JP2004140393A - Group iii nitride semiconductor light emitting diode - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a group III nitride semiconductor light emitting diode made of an n side rise type laminate having a light emitting layer made of a gallium nitride-indium and having a high output and excellent monochromaticity. <P>SOLUTION: The group III nitride semiconductor light emitting diode includes a laminate of a first Al<SB>x</SB>Ga<SB>y</SB>In<SB>z</SB>V<SB>a</SB>N<SB>1-a</SB>intermediate layer 102 inserted intermediate between a lattice non-matching substrate and a clad layer containing a p-type impurity, and a second Al<SB>x</SB>Ga<SB>y</SB>In<SB>z</SB>V<SB>a</SB>N<SB>1-a</SB>intermediate layer 107 inserted intermediately between a clad layer containing p-type impurity and a light emitting layer. Particularly, the first intermediate layer is a layer mainly made of a single crystal aggregate, and the second intermediate layer is made of a group III nitride semiconductor in which a disorderliness in an orientation in the second intermediate layer is smaller than that in the first intermediate layer. <P>COPYRIGHT: (C)2004,JPO

Description

The present invention relates to a group III nitride semiconductor light emitting device having a light emitting layer made of a group III nitride semiconductor, and more particularly to a light emitting device including a light emitting layer suitable for providing a light emitting device that emits short-wavelength visible light at high output.

(Target group III nitride semiconductors and examples of their use in the past)
Elements belonging to Group III of the periodic table include aluminum (element symbol: Al), gallium (element symbol: Ga), and indium (element symbol: In). On the other hand, the elements belonging to Group V include arsenic (Element symbol: As) and phosphorus (Element symbol: P) in addition to nitrogen (Element symbol: N) (here, the Group V elements other than nitrogen are represented by the symbol V Are collectively represented by.). Group III nitride semiconductors composed of these elements are used in light-emitting devices such as light-emitting diodes (LEDs) and visible laser diodes (LDs) that emit short-wavelength visible light such as blue, blue-green or green. It is used as a functional layer for manifesting electrical characteristics of the element, a buffer layer for improving the crystallinity of the functional layer, and the like. Group III nitride semiconductors are also used for light receiving elements such as photodiodes (PDs) and optical / electronic (Opto-electronic) devices obtained by combining an optical device and an electronic device. Group III nitride semiconductor crystal layer which is conventionally used in these devices (device) has the general formula _{Al x Ga y In z V a} N 1-a ( where, V is a group V element other than nitrogen, 0 ≦ x , Y, z ≦ 1, x + y + z = 1, 0 ≦ a <1). In particular, the general formula in at with a = 0 to the corresponding nitride aluminum gallium indium mixed crystal _{(Al x Ga y In z N} ( where, 0 ≦ x, y, z ≦ 1, x + y + z = 1) the LED are conventionally employed as, for example, light-emitting layer conventionally be in (see Japanese specification No. Kokoku 6-101587). Further, for example, modulation doped (m Odulation
d OPED) field effect transistor (In the MOFET) is conventionally used as an electron supply layer (Appl.Phys.Lett., 69 (25) (1996), pp. 3,872 to 3,874).

(Practical arrangement example of conventional AlGaInN mixed crystal)
In on the device structure Al _x Ga _y In _z N (where, 0 ≦ x, y, z ≦ 1, x + y + z = 1) and the conventional arrangement of the blue, an example short-wavelength LED that emits blue-green or green Will be explained. First, an Al _x Ga _y N mixed crystal (0 ≦ x, y ≦ 1, x + y = 1) having a composition corresponding to z = 0 can be used in the clad layer as well. There is an example in which the n-type cladding layer is composed of gallium nitride (GaN; x = z = 0 in the above general formula) as seen in the structure of an LED already in practical use. 6-260682). On the other hand, p-type cladding layer is made of aluminum-gallium nitride mixed crystal; they are generally composed of _{(Al x Ga y N 0 ≦} x, y ≦ 1, x + y = 1) ( JP-A-6- No. 260683). Looking at an example of such an arrangement as a clad layer made of a group III nitride semiconductor material, it is practical that the n-type gallium nitride clad layer is arranged on the substrate side below the light emitting layer made of gallium nitride / indium mixed crystal. (Jpn. J. Appl. Phys., 34 (10B) (1995), pages L1332 to L1335). Conversely, it is customary in practice that a cladding layer made of a p-type aluminum nitride-gallium mixed crystal is disposed on a light emitting layer (see Japanese Patent Application Laid-Open No. Hei 6-177423). That is, if the arrangement of the conventional cladding layer is summarized from the relative positional relationship of the light emitting layer, the n-type cladding layer is disposed on the substrate side below the light emitting layer as the lower cladding layer, and the p-type Al _x A practical arrangement method is to dispose the Ga _y N mixed crystal layer as an upper clad layer on the light emitting layer and to form a so-called double hetero structure.

(Problems associated with the conventional arrangement example of the p-type Al _x Ga _y N mixed crystal layer)
As is actually seen on the configuration of the base material of the light-emitting element such as an LED is put to practical use, p-type Al _x Ga _y N mixed crystal layer (0 ≦ x, y ≦ 1 , x + y = 1) is such as described above In addition to the upper cladding layer, it has been used as a p-type contact layer. The contact layer is provided on the p-type upper cladding layer with good electrical contact with the p-type electrode (positive electrode), so to speak, is a low-resistance layer for electrode contact (Japanese Patent Application Laid-Open No. 6-268259). Book). The p-type electrode on the p-type contact layer has a light-transmitting thin-film flat electrode that transmits light emitted from the light-emitting layer in half, and a connection (bonding) that is electrically connected to the thin-film flat electrode to supply an operation current. At present, the positive electrode is gold (element symbol: Au) or an alloy thereof, nickel (element symbol: Ni), chromium (element symbol: Cr). ), Etc. In addition, a wide variety of metal materials such as aluminum (element symbol: Al) and its alloys, silver (element symbol: Ag) are used for p-type electrodes. Disclosed as material (UK
Patent GB 2250635A). However, it remains as a common problem that a p-type electrode having a sufficiently low resistance value and a good ohmic characteristic and having a sufficiently low resistance value can not be formed stably irrespective of the type of the electrode material. In particular, in the case of a light-transmitting thin-film electrode in which it is inevitable that the thickness of the thin-film electrode constituting the electrode is tens of nanometers (nm) or less in order to efficiently transmit the light emitted from the light-emitting layer, it is even more difficult. It is difficult to stably realize an ohmic electrode having good characteristics. p-type Al _x Ga _y N mixed crystal layer (0 ≦ x, y ≦ 1 , x + y = 1) if the deposition conditions of the electrode into or alloy (alloy) condition is defective, Schottky (Schottky) junction manner Some rectifying electrodes may be formed. In other words, p-type having good ohmic characteristics for the type of irrespective p-type Al _x Ga _y N mixed crystal layer of the material constituting the electrodes (0 ≦ x, y ≦ 1 , x + y = 1) In order to stably form the transparent electrode, there is a technical difficulty. Originally, it is not an unreasonable method of producing light transmissivity by relying on a thin layer of a metal material having a small transmittance to visible light, but originally a transparent indium tin (Sn) oxide (ITO) or the like. If used as a transparent electrode instead of a translucent one, the conventional problem can be solved (see Japanese Patent Publication No. 53-11439). At present, ITO having a low specific resistance of about several ohm-cm (Ω · cm) sufficient to enable formation of an ohmic electrode having low contact resistance has been manufactured. However, ITO is a substance exhibiting n-type conductivity. Therefore, it cannot be used as a transparent electrode that contacts the p-type contact layer. This is because a pn junction is formed between the p-type contact layer and the n-type ITO electrode, and in an LED structure, the flow of an operating current or the like to the light emitting layer is hindered.

(Arrangement example of n-type Al _x Ga _y N mixed crystal layer as an alternative to p-type Al _x Ga _y N mixed crystal layer)
Instead of arranging a p-type Al _x Ga _y N mixed crystal layer as a layer for forming a p-type electrode on the outermost layer of the laminated structure, which has a difficulty in stably forming a low-resistance translucent electrode made of a single metal or alloy. , n-type Al _x Ga _y n mixed crystal layer (0 ≦ x, y ≦ 1 , x + y = 1) are also disclosed an example stackup arranged as the surface layer (JP-a 5-63236 JP herein reference ). The configuration in which an n-type layer for forming an electrode is disposed on the outermost layer of the multilayer structure is n-side-up
Side up) structure. In the case of a pn-junction type laminated structure adopted in order to increase the light emission intensity, a p-type layer is first deposited on a p-type substrate, and a light-emitting layer, a clad layer and a contact layer, which are the outermost layers, are formed above the p-type layer. It is usual that an n-type layer forming a layer or the like is overlaid. Al _x Ga _y N if mixed crystal layer specifically described with called n-side up laminated structure of the pn junction type LED applications structure _{with, Ga a Al 1-a N} (0 on a sapphire substrate <a ≦ 1 A) a buffer layer, a structure in which Ga _a Al _1-a N doped with p-type impurities (0 <a ≦ 1) and an n-type Ga _a Al _1-a N (0 <a ≦ 1) are sequentially stacked; (See Japanese Patent Application Laid-Open No. 5-63236). In this laminated structure, a p-type impurity doped Ga _a Al _1-a N (0 <a ≦ 1) layer is used as a light emitting layer, and an n-type Ga _a Al _1-a N (0 <A ≦ 1) is a layer that is overlaid as a cladding layer also serving as a cap (protection) layer for preventing thermal decomposition during the growth process of the light emitting layer (see JP-A-5-63236).

Conventional n-side up structure type in the p-type Al _x Ga _y N mixed crystal layer (0 ≦ x, y ≦ 1 , x + y = 1) a layered relationship relating to more Stated in more detail, in this structure p The mold layer is overlaid on the buffer layer. The buffer layer described above, which also serves as an underlayer for the p-type growth layer, is usually formed at a lower temperature than the growth temperature of the group III nitride semiconductor single crystal layer at about 400 ° C. to about 600 ° C. Because it is filmed, it is commonly referred to as a low temperature buffer layer. The buffer layer is generally made of aluminum nitride (AlN), gallium nitride, indium nitride (InN), a mixed crystal of aluminum nitride / gallium, or aluminum / gallium / indium nitride (Japanese Patent Application Laid-Open No. 2-229476). Etc.).

