An optimization design method for a defective silicon nitride ceramic
By constructing a multi-scale simulation optimization design method for β-silicon nitride ceramics, and combining first-principles calculations and molecular dynamics simulations, the coupling effect of point defects and vacancy defects was quantified, the quantitative correlation problem of thermal conductivity regulation was solved, and the controllable design and performance optimization of β-silicon nitride ceramics were realized.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2026-05-12
- Publication Date
- 2026-06-09
AI Technical Summary
In the existing technology, the thermal conductivity control method of β-silicon nitride ceramics lacks systematic research on the coupling system of point defects and vacancy defects, making it difficult to quantitatively establish the correlation between defect concentration, distribution pattern and thermal conductivity. Moreover, traditional experimental methods are time-consuming and costly, and cannot decouple the contribution of defects.
A multi-scale simulation optimization design method combining first-principles calculations and molecular dynamics simulations was adopted to construct structural models with different defect types, concentrations and distribution patterns. The influence of defects on thermal conductivity was quantified by simulating the potential function of machine learning and a quantitative correlation was established.
It enables controllable design of the thermal conductivity of β-silicon nitride ceramics, breaking through the limitations of traditional trial-and-error methods, which are characterized by long cycles and high costs. It provides a theoretical basis and design tools, and is applicable to the design of materials related to thermal insulation and thermal management.
Smart Images
Figure CN122177319A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of ceramic material preparation technology, specifically relating to an optimized design method for defective silicon nitride ceramics. Background Technology
[0002] β-Silicon nitride ceramics (β-Si3N4) have important applications in thermal management and insulation ceramic components due to their high melting point, high hardness, and high-temperature stability. Thermal conductivity, as a key performance parameter, is significantly affected by lattice vibration (phonon) scattering behavior, and lattice defects are the main means to enhance phonon scattering and control thermal conductivity.
[0003] In existing technologies, point defects are often formed by introducing elements such as Al and O to replace Si and N in the crystal lattice, or by introducing Si vacancies and N vacancies through processes to reduce thermal conductivity. However, current research and engineering practice mainly focus on the impact of a single defect type (doping only or vacancy only) on thermal conductivity, and lack systematic research on the "doped point defect-vacancy defect coupling system". The specific shortcomings are: (1) it is difficult to distinguish the contribution of point defects and vacancy defects to the decrease in thermal conductivity under uniform conditions; (2) it is impossible to quantitatively establish the correlation between defect concentration, distribution pattern and thermal conductivity; (3) the traditional experimental trial-and-error method of "preparation-testing-adjustment" is long-cycled and costly, and it is difficult to reveal the microscopic mechanism of defect coupling at the atomic scale.
[0004] Therefore, there is an urgent need to establish a multi-scale simulation optimization method that can decouple the contributions of point defects and vacancy defects and quantitatively predict thermal conductivity, so as to provide a theoretical basis and design tool for the predictable control of the thermal conductivity of β-silicon nitride ceramics. Summary of the Invention
[0005] To address the aforementioned problems, the present invention aims to provide an optimized design method for defective silicon nitride ceramics. This invention is a multi-scale simulation optimization design method based on a combination of first-principles calculations and molecular dynamics simulations, used to systematically study the influence mechanism of Al / O point defects and vacancy defects on the lattice thermal conductivity of β-silicon nitride ceramics. By constructing structural models with different defect types, concentrations, and distribution patterns, and combining the training dataset obtained from first-principles calculations with molecular dynamics simulations using machine learning potential functions, the method reveals the defect-induced phonon scattering behavior and the regulation of thermal transport properties at the atomic scale, establishing a quantitative correlation between defect parameters and thermal conductivity, thereby determining the defect parameter range that meets the target thermal conductivity regulation requirements.
