An epitaxial wafer for a high-efficiency infrared light-emitting diode and its fabrication method

Abstract

This invention provides an epitaxial wafer for a high-efficiency infrared light-emitting diode and a method for fabricating the same. The epitaxial wafer includes a GaAs substrate, an N-type semiconductor structure formed on the substrate, a quantum well active region, and a P-type semiconductor structure. The N-type semiconductor structure contains an N-type AlGaAs stacked layer, co-doped with diethyltellurium (DETe) and disilane (Si₂H₆), and designed as a superlattice structure with periodically alternating high and low doping concentrations in the vertical direction; a doping concentration gradient is also set near the quantum well region, causing the tellurium doping concentration to gradually increase. Through this design, this invention reduces the carbon impurity content and non-radiative recombination probability in the material, achieving better current spreading and quantum well stress release, and significantly improving the luminous efficiency and performance of the infrared LED.

Description

Technical Field

[0001] This invention relates to the field of light-emitting diode (LED) semiconductor technology, and in particular to an epitaxial wafer of a high-efficiency infrared light-emitting diode and its fabrication method. Background Technology

[0002] Infrared light-emitting diodes (LEDs) are semiconductor devices that emit near-infrared light, typically with wavelengths ranging from 0.70 μm to 1.0 μm. Due to their advantages such as small size, low power consumption, long lifespan, high stability, and good directivity, infrared LEDs are widely used in fields such as camera-assisted lighting, material composition analysis, remote control sensing, biomedicine, and plant photography. With the emergence of new demands such as 5G / 6G communication, autonomous driving, and precision medicine, the application prospects of infrared LEDs are even broader. However, improving the luminous efficiency of infrared LEDs remains a key technical focus and challenge for the industry.

[0003] Traditional infrared light-emitting diode epitaxial structures, such as Figure 1 As shown, it includes: a substrate, and N-type GaAs buffer layer, N-type GaInP etch stop layer, N-type GaAs ohmic contact layer, N-type AlGaAs extension layer, N-type AlGaAs confinement layer, quantum well active layer, P-type AlGaAs confinement layer, P-type AlGaAs window layer, P-type GaInP transition layer, and P-type GaP ohmic contact layer sequentially grown on the substrate. In conventional epitaxial growth processes, the aforementioned N-type AlGaAs extension layer and N-type AlGaAs confinement layer are typically used as the current extension layer and carrier confinement layer, respectively, and their N-type doping generally uses diethyltellurium (DETe) or disilane (Si2H6) as the doping source alone. However, this traditional single doping method has many shortcomings and is difficult to significantly improve the luminous efficiency of infrared LEDs. For example, when only DETe is used as the N-type doping source, due to the large atomic radius of tellurium, it may introduce high lattice stress and defect density in the AlGaAs crystal, reducing material quality and not conducive to luminous efficiency. Furthermore, DETe doping increases the carbon (C) impurity content in AlGaAs materials, requiring the introduction of excessive Group V source gases such as AsH3 to suppress carbon impurities, which increases growth costs. When only Si2H6 is used as the N-type doping source, silicon doping easily forms deep-level trap centers, increasing the probability of nonradiative recombination, thereby reducing carrier recombination luminous efficiency and significantly affecting carrier lifetime and device performance. In summary, both pure Te-doped and pure Si-doped schemes in the existing technology have limitations and cannot achieve further breakthroughs in luminous efficiency in infrared LED epitaxial structures. Summary of the Invention

[0004] To address the shortcomings of the existing technology, this invention proposes an optimized epitaxial wafer structure for infrared light-emitting diodes and its growth method, which significantly improves the luminous efficiency of infrared LEDs by improving the doping strategy of the N-type layer without changing the material system.

[0005] The high-efficiency infrared light-emitting diode epitaxial wafer of the present invention includes a GaAs substrate, an N-type semiconductor layer structure, a quantum well active layer, and a P-type semiconductor layer structure. The N-type semiconductor layer structure comprises multiple layers of arsenide material, with the core being an N-type AlGaAs stacked layer. This N-type AlGaAs stacked layer consists of two adjacent parts: a lower first N-type AlGaAs layer (closer to the N-type GaAs ohmic contact layer) and an upper second N-type AlGaAs layer (closer to the quantum well active layer). Both parts are co-doped with DETe and Si₂H₆, but with different doping distributions.

