fine concavo-convex structure 20, but may be partially provided on the surface of the fine concavo-convex structure 20. In particular, the buffer layer 33 can be preferentially provided at the bottom of the concave portion of the fine concavo-convex structure 20.

  Furthermore, the transparent conductive film 60 can be provided on the second semiconductor layer 50, the anode electrode 70 can be provided on the surface of the transparent conductive film 60, and the cathode electrode 80 can be provided on the surface of the first semiconductor layer 30. The arrangement of the transparent conductive film 60, the anode electrode 70, and the cathode electrode 80 is not limited because it can be appropriately optimized by the semiconductor light emitting device, but is generally provided as illustrated in FIG. 1.

  The semiconductor light emitting device 100 shown in FIG. 1 is an example applied to a semiconductor light emitting device having a double hetero structure, but the stacked structure of the first semiconductor layer 30, the light emitting semiconductor layer 40, and the second semiconductor layer 50 is limited to this. It is not something.

  In the semiconductor light emitting device 100 according to the present embodiment, the fine concavo-convex structure 20 of the optical substrate 10 satisfies the following conditions (1) to (3).

  (1) The average pitch Pav of the fine concavo-convex structure 20 and the first semiconductor layer 30 provided on the surface of the optical substrate 10 on which the fine concavo-convex structure 20 is formed (hereinafter also referred to as the fine concavo-convex structure surface). The ratio (Pav / λ) to the wavelength λ of the emitted light is 2.5 or more and 8 or less.

  (2) The ratio between the in-plane average height Hav of the fine concavo-convex structure 20 and the wavelength λ of the emitted light is 0.1 or more and 1 or less.

  (3) The aspect ratio of the fine relief structure 20 is 0.2 or more and 1 or less.

  The semiconductor light emitting device 100 according to the present embodiment has the following effects when the fine concavo-convex structure 20 satisfies the above conditions (1) to (3).

  First, the internal quantum efficiency IQE can be improved. Specifically, when the average pitch Pav of the fine concavo-convex structure 20 is sufficiently small, it becomes possible to disturb the growth mode at the time of film formation of the first semiconductor layer 30 on the optical substrate 10, and the first semiconductor Dislocations generated in the layer 30 can be reduced.

  On the other hand, the in-plane average height Hav of the fine concavo-convex structure 20 is not less than 0.1 times and not more than 1 time the wavelength λ of the emitted light in the first semiconductor layer 30, and the aspect ratio of the fine concavo-convex structure 20 is By being 0.2 or more and 1 or less, the volume change of the fine concavo-convex structure 20 viewed from light is sufficiently large, and the diffraction angle can advantageously break the waveguide mode, and the light extraction efficiency LEE is improved.

  Further, when the ratio (Pav / λ) between the average pitch Pav and the wavelength λ of the emitted light is 2.5 or more and 8 or less, the wave nature of the light is emphasized, and the emitted light has directivity. This increases the degree of freedom in designing the electrode area and material. For example, in the case where the diffracted light is emitted six times symmetrically, a portion that is a shadow of the electrode portion is reduced by appropriately selecting the rotation angle formed by the semiconductor light emitting element 100 and the substrate.

  FIG. 2 is a schematic diagram for explaining the diffraction azimuth angle of the diffracted light in the semiconductor light emitting device 100 according to the present embodiment. 2A shows the z-axis when the direction in which the first semiconductor layer 30, the light-emitting semiconductor layer 40, and the second semiconductor layer 50 are sequentially stacked (hereinafter also referred to as a stacking direction) is defined as the z-axis in the semiconductor light emitting device 100. The cross section along is shown. In FIG. 2A, reference numeral 201 indicates diffracted light emitted from the fine concavo-convex structure 20. 2A shows the x-axis and the y-axis included in a plane that is orthogonal to the z-axis and is parallel to the surface of the fine concavo-convex structure 20 on which the fine concavo-convex structure 20 is formed.

  FIG. 2B shows the diffracted light 201 viewed from the upper side in the z-axis direction in FIG. 2A. In this case, the diffracted light 201 is emitted six times symmetrically. One radiating portion 202 of the diffracted light 201 forms a rotation angle θ with respect to the x axis. This rotation angle θ is called a “diffraction azimuth angle”.

In the semiconductor light emitting device 100 according to the present embodiment, the wavelength λ of the emitted light of the first semiconductor layer 30 is obtained by dividing the vacuum wavelength λ ₀ of the emitted light by the refractive index n of the first semiconductor layer 30.

  Here, the emitted light refers to light emitted from the light emitting semiconductor layer 40 of the semiconductor light emitting device 100.

As described above, when the first semiconductor layer 30 has the buffer layer 33 and is formed by sequentially laminating the undoped first semiconductor layer 31 and the doped first semiconductor layer 32, the wavelength λ of the emitted light is The wavelength in the undoped first semiconductor layer 31 provided so as to flatten the fine concavo-convex structure 20. Thus, the wavelength λ of the emitted light is determined depending on the layer serving as the medium of the emitted light provided on the fine concavo-convex structure 20. This layer is called an “optical medium layer”. Therefore, the wavelength λ of the emitted light can be calculated from the vacuum wavelength λ ₀ and the refractive index n of the optical medium layer.

Note that the vacuum wavelength λ ₀ of the emitted light is the vacuum wavelength of the light emitted by the light emitting semiconductor layer according to the band gap, and can be obtained by, for example, a wavelength meter or a spectral radiance meter.

  In the semiconductor light emitting device 100 according to the present embodiment, it is preferable that the bottom of the concave portion constituting the fine concavo-convex structure 20 is a flat surface. The term “flat” as used herein means that when the optical substrate 10 is observed with a scanning electron microscope (SEM) at a relatively low magnification, for example, when observed at 20000 to 100000 times, the bottom of the recess in the cross-sectional image is linear. Or when the optical substrate 10 is tilted (tilted), for example, when the optical substrate 10 is tilted by 10 degrees, the bottom of the recess is observed flat in the cross-sectional image. That is, in these cases, a terrace and a step determined by an offset angle, which will be described later, are not observed, and are flat. Even when the terrace and the step are included at the bottom of the concave portion, the flat surface is obtained.

  According to this configuration, the internal quantum efficiency IQE can be improved. In order to improve the internal quantum efficiency IQE in the semiconductor light emitting device 100, the chemical vapor deposition (CVD) growth mode of the semiconductor crystal layer is disturbed, the dislocations inside the semiconductor crystal layer are dispersed, and the local and macroscopic dislocation density. Need to be reduced. Here, initial conditions of these physical phenomena are nucleation and growth when the semiconductor crystal layer is formed by chemical vapor deposition (CVD). Since a flat surface is provided at the bottom of the concave portion of the fine concavo-convex structure 20, nucleation can be suitably generated, and therefore, the effect of reducing dislocations in the semiconductor crystal layer due to the density of the fine concavo-convex structure 20 can be expressed more. Is possible. As a result, the internal quantum efficiency IQE can be further increased.

  Further, in the semiconductor light emitting device 100 according to the present embodiment, the convex portion constituting the fine concavo-convex structure 20 has a smaller diameter at the top than at the bottom, that is, the cross section of the convex is inclined from the top to the bottom. It is preferable to have a side part.

  According to this configuration, it is possible to reduce dislocations generated in the first semiconductor layer 30 and to disperse dislocations according to the fine concavo-convex structure 20. That is, it is possible to produce the first semiconductor layer 30 having a low dislocation density both macroscopically and microscopically, and the internal quantum efficiency IQE is improved.

  Moreover, the optical substrate 10 according to the present embodiment is beneficial for manufacturing the semiconductor light emitting device 100 according to the present embodiment by satisfying the conditions (1) to (3).

  The fine concavo-convex structure 20 should just be provided in at least one part on the surface of the optical base material 10. FIG. The area of the preferred region is a region corresponding to one of the semiconductor light emitting elements 100, most preferably the entire surface of the optical substrate 10. In addition, the optical base material 10 refers to the board | substrate used in order to laminate | stack a semiconductor layer on the surface and to form a semiconductor light-emitting device.

  Hereinafter, the principle that the semiconductor light emitting device 100 according to the present embodiment exerts an effect will be described.

  In general, the light emission efficiency of a semiconductor light emitting device is determined by the internal quantum efficiency IQE and the light extraction efficiency LEE. In order to improve both, it is necessary to achieve both reduction of crystal defects by the nanoscale uneven structure, collapse of the waveguide mode, and increase of the effective light extraction area. In particular, in order to break down the waveguide mode, the light extraction efficiency has been improved by scattering with a micro-order uneven structure, but we are now considering the light diffraction that occurs in the nano order, especially the number of diffraction modes and the diffraction angle. noticed. This makes it possible to improve both the light extraction efficiency due to the reduction of crystal defects, the action of effectively breaking the waveguide mode due to diffraction, and the increase of the effective light extraction area.

  Further, the interaction between the emitted light generated in the light emitting layer of the semiconductor light emitting element and the uneven structure is determined by the relationship between the wavelength λ of the emitted light in the semiconductor light emitting element and the size of the uneven structure.

  The micro-order concavo-convex structure conventionally used is about 10 times as large as the wavelength λ of the emitted light in the semiconductor light emitting device, and the interaction between the light and the concavo-convex structure is scattering. However, when the size of the concavo-convex structure becomes nanoscale in order to improve the internal quantum efficiency IQE, it becomes the same size as the wavelength λ of the emitted light in the semiconductor light emitting device to several times, and the light wave nature is emphasized. As an interaction, light diffraction occurs.

  The quantities that characterize light diffraction are the mode, which is the number of outgoing light rays after diffraction, the outgoing angle of each mode, and the intensity of each mode. In light diffraction, due to its wave nature, the interaction between the interface having the refractive index difference and the incident light splits the reflected light and transmitted light when viewed as a light beam, generating a mode.

  The number of modes of light diffraction is small when the average pitch is small, and increases as the average pitch is large. That is, the number of modes is limited under conditions where the average pitch is small. This is observed when the diffracted light has directivity.

  On the other hand, the number of modes of light diffraction increases as the average pitch increases. At this time, the strength of each mode decreases as the number of modes increases. In addition, the distribution of the light emission angle gradually becomes wider. Its behavior is observed as scattered.

  Therefore, in order to maximize the luminous efficiency of the semiconductor light emitting device, a nanoscale fine concavo-convex structure is provided to disturb the crystal growth mode and improve the internal quantum efficiency, and the optical diffraction generated by the nanoscale concavo-convex structure. It is necessary to provide the diffraction angle and diffraction intensity so as to effectively destroy the waveguide mode.

  Next, the reason why the above conditions (1) to (3) are preferable in breaking the waveguide mode in the nanoscale fine uneven structure will be described.

  The number of modes of light diffraction is small when the average pitch Pav of the fine concavo-convex structure 20 is small, and increases as the average pitch Pav is large. If the pitch of the fine concavo-convex structure is too large, light behaves in a scattering manner. At this time, the following loss is generated.

  First, the emitted light is newly introduced by acting in a scattering manner on components that are not originally desired to be scattered, for example, light having an incident angle equal to or less than the critical angle of the soot that can be extracted without the fine relief structure 20. Wave mode (light with an incident angle larger than the critical angle).

  In addition, it acts on the incident light as a waveguide mode in a scattering manner. That is, since the light is emitted at an angle that does not contribute to light extraction, the waveguide mode cannot be effectively broken.

