JP2010103429A - Method of manufacturing gallium nitride-based light-emitting device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To increase the light emission output of a gallium nitride-based light-emitting device of ≤405 nm in wavelength. <P>SOLUTION: The method of manufacturing the gallium nitride-based light-emitting device includes steps of: forming buffer layers 12 and 14 on a substrate 10; forming an n-contact layer 16 on the buffer layer 14; forming an n-clad layer 18 on the n-contact layer 16; forming an n-block layer 20 on the n-clad layer 18; forming an active layer 22 on the n-block layer 20; forming a p-block layer 24 on the active layer 22; forming a p-clad layer 26 on the p-block layer 24; and forming a p-contact layer 26 on the p-clad layer 26. The step of forming the n-contact layer 16 is composed of a step of laminating an Si layer and a u-Al<SB>x</SB>In<SB>y</SB>Ga<SB>1-x-y</SB>N layer alternately. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a method for manufacturing a gallium nitride based light emitting device, and more particularly to a method for manufacturing a light emitting device having a wavelength of 405 nm or less.

  LEDs emitting light with a wavelength of 405 nm or less are used as light sources for sensors, light sources for chemical reactions, light sources for air purifiers in combination with photocatalysts, DNA analysis in the bio field, inspection in the medical field, and next-generation lighting. There are uses such as white illumination with high color rendering properties in combination with RGB phosphors.

  However, when the emission wavelength is particularly 375 nm or less, the In composition of the light emitting layer (active layer) is easily reduced, so that it is easily captured by the non-light emitting center, and the light emission efficiency is significantly reduced. In order to improve the luminous efficiency, it is necessary to promote the bonding of electrons and holes in the light emitting layer (active layer), which reduces non-luminescent centers.

JP 2004-186509 A JP 2005-31119 A JP 2005-317823 A Japanese Patent No. 3403665

  Conventionally, various attempts have been proposed to increase the light emission output of LEDs that emit light having a wavelength of 405 nm or less, but the actual situation is that they have not been sufficiently improved.

  An object of the present invention is to increase light emission output in a gallium nitride based light emitting device that emits light having a wavelength of 405 nm or less.

The method for manufacturing a gallium nitride based light-emitting device of the present invention includes a step of forming a buffer layer on a substrate, a step of forming an n-type contact layer on the buffer layer, and an n-type cladding layer on the n-type contact layer. Forming an n-type block layer on the n-type cladding layer, forming an active layer on the n-type block layer, and forming a p-type block layer on the active layer A step of forming a p-type cladding layer on the p-type block layer, and a step of forming a p-type contact layer on the p-type cladding layer, and forming the n-type contact layer, , Si layers and undoped Al _x In _y Ga _{1 -xy} N layers (where 0 ≦ x <1, 0 ≦ y <1, 0 ≦ x + y <1) are alternately stacked.

  In one embodiment of the present invention, in the step of forming the n-type contact layer, the layers are alternately stacked at 900 degrees or more.

  According to the present invention, it is possible to obtain a gallium nitride-based light emitting device having excellent light emission output.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows a configuration of an LED having a wavelength of 405 nm or less in the present embodiment. On the transparent substrate 10 such as sapphire, a low temperature buffer layer 12, a high temperature buffer layer 14, and an n-contact layer 16 are sequentially formed. An n-cladding layer 18, an n-blocking layer 20, an active layer 22, a p-blocking layer 24, a p-cladding layer 26, and a p-contacting layer 28 are sequentially formed on a part of the surface of the n-contact layer 16. An n-electrode 30 is formed on another part of the surface of the n-contact layer 16, and a p-electrode 32 is formed on the p-contact layer 28.

The n-contact layer 16 is a layer for forming the n-electrode 30 and has a structure in which Si layers and undoped Al _x In _y Ga _1-xy N layers are alternately stacked in this embodiment. Here, 0 ≦ x <1, 0 ≦ y <1, and 0 ≦ x + y <1.

FIG. 2 shows the configuration of the n-contact layer 16. It is configured by alternately stacking Si layers 16a and Al _x In _y Ga _{1 -xy} N layers 16b. An example of the Al _x In _y Ga _{1 -xy} N layer 16b is an undoped GaN layer, and another example is an undoped AlInGaN layer. At the time of LED manufacturing, Si layers 16a and AlInGaN layers 16b are alternately stacked as shown in FIG. 2, but Si constituting the Si layer 16a is diffused by stacking at a predetermined temperature or higher, specifically 900 degrees or higher. The AlInGaN layer 16b can be doped with Si.