[0006] In the first N-type AlGaAs layer, a superlattice doping structure design is employed. This involves adjusting the supply of the DETe doping source to create periodically alternating high-doped and low-doped regions along the growth direction (i.e., the DETe concentration alternates between high and low values). Simultaneously, an appropriate amount of Si₂H₆ is introduced to provide background n-type doping. This alternating high- and low-doping concentration design allows for different carrier concentration distributions at different locations within the AlGaAs layer, thereby enhancing lateral current diffusion capability. While ensuring sufficient current diffusion, the thickness of the first N-type AlGaAs layer can be appropriately reduced to minimize the material's own absorption loss for luminescence.

[0007] In the second N-type AlGaAs layer, a doping gradient design is adopted, that is, a higher concentration of DETe doping is introduced near the quantum well, while a lower concentration of DETe doping is maintained away from the quantum well, so that the concentration of DETe gradually increases along the growth direction. Si2H6 also provides N-type co-doping in this layer, but the concentration is relatively low and remains basically constant. By increasing the Te doping concentration near the quantum well, more electron carriers can be gathered near the active region, increasing the radiative recombination probability; on the other hand, since Te has a large atomic mass and radius, higher Te doping will introduce a slight change in lattice constant in the region near the quantum well, which can relieve the stress of the quantum well, thereby helping to stabilize the quantum well structure and improve the luminescence efficiency.

[0008] Furthermore, the active quantum well layer of this invention preferably features a multi-quantum-well structure, for example, consisting of several pairs of alternating InGaAs quantum wells and AlGaAsP barriers to achieve the desired emission wavelength and intensity. The P-type semiconductor layer structure can employ conventional designs, including, for example, a P-type AlGaAs confinement layer (electron blocking layer) adjacent to the quantum well, a relatively thick P-type AlGaAs window layer (to reduce surface recombination and absorption losses), a P-type GaInP transition layer, and a top-layer P-type GaP contact layer. High-concentration hole-type doping can be achieved using organic magnesium sources (Cp₂Mg) or carbon halide sources (CBr₄, CCl₄), etc. Through the above structural design, the epitaxial wafer of this invention, while maintaining compatibility with the original material system, significantly improves the doping distribution of the N-type layer, thereby enhancing the overall device performance.

[0009] Compared with existing infrared LED epitaxial structures, this invention offers significant advantages. First, by co-doping with DETe and Si₂H₆, the carbon impurity content in the AlGaAs material is effectively reduced. Background carbon levels can be controlled without the need for large-scale introduction of arsine (AsH₃), thus simplifying the process and reducing production costs. Simultaneously, the introduction of Si increases the ionization activation rate of N-type impurities and reduces the generation of deep-level defect centers. The synergistic effect of the two impurities improves the crystal quality and carrier transport characteristics of the material. Second, the superlattice doping design in the first N-type AlGaAs layer makes the lateral diffusion of carriers more uniform, effectively reducing the decrease in local luminous efficiency caused by current congestion and allowing for thinning of this layer to reduce infrared light absorption losses in the N-type layer. Third, the doping gradient design in the second N-type AlGaAs layer alleviates the stress of the quantum well and improves quantum efficiency. These improvements work together to enhance both the light output power and electrical performance of the infrared LED. According to test results, under conditions of a wavelength of approximately 850nm, a chip size of 42mil, and an injection current of 350mA, compared with the traditional structure, the epitaxial wafer of this invention reduces the forward voltage of the LED from approximately 1.62V to approximately 1.60V, while increasing the output optical power from approximately 385mW to approximately 405mW under the same driving current, an improvement of more than 5%. Therefore, this invention, through material doping and structural innovation, significantly improves the luminous efficiency and output power of infrared LEDs, providing reliable technical support for the high-performance applications of infrared LEDs in night vision surveillance, communication, medical fields, and other areas. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of the epitaxial structure of an infrared light-emitting diode in the prior art;

