Ultraviolet light-emitting diodes and electrical equipment equipped therewith

Abstract

In order to improve the luminous efficiency of an ultraviolet light emitting diode in a Far-UVC wavelength range, an ultraviolet light emitting diode 100A according to one embodiment of the present disclosure contains an AlGaN crystal or an InAlGaN crystal, while comprising a light emitting layer 134, a spacer layer 136A and an electron blocking layer 138, which are stacked upon each other in the direction of the electron flow; and the Al composition ratio in the spacer layer changes corresponding to the position in the thickness direction of the stacked layers. This ultraviolet light emitting diode according to one embodiment of the present disclosure is simultaneously or optionally provided with a light emitting layer 134 which comprises at least one barrier layer 13B and at least two quantum well layers 13W that holds the barrier layer therebetween; and the barrier layer is formed thin.

Description

[Technical Field]

[0001] This disclosure relates to ultraviolet light-emitting diodes and electrical equipment comprising them. More specifically, this disclosure relates to ultraviolet light-emitting diodes that emit far ultraviolet light and electrical equipment comprising them. [Background technology]

[0002] Solid-state light-emitting devices utilizing nitride semiconductors are widely used, for example, as blue light-emitting diodes (LEDs). Solid-state light sources are also needed in the ultraviolet region, and ultraviolet light-emitting diodes (UVLEDs) using materials similar to those used for blue LEDs are being developed. The wavelength range below 350 nm in the ultraviolet region is also called the deep ultraviolet (DUV) region, and within that, approximately 200 nm to 280 nm is called the UVC wavelength band. A portion of this, the wavelength range of about 260 to 280 nm, is called the germicidal wavelength, and intensive technological development of UVLEDs for this wavelength range is underway. In recent years, the wavelength range of about 210 to 230 nm has also been called far ultraviolet (Far-UVC) and is attracting particular attention. Far-UVC offers germicidal and virus inactivation capabilities while preventing adverse effects on the human body, so there is a demand for practical solid-state light sources in this wavelength range.

[0003] Generally, deep ultraviolet (DUV) LEDs are fabricated from AlGaN nitride semiconductor crystals. In AlGaN, including AlN and GaN, the band gap corresponding to the wavelength range of 210 nm (AlN) to 340 nm (GaN) can be achieved depending on the composition ratio of Al and Ga. Focusing solely on the principle, it is not impossible to manufacture LEDs that emit ultraviolet light within the 210 nm to 340 nm wavelength range.

[0004] In GaN / AlGaN / InGaN nitride semiconductor light-emitting devices, the properties of a layer also called the LQB (last quantum barrier) have been investigated. For example, it has been proposed to adopt a compositional gradient in the LQB for the electron blocking function in blue LEDs (Non-Patent Literature 1, Non-Patent Literature 2). These proposals investigate the effect of the compositional gradient of the LQB on the function of the electron blocking layer (EBL) from the viewpoint of electron-blocking properties and hole injection efficiency. Similar attempts have been made in ultraviolet LEDs that emit near-ultraviolet light at a wavelength of approximately 360 nm (Non-Patent Literature 3), and theoretical investigations have also been conducted for wavelengths of approximately 290 nm in the DUV region (Non-Patent Literature 4). The role that the thickness of the LQB plays in the DUV region has also been investigated (Non-Patent Literature 5). Furthermore, methods for introducing a superlattice into the LQB of an LED element that emits light at 285 nm (Non-Patent Literature 6) and a configuration in which a high barrier layer is placed in addition to the EBL in an LED element that emits light at 270 nm have been reported (Non-Patent Literature 7).

[0005] Furthermore, in GaN / AlGaN / InGaN nitride semiconductor light-emitting devices, a method has been employed to increase internal quantum efficiency by employing a multiple quantum well structure, thereby increasing the overlap of electron and hole distributions through wave function confinement and generating a large number of such electron-hole pairs. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] Chang Sheng Xia et al, "Advantages of GaN-based LEDs with two-step graded AlGaN last quantum barrier", Opt Quant Electron (2016) 48:509 DOI: 10.1007 / s11082-016-0785-6 [Non-Patent Document 2] Liwen Cheng et al, "Electron Confinement and Hole Injection Improvement in InGaN / GaN Light-Emitting Diodes With Graded-Composition Last Quantum Barrier and Without Electron Blocking Layer", Journal of Display Technology, Vol. 11, No. 9, 2015, DOI: 10.1109 / JDT.2015.2437454 [Non-Patent Document 3] Longfei He et al, "Performance enhancement of AlGaN-based 365 nm ultraviolet light-emitting diodes with a band-engineering last quantum barrier", Optics Letters Vol. 43, No. 3, pp. 515-518 (2018), DOI: 10.1364 / OL.43.000515 [Non-Patent Document 4] Longfei He et al, "Marked enhancement in the efficiency of deep ultraviolet light-emitting diodes by using a AlxGa1-xN carrier reservoir layer", Applied Physics Express 12, 062013 (2019), DOI: 10.7567 / 1882-0786 / ab22df [Non-Patent Document 5] Xianglong Bao et al, "Performance Improvements for AlGaN-Based Deep Ultraviolet Light-Emitting Diodes With the p-Type and Thickened Last Quantum Barrier", IEEE Photonics Journal, Volume: 7, Issue: 1, 2015; DOI: 10.1109 / JPHOT.2014.2387253 [Non-Patent Document 6] Qian Chen et al, "Improved the AlGaN-Based Ultraviolet LEDs Performance With Super-Lattice Structure Last Barrier", IEEE Photonics Journal, Volume: 10, Issue: 4, 2018; DOI: 10.1109 / JPHOT.2018.2852660 [Non-Patent Document 7] Ravi Teja Velpula et al, "Improving carrier transport in AlGaN deep-ultraviolet light-emitting diodes using a strip-in-a-barrier structure", Applied Optics, Vol. 59, Issue 17, pp. 5276-5281, 2020; DOI: 10.1364 / AO.394149 [Overview of the project] [Problems that the invention aims to solve]

[0007] The luminous efficiency of light-emitting devices in the actual UVC wavelength range (200nm~280nm) decreases exponentially as the emission wavelength shortens. Far-UVC (210~230nm) light-emitting diodes (called "Far-UVCLEDs") have only achieved an extremely low external quantum efficiency of less than 0.03%. To improve the luminous efficiency of Far-UVCLEDs that emit light at any wavelength in the Far-UVC region, applying the technological concepts used to improve efficiency at longer wavelengths than Far-UVC will not yield good results. One of the fundamental reasons for this is that, given the high Al composition ratio of AlGaN (AlGaN) of 0.8 or more in Far-UVCLEDs, there is little room to increase the Al composition ratio. For example, at emission wavelengths of around 260~280nm, AlGaN with an Al composition ratio of around 0.6~0.7 has been used. However, in the Far-UVC region, the Al composition ratio of AlGaN must be 0.8 or higher. As a result, the difference from the upper limit of the Al composition ratio, 1.0 (i.e., AlN), is small, and methods that rely solely on increasing the Al composition ratio are insufficient. More specifically, the method of increasing the Al composition ratio, which has been used in light-emitting diodes that emit light at long wavelengths, is insufficient for the function of electron blocking as a means of suppressing serious electron overflow.

