Dynamic evolution analysis method based on silicon carbide polycrystalline nucleus crystal orientation identification data
By using molecular dynamics simulations and polycrystalline material model analysis, the problem of accurately capturing the polycrystalline nucleus orientation competition growth process of silicon carbide was solved, thereby improving the epitaxial quality and device performance of silicon carbide heterojunctions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-04
- Publication Date
- 2026-04-07
AI Technical Summary
Existing technologies struggle to accurately capture the competition and growth process of polycrystalline nuclei in silicon carbide, affecting the epitaxial quality of heterojunction films and device performance.
By creating a molecular dynamics simulation model of silicon carbide heterojunction, the interatomic interactions at different temperatures are simulated, a polycrystalline material model is constructed, the crystal orientation of polycrystalline nuclei is identified, and dynamic evolution analysis is performed by combining radial distribution function, surface roughness and dislocation identification techniques.
It enables precise capture of the dynamic evolution state of polycrystalline nuclei and crystal orientation in silicon carbide, improves the quality analysis and optimization capabilities of silicon carbide materials, optimizes the crystal growth process, and enhances device performance.
Smart Images
Figure CN121811984A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of material science, in particular to a dynamic evolution analysis method based on silicon carbide polycrystalline nucleus crystal direction identification data. BACKGROUND
[0002] Silicon carbide heterojunction technology is the core direction of the third generation of semiconductors. With the advantages of wide band gap, high thermal conductivity and high breakdown field strength of silicon carbide material, it has promoted the application breakthrough of high-voltage power devices in the fields of new energy vehicles, smart grids and the like. The research on silicon carbide heterojunction epitaxial structure has shown important application potential, and has been widely used in electric vehicle inverters, power conversion and photovoltaic inverter systems.
[0003] The epitaxial quality of silicon carbide heterojunction film is closely related to the initial growth of the crystal nucleus, and the competitive growth of the crystal nucleus has a crucial influence on the performance of the device. It is found that as the temperature increases (from 2800K to 3250K), the crystal growth rate significantly accelerates; the cubic and hexagonal structures at the heterojunction solid-liquid interface show intense competitive growth, and the dominant mode and interface integration effect are significantly affected by temperature regulation; defect analysis shows that the hexagonal structure is more prone to lattice disorder, and the initial microcavity expands during growth, especially inducing large-size cavity defects at the heterojunction interface. This provides an atomic-level understanding of the growth mechanism of silicon carbide heterojunction, helps to reveal the dynamic process of crystal competitive evolution and the defect generation rule, and provides a scientific basis for the controllable preparation of high-quality semiconductor heterojunction, thereby optimizing the crystal growth process design and improving the performance and reliability of the device.
[0004] Although theoretical calculations and simulations have been widely used in crystal nucleus growth prediction, due to the limitation of calculation scale, it is difficult to accurately capture the crystal nucleus competition and growth process. Therefore, a technical scheme is needed to realize the dynamic evolution analysis of silicon carbide polycrystalline nucleus crystal direction identification data. SUMMARY
[0005] To achieve the above purpose, the present application provides a dynamic evolution analysis method based on silicon carbide polycrystalline nucleus crystal direction identification data, which is used for silicon carbide heterojunction semiconductor material analysis, comprising: creating a molecular dynamics simulation model of a silicon carbide heterojunction, generating a heterojunction model according to environmental parameters; performing molecular dynamics simulation to simulate the thermodynamic behavior of atomic interaction at different temperatures, and obtaining atomic behavior data during simulation; constructing a polycrystalline material model containing different grain orientations, identifying the crystal direction of the polycrystalline nucleus according to the heterojunction surface and internal structure of the polycrystalline material model, and extracting the directionality information of the grain; According to the directionality information of the crystal grains, data statistics and optimization are performed to obtain and display the grain distribution law.