  The manufacturing procedure of this element and details of each layer are as follows.

  In a normal pressure MOCVD apparatus, the sapphire c-plane substrate 10 is heat-treated at 1100 degrees in a hydrogen atmosphere for 10 minutes, and the temperature is lowered to 450 degrees. Then, a discontinuous SiN film is produced by flowing monomethylsilane gas and ammonia gas for 100 seconds. Subsequently, a low temperature buffer layer 12 having a GaN layer thickness of 25 nm is deposited by flowing trimethylgallium and ammonia gas at the same temperature.

  Next, the temperature is raised to 1100 degrees, and trimethylgallium and ammonia gas are flowed again to deposit an undoped GaN layer having a thickness of 3 μm. Subsequently, a thin SiN film is produced by flowing monomethylsilane gas and ammonia gas at the same temperature for 8 seconds. An undoped GaN layer having a thickness of 3 μm is deposited by flowing trimethylgallium and ammonia gas again. Thus, the high-temperature buffer layer 14 of GaN layer (3 μm thickness) / SiN layer / GaN layer (3 μm thickness) is formed.

  Next, only monomethylsilane gas was flowed to deposit Si having a thickness of 1 nm, and trimethylgallium and ammonia gas were flowed thereon to grow undoped GaN, that is, u-GaN (10 nm thickness). A stacked structure in which Si (1 nm thickness) / u-GaN (10 nm thickness) is repeated and grown 150 times is used as an n-contact layer 16.

  Subsequently, an AlInGaN (1.5 nm thickness) / InGaN (1.5 nm thickness) 50 pair n-cladding layer 18 is grown at the same temperature. Trimethylaluminum (TMA) is used as the Al material, and trimethylindium (TMI) is used as the In material.

  Next, an n-block layer 20 of AlGaN (25 nm thick) is grown at the same temperature. Thereafter, the temperature is lowered to about 820 degrees to grow a 20 nm thick AlGaN barrier layer, and a 0.9 nm thick u-AlInGaN well layer, 8 nm thick u-AlInGaN barrier layer, 1.1 nm thick u-AlInGaN well layer, 8 nm thick An MQW (Multiple Quantum Well Structure) active layer 22 comprising a u-AlInGaN barrier layer, a 1.3 nm thick u-AlInGaN well layer, and an 8 nm thick u-AlInGaN barrier layer is grown. Note that a thin SiN film is grown for 4 seconds using monomethylsilane and ammonia gas before each well layer is grown.

  Subsequently, after raising the temperature to 1025 degrees and depositing a p-block layer 24 of Mg-doped p-AlGaN (25 nm thickness), u-AlGaN (0.5 nm thickness) / Mg-doped p-GaN (0.5 nm thickness). 30 pairs of SLS (p-clad layer) 26 are grown, and then a 15 nm thick Mg-doped p-GaN contact layer 28 is grown.

  Next, the wafer is taken out from the MOCVD apparatus, and a transparent electrode is deposited on the growth layer surface. Photoresist was applied to the entire surface, and etching for forming the n-electrode was performed using the photoresist as an etching mask. An n-electrode 30 and a p-electrode 32 are formed on the n-contact layer 16 and the p-contact layer 28 exposed by etching, respectively, and a gold pad for wire bonding is formed on the electrodes. The back surface of the substrate is polished to 100 μm, the chip is cut out by scribe, and mounted to complete the device.

  In the LED having such a configuration, as the n-contact layer 16, a structure in which Si layers 16a and undoped u-GaN layers 16b are alternately stacked is used, and the thickness of the Si layer 16a is changed variously. Output comparison was performed.

  The results are shown in FIGS. 3 shows an AFM observation result in the case of a Si-doped n-GaN layer (1.5 μm thickness) as a comparative example (comparative example), and FIG. 4 shows an Si layer (1 nm thickness) / u-GaN layer (10 nm thickness) 150. FIG. 5 shows an AFM observation result in the case of a pair (Example 1) and FIG. 5 shows an AFM observation result in the case of 150 pairs of Si layer (1.6 nm thickness) / u-GaN layer (10 nm thickness) (Example 2). In FIG. 3, the surface step is not clear, but in FIG. 4, the step is clearly seen and the mountain is smooth, and the surface condition is good. However, in FIG. 5, the surface roughness is conspicuous. From this, the thickness d of the Si layer 16a is preferably 0 <d <1.6 nm.