[0011] Figure 2 This is a schematic diagram of the epitaxial wafer structure of the high-efficiency infrared light-emitting diode provided by the present invention. Wherein, in Figure 1The labeling of each layer is as follows: 500 is the GaAs substrate; 501 is the N-type GaAs buffer layer; 100 is the N-type GaInP etching stop layer; 101 is the N-type GaAs ohmic contact layer; 102 is the N-type AlGaAs extension layer; 103 is the N-type AlGaAs confinement layer; 104 is the multiple quantum well; 105 is the P-type AlGaAs confinement layer; 106 is the P-type AlGaAs window layer; 107 is the P-type GaInP transition layer; 108 is the P-type GaP ohmic contact layer; Figure 2 The labels for each layer are as follows: 600 is the GaAs substrate; 601 is the N-type GaAs buffer layer; 200 is the N-type GaInP etching stop layer; 201 is the N-type GaAs ohmic contact layer; 202 is the first N-type AlGaAs layer; 203 is the second N-type AlGaAs layer; 204 is the multi-quantum-well active layer; 205 is the P-type AlGaAs confinement layer; 206 is the P-type AlGaAs window layer; 207 is the P-type GaInP transition layer; and 208 is the P-type GaP ohmic contact layer. The thickness ratios of the layers in the figure are not drawn to actual scale and are only used to illustrate the structure of the present invention. Detailed Implementation

[0012] The epitaxial structure and its manufacturing method of the present invention will be further described below with reference to the accompanying drawings and specific embodiments, but the present invention is not limited to these embodiments.

[0013] like Figure 2 As shown, the high-efficiency infrared light-emitting diode epitaxial wafer provided in this embodiment includes, from bottom to top: a GaAs substrate (600), an N-type GaAs buffer layer (601), an N-type GaInP etching stop layer (200), an N-type GaAs ohmic contact layer (201), a first N-type AlGaAs layer (202), a second N-type AlGaAs layer (203), a multi-quantum-well active layer (204), a P-type AlGaAs confinement layer (205), a P-type AlGaAs window layer (206), a P-type GaInP transition layer (207), and a topmost P-type GaP ohmic contact layer (208). The material composition, doping, and growth conditions of each layer are described in detail below.

[0014] N-type layer growth: An N-type GaAs buffer layer (601) was first grown on a clean GaAs substrate (600) using MOCVD. The growth temperature was controlled at approximately 700℃ (preferably within the range of 680–740℃) to ensure crystal quality. Arsenic was provided by arsine (AsH3) and gallium by trimethylgallium (TMGa) as the vapor source, while a small amount of disilane (Si2H6) was introduced as an n-type dopant source to achieve a buffer layer doping concentration of approximately 1 × 10⁻⁶. 18 ~3×10 18 carriers·cm -3(Silicon is the donor impurity). The thickness of this GaAs buffer layer is approximately... (scope ).

[0015] Next, an N-type GaInP etching stop layer (200) is grown on the buffer layer, using pentaalkyl gallium (TMGa), pentaalkyl indium (TMIn), and phosphine (PH3) to co-grow the GaInP alloy. The growth temperature can be the same as or slightly adjusted from the previous layer (approximately 680–700 °C). Similarly, a Si2H6-doped GaInP layer is used to reach 1 × 10⁻⁶. 18 ~3×10 18 carriers·cm -3 The n-type concentration is adjusted to ensure the layer remains conductive in subsequent device fabrication processes and can serve as a stop layer for selective etching. The GaInP layer thickness is approximately... (scope ).

[0016] An N-type GaAs ohmic contact layer (201) was grown on top of a GaInP layer. Heavy doping of N-type GaAs was achieved by continuing to introduce TMGa and AsH3 while simultaneously increasing the Si2H6 flow rate. The doping concentration was controlled to be greater than 5 × 10⁻⁶. 18 carriers·cm -3 This process forms a highly doped contact layer, thereby reducing the contact resistance between the N-electrode and the N-type layer. The layer is grown at approximately 700°C and has a thickness of approximately [missing information]. (scope ).

