WO2005106979A1

WO2005106979A1 - Nitride semiconductor light emitting element

Info

Publication number: WO2005106979A1
Application number: PCT/JP2005/008389
Authority: WO
Inventors: Hiromitsu Kudo; Tomoo Yamada; Kazuyuki Tadatomo; Yoichiro Ouchi; Hiroaki Okagawa
Original assignee: Mitsubishi Cable Industries, Ltd.
Priority date: 2004-04-28
Filing date: 2005-04-26
Publication date: 2005-11-10
Also published as: JP5082444B2; TW200603442A; TWI295859B; JPWO2005106979A1

Abstract

In a GaN-based light emitting element using Mg as a p-type impurity of a p-type layer, a p-type clad layer (31) composed of AlxGa1-xN (0≤x≤1) is provided at the lowermost part of the p-type layer, an Mg high concentration layer (33) composed of AlyGa1-yN (0≤y≤1) is provided on an upper part of the clad layer through at least one hetero interface, and directly above the Mg high concentration layer, a p-type contact layer (34) composed of AlzGa1-zN (y<z≤1) is provided. The p-type contact layer has a layer thickness of 10nm or less and an Mg concentration (a) of 5×1019cm-3≤(a), the Mg high concentration layer has a film thickness of 5nm-20nm and an Mg concentration (b) of 2×1019cm-3<(b), and a layer between the Mg high concentration layer and a light emitting layer has an Mg concentration (c) of 1×1019cm-3≤(c)<(b). An Mg concentration average of the p-type layers excluding the p-type contact layer is less than 5×1019cm-3. Light absorption due to Mg can be reduced, while suppressing an increase of an operating voltage, by reducing the total quantity of Mg added to the p-type layer, while maintaining the high Mg concentration of a front layer of the p-type contact layer.

Description

Specification

Nitride semiconductor light emitting device

Technical field

The present invention relates to a nitride semiconductor light emitting device such as a light emitting diode (LED) or a laser diode (LD), which is composed of a nitride semiconductor as a material of a light emitting layer, and more particularly to a p-type semiconductor in the light emitting device structure. It relates to the configuration of the part.

Background art

A nitride semiconductor is a group III nitride determined by the formula A 1 _α I n ^ G a _y N (0 ≤ a ≤ l 0 ≤ β ≤ 1, 0 ≤ y ≤ 1, α + β + γ = 1) Compound semiconductors of any composition, for example, G a N, In G a N, A l G a N, A l In g G a N, A l N, In n, etc. . In the present invention, the composition ratios of the group III elements constituting the nitride semiconductor, such as α, β, and γ in the above formula, are also referred to as A 1 ratio, I η ratio, and Ga ratio, respectively. By selecting the material composition of the nitride semiconductor used for the light emitting layer, it is possible to generate short wavelength light ranging from blue to ultraviolet.

Hereinafter, “nitride semiconductor” is also referred to as “GaN based light emitting device”. For example, “nitride semiconductor light emitting device” is also referred to as “GaN based light emitting device”. explain.

FIG. 3 is a diagram showing an example of a general element structure of a GaN-based LED, which is one of the GaN-based light-emitting elements, in which a G-N LED is formed on a crystal substrate 100 such as a sapphire substrate. The laminated body S1 composed of the GaN-based crystal layer is formed via the aN-based low-temperature growth buffer layer 100b. The laminate S1 has a pn junction structure including an n-type layer and a p-type layer, and a light emitting layer 120 is formed at the junction. Specifically, in order from the lower side (from the crystal substrate side), an n-type cladding layer 110 (also serving as an n-type contact layer in this example) and a light-emitting layer (a multilayer structure such as a multiple quantum well) are used. 1 20, a p-type cladding layer 130, and a p-type contact layer 140 are stacked by vapor phase growth. P 10 and P 20 are an n-side electrode and a p-side electrode, respectively. In a light emitting device having a double-headed structure, the light-emitting layer 120 has a II-type cladding layer 110 and a p-type cladding layer. It is made of a crystal having a smaller band gap than that of the layer 130. It is said that a light emitting element having a double hetero structure has a light emission output that is at least 10 times higher than that of a light emitting element having a homojunction (for example, Reference 1: Japanese Patent Application Laid-Open No. H8-33069). .

The n-type cladding layer 110 is formed to have n-type conductivity by adding an n-type impurity. The p-type cladding layer 130 and the p-type contact layer 140 are doped with a p-type impurity and, if necessary, subjected to a resistance reduction treatment such as a p-type annealing treatment. , Formed into p-type conductivity. The light-emitting layer 120 can be formed to have either n-type conductivity or p-type conductivity, or a mode in which these conductive layers are mixed. In some cases, the layer is undoped without intentionally adding impurities (an undoped layer that does not contain any impurities usually shows weak n-type conductivity).

Mg is used as a preferred p-type impurity for making the GaN-based crystal layer p-type conductive.

The lower the operating voltage of the light emitting element, the more practically desirable. To lower the operating voltage, it is better to lower the series resistance of the p-type layer.Therefore, it is preferable to increase the hole concentration of the p-type layer. The mobility deteriorates due to the deterioration of the hole mobility, and the resistance does not decrease sufficiently. Therefore, it is preferable that the amount of Mg added is such that the hole concentration is sufficiently high and the crystallinity is not significantly reduced.

To lower the operating voltage of the light emitting device, it is also important to reduce the contact resistance between the p-type layer and the P-side electrode.

In order to reduce the contact resistance, it is known that the hole concentration of a P-type contact layer, which is a layer on which a p-side electrode is formed, is increased. When G n with a high carrier concentration is used for the p-type contact layer, the uniformity with the p-side electrode is improved, and the forward voltage of the LED (applied voltage required to flow a predetermined forward current) is reduced. It is said to decrease (Ref. 2; Japanese Patent Application Laid-Open No. 7-15041).

Alternatively, it is said that increasing the Mg concentration in the p-type contact layer is not effective through increasing the carrier concentration, but is itself effective in reducing the contact resistance. Therefore, a method is known in which the operating voltage and the forward voltage of the light emitting element are reduced by increasing the Mg concentration on the surface layer side of the p-type contact layer (see References 1, 3; Japanese Patent Application Laid-Open No. 8-97471). No.). According to Reference 3, when the Mg concentration is higher than about 5 × 10 ¹⁹ / cm ³ , the rate of increase of the carrier concentration with respect to the increase of the Mg concentration decreases, and finally, the Mg concentration and the carrier concentration are inversely proportional. However, when the Mg concentration of the p-type contact layer is set to such a high concentration region, the ohmic property with the p-side electrode is improved, and the operating voltage is reduced.

It has been proposed that when Mg is used as a p-type impurity, the luminous efficiency is reduced due to light absorption by the Mg (for example, Reference 4; Japanese Patent Application Laid-Open No. H10-125956). Light absorption by Mg is a phenomenon in which Mg forms a deep level acceptor level in a GaN-based crystal, and light with a wavelength of 430 nm or less is absorbed by the band gap between the deep level and the conduction band. Has been described.

In the invention described in Reference 4, in order to solve this problem, the p-type cladding layer is made of AI GaN having an A 1 ratio of 8% or more, so that the above-mentioned band gap is enlarged, and the wavelength of the absorbed light is increased. Is reduced to less than 430 nm, and the thickness of the p-type cladding layer and p-type contact layer is reduced, thereby reducing the problem of light absorption as described above. In order to achieve this, it is known to reduce the thickness of the p-type layer (the entire p-type conductive layer formed above the light-emitting layer, including the p-type cladding layer and the p-type contact layer).