[0006] Specifically, the present invention provides the following technical solution: In a first aspect, the present invention provides an optimized design method for defective silicon nitride ceramics, comprising the following steps: A β-silicon nitride supercell model was constructed, in which point defects with Al atoms substituting Si sites and O atoms substituting N sites, as well as vacancy defects with Si vacancies and / or N vacancies, were simultaneously introduced to form a point defect-vacancy defect coupling system; different defect concentrations and defect distribution patterns were set to obtain multiple sets of coupled defect structure models. First-principles density functional theory calculations are performed on the multiple sets of coupled defect structure models to obtain cell parameters, total energy, atomic force and / or stress data, forming a training dataset. Then, machine learning potential function training is performed based on the training dataset. Molecular dynamics simulations were performed on the coupled defect structure model based on the trained machine learning potential function, and the lattice thermal conductivity was calculated using a uniform non-equilibrium molecular dynamics method. Based on the calculation results of the lattice thermal conductivity, the defect coupling effect is quantified, a quantitative correlation between defect concentration and defect distribution pattern and thermal conductivity is established, and the defect parameter range that meets the target thermal conductivity control requirements is determined.
[0007] Preferably, the β-silicon nitride supercell model is a 3×3×3 supercell structure; the trained machine learning potential function is a neural network potential function (Si-Al-ON system) or a Gaussian approximation potential function for the interatomic interactions of the system.
[0008] Preferably, the first-principles density functional theory calculation uses PBE functionals, the plane wave cutoff energy is 500 eV, and the energy convergence accuracy is no higher than 1 × 10⁻⁶. -6 eV, the atomic force convergence accuracy is no higher than 0.01 eV / Å.
[0009] Preferably, the molecular dynamics simulation is performed in the temperature range of 300~1500 K, using NPT ensemble equilibrium relaxation, with a time step of 1 fs and an equilibrium simulation duration of not less than 20 ps.
[0010] Preferably, the external driving force F applied in the uniform nonequilibrium molecular dynamics method e 1×10 -5 Å -1 And under the driving force conditions, the linear response relationship between heat flow and driving force is obtained to calculate thermal conductivity.
[0011] Preferably, the concentration of the point defects is 0.5~3 at.%, and the concentration of the vacancy defects is 0.5~2 at.%.
[0012] Preferably, the concentration of point defects is 3 at.% and the concentration of vacancy defects is 1 at.%.
[0013] Preferably, the defect distribution pattern includes random distribution, clustered distribution, and uniform distribution.
[0014] Preferably, the calculation formula for the quantification defect coupling effect is: Δκ couple =κ couple (κ doped +κ vacant κ pristine ); Among them κ pristine κ represents the thermal conductivity of a defect-free system. doped For the thermal conductivity of a system containing only Al / O point defects, κ vacant For the thermal conductivity of a system containing only vacancy defects, Δκ couple The thermal conductivity is the value of the system containing both Al / O point defects and vacancy defects.
[0015] In a first aspect, the present invention provides the application of the optimized design method for defective silicon nitride ceramics described in the first aspect in the preparation of β-silicon nitride ceramic components for thermal insulation or thermal management.
[0016] One or more embodiments of the present invention have at least the following beneficial effects: (1) This invention constructs four types of comparative models (defect-free model, point defect-only model, vacancy defect-only model, and point defect-vacancy defect coupling model), and uses the variable control method to perform comparative calculations under the same simulation conditions. For the first time, it quantitatively distinguishes and quantifies the contribution of Al / O point defects and Si / N vacancy defects to the decrease in thermal conductivity, and solves the technical problem in the prior art that it is impossible to decouple and analyze the contribution of the two types of defects and to quantitatively control the thermal conductivity.
[0017] (2) This invention quantifies the coupling effect by defining the formula Δκ. couple =κ couple (κ doped +κ vacant κ pristine This study is the first to quantitatively demonstrate a significant nonlinear synergistic enhancement effect when Al / O point defects and Si / N vacancy defects coexist, with the coupling effect term Δκ... couple The thermal conductivity can be as high as 68 W / (m·K), meaning that when both defects are present simultaneously, the thermal conductivity is reduced by an additional 68 W / (m·K) compared to the linear superposition of their individual effects. This discovery provides a new theoretical basis for defect engineering design of silicon nitride ceramics with low thermal conductivity.
