Sapphire gallium nitride heterogeneous material epitaxial structure and epitaxial growth method thereof

Abstract

The invention provides a sapphire gallium nitride heterogeneous material epitaxial structure and an epitaxial growth method thereof, and relates to the technical field of semiconductor devices, the epitaxial structure comprises a sapphire substrate, an AlN buffer layer, a GaN channel layer and an AlGaN barrier layer; the sapphire substrate is the foundation of the epitaxial structure; the sapphire substrate is provided with a nano pattern, the nano pattern is specifically a pattern array formed by a plurality of pattern units according to an array mode, and each pattern unit is provided with a nanosphere; the AlN buffer layer is arranged on the sapphire substrate; the GaN channel layer is arranged on the AlN buffer layer and is used for electron transmission; the AlGaN barrier layer is arranged on the GaN channel layer and is used for forming two-dimensional electron gas together with the GaN channel layer; according to the invention, the growth direction of the crystal can be guided and the propagation of dislocation can be limited through the nano pattern, so that the dislocation density is obviously reduced, the defects caused by lattice mismatch are reduced, the GaN epitaxial quality is improved, and the performance and the reliability of a device are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor devices, and particularly to a sapphire gallium nitride heteromaterial epitaxial structure and an epitaxial growth method thereof. Background Art

[0002] Gallium nitride (GaN) semiconductor chips are semiconductor materials widely used in high-power, high-frequency, and high-temperature environments. Due to the wide bandgap, high thermal conductivity, and electrical conductivity of GaN, it can operate stably at higher voltages and frequencies, and is particularly suitable for high-efficiency, long-life power electronic devices and radio frequency (RF) applications.

[0003] The epitaxial structure is of crucial significance in gallium nitride (GaN) semiconductor chips. The epitaxial structure can reduce defects and dislocations caused by lattice mismatch between the substrate and the material. Currently, most epitaxial structures use Si, sapphire, etc. as substrates.

[0004] Currently, whether using Si or sapphire as the substrate, a planar substrate is often used. There is usually a large lattice mismatch between GaN and the substrate material. Due to the lattice mismatch, the planar substrate often accumulates a large number of dislocations and other defects, which directly affect the quality of the epitaxial material and reduce the performance and reliability of the device. Summary of the Invention

[0005] In order to solve the technical problems that traditional planar substrates often accumulate a large number of dislocations and other defects, which directly affect the quality of the epitaxial material and reduce the performance and reliability of the device, the present invention provides a sapphire gallium nitride heteromaterial epitaxial structure and an epitaxial growth method thereof.

[0006] The technical solutions provided by the embodiments of the present invention are as follows: A first aspect of the embodiments of the present invention provides a sapphire gallium nitride heteromaterial epitaxial structure, including: a sapphire substrate, an AlN buffer layer, a GaN channel layer, and an AlGaN barrier layer; The sapphire substrate is the basis of the epitaxial structure; The sapphire substrate has a nano-pattern, and the nano-pattern is specifically a pattern array composed of a plurality of pattern units in an array manner, and nano-spheres are provided on each pattern unit; The AlN buffer layer is disposed on the sapphire substrate; The GaN channel layer is disposed on the AlN buffer layer and is used for electron transport; The AlGaN barrier layer is disposed on the GaN channel layer and is used to jointly form a two-dimensional electron gas with the GaN channel layer.

[0007] The second aspect of the embodiments of the present invention provides a method for epitaxial growth of a sapphire gallium nitride heteromaterial, which is used to prepare the sapphire gallium nitride heterostructure as described in the first aspect. The method includes: S1: Use self-assembled SiO 2 nanospheres as a sacrificial layer etching mask; S2: Transfer the nanosphere pattern to the sapphire surface through dry etching to prepare a sapphire substrate with a nanoscale pattern; S3: Deposit an AlN buffer layer on the sapphire substrate through metalorganic chemical vapor deposition technology; S4: Prepare a stepped channel mask on the AlN buffer layer; S5: Grow a GaN channel layer in the stepped channel mask through metalorganic chemical vapor deposition technology; S6: Deposit an AlGaN barrier layer on the GaN channel layer through metalorganic chemical vapor deposition technology, and form a two-dimensional electron gas together with the GaN channel layer.