Therefore, the present inventors consider that it is most important to reduce the thickness of the p-type contact layer, and formed a p-type contact layer doped with Mg at a high concentration to a thickness of 10 nm or less, and formed the p-type A GaN-based LED with a rapidly reduced Mg concentration in the layer was fabricated and its evaluation was attempted.

However, it has been found that the operating voltage of such a GaN LED increases. The reason is considered to be that the Mg concentration in the surface layer of the p-type contact layer decreases due to diffusion due to the concentration gradient, and the contact resistance with the p-side electrode increases. The Conversely, in order to sufficiently increase the Mg concentration in the surface layer of the p-type contact layer and reduce the contact resistance with the p-side electrode, the p-type contact layer must be thickened to a certain degree or more. Inevitably, light absorption was involved.

In other words, in the prior art, it has not been possible to provide a p-type contact layer having a high Mg concentration to lower the operating voltage of the light emitting device and to sufficiently suppress light absorption in the p-type layer. Was.

Disclosure of the invention

An object of the present invention is to solve the above-mentioned problems of the p-type layer and reduce the total amount of Mg added to the p-type layer while maintaining a high Mg concentration in the surface layer of the p-type contact layer. An object of the present invention is to provide a GaN-based light emitting device with a p-type layer structure that can reduce light absorption due to Mg while suppressing an increase in operating voltage.

The inventors of the present invention have proposed the unique change of the composition of the GaN-based material, the change of the Mg concentration in the stacking direction, and the thickness of each layer from the p-type contact layer to the layer closest to the light emitting layer in the p-type layer. It has been found that the above object can be achieved by adding a limitation, and the present invention has been completed.

That is, the present invention has the following features.

(1) It has a stacked body composed of a nitride semiconductor crystal layer, and the stacked body includes an n-type layer, a light-emitting layer, and a p-type layer in order from the lower layer side, and the p-type layer contains Mg as a p-type impurity. A doped nitride semiconductor light emitting device,

At the bottom of the p-type layer, there is provided a p-type cladding layer composed of A 1 _X G a _X N (0≤x≤1), and above the p-type cladding layer via one or more heterointerfaces. _{_{_{a 1 y G a x _ y}}} N consists (0≤y≤ l) M g high concentration layer is provided on the upper straight of the Mg-rich layer! A) _Z G a (y <z≤ 1) as the top of the) type layer! ) Type contact layer is provided,

The p-type contact layer has a thickness of 10 nm or less, a Mg concentration a of 5 × 10 ¹⁹ cm− ³ ≤ a,

Mg high concentration layer has a thickness of 5 nm~20 nm, M _g concentration b is 2 X 10 ¹⁹ cm one ³ B

of the p-type layer, a layer between the Mg-enriched layer and the light emitting layer, Mg concentration c is ^{^{^{1 X 10 1 9 cm- 3 ≤}}} c <b,

A nitride semiconductor light emitting device, characterized in that the average value of the Mg concentration of the p-type layer excluding the p-type contact layer is less than 5 × 10 ¹⁹ cm— ³ .

(2) The nitride semiconductor light-emitting device according to (1), wherein the Mg-rich layer includes a portion having a Mg concentration of 5 × 10 ¹⁹ cm- ³ or more at least on a side in contact with the p-type contact layer.

(3) The nitride semiconductor light-emitting device according to the above (2), wherein the high Mg concentration layer includes a portion having a Mg concentration of 5 XI 0 ¹⁹ cm- ³ or more only on the side in contact with the p-type contact layer.

(4) The nitride semiconductor light-emitting device according to (1), wherein the Mg concentration c is c <5 × 10 ¹⁹ cm− ³ .

(5) The nitride semiconductor light-emitting device according to (1), wherein b <a.

(6) The nitride semiconductor light-emitting device according to (1), wherein a layer between the Mg-rich layer and the light-emitting layer in the p-type layer includes a portion having a Mg concentration of less than 2 × 10 ¹⁹ cm− ^3. element.

(7) The nitride semiconductor light-emitting device according to (6), wherein an average value of Mg concentration in a layer between the Mg-rich layer and the light emitting layer is less than 2 × 10 ¹⁹ cm− ³ .

(8) The nitride semiconductor light emitting device according to (1), wherein the p-type layer has a thickness of 100 nm to 300 nm.

(9) p-type contact layer has a thickness of 0. 5~: 1 0 nm, Mg concentration a is ^{^{^{5 X 10 1 9 cm _3 ≤}}} a≤ 1 X 10 21 cm - ^3, and

Mg high concentration layer has a thickness is 5 to 20 nm, 1 § concentration 1> is ^{^{2 1 0 19 cm- 3 <(}} 0. 5 X a) ≤b≤ 1 X 10 21 cm- 3,

The nitride semiconductor light-emitting device according to the above (1), wherein a layer between the high Mg concentration layer and the light emitting layer among the p-type layers has a Mg concentration c of (0.2 X b) ≤ c <b.

Brief Description of Drawings

FIG. 1 is a schematic view showing an element structure of a GaN-based light emitting element according to the present invention. C Touching is performed for the purpose of distinguishing the areas.

FIG. 2 is a graph showing the relationship between the thickness of the Mg high-concentration layer, V f, and emission output. FIG. 3 is a schematic diagram showing the element structure of a conventional GaN-based light emitting element.

Each symbol in FIG. 1 indicates the following, respectively. 1; undoped layer, 2; n-type layer, 3; light-emitting layer, 4; p-type layer, 41; p-type cladding layer, 42; low Mg concentration layer, 43; high Mg concentration layer, 44 ; _P- type contact layer

BEST MODE FOR CARRYING OUT THE INVENTION

In this specification, in order to describe the position of each layer in the layered structure of the light-emitting element, terms indicating a vertical relationship such as “lower layer side”, “bottom”, and “directly above” are used. This is for convenience based on the stacking order, such as forming the light emitting layer and the p-type layer on the n-type layer formed first on the crystal substrate in the process of forming the stacked structure with the crystal substrate on the lower side. It is an expression, and does not limit the absolute vertical direction of the device or the mounting direction (posture at mounting) of the device. "Directly above" means directly above and directly below, and "directly below" means directly below and below.

Hereinafter, the present invention will be described using a GaN-based LED, which is one of the GaN-based light-emitting elements, as an example. The aspect of the semiconductor laser will be referred to as necessary.

The GaN-based LED device structure according to the present invention has a device structure in which a GaN-based crystal layer is sequentially grown on a crystal substrate B1 to form a laminate S, as shown in an example in FIG. The laminate S includes an undoped layer 1, an n-type layer 2, a light-emitting layer 3, and a p-type layer 4 in this order from the lower layer side. As described in the above (1), in the present invention, the p-type impurity doped into the p-type layer is Mg.

In some cases, the n-type layer 2 includes an n-type contact layer and an n-type cladding layer independently, but in the example shown in the figure, only one layer is used for both layers.

The light-emitting layer 3 is a layer for generating light emission due to recombination of carriers, and may have not only a single-layer form but also a laminated structure as described later.