[0006] The molecular dynamics simulation model for creating the silicon carbide heterojunction is implemented in a molecular dynamics simulation software, and includes the following steps: The silicon carbide cell structure is defined, and a heterojunction model composed of 2H-SiC and 3C-SiC is established; the silicon carbide cell structure supports the layered arrangement of 2H-SiC and 3C-SiC in a 3:2 ratio, and the size is adjusted by functional cell expansion; The force field parameters are set.
[0007] The molecular dynamics simulation is performed by setting the simulation dimension, boundary condition and potential function; The atomic behavior data includes the three-dimensional coordinates, velocity and force data of the atoms.
[0008] The thermodynamic behavior of the atomic interaction at different temperatures is simulated by using the NPT ensemble, including: By setting appropriate periodic boundary conditions, the behavior simulation of the material at a large scale is realized; According to the simulation requirements, appropriate potential functions are selected to describe the atomic interaction, and the temperature change is controlled to simulate the thermodynamic behavior at different temperatures; The atomic behavior data is obtained during the simulation process.
[0009] Further, the interaction between the atoms is realized by the Tersoff potential function; The Tersoff potential function is represented as: , , Wherein, E is the total energy of the SiC material, which is the sum of the contributions of all atomic pairs, i N is the number of atoms, j N is the number of interacting neighbor atoms, represents the contribution of the i th atom; is a cutoff function, represents the repulsive term, represents the attractive term, and the constant is used to adjust the relative strength of the repulsive and attractive terms.
[0010] Further, the polycrystalline material model containing different grain orientations is constructed by radial distribution function, surface roughness and dislocation identification; The radial distribution function is used to reflect the interaction between particles and the structural characteristics of the system. Surface roughness is used to reflect the morphology of the surface; Dislocation identification is used to reveal the influence of factors such as temperature, strain rate, crystal orientation, etc. on the behavior of dislocations.
[0011] wherein the radial distribution function is expressed as: , wherein, is the area density, is the total volume of the crystal, is the total number of atoms in the crystal, is the radius of the spherical shell, is the average number of atoms distributed in the spherical shell volume from to .
[0012] The calculation method of surface roughness includes: calculating a set of surface atomic height values along a given reference direction, denoted as: wherein, is the set of surface atomic height values, is the surface atomic height value along the given reference direction, is the number of elements in the set of surface atomic height values; calculating the average height of the surface atoms , denoted as: ; calculating the mean square deviation of and , denoted as: wherein, is the mean square deviation.
[0013] wherein, data statistics and optimization include the following steps: adopting a grain identification algorithm to extract the crystal orientation parameters of each grain, and calculating the orientation difference between adjacent grains; wherein, the crystal orientation parameters include Euler angles, principal axis directions; statistically analyzing the proportion of low-angle and high-angle grain boundaries, and analyzing the influence on grain boundary energy and migration behavior; combining grain size distribution, grain boundary type and orientation dense area for statistics, performing image visualization, and realizing dynamic evolution analysis.
[0014] Further, the dynamic evolution analysis includes: statistically analyzing the atomic data in the relaxation growth process under isothermal conditions corresponding to different temperatures, realizing dynamic evolution process analysis of SiC heterojunction crystal structure, and visualizing the dynamic evolution process; The changes in the number of atoms of a single atom type during relaxation growth under isothermal conditions at different temperatures were statistically analyzed, and the changes were visualized. The crystallization rate of SiC heterojunction during relaxation growth under different isothermal conditions was statistically analyzed, and the crystallization rate under different conditions was visualized. The surface morphology and size changes of the atomic model of the growth interface during relaxation growth under different isothermal conditions were statistically analyzed, and the dynamic evolution of height evolution and crystal surface roughness was visualized. Atomic data at the solid-liquid interface during relaxation under isothermal conditions at different temperatures were statistically analyzed, and atomic slice diagrams were generated to reflect the structural states of different regions in the liquid phase near the solid-liquid interface. Growth rules were analyzed and visualized. The process of defect state change under isothermal conditions at different temperatures is calculated, and the location, size and shape of defects are observed by removing crystal atoms through visualization software DXA. The structure of wurtzite and zincblende atoms during relaxation under isothermal conditions at different temperatures, as well as their different stacking patterns at grain boundaries, are statistically analyzed and visualized using visualization software.