[0008] Another fundamental reason for the low luminous efficiency of Far-UVCLEDs is that AlGaN with a high Al composition ratio undergoes a significant change in electronic structure compared to AlGaN with a low Al composition ratio. AlGaN exhibits a drastic change in properties above an Al composition ratio of approximately 0.5. Specifically, it is known that in ultraviolet light-emitting diodes where AlGaN has an Al composition ratio exceeding 0.5, TM emission becomes dominant over TE emission. TM emission is more difficult to extract than TE emission. Nevertheless, if the emission wavelength is 240 nm or longer, a method of increasing the ratio of TE emission by creating quantum wells in the light-emitting layer and utilizing the quantum confinement effect has been effective. However, in Far-UVCLEDs, even with the use of quantum wells, it is not possible to sufficiently increase the ratio of TE emission.

[0009] The present disclosure aims to solve at least one of the above problems. To improve the luminous efficiency of Far-UVC LEDs, it can be said that a new design concept suitable for the material characteristics of AlGaN with a high Al composition is essential. The present disclosure contributes to the development of various applications that employ Far-UVC LEDs as a light source by providing a new design concept that can also be adapted to the properties of AlGaN with a high Al composition employed in Far-UVC LEDs.

Means for Solving the Problem

[0010] The inventors of the present invention have found that, on the premise of the inherent constraints of AlGaN with a high Al composition and its material characteristics, the efficiency of Far-UVC LEDs can be increased by making full use of precise control (band engineering) of the electron band structure, and have completed the invention related to the present application.

[0011] As a result of intensively and theoretically studying the details of the electron injection operation, the inventors of the present invention have found that overflow can be suppressed by devising a layer (spacer layer) disposed immediately before the electron blocking layer in the electron flow, and have confirmed the effect through experiments. That is, in an embodiment of the present disclosure, there is provided an ultraviolet light-emitting diode including an AlGaN-based crystal or an InAlGaN-based crystal, which is provided with a light-emitting layer, a spacer layer, and an electron blocking layer laminated in this order from the upstream to the downstream of the electron flow, and the ultraviolet light-emitting diode in which the Al composition ratio in the spacer layer changes according to the position in the thickness direction of the lamination.

[0012] In addition, the inventor has predicted through theoretical calculations that if the barrier layer sandwiched between a plurality of quantum well layers is thinned, an electron band structure advantageous for TE emission will be realized, and has actually confirmed an increase in luminous efficiency through experiments. That is, in an embodiment of the present disclosure, there is provided an ultraviolet light-emitting diode including an AlGaN-based crystal or an InAlGaN-based crystal, which is provided with a light-emitting layer including at least one barrier layer and at least two quantum well layers sandwiching the barrier layer, and the ultraviolet light-emitting diode in which the barrier layer is thinned.

[0013] Furthermore, in the present disclosure, it is also preferable to simultaneously adopt making the spacer layer have a compositional gradient and thinning the barrier layer.

[0014] In these ultraviolet light-emitting diodes, it is preferable that the compositional distribution of the spacer layer is inclined such that the Al composition ratio decreases from the light-emitting layer toward the electron blocking layer. Furthermore, it is also preferable that the compositional distribution of the spacer layer is inclined such that the Al composition ratio increases from the light-emitting layer toward the electron blocking layer. Also, in these ultraviolet light-emitting diodes, it is preferable that the thickness of the barrier layer is 0.2 nm or more and 4 nm or less, and it is more preferable that the thickness of the barrier layer is 1 nm or more and 3 nm or less. The thickness of the barrier layer in these ultraviolet light-emitting diodes is preferably a thickness that makes the light emission in the light-emitting layer such that the TE emission is stronger than the TM emission, and it is also preferable that it is not more than the thickness of the quantum well layer. In these ultraviolet light-emitting diodes, it is preferable that the main wavelength of the emitted ultraviolet light is 210 to 230 nm. Furthermore, in an embodiment of the present disclosure, an electric device including the above-described ultraviolet light-emitting diode as an ultraviolet light emission source is also provided.

[0015] In this application, ultraviolet light in the Far-UVC region generally refers to ultraviolet light in the wavelength range of approximately 210-230 nm. The "major wavelength of emitted ultraviolet light" is generally the wavelength that characterizes the emission spectrum of a light-emitting diode, which is not necessarily a single wavelength, and typically includes the peak wavelength of a single-peak, bell-shaped emission spectrum. However, the fact that a wavelength range is described for the major wavelength does not mean that the wavelength range described for that major wavelength should encompass the entire emission spectrum. Furthermore, the description in this application may use technical terms adapted or borrowed from the fields of electronic devices and physics dealing with visible light and ultraviolet light to explain the device structure and function. For this reason, even when describing electromagnetic waves in the ultraviolet region (ultraviolet light) that are not visible light, terms such as "photon," "emission," and further terms such as "optical," "photo," etc., may be used to explain the operation of LEDs and radiation phenomena. The light-emitting layer includes a quantum well layer and a barrier layer. The quantum well layer is the layer that provides electrons with the conduction band edge potential that forms a quantum well, and the barrier layer is the layer that provides a relatively high conduction band edge potential in relation to the quantum well layer. The electron blocking layer is a layer with a high conduction band edge potential provided to prevent electron leakage, and the spacer layer is placed between the quantum well layer and the electron blocking layer, which are the downstream layers in the electron flow of the light-emitting layer, and is the layer that has the conduction band edge potential height between them. [Effects of the Invention]

[0016] In any embodiment of this disclosure, ultraviolet light-emitting diodes achieve light emission operation with higher efficiency than conventional diodes. [Brief explanation of the drawing]

[0017] [Figure 1] Figure 1 is a perspective view showing a schematic configuration common to both conventional LEDs and LEDs of the embodiments of this disclosure. [Figure 2] Figure 2 is an explanatory diagram illustrating the problems solved by this embodiment of the disclosure, and shows the electron conduction band edge profiles of the n-type layer, light-emitting layer, spacer layer, and electron blocking layer. [Figure 3] Figure 3 is a schematic diagram showing the relationship between the position in the thickness direction of the stacking of the barrier layer, quantum well layer, spacer layer, and electron block layer in the conventional LED and the LED of the embodiment of this disclosure, and the Al composition ratio (vertical direction). [Figure 4A-C] Figures 4A-C are explanatory diagrams (Figure 4A) showing the thickness-direction distribution of the Al composition ratio of the structure used in the calculations for the LEDs of the embodiments of this disclosure, a graph of the current-internal quantum efficiency obtained by calculation (Figure 4B), and a graph of the injection efficiency obtained by changing the composition gradient (Figure 4C). [Figure 5A-C] Figures 5A-C are illustrative band diagrams for illustrating the calculation of the block height of the LED in the embodiments of this disclosure, showing the band diagram during operation (Figure 5A) and enlarged views of the band diagrams in the portion that gives the block height of electrons and holes (Figures 5B and 5C). [Figure 6A-B] Figures 6A-B show the calculated block heights for electrons (Figure 6A) and holes (Figure 6B) for Far-UVCLEDs in the cases of flat (conventional), decreased composition, and increased composition (all embodiments of the present disclosure). [Figure 7] Figure 7 is a graph of the luminescence recombination rates calculated at each position in the thickness direction of the LED in the embodiment of this disclosure. [Figure 8A-B] Figures 8A and 8B show the measured electroluminescence (EL) spectra (Figure 8A) and current-light output characteristics (Figure 8B) of LED samples from conventional and embodiments of the present disclosure. [Figure 9] Figure 9 shows the EL spectra obtained from each sample with a different starting Al composition ratio in the LED according to the embodiment of this disclosure. [Figure 10A-B] Figures 10A and 10B are band diagrams illustrating the polarization of UV emitted by a conventional LED and an LED according to the embodiments of this disclosure, respectively. [Figure 11] Figure 11 is a graph showing measured emission spectra of samples of conventional LEDs and LEDs according to embodiments of this disclosure. [Figure 12] Figure 12 is a graph showing the relationship between the thickness of the barrier layer and the intensity of TE emission, obtained in a simulation to confirm the technical concept of the embodiment of this disclosure. [Figure 13A-B] Figures 13A and 13B are graphs of the emission spectrum (Figure 13A) and current-optical output characteristics (Figure 13B) of an LED sample in an embodiment of this disclosure. [Modes for carrying out the invention]