P-type layer 4 has at least the bottom; type cladding layer 4 1, M above it g high-concentration layer 43, and directly above the layer 43; a type contact layer 44 is included. The characteristics of the P-type layer composed of the p-type cladding layer 41, the Mg-rich layer 43, and the p-type contact layer 44 are as described in (1) above.

In the example of FIG. 1, the Mg low concentration layer exists between the p-type cladding layer 41 and the Mg high concentration layer 43, which will be described later.

Although the upper surface of the crystal substrate may be flat as in the example of FIG. 3, in the example of the element structure of FIG. 1, irregularities (described later) are formed on the upper surface of the crystal substrate B1, and G a N A buffer layer B2 made of a base material is formed, and an undoped GaN layer 1 and an n-type GaN cladding layer (also serving as an n-type contact layer) 2 are grown to cover the irregularities. The laminate S is etched from the p-type layer side so that the n-type GaN cladding layer 2 is partially exposed, and the exposed portion is provided with an n-side electrode P1 made of A1 (aluminum). Further, on the upper surface of the p-type contact layer 44, a p-side electrode P2 formed by laminating Ni (nickel) and Au (gold) in this order from the side in contact with the upper surface is provided.

It is optional to extract the light emitted from the light emitting layer 3 from above (from the p-side electrode side) or from below (through the back side of the substrate) through the crystal substrate. Or, it is sufficient to adopt a structure that allows mounting in a normal posture and flip-chip mounting.

As described above, according to the present invention, at least the p-type cladding layer, the (1> type) ^ 18 high-concentration layer, and the p-type contact layer are formed in order from the lower layer side as a stacked structure of the p-type layer 內. The structure is included. Other layers may be present between the Mg-rich layer and the p-type cladding layer (described later). Next, the material composition, Mg addition concentration, layer thickness, etc. of these layers are defined as described in (1) above (preferable embodiments are defined as in (2) to (7) above). The above purpose is achieved by generating the following three effects.

(i) Improvement of crystallinity

The p-type cladding layer is a binary or ternary crystal of A 1 _X G a, _-X N (0≤x≤ 1), M g High-concentration layer is binary or ternary crystal A 1 _y G _ay N (0≤ y≤ 1), p-type contact layer is ternary crystal A 1 _Z G a i_ _z N (y <z ≤ 1) because it is easy to obtain high-quality crystals with low dislocation density.

Since the mobility of holes is improved by reducing the dislocation density, the resistance of the p-type layer is reduced, and the operating voltage of the light-emitting element can be reduced. In addition, the diffusion of Mg along the dislocation defect region is suppressed.

(ii) Reduction of contact resistance

Let the p-type contact layer be A 1 _Z G a (y <z ≤ 1) and let its Al ratio z be

The A1 ratio of the Mg-rich layer is set higher than y because the p-type contact layer formed at the top of the p-type layer is formed by increasing the ratio of A1, which has a stronger binding force with nitrogen than Ga. This is because, during crystal growth, during cooling after crystal growth, during p-type annealing, during electrode annealing, and the like, nitrogen is prevented from being released when exposed to a high-temperature atmosphere.

When nitrogen escape occurs in the GaN crystal layer, the remaining nitrogen vacancies give n-type conductivity, hindering the development of p-type conductivity, causing contact resistance with the electrode and This will increase the series resistance inside.

In the present invention, by increasing the A 1 ratio of the p-type contact layer higher than that of the layer immediately below, the nitrogen release is suppressed, and the increase in the contact resistance with the electrode and the increase in the series resistance in the layer are suppressed. When the Mg concentration a of the layer is set to 5 × ^{10 19} cm— ³ ≤ a, a preferable ohmic contact with the p-side electrode can be achieved within this concentration range, and the contact with the electrode can be achieved. This is because the resistance can be made sufficiently low.

(iii) Reduction of Mg content in p-type layer and suppression of Mg diffusion from p-type contact layer

What makes the p-type contact layer extremely thin, less than 10 nm, is this layer! Since Mg is added at a high concentration in order to obtain ohmicity with the side electrode, it is most effective to reduce the thickness of this layer in suppressing light absorption due to Mg. is there. The wavelength of light absorbed by Mg-added p-type layer is Mg This is because the higher the concentration, the longer the wavelength of Mg to form a deeper level, and the longer the absorption wavelength, the more easily the light generated in the light emitting layer is absorbed.

However, simply reducing the thickness of the p-type contact layer containing a high concentration of Mg raises the problem of increasing the contact resistance.

In contrast, in the present invention, the amount of Mg contained in the p-type layer is reduced while preventing an increase in the contact resistance of the p-type contact layer by devising the configuration of the entire p-type layer. First, at least two hetero interfaces, which are bonding interfaces of crystals having different compositions, are provided between the p-type cladding layer and the p-type contact layer. That is, at least one provided are hetero interface between the p-type cladding layer and the Mg-enriched layer, _{_{and, A l y Ga! _ Y}} N and Mg high concentration layer made of (0≤y≤ 1), A l _z G _ai - is a hetero-interface is at least one formed between the _z N (y _<z≤ 1) made of p-type contact layer.

Since an electric field due to the polarization of the crystal exists at the hetero interface, ionized impurities are easily captured, which has the effect of suppressing the diffusion of Mg.

Next, the Mg concentration b of the Mg-rich layer is set to 2 × 10 ¹⁹ cm− ³ <b, the layer thickness is set to 5 nm to 20 nm, and the p-type conduction layer (p-type In other words, the Mg concentration c of the p-type conductive layer between the high Mg concentration layer and the light emitting layer is set to 1 × 10 ¹⁹ cm ⁻³ ≤ c <b. This inequality c <b means that the Mg concentration b in the Mg-rich layer is higher than the Mg concentration in such a portion of the p-type conduction layer between the Mg-rich layer and the light emitting layer. I have. Therefore, the limitation that the layer thickness of the Mg-rich layer is set to 20 nm or less and c <b is limited to the region below the p-type contact layer where the distance from the lower surface of the p-type contact layer is within 20 nm. This means setting the Mg concentration in the p-type layer so that the Mg concentration is the highest.

Furthermore, the average value of the Mg concentration of the p-type layer excluding the p-type contact layer, that is, the average value of the Mg concentration in the p-type conduction layer including the Mg-rich layer and the p-type cladding layer is 5 × 10 ¹⁹ cm less than ³ cm. By setting the Mg concentration of the _P- type conductive layer between the p-type contact layer and the light emitting layer in this way, it is possible to prevent the increase in the contact resistance of the p-type contact layer and to increase the Mg added to the p-type layer. Can be kept low.

The crystal substrate may be any as long as a GaN-based crystal can be grown. Preferred crystal substrates include, for example, sapphire (C-plane, A-plane, R-plane), SiC (6H, 4H, 3C), GaN, A1N, Si, spinel, ZnO, GaAs, NGO, etc. No. Further, a substrate having these crystals as a surface layer may be used. The plane orientation of the substrate is not particularly limited, and may be a just substrate or a substrate having an off angle.

It is preferable to interpose a buffer layer between the crystal substrate and the GaN-based crystal layer. This is because the dislocation density of the p-type layer formed above is reduced and the crystallinity is improved. As described above, when the dislocation density of the p-type layer is reduced, favorable effects such as reduction in electric resistance and suppression of Mg diffusion are generated.