[0008] The beneficial effects brought by the technical solutions provided by the embodiments of the present invention at least include: In the present invention, the sapphire substrate has a nanoscale pattern, which can guide the crystal growth direction and limit the propagation of dislocations through the nanoscale pattern, significantly reduce the dislocation density, reduce the defects caused by lattice mismatch, improve the GaN epitaxial quality, and improve the performance and reliability of the device. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0010] Figure 1 FIG. is a schematic structural diagram of a sapphire gallium nitride heterostructure provided by an embodiment of the present invention.

[0011] Figure 2 FIG. is a schematic flow chart of a method for epitaxial growth of a sapphire gallium nitride heteromaterial provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0012] The following will describe the technical solutions in the present invention with reference to the drawings.

[0013] Referring to the attached Figure 1 to the specification, a schematic structural diagram of a sapphire gallium nitride heterostructure provided by an embodiment of the present invention is shown.

[0014] An embodiment of the present invention provides a sapphire gallium nitride heteromaterial epitaxial structure, including: a sapphire substrate 1, an AlN buffer layer 2, a GaN channel layer 3, and an AlGaN barrier layer 4.

[0015] The sapphire substrate 1 is the basis of the epitaxial structure. The sapphire substrate 1 provides a support platform and ensures the epitaxial growth of gallium nitride (GaN) material.

[0016] The sapphire substrate 1 has a nano-pattern, and the nano-pattern is specifically a pattern array composed of a plurality of pattern units in an array manner, and nano-spheres are arranged on each pattern unit.

[0017] It should be noted that the nano-pattern can guide the directional growth of crystals, reduce dislocations and defects caused by lattice mismatch, thereby improving the crystal quality of the GaN layer. The introduction of nano-spheres can form local growth regions during the growth process, promote better fusion of GaN materials in these regions, and reduce the accumulation of cracks and stress.

[0018] The AlN buffer layer 2 is disposed on the sapphire substrate 1. The function of the AlN buffer layer 2 is to relieve the lattice mismatch problem between the sapphire substrate and the GaN channel layer. The AlN buffer layer 2 can provide a smoother interface for the epitaxial growth of GaN, reduce stress, and prevent cracks or dislocations from occurring in GaN during the growth process due to thermal expansion differences.

[0019] The GaN channel layer 3 is disposed on the AlN buffer layer 2 and is used for electron transport. The GaN channel layer 3 is the core layer in the entire heterostructure. GaN has excellent electron mobility and high electron saturation velocity, so it has important applications in high-frequency, high-power electronic and optoelectronic devices.

[0020] The AlGaN barrier layer 4 is disposed on the GaN channel layer 3 and is used to jointly form a two-dimensional electron gas with the GaN channel layer 3. The high energy bandgap characteristic of AlGaN enables it to effectively form a high-density two-dimensional electron gas above the GaN layer, enhancing the conductivity and working efficiency of the device.

[0021] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include: In the present invention, the sapphire substrate has a nano-pattern, which can guide the growth direction of crystals and limit the propagation of dislocations through the nano-pattern, significantly reduce the dislocation density, reduce the defects caused by lattice mismatch, improve the GaN epitaxial quality, and improve the performance and reliability of the device.

[0022] Optionally, the pattern parameters of the nano-pattern include: the shape of the pattern unit, the width of the pattern unit, the number of nano-spheres arranged on each pattern unit, and the radius of the nano-spheres.

[0023] Optionally, the pattern unit shape can be a triangle, a square, or a regular hexagon.

[0024] In a possible implementation, the pattern parameters of the nano-pattern are optimized and determined by the Seagull Optimization Algorithm.