[0018] The Far-UVC light-emitting diode (Far-UVC LED, also referred to as LED) according to this disclosure will be described below. Unless otherwise specified in this description, common parts or elements are denoted by common reference numerals. In addition, in the figures, the elements of each embodiment are not necessarily shown while maintaining their relative scales.

[0019] 1. Embodiment In the LED 100A of this embodiment, electron overflow is suppressed by employing a compositionally graded spacer layer 136A. Furthermore, in the LED 100A of this embodiment, a band structure favorable to TE light emission is realized by thinning the barrier layer 13B located between multiple quantum well layers 13W, either in combination with or independently. The structure of an LED employing both the compositionally graded spacer layer 136A and the thinned barrier layer 13B will be described below (1-1), then the details of the compositionally graded spacer layer will be described (1-2), the details of the thinned barrier layer will be described (1-3), combinations thereof will be described (1-4), and finally, modified examples will be described (1-5).

[0020] 1-1. Structure of an LED Figure 1 is a perspective view showing a schematic configuration common to both the conventional LED 100 and the LED 100A of this embodiment. In the typical configuration of LED 100 and 100A, a buffer layer 120 is epitaxially grown on one surface 104 of a substrate 110 which is a flat α-Al2O3 single crystal (sapphire) using a material such as AlN crystal. An n-type conductive layer 132 is then stacked and formed in this order from the buffer layer 120 side.

[0021] In a conventional LED 100, the n-type conductive layer 132 is followed by a light-emitting layer 134, a spacer layer 136, an electron-blocking layer 138, a p-type contact layer 150, and an electrode 160 acting as a second electrode, all stacked in this order. The direction of electron flow from upstream to downstream during operation is also the same as this stacking order. The materials of the n-type conductive layer 132 to the spacer layer 136 are typically AlGaN or InAlGaN, or one of them doped with trace elements (Si for n-type, Mg for p-type, etc.) as needed. The electron-blocking layer 138 is a single layer that provides a high barrier to electrons, and its material is AlGaN with a high Al composition ratio, or AlN. The first electrode 140 is electrically connected to the n-type conductive layer 132. The electrode 160 establishes an electrical connection with the electron-blocking layer 138 via the p-type contact layer 150. The light output L, which is emitted UV light, is emitted from the other side, the light extraction surface 102, through the substrate 110.

[0022] The structure of each layer will be described in more detail. The substrate 110 is a growth substrate capable of epitaxially growing the n-type conductive layer 132 to the p-type contact layer 150, and is selected from any material that satisfies conditions such as crystal orientation and heat resistance for growth. When the substrate 110 is finally left, the substrate 110 is also required to have permeability to radiated UV. Typical materials for the substrate 110 include, in addition to the above-described α-Al2O3 single crystal (sapphire), AlN single crystal substrates, and Ga2O3 substrates in the case of radiated UV with a wavelength of 300 nm or more, and those having appropriate crystal orientation and off-angle according to each material are appropriately selected. When a material with expected conductivity such as a Ga2O3 substrate is adopted for the substrate 110, the arrangement of the first electrode 140 can be different from that in FIG. 1 as long as it can be electrically connected to the n-type conductive layer 132. The buffer layer 120 is formed as a single layer or multiple layers as necessary to satisfy the requirements for crystal growth of enhancing the internal light emission efficiency η IQE and is formed, for example, to a thickness of about 2 μm. In the n-type conductive layer 132 to the spacer layer 136, a typical configuration when an AlGaN layer is adopted is that the n-type conductive layer 132 is, for example, an Al 0.85 Ga 0.15 N layer doped with Si to be n-type, that is, an Al 0.85 Ga 0.15 N;Si layer. The light-emitting layer 134 has barrier layers 13B and quantum well layers 13W alternately laminated, and has a structure of a MQW (multiple quantum well) laminate of barrier layer 13B, quantum well layer 13W, barrier layer 13B,... quantum well layer 13W from the side of the n-type conductive layer 132. Therefore, at least two quantum well layers 13W are included in the light-emitting layer 134, and some of the barrier layers 13B are sandwiched between at least two quantum well layers 13W. The material of the light-emitting layer 134 is Al 0.94 Ga 0.06 N for the barrier layer 13B and Al 0.82 Ga 0.18The composition is made to be N. The typical number of quantum wells is, for example, around 3. The spacer layer 136 is an undoped AlGaN layer. A similar configuration can be adopted when an InAlGaN layer is used for the ultraviolet emitting layer 130.

[0023] The p-type contact layer 150 is p-type AlGaN or p-type InAlGaN, which is an AlGaN or InAlGaN material doped with Mg. By appropriately selecting the Al composition ratio, the p-type contact layer 150 can be given high transmittance to emitted UV. The first electrode 140 is a metal electrode with a Ni / Au laminated structure from the substrate side. This Ni is a layer, for example, 25 nm thick, inserted between the Au and the underlying semiconductor layer to achieve ohmic contact. For the second electrode, a UV reflective film 164 that exhibits high reflectivity to emitted UV is used for the reflective electrode 160. This UV reflective film 164 is a film made of a material mainly composed of Al, Mg, and Rh. For ohmic contact, Ni is also inserted on the substrate side of the reflective electrode 160, forming an inserted metal layer 162 that becomes part of the reflective electrode.

[0024] The LED 100A of this embodiment is the same as the conventional LED 100, except for the internal configuration of the light-emitting layer 134 and the specific configuration of the spacer layer 136A. In an LED employing a quantum well in a nitride semiconductor, electrons are injected from the n-type conductive layer 132 through the conduction band, and holes are injected from the p-type contact layer 150 through the valence band, into the quantum confinement state of the quantum well layer 13W formed in the light-emitting layer 134. The electrons and holes recombine in the quantum well through interband transitions and emit ultraviolet light. Among the mechanisms governing this radiation efficiency, the inventors focused on two issues for LEDs in the Far-UVC region: firstly, electron overflow, and secondly, the polarization state of the light emission.