Preferred buffer layers include GaN-based buffer layers. The material, formation method, and formation conditions of the buffer layer may refer to a known technique.Examples of GaN-based buffer layer materials include GaN, AlGaN, A1N, InN, and the like. It is sufficient that the temperature is lower than the growth temperature of the GaN-based crystal layer grown thereon, for example, 300 ° C. to 700 ° C.

The thickness of the buffer layer is preferably 10 nm to 50 nm, particularly preferably 20 nm to 40 nm. It is preferable to use GaN as the material of the buffer layer because the dislocation density of the crystal layer grown upward is lowest and the optimum thickness range of the buffer layer is wide.

When a substrate made of GaN, A 1 N crystal, or the like is used as the crystal substrate, the buffer layer is not essential.

The GaN-based crystal layer may be grown on a flat crystal substrate, but a structure for reducing the dislocation density in the crystal may be appropriately introduced on the crystal substrate surface. Along with this, a portion made of a different material such as SiO ₂ may be included in the laminate composed of the GaN-based crystal layer. As a structure for reducing the dislocation density, (i) a mask layer (for example, Sio ₂ is used) is formed on a top surface of a crystal substrate by a stripe pattern so that a conventionally known selective growth method (ELO method) can be performed. And (3) a structure in which the top surface of the crystal substrate is processed to have irregularities such as dots or stripes so that the GaN crystal can grow laterally or facetly. .

These structures and the buffer layer may be appropriately combined.

Among the structures for reducing the dislocation density, the structure in which the above-mentioned surface roughness is applied to the upper surface of the substrate does not use a mask layer, so the diffusion of the mask layer material contaminates the GaN crystal. This is preferred because it is prevented. In addition, when a GaN crystal is grown so as to bury the convexity, when a crystal substrate made of a material different from the GaN material, such as a sapphire substrate, is used, a crystal substrate having a different refractive index is used. a Since the interface with the N-based crystal becomes light-scattering, a favorable effect of improving the light extraction efficiency of the LED (an effect independent of dislocation density reduction) is produced.

If the GaN-based crystal grown by embedding the irregularities is GaN, especially undoped GaN, it grows upward because it is easy to obtain high-quality crystals with good growth surface flatness and low dislocation density. ! ) It is preferable for improving the quality of the mold layer.

The method of processing the unevenness on the upper surface of the crystal substrate, the arrangement pattern of the unevenness, the cross-sectional shape of the unevenness, the growth process of the GaN-based crystal on the unevenness, etc. are described in JP-A-2000-331947, JP-A-2002-164296. References may be referred to. In addition, in the case of forming the grooves as stripes in the form of stripes, the longitudinal direction of the grooves, the width of the grooves, the width of the convex ridges, the amplitude of the unevenness (the depth of the grooves), and the like are also described in these documents and publicly known. You may refer to the technology.

GaN-based crystal layer growth methods include HVPE, MOVPE, and MBE. The MO VPE method is particularly preferable because a high-quality crystal thin film can be formed at a practical growth rate.

The light-emitting layer has a multilayer structure such as a single quantum well (SQW) structure or a multiple quantum well (MQW) structure consisting of multiple layers with different band gaps, even if it has a structure composed of a single composition crystal layer. There may be. Strictly speaking, in a quantum well structure, light emission occurs In the present invention, the entire stacked structure of the barrier layer and the well layer is referred to as a light emitting layer as one unit.

The composition of the GaN crystal used for the light emitting layer may be appropriately selected from compositions having a band gap according to the wavelength of the light to be generated, but the band gap is wider than that of the P-type cladding layer and the n-type cladding layer. It is particularly preferable to use a composition having a small value because a double hetero structure is obtained. When the light emitting layer has a quantum well structure, the band gap of the well layer may be smaller than the band gap of the p-type cladding layer and the n-type cladding layer.

When the light-emitting layer is composed of InGaN, the emission wavelength can be widened from about 360 nm (zero In-ratio) to the infrared wavelength range by adjusting the In ratio of the crystal. It can be controlled throughout. The emission wavelength can also be controlled by adding an n-type impurity or a p-type impurity to the light-emitting layer.

When the light-emitting layer is composed of InGaN, a region in which In is locally distributed at a high concentration is formed, and the region functions as a luminescence recombination center, so that the light-emitting layer has a relatively high dislocation density. , High luminous efficiency can be obtained.

However, if the In ratio is reduced to reduce the emission wavelength to a range of less than 420 nm in the violet-ultraviolet region, such effects are unlikely to occur, and the dislocation density in the emission layer greatly affects luminous efficiency. I will do it. Therefore, when the emission wavelength is in such a short wavelength region, it is preferable to employ a structure effective for reducing the dislocation density described above in a structure below the light emitting layer.

Carriers can be effectively confined in the light emitting layer by using a material having a larger band gap than the material of the light emitting layer as the material of the n-type cladding layer. In the case of LEDs, since the current density during use is relatively small, it is not necessary to make the difference in the gap between the light emitting layer and the light emitting layer too large. It is also possible to use one with no gap difference (of the same composition) or one with a smaller band gap.

The reason is that the injection from the p-type cladding layer into the light-emitting layer The diffusing holes recombine with the electrons with a high probability before reaching the n-type cladding layer, because the diffusing holes have a lower mobility than the electrons diffusing in the opposite direction.

Therefore, when using InGaN for the light emitting layer of the LED, the material of the n-type cladding layer includes a binary crystal G aN, a ternary crystal A 1 G a N It is preferable to use InG aN, and particularly preferable to use G aN. In the case of A 1 GaN, the crystallinity tends to decrease when the A 1 ratio is increased. Therefore, the A 1 ratio is preferably set to 0.2 or less, more preferably 0.1 or less.

In the present invention, it is used A 1 _X G a nN as the material of the p-type cladding layer (0≤ x≤ 1). That is, the p-type cladding layer is composed of binary crystal GaN or ternary crystal A 1 GaN. It is desirable to select the composition of the p-type cladding layer so that the band gap is larger than that of the light emitting layer. Specifically, in order to effectively confine electrons diffusing in the light emitting layer in the direction of the p-type cladding layer into the light emitting layer, a band gap with the light emitting layer (or a well layer when the light emitting layer has a quantum well structure) is used. The composition is preferably such that the difference is 0.3 eV or more.

The relationship between the emission wavelength λ (nm) and the band gap Eg (eV) of the light emitting layer (well layer when the light emitting layer has a quantum well structure) is expressed by the following equation.

XE g = 1240 (Equation 1) On the other hand, the band gap E g _A1GaN of A l _x G _ai — _X N (0≤ x≤ 1) is expressed by the following equation.

E g _A1GaN = _6.16 x _+3.4 (1 一 X) _-χ (1— χ) (Equation 2) Therefore, using these relational expressions, the minimum A 1 ratio of A 1 _X G ai— The value can be determined.

For example, when the emission wavelength is 400 nm, since the band gap of the light emitting layer is 3.1 eV according to (Equation 1), the preferable band gap of the p-type cladding layer is 3.4 eV or more. ) Is preferable in this case: The A 1 ratio X of the mold clad is 0.06 or more. When the ratio A 1 is increased, A l „G _ai — _Since the crystallinity of XN tends to decrease, X is preferably 0.2 or less, more preferably 0.1 or less. As described above, when the crystallinity is lowered, Mg is easily diffused along dislocation defects.