[0017] Next, an N-type AlGaAs stacked layer is grown according to the key scheme of the present invention. First, a first N-type AlGaAs layer (202) is grown. TMGa, TMAl, and AsH3 are introduced to co-deposit an AlGaAs alloy, wherein the Al composition ratio is controlled at approximately 20% (range 15%–25%). Throughout the AlGaAs growth process, DETe and Si2H6 are simultaneously introduced as n-type doping sources to achieve co-doping of Te and Si. By adjusting the supply amount of the DETe source, a superlattice doped structure is formed in the first N-type AlGaAs layer along the growth direction—that is, whenever a certain thickness is reached (e.g., when the growth reaches a certain point...). (Approximately), the DETe supply is appropriately reduced to form a low-doped layer segment, and then the DETe supply is increased to form a high-doped layer segment. This process is repeated alternately to achieve a thickness of 6–8 μm. Several pairs of highly / lowly doped sublayers are formed within the AlGaAs layer. Preferably, the carrier concentration of DETe in the highly doped region is approximately 3 × 10⁻⁶. 17 ~5×10 17 carriers·cm -3The carrier concentration in the low-doped region is approximately 1 × 10⁻⁶. 17 ~2×10 17 carriers·cm -3 The Si2H3 doping concentration in the entire first AlGaAs layer can be maintained at approximately 1 × 10⁻⁶. 17 ~3×10 17 carriers·cm -3 The superlattice co-doping design described above can significantly optimize carrier distribution and current spread, and reduce the absorption of light emission by the thick AlGaAs layer.

[0018] A second N-type AlGaAs layer (203) is then grown, with a slightly higher Al content, controlled at approximately 30% (range 25%–35%), to form a potential barrier to the quantum well in the energy band (restricting electron overflow into the N-region). This layer is relatively thin, approximately [missing information - likely a number]. (scope DETe and Si2H+ were continued for co-doping, but instead of alternating periods, the DETe doping concentration was gradually increased along the growth direction, maintaining approximately 5 × 10⁻⁶ at the interface near the first AlGaAs layer below. 17 ~6×10 17 carriers·cm -3 The Te doping level was gradually increased by increasing the DETe supply, so that the Te doping concentration rose to about 7 × 10⁻⁶ near the quantum well layer (204). 17 ~8×10 17 carriers·cm -3 The Si2H+ doping concentration in the second AlGaAs layer is still controlled at approximately 1 × 10⁻⁶. 17 ~3×10 17 carriers·cm -3 And it remains constant. Through this gradual doping design, the second AlGaAs layer has a higher electron concentration and a slightly larger lattice constant near the active region, which helps to reduce nonradiative recombination in the quantum well caused by stress, thereby improving quantum efficiency.

[0019] Active region growth: After completing the above N-type layer structure, stop introducing any doping source and grow a multi-quantum-well active layer (204) on the second AlGaAs layer. For example, InGaAs quantum well layers and AlGaAsP barrier layers can be grown alternately to form a light-emitting active region. In this embodiment, each quantum well structure has a thickness of approximately InGaAs well layers and thickness approximately The active region consists of an AlGaAsP barrier layer with an Al content of approximately 0.1% and a P content of approximately 0.1%. Preferably, the number of quantum well pairs is 8 (selectable range 4 to 15 pairs) to balance luminescence intensity and material stress. The entire active region growth process remains undoped to avoid the influence of impurities on radiative recombination.

[0020] P-type layer growth: After the active region is completed, a P-type semiconductor layer structure is grown on it. First, a P-type AlGaAs confinement layer (205) is grown, with an Al composition of approximately 30% (25%–35%) and a thickness of approximately [missing information]. Its main function is to act as an electron blocking layer to limit the outflow of charge carriers in the active region. During growth, Cp2Mg (organic magnesium source) or a carbon halide source (CBr4, CCl4) is introduced for p-type doping, so that the hole concentration of this layer reaches approximately 1 × 10⁻⁶. 18 carriers·cm -3 Left and right (typical range 9×10) 17 ~1.5×10 18 carriers·cm -3 ).