(Problems of the prior art when p-type AlGaInN mixed crystal layer is overlaid on low-temperature buffer layer)
Light emitting elements such as LEDs having a configuration in which a p-type crystal layer is disposed on a low-temperature buffer layer as seen in the above-described conventional example have not been put to practical use. This doped with p-type impurities on the previous low-temperature buffer p-type Al _x Ga _y In _z N (where, 0 ≦ x, y, z ≦ 1, x + y + z = 1) p -type dopant that occurs when stacking the This is mainly due to denaturation of the low-temperature buffer layer due to (dopant). Conventionally, a crystal layer called a low-temperature buffer layer is mainly composed of amorphous or amorphous, and a layer in which single crystal grains or polycrystal grains are scattered in the layer has been considered to be the best (particularly). See Japanese Unexamined Patent Publication No. 2-229476. It has been determined that a layer made of a single crystal is not suitable for a low-temperature buffer layer (see Japanese Patent No. 229476). As a matter of course, in the case of an amorphous body, the bonding force between the atoms constituting the amorphous body is much weaker than that of the single crystal body. For this reason, the low-temperature buffer layer mainly composed of an amorphous body has a drawback such as easy volatilization in a high-temperature environment. Nevertheless, heretofore, a thin layer exhibiting an amorphous state or a crystal form mainly composed of an amorphous state has been optimized as a low-temperature buffer layer, and has been conventionally and continuously used.

For example, magnesium (element symbol: Mg), beryllium (element symbol: Be), or p-type impurities conventionally used as a p-type impurity for vapor phase (vapor phase) growth of a p-type Al _x Ga _y In _z N mixed crystal are used. It is an element belonging to Group II of the periodic table, such as zinc (element symbol: Zn) (see Japanese Patent Application Laid-Open No. 56-80183). Magnesium and calcium (element symbol: Ca) are used to form p-type Al _x Ga _y In _z N using a so-called ion implantation method in which an ionized group II element is implanted (Japanese Patent Application Laid-Open No. Sho 54). -71589 and Appl. Phys. Lett., 68 (14) (1996), pp. 1945 to 1947). In any means for introducing a p-type impurity, magnesium is preferably used as a p-type impurity. However, magnesium smoothness of the surface, known as a dopant to inhibit the growth of Al _x Ga _y In _z N crystal layer excellent in flatness (Jpn.J.Appl.Phys., 33 (1994) , L1367~ L1369). In particular, when doped with magnesium for the sake of forming the p-type Al _x Ga _y In _z N crystal layer having a low resistance to a high concentration, see the crack (crack) in the Al _x Ga _y In _z N growth layer The more the surface condition becomes worse. Further problem becomes complicated when deposited on low-temperature buffer layer a p-type Al _x Ga _y In _z N growth layer surface state is deteriorated even by doping magnesium as the lower cladding layer. On the other hand, if an impurity such as magnesium is intentionally added (= doped) to the amorphous buffer layer, it is said that it is advantageous to grow a p-type group III nitride semiconductor layer on that layer (Japanese Unexamined Patent Publication No. No. 206519), it is no longer common that the low-temperature buffer layer undergoes alteration due to the presence of magnesium due to the strong chemical reactivity of magnesium itself as judged from the high redox reactivity. It has been experienced as a phenomenon. When magnesium penetrates into the low-temperature buffer layer, the magnesium has a smaller atomic radius than the Group III elements constituting the low-temperature buffer layer, such as aluminum and gallium, so that the low-temperature buffer layer shrinks. This reduction causes cracks in the low temperature buffer layer. The low-temperature buffer layer originally reduces the lattice mismatch between the crystal serving as the substrate such as sapphire and the group III nitride semiconductor crystal layer, and is excellent in the surface state and crystallinity. It is provided with the intention of forming a film. In addition, when the low-temperature buffer layer is mainly composed of amorphous, the low-temperature buffer layer is volatilized during the transition from the weak bonding between the constituent atoms of the amorphous body to the high-temperature growth process. Disappearance is inevitable. When the crystal of the low-temperature buffer layer is exposed to a lattice mismatch with the growth layer due to the loss of the low-temperature buffer layer in parallel with the modification and cracking of the low-temperature buffer layer due to the incorporation of magnesium, the low-temperature buffer layer is no longer used. The effect of alleviating the lattice mismatch is not sufficiently achieved, and only a group III nitride semiconductor growth layer having a poor surface state occurs. There is an urgent need to clarify the constituent requirements of a new low-temperature buffer layer suitable for growing a group III nitride semiconductor containing a p-type impurity.

(Ripple effect on crystallinity of light emitting layer)
Further, consider the case of overlaying the light emitting layer p-type Al _x Ga _y In _z N growth layer over the layer as a lower cladding layer. As a matter of course, a light emitting layer having excellent smoothness cannot be obtained on the lower cladding layer having irregularities on the surface due to cracks generated due to the modification and disappearance of the low-temperature buffer layer. Further, in the region where cracks are generated due to invasion of magnesium or loss at a high temperature and are opened, the lower cladding layer is directly deposited on the exposed surface of the lattice mismatched substrate. For this reason, the lower clad layer in this region is inferior in surface state and has poor crystallinity. The degree of crystallinity affects the crystallinity of the light emitting layer deposited on the lower cladding layer. That is, since a light emitting layer having excellent crystallinity cannot be obtained in this region, it is not possible to uniformly obtain a light emitting layer providing high-output light emission over a wide range. This is the main reason for not yet reached the p-type Al _x Ga _y In _z N crystal layer even until now the LED that is a lower cladding layer is put into practical use.

(Constituent material of light emitting layer)
Group III gallium indium nitride is the light-emitting element composed of a nitride semiconductor material _{(Ga y In z N: 0} ≦ y, z ≦ 1, y + z = 1) is commonly used as a light emitting layer (for example, JP-B 55-3834, etc.). The specification required for the gallium indium nitride as the light emitting layer was omitted mainly from the composition ratio (z) of indium (In). This is based on the conventional insight that the main factor influencing the emission wavelength from gallium indium nitride is the indium composition ratio (see Japanese Patent Publication No. 55-3834 described above). For example, in a light-emitting layer intended to emit blue light having a high visibility near 450 nm (nm), the indium composition ratio (z) is adjusted to a range of about 5% to about 20%. It is customary. The wavelength of light emitted from the light emitting layer having the mixed crystal form can be uniquely determined by the mixed crystal ratio. However, such a certain allowable range of the indium composition ratio in obtaining light emission of the same wavelength depends on the presence or absence of the dopant in the gallium indium nitride crystal layer. In particular, when zinc which is a Group II impurity is present in the gallium indium nitride layer, a deep impurity level (deep) formed by zinc is formed.
level), light emission of a long wavelength can be obtained with a small amount of indium (see Japanese Patent Publication No. 55-3834). On the other hand, in the case of undoped gallium indium nitride, the emission wavelength directly depends on the concentration of indium, so that more indium is used than in the case of gallium indium nitride containing impurities that form deep levels. It must be present in the layer. In obtaining the above-described blue light emission having a wavelength of, for example, 450 nm, an appropriate range of the indium concentration exists because of the doping state of gallium / indium nitride. The thickness of the light emitting layer is generally set to about 100 nm or less. Recently, it has been known that the monochromaticity (color purity of light emission) can be improved by reducing the thickness of the light-emitting layer, and an LED which has a narrow half bandwidth of an emission spectrum and has excellent monochromaticity (color purity) of light emission. In some cases, the thickness of the light emitting layer may be reduced to about 1/10 of the previous thickness. Specifically, a blue band LED having a thin emission layer having a thickness of about 10 nm and a half-width of an emission spectrum of about 15 to 30 nm, which is about 1/2 or less of the conventional one, is proposed (J. Appl. Phys.). , Vol.34 (1995), pages L1332 to L1335). When gallium indium nitride is used as the light-emitting layer, the prior art has already pointed out as a factor that affects the half-width of the emission spectrum that affects the emission wavelength and emission color purity from the emission layer, It was a comprehensive and average indium concentration (mixed crystal ratio) and the thickness of the light emitting layer.

(Crystal morphology of gallium indium indium light emitting layer)
From the viewpoint of the crystal form, the form conventionally required of gallium indium nitride as the light emitting layer has been a single crystal. In addition, it was a crystal layer composed of a so-called single composition phase having a uniform crystallinity and a uniform mixed crystal composition ratio. This requirement for crystal morphology and homogeneity merely corresponds to the reason that it has been vaguely inferred that it is convenient for improving the emission intensity from gallium indium nitride. Recent progress in research on the crystal morphology of gallium-indium nitride teaches that the growth of gallium-indium nitride mixed crystals with a homogeneous mixed crystal composition is rather difficult (1996 (1996) Autumn 57th). Proceedings of the Japan Society of Applied Physics Academic Lectures No. 1, Lecture No. 8p-ZF-14, 209 pages.) This is mainly due to the fact that the gallium indium nitride crystal layer potentially has the property of separating into phases having different mixed crystal composition ratios (Jpn. J. Appl. Phys., 46 ( 8) (1975), 3432.). Furthermore, the results show that a uniform, single-phase gallium indium nitride layer with a uniform indium concentration is not necessarily a prerequisite for producing light emission. Concerning the emission from the gallium indium nitride emission layer, factors that influence the emission wavelength and the color purity of the emission are being elucidated, and measures are being taken to control them. At present, in addition to the fact that the main factors influencing have not been clearly determined, even the mechanism of light emission has not been clarified.