The thickness of the p-type cladding layer is not particularly limited, and may be determined by appropriately referring to a known technique, but is preferably 10 nm to 100 nm, and more preferably 20 ηπ! ~ 70 nm.

The Mg-rich layer can be said to be the second contact layer located immediately below the thin p-type contact layer.

The material composition of the Mg-rich layer is A _y G _ai — _y N (0≤y≤l), and y = 0 (that is, G a N).

The present invention is characterized in that at least one hetero interface is provided between the p-type cladding layer and the Mg-rich layer.

Therefore, when the Mg-rich layer is placed directly above the p-type cladding layer, the A1 ratio X of the p-type cladding layer should be different from the A1 ratio y of the Mg-rich layer. . In that case, it is considered that the larger the difference between X and y, the greater the effect of suppressing the diffusion of Mg.Therefore, the difference between X and y is preferably 0.01 or more, more preferably 0.03 or more, More preferably, it is 0.05 or more.

If the difference between X and y is too large, the dislocation density tends to increase due to lattice mismatch. Therefore, the difference between X and y is preferably 0.2 or less, more preferably 0.15 or less. Preferably it is 0.1 or less.

Since the crystallinity tends to decrease as the A 1 ratio y of the Mg-rich layer increases, y is preferably set to 0.2 or less. Further, the A 1 ratio y of the Mg-rich layer may be smaller than the A 1 ratio X of the p-type clad layer. As y becomes smaller, the band gap of the Mg-rich layer becomes smaller, so that it becomes easier to activate as a Mg-force Sp-type impurity, and a decrease in series resistance due to an increase in carrier concentration can be expected. Therefore, the more preferable range of y is 0≤y≤0.05, and it is more preferable to form the Mg-rich layer with G a N. the material composition of the p-type contact layer, A l _z G _ai - a _{z N (y <z≤ 1)} , A 1 ratio z of that is larger than A 1 ratio y of Mg high concentration layer, the As described in Effect of the Invention, the effect of suppressing nitrogen release from the surface in a high-temperature atmosphere and the effect of suppressing Mg diffusion by the hetero interface are generated. From the viewpoint of suppressing nitrogen elimination in a high-temperature atmosphere, 0.01 ≤ z is preferable, and particularly preferably 0.03 ≤ z.

On the other hand, when z is large, crystallinity tends to decrease, and activation of Mg as a p-type impurity is unlikely to occur.From this viewpoint, z is preferably 0.2 or less, and more preferably z. 0.1 or less, more preferably 0.05 or less.

The Mg concentration a of the p-type contact layer may be set to a concentration high enough to form a good ohmic contact with the electrode, and a known technique can be referred to. Specifically, if 5 x 10 ¹⁹ cm- ³ ≤ a, especially 1 x 10 ² ° cm- ³ ≤ a, the distance between various electrodes known as p-type common electrodes As a result, a good ohmic contact is formed.

On the other hand, in order to further enhance the effect of suppressing light absorption caused by Mg, the Mg concentration a is preferably 5 × 10 ¹⁹ cm— ³ to 1 × 10 ²¹ cm— ³ , more preferably 5 × 10 ²¹ cm— ^3. ¹⁹ cm— ^{3 to} 5 X 10 ² . cm- ^3, and particularly preferably 5 X 1 0 ¹⁹ cm one ³ to 1 X 10 ²⁰ cm one ^3.

If the thickness of the p-type contact layer is 0.5 nm or more, the operating voltage of the light-emitting device will be lower than in the case where the p-side electrode is formed directly on the Mg-rich layer, probably due to the reduction in contact resistance.

On the other hand, since the p-type contact layer is added with a high concentration of Mg, light absorption due to Mg is suppressed as the layer thickness is reduced.

Therefore, the layer thickness of the p-type contact layer is preferably from 0.5 nm to 10 nm, more preferably from 0.5 11111 to 5 11111.

When the p-type layer is formed by the MOVP E method, if a large amount of hydrogen that inhibits activation of Mg as a p-type impurity enters the p-type layer during crystal growth, the series resistance of the p-type layer increases. And the operating voltage of the device is estimated to increase due to an increase in the contact resistance between the _P- type contact layer and the p-side electrode. Alternatively, it is necessary to perform a process such as annealing for activating Mg under more severe conditions, and there is a possibility that the device may be deteriorated due to the process.

Therefore, it is preferable to use an inert gas such as nitrogen among the gases supplied into the reaction vessel during the growth of the p-type layer, except for the carrier gas used as the organometallic compound raw material. The growth rate of the material is significantly reduced.

Therefore, if the thickness of the p-type layer is large, the time required for the growth is long, and as a result, the time required for the crystal layer to be exposed to high temperatures is prolonged. Degradation is promoted, and the device characteristics tend to deteriorate. Reducing the thickness of the Mg-rich layer or the ρ-type contact layer is also preferable in suppressing such a decrease in device characteristics.

The Mg concentration b of the Mg-rich layer should be 2 × 10 ¹⁹ cm— ³ <b. In particular, the Mg concentration at the part in contact with the p-type contact layer should be 5 × 10 ¹⁹ cm— ³ or more. Is preferred.

against Mg concentration a of the p-type contact layer, (0. 5 X a) be ≤b≤ 1 X 10 ²¹ cm one ^3, one preferred embodiment for achieving the object of the present invention Tsudea You.

From the viewpoint of suppressing light absorption due to Mg, it is preferable that the Mg concentration b of the Mg high concentration layer be lower than that of the p-type contact layer. Since the Mg-rich layer is formed of a GaN-based crystal with a smaller A1 ratio than the p-type contact layer, that is, a GaN-based crystal with a smaller bandgap, the Mg-rich layer and the p-type If the same layer is doped with the same concentration of Mg, the light absorption in the Mg-rich layer becomes larger than that in the p-type contact layer.

When Mg concentration b is higher than 1 X 10 ²¹ cm one ^3, there is a tendency that the crystallinity of the Mg-rich layer is low down.

The layer thickness of the high Mg concentration layer is 5 nm to 20 nm. 5x10 Mg high concentration layer If a part containing Mg at a concentration of ¹⁹ cm- ³ or more is included, only the side of the Mg-rich layer that is in contact with the p-type contact layer has the Mg concentration, since light absorption by Mg is suppressed. it is preferable that the 5 X 1 0 ¹⁹ cm one ³ or more. When this is done, the p-type layer of the lower than the p-type contact layer, the thickness of the portion is Mg concentration 5 X 1 0 ¹⁹ cm one ³ or more, and 20 nm or less. A more preferred range for this thickness is 5 nrr! 1515 nm, more preferably 5 nrr! ~ 10 nm.

The Mg concentration and thickness in terms of providing the M _g high concentration layer specified as described above, p-type conductive layer between the Mg-enriched layer and the light-emitting layer in the p-type layer (e.g., in Figure 1 In the example, the Mg concentration c of the p-type cladding layer 31 and the Mg low concentration layer 32) is set to 1 X 10 ¹⁹ c'm ^1-3 ≤ c, so that the p-type The diffusion of Mg from the surface of the contact layer is suppressed, and the rise in operating voltage of the device is suppressed. In addition, setting (0.2 Xb) ≤ c is one of preferred embodiments for achieving the object of the present invention.