[0025] Figure 2 is an explanatory diagram illustrating the problems solved by this embodiment. The thickness direction of the stacking is oriented in the left-right direction of the drawing, and the electron conduction band edge profiles of each layer from left to right are shown: the n-type conductive layer 132, the light-emitting layer 134, the spacer layer 136, and the electron blocking layer 138. Overflow is a phenomenon in which some carriers pass through the light-emitting layer 134, and current is wasted without the intended recombination occurring. This is a problem for electrons, which exhibit high mobility in nitride semiconductor carriers. To suppress electron overflow, in nitride semiconductor LEDs, a high barrier layer called the electron blocking layer 138 is generally provided at a position downstream of the electron flow relative to the light-emitting layer 134. However, in Far-UVC LEDs, even if the Al composition ratio is increased to raise the conduction band edge and thus increase the barrier for the electron blocking layer 138, the Al composition ratio can only be increased to its upper limit (100%, i.e., AlN). As a result, the conduction band edge of the electron blocking layer 138 does not become sufficiently high relative to the light-emitting layer 134 made of AlGaN with a high Al composition, making it difficult to perform the electron-blocking function. As a result of not being able to sufficiently suppress overflow, even if a Far-UVC LED is made with the conventional LED 100 structure, the luminous efficiency of the LED decreases due to reactive current that does not contribute to light emission.

[0026] The polarization state of the emitted light depends on whether the emitted ultraviolet light is in TM (transverse magnetic) mode or TE (transverse electric) mode. When the electric field oscillates in the thickness direction of the light-emitting layer 134, it is called TM mode, and when the electric field oscillates in the plane of the light-emitting layer 134, it is called TE mode. In this application, TE mode light and the radiation operation that emits that light are also referred to as TE light and TE emission, respectively, and the same applies to TM. Optical transitions that emit TM light have a profile that radiates in the plane of the laminated structure, such as the quantum well layer 13W and the barrier layer 13B. For this reason, they are easily scattered or absorbed as they propagate inside an LED of about mm in size, and tend to attenuate before being emitted to the outside. In contrast, optical transitions that emit TE light have a profile where the radiation direction is oriented in the thickness direction of the laminated structure. For this reason, TE light is easily extracted from the LED, either by being directly emitted to the outside or by having its propagation direction reversed and emitted with the help of, for example, the reflective electrode 160. Increasing the proportion of TE light is particularly advantageous for improving the light extraction efficiency. However, when the Al composition ratio is increased to 0.8 or higher in AlGaN for the Far-UVC region, TM emission becomes dominant due to the influence of the band structure, particularly the influence of the hole-forming electronic state, resulting in a decrease in light extraction efficiency.

[0027] 1-2. Suppression of electron overflow by composition-graded spacer layers In the LED 100A of this embodiment, the Al composition ratio of the spacer layer 136A is continuously varied according to its position in the thickness direction. This continuous variation is typically sloped to increase or decrease depending on the position, and is therefore called "composition slope" in this embodiment. Figure 3 is a schematic diagram showing the relationship between the position in the thickness direction of the stacking of the barrier layer 13B, quantum well layer 13W, spacer layers 136, 136A, and electron block layer 138 and the Al composition ratio. The horizontal and vertical directions in the drawing represent the position in the stacking direction and the AlN composition ratio at each position in the LEDs 100 and 100A, respectively. The spacer layers 136 and 136A are sandwiched between the downstreammost of the multiple quantum well layers 13W in the direction of electron flow (to the right in the drawing) and the electron block layer 138. The Al composition ratio is small in the quantum well layer 13W and large in the electron block layer 138, while the spacer layers 136 and 136A are configured to have a value between that of the quantum well layer 13W and the electron block layer 138. The spacer layer 136 of the conventional LED 100 is manufactured to have an Al composition ratio that is distributed flatly regardless of the position in the thickness direction (dashed line 136F in the figure). In contrast, the spacer layer 136A of the LED element 100A of this embodiment has a gradient 136U that increases the Al composition ratio or a gradient 136D that decreases it depending on the position in the thickness direction from the light-emitting layer 134 toward the electron block layer 138. In this application, the configuration in which the Al composition ratio has gradients that increase and decrease in the direction from the light-emitting layer 134 toward the electron block layer 138 is also referred to as composition increase and composition decrease, respectively. Note that the conduction band edge profile of the spacer layer 136 shown in Figure 2 may experience piezoelectric polarization at the position where the composition changes due to the polarity of the crystal. As a result, the conduction band edge profiles of spacer layers 136 with a flat Al composition ratio also generally exhibit slope and bending.

[0028] To confirm the effect of the compositionally graded spacer layer 136A, the inventors investigated the electron injection efficiency through theoretical calculations and then confirmed the actual light emission operation through experiments. The theoretical calculations were performed using the simulation software SiLENSe (Semiconductor Technology Research (STR), Saint Petersburg, Russia). The calculations were performed using a realistic configuration, and the typical conditions for both the conventional LED 100 and the LED 100A of this embodiment were as follows: the quantum well layer 13W had a thickness of 3 nm, an Al composition ratio of 0.82, and 4 layers. The electron blocking layer 138 had a thickness of 9 nm and an Al composition ratio of 1.0 (i.e., AlN). In the conventional flat spacer layer 136 for LED100, the Al composition ratio was kept constant at 0.94. In the composition-graded spacer layer 136A for LED100A of this embodiment, the Al composition ratio was configured to increase linearly from 0.94 to 1.0 depending on the position in the thickness direction from the quantum well layer to the electron block layer, and to decrease linearly from 0.94 to 0.82. These calculations reflect the band gradient and bending effects due to polarization when a polar substrate is used. The effect of electron leakage was evaluated by the calculation results of the electron injection efficiency.

[0029] Figures 4A-C are explanatory diagrams (Figure 4A) showing the thickness-direction distribution of the Al composition ratio of the structure used in the calculations for the LED element of this embodiment, a graph of the current-internal quantum efficiency obtained by calculation (Figure 4B), and a graph of the injection efficiency obtained by changing the gradient of the composition (Figure 4C). In Figures 4A and B, "Flat (-)", "Graded (-)", and "Graded (+)" are denoted for flat (i.e., spacer layer 136), a composition decrease configuration in which the Al composition ratio x decreases from 0.94 to 0.82, and a composition increase configuration in which it increases from 0.94 to 1.0, respectively. Note that the Al composition ratio x is Al x Ga 1-x N, or (AlN) x (GaN) 1-xThis is a fractional value that can also be expressed as [this]. As shown in Figure 4B, in all configurations, the internal quantum efficiency increased with increasing current, exhibiting an upward-sloping characteristic. Among these, the configuration with a gradient showed higher internal quantum efficiency compared to the flat composition distribution, regardless of whether the gradient was increasing or decreasing. In other words, the spacer layer 136A with a gradient achieved recombination as intended, more so than the spacer layer 136 with a flat composition. Figure 4C shows a more detailed investigation into what degree of gradient should be applied to the composition distribution. Here, the Al composition ratio of the spacer layer 136A was fixed at 0.94 at the position in contact with the quantum well layer 13W (referred to as the "starting Al composition ratio"), as before, and the value at the position reaching the electron block layer 138 (referred to as the "reached Al composition ratio") was set to x, and the injection efficiency for x was calculated. Note that the reached Al composition ratio x = 0.94 corresponds to the configuration of the flat spacer layer 136. As shown in the figure, the injection efficiency was minimal in the flat spacer layer 136, and the injection efficiency increased with compositional gradients, regardless of whether the gradient increased or decreased.