(Optical action of p-type impurity in light emitting layer)
When obtaining a Group III nitride compound semiconductor crystal layer such as an aluminum nitride-gallium mixed crystal having a p-type conductivity type, magnesium has been conventionally used as a preferred element as a p-type impurity as described above. is there. Magnesium is known to have an optical effect on the p-type in addition to serving as an impurity. When omitting the electrical action of magnesium as well as its optical action, magnesium emits a light-emitting center (color center: color
center (see JP-A-54-71589 and J. Appl. Phys., 47 (12) (1976), pp. 5387-5390). The emission emitted from the magnesium emission center changes according to the concentration of magnesium in the group III nitride semiconductor layer. In a range where the concentration of magnesium in the layer is not so large, a pair spectrum with a donor-acceptor impurity (so-called DA)
purple (blue-violet) color development is observed corresponding to the wavelength at which Pair) appears. When the concentration of magnesium in the layer increases and a deep level of magnesium is formed, a blue color is developed. It is said that the concentration of magnesium required to obtain blue light emission exceeds about 10 ¹⁹ cm ^-3 (see "Applied Physics", Vol . 60, No. 2, (1991), pp. 163-166). The wavelength of blue light emission corresponding to the deep impurity level of magnesium is generally in the range of about 410 nm to about 460 nm. This wavelength band is a wavelength band that includes a target emission wavelength of about 450 nm when producing a blue LED because of its high visibility. At first glance, the optical function of magnesium is thought to be that if the concentration of magnesium giving a desired blue band wavelength, for example, 450 nm light emission, coexists in a light emitting layer made of, for example, gallium indium nitride, the emission intensity can be enhanced. Let it.

(Technical problems due to the p-type impurity concentration of the light emitting layer)
However, it is necessary that the wavelength of light emitted from impurities forming a color center, such as magnesium, be exactly the same as the wavelength of the main emission spectrum emitted by the light emitting layer. If the emission wavelengths do not match, the effect of the intensity of the main emission spectrum emitted by the emission layer cannot be achieved at all. Conversely, if the emission wavelengths do not match, a secondary emission spectrum appears in view of the wavelength in addition to the main emission having the desired wavelength, and as a result, the monochromaticity of the emission deteriorates. Therefore, in order to conveniently increase only the emission intensity from the light emitting layer without deteriorating the color purity of the light emission, (a) precise control of the emission wavelength based on magnesium impurities to match the original emission wavelength of the light emitting layer; (B) Measures are required to precisely control the monochromaticity of light emission so that the half width of the emission spectrum based on the magnesium impurity is less than the half width from the light emitting layer itself. The wavelength of the light emission due to the impurity forming the color center and the half width of the emission spectrum change sensitively depending on the concentration of the impurity in the light emitting layer. Therefore, a special and advanced technique for controlling the concentration of the color center impurity such as magnesium taken into the light emitting layer very precisely is required. Against the background of the difficulty in developing such advanced control technology and the demand for a light-emitting element which emits three primary colors of light with the development of display technology and which has excellent color purity, recently, light emission having excellent monochromaticity has been developed. One of the technical trends at present is how to obtain an element. That is, as described above, it has been generally recognized that an undoped layer in which the amount of impurities in the light emitting layer is as small as possible is not intentionally added, and that it is superior in providing light emission with excellent monochromaticity.

n is not limited to the side-up structure type, the light emitting layer has an impurity for forming a color center is deliberately added, of the so-called doped p-type Al _x Ga _y In _z N (where, 0 ≦ x, y, z ≦ 1, x + y + z = 1) in most cases (see, for example, JP-A-7-15041). In a conventional common arrangement example of such a light-emitting layer and a p-type layer doped with a p-type impurity, the p-type impurity present in the p-type layer is diffused into the undoped light-emitting layer. Intrusion is more than likely. In other words, the light emitting layer which is intentionally made an undoped layer for monochromaticity of light emission is contaminated by p-type impurities. In the prior art, a p-type junction with a light emitting layer
No effective measures have been taken to prevent intrusion of p-type impurities from the group III nitride semiconductor layer into the undoped light emitting layer.

In a light-emitting element using an opaque sapphire substrate, an n-side-up structure is advantageous for forming a transparent electrode having good ohmic characteristics on a light extraction surface. In the case of the n-side-up structure, a clad layer doped with a p-type impurity is disposed on the low-temperature buffer layer side, and diffusion of the p-type impurity into the low-temperature buffer layer and the light emitting layer occurs in the subsequent epitaxial growth process. This impairs the function of the buffer layer and prevents the light emission output from increasing. In the laminated structure comprising a base material of the group III nitride semiconductor light-emitting device, In _x Ga _y Al _z N of the substrate crystal lattice-matched to the substrate crystal (0 ≦ x, y, z ≦ 1) thin film A laminated structure provided is also disclosed (see Japanese Patent Publication No. 6-101587). However, in a very general laminated structure that is put into practical use, the constituent layers of the laminated structure are not lattice-matched with the substrate crystal. A simple example is to use sapphire (Al ₂ O ₃ single crystal) as the substrate crystal as described above. Therefore, measures have been taken to insert a low-temperature buffer layer in order to alleviate the lattice mismatch between the substrate crystal and the group III nitride semiconductor deposition layer. However, in the conventional low-temperature buffer layer, the requirements to be provided, such as the crystal form and layer thickness, are merely determined for the main purpose of reducing lattice mismatch with the substrate crystal. The crystal form to be provided in the low-temperature buffer layer for solving the problem of depositing a group III nitride semiconductor clad layer containing a p-type impurity described in the present invention is not clear. A first object of the present invention is to clarify an internal crystal structure to be provided in a low-temperature buffer layer suitable for depositing a group III nitride semiconductor clad layer containing a p-type impurity. A second object is to provide a measure for suppressing the diffusion and intrusion of a p-type impurity into a group III nitride semiconductor light-emitting layer that is internally heterogeneous in crystal growth. A good n-side up laminated structure is achieved by preventing the adverse effect of p-type impurities, and an n-type electrode having good transparency and good ohmic characteristics is stably provided.

That is, the present invention general formula Al _x Ga a lattice mismatch substrate Periodic Group V element other than nitrogen atoms on the crystal as _{V y In z V a N 1} -a ( where, x + y + z = 1,0 ≦ x , y, stackup z ≦ 1, and that a _{0 ≦ a <1) Al x} Ga y in z V a N 1-a lower cladding layer underlying the p-type impurity that is represented in the light-emitting layer were sequentially deposited in the group III nitride semiconductor light-emitting device comprising, Al _x endogenously containing p-type impurities Ga _y in _z V _a Al N in _1-a substrate crystal side of the cladding layer _{x Ga y in z V a N} 1-a place the first intermediate layer comprising a layer, the light emitting layer side is arranged a second intermediate layer composed of _{Al x Ga y in z V a} N 1-a, p -type impurities into the second intermediate layer Gallium indium nitride (Ga _x In _y N; x +) composed of a multiphase structure having a different indium content concentration of less than 1 × 10 ¹⁸ cm ^−3. y = 1,0 ≦ x, y ≦ 1) light-emitting layer are _{stacked, Al x Ga y In z V} a N 1-a on the cladding layer group III nitride provided with the with the structures laminated n-type on the light emitting layer An object semiconductor light emitting device is provided. In particular, the first intermediate layer is a layer mainly composed of an aggregate of single crystals, and the second intermediate layer is a group III nitride having a smaller degree of disorder in orientation than the first intermediate layer. An object semiconductor light emitting device is provided.

Since the first intermediate layer prevents diffusion of the p-type impurity from the lower cladding layer to the buffer _{layer, Al x Ga y In z V} a N 1 inherent p-type impurity which is excellent in surface conditions that continuous smooth _-a Has the effect of contributing to the growth layer. The second intermediate layer has a function as a diffusion blocking layer that suppresses diffusion of the p-type impurity from the p-type impurity-containing layer into the light-emitting layer and a base layer that provides a light-emitting layer having good surface morphology. By configuring the layers sandwiching the p-type cladding layer so as to satisfy certain requirements, a good n-side up structure is obtained, and as a result, a transparent n-type ohmic electrode can be reliably obtained. .
It is possible to provide an n-side-up structure suitable for a light-emitting element having a high light-emission output. Further, it is possible to avoid the conventional technical complexity associated with the formation of the translucent electrode, and it is possible to provide a light-emitting element having excellent monochromaticity.

_{Al x Ga y In z V a} N 1-a ( where, x + y + z = 1,0 ≦ x, y, z ≦ 1 first feature inherent p-type impurity which acts as a lower cladding layer of the present invention and, is to place a 0 ≦ a <1) layer intermediate layer made of the first and second _{Al x Ga y in z V a} N 1-a layer on both the surface side of the. A layer containing a p-type impurity is a layer formed by intentionally adding, by doping, magnesium, zinc, beryllium, calcium, or the like belonging to Group II of the periodic table, which is assumed to act as an acceptor in a Group III nitride semiconductor. Or a so-called cladding layer containing these elements as residual impurities. Usually the first intermediate layer is disposed intermediate the _{Al x Ga y In z V a} N 1-a layer underlying the substrate crystal and a p-type impurity. A typical function is an example in which a first intermediate layer is arranged on a substrate crystal surface and used as a buffer layer. The first intermediate layer may be provided directly on the substrate surface via the deposited low-temperature buffer layer. without directly joined to the _{Al x Ga y In z V a} N 1-a cladding layer underlying the p-type impurity is also an example of arranging the first intermediate layer to the bottom of any intervening layer. An example of the arrangement of the first intermediate layer according to the present invention will be described below in accordance with the order of lamination on the substrate crystal.
(1) a sapphire substrate _{/ Al x Ga y In z V} a N 1-a cladding layer where the first intermediate layer / p-type impurities inherent.
(2) a sapphire or substrate / gallium nitride (GaN) formed of silicon carbide low-temperature buffer layer _{/ Al x Ga y In z V} a N 1-a cladding layer where the first intermediate layer / p-type impurities inherent.
(3) _{sapphire, Al x Ga y In z V} a N 1-a cladding layer of silicon carbide or the III -V compound comprising a semiconductor substrate / first intermediate layer / intervening layer / p-type impurities inherent.
Representative materials for forming the first intermediate layer, denoted by the general formula _{Al x Ga y In z V a} N 1-a, gallium nitride (GaN), aluminum nitride (AlN) or aluminum gallium nitride mixed (Al _x Ga ₁ -xN: 0 <x <1) and mixed crystals of these and indium nitride (InN).