Furthermore, with the c <b, by a set of average values of the Mg concentration in the p-type layer except for the p-type contact layer to less than 5 X 10 ¹⁹ cm one ^3, suppression of increase in operating voltage of the device However, light absorption due to Mg in the entire p-type layer can be suppressed.

It is preferable to change the Mg concentration stepwise between the high Mg concentration layer and the p-type conductive layer immediately below. In this case, across the boundary between both layers, l X 10 ¹⁹ cm one ³ or more, further can be provided with 3 X 10 ¹⁹ cm one ³ or more Mg concentration difference. Contact name by the condition that c <b, formed in the step of such a Mg concentration is also essential for a not, for example, a p-type cladding layer toward M _g high concentration layer, Mg concentration linearity to increase Is also acceptable.

Further, the Mg concentration c, it is preferably Ri yo an upper limit value of 5 X 10 ¹⁹ cm one ^3, because, in the p-type conductive layer doped with Mg, the Mg concentration 5 X 10 ¹⁹ cm one ^{If the} concentration is higher than ³ (preferable Mg concentration for the p-type contact layer), light absorption caused by Mg becomes a particular problem. Also, in order to reduce the total amount of Mg contained in the p-type layer, the P-type layer between the Mg-rich layer and the light-emitting layer contains a portion where the Mg concentration is less than 2 × 10 ¹⁹ cm- ^3. preferably to so that Re, further more preferably an average value of the Mg concentration in the!) type layer is made to be less than 2 X 10 ¹⁹ cm one ^3.

The p-type contact layer, the Mg-rich layer, and the p-type cladding layer may be continuous three layers, but between the Mg-rich layer and the p-type cladding layer, another Mg-rich layer is provided. A low Mg concentration layer having a lower Mg concentration may be further provided.

The Mg low concentration layer is preferably provided so that the entire p-type layer has a thickness of 100 nm or more, preferably 100 nm to 30 O nm, and more preferably 100 nm to 200 nm. It is preferable that the total thickness of the p-type layer is larger than 100 nm because the balance with the n-type layer is improved, during cooling after crystal growth, during p-type annealing, and during electrode formation. This is because deterioration of the light emitting layer at the time of annealing treatment or the like is suppressed. If the total thickness of the p-type layer is larger than 30 O nm, these effects are saturated, and if the layer thickness is further increased, the problem of light absorption by Mg increases, and the manufacturing time increases. Problems such as reduced efficiency and wasted material occur. Also, if the growth time of the p-type layer is too long, there is a problem that the device is deteriorated due to the high temperature during the growth.

The composition of the low Mg concentration layer may be determined so that the interface with the high Mg concentration layer and / or the p-type cladding layer is a hetero interface.

From the viewpoint of crystallinity, it is preferable to use binary GaN or ternary A 1 GaN for the low Mg concentration layer. When A 1 G aN is used, since the crystallinity tends to decrease when the A 1 ratio is high, the A 1 ratio is preferably 0.2 or less, particularly preferably 0.1 or less.

M g low concentration layer directly above the M g high concentration layer material having the same crystal composition as A l _{_y} G _ai _ _y N a (0≤y≤ 1) may be used. If the low Mg concentration layer has the same crystal composition as the Mg high concentration layer directly above, or has a superlattice structure, the crystallinity of the Mg high concentration layer immediately above and, by extension, the P-type contact layer above it Can be expected to improve This is preferable from the viewpoint of lowering the resistance of the Mg-rich layer or the p-type contact layer and suppressing the diffusion of Mg that occurs along dislocation defects.

An additional layer may be inserted between the Mg-rich layer and the p-type cladding layer, if necessary, in view of the element structure. For example, in the case of forming a surface emitting laser (a laser in which the emission direction ′ matches the emission direction and the stacking direction), a plug reflection layer as one reflection layer of the resonator, and an edge emission laser (an oscillation direction · A light confinement layer in the case of forming a laser whose emission direction is perpendicular to the lamination direction).

The Bragg reflection layer or the light confinement layer may be an independent layer, or may be a layer partially or entirely serving as a high Mg concentration layer or a low Mg concentration layer.

From the viewpoint of crystallinity, it is preferable to use a binary crystal GaN or a ternary crystal A 1 GaN, and to use A 1 GaN when using A 1 GaN. The A 1 ratio of the GaN crystal is preferably 0.2 or less, particularly preferably 0.1 or less. Further, each of the layers such as the p-type contact layer, the Mg-rich layer, the Mg-poor layer, and the p-type clad layer is provided that the crystal composition and the Mg concentration fall within a predetermined range defined by the present invention. A multilayer structure (including a superlattice structure) may be used.

Example

Hereinafter, the GaN-based LED as an actual example of the GaN-based light emitting device of the present invention, including an experiment for specifying the preferable thickness of the p-type contact layer and the Mg concentration of the Mg-rich layer, will be described. A specific configuration will be described.

The Mg concentration and film thickness of the Mg-doped layer shown in the following experiments are all design values, and the measured values of the actually obtained product may include manufacturing errors.

The growth of the Mg-doped layer having a specific Mg concentration (design value) in each experiment was performed by the following procedure.

(A) Supply of Mg raw material (biscyclopentagenenylmagnesium) and supply of Group 3 raw materials (trimethylgallium, trimethylaluminum) when growing GaN-based crystal layer to be grown by MOVPE method Ratio to quantity (Mg, 3 group The relationship between the ratio] and the Mg concentration in the crystal actually obtained is examined in advance. The film thickness of the crystal layer grown on the other Me was about 300 nm, Mg concentration S IMS (Secondary I on mass spectrometry: Secondary Ion Mass Spectroscopy) measured by ₀

From (mouth) above relationship, Mg concentration of predetermined design value sought [Mg Roh group III ratio], while supplying the Mg raw material and group III material at its [M _{g /} group III ratio], MOVPE A GaN-based crystal layer is grown by the method.

The film thickness of each layer was controlled by determining the growth time required to grow the film to a predetermined thickness from the growth rate when the film having the composition was grown alone, and growing the film for the growth time.

The thickness of the A 1 G aN layer and G a N layer fabricated in each experiment was measured to be approximately the designed values by measuring the distribution of G a and A 1 in the depth direction by using SIMS. confirmed. Especially when the film thickness was small, XPS (X-ray Photoelectron Spectroscopy), which is an analysis method with higher resolution in the thickness direction, was also confirmed.

In addition, it was confirmed by SIMS that the Mg concentration in each layer was almost as designed. In particular, when performing SIMS measurement near the surface of the crystal layer, the resolution in the depth direction was increased by lowering the etching rate.

When the thickness of the p-type contact layer or the high-concentration Mg layer is small, the influence of the diffusion and outflow of Mg into the adjacent layer or the inflow of Mg from the adjacent layer becomes large. Was sometimes inclined. In some cases, the thickness of the Mg-rich layer was smaller than the design value due to the outflow of Mg to the adjacent Mg-poor layer.

Experiment 1

In this experiment, a GaN LED having the structure shown in Fig. 1 was fabricated. The MOVP E method was used to grow the GaN-based crystal layer.

In order to specify the preferred thickness of the p-type contact layer (AlGaN) 44, the thickness of the contact layer is assumed to be A, and the p-type GaN layer (Mg low concentration layer 42 and M The light emission characteristics were evaluated by setting the layer thickness of the high-concentration layer 43 as one layer) as B, and changing the thicknesses A and B of both layers.