  This new waveguide mode reaches the light emitting semiconductor layer 40 and the second semiconductor layer 50, and then is connected to the interface between the second semiconductor layer 50 and the transparent conductive film 60, or to the outside such as the transparent conductive film 60 and the sealing material. Due to total reflection at the interface, the light again passes through the second semiconductor layer 50 and the light emitting semiconductor layer 40, and then enters the fine concavo-convex structure 20 again and is scattered again, resulting in a component that contributes to light extraction. However, it is known that the light emitting semiconductor layer 40 has a very large absorption with respect to the wavelength λ of the emitted light because it has a band gap corresponding to the emitted light. That is, it is absorbed when passing through the light emitting semiconductor layer 40 in a new waveguide mode, resulting in a loss.

  Due to the directivity of light diffraction, generation of a new guided mode can be avoided by using a pattern having a diffraction angle that contributes to extraction with respect to the incident angle. Thereby, the light extraction efficiency LEE can be improved.

  As described above, it is difficult to improve the internal quantum efficiency with a micro-order uneven structure. This is because the density of the uneven structure is overwhelmingly insufficient to disturb the mode of crystal growth. In order to reduce crystal defects, a nanoscale fine uneven structure is preferable.

  In addition, in the conventional micro-order concavo-convex structure, since the interaction with light is scattering, it becomes a shadow regardless of the arrangement of the electrodes provided on the light extraction surface, and the effective light extraction surface The area of will decrease. That is, the electrode area cannot be increased in order to improve the current injection efficiency, and the shape and size of the electrode portion are significantly limited. However, the nano-order fine concavo-convex structure enhances the wave nature of light, and the emitted light has directivity. Thus, by appropriately selecting the diffraction azimuth angle θ of the diffracted light derived from the arrangement of the fine concavo-convex structure, the degree of freedom in designing the area and material of the electrode portion is increased. For example, in the case where the diffracted light is emitted six times symmetrically, a region where the electrode portion is shaded is reduced by appropriately selecting the rotation angle between the semiconductor light emitting element and the substrate.

  FIG. 3 is a schematic diagram for explaining the relationship between the diffraction azimuth angle and the electrode portion in the semiconductor light emitting device according to the present embodiment.

  The diffraction azimuth angle is a plane (x-axis in FIG. 2) in which the diffracted light emitted upward from the fine concavo-convex structure 20 is orthogonal to the stacking direction of the semiconductor layers (z-axis in FIG. 2) in the first semiconductor layer 30. And the rotation angle of the radiation part of the diffracted light within the plane including the y-axis). Here, the diffracted light is incident on the fine concavo-convex structure 20 from the first semiconductor layer 30 side, reflected and diffracted light, and reflected from the back surface of the optical substrate 10, and then the fine concavo-convex structure from the optical substrate 10 side. Light that re-enters 20 and passes through. This angle varies depending on the arrangement of the fine concavo-convex structure 20, the average pitch, or the height.

  FIGS. 3A to 3C illustrate a diffraction azimuth angle of diffracted light that interacts with a part of the optical base material 10 by paying attention to the part. This means that the light emission pattern is observed when the semiconductor light emitting device 100 is viewed from the electrode side. FIG. 3A shows a case where the diffracted light 201 is six-fold symmetric. FIG. 3B shows a case where the diffracted light is 6-fold symmetric as in FIG. 3A and the radiating portion 202 of the diffracted light 201 is rotated 90 degrees by rotating the optical substrate 10 by 90 degrees. FIG. 3C shows a case where the fine relief structure 20 is microscale. In this case, the interaction between the fine concavo-convex structure 20 and the incident light becomes scattering, and the diffracted light 201 is observed so as not to be anisotropic because it emits light in all directions.

  Here, it is considered that the diffracted light diffracted by the fine concavo-convex structure 20 is extracted from the electrode side. The diffracted light 201 has a certain diffraction azimuth because of its wave nature. On the other hand, there is a distribution with respect to the polar angle, that is, the angle formed by the z-axis, which is the stacking direction of the first semiconductor layer 30, the light emitting semiconductor layer 40, and the second semiconductor layer 50, and the emission direction. That is, the diffracted light 201 interacting with a part of the optical substrate 10 expects a region (hereinafter referred to as a prospective angle region) whose light output direction is surrounded by a diffraction azimuth angle and a polar angle. When the anode electrode 70 or the cathode electrode 80 is in this expected angle region, the light extraction efficiency is not effectively improved. This is because these electrode portions are made of metal and do not transmit light, so that they are substantially shaded.

  3D to FIG. 3F illustrate portions in which the electrode portion enters the expected angle region in each region of the fine concavo-convex structure 20 as described above.

  FIG. 3D shows a case where the diffracted light 201 is symmetric six times as shown in FIG. 3A. When the electrode part 203 comes on an extension line 204 (indicated by a one-dot broken line in the figure) 204 of the radiating portion 202 of the diffracted light 201 at a certain point on the optical base material 10, it becomes a shadow. Is as shown in FIG. 3D. It can also be seen that a region that can be shaded in the longitudinal direction of the semiconductor light emitting device 100 is formed.

  FIG. 3E shows a case where the electrode part 203 is provided by rotating the optical substrate 10 of FIG. 3D by 90 degrees. It can be seen that the expected angle region 205 of the electrode part 203 is smaller than that in FIG. 3D. That is, as the arrangement of the optical base material 10, it is preferable that the expected angle region 205 where the electrode portion 203 comes on the extended line 204 of the radiation portion 202 of the diffracted light 201 is smaller.

  FIG. 3F shows a case where the conventional uneven structure is a microscale. In this case, since the behavior is scattering, there are many areas that can be shaded. For this reason, in the conventional semiconductor light emitting device 200, the area of the electrode portion 203 is reduced, a transparent conductive film is provided to increase the area where current can be injected, and the area contributing to light extraction is increased.

  However, as shown in FIG. 3E, if the expected angle region 205 is a shape that avoids the diffraction azimuth angle, the electrode area can be increased, or a non-transparent electrode material can be used. This increases the degree of freedom in designing the position, material, or shape of the electrode portion 203. In addition, light can be effectively extracted from the semiconductor light emitting device 100. In addition, the current dispersibility is improved by the efficient arrangement of the electrode portion 203, and an element having a larger area can be obtained. This effect is remarkable when the cross section of the semiconductor light emitting device 100 is rectangular, as shown in the comparison between FIG. 3D and FIG. 3E, and the aspect ratio of the cross sectional shape (length in the longitudinal direction / length in the short direction). Greater than 1, the more prominent.

  As described above, in the semiconductor light emitting device 100 according to the present embodiment, the nanoscale fine concavo-convex structure 20 improves both the internal quantum efficiency IQE that determines the light emission efficiency of the semiconductor light emitting device 100 and the light extraction efficiency LEE. Can be maximized.

  When Pav / λ is 2.5 or more and 8 or less, the internal quantum efficiency IQE can be improved. Specifically, by making the pitch of the fine concavo-convex structure 20 sufficiently small, it becomes possible to disturb the growth mode during the film formation of the first semiconductor layer 30 on the optical substrate 10, and the first semiconductor layer 30. Can be reduced. This makes it difficult for cracks to enter the semiconductor layer. Further, the wave nature of light is emphasized, and the emitted light has directivity. This increases the degree of freedom in designing the electrode area and material.

  From the above viewpoint, the ratio Pav / λ between the average pitch Pav and the wavelength λ in the first semiconductor layer 30 is preferably 2.7 or more, and more preferably 3 or more. Moreover, in order to suppress scattering, the ratio is more preferably 7 or less, and further preferably 6.5 or less.

The ratio (Pav / λ) between the average pitch Pav and the wavelength λ of the emitted light in the first semiconductor layer 30 is determined using the vacuum wavelength λ ₀ of the emitted light and the refractive index n of the first semiconductor layer 30. It can also be expressed as Pav * n / λ ₀ .

  On the other hand, the in-plane average height (Hav) of the fine uneven structure 20 is not less than 0.1 times and not more than 1 time the wavelength λ of the emitted light in the first semiconductor layer 30, and the aspect ratio of the fine uneven structure 20 is By being 0.2 or more and 1 or less, the volume change of the fine concavo-convex structure 20 viewed from light is sufficiently large, and the diffraction angle can advantageously break the waveguide mode, and the light extraction efficiency LEE is improved.

The in-plane average height Hav is the convex volume per unit area when the fine concavo-convex structure is a dot structure. In particular, in the case of a regular arrangement, it is represented by the ratio of the volume of convex portions contained in the unit cell. For example, when the convex portions of the convex volume V ₀ are arranged in a hexagonal manner with a pitch P, the in-plane average height Hav is expressed by the following formula (1). In addition, when the convex portions of the convex volume V ₀ are arranged in a regular tetragon with a pitch P, the in-plane average height Hav is expressed by the following equation (2). On the other hand, when the fine concavo-convex structure is a hole structure, the in-plane average height Hav is a recess volume per unit area, that is, an in-plane average depth.

  A large in-plane average height Hav means that the convex part volume is large. In the case of a dot structure, the aspect ratio is a ratio of the convex height H to the convex bottom width, and a convex height H / convex bottom width ratio. On the other hand, in the case of the hole structure, the ratio between the recess depth and the recess bottom width, the recess depth / the recess bottom width. An aspect ratio of 1 or less means that the height of the convex portion is not larger than the width of the bottom portion of the convex portion.

  The reason why it is preferable that the ratio of these parameters, the in-plane average height Hav and the wavelength λ of the emitted light, and the aspect ratio are in the above ranges will be described. Therefore, the effect when the height H constituting these parameters and the duty break the waveguide mode will be described. In addition, although the case of a dot structure is demonstrated below, it can be read by the above definition also in a hole structure.

  The larger the height H, the more modes interact with the incident light from the group of diffraction modes determined by the average pitch Pav. That is, when the height H is low, incident light acts through only the diffracted light of a limited mode from the group of diffraction modes determined by the average pitch Pav through the fine concavo-convex structure 20, but the height H is If it is high, more diffraction modes in the group of diffraction modes determined by the average pitch Pav interact with the incident light. That is, as the height H is higher, the interaction between the incident light and the fine concavo-convex structure 20 becomes more scattering in behavior. The higher the height H, the more modes interact, but some of these modes do not contribute to extraction. Therefore, if it is scattering, a loss occurs. On the other hand, when the height H is low, the volumetric change of the fine concavo-convex structure 20 is small, so that the diffraction efficiency is lowered. Since incident light acts through only the diffracted light of a limited mode from the group of diffraction modes determined by the average pitch Pav through the fine concavo-convex structure 20, it is difficult to break the waveguide mode.

  Next, the reason why the duty contributes to the improvement of the internal quantum efficiency IQE and the light extraction efficiency LEE will be described.

  When the duty is large, the volume of the convex portion becomes large. For this reason, the intensity of the 0th-order light that is reflected or transmitted without being diffracted with respect to the light incident on the fine concavo-convex structure 20 is lowered, and the intensity other than the 0th-order light and the intensity around the diffracted light is increased. Thus, after setting the height H so that the mode contributing to light extraction from the diffraction mode group determined by the average pitch Pav has an intensity, the intensity can be adjusted more effectively by adjusting the duty. It becomes possible to break the waveguide mode by diffraction.

  It can be seen that there are preferable ranges for the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio of the convex portion, depending on the above two parameters, the height of the convex portion and the duty of the convex portion. What arranged this becomes the range of said (2) and (3).

  The in-plane average height Hav of the fine concavo-convex structure 20 is more preferably 0.12 times or more, and further preferably 0.15 times or more the wavelength λ of the emitted light in the first semiconductor layer 30. Moreover, 0.9 times or less are more preferable and 0.8 times or less are still more preferable.