The first intermediate layer is one in which also plays the role of the base layer when depositing the _{Al x Ga y In z V a} N 1-a cladding layer underlying the p-type impurity. Thus, the first intermediate layer withstands _{Al x Ga y In z V a} N 1-a magnesium at the time of deposition of the cladding layer, sufficient erosion and crack generation due to p-type impurity of beryllium inherent the p-type impurity Must be something. Also, it must not volatilize under the high temperature environment associated with conventional low temperature buffer layers. Shall relaxing the lattice mismatch between the _{Al x Ga y In z V a} N 1-a cladding layer underlying the substrate crystal and a p-type impurity. In the present invention, the first intermediate layer is mainly composed of a single crystal in an as-grown state, instead of being mainly composed of an amorphous material as in the conventional low-temperature buffer layer. The single crystal body is a general term for a single crystal layer existing in a layered manner in a direction in which the thickness of the first intermediate layer increases and fine grains of a single crystal growing on the single crystal layer. The layer mainly composed of a single crystal refers to a crystal form in which about 50% or more of the constituent elements of the layer, whether layered or granular, are a single crystal. The component other than the single crystal is an amorphous body or the like. The preferred crystal form of the first intermediate layer of the present invention may be a state in which the lower portion is a single crystal layer and the upper portion is scattered with single crystal grains having a lattice arrangement substantially parallel to the substrate crystal. When the first intermediate layer is directly deposited on the substrate crystal surface, by appropriately selecting the growth conditions of the first intermediate layer, the lattice arrangement of the substrate crystal existing over a certain thickness on the substrate surface is achieved. And a first intermediate layer mainly composed of a single crystal having a structure in which single crystal fine grains are crowned on the same layer as a single crystal layer having a lattice arrangement substantially parallel to the above.

The atoms that make up the single crystal are more rigid and tightly bound to each other than amorphous. Therefore, a substance mainly composed of a single crystal does not volatilize easily even when exposed to a high temperature and remains. This, the first intermediate layer is directly deposited on the substrate surface, there the construction of arranging the _{Al x Ga y In z V a} N 1-a cladding layer underlying the p-type impurity to the first intermediate layer Te, p-type _{Al x Ga y in z V a} N 1-a cladding layer and the surface of the substrate crystal in the relation of the lattice mismatch is avoided bugs exposed by the remaining of the first intermediate layer. That is, born advantage that deterioration of the surface state of the grown layer due to lattice mismatch with the _{Al x Ga y In z V a} N 1-a cladding layer underlying the substrate crystal and a p-type impurity is prevented. If an element for hardening a crystal such as silicon or indium is added to the first intermediate layer, the effect of preventing the loss of the first intermediate layer at a high temperature can be further improved. Another advantage of forming the first intermediate layer of single crystal mainly includes a p-type impurity such as zinc or calcium from _{Al x Ga y In z V a} N 1-a cladding layer underlying the p-type impurity It is possible to suppress the occurrence of cracks and the like in the low-temperature buffer layer due to intrusion. This also depends on the toughness of the bonding force between the constituent atoms because it is a single crystal. That is, the problem associated with the conventional buffer layer is solved by changing the first intermediate layer disposed between the light emitting layer and the substrate crystal to a crystal form mainly composed of a single crystal which is not present in the conventional buffer layer. It is possible to do.

The first intermediate layer mainly composed of a single crystal according to the present invention is formed by a normal pressure or reduced pressure metalorganic pyrolysis vapor deposition method (a method abbreviated as so-called MOCVD or MOVPE) or a molecular beam epitaxial growth (MBE) method. Alternatively, it can be grown using a vapor phase growth method such as a halogen or hydride (hydride) vapor phase growth method (a so-called VPE method). For example, in the case of forming the first intermediate layer made of aluminum nitride by the normal pressure MOCVD method, if a growth condition such as a growth temperature, a growth atmosphere, and a supply amount of a nitrogen source is precisely controlled, a single crystal is mainly used. A first intermediate layer can be formed. Incidentally, on a c-plane sapphire substrate crystal having a plane orientation of (0001), the growth temperature is 450 ° C., the growth atmosphere is composed only of hydrogen, and an ammonia (NH ₃ ) gas as a nitrogen source is one of hydrogen gas constituting the atmosphere. The hydrogen gas was supplied at a rate of about / 8, and the flow rate of hydrogen gas for bubbling (foaming) trimethyl aluminum ((CH ₃ ) ₃ Al) as an aluminum source maintained at a constant temperature of 20 ° C. was set to 20 cc / min. Is set. If the supply of the nitrogen source and the aluminum source is continued for 3 minutes under these growth conditions, an aluminum nitride layer having a layer thickness of about 20 nanometers (nm) is obtained. According to observation by a cross-sectional TEM technique using a general transmission electron microscope (TEM), this aluminum nitride layer has a lattice plane substantially parallel to the crystal lattice of a sapphire c-plane in a region of about 5 nm on the surface of the sapphire substrate. It is determined that a single crystal grain having a spherical shape, a spindle shape, a polygonal shape, or the like is present above the single crystal layer. That is, the first intermediate layer satisfying the crystal form according to the present invention can be obtained by precisely controlling the growth conditions as described above. If the growth temperature is increased without changing the flow rates of the atmosphere constituent gas and the source gas, the thickness of the first intermediate layer obtained in the same growth time increases. While the overall layer thickness increases, the region width of the layered single crystal immediately above the substrate surface remains at approximately 5 nm. That is, as the growth temperature increases, the size, height or quantity of the granular single crystal growing on the layered single crystal increases, and the overall layer thickness increases. As the growth of the granular single crystal progresses as the layer thickness increases, the amount of single crystal grains whose lattice is not substantially parallel to the substrate crystal in the single crystal grains increases. If the thickness of the layer increases and there are many single crystal grains on the surface of the layer that do not have lattice planes parallel to each other, the growth direction of the layer deposited on it will not be fixed in a certain direction, and the difference due to anisotropy in the growth rate As a result, the surface smoothness deteriorates. The thickness of the first intermediate layer mainly composed of a single crystal that provides a deposited layer having an excellent surface condition of a smooth growth surface is generally in the range of 5 to 40 nm when made of aluminum nitride.

Whether or not to include an n-type or p-type impurity in the first intermediate layer can be appropriately selected. _{Al x Ga y In z V a} N 1-a cladding layer and the first intermediate layer conduction type by doping if if p-type impurity achieve unification of electrical conduction type of the underlying p-type impurities deposited over May be a p-type. As the p-type impurities, known Group II elements such as magnesium, beryllium, zinc, and calcium can be used. Particularly, magnesium can be preferably used for imparting p-type conductivity because of the shallow acceptor level of the group III nitride semiconductor. A donor impurity such as silicon which exhibits a hardening effect may be doped together with the group II impurity. Other donor impurities include Group IV elements such as tin (element symbol: Sn) and germanium (element symbol: Ge) and Group VI elements such as sulfur (element symbol: S) and selenium (element symbol: Se). Can be used. Even if undoping is performed intentionally without adding an impurity, the essential effect of the first intermediate layer is not reduced.

Sum up words, a first intermediate layer constituting the single crystal body as a principal because excellent in thermal resistance and chemical _{resistance, Al x Ga y In z V} a N inherent p-type impurity to be subsequently deposited _It can contribute to improvement of the surface state of the _1-a cladding layer. _{Al x Ga y In z V a} N 1-a improvement of the surface condition of the cladding layer underlying the p-type impurity is spread to help improve the surface state of the light emitting layer deposited thereover. A second aspect of the present invention, p-type impurity to suppress the penetration of the p-type impurity into the light-emitting layer from the _{Al x Ga y In z V a} N 1-a cladding layer underlying the p-type impurity as a primary objective the _{Al x Ga y in z V a} N 1-a is located intermediate the second intermediate layer made of _{Al x Ga y in z V a} N 1-a of the cladding layer and the light emitting layer inherent. _{Al x Ga y In z V a} N illustrated below an example of the arrangement of the second intermediate layer in the _1-a cladding layer and the light emitting layers underlying the p-type impurity.
_{(A) Al x Ga y In} z V a N 1-a cladding layer underlying the p-type impurity / second intermediate layer / light-emitting layer.
_{(B) Al x Ga y In} z V a N 1-a clad layer / second intermediate layer / light-emitting layer and the bonding layer / light-emitting layer constituting the bonding inherent p-type impurity.
(C) Al inherent p-type impurity _{x Ga y In z V a N} 1-a clad layer / second intermediate layer 1 / second intermediate layer 2 / emission layer, such as.

The second intermediate layer is composed of gallium nitride, aluminum nitride-gallium mixed crystal, aluminum nitride-gallium-indium mixed crystal, or arsenic (element symbol: As) or phosphorus (element symbol: P) which is a Group V element other than nitrogen. ) Containing gallium nitride arsenide (GaAsN) mixed crystal. Layer and the second intermediate layer to be installed in order to suppress the penetration caused by the thermal diffusion of p-type impurities into the _{Al x Ga y In z V a} N 1-a cladding layer from the light-emitting layer underlying the p-type impurity Therefore, it is desirable to use a material having a strong affinity for p-type impurities. The second intermediate layer according to the _{Al x Ga y In z V a} N 1-a cladding layer underlying the magnesium which is a typical p-type impurity includes aluminum, such as aluminum gallium nitride mixed crystal as an element The material is preferred. Aluminum and magnesium have high affinity and are effective in capturing magnesium. A good example is magnesium-doped aluminum-gallium nitride mixed crystal (Al _x Ga ₁ -xN: 0 ≦ x ≦ 1).