The configuration of each layer of the GaN-based LED fabricated in this experiment is shown in order from the upper layer to the lower layer.

(P-type contact layer)

Materials; A1. . ₃ G a. . ₉₇ N, Mg ^{concentration; 1 X 10 2 ° (cm-} 3), the layer thickness; A (nm).

(P-type G a N layer)

The p-type GaN layer should be divided into a high-concentration Mg layer and a low-concentration Mg layer. Here, for the purpose of experiments, the Mg concentration was set to 5 × 10 ¹⁹ (cm ⁻³ ) over the entire layer. The layer thickness is B (nm).

(P-type cladding layer)

Materials; A1. . ₈ G a. ₉₂ N, Mg concentration; 2 × 10 ¹⁹ (cm- ³ ), layer thickness: 50 (nm).

(Light-emitting layer)

An MQW structure is used in which a GaN barrier layer and an In GaN well layer (emission wavelength: 405 nm) are stacked for six periods. The top layer was a GaN barrier layer. The layer thickness is 100 (nm). (N-type cladding layer)

Material: GaN, also used as n-type contact layer, layer thickness is 4 (m).

[Undoped GaN layer]

Material; GaN. The layer thickness is 2 μm), which is the thickness of the part above the top surface of the projection of the substrate.

[Crystal substrate, buffer layer]

The surface of the C-plane sapphire substrate was subjected to a stripe-shaped pattern and processed to have a rectangular wave-shaped cross section. A GaN buffer layer was grown on the top surface of the projection and the bottom surface of the depression in the irregularities.

Regarding the GaN-based LED having the above configuration, the layer thickness A of the p-type contact layer and the p-type Ga Device samples were manufactured in which the layer thickness B of the N layer was changed variously so that A + B = 1 O Onm at all times, and the emission output was measured at a current of 20 (mA). It was found that when the layer thickness A of the p-type contact layer exceeded 10 nm, the output decreased significantly. This is thought to be because the effect of light absorption by the p-type contact layer added with a high concentration of Mg (5 times the p-type cladding layer and 2 times the ρ-type GaN layer) increases. .

From this, it was found that the upper limit of the thickness of the p-type contact layer should be 10 nm.

The Vf (forward voltage required to pass a forward current of 2 OmA) when the thickness of the p-type contact layer was 10 nm was 3.5 V.

Experiment 2

From the results of Experiment 1, the thickness of the p-type contact layer was fixed at 1 O nm. For the p-type GaN layer immediately below, the layer thickness was 9 Onm, and the Mg concentration was 1 × 10 ¹⁹ (cm " ³ ) to further suppress light absorption by Mg.

Except for these changes, a GaN LED was manufactured with the same specifications as in Experiment 1, and the luminescence output was measured.The luminescence output was improved by about 40% compared to the result of Experiment 1, and 1S £ was about 0.3 V rose.

According to the studies by the present inventors, the cause is that Mg diffuses and flows out from the p-type contact layer to the p-type GaN layer immediately below, thereby lowering the Mg concentration in the surface layer of the p-type contact layer. This is presumed to be because the contact resistance with the p-side electrode increased. Since the p-type contact layer is as thin as 1 O nm, the total amount of Mg retained in this layer is small, and it is probable that the effect of the outflow of Mg appears remarkably.

Experiment 3

In order to improve the results of Experiment 2 above, and to keep V f from rising while keeping the light emission output high, the aspect of the Mg addition to the p-type GaN layer was changed. Specifically, the Mg concentration at the bottom of the p-type GaN layer (the region in contact with the p-type cladding layer) is 1 × 10 ¹⁹ cm ^3, and the Mg concentration at the top (the region in contact with the p-type contact layer) is 5 × 5. X And 10 ¹⁹ cm one ^3, Mg concentration was varied so that to increase substantially continuously towards the layer from a lower portion to an upper portion layer (gradient doped).

Except for this change, a GaN LED was manufactured with the same specifications as in Experiment 1, and the luminescence output was measured. As a result, both V f and luminescence output were at the same level as in Experiment 1.

According to the study by the present inventors, Vf decreased to the same level as in Experiment 1, and it is considered that the suppression of the diffusion outflow of Mg from the p-type contact layer was achieved. On the other hand, the light emission output dropped to the same level as in Experiment 1, suggesting that the Mg-gradient doping in Experiment 3 may not sufficiently suppress the light absorption caused by Mg.

Experiment 4

In order to improve the results of Experiment 2 above and increase the light emission output while keeping Vf low, the thickness X of the p-type GaN layer (layer thickness 90 nm) from the interface with the p-type contact layer was increased. (nm) a portion up Mg high concentration layer. (Mg concentration 5 X 10 ¹⁹ cm one ^3), and the remaining thickness Y (nm) (Y = 9 Ο-Χ) Mg low concentration layer portion (Mg conc. X and Y were variously changed at a degree of 1 × 10 ¹⁹ cm— ³ ).

Except for this change, a GaN LED was manufactured with the same specifications as in Experiment 2, and V f and emission output were measured.

The results of the measurement are shown in the graph of FIG. In the graph in this figure, the horizontal axis is the layer thickness X of the Mg-rich layer, and the values of V f at X = 0, 5, 10, 20, 90 (nm) are plotted with black circles, and the emission output value Is plotted with a white square.

As is evident from the graph of the figure, Vf was 3.5 V in the range of 5≤X≤90 (nm), which was the same as Experiment 1 described above, and was a preferable value. This indicates that the value of V f can be kept low when the thickness of the Mg-rich layer is 5 nm. Note that the case of X = 0 (nm) in Experiment 4 corresponds to Experiment 2 above.

In addition, a remarkable improvement in luminescence output was observed when X≤20 nm.

From Experiment 4, from the viewpoint of keeping V f low and further improving the light emission output, M It is understood that the thickness of the high concentration layer is preferably 5 to 20 nm.

In Experiment 4, there was no change in Vf even when the thickness Y of the Mg low concentration layer was varied over a wide range from 0 nm to 85 nm. A high-concentration Mg layer of 5 × 10 ¹⁹ cm ³ is provided, and the p-type layer below it has a Mg concentration of at least 1 × 10 ¹⁹ cm− ³ . This indicates that the diffusion runoff can be suppressed to a level that does not affect V f.

Further, when the design value of the Mg concentration of the p-type contact layer is set to a higher concentration, the problem of insufficient Mg concentration does not occur, so that similarly low V f can be obtained. In addition, when the design value is set to a lower concentration, the concentration difference between the p-type contact layer and the p-type high concentration layer becomes smaller, and the diffusion and outflow of Mg from the p-type contact layer are more effectively suppressed. Therefore, a similarly low V f is obtained.

Experiment 5

In addition to the results of Experiment 4, an experiment was conducted to further reduce the thickness of the p-type contact layer for the purpose of further reducing V f.