[0021] Next, a p-type AlGaAs window layer (206) was grown, with the Al content reduced to about 20% (15%–25%) and the thickness increased to about 2 μm. The window layer serves to widen the emission window and reduce surface recombination. This layer is also doped with Mg or C, with the concentration controlled at approximately 1 × 10⁻⁶. 18 ~2×10 18 carriers·cm -3 This gives it good electrical conductivity.

[0022] Then, a p-type GaInP transition layer (207) is grown, with a thickness of approximately [missing information]. Due to the large bandgap of GaInP and its lattice-matching effect with GaP, introducing this transition layer can improve the p-type contact characteristics. This layer is doped with a Mg source (Cp₂Mg) to achieve p-type doping, with a concentration of approximately 2 × 10⁻⁶. 18 ~3×10 18 carriers·cm -3 .

[0023] Finally, a p-type GaP ohmic contact layer (208) is grown, with a thickness of approximately [missing information]. GaP possesses a high bandgap and good surface stability, which can improve the transparency of the outermost surface region and serve as a low-resistivity contact layer for metal electrodes. By introducing high concentrations of CBr4 or CCl4 for doping, the hole concentration can be increased to >5 × 10⁻⁶. 19 carriers·cm -3This ensures the formation of a good P-type ohmic contact.

[0024] Thus, the entire infrared light-emitting diode epitaxial structure has been grown according to the design of this invention. The resulting epitaxial wafer structure is as follows: Figure 2 As shown in the table below, LED devices were fabricated using conventional chip fabrication processes (such as photolithography, electrode evaporation, dicing and encapsulation) and then tested.

[0025]

[0026] Tests show that the infrared LED of this invention, while maintaining stable operation, exhibits lower forward voltage drop and higher output optical power, demonstrating significant advantages compared to conventional structures in the prior art. This invention, through innovative design at the material structure level, achieves improved photoelectric performance of infrared LEDs and has promising prospects for industrial application.

Claims

1. An epitaxial wafer for a high-efficiency infrared light-emitting diode, comprising a GaAs substrate, an N-type semiconductor structure, a quantum well active region, and a P-type semiconductor structure sequentially formed on the GaAs substrate, characterized in that: The N-type semiconductor structure includes an N-type AlGaAs stack layer, which consists of a first N-type AlGaAs layer and a second N-type AlGaAs layer adjacent to the first N-type AlGaAs layer. The first N-type AlGaAs layer is co-doped with diethyltellurium (DETe) and disilane (Si₂H₆) and grown in a superlattice structure. The DETe doping concentration in the first N-type AlGaAs layer exhibits a periodic distribution of alternating high and low concentrations along the growth direction. The second N-type AlGaAs layer is co-doped with the DETe and Si₂H₆ and has a gradually increasing DETe doping concentration gradient on the side closer to the active region of the quantum well.

2. The epitaxial wafer according to claim 1, characterized in that: The first N-type AlGaAs layer has an Al molar fraction of 0.15–0.25 and a thickness of 60,000–80,000 Å. During the growth of the first N-type AlGaAs layer, DETe and Si₂H₆ are introduced for co-doping; wherein the Si₂H₆ doping concentration is 1 × 10⁻⁶. 17 ~3×10 17 carriers·cm -3 The low doping concentration of DETe is 1×10⁻⁶. 17 ~2×10 17 carriers·cm -3 The high-concentration doping concentration is 3×10 17 ~5×10 17 carriers·cm -3 .

