TWI907573B - Nitride semiconductor ultraviolet light emitting element and its manufacturing method - Google Patents

Abstract

In an ultraviolet-emitting device of an AlGaN-based semiconductor with a zincblende structure, consisting of an n-type layer, an active layer, and a p-type layer, each semiconductor layer has an epitaxial growth layer on the surface of a multi-segmented platform parallel to the (0001) plane. The n-type layer has an extension direction where the local AlN molar fraction is low. The average AlN molar fraction is lower than that of the (n-) type layer in the inclined layered region that extends throughout the depth direction of the aforementioned n-type layer. 0.25)/12 is large, less than (n+0.25)/12. Above at least one specific depth where the average AlN molar fraction of the second depth d above the aforementioned n-type layer is more than 1/n/12, there exists a region larger than n/12. Below the aforementioned specific depth, there exists a region less than n/12. Within the layered regions, an intermediate AlGaN region with an AlN molar fraction of (n-0.5)/12 is formed.

Description

Nitride semiconductor ultraviolet light-emitting element and its manufacturing method

This invention relates to a nitride semiconductor ultraviolet emitting element formed by depositing an n-type layer, an active layer, and a p-type layer of an AlGaN-based semiconductor with a zincblende structure in a light-emitting element structure in the vertical direction, and a method for manufacturing the same.

Generally, most light-emitting devices with nitride semiconductors are constructed by epitaxial growth on substrates such as sapphire, forming multiple nitride semiconductor layers. A nitride semiconductor layer is represented by the general formula Al _1-xy Ga _x In _y N (0≦x≦1,0≦y≦1,0≦x+y≦1).

The light-emitting element structure of a light-emitting diode is a bi-heterogeneous structure that holds an active layer formed by the nitride semiconductor layer between two cladding layers: an n-type nitride semiconductor layer and a p-type nitride semiconductor layer. When the active layer is an AlGaN-based semiconductor, by adjusting the AlN molar fraction (also known as the Al composition ratio), the band gap energy is adjusted to a range where the band gap energies obtained by GaN and AlN (approximately 3.4 eV and approximately 6.2 eV, respectively) are the lower and upper limits, respectively, resulting in an ultraviolet light-emitting element with an emission wavelength of approximately 200 nm to approximately 365 nm. Specifically, by flowing a current in the same direction from the p-type nitride semiconductor layer to the n-type nitride semiconductor layer, light emission corresponding to the aforementioned band gap energy is generated in the active layer through the recombination of carriers (electrons and holes). Since this current is supplied from the outside, a p-electrode is provided on the p-type nitride semiconductor layer, and an n-electrode is provided on the n-type nitride semiconductor layer.

When the active layer is an AlGaN-based semiconductor, the n-type nitride semiconductor layer and the p-type nitride semiconductor layer containing the active layer are constructed using AlGaN-based semiconductors with a higher AlN molar fraction than the active layer. However, it is difficult to form a good ohmic contact with the p-electrode in the p-type nitride semiconductor layer with a high AlN molar fraction. Therefore, a p-type interconnect layer that can form a good ohmic contact with the p-electrode of a p-type AlGaN-based semiconductor (specifically p-GaN) with a low AlN molar fraction is generally formed on the top layer of the p-type nitride semiconductor layer. Because the molar fraction of AlN in this p-type interconnect layer is smaller than that of the AlGaN-based semiconductor constituting the active layer, ultraviolet light emitted from the active layer toward the p-type nitride semiconductor layer is absorbed in this p-type interconnect layer and cannot be effectively extracted to the outside of the device. Therefore, a typical ultraviolet-emitting diode system with an AlGaN-based semiconductor active layer adopts the schematic display device structure shown in Figure 18, which can effectively extract ultraviolet light emitted from the active layer toward the n-type nitride semiconductor layer to the outside of the device (see, for example, patents 1 and 2 below).

As shown in Figure 18, a typical ultraviolet-emitting diode (UV-LED) system is formed on a substrate 100 such as a sapphire substrate, on a template 102 formed by depositing an AlGaN-based semiconductor layer 101 (e.g., an AlN layer), and then sequentially depositing an n-type AlGaN-based semiconductor layer 103, an active layer 104, a p-type AlGaN-based semiconductor layer 105, and a p-type interconnect layer 106. The active layer 104, a portion of the p-type AlGaN semiconductor layer 105 and the p-type interconnect layer 106 are etched away until the n-type AlGaN semiconductor layer 103 is exposed. On the exposed surface of the n-type AlGaN semiconductor layer 103, n electrodes 107 are formed on the surface of the p-type interconnect layer 106, and p electrodes 108 are formed respectively.

Furthermore, in order to improve the luminescence efficiency (internal quantum efficiency) formed by carrier recombination within the active layer, an active layer is constructed as a multiple quantum well structure, and an electron barrier layer is set on the active layer.

On the other hand, reports indicate that compositional modulation caused by Ga segregation (segregation accompanied by Ga mass migration) within a coating layer composed of an n-type AlGaN semiconductor layer results in layered regions with low AlN molar fractions on the surface of the coating layer extending in the synchrotropic direction (see, for example, Patent 3, Non-Patent Documents 1 and 2 below). Because the band gap energy of the AlGaN semiconductor layer with locally low AlN molar fractions is also locally reduced, Patent 3 reports that carriers within this coating layer are easily localized in the layered regions, providing a low-impedance current path to the active layer and thus improving the luminous efficiency of the ultraviolet-emitting diode. [Previous Art Documents] [Patent Documents]

[Patent Document 1] International Publication No. 2014/178288 [Patent Document 2] International Publication No. 2016/157518 [Patent Document 3] International Publication No. 2019/159265 [Patent Document 4] Japanese Patent Application Publication No. 2012-089754 [Non-Patent Documents]

[Non-patent document 1] Y. Nagasawa, et al., “Comparison of Al _x Ga _1−x N multiple quantum wells designed for 265 and 285nm deep-ultraviolet LEDs grown on AlN templates having macrosteps”, Applied Physics Express 12, 064009 (2019) [Non-patent document 2] K. Kojima, et al., “Carrier localization structure combined with current micropaths in AlGaN quantum wells grown on an AlN template with macrosteps", Applied Physics letter 114, 011102 (2019) [Non-patent document 3] Shigeaki Sumiya, et al., "AlGaN-Based Deep Ultraviolet Light-Emitting Diodes Grown on Epitaxial AlN/Sapphire Templates", Journal of Applied Physics, Vol. 47, NO. 1, 2008, pp.43-46

[The problem the invention aims to solve]

Ultraviolet light-emitting devices (UV emitting devices) made of AlGaN semiconductors are fabricated on substrates such as sapphire substrates using well-known epitaxial growth methods, such as metal vapor phase growth (MOVPE). However, during the production of UV emitting devices, the characteristics of the UV emitting devices (emission wavelength, socket efficiency, forward bias, etc.) vary due to the drift of the crystal growth apparatus, making it difficult to produce them at a stable yield.

Drift in crystal growth equipment is caused by deposits on the support plate or the walls of the processing chamber, and is generated by changes in the effective temperature of the crystal growth area. To suppress drift, the growth history has been reviewed in the past. Although experienced operators have made efforts such as subtly changing the set temperature or the composition of the raw material gas, or fixing the growth process for a certain period of time, and performing cleaning and other maintenance at the same time, it is still difficult to completely eliminate drift.

Here, the inventors of this invention have discovered that within the layered region of an n-type AlGaN semiconductor layer, formed by Ga segregation, a localized AlN molar fraction with low impedance current path for the active layer is formed. By having "quasi-stable AlGaN" (described later) with an integer AlGaN composition ratio dominate the n-type AlGaN semiconductor layer, characteristic variations caused by crystal growth device drift are suppressed, enabling the stable production of nitride semiconductor ultraviolet-emitting devices with desired light-emitting characteristics. This is proposed within the layered region of an n-type AlGaN semiconductor layer, utilizing quasi-stable AlGaN (see PCT/JP2020/024827, PCT/JP2020/). (Explanation of international applications such as 024828, PCT/JP2020/026558, PCT/JP2020/031620, etc.)

<Characteristics of Quasi-Stable AlGaN> For ease of explanation, we will first explain "quasi-stable AlGaN" represented by a specific integer ratio of AlGaN components.

When quasi-stable AlGaN is not considered, ternary mixed crystals such as AlGaN are crystal states of random mixing of group 3 elements (Al and Ga), which can be described as "random configuration". However, because the common bonding radius of Al and Ga is different, in the crystal structure, the one with higher symmetry in the atomic arrangement of Al and Ga generally becomes a stable structure.

AlGaN-based semiconductors with a zincblende structure exhibit two types of arrangements: random arrangements with asymmetry and stable symmetrical arrangements. Here, symmetrical arrangements become dominant at certain ratios. As will be discussed later, in "quasi-stable AlGaN" where the Al composition ratio (the ratio of Al to Ga to N) is expressed as a specific integer ratio, a symmetrical arrangement of Al and Ga is found.

In this symmetrical arrangement, the Ga supply to the crystal growth plane increases or decreases only slightly. Due to its high symmetry, it becomes an energy-stable mixed crystal molar fraction, which can prevent the concentration of Ga, which is prone to mass transfer, from being out of control.

<Symmetrical Al and Ga Arrangement Structure of Quasi-Stable AlGaN> Next, in the section on "Quasi-Stable AlGaN," the discovery of the symmetrical Al and Ga arrangement structure will be explained.

Figure 1 shows a schematic diagram of a single unit cell (2 monolayers) of AlGaN along the c-axis. In Figure 1, white circles indicate the positions of atoms of group 3 elements (Al, Ga), and black circles indicate the positions of atoms of group 5 elements (N). In the following description, a monolayer is denoted as ML. In Figure 1, a single unit cell is denoted as 2ML.

In Figure 1, the positional faces of the three groups of elements (A3 face, B3 face) and the five groups of elements (A5 face, B5 face) shown by the hexagon are all parallel to the (0001) face. Of the positions on faces A3 and A5 (collectively referred to as faces A), there are six positions at each vertex of the hexagon and one position at the center of the hexagon. The same applies to faces B3 and B5 (collectively referred to as faces B). In Figure 1, only three positions within the hexagon on face B are shown. The positions on face A overlap along the c-axis, and the positions on face B overlap along the c-axis. However, the atom (N) at one position on the B5 face forms a 4-coordinate bond with the atom (Al, Ga) at three positions on the A3 face above the B5 face and the atom (Al, Ga) at one position on the B3 face below the B5 face. The atom (Al, Ga) at one position on the B3 face forms a 4-coordinate bond with the atom (N) at one position on the B5 face above the B3 face and the atom (N) at three positions on the A5 face below the B3 face. Therefore, as shown in Figure 1, the positions on the A face do not overlap with the positions on the B face in the c-axis direction.

Figure 2 illustrates the positional relationship between the positions on plane A and plane B, showing the planes A3 and B3 as viewed from the c-axis. The black and white circles in Figure 2 distinguish the positions of plane A3 and plane B3, respectively. On planes A3 and B3, the six vertices of each hexagon are shared by two adjacent hexagons, while the center position is not shared with any of the other hexagons. Therefore, within each hexagon, there are substantially three atomic positions. Thus, in each unit cell (2ML), there are six positions for atoms of group 3 elements (Al, Ga) and six positions for atoms of group 5 elements (N). Therefore, for the AlGaN composition ratio, excluding GaN and AlN expressed as an integer ratio, there are the following five cases. 1)Al ₁ Ga ₅ N ₆ , 2)Al ₂ Ga ₄ N ₆ (=Al ₁ Ga ₂ N ₃ ), 3)Al ₃ Ga ₃ N ₆ (=Al ₁ Ga ₁ N ₂ ), 4)Al ₄ Ga ₂ N ₆ (=Al ₂ Ga ₁ N ₃ ), 5)Al ₅ Ga ₁ N ₆ .

Here, the Al ₁ Ga ₂ N ₃ , Al ₁ Ga ₁ N ₂ , and Al ₂ Ga ₁ N ₃ in 2) to 4) above are as shown in Figure 3. Because the A3 and B3 planes have identical symmetrical Al and Ga arrangements, the AlGaN composition ratio of 2) to 4) above is formed in 1 mL units along the c-axis. Figure 3 illustrates the arrangement on only one side of the A3 and B3 planes. However, in Figure 3, Ga is shown as a large black circle, and Al as a small black circle.

On the other hand, let Al ₁ Ga ₅ N ₆ in 1) above be a symmetrical arrangement of Al ₁ Ga ₂ N ₃ in 2) above, with one of the A3 and B3 faces being an arrangement of GaN (all positions of group 3 elements being Ga). Furthermore, let Al ₁ Ga ₅ N ₆ in 5) above be a symmetrical arrangement of Al ₂ Ga ₁ N ₃ in 4) above, with one of the A3 and B3 faces being an arrangement of AlN (all positions of group 3 elements being Al).

Furthermore, as four quasi-stable AlGaNs with an integer composition ratio located between each of the above 1) and 2), 2) and 3), 3) and 4), and 4) and 5), the following are envisioned as 6) to 9): 6) Al ₃ Ga ₉ N ₁₂ (= Al ₁ Ga ₃ N ₄ ), 7) Al ₅ Ga ₇ N ₁₂ , 8) Al ₇ Ga ₅ N ₁₂ , 9) Al ₉ Ga ₃ N ₁₂ (= Al ₃ Ga ₁ N ₄ ).

Here, Al ₁ Ga ₃ N ₄ in 6) above is a symmetrical arrangement of 2ML units where one of the A3 and B3 faces is arranged as Al ₁ Ga ₁ N ₂ in 3) above, and the other is arranged as GaN (all elements of group 3 are Ga). Al ₅ Ga ₇ N ₁₂ in 7) above is a symmetrical arrangement of 2ML units where one of the A3 and B3 faces is arranged as Al ₁ Ga ₂ N ₃ in 2) above, and the other is arranged as Al ₁ Ga ₁ N ₂ in 3) above. Al ₇ Ga ₅ N ₁₂ in 8) above is a symmetrical arrangement of 2ML units where one of the A3 and B3 faces is arranged as Al ₁ Ga ₁ N ₂ in 3) above, and the other is arranged as Al ₂ Ga ₁ N ₃ in 4) above. The Al ₃ Ga ₁ N ₄ in 9) above is a symmetrical arrangement of 2ML units where one of the A3 and B3 faces is arranged as Al ₁ Ga ₁ N ₂ in 3) above, and the other is arranged as AlN (all three group elements are Al).