[0018] (3) Through systematic concentration gradient experiments, this invention establishes a quantitative correlation between doping concentration, vacancy concentration and thermal conductivity. Based on this quantitative relationship, the optimal defect parameter range can be determined in reverse according to the target thermal conductivity, realizing a fundamental shift from "trial and error" to "design".
[0019] (4) The present invention adopts a multi-scale simulation method of first principle + molecular dynamics + machine learning potential function. Compared with the traditional experimental trial and error method, it has a shorter cycle, lower cost and clearer mechanism, and can provide a reliable theoretical basis for material design. In addition, the optimization method of the present invention has strong versatility and can be extended to the defect design and performance improvement of other nitride ceramics. The process idea is clear and the theoretical support is complete, which has high engineering application value.
[0020] (5) The optimization design method of this invention breaks through the limitations of traditional trial and error methods, which rely on long experimental cycles, high costs and unclear microscopic mechanisms. It realizes the controllable design and precise optimization of the thermal conductivity of β-silicon nitride ceramics. It is applicable to the design and performance optimization of β-silicon nitride ceramic materials related to thermal insulation and thermal management. Moreover, this method has the outstanding features of clear theoretical system, precise control methods and wide applicability. It can effectively guide the integrated design of material composition-structure-performance and can be extended to the defect control and thermal performance optimization of other nitride ceramics. Attached Figure Description
[0021] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.
[0022] Figure 1 This is a schematic diagram of the β-silicon nitride supercell structure of the coupling defect model in Embodiment 1 of the present invention; Figure 2 The diagram shows the energy error convergence of the machine learning potential function training process in Embodiment 1 of the present invention, where (a) is the loss function convergence curve, (b) is the energy correlation verification diagram, (c) is the atomic force correlation verification diagram, and (d) is the stress correlation verification diagram. Figure 3 This is a molecular dynamics verification diagram of the heating process of the coupled defect system in this invention, where (a) is the temperature evolution curve with simulation time, and (b) is the pressure (P) in three directions. x P y P z (c) shows the fluctuation curves of the system's potential energy (PE) and kinetic energy (KE) over simulation time; (d) shows the lattice parameters (L) in three directions. x L y L z(e) is the evolution curve of the simulated system volume with simulation time, and (f) is the evolution curve of the crystal axis angle (α, β, γ) with simulation time. Detailed Implementation
[0023] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0024] The present invention will be further described in detail below with reference to specific embodiments. It should be noted that the specific embodiments are explanations of the present invention and not limitations thereof.
[0025] Example 1: This embodiment provides an optimized design method for silicon nitride ceramics containing defects, including the following steps: (1) Constructing four types of contrast defect supercell models A 3×3×3 β-silicon nitride supercell structure was established, and based on this supercell, the following four types of comparative models were constructed: (i) Defect-free model: Maintaining the complete β-silicon nitride lattice structure without introducing any defects, used to obtain the reference thermal conductivity κ. pristine ; (ii) Vacancy-only defect model: Only Si vacancies and N vacancies are introduced, without Al / O doping, and the total vacancy concentration is controlled at 1 at.%, with defect sites randomly distributed, to obtain the thermal conductivity κ of the vacancy-only system. vacant ; (iii) Point Defect Only Model: Only Al atoms are introduced to replace Si sites and O atoms to replace N sites, without introducing vacancy defects. The total Al / O doping concentration is controlled at 3 at.%, and the defect sites are randomly distributed. This model is used to obtain the thermal conductivity κ of the doped system. doped ; (iv) Coupling defect model (β-silicon nitride supercell structure, such as...) Figure 1 As shown): Al / O point defects (3 at.%) and Si / N vacancy defects (1 at.%) are introduced simultaneously to obtain the thermal conductivity κ of the coupled system. couple Based on the defect distribution pattern, the coupled defect model is further divided into three subclasses: random distribution, clustered distribution, and uniform distribution; For each type of model, three initial computational models with different random seeds are constructed to eliminate the random errors caused by random distribution.