The thickness of the second intermediate layer is set so that the intrusion of the p-type impurity into the light emitting layer can be suppressed in order to maintain the total amount of the p-type impurity inside the light emitting layer at less than 10 ¹⁸ cm ⁻³ . When the second intermediate layer is composed of an aluminum nitride-gallium mixed crystal (Al _x Ga ₁ -xN: 0 ≦ x ≦ 1), the aluminum composition ratio (x) of the aluminum nitride-gallium mixed crystal is approximately 0. .25 (25%) or less. A preferred range for x is from about 0.02 (2%) to about 0.20 (20%). In the case where the second intermediate layer is made of aluminum / gallium nitride, if the aluminum composition ratio is extremely large, the stress that causes strain and the like spreads to the gallium / indium nitride light emitting layer disposed thereabove. In some cases, the wavelength of light emitted from the light-emitting layer changes and shifts to longer wavelengths. In addition, as the aluminum composition ratio increases, the surface state of the gallium / indium indium light emitting layer deteriorates. A second aluminum-gallium mixed crystal (Al _0.05 Ga _0.95 N) having an aluminum composition ratio of 0.05 (5%) to a gallium nitride layer containing magnesium of about 10 ¹⁹ cm ^-3 in aluminum atom concentration. As an example, the desired layer thickness is about 1 nm to about 50 nm. A preferred layer thickness range for the second intermediate layer is from about 5 nm to about 40 nm. Further preferred is a layer thickness range of about 10 nm to 30 nm. When the thickness of the second intermediate layer is extremely thin, about 1 nm, it is not enough to suppress the intrusion of the p-type impurity into the light emitting layer. Conversely, when the second intermediate layer is extremely thick, generally exceeding 50 nm, the probability of the appearance of single crystals having different crystal lattice plane orientations as described above for the first intermediate layer will increase. Increase. The second intermediate layer becomes a base layer when the light emitting layer is grown. Therefore, a situation in which single crystal grains having different orientation directions are exposed on the surface of the second intermediate layer having an extremely increased layer thickness, which results in deterioration of the surface state of the light emitting layer. In addition, taking a case where the light emitting layer is arranged so as to be bonded to the second intermediate layer as an example, an appropriate layer thickness of the second intermediate layer is illustrated from the viewpoint of electrical requirements. In particular, if the second intermediate layer is a crystalline layer having a high resistance, the thickness of the layer is about 40 nm or less, which is expected to have a tunnel effect in consideration of the easiness of injecting an operating current into the light emitting layer. Should. If the doped crystal layer has good electrical conductivity, the layer thickness can be set to a thickness that does not exhibit the tunnel effect. Worsened and not preferred.

And due to the insertion of the second intermediate layer in the lower portion of the light-emitting layer, from entering the _{Al x Ga y In z V a} N 1-a cladding of p-type impurity in the layer emitting layer inherent the p-type impurity Can be suppressed. The effect of ^reducing the p-type impurity concentration in the light emitting layer to 10 ¹⁸ cm ⁻³ or less is remarkably realized in monochromaticity of color development. For example, taking magnesium as a representative example of the p-type impurity, if the atomic concentration of magnesium is limited to 10 ¹⁸ cm ⁻³ or less, blue light emission caused by magnesium atoms becomes negligibly small. Further, even if it is assumed that secondary spectra appear vertically in a region other than the wavelength of the main emission spectrum from the light emitting layer, the intensity of those secondary spectra is weak, so that light emission of multiple wavelengths is caused. It is rare to reach.

In the present invention, in order to obtain a light-emitting layer having a good surface state (surface morphology), the orientation of the second intermediate layer is defined in addition to the above-mentioned layer thickness. Orientation of the _{Al x Ga y In z V a} N 1-a layer of crystals constituting the second intermediate layer, the orientation degree of the first intermediate layer on the basis, and is better than that. In other words, the second intermediate layer has a smaller degree of disorder in the orientation of the crystal than the first intermediate layer. This is a specification required from the viewpoint of the internal crystal structure of the light emitting layer deposited on the second intermediate layer. The degree of disorder in the orientation of the second intermediate layer affects the crystallinity of the light emitting layer regardless of the presence or absence of an intervening layer between the light emitting layer and the second intermediate layer. The degree of randomness of the light-emitting layer deposited with the second intermediate layer, which makes the degree of alignment randomness extremely large, also becomes large. A large number of domains are generated in the light emitting layer where the degree of disorder is large. In particular, when the light-emitting layer is formed of a group III nitride semiconductor layer containing indium which easily diffuses thermally, the light-emitting layer tends to be selectively accumulated at the boundary of each domain of indium. That is, uneven distribution of the indium-containing phase inside the light emitting layer is caused. In the first place, indium accumulated in the grain boundaries and the like hardly contributes to the improvement of the light emission characteristics such as the light emission intensity, and causes a problem of reducing the generation density of nuclei required for forming the indium-containing phase which controls light emission. The randomness of the crystal orientation can be quantified by a general X-ray double crystal diffraction method. In the X-ray double crystal diffraction method, the degree of orientation disorder is determined by an X-ray rocking curve.
curve) (FWHM). If the small angle scattering X-ray diffraction method represented by the X-ray four crystal diffraction method is used, the degree of orientation disorder can be measured more precisely. As a typical value, the degree of randomness of the orientation of the first intermediate layer is about 600 seconds (sec.) To about 1000 seconds in FWHM. That of the second intermediate layer is from about 300 seconds to about 500 seconds. Introducing the measurement results by the small angle scattering X-ray diffraction method, the deviation of the diffraction azimuth angle (generally denoted by ω and called mosaic degree), which is an index of the degree of orientation disorder, is the first. Is about ± 1.0 degrees (°), and that of the second intermediate layer is about ± 0.8 degrees. When the degree of randomness of the second intermediate layer is at this level, even in the indium-containing light-emitting layer directly overlaid thereon, the rate at which indium atoms that promote the generation of the indium-containing phase are captured at the grain boundaries is reduced. Can be kept low.

In the formation of the second intermediate layer, no remarkable difference is observed in the degree of orientation disorder depending on the kind of raw material used for growing the layer. III group material as trimethyl gallium _{((CH 3) 3 Ga)} , triethyl gallium _{((C 2 H 5) 3} Ga), trimethyl aluminum _{((CH 3) 3 Al)} , triisobutylaluminum (i- (C ₄ H _{9 3} ) The use of hydrazines such as ammonia and methylhydrazine or amines such as tert-butylamine as group V elements does not produce a significant difference in the degree of disorder in the orientation of the second intermediate layer. . If the second missing an intermediate layer smoothness of the surface, is disposed above the _{Al x Ga y In z V a} N 1-a cladding layer underlying the such discrete p-type impurity cracks there, the It is difficult to stably control the degree of disorder in the orientation of the intermediate layer 2 by using factors related to growth such as the growth rate. Base on _{Al x Ga y In z V a} N 1-a cladding layer of a (the deposited layer) uneven, non-smoothness of the surface due to cracks or the like on the crystal orientation of the deposited layer serving the second intermediate layer This is because the degree of randomness is determined to be poor. On _{Al x Ga y In z V a} N 1-a cladding layer surface inherent the p-type impurity improve the surface condition is fulfilled by the insertion of the first intermediate layer of the present invention, the second intermediate layer In the case of direct superposition, the degree of disorder of the second intermediate layer can be controlled mainly by the growth temperature, the supply ratio of the group V element to the group III element raw material, so-called V / III ratio, and the growth rate. In particular, the growth rate is an important factor affecting the degree of randomness of the second intermediate layer. As the growth rate is reduced, the degree of disorder in the orientation of the second intermediate layer tends to decrease. Therefore, the growth (deposition) rate during the growth of the second intermediate layer is set to be lower than the growth rate of the first intermediate layer. When the second intermediate layer is made of an aluminum gallium nitride mixed crystal (Al _0.10 Ga _0.90 N) having an aluminum composition ratio of about 0.10 (10%), when the growth rate is reduced to half, the above ω A decrease of about ± 0.2 degrees or more is observed.

First and second intermediate layer may be composed of the same _{Al x Ga y In z V a} N 1-a material. An example in which both the first and second intermediate layers are made of aluminum nitride corresponds to this. In addition, it is also allowed to be composed of different materials. A first intermediate layer composed of aluminum nitride, the second intermediate layer aluminum gallium nitride mixed crystal _{(Al c Ga 1-c N} : 0 ≦ c <1) to consist of first and second This is an example in which the material of the intermediate layer is different. _{Al x Ga y In z V a} N 1-a (x + y + z = 1,0 ≦ x, y, z ≦ 1, and 0 ≦ a <1) as well as selection in materials to be notation allowed in The growth of the first and second intermediate layers need not necessarily be performed at the same temperature. Making the growth temperature the same is also an option. What is important is that the first intermediate layer is mainly composed of a single crystal as described above, and that the composition ratio of the single crystal and the single crystal grains generated thereon is appropriately adjusted, That is, adjustment of the layer thickness. The purpose of the second intermediate layer is to reduce the degree of disorder in crystal orientation as compared with the first intermediate layer. It is well known that the growth temperature that satisfies these structural requirements slightly changes depending on the growth method, growth atmosphere, and more particularly the structure of the growth apparatus. Therefore, there is time to appropriately select, select, and determine the growth temperature in view of the growth conditions and the growth equipment environment that can satisfy the above requirements.

発 光 The light emitting layer deposited as an upper layer via the second intermediate layer does not have to be a single composition gallium indium nitride layer. In addition, a layer whose internal crystal structure is not specified with any uncertainty such as containing gallium and indium (see Japanese Patent Application Laid-Open No. 6-196775) is not used as a light emitting layer. The light emitting layer of the present invention employs a gallium indium nitride layer composed of a plurality of phases having different indium concentrations as the light emitting layer (see Japanese Patent Application No. 8-261444). That is, in the present invention, the light emitting layer is constituted by a layer having a plurality of phases and having a non-homogeneous internal crystallographic structure, instead of using the conventional homogeneous layer formed of a single phase as the light emitting layer. From the viewpoint of the crystal structure inside the light emitting layer, the structure is clearly different from the conventional structure.