According to the results of Experiment 4 above, of the p-type GaN layer (layer thickness 90 nm), the upper side is a Mg high concentration layer (layer thickness 5 nm, Mg concentration 5 × 10 ¹⁹ cm- ³ ), and the remaining lower side is M g low concentration layer (thickness 85 nm, M g concentration 1 X 10 ¹⁹ cm- ³⁾ is fixed to, changes in the thickness of the p-type co Ntakuto layer T (nm) of the various thin range than 10 nm I let it. In this experiment 5, in forming the p-type contact layer, hydrogen gas and nitrogen gas were used as carrier gas used to supply the organometallic compound raw materials trimethy / regarium, trimethy / rhenoleminium, and biscyclopentageninolemagnesium to the reaction vessel. Except for using a mixed gas of, the only gas supplied to the reaction vessel was ammonia, which is a Group 3 raw material, and nitrogen gas.

The flow rate ratio of hydrogen gas to the hydrogen gas, ammonia, and nitrogen gas supplied to the reaction vessel was 30% or less. The growth rate of the p-type contact layer was about one-fifth of that in Experiments 1 to 4 in which the flow rate of hydrogen gas was about 60%. Except for this change, a GaN LED was manufactured with the same specifications as in Experiment 4, and V f and emission output were measured.

As a result of the measurement, from 10 to 1, the output remained at the same level, and a tendency for V f to decrease was observed. At T = 0.5, V f increased slightly, but still was about 0.1 leV lower than T = 0 (no p-type contact layer). From these results, it was found that it is preferable that the p-type contact layer has a higher A 1 ratio than the lower layer, and that the layer thickness is in a preferable range of 0.5 to 10 nm. The reason why V f showed a tendency to decrease from D = 10 to 1 was that the growth time of the p-type contact layer was shortened and the time during which the crystal was exposed to high temperatures was shortened, resulting in thermal degradation. It is presumed that the influence of was reduced.

Experiment 6

In addition to the results of Experiment 4, an experiment was conducted to determine the lower limit of the Mg concentration in the Mg-rich layer necessary to suppress the increase in V f.

In Experiment 4 above, the layer thickness X of the Mg-rich layer was fixed at 5 nm, and the layer thickness Y of the Mg-low layer was 8511 m. ^ A ^^ concentration 1§ high concentration layer with 5 X 10 ¹⁹ cm- ³ was varied between ¹ X 10 1 ⁹ cm one ^3. Except for this change, a GaN-based LED was manufactured with the same specifications as in Experiment 4, and V f was measured.

As a result, V f was at the same level as in Experiment 2 when the Mg concentration in the Mg-rich layer was between 1 × 10 ¹⁹ cm— ³ (corresponding to Experiment 2 above) and 2 × 10 ¹⁹ cm— ³ . Mg started to decrease with the concentration of more than 2 X 10 ¹⁹ cm one ^3, close to the 5 X 10 ¹⁹ cm- is a value at ³ 3. 5 V.

From this experiment 6, the thickness of the Mg-enriched layer is provided immediately below the p-type contact layer 5 nm or more, if the Mg concentration is 2 X 10 ¹⁹ cm one ³ above, the thickness of the p-type contact layer It has been found that there is an effect of suppressing an increase in operating voltage at the time.

Industrial applications

From the practical point of view, the light emitting element does not simply have to have a high output, but there is a demand for low power consumption from the device / apparatus side in which the element is incorporated. 9 Further, the operating voltage of a light emitting element is directly related to the amount of heat generated by the element, and thus greatly affects the life of the element. In addition, since there is a need for a mounting structure that gives priority to heat radiation as the heat generation increases, there is also a problem that various design restrictions are generated. In particular, in GaN-based semiconductor light-emitting devices, short-wavelength light is generated, so that the operating voltage must be increased in principle, and sapphire, which is currently the most suitable substrate for crystal growth, There is also a problem that the thermal conductivity of the material is extremely low.

Under these circumstances, extremely strong demands are placed on the operating voltage of GaN-based semiconductor light-emitting devices, such as the forward voltage (V f) in LEDs and the reduction in oscillation threshold voltage in laser diodes. Yes, it is desirable to reduce even the IV.

According to the present invention, the problem of light absorption can be improved by effectively reducing the total amount of Mg in the p-type layer while suppressing an increase in the operating voltage of the GaN-based semiconductor light emitting device. .

This application is based on a patent application No. 2004-134704 filed in Japan, the contents of which are incorporated in full herein.

Claims

The scope of the claims

1. It has a stacked body composed of a nitride semiconductor crystal layer. The stacked body includes an n-type layer, a light-emitting layer, and a p-type layer in order from the lower layer side, and the p-type layer is doped with Mg as a p-type impurity. A nitride semiconductor light emitting device,

At the bottom of the p-type layer is A 1 _X G a _X N (0≤ x≤ 1)! ) Type cladding layer is provided, M g high concentration layer made of the above the P-type clad layer via one or more hetero interface _{_{_{A 1 y G a x _ y}}} N (0≤y≤ 1) is provided p-type con tact layer made of the Mg height above linear density layer p-type layer a 1 _Z G a as the top of the (y <z≤ 1) is provided,

The p-type contact layer has a layer thickness of 10 nm or less, an Mg concentration a of 5 × 10 ¹⁹ cm ^-3 ≤ a,

Mg high concentration layer has a thickness of 5 nm ~ 20 nm, a M _g concentration b is 2 X 10 ¹⁹ cm one ³ <b,

of the p-type layer, a layer between the Mg-enriched layer and the light emitting layer, Mg concentration c is ^{^{^{1 X 10 1 9 cm- 3 c}}} <b,

wherein the average value of the Mg concentration in the p-type layer except for the p-type contact layer is less than 5 X 1 0 ¹⁹ cm_ ^3, the nitride semiconductor light emitting device.

2. At least high Mg concentration layer! 2. The nitride semiconductor light-emitting device according to claim 1, further comprising a portion having a Mg concentration of 5 × 10 ¹⁹ cm− ³ or more on a side in contact with the ()) type contact layer.

3. The nitride semiconductor light emitting device according to claim 2, wherein the high Mg concentration layer includes a portion having a Mg concentration of 5 × 10 ¹⁹ cm− ³ or more only on the side in contact with the p-type contact layer.

4. The nitride semiconductor phosphor device according to claim 1, wherein the Mg concentration c is c <5 × 10 ¹⁹ cm— ³ .

5. The nitride semiconductor light emitting device according to claim 1, wherein b <a.

6. Of the p-type layer, a layer between the Mg-enriched layer and the light emitting layer, Mg concentration include portions of less than 2 X 10 ¹ ⁹ cm- ^3, the nitride semiconductor light emitting device according to claim 1, wherein .

7. The nitride semiconductor light emitting device according to claim 6, wherein an average value of Mg concentration in a layer between the Mg high concentration layer and the light emitting layer is less than 2 × 10 ¹⁹ cm ⁻³ .

8. The nitride semiconductor light emitting device according to claim 1, wherein the layer thickness of the p-type layer is 100 nm to 300 nm.

9. The r> -type contact layer has a layer thickness of 0.5 to: L 0 nm, Mg concentration a is 5 × 10 ¹⁹ cm ^_3 ≤ a≤ 1 X 10 ²¹ cm— ³ ,

The Mg-rich layer has a layer thickness of 5 to 20 nm and a Mg concentration b of 2 × 10 ¹⁹ cm — ³ <(0.5 X a) ≤ b ≤ 1 × 10 ²¹ cm — ³ ,

2. The nitride semiconductor light-emitting device according to claim 1, wherein a layer between the Mg-rich layer and the light-emitting layer in the p-type layer has a Mg concentration c of (0.2 Xb) ≤ c <b.