A single photon avalanche diode based on multi-element rare earth co-doping and a preparation method thereof

Abstract

This invention discloses a single-photon avalanche diode based on multi-element rare-earth co-doping and its fabrication method, aiming to solve the problems of excessively high room-temperature dark count rate caused by bulk and interface defects in existing InP / InGaAs-based SPADs, and the trade-off between core performance indicators. The core structure of the device includes an InP substrate and sequentially epitaxial layers of a buffer layer, an absorption layer, a gradient layer, a charge layer, and a multiplication layer, with at least one functional layer using two or more rare-earth co-doped materials selected from Gd. 3+ La 3+ Pr 3+ 、Nd 3+ Eu 3+ Rare-earth ion co-doping enables targeted cross-scale defect passivation and the construction of shallow trap energy level networks. This invention suppresses the dark count core generation mechanism at the intrinsic material level, synergistically improving photon detection efficiency and reducing afterpulse probability. Furthermore, the fabrication process is fully compatible with existing semiconductor production lines, facilitating industrial-scale mass production and making it suitable for fields such as quantum communication, lidar, and fluorescence lifetime imaging.

Description

Technical Field

[0001] This invention relates to the field of semiconductor device technology, specifically to a single-photon avalanche diode based on multi-element rare-earth co-doping and its fabrication method. Background Technology

[0002] SPAD (Single-Photon Avalanche Diode) is a semiconductor detector operating in Geiger mode, achieving high-sensitivity detection at the single-photon level through the avalanche multiplication effect. Its core materials include semiconductors such as InGaAs, exhibiting short-wave infrared response, low power consumption, and small size, making it suitable for quantum communication, lidar, and 3D imaging. SPADs are core detectors in cutting-edge technologies such as quantum communication, lidar, and fluorescence lifetime imaging. SPADs based on InP / InGaAs materials are the mainstream technology for detecting photons in the 1550nm band, but their performance is limited by a high dark count rate (DCR). The main reason is: 1. Dark count rate is a core indicator that determines the detection accuracy, signal-to-noise ratio and practical application level of SPAD. When existing InP / InGaAs-based SPADs operate at room temperature, the narrow bandgap of their InGaAs absorption layer has a significant thermal excitation effect at room temperature, forming a high intrinsic thermal noise floor, which directly increases the dark count background and keeps the dark count rate at a high level. Moreover, during the epitaxial growth process, the InP multiplication layer and InGaAs absorption layer of existing InP / InGaAs-based SPADs inevitably introduce bulk defects such as point defects and dislocations. These defects form highly efficient Shockley-Read-Hall (SRH) recombination centers, significantly increasing the probability of thermally generated carriers. In addition, due to lattice mismatch, the interfaces of heterojunctions such as InGaAs / InP, InGaAs / InGaAsP, and InGaAsP / InP have high-density interface states and dangling bonds, which further exacerbate the nonradiative recombination of carriers. The above defects, especially the deep-level defects located in the high electric field region of the InP multiplication layer, will induce a strong trap-assisted tunneling (TAT) effect when the device is operating in Geiger mode. This is the main reason for the high dark count rate at room temperature, which seriously limits the high performance application of the device at room temperature.

[0003] 2. To address the industry pain point of excessively high dark count rates, traditional solutions for improving single-photon avalanche diodes mainly focus on optimizing epitaxial growth conditions, designing guard ring structures, and performing surface passivation treatments. However, these solutions typically only achieve limited improvement in surface or local defects, failing to achieve comprehensive and effective passivation of deep-level defects within InGaAs and InP bulk materials and at heterojunction interfaces at the intrinsic material level. They cannot fundamentally eliminate carrier generation-recombination centers and thus cannot achieve a fundamental reduction in dark count rates. Furthermore, some optimization solutions introduce new problems such as increased process complexity, higher fabrication costs, and decreased photon detection efficiency, making it difficult to balance device performance and fabrication feasibility.

[0004] 3. In existing technical solutions, there is a clear trade-off between the three core performance indicators: dark count rate, photon detection efficiency (PDE), and afterpulse probability. Optimizing a single indicator often comes at the expense of other indicators. For example, adjusting the device structure and doping design to suppress dark count can easily lead to uncontrolled electric field distribution in the multiplication layer and a decrease in photon absorption efficiency, thereby reducing the device's photon detection efficiency. On the other hand, the presence of deep-level defects not only increases the dark count rate but also exacerbates carrier trapping and delayed release, resulting in an increased afterpulse probability. This severely affects the device's detection stability and imaging resolution, making it difficult to meet the stringent requirements of high-end applications for overall device performance.

[0005] Therefore, how to overcome the inherent limitations of existing technologies and develop a technical solution that can actively passivate bulk and interface defects at the intrinsic level of materials, suppress the dark count core generation mechanism from the root, and simultaneously improve core performance such as photon detection efficiency and afterpulse probability, while being compatible with existing semiconductor fabrication processes, to achieve a significant reduction in dark count rate and a leapfrog improvement in overall performance of SPAD devices at room temperature, and promote their large-scale and batch application in various cutting-edge fields, has become a key technical issue that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] This invention addresses the technical bottlenecks of existing InP / InGaAs single-photon avalanche diodes, such as excessively high dark count rates due to intrinsic material defects, bulk defects, and heterojunction interface defects, the difficulty of fundamentally solving this problem with traditional passivation techniques, and the challenge of synergistic optimization of core performance. It provides a single-photon avalanche diode based on multi-element rare-earth co-doping and its fabrication method.

[0007] This invention achieves comprehensive passivation of cross-scale defects and optimization of the energy level system by implementing targeted multi-element rare earth ion co-doping combinations in each epitaxial functional layer of the device, and synergistically utilizing the dual physical mechanisms of strain field defect pinning and shallow trap state carrier competition. At the same time, it is compatible with existing semiconductor manufacturing processes, fundamentally reducing the dark count rate and synergistically improving photon detection efficiency and reducing the afterpulse probability, ultimately achieving a leapfrog improvement in the overall performance of the device. Moreover, the fabrication process is highly compatible and easy to industrialize and mass-produce.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: A single-photon avalanche diode based on multi-element rare-earth co-doping includes, from bottom to top, an InP substrate (1), an n-type InP buffer layer (2), an intrinsic InGaAs absorber layer (3), an n-type InGaAsP gradient layer (4), an n-type InP charge layer (5), and an intrinsic InP multiplication layer (6); at least one functional layer among the n-type InP buffer layer (2), intrinsic InGaAs absorber layer (3), n-type InGaAsP gradient layer (4), n-type InP charge layer (5), and intrinsic InP multiplication layer (6) is co-doped with two or more rare-earth ions; the rare-earth ions are selected from gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Praseodymium ion Pr 3+ Neodymium ions (Nd) 3+ Europium ions (Eu) 3+ .

[0009] Furthermore, the n-type InP buffer layer (2) is made of gadolinium ions (Gd). 3+ With lanthanum ion La 3+ Co-doped; the intrinsic InGaAs absorber layer (3) is made of gadolinium-ionized Gd 3+ Praseodymium ion Pr 3+ With europium ions Eu 3+ Co-doped; the n-type InGaAsP graded layer (4) uses gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Neodymium ions (Nd) 3+ With europium ions Eu 3+ Co-doping; the n-type InP charge layer (5) is made of gadolinium ions (Gd). 3+ With lanthanum ion La 3+ Co-doped; the intrinsic InP multiplication layer (6) uses gadolinium ions (Gd). 3+ Praseodymium ion Pr 3+ With neodymium ions Nd 3+ Co-doping.

[0010] Furthermore, the total rare earth ion doping concentration of each functional layer in the n-type InP buffer layer (2), intrinsic InGaAs absorber layer (3), n-type InGaAsP graded layer (4), n-type InP charge layer (5), and intrinsic InP multiplication layer (6) is 1×10⁻⁶. 16 cm -3 ~2×10 18 cm -3 .