[0015] This invention constructs a crystal orientation identification and visualization analysis model for silicon carbide heterojunction semiconductor materials based on polycrystalline nucleus orientation identification data analysis and image visualization technology. By accurately extracting the orientation information of the grains and combining methods such as radial distribution function and surface roughness analysis, a comprehensive analysis is conducted on the growth, competition, and evolution process of polycrystalline nuclei with different crystal orientations on the substrate. Simultaneously, image visualization technology is used to transform the directional data of the grains into intuitive anisotropic images, and quantitative analysis is used to study the interrelationships between grains, the formation mechanism of grain boundaries, and their impact on material properties, thereby providing a scientific basis for the optimized processing of silicon carbide heterojunction materials. Attached Figure Description
[0016] Figure 1 This is a flowchart illustrating the dynamic evolution analysis method for silicon carbide polycrystalline nucleus orientation identification data according to an embodiment of the present invention. Figure 2 This is a molecular dynamics simulation model diagram provided according to an embodiment of the present invention; Figure 3 This is a diagram of the average atomic energy at different crystallization rates and different relaxation temperatures provided according to an embodiment of the present invention; Figure 4 This is a graph showing the change in the number of sphalerite and wurtzite atoms with relaxation time at different temperatures, according to an embodiment of the present invention. Figure 5Snapshot graphs of 10, 25, 40, 55 and 70 ns of each temperature once under isothermal relaxation are provided according to the embodiment of the present application; Figure 6 The morphological change process schematic diagram of wurtzite and zinc blende is provided according to the embodiment of the present application; Figure 7 The slice graph at different X coordinates in the liquid phase region near the solid-liquid interface at different times at a temperature of 3250K is provided according to the embodiment of the present application; Figure 8 The crystal front and top surface defect analysis graph under each temperature state is provided according to the embodiment of the present application; Figure 9 The type graph of crystal growth defects in wurtzite and zinc blende is provided according to the embodiment of the present application; Figure 10 The formation process schematic diagram of different stacking structures of wurtzite and zinc blende at the interface is provided according to the embodiment of the present application. DETAILED DESCRIPTION
[0017] The present application provides a dynamic evolution analysis method based on silicon carbide polycrystal nucleus crystal recognition data, which realizes the capture of the dynamic evolution state in the silicon carbide polycrystal material nucleus crystal recognition process, realizes the data analysis of the dynamic evolution state, and uses image visualization technology to display, and further improves the quality analysis and optimization ability of silicon carbide material.
[0018] The specific implementation mode of the present application will be described in detail below in conjunction with the drawings of the specification.
[0019] The dynamic evolution analysis method provided by the present application, as shown in Figure 1 includes the following steps: The dynamic evolution analysis method provided by the present application is used for silicon carbide heterojunction semiconductor material analysis, including: Step S100: Create a molecular dynamics simulation model of silicon carbide heterojunction, and generate a heterojunction model according to environmental parameters; In the molecular dynamics simulation software (such as Lammps software), select a suitable crystal structure and force field to model and generate a molecular dynamics simulation model of silicon carbide heterojunction, which specifically includes the following steps: 1) Define the silicon carbide unit cell structure in the Lammps software, and establish a heterojunction model composed of 2H-SiC and 3C-SiC two phases; In terms of composition, 2H-SiC and 3C-SiC in the heterojunction can be arranged in multiple layers according to different proportions as required; such as Figure 2 The molecular dynamics simulation model is arranged in layers according to a proportion of 3:2 to ensure that the physical properties of the model are consistent with the experimental results; In terms of shape, the silicon carbide heterojunction can adopt a cuboid, or other shaped substrates; In terms of size, by performing functional expansion of the 2H-SiC and 3C-SiC layers and adjusting the size of the crystal, a heterojunction model of a specified size is generated to meet different simulation needs; for example, using the "Supercell" function of Material Studio to expand the cell, a heterojunction cuboid model with a size of 67.8x75.6x53.4nm³ is generated, and is spliced with a silicon carbide liquid model of the same size.