Therefore, the Al1Ga5N6, _Al5Ga1N6 , _{Al3Ga9N12 ( =Al1Ga3N4 )} , _Al5Ga7N12 , Al7Ga5N12, or _Al9Ga3N12 (= _Al3Ga1N4 ) mentioned above in ₁ ), ₅ ) to ₉ ) are arranged in a symmetrical configuration with different _2ML units on the _{A3 and B3} planes _, as described above. The _AlGaN composition ratio of the quasi-stable _{AlGaN in 1} ), _{5 ) to 9} ) mentioned above is formed in 2ML units in the c-axis direction.

However, although specific examples are not given for the AlGaN composition ratios in 1), 5) to 9) above, for instance, by synthesizing different symmetrical arrangements on the A3 and B3 planes within the same plane, similar to Al ₁ Ga ₂ N 3, Al 1 Ga 1 N ₂ , and Al ₂ Ga ₁ N ₃ in 2) to 4) above, a symmetrical arrangement of Al _and Ga with identical _A3 and _B3 planes is obtained. In this case, the quasi-stable AlGaN with AlGaN composition ratios in 1), 5) to 9) above is the same as the quasi-stable AlGaN with AlGaN composition ratios in 2) to 4) above, and can be formed in 1 mL units along the c-axis direction.

Based on the above, the quasi-stable AlGaN shown in 1)~9) is an energy-stable AlGaN with Al and Ga atoms arranged in a symmetrical configuration. However, to maintain a certain crystallinity in AlGaN, crystal growth at temperatures above 1000℃ is required. However, Ga atoms, located on the crystal surface, tend to move randomly at temperatures above 1000℃. On the other hand, Al, unlike Ga, readily adsorbs onto the surface, and its movement after entering its position, while somewhat limited, is strongly restricted. Therefore, even for quasi-stable AlGaN, the _Al1Ga5N6 composition in 1) above is high because of _the high Ga composition _ratio . At growth temperatures around 1000℃, Ga moves more vigorously, and the symmetry of atomic arrangement is disordered. The atomic arrangement of Al and Ga is close to random, and the stability mentioned above is lower than that of other quasi-stable AlGaN.

Here, the quasi-stable AlGaN with AlGaN composition ratios of 1) to 9) mentioned above can be represented by an integer n (n=2~10) as Al _n Ga _12-n N _12. While the AlN mole fraction of Al _n Ga _12-n N ₁₂ is expressed as n/12 when expressed as a fraction, it will have decimal places when expressed as a percentage. Therefore, for ease of explanation, the following six AlN mole fractions, expressed as 2/12 (=1/6), 4/12 (=1/3), 5/12, 7/12, 8/12 (=2/3), and 10/12 (=5/6), are approximately represented as 16.7%, 33.3%, 41.7%, 58.3%, 66.7%, and 83.3%.

<Problems with the Use of Quasi-Stable AlGaN in Layered Regions> As described above, the inventors of this invention propose using quasi-stable AlGaN in the layered regions of an n-type AlGaN semiconductor layer to suppress characteristic variations caused by drift and other factors in the crystal growth apparatus. Furthermore, the problems caused by the dominant formation of quasi-stable AlGaN in the layered regions are explained.

In a general ultraviolet-emitting diode (UVD) that takes ultraviolet light emitted from the active layer toward the n-type nitride semiconductor layer (n-type AlGaN semiconductor layer) and takes it out to the outside of the device, as long as the wavelength difference (λp-λae) between the peak emission wavelength (λp) of the light emitted from the active layer and the absorption end wavelength (λae) determined by the average AlN molar fraction of the absorption spectrum of the n-type AlGaN semiconductor layer is 10 nm or more, the absorption within the n-type AlGaN semiconductor layer emitted from the active layer can be suppressed.

The light emission from the active layer causes the transmittance (the ratio of incident light intensity I0 to transmitted light intensity I, I/I0) of the n-type AlGaN semiconductor layer to decrease exponentially with respect to the thickness of the n-type AlGaN semiconductor layer along the optical path. The thickness of the n-type AlGaN semiconductor layer can be as large as 1~4μm, while the thickness of each layer in the depth direction of the layered region is on average as short as 20nm. Therefore, since the AlN molar fraction in the layered region is about 4% lower than the average AlN molar fraction in the n-type AlGaN semiconductor layer, although the absorption wavelength of the layered region is only shifted to a longer wavelength side compared to the overall absorption wavelength λae of the n-type AlGaN semiconductor layer, the optical density (the commonly used logarithm of the reciprocal of transmittance) of the layered region is smaller than that of the n-type AlGaN semiconductor layer as a whole. Therefore, the absorption on the longer wavelength side than the absorption wavelength λae is extremely limited.

To achieve dominant formation of quasi-stable AlGaN within the layered region, the average AlN molar fraction of the n-type AlGaN semiconductor layer is set to be approximately 4% higher than that of the quasi-stable AlGaN. On the other hand, since the AlGaN composition ratio in the quasi-stable AlGaN is an integer ratio, this AlN molar fraction achieves a dispersion value of approximately 8.33%. Therefore, the range of the average AlN molar fraction of the n-type AlGaN semiconductor layer is also set to a dispersion range approximately 4% higher than that of the quasi-stable AlGaN.

In an n-type AlGaN semiconductor layer where quasi-stable AlGaN is dominantly formed within a layered region, if the absorption wavelength λae determined by the average AlN molar fraction of the n-type AlGaN semiconductor layer cannot guarantee a wavelength difference (λp-λae) of more than 10 nm for the peak emission wavelength λp, then a portion of the emission spectrum (especially the portion distributed on the short-wavelength side) is absorbed within the n-type AlGaN semiconductor layer, leading to a decrease in external quantum efficiency. Therefore, in order to avoid this decrease in external quantum efficiency, it is necessary to increase the discrete AlN moiré fraction of quasi-stable AlGaN step by step (approximately 8.33%) so that the wavelength difference (λp-λae) becomes above 10nm, and the average AlN moiré fraction of the n-type AlGaN semiconductor layer is also set at a high level in stages.

On the other hand, when the average AlN moiré fraction of the n-type AlGaN semiconductor layer increases, the contact impedance between the n-electrode and the n-type AlGaN semiconductor layer, and the bulk resistivity for the current flowing in the n-type AlGaN semiconductor layer, increase. Consequently, the parasitic impedance of the current path between the n-electrode and the active layer increases, and the socket efficiency decreases. In particular, the contact resistance between the n-type AlGaN semiconductor layer and the n-electrode (e.g., a Ti/Al/Ti/Au layer structure: Ti at the bottom and Au at the top) tends to increase as the AlN molar fraction of the n-type AlGaN semiconductor layer increases, especially when it exceeds 60%. Therefore, it is preferable to set the average AlN molar fraction to no more than 60% (see Non-Patent Reference 3). When the AlN molar fraction of the n-type AlGaN semiconductor layer is below 60%, the contact resistance can be adjusted to below 0.01 ^Ωcm² by appropriately selecting the heat treatment temperature, so that the forward voltage Vf becomes practically problem-free (see Patent Reference 4).

Therefore, to avoid a decrease in external quantum efficiency, when the aforementioned wavelength difference (λp-λae) exceeds 10 nm and becomes excessively large, the contact impedance and bulk resistivity increase, leading to a decrease in socket efficiency. To address this, for the target peak emission wavelength λp, to avoid a decrease in external quantum efficiency, the average AlN molar fraction is increased by one step (approximately 8.33%). Since the wavelength difference (λp-λae) changes from less than 10 nm to more than 10 nm, and the socket efficiency decreases, it is preferable to keep this wavelength difference no more than 10 nm. This prevents the dominant formation of quasi-stable AlGaN within the layered region. Thus, corresponding to the target peak emission wavelength λp, the decrease in external quantum efficiency can be avoided while simultaneously suppressing a decrease in socket efficiency.

This invention addresses the aforementioned problems by forcefully suppressing the formation of quasi-stable AlGaN within layered regions of locally low AlN molar fraction in n-type AlGaN semiconductor layers, thereby preventing a decrease in luminous efficiency caused by the dominant formation of quasi-stable AlGaN within these layered regions. [Means for solving the problem]

This invention, to achieve the aforementioned objective, provides a nitride semiconductor ultraviolet-emitting element formed by stacking an n-type layer, an active layer, and a p-type layer of an AlGaN-based semiconductor with a zincblende structure in a vertically oriented emitting element structure. The aforementioned n-type layer is composed of an n-type AlGaN-based semiconductor. The aforementioned active layer, disposed between the aforementioned n-type layer and the aforementioned p-type layer, has a quantum well structure comprising one or more well layers composed of AlGaN-based semiconductors. The aforementioned p-type layer is composed of a p-type AlGaN-based semiconductor. Each semiconductor layer within the aforementioned n-type layer, the aforementioned active layer, and the aforementioned p-type layer has an epitaxial growth layer forming a multi-segmented platform surface parallel to the (0001) plane. The aforementioned n-type layer comprises layered regions with locally low AlN molar fractions dispersed within it. Each extension direction of these layered regions on a first plane orthogonal to the upper surface of the aforementioned n-type layer has a portion inclined with respect to the intersection of the upper surface of the aforementioned n-type layer and the first plane. The integer n is 6 or 7. The first average AlN molar fraction Xna1, covering the entire depth direction of the aforementioned n-type layer, falls within the range of (n-0.25)/12 < Xna1 < (n+0.25)/12. The second average AlN molar fraction Xna2(d) at depth d above the aforementioned n-type layer varies corresponding to the aforementioned depth d. Within the aforementioned n-type layer, at least one specific depth having a depth of 1 or more where Xna2(d) = n/12, along with the aforementioned specific depth, there exists a region on the upper side where Xna2(d) < n/12, and a region on the lower side where Xna2(d) > n/12. Within the aforementioned layered regions, a nitride semiconductor ultraviolet-emitting element characterized by an AlGaN region with an AlN molar fraction of (n-0.5)/12 is formed.

Furthermore, it is preferable that the nitride semiconductor ultraviolet emitting element of the first characteristic described above has an average AlN molar fraction Xna3 that is within the range of (n-0.25)/12 < Xna3 < (n+0.25)/12, extending from the upper end of the aforementioned n-type layer to the aforementioned specific depth.

Furthermore, the nitride semiconductor ultraviolet emitting element with the first characteristic mentioned above, and the average AlN molar fraction Xna2(d) mentioned in the second point, preferably falls within the range of (n-0.25)/12 < Xna2(d) < (n+0.25)/12 throughout the depth direction of the aforementioned n-type layer.

This invention aims to achieve the aforementioned objectives by providing a method for manufacturing a nitride semiconductor ultraviolet-emitting element. This method involves epitaxially growing an n-type AlGaN semiconductor, comprising an n-type layer, an active layer, and a p-type layer, formed by stacking these layers in a vertically oriented light-emitting element structure. Specifically, regarding the (0001) plane, on a sapphire substrate including a main surface tilted at a specific angle, the aforementioned n-type layer of the n-type AlGaN semiconductor is epitaxially grown. On the surface of the aforementioned n-type layer, a first process is described, exhibiting a multi-segmented platform parallel to the (0001) plane. Above the aforementioned n-type layer, a quantum well structure containing one or more well layers composed of AlGaN-based semiconductors is epitaxially grown to form the aforementioned active layer. A second process is performed on the surface of the aforementioned well layer, exhibiting multi-segmented platforms parallel to the (0001) plane. Above the aforementioned active layer, a third process is performed, forming the aforementioned p-type layer of p-type AlGaN-based semiconductors through epitaxial growth. In the aforementioned first process, the integer n is 6 or 7, the first average AlN molar fraction Xna1, spanning the entire depth direction of the aforementioned n-type layer, is within the range of (n-0.25)/12 < Xna1 < (n+0.25)/12 The second average AlN molar fraction Xna2(d) at depth d above the aforementioned n-type layer varies corresponding to the aforementioned depth d. Within the aforementioned n-type layer, at least one specific depth having a depth of 1 or more where Xna2(d) = n/12, along with the aforementioned specific depth, there exists a region on the upper side where Xna2(d) < n/12, and a region on the lower side where Xna2(d) > n/12. Layered regions with locally low AlN molar fractions, similarly dispersed within the aforementioned n-type layer, extend obliquely upwards to form, forming an intermediate AlGaN region with an AlN molar fraction of (n-0.5)/12 within the aforementioned layered regions. A method for manufacturing a nitride semiconductor ultraviolet-emitting element characterized by forming the aforementioned n-type layer.

Furthermore, in the manufacturing method of the nitride semiconductor ultraviolet emitting element with the first characteristic described above, it is preferable that, in the aforementioned first process, the average AlN molar fraction Xna3, covering the region from the upper end of the aforementioned n-type layer to the aforementioned specific depth, is within the range of (n-0.25)/12 < Xna3 < (n+0.25)/12, to form the aforementioned n-type layer.

Furthermore, in the manufacturing method of the nitride semiconductor ultraviolet emitting element with the first characteristic mentioned above, it is preferable that the average AlN molar fraction Xna2(d) in the second step of the aforementioned process is within the range of (n-0.25)/12 < Xna2(d) < (n+0.25)/12 across the entire depth direction of the aforementioned n-type layer, and that the aforementioned n-type layer is formed accordingly.