  FIG. 6 shows the relationship between the thickness of the Si layer and the light emission output. The horizontal axis is the thickness (nm) of the Si layer, and the vertical axis is the light emission output power. Comparative Example 1 corresponds to the case where the thickness of the Si layer 16a is 0, Example 1 corresponds to the case where the thickness of the Si layer 16a is 1 nm, and Example 2 corresponds to the case where the thickness of the Si layer is 1.6 nm. . In each example, the light emission outputs of the edge portion, middle portion, and center portion of the LED are shown. In the case of Example 1, the light emission output is increased as compared with Comparative Example 1, while in Example 2, the light emission output is decreased by about 10% rather than Comparative Example 1. Therefore, also from the viewpoint of light emission output, the thickness d of the Si layer 16a is preferably 0 <d <1.6 nm, and for example, d = 1 nm is preferable.

7, when alternately laminated Si layer 16a and the _{u-Al x In y Ga 1} -xy N layer 16b as n- contact layer 16, showing the relationship of the luminous output and the Al composition ratio x. The horizontal axis represents the flow rate of TMA containing Al. The In composition ratio y fixes the In supply amount to 76.9 μmol / min, and the Al composition ratio x sets the Al supply amount to 0.4, 0.8, 1.6, 2.4, 3.2 (all μmol / Min). As can be seen from the figure, the light emission output of the Si layer / u-AlInGaN layer is greater than that of the Si layer / u-GaN, and the light emission output is maximized when the Al supply amount is 2.4 μmol / min. Compared to the case of the Si layer / u-GaN layer (corresponding to the case where the Al supply amount is 0), it is improved by about 35%. From this result, as the n-contact layer 16, the Si layer / u-AlInGaN layer is more preferable than the Si layer / u-GaN layer from the viewpoint of improving the light emission output.

Thus, n- as a contact layer 16, using an Si layer _{/ u-Al x In y Ga} 1-xy N layer, when the Si layer is formed about 1.0nm thickness, a contact layer of the conventional single-layer (Si-doped Compared to the n-GaN layer), the step state of the surface can be clearly confirmed, and the mountain portion is formed more smoothly than the conventional step. This is because, as compared with the case where n-GaN doped with Si is used as the n-contact layer 16, the crystallinity is improved while securing carriers as the n-contact layer 16 due to the diffusion effect of Si. Conceivable. However, if the Si layer 16a is too thick, carriers can be secured by diffusion of Si, but the surface becomes uneven, resulting in poor crystallinity and a reduction in output.

  In the case of doping, group IV Si and group III-V AlInGaN are flowed simultaneously in doping, and it is not known whether group III-V enters a proper position in group IV movement. On the other hand, in the case of the Si layer / u-AlInGaN layer, since Si and AlInGaN flow alternately, each moves independently and enters the lattice position. It is considered that this difference appears in the difference in the surface state between the two. In the present embodiment, the n-contact layer 16 is grown at a high temperature of 900 degrees or more. If it is 900 degrees or more, Si does not stay and diffuses, and the Si layer / u-AlINGaN layer is considered to be a Si-doped n-AlInGaN layer as a result of Si diffusion. That is, as a method for forming the n-contact layer 16, the Si layer and the u-AlInGaN layer are alternately stacked. As a result, the Si layer and the u-AlInGaN layer are stacked. Rather, it can be said to be a Si-doped n-AlInGaN layer. In this sense, the present embodiment forms a Si-doped n-AlInGaN layer having a smooth surface and excellent crystallinity by alternately laminating Si layers and u-AlInGaN layers at 900 degrees or more. It can be said.

  Note that the light emission output of the Si layer / u-AlInGaN layer is larger than that of the Si layer / u-GaN layer because the Si layer / AlInGaN layer has a wider band gap, so that diffusion of holes is prevented. This is to prevent holes from flowing over the active layer toward the n-type layer.

  In the present embodiment, the active layer 22 has an AlInGaN barrier layer / u-AlInGaN well layer structure. By using AlInGaN as the barrier layer, the In composition fluctuation can be promoted and the non-luminescent center can be reduced.

