JP6245380B2 - Light emitting device and method for manufacturing light emitting device - Google Patents

Description

  The present invention relates to a light-emitting element and a method for manufacturing the light-emitting element, and in particular, a first semiconductor layer, an active layer, a second semiconductor layer, and a window layer / support substrate are formed on a substrate by epitaxial growth. The present invention relates to a structure and a manufacturing method for performing a rough surface treatment on a formed light emitting element substrate.

  In recent years, the efficiency of light emitting diodes (LEDs) has increased, and the application to lighting fixtures has progressed. Most conventional lighting fixtures are a combination of an InGaN blue LED and a fluorescent agent. However, when a fluorescent agent is used, the occurrence of Stokes loss is unavoidable in principle, and there is a problem that it is impossible to convert all the light received by the fluorescent agent into another wavelength. In particular, this problem is remarkable in a region such as yellow or red having a longer wavelength than blue.

  In order to solve this problem, a technique of combining a yellow or red LED and a blue LED has recently been adopted. At that time, instead of taking out light on one surface as in the case of COB (chip on board) type, a light bulb type luminaire in which LEDs are arranged on a board to emit light in a filament type is becoming widespread. Since the LED element applied to this type of device needs to extract light over the entire surface of the filament, the type that extracts light to one of the elements is not suitable, and an element having a light distribution that extracts light to the entire chip is ideal. It is.

  The InGaN-based LED, which is a blue LED, generally uses a sapphire substrate, and the sapphire substrate is transparent to the emission wavelength, and thus is an ideal form for the above-described lighting fixture. However, in yellow and red LEDs, GaAs or Ge serving as a light-absorbing substrate with respect to the emission wavelength is used as a starting substrate, which is not suitable for the above-described use.

  In order to solve this problem, a window layer is grown to a thickness that can be used for a support substrate as shown in Patent Document 1 or a method of bonding a transparent substrate to a light emitting portion as shown in Patent Document 2, and absorbs light. A technique for removing a starting substrate, which is a substrate, to make an LED is disclosed.

  In the method disclosed in Patent Document 1, it is necessary to join a transparent substrate having a thickness greater than a necessary thickness, and it is necessary to cut the substrate to a predetermined thickness after joining. This causes an increase in cost. In general, the substrate used for bonding has a thickness of 200 μm or more. The film thickness required for the LED element is at most about 100 μm in consideration of the light distribution characteristic and the assemblability with other elements. Therefore, it is necessary to thin the film to such a thickness. In thinning processing, the number of man-hours due to processing and the risk of wafer breakage also increase, leading to increased costs and reduced yield.

  On the other hand, in the method of using a window layer grown by crystal growth to a thickness that can be used for the support substrate disclosed in Patent Document 2 as the support substrate, the window layer may be grown to a desired thickness. Since the step of bonding and bonding the substrates is not necessary, it can be formed at a low cost and is an excellent method.

  In addition, in a light emitting device having a transparent support substrate as described above, it is common to take a method of increasing luminous efficiency by preventing multiple reflections inside the light emitting device and suppressing light absorption. In Patent Document 3, in a structure in which a thick window layer / current diffusion layer and a thick window layer / support substrate sandwich a light emitting portion, the window layer / current diffusion layer and the window layer / support substrate are roughened, and the light emitting portion is rough. A method that does not face the surface has been proposed. However, this method requires the formation of a deep trench that penetrates the window / current diffusion layer portion, which is not only costly, but the height difference between the upper and lower electrodes is too large, so that wire bonding can be performed. difficult. Even when applied to the flip chip type, it is necessary to form a thick insulating film and a very long metal via, which causes an increase in cost. Therefore, it is desirable that the window layer / current diffusion layer portion which is the upper electrode portion is thin.

  Patent Documents 4 and 5 are disclosed as disclosure techniques in which the window layer / current diffusion layer is thin, the height difference between the upper electrode part and the lower electrode part is small, and the light extraction part or the light reflection part has a rough surface. It is done. In Patent Document 4, a rough surface is formed on the surface of the n-type semiconductor layer on the side opposite to the light extraction surface side, but this is a technical disclosure of a flip-chip type, and efficient light from the electrode side to the window layer side. The purpose is reflection. Further, it is disclosed that it is difficult to apply a rough surface to both the window layer / support substrate and the light emitting portion.

  Patent Document 5 discloses a technique in which a light emitting portion surface has a rough surface and the light emitting portion side surface has a mesa shape with a different angle or a simple mesa shape. In this case, the substrate has a reflective structure that does not require a rough surface. Further, a technique is disclosed in which the surface of the light emitting part is formed with irregularities having a period of 2 μm by photolithography.

  On the other hand, when the window layer / support substrate is formed by epitaxial growth, the substrate is greatly warped due to lattice irregularity, and even if the contact exposure method is adopted, the photolithography method has a uniform pitch of 2 μm or less on the light emitting portion surface. It is very difficult to form a pattern. Therefore, when the window layer / supporting substrate is formed by epitaxial growth, the surface of the light emitting portion is roughened by using a rough surface liquid.