However, although AlGaN-based semiconductors are represented by the general formula Al _1-x Ga _x N (0≦x≦1), within the range where the band gap energy of GaN and AlN respectively becomes the lower and upper limits, they can contain trace amounts of impurities such as group 3 elements like B or In, or group 5 elements like P. Furthermore, although GaN-based semiconductors are essentially nitride semiconductors composed of Ga and N, they can also contain trace amounts of impurities such as group 3 elements like Al, B, or In, or group 5 elements like P. Similarly, although AlN-based semiconductors are essentially nitride semiconductors composed of Al and N, they can also contain trace amounts of impurities such as group 3 elements like Ga, B, or In, or group 5 elements like P. Therefore, in this invention, GaN-based semiconductors and AlN-based semiconductors are each a part of AlGaN-based semiconductors.

Furthermore, n-type or p-type AlGaN semiconductors are AlGaN semiconductors doped with Si or Mg as donor or acceptor impurities. In this invention, the terms p-type and n-type AlGaN semiconductors do not necessarily mean undoped AlGaN semiconductors, but even if undoped, they may contain trace amounts of donor or acceptor impurities to an unavoidable degree. Also, the first plane is not the interface between the actual exposed surface or other semiconductor layers formed during the manufacturing process of the aforementioned n-type layer, but an imaginary plane extending parallel to the vertical direction within the aforementioned n-type layer. Furthermore, in this specification, the AlGaN-based semiconductor layer, the GaN-based semiconductor layer, and the AlN-based semiconductor layer are semiconductor layers respectively constructed from AlGaN-based semiconductors, GaN-based semiconductors, and AlN-based semiconductors.

In the aforementioned nitride semiconductor ultraviolet emitting element with the first characteristic, the first average AlN molar fraction Xna1 is centered on n/12 of the discrete AlN molar fraction of quasi-stable AlGaN with an integer n of 6 or 7, and controlled within the range of ±0.25/12 (± approximately 2.08%). This first quasi-stable AlGaN region (Al _n Ga _{12 - n} N ₁₂ ) with an AlN molar fraction of n/12 is stably formed in the region outside the layered region within the n-type layer (hereinafter appropriately referred to as the "n-type body region"). Furthermore, the second average AlN molar fraction Xna2(d) also follows n/12 and is distributed in the vicinity. As a result, within the layered regions, the second quasi-stable AlGaN region (Al _n-1 Ga ₁₃ -n N 12) with an AlN molar fraction one stage smaller than n/12 (approximately 8.33%) (n- ₁ )/12 is not formed. Instead, an intermediate AlGaN region (Al _n-0.5 Ga _12.5-n N ₁₂ ) is formed between the first and second quasi-stable AlGaN regions.

Therefore, in the ultraviolet light-emitting element of a quasi-stable AlGaN nitride semiconductor within the n-type layer, when the quasi-stable AlGaN is dominantly formed in the layered region, the target peak emission wavelength λp becomes a value of approximately 8.33% of the AlN molar fraction. In a specific wavelength range where it is difficult to simultaneously avoid the decrease in external quantum efficiency and the decrease in socket efficiency, by forming an intermediate AlGaN region in the layered region and forming a first quasi-stable AlGaN region in the n-type body region, the decrease in external quantum efficiency can be avoided while the decrease in socket efficiency can be suppressed.

Furthermore, in the manufacturing method of the nitride semiconductor ultraviolet light-emitting element with the first feature described above, the decrease in luminous efficiency caused by the dominant formation of quasi-stable AlGaN in the layered region can be prevented in the manufacturing of the nitride semiconductor ultraviolet light-emitting element with the first feature described above.

Furthermore, in addition to the first feature described above, this invention provides a nitride semiconductor ultraviolet-emitting element with the second feature, wherein the average AlN molar fraction Xna2(d) is within the range of Xna2(d) ≦ n/12 from the upper end of the aforementioned n-type layer to the aforementioned specific depth.

Furthermore, in addition to the first feature described above, this invention provides a method for manufacturing a nitride semiconductor ultraviolet-emitting element with the aforementioned second feature, wherein the average AlN molar fraction Xna2(d) in the aforementioned second feature is within the range of Xna2(d) ≦ n/12 from the upper end of the aforementioned n-type layer to the aforementioned specific depth.

When manufacturing a nitride semiconductor ultraviolet light-emitting element according to the second feature described above or the manufacturing method of the nitride semiconductor ultraviolet light-emitting element according to the second feature described above, the contact impedance between the n-electrode and the n-type layer and the bulk resistivity for the current flowing through the upper part of the n-type layer can be further reduced, thus avoiding the decrease in external quantum efficiency and suppressing the decrease in socket efficiency.

Furthermore, the nitride semiconductor ultraviolet emitting element with the second characteristic described above preferably has an average AlN molar fraction Xna2(d) that is within the range of Xna2(d) ≥ n/12 in a region deeper than the previously mentioned specific depth of the n-type layer.

Furthermore, in the manufacturing method of the nitride semiconductor ultraviolet emitting element with the second characteristic mentioned above, in the first process described above, it is preferable that the average AlN molar fraction Xna2(d) of the second characteristic is within the range of Xna2(d) ≥ n/12 in the depth direction of the n-type layer, and that the n-type layer is formed accordingly.

According to any appropriate implementation form mentioned above, in the region deeper than the specific depth of the n-type layer, the absorption of light emitted from the active layer can be suppressed even more, and the decrease in external quantum efficiency can be suppressed even more.

Furthermore, it is preferable that the peak emission wavelength of the nitride semiconductor ultraviolet light-emitting element with the first or second feature described above is set to a specific value within the range of 280nm to 315nm when the aforementioned integer n is 7, and to a specific value within the range of 300nm to 330nm when the aforementioned integer n is 6.

Furthermore, the manufacturing method of the aforementioned nitride semiconductor ultraviolet emitting element is carried out in the second step described above. When the aforementioned nitride semiconductor ultraviolet emitting element has a peak emission wavelength of 280nm to 315nm (when the integer n is 7), and 300nm to 330nm (when the integer n is 6), it is preferable to form the aforementioned active layer.

According to any appropriate embodiment described above, when the integer n is 7, a specific peak emission wavelength is set within the range of 280nm to 315nm; when the integer n is 6, a specific peak emission wavelength is set within the range of 300nm to 330nm. These measures can respectively prevent a decrease in luminous efficiency caused by the dominant formation of quasi-stable AlGaN within the layered region. However, the specific peak emission wavelength is within the aforementioned specific wavelength range, and the peak emission wavelength that would cause a decrease in luminous efficiency (external quantum efficiency or socket efficiency) by the dominant formation of quasi-stable AlGaN within the layered region is a peak emission wavelength within the aforementioned specific wavelength range.

Furthermore, the nitride semiconductor ultraviolet emitting element with the first or second feature mentioned above has a multi-quantum well structure in which the active layer has two or more well layers as described above, and it is preferable that a barrier layer made of AlGaN semiconductor exists between the two well layers.

Furthermore, the manufacturing method of the nitride semiconductor ultraviolet light-emitting element with the first or second feature mentioned above is preferably carried out in the second process described above by stacking the aforementioned well layer composed of AlGaN semiconductor and the barrier layer composed of AlGaN semiconductor through alternating epitaxial growth, so that the aforementioned well layer forms the aforementioned active layer containing a multi-quantum well structure of two or more layers.

According to any of the above-mentioned appropriate implementation forms, the active layer becomes a multiple quantum well structure, and the luminescence efficiency can be expected to be improved compared to when there is only one well layer.

Furthermore, the nitride semiconductor ultraviolet light-emitting element with the first or second feature mentioned above is further provided with a substrate portion including a sapphire substrate. The sapphire substrate has a main surface that is tilted at a specific angle relative to the (0001) surface. The light-emitting element structure portion is formed above the main surface. It is preferable that each semiconductor layer from the aforementioned main surface of the sapphire substrate to the aforementioned p-type layer has an epitaxial growth layer with a surface that forms a multi-segmented platform parallel to the (0001) surface.

According to the above-described appropriate embodiment, epitaxial growth can be performed on the surfaces of each layer from the main surface of the sapphire substrate to the p-type layer, exhibiting multi-segmented plateaus, thereby realizing a nitride semiconductor ultraviolet-emitting device with the aforementioned characteristics. [Invention Effects]

Furthermore, in the manufacturing method of the nitride semiconductor ultraviolet emitting element or the nitride semiconductor ultraviolet emitting element according to the first or second feature mentioned above, the formation of quasi-stable AlGaN in the layered region with low local AlN molar fraction within the n-type AlGaN semiconductor layer is strictly suppressed at a specific peak emission wavelength. This prevents the decrease in luminous efficiency caused by the dominant formation of quasi-stable AlGaN in the layered region. Moreover, when the target peak emission wavelength is set within a specific wavelength range, the embodiment utilizing quasi-stable AlGaN and the embodiment not utilizing it are used separately in the layered region of the n-type layer, corresponding to the target peak emission wavelength. This increases the freedom of setting the peak emission wavelength while preventing a decrease in luminous efficiency.

The nitride semiconductor ultraviolet light-emitting element (hereinafter simply referred to as "light-emitting element") of the present invention will be described with reference to the figures. However, in the schematic diagrams used in the following description, for ease of understanding and to emphasize the main parts, the dimensions of each part may not be the same as those of the actual element. In the following description, the light-emitting element is assumed to be a light-emitting diode.

[First Embodiment] <Element Structure of Light Emitting Element> As shown in FIG4, the light emitting element 1 of the first embodiment has a substrate portion 10 including a sapphire substrate 11, and a plurality of AlGaN-based semiconductor layers 21-24, and a light emitting element structure portion 20 including a p-electrode 26 and an n-electrode 27. The light emitting element 1 is mounted (flip-chip mounting) with the light emitting element structure portion 20 side (upper side in FIG4) facing the mounting base (sub-fixed base, etc.), and the light extraction direction is the substrate portion 10 side (lower side in FIG4). However, for ease of explanation in this specification, the direction perpendicular to the main surface 11a of the sapphire substrate 11 (or the top surface of the substrate portion 10 and each AlGaN-based semiconductor layer 21-24) is referred to as the "vertical direction" (or "longitudinal direction"). The direction from the substrate portion 10 towards the light-emitting element structure portion 20 is called the upward direction, and the opposite direction is called the downward direction. Furthermore, the plane parallel to the vertical direction is referred to as the "first plane." Moreover, the plane parallel to the main surface 11a of the sapphire substrate 11 (or the top surface of the substrate portion 10 and each AlGaN-based semiconductor layer 21-24) is referred to as the "second plane," and the direction parallel to this second plane is referred to as the "horizontal direction."

The substrate 10 is composed of a sapphire substrate 11 and an AlN layer 12 formed directly on the main surface 11a of the sapphire substrate 11. The sapphire substrate 11 is a slightly tilted substrate in which the main surface 11a is tilted relative to the (0001) surface at an angle (offset) within a certain range (e.g., 0.3° to 6°), and multi-segmented platforms are exhibited on the main surface 11a.

The AlN layer 12 is formed by epitaxially growing AlN crystals from the main surface 11a of the sapphire substrate 11. These AlN crystals have an epitaxial crystal orientation relative to the main surface 11a of the sapphire substrate 11. Specifically, for example, AlN crystals are grown so that the C-axis direction (<0001> direction) of the sapphire substrate 11 is aligned with the C-axis direction of the AlN crystals. However, the AlN crystals constituting the AlN layer 12 may contain trace amounts of Ga or other impurities, and may also be an AlN-based semiconductor layer. In this embodiment, the film thickness of the AlN layer 12 is assumed to be around 2 μm to 3 μm. However, the structure of the substrate portion 10 and the substrate used are not limited to the above configuration. For example, between AlN layer 12 and AlGaN semiconductor layer 21, there may be an AlGaN semiconductor layer with an AlN molar fraction greater than that of AlGaN semiconductor layer 21.

The AlGaN semiconductor layers 21-24 of the light-emitting element structure 20 are constructed by epitaxial growth of the following layers sequentially from the substrate 10: n-type cladding layer 21 (n-type layer), active layer 22, electron barrier layer 23 (p-type layer), and p-type interconnect layer 24 (p-type layer).

In this embodiment, the AlN layer 12 of the substrate portion 10, which is epitaxially grown sequentially from the main surface 11a of the sapphire substrate 11, and the semiconductor layers and electron barrier layers 23 within the n-type cladding layer 21 and active layer 22 of the light-emitting element structure portion 20, have a surface with a multi-segmented platform formed parallel to the (0001) plane on the main surface 11a of the sapphire substrate 11 by step flow growth. However, for the p-type interconnect layer 24 of the p-type layer, since it is formed by epitaxial growth on the electron barrier layer 23, although the same multi-segmented platform can be formed, it may not have a surface with the same multi-segmented platform.

However, as shown in Figure 4, the portion of the active layer 22, electron barrier layer 23, and p-type interconnect layer 24 that are deposited on the second region R2 above the n-type cladding layer 21 within the light-emitting element structure 20 is removed by etching or the like, and formed on the first region R1 above the n-type cladding layer 21. Then, the top of the n-type cladding layer 21 is exposed in the second region R2, excluding the first region R1. As shown schematically in Figure 4, there is a height difference between the first region R1 and the second region R2, and in this case, the top of the n-type cladding layer 21 is defined separately in the first region R1 and the second region R2.

In the n-type cladding layer 21 composed of n-type AlGaN semiconductor, during epitaxial growth, multi-segmented platforms parallel to the (0001) plane are formed on the growth surface. Ga, which is easily transported by mass, is concentrated in inclined regions that are tilted relative to the (0001) plane between the adjacent platforms, forming Ga-rich n-type regions with a lower AlN molar fraction than the platform regions. As epitaxial growth continues, these inclined regions extend obliquely upwards, resulting in the formation of layered regions 21a with locally low AlN molar fractions within the n-type cladding layer 21. Each extension direction of the layered regions 21a on the first plane has a portion that is tilted relative to the intersection of the top surface of the n-type cladding layer 21 and the first plane.

In one embodiment, during the epitaxial growth of the n-type cladding layer 21, accompanied by the formation of Ga-enriched n-type regions within the tilted region, that is, through the mass movement of Ga from the plateau region to the tilted region, the Al density relatively increases in a portion of the plateau region, thus forming Al-enriched n-type regions with a higher AlN molar fraction than the average AlN molar fraction.