[0011] Furthermore, the rare earth ions in the intrinsic InGaAs absorber layer (3) are doped using a V-shaped gradient. From the interface between the intrinsic InGaAs absorber layer (3) and the n-type InP buffer layer (2) towards the middle of the intrinsic InGaAs absorber layer (3), the total doping concentration of rare earth ions gradually decreases; from the middle of the intrinsic InGaAs absorber layer (3) towards the interface between the intrinsic InGaAs absorber layer (3) and the n-type InGaAsP gradient layer (4), the total doping concentration of rare earth ions gradually increases. The rare earth ions in the n-type InGaAsP gradient layer (4) are doped using a V-shaped gradient. From the interface between the n-type InGaAsP gradient layer (4) and the intrinsic InGaAs absorber layer (3) towards the middle of the n-type InGaAsP gradient layer (4), the total doping concentration of rare earth ions gradually decreases. From the middle of the n-type InGaAsP gradient layer (4) towards the interface between the n-type InGaAsP gradient layer (4) and the n-type InP charge layer (5), the total doping concentration of rare earth ions gradually increases.

[0012] Furthermore, the thickness of the n-type InP buffer layer (2) is 500 nm; the thickness of the intrinsic InGaAs absorption layer (3) is 1.5 μm to 3 μm; the thickness of the n-type InGaAsP gradient layer (4) is 100 nm to 200 nm; the thickness of the n-type InP charge layer (5) is 100 nm to 300 nm; and the thickness of the intrinsic InP multiplication layer (6) is 2.5 μm to 3.5 μm.

[0013] The second aspect of this invention provides a method for fabricating a single-photon avalanche diode based on multi-element rare-earth co-doping, used to fabricate the single-photon avalanche diode described in the first aspect, comprising the following steps: S1. Substrate preparation: Provide an InP substrate (1); S2. Epitaxial growth and in-situ rare earth co-doping: An n-type InP buffer layer (2), an intrinsic InGaAs absorber layer (3), an n-type InGaAsP gradient layer (4), an n-type InP charge layer (5), and an intrinsic InP multiplication layer (6) are epitaxially grown sequentially on the InP substrate (1). During the growth of at least one functional layer, two or more rare earth ion precursors are introduced in-situ through an independent metal-organic source pipeline to achieve rare earth ion co-doping. The rare earth ions are selected from gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Praseodymium ion Pr 3+ Neodymium ions (Nd) 3+ Europium ions (Eu) 3+ ; S3. Device back-end fabrication: A p-type contact region (7) is fabricated on the epitaxially grown structure, a p-type ohmic contact electrode (8) is fabricated, and an n-type ohmic contact electrode (9) is fabricated on the back side of the InP substrate (1) to obtain the low-noise single-photon avalanche diode.

[0014] Furthermore, the epitaxial growth described in step S2 is achieved using an MOCVD or MBE device. The flow rate of the rare earth ion precursor is controlled by an independent metal-organic source pipeline to achieve uniform or gradient doping of rare earth ions in the functional layer.

[0015] Furthermore, in step S2, the in-situ rare earth co-doping process parameters for each functional layer are as follows: During the growth of the n-type InP buffer layer (2), gadolinium ions (Gd) were co-doped in situ. 3+ With lanthanum ion La 3+ The growth temperature is 550℃~640℃, and the total rare earth ion doping concentration is 6×0. 16 cm -3 ~8×0 16 cm -3 Doping flux ratio of gadolinium ions (Gd) 3+ ≥ Lanthanum ion La 3+ ; During the growth of the intrinsic InGaAs absorber layer (3), gadolinium ions (Gd) are co-doped in situ. 3+ Praseodymium ion Pr 3+ With europium ions Eu 3+ The growth temperature is 600℃~680℃, and the total rare earth ion doping concentration is 3×10⁻⁶. 17 cm -3 ~5×10 17 cm -3 Doping flux ratio of gadolinium ions (Gd) 3+ ≥Pr ion 3+ ≥Eu ion 3+ ; During the growth of n-type InGaAsP graded layers (4), gadolinium ions (Gd) were co-doped in situ. 3+ Lanthanum ion La 3+ Neodymium ions (Nd) 3+ With europium ions Eu 3+ The growth temperature is 600℃~680℃, and the total rare earth ion doping concentration is 9×10⁻⁶. 17 cm -3 ~2×10 18 cm -3 Doping flux ratio of gadolinium ions (Gd) 3+ ≥ Lanthanum ion La 3+ ≥Nd ions 3+ ≥Eu ion 3+ The n-type InGaAsP gradient layer (4) serves as a high-concentration doped region. During the growth of the n-type InP charge layer (5), gadolinium ions (Gd) were co-doped in situ. 3+ With lanthanum ion La 3+ The growth temperature is 550℃~640℃, and the total rare earth ion doping concentration is 6×10⁻⁶. 17 cm -3 ~9×10 17 cm -3 Doping flux ratio of gadolinium ions (Gd) 3+ ≥ Lanthanum ion La 3 + ; During the growth of intrinsic InP multiplication layers (6), gadolinium ions (Gd) were co-doped in situ. 3+ Praseodymium ion Pr 3+ With neodymium ions Nd 3+ The growth temperature is 550℃~640℃, and the total rare earth ion doping concentration is 1×10⁻⁶. 17 cm -3 ~3×10 17 cm -3 Doping flow rate ratio Gd 3+ ≥Pr 3+ ≥Nd 3+ .

[0016] Furthermore, in step S2, when growing the intrinsic InGaAs absorber layer (3), a V-shaped gradient doping is performed on rare earth ions. By controlling the inlet flow rate of the rare earth ion precursor, the total doping concentration of rare earth ions at the interface between the intrinsic InGaAs absorber layer (3) and the n-type InP buffer layer (2) and the n-type InGaAsP gradient layer (4) is higher than that in the middle of the intrinsic InGaAs absorber layer (3). When growing the n-type InGaAsP graded layer (4), V-shaped gradient doping is performed on rare earth ions. By controlling the influx flow rate of the rare earth ion precursor, the total doping concentration of rare earth ions at the interface between the intrinsic InGaAs absorber layer (3) and the n-type InP charge layer (5) of the n-type InGaAsP graded layer (4) is higher than that at the middle of the n-type InGaAsP graded layer (4).

[0017] Further, step S3 includes the following steps: S31. A secondary Zn diffusion process is used to selectively diffuse Zn onto the surface of the intrinsic InP multiplication layer (6) to form a p-type contact region (7). S32. A dielectric film is deposited on the surface of the intrinsic InP multiplication layer (6) as a passivation anti-reflection layer (10). S33. An electrode contact window (101) is formed on the passivation anti-reflection layer (10) by photolithography and etching processes. S34. Using electron beam evaporation or magnetron sputtering, a Ti / Pt / Au or Cr / Au metal stack is deposited at the electrode contact window (101) to form a p-type ohmic contact electrode (8); and after thinning and polishing the InP substrate (1), an n-type ohmic contact electrode (9) is formed.

[0018] Compared with the prior art, the beneficial effects of the present invention are: 1. More comprehensive synergistic passivation effect: By designing targeted multi-element rare earth ion combinations in different functional layers (including n-type InP buffer layer (2), intrinsic InGaAs absorption layer (3), n-type InGaAsP graded layer (4), n-type InP charge layer (5), and intrinsic InP multiplication layer (6), etc.), synergistic passivation of various defect types such as point defects, bulk defects, dislocations, and interface dangling bonds is achieved. Especially in the InGaAsP graded layer and InP multiplication layer, rare earth co-doping effectively suppresses lattice mismatch stress and optimizes interface bonding strength, making the passivation effect more thorough and comprehensive than that of a single ion.

[0019] 2. Multidimensional Dark Count Suppression Mechanism: The strain field is modulated using ionic radius differences to introduce defect trapping centers in the InGaAs layer; shallow trap networks are constructed using 4f electron configuration differences to optimize carrier trapping at the InGaAsP interface layer; and chemical property differences are used to saturate dangling bonds to optimize doping distribution in the InP charge layer. This multidimensional synergistic suppression mechanism effectively reduces dark count generation mechanisms such as SRH recombination and TAT tunneling.

[0020] 3. Significant performance improvement: Especially in the interface regions of multiple heterojunctions such as InGaAs / InP, InGaAs / InGaAsP, and InGaAsP / InP, by implementing V-shaped gradient doping with multiple rare earth ions, combined with the dual effects of strain compensation and chemical bonding, the interface transition is effectively smoothed, lattice mismatch is compensated, and the interface state density is significantly reduced, thereby suppressing the trap-assisted tunneling effect that dominates dark counting at room temperature.

