An intelligent control method for electric energy barrier control of flexible electronic devices

Abstract

The present invention discloses an intelligent control method for electric energy barrier control of flexible electronic devices, including precise control of device barrier parameters under radiation by establishing a mathematical model and finite element simulation, combined with a machine learning algorithm and first principles calculation. The method of the present invention includes establishing a barrier control model, analyzing test data, studying radiation-induced defects, online detecting internal parameters, screening radiation-resistant materials and optimizing device structure, and establishing a device failure mechanism model. The present invention effectively improves the long-term reliability of the device in an irradiated environment, solves the defects and performance degradation problems introduced by radiation, and provides a new technical path for the development of flexible electronic devices.

Description

Technical Field

[0001] The present invention relates to the technical field of radio energy conversion, and in particular to an intelligent control method for electric energy barrier control of a flexible electronic device. Background Art

[0002] Precise regulation of the potential barrier of flexible electronic devices has always been a challenging technical problem. Although the characteristics of the potential barrier can be modified by radiation irradiation, there are still many contradictions and limitations in practical applications. First, the radiation irradiation process introduces additional defects or unstable factors, resulting in reduced device reliability. Second, radiation irradiation has high requirements on the selectivity and positioning accuracy of the device, and requires precise control of radiation energy, dose and incident angle. Otherwise, it may cause device performance degradation outside the barrier area. It is urgent to achieve precise adjustment and optimization of the barrier parameters while ensuring device performance and stability.

[0003] The present invention proposes a solution to the above-mentioned shortcomings:

[0004] By establishing a mathematical model and conducting refined simulations, combining machine learning algorithms to optimize barrier parameters, and establishing a correlation model between radiation irradiation and device performance degradation, first-principles calculations and molecular dynamics simulations are used to study the impact of radiation on defect states, and real-time monitoring and regulation are achieved. Further screening of radiation-resistant materials and optimization of device structures are carried out, and a multi-physical field coupling model is established through orthogonal experiments and big data analysis, thereby effectively improving the long-term reliability of devices in irradiated environments, reducing defects, optimizing performance, and providing a scientific basis for device design and manufacturing. Summary of the invention

[0005] In order to solve the above-mentioned technical problems, the present invention provides an intelligent control method for electric energy barrier control of flexible electronic devices.

[0006] The technical solution of the present invention is implemented as follows: a method for intelligently controlling electric energy barrier control of a flexible electronic device comprises the following steps:

[0007] S1. According to the material properties and device structure of the flexible electronic device, a mathematical model of barrier regulation is established, and the barrier region is simulated in detail using the finite element analysis method to obtain a quantitative relationship model between the device structure parameters and the device performance. The device structure parameters include barrier height and barrier width. The quantitative relationship model is trained using a vector machine algorithm to obtain the optimized barrier height and barrier width.

[0008] S2. Using mathematical models and simulation results, the optimized barrier height and barrier width obtained in step S1 are mined and analyzed through machine learning algorithms, and a correlation model is established with inputs of radiation energy, dose, and incident angle irradiation parameters and outputs of quantitative indicators of device performance degradation, so as to evaluate the impact of radiation irradiation on the quantitative indicators of device performance degradation;

[0009] S3. Based on the correlation model of step S2, the first principle calculation and molecular dynamics simulation methods are used to obtain the position and concentration distribution of the defect state in the bandgap and the influence on the height and width of the barrier in view of the changes in the material structure and chemical composition caused by the radiation.

[0010] S4. According to the calculation and simulation in step S3, during the radiation irradiation process, the internal barrier parameters of the device are detected online by means of capacitance-voltage measurement and deep level transient spectrum detection, which are used for the control of the radiation irradiation process and the real-time feedback adjustment of the barrier parameters;

[0011] S5. According to the calculation and simulation of step S3, a barrier control material system is used, barrier layer materials are screened by first-principles calculation, and the composition and thickness of the barrier layer are controlled by atomic layer deposition and molecular beam epitaxy technology to obtain a screened device structure, where the screened device structure is used to suppress the formation of radiation-induced defects;

[0012] S6. Test the screened device structure, regulate the technical parameters of irradiation dose, irradiation time and annealing temperature through orthogonal test, and establish a mapping relationship between barrier characteristics and regulation parameters for the regulation of barrier parameters and coordinated optimization of device performance;

[0013] S7. According to the calculation and simulation of step S3, a device failure mechanism model is established. By establishing a multi-physical field coupling model, the influence weights of different stress factors on device performance are analyzed. Accelerated life testing and big data analysis methods are used to record the performance degradation data of the device, including leakage current and breakdown voltage.

[0014] Furthermore, in step S1, a mathematical model of barrier control based on the material properties and structural parameters of the electronic device is established, and the value range of the barrier height and the barrier width is set; a finite element analysis method is used to perform a refined simulation of the grid division of the barrier area to obtain the current-voltage characteristic curve, threshold voltage and switching ratio performance parameters under different barrier parameters, wherein the grid division accuracy is 0.1 nm; a quantitative relationship model between barrier parameters and device performance is established using a polynomial fitting method, and training and optimization are performed through a support vector machine algorithm to obtain the optimal barrier height and barrier width.