As described above in the prior art section, the layered region 21a extends obliquely to the surface of the n-type cladding layer 21, and because the band gap energy locally decreases, carriers are easily localized. The region outside the layered region 21a within the n-type cladding layer 21, which operates as a low-impedance current path, is called the n-type body region 21b.

In this embodiment, the first average AlN molar fraction Xna1, covering the entire depth direction of the n-type coating layer 21, is within the range expressed by the inequality in equation (1) below. However, n in equation (1) below is an integer of 6 or 7. (n-0.25)/12＜Xna1＜(n+0.25)/12　　(1)

Furthermore, in this embodiment, the second average AlN molar fraction Xna2(d) at depth d (nm) above the n-type coating layer 21 varies with depth d, and there exists a depth d0 where Xna2(d) = 0 and is 1 or higher. Then, within at least one specific depth dx within that depth d0, below that specific depth dx, there exists a region on the upper side where Xna2(d) < n/12, and on the lower side where Xna2(d) > n/12.

As a typical example of depth d0 and a specific depth dx, imagine a case where depth d0 is 1 and d0 = dx. Furthermore, in this typical example, as shown in the embodiment of Figure 12 described later, imagine that at the beginning of the growth of the n-type covering layer 21, Xna2(d) > n/12, and as it grows, it becomes Xna2(d) < n/12, causing Xna2(d) to decrease. However, at least one depth d0 includes cases where Xna2(d0) = 0 at a non-uniform point, and Xna2(d0) = 0 over a continuous range of depths d. The same applies to the specific depth dx.

Furthermore, as an implementation, it is preferable that the third average AlN molar fraction Xna3, extending from the top of the n-type covering layer 21 to a specific depth dx, falls within the range represented by the inequality in equation (2) below. (n-0.25)/12＜Xna3＜(n+0.25)/12　　(2) However, when a complex number exists at a specific depth dx, it is preferable that the above relationship is satisfied for each specific depth dx.

Furthermore, as an implementation, the second average AlN molar fraction Xna2(d) preferably falls within the range represented by the inequality in equation (3) below, across the entire depth direction of the n-type coating layer 21. (n-0.25)/12＜Xna2(d)＜(n+0.25)/12　　(3)

Furthermore, as an embodiment, it is preferable that Xna2(d) lies within the range represented by the inequality in equation (4) below, in the region from the top of the n-type covering layer 21 to a specific depth dx. Furthermore, in the aforementioned preferred embodiment where Xna2(d) lies within the range represented by equation (3), it is preferable that Xna2(d) lies within the range represented by the inequality in equation (5) below, in the region from the top of the n-type covering layer 21 to a specific depth dx. Xna2(d)≦n/12　　(4) (n-0.25)/12＜Xna2(d)≦n/12　　(5)

Furthermore, in the preferred embodiments described above, where Xna2(d) is within the range expressed by equation (4) or (5), and where Xna2(d) is in a region deeper than a specific depth dx, it is more preferably within the range expressed by the inequality in equation (6) above, and even more preferably within the range expressed by the inequality in equation (7) below. n/12≦Xna2(d)　　(6) n/12≦Xna2(d)＜(n+0.25)/12　　(7)

Furthermore, the first average AlN molar fraction Xna1 is based on the AlN molar fraction (n/12) of the first quasi-stable AlGaN region where the Al _n Ga _12-n N ₁₂ composition ratio is an integer ratio. It is controlled within ±0.25/12 (±approximately 2.08%) and is similarly formed within the n-type body region 21b. The second average AlN molar fraction Xna2(d) is in the region where the AlN molar fraction (n/12) of the first quasi-stable AlGaN region is lower. Within the Al-enriched n-type region of the n-type body region 21b, the first quasi-stable AlGaN region is formed.

Within the layered region 21a, the AlN molar fraction is one level lower (approximately 8.33%) than that of the first quasi-stable AlGaN region. The second quasi-stable AlGaN region, with an AlN composition ratio of an integer ratio of Al _n-1 Ga _13-n N ₁₂ , is formed along with the mass movement of Ga during the formation of the layered region 21a. However, the second quasi-stable AlGaN region does not dominate the formation of the layered region 21a. Instead, an intermediate AlGaN region (Al _n-0.5 Ga _12.5-n N ₁₂ ) is formed, with an AlN molar fraction located between the first and second quasi-stable AlGaN regions.

In this embodiment, the thickness of the n-type coating layer 21 is the same as that used in general nitride semiconductor ultraviolet light-emitting devices, which is envisioned to be 1μm to 2μm, but the thickness can also be 2μm to 4μm.

The active layer 22 is a multi-quantum well structure having two or more well layers 220 composed of AlGaN-based semiconductors (excluding AlN-based semiconductors) and one or more barrier layers 221 composed of AlGaN-based semiconductors (excluding GaN-based semiconductors) or AlN-based semiconductors. There is no need to set a barrier layer 221 between the bottom well layer 220 and the n-type cladding layer 21. Furthermore, in this embodiment, as a preferred embodiment, although no barrier layer 221 is provided between the uppermost well layer 220 and the electronic barrier layer 23, as a preferred embodiment, a thinner AlGaN layer or an AlN layer with a higher molar content than the barrier layer 221 may be provided.

The electronic barrier layer 23 is made of p-type AlGaN semiconductor. The p-type interconnect layer 24 is made of p-type AlGaN semiconductor or p-type GaN semiconductor. The p-type interconnect layer 24 is typically made of p-GaN.

Figure 5 shows an example of a multi-quantum-well structure (multiple quantum well structure) of the well layer 220 and the barrier layer 221 of the active layer 22. Figure 5 illustrates a case where the well layer 220 and the barrier layer 221 each have three layers. On the n-type coating layer 21, three layers are stacked in the order of barrier layer 221 and well layer 220. On the topmost well layer 220, an electron barrier layer 23 is positioned.

The multi-segment structure of the well layer 220, barrier layer 221, and electronic barrier layer 23 shown in Figure 5 is a known structure as disclosed in Non-Patent Documents 1 and 2. In each layer, adjacent platforms T in the lateral direction form inclined regions IA tilted towards the (0001) plane, as described above. The regions outside the inclined regions IA, held above and below by the platforms T, are called platform regions TA. In this embodiment, the depth of one platform T (the distance between adjacent inclined regions IA) is assumed to be several tens of nm to several hundred nm. Therefore, within the inclined regions IA, a stepped (0001) plane system is distinguished from the platform surfaces of the multi-segment platform T. Figure 6, for example, schematically shows a stepped structure (giant stepped structure) on the surface of the inclined domain IA of a trapezoidal layer 220.

As shown schematically in Figure 5, the well layer 220 is constructed of AlGaN-based semiconductor. When the AlN molar fraction is not 0%, in each layer of the well layer 220, through the mass movement of Ga from the plateau region TA to the inclined region IA, a Ga-enriched well region 220a with a lower AlN molar fraction than the average AlN molar fraction Xwa within the well layer 220 is formed in the inclined region IA. Furthermore, in one embodiment, with the formation of the Ga-enriched well region 220a in the inclined region IA, that is, through the mass movement of Ga from the plateau region TA to the inclined region IA, the Al density in a portion of the plateau region TA relatively increases, thus forming an Al-enriched well region with a higher AlN molar fraction than the average AlN molar fraction Xwa.

Furthermore, as a preferred embodiment, the well layer 220 is constructed of an AlGaN-based semiconductor. When the AlN molar fraction is not 0%, the average AlN molar fraction Xwa of the well layer 220 is preferably adjusted to the range of Xw0+2%~Xw0+3% when forming a quasi-stable AlGaN with an AlN molar fraction Xw0 in the Ga-rich well region 220a. Through this preferred embodiment, the difference in AlN molar fraction between the tilt region IA and the plateau region TA of the well layer 220 can suppress the generation of double emission spikes caused by the difference in AlN molar fraction by less than 4%. However, even if the AlN molar fraction Xwa of the platform region of the well layer 220 exceeds the range of Xw0+2%~Xw0+3%, when a Ga-rich well region 220a with a localized low AlN molar fraction is formed only in the inclined region IA of the well layer 220, the AlN molar fraction in the Ga-rich well region 220a corresponding to the target value of the peak emission wavelength can be set to Xw1%, which can be any value within the range of Xw1+2%~Xw1+3%.

In this embodiment, when the barrier layer 221 is constructed of an AlGaN-based semiconductor (excluding AlN-based semiconductors), a Ga-enriched barrier region 221a with a lower AlN molar fraction than the average AlN molar fraction Xba of the barrier layer 221 is also formed within the inclined region IA of the barrier layer 221. Furthermore, in another embodiment, similar to the well layer 220, an Al-enriched barrier region with a higher AlN molar fraction than the average AlN molar fraction Xba of the barrier layer 221 may also be formed within a portion of the platform region TA.

As a preferred embodiment, when forming a quasi-stable AlGaN region with an AlN molar fraction Xb0 within the Ga-enriched barrier region 221a of the barrier layer 221, the average AlN molar fraction Xba of the barrier layer 221 is preferably adjusted to be approximately within the range of Xb0+2% to Xb0+8%. Thus, the difference in AlN molar fraction between the Ga-enriched barrier region 221a of the barrier layer 221 and the platform region TA can be ensured to be at least approximately 2%.

Furthermore, as a preferred embodiment, the AlN molar fraction of the platform region TA in the barrier layer 221 is approximately in the range of 51% to 90%, which is set to be at least 1% higher than the AlN molar fraction of the Ga-enriched barrier region 221a, preferably at least 2%, and more preferably at least 4%. In order to fully ensure the localization effect of carriers in the Ga-enriched barrier region 221a, although it is preferable that the difference in AlN molar fraction between the Ga-enriched barrier region 221a and the platform region TA in the barrier layer 221 is at least 4% to 5%, a localization effect of carriers can be expected at a level of 1% to 2%.

In this embodiment, within the electron barrier layer 23, in the tilted region IA, a Ga-enriched EB region 23a with a lower AlN molar fraction than the average AlN molar fraction Xea of the electron barrier layer 23 is also formed. Furthermore, in another embodiment, similar to the well layer 220, in a portion of the platform region TA, an Al-enriched EB region with a higher AlN molar fraction than the average AlN molar fraction Xea of the electron barrier layer 23 may also be formed.

The AlN molar fraction in the platform region TA of the electron barrier layer 23 is approximately in the range of 69% to 90%, which is set to above 20%, preferably above 25%, and even more preferably above 30% compared to the AlN molar fraction in the well layer 220, thus setting it at a high level. Furthermore, the AlN molar fraction in the Ga-enriched EB region 23a of the electron barrier layer 23 is set to above 20%, preferably above 25%, and even more preferably above 30% compared to the AlN molar fraction in the Ga-enriched well region 220a of the well layer 220, thus setting it at a high level.

As a preferred embodiment, when forming a quasi-stable AlGaN region with an AlN molar fraction of Xe0 within the Ga-enriched EB region 23a of the electron barrier layer 23, the average AlN molar fraction Xea of the electron barrier layer 23 is preferably adjusted to be approximately within the range of Xe0+2% to Xe0+8%. Thus, the difference in AlN molar fraction between the Ga-enriched EB region 23a, which serves as the electron barrier layer 23, and the platform region TA can be ensured to be at least approximately 2%.

In the layered structure (giant ladder structure) of the light-emitting element structure 20 shown in Figure 5, in the n-type coating layer 21, carriers are readily localized in the layered region 21a where the local AlN molar fraction is low. In the active layer 22, in the Ga-enriched well region 220a where the local AlN molar fraction is low in the inclined region IA of the well layer 220, and in the Ga-enriched barrier region 221a where the local AlN molar fraction is low in the inclined region IA of the barrier layer 221, carriers are readily localized. In the electronic barrier layer 23, in the Ga-enriched EB region 23a where the local AlN molar fraction is low in the inclined region IA, carriers are locally localized. Therefore, with the layered region 21a separating the n-type cladding layer 21 and the Ga-enriched EB region 23a separating the electron barrier layer 23, the Ga-enriched well region 220a of the well layer 220 can be efficiently supplied with carriers, thus forming a device structure that can improve the light emission efficiency of the carriers (electrons and holes) in the well layer 220 by recombination.

In this embodiment, the thickness of the well layer 220 includes the plateau region TA and the tilt region IA, and is set, for example, in the range of 3ML to 14ML, corresponding to the target value of the peak emission wavelength λp of the light-emitting element 1. Furthermore, the thickness of the barrier layer 221 includes the plateau region TA and the tilt region IA, and is set, for example, in the range of 6nm to 8nm. Moreover, the thickness of the electron barrier layer 23 includes the plateau region TA and the tilt region IA, and is set, for example, in the range of 15nm to 30nm (the optimal value is approximately 20nm).

The well layer 220 is made of AlGaN semiconductor. When the AlN molar fraction is not 0%, the AlN molar fraction and film thickness of the well layer 220 (especially the Ga-enriched well region 220a in the tilted region IA), as well as the AlN molar fraction of the barrier layer and electronic barrier layer 23 adjacent to the well layer 220 (especially the Ga-enriched barrier region 221a and Ga-enriched EB region 23a in the tilted region IA) are set according to the target value of the peak emission wavelength λp of the light-emitting element 1.

The well layer 220 is made of GaN-based semiconductor. When the AlN molar fraction is 0%, the AlN molar fraction of the barrier layer and electronic barrier layer 23 adjacent to the well layer 220 (especially the Ga-enriched barrier region 221a and Ga-enriched EB region 23a in the tilted region IA) and the film thickness of the well layer 220 are set according to the target value of the peak emission wavelength λp of the light-emitting element 1.