Japanese Patent No. 5427585 Japanese Patent No. 4569858 Japanese Patent No. 4715370 JP 2007-059518 A JP 2011-198992 A

  However, when the element separation is performed after roughening the surface of the light emitting portion with a rough surface liquid, the shape of the side surface of the light emitting portion after the element separation reflects the unevenness of the surface of the light emitting portion that has been roughened. Become a shape. In the element isolation, the recesses are thin, so that the etching of the light emitting portion is completed early and reaches the window layer / support substrate portion, whereas the protrusions are thick, so that they reach the window layer / support substrate portion later. Therefore, during etching of the convex portion, the light emitting portion is over-etched with the concave portion, and unevenness is formed on the side surface of the light emitting portion in a projection view in the etching direction. Thus, when element isolation is performed after the surface of the light emitting portion is roughened with a rough surface liquid, the electric field tends to concentrate on the convex portions in the shape of the side surface of the light emitting portion that has been element-isolated, leading to leakage defects and ESD (electrostatic discharge). There was a problem of generating a discharge.

  The present invention has been made in view of the above problems, and has a window layer / supporting substrate and a light emitting part, and after the light emitting part is roughened with a rough surface liquid, the light emitting element is subjected to element separation and has a leakage defect. Another object of the present invention is to provide a light-emitting element in which the occurrence of ESD defects and the production of the light-emitting element are suppressed.

To achieve the above object, according to the present invention, a window layer / support substrate, a second conductivity type second semiconductor layer, an active layer, and a first conductivity layer provided on the window layer / support substrate. In a light emitting device having a light emitting part including a first semiconductor layer of the type in this order,
The light emitting element includes: a removed portion from which the light emitting portion has been removed; a non-removed portion other than the removed portion; a first ohmic electrode provided on the first semiconductor layer of the non-removed portion; A second ohmic electrode provided on the window layer and supporting substrate;
At least a part of the surface of the first semiconductor layer and the side surface of the light emitting portion is covered with an insulating protective film, the surface excluding the outer peripheral portion of the first semiconductor layer and the surface of the window layer and supporting substrate are roughened, And the _Rz of the side surface of the said light emission part is less than 2 micrometers, The light emitting element characterized by the above-mentioned is provided.

  With such a light emitting device, the light emitting efficiency is improved by roughening the surface, and the light emitting device in which the occurrence of leakage failure and ESD failure depending on the uneven shape on the side surface of the light emitting portion is suppressed.

At this time, the window layer / support substrate is made of any one of GaP, GaAsP, AlGaAs, sapphire (Al ₂ O ₃ ), quartz (SiO ₂ ), and SiC, and includes the first semiconductor layer, the active layer, and the second layer. The semiconductor layer is preferably made of AlGaInP or AlGaAs.
Thus, the above materials can be suitably used as the window layer / support substrate, the first semiconductor layer, the active layer, and the second semiconductor layer.

In addition, according to the present invention, a step of forming a light emitting portion by sequentially growing a first semiconductor layer, an active layer, and a second semiconductor layer on a substrate with a lattice-matching material with the substrate, and the light emission Forming a window layer / support substrate by epitaxial growth on the substrate with a non-lattice-matched material, removing the substrate, and providing a first ohmic electrode on the surface of the first semiconductor layer. A step of forming, a first rough surface treatment step of performing a rough surface treatment on the surface of the first semiconductor layer, a removal portion for removing a part of the light emitting portion, and an element separation for forming other non-removal portions A step of forming a second ohmic electrode on the window layer / support substrate from which the light emitting portion has been removed, and covering at least a part of the surface of the first semiconductor layer and the side surface of the light emitting portion with an insulating protective film. Process of the window layer and supporting substrate Made from the second surface roughening step of roughening the surface and side surfaces,
In the first rough surface treatment step, the first ohmic electrode periphery and the region that becomes the outer peripheral portion of the surface of the first semiconductor layer of the non-removed portion in the subsequent element isolation step are not roughened. A method for manufacturing a light emitting device is provided.

  With such a manufacturing method, it is possible to manufacture a light emitting element in which the light emission efficiency is improved by roughening, and the occurrence of leakage failure and ESD failure depending on the uneven shape on the side surface of the light emitting portion is suppressed.

At this time, the substrate is GaAs or Ge , and the window layer / support substrate is any one of GaP , GaAsP , AlGaAs, sapphire (Al ₂ O ₃ ), quartz (SiO ₂ ), SiC , the first semiconductor layer, The active layer and the second semiconductor layer are preferably made of AlGaInP or AlGaAs.
Thus, the above materials can be suitably used for the substrate, the window / support substrate, the first semiconductor layer, the active layer, and the second semiconductor layer.

At this time, the first rough surface treatment step,
A mixed solution of an organic acid and an inorganic acid is used, and the organic acid contains at least one of citric acid, malonic acid, formic acid, acetic acid, and tartaric acid, and the inorganic acid includes hydrochloric acid, sulfuric acid, nitric acid, and hydrofluoric acid. Using a solution containing any one or more of them,
The second rough surface treatment step includes
Contains at least one organic acid such as citric acid, malonic acid, formic acid, acetic acid, and tartaric acid, and contains one or more inorganic acids such as hydrochloric acid, sulfuric acid, nitric acid, and hydrofluoric acid, and contains iodine. It is preferable to use a solution.

  In this way, the surface can be reliably roughened.