The ratio (Hav / λ) between the in-plane average height Hav and the wavelength λ of the emitted light in the first semiconductor layer 30 is determined by the vacuum wavelength λ ₀ of the emitted light and the refractive index n of the first semiconductor layer 30. It can also be expressed as Hav * n / λ ₀ .

  In addition, when the aspect ratio is large, the height H is large, that is, many diffraction modes are established, and as a result, a mode contributing to light extraction is likely to be established. For this reason, the aspect ratio is more preferably 0.25 or more, and further preferably 0.3 or more. However, if the aspect ratio is too large, that is, the convex portion is too high, it is not preferable because it is excessively scattered. In this respect, the aspect ratio is more preferably 0.9 or less, and further preferably 0.85 or less.

  The fine concavo-convex structure 20 provided on the optical substrate 10 is composed of a plurality of convex portions or concave portions. If the fine concavo-convex structure 20 has a convex part or a concave part, the shape and arrangement thereof are not limited, and the light extraction efficiency LEE can be increased while maintaining the improvement of the internal quantum efficiency IQE. For this reason, for example, a line-and-space structure in which a plurality of fence-like bodies are arranged, a dot structure in which a plurality of dot (convex portions, protrusions) -like structures are arranged, a hole structure in which a plurality of hole (concave) -like structures are arranged, etc. Can be adopted. Examples of the dot structure and the hole structure include a cone, a cylinder, a quadrangular pyramid, a quadrangular prism, a hexagonal pyramid, a hexagonal pyramid, a polygonal pyramid, a double ring shape, and a multiple ring shape. These shapes include a shape in which the outer diameter of the bottom surface is distorted and a shape in which the side surface is curved. The dot structure is a structure in which a plurality of convex portions are arranged independently of each other. That is, each convex part is separated by a continuous concave part. In addition, each convex part may be smoothly connected by the continuous recessed part. On the other hand, the hole structure is a structure in which a plurality of recesses are arranged independently of each other. That is, each recessed part is separated by the continuous convex part. In addition, each recessed part may be smoothly connected by the continuous convex part. The hole structure is more preferable when the base material is peeled to form a thin film semiconductor light emitting device because the semiconductor layer having a high refractive index becomes convex and the light emission efficiency of the thin film semiconductor light emitting device is improved.

  As the arrangement of the fine concavo-convex structure 20, a regular hexagonal arrangement, a regular tetragonal arrangement, a quasi-hexagonal arrangement, a quasi-tetragonal arrangement, or the like can be adopted. Furthermore, a combination of these sequences can also be employed. For example, an array in which a regular hexagonal array and a regular tetragonal array are alternately arranged, an array that gradually changes from a regular hexagonal array to a regular tetragonal array, and gradually returns from the regular tetragonal array to the regular hexagonal array, and the like. From the viewpoint of improving the internal quantum efficiency IQE, it is preferable to employ an array including a regular hexagonal array or a quasi-hexagonal array. Here, the quasi-hexagonal array is defined as an array in which the distance (pitch) between adjacent convex portions of the regular hexagonal array has a deviation of ± 25% or less. This deviation amount may be a uniform deviation amount in the array, or a distribution may be provided for the deviation amount.

  In addition, a line-and-space structure or a line-and-space arrangement of continuous protrusions in a one-dimensional or two-dimensional array is more efficient because the volume of the protrusions is large in addition to the arrangement. Therefore, it is possible to generate diffracted light that disrupts the waveguide mode.

  FIG. 4A to FIG. 4G and FIG. 5 are schematic views showing an example of an array of continuous convex portions in the semiconductor light emitting device according to the present embodiment. The solid line in the figure indicates the top of the convex portion. FIG. 4A shows a line and space structure. FIG. 4B shows a tatami arrangement. FIG. 4C shows a structure in which the tops of continuous convex portions are orthogonal. FIG. 4D shows a structure in which continuous convex portions form a parallelogram. FIG. 4E shows a structure in which continuous convex portions constitute a triangle. FIG. 4F shows a case where the line and space structure is deviated from parallel. FIG. 4G shows a structure in which consecutive convex portions having different lengths are orthogonal to each other. FIG. 5 shows a case where continuous convex portions constitute a kagome lattice.

<Shape of fine uneven structure>
The shape of the fine relief structure 20 is not particularly limited. In case of a single convex side slope angle, for example, in the case of a dot structure, a conical shape, and in order to increase the volume of the convex part, the convex side slope angle changes in multiple steps from the convex top to the convex bottom. It is preferable. For example, when the inflection point where the convex side surface bulges up draws one curve, the inclination angle becomes two. By having such a multi-step inclination angle, it becomes easy to control the angle and intensity by the light diffraction property of the fine concavo-convex structure 20, and the light extraction efficiency LEE can be improved. Further, the inclination angle of the side surface of the convex portion can be selected from the crystal plane appearing on the side surface of the convex portion depending on the material of the growth base and the semiconductor crystal layer.

  Next, terms used to describe the fine concavo-convex structure 20 are defined. Further, in the following description, the case where the fine concavo-convex structure 20 is a convex portion will be described. However, in the description of the case of a concave portion, the concave portion shape is an inversion of the convex portion shape. The following terms can be used in the same way.

<Pitch P and average pitch Pav>
FIG. 6 is a top view of the first semiconductor layer of the semiconductor light emitting device according to the present embodiment as viewed from the surface side where the fine concavo-convex structure is formed. When the fine concavo-convex structure 20 is a dot structure in which a plurality of convex portions 20a are arranged in the concave portions 20b, the center of a certain convex portion A1 and the convex portions B1-1 to B1- adjacent to the convex portion A1. the distance _{P A1B1-1} ~ distance _{P A1B1-6} between the center of 6, is defined as a pitch P. However, as shown in FIG. 6, when the pitch P differs depending on the adjacent convex portions, the average pitch Pav is determined according to the following procedure. (1) A plurality of arbitrary convex portions A1, A2,... AN are selected. (2) The pitches P _{AMBM-1 to} P _AMBM-k between the convex portions AM and the convex portions (BM-1 to BM-k) adjacent to the convex portions AM (1 ≦ M ≦ N) are measured. (3) For the convex portions A1 to AN, the pitch P is measured as in (2). (4) An arithmetic average value of the pitches P _{A1B1-1 to} P _ANBN-k is defined as an average pitch Pav. However, N is 5 or more and 10 or less, and k is 4 or more and 6 or less. In the case of the hole structure, the average pitch Pav can be defined by replacing the convex portion 20a described in the dot structure with a concave opening.

As shown in FIG. 7, when the fine concavo-convex structure 20 is a line and space structure, the center line of a certain convex line A1, and the centers of the convex line B1-1 and the convex line B1-2 adjacent to the convex line A1. the arithmetic mean of the shortest distance _{P A1B1-1} and the shortest distance _{P A1B1-2} between lines is defined as the pitch P. However, as shown in FIG. 7, when the pitch P differs depending on the selected convex line, the average pitch Pav is determined according to the following procedure. (1) An arbitrary plurality of convex lines A1, A2,... AN are selected. (2) The pitches P _AMBM-1 and P _AMBM-2 between the convex line AM and the convex line (BM-1, BM-2) adjacent to the convex line AM (1 ≦ M ≦ N) are measured. (3) For the convex lines A1 to AN, the pitch P is measured as in (2). (4) An arithmetic average value of the pitches P _{A1B1-1-1 to} P _{ANBN -2} is defined as an average pitch Pav. However, N is 5 or more and 10 or less.

<Convex top width lcvt, concave opening width lcct, convex bottom width lcvb, concave bottom width lccb>
FIG. 8 shows a top view when the fine uneven structure 20 has a dot structure. A line segment indicated by a broken line in FIG. 8 is a distance between the center of a certain convex portion 20a and the center of the convex portion 20a closest to the convex portion 20a, and means the pitch P described above. FIGS. 9A and 9B show schematic cross-sectional views of the fine concavo-convex structure 20 at the line segment position corresponding to the pitch P shown in FIG.

  As shown in FIG. 9A, the convex portion top width lcvt is defined as the width of the top surface of the convex portion 20a, and the concave portion opening width lcct is a difference value (P−lcvt) between the pitch P and the convex portion top width lcvt. Defined.

  As shown in FIG. 9B, the convex bottom width lcvb is defined as the width of the convex portion 20a, and the concave bottom width lccb is defined as a difference value (P-lcvb) between the pitch P and the convex bottom width lcvb. Is done.

  FIG. 10 shows a top view when the fine uneven structure 20 has a hole structure. A line segment indicated by a broken line in FIG. 10 is a distance between the center of a certain recess 20b and the center of the recess 20b closest to the recess 20b, and means the pitch P described above. 11A and 11B show schematic cross-sectional views of the fine concavo-convex structure 20 at the position of the line segment corresponding to the pitch P shown in FIG.

  As shown in FIG. 11A, the recess opening width lcct is defined as the opening width of the recess 20b, and the protrusion top width lcvt is defined as a difference value (P-lcct) between the pitch P and the recess opening width lcct.

  As shown in FIG. 11B, the convex bottom width lcvb is defined as the width of the convex portion 20a bottom, and the concave bottom width lccb is defined as a difference value (P-lcvb) between the pitch P and the convex bottom width lcvb. Is done.

<Duty>
In the case of the dot structure, the duty is defined by a ratio (lcvb / P) between the convex bottom width lcvb and the pitch P. On the other hand, in the case of a hole structure, it is defined by the ratio between the bottom width of the recess and the pitch P (lccb / P).

  A small duty indicates that there are many flat portions of the concavo-convex structure, that is, flat surfaces at the bottom of the recesses. This is advantageous in terms of crystal growth because nucleation is likely to occur, but from the light extraction surface, it is preferable that diffraction does not occur and there are only a few flat portions that are totally reflected. In this respect, the duty is preferably 0.2 or more, more preferably 0.3 or more, and most preferably 0.5 or more.

  On the other hand, the upper limit of the duty is determined from the crystal growth of the semiconductor layer. If the area of the flat portion is too small, nucleation during crystal growth is hindered, resulting in a decrease in crystal quality. In this respect, the duty is preferably 0.9 or less, more preferably 0.83 or less, and most preferably 0.73 or less.

  The duty also affects the in-plane average height Hav that characterizes diffraction. In the dot structure or hole structure, which is an array of independent protrusions or recesses, the duty works as a square with respect to the volume of the protrusions, so the duty is 0.3 for the array of dot-shaped protrusions. The above is more preferable, and 0.5 or more is most preferable. The upper limit is preferably 0.9 or less, and more preferably 0.8 or less.

  Furthermore, in a fine concavo-convex structure having continuous convex portions such as a line and space structure, the volume of the convex portions is linear with respect to the duty. Therefore, the duty is more preferably 0.3 or more, and most preferably 0.5 or more. The upper limit is preferably 0.8 or less, and more preferably 0.6 or less.

<Height H>
The height H of the fine concavo-convex structure is defined as the shortest distance between the average position of the concave bottom of the concavo-convex structure and the average position of the convex vertices of the concavo-convex structure. The number of sample points for calculating the average position is preferably 10 or more.

  Since the upper limit of the preferred aspect ratio of the height H of the convex portion 20a of the fine concavo-convex structure 20 is 1, the height H of the convex portion 20a is equal to or less than the average pitch Pav. When the average pitch is Pav or less, the refractive index distribution of the fine concavo-convex structure 20 is appropriate in view of the emitted light, so that the light extraction efficiency LEE can be further improved. Further, the thickness of the semiconductor layer required for flattening the convex portion 20a can be reduced. Therefore, warpage of the semiconductor light emitting device 100 can be suppressed. From this viewpoint, the height H of the fine concavo-convex structure 20 is preferably 1 time or less, more preferably 0.7 times or less of the average pitch Pav.