[0030] Based on common technical knowledge, it can be said that compositional decrease and compositional increase in a compositional gradient would have opposite effects. However, the simulation results show a more complex situation, suggesting that multiple mechanisms are at work. The inventors believe that the following multiple physical mechanisms acting from a structural perspective may contribute to the improvement of properties: A) In compositional gradients, the electron block height is substantially increased compared to flat configurations, and electron overflow is suppressed. B) In compositional gradients, compared to flat configurations, the hole block height is reduced, resulting in increased hole injection efficiency. C) In compositional gradients, compared to flat configurations, carriers accumulated at the interface between the spacer layer and the electron block layer contribute to luminescence. This effectively acts as if an additional quantum well is added. This effective quantum well is called an "interface quantum well (IQW)". In reality, it is predicted that at least one of these physical mechanisms is at work, and that it is possible that several of them are working simultaneously.

[0031] Therefore, in order to investigate the extent of the influence of each mechanism AC, the block heights of electrons and holes were calculated from the band diagram during 100mA operation. Figures 5A-C are illustrative band diagrams to explain the calculation of the block height of the LED in this embodiment, showing the band diagram during 100mA operation (Figure 5A) and enlarged views of the band diagram that gives the block heights of electrons and holes (Figures 5B and 5C). Figures 6A-B show the calculated block heights for electrons (Figure 6A) and holes (Figure 6B) for conventional and this embodiment's Far-UVC LEDs, in the cases of flat, decreased composition, and increased composition. This block height represents the energy difference between the pseudo-Fermi level and the maximum potential of the block layer 138 in eV (or meV). In Figures 5A-C, the pseudo-Fermi level is shown by the dashed line, and the potentials for electrons and holes are shown by the solid lines. In Figures 5B-C, the pseudo-Fermi level is shown by E Fn (Electrons) and E Fp The sign of (hole) is specified, and the potential is the conduction band edge E. C and valence band edge E V The signs are assigned to each. The energy difference is determined by measuring the peak (electrons) or bottom (holes) of the potential in the range of position (z=340~346nm) of block layer 138 from the pseudo-Fermi level. That is, the pseudo-Fermi level E for electrons flowing from upstream to downstream (left to right on the page) Fn (In the example in Figure 5B, the potential E is 0.0 eV) C It is determined as the difference leading to the maximum value (same, 0.26 eV). Similarly, for holes, the pseudo-Fermi level E for the hole flows from upstream to downstream (right to left on the page). Fp (Same, -5.58eV) from the potential E VIt is determined as the difference leading to the minimum value (same, -6.08 eV). The absolute value of these differences becomes the block height for each carrier. For electrons, a large block height is preferable due to the function of the electron blocking layer, but a small value is preferable for hole injection. From Figures 6A and 6B, compared with a flat composition, the electron block height increases and the hole block height decreases with decreasing composition. In other words, decreasing composition improves efficiency for both electron and hole carriers, so that electron blocking functions as intended and hole injection is smooth. In contrast, with increasing composition, the electron block height decreases, and the hole block height decreases even more. In other words, compared with a flat composition, electron blocking becomes insufficient with increasing composition, but the simultaneously improved hole injection improves injection efficiency even more.

[0032] Furthermore, the luminescence recombination rate at each well was calculated. Figure 7 is a graph of the luminescence recombination rates calculated at each position in the thickness direction of the LED 100A in this embodiment. The symbols QW1 to QW4 are identifiers assigned to the quantum well layer 13W in the stacking order toward the electron block layer 138, and IQW refers to the interfacial quantum well substantially occurring at the interface between the spacer layer 136A and the electron block layer 138. The symbols "Flat (-)", "Graded (-)", and "Graded (+)" are the same as in Figures 4A and B, and the Al composition ratio was 0.94 for flat, 0.94 for composition decrease and 0.82 for destination Al composition, and 0.94 for composition increase and 1.0 for destination Al composition. From Figure 7, it can be seen that the luminescence rate at each well is higher than that of the flat layer in both composition decrease and composition increase. In addition, the luminescence component of the interfacial quantum well (IQW) is also larger in the composition decrease.

[0033] Figures 4A-C, 5A-C, 6A-B, and 7 show that composition reduction exhibits superior characteristics compared to the flat structure for all mechanisms A, B, and C. The emission peak wavelength of the IQW depends on the achieved Al composition ratio x during composition reduction. If the achieved Al composition ratio x falls below 80% and the IQW emission peak wavelength becomes separated from the wavelengths of other quantum wells, it may result in characteristics unsuitable for LEDs. Whether or not to incorporate the IQW into the emission region, and whether the IQW emission peak wavelength is appropriate, are determined according to the application. The optimal value and range of the achieved Al composition ratio x can be determined from the perspective of the necessity and appropriateness of IQW emission. For example, setting the achieved Al composition ratio x to achieve an IQW emission peak wavelength that matches the emission wavelength of the emission layer 134 is a typical example of this optimization. Another typical example of this optimization is adjusting the IQW emission peak wavelength so that the suppression of adverse effects on the human body, which is one of the characteristics of Far-UVC, is maintained.

[0034] As described above, theoretical calculations showed that the electron injection efficiency, as a relative ratio to the flat spacer layer 136, was approximately 1.7 times for spacer layer 136A with a composition gradient that increased the Al composition ratio, and approximately 2.1 times for spacer layer 136A with a composition gradient that decreased the Al composition ratio.

[0035] In the light emission operation verification experiment, LED samples were fabricated corresponding to the conventional and the LEDs 100 and 100A of this embodiment, with a flat spacer layer 136 and a spacer layer 136A having compositional gradients of increasing and decreasing Al composition ratios, and their electroluminescence was measured. The LED samples were fabricated using the MOVPE (metal-organic vapor phase epitaxy) method, similar to conventional nitride semiconductor ultraviolet LEDs. The only difference was that, in order to form the compositionally gradient spacer layer 136A, the ratio of the Al source gas TMAl (tri-methyl-aluminum) and the Ga source gas TMGa (tri-methyl-gallium) was changed to match whether the spacer layer had a flat Al composition or a compositional gradient. Other structural fabrication conditions were kept the same. The electrode size was 0.4 mm square, and steady-state current operation was performed. Figures 8A and 8B show the measured electroluminescence (EL) spectra (Figure 8A) and current-light output characteristics (Figure 8B) of conventional and this embodiment LED samples. As shown in Figure 8, a significant improvement in output was observed with decreasing composition compared to a flat result. On the other hand, increasing composition resulted in a slight decrease in output compared to a flat result. The differences from the simulation can be attributed to a wide range of factors, including differences in physical properties and structure.

[0036] Regarding the choice between increasing and decreasing the composition, the inventors believe that decreasing the composition is a more practical structure, offering greater ease of use and reproducibility, considering both tolerance to deviations from the ideal structure that can actually occur, such as structural and parameter fluctuations, and the ability to stably improve output. Furthermore, decreasing the composition can be advantageous from another perspective as well. Referring to the Al composition dependence of the injection efficiency obtained from simulation (Figure 4C), it can be seen that the efficiency is lowest when the Al composition ratio of the spacer layer is flat, and that the efficiency improves as the amount of change in composition (composition gradient) increases from there. However, in the case of actual Far-UVC LEDs with a wavelength of 230 nm, it is necessary to use AlGaN close to AlN in the light-emitting layer, so even if one tries to adopt increasing the composition, the achievable Al composition ratio plateaus at 100%. Conversely, with decreasing the composition, a wide range of achievable Al composition ratios can be selected, so it can be said that decreasing the composition offers greater room for efficiency improvement in short-wavelength LEDs using high Al compositions.