[0025] Among them, the Seagull Optimization Algorithm (SOA) is a heuristic optimization algorithm that mimics the foraging behavior of seagulls. This algorithm simulates the cooperative and competitive behaviors exhibited by seagull groups during the foraging process, and uses the path selection, group collaboration, and distribution strategies of seagulls when searching for food to find the optimal solution to the problem. The Seagull Optimization Algorithm simulates the movement of seagulls and the interaction between groups in the search space, continuously adjusts the search strategy, and quickly finds the global optimal solution or approximate optimal solution to the problem. Due to its strong global search ability and high convergence speed, the Seagull Optimization Algorithm has a wide range of application prospects in solving complex optimization problems.

[0026] Optionally, the fitness function of the Seagull Optimization Algorithm when optimizing the pattern parameters of the nano-pattern is specifically: ; Among them, f( ) represents the fitness function, X represents the set of pattern parameters of the nano-pattern, μ represents the carrier mobility, ρ 2DEG represents the electron density of the two-dimensional electron gas, E br represents the breakdown voltage, λ 1 represents the weight coefficient of the device performance, R represents the GaN interface roughness, δ represents the lattice mismatch, λ 2 represents the weight coefficient of the interface quality.

[0027] Among them, those skilled in the art can set the weight coefficient λ 1 of the device performance and the weight coefficient λ 2 of the interface quality according to the actual situation, and the present invention does not make any limitations.

[0028] In the actual application process, the weight coefficient λ 1 of the device performance can be set to 0.7 and the weight coefficient λ 2 of the interface quality can be set to 0.3.

[0029] In the present invention, the fitness function combines factors related to device performance such as carrier mobility, electron density of two-dimensional electron gas, breakdown voltage, etc., and also considers interface quality factors such as GaN interface roughness and lattice mismatch. By adjusting the weight coefficients of these parameters, the optimization algorithm can effectively improve the electrical performance and structural quality of the device, reduce defects, and ensure the best performance under different working conditions. This can more precisely design the nano-patterns, maximize the device efficiency, and promote the development of high-performance GaN-based devices.

[0030] Optionally, through the seagull optimization algorithm, the pattern parameters of the nano-patterns are optimized and determined, specifically including: Initialize seagull individuals, each seagull individual represents a set of feasible model parameters, each seagull individual consists of multiple dimensional components, and each component represents a model parameter.

[0031] In the global search stage, avoid collisions and move towards the optimal individual: ; ; ; where, represents the position of the i-th seagull individual after the global search stage at the t-th iteration, represents the position of the i-th seagull individual after anti-collision processing at the t-th iteration, A represents the control factor, represents the position of the i-th seagull individual at the t-th iteration, represents the displacement of the i-th seagull individual moving towards the optimal individual at the t-th iteration, B represents the search balance factor, represents the position of the optimal individual at the t-th iteration.

[0032] In the present invention, by introducing a collision avoidance mechanism in the global search stage and making the seagull individuals move towards the optimal individual, the efficiency and accuracy of the search process can be effectively improved. Specifically, by adjusting the control factor A and the search balance factor B, the exploration and exploitation behaviors of the seagull individuals can be balanced, collisions between individuals in the search process can be avoided, unnecessary repeated calculations can be reduced, and thus the search process can be accelerated. At the same time, by guiding the individuals to move towards the optimal solution direction, the algorithm can converge to the global optimal solution or close to the optimal solution more quickly. This method enhances the global search ability and local optimization ability of the seagull optimization algorithm, and improves its effect in solving complex optimization problems.

[0033] A = [ 1 − lg ( 9 t T ) ] × f c ; where, t represents the current iteration number, T represents the maximum iteration number, f c represents the linearly decreasing frequency.

[0034] In the present invention, the value of the control factor is gradually decreased as the number of iterations increases. As the algorithm progresses, the exploration phase gradually turns to a more refined local search, helping the optimization algorithm to converge to the optimal solution or a solution close to the optimal solution more quickly. A larger control factor in the initial stage helps to search the space widely, while in the later stage, a smaller control factor promotes more precise local optimization, avoiding premature convergence and ensuring the balance between the global search ability and the local search ability. This dynamic adjustment strategy improves the search efficiency and accuracy of the algorithm.