[0020] At the same time, the environment can also be adjusted, with temperature, pressure or other conditions in the silicon carbide heterojunction as variables to set the environmental conditions, to ensure the scalability and flexibility of the model.
[0021] 2) After the molecular dynamics simulation model is created, necessary force field parameter settings are performed: In terms of force field setting, the Tersoff potential function is selected to describe the interaction between atoms, ensuring that the interaction force and interface effect between different crystal phases can be accurately simulated. According to specific needs, the parameters in the force field can also be fine-tuned to improve the accuracy of the model and ensure its high reliability and stability in predicting the behavior of silicon carbide materials. By reasonably setting the simulation parameters and optimizing the force field, the accuracy of the molecular dynamics simulation results is ensured, providing a solid foundation for subsequent crystal orientation recognition and structure analysis.
[0022] Step S110: Perform molecular dynamics simulation to simulate the thermodynamic behavior of atomic interaction at different temperatures, and obtain atomic behavior data during simulation; In this step, by setting simulation parameters such as simulation dimension, boundary condition, potential function and temperature control, the processing of silicon carbide materials under different environments is simulated, and relevant structure data is collected.
[0023] Specifically, by setting appropriate periodic boundary conditions, the behavior of materials at large scales is simulated; according to simulation needs, appropriate potential functions are selected to describe atomic interactions, and temperature changes are controlled to simulate thermodynamic behavior at different temperatures; and atomic behavior data is obtained during simulation, including atomic three-dimensional coordinates, velocity and force data, etc.; atomic behavior data can support subsequent crystal orientation recognition and grain evolution analysis, and further optimize the processing of silicon carbide materials.
[0024] 1) Determine the interaction between atoms in the zincite structure of SiC fibers, and calculate atomic behavior data by Tersoff potential function; The Tersoff potential function is expressed as: , , in, The total energy of SiC material is the sum of the contributions from the interactions between all atomic pairs. i Indicates the number of atoms, j This indicates the number of neighboring atoms that interact with each other. Indicates the first i The contribution of each atom; due to the total energy It is calculated by summing the interaction energies between all atomic pairs, while avoiding duplicate calculations (when...). i equal j hour).
[0025] It is a cutoff function. Indicates the exclusion term. Represents the attraction term, a constant. Used to adjust the relative strength of repulsive and attractive terms.
[0026] Atomic behavior data simulation can effectively predict the growth process and structural evolution of materials, and the simulation results can be visualized using open visualization tools (such as Ovito), enabling the analysis of the dynamic behavior of heterogeneous crystal boundaries during crystal growth.
[0027] 2) Use the NPT ensemble to simulate the thermodynamic behavior of interatomic interactions at different temperatures, and adjust the pressure and temperature during the simulation to maintain equilibrium conditions that reflect real-world applications; By replicating the simulation unit, the infinite system with periodic boundary conditions can be simulated, thus reducing the influence of the boundary on the simulation results. The NPT ensemble controlled the solid-liquid hybrid system to relax for 30 ps at 300 K to release initial stress and achieve contact at the phase interface; it was then relaxed for 70 ns at 2800, 3000, 3100 and 3250 K to induce growth; a time step of 1 fs was selected to accurately capture rapid changes in atomic motion and ensure the accuracy of the simulation results.
[0028] Step S120: Construct a polycrystalline material model containing different grain orientations. Based on the heterojunction surface and internal structure of the polycrystalline material model, identify the crystal orientation of the polycrystalline nuclei and extract the directional information of the grains. This step is achieved through methods such as radial distribution function, surface roughness, and dislocation identification.