[0026] (2) First-principles DFT calculation and training dataset construction First-principles density functional theory (DFT) calculations were performed using VASP software, with the following parameter settings: The PBE functional (Perdew-Burke-Ernzerhof functional) was selected, the plane wave cutoff energy was set to 500 eV, and the energy convergence accuracy was 1×10⁻⁶. -6 eV, the atomic force convergence accuracy is 0.01 eV / Å; perform structural relaxation and single-point energy calculation on the various defect cells constructed in step (1) to obtain system energy, atomic force and stress data, and form a high-precision training dataset.
[0027] (3) Machine learning potential function training Import the DFT training dataset obtained in step (2) into the GPUMD simulation platform and train the potential function using the Neuroevolution Potential (NEP) method. The training parameters are set as follows: the maximum number of training epochs is 10. 5 The population size was 50, the energy weight was 1.0, the force weight was 1.0, and the virial weight was 0.1. During training, the root mean square error of energy (RMSE) and the root mean square error of force were used as convergence criteria. When the energy RMSE was lower than 5 meV / atom and the force RMSE was lower than 150 meV / Å, the training was considered to have converged.
[0028] The actual accuracy after training is as follows Figure 2 As shown: the continuous decrease of various loss functions during training indicates that the model is converging well (e.g. Figure 2 As shown in (a), the root mean square error of energy prediction is 3.50 meV / atom (as shown in (a)). Figure 2 As shown in (b), the root mean square error of the atomic force prediction is 132.28 meV / Å (as shown in [reference]). Figure 2 As shown in (c), the root mean square error of stress prediction is 0.9297 GPa (as shown in the figure). Figure 2 As shown in (d), the NEP predictions and DFT calculations of all physical quantities exhibit a high degree of linear correlation. For Si-Al-ON, a complex quaternary system containing a mixture of covalent and ionic bonds, this fully meets the accuracy requirements for subsequent molecular dynamics simulations and thermal conductivity calculations. Therefore, by utilizing a trained machine learning potential function, it is possible to describe the interatomic interactions of the Si-Al-ON quaternary system with accuracy close to that of the DFT.
[0029] (4) Molecular dynamics simulation settings Based on the machine learning potential function trained in step (3), molecular dynamics simulations are performed on the GPUMD platform.
[0030] Five temperature points were set: 300 K, 600 K, 900 K, 1200 K, and 1500 K, to simulate the actual operating conditions of the material in high-temperature service environments. Specifically, an NPT isothermal and isobaric ensemble was used, with target temperatures at the five temperature points mentioned above, a target pressure of 0 GPa, a time step of 1 fs, an equilibrium simulation duration (relaxation time) of 20 ps, a Nose-Hoover temperature control method, a temperature coupling time constant of 100 fs, and a pressure coupling time constant of 1000 fs. The stable configuration and heat flux data of the system at different temperatures were obtained for structural stability verification in subsequent thermal conductivity calculations.
[0031] In this invention, all embodiments are verified through molecular dynamics heating processes, such as... Figure 3 As shown: Figure 3 (a) shows the complete process of the system temperature continuously increasing from 300 K to 1750 K. Figure 3 (b) shows the pressure fluctuations in three directions during the heating process. Figure 3 (c) shows the evolution trend of the system's potential energy and kinetic energy as temperature increases. Figure 3 (d) and Figure 3 (f) shows that the lattice parameters and the crystal axis angle remain stable throughout the entire heating range. Figure 3 Figure (e) shows the thermal expansion of the system volume as the temperature increases. The above verification results show that the crystal structure remains intact throughout the temperature range, ensuring the structural reliability of subsequent thermal conductivity calculations at different temperature points.
[0032] (5) Calculation of thermal conductivity using homogeneous nonequilibrium molecular dynamics (HNEMD) The lattice thermal conductivity was calculated using the Uniform Nonequilibrium Molecular Dynamics (HNEMD) module built into the GPUMD software package. The specific steps are as follows: Based on the equilibrium configuration after NPT relaxation in step (4) (the system reaches a thermodynamically stable state), a uniform external driving force F is applied along the heat transport direction. e =1×10 -5 Å -1 A Nose-Hoover chain thermostat was used to maintain the system temperature at the target temperatures. The formal simulation duration was set to 1 ns, with a time step of 1 fs.