<Convex side inclination angle Θ>
The inclination angle Θ of the side surface of the convex portion is determined by the shape parameter of the fine concavo-convex structure 20 described above. In particular, it is preferable that the inclination angle changes in multiple steps from the top of the convex portion toward the bottom of the convex portion. For example, when the inflection point where the convex side surface bulges up draws one curve, the inclination angle becomes two. By having such a multi-step inclination angle, it becomes easy to control the angle and intensity by the light diffraction property of the fine concavo-convex structure 20, and the light extraction efficiency LEE can be improved. Further, the inclination angle of the side surface of the convex portion can be selected from the crystal plane appearing on the side surface of the convex portion depending on the material of the growth base and the semiconductor crystal layer.

(Arithmetic mean)
When N measured values of the distribution of a certain element (variable) are x1, x2,..., Xn, the arithmetic mean value is defined by the following equation (3).

  The number N of sample points when calculating the arithmetic mean is defined as 10 or more. The number of sample points when calculating the standard deviation is the same as the number N of sample points when calculating the arithmetic mean.

  Here, the local range used for observation is defined as a range of about 5 to 50 times the average pitch of the fine relief structure 20. For example, if the average pitch is 700 nm, the observation is performed within the observation range of 3500 nm to 35000 nm. Therefore, for example, a field image of 7500 nm is captured, and an arithmetic average is obtained using the captured image.

The shape of the fine concavo-convex structure 20 provided on the optical substrate 10 will be described.
The ratio (lcvt / lcct) between the convex top width lcvt and the concave opening width lcct of the fine concavo-convex structure 20 is preferably as small as possible, and is most preferably substantially zero. Here, “lcvt / lcct = 0” means that lcvt = 0 nm. However, for example, even when lcvt is measured by a scanning electron microscope, 0 nm cannot be measured accurately. Therefore, lcvt here includes all cases where the resolution is less than the measurement resolution. When the ratio (lcvt / lcct) is 3 or less, the internal quantum efficiency IQE can be effectively improved. This is because the dislocation generated from the top of the convex portion is suppressed, the dispersibility of the dislocation is improved, and the microscopic and macroscopic dislocation density is decreased. Furthermore, when (lcvt / lcct) is 1 or less, the light extraction efficiency LEE can be improved. This is because the refractive index distribution of the fine concavo-convex structure 20 formed by the optical base material 10 and the semiconductor crystal layer is appropriate in view of the emitted light. From the viewpoint of greatly improving both the internal quantum efficiency IQE and the light extraction efficiency LEE described above, (lcvt / lcct) is preferably 0.4 or less, more preferably 0.2 or less, and further preferably 0.15 or less.

  Further, it is preferable that the bottom of the concave portion 20b of the fine concavo-convex structure 20 has a flat surface because the internal quantum efficiency IQE can be improved and the difference between the semiconductor crystal film forming apparatuses can be reduced. In order to improve the internal quantum efficiency IQE in the LED element, it is necessary to disperse dislocations inside the semiconductor crystal layer and reduce the local and macroscopic dislocation density. Here, initial conditions of these physical phenomena are nucleation and growth when the semiconductor crystal layer is formed by chemical vapor deposition (CVD). Since a flat surface is provided at the bottom of the concave portion 20b of the fine concavo-convex structure 20, nucleation can be suitably generated, so that the dislocation reduction effect in the semiconductor crystal layer due to the density of the fine concavo-convex structure 20 is more manifested. It becomes possible to make it. As a result, the internal quantum efficiency IQE can be further increased. In addition, since dislocations are reduced, cracks are less likely to occur in the semiconductor crystal.

  From the above viewpoint, it is preferable that the ratio (lcvb / lccb) between the convex bottom width lcvb and the concave bottom width lccb of the fine concavo-convex structure 20 is 5 or less. In particular, (lcvb / lccb) is more preferably 2 or less, and most preferably 1 or less, from the viewpoint of further promoting the growth of the semiconductor crystal layer using the bottom of the recess of the fine relief structure 20 as a reference plane.

  On the other hand, in order to satisfy the improvement of the electron injection efficiency EIE and the improvement of the light extraction efficiency LEE at the same time, it is preferable that the flat surface at the bottom of the concave portion of the fine concavo-convex structure 20 is asymptotic to zero. In order to improve the electron injection efficiency EIE in the LED element, it is necessary to effectively increase the specific surface area and decrease the contact resistance. On the other hand, in order to improve the light extraction efficiency LEE, it is necessary to effectively disturb the waveguide mode. From the above viewpoint, the ratio (lcvb / lccb) between the convex bottom width lcvb and the concave bottom width lccb of the fine concavo-convex structure 20 is preferably 0.33 or more. In particular, from the viewpoint of increasing the specific surface area and improving the scattering property, (lcvb / lccb) is more preferably 0.6 or more, and most preferably 3 or more.

  Further, if the convex top width lcvt is smaller than the convex bottom width lcvb, it becomes easy to satisfy the ratio (lcvt / lcct) and the ratio (lcvb / lcccb) described above. With this mechanism, the internal quantum efficiency IQE or electron injection efficiency EIE and the light extraction efficiency LEE can be increased simultaneously.

  Further, if the fine concavo-convex structure 20 is a dot structure, it is easy to control the convex top width lcvt and the convex bottom width lcvb, and it is easy to satisfy the ratio (lcvt / lcct) and the ratio (lcvb / lccc) at the same time. For this reason, the internal quantum efficiency IQE or electron injection efficiency EIE and the light extraction efficiency LEE can be simultaneously increased by the mechanism described above.

  Next, materials and the like of each layer constituting the semiconductor light emitting element 100 will be described. The semiconductor light emitting device 100 according to the present embodiment improves both the internal quantum efficiency IQE and the light extraction efficiency LEE because the fine concavo-convex structure 20 provided on the surface of the optical substrate 10 has a shape that satisfies the above range. As long as this effect is exhibited, the material, state, number of layers or thickness of each semiconductor layer, electrode material, number of layers, arrangement or thickness, material of the light emitting semiconductor layer, number of layers, or The thickness, the material of the growth substrate, the plane orientation, the thickness, etc. can be selected as appropriate and are not particularly limited.

  As shown in FIG. 1, a first semiconductor layer 30, a light emitting semiconductor layer 40, and a second semiconductor layer 50 are sequentially stacked on a fine concavo-convex structure 20 of an optical substrate 10 having a fine concavo-convex structure 20 on a surface, and a semiconductor. The light emitting element 100 is configured. Here, the emitted light generated in the light emitting semiconductor layer 40 is extracted from the second semiconductor layer 50 side or the optical substrate 10. Further, the first semiconductor layer 30 and the second semiconductor layer 50 are different semiconductor layers. Here, it is preferable that the first semiconductor layer 30 planarizes the fine concavo-convex structure 20. By providing the first semiconductor layer 30 so as to planarize the fine concavo-convex structure 20, the performance of the first semiconductor layer 30 as a semiconductor can be reflected in the light emitting semiconductor layer 40 and the second semiconductor layer 50. Therefore, the internal quantum efficiency IQE is improved. The first semiconductor layer 30 may be composed of an undoped first semiconductor layer 31 and a doped first semiconductor layer 32.

  Further, the light emitting semiconductor layer 40 and the second semiconductor layer 50 are stacked following the first semiconductor layer 30. Thereafter, the transparent conductive film 60 may be provided on the second semiconductor layer 50 of the semiconductor light emitting device 100, the anode electrode 70 may be provided on the surface of the transparent conductive film 60, and the cathode electrode 80 may be provided on the surface of the first semiconductor layer 30. it can. The arrangement of the transparent conductive film 60, the anode electrode 70, and the cathode electrode 80 is not limited because it can be optimized as appropriate by the semiconductor light emitting device, but is generally provided as illustrated in FIG.

First semiconductor layer 30
The first semiconductor layer 30 is not particularly limited as long as it can be used as an n-type semiconductor layer suitable for a semiconductor light emitting device (LED). For example, elemental semiconductors such as silicon and germanium, and compound semiconductors such as III-V, II-VI, and VI-VI can be appropriately doped with various elements.

  The first semiconductor layer 30 and the fine relief structure 20 can be appropriately combined from the viewpoint of reducing dislocations in the first semiconductor layer 30. When the flat surface of the bottom of the concave portion 20b of the fine concavo-convex structure 20 and the surface substantially parallel to the stable growth surface of the first semiconductor layer 30 are parallel to each other, in the vicinity of the concave portion 20b of the fine concavo-convex structure 20 Since the disorder of the growth mode of the first semiconductor layer 30 is increased and the dislocations in the first semiconductor layer 30 can be effectively dispersed according to the fine concavo-convex structure 20, the internal quantum efficiency IQE is improved. The stable growth surface is the surface with the slowest growth rate in the material to be grown.

  Further, since the fine concavo-convex structure 20 is nanoscale, the thickness necessary for planarizing the fine concavo-convex structure 20 in the first semiconductor layer 30 is reduced. For this reason, since the semiconductor layer that absorbs light from the light emitting semiconductor layer 40 is thinned, the light extraction efficiency LEE is expected to be further improved, and the first semiconductor layer 30 and the light emitting semiconductor layer 40 sequentially stacked thereon are also provided. And it becomes possible to suppress the curvature of the 2nd semiconductor layer 50, and it can be set as a semiconductor light emitting element of a larger area than before. For this reason, the thickness of the first semiconductor layer 30 is preferably 5 μm or less, more preferably 4 μm or less, further preferably 3.5 μm or less, still more preferably 2.5 μm or less, and most preferably 1.5 μm or less.

-Light emitting semiconductor layer 40
The light emitting semiconductor layer 40 is not particularly limited as long as it has a light emitting characteristic as a semiconductor light emitting element (LED). For example, as the light emitting semiconductor layer 40, a semiconductor layer such as AsP, GaP, AlGaAs, InGaN, GaN, AlGaN, ZnSe, AlHaInP, or ZnO can be applied. Further, the light emitting semiconductor layer 40 may be appropriately doped with various elements according to characteristics.

Second semiconductor layer 50
The second semiconductor layer 50 is not particularly limited as long as it can be used as a p-type semiconductor layer suitable for a semiconductor light emitting device (LED). For example, elemental semiconductors such as silicon and germanium, and compound semiconductors such as III-V, II-VI, and VI-VI can be appropriately doped with various elements.

  In the semiconductor light emitting device 100 according to the present embodiment, for example, a III-V group semiconductor is preferably used for the first semiconductor layer 30, the light emitting semiconductor layer 40, and the second semiconductor layer 50.

  Moreover, it is preferable to use a GaN-based semiconductor for the first semiconductor layer 30, the light emitting semiconductor layer 40, and the second semiconductor layer 50. The GaN system is preferable because the band gap can be changed relatively easily by doping impurities such as In. In addition, since the band gap is large and the emission wavelength region is wide from blue to ultraviolet, a wider range of phosphors can be excited, which is preferable.