At this time, the element isolation step includes:
It can be performed by a wet etching method using a wet etching solution containing hydrochloric acid, and a non-Al-containing layer can be left in a region that is not roughened in the first roughening treatment step, and can be used as an etching mask.
In this way, it is possible to obtain a clear mesa shape on the side surface of the light emitting portion where element isolation has been performed.

At this time, it is preferable that the non-Al-containing layer includes at least one of GaAs, InGaP, InGaAs, and Ge, and a step of removing the non-Al-containing layer after the wet etching is performed.
In this way, a clear mesa shape can be more reliably obtained on the side surface of the light emitting part that has been subjected to element isolation.

The element isolation step includes
It can be performed by a dry etching method using a gas containing hydrogen chloride.
In this way, it is possible to obtain a shape with no constriction on the side surface of the light emitting portion from which element isolation has been performed.

  The light-emitting element and the method for manufacturing the light-emitting element of the present invention provide a light-emitting element in which light emission efficiency is improved by roughening and the occurrence of leakage failure and ESD failure depending on the uneven shape on the side surface of the light-emitting portion is suppressed. Can do.

It is the schematic which showed an example of the light emitting element of this invention. It is process drawing which showed an example of the manufacturing method of the light emitting element of this invention. It is the schematic which shows the epitaxial substrate which grew the light emission part and the window layer and support substrate on the board | substrate in the manufacture process of the manufacturing method of the light emitting element of this invention. It is the schematic which shows the light emitting element substrate which removed the board | substrate from the epitaxial substrate in the manufacture process of the manufacturing method of the light emitting element of this invention. It is the schematic of the light emitting element substrate in which the 1st ohmic electrode in the manufacturing process of the manufacturing method of the light emitting element of this invention was formed. It is the schematic of the light emitting element substrate in which the 1st roughening process in the manufacture process of the manufacturing method of the light emitting element of this invention was performed. It is the schematic of the light emitting element substrate which performed the element separation process in the manufacture process of the manufacturing method of the light emitting element of this invention. It is the schematic of the light emitting element substrate which formed the 2nd ohmic electrode in the manufacture process of the manufacturing method of the light emitting element of this invention, and formed the insulating protective film. 1 is a schematic view of a light emitting device of Example 1. FIG. 6 is a schematic view of a light emitting device of Example 2. FIG. 2 is a photograph of an element isolation end in Example 1. FIG. It is a photograph of an element isolation end in a comparative example. It is the figure which showed the frequency of the reverse voltage (VR) of the light emitting element in an Example and a comparative example. It is the figure which showed the relationship between the ESD voltage and ESD defect rate of the light emitting element in an Example and a comparative example. It is the figure which showed the relationship between _Rz of the side surface of the light emission part in an experiment, and VR defect rate.

Hereinafter, although an embodiment is described about the present invention, the present invention is not limited to this.
As described above, when element separation is performed after the surface of the light-emitting portion is uniformly roughened with a rough surface liquid, there is a problem in that leakage failure and ESD failure occur.

Therefore, the present inventors have intensively studied to solve such problems. As a result, when the surface of the light emitting portion is roughened, the region that becomes the outer peripheral portion of the surface of the first semiconductor layer in the subsequent element separation is not roughened, so that the R _z on the side surface of the light emitting portion is less than 2 μm. Thus, the inventors have conceived that leakage defects and ESD defects can be suppressed. And the best form for implementing these was scrutinized and the present invention was completed.

First, the light-emitting element of the present invention will be described with reference to FIG.
As shown in FIG. 1, the light emitting device 1 of the present invention is provided on a window layer / support substrate 107, a window layer / support substrate 107, a second conductivity type second semiconductor layer 105, an active layer 104, The light emitting unit 108 includes the first conductivity type first semiconductor layer 103 in this order.

The window layer / support substrate 107 is made of GaP, GaAsP, AlGaAs, sapphire (Al ₂ O ₃ ), quartz (SiO ₂ ), SiC, etc., and the first semiconductor layer 103, the active layer 104, and the second semiconductor layer 105 are made of AlGaInP or It can consist of AlGaAs.

  The light emitting element 1 includes a removal unit 170 from which at least the first semiconductor layer 103 and the active layer 104 of the light emitting unit 108 are removed, a non-removal unit 180 other than the removal unit 170, and the first semiconductor layer 103 of the non-removal unit 180. And a second ohmic electrode 122 provided on the window layer / supporting substrate 107 of the removing portion 170.

At least a part of the surface of the first semiconductor layer 103 and the side surface of the light emitting unit 108 is covered with an insulating protective film 150, and the surface of the first semiconductor layer 103 excluding the outer peripheral part (second region 131) and the window layer / supporting substrate 107. The surface is roughened, and the R _{z on the} side surface of the light emitting unit 108 is less than 2 μm.
In addition, _Rz in this application shall show the 10-point average roughness (JIS B0601-1994) of the surface.

With such a light emitting device 1 of the present invention, the light emission efficiency is improved by roughening the surface, and the _{Rz on the} side surface of the light emitting unit 108 is less than 2 μm. Thus, the light emitting element is suppressed.

Next, a method for manufacturing a light emitting device of the present invention will be described with reference to FIGS.
First, as shown in FIG. 3, a substrate 101 is prepared as a starting substrate (SP1 in FIG. 2).

As the substrate 101, GaAs or Ge can be preferably used.
In this way, since the material of the active layer 104 to be described later can be epitaxially grown in a lattice matching system, the quality of the active layer 104 can be easily improved, and the luminance can be increased and the life characteristics can be improved.