[0033] According to HNEMD theory, the lattice thermal conductivity tensor κ is calculated using the following formula: J ne =T×V×κ×F e ; Among them, J ne Non-equilibrium heat flux vector (unit: W / m) 2T is the thermodynamic temperature of the system (unit: K), and V is the volume of the simulated system (unit: m³). 3 ), F e The applied external driving force vector (unit: Å) -1 ), where κ is the lattice thermal conductivity tensor to be determined (unit: W / (m·K)).
[0034] In actual calculations, to reduce statistical errors, the κ value was calculated and averaged for the three sets of models constructed with different random seeds.
[0035] (6) Calculation results Based on the 3 at.% Al / O doped + 1 at.% vacancy + randomly distributed β-Si3N4 coupled defect system constructed in this embodiment, HNEMD thermal conductivity was calculated at different temperature points. The results show that the thermal conductivity of the system generally decreases with increasing temperature. The specific calculation results are shown in Table 1. Table 1
[0036] Example 2: The only difference between this embodiment and Embodiment 1 is that the spatial distribution pattern of the defect sites in the coupled defect model is an aggregated distribution rather than a random distribution; the other steps are the same as in Embodiment 1.
[0037] Based on the β-Si3N4 coupled defect system with 3 at.% Al / O doping + 1 at.% vacancy + aggregate distribution constructed in this embodiment, HNEMD thermal conductivity was calculated at different temperature points. The results show that the thermal conductivity of the system also decreases as the temperature increases, and the thermal conductivity at each temperature point is lower than that of the randomly distributed system with the same defect concentration (Example 1). The specific calculation results are shown in Table 2: Table 2
[0038] Example 3: The difference between this embodiment and Embodiment 1 is that the spatial distribution pattern of the defect sites in the coupled defect model is uniform, while the remaining steps are the same as in Embodiment 1.
[0039] Based on the β-Si3N4 coupled defect system with 3 at.% Al / O doping + 1 at.% vacancy + spatial distribution constructed in this embodiment, HNEMD thermal conductivity calculations were performed at different temperature points. The results show that the thermal conductivity of the system generally decreases with increasing temperature, and the thermal conductivity at each temperature point is higher than that of the randomly distributed system (Example 1) and the aggregated distributed system (Example 2) with the same defect concentration. The specific calculation results are shown in Table 3. Table 3
[0040] Example 4: The difference between this embodiment and Embodiment 1 is that the Al / O atom substitution concentrations are set to 1 at.%, 2 at.%, and 3 at.%, while the Si / N vacancy concentration is fixed at 1 at.%, and the remaining steps are the same as in Embodiment 1.
[0041] Based on the Al / O atom doping system constructed in this embodiment, HNEMD thermal conductivity calculations were performed at different temperature points. The results show that, under the condition of a fixed Si / N vacancy concentration of 1 at.%, the thermal conductivity of the system monotonically decreases with increasing Al / O doping concentration. This is because Al / O doping introduces additional mass perturbation and force constant perturbation, which enhances phonon scattering, leading to a further decrease in thermal conductivity. Simultaneously, the thermal conductivity monotonically decreases with increasing temperature due to the shortening of the mean free path caused by enhanced phonon-phonon scattering; specific calculation results are shown in Table 4. Table 4
[0042] Example 5: The difference between this embodiment and Embodiment 1 is that the Al / O doping concentration is fixed at 3 at.%, and the Si / N vacancy concentrations are set to 0.5 at.%, 1.0 at.%, 1.5 at.%, and 2.0 at.%, respectively. The remaining steps are the same as in Embodiment 1.
[0043] Based on the different vacancy doping concentration systems constructed in this embodiment, HNEMD thermal conductivity calculations were performed at different temperature points. The results show that the thermal conductivity of the system decreases monotonically with increasing Si / N vacancy concentration. This is because vacancy defects disrupt the lattice periodicity and introduce stronger phonon scattering centers (mass deficiency + force constant deficiency), which have a more significant effect on reducing thermal conductivity compared to point defects. Specific calculation results are shown in Table 5. Table 5
[0044] Comparative Example 1: The difference between this comparative example and Example 1 is that Al / O point defects are not introduced; only Si vacancies and N vacancies are constructed, with a total vacancy concentration of 1 at.% and random distribution of defect sites. The remaining steps are the same as in Example 1.