[0035] ; where r 1 represents a random number between 0 and 1.

[0036] In the present invention, since the search balance factor is proportional to the square of the control factor, this means that in the initial iteration, a larger control factor will make the search balance factor larger, thereby increasing the breadth of the exploration space and promoting the algorithm to conduct a wider search. As the iteration progresses, the decrease of the control factor will make the search balance factor gradually decrease, prompting the algorithm to conduct a more concentrated local search, thereby improving the search accuracy and accelerating the convergence speed. Through this dynamic adjustment, the stability and efficiency of the algorithm can be maintained, which helps to improve the ability to find the optimal solution or an approximate optimal solution.

[0037] In the local search phase, a random number r 2 is generated. According to the random number r 2 , a parallel selection is made between the spiral search strategy and the surrounding strategy, and the displacement is carried out in a spiral motion manner: ; ; ; ; ; where represents the position of the i-th seagull individual after spiral motion at the t-th iteration, x represents the spiral flight coefficient in the x direction, y represents the spiral flight coefficient in the y direction, z represents the spiral flight coefficient in the z direction, r represents the radius of the spiral flight trajectory, θ represents a random number between 0 and 2π, u and v represent spiral constants, and e represents the natural constant.

[0038] Optionally, the spiral constant u can be 2, and the spiral constant v can be 3.

[0039] In the present invention, by introducing a parallel selection of the spiral search strategy and the surrounding strategy in the local search phase, the flexibility and accuracy of the search can be effectively improved. When the random number When, the seagull individuals displace through spiral motion, making the search more diverse and enabling exploration of a wider area in the search space. When it is, the encirclement strategy is adopted to make the individuals approach the optimal solution more concentratedly. This dynamic selection mechanism can enhance the local search accuracy while ensuring the global search ability. The introduction of the spiral flight trajectory further increases the diversity of the search process. Through adjustment by factors such as the natural constant and the spiral constant, progressive optimization of the search process can be achieved, avoiding the dilemma of falling into local optimal solutions. Overall, this strategy enables the seagull optimization algorithm to exhibit higher flexibility, accuracy, and convergence efficiency in complex optimization problems.

[0040] Perform mutation operations on each seagull individual: .

[0041] Among them, represents the position of the i-th seagull individual after mutation at the t-th iteration, P r represents a random individual, and ω represents an adaptive scaling factor.

[0042] In the present invention, by introducing the mutation operation, each seagull individual adjusts its position based on the difference between the optimal individual and the random individual on the basis of the current position, thereby avoiding falling into local optimal solutions.

[0043] ω = ω max − ( ω max − ω min ) sin [ π 2 ( t T ) 2 ] ; Among them, ω max represents the maximum scaling factor, and ω min represents the minimum scaling factor.

[0044] Optionally, the maximum scaling factor ω max can be 2, and the minimum scaling factor ω min can be 0.5.

[0045] In the present invention, as the number of iterations increases, the scaling factor gradually changes from the maximum value to the minimum value. This change enables the algorithm to conduct extensive searches in the initial stage, exploring more potential solution spaces, while in the later stage, by gradually reducing the scaling factor, the algorithm conducts fine searches in the local area, thereby improving the convergence speed and reducing the jumpiness. This adaptive adjustment strategy helps to avoid premature convergence and enhance the global search ability of the algorithm, while ensuring accurate local search when approaching the optimal solution.

[0046] Judge whether the fitness value of the position after mutation is greater than the fitness value of the position before mutation. If so, replace the position before mutation with the position after mutation. Otherwise, keep the position before mutation unchanged.

[0047] Update the fitness values of each seagull individual and the globally optimal individual.

[0048] Determine whether the current iteration count has reached the maximum iteration count. If so, output the set of model parameters represented by the seagull individual with the highest current fitness. Otherwise, return for continued iteration.