Next, a first conductive type first semiconductor layer 103 having a lattice constant different from that of the substrate 101, an active layer 104, and a second conductive type second semiconductor layer 105 are sequentially grown on the substrate 101 by epitaxial growth, so that the light emitting unit 108 is obtained. (SP2 in FIG. 2).
Although not shown, it is preferable that a substrate removal selective etching layer is inserted between the substrate 101 and the first semiconductor layer 103 for the step of removing the substrate 101 described later.

  Next, the window layer / support substrate 107 is formed by epitaxial growth on the light emitting portion 108 with a material of a non-lattice matching system with respect to the substrate 101 to produce the epitaxial substrate 109 (SP3 in FIG. 2).

In SP2 and SP3, specifically, as shown in FIG. 3, on the substrate 101, for example, by MOVPE (metal organic chemical vapor deposition), MBE (molecular beam epitaxy), or CBE (chemical beam epitaxy). An epitaxial substrate in which a buffer layer 106 and a window layer / support substrate 107 are epitaxially grown in this order on a light emitting portion 108 including a first semiconductor layer 103 of the first conductivity type, an active layer 104, and a second semiconductor layer 105 of the second conductivity type. 109 can be produced.
Note that the window layer / support substrate 107 may be formed by HVPE (hydride vapor phase epitaxy).

The active layer 104 has (Al _x Ga _1-x ) _y In _1-y P (0 ≦ x ≦ 1, 0.4 ≦ y ≦ 0.6) or Al _z Ga _1-z As (0 ≦ z ≦ 0.45). When applying to visible light illumination, it is preferable to select AlGaInP, and when applying to infrared illumination, it is preferable to select AlGaAs. However, the design of the active layer 104 is not limited to the above materials because the wavelength can be adjusted by using a superlattice or the like other than the wavelength resulting from the material composition.

  AlGaInP or AlGaAs is selected for the first semiconductor layer 103 and the second semiconductor layer 105, and the selection may not necessarily be made of the same material as that of the active layer 104.

  In the present embodiment, the first semiconductor layer 103, the active layer 104, and the second semiconductor layer 105 having the simplest structure are exemplified by AlGaInP, which is the same material, but the first semiconductor layer 103 or the second semiconductor layer is exemplified. 105 generally includes a plurality of layers in order to improve characteristics, and the first semiconductor layer 103 or the second semiconductor layer 105 is not limited to a single layer.

As the window layer / supporting substrate 107, GaP, GaAsP, AlGaAs, sapphire (Al ₂ O ₃ ), quartz (SiO ₂ ), SiC, or the like can be suitably used. When the window layer / support substrate 107 is formed of GaAsP or GaP, the buffer layer 106 is most preferably formed of InGaP.
The window layer / support substrate 107 can also be formed of AlGaAs, which is a lattice matching material. Further, when GaAsP is selected as the window layer / supporting substrate 107, the weather resistance is good.
However, since there is a large lattice mismatch between GaAsP and an AlGaInP-based material or AlGaAs-based material, high-density strain and threading dislocations enter GaAsP. As a result, the epitaxial substrate 109 has a large warp.

  Here, in order to prevent a wavelength shift due to the formation of the natural superlattice, the light emitting portion 108 is preferably grown with a crystallographic inclination of 12 degrees or more with respect to the growth surface. This tilt direction can be selected in any direction, but when adopting the process of separating the elements in the scribe and braking process, select the direction in which the crystal axis does not tilt and is orthogonal to one of the scribe lines, If the direction in which the crystal axis is inclined is selected for the other of the scribe lines, it is possible to reduce the number of elements whose side surfaces are inclined with respect to the element surface and the back surface. Therefore, a direction in which one of the scribe lines is not inclined is usually selected, but the inclination of the element side surface of about 20 degrees is not a big problem on the assembly. Therefore, the orthogonal directions do not have to coincide exactly, and an angular range of about ± 20 degrees from the orthogonal directions is conceptually included in the orthogonal directions.

Next, the substrate 101 is removed from the epitaxial substrate 109, and the light emitting element substrate 110 is manufactured as shown in FIG. 4 (SP4 in FIG. 2).
Specifically, the light emitting element substrate 110 can be obtained by removing the substrate 101 from the epitaxial substrate 109 by wet etching.

  Next, as shown in FIG. 5, a first ohmic electrode 121 for supplying a potential to the light emitting element is formed on the substrate removal surface 120 of the first semiconductor layer 103 of the light emitting element substrate 110 (SP5 in FIG. 2). ).

  Next, as shown in FIG. 6, a first rough surface treatment step is performed to roughen the surface of the first semiconductor layer 103 (SP6 in FIG. 2). In the first rough surface treatment step, the third region 130 around the first ohmic electrode 121 and the second region 131 on the surface of the first semiconductor layer 103 are not roughened.

In the first rough surface treatment step, a mixed solution of an organic acid and an inorganic acid is used, and the organic acid contains carboxylic acid, in particular, any one or more of citric acid, malonic acid, formic acid, acetic acid, tartaric acid, The inorganic acid can be used using a solution containing one or more of hydrochloric acid, sulfuric acid, nitric acid, and hydrofluoric acid.
In this way, the surface can be reliably roughened.