Figures 7, 8, and 9 graphically represent the simulation results (equivalent to peak emission wavelengths) of the emission wavelength obtained by varying the well layer thickness within the range of 3ML~14ML or 4ML~14ML for a quantum well structure model composed of AlGaN in well layer 220 and barrier layer 221. As a condition for the above simulation, assuming the dominant presence of quasi-stable AlGaN with an integer AlGaN composition ratio in the Ga-rich well region 220a of well layer 220, the AlN molar fraction in the Ga-rich well region 220a of well layer 220 is set to 50% (half) of the AlN molar fraction of quasi-stable AlGaN in Figure 7, and to become quasi-stable AlGaN in Figure 8. The AlN molar fraction of the well layer 220 is 41.7% (5/12), which, in Figure 9, becomes 33.3% (1/3) of the AlN molar fraction of the quasi-stable AlGaN. In Figures 7-9, the AlN molar fraction in the Ga-enriched barrier region 221a of the barrier layer 221 is 66.7% (2/3), 75% (3/4), and 83.3% (5/6). In the simulation results shown in Figures 7-9, the ultraviolet luminescence of the well layer 220 occurs significantly in the tilted region IA. Therefore, it is important that the film thickness of the well layer 220 meets the requirements of this tilted region IA.

From Figures 7 to 9, it can be seen that when the thickness of the well layer 220 is in the range of 3ML to 14ML, the smaller the thickness of the well layer 220, the greater the quantum sealing effect of the well layer 220, and the shorter the emission wavelength. Furthermore, the greater the AlN molar fraction of the barrier layer 221, the greater the degree of change in emission wavelength with respect to the change in the thickness of the well layer 220. Also, as shown in Figure 7, when the AlN molar fraction of the Ga-enriched well region 220a is 50%, within the above-mentioned range of the thickness of the well layer 220 and the AlN molar fraction of the barrier layer 221, the emission wavelength varies in the approximate range of 246nm to 295nm. As shown in Figure 8, when the AlN molar fraction of the Ga-rich well region 220a is 41.7%, within the aforementioned range of the film thickness of the well layer 220 and the AlN molar fraction of the barrier layer 221, the emitted wavelength varies in the approximate range of 249 nm to 311 nm. As shown in Figure 9, when the AlN molar fraction of the Ga-rich well region 220a is 33.3%, within the aforementioned range of the film thickness of the well layer 220 and the AlN molar fraction of the barrier layer 221, the emitted wavelength varies in the approximate range of 261 nm to 328 nm. Furthermore, when the barrier layer 221 is composed of AlN (AlN molar fraction = 100%), the emitted wavelength can be further expanded.

As shown in Figures 7-9, a quasi-stable AlGaN with an AlGaN composition ratio of Al ₁ Ga ₁ N ₂ , Al ₅ Ga ₇ N ₁₂ , or Al ₁ Ga ₂ N ₃ is formed in the Ga-enriched well region 220a of the well layer 220. Corresponding to the AlN molar fraction of the quasi-stable AlGaN, the film thickness of the well layer 220 is adjusted to the range of 3ML to 14ML, and the AlN molar fraction of the Ga-enriched barrier region 221a of the barrier layer 221 is adjusted to the range of 66.7% to 100%, so that the peak emission wavelength can be set in the range of 246nm to 328nm.

Figure 10 shows the simulation results (equivalent to peak emission wavelengths) of the emission wavelength obtained by varying the well layer thickness in the range of 4ML to 10ML for two cases: the well layer is GaN with a barrier layer composed of AlGaN or AlN. The barrier layer has a molar fraction of 66.7% (AlGaN) and 100% (AlN). From Figure 10, it can be seen that the emission wavelength varies in the approximate range of 270nm to 325nm. Therefore, even if the well layer is composed of GaN (AlN molar fraction = 0%), by adjusting the thickness of the well layer 220 to the range of 4ML~10ML, and by adjusting the AlN molar fraction of the Ga-enriched barrier region 221a of the barrier layer 221 to the range of 66.7%~100%, the peak emission wavelength can be set to the range of 270nm~325nm.

As a preferred embodiment of this embodiment, the target value of the peak emission wavelength λp of the light-emitting element 1 is, for example, set in the range of 280nm~315nm when n=7, and in the range of 300nm~330nm when n=6.

When n=7, the first average AlN molar fraction Xna1 is within the range shown in Equation (1) above, that is, within the range of about 56.3% to about 60.4%. The absorption end wavelength (λae) of the n-type coating layer 21 determined by the first average AlN molar fraction Xna1 is located in the range of about 263nm to about 269nm. Therefore, when the target value of the peak emission wavelength λp is within the range of 280nm to 315nm above, the wavelength difference (λp-λae) between the peak emission wavelength (λp) and the absorption end wavelength (λae) is ensured to be above 10nm, which sufficiently suppresses the emission absorption within the n-type coating layer 21. Here, as a comparison when n=7, in the layered region 21a, assuming that the second quasi-stable AlGaN region with a dominant AlN molar fraction of 50% is formed, the first average AlN molar fraction Xna1 becomes a value of about 54.2%, and the absorption end wavelength (λae) becomes about 273nm. As the target value for the peak emission wavelength λp becomes less than 283nm, the above wavelength difference is less than 10nm, resulting in emission absorption within the n-type coating layer 21, which may lead to a decrease in external quantum efficiency.

When n=6, the first average AlN molar fraction Xna1 is within the range shown in Equation (1) above, that is, within the range of approximately 47.9% to approximately 52.1%. The absorption end wavelength (λae) of the n-type coating layer 21 determined by the first average AlN molar fraction Xna1 is located in the range of approximately 276 nm to approximately 283 nm. Therefore, when the target value of the peak emission wavelength λp is within the range of 300 nm to 330 nm above, the wavelength difference (λp-λae) between the peak emission wavelength (λp) and the absorption end wavelength (λae) is sufficiently ensured to be above 10 nm, thus sufficiently suppressing the emission absorption within the n-type coating layer 21. Furthermore, the lower limit of the above target value can be reduced from 300 nm to approximately 293 nm. Here, as a comparison when n=6, in the layered region 21a, assuming that the second quasi-stable AlGaN region with a dominant AlN molar fraction of 41.7% is formed, the first average AlN molar fraction Xna1 becomes a value of about 45.8%, and the absorption end wavelength (λae) becomes about 286nm. As the target value for the peak emission wavelength λp becomes less than 296nm, the above wavelength difference is less than 10nm, resulting in emission absorption within the n-type coating layer 21, which may lead to a decrease in external quantum efficiency.

The AlN molar fraction (n/12) of the first quasi-stable AlGaN region formed within the n-type body region 21b is approximately 58.3% at n=7 and 50% at n=6. Although the contact resistance between the n-type cladding layer 21 and the n-electrode 27 formed on the exposed surface is higher than when the AlN molar fraction is less than 50%, it is lower than when the AlN molar fraction is 60%, thus suppressing a significant increase in contact resistance.

Since the first average AlN molar fraction Xna1 is controlled within the range shown in Equation (1) above, there is also a portion of the region with an AlN molar fraction of 60% in the n-type body region 21b when n=7. However, since the n-electrode 27 is formed on the exposed surface in the second region R2 of the n-type cladding layer 21, and the contact area with the n-type cladding layer 21 is sufficiently wide, in addition to the region with an AlN molar fraction exceeding 60%, this contact area also includes the first quasi-stable AlGaN region with an AlN molar fraction of approximately 58.3% and layered regions with a smaller AlN molar fraction. Therefore, the average contact impedance between the n-electrode 27 and the n-type cladding layer 21 is reduced. When n=7, in the preferred embodiment where Xna2(d) extends from the top of the n-type cladding layer 21 to a specific depth dx, within the range represented by the inequality in equation (4) above, the average contact resistance between the n-electrode 27 and the n-type cladding layer 21 is further reduced. Furthermore, when n=6, the average contact resistance between the n-electrode 27 and the n-type cladding layer 21 is reduced even more than when n=7.

Furthermore, in this embodiment, in order to control the light absorption within the n-type coating layer 21, the first average AlN molar fraction Xna1 and the second average AlN molar fraction Xna2(d) are not set at unnecessarily high points, thus suppressing an unnecessary increase in bulk resistivity.

The p-electrode 26 is formed on top of the p-type interconnect layer 24, for example, by a multilayer metal film such as Ni/Au. The n-electrode 27 is formed on a portion of the exposed surface of the second region R2 of the n-type coating layer 21, for example, by a multilayer metal film such as Ti/Al/Ti/Au. However, the p-electrode 26 and the n-electrode 27 are not limited to the aforementioned multilayer metal films, and the electrode structure, such as the metal composition, number of layers, and layering order, can be appropriately changed. Figure 11 shows an example of the shape of the light-emitting element 1 viewed from above by the p-electrode 26 and the n-electrode 27. In Figure 11, the line BL between the p electrode 26 and the n electrode 27 shows the boundary line between the first territory R1 and the second territory R2, which coincides with the outer peripheral sidewall of the active layer 22, the electronic barrier layer 23, and the p-type interconnect layer 24.

In this embodiment, as shown in FIG11, the planar shape of the first domain R1 and the p electrode 26 is an example. Although a comb shape is adopted, the planar shape and arrangement of the first domain R1 and the p electrode 26 are not limited to the example shown in FIG11.

When a forward bias voltage is applied between the p-electrode 26 and the n-electrode 27, holes are supplied from the p-electrode 26 toward the active layer 22, and electrons are supplied from the n-electrode 27 toward the active layer 22. The supplied holes and electrons then reach the active layer 22 and recombine to emit light. Furthermore, a forward current flows between the p-electrode 26 and the n-electrode 27.

<Manufacturing Method of Light-Emitting Element> Next, an example of the manufacturing method of the light-emitting element 1 illustrated in Figure 4 will be described.

First, the AlN layer 12 contained in the substrate portion 10 and the nitride semiconductor layers 21-24 contained in the light-emitting element structure portion 20 are sequentially epitaxially grown on the sapphire substrate 11 by metal vapor phase growth (MOVPE). At this time, in the n-type cladding layer 21, Si is used as a donor impurity, and in the electron barrier layer 23 and the p-type interconnect layer 24, Mg is used as an acceptor impurity.

In this embodiment, at least on each surface of the AlN layer 12, the n-type cladding layer 21, the active layer 22 (well layer 220, barrier layer 221), and the electronic barrier layer 23, a multi-segmented platform is formed parallel to the (0001) plane. The sapphire substrate 11 is tilted with respect to the (0001) plane at an angle (offset) within a certain range (e.g., 0.3° to 6°) on the main surface 11a, and a micro-tilted substrate forming a multi-segmented platform is used on the main surface 11a.

In addition to the use of the (0001) sapphire substrate 11 with a micro-tilted substrate, as well as other conditions for epitaxial growth, such as growth rates that easily exhibit multi-segmented plateaus (specifically, such growth rates are achieved by appropriately setting conditions such as growth temperature, the supply amount or flow rate of the raw material gas or carrier gas), these conditions can be specified by actually producing several samples in the film deposition apparatus.

As growth conditions for the n-type coating layer 21, after growth begins, in the stepped regions (inclined regions) between the multi-segmented platforms formed on the AlN layer 12, the mass movement of Ga forms the growth starting point of the layered region 21a. Then, with the epitaxial growth of the n-type coating layer 21, the layered region 21a, accompanied by the mass movement of Ga, can grow obliquely upwards through segregation. The growth temperature, growth pressure, and donor impurity concentration can be selected.

Specifically, as a growth temperature, it is preferable to be above 1050°C, where Ga mass migration is easily generated, and below 1150°C, which allows for the formation of well-formed n-mold AlGaN. Furthermore, as a preferred embodiment, when forming a quasi-stable AlGaN region with an AlN molar fraction of n/12 within the n-type body region 21b, at growth temperatures exceeding 1170°C, Ga mass migration becomes excessive. Even with quasi-stable AlGaN, the AlN molar fraction can easily fluctuate, making it difficult to stably form quasi-stable AlGaN regions with an AlN molar fraction of 50%~58.3%. As for growth pressure, below 75 Torr is preferred for good AlGaN growth conditions, and practically, above 10 Torr is preferable as a control limit for the film-forming device. The optimal concentration of donor impurities is 1× ^10¹⁸ to 5× ^{10¹⁸ cm⁻³} . However, the growth temperature and growth pressure mentioned above are just examples; the specific and optimal conditions should be determined for the film-forming device used.

The supply amount and flow rate of the feed gas (trimethylaluminum (TMA) gas, trimethylgallium (TMG) gas, ammonia gas) or carrier gas used in the gas phase growth method of organometallic compounds are set with the first average AlN molar fraction Xna1 of the above-mentioned n-type coating layer 21 as the target value.

In this embodiment, the second average AlN molar fraction Xna2(d) varies with the depth d, along with the first quasi-stable AlGaN molar fraction (n/12). This variation in AlN molar fraction Xna2(d) can be achieved by actively adjusting the supply amount and flow rate of the aforementioned raw material gas or carrier gas, or by adjusting the growth temperature. Furthermore, it can utilize the natural change in substrate surface temperature that occurs as the film thickness increases during the growth of the n-type coating layer 21. In this case, the supply amount and flow rate of the aforementioned raw material gas or carrier gas are set according to the tendency of the substrate surface temperature change.

However, the donor impurity concentration refers to the film thickness of the n-type coating layer 21 and does not need to be uniformly controlled in the vertical direction. For example, the impurity concentration in a specific film thickness portion within the n-type coating layer 21 can be lower than the aforementioned set concentration, for example, it can be controlled to be less than 1× ^{10¹⁸ cm⁻³} , and even more so to be a low impurity concentration layer below 1× ^{10¹⁷ cm⁻³} . The film thickness of this low impurity concentration layer is preferably between 200 nm and 0 nm, even better if it is between 10 nm and 100 nm, and even more ideally between 20 nm and 50 nm. Furthermore, the donor impurity concentration of the low impurity concentration layer only needs to be lower than the above-mentioned set concentration, and the undoped layer (0 ^cm⁻³ ) may also be included in a portion. Moreover, it is preferable that part or all of the low impurity concentration layer exists in the upper layer region at a depth of less than 100 nm from the n-type coating layer 21 downwards.

In accordance with the above-mentioned principles, when forming an n-type cladding layer 21 having layered regions 21a and n-type body regions 21b, an active layer 22 (well layer 220, barrier layer 221), an electronic barrier layer 23, and a p-type interconnect layer 24 are formed on the entire surface of the n-type cladding layer 21 by a known epitaxial growth method such as organometallic vapor phase growth (MOVPE).