[0049] In the present invention, by optimizing the pattern parameters of the nano-pattern through the seagull optimization algorithm, an efficient balance and optimization can be achieved among multi-dimensional performance indicators. By simulating the behavior of a seagull population, this algorithm can quickly search for the globally optimal solution or an approximate optimal solution, thereby optimizing the design of the nano-pattern and improving the performance of the device. In device design, this method can effectively improve key parameters such as carrier mobility, density of two-dimensional electron gas, breakdown voltage, etc., while reducing interface roughness and lattice mismatch, and optimizing the overall device performance and structural quality. This optimization strategy helps to precisely control the pattern parameters and promotes the development of high-efficiency, low-defect, and high-performance GaN-based devices.

[0050] Refer to the attached Figure 2 illustrates a schematic flowchart of a method for epitaxial growth of a sapphire gallium nitride heteromaterial provided by an embodiment of the present invention.

[0051] An embodiment of the present invention provides a method for epitaxial growth of a sapphire gallium nitride heteromaterial for preparing the above-mentioned sapphire gallium nitride heterostructure. The method includes: S1: Using self-assembled SiO 2 nanospheres as a sacrificial layer etching mask. S2: Through dry etching, transfer the nanosphere pattern to the sapphire surface to prepare a sapphire substrate 1 with a nano-pattern.

[0052] S3: Deposit an AlN buffer layer 2 on the sapphire substrate 1 through metalorganic chemical vapor deposition technology.

[0053] Among them, metalorganic chemical vapor deposition (MOCVD) is a technology widely used in the preparation of semiconductor materials, optoelectronic devices, and other functional thin film materials. This method deposits the required thin film material by introducing metalorganic precursors and gas reactants into the reaction chamber and carrying out chemical reactions under high-temperature conditions.

[0054] S4: Prepare a stepped channel mask on the AlN buffer layer 2.

[0055] It should be noted that the stepped-channel mask enables GaN to experience multiple turns during growth above the mask layer region. During the turning process, dislocations are effectively filtered out, and thus high-quality GaN is grown. Through the geometric shape of the stepped channel, the growth direction and nucleation position of GaN are controlled, effectively promoting the uniform growth of GaN crystals and reducing the generation of defects. The design of the stepped channel can improve the quality of the GaN layer by improving the nucleation and growth rate of GaN, while optimizing the surface flatness of the layer and reducing the surface roughness.

[0056] In a possible implementation, S4 specifically includes sub-steps S401 to S407: S401: Deposit and form a first Si 3 N 4 mask layer on the AlN buffer layer 2 through low-pressure chemical vapor deposition technology.

[0057] Among them, low-pressure chemical vapor deposition (LPCVD) is a technology for thin-film deposition under low-pressure conditions. By introducing gaseous precursors into the reaction chamber, chemical reactions occur at high temperatures, forming a thin film and depositing it on the substrate.

[0058] S402: Remove part of the material on the first Si 3 N 4 mask layer through standard photolithography and reactive ion etching technologies, and etch out a first parallel strip window array.

[0059] S403: Deposit a SiO 3 N 4 layer on the first Si 2 mask layer through plasma-enhanced chemical vapor deposition technology.

[0060] Among them, plasma-enhanced chemical vapor deposition (PECVD) is a technology that uses low-temperature plasma to excite gaseous precursors, causing them to undergo chemical reactions and deposit into thin films.

[0061] S404: Deposit a second Si3N4 mask layer on the SiO 2 layer through low-pressure chemical vapor deposition technology.

[0062] S405: Remove part of the material on the second Si 3 N 4 mask layer through standard photolithography and reactive ion etching technologies, and etch out a second parallel strip window array.

[0063] Optionally, the second parallel strip window array has the same size as the first parallel strip window array.

[0064] Optionally, the width of each strip window in the first parallel strip window array is 2 μm, the distance between adjacent strip windows is 12 μm, the width of each strip window in the second parallel strip window array is 2 μm, and the distance between adjacent strip windows is 12 μm.

[0065] S406: When the second parallel strip window array is laterally displaced relative to the first parallel strip window array, a stepped channel is formed.