[0021] 4. Synergistic optimization of functions of each layer: In the intrinsic InGaAs absorption layer (3), rare earth doping effectively suppresses thermal noise sources; in the n-type InGaAsP gradient layer (4), strain compensation and interface passivation of multiple heterojunctions are realized; in the n-type InP buffer layer (2), n-type InP charge layer (5) and intrinsic InP multiplication layer (6), the material quality and electric field distribution are synergistically optimized, a complete shallow trap network and defect suppression system are constructed, and the device performance is synergistically improved from multiple levels.

[0022] 5. Good process compatibility: Multi-element rare earth ions can be introduced in situ during standard MOCVD or MBE epitaxy without adding complex process steps. It is fully compatible with existing semiconductor manufacturing processes and is easy to industrialize. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the layer structure of a single-photon avalanche diode according to Embodiment 1 of the present invention; Figure 2 This is a schematic flowchart of the steps in the fabrication method of a single-photon avalanche diode according to Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of the structure of a portion of the Zn diffusion mask in step S31 of the present invention. Detailed Implementation

[0024] The following examples further illustrate the features and other related characteristics of the present invention in detail, to facilitate understanding by those skilled in the art: Example

[0025] like Figure 1 As shown, this embodiment is a single-photon avalanche diode based on multi-element rare earth co-doping, which includes the following stacked from bottom to top: InP substrate (1), n-type InP buffer layer (2), intrinsic InGaAs absorption layer (3), n-type InGaAsP gradient layer (4), n-type InP charge layer (5) and intrinsic InP multiplication layer (6). At least one functional layer selected from the n-type InP buffer layer (2), intrinsic InGaAs absorber layer (3), n-type InGaAsP graded layer (4), n-type InP charge layer (5), and intrinsic InP multiplication layer (6) is co-doped with two or more rare earth ions; the rare earth ions are selected from gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Praseodymium ion Pr 3+ Neodymium ions (Nd) 3+ Europium ions (Eu) 3+ In specific implementation, in one or more layers of the above-mentioned epitaxial structure, gadolinium (Gd) is used as a component. 3+ lanthanum (La) 3+ ), Praseodymium (Pr 3+ ), neodymium (Nd) 3+ Europium (Eu) 3+ Two or more rare earth ions are co-doped with each other, and specific defect passivation and energy level control targets are achieved in each functional layer through ion combination design and concentration control.

[0026] Due to La 3+ (1.03Å)>Pr 3+ (0.99Å)>Nd 3+ (0.98Å)>Eu 3+ (0.95Å)>Gd 3+ (0.94Å)>In 3+ (0.80Å)>Ga 3+ (0.62 Å), therefore, the above rare earth ions will preferentially replace In when they enter GaInAs, GaInAsP, or InP materials. 3 + Its ionic radius is similar to that of In 3+ The differences cause local lattice distortion and elastic strain fields, which can effectively attract and "pin" point defects such as vacancies and interstitial atoms migrating around them. Meanwhile, besides La... 3+ (4f) 0 Apart from the configuration, other rare earth ions (Gd) 3+ 4f 7 Pr 3+ 4f 2 Nd 3+ 4f 3 Eu 3+ 4f 6 It has a stable unfilled 4f electron configuration, which can introduce localized shallow trap states in the material bandgap to competitively capture charge carriers.

[0027] As described above, this invention designs and combines Gd-doped layers in different epitaxial functional layers such as n-type InP buffer layers and intrinsic InGaAs absorber layers. 3+La 3+ By employing two or more rare-earth ions, a cross-scale defect co-passivation system is constructed through the synergistic utilization of strain field defect pinning and shallow-trap carrier competition mechanisms. This allows for the generation of a strong lattice strain field using the ionic radius differences of different rare-earth ions, efficiently pinning bulk defects such as vacancies and dislocations. Simultaneously, the unique 4f electron configuration of these ions forms a shallow-trap network with a reasonable energy level distribution, competitively capturing carriers to suppress SRH recombination and TAT effects dominated by deep-level defects. This comprehensively passivates bulk defects and heterojunction interface defects at the intrinsic level of the material, fundamentally reducing the thermal noise substrate and dark count rate at room temperature. Furthermore, this multi-element rare-earth ion system... The co-doped design is deeply compatible with the classic InP / InGaAs stacked structure, optimizing the defect passivation effect without compromising the core function and electric field distribution balance of each functional layer. It effectively breaks the trade-off between dark count rate, photon detection efficiency and afterpulse probability, achieving a synergistic improvement in the three core performances. At the same time, the device structure is fully compatible with existing semiconductor epitaxial processes, without the need to reconstruct the core architecture, providing favorable conditions for large-scale and mass production. It provides a brand-new solution for fundamentally reducing the dark count rate of SPAD devices, significantly promoting its high-performance application in cutting-edge fields such as quantum communication and lidar.

[0028] As a preferred implementation, the specific co-doping schemes for each functional layer are as follows: 1. The n-type InP buffer layer (2) is preferably made of gadolinium ions (Gd). 3+ With lanthanum ion La 3+ Co-doping; thus, La 3+ The large ionic radius of Gd generates a strong lattice strain field, effectively "anchoring" and passivating dislocations extending from the substrate; 3+ This provides shallow electron trap states, initially passivating point defects. Together, they provide a high-quality crystal template for subsequent epitaxy.

[0029] 2. The intrinsic InGaAs absorber layer (3) is preferably made of gadolinium ions (Gd). 3+ Praseodymium ion Pr 3+ With europium ions Eu 3+ Co-doping; has the following effects: Gd 3+ Passivation of intrinsic point defects such as As / Ga vacancies; Pr 3+ (4f) 2 Introducing shallow hole traps: Specifically, Pr 3+ The 4f orbital has two electrons, which, upon entering InGaAs material, will replace In. 3+ The lattice position of its 4f 2 The configuration creates shallow trap energy levels near the valence band in the material's band gap, introducing shallow hole traps.

[0030] Eu 3+ (4f) 6 ) Auxiliary saturated anion vacancies: Specifically, Eu 3+ 4f 6 The configuration has a strong chemical bonding ability. After entering the material, it will preferentially combine with these anion vacancies, which is equivalent to "blocking" the "capture channel" of deep energy level defects, reducing the probability of electrons being captured by vacancies, thereby suppressing nonradiative recombination.

[0031] Thus, different rare earth ions (Pr 3+ Eu 3+ Gd 3+ The number of 4f electrons in (etc.) is different (4f 2 4f 6 4f 7 The shallow trap energy level positions and the types of trapped carriers (electrons or holes) are also different; the Pr in the intrinsic InGaAs absorption layer (3) 3+ hole trapping, Gd 3+ Electron trapping, Eu 3+ Saturated vacancies, these traps are evenly distributed in the material, forming an all-round shallow trap network, competitively suppressing deep level recombination and reducing thermally generated dark current.

[0032] 3. The n-type InGaAsP graded layer (4) serves as a defect enrichment region, preferably using gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Neodymium ions (Nd) 3+ With europium ions Eu 3+ Co-doping. Specifically, the InGaAsP graded layer is the "transition region" between the InGaAs absorption layer and the InP charge layer. Due to the large differences in lattice constants and band gaps between the two materials, the interface contains a large number of dangling bonds, lattice stress, and Fermi level mismatch, making it the region with the highest defect density and the most severe area for the TAT effect; while Gd 3+ (4f) 7 Eu 3+ (4f) 6 The outer electrons of InGaAs have strong coordination bonding capabilities. After entering the interface region, they form stable chemical bonds with the unpaired electrons of the dangling bonds, causing the original deep-level defects to disappear. The charge carriers lose the tunneling trap channels, eliminating the core cause of the TAT effect. Moreover, although the lattice constants of InGaAs (lattice constant ≈ 5.868 Å) and InP (lattice constant ≈ 5.869 Å) are close, there are still slight differences. In addition, the compositional gradient of the InGaAsP graded layer itself will lead to lattice mismatch stress at the interface. This stress will induce lattice distortion and generate dislocation defects. These dislocations will form continuous deep-level defect chains, becoming long-distance tunneling channels for the TAT effect.