For example, the acceptor impurity concentration of the electron barrier layer 23 is preferably between 1.0 × ^10¹⁶ and 1.0 × ^{10¹⁸ cm⁻³} , and the acceptor impurity concentration of the p-type interconnect layer 24 is preferably between 1.0 × ^10¹⁸ and 1.0 × ^{10²⁰ cm⁻³} . However, the acceptor impurity concentration is relative to the thickness of each film of the electron barrier layer 23 and the p-type interconnect layer 24, and does not need to be uniformly controlled in the vertical direction.

During the formation of the active layer 22, following the same principles as the n-type coating layer 21, and under growth conditions that easily exhibit the aforementioned multi-segmented platform, the average AlN molar fraction Xwa of the well layer 220 is used as the target value for growth, and the average AlN molar fraction Xba of the barrier layer 221 is used as the target value for growth, and the barrier layer 221 is grown. The average AlN molar fractions Xwa and Xba of the well layer 220 and the barrier layer 221 are as described above, and will not be repeated.

In the formation of the electronic barrier layer 23, the same principles as for the n-type coating layer 21 are followed. Under growth conditions that easily exhibit the aforementioned multi-segmented platform, the average AlN molar fraction Xea of the electronic barrier layer 23 is used as the target value for growth. The average AlN molar fraction Xea of the electronic barrier layer 23 is as described above, and will not be repeated here.

In this embodiment, the optimal growing temperatures of the active layer 22 (well layer 220, barrier layer 221), the electron barrier layer 23, and the p-type interconnect layer 24 are such that the growth temperature of the n-type coating layer 21 is T1, the growth temperature of the active layer 22 is T2, the growth temperature of the electron barrier layer 23 is T3, and the growth temperature of the p-type interconnect layer 24 is T4. Within these preferred temperature ranges (1050℃~1170℃), the relationships shown in equations (8) and (9) are preferably satisfied. T3≧T2　　　　　(8) T3＞T1＞T4　　(9)

The growth temperature T3 of the electron barrier layer 23 can be reduced, for example, by increasing the flow rate of nitrogen feed gas, thereby decreasing the growth rate.

However, when the growth temperature T3 of the electron barrier layer 23 is increased from the growth temperature T2 of the active layer 22, GaN decomposition occurs in the well layer 220 below during the migration process of the growth temperature. This GaN decomposition may degrade the characteristics of the light-emitting element 1. Therefore, in order to suppress the decomposition of GaN, it is preferable to form an AlGaN layer or an AlN layer with a higher AlN molar ratio than the barrier layer 221 and the electron barrier layer 23, with a thinner film (e.g., less than 3 nm, preferably less than 2 nm) between the uppermost well layer 220 and the electron barrier layer 23.

In accordance with the above-mentioned principles, when an active layer 22 (well layer 220, barrier layer 221), an electronic barrier layer 23, and a p-type interconnect layer 24 are formed over the entire surface of the n-type coating layer 21, then, using a known etching method such as reactive ion etching, the second region R2 of the nitride semiconductor layers 21-24 is selectively etched to expose the second region R2 portion above the n-type coating layer 21. Then, using a known film formation method such as electron beam evaporation, a p-electrode 26 is formed on the p-type interconnect layer 24 in the unetched first region R1, while an n-electrode 27 is formed on the n-type coating layer 21 in the etched second region R2. However, after the formation of one or both of the p-electrode 26 and the n-electrode 27, heat treatment can be performed using a known heat treatment method such as RTA (instantaneous thermal annealing).

However, the light-emitting element 1 is used as an example, after being flip-chip mounted on a base such as a sub-fixed seat, and is sealed in a specific resin (such as a lens-shaped resin) such as polysiloxane resin or amorphous fluororesin.

The cross-sectional structure of the AlGaN semiconductor layers 21-24 of the light-emitting element 1 fabricated according to the above-described method is a sample before etching the second region R2 and forming the p-electrode 26 and n-electrode 27. A sample sheet with a cross-section perpendicular (or slightly perpendicular) to the surface of the sample is processed using focused ion beam (FIB) and observed through a HAADF-STEM image. The HAADF-STEM image provides a comparison of scale relative to atomic weights, with heavy elements being brightly displayed. Therefore, regions with low AlN molar fractions are relatively brightly displayed. The HAADF-STEM image is more suitable for observing differences in AlN molar fractions than a typical STEM image (bright field image).

Furthermore, the compositional analysis of specific semiconductor layers 21-24 in the AlGaN semiconductor system can be performed using the aforementioned sample using energy-dispersive X-ray spectrometry (TEM-EDX) or cathode luminescence (CL). Detailed descriptions of the compositional analysis using TEM-EDX and CL methods can be found in the specifications of the inventor's previous applications (PCT/JP2020/024827, PCT/JP2020/024828, PCT/JP2020/026558, PCT/JP2020/031620, etc.).

<Compositional Analysis Results of the n-type Coating Layer> Next, the results of measuring the second average AlN molar fraction Xna2(d) of the n-type coating layer 21 using Rashof backscattering (RBS) analysis will be explained; the results of measuring the AlN molar fraction of the layered regions 21a and n-type bulk regions 21b within the n-type coating layer 21 using CL (cathode luminescence) method will be explained; and the results of measuring the AlN molar fraction of the layered regions 21a within the n-type coating layer 21 using profile TEM-EDX method will be explained.

To prepare a sample for compositional analysis of the n-type coating layer 21, a sample sheet having a cross section perpendicular (or slightly perpendicular) to the top of the n-type coating layer 21 is processed by focused ion beam (FIB) to prepare a sample sheet for measurement.

The aforementioned sample was fabricated according to the aforementioned manufacturing method of the n-type cladding layer 21, etc., on a substrate portion 10 formed from the aforementioned sapphire substrate 11 and AlN layer 12, by sequentially depositing the n-type cladding layer 21, an AlGaN layer with a higher AlN molar fraction than the n-type cladding layer 21, an AlGaN layer for sample surface protection, and a protective resin film. However, in the fabrication of this sample, a sapphire substrate 11 with an angle between its main surface and the (0001) surface was used, and a substrate portion 10 exhibiting a multi-segmented platform was used on the surface of the AlN layer 12. Furthermore, the donor impurity concentration was set to approximately 3 × ^{10¹⁸ cm⁻³} , and the implantation amount of donor impurity (Si) was controlled.

The first average AlN molar fraction Xna1, the second average AlN molar fraction Xna2(d) of the depth d (nm), and the third average AlN molar fraction Xna3 of the n-type coating layer 21 of the above-mentioned sample satisfy the above-mentioned equations (1), (3), and (2) when n=7. Therefore, a first quasi-stable AlGaN region with an AlN molar fraction of 7/12 (approximately 58.3%) is formed in the n-type body region 21b, and an intermediate AlGaN region with an AlN molar fraction of 6.5/12 (approximately 54.2%) is formed in the layered region 21a.

Figure 12 shows the RBS analysis results of the second average AlN molar fraction Xna2(d). The AlN molar fraction Xna2(d) varies with depth d, ranging from approximately 57.8% to approximately 59.9%, satisfying the above equation (3). The first average AlN molar fraction Xna1 is calculated from Xna2(d) shown in Figure 12 to be approximately 58.4%, which is slightly higher than the AlN molar fraction in the first quasi-stable AlGaN field (approximately 58.3%), but roughly the same. As will be discussed later, the specific depth dx is 1, approximately 1061 nm. The third average AlN molar fraction Xna3 is calculated from Xna2(d) shown in Figure 12 to be approximately 57.9%, which is smaller than the first average AlN molar fraction Xna1, satisfying equation (2). When the second average AlN molar fraction Xna2(d) satisfies equation (3), the first average AlN molar fraction Xna1 naturally satisfies the above equation (1), and the third average AlN molar fraction Xna3 naturally satisfies equation (2).

As shown in Figure 12, in the second average AlN molar fraction Xna2(d) system, in the upper region with a depth d (nm) of 0 nm to approximately 750 nm, approximately 57.8% remains constant. As the depth d (nm) increases from approximately 750 nm to approximately 1900 nm, it gradually increases from approximately 57.8% to approximately 59.9%, reaching a depth d (nm) of approximately 1061 nm, which is equal to the AlN molar fraction (approximately 58.3%) in the first quasi-stable AlGaN region. Therefore, in the example shown in Figure 12, the specific depth dx is approximately 1061 nm. The second average AlN molar fraction Xna2(d) is smaller than the AlN molar fraction (approximately 58.3%) in the first quasi-stable AlGaN region in the depth d (nm) range of 0 nm to approximately 1061 nm, satisfying equation (5). In the depth d (nm) range of approximately 1061 nm to approximately 1900 nm, the AlN molar fraction is larger than the AlN molar fraction (approximately 58.3%) in the first quasi-stable AlGaN region, satisfying equation (7). Therefore, the second average AlN molar fraction Xna2(d) shown in Figure 12 is a typical example satisfying equations (5) and (7). That is, d = approximately 1061 nm is a specific depth dx.

However, in the RBS analysis method, for example, although a He ²⁺ ion beam (beam diameter: 2.2 mm) is used to irradiate the n-type coating layer 21 of the sample vertically at an accelerating voltage of 2.3 MeV, the measurement range in the vertical direction is as large as 300 nm, so the film thickness of the object to be analyzed needs to be larger than 300 nm.

Figure 13 is a scanning electron microscope (SEM) image showing the main portion of the n-type coating layer 21 on the measurement cross-section of the sample specimen. The measurement range of the sample specimen (the range of the incident point of the electron beam used for measurement) is 6.25 μm in the X direction (parallel to the second plane) and 2.2 μm in the Y direction (orthogonal to the second plane), forming a 121-mesh × 41-mesh grid to set the incident point of the electron beam. The mesh spacing is approximately 52 nm in the X direction and approximately 55 nm in the Y direction.

The Y values (Y coordinates) recorded in the measurement range of the sample sheet shown in Figure 13 represent the mesh count from the top of each measurement range, where Y=0. In Figure 13, Y=4 and Y=38 are located near the top and bottom of the n-type coating layer 21, respectively. Therefore, the film thickness of the n-type coating layer 21 is approximately 1.9 μm.

At the incident point of the lattice-shaped electron beam within the measurement range of the sample, an electron beam with a beam diameter of 50 nm (diameter) is irradiated once, and the CL spectrum of each irradiation is measured.

Figure 14 shows the 121 CL spectra obtained by scanning the X direction from the six Y coordinates (Y=10, Y=16, Y=21, Y=26, Y=31, Y=35) of each sample A. The first CL spectrum (solid line) and the second CL spectrum (dashed line) are derived below. The six Y coordinates (Y=10~35) are converted to a depth d from the top of the n-type coating layer 21, which is approximately 330 nm to approximately 1700 nm. The first and second CL spectra of the six Y coordinates are offset from their respective origins in the longitudinal direction and can be mutually identified on the same chart. The vertical axis of Figure 14 displays the luminous intensity (in arbitrary units). Furthermore, the luminous intensity of the two Y axes (Y=10, 35) is set to 0.5 times (Y=10) and 1.5 times (Y=35) for easy identification. The horizontal axis of Figure 14 displays the wavelength (nm).

Furthermore, in Figure 14, for reference, the CL wavelengths (approximately 253 nm, approximately 266 nm, and approximately 279 nm) of the three quasi-stable AlGaNs (AlN molar fractions of 50%, 58.3%, and 66.7%) are plotted with a vertical line of a dotted chain.

The first CL spectrum shown in Figure 14 is obtained by extracting 6 to 7 points or more from the CL spectrum of the same Y coordinate, which are shifted to the same wavelength near the long wavelength side of the AlN molar fraction (58.3%) of the first quasi-stable AlGaN region, from the peak of the luminescence intensity of the CL spectrum of the same Y coordinate. The extracted CL spectrum is then calculated by averaging these points. Therefore, within the measurement domain of the first CL spectrum, there are many layered domains 21a in other measurement domains of the same Y coordinate.

On the other hand, the second CL spectrum shown in Figure 14 is obtained by extracting 6 to 7 points or more from the CL spectrum of the same Y coordinate, where the peak of luminescence intensity is shifted to the same wavelength near the shorter wavelength side of the AlN molar fraction (58.3%) of the first quasi-stable AlGaN region, and averaging the extracted CL spectrum. Therefore, within the measurement domain of the second CL spectrum, through other measurement domains of the same Y coordinate, there are many n-type body domains 21b (especially Al-enriched n-type domains).

The first CL spectral system of each Y coordinate shown in Figure 14 shows that the wavelength λ0(Y) of the maximum signal intensity In0(Y) exists in the range of approximately 268 nm to approximately 271 nm. It exists between the CL wavelength (approximately 273 nm) corresponding to the AlN molar fraction ((n-0.5)/12) of the middle AlGaN region (Al _n-0.5 Ga _12.5-n N ₁₂ ) when n=7, and the CL wavelength (approximately 266 nm) corresponding to the AlN molar fraction (n/12) of the first quasi-stable AlGaN region (Al _n Ga _12-n N ₁₂ ) when n=7. This refers to the first CL spectrum, which includes the CL spectrum from the layered region 21a, the CL spectrum from the first quasi-stable AlGaN region within the n-type body region 21b, and the CL spectrum from the Al-enriched n-type region within the n-type body region 21b, and especially mainly includes the first two CL spectra, thus forming the composite spectrum.

Here, the Al-enriched n-type region within the n-type body region 21b is formed as described above during the epitaxial growth of the n-type cladding layer 21 by the mass movement of Ga from the plateau region to the inclined region, accompanied by the formation of the layered region 21a (Ga-enriched n-type region). Similarly, within the layered region 21a, a second intermediate AlGaN region (Al _n+0.5 Ga _11.5-n N ₁₂ ) is also formed within the Al-enriched n-type region, just as an intermediate AlGaN region (Al _n-0.5 Ga _12.5-n N ₁₂ ) is formed. The second intermediate AlGaN region has an AlN molar fraction that falls between the first quasi-stable AlGaN region and the third quasi-stable AlGaN region (Al _n+1 Ga _11-n N ₁₂ ), which has an AlN molar fraction one step larger (approximately 8.33%) than the first quasi-stable AlGaN region. However, the CL wavelength of the AlN molar fraction ((n+0.5)/12) in the second intermediate AlGaN region corresponding to n=7 is approximately 259 nm.