[0066] S407: Partially remove the SiO 2 layer with buffered oxide etchant until the bottom window exposes the AlN buffer layer 2, obtaining a stepped channel mask.

[0067] S5: Grow a GaN channel layer 3 in the stepped channel mask by metal organic chemical vapor deposition technology.

[0068] In a possible implementation, S5 specifically includes sub-steps S501 and S502: S501: Using TMGa and NH 3 as precursors and H 2 as the carrier gas, maintain the substrate temperature at 530 °C and grow GaN in the stepped channel mask.

[0069] S502: Raise the substrate temperature to 1060 °C and hold for 240 minutes, and continue to grow GaN on the existing nucleation sites to form the GaN channel layer 3.

[0070] S6: Deposit an AlGaN barrier layer 4 on the GaN channel layer 3 by metal organic chemical vapor deposition technology, and jointly form a two-dimensional electron gas with the GaN channel layer 3.

[0071] In the present invention, first, using SiO 2 nanospheres as a sacrificial layer etching mask can form high-precision nano-patterns on the sapphire surface, reduce dislocations and defects, and improve the material quality. Then, depositing the AlN buffer layer helps to relieve the lattice mismatch between sapphire and GaN and improve the quality of the GaN channel layer. The preparation of the stepped channel mask further optimizes the growth environment of GaN and helps to form a high-quality two-dimensional electron gas 2DEG. Through this multi-step fine control, the performance of the device can be significantly improved, including higher electron mobility and lower defect density, which is suitable for high-frequency and high-performance nitride-based electronic devices.

[0072] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention.

Claims

1. A sapphire gallium nitride heterogeneous material epitaxial structure, characterized in that: include: Sapphire substrate, AlN buffer layer, GaN channel layer and AlGaN barrier layer; The sapphire substrate is the basis of the epitaxial structure; The sapphire substrate has a nano pattern, and the nano pattern is specifically a pattern array composed of a plurality of pattern units in an array manner, and each of the pattern units is provided with a nano ball; The AlN buffer layer is disposed on the sapphire substrate; The GaN channel layer is disposed on the AlN buffer layer for electron transmission; The AlGaN barrier layer is disposed on the GaN channel layer, and is used to form a two-dimensional electron gas together with the GaN channel layer.

2. The sapphire gallium nitride heterogeneous material epitaxial structure according to claim 1, characterized in that: The pattern parameters of the nano-pattern include: pattern unit shape, pattern unit width, the number of nano-balls arranged on each pattern unit, and the radius of the nano-balls.

3. The sapphire gallium nitride heterogeneous material epitaxial structure according to claim 2, characterized in that: The pattern parameters of the nanopattern are optimized and determined by the Seagull optimization algorithm.

4. The sapphire gallium nitride heterogeneous material epitaxial structure according to claim 3, characterized in that: The fitness function of the Seagull optimization algorithm when optimizing the pattern parameters of the nanopattern is specifically: ； Where f( ) represents the fitness function, X represents the pattern parameter set of the nanopattern, μ represents the carrier mobility, and ρ 2DEG represents the electron density of the two-dimensional electron gas, E br represents the breakdown voltage, λ1 represents the weight coefficient of device performance, R represents the GaN interface roughness, δ represents the lattice mismatch, and λ2 represents the weight coefficient of interface quality.