[0033] La 3+ The ionic radius (1.03 Å) is much larger than that of In. 3+ (0.80 Å) After entering the InGaAsP lattice, the size difference generates a strong lattice strain field. This strain field can counteract the stress caused by lattice mismatch, alleviating lattice distortion, suppressing the generation and extension of dislocation defects, and structurally disrupting the continuous tunneling channel of the TAT effect, thus reducing the density of deep-level defects. Furthermore, at the heterojunction interface, the Fermi level mismatch between the two materials leads to uneven carrier distribution at the interface (such as local accumulation of electrons / holes), increasing the probability of carriers being trapped by defects. Simultaneously, the Fermi level mismatch causes level bending in the band gap, lowering the energy barrier for carrier tunneling and making the TAT effect more likely to occur. Meanwhile, Nd... 3+ 4f 3 The configuration creates shallow trap levels near the conduction band in the band gap of InGaAsP materials. Unlike deep defects, these shallow levels do not trap carriers. Instead, they adjust the position of the Fermi level in the interface region through energy level coupling, thereby matching the Fermi level of the InGaAs absorption layer with that of the InP charge layer. This reduces local carrier accumulation and flattens the band bend at the interface, raising the energy barrier for carrier tunneling and suppressing the probability of the TAT effect from an energy perspective.

[0034] Thus, Gd 3+ With Eu 3+ Through chemical bonding of dangling bonds at saturated interfaces; La 3+ Strong strain alleviates lattice mismatch stress; Nd 3+ (4f) 3 The shallow energy level of the interface Fermi level is adjusted. The rare earth ion doping setting of this n-type InGaAsP graded layer (4) is key to suppressing the TAT effect by synergistically passivating the interface from multiple dimensions such as chemistry, strain and energy band.

[0035] 4. The n-type InP charge layer (5) is preferably made of gadolinium ions (Gd). 3+ With lanthanum ion La 3+ Co-doping. Specifically, the core function of the n-type InP charge layer (5) is to precisely control the electric field distribution through n-type doping (such as Si doping). However, in the traditional doping process, n-type dopants (such as Si atoms) are prone to local aggregation due to factors such as lattice vibration and interatomic forces, resulting in uneven doping concentration and thus causing electric field distortion. The high electric field region may shift to the absorption layer, and the low electric field region cannot effectively confine the charge carriers, ultimately affecting the stability of avalanche multiplication. Gd 3+ The ionic radius (0.94 Å) of In 3+ The difference (0.80 Å) is moderate, and 4f 7The electronic configuration exhibits a strong localization electric field effect. Upon entering the InP lattice, it exerts a dispersing force on the surrounding n-type dopant atoms through the electric field, preventing them from aggregating and ensuring a uniform distribution of dopant atoms within the charge layer, thus guaranteeing consistent doping concentration. Furthermore, La... 3+ The ionic radius (1.03 Å) is much larger than that of In. 3+ When dopants enter the InP lattice, they cause local lattice distortion, forming a weak elastic strain field. This strain field changes the potential energy distribution of the lattice, making the aggregation of n-type dopant atoms thermodynamically unstable. If dopant atoms attempt to form clusters, they will be repelled by the strain field, thus inhibiting the formation and growth of clusters and further ensuring the uniformity of charge layer doping.

[0036] Thus, the n-type InP charge layer (5) of this application is generated by Gd 3+ Homogenize the n-type dopant distribution; through La 3+ Suppress dopant clusters. The two work together to ensure a highly uniform and controllable doping concentration in the charge layer, thereby achieving precise electric field confinement and optimizing the longitudinal electric field profile within the device.

[0037] 5. The intrinsic InP multiplication layer (6) is preferably made of gadolinium ions (Gd). 3+ Praseodymium ion Pr 3+ With neodymium ions Nd 3+ Co-doping. Specifically, the intrinsic InP multiplication layer (6) is the core region of the avalanche multiplication effect and needs to operate under a high electric field. The point defects, lattice distortions and other bulk defects inside it will become deep-level traps, triggering trap-assisted tunneling (TAT) effect, which leads to an increase in dark count rate. Gd 3+ 4f 7 The electronic configuration exhibits strong localization bonding characteristics. Upon entering the InP lattice, it forms stable chemical bonds with bulk defects such as phosphorus vacancies and indium interstitials, passivating these deep-level defects and reducing carrier trapping, thus providing a low-noise lattice basis for avalanche multiplication. Furthermore, Pr 3+ The 4f orbital contains two electrons, replacing In in the InP lattice. 3+ Subsequently, a shallow trap level close to the valence band will form in the band gap. This level is specifically designed to temporarily trap holes, thus avoiding the permanent trapping of holes as seen with deep defects. It also prioritizes the trapping channel for holes, preventing the trapping of holes by deep defects from triggering the TAT effect. Furthermore, Nd... 3+ The 4f orbital contains 3 electrons, and the shallow trap level it forms is close to the conduction band, primarily used for temporarily trapping electrons. This level is similar to Pr. 3+ The shallow hole trap energy levels formed are distributed in a "stepped" manner in the band gap, meaning that the positions of the two energy levels are different, and together they form a stepped shallow trap energy level network covering electrons and holes.

[0038] Thus, in Gd 3+ While basic passivation, Pr 3+ (4f) 2 ) and Nd 3+ (4f) 3 This forms a stepped shallow trap level network. This network can more efficiently "consume" carriers that might trigger an avalanche through a competitive trapping mechanism, significantly reducing the probability of deep-level assisted tunneling. At the same time, this multi-level gradual release of carriers also helps reduce the probability of afterpulse.

[0039] In a preferred embodiment, the total rare earth ion doping concentration of each functional layer in the n-type InP buffer layer (2), intrinsic InGaAs absorber layer (3), n-type InGaAsP graded layer (4), n-type InP charge layer (5), and intrinsic InP multiplication layer (6) is 1×10⁻⁶. 16 cm -3 ~2×10 18 cm -3 Specifically, the concentration of co-doping with two or more rare earth ions is controlled at 1×10⁻⁶. 16 cm -3 ~2×10 18 cm -3 Within this range, the main goal is to achieve a balance between crystal quality, energy transfer efficiency, and luminescence stability. 1. Preventing lattice distortion: Rare earth ions have large radii, and excessively high concentrations can lead to mismatched dislocations in the GaInAs epitaxial layer.

[0040] 2. Suppress cross-relaxation: By controlling the upper limit of total concentration and combining the isolation effect of co-doped ions, non-radiative recombination is reduced.

[0041] 3. Optimized energy transfer: Within this concentration range, the spacing between rare earth ions is moderate, which can achieve effective sensitization (energy transfer) without causing energy dissipation due to excessive proximity.

[0042] As mentioned above, this application uniformly limits the total rare earth ion doping concentration of each functional layer to 1×10⁻⁶. 16 cm -3 ~2×10 18 cm -3Within this range, it can achieve the best balance between crystal quality, defect passivation efficiency, carrier transport and device stability: it can effectively avoid lattice distortion, mismatch dislocation and cross-relaxation effects caused by excessive rare earth ion concentration, reduce non-radiative recombination channels, and ensure that the rare earth ion spacing is moderate and the distribution is uniform, so as to achieve efficient defect sensitization and energy transfer, give full play to the synergistic passivation effect of multi-element rare earth co-doping, fundamentally suppress dark current and trap-assisted tunneling (TAT), while not damaging the crystal quality and carrier transport characteristics of the epitaxial layer. Ultimately, it enables the device to achieve simultaneous improvement in three core indicators: low dark count rate, high photon detection efficiency and low afterpulse probability, and significantly enhance the device's operational reliability and lifespan.

[0043] In a preferred embodiment, the rare earth ions in the intrinsic InGaAs absorber layer (3) are doped using a V-shaped gradient. From the interface between the intrinsic InGaAs absorber layer (3) and the n-type InP buffer layer (2) towards the middle of the intrinsic InGaAs absorber layer (3), the total doping concentration of rare earth ions gradually decreases; from the middle of the intrinsic InGaAs absorber layer (3) towards the interface between the intrinsic InGaAs absorber layer (3) and the n-type InGaAsP gradient layer (4), the total doping concentration of rare earth ions gradually increases. The rare earth ions in the n-type InGaAsP gradient layer (4) are doped using a V-shaped gradient. From the interface between the n-type InGaAsP gradient layer (4) and the intrinsic InGaAs absorber layer (3) towards the middle of the n-type InGaAsP gradient layer (4), the total doping concentration of rare earth ions gradually decreases. From the middle of the n-type InGaAsP gradient layer (4) towards the interface between the n-type InGaAsP gradient layer (4) and the n-type InP charge layer (5), the total doping concentration of rare earth ions gradually increases.