  The third region 130 has an effect of suppressing over-etching in the first rough surface treatment step and suppressing electrode peeling of the first ohmic electrode 121. On the other hand, the width of the third region 130 needs to be set to an appropriate width that is not too wide so that the antireflection effect due to the roughening effect is not reduced. It is preferable to have a width of about 0.5 to 15 times the depth roughened by the first roughening treatment as a width that has a small decrease in the antireflection effect and is effective in suppressing electrode peeling. . The uneven height of the rough surface is preferably an integral multiple of 1/2 of the emission wavelength.

The width of the second region 131 is preferably about 0.5 to 15 times the depth roughened by the first roughening process.
Here, when the above-described substrate removal selective etching layer is provided on the first semiconductor layer 103, a first rough surface treatment selective etching layer (non-etching layer) is provided between the substrate removal selective etching layer and the first semiconductor layer 103. It is preferable to provide the first rough surface treatment by leaving the first rough surface treatment selective etching layer in the second region 131 after removing the substrate. In this way, over-etching under the photoresist pattern in the first rough surface treatment step can be reduced.

  Since the first rough surface liquid has a selective etching property with respect to the layer made of the non-Al-containing material, the first rough surface treatment selective etching layer is preferably composed of GaAs, InGaP, InGaAs, or Ge. In this way, facets are formed starting from the edge portion of the first rough surface treatment selective etching layer, and over-etching under the pattern can be suppressed. However, when the first rough surface treatment selective etching layer is used, the non-Al-containing material absorbs light with respect to the emission wavelength. Therefore, after the first rough surface treatment, the first surface is treated with a hydrogen peroxide solution containing SPM or the like. It is preferable to add a step of selectively removing the rough surface treatment selective etching layer.

  The width of the second region 131 is limited not only by the depth roughened by the first roughening process but also by the alignment accuracy of photolithography. If an aligner or stepper with high alignment accuracy is used, the additional width in anticipation of alignment accuracy can be reduced. On the other hand, when an aligner with low alignment accuracy is used, it is preferable to increase the additional width in anticipation of accuracy.

  By performing the above steps, a flat second region 131 that is not roughened can be obtained on the surface of the first semiconductor layer 103. As a result, it is possible to suppress the occurrence of unevenness on the side surface of the light emitting unit 108 that has been subjected to element isolation in the element isolation step described later.

  Next, as shown in FIG. 7, an element isolation process is performed to form a removal unit 170 that removes a part of the light emitting unit 108 and a non-removal unit 180 other than that (SP7 in FIG. 2).

In the element isolation step, for example, a pattern in which predetermined regions (second ohmic electrode formation region 140 and scribe region 141 in FIG. 6) are opened with a resist by a photolithography method is formed. This can be performed by etching using a resist as an etching mask.
The etching is performed by a wet etching method using a wet etching solution containing hydrochloric acid to remove the second semiconductor layer 105, the buffer layer 106, or the window layer / support substrate 107, and a non-removed portion other than the removed portion 170. 180 can be formed.

At this time, a non-Al-containing layer (not shown) can remain in the second region 131 that is not roughened in the first rough surface treatment step, and can be used as an etching mask.
The non-Al-containing layer may include one or more of GaAs, InGaP, InGaAs, and Ge. Since the non-Al-containing layer is not etched with an etching solution containing hydrochloric acid, facets are formed starting from the end of the non-Al-containing layer, so that a clear mesa shape is obtained on the side surface of the light-emitting portion 108 where element isolation has been performed. be able to. However, when the non-Al-containing layer is used, it is preferable to perform a step of selectively removing the non-Al-containing layer with a hydrogen peroxide solution-containing liquid such as sulfuric acid / hydrogen peroxide after the element isolation step.

In addition to the wet etching method described above, the element isolation step can also be performed by a dry etching method by a method using a halogen gas, preferably a gas containing hydrogen chloride.
By doing so, it is possible to obtain a shape without constriction (over-etching) on the side surface of the light-emitting portion from which element isolation has been performed.

  Next, as shown in FIG. 8, the second ohmic electrode 122 is formed on the removal portion 170 on the window layer / support substrate 107 from which the light emitting portion 108 has been removed (SP8 in FIG. 2).

Next, as shown in FIG. 8, at least a part of the surface of the first semiconductor layer 103 and the side surface of the light emitting unit 108 is covered with an insulating protective film 150 (SP9 in FIG. 2).
The insulating protective film 150 can be any material as long as it is transparent and has insulating properties. As the insulating protective film 150, for example, SiO ₂ or SiN _x is preferably used. If it is such, the process which opens the upper part of the 1st ohmic electrode 121 and the 2nd ohmic electrode 122 with the photolithographic method and the etching liquid containing a hydrofluoric acid can be performed easily.

  Next, as shown in FIG. 1, a second rough surface treatment step is performed to roughen the surface and side surfaces of the window layer / support substrate 107 (SP10 in FIG. 2).

  Before performing the second rough surface treatment, it is preferable to first scribe lines along the removal portion 170 and perform braking to separate the light emitting elements to form a light emitting element die. After forming the light emitting element die, it is preferable to transfer the light emitting element die to the holding tape so that the window layer / supporting substrate 107 is on the upper surface, and then perform the following second roughening treatment.