3. The epitaxial wafer according to claim 1, characterized in that: The second N-type AlGaAs layer has an Al molar fraction of 0.25–0.35 and a thickness of 3000–4000 Å. During the growth of the second N-type AlGaAs layer, DETe and Si₂H₆ are introduced for co-doping, and the DETe doping concentration is increased from 5 × 10⁻⁶ on the side furthest from the quantum well active region. 17 ~6×10 17 carriers·cm -3 Gradually increase to 7×10 near the active region of the quantum well. 17 ~8×10 17 carriers·cm -3 The doping concentration of the Si2H6 is 1×10⁻⁶. 17 ~3×10 17 carriers·cm -3 .

4. The epitaxial wafer according to claim 1, characterized in that: The high-concentration doped region and low-concentration doped region of DETe in the first N-type AlGaAs layer alternately form a superlattice structure, and the superlattice period of high and low doping concentration is repeated 10 to 30 pairs in the first N-type AlGaAs layer.

5. The epitaxial wafer according to claim 1, characterized in that: The active region of the quantum well is a multi-quantum-well structure, consisting of alternating InGaAs quantum well layers and AlGaAsP barrier layers. The thickness of the quantum well layers is 55–75 Å, the thickness of the barrier layers is 230–280 Å, the molar fraction of Al is 0.05–0.15, the molar fraction of P is 0.08–0.15, the number of quantum well pairs is 4–15, and the active region of the quantum well is an intrinsic layer that is undoped.

6. The method for manufacturing an epitaxial wafer according to any one of claims 1-5, characterized in that, The layered structure is grown using a metal-organic chemical vapor deposition process at a reaction chamber temperature of 680–740°C, including the following steps: Step 1: An N-type GaAs buffer layer, an N-type GaInP etching stop layer, and an N-type GaAs ohmic contact layer are epitaxially grown sequentially on a GaAs substrate. During the growth, Si2H6 gas is introduced as an N-type dopant source. Step 2, continue the epitaxial growth of the N-type AlGaAs stacked layer, including the sequential growth of the first N-type AlGaAs layer and the second N-type AlGaAs layer. During the growth process, DETe and Si2H6 gases are simultaneously introduced for N-type co-doping. During the growth of the first N-type AlGaAs layer, the doping concentration of DETe is controlled to alternate in a high-low-high-low cycle to form a superlattice doped structure. During the growth of the second N-type AlGaAs layer, the doping concentration of DETe is controlled to gradually increase along the growth direction. Step 3: Epitaxially grow the active region of the quantum well, which is an InGaAs / AlGaAsP multi-quantum-well structure without introducing doping; Step 4: Epitaxial growth of a P-type semiconductor structure, including sequential growth of a P-type AlGaAs confinement layer, a P-type AlGaAs window layer, a P-type GaInP transition layer, and a P-type GaP ohmic contact layer. During the growth process, a P-type doping source is introduced, wherein the P-type doping source is selected from magnesium pyrocene, carbon tetrabromide, or carbon tetrachloride.

7. The manufacturing method according to claim 6, characterized in that: In step 1, the thickness of both the N-type GaAs buffer layer and the N-type GaInP etching stop layer is 1500–2500 Å, and the doping concentration is 1 × 10⁻⁶. 18 ~3×10 18 carriers·cm -3 The thickness of the N-type GaAs ohmic contact layer is 500–800 Å, and the doping concentration is greater than 5 × 10⁻⁶. 18 carriers·cm -3 .

8. The manufacturing method according to claim 6, characterized in that: In step 4, the molar fraction of Al in the p-type AlGaAs confinement layer is 0.25–0.35, the thickness is 3000–4000 Å, and the doping concentration is 9 × 10⁻⁶. 17 ~1.5×10 18 carriers·cm -3 The p-type AlGaAs window layer has an Al molar fraction of 0.15–0.25, a thickness of 15,000–25,000 Å, and a doping concentration of 1 × 10⁻⁶. 18 ~2×10 18 carriers·cm -3 The thickness of the p-type GaInP transition layer is 150–250 Å, and the doping concentration is 2 × 10⁻⁶. 18 ~3×10 18 carriers·cm -3 The thickness of the p-type GaP ohmic contact layer is 300–500 Å, and the doping concentration is greater than 5 × 10⁻⁶. 19 carriers·cm -3 .

Images