[0044] As described above, this application employs a V-shaped gradient doping method in both the intrinsic InGaAs absorber layer and the n-type InGaAsP graded layer, resulting in a rare earth ion concentration distribution characterized by "low concentration in the middle and high concentration at the interface" within both layers. On one hand, maintaining a high rare earth ion concentration at the heterojunction interface fully utilizes the chemical bonding and lattice matching regulation effects of rare earth ions to smooth the interface barrier, compensate for lattice mismatch, and significantly reduce the interface state density. On the other hand, gradually reducing the concentration in the middle of the layer avoids crystal quality degradation and an increase in non-radiative recombination centers due to excessive doping, achieving an optimal balance between defect passivation and crystal integrity. Thus, through targeted enhanced passivation in multiple heterojunction interface regions such as InGaAs / InP, InGaAs / InGaAsP, and InGaAsP / InP, the combined effect of strain compensation and chemical bonding significantly suppresses the trap-assisted tunneling (TAT) effect that dominates dark counting at room temperature, ultimately achieving a comprehensive performance improvement with a significant reduction in dark count rate, a significant increase in detection efficiency, and a significant enhancement in device stability.

[0045] In a preferred embodiment, the n-type InP buffer layer (2) has a thickness of 500 nm. This fixed thickness ensures sufficient coverage of substrate defects while avoiding excessive thickness that could lead to increased series resistance. The intrinsic InGaAs absorption layer (3) has a thickness of 1.5 μm to 3 μm. 2.0 μm is selected for conventional detection scenarios, while 2.5 μm to 3 μm can be used for long-distance low-light detection scenarios, ensuring sufficient absorption of photons in the target wavelength band. The thickness of the n-type InGaAsP gradient layer (4) is 100nm~200nm. The n-type InGaAsP gradient layer is the core transition layer between the intrinsic InGaAs absorption layer and the n-type InP charge layer. It can smooth the band transition of the heterojunction, alleviate the lattice mismatch stress, and ensure the efficient transport of photogenerated carriers. It is also the core functional layer of rare earth multi-element co-doping and V-shaped gradient doping passivation interface defects in this application. By setting this thickness range, the composition of InGaAs to InP can be linearly gradientd, so as to eliminate the band peaks and valence band barriers at the interface of narrow bandgap and wide bandgap heterojunction, and avoid the accumulation and obstruction of photogenerated holes at the interface. This ensures the efficient transport of photogenerated carriers from the absorption layer to the multiplication layer, prevents the photon detection efficiency from decreasing, and avoids the accumulation of carriers at the interface from aggravating the trap-assisted tunneling (TAT) effect, thereby reducing the dark count inducement from the structural level. If the thickness is less than 100 nm, the composition gradient is too steep, making it impossible to achieve a smooth band transition and easily forming an interface barrier and local electric field concentration; if the thickness is greater than 200 nm, it will needlessly increase the carrier transport path, increasing the device series resistance and transmission loss. The thickness of the n-type InP charge layer (5) is 100 nm to 300 nm. The n-type InP charge layer is the core of electric field control in the SAGCM structure. Through n-type doping, electric field clamping can be achieved, and an electric field distribution with a low electric field in the absorption region and a high electric field in the multiplication region can be constructed. If the thickness is less than 100 nm, the electric field clamping ability is insufficient, and the high electric field is easy to leak from the multiplication layer to the absorption layer, directly causing the dark count rate of the absorption layer to soar sharply; if the thickness is greater than 300 nm, it will significantly increase the device series resistance, leading to an abnormal increase in breakdown voltage and device power consumption, while also causing problems such as slower avalanche response speed and decreased time resolution. The intrinsic InP multiplication layer (6) has a thickness of 2.5 μm to 3.5 μm. The intrinsic InP multiplication layer is the core functional area for realizing single-photon avalanche multiplication in the device. It can achieve high-gain amplification of single-photon signals through carrier collision ionization under high electric field. The thickness range specified in this application provides sufficient and suitable multiplication paths for carrier collision ionization. It can achieve stable and high-gain avalanche multiplication effect, ensure effective detection of single-photon level signals, and avoid the avalanche response delay and time resolution reduction caused by excessively long multiplication paths. It is suitable for the dual requirements of detection speed and sensitivity in scenarios such as quantum communication and lidar.If the thickness is less than 2.5 μm, the critical electric field required for avalanche breakdown increases sharply. The defect-assisted tunneling (TAT) effect under high electric field is significantly amplified. Even a small number of volume defects can cause the dark count rate to soar. At the same time, the high electric field of thin layer can easily cause premature breakdown at the edge, resulting in extremely poor device stability. If the thickness is greater than 3.5 μm, the breakdown voltage of the device will increase significantly, increasing the difficulty of driving circuit design and device power consumption. At the same time, the excessively long carrier multiplication path will slow down the response speed and exacerbate the carrier trapping effect, increasing the probability of afterpulse. Example

[0046] This second embodiment is a method for fabricating a single-photon avalanche diode based on multi-element rare-earth co-doping, including the following steps: like Figure 2-3 As shown, a method for fabricating a single-photon avalanche diode based on multi-element rare-earth co-doping includes the following steps: S1. Substrate preparation: Select a 2-4 inch InP substrate (1) and transfer it into the MOCVD equipment.

[0047] S2. Epitaxial growth and in-situ rare earth co-doping: Each functional layer is epitaxially grown sequentially from bottom to top on an InP substrate (1) using an MOCVD device; during the growth of at least one functional layer, two or more rare earth ion precursors are introduced in-situ through an independent metal-organic source pipeline to achieve rare earth ion co-doping, wherein the rare earth ions are selected from gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Praseodymium ion Pr 3+ Neodymium ions (Nd) 3+ Europium ions (Eu) 3+ Specifically, the in-situ rare earth co-doping process is carried out simultaneously during the growth of each layer, and the specific steps are as follows: 2.1 Growth of n-type InP buffer layer (2): Temperature 550~640℃, thickness 500nm, Si doping concentration 1~5×10 18 cm -3 During the growth process, Gd is co-doped through an independent metal-organic source pipeline. 3+ with La 3+ Ions, total doping concentration 6~8×10 16 cm -3 Flow ratio: Gd 3+ ≥La 3+ ;La 3+ The large ionic radius of Gd generates a strong lattice strain field, effectively "anchoring" and passivating dislocations extending from the substrate; 3+ This provides shallow electronic trap states and initially passivates point defects; the two work together to provide a high-quality crystal template for subsequent epitaxy.

[0048] 2.2 Growth of intrinsic InGaAs absorber layer (3): An intrinsic InGaAs absorber layer (3) is epitaxially grown on an n-type InP buffer layer (2) at a growth temperature of 600~680℃ and a thickness of 1500~3000nm. Gd is co-doped during the growth process. 3+ Pr 3+ With Eu 3+ Ions, total doping concentration 3~5×10 17 cm -3 Flow ratio: Gd 3+ ≥Pr 3+ ≥Eu 3+ Prioritize Gd 3+ Slightly higher; Gd 3+ Passivation of intrinsic point defects such as As / Ga vacancies; Pr 3+ (4f) 2 Introducing shallow cavity traps; Eu 3+ (4f) 6 The auxiliary saturated anion vacancies and multiple 4f configurations jointly construct a shallow trap network, competitively suppressing deep-level recombination and reducing thermally generated dark current.

[0049] 2.3 Growth of n-type InGaAsP graded layer (4): An n-type InGaAsP graded layer (4) is epitaxially grown on the intrinsic InGaAs absorber layer (3) at a growth temperature of 600~680℃ and a thickness of 100~200nm. Gd is co-doped during the growth process. 3+ La 3+ 、Nd 3+ With Eu 3+ Ions, total doping concentration 9×10 17 ~2×10 18 cm -3 Flow ratio: Gd 3+ ≥La 3+ ≥Nd 3+ ≥Eu 3+ This layer is a highly doped region to achieve strong interface passivation; Gd 3+ With Eu 3+ Through chemical bonding of dangling bonds at saturated interfaces; La 3+ Strong strain alleviates lattice mismatch stress; Nd 3+ (4f) 3 The shallow energy level adjustment of the interface Fermi level, and the multi-dimensional passivation of the interface from chemical, strain and energy band dimensions, are the key to suppressing the TAT effect.