[0037] Furthermore, the inventors have confirmed through experimental results that the current structure is nearly optimal in terms of changes in spacer layer thickness and the resulting Al composition ratio. This was confirmed by more comprehensively changing the parameters through the experiments described below.

[0038] First, we investigated the dependence of the spacer layer 136 thickness (spacer layer thickness) on the adoption of a composition reduction. As a result, when the initial Al composition ratio was fixed at 0.94 and the final Al composition ratio at 0.82, and the thickness of the spacer layer 136 was varied to 3, 6, and 9 nm, the optical output was 0.15 mW, 0.48 mW, and 0.33 mW, respectively.

[0039] Next, we investigated the dependence of the target Al composition ratio on the initial Al composition ratio being fixed at 0.94. As a result, in the range where the target Al composition ratio x was varied to 0.78, 0.82, and 0.86, the optical output was 0.20 mW, 0.48 mW, and 0.10 mW, respectively.

[0040] Furthermore, the dependence of the initial Al composition ratio on the target Al composition ratio, when fixed at 0.82, was investigated. As a result, in the range where the initial Al composition ratio was varied to 0.90 and 0.94, the light output was 0.45 mW and 0.90 mW, respectively. Figure 9 shows the EL spectra obtained from samples of the LED of this embodiment, where the initial Al composition ratio was varied to 0.90 and 0.94.

[0041] From this, we confirmed that a spacer layer 136A with a spacer layer thickness of 6 nm, an initial Al composition ratio of 0.94, and an ultimate Al composition ratio of 0.82 can achieve good properties.

[0042] It should be noted that LQB layers with compositional gradients have been investigated for LEDs with emission wavelengths longer than 240 nm in the past (Non-Patent Documents 1-7). In particular, Non-Patent Document 4 theoretically investigates the compositional gradient of the spacer layer. However, the disclosures in all of these documents are limited to techniques for wavelengths longer than Far-UVC (210 nm-230 nm) as in the present application. Since the wavelength range is not one in which the limit of the Al composition ratio is a problem as in the present application, the constraints imposed by the Al composition ratio have not been considered in the conventional methods. In particular, special attention must be paid to the fact that the effect of compositional gradients differs greatly depending on the emission wavelength and material composition adopted. For example, whether increasing or decreasing the Al composition ratio towards the electron block layer increases the emission intensity cannot be determined without considering the emission wavelength. For example, at an emission wavelength of 280 nm, there is ample room to increase the Al composition ratio in the composition of the electron block layer, so the Al composition ratio of the electron block layer can be widely adjusted. For emission wavelengths longer than 240 nm, the contribution of the LQB layer and the spacer layer itself to electron leakage is relatively limited. Therefore, findings at emission wavelengths longer than 240 nm cannot be applied to Far-UVC. In fact, in this embodiment, which typically relates to operation in the 210-230 nm wavelength range, the effectiveness of the composition gradient has been confirmed for the first time in the short wavelength range where AlGaN exhibits significantly different properties from longer wavelengths, as shown in the experimental results in Figure 8A. In the 210-230 nm wavelength range, in a conventional configuration combining an electron blocking layer with a spacer layer having a flat Al composition ratio, even if an AlN electron blocking layer 138 with an Al composition ratio maximized is used, it is predicted that a serious decrease in injection efficiency due to electron overflow will occur under conditions where a current for driving is applied. Thus, in this embodiment, it is predicted, and experimentally confirmed, that by devising the composition distribution of the spacer layer 136A, it is possible to suppress electron leakage under conditions where a current for driving is applied. The inventors speculate that this operating principle is due to the fact that, by creating a compositional gradient, the barrier height of the electron blocking layer 138 increases from the perspective of the pseudo-Fermi level of electrons, thereby suppressing overflow.The effect of increasing the barrier height of the electron blocking layer 138 relative to this pseudo-Fermi level can occur with both a decrease in the Al composition ratio gradient 136D and an increase in the Al composition ratio gradient 136U (both shown in Figure 3).

[0043] 1-3. Polarization control by thinning the barrier layer Figures 10A and 10B are band diagrams used to explain the technical concept of this embodiment by comparing a conventional LED (Figure 10A) with the LED of this embodiment (Figure 10B) and focusing on the polarization of the emitted UV light. The simulation software nextnano (nextnano GmbH, Munich, Germany) was used to calculate the band diagrams, and the Al composition of the barrier layer and well layer used in the calculations was set to 0.94 and 0.82, respectively. These figures are magnified views of a portion of the quantum well layer 13W, showing the conduction band edge (E C ) and valence band edge (E V Figure 10A shows two band profiles near the valence band edge and the envelope of the wave function based on the amplitude at each position. Each figure also shows an enlarged view of the two wave function envelopes near the valence band edge within a pair of circles. Figure 10A shows a configuration with a quantum well layer 13W of 2 nm thickness and a barrier layer 13B of 6 nm thickness, while Figure 10B shows a configuration with a quantum well layer 13W of 2 nm thickness and a barrier layer 13B of 1 nm thickness, both under crystal growth conditions on a polar substrate. The left-right direction of the figures represents the thickness direction, but as indicated by the scale, the thickness direction scale in Figure 10B is larger than in Figure 10A. The ranges of the quantum well layer 13W and barrier layer 13B are shown above the graphs in Figures 10A and 10B. Specifically, in the thickness direction, in conventional LEDs, the barrier layer 13B is located from the origin (0nm) to 6nm and from 8nm to 14nm, while the quantum well layer 13W is located from -2nm to 0nm, from 6nm to 8nm, and from 14nm to 16nm. In contrast, in this embodiment, where the barrier layer is 1nm thick, the barrier layer 13B is located from the origin (0nm) to 1nm and from 3nm to 4nm, while the quantum well layer 13W is located from 1nm to 3nm and from 4nm to 6nm.

[0044] Furthermore, setting the barrier layer thickness to 6 nm has been commonly adopted in the configuration of conventional LEDs with emission wavelengths of around 280 nm. Also, as mentioned above, TE emission is radiated in the thickness direction of the stacked structure of the LED in Figure 1 and can be easily extracted from the LED itself, while TM emission is radiated in the plane of the stacked structure and does not easily escape.

[0045] As shown in Figure 1, in the configuration of LED 100A of this embodiment, multiple quantum well layers 13W are provided, and adjacent quantum well layers 13W are separated by a barrier layer 13B. Conduction band edge E in Figures 10A and 10B C In the profile, the quantum well layer 13W and the barrier layer 13B are, respectively, at the conduction band edge E C These are drawn at low-energy and high-energy positions. Valence band edge E V The two profiles in the vicinity are the opposite. This figure was obtained by simultaneously calculating the Schrödinger equation and the Poisson equation for a structure in which well layers and barrier layers repeat periodically. The two band profiles near the conduction band edge and the valence band edge each have a slope corresponding to the crystal formed on the polar substrate.