[0045] Based on the 1 at.% vacancy + randomly distributed β-Si3N4 coupled defect system constructed in this embodiment, HNEMD thermal conductivity calculations were performed at different temperature points. The results show that the thermal conductivity of the β-Si3N4 system containing only 1 at.% vacancy defects is significantly lower than that of the defect-free system (Comparative Example 3), but significantly higher than that of the coupled defect system (Example 1). This indicates that vacancy defects can effectively reduce thermal conductivity, but the scattering intensity of a single vacancy defect is limited, and its thermal conductivity modulation effect is weaker than that of the point defect-vacancy coupled system. The specific calculation results are shown in Table 6. Table 6
[0046] Comparative Example 2: The difference between this comparative example and Example 1 is that no vacancy defects are introduced; only an Al / O point defect system is constructed. The total Al / O doping concentration remains 3 at.%, and the defect sites are randomly distributed. All other steps are the same as in Example 1.
[0047] Based on the 3 at.% Al / O doped β-Si3N4 coupled defect system constructed in this example, HNEMD thermal conductivity calculations were performed at different temperature points. The results show that the thermal conductivity of the β-Si3N4 system containing only 3 at.% Al / O point defects is significantly lower than that of the defect-free system (Comparative Example 3) and the vacancy-only system (Comparative Example 1), but significantly higher than that of the coupled defect system (Example 1). This indicates that Al / O point defects have a more significant effect on reducing thermal conductivity than vacancy defects, but the control effect of a single defect is still weaker than that of the point defect-vacancy coupled system. The specific calculation results are shown in Table 7. Table 7
[0048] Comparative Example 3: The difference between this comparative example and Example 1 is that no defects are introduced, the complete β-silicon nitride lattice structure is maintained, and the remaining steps are the same as in Example 1.
[0049] Based on the β-Si3N4 system constructed in this embodiment, HNEMD thermal conductivity calculations were performed at different temperature points. The results show that the thermal conductivity of the defect-free β-Si3N4 system at 300 K is significantly higher than that of all defective systems (the thermal conductivity of the defect-free system decreases significantly with increasing temperature, which is consistent with the phonon scattering theory). The specific calculation results are shown in Table 8. Table 8
[0050] Experimental Example: Quantitative Verification of Defect Coupling Effect To further verify the accuracy of the optimization design method proposed in this invention in quantifying the coupling effect of defects, this experimental example, based on the calculation results of Example 1 and Comparative Examples 1-3, quantitatively calculates and analyzes the coupling effect of Al / O point defects and vacancy defects.
[0051] Using the thermal conductivity calculation results at 300 K as the baseline data, the non-additive coupling term Δκ is calculated according to the coupling effect formula defined in this invention. couple : Δκ couple =κ couple (κ doped +κ vacant κ pristine ); Substitute the data, Δκ couple =28 (45+85 170) = 28 ( 40) = 68 W / (m·K); Calculation results show that, considering only the linear superposition effect of a single defect, the thermal conductivity should be 40% lower than that of the defect-free system. That is, if Al / O point defects and vacancy defects act independently without coupling effects, the thermal conductivity when both coexist should be κ. doped +κ vacant κ pristine =45+85 170 = The negative value of 40 W / (m·K) indicates a 40 W / (m·K) decrease in thermal conductivity relative to the defect-free system. However, the actual thermal conductivity of the coupled system is 28 W / (m·K), far lower than the linear superposition prediction mentioned above. At the same time, Δκ couple =68>0, this positive value indicates that there is a significant nonlinear synergistic effect between Al / O point defects and vacancy defects. That is, when the two defects coexist, the resulting phonon scattering intensity is not simply added together, but produces a coupling enhancement effect of "1+1>2", which makes the thermal conductivity decrease by an additional 68 W / (m·K).