[0050] 2.4. Growth of n-type InP charge layer (5): An n-type InP charge layer (5) is epitaxially grown on the n-type InGaAsP graded layer (4) at a growth temperature of 550~640℃ and a thickness of 150~300nm. The areal density is achieved by Si doping to reach 2~2.5×10⁻⁶.12 cm -2 Gd is co-incorporated during the growth process. 3+ with La 3+ Ions, total doping concentration 6~9×10 17 cm -3 Flow ratio: Gd 3+ ≥La 3+ To ensure uniform charge distribution; Gd 3+ Homogenize the distribution of n-type dopant; La 3+ Suppressing dopant clusters, the two work together to ensure that the charge layer has a highly uniform and controllable doping concentration, thereby achieving precise electric field confinement and optimizing the longitudinal electric field distribution profile inside the device.

[0051] 2.5 Growth of intrinsic InP multiplication layer (6): An intrinsic InP multiplication layer (6) is epitaxially grown on an n-type InP charge layer (5) at a growth temperature of 550~640℃ and a thickness of 2500~3500nm. Gd is co-doped during the growth process. 3+ Pr 3+ With Nd 3+ Ions, total doping concentration 1~3×10 17 cm -3 Flow ratio: Gd 3+ ≥Pr 3+ ≥Nd 3+ ; in Gd 3+ While basic passivation, Pr 3+ (4f) 2 ) and Nd 3+ (4f) 3 This forms a stepped shallow trap energy level network, which can more efficiently "consume" carriers that may trigger an avalanche through a competitive capture mechanism, greatly reducing the probability of deep-level assisted tunneling. At the same time, this multi-level gradual release of carriers mechanism also helps to reduce the probability of afterpulse.

[0052] S3. Device back-end fabrication: A p-type contact region (7) is fabricated on the epitaxially grown structure, a p-type ohmic contact electrode (8) is fabricated, and an n-type ohmic contact electrode (9) is fabricated on the back side of the InP substrate (1) to obtain the low-noise single-photon avalanche diode. Specifically, the following steps are included: S31. Secondary Zn diffusion is performed using a Zn diffusion process. Figure 3 The mask (100) shown is selectively diffused to form a p-type contact region (7) on the surface of the intrinsic InP multiplication layer (6). S32. A dielectric film (specifically SiNx or SiO2 dielectric film) is covered on the surface of the intrinsic InP multiplication layer (6) using PECVD process as a passivation and anti-reflection layer (10). This dielectric film has both passivation and anti-reflection functions, saturates surface dangling bonds and reduces photon reflection loss. S33. An electrode contact window (101) is formed on the passivation anti-reflection layer (10) by photolithography and etching processes. S34. Using electron beam evaporation or magnetron sputtering, a Ti / Pt / Au or Cr / Au metal stack is deposited at the electrode contact window (101) to form a p-type ohmic contact electrode (8); and after thinning and polishing the InP substrate (1), an n-type ohmic contact electrode (9) is formed on the back side of the InP substrate (1) to complete the fabrication of the single-photon avalanche diode.

[0053] As described above, this embodiment two is based on the fabrication method of a single-photon avalanche diode using multi-element rare earth co-doping. This involves sequentially epitaxially growing an n-type InP buffer layer (2), an intrinsic InGaAs absorption layer (3), an n-type InGaAsP gradient layer (4), an n-type InP charge layer (5), and an intrinsic InP multiplication layer (6) on an InP substrate (1). Simultaneously, multi-element rare earth ions are co-doped in situ during the growth of each functional layer to ensure lattice matching and good interface quality between layers, effectively reducing interlayer defects and stress accumulation, and avoiding problems such as epitaxial layer performance degradation and abnormally high dark counts in subsequent processes. Specifically, for the n-type InP buffer layer (2), the growth temperature is controlled between 550℃ and 640℃ to ensure good crystal quality of the InP epitaxial layer and reduce the introduction of dislocations and defects; using a 6×10⁻⁶ temperature... 16 cm -3 ~8×10 16 cm -3 The relatively low rare-earth doping concentration can effectively passivate extended dislocations on the substrate without compromising lattice integrity due to excessive doping; setting the doping flux ratio Gd 3+ ≥La 3+ This can make Gd 3+ La acts as the host passivating ion to achieve stable passivation of point defects. 3+ As auxiliary ions enhance the pinning effect on dislocations, the two work together to provide a high-quality crystal template for subsequent epitaxy. For the intrinsic InGaAs absorber layer (3), the growth temperature is controlled at 600℃~680℃, which matches the optimal epitaxial window of the InGaAs material, which is beneficial to suppress component segregation and improve film uniformity; using 3×10 17 cm -3 ~5×10 17 cm -3With moderate doping concentration, efficient passivation of bulk defects can be achieved without affecting light absorption and carrier transport; by setting the doping flux ratio Gd 3+ ≥Pr 3+ ≥Eu 3+ , making Gd 3+ Dominant defect passivation, Pr 3+ With Eu 3+ A shallow trap network is constructed stepwise to competitively suppress deep-level recombination, thereby significantly reducing thermally generated dark current. For the n-type InGaAsP graded layer (4), controlling the growth temperature at 600℃~680℃ is beneficial to achieve a smooth gradient of composition and improve the heterojunction interface quality; using 9×10 17 cm -3 ~2×10 18 cm -3 The highest doping concentration was set for this region, which is designated as a high-concentration doping area. This is because the layer is located in the multiple heterojunction region between InGaAs and InP, characterized by large lattice mismatch and high interface state density, making it the region where trap-assisted tunneling (TAT) effects are most prominent. Higher concentrations of multi-element rare-earth co-doping can achieve strong interface passivation from multiple dimensions, including strain compensation, chemical bonding, and Fermi level modulation, thereby suppressing dark counts at their source. The doping flux ratio Gd was set accordingly. 3+ ≥La 3+ ≥Nd 3 + ≥Eu 3+ This allows various rare earth ions to work in an orderly and hierarchical manner according to their functions, ensuring passivation strength while avoiding inter-ion interference, thus improving doping uniformity and process stability. For the n-type InP charge layer (5), the growth temperature is controlled at 550℃~640℃ to ensure a uniform and dense epitaxial layer and stable dopant distribution; using 6×10 17 cm -3 ~9×10 17 cm -3 The rare earth doping concentration is adjusted to achieve defect passivation without affecting the carrier control capability of the n-type doped material; the doping flux ratio Gd is set accordingly. 3+ ≥La 3+ This can make Gd 3+ Homogenize dopant distribution, La 3+ Suppressing doped clusters, the two work together to ensure precise electric field confinement and uniform longitudinal electric field distribution, avoiding dark count increases and device breakdown anomalies caused by local electric field concentration. For the intrinsic InP multiplication layer (6), controlling the growth temperature at 550℃~640℃ is beneficial to obtaining an intrinsic InP epitaxial layer with high crystal quality and low defect density, ensuring the stability and consistency of avalanche multiplication; using 1×10 17 cm -3 ~3×10 17 cm-3 The relatively low rare-earth doping concentration, while achieving defect passivation, avoids the introduction of additional carrier traps or impact on the avalanche electric field distribution by excessive doping; the doping flux ratio Gd is set accordingly. 3+ ≥Pr 3+ ≥Nd 3+ , with Gd 3+ As the basic passivation ion, while utilizing Pr 3+ With Nd 3+ A stepped shallow trap energy level network is formed, which suppresses the deep level-assisted tunneling (TAT) effect through competitive trapping and reduces the afterpulse probability by gradually releasing charge carriers. This significantly improves the device noise performance and operational reliability while ensuring high avalanche gain. Example

[0054] Example 3, based on Example 2, further optimizes the total rare earth ion doping concentration distribution in some key epitaxial functional layers by employing V-gradient doping to achieve precise enhancement and passivation of high defect density interface regions. The specific implementation method is as follows: 1. Intrinsic InGaAs absorber layer (3), using Gd 3+ Pr 3+ With Eu 3+ Ion co-doping was performed, and a V-shaped gradient doping was implemented. By controlling the influx flow rate of the rare earth ion precursor, the total rare earth ion doping concentration at the interface between the intrinsic InGaAs absorber layer (3) and the n-type InP buffer layer (2) and the n-type InGaAsP gradient layer (4) was higher than that at the middle of the intrinsic InGaAs absorber layer (3). Specifically, the total doping concentration at the interface near the n-type InP buffer layer (2) was set to approximately 5 × 10⁻⁶. 17 cm -3 The absorption layer gradually decreases towards the middle to approximately 1×10⁻⁶. 17 cm -3 Then, the total doping concentration continued to increase, reaching approximately 5 × 10⁻⁶ at the interface near the n-type InGaAsP graded layer (4). 17 cm -3 Thus, the main technical effect of this V-shaped gradient design is that it forms a high-concentration rare-earth doped region at the two key heterojunction interfaces between the absorption layer and the adjacent buffer layer and gradient layer, achieving the most effective passivation of interface defects; at the same time, it maintains a low doping concentration in the bulk region of the absorption layer material, avoiding new lattice disturbances or electrical performance degradation caused by the introduction of excessive rare earth elements, thus achieving the best balance between interface passivation and bulk quality.