Optical substrate 10
The material of the optical substrate 10 is not particularly limited as long as it can be used as a substrate for a semiconductor light emitting device. Sapphire, SiC, SiN, GaN, W-Cu, silicon, zinc oxide, magnesium oxide, manganese oxide, zirconium oxide, manganese zinc iron oxide, magnesium aluminum oxide, zirconium boride, gallium oxide, indium oxide, lithium gallium oxide, oxidation Substrates such as lithium aluminum, neodymium gallium oxide, lanthanum strontium aluminum tantalum, strontium titanium oxide, titanium oxide, hafnium, tungsten, molybdenum, GaP, and GaAs can be used. Of these, sapphire, GaN, GaP, GaAs, SiC substrate, Si substrate, spinel substrate and the like are preferably used from the viewpoint of lattice matching with the semiconductor layer. Furthermore, it may be used alone, or may be a heterostructure substrate in which another substrate is provided on the substrate body using these. For example, a sapphire substrate having a C plane (0001) as the main surface can be used as the optical base material 10.

  In particular, when a sapphire substrate is used, a substrate having an inclination of more than 0 ° and not more than α °, that is, an offset angle with respect to the c-plane is preferable.

  Since the sapphire substrate has the fine concavo-convex structure 20 that satisfies the above conditions (1) to (3) and has the offset angle in the above range, the occurrence of defects during the formation of the first semiconductor layer is suppressed, and the internal The quantum efficiency IQE can be further improved. This is presumed to be the following reason.

  In general, a sapphire substrate used for epitaxial growth is polished and has smoothness at the level of several atoms. In the c-plane substrate, the inclination of the c-axis from the normal of the substrate is the offset angle α. The substrate having the offset angle α has a terrace and a step.

  When the sapphire substrate has an offset angle α, a substrate having flatness suitable for epitaxial growth can be obtained by etching the substrate under predetermined conditions (gas flow rate, gas pressure, plasma output). At this time, it is assumed that a terrace and a step determined by the offset angle are formed in the flat portion of the concavo-convex structure, that is, the flat surface of the bottom of the concave portion.

  FIG. 12 is a schematic cross-sectional view showing a sapphire substrate having an offset angle according to the present embodiment. The sapphire substrate 10 having the offset angle α normally forms a single step having a height corresponding to one atomic layer, and has a predetermined step height S and terrace width W depending on the offset angle α. In the case of a sapphire substrate, the single step height is approximately 0.2 nm. The step height S has an upper limit from the flatness of the sapphire substrate 10 and is estimated to be several times the height of the single step. The terrace width W, the offset angle α, and the step height S are expressed by the following equation (4).

  13 and 14 are explanatory views for explaining a sapphire substrate having an offset angle and a fine concavo-convex structure. The symbols P, lcvb, and lccb are as described with reference to FIG. 9B. When the fine concavo-convex structure 20 having a dot structure is provided on the sapphire substrate 10, since the step height S is the number of atomic layers, if the offset angle α is very small, the terrace width W is larger than the concave bottom width lccb. (FIG. 13). At this time, since there is no step in the range of the recess bottom width lccb, it is presumed that it behaves as if the substrate had an offset angle α of 0 degrees. Therefore, if the offset angle α is such that the terrace width W is smaller than the recess bottom width lccb (FIG. 14), the occurrence of defects is suitably suppressed by the steps at the recess bottom, and the internal quantum efficiency IQE is further improved. 13 and 14 are examples in which the fine concavo-convex structure 20 having the dot structure is applied to the sapphire substrate 10 having the offset angle α, but the present invention is not limited to this.

  When providing the fine concavo-convex structure 20, if the duty is increased from the viewpoint of light extraction efficiency LEE, in the fine concavo-convex structure 20, there are few flat portions, that is, the concave bottom width lccb is small, or the etching accuracy of the substrate is not good. The step height S increases. In these cases, in order to prevent the terrace width W from exceeding the recess bottom width lccb, the offset angle α is preferably 0.05 degrees or more, and more preferably 0.1 degrees or more.

  Further, as the average pitch Pav decreases, the ratio of the in-plane average height Hav to the wavelength λ of the emitted light and the aspect ratio are set in a desired range while preventing the terrace width W from exceeding the recess bottom width lccb. Is difficult. This is because the average in-plane height Hav, the height H that defines the upper limit of the aspect ratio, and the convex bottom width lcvb are limited by the average pitch Pav and the concave bottom width lccb. For this reason, in order to make the ratio of the in-plane average height Hav and the wavelength λ of the emitted light, the aspect ratio, within a preferable range while maintaining the size relationship between the terrace width W and the recess bottom width lccb, the average pitch Pav is particularly 500 nm. In this case, the offset angle α is preferably 0.2 degrees or more.

  On the other hand, if the offset angle α is too large and the terrace width W becomes small even at the same step height S, and there are many terraces and steps in the flat portion in the fine concavo-convex structure 20, crystal growth is hindered. It will cause defects. Therefore, from the viewpoint of sufficiently improving the internal quantum efficiency IQE by providing the fine concavo-convex structure 20, the offset angle α is preferably 0.8 degrees or less, and more preferably 0.5 degrees or less.

  Further, the optical substrate 10 may be removed at least after the first semiconductor layer 30 is stacked. By removing the optical base material 10, the effect of disturbing the waveguide mode is increased, so that the light extraction efficiency LEE is greatly improved. In this case, the light emission surface of the emitted light of the semiconductor light emitting device 100 is preferably on the first semiconductor layer 30 side when viewed from the light emitting semiconductor layer 40.

  At this time, a fine pattern in which the nanoscale concavo-convex structure on the optical base material 10 is reversed is formed on the peeled surface of the first semiconductor layer 30, but the fine pattern of the semiconductor layer and the concavo-convex structure of the optical base material 10 are included. Need not be exactly the same. That is, when the optical substrate 10 is peeled from the semiconductor layer, even if a part of the concavo-convex structure partially remains in the fine pattern, a part of the fine pattern remains partially in the concavo-convex structure. May be. Even in such a case, the effect of the present invention that the internal quantum efficiency IQE and the light extraction efficiency LEE are compatible is exhibited without problems.

  In addition, the peeling of the above-mentioned optical base material 10 can use the lift-off using a laser beam, and the total dissolution or partial dissolution of an optical base material. In particular, removal by dissolution is preferable in terms of the accuracy of the pattern formed on the release surface.

In the optical substrate according to the present embodiment, the material of the transparent conductive film is not particularly limited as long as it can be used as a transparent conductive film suitable for LEDs. For example, a metal thin film such as a Ni / Au electrode or a conductive oxide film such as ITO, ZnO, In ₂ O ₃ , SnO ₂ , IZO, or IGZO can be applied. In particular, ITO is preferable from the viewpoints of transparency and conductivity.

-Other uneven structure As shown in FIG. 1, in the semiconductor light-emitting device 100, although the fine uneven structure 20 is provided between the optical base material 10 and the 1st semiconductor layer 30, as shown in FIG. Further, another uneven structure can be provided. FIG. 15 is a schematic cross-sectional view showing another example of the semiconductor light emitting device according to the present embodiment. As shown in FIG. 15, the uneven structure provided separately in the semiconductor light emitting device 800 includes the following.
The uneven structure 801 provided on the surface of the anode electrode 70 and the surface of the cathode electrode 80
An uneven structure 802 provided between the anode electrode 70 and the transparent conductive film 60 or between the cathode electrode 80 and the first semiconductor layer 30.
Uneven structure 803 provided on the surface of the transparent conductive film 60
The concavo-convex structure 804 provided between the second semiconductor layer 50 and the transparent conductive film 60
The uneven structure 805 provided on the side surfaces of the first semiconductor layer 30, the light emitting semiconductor layer 40, the second semiconductor layer 50, and the transparent conductive film 60.
The concavo-convex structure 806 provided on the surface facing the semiconductor laminate surface of the optical substrate 10

  As described above, in addition to the fine concavo-convex structure 20, by providing at least one of the concavo-convex structures 801 to 806, the effects corresponding to the concavo-convex structures 801 to 806 described below can be exhibited. .

  By providing the concavo-convex structure 801, a contact area at the time of wiring increases, and peeling is suppressed. Furthermore, the electron injection efficiency EIE can be increased.

  By providing the concavo-convex structure 802, the contact area between the anode electrode 70 and the transparent conductive film 60 or the contact area between the cathode electrode 80 and the first semiconductor layer 30 increases, so that the electron injection efficiency EIE can be improved. The quantum efficiency EQE can be improved.

  By providing the concavo-convex structure 803, the waveguide mode is easily destroyed even at the interface between the transparent conductive film 60 having a high refractive index and the air, so that the light extraction efficiency LEE can be improved.

  By providing the concavo-convex structure 804, the waveguide mode in the second semiconductor layer 50 is easily disturbed, and the light extraction efficiency LEE can be improved. Furthermore, since the contact area between the second semiconductor layer 50 and the transparent conductive film 60 can be increased, the electron injection efficiency EIE can be increased.

  By providing the concavo-convex structure 805, the amount of emitted light emitted from the side surfaces of the first semiconductor layer 30, the light emitting semiconductor layer 40, and the second semiconductor layer 50 can be increased. Can be reduced. For this reason, the light extraction efficiency LEE is improved, and the external quantum efficiency EQE can be increased.

  By providing the concavo-convex structure 806, the waveguide mode in the optical substrate 10 having a high refractive index with respect to air can be disturbed on the back surface of the optical substrate 10 and the light extraction efficiency LEE can be improved. it can. This is particularly advantageous in the flip chip type.

  As the concavo-convex structure 801 to 806, the shape, arrangement, size, etc. of the fine concavo-convex structure 20 described above can be adopted, and thereby the effect according to the fine concavo-convex structure 20 (improvement of electron injection efficiency EIE, light The improvement of the extraction efficiency LEE, the increase in the area of the semiconductor light emitting device 800, and the suppression of electrode peeling) can be exhibited.

  In order to form the uneven structures 801 to 806, a known etching method can be used. This type of etching method is, for example, wet chemical etching or dry etching. Such etching methods include reactive ion etching, ion beam etching or chemically assisted ion beam etching.

  Next, each process of the manufacturing method of the semiconductor light emitting device 100 according to the present embodiment will be described.

-Optical base material preparatory process The optical base material 10 will not be limited if the fine concavo-convex structure 20 demonstrated above can be formed, its production method is not limited, transfer method, photolithography method, thermal lithography method, electron beam drawing method, It can be produced by an interference exposure method, a lithography method using nanoparticles as a mask, a lithography method using a self-organized structure as a mask, or the like. In particular, it is preferable to employ a transfer method from the viewpoint of processing accuracy and processing speed of the fine concavo-convex structure 20 of the optical substrate 10.