[0046] Generally, the low-energy position of the conduction band edge of the quantum well layer 13W is located below the high-energy position of the conduction band edge of the adjacent barrier layer 13B, and when traced in the thickness direction, it forms a downward convex shape, acting as a quantum well. Electrons contributing to light emission undergo optical transitions from the state formed in that quantum well. The optical transition of electrons is a recombination with a valence band edge state (hole state) that satisfies the selection rule due to spatial symmetry. In this case, in addition to the selection rule due to spatial symmetry, the optical transition with the hole with the smallest energy difference becomes dominant, and a photon with that energy difference is emitted. The emission pattern depends on the direction of the electric dipole moment of the optical transition, and the direction of the electric dipole moment of the optical transition is determined by the pair of wave functions that yield the initial and final state pair allowed by the selection rule reflecting spatial symmetry. In Figures 10A and 10B, the envelopes of the electron probability states near the valence band edge are denoted with TE if TE emission is dominant in the optical transition between the quantum well layer states, and with TM if TM emission is dominant.

[0047] In Figures 10A and 10B, near the conduction band edge and valence band edge, one wavefunction envelope is drawn at the conduction band edge and two at the valence band edge, roughly aligned with the quantum well layer 13W. The number of these envelopes corresponds to the number of profiles at the conduction band edge and valence band edge, respectively. The vertical position of the straight line drawn as a reference for the amplitude of the envelopes, except in the enlarged view, is drawn to coincide with the energy values ​​of the scale for the band profile shown on the left axis of each figure. When the inventors investigated the band structure in detail using theoretical calculations with the simulation software nextnano, it was found that in the case of the 6nm barrier layer shown in Figure 10A, the wavefunction of electrons that become TM polarized gives high energy at the valence band edge. In other words, although the quantum well layer 13W is sufficiently thin at 2nm, when the Al composition ratio is adjusted for an emission wavelength of about 210-230nm, TM emission becomes dominant if the thickness of the barrier layer 13B remains at the conventional 6nm. This was found to be one of the factors contributing to the low luminescence efficiency. In conventional LEDs, such as blue light-emitting diodes, which use GaN-based materials (including InGaN-based materials) with an Al composition ratio of 0, TE emission is achieved. In conventionally designed LEDs, as the Al composition ratio increases to shorten wavelengths using AlGaN-based crystals, the relative ratio of TM emission increases, and in materials with an Al composition ratio of approximately 0.5, TE emission and TM emission become balanced. If the Al composition ratio continues to increase, eventually almost 100% of the emission becomes TM emission. In deep ultraviolet LEDs operating at 260-280 nm, for example, which use an Al composition ratio exceeding approximately 0.5, a method has been employed to address this problem by using a thin quantum well layer of about 2 nm and increasing the relative ratio of TE emission through the confinement of its quantum effect. However, the quantum confinement effect achieved by thinning the quantum well layer cannot cause a reversal of the energy values ​​of the hole state. In other words, if a high Al composition ratio is adopted for emission wavelengths of 210-230 nm, conventional LED design methods or simply using a thin quantum well layer will inevitably result in a high ratio of TM emission, as shown in Figure 10A, making efficiency improvements unlikely. Even if the quantum well layer is made thinner, there are concerns about the negative effects of carrier detachment from the quantum well layer.

[0048] To address this challenge, the inventors' research has shown that, even with a high Al composition ratio for 210-230 nm in this embodiment, thinning the barrier layer 13B as shown in Figure 10B allows for the realization of an electron wave function such that the wave function with the smaller energy difference from the conduction band edge among the two wave functions at the valence band edge becomes the TE emission, thereby enabling the dominant generation of TE emission. Figure 11 is a graph showing the measured emission spectra of LED samples fabricated with barrier layer thicknesses of 6 nm (curve C1) and 1 nm (curve C2) to confirm the technical concept of this embodiment. The operating conditions were room temperature (300 K) and 50 mA (element size 0.16 mm). 2) was used. Here, a spacer layer 136A with a composition gradient was adopted, the quantum well layer 13W was fixed to a thickness of 2 nm, and LED samples with barrier layers 13B of 6 nm and 1 nm were compared. The emission wavelength was 227 nm in both cases, and the number of quantum well layers 13W was 4. Figure 12 is a graph showing the relationship between the thickness of the barrier layer 13B and the intensity of TE emission, obtained in a simulation to confirm the technical concept of this embodiment, with the thickness of the quantum well layer fixed at 3 nm and the number of layers fixed at 4. As shown in Figure 11, by changing the barrier layer 13B from the conventional 6 nm to 1 nm, the emission intensity is increased by approximately 2 times. Also, as can be seen from Figure 12, compared with the case where the thickness of the barrier layer 13B is 6 nm, the emission intensity begins to increase significantly when the thickness of the barrier layer 13B is 4 nm or less. The thickness of the barrier layer 13B is preferably 4 nm or less, more preferably 3 nm or less, even more preferably 2 nm or less, and even more preferably 1 nm or less. The thickness of the barrier layer 13B can be greater than the monolayer thickness of 0.25 nm without any particular manufacturing issues. Furthermore, the thickness of the barrier layer 13B can be determined based on the thickness of the quantum well layer. Specifically, it is preferable to set the thickness of the barrier layer 13B to be less than or equal to the thickness of the quantum well layer. It is also useful to determine the thickness based on the relative ratio of TE emission to TM emission. If the barrier layer 13B is 3 nm to 4 nm or less, the TE emission will be about 50% of the total emission intensity (TE emission + TM emission), exceeding the TM emission. Therefore, it is preferable to determine the thickness of the barrier layer 13B so that the TE emission is stronger than the TM emission. However, the TE emission ratio exceeds 90% when the barrier layer 13B is about 1 nm thick. Also, even if the barrier layer 13B is made even thinner to 0.5 nm (equivalent to two monolayers), the TE emission ratio does not differ significantly from the 1 nm case. Therefore, by determining the lower limit of the barrier layer 13B thickness to a thickness of 1 nm or less that facilitates stable manufacturing, high practicality can be expected. However, considering the possibility of distortion occurring in the crystals constituting the light-emitting layer 134 and the barrier layer 13B, in the implemented ultraviolet light-emitting diode, the thickness of the barrier layer 13B may be less than 0.25 nm, for example, to about 0.2 nm.In other words, a desirable thickness for the barrier layer 13B corresponding to a monolayer is 0.2 nm or more. Thinning both the quantum well layer 13W and the barrier layer 13B not only allows for the generation of TE emission but also maintains carrier trapping, thus avoiding concerns about efficiency reduction due to carrier detachment. Furthermore, while there is generally a concern that carrier mobility decreases with high Al composition ratios, thinning the barrier layer 13B also has the advantage of homogenizing the carrier distribution among multiple quantum well layers 13W. Note that if the thickness of the barrier layer 13B is 6 nm, the TE emission ratio will be approximately 20%.