  FIG. 8 shows the relationship between the In composition ratio in the AlInGaN barrier layer of the active layer 22 and the light emission output. The horizontal axis represents the flow rate of TMI including In. The Al composition ratio was such that the Al supply amount was fixed at 1.0 μmol / min, and the In supply amount was changed. As can be seen from the figure, the light emission output becomes maximum when the In supply amount is 19.2 μmol / min, which is improved by about 20% compared to the case of the AlGaN barrier layer (corresponding to the case where the In supply amount is 0). From this result, it is preferable to use AlInGaN as the barrier layer of the active layer 22 from the viewpoint of improving the light emission output.

  In this embodiment, when forming the n-cladding layer 18, In is used at a high temperature of 1100 ° C., and it is considered that the light emission intensity increases because In is mainly caught by defects. In general, the n-cladding layer 18 is preferably formed at a high temperature of 1050 degrees or more.

  Furthermore, in this embodiment, the well layer of the active layer 22 is formed so that the thickness is smaller as it is closer to the n-type layer. Thereby, compared with the case where all are formed in the same thickness, it becomes easy to flow an electron and can improve the emitted light intensity.

FIG. 9 shows the results when the number of active layers 22 is 3 MQW and the thickness of each well layer is changed. sample,
No. 1 (1.1 nm / 1.1 nm / 1.1 nm)
No. 2 (1.3nm / 1.1nm / 0.9nm)
No. 3 (0.9nm / 1.1nm / 1.3nm)
No. 4 (1.2 nm / 0.9 nm / 1.2 nm)
No. 5 (0.7nm / 1.1nm / 1.5nm)
No. 6 (1.5nm / 1.1nm / 0.7nm)
6 samples. In each sample, the order of the layer thicknesses is the order close to the substrate 10. For example, sample no. No. 1 is a comparative example and all have the same layer thickness. 3 is 0.9 nm, 1.1 nm, and 1.3 nm in order from the substrate 10, and the closer to the substrate 10, the thinner the well layer becomes. In other words, the closer to the p-type layer, the thicker the well layer. As can be seen from the drawing, the sample No. 2 is thinner as the thickness of the well layer of the active layer 22 is closer to the substrate 10. 3 and sample no. In FIG. 5, the light emission output increases. From this, the thickness of the well layer of the active layer 22 is not constant, and the emission intensity can be further improved by changing the thickness so as to become thinner as it approaches the substrate 10 side or the n-side layer. it can.

It is a block diagram of embodiment. It is a block diagram of the n-contact layer of embodiment. It is a figure of the AFM micrograph of Si-doped n-GaN. It is a figure of the AFM micrograph of a Si layer (1.1 nm thickness) / u-GaN layer. It is a figure of the AFM micrograph of a Si layer (1.6 nm thickness) / u-GaN layer. It is a figure which shows the relationship between the thickness of Si layer, and light emission output. It is a figure which shows the relationship between Al supply amount and light emission output. It is a figure which shows the relationship between In supply amount and light emission output. It is a figure which shows the relationship between the sample which changed the well layer of the active layer, and light emission output.

Explanation of symbols

  10 substrate, 12 low-temperature buffer layer, 14 high-temperature buffer layer, 16 n-contact layer, 18 n-cladding layer, 20 n-blocking layer, 22 active layer, 24 p-blocking layer, 26 p-cladding layer, 28 p- Contact layer, 30 n-electrode, 32 p-electrode.

Claims (5)

Forming a buffer layer on the substrate;
Forming an n-type contact layer on the buffer layer;
Forming an n-type cladding layer on the n-type contact layer;
Forming an n-type block layer on the n-type cladding layer;
Forming an active layer on the n-type block layer;
Forming a p-type block layer on the active layer;
Forming a p-type cladding layer on the p-type block layer;
Forming a p-type contact layer on the p-type cladding layer;
Have
The step of forming the n-type contact layer includes alternately stacking Si layers and undoped Al _x In _y Ga _1-xy N layers (where 0 ≦ x <1, 0 ≦ y <1, 0 ≦ x + y <1). A method of manufacturing a gallium nitride based light-emitting device, characterized by comprising the steps of: The method of claim 1, wherein
In the step of forming the n-type contact layer, the gallium nitride light-emitting device manufacturing method is characterized in that the layers are alternately stacked at 900 degrees or more. The method according to any one of claims 1 and 2,
In the step of forming the active layer, an AlInGaN barrier layer and an AlInGaN well layer are alternately stacked. The method of claim 3, wherein
In the step of forming the active layer, the thickness of the AlInGaN well layer is formed so as to become gradually smaller toward the substrate side. The method according to any one of claims 1 and 2,
In the step of forming the n-type cladding layer, a layer containing In is formed.