  The transfer method in this specification is defined as a method including a step of transferring a texture of a mold having a texture on the surface to an object to be processed (an optical base material before producing a concavo-convex structure). That is, it is a method including at least a step of bonding a texture of a mold and an object to be processed through a transfer material and a step of peeling the mold. More specifically, the transfer method can be classified into two. First, the transfer material transferred to the object to be processed is used as a permanent agent. In this case, the materials constituting the optical base body and the fine concavo-convex structure are different. The fine uneven structure remains as a permanent agent and is used as a semiconductor light emitting device. Since the semiconductor light emitting element is used for a long period of tens of thousands of hours, when the transfer material is used as a permanent agent, the material constituting the transfer material preferably contains a metal element. In particular, it is preferable to include a metal alkoxide that generates a hydrolysis / polycondensation reaction or a metal alkoxide condensate as a raw material because the performance as a permanent agent is improved. Secondly, there is a nanoimprint lithography method. The nanoimprint lithography method includes a step of transferring a texture of a mold onto a target object, a step of providing a mask for processing the target object by etching, and a step of etching the target object. For example, when one type of transfer material is used, first, the object to be processed and the mold are bonded via the transfer material. Subsequently, the transfer material is cured by heat or light (UV), and the mold is peeled off. Etching typified by oxygen ashing is performed on the concavo-convex structure made of a transfer material to partially expose the object to be processed. Thereafter, the object to be processed is processed by etching using the transfer material as a mask. As a processing method at this time, dry etching and wet etching can be employed. Dry etching is useful for increasing the height of the concavo-convex structure. For example, when two types of transfer materials are used, a first transfer material layer is first formed on the object to be processed. Subsequently, the first transfer material layer and the mold are bonded via the second transfer material. Thereafter, the transfer material is cured by heat or light (UV), and the mold is peeled off. Etching typified by oxygen ashing is performed on the concavo-convex structure formed of the second transfer material to partially expose the first transfer material. Subsequently, the first transfer material layer is etched by dry etching using the second transfer material layer as a mask. Thereafter, the object to be processed is processed by etching using the transfer material as a mask. As a processing method at this time, dry etching and wet etching can be employed. Dry etching is useful for increasing the height of the fine relief structure.

  As described above, by adopting the transfer method, the texture of the mold can be reflected on the object to be processed, so that a good optical substrate 10 can be obtained.

  The material of the imprint mold is not particularly limited, and non-flexible glass, quartz, sapphire, nickel, or flexible resin can be used. Among them, it is preferable to use a flexible mold because the transfer accuracy of the texture of the mold is improved and the accuracy of the fine concavo-convex structure 20 of the optical substrate 10 is improved.

-Semiconductor layer and transparent conductive film lamination process On the fine uneven structure 20 of the optical base material 10, the 1st semiconductor layer 30, the light emitting semiconductor layer 40, the 2nd semiconductor layer 50, and the transparent conductive film 60 are formed into a film in order. The method for forming each semiconductor layer is not particularly limited, but the well-known metal organic chemical vapor deposition method (MOCVD method), molecular beam crystal growth method (MBE method), halide vapor phase epitaxy method (HVPE method), sputtering method, It can be formed by an ion plating method, an electron shower method, or the like.

  Since the fine concavo-convex structure 20 is a nanoscale structure, the thickness of the first semiconductor layer 30 can be reduced. This means that the manufacturing time and cost are reduced as compared with the conventional method. In LED manufacturing, the (MO) CVD process, which is a semiconductor crystal layer deposition process, is rate-limiting, lowering throughput and raising material costs. The ability to reduce the amount of semiconductor crystals means an improvement in throughput of the (MO) CVD process and a reduction in materials used, which is an important requirement for manufacturing.

-Electrode formation process The 2nd electrode layer 90 is formed in the surface (peeling surface) of the exposed 1st semiconductor layer 30. As shown in FIG. The material of the second electrode layer 90 may be any material that can contact the first semiconductor layer 30 with low resistance. The second electrode layer 90 may have a structure including only the pad portion. However, a wiring electrode having a wiring pattern such as a lattice shape or a radial shape is provided on the pad portion so as to improve current diffusion in the element surface direction. It may be. In addition, when a transparent electrode such as ITO is provided between the first semiconductor layer 30 and the second electrode layer 90, it is possible to create a large-area LED by increasing the contact area, and to improve current diffusibility. Therefore, improvement of the electron injection efficiency EIE is expected.

-Cutting process Using a laser scrip or a laser dicer, individual change processing is performed to divide into light emitting elements.

Examples performed to confirm the effects of the present invention will be described below. The symbols used in the following description have the following meanings.
・ DACHP: Fluorine-containing urethane (meth) acrylate (OPTOOL DAC HP (manufactured by Daikin Industries))
M350: trimethylolpropane (EO-modified) triacrylate (M350, manufactured by Toagosei Co., Ltd.)
・ I. 184 ... 1-hydroxycyclohexyl phenyl ketone (Irgacure (registered trademark) 184, manufactured by BASF)
・ I. 369 ... 2-benzyl-2-dimethylamino-1- (4-morpholinophenyl) -butanone-1 (Irgacure (registered trademark) 369, manufactured by BASF)
-TTB: Titanium (IV) tetrabutoxide monomer (Wako Pure Chemical Industries, Ltd.)
SH710: Phenyl-modified silicone (Toray Dow Corning)
・ 3APTMS ... 3-acryloxypropyltrimethoxysilane (KBM5103 (manufactured by Shin-Etsu Silicone))
DIBK: diisobutyl ketone MEK: methyl ethyl ketone MIBK: methyl isobutyl ketone DR833: tricyclodecane dimethanol diacrylate (SR833 (manufactured by SARTOMER))
SR368 ... Tris (2-hydroxyethyl) isocyanurate triacrylate (SR833 (manufactured by SARTOMER)

(Example)
An optical base material having a fine concavo-convex structure on the surface is manufactured, a first semiconductor layer, a light emitting semiconductor layer, and a second semiconductor layer are sequentially stacked on the optical base material to manufacture a semiconductor light emitting device (LED). The efficiency was compared. At this time, the arrangement and shape of the fine concavo-convex structure were changed.

  In the following examination, in order to produce an optical substrate having a fine concavo-convex structure on the surface, first, (1) a cylindrical master mold is produced, and (2) the optical transfer method is applied to the cylindrical master mold. Thus, a reel-shaped resin mold was produced. (3) Thereafter, the reel-shaped resin mold was processed into a nano-processing member (nano-processing film) of an optical substrate. Subsequently, (4) using a film for nano-processing, forming a mask on the optical substrate, and performing dry etching through the obtained mask, an optical substrate having a fine concavo-convex structure on the surface Produced. (5) Each semiconductor layer was laminated | stacked on the optical base material, and the transparent conductive film was vapor-deposited. Finally, (6) electrodes were attached and performance was evaluated.

(1) Production of cylindrical master mold A texture was formed on the surface of cylindrical quartz glass by a direct drawing lithography method using a semiconductor laser. First, a resist layer was formed on the surface of the cylindrical quartz glass by a sputtering method. The sputtering method was carried out using φ3 inch CuO (containing 8 atm% Si) as a target (resist layer) with a power of RF 100 W to form a 20 nm resist layer. Thereafter, the entire surface of the cylindrical quartz glass was exposed once. Subsequently, exposure was performed using a semiconductor laser having a wavelength of 405 nm while rotating the cylindrical quartz glass. Next, the resist layer after exposure was developed. The development of the resist layer was performed for 240 seconds using a 0.03 wt% glycine aqueous solution. Next, using the developed resist layer as a mask, the etching layer (quartz glass) was etched by dry etching. Dry etching was performed using SF ₆ as an etching gas under the conditions of a processing gas pressure of 1 Pa, a processing power of 300 W, and a processing time of 5 minutes. Finally, only the resist layer residue was peeled off from the cylindrical quartz glass having a texture on the surface using hydrochloric acid having a pH of 1. The peeling time was 6 minutes.

  Durasurf HD-1101Z (made by Daikin Chemical Industries), which is a fluorine-based mold release agent, is applied to the texture of the obtained cylindrical quartz glass, heated at 60 ° C. for 1 hour, and then allowed to stand at room temperature for 24 hours. Immobilized. Then, it wash | cleaned 3 times by Durasurf HD-ZV (made by Daikin Chemical Industries), and the cylindrical master mold was obtained.

(2) Production of reel-shaped resin mold Using the produced cylindrical master mold as a mold, the optical nanoimprint method was applied to continuously produce a reel-shaped resin mold G1. Subsequently, a reel-shaped resin mold G2 was continuously obtained by an optical nanoimprint method using the reel-shaped resin mold G1 as a template.

The material 1 shown below was apply | coated to the easily bonding surface of PET film A-4100 (Toyobo Co., Ltd .: width 300mm, thickness 100micrometer) by the micro gravure coating (manufactured by Yurai Seiki Co., Ltd.) so that it might become a coating film thickness of 5 micrometers. . Next, the PET film coated with the material 1 is pressed against the cylindrical master mold with a nip roll so that the integrated exposure amount under the center of the lamp is 1500 mJ / cm ^{2 at} 25 ° C. and 60% humidity in the air. In addition, a UV-irradiated UV exposure apparatus (H bulb) manufactured by Fusion UV Systems Japan Co., Ltd. was used to irradiate ultraviolet rays, and continuous photo-curing was performed. , Width 300 mm).

  Next, the reel-shaped resin mold G1 was regarded as a template, and the optical nanoimprint method was applied to continuously produce the reel-shaped resin mold G2.

Material 1 was applied to an easily adhesive surface of PET film A-4100 (manufactured by Toyobo Co., Ltd .: width 300 mm, thickness 100 μm) by microgravure coating (manufactured by Yurai Seiki Co., Ltd.) so as to have a coating thickness of 3 μm. Next, the PET film coated with the material 1 is pressed against the textured surface of the reel-shaped resin mold G1 with a nip roll (0.1 MPa), and integrated exposure under the center of the lamp at 25 ° C. and 60% humidity in the air. Ultraviolet rays were irradiated using a UV exposure apparatus (H bulb) manufactured by Fusion UV Systems Japan Co., Ltd. so that the amount was 1200 mJ / cm ^2, and photocuring was continuously performed, and the texture was transferred to the surface. A plurality of reel-shaped resin molds G2 (length 200 m, width 300 mm) were obtained.
Material 1 ... DACHP: M350: I. 184: I.D. 369 = 17.5 g: 100 g: 5.5 g: 2.0 g

(3) Production of Nano-Processing Film A diluent of the following material 2 was applied to the textured surface of the reel-shaped resin mold G2. Subsequently, a diluted solution of the following material 3 was applied on the textured surface of the reel-shaped resin mold G2 enclosing the material 2 inside the texture to obtain a nano-processing film.
Material 2 ... TTB: 3APTMS: SH710: I. 184: I.D. 369 = 65.2 g: 34.8 g: 5.0 g: 1.9 g: 0.7 g
Material 3 ... Binding polymer: SR833: SR368: I.I. 184: I.D. 369 = 77.1 g: 11.5 g: 11.5 g: 1.47 g: 0.53 g
Binding polymer: Methyl ethyl ketone solution of binary copolymer of 80% by mass of benzyl methacrylate and 20% by mass of methacrylic acid (solid content 50%, weight average molecular weight 56000, acid equivalent 430, dispersity 2.7)

  (2) Using a device similar to the production of the reel-shaped resin mold, the material 2 diluted with PGME was directly applied onto the texture surface of the reel-shaped resin mold G2. Here, the dilution concentration was set such that the solid content contained in the coating raw material per unit area (material 2 diluted with PGME) was 20% or more smaller than the texture volume per unit area. After coating, the material was passed through an air-drying oven at 80 ° C. for 5 minutes, and the reel-shaped resin mold G2 containing the material 2 inside the texture was wound up and collected.

  Subsequently, the reel-shaped resin mold G2 that encloses the material 2 in the texture is unwound, and (2) the material 3 diluted with PGME and MEK is used for the texture using the same device as the production of the reel-shaped resin mold. Direct coating on the surface. Here, the dilution concentration was set such that the distance between the interface between the material 2 disposed inside the texture and the coated material 3 and the surface of the material 3 was 400 nm to 800 nm. After coating, the material was passed through an air-drying oven at 80 ° C. for 5 minutes, and a cover film made of polypropylene was put on the surface of the material 3 and wound up and collected.