[0049] The inventors believe that thinning the barrier layer 13B increases TE emission for the following reasons. As described above, in the optical transition that causes emission, a transition occurs in which the electron is at the conduction band edge as the initial state and at the valence band edge as the final state. At this time, whether TM emission or TE emission occurs is determined by the direction of the electric dipole moment realized by the value of the large dipole matrix element between the initial and final states. There is only one electronic state at the conduction band edge that becomes the initial state, regardless of whether the barrier layer 13B is thinned or not, and there can be two states (hole states) at the valence band edge that can become the final state. That is, the state that contributes to emission in the quantum well must have an energy less than the maximum energy of the barrier layer, and generally this state is a bound state. A state with an energy greater than the maximum energy is an unbound state. The number of bound states depends on the magnitude of the effective mass and the magnitude of the confinement potential (energy difference between the barrier layer and the well layer), and the lighter the effective mass and the smaller the confinement potential, the fewer the number of bound states. In Far-UVC LEDs, the confinement potential cannot be very large in order to achieve short wavelengths. Therefore, in this calculation, there is only one bound state for electrons. This is in contrast to LEDs that emit light at 280 nm, where there can be two or more bound states for electrons. On the other hand, the number of bound states for holes is quite large, around 10 under the calculation conditions. However, whether or not they contribute to emission depends on how well the carriers are distributed in those bound states. As the number of carriers decreases exponentially to deeper bound states (lower in the band diagram), the emission characteristics are largely determined by comparing the shallowest bound state with the second shallowest bound state. In this sense, Figures 10A and 10B show only two levels. Note that the TE / TM emission intensity ratio in Figure 12 is calculated considering up to the fourth bound state, but in reality, the number of carriers in the fourth state is less than 1% of that of the first state. This is shown in Figures 10A and 10B. If we consider a change in which the barrier layer 13B is gradually thinned, the quantum confinement effect causes the state at the valence band edge to shift in the direction that widens the energy difference between it and the state at the conduction band edge, that is, towards the lower energy side (downward in the diagrams in Figures 10A and B).This shift is due to quantum confinement and, more fundamentally, is a manifestation of the uncertainty principle. Of the two states in question at the valence band edge, the electron responsible for TM emission has a smaller effective mass than the electron responsible for TE emission. Therefore, the state responsible for TM emission, which has a smaller effective mass, shifts significantly downward, while the state responsible for TE emission, which has a larger effective mass, shifts relatively less. By thinning the barrier layer 13B, the state responsible for TM emission, which has a smaller effective mass, shifts significantly downward, and can even pass over the state responsible for TE emission. This is likely the reason for the change from TM emission to TE emission between Figure 10A and Figure 10B. Conventionally, the quantum confinement effect due to thinning the quantum well layer 13W was considered, but in this embodiment, it was discovered for the first time that a shift can be induced by thinning the barrier layer 13B, and it was experimentally confirmed that this shift is also effective for optical transitions at wavelengths less than 240 nm.

[0050] Furthermore, one possible reason why the barrier layer 13B contributes to the quantum confinement effect is related to the sawtooth-shaped valence band edge profile caused by inclination and bending in the quantum well and barrier. Such a valence band edge profile reflects the polar orientation of the crystal. However, the quantum well layer 13W and the thinned barrier layer 13B together affect the valence band edge profile, resulting in the aforementioned shift. In this case, the effect of thinning the layer together is not limited to cases where the crystal has a polar orientation; even when the substrate for epitaxial growth is not a polar substrate but a semi-polar or non-polar substrate, thinning the barrier layer 13B can still be effective.

[0051] 1-4. Overlapping application of spacer layers with compositional gradients and thinned barrier layers. In the LED sample shown in Figure 11, where the emission spectrum is indicated by curve C2, the difference in the thickness of the barrier layer 13B was investigated after employing a compositionally graded spacer layer 136A. Furthermore, as mentioned above in relation to Figure 8A, the electron injection efficiency is improved in the compositionally graded spacer layer 136A. In other words, there are no technical obstacles to combining the compositionally graded spacer layer 136A and the thinned barrier layer 13, both of which are related to light emission. Therefore, adopting these technical concepts in overlapping manner is advantageous from a practical standpoint. Figures 13A and 13B are graphs of the emission spectrum (Figure 13A) and current-optical output characteristics (Figure 13B) of the Far-UVC LED sample of this embodiment, in which a compositionally graded spacer layer and a thinned barrier layer are applied in overlapping manner, similar to Figure 8A. As shown in Figure 13A, during continuous operation at room temperature, emission of a narrow single peak with an emission wavelength of 227 nm and a full width at half maximum of 10 nm was observed in the Far-UVC region. Furthermore, as shown in Figure 13B, it was confirmed that a practical Far-UVC LED with an external quantum efficiency of 0.2% and an optical output of 1.2 mW can be realized during pulsed operation.

[0052] 1-5. Variations The technical concepts of this embodiment, such as the compositional gradient of the spacer layer and the thinning of the barrier layer for TE emission enhancement, are applicable not only to AlGaN-based crystals but also to structures made of InAlGaN-based crystals, for example, in which In is included as part of the composition of one of the layers. These technical concepts are also applicable to LEDs whose emission wavelength is outside the 210-230 nm range. Furthermore, the manufacturing method of the LED that can be used in this embodiment is not particularly limited; for example, in addition to the MOVPE method, the MBE (Molecular Beam Epitaxy) method can be used. Although the compositional distribution of the Al composition ratio in the spacer layer 136 has been described as linearly increasing or linearly decreasing, various changes depending on the position in the thickness direction of the stacking can also be adopted as a variation of this embodiment.

[0053] The embodiments of this disclosure have been described in detail above. Each of the embodiments and configurations described above is provided for illustrative purposes, and the scope of the invention of this application should be determined based on the claims. Modifications that exist within the scope of this disclosure, including other combinations of each embodiment, are also included in the claims. [Industrial applicability]

[0054] The Far-UVCLED with improved luminous efficiency described herein can be used in any device that incorporates it as a source of ultraviolet light. [Explanation of Symbols]

[0055] 100V, 100A Light-Emitting Diode (Far-UVC LED) 102 Light extraction surface 104 One side of the substrate 110 circuit boards 120 buffer layers 132 n-type conductive layer 134 Emitting layer 13W quantum well layer 13B Barrier layer 136 Spacer layer 136A Spacer layer (composition gradient) 138 Electron Block Layer 140 1st electrode 150 p-type contact layer 160 Reflecting electrode 162 Inserted metal layer 164 UV reflective film

Claims

1. An ultraviolet light-emitting diode comprising an AlGaN-based crystal or an InAlGaN-based crystal, The light-emitting layer, Spacer layer and Electron block layer and These are stacked in this order from upstream to downstream of the electron flow, The Al composition ratio in the spacer layer changes according to its position in the thickness direction of the lamination. The compositional distribution of the spacer layer is sloped such that the Al composition ratio decreases from the light-emitting layer toward the electron-blocking layer. The value of the reached Al composition ratio, which is the Al composition ratio of the spacer layer at the position where it reaches the electron block layer, is determined according to the emission peak wavelength of the interfacial quantum well formed at the interface between the spacer layer and the electron block layer. Ultraviolet light-emitting diode.

2. The light-emitting layer includes at least one barrier layer and at least two quantum well layers flanking the barrier layer, The barrier layer is thinned. The ultraviolet light-emitting diode according to claim 1.

3. The thickness of the barrier layer is 0.2 nm or more and 4 nm or less. The ultraviolet light-emitting diode according to claim 2.

4. The thickness of the barrier layer is 1 nm or more and 3 nm or less. The ultraviolet light-emitting diode according to claim 3.

5. The thickness of the barrier layer is such that the light emission in the light-emitting layer is such that the TE emission is stronger than the TM emission. The ultraviolet light-emitting diode according to claim 2.

6. The thickness of the barrier layer is less than or equal to the thickness of the quantum well layer. The ultraviolet light-emitting diode according to claim 2.

7. The main wavelength of the emitted ultraviolet light is 210-230 nm. The ultraviolet light-emitting diode according to claim 1.

8. An electrical device comprising the ultraviolet light-emitting diode described in claim 1 as a source of ultraviolet light.

Images