[0052] Furthermore, based on the concentration gradient experimental results of Examples 4 and 5, this invention establishes a quantitative correlation between doping concentration, vacancy concentration, and thermal conductivity. According to the target thermal conductivity control requirements, an optimal range of defect parameters can be determined.
[0053] If the target thermal conductivity is 25~35 W / (m·K) (suitable for medium thermal insulation requirements), then the preferred Al / O doping concentration is 2~3 at.%, the Si / N vacancy concentration is 0.5~1.0 at.%, and the defect distribution pattern is random or uniform. If the target thermal conductivity is less than 20 W / (m·K) (suitable for strong thermal insulation requirements), the preferred Al / O doping concentration is 3 at.%, the Si / N vacancy concentration is 1.5~2.0 at.%, and the defect distribution pattern is aggregated distribution.
[0054] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. An optimized design method for defective silicon nitride ceramics, characterized in that, Includes the following steps: A β-silicon nitride supercell model was constructed, in which point defects with Al atoms substituting Si sites and O atoms substituting N sites, as well as vacancy defects with Si vacancies and / or N vacancies, were simultaneously introduced to form a point defect-vacancy defect coupling system; different defect concentrations and defect distribution patterns were set to obtain multiple sets of coupled defect structure models. First-principles density functional theory calculations are performed on the multiple sets of coupled defect structure models to obtain cell parameters, total energy, atomic force and / or stress data, forming a training dataset. Then, machine learning potential function training is performed based on the training dataset. Molecular dynamics simulations were performed on the coupled defect structure model based on the trained machine learning potential function, and the lattice thermal conductivity was calculated using a uniform non-equilibrium molecular dynamics method. Based on the calculation results of the lattice thermal conductivity, the defect coupling effect is quantified, a quantitative correlation between defect concentration and defect distribution pattern and thermal conductivity is established, and the defect parameter range that meets the target thermal conductivity control requirements is determined.
2. The method according to claim 1, characterized in that, The β-silicon nitride supercell model is a 3×3×3 supercell structure; the trained machine learning potential function is a neural network potential function or a Gaussian approximation potential function for the interatomic interactions of the system.
3. The method according to claim 1, characterized in that, The first-principles density functional theory calculations used PBE functionals, with a plane wave cutoff energy of 500 eV and an energy convergence accuracy not exceeding 1 × 10⁻⁶ eV. -6 eV, the atomic force convergence accuracy is no higher than 0.01 eV / Å.
4. The method according to claim 1, characterized in that, The molecular dynamics simulations were conducted in the temperature range of 300–1500 K, using NPT ensemble equilibrium relaxation with a time step of 1 fs and an equilibrium simulation duration of no less than 20 ps.
5. The method according to claim 1, characterized in that, The external driving force F applied in the homogeneous nonequilibrium molecular dynamics method e 1×10 -5 Å -1 And under the driving force conditions, the linear response relationship between heat flow and driving force is obtained to calculate thermal conductivity.
6. The method according to claim 1, characterized in that, The concentration of point defects is 0.5~3 at.%, and the concentration of vacancy defects is 0.5~2 at.%.
7. The method according to claim 6, characterized in that, The concentration of point defects is 3 at.%, and the concentration of vacancy defects is 1 at.%.
8. The method according to claim 1, characterized in that, The defect distribution patterns include random distribution, clustered distribution, and uniform distribution.
9. The method according to claim 1, characterized in that, The calculation formula for the quantification defect coupling effect is as follows: Dk couple =k couple (k doped +k vacant k pristine ); Among them κ pristine κ represents the thermal conductivity of a defect-free system. doped For the thermal conductivity of a system containing only Al / O point defects, κ vacant For the thermal conductivity of a system containing only vacancy defects, Δκ couple The thermal conductivity is the value of the system containing both Al / O point defects and vacancy defects.
10. The application of the optimized design method of defective silicon nitride ceramics according to any one of claims 1 to 9 in the preparation of β-silicon nitride ceramic components for thermal insulation or thermal management.