(4) Nano-processing of growth base material Using the prepared nano-processing film, processing of the growth base material was attempted. A c-plane sapphire substrate was used as the growth substrate. In addition, seven types of offset angles of the sapphire substrate were prepared: 0 degree, 0.05 degree, 0.1 degree, 0.2 degree, 0.5 degree, 0.8 degree, and 1 degree.

The sapphire substrate was subjected to UV-O ₃ treatment for 5 minutes to remove surface particles and to make it hydrophilic. Then, the material 3 surface of the film for nano processing was bonded with respect to the sapphire base material. At this time, the sapphire substrate was bonded in a state heated to 80 ° C. Subsequently, using a high-pressure mercury lamp light source, light was irradiated through the reel-shaped resin mold G2 so that the integrated light amount was 1200 mJ / cm ² . Thereafter, the reel-shaped resin mold G2 was peeled off.

Etching using oxygen gas is performed from the material 2 surface side of the obtained laminate (material 2 / material 3 / substrate laminate), and the material 3 is nano-processed using the material 2 as a mask. Partially exposed. Oxygen etching was performed under conditions of a pressure of 1 Pa and a power of 300 W. Subsequently, reactive ion etching using BCl ₃ gas was performed from the material 2 surface side to nano-process the sapphire substrate. Etching using BCl 3 was performed at ICP: 150 W, BIAS: 50 W, and pressure 0.2 Pa, and a reactive ion etching apparatus (RIE-101iPH, manufactured by Samco Corporation) was used.

  Finally, it was washed with a solution in which sulfuric acid and hydrogen peroxide solution were mixed at a weight ratio of 2: 1 to obtain a sapphire substrate having a fine relief structure on the surface. In addition, the shape of the fine concavo-convex structure produced on the sapphire substrate was controlled mainly by the filling rate of the material 2 of the film for nano-processing and the film thickness of the material 3.

  The shape of the fine concavo-convex structure produced on the surface of the sapphire substrate was appropriately controlled according to the shape of the texture produced in the cylindrical master mold, the nip pressure conditions when producing the resin mold, and the dry etching treatment conditions.

(5) Fabrication of Semiconductor Light-Emitting Device On the obtained sapphire substrate, 100 × low temperature growth buffer layer of Al _x Ga _1-x N (0 ≦ x ≦ 1) was formed as a buffer layer. Next, undoped GaN was formed as the undoped first semiconductor layer 31, and Si-doped GaN was formed as the doped first semiconductor layer. Subsequently, a strain absorption layer is provided, and then, as the light emitting semiconductor layer 40, a multi-quantum well active layer (well layer, barrier layer = undoped InGaN, Si-doped GaN) is formed with the respective film thicknesses (60 mm, 250 mm). The layers were alternately stacked so that there were 6 layers and 7 barrier layers. On the light emitting semiconductor layer, Mg-doped AlGaN, undoped GaN, and Mg-doped GaN were stacked as a second semiconductor layer so as to include an electroblocking layer. Subsequently, an ITO film was formed and etched, and then an electrode pad was attached as an electrode part. In this state, a 20 mA current was passed between the p electrode pad and the n electrode pad using a prober, and the light emission output was measured. In addition, about the output directivity, it measured by the light distribution measuring apparatus whether it could arrange | position so that an electrode might not enter into a diffraction azimuth angle, and evaluated. ○ When the electrode can be provided so as to avoid the diffraction azimuth, the light emission distribution is somewhat scattering, and even if the electrode is provided so as to avoid the diffraction azimuth, a part of the electrode is applied to the electrode, The case where the shadow was thinly observed was indicated by Δ, and the case where the output light distribution was scattering, the electrode could not be provided so as to avoid the diffraction azimuth angle, and the electrode portion became a shadow was indicated by ×. The light emission output ratio was evaluated assuming that the output when a sapphire substrate having no uneven structure described in Comparative Example 3 in Table 1 was used was 1.

  In addition, the comparative example 3 of Table 1 is a case where the semiconductor light-emitting device is manufactured using the flat sapphire substrate which does not comprise a fine uneven structure.

  In the examples, in order to improve the internal quantum efficiency IQE, a sapphire substrate having a fine concavo-convex structure having a dot-like convex portion or a line-and-space structure was used. Therefore, when the fine concavo-convex structure was observed with a scanning electron microscope, it had a plurality of convex portions. The shape of one convex portion is a shape that is curved so that the side surface from the top of the convex portion to the bottom of the convex portion swells, and the inclination angle of the side surface changes in multiple steps. The central wavelength of light emission was 450 nm, and the refractive index n of the semiconductor layer was 2.39.

  The internal quantum efficiency IQE was determined from the PL intensity. The internal quantum efficiency IQE is defined by (number of photons emitted from the light emitting semiconductor layer 40 per unit time / number of electrons injected into the semiconductor light emitting device 100 per unit time). In this example, (PL intensity measured at 300K / 10 PL intensity measured at 10K) was adopted as an index for evaluating the internal quantum efficiency IQE.

  In Table 1, H represents the height of the convex portion, D represents the duty, α represents the offset angle, and A represents the aspect ratio. As can be seen from Table 1, with respect to various arrangements and pitches, compared with the case where the fine uneven structure is not provided (Comparative Example 3), the case where the fine uneven structure is provided (Comparative Example 1, Comparative Example 2, Comparative Example 4). It can be seen that the light emission output of Comparative Example 13 and Examples 1 to 22 is improved.

  Comparative Example 1 is a dot array arranged in a regular hexagon, in which the average pitch is 300 nm, the duty is 0.48, the height is 89 nm, and the offset angle is 0.2 degrees. In this case, it can be seen that the light emission output hardly increases. This is because the average pitch is as small as 300 nm, the dislocation density is reduced, the internal quantum efficiency IQE is improved, and the wave nature of light is emphasized, so that an electrode is provided so as to avoid the diffraction azimuth. However, it can be considered that light projection is not effectively generated and the light extraction efficiency LEE is hardly improved because the height of the convex portion is low and the convex portion volume is not large.

  Comparative Example 2 is a dot arrangement arranged in a regular hexagon, where the average pitch is 3000 nm, the height is 1403 nm, in the micro order, and the offset angle is 0.2 degrees. In this case, since the waveguide mode is broken, the light emission output is improved as compared with the case where there is no concavo-convex structure. However, since the average pitch is large, the improvement of the internal quantum efficiency IQE is small, and the light and the concavo-convex structure. The interaction was scattering, and the electrode could not be provided so as to avoid the diffraction azimuth angle, and the electrode portion was shaded.

  Comparative Example 4 is a dot array arranged in a regular hexagon, in which the average pitch is 700 nm, the duty is 0.63, the height is 53 nm, and the offset angle is 0.2 degrees. In this case, it can be seen that the light emission output hardly increases. This is because the average pitch is as small as 700 nm, so the dislocation density is reduced and the internal quantum efficiency IQE is improved. However, since the convex height is low and the convex volume is not large, optical diffraction is effectively generated. This is probably because the light extraction efficiency LEE is hardly improved.

  Comparative Example 5 is a case where the offset angle is 0 degrees with substantially the same dimensions and the same arrangement as in Example 1. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but the offset angle is 0 degree, so that the improvement of the internal quantum efficiency IQE by the fine uneven structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  Comparative Example 6 is a case where the offset angle is 1 degree with substantially the same dimensions and the same arrangement as in Example 1. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but since the offset angle is 1 degree, the improvement of the internal quantum efficiency IQE by the fine concavo-convex structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  Comparative Example 7 is a case where the dimensions and arrangement are substantially the same as in Example 3, and the offset angle is 1 degree. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but since the offset angle is 1 degree, the improvement of the internal quantum efficiency IQE by the fine concavo-convex structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  Comparative Example 8 is a case where the dimensions and arrangement are substantially the same as those of Example 4, and the offset angle is 0 degree. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but the offset angle is 0 degree, so that the improvement of the internal quantum efficiency IQE by the fine uneven structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  The comparative example 9 is a case where it has substantially the same dimensions and the same arrangement as the fourth embodiment and the offset angle is 1 degree. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but since the offset angle is 1 degree, the improvement of the internal quantum efficiency IQE by the fine concavo-convex structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  Comparative Example 10 is the case where the dimensions and arrangement are substantially the same as in Example 8, and the offset angle is 0 degrees. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but the offset angle is 0 degree, so that the improvement of the internal quantum efficiency IQE by the fine uneven structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  Comparative Example 11 is a case where the dimensions and arrangement are almost the same as those of Example 8, and the offset angle is 1 degree. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but since the offset angle is 1 degree, the improvement of the internal quantum efficiency IQE by the fine concavo-convex structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  Comparative example 12 is a case where the dimensions and arrangement are substantially the same as in example 5, and the offset angle is 0 degrees. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but the offset angle is 0 degree, so that the improvement of the internal quantum efficiency IQE by the fine uneven structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  The comparative example 13 is a case where it has substantially the same dimensions and the same arrangement as the fifth embodiment and the offset angle is 1 degree. In this case, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light and the aspect ratio are within a predetermined range, but since the offset angle is 1 degree, the improvement of the internal quantum efficiency IQE by the fine concavo-convex structure is effective. Therefore, it is considered that the improvement of the light emission output ratio is small.

  Subsequently, Example 1 is a dot arrangement arranged in a regular hexagon, in which the average pitch is 700 nm, the height is 286 nm, the duty is 0.79, and the offset angle is 0.2 degrees. Since it is a regular hexagonal arrangement, the in-plane average height Hav is expressed by the formula (1). In the case of Example 1, it turns out that the light emission output is increasing compared with the comparative example. This is because the average pitch is nano-order, the internal quantum efficiency is improved, the ratio between the in-plane average height Hav and the wavelength λ of the emitted light, the aspect ratio suitably acts, and the mode that contributes to light extraction more This is presumed to be due to the occurrence of diffraction. The radiated part of the diffracted light was observed symmetrically six times. The electrode part could be provided so as to avoid the diffraction azimuth angle.

  In the second embodiment, the same average pitch and the same offset angle as in the first embodiment are arranged in a square. The in-plane average height Hav is expressed by the formula (2). Also in this case, as in Example 1, improvements in internal quantum efficiency and light emission efficiency were obtained.