[0029] 1) The internal structure is represented by a radial distribution function: The radial distribution function can describe the spatial distribution of particles in a multi-particle system at the microscopic level, reflecting the interactions between particles and the structural characteristics of the system. It helps researchers understand and analyze the physical and chemical properties of the system at the microscopic level, and is expressed as: , in, For regional density, The total volume of the crystal. It is the total number of atoms in the crystal. Let be the radius of the spherical shell. It is evenly distributed from arrive The number of atoms within the volume of the spherical shell.
[0030] The height of the regional density peaks reflects the strength of the interaction between particles. The higher the peak, the stronger the interaction, and the more the particles tend to be distributed around that distance. In addition, the number, location, and spacing of the peaks can reveal the microstructure of the system, such as the periodic structure of crystals or the short-range ordered structure of liquids.
[0031] 2) Surface roughness is used to reflect the morphology of a surface: On the one hand, there is a very close correlation between the zincblende nucleation of polycrystalline silicon carbide and its subsequent evolution to wurtzite; on the other hand, crystal growth exhibits a domino-like chain effect. The morphology of the induced surface plays a decisive role in obtaining high-quality crystals, therefore, in-depth research into crystal morphology is of paramount importance. In this invention, specific crystal surfaces are selected, and the surface roughness of zincblende and wurtzite is calculated using the following method: the average atomic height of the surface: The set of surface atomic height values calculated along a given reference direction is represented as: ,in, It is a set of surface atomic height values. The surface atomic height value along a given reference direction. The number of elements in the set of surface atomic height values; Calculate the average height of surface atoms , is represented as: ; calculate and The mean squared error is expressed as: ,in, The mean squared error is denoted as .
[0032] In this invention, the crystal growth direction is set along the x-axis, so the height direction involved also corresponds to the x-axis direction.
[0033] 3) Using the dislocation identification method: In this invention, a dislocation identification method is used to track the position and movement of atoms, and combined with crystal structure analysis algorithms, the nucleation, slip, climb, cross-slip, reaction, multiplication and annihilation processes of dislocations at the atomic scale are dynamically visualized and quantified.
[0034] The implementation of dislocation identification involves applying stress and strain to a crystal model and capturing in real time the precise atomic configuration of dislocation lines in the lattice, the evolution of the core structure, and their interaction with defects such as vacancies, solutes, and grain boundaries.
[0035] Dislocation identification methods can reveal the influence mechanism of factors such as temperature, strain rate, and crystal orientation on dislocation behavior, providing dynamic evidence for further research.
[0036] In this invention, the simulation process of the model is observed to have fewer dislocation lines by using visualization DXA technology.
[0037] Step S130: Perform data statistics and optimization based on the orientation information of the grains to obtain and display the grain distribution pattern.
[0038] In step S120, the orientation information of the grains is accurately extracted using methods such as radial distribution function, surface roughness analysis, and dislocation identification. In this step, statistical analysis of the grain distribution pattern is performed based on the orientation information to further identify the orientation relationship and influencing factors between grains, optimize the crystal orientation data, and ensure the accuracy of the grain orientation information.
[0039] Data statistics and optimization include the following steps: A grain identification algorithm is used to extract the crystal orientation parameters of each grain and calculate the orientation difference between adjacent grains; the crystal orientation parameters include information such as Euler angles and principal axis directions. The ratio of small-angle to large-angle grain boundaries was statistically analyzed to determine its impact on grain boundary energy and migration behavior. By combining statistical analysis of grain size distribution, grain boundary type, and orientation-dense regions, it is possible to achieve, for example... Figures 3 to 10 The image visualizations shown support various display methods, including grain pole figures, orientation maps, and snapshots. Image visualization can be used to reveal intergranular interaction mechanisms and material texture characteristics, perform dynamic evolution analysis of grain distribution patterns, and provide a theoretical basis for the microstructure control and process optimization of silicon carbide materials.