5. The sapphire gallium nitride heterogeneous material epitaxial structure according to claim 3, characterized in that: The step of optimizing and determining the pattern parameters of the nanopattern by using the Seagull optimization algorithm specifically includes: Initialize individual seagulls. Each individual seagull represents a feasible set of model parameters. Each individual seagull consists of multiple dimensional components, and each component represents a model parameter. In the global search phase, avoid collisions and move towards the optimal individual: ； ； ； in, represents the position of the i-th seagull individual after the global search phase at the t-th iteration, represents the position of the i-th seagull after anti-collision processing at the t-th iteration, A represents the control factor, represents the position of the i-th seagull at the t-th iteration, represents the displacement of the i-th seagull individual towards the optimal individual at the t-th iteration, B represents the search balance factor, represents the optimal individual position at the tth iteration; In the local search phase, a random number r2 is generated. Based on the random number r2, the spiral search strategy and the encirclement strategy are selected in parallel, and the displacement is performed in a spiral motion: ； ； ； ； ； in, represents the position of the i-th seagull after the spiral motion at the t-th iteration, x represents the spiral flight coefficient in the x direction, y represents the spiral flight coefficient in the y direction, z represents the spiral flight coefficient in the z direction, r represents the radius of the spiral flight trajectory, θ represents a random number between 0 and 2π, u and v represent spiral constants, and e represents a natural constant; Perform mutation operations on each seagull individual: ； in, represents the position of the ith seagull individual after mutation at the tth iteration, P r represents a random individual, ω represents an adaptive scaling factor; Determine whether the fitness value of the position after mutation is greater than the fitness value of the position before mutation; if so, use the position after mutation to replace the position before mutation; otherwise, keep the position before mutation unchanged; Update the fitness value of each seagull individual and the global optimal individual; Determine whether the current number of iterations has reached the maximum number of iterations; if so, output the model parameter set representing the seagull individual with the highest current fitness; otherwise, return to continue iterating.

6. A method for epitaxial growth of sapphire gallium nitride heterogeneous material, characterized in that: For preparing the sapphire gallium nitride heterogeneous material epitaxial structure according to any one of claims 4 to 5, the method comprises: S1: using self-assembled SiO2 nanospheres as sacrificial etching mask; S2: transferring the nanosphere pattern to the sapphire surface by dry etching to prepare the sapphire substrate with the nano pattern; S3: depositing the AlN buffer layer on the sapphire substrate by metal organic chemical vapor deposition technology; S4: preparing a stepped channel mask on the AlN buffer layer; S5: growing a GaN channel layer in the stepped channel mask by metal organic chemical vapor deposition technology; S6: depositing the AlGaN barrier layer on the GaN channel layer by metal organic chemical vapor deposition technology, and forming a two-dimensional electron gas through the AlGaN barrier layer and the GaN channel layer.

7. The sapphire gallium nitride heterogeneous material epitaxial growth method according to claim 6, characterized in that: The S4 specifically includes: S401: depositing a first Si3N4 mask layer on the AlN buffer layer by low pressure chemical vapor deposition technology; S402: removing part of the material on the first Si3N4 mask layer by standard photolithography and reactive ion etching technology, and etching out a first parallel strip window array; S403: depositing a SiO2 layer on the first Si3N4 mask layer by plasma enhanced chemical vapor deposition technology; S404: depositing a second Si3N4 mask layer on the SiO2 layer by low pressure chemical vapor deposition technology; S405: removing part of the material on the second Si3N4 mask layer by standard photolithography and reactive ion etching technology, and etching out a second parallel strip window array; S406: When the second parallel strip window array and the first parallel strip window array are laterally phase-shifted, a stepped channel is formed; S407: partially removing the SiO2 layer by using a buffered oxide etchant until the bottom window exposes the AlN buffer layer, thereby obtaining the stepped channel mask.

8. The sapphire gallium nitride heterogeneous material epitaxial growth method according to claim 7, characterized in that: The second parallel strip window array has the same size as the first parallel strip window array.

9. The sapphire gallium nitride heterogeneous material epitaxial growth method according to claim 8, characterized in that: The width of each strip window in the first parallel strip window array is 2 μm, and the distance between adjacent strip windows is 12 μm. The width of each strip window in the second parallel strip window array is 2 μm, and the distance between adjacent strip windows is 12 μm.

10. The sapphire gallium nitride heterogeneous material epitaxial growth method according to claim 6, characterized in that: The S5 specifically includes: S501: using TMGa and NH3 as precursors and H2 as carrier gas, maintaining the substrate temperature at 530° C., and growing GaN in the stepped channel mask; S502: Raise the substrate temperature to 1060° C. and maintain it for 240 minutes, and continue to grow GaN on the existing nucleation points to form a GaN channel layer.

Images