The heterogeneous layer composed of a plurality of phases means that a plurality of phases having different indium composition ratios are mixed. Domo specifically be said to collectively as if if gallium indium nitride light emitting layer shows an example, the layer GaN (corresponding to the indium composition ratio = 0.), Such as indium Ga _0.10 In _0.90 N and Ga _0.80 In _0.20 N This means that the layer is not uniform with respect to the indium composition ratio in which a plurality of phases having different composition ratios are mixed. From a morphological point of view, in most cases, a phase having a different indium composition ratio is present in a main phase having a certain indium composition ratio as a substantially spherical or island-like microcrystal. . The gallium-indium nitride layer having such a systematically inhomogeneous layer is used as the light-emitting layer mainly due to the following reasons.
(B) In the case of strongly immiscible materials such as gallium indium and indium, etc. (1996 (1996) Autumn 57th Annual Meeting of the Japan Society of Applied Physics, Proceedings No. 1, Lecture No. 8p- ZF-14, page 209), and phase separation is likely to occur. Therefore, the gallium-indium nitride light-emitting layer having such a multi-phase structure can be reasonably determined to be supported by the thermodynamic properties of the gallium-indium nitride mixed crystal (Solid).
State Commun. , 11 (1972), 617. ).
(B) This is because gallium / indium nitride having a non-uniform composition results in emission with excellent intensity (see Japanese Patent Application No. 8-208486).
That is, the structure of the light emitting layer of the present invention on the crystal structure is defined after appropriately considering a phenomenon in which the group III nitride semiconductor forming the light emitting layer changes its form under thermal influence. is there. This provision relieves the technical difficulty of trying to achieve unnatural homogenization. Further, the inhomogeneous gallium-indium nitride light-emitting layer is formed, for example, by using an indium vapor-phase composition ratio, which is one of the conditions set when forming a gallium-indium nitride crystal layer by MOCVD (metal organic chemical vapor deposition). (Generally, it refers to the concentration ratio of the indium raw material to the total amount of the group III element raw material supplied to the growth environment.) By adjusting factors such as the film formation temperature and the amount of heat received in the process after the film formation, it is almost unambiguous. Can be controlled. In summary, the light-emitting layer as defined in the present invention possesses a crystalline structure that can be formed more routinely as compared to the previous unnatural "homogeneous layer consisting of a single phase". .

A light emitting device having a laminated structure including the first and second intermediate layers and the light emitting layer of the present invention is exemplified below, and the laminated structure is schematically shown in FIGS.
(A) a first intermediate layer (102) made of gallium nitride mainly composed of a single crystal and having a layer thickness of about 5 to about 30 nm / a GaN lower cladding having a layer thickness of μm unit doped with magnesium as a p-type impurity; Layer (106) / second intermediate layer (107) composed of an aluminum gallium nitride mixed crystal layer having an undoped layer thickness of about 10 nm to 30 nm / gallium / indium nitride light emitting layer (108) / n-type GaN upper cladding layer A light-emitting element including a DH structure light-emitting portion made of (109) (see FIG. 7).
(A) Low-temperature buffer layer (113) mainly composed of a single crystal of aluminum nitride having a layer thickness of about 20 nm / first intermediate layer (102) of gallium nitride having a layer thickness of about 10 nm / magnesium Gallium nitride lower cladding layer (106) with a layer thickness of about 200 nm doped with Al / second intermediate layer (107) of undoped aluminum / gallium nitride with a layer thickness of about 15 to 25 nm / undoped n-type nitride An n-side-up type light emitting device having a stacked structure of a gallium / indium light emitting layer (108) / an undoped aluminum / gallium nitride intervening layer (114) having a layer thickness of about 10 nm or less / n-type GaN upper cladding layer (109). Element (see FIG. 8).
(C) In the laminated structure of (a), light emission is obtained by using a transition mechanism between quantum levels different from the quantum level of the quantum well in which the light emitting layer has a layer thickness of about 2 nm to about 200 nm. A light emitting element having an undoped gallium indium nitride layer.
(D) In the laminated structure of (A), the total layer thickness of the multilayer structure of the undoped gallium indium nitride layer and the impurity-doped gallium indium nitride layer is about 80 nm to about 100 nm, or about 100 nm. A light-emitting element comprising a light-emitting layer having a thickness exceeding the above range.

The present invention will be described by taking as an example a light-emitting diode comprising a laminated structure having first and second intermediate layers. The respective layers constituting the laminated structure were formed by the following procedure using a general atmospheric pressure (atmospheric pressure) type MOCVD growth apparatus. (0001) (c-plane) -sapphire (α-Al ₂ O ₃ single crystal) having a thickness of about 90 μm and a diameter of about 1 inch (about 25 mm in diameter) was used as the substrate (101). This crystal substrate (101) was horizontally mounted on a support (susceptor) in a high-purity quartz reaction tube for semiconductor industry having a low content of alkali metals for forming a film on the substrate (101). . The support base is a wedge-shaped vertical section formed by processing a high-purity graphite material. The vertical cross section of the reactor is rectangular and has a cross-sectional area of about 10.5 cm ² . The inside of the reactor was evacuated to a vacuum via a vacuum evacuation path equipped with a common oil rotary vacuum pump. After reaching a degree of vacuum of about 10 ⁻³ torr (Torr), the state of the vacuum is maintained for about 10 minutes, and then purified argon gas (Ar) at a flow rate of about 3 liter / minute is passed through the reaction furnace. The pressure in the furnace was returned to almost the atmospheric pressure.

After purging the inside of the reaction furnace with purified high-purity argon gas for about 5 minutes, the supply of the argon gas to the reaction furnace was stopped. Instead, the supply of purified hydrogen gas (H ₂ ) having a dew point of about −90 ° C. into the reactor was started. The flow rate of hydrogen gas was maintained at about 8 liters / minute by an electronic mass controller (so-called mass flow controller (MFC)). Thereafter, a high-frequency power source having a frequency of 400 kilohertz (kHz) was supplied from a general high-frequency power source to a high-frequency heating coil wound in a circular shape and provided on the outer periphery of a reactor having a vertical vertical cross section. Thereby, the temperature of the substrate (101) was increased from room temperature (about 25 ° C.) to 450 ° C. The temperature of the substrate (101) is controlled by a molybdenum (Mo) sheath-type platinum-platinum-rhodium alloy thermocouple (Japanese Industrial Standard JIS-R standard) inserted into a through hole having a diameter of about 5 mm formed in the middle of the support. The temperature was measured with a thermocouple according to The substrate temperature was precisely controlled to within ± 1 ° C. by a commercially available PID type temperature controller that inputs a thermoelectromotive force signal generated from a thermocouple. About 20 minutes after the temperature of the substrate (101) reached 450 ° C., ammonia gas (NH ₃ ) generated by vaporization of liquefied ammonia gas used as a nitrogen source was supplied to the reactor at a rate of 1 liter per minute. Started to supply. At the same time, trimethyl aluminum ((CH ₃ ) ₃ Al) was supplied to the reaction furnace as an aluminum (Al) source. A bubbler (foam) container made of 316 stainless steel containing trimethylaluminum was kept at 20 ° C. in an electronic thermostat utilizing the Peltier effect. Trimethylaluminum in this vessel was bubbled with hydrogen gas at a flow rate of 20 cc / min, and hydrogen bubbling gas accompanied by trimethylaluminum vapor was supplied to the reaction furnace to supply trimethylaluminum. The supply of the hydrogen gas and the ammonia gas accompanying the trimethylaluminum to the reactor was continued for exactly 6 minutes. Thus, a first intermediate layer (102) made of aluminum nitride having a layer thickness of 20 nm was formed. The growth of the first intermediate layer (102) was terminated by stopping the supply of hydrogen gas accompanying the vapor of trimethylaluminum to the reactor. The laminate comprising the substrate and the first intermediate layer was once cooled to room temperature, and some of the slices were subjected to the identification of the internal crystal structure of the first intermediate layer and the measurement of the degree of disorder in the crystal orientation.

In the as-grown state, the internal crystal structure of the first intermediate layer (102) was observed by a general cross-sectional TEM technique. Prior to the cross-sectional TEM observation, the observation sample was thinned by sputtering using argon (Ar) ions. FIG. 1 schematically shows a cross-sectional TEM image taken at an acceleration voltage of 200 kilovolts (KV). The single crystal layer (103) was present in a region having a thickness of about 5 nm from the surface of the substrate (101). It was recognized that the lattice image (104) observed inside the single crystal layer (103) was arranged substantially parallel to the lattice image of the substrate (101). Above the single crystal layer (103), there is a single crystal grain (105) including a lattice image (104) substantially parallel to the lattice image (104) in the substrate (101) and the single crystal layer (103). Was. Based on the above observation results, the single crystal layer occupies about 25% of the total thickness, and the remaining thickness of about 15 nm is a region mainly occupied by single crystal grains. This configuration differs from the configuration in that almost all of the layer observed in some conventional low-temperature buffer layers is a layered single crystal (J. Electron. Mater., 24 ( 4) (1995), pp. 241-247).

Similarly, in the as-grown state, the degree of randomness of the orientation of the aluminum nitride layer constituting the first intermediate layer was measured by using the four crystal X-ray diffraction method. The mosaic degree was approximately ± 1.0 degree (°) as a scattering azimuth angle (generally represented by the symbol ω).