[0055] In the 2.n-type InGaAsP graded layer (4), Gd is used. 3+ La 3+ 、Nd 3+ With Eu3+ Ion co-doping was also performed using a V-gradient gradient doping method. By adjusting the influx flow rate of the rare earth ion precursor, the total rare earth ion doping concentration at the interface between the intrinsic InGaAs absorber layer (3) and the n-type InP charge layer (5) of the n-type InGaAsP gradient layer (4) was higher than that in the middle of the n-type InGaAsP gradient layer (4). Specifically, the total doping concentration at the interface near the intrinsic InGaAs absorber layer (3) was set to approximately 1 × 10⁻⁶. 18 cm -3 The gradient of the n-type InGaAsP gradient layer (4) is gradually reduced to approximately 5 × 10⁻⁶. 17 cm -3 Then, the total doping concentration continued to increase, reaching approximately 1 × 10⁻⁶ at the interface near the n-type InP charge layer (5). 18 cm -3 The V-shaped gradient design aims to form rare earth enrichment layers in the interface regions of the upper and lower heterojunctions of the InGaAsP gradient layer. It fully leverages the synergistic effects of multi-element rare earth ions in chemical bonding, strain compensation, and shallow trap construction, and provides the most effective site passivation for high-density defects such as interface dangling bonds and mismatch dislocations. This significantly suppresses the trap-assisted tunneling effect caused by interface states, further improving the dark counting performance of the device at room temperature.

[0056] In summary, this invention discloses a single-photon avalanche diode based on multi-element rare-earth co-doping and its fabrication method. The method involves sequentially epitaxially growing an n-type InP buffer layer (2), an intrinsic InGaAs absorption layer (3), an n-type InGaAsP gradient layer (4), an n-type InP charge layer (5), and an intrinsic InP multiplication layer (6) on an InP substrate (1). During the epitaxial growth of each functional layer, the system utilizes La... 3+ Pr 3+ 、Nd 3+ Eu 3+ Gd 3+ The differences in ionic radius and 4f electronic configuration of different rare earth ions, combined with the core functions, defect types and working environment of each layer, are used to implement targeted multi-element rare earth ion in-situ co-doping combinations and precise process control: among which the n-type InP buffer layer (2) and the n-type InP charge layer (5) adopt Gd 3+ with La 3+ Co-doped, the intrinsic InGaAs absorber layer (3) uses Gd 3+ Pr 3+ With Eu 3+ Co-doped, n-type InGaAsP graded layer (4) using Gd 3+ La 3+ 、Nd 3+ With Eu 3+Co-doped and set as a high-concentration doped region, the intrinsic InP multiplication layer (6) uses Gd 3+ Pr 3+ With Nd 3+ Co-doping is carried out, and rare earth ion V-shaped gradient doping is performed on the intrinsic InGaAs absorber layer (3) and the n-type InGaAsP gradient layer (4) to precisely control the growth temperature, total rare earth ion doping concentration and doping flow rate ratio of each layer, so as to achieve the optimal match between doping effect and interlayer function. After epitaxial growth is completed, a secondary Zn diffusion process is used to selectively diffuse Zn onto the intrinsic InP multiplication layer (6) to form a p-type contact region (7). A SiNx or SiO2 dielectric film is covered on the intrinsic InP multiplication layer (6) as a passivation anti-reflection layer (10) using a PECVD process. An electrode contact window (101) is formed on the passivation anti-reflection layer (10) using photolithography and etching processes. Then, a Ti / Pt / Au or Cr / Au metal stack is deposited at the electrode contact window (101) using electron beam evaporation or magnetron sputtering processes to form a p-type ohmic contact electrode (8). After thinning and polishing the InP substrate (1), an n-type ohmic contact electrode (9) is fabricated, and the fabrication of a low-noise single-photon avalanche diode is finally completed.

[0057] Thus, this invention specifically addresses the core technical problem of excessively high dark count rate in existing InP / InGaAs single-photon avalanche diodes due to intrinsic point defects, bulk defects, and heterojunction interface defects in the material, and constructs a complete defect engineering solution based on multi-element rare earth ion synergistic doping. This invention achieves point defect pinning through the strong lattice strain field of large-radius rare-earth ions and establishes a cross-scale defect synergistic passivation system from microscopic point defects to macroscopic interface defects through the chemical bonding of medium-radius rare-earth ions to saturate interface dangling bonds. It utilizes rare-earth ions with different 4f electron configurations to form a gradient shallow trap state network with a reasonable energy level distribution in the band gap, effectively suppressing nonradiative recombination and trap-assisted tunneling processes dominated by deep-level defects through a competitive carrier trapping mechanism. In the key interface region of the InGaAsP graded layer, high-concentration gradient co-doping, combined with strain compensation and chemical passivation capabilities, achieves systematic control of the heterojunction interface states, significantly reducing the interface state density. Simultaneously, rare-earth co-doping of the charge layer improves doping uniformity, optimizes the longitudinal electric field profile of the device, ensures precise confinement of the high electric field region to the multiplication layer, and suppresses interband tunneling. Ultimately, through the synergistic effect of the aforementioned multiple mechanisms, a systematic optimization is achieved from the defect generation source and carrier transport channel to the electric field triggering environment. This reduces the dark count rate of SPAD devices by more than an order of magnitude, while maintaining or improving photon detection efficiency (PDE) and reducing the afterpulse probability. Furthermore, the overall technical solution of this invention is fully compatible with existing commercial MOCVD / MBE epitaxial processes and standard semiconductor back-end fabrication processes, requiring no additional dedicated equipment or production line modifications. It offers strong process controllability, high device performance consistency, and is suitable for wafer-level mass production, significantly reducing the industrialization threshold and fabrication cost of the devices. This provides solid technical support for the large-scale and market-oriented application of InP / InGaAs single-photon avalanche diodes in high-end cutting-edge fields such as quantum communication, lidar, and fluorescence lifetime imaging.

[0058] As stated above, this case protects a single-photon avalanche diode based on multi-element rare earth co-doping and its preparation method. All technical solutions that are the same as or similar to this case should be considered to fall within the protection scope of this case.

Claims

1. A single-photon avalanche diode based on multi-element rare-earth co-doping, characterized in that, The structure includes, from bottom to top, an InP substrate (1), an n-type InP buffer layer (2), an intrinsic InGaAs absorption layer (3), an n-type InGaAsP gradient layer (4), an n-type InP charge layer (5), and an intrinsic InP multiplication layer (6). At least one functional layer selected from the n-type InP buffer layer (2), intrinsic InGaAs absorber layer (3), n-type InGaAsP graded layer (4), n-type InP charge layer (5), and intrinsic InP multiplication layer (6) is co-doped with two or more rare earth ions; the rare earth ions are selected from gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Praseodymium ion Pr 3+ Neodymium ions (Nd) 3+ Europium ions (Eu) 3+ .