Furthermore, in the first CL spectrum of each Y coordinate, the signal intensity In11(Y) at the wavelength λs1 (approximately 266 nm) corresponding to the AlN molar fraction (approximately 58.3%) in the first quasi-stable AlGaN region is approximately 67% to approximately 96% of the maximum signal intensity In0(Y). Within the measurement domain of the first CL spectrum of each Y coordinate, it can be seen that the n-type body domain 21b, which dominates the formation of the first quasi-stable AlGaN region, is included.

Furthermore, in the first CL spectrum of each Y coordinate, the signal intensity In1m(Y) at the wavelength λsm (approximately 273 nm) corresponding to the AlN molar fraction (approximately 54.2%) in the intermediate AlGaN region is approximately 73% to approximately 93% of the maximum signal intensity In0(Y). Within the measurement domain of the first CL spectrum of each Y coordinate, it can be seen that there is a layered region 21a that dominates the formation of the intermediate AlGaN region (Al _n-0.5 Ga _12.5-n N ₁₂ ).

Based on the above measurement results, when the signal intensity In1m(Y) is 70% or more of the maximum signal intensity In0(Y), it is determined that the intermediate AlGaN region (Al _n-0.5 Ga _12.5-n N ₁₂ ) is dominantly formed within the layered region 21a. The same applies when n=6. However, to ensure the accuracy of the measurement obtained by the CL method, the Y coordinate of the first CL spectrum is preferably limited to the Y coordinate in the middle region excluding the upper region 150 nm below the upper end of the n-type coating layer 21 and the lower region 150 nm above the lower end, excluding the influence of the active layer 22 existing on the upper side of the n-type coating layer 21 and the influence of the substrate portion 10 (in one embodiment, the uppermost layer is AlN12) existing on the lower side of the n-type coating layer 21.

Furthermore, in the first CL spectrum of each Y coordinate, the signal intensity In12(Y) at the wavelength λs2 (approximately 279 nm) corresponding to the AlN molar fraction (50%) in the second quasi-stable AlGaN region is approximately 13% to approximately 40% of the maximum signal intensity In0(Y). Although the second quasi-stable AlGaN region exists within the measurement domain of the first CL spectrum of each Y coordinate, it is not dominant.

Regarding the above measurement results, when the signal intensity In12(Y) is less than 50% of the maximum signal intensity In0(Y), it is determined that a second quasi-stable AlGaN region is not dominantly formed within the layered region 21a. The same applies when n=6. However, to ensure the measurement accuracy obtained by the CL method, it is preferable that the Y coordinate of the first CL spectrum is limited to the Y coordinate within the intermediate region.

Figure 14 shows that the wavelength λ1(Y) of the maximum signal intensity In1(Y) in the second CL spectral system for each Y coordinate is in the range of approximately 258 nm to approximately 261 nm. The AlN molar fraction Xn1(d) corresponding to the depth d at each wavelength λ1(Y) is in the range of approximately 61.5% to approximately 63.5%, which is higher than the average AlN molar fraction Xna2(d) of the second CL spectral system. Therefore, with the mass movement of Ga during the formation of the layered region 21a, Al-enriched n-type regions are formed within the n-type body region 21b. However, near the upper end of the n-type cladding layer 21, the AlN molar fraction Xn1(d) is approximately 61.5%.

Furthermore, in the second CL spectrum of each Y coordinate, the signal intensity In21(Y) at the wavelength λs1 (approximately 266 nm) corresponding to the AlN molar fraction (approximately 58.3%) in the first quasi-stable AlGaN region is approximately 55% to approximately 84% of the maximum signal intensity In1(Y). Within the measurement domain of the second CL spectrum of each Y coordinate, it can be seen that the n-type body domain 21b, which dominates the formation of the first quasi-stable AlGaN region, is included.

Furthermore, since the signal intensity In21(Y) of the wavelength λs1 (approximately 266 nm) corresponding to the AlN molar fraction (approximately 58.3%) in the first and second CL spectra of each Y coordinate is the same, it can be concluded that the first quasi-stable AlGaN region is also formed within the n-type body region 21b.

Furthermore, in the second CL spectrum of each Y coordinate, the signal intensity In2m(Y) at the wavelength λsm (approximately 259 nm) corresponding to the AlN molar fraction (62.5%) in the second intermediate AlGaN region is approximately 90% to approximately 100% of the maximum signal intensity In2(Y). Within the measurement domain of the second CL spectrum of each Y coordinate, it can be seen that it includes the Al-enriched n-type region within the n-type body region 21b that dominates the formation of the second intermediate AlGaN region (Al _n+0.5 Ga _11.5-n N ₁₂ ).

Therefore, on the surface of the n-type coating layer 21 (especially the exposed surface within the second domain R2), within the n-type body domain 21b, there is a first quasi-stable AlGaN domain with an AlN molar fraction of approximately 58.3% and a layered domain 21a with a low AlN molar fraction. Even if there are some Al-enriched n-type domains with an AlN molar fraction slightly exceeding 60%, it can be seen that the contact resistance between the surface of the n-type coating layer 21 and the n-electrode 27 can be reduced.

Figure 15 shows the AlN molar fraction in the layered region 21a of the above-mentioned sample. The HAADF-STEM images of the measurement regions A to D at the four locations are shown by the line analysis of the cross-sectional TEM-EDX.

In the compositional analysis (EDX measurement) performed by the cross-sectional TEM-EDX method, firstly, in the overall measurement area covering the four locations shown in Figure 15 (A~D), an electron beam probe (diameter: approximately 2nm) is scanned in the longitudinal (vertical) and transverse (parallel to the second plane) directions to form a 512×512 matrix. In the longitudinal and transverse directions, the detection data of each probe location distributed at intervals of approximately 4nm (corresponding to the X-ray intensity of each component of Al and Ga) are obtained.

Next, in order to perform line analysis for EDX measurement on the layered regions 21a scattered throughout the overall measurement area, measurement regions A to D (shown as dashed lines in Figure 15) are set at the above four locations within the overall measurement area. Each measurement region A to D is rectangular, and its inclination and angle are determined by the extension direction of at least one layered region 21a within the measurement region, orthogonal to the scanning direction of the line analysis, and are set in each measurement region. Furthermore, the inclination of each measurement region A to D (the angle between the longitudinal direction of the overall measurement area and the longitudinal direction of each measurement region) is approximately equal to 20°, although strictly speaking, they are not necessarily exactly the same. Here, in addition to the overall measurement area in both the vertical and horizontal directions, within each measurement area A-D in Figure 15, for ease of explanation, the scanning direction of the line analysis is taken as the vertical direction, and the direction orthogonal to the scanning direction is taken as the horizontal direction. The vertical line shown in the center of each measurement area indicates the scanning direction, and the × mark on the vertical line indicates the position in the vertical direction of the layered area 21a that is the object of AlN mole fraction measurement. However, the positions of the × marks in measurement areas A-D approximately correspond to the Y coordinates 10, 21, 31, and 36 in the measurement range of AlN mole fraction obtained by the CL method shown in Figure 13.

In EDX measurements, the diameter of the irradiated electron beam probe is as small as 2 nm. Although the spatial resolution energy is high, the X-rays emitted from each probe are weak. In the line analysis of this embodiment, the detection data obtained from multiple probes arranged in the horizontal direction are accumulated at each scan position to form the detection data for each scan position. However, "arranged in the horizontal direction" means that the irradiation range of the electron beam probe overlaps with the horizontal lines that intersect the aforementioned vertical lines and extend in the horizontal direction at each scan position.

The AlN molar fraction within the layered regions 21a marked with ×, derived from the accumulated detection data obtained according to the above-described principles, is shown below. However, within the parentheses to the right of the AlN molar fraction measurement results, the AlN molar fraction difference Δ is displayed, calculated by subtracting the AlN molar fraction of the intermediate AlGaN region (approximately 54.17%) from the AlN molar fraction measurement results within each of the layered regions 21a of measurement regions A through D. Measurement Area A (Y = approx. 10): 52.62% (Δ = -1.55%) Measurement Area B (Y = approx. 21): 54.52% (Δ = 0.35%) Measurement Area C (Y = approx. 31): 54.63% (Δ = 0.46%) Measurement Area D (Y = approx. 36): 54.05% (Δ = -0.12%)

Based on the measurement results of the AlN molar fraction within the layered regions 21a of each measurement region A to D, it can be seen that the intermediate AlGaN region with a dominant AlN molar fraction of 6.5/12 (approximately 54.2%) is formed within the layered regions 21a. This is consistent with the results derived from the first CL spectrum in Figure 14.

[Second Embodiment] In the light-emitting element 1 of the first embodiment, the p-type layer constituting the light-emitting element structure 20 consists of two layers: an electron barrier layer 23 and a p-type interconnect layer 24. However, in the light-emitting element 2 of the second embodiment, the p-type layer is a p-type cladding layer 25 composed of one or more p-type AlGaN semiconductor layers, located between the electron barrier layer 23 and the p-type interconnect layer 24.

Therefore, in the second embodiment, as shown in FIG16, the AlGaN semiconductor layers 21-25 of the light-emitting element structure 20 have a structure in which the layers are epitaxially grown sequentially from the substrate 10 side in the order of n-type cladding layer 21 (n-type layer), active layer 22, electron barrier layer 23 (p-type layer), p-type cladding layer 25 (p-type layer) and p-type interconnect layer 24 (p-type layer).

The substrate portion 10 of the light-emitting element 2 and the AlGaN semiconductor layers 21-24, p-electrode 26, and n-electrode 27 of the light-emitting element structure portion 20 in the second embodiment are the same as those of the substrate portion 10 of the light-emitting element 1 and the AlGaN semiconductor layers 21-24, p-electrode 26, and n-electrode 27 of the light-emitting element structure portion 20 in any of the first to third embodiments, so repeated descriptions are omitted.

The p-type cladding layer 25 is the same as the AlN layer 12 of the substrate portion 10 epitaxially grown sequentially on the main surface 11a of the sapphire substrate 11, and the semiconductor layers and electron barrier layers 23 in the n-type cladding layer 21 and active layer 22 of the light-emitting element structure portion 20. It has a surface of a multi-segmented platform formed parallel to the (0001) plane from the formation of the main surface 11a of the sapphire substrate 11.

Figure 17 shows an example of a multi-quantum-well structure (multiple quantum well structure) of a well layer 220 and a barrier layer 221 of an active layer 22. In Figure 17, in the first embodiment, a p-type cladding layer 25 is formed on the electronic barrier layer 23 of the multilayer structure illustrated in Figure 5.

In the p-type cladding layer 25, the platforms T adjacent to each other in the lateral direction form a tilted region IA tilted relative to the (0001) plane, as described above. The region above and below the tilted region IA, held by the platforms T, is called the platform region TA. The film thickness of the p-type cladding layer 25 includes the platform region TA and the tilted region IA, and is, for example, adjusted to the range of 20nm to 200nm.

As shown in the schematic diagram of Figure 17, in the p-type coating layer 25, through the mass movement of Ga from the plateau region TA to the inclined region IA, a Ga-enriched p-type region 25a with a lower AlN molar fraction than the plateau region TA is formed in the inclined region IA.

The AlN molar fraction in the platform region TA of the p-type coating layer 25 is above 51%, which is within the range of the AlN molar fraction in the platform region TA of the deficient electron barrier layer 23. Furthermore, the AlN molar fraction in the Ga-enriched p-type region 25a of the p-type coating layer 25 is set to the AlN molar fraction in the Ga-enriched EB region 23a of the deficient electron barrier layer 23.

Furthermore, the AlN molar fraction of the platform region TA in the p-type coating layer 25 is set to be 1% or more, preferably 2% or more, and more preferably 4% or more, within the aforementioned range, compared to the AlN molar fraction of the Ga-enriched p-type region 25a. This setting is at a high level. In order to fully ensure the localization effect of the carrier in the Ga-enriched p-type region 25a, although it is preferable that the difference in AlN molar fraction between the Ga-enriched p-type region 25a and the platform region TA within the p-type coating layer 25 is 4-5% or more, a localization effect of the carrier can be expected at a level of 1-2%.

Next, the growth method of the p-type coating layer 25 will be briefly explained. In the formation of the p-type coating layer 25, the same principle as that used for the n-type coating layer 21 and the electron barrier layer 23 described in the first embodiment is followed. Under growth conditions that easily exhibit the aforementioned multi-segmented platform, the AlN molar fraction Xpa of the p-type coating layer 25 is used as the target value to grow the p-type coating layer 25.

[Other Embodiments] The following describes variations of the first and second embodiments described above.

(1) In the above embodiments, the active layer 22 is envisioned as a multi-quantum well structure consisting of two or more well layers 220 made of AlGaN semiconductor and one or more barrier layers 221 made of AlGaN semiconductor or AlN semiconductor. However, the active layer 22 may also be a single quantum well structure with only one well layer 220 and may not have a barrier layer 221 (quantum barrier layer). For the related single quantum well structure, it is also clear that the effect caused by the well layer 220 used in the above embodiments can be achieved.

(2) For each of the above embodiments, an n-type AlGaN semiconductor layer (hereinafter referred to as "n-type substrate layer") with a higher AlN molar fraction than the n-type cladding layer 21 may be provided between the n-type cladding layer 21 and the substrate portion 10. Thus, the average AlN molar fraction throughout the depth direction of the combined n-type cladding layer 21 and the n-type substrate layer may be higher than the range of the above formula (1).

The n-type substrate layer located below the n-type cladding layer 21 has a higher AlN molar fraction than the n-type cladding layer 21, therefore it does not absorb light emitted from the active layer. Furthermore, even though the n-type substrate layer has a higher AlN molar fraction than the n-type cladding layer 21, it does not contact the n-electrode 27 and does not form a current path between the n-electrode 27 and the active layer. Therefore, it does not increase the parasitic impedance between the n-electrode 27 and the active layer and will not be a factor in the decrease in socket efficiency. In other words, since the n-type substrate layer does not actually function as part of the light-emitting element structure 20, there are neither any particular advantages nor significant disadvantages to having it.