  The second rough surface treatment step includes one or more organic acids such as citric acid, malonic acid, formic acid, acetic acid, and tartaric acid, and one or more inorganic acids such as hydrochloric acid, sulfuric acid, nitric acid, and hydrofluoric acid. And a solution containing iodine.

The first rough surface liquid applied to the first semiconductor layer 103 used in the first rough surface treatment step described above and the second rough surface liquid applied to the window layer / support substrate 107 in the second rough surface treatment step have liquid compositions. Different. Therefore, since the etching characteristics are different, the shape of the rough surface and R _a (arithmetic average roughness) of the first semiconductor layer 103 and the window layer / support substrate 107 inevitably differ.

In the method for manufacturing a light emitting device of the present invention described above, overetching occurs in the light emitting portion 108 in the device isolation process by providing the second region 131 that is not roughened on the surface of the first semiconductor layer 103. Therefore, the shape of the side surface of the light emitting unit 108 separated from the element substantially matches the shape of the second region 131. Therefore, _Rz on the side surface of the light-emitting portion 108 separated from the element can be less than 2 μm, and thus it is possible to prevent electric field concentration from occurring in the convex portion in the shape of the side surface of the light-emitting portion 108 separated from the element. As a result, it is possible to manufacture a light emitting device in which the luminous efficiency is improved by roughening and the occurrence of leakage defects and ESD defects depending on the uneven shape on the side surface of the light emitting part is suppressed.

  EXAMPLES Hereinafter, the present invention will be described more specifically with reference to examples and comparative examples of the present invention, but the present invention is not limited to these.

Example 1
An n-type AlInP substrate removal selective etching layer is formed on an n-type GaAs buffer layer (not shown) of 0.5 μm on a 280 μm thick n-type GaAs substrate 101 whose crystal axis is inclined by 15 ° in the [110] direction from the [001] direction. After growing 1 μm (not shown), a light-emitting unit composed of an n-type cladding layer (first semiconductor layer 103), an active layer 104, and a p-type cladding layer (second semiconductor layer 105) made of AlGaInP by the MOVPE method. 108 μm was formed, a buffer layer 106 made of p-type InGaP was formed 0.3 μm, and a p-type GaP layer 1 μm was formed as a part of the GaP window layer / support substrate 107. Next, the substrate was transferred to an HVPE furnace, and a window layer / support substrate 107 made of p-type GaP was grown by 120 μm to obtain an epitaxial substrate 109 (see FIG. 3).

  Next, the GaAs substrate 101, the GaAs buffer layer, and the n-type AlInP substrate removal selective etching layer were removed to produce a light emitting element substrate 110 (see FIG. 4).

  Next, a first ohmic electrode 121 is formed on the substrate removal surface 120 of the first semiconductor layer 103 of the light emitting element substrate 110 (see FIG. 5), and the third region 130 and the second region 131 are formed of a resist by photolithography. A pattern to be coated was formed. As shown in FIG. 9, the third region 130 is provided along the first ohmic electrode, and the second region 131 is provided along the element isolation planned line 160. The width of the third region 130 was 2 μm, which is four times the etching depth. The width of the second region 131 was 6 μm.

  Next, a first rough surface treatment step was performed on the surface of the first semiconductor layer 103 (see FIG. 6). As the first rough surface liquid, a mixed liquid of acetic acid and hydrochloric acid was prepared, and the rough surface treatment was realized by etching at room temperature for 1 minute.

  Next, by a photolithography method, a region other than the second ohmic electrode formation region 140 and the scribe region 141 (see FIG. 6) is covered with a resist, and an element isolation step is performed by a wet etching method using a wet etching solution containing hydrochloric acid. The light emitting portion 108 was removed to form a removal portion 170 where the window layer / support substrate 107 was exposed, and other non-removal portions 180 (see FIG. 7).

As a result of performing the above steps, the R _{z on} the side surface of the light emitting portion from which the element was isolated showed about 0.5 μm corresponding to the photoresist pattern accuracy.

Next, the 2nd ohmic electrode 122 was formed in the removal part 170 (refer FIG. 8). Next, an insulating protective film 150 made of SiO ₂ was laminated, and the surface of the first semiconductor layer 103 and the side surfaces of the light emitting unit 108 were covered with the insulating protective film 150. Then, openings were formed in the insulating protective film 150 at the first ohmic electrode 121 and the second ohmic electrode 122 by photolithography and hydrofluoric acid etching.

  Next, the scribe line was scribed along the exposed removal part 170, the crack line was extended along the scribe line, and then the elements were separated by braking to form a light emitting element die.

  After the light emitting element die was formed, the light emitting element die was transferred to the holding tape so that the surface on which the first ohmic electrode was provided was on the tape surface side, and then the second rough surface treatment step was performed. As a rough surface solution used for roughening the window layer / supporting substrate in the second rough surface treatment step, a mixed solution of acetic acid, hydrofluoric acid and iodine was prepared. And the 2nd roughening process was performed by etching for 1 minute at normal temperature.

  A light emitting device as shown in FIG. 9 was manufactured as described above.

  FIG. 11 shows a photograph of the end of the element isolation portion in Example 1. FIG. 11 shows that the end portion of the second region 131 is formed in a straight line.

  A lamp was manufactured with the light emitting element manufactured as described above, and measurement and evaluation were performed.