2. The single-photon avalanche diode according to claim 1, characterized in that, The n-type InP buffer layer (2) uses gadolinium ions (Gd). 3+ With lanthanum ion La 3+ Co-doped; the intrinsic InGaAs absorber layer (3) uses gadolinium ions (Gd). 3+ Praseodymium ion Pr 3+ With europium ions Eu 3+ Co-doped; the n-type InGaAsP graded layer (4) uses gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Neodymium ions (Nd) 3+ With europium ions Eu 3+ Co-doping; the n-type InP charge layer (5) is made of gadolinium ions (Gd). 3+ With lanthanum ion La 3+ Co-doped; the intrinsic InP multiplication layer (6) uses gadolinium ions (Gd). 3+ Praseodymium ion Pr 3+ With neodymium ions Nd 3+ Co-doping.

3. The single-photon avalanche diode according to claim 1 or 2, characterized in that, The total rare earth ion doping concentration of each functional layer in the n-type InP buffer layer (2), intrinsic InGaAs absorber layer (3), n-type InGaAsP graded layer (4), n-type InP charge layer (5), and intrinsic InP multiplication layer (6) is 1×10⁻⁶. 16 cm -3 ~2×10 18 cm -3 .

4. The single-photon avalanche diode according to claim 1 or 2, characterized in that, The rare earth ions in the intrinsic InGaAs absorber layer (3) are doped using a V-shaped gradient. From the interface between the intrinsic InGaAs absorber layer (3) and the n-type InP buffer layer (2) towards the middle of the intrinsic InGaAs absorber layer (3), the total doping concentration of rare earth ions gradually decreases; from the middle of the intrinsic InGaAs absorber layer (3) towards the interface between the intrinsic InGaAs absorber layer (3) and the n-type InGaAsP gradient layer (4), the total doping concentration of rare earth ions gradually increases. The rare earth ions in the n-type InGaAsP gradient layer (4) are doped using a V-shaped gradient. From the interface between the n-type InGaAsP gradient layer (4) and the intrinsic InGaAs absorber layer (3) towards the middle of the n-type InGaAsP gradient layer (4), the total doping concentration of rare earth ions gradually decreases. From the middle of the n-type InGaAsP gradient layer (4) towards the interface between the n-type InGaAsP gradient layer (4) and the n-type InP charge layer (5), the total doping concentration of rare earth ions gradually increases.

5. The single-photon avalanche diode according to claim 1, characterized in that, The thickness of the n-type InP buffer layer (2) is 500 nm; the thickness of the intrinsic InGaAs absorption layer (3) is 1.5 μm to 3 μm; the thickness of the n-type InGaAsP gradient layer (4) is 100 nm to 200 nm; the thickness of the n-type InP charge layer (5) is 100 nm to 300 nm; and the thickness of the intrinsic InP multiplication layer (6) is 2.5 μm to 3.5 μm.

6. A method for fabricating a single-photon avalanche diode based on multi-element rare-earth co-doping, characterized in that, Includes the following steps: S1. Substrate preparation: Provide an InP substrate (1); S2. Epitaxial growth and in-situ rare earth co-doping: An n-type InP buffer layer (2), an intrinsic InGaAs absorber layer (3), an n-type InGaAsP gradient layer (4), an n-type InP charge layer (5), and an intrinsic InP multiplication layer (6) are epitaxially grown sequentially on the InP substrate (1). During the growth of at least one functional layer, two or more rare earth ion precursors are introduced in-situ through an independent metal-organic source pipeline to achieve rare earth ion co-doping. The rare earth ions are selected from gadolinium ions (Gd). 3+ Lanthanum ion La 3+ Praseodymium ion Pr 3 + Neodymium ions (Nd) 3+ Europium ions (Eu) 3+ ; S3. Device back-end fabrication: A p-type contact region (7) is fabricated on the epitaxially grown structure, a p-type ohmic contact electrode (8) is fabricated, and an n-type ohmic contact electrode (9) is fabricated on the back side of the InP substrate (1) to obtain the single-photon avalanche diode.

7. The method for fabricating a single-photon avalanche diode according to claim 6, characterized in that, The epitaxial growth described in step S2 is achieved using an MOCVD or MBE device. The flow rate of the rare earth ion precursor is controlled by an independent metal-organic source pipeline to achieve uniform or gradient doping of rare earth ions in the functional layer.

8. The method for fabricating a single-photon avalanche diode according to claim 6, characterized in that, In step S2, the in-situ rare earth co-doping process parameters for each functional layer are as follows: During the growth of the n-type InP buffer layer (2), gadolinium ions (Gd) were co-doped in situ. 3+ With lanthanum ion La 3+ The growth temperature is 550℃~640℃, and the total rare earth ion doping concentration is 6×10⁻⁶. 16 cm -3 ~8×10 16 cm -3 Doping flux ratio of gadolinium ions (Gd) 3+ ≥ Lanthanum ion La 3+ ; During the growth of the intrinsic InGaAs absorber layer (3), gadolinium ions (Gd) are co-doped in situ. 3+ Praseodymium ion Pr 3+ With europium ions Eu 3+ The growth temperature is 600℃~680℃, and the total rare earth ion doping concentration is 3×10⁻⁶. 17 cm -3 ~5×10 17 cm -3 Doping flux ratio of gadolinium ions (Gd) 3+ ≥Pr ion 3+ ≥Eu ion 3+ ; During the growth of n-type InGaAsP graded layers (4), gadolinium ions (Gd) were co-doped in situ. 3+ Lanthanum ion La 3+ Neodymium ions (Nd) 3+ With europium ions Eu 3+ The growth temperature is 600℃~680℃, and the total rare earth ion doping concentration is 9×10⁻⁶. 17 cm -3 ~2×10 18 cm -3 Doping flux ratio of gadolinium ions (Gd) 3+ ≥ Lanthanum ion La 3+ ≥Nd ions 3+ ≥Eu ion 3+ The n-type InGaAsP gradient layer (4) serves as a high-concentration doped region. During the growth of the n-type InP charge layer (5), gadolinium ions (Gd) were co-doped in situ. 3+ With lanthanum ion La 3+ The growth temperature is 550℃~640℃, and the total rare earth ion doping concentration is 6×10⁻⁶. 17 cm -3 ~9×10 17 cm -3 Doping flux ratio of gadolinium ions (Gd) 3+ ≥ Lanthanum ion La 3+ ; During the growth of intrinsic InP multiplication layers (6), gadolinium ions (Gd) were co-doped in situ. 3+ Praseodymium ion Pr 3+ With neodymium ions Nd 3+ The growth temperature is 550℃~640℃, and the total rare earth ion doping concentration is 1×10⁻⁶. 17 cm -3 ~3×10 17 cm -3 Doping flow rate ratio Gd 3+ ≥Pr 3+ ≥Nd 3 + .

9. The method for fabricating a single-photon avalanche diode according to claim 6 or 7, characterized in that, In step S2, when growing the intrinsic InGaAs absorber layer (3), V-shaped gradient doping is performed on rare earth ions. By controlling the inlet flow rate of the rare earth ion precursor, the total doping concentration of rare earth ions at the interface of the intrinsic InGaAs absorber layer (3) near the n-type InP buffer layer (2) and the n-type InGaAsP gradient layer (4) is higher than that in the middle of the intrinsic InGaAs absorber layer (3). When growing the n-type InGaAsP graded layer (4), V-shaped gradient doping is performed on rare earth ions. By controlling the influx flow rate of the rare earth ion precursor, the total doping concentration of rare earth ions at the interface between the intrinsic InGaAs absorber layer (3) and the n-type InP charge layer (5) of the n-type InGaAsP graded layer (4) is higher than that at the middle of the n-type InGaAsP graded layer (4).

10. The method for fabricating a single-photon avalanche diode according to claim 6, characterized in that, Step S3 includes the following steps: S31. A secondary Zn diffusion process is used to selectively diffuse Zn onto the surface of the intrinsic InP multiplication layer (6) to form a p-type contact region (7). S32. A dielectric film is deposited on the surface of the intrinsic InP multiplication layer (6) as a passivation anti-reflection layer (10). S33. An electrode contact window (101) is formed on the passivation anti-reflection layer (10) by photolithography and etching processes. S34. Using electron beam evaporation or magnetron sputtering, a Ti / Pt / Au or Cr / Au metal stack is deposited at the electrode contact window (101) to form a p-type ohmic contact electrode (8); and after thinning and polishing the InP substrate (1), an n-type ohmic contact electrode (9) is formed.

Images