[0040] The data statistics and optimizations provided in this invention include the following: 1) Collect atomic data during the relaxation growth process under isothermal conditions at different temperatures to analyze the dynamic evolution process of SiC heterocrystalline structure and visualize the dynamic evolution process; Atomic data is provided via LAMMPS, atomic energies are obtained using Ovito visualization software, and the time interval for crystallization calculations is set (e.g., 1000 ps). Figure 3 As shown: Part (a) is a schematic diagram of the crystallization rate at different temperatures, and Part (b) is a diagram of the average atomic energy during relaxation at different temperatures. As can be seen in the diagram, when isothermal relaxation begins, atoms at the interface drastically adjust their positions to lower the system energy and reach a more stable state. Due to the directionality and saturation of the covalent bonds between silicon carbide atoms, this drastic adjustment may cause some atomic bonds to break, allowing the atoms to gain enough energy to escape the original lattice constraints. Furthermore, the heterostructure may have local defects during construction, further weakening the interatomic bonding forces.
[0041] 2) Statistically analyze the changes in the number of atoms of a single atom type during relaxation growth under isothermal conditions at different temperatures, and visualize the changes. Figure 4 As shown, the atomic number of zinc sphalerite and wurtzite varies with relaxation time at different temperatures; (a) is zinc sphalerite, and (b) is wurtzite. From the perspective of crystal structure, zinc sphalerite has a face-centered cubic structure with high symmetry, while wurtzite has a hexagonal close-packed structure. When the two combine to form a heterojunction, lattice mismatch will occur at the interface, which will generate specific interfacial stress and strain field, promote the redistribution of electronic state density, and thus affect the electrical properties of the material, effectively regulating the migration and distribution of charge carriers.
[0042] 3) The crystallization rate of SiC heterojunction during the relaxation growth process under different isothermal conditions at different temperatures is statistically analyzed, which can be used to analyze the competitive ability during the atomic growth process; the crystallization rate under different states is visualized using Ovito visualization software; like Figure 5 This diagram illustrates the isothermal relaxation states of atoms at 10, 25, 40, 55, and 70 ns at 2800, 3000, 3100, and 3250 K, denoted by a, b, c, and d, respectively. Figure 5 The visualization clearly shows that, under the current induction conditions, the growth rate of wurtzite is lower than that of sphalerite (the blue part of sphalerite and the orange part of wurtzite, with white representing the amorphous liquid state). This also reflects that sphalerite atoms have a stronger competitive advantage during the growth of both.
[0043] 4) Statistically analyze the surface morphology and size changes of the atomic model of the growth interface during the relaxation growth process under isothermal conditions corresponding to different temperatures, and visualize the dynamic evolution of height evolution and crystal surface roughness. The dimensions of the atomic model of the growth interface can be obtained through LAMMPS software and displayed through a 3D plot in MATLAB. like Figure 6 The changes in the surface morphology of sphalerite and wurtzite are presented. Part (a) shows the absolute height evolution and crystal surface roughness of sphalerite and wurtzite. Parts (b) and (c) are snapshots at 0 and 57 ns after the removal of liquid atoms, respectively. Parts (d) and (e) are three-dimensional snapshot images plotted by Matlab at 22 and 51 ns after the removal of liquid atoms, respectively. The roughness comparison of the growth surface is clearly visible from the figures.
[0044] 5) Collect atomic data at the solid-liquid interface during the relaxation process under isothermal conditions at different temperatures, generate atomic slice diagrams to reflect the structural state of different regions in the liquid phase near the solid-liquid interface, analyze the growth rules, and visualize them. Atomic slice diagrams can be obtained using LAMMPS; like Figure 7 As shown, at a temperature of 3250K, parts (a) to (c) and (d) to (f) represent slice images at different x-coordinates in the liquid phase region near the solid-liquid interface at 0 ns and 70 ns, respectively.