FIG. 13 shows the results of reverse voltage (VR) when the reverse direction applied current value of the lamp manufactured by the light emitting element manufactured in Example 1 is 10 μA. FIG. 13 also shows the results of Example 2 and Comparative Example described later.
As a result, as shown in FIG. 13, in the lamps manufactured in Example 1 and Example 2 described later, when the reverse direction applied current value is 10 μA, VR is required to be 30V or more, and the lamp showing VR value of less than 30V is Did not occur. On the other hand, the VR value of about half of the lamps produced in the comparative examples described later was less than 30V. Thus, it was found that the light emitting device of the present invention has excellent reverse voltage characteristics due to the presence of the second region.

Next, FIG. 14 shows the results of an ESD test using a lamp manufactured using the light-emitting element manufactured in Example 1. The ESD execution conditions were HBM (Human Body Model). FIG. 14 also shows the results of Example 2 and Comparative Example described later.
As a result, as shown in FIG. 14, in the lamps manufactured in Example 1 and Example 2 described later, ESD breakdown did not occur in the ESD test up to 2000V. On the other hand, in a comparative example to be described later, an ESD breakdown occurred at an ESD voltage of about 100V, and by 600V, all the elements that were put into the test were destroyed.

(Example 2)
First, it carried out similarly to Example 1 to the 1st roughening process process.

Next, in order to perform the dry etching method, a SiO ₂ film was coated with a thickness of 300 nm so as to cover the first semiconductor layer and the first ohmic electrode, and a resist pattern having an element isolation shape was formed by a photolithography method. Next, the pattern opening was etched with hydrofluoric acid.
Then, a dry etching method was performed using an SiO ₂ film having an opening pattern as an etching mask. In dry etching, element isolation was performed by the RIE method or ICP method in which a chlorine-containing gas was introduced, and the light emitting portion was removed to form a removed portion exposing the window layer / support substrate.

As a result of performing the above steps, the R _{z on} the side surface of the light emitting portion from which element isolation was performed showed about 0.5 μm corresponding to the pattern accuracy of the SiO ₂ mask.

  Then, it carried out similarly to Example 1, and formed from the formation of the 2nd ohmic electrode to the 2nd roughening process process, and manufactured the light emitting element as shown in FIG.

  A lamp was manufactured with the light emitting element manufactured as described above, and measurement and evaluation were performed.

FIG. 13 shows the results of reverse voltage (VR) when the reverse direction applied current value of the lamp manufactured by the light emitting device manufactured in Example 2 is 10 μA.
As a result, as shown in FIG. 13, in the lamp manufactured in Example 2, when the reverse direction applied current value was 10 μA, VR was required to be 30 V or more, and a lamp showing a VR value of less than 30 V was not generated. It has been found that the light emitting device of the present invention has excellent reverse voltage characteristics due to the presence of the second region.

Next, FIG. 14 shows a result of an ESD test using a lamp manufactured using the light-emitting element manufactured in Example 2. The ESD execution conditions were HBM (Human Body Model).
As a result, as shown in FIG. 14, in the lamp manufactured in Example 2, ESD breakdown did not occur in the ESD test up to 2000V.

(Comparative example)
In the first rough surface treatment step, a light emitting device was manufactured in the same manner as in Example 1 except that the second region was not provided.
As a result, in the comparative example, the unevenness (R _z = 0.6 μm on the surface of the first semiconductor layer) generated by the first rough surface treatment is increased by the element isolation process, and the R _{z on} the side surface of the light emitting part that has performed element isolation is increased. Reached 3-4 μm.
FIG. 12 shows a photograph of the end portion of the element isolation portion in the comparative example.

  A lamp was manufactured with the light emitting element manufactured as described above, and measurement and evaluation were performed.

FIG. 13 shows the results of reverse voltage (VR) when the reverse direction applied current value of the lamp manufactured with the light emitting device manufactured in the comparative example is 10 μA.
As a result, as shown in FIG. 13, the VR value of about half of the lamps produced in the comparative example was less than 30V.

Next, FIG. 14 shows the results of an ESD test using a lamp manufactured using the light-emitting element manufactured in the comparative example. The ESD execution conditions were HBM (Human Body Model).
As a result, as shown in FIG. 14, in the lamp manufactured in the comparative example, ESD breakdown occurred at an ESD voltage of about 100V, and all the devices that were put into the test were ESD destroyed by 600V.

(Experiment)
A plurality of lamps were manufactured using the light-emitting element manufactured by the method of Example 1 except that the mask pattern at the time of forming the element separation process was changed and the _Rz on the side surface of the light-emitting portion was changed (Experiment 1).
A plurality of lamps were manufactured using the light emitting device manufactured by the method of Example 2 except that the mask pattern at the time of forming the device isolation process was changed and the _Rz on the side surface of the light emitting portion was changed (Experiment 2).

  FIG. 15 shows the result of measurement of the VR defect rate when the reverse applied current value of the lamps in Experiments 1 and 2 was 10 μA. In addition, the VR value at the time of reverse direction applied current of 10 μA was measured as VR failure when it was less than 30V.

As a result, as shown in FIG. 15, VR failure hardly occurs when _Rz is less than 2 μm in both experiments 1 and 2, but when _Rz is 2 μm or more, VR failure is found in both experiments 1 and 2. You can see that it starts to increase.