  In Example 3, the arrangement is a line and space structure with the same average pitch as in Example 1. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 4 is a case of a regular hexagonal array having an average pitch larger than those of Examples 1 to 3 and a pitch of 900 nm. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In Example 5, an average pitch is 1200 nm and the arrangement is a regular hexagonal dot structure. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 6 is a case of the tatami arrangement shown in FIG. 4B in which line and space are arranged orthogonally for each region with an average pitch of 700 nm. In addition, each area | region was made into the square of 14 micrometers square, and it was made for 20 convex parts to be included. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 7 is the case of the arrangement shown in FIG. 4C in which the lines and spaces are superimposed so as to be orthogonal at an average pitch of 700 nm. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 8 is a case of a regular hexagonal array having an average pitch smaller than those of Examples 1 to 3 and a pitch of 500 nm. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 9 is a dot arrangement arranged in a regular hexagon, with an average pitch of 700 nm, a height of 321 nm, a duty of 0.56, and an offset angle of 0.05 degrees. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 10 is a dot arrangement arranged in a regular hexagon, with an average pitch of 700 nm, a height of 312 nm, a duty of 0.60, and an offset angle of 0.1 degree. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  The eleventh embodiment has substantially the same dimensions and arrangement as the third embodiment, and an offset angle of 0.5 degrees. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  The twelfth embodiment has substantially the same dimensions and arrangement as the tenth embodiment and an offset angle of 0.8 degrees. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 13 is a regular hexagonal dot arrangement with an average pitch of 900 nm, a height of 393 nm, a duty of 0.53, and an offset angle of 0.05 degrees. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  The fourteenth embodiment has substantially the same dimensions and arrangement as the fourth embodiment and an offset angle of 0.1 degree. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  The fifteenth embodiment has substantially the same dimensions and arrangement as the fourth embodiment and an offset angle of 0.5 degrees. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 16 is a line-and-space arrangement in which the average pitch is 900 nm, the height is 333 nm, the duty is 0.48, and the offset angle is 0.8 degrees. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In the seventeenth embodiment, the offset angle is 0.5 degrees with the same dimensions and arrangement as in the eighth embodiment. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  The eighteenth embodiment has substantially the same dimensions and arrangement as the eighth embodiment, and an offset angle of 0.8 degrees. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In the nineteenth embodiment, the offset angle is 0.05 degrees with the same dimensions and arrangement as the fifth embodiment. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  Example 20 is a line-and-space arrangement in which the average pitch is 1200 nm, the height is 428 nm, the duty is 0.47, and the offset angle is 0.1 degree. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In Example 21, Example 5 is substantially the same size and arrangement, and the offset angle is 0.5 degrees. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In the twenty-second embodiment, the offset angle is set to 0.8 degrees with the same dimensions and arrangement as in the fifth embodiment. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  From the above, it can be seen that, regardless of the average pitch and arrangement, if the fine concavo-convex structure of the sapphire substrate satisfies the above conditions (1) to (3), an effect can be obtained. Next, the range of each parameter was examined.

  Table 2 shows the result of changing the average pitch Pav using a substrate with a regular hexagonal arrangement and an offset angle of 0.2 degrees. In Comparative Example 14, the average pitch Pav is 450 nm. Since the average pitch Pav is small and the wave nature of light is emphasized, electrodes can be provided so as to avoid the diffraction azimuth angle. However, since the diffraction angle determined from the average pitch Pav is not appropriate, the luminous efficiency is improved. It is estimated that the degree is small.

  In Example 23, the average pitch Pav is 550 nm. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In Example 24, the average pitch Pav is 600 nm. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In Example 25, the average pitch Pav is 1300 nm. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In Example 26, the average pitch Pav is 1500 nm. The ratio of the in-plane average height Hav and the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges, so that the light emission efficiency is improved, but the light emission distribution is somewhat scattering, Even when the electrode was provided so as to avoid the diffraction azimuth angle, a part of the electrode was applied to the electrode, and the shadow of the electrode portion was observed to be thin. For this reason, although the improvement of luminous efficiency is seen, it is estimated that the improvement degree of luminous efficiency is small compared with the case where average pitch Pav is smaller.

  In Comparative Example 15, the average pitch Pav is 1600 nm. The interaction between the light and the concavo-convex structure was scattering, and the electrode portion was shaded.

  Table 3 shows the results of examining the influence of the aspect ratio using a regular hexagonal array with an average pitch Pav of 700 nm and a substrate with an offset angle of 0.2 degrees. In Comparative Example 16, since the aspect ratio is small and the ratio between the in-plane average height Hav and the wavelength λ of the emitted light is small, it is considered that light diffraction does not effectively occur and the light extraction efficiency LEE is hardly improved. Since Example 27, Example 28, Example 29, and Example 30 are in a desired range, it can be seen that the luminous efficiency is improved. In Example 31, although the light emission efficiency is improved, the aspect ratio is close to 1, and the angle of light diffraction is difficult to contribute to light extraction, so that the light emission output ratio is slightly improved. Inferred. In Example 32, although the luminous efficiency is improved, the aspect ratio is close to 1, the angle of light diffraction is less likely to contribute to light extraction, and the height H is relatively large. It is presumed that the improvement of the light emission output ratio is slightly reduced because the behavior of is slightly scattered and the electrode part is slightly shaded. In Comparative Example 17, the aspect ratio is greater than 1, and the height H of the protrusion is larger than the width of the bottom of the protrusion, so that light diffraction does not occur effectively and the light behavior is also scattered. It is considered that the improvement in the extraction efficiency LEE is small.

  Table 4 shows the results of examining the influence of the aspect ratio at an average pitch Pav of 900 nm, as in Table 3. It can be seen that the tendency is similar to the case of an average pitch of 700 nm.

  Table 5 shows the result of changing the in-plane average height Hav. A substrate with a hexagonal array and an offset angle of 0.2 degrees was used. In Example 39, the average pitch is 900 nm, and the ratio of the in-plane average height Hav to the wavelength λ of the emitted light is 0.157. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In Example 40, the average pitch is 900 nm, and the ratio of the in-plane average height Hav to the wavelength λ of the emitted light is 0.598. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In Example 41, the average pitch is 1200 nm, and the ratio of the in-plane average height Hav to the wavelength λ of the emitted light is 0.793. It can be seen that the luminous efficiency is improved when the ratio of the in-plane average height Hav to the wavelength λ of the emitted light, the aspect ratio, and the offset angle are within the desired ranges.

  In Example 42, the average pitch is 1300 nm, and the ratio between the in-plane average height Hav and the wavelength λ of the emitted light is 0.818. At this time, although the light emission efficiency is improved, the light emission distribution is somewhat scattering compared to the case where the ratio is less than 0.8, and even if an electrode is provided so as to avoid the diffraction azimuth, a part of Was applied to the electrode, and the shadow of the electrode portion was observed to be thin. This is considered to be because the in-plane average height Hav is large, that is, the convex portion volume is large, and it is estimated that the degree of improvement in luminous efficiency is small due to the occurrence of shadows.

  In Example 43, the average pitch is 1200 nm, and the ratio of the in-plane average height Hav to the wavelength λ of the emitted light is 0.947. At this time, as in Example 42, the light emission efficiency is improved, but the light emission distribution is somewhat scattering, and even if an electrode is provided so as to avoid the diffraction azimuth angle, a part of the electrode is applied to the electrode portion. The shadow of was thinly observed. For this reason, it is estimated that the improvement in luminous efficiency is small.

  In Comparative Example 20, the average pitch is 1200 nm and the ratio between the in-plane average height Hav and the wavelength λ of the emitted light exceeds 1. At this time, as with the substrates having the same average pitch Pav, the internal quantum efficiency IQE was improved, but the improvement of the light emission output ratio was small. Further, since the volume of the convex portion is large, the light behaves in a scattering manner, and the electrode cannot be provided so as to avoid the diffraction azimuth angle, and the electrode portion becomes a shadow.

  In addition, this invention is not limited to the said embodiment, It can change and implement variously. In the above-described embodiment, the size, shape, and the like illustrated in the drawings are not limited to this, and can be appropriately changed within a range in which the effects of the present invention are exhibited.

  The present invention can be suitably applied to semiconductor light emitting devices such as OLEDs, phosphors, and light emitting diodes (LEDs).

10 Optical base material (sapphire substrate)
DESCRIPTION OF SYMBOLS 20 Fine uneven structure 30 1st semiconductor layer 40 Light emitting semiconductor layer 50 2nd semiconductor layer 60 Transparent conductive film 70 Anode electrode 80 Cathode electrode 100 Semiconductor light emitting element

Claims (18)

A semiconductor light emitting device having a structure in which at least a first semiconductor layer, a light emitting semiconductor layer, and a second semiconductor layer are laminated on an optical substrate having a fine relief structure formed on a part or the entire surface thereof,
The ratio between the average pitch Pav of the fine concavo-convex structure and the wavelength λ of the emitted light emitted by the light-emitting semiconductor layer in the optical medium layer provided on the surface of the optical substrate on which the fine concavo-convex structure is formed (Pav / λ) is 2.5 or more and 8 or less,
A ratio (Hav / λ) of an in-plane average height Hav of the fine concavo-convex structure to a wavelength λ of the emitted light is 0.1 or more and 1 or less; and
An aspect ratio of the fine concavo-convex structure is 0.2 or more and 1 or less, a semiconductor light emitting device. 2. The semiconductor light emitting element according to claim 1, wherein the wavelength [lambda] of the emitted light is obtained by dividing the vacuum wavelength [lambda] ₀ of the emitted light by the refractive index n of the first semiconductor layer.   The semiconductor light-emitting element according to claim 1, wherein the optical medium layer is the first semiconductor layer.   The semiconductor light emitting element according to any one of claims 1 to 3, wherein a bottom portion of the concave portion constituting the fine concavo-convex structure is a flat surface.   The semiconductor light emitting element according to any one of claims 1 to 4, wherein the convex portion constituting the fine concavo-convex structure has a smaller diameter at the top than at the bottom.   6. The semiconductor light emitting device according to claim 1, wherein the optical base material is a sapphire substrate, a Si substrate, a SiC substrate, a GaN substrate, a GaAs substrate, or a spinel substrate.   7. The semiconductor light-emitting element according to claim 1, wherein the optical base material is a c-plane sapphire substrate, and an offset angle is 0.05 degree or more and 0.8 degree or less. .   The semiconductor light-emitting element according to claim 1, wherein the first semiconductor layer, the light-emitting semiconductor layer, and the second semiconductor layer are III-V group semiconductors.   The semiconductor light-emitting element according to claim 1, wherein the first semiconductor layer, the light-emitting semiconductor layer, and the second semiconductor layer are GaN-based semiconductors.   10. The method for manufacturing a semiconductor light emitting device according to claim 1, wherein the step of preparing the optical substrate and at least a surface of the optical substrate on which the fine concavo-convex structure is formed are provided. And sequentially stacking the first semiconductor layer, the light emitting semiconductor layer, and the second semiconductor layer. An optical substrate having a fine concavo-convex structure formed on a part of or the entire surface thereof, which is used in a semiconductor light emitting device having a structure in which a first semiconductor layer, a light emitting semiconductor layer, and a second semiconductor layer are laminated,
Ratio (Pav / λ) between the average pitch Pav of the fine concavo-convex structure and the wavelength λ of the emitted light emitted by the light-emitting semiconductor layer in the optical medium layer provided on the surface where the fine concavo-convex structure is formed Is 2.5 or more and 8 or less,
A ratio (Hav / λ) of an in-plane average height Hav of the fine concavo-convex structure to a wavelength λ of the emitted light is 0.1 or more and 1 or less; and
An optical base material, wherein the fine concavo-convex structure has an aspect ratio of 0.2 to 1. The optical substrate according to claim 11, wherein the wavelength λ of the emitted light is obtained by dividing the vacuum wavelength λ ₀ of the emitted light by the refractive index n of the first semiconductor layer.   The optical substrate according to claim 11 or 12, wherein the optical medium layer is the first semiconductor layer.   The optical substrate according to any one of claims 11 to 13, wherein a bottom of the concave portion constituting the fine concavo-convex structure is a flat surface.   The optical base material according to any one of claims 11 to 14, wherein the convex portion constituting the fine concavo-convex structure has a smaller diameter at the top than at the bottom.   The optical substrate according to any one of claims 11 to 15, wherein the optical substrate is a sapphire substrate, a Si substrate, a SiC substrate, a GaN substrate, a GaAs substrate, or a spinel substrate.   The optical substrate according to any one of claims 11 to 16, which is a c-plane sapphire substrate and has an offset angle of 0.05 degrees or more and 0.8 degrees or less.   A step of preparing the optical substrate according to any one of claims 11 to 17, and at least the first semiconductor layer, the light emitting semiconductor layer, and the surface of the optical substrate on which the fine concavo-convex structure is formed. And sequentially stacking the second semiconductor layers. A method for manufacturing a semiconductor light-emitting element.

Images