The remaining part of the above laminated body was placed on a susceptor and inserted again into the reactor. Thereafter, the supply of argon gas was started at a flow rate of 4 liters / minute to the reactor. The amount of power applied to the high-frequency coil was increased, and the temperature of the substrate (101) was increased from room temperature to 1050 ° C. at an average rate of about 100 ° C./min. On the way, when the temperature of the substrate (101) passed 450 ° C., the supply of ammonia gas to the reaction furnace was started at a flow rate of 1 liter / min. As soon as the substrate temperature measured by the thermocouple reached 1050 ° C., the supply amount of ammonia to the reactor was increased from 1 liter per minute to 3.5 liters per minute using an electronic mass flow meter. At the same time, the supply rate of argon gas was reduced from 4 liters / minute to 1.3 liters / minute, and the supply of hydrogen gas to the reactor was restarted at a flow rate of 1.3 liters / minute. As a result, a total of 6.1 liter / min of a mixed gas composed of hydrogen, argon, and ammonia flows into the reaction furnace including the high-purity quartz tube. After the temperature of the substrate (101) reached 1050 ° C. and waited for 5 minutes, an upper clad layer (106) made of gallium nitride doped with magnesium was overlaid on the first intermediate layer (102). Trimethylgallium was used as the gallium source. The 316 stainless steel bubbler container containing trimethylgallium was maintained at exactly 0 ° C. in an electronic thermostat. The flow rate of hydrogen gas for bubbling and entrainment of trimethylgallium vapor was 30 cc / min. The doping source magnesium bis - using cyclopentadienyl magnesium _{(bis- (CH 3 C 5 H} 4) 2 Mg). The stainless steel cylinder container containing the magnesium doping source was kept at a constant temperature of 45 ° C. by an electronic constant temperature bath. In the bis-methylcyclopentadienyl magnesium liquefied while maintaining the same temperature, a hydrogen gas controlled as a bubbling gas, the flow rate of which was precisely controlled to 20 cc / min by an electronic mass flowmeter, was passed. The supply of hydrogen gas accompanying the vapors of the gallium source and the magnesium source to the reactor was continued for 60 minutes to form a gallium nitride layer (106) containing magnesium having a layer thickness of 3.2 μm. The growth of the magnesium-doped gallium nitride layer (106) was terminated by stopping the supply of the gallium source and the magnesium source to the reactor.

After the growth of the gallium nitride layer (106) containing the p-type impurity is completed, the temperature of the substrate (101) is maintained at 1050 ° C., and the mixture composition and the flow rate of the carrier gas and the flow rate of the ammonia gas remain unchanged. And waited for 3 minutes. After the standby, the above-mentioned aluminum source and gallium source were supplied to the reactor, and the growth of the second intermediate layer (107) was started. At this time, trimethylgallium was bubbled, and the flow rate of hydrogen gas accompanying the vapor of trimethylgallium was 5 cc / min. Trimethylaluminum was bubbled, and the flow rate of hydrogen gas accompanying the vapor of trimethylaluminum was set at 10 cc / min. The supply of hydrogen gas accompanying the vapors of trimethylgallium and trimethylaluminum to the reactor was continued for exactly 3 minutes to form an undoped second intermediate layer (107) having a layer thickness of 25 nm. The growth of the second intermediate layer (107) was terminated by stopping the supply of the hydrogen bubbling gas accompanying the vapor of the group III raw material to the reactor.

After the growth of the second intermediate layer (107) was completed, the layer was cooled to room temperature in order to measure the degree of disorder of the crystal orientation of the same layer (107) and the aluminum composition ratio. The degree of disorder and the aluminum composition ratio were measured by using the above-mentioned four-crystal X-ray diffraction method as in the case of the first intermediate layer. The mosaic degree was within a range of ± 0.5 degrees as ω. The aluminum composition ratio was determined to be about 0.06 (6%) from the degree of separation based on the position where the diffraction X-ray peak from the gallium nitride layer (106) appears. Based on this result, the second intermediate layer (107) has a smaller degree of disorder in orientation than the first intermediate layer (106), and an aluminum nitride-gallium mixed crystal with an aluminum composition ratio of 0.06 ( Al _0.06 Ga _0.94 N).

After completion of the measurement, the laminate having the second intermediate layer (107) as the outermost layer is inserted into the reaction furnace, and the temperature is increased from room temperature to 890 ° C. on average in an argon atmosphere at a rate of 100 ° C./min. did. It waited for 3 minutes until the temperature of the substrate (101) was stabilized at 890 ° C. Thereafter, supply of a gallium source and an indium source into a growth atmosphere composed of an argon gas having a flow rate of 3.0 liters per minute and an ammonia gas having a flow rate of 3.0 liters per minute was started, and a second gas supply was started. The growth of the undoped gallium indium nitride (Ga _0.94 In _0.06 N) light emitting layer (108) on the intermediate layer (107) was started. The above-mentioned trimethylgallium was used as a gallium (Ga) source, and the temperature of the bubbler container of the gallium source was set to 0 ° C. The flow rate of hydrogen gas for bubbling to accompany the vapor of trimethylgallium was controlled at 5 cc / min by an electronic mass flow meter. Trimethylindium ((CH ₃ ) ₃ In) was used as a source of indium. Trimethylindium was stored in a stainless steel cylinder container having an internal volume of about 100 cc, and the cylinder container was maintained at exactly 35 ° C. using an electronic thermostat. Hydrogen gas at a flow rate of 13.3 cc / min was passed through the cylinder vessel containing the indium source to entrain the sublimated trimethylindium vapor into the reactor. The ratio of the supply amount of the indium source to the total amount of the gallium source and the indium source supplied into the reactor, that is, the so-called gas phase composition ratio of indium was 0.10. The vapor phase composition ratio of indium was determined by setting the vapor pressure of trimethylgallium at 0 ° C. to about 64.4 torr (Torr) and the vapor pressure of trimethylindium at 35 ° C. to 3.0 torr. While maintaining the flow rate of ammonia gas as a nitrogen (N) source at 3.5 liters per minute and maintaining the flow rate of argon gas at 2.6 liters / minute, the raw materials of the group III and V elements are accurately taken for 15 minutes. Then, an undoped gallium-indium nitride light-emitting layer (108) having a layer thickness of 6 nm by continuously supplying hydrogen and an argon carrier gas was formed.

The supply of the gallium source and the indium source to the reaction furnace was interrupted to terminate the formation of the light emitting layer (108). While adjusting the amount of power from the high-frequency power source applied to the high-frequency coil based on the thermo-electromotive force signal from the thermocouple inserted in the middle of the support while maintaining the flow rates of both the argon and ammonia gases at the above values, the substrate is adjusted. The temperature of (101) was increased from 890 ° C. to 1050 ° C. The temperature was raised from 890 ° C. to 1050 ° C. for 3 minutes in order to suppress volatilization of the indium-containing light emitting layer (108) due to a carelessly gradual rise in temperature during the temperature rise process. Immediately after the temperature of the substrate (101) reached 1050 ° C., the flow rate of argon was reduced to 1.3 liters per minute, and at the same time, hydrogen gas was supplied at a flow rate of 1.3 liters per minute. Thus, a growth atmosphere composed of a hydrogen-argon-ammonia mixed gas having a total flow rate of 6.1 liter / min was created. After instantaneously adjusting the flow rates of hydrogen and argon, a silicon (Si) -doped n-type gallium nitride layer (109) was grown on the light emitting layer (108) without interruption. During the growth of the n-type gallium nitride layer (109), the liquefied trimethylgallium kept at 0 ° C. is bubbled with hydrogen gas at a flow rate of 30 cc / min, and a hydrogen bubbling gas accompanied by trimethylgallium is supplied into the reactor. did. As silicon, disilane (Si ₂ H ₆ ) diluted to about 1 ppm by volume with high-purity hydrogen was added as a doping source. The flow rate of the disilane doping gas was set to 20 cc / min by an electronic mass flow meter. Supply of the hydrogen bubbling gas accompanying the vapor of the gallium source and the silicon doping source gas to the reaction furnace was continued for exactly 6 minutes to obtain a Si-doped n-type gallium nitride layer (109) having a layer thickness of 300 nm. .

(5) The flow of the gallium source into the reactor was interrupted to stop the formation of the n-type gallium nitride layer (109). At the same time, the supply of the silicon doping gas and the hydrogen gas into the reactor was stopped. The flow rate of ammonia gas was maintained at 3.5 liters per minute. Thus, the gas species flowing in the reaction furnace were argon gas and ammonia. While the atmosphere in the reactor was composed of argon and ammonia gas, the amount of high frequency power applied to the high frequency coil was reduced to lower the temperature of the substrate (101) from 1050 ° C. to 950 ° C. in about 2 minutes. When the temperature of the substrate (101) reached about 950 ° C., the flow rate of argon flowing through the reactor was increased from 1.3 liters per minute to 2.6 liters per minute. In a mixed atmosphere composed of argon and ammonia, the temperature was lowered from 950 ° C. to 650 ° C. at a rate of 15 ° C./minute for 20 minutes. At the time when the temperature was lowered to 650 ° C., the supply of the ammonia gas into the reactor was cut off, and only argon gas was passed through the reactor. In this state, it was cooled to room temperature.

After cooling, the carrier concentration on the surface of the n-type gallium nitride layer (109) measured by a general electrolytic CV method was about 1 × 10 ¹⁸ cm ⁻³ . Concentration distributions of the constituent elements and the dopant in the depth direction were measured by secondary ion mass spectrometry (SIMS) using a part of the laminated structure. In particular, the measurement was performed by paying attention to the region near the interface between the second intermediate layer and the light emitting layer and the distribution and concentration of magnesium inside the light emitting layer. As shown in FIG. 2, magnesium in the magnesium-doped gallium nitride layer (106) was distributed almost uniformly in the thickness direction, and its concentration was about 2 × 10 ¹⁹ cm ⁻³ . In the second intermediate layer (107), it decreases exponentially from the junction interface between the magnesium-doped gallium nitride layer (106) and the second intermediate layer (107) toward the junction interface with the light emitting layer (108). Was recognized. At the joint interface between the second intermediate layer (107) and the light emitting layer (108), the magnesium concentration was reduced to about 6 × 10 ¹⁷ cm ⁻³ . Inside the light-emitting layer (108), the concentration of magnesium decreases toward the silicon-doped n-type gallium nitride layer (109), and is generally about 4 × 10 ¹⁷ cm ⁻³ .