Since the shape of the side surface of the light emitting portion after the element isolation substantially follows the shape of the outer periphery of the first semiconductor layer in the element isolation process, the larger the irregularity of the outer periphery of the first semiconductor layer during the element isolation process, the greater the VR. FIG. 15 shows that defects are likely to occur.
Therefore, the _{Rz on the} side surface of the light emitting unit needs to be less than 2 μm. In FIG. 15, only the VR failure is shown, but the same tendency was observed with respect to the IR characteristic indicating the leakage current value when the reverse bias was applied.

Further, the unevenness on the side surface of the light emitting unit was not uniform on the side surface of the light emitting unit, and a pattern was prepared by halving the amount of unevenness on the side surface of the light emitting element, and the measurement was performed. It was the same as the case of the pattern comprised by the unevenness | corrugation amount. Therefore, it can be seen that the relationship between the unevenness on the side surface of the light emitting part and the occurrence of the VR defect rate indicates that the VR defect rate increases when _{Rz on the} side surface of the light emitting part is 2 μm or more. Therefore, in order to suppress leakage failure or ESD failure, it is necessary that _{Rz on the} side surface of the light emitting portion is less than 2 μm.

  The present invention is not limited to the above embodiment. The above-described embodiment is an exemplification, and the present invention has any configuration that has substantially the same configuration as the technical idea described in the claims of the present invention and that exhibits the same effects. Are included in the technical scope.

Claims (8)

A window layer / supporting substrate; a light emitting unit provided on the window layer / supporting substrate, including a second conductive type second semiconductor layer, an active layer, and a first conductive type first semiconductor layer in this order; In a light emitting device having
The light emitting element includes: a removed portion from which the light emitting portion has been removed; a non-removed portion other than the removed portion; a first ohmic electrode provided on the first semiconductor layer of the non-removed portion; A second ohmic electrode provided on the window layer and supporting substrate;
At least a part of the surface of the first semiconductor layer and the side surface of the light emitting portion is covered with an insulating protective film, the surface excluding the outer peripheral portion of the first semiconductor layer and the surface of the window layer and supporting substrate are roughened, In addition, the light-emitting element, wherein R _{z on the} side surface of the light-emitting portion is less than 2 μm. The window layer / support substrate is made of any one of GaP, GaAsP, AlGaAs, sapphire (Al ₂ O ₃ ), quartz (SiO ₂ ), and SiC, and the first semiconductor layer, the active layer, and the second semiconductor layer include 2. The light emitting device according to claim 1, wherein the light emitting device is made of AlGaInP or AlGaAs. Forming a light emitting portion on the substrate by sequentially growing a first semiconductor layer, an active layer, and a second semiconductor layer using a lattice-matching material with the substrate by epitaxial growth; and Forming a window layer and supporting substrate by epitaxial growth using a non-lattice matching material, removing the substrate, forming a first ohmic electrode on the surface of the first semiconductor layer, and the first A first rough surface treatment step for performing a rough surface treatment on the surface of the semiconductor layer; a removal portion for removing a part of the light emitting portion; an element separation step for forming other non-removed portions; and the light emitting portion removed. A step of forming a second ohmic electrode on the window / cumulative substrate thus formed; a step of covering at least part of the surface of the first semiconductor layer and the side surface of the light emitting portion with an insulating protective film; Roughen the surface and side of Made from the second surface roughening step,
In the first rough surface treatment step, the first ohmic electrode periphery and the region that becomes the outer peripheral portion of the surface of the first semiconductor layer of the non-removed portion in the subsequent element isolation step are not roughened. A method for manufacturing a light emitting device. The substrate is GaAs or Ge, the window layer / support substrate is one of GaP, GaAsP, AlGaAs, sapphire (Al ₂ O ₃ ), quartz (SiO ₂ ), SiC, the first semiconductor layer, the active layer The method for manufacturing a light emitting device according to claim 3, wherein the second semiconductor layer is made of AlGaInP or AlGaAs. The first rough surface treatment step includes
A mixed solution of an organic acid and an inorganic acid is used, and the organic acid contains at least one of citric acid, malonic acid, formic acid, acetic acid, and tartaric acid, and the inorganic acid includes hydrochloric acid, sulfuric acid, nitric acid, and hydrofluoric acid. Using a solution containing any one or more of them,
The second rough surface treatment step includes
Contains at least one organic acid such as citric acid, malonic acid, formic acid, acetic acid, and tartaric acid, and contains one or more inorganic acids such as hydrochloric acid, sulfuric acid, nitric acid, and hydrofluoric acid, and contains iodine. The method for producing a light-emitting element according to claim 4, wherein the method is performed using a solution. The element isolation step includes
The wet etching method using a wet etching solution containing hydrochloric acid is used, and a non-Al-containing layer is left in a region that is not roughened in the first roughening treatment step, and is used as an etching mask. The manufacturing method of the light emitting element as described in any one of Claims 5.   The light emitting device according to claim 6, wherein the non-Al-containing layer includes at least one of GaAs, InGaP, InGaAs, and Ge, and the non-Al-containing layer is removed after the wet etching. Manufacturing method. The element isolation step includes
The method for manufacturing a light-emitting element according to claim 3, wherein the method is performed by a dry etching method using a gas containing hydrogen chloride.