[0045] 6) Calculate the change process of defect state under isothermal conditions corresponding to different temperatures. The location, size and shape of defects can be observed by removing crystal atoms through Ovito visualization software DXA. like Figure 8 As shown, (a) and (b) are the defect analysis diagrams of the front and top surfaces at 2800K and 3250K at 1ns and 70ns relaxation, respectively. The red dashed box indicates the defect size at different temperatures at the same 70ns, and the purple box indicates the porosity defect.
[0046] 7) Based on the crystallization quality of the heterojunction during the solid-liquid model crystal growth process, the types of growth defects are statistically analyzed, and the Ovito visualization software is used to visualize the state of different crystal defects, providing an observation environment; like Figure 9 As shown, (a1) to (a4) and (b1) to (b6) are the types of crystal growth defects in sphalerite and wurtzite, respectively.
[0047] 8) Statistically analyze the structure of wurtzite and zincblende atoms and their different stacking patterns at grain boundaries during isothermal relaxation at different temperatures, and provide visualization through Ovito visualization software; like Figure 10As shown, the structures with wurtzite atoms below zincblende atoms and vice versa are illustrated. (a1) to (a3) and (b1) to (b3) represent different stacking structures of zincblende and wurtzite at the interface, allowing for further observation of their evolution. Different stacking configurations are circled in yellow, and amorphous clusters are circled in red. Based on the stability differences between these two crystal types, the relationship between the competitive ability of free atoms and the stacking configuration is analyzed.
[0048] This technology proposes a dynamic analysis and image visualization scheme for silicon carbide polycrystalline nucleus orientation identification data. By employing techniques such as radial distribution function, surface roughness, and dislocation identification, it accurately extracts grain orientation information and visualizes the orientation data. This scheme can be used to optimize silicon carbide crystal orientation control and reduce defect density, simplifying the simulation process. Based on tools such as Lammps, Python, Ovito, and Origin, it effectively improves the accuracy of crystal orientation analysis, providing support for performance optimization and fabrication process improvement of silicon carbide devices. This method is applicable to both Windows and Linux systems, offering flexible computing environments and effectively saving experimental time and costs.
[0049] The above-disclosed embodiments are merely a few specific examples of the present invention. However, the present invention is not limited thereto, and any variations that can be conceived by those skilled in the art should fall within the protection scope of the present invention.
Claims
1. A dynamic evolution analysis method based on silicon carbide polycrystalline nucleus orientation identification data, characterized in that, Used for the analysis of silicon carbide heterojunction semiconductor materials, including: Create a molecular dynamics simulation model of silicon carbide heterojunction and generate the heterojunction model based on environmental parameters; Perform molecular dynamics simulations to simulate the thermodynamic behavior of interatomic interactions at different temperatures and obtain atomic behavior data during the simulation process; A polycrystalline material model containing different grain orientations is constructed. Based on the heterojunction surface and internal structure of the polycrystalline material model, the crystal orientation of the polycrystalline nuclei is identified, and the directional information of the grains is extracted. Data statistics and optimization are performed based on the orientation information of the grains to obtain and display the grain distribution pattern.
2. The dynamic evolution analysis method according to claim 1, characterized in that, The molecular dynamics simulation model for creating the silicon carbide heterojunction is implemented in molecular dynamics simulation software, including: A silicon carbide cell structure is defined, and a heterojunction model composed of two phases, 2H-SiC and 3C-SiC, is established. The silicon carbide cell structure supports the layered arrangement of 2H-SiC and 3C-SiC in a 3:2 ratio, and the size is adjusted by functional cell expansion. Configure the force field parameters.
3. The dynamic evolution analysis method according to claim 1, characterized in that, The molecular dynamics simulation is performed by setting the simulation dimension, boundary conditions, and potential function. The atomic behavior data includes: the three-dimensional coordinates, velocity, and force data of the atom.