(3) In the above embodiments, the planar shape of the first domain R1 and the p electrode 26 is taken as an example. Although a comb shape is adopted, the planar shape is not limited to a comb shape. Furthermore, there may be multiple first domains R1, and each second domain R2 may be surrounded by a planar shape.

(4) In the above embodiments, although a sapphire substrate 11 with an angle to the (0001) surface is used, a substrate portion 10 exhibiting a multi-segmented platform is used on the surface of the AlN layer 12. The size of the angle or the direction of the angle (specifically, the direction of the inclination to the (0001) surface, such as the m-axis direction or the a-axis direction) is on the surface of the AlN layer 12 to exhibit a multi-segmented platform. As long as the growth starting point of the layered region 21a is formed, it can be arbitrarily determined.

(5) In each of the above embodiments, although the light-emitting element 1 is illustrated in FIG. 1 as having a substrate portion 10 including a sapphire substrate 11, the sapphire substrate 11 (and moreover, a portion or all of the layer contained in the substrate portion 10) can be removed by peeling or the like. Furthermore, the substrate constituting the substrate portion 10 is not limited to a sapphire substrate. [Industrial Applicability]

This invention relates to a nitride semiconductor ultraviolet emitting element formed by depositing an n-type layer, an active layer, and a p-type layer of an AlGaN-based semiconductor with a zincblende structure in the vertical emitting element structure.

1,2: Nitride semiconductor ultraviolet light-emitting element 10: Substrate 11: Sapphire substrate 11a: Main surface of sapphire substrate 12: AlN layer 20: Light-emitting element structure 21: n-type cladding layer (n-type layer) 21a: Layered region (n-type layer) 21b: n-type body region (n-type layer) 22: Active layer 220: Well layer 220a: Ga-enriched well region 221: Barrier layer 221a: Ga-enriched barrier region 23: Electron barrier layer (p-type layer) 23a: Ga-enriched EB region 24: p-type interconnect layer (p-type layer) 25: p-type cladding layer (p-type layer) 25a: Ga-enriched p-type region 26: p-electrode 27: n-electrode 100: Substrate 101: AlGaN semiconductor layer 102: Template 103: n-type AlGaN semiconductor layer 104: Active layer 105: p-type AlGaN semiconductor layer 106: p-type interconnect layer 107: n-electrode 108: p-electrode BL: Boundary between the first and second regions IA: Tilted region R1: First region R2: Second region T: Plateau TA: Plateau region

[Figure 1] A schematic diagram showing the zincblende crystal structure of AlGaN. [Figure 2] A plan view showing the positional relationship between the positions on the A-plane and the B-plane as viewed along the c-axis of the zincblende crystal structure shown in Figure 1. [Figure 3] A schematic diagram showing the symmetrical arrangement of Al _and Ga _{on the} position planes ( _A3 plane, _B3 plane ₎ of the group 3 elements of each quasi-stabilized AlGaN represented by _Al1Ga2N3 , _Al1Ga1N2 , and _Al2Ga1N3 in an integer ratio. [Figure 4] A schematic main cross-sectional view showing an example of the structure of a nitride semiconductor ultraviolet light-emitting element of the first embodiment. [Figure 5] A schematic cross-sectional view of an example of the stacked structure of the active layer of the nitride semiconductor ultraviolet light-emitting element shown in Figure 4. [Figure 6] A schematic view of the detailed structure through the tilted region IA shown in Figure 5. [Figure 7] A graph showing the relationship between the emission wavelength of the quantum well structure formed by the AlGaN well layer and the AlGaN barrier layer when the AlN molar fraction in the Ga-rich well region 220a is 50%, the film thickness of the well layer, and the AlN molar fraction of the barrier layer. [Figure 8] A graph showing the relationship between the emission wavelength of the quantum well structure formed by the AlGaN well layer and the AlGaN barrier layer in Ga-rich well region 220a with an AlN molar fraction of 41.7%, and the film thickness of the well layer and the AlN molar fraction of the barrier layer. [Figure 9] A graph showing the relationship between the emission wavelength of the quantum well structure formed by the AlGaN well layer and the AlGaN barrier layer in Ga-rich well region 220a with an AlN molar fraction of 33.3%, and the film thickness of the well layer and the AlN molar fraction of the barrier layer. [Figure 10] A graph showing the relationship between the emission wavelength of the quantum well structure formed by the GaN well layer and the AlGaN barrier layer, and the film thickness of the well layer and the AlN molar fraction of the barrier layer. [Fig. 11] A schematic plan view showing an example of the structure of the nitride semiconductor UV-emitting element shown in Fig. 4, viewed from the top of Fig. 4. [Fig. 12] A chart showing an example of the depth d variation of the second average AlN molar fraction Xna2(d) corresponding to depth d from the top of the n-type layer. [Fig. 13] A SEM image of the main part of the measurement profile of AlN molar fraction obtained by the CL method of the sample. [Fig. 14] A CL spectrum of the first and second at the six Y coordinates on the measurement profile of the sample shown in Fig. 13. [Fig. 15] HAADF-STEM images of the measurement areas A to D of the four locations of the AlN molar fraction in the layered region of the sample, shown by line analysis of the profile TEM-EDX. [Fig. 16] A schematic cross-sectional view showing a main portion of an example of the structure of a nitride semiconductor ultraviolet light-emitting element of the second embodiment. [Fig. 17] A schematic cross-sectional view showing a main portion of an example of the laminated structure including the main portion of the active layer of the nitride semiconductor ultraviolet light-emitting element shown in Fig. 16. [Fig. 18] A schematic cross-sectional view showing a main portion of an example of the structure of a general ultraviolet light-emitting diode.

1: Nitride semiconductor ultraviolet light-emitting element

10: Substrate Section

11: Sapphire substrate

11a: Main surface of sapphire substrate

12: AlN layer

20: Light-emitting element structure section

21: n-type coating layer (n-type layer)

21a: Layered Domain (n-type Layer)

21b: n-type body domain (n-type layer)

22: Active layer

23: Electron barrier layer (p-type layer)

24: P-type connection layer (P-type layer)

26:p electrode

27:n electrode

220: Trap Layer

221: Barrier Layer

R1: First Domain

R2: Second Domain

Claims (15)

A nitride semiconductor ultraviolet light-emitting device comprises an n-type layer, an active layer, and a p-type layer formed by an AlGaN-based semiconductor with a zincblende structure, stacked in the vertical direction of the light-emitting device structure. The characteristic feature is that the aforementioned n-type layer is composed of an n-type AlGaN-based semiconductor. The aforementioned active layer, disposed between the n-type layer and the p-type layer, has a quantum well structure comprising one or more well layers composed of AlGaN-based semiconductors. The aforementioned p-type layer is composed of a p-type AlGaN-based semiconductor. Composed of lGaN-based semiconductors, each semiconductor layer within the aforementioned n-type layer, active layer, and p-type layer has an epitaxial growth layer forming a multi-segmented platform surface parallel to the (0001) plane. The aforementioned n-type layer has layered regions with locally low AlN molar fractions dispersed within it. The extension directions of these layered regions on a first plane orthogonal to the upper surface of the aforementioned n-type layer have a portion inclined relative to the intersection of the upper surface of the aforementioned n-type layer and the first plane. When n is 6 or 7, the first average AlN molar fraction Xna1, which extends throughout the depth direction of the aforementioned n-type layer, falls within the range of (n-0.25)/12 < Xna1 < (n+0.25)/12. The second average AlN molar fraction Xna2(d) at depth d above the aforementioned n-type layer varies corresponding to the aforementioned depth d. Within the aforementioned n-type layer, at least one specific depth at which Xna2(d) = n/12 or higher is observed. At the aforementioned specific depth, there exists a region on the upper side where Xna2(d) < n/12, and a region on the lower side where Xna2(d) > n/12. Within these layered regions, an intermediate AlGaN region with an AlN molar fraction of (n-0.5)/12 is formed. The aforementioned second average AlN molar fraction Xna2(d) falls within the range of (n-0.25)/12 < Xna2(d) < (n+0.25)/12 across the entire depth direction of the aforementioned n-type layer. As described in claim 1, the average AlN molar fraction Xna3, extending over the region from the upper end of the aforementioned n-type layer to the aforementioned specific depth, is in the range of (n-0.25)/12 < Xna3 < (n+0.25)/12. As described in claim 1 or 2, the aforementioned average AlN molar fraction Xna2(d) is within the range of Xna2(d) ≤ n/12 in the region from the upper end of the aforementioned n-type layer to the aforementioned specific depth. As described in claim 3, the average AlN molar fraction Xna2(d) of the aforementioned second step is within the range of Xna2(d) ≥ n/12 in a region deeper than the previously described specific depth of the n-type layer. As described in claim 1 or 2, the nitride semiconductor ultraviolet light-emitting element, wherein the aforementioned integer n is 7, and the peak emission wavelength is set to a specific value within the range of 280 nm to 315 nm. As described in claim 1 or 2, the nitride semiconductor ultraviolet light-emitting element, wherein the aforementioned integer n is 6, and the peak emission wavelength is set to a specific value within the range of 300nm to 330nm. As described in claim 1 or 2, the nitride semiconductor ultraviolet emitting element comprises an active layer having a multiple quantum well structure including two or more of the aforementioned well layers, with a barrier layer made of AlGaN-based semiconductor existing between the two aforementioned well layers. The nitride semiconductor ultraviolet light-emitting element as described in claim 1 or 2 further comprises a substrate portion including a sapphire substrate. The sapphire substrate has a main surface tilted at a specific angle relative to the (0001) plane. The light-emitting element structure is formed above this main surface. Each semiconductor layer from the aforementioned main surface of the sapphire substrate to the aforementioned p-type layer has an epitaxial growth layer with a surface forming a multi-segmented platform parallel to the (0001) plane. A method for manufacturing a nitride semiconductor ultraviolet light-emitting element comprises an n-type layer, an active layer, and a p-type layer formed by an AlGaN-based semiconductor with a zincblende structure, which are deposited in the upper and lower light-emitting element structure. The method is characterized by: for the (0001) plane, on a sapphire substrate including a main surface having a specific tilt angle, the aforementioned n-type layer of the n-type AlGaN-based semiconductor is epitaxially grown; and on the surface of the aforementioned n-type layer, multiple segments parallel to the (0001) plane are observed. The first process of the platform involves epitaxial growth of a quantum well structure, comprising one or more well layers composed of AlGaN semiconductors, on the aforementioned n-type layer to form the aforementioned active layer. The second process involves the epitaxial growth of a p-type AlGaN semiconductor p-type layer on the surface of the aforementioned well layer, exhibiting a multi-segmented platform parallel to the (0001) plane. The third process involves epitaxial growth of a p-type AlGaN semiconductor p-type layer on the aforementioned active layer. In the first process, the integer n is 6 or 7, and the first average AlN molar fraction Xna is distributed throughout the depth direction of the aforementioned n-type layer. Within the range of (n-0.25)/12 < Xna1 < (n+0.25)/12, the second average AlN molar fraction Xna2(d) at depth d above the aforementioned n-type layer varies corresponding to the aforementioned depth d. Within the aforementioned n-type layer, at least one specific depth at which Xna2(d) = n/12 or higher exists, and below the aforementioned specific depth, there exists an upper region where Xna2(d) < n/12, and a lower region where Xna2(d) > n/12. Within the n-type layer, layered regions with low local AlN molar fractions are dispersed and extend obliquely upwards to form an intermediate AlGaN region with an AlN molar fraction of (n-0.5)/12 within the aforementioned layered regions, thus forming the aforementioned n-type layer. In the first process described above, the average AlN molar fraction Xna2(d) of the second process is such that the aforementioned n-type layer is formed within the range of (n-0.25)/12 < Xna2(d) < (n+0.25)/12 across the entire depth direction of the aforementioned n-type layer. As described in claim 9, in the method for manufacturing a nitride semiconductor ultraviolet light-emitting element, the aforementioned n-type layer is formed such that the third average AlN molar fraction Xna3, extending from the upper end of the aforementioned n-type layer to the aforementioned specific depth, falls within the range of (n-0.25)/12 < Xna3 < (n+0.25)/12. As described in claim 9 or 10, in the method of manufacturing a nitride semiconductor ultraviolet light-emitting element, wherein, in the aforementioned first process, the aforementioned average AlN molar fraction Xna2(d) in the aforementioned second process is within the range of Xna2(d) ≤ n/12 in the region from the upper end of the aforementioned n-type layer to the aforementioned specific depth, the aforementioned n-type layer is formed. As described in claim 11, in the method for manufacturing a nitride semiconductor ultraviolet light-emitting element, the average AlN molar fraction Xna2(d) of the aforementioned second step in the first process is formed within the range of Xna2(d) ≥ n/12 in a region deeper than the previously described specific depth of the n-type layer. As described in claim 9 or 10, in the method for manufacturing a nitride semiconductor ultraviolet light-emitting element, wherein the aforementioned integer n is 7, and in the aforementioned second step, the peak emission wavelength of the aforementioned nitride semiconductor ultraviolet light-emitting element is made to a specific value within the range of 280 nm to 315 nm, thereby forming the aforementioned active layer. As described in claim 9 or 10, in the method for manufacturing a nitride semiconductor ultraviolet light-emitting element, wherein the aforementioned integer n is 6, and in the aforementioned second step, the peak emission wavelength of the aforementioned nitride semiconductor ultraviolet light-emitting element is made to a specific value within the range of 300 nm to 330 nm, thereby forming the aforementioned active layer. As described in claim 9 or 10, in the method for manufacturing a nitride semiconductor ultraviolet light-emitting element, in the second process described above, the aforementioned active layer of a multiple quantum well structure comprising two or more of the aforementioned well layers is formed by layering a well layer composed of an AlGaN-based semiconductor and a barrier layer composed of an AlGaN-based semiconductor through alternating epitaxial growth.