The cross section of the laminated body obtained by the above growth operation was subjected to a sputtering treatment with argon ions to have a layer thickness suitable for cross-sectional TEM observation at an acceleration voltage of 200 KV. Observation of the internal structure of the first intermediate layer (102) reveals that, in the as-grown state, a region only in the vicinity of the interface with the sapphire substrate (101) is recognized as a layered single crystal. After the completion of the above, the entire surface of the first intermediate layer (102) was converted into a layered single crystal. A region where the continuity of the layer is impaired due to a crack or the like is not recognized inside the first intermediate layer (102), and the junction interface with the magnesium-doped gallium nitride layer (106) is almost flat. Met. Many dislocations, which are imaged as a linear black contrast on a bright-field cross-sectional TEM image that appears to originate at the junction interface between the first intermediate layer (102) and the magnesium-doped gallium nitride layer (106), are second. At the bonding interface between the intermediate layer (107) and the magnesium-doped gallium nitride layer (106). The second intermediate layer (107) was composed of a layered single crystal over almost the entire area in the layer thickness direction. From a cross-sectional TEM image of a portion of the gallium indium nitride (Ga _0.94 In _0.06 N) light emitting layer (108), it was confirmed that the layer (108) had a heterogeneous structure. Describing morphologically, a substantially spherical microcrystal was present in the layered body. The diameter of the substantially spherical crystal was at most about 3 nm based on the captured circular contrast. In addition, the crystals did not segregate in a specific region inside the light emitting layer and were distributed almost uniformly. The number of microcrystals imaged in the cross-sectional TEM image was approximately two in an imaging area of 50 nm in width and about 6 nm in height (layer thickness).

The cross section of the thinned sample was observed with a general analytical electron microscope to analyze the composition inside the light emitting layer (108). EPMA was performed to analyze the concentration of indium (e lectron- p robe
m icro- a nalysis) the fine crystals from the analysis of other regions so-called, the more indium compared to the interior of the mother phase is contained was observed. Although a signal (signal) presumed to be caused by the characteristic X-ray of indium was detected from the mother phase, the S / N ratio was low and the signal was not clearly recognized as a characteristic X-ray signal of indium. That is, it was determined that the matrix phase was very close to the composition of gallium nitride because of the low indium concentration. Judging from the intensity of the detected kα characteristic X-ray of indium, a result was obtained which teaches that a difference in the indium concentration exists between the microcrystals in units of%. Accurate quantification did not reach the detection performance of the EPMA analyzer. However, the concentration of indium contained in the microcrystal was at most several percent, and the microcrystal containing indium exceeding 10% was rare. At the boundary between the microcrystal and the surrounding parent phase, there was a region in which the arrangement of the crystal lattice was disturbed, which is considered to be caused by distortion or the like.

A light emitting diode was manufactured using the above-described stacked structure of the n-side-up structure type as a base material. A second intermediate layer (107), a light emitting layer (108) and an n-type gallium nitride layer (109) on the magnesium-doped gallium nitride layer (106) in a region where a p-type electrode (positive electrode) (110) is to be formed. ) Was removed by etching using a microwave plasma etching technique using an argon-methane (CH ₄ ) -hydrogen mixed gas. This etching was continued until the surface of the magnesium-doped gallium nitride layer (106) was removed by about 150 nm. Thereafter, on the surface of the magnesium-doped gallium nitride layer (106) exposed by etching, a pad electrode (110) having no light-transmitting or light-transmitting electrode was formed. The pad electrode (110) was a multilayer electrode of a gold-beryllium (Au-Be) alloy and gold (Au). On the other hand, on the Si-doped n-type gallium nitride layer (109), which is the outermost layer, a translucent electrode (111) made of an indium tin oxide (ITO) film exhibiting n-type conduction and having a thickness of about 200 nm. Was adhered. The translucent electrode (111) was formed on almost the entire region of the n-type layer (109) left in the mesa shape by the above-described etching. Although there is a method of forming a pad electrode directly on the surface of the translucent electrode (111), in this embodiment, an n-type (negative electrode) made of aluminum by removing a part of the translucent electrode (111) Pad electrode (112) was formed. That is, the bottom of the aluminum pad electrode (112) was grounded to the surface of the n-type gallium nitride layer (109). The peel strength between the aluminum pad electrode (112) and the ITO translucent electrode (111) was almost equal. Incidentally, the transmittance at a wavelength of 450 nm of an ITO translucent electrode having a thickness of about 200 nm is about 43% of that of a conventional gold-nickel (Au.Ni) translucent metal thin film having a thickness of about 25 nm. It showed twice as high permeability. FIG. 3 is a schematic plan view of the LED. FIG. 4 schematically shows the cross-sectional structure.

直流 DC voltage was applied between adjacent electrodes ((110) and (112)). Blue light emission was already obtained by applying a DC voltage of less than 1 volt (V), for example, 0.4 V. The intensity of blue light emission increased with an increase in the applied voltage value. In the measurement using a general integrating sphere, a DC voltage of 6.0 V was applied, and the emission output when a forward current of 38 milliamperes (mA) was passed was 2.0 milliwatts (mW). On the other hand, in the measurement of the emission spectrum using a spectroscope attached to an ordinary photoluminescence photometer, it was found that the emission wavelength of the main emission spectrum was 452 nm. In addition to the main emission spectrum, the appearance of any spectrum other than the spectrum considered to be caused by gallium nitride band-edge emission near the wavelength of 365 nm was not observed. In other words, the emission intensity of the secondary spectrum was very weak, less than 1/100 of that of the main spectrum. Further, the half-width of the main emission spectrum was narrowed to 130 millielectron volts (meV) at room temperature. For this reason, the LED was excellent in both intensity and monochromaticity.

整流 Rectification was observed in the current-voltage characteristics (IV characteristics) between both electrodes. The reverse breakdown voltage (voltage) when the current in the reverse direction was 10 microamps (μA) exceeded 10 V. This suggests that the magnesium-doped gallium nitride layer (106) has a lower resistance during the growth period, resulting in a pn junction sufficient to provide rectification.

(Comparative Example) A laminated structure having the same configuration as that of the above embodiment was formed except that only the second intermediate layer was deleted. In other words, the second intermediate layer is omitted according to the conditions described in the above embodiment, and the laminated structure includes a direct junction between the magnesium-doped gallium nitride layer (106) and the gallium-indium nitride light-emitting layer (108). Got a body. FIG. 5 schematically shows a cross-sectional structure of the laminated structure. The reason for removing the second intermediate layer is to show the difference in the surface state of the laminated structure and the degree of penetration of p-type impurities (magnesium) into the light emitting layer depending on the presence or absence of the second intermediate layer. According to observation by a general differential interference type optical microscope, many hemispherical projections were present on the surface of the laminated structure. In order to find the origin of the projections, step etching by plasma etching was repeated from the outermost n-type gallium nitride layer (109) of the laminated structure, and the surface was sequentially observed. As a result, it was visually recognized that hemispherical projections had already occurred in the region of the light emitting layer, and the surface of the light emitting layer was not smooth but had wavy “undulations”. However, in the region near the surface of the magnesium-doped gallium nitride layer (106) immediately below the light-emitting layer (108), almost no hemispherical projection was recognized, and the surface was smooth and flat without cracks. From this, it was determined that many protrusions were generated from the light emitting layer (108), and therefore, the surface state itself of the light emitting layer was also impaired. FIG. 6 shows the distribution of magnesium atoms in the depth direction by SIMS. Unlike the case of the above example in which the second intermediate layer was inserted, remarkable intrusion of magnesium into the light emitting layer (108) was observed. The concentration of magnesium in the light emitting layer (108) was determined to be about 6 × 10 ¹⁸ cm ⁻³ .

An LED having the same electrode configuration as that of the example was manufactured. The luminous intensity itself was weak, about 0.2 to 0.6 mW, as compared with the example, and less than about 1/3 as compared with that of the LED of the example. Although the wavelength of the main emission spectrum was about 440 nm, which is almost the same as that of the example, secondary light emission occurred in the wavelength range of about 425 to 430 nm and the wavelength range of about 400 nm. For this reason, the light emission of the LED of the example was observed to be blue, but the LED of the comparative example was rather white-blue-purple and monochromatic.

4 is an example of a cross-sectional TEM image showing a crystal structure inside a first intermediate layer. It is a SIMS analysis result which shows the distribution of the magnesium concentration of an Example in the depth direction. It is a plane schematic diagram of the LED described in an Example. FIG. 4 is a schematic diagram showing a cross-sectional structure along a dashed line A-A ′ in the schematic plan view shown in FIG. 3. It is a cross section of a laminated structure in a comparative example. 7 is a SIMS analysis result showing a distribution of magnesium in a depth direction in a comparative example. It is a cross section of another embodiment of the present invention. FIG. 6 is a view schematically showing a cross section of a laminated structure including an intervening layer on a light emitting layer, which is another embodiment of the present invention.

Explanation of reference numerals

(101) Substrate (102) First intermediate layer (103) Single crystal layer inside first intermediate layer (104) Crystal lattice image captured in cross-sectional TEM image (105) Single crystal inside first intermediate layer Grain (106) Lower cladding layer (Mg-doped p-type gallium nitride)
(107) Second intermediate layer (aluminum / gallium nitride)
(108) Light emitting layer (gallium indium nitride)
(109) Upper cladding layer (Si-doped n-type gallium nitride)
(110) p-type pad electrode (111) translucent n-type electrode (112) n-type pad electrode (113) low-temperature buffer layer (114) intervening layer

Claims (1)

_{Al x Ga y In z V a} N 1-a layer on a substrate crystal (where, V is the Periodic Group V element other than nitrogen atoms, x + y + z = 1,0 ≦ x, y, z ≦ 1,0 ≦ a <1 first intermediate layer consisting of), generally doped with a p-type impurity on the intermediate layer type _{Al x Ga y in z V a} N 1-a ( where, V is the periodic other than nitrogen atom V group element, x + y + z = 1,0 ≦ x, y, cladding layer represented by z ≦ 1,0 ≦ a <1) , Al x over the cladding layer _{Ga y in z V a N 1} -a ( where , V is a second intermediate layer made of a group V element of the periodic rule other than a nitrogen atom, x + y + z = 1, 0 ≦ x, y, z ≦ 1, 0 ≦ a <1); the concentration of the p-type impurity 1 × 10 ¹⁸ cm ^-3 under that indium-containing density gallium indium nitride consisting multiphase structure having different above _{(Ga x in y N; x} + y 1,0 ≦ x, y ≦ 1) light emitting layer and the n-type _{Al x Ga y In z V a} N 1-a layer of the light emitting layer (where, V is a group V element periodic table other than nitrogen atom , X + y + z = 1, 0 ≦ x, y, z ≦ 1, 0 ≦ a <1), and a group III nitride semiconductor light emitting device having a configuration in which electrodes are provided.

Images