4. The dynamic evolution analysis method according to claim 1, characterized in that, The simulated thermodynamic behavior resulting from interatomic interactions at different temperatures was performed using the NPT ensemble, including: By setting appropriate periodic boundary conditions, the behavior of materials on a large scale can be simulated. Based on the simulation requirements, an appropriate potential function is selected to describe the interatomic interactions, and the temperature change is controlled to simulate the thermodynamic behavior at different temperatures; Data on atomic behavior is acquired during the simulation process.
5. According to the dynamic evolution analysis method of claim 4, the interaction between the atoms is determined to be realized through the Tersoff potential function; The Tersoff potential function is expressed as: ， ， in, The total energy of SiC material is the sum of the contributions from the interactions between all atomic pairs. i Indicates the number of atoms, j This indicates the number of neighboring atoms that interact with each other. Indicates the first i The contribution of each atom; It is a cutoff function. Indicates the exclusion term. Represents the attraction term, a constant. Used to adjust the relative strength of repulsive and attractive terms.
6. The dynamic evolution analysis method according to claim 1, characterized in that, The construction of polycrystalline material models with different grain orientations is achieved through radial distribution function, surface roughness, and dislocation identification; The radial distribution function is used to reflect the interactions between particles and the structural characteristics of the system. Surface roughness is used to describe the shape of a surface; Dislocation identification is used to reveal the influence of factors such as temperature, strain rate, and crystal orientation on dislocation behavior.
7. The dynamic evolution analysis method according to claim 6, characterized in that, The radial distribution function is expressed as: , in, For regional density, The total volume of the crystal. It is the total number of atoms in the crystal. Let be the radius of the spherical shell. It is evenly distributed from arrive The number of atoms within the volume of the spherical shell.
8. The dynamic evolution analysis method according to claim 6, characterized in that, The method for calculating the surface roughness includes: The set of surface atomic height values calculated along a given reference direction is represented as: ,in, It is a set of surface atomic height values. The surface atomic height value along a given reference direction. The number of elements in the set of surface atomic height values; Calculate the average height of surface atoms , represented as: ; calculate and The mean squared error is expressed as: ,in, The mean squared error is denoted as .
9. The dynamic evolution analysis method according to claim 1, characterized in that... The data statistics and optimization include the following steps: A grain identification algorithm is used to extract the crystal orientation parameters of each grain and calculate the orientation difference between adjacent grains; the crystal orientation parameters include Euler angles and principal axis directions. The ratio of small-angle to large-angle grain boundaries was statistically analyzed to determine its impact on grain boundary energy and migration behavior. By combining statistical analysis of grain size distribution, grain boundary type, and orientation-dense regions, and performing image visualization, dynamic evolution analysis can be achieved.
10. The dynamic evolution analysis method according to claim 9 is characterized in that... The implementation of dynamic evolution analysis includes: By statistically analyzing atomic data during the relaxation growth process under isothermal conditions at different temperatures, the dynamic evolution process of SiC heterocrystalline structures can be analyzed and visualized. The changes in the number of atoms of a single atom type during relaxation growth under isothermal conditions at different temperatures were statistically analyzed, and the changes were visualized. The crystallization rate of SiC heterojunction during relaxation growth under different isothermal conditions was statistically analyzed, and the crystallization rate under different conditions was visualized. The surface morphology and size changes of the atomic model of the growth interface during relaxation growth under different isothermal conditions were statistically analyzed, and the dynamic evolution of height evolution and crystal surface roughness was visualized. Atomic data at the solid-liquid interface during relaxation under isothermal conditions at different temperatures were statistically analyzed, and atomic slice diagrams were generated to reflect the structural states of different regions in the liquid phase near the solid-liquid interface. Growth rules were analyzed and visualized. The process of defect state change under isothermal conditions at different temperatures is calculated, and the location, size and shape of defects are observed by removing crystal atoms through visualization software DXA. The structure of wurtzite and zincblende atoms during relaxation under isothermal conditions at different temperatures, as well as their different stacking patterns at grain boundaries, are statistically analyzed and visualized using visualization software.