[0015] Furthermore, in step S2, test data of the radiation irradiated device is collected, the test data including performance parameters of current, voltage, response time and radiation energy, dose, and incident angle radiation parameters; data preprocessing is performed on the test data, the data preprocessing includes data cleaning to remove outliers, fill in missing values ​​and normalize; key characteristic parameters in the test data are extracted, the key characteristic parameters include average response time and maximum current change rate, and a correlation model between device performance degradation and radiation irradiation parameters is established using a support vector machine (SVM) algorithm, and the correlation model is evaluated through cross-validation to avoid overfitting.

[0016] Furthermore, in step S3, semiconductor materials including silicon or germanium are selected through the first principles calculation method to simulate the influence of radiation irradiation on their structure and chemical composition, and the density functional theory DFT calculation is used to determine the defect state type generated by radiation irradiation, as well as the position and concentration distribution of the defect state in the band gap; the formation mechanism and evolution behavior of radiation-induced defects are simulated by molecular dynamics to form an irradiation-induced defect correlation model, and the influence of defect state density and energy level distribution on electronic state density distribution is analyzed, and the calculation results reflect that the increase in defect state density leads to the degradation of material barrier characteristics and conductivity characteristics; based on the radiation-induced defect correlation model, the change of material barrier characteristics when the defect state density exceeds the eigenvalue is predicted, and the chemical composition, including the doping concentration, is adjusted through molecular dynamics simulation to regulate the defect state density distribution, and the optimized parameters are input into the model.

[0017] Furthermore, in step S4, according to preset irradiation parameters, an X-ray source is used to irradiate the semiconductor device, and electrical characteristic data is collected in real time. The data is transmitted to the data processing unit, and the built-in capacitance-voltage measurement algorithm and deep energy level transient spectrum detection algorithm respectively calculate the internal barrier parameters and defect state density of the device, and the calculation results are compared with the standard parameter range to judge the irradiation intensity; the decision tree algorithm is used to analyze the electrical characteristics and barrier parameter data, identify the key factors affecting the device performance, and construct a data-driven statistical evaluation model to predict the performance change trend, and the prediction results of the performance change trend of the statistical evaluation model are fed back to the control unit and the data processing unit, the ray source control unit adjusts the parameters according to the prediction results, and performs the next round of irradiation and detection. By repeating the optimization process, the effectiveness of the measures is verified, and the evaluation results are compared with the performance lower limit threshold to determine the triggering of the sound and light alarm.

[0018] Furthermore, in step S5, the target performance indicators of the barrier layer material are determined according to the device performance requirements, including the ability to resist radiation damage, and the candidate materials are simulated and the performance is predicted by first-principles calculation software to select the optimal material system; the functional layer material is epitaxially grown on the surface of the optimal material system using molecular beam epitaxy technology, and the barrier layer material is precisely grown at a specified position using atomic layer deposition technology, and a radiation damage test is performed, and the device performance is optimized according to the test results. The first-principles calculation is used to simulate the performance of the optimized device structure.

[0019] Furthermore, in step S6, the material, dimensional parameters and electrode arrangement parameters of the screened device structure are obtained, wherein the dimensional parameters include layer thickness and electrode spacing; the barrier characteristics of the device are regulated by using ray irradiation technology, the electronic structure and defect state density are changed by controlling the irradiation dose and time, and multiple groups of control parameter combinations are generated by using an orthogonal experimental method; for each group of parameter combinations, the device performance indicators are tested, and data including current density and photoelectric conversion efficiency are recorded, and a mapping relationship model between barrier characteristics and control parameters is established using a support vector regression SVR algorithm. The mapping relationship model is used to capture the nonlinear relationship between parameters and performance, and the target performance is input. The SVR model solves the optimal barrier control parameter combination through an optimization algorithm.

[0020] Furthermore, in step S7, multiple stress factor data of radiation irradiation, electrical stress and thermal stress are collected to provide data input for device failure mechanism analysis, and a multi-physical field coupling model is established through a finite element analysis method to determine the maximum stress concentration area; the principal component analysis method is used to determine the weight of each stress factor on device performance, and an orthogonal test is performed after sorting, and a support vector machine regression algorithm is used to establish a mathematical model between performance degradation and key stress factors.

[0021] The invention discloses an intelligent control method for controlling the electric energy barrier of a flexible electronic device, which has the advantages that the invention realizes precise control of the electric energy barrier of the flexible electronic device through the intelligent control method, effectively improves the long-term reliability of the flexible device in the irradiation environment, and has the following specific beneficial effects: through refined simulation and machine learning algorithms, a quantitative relationship between the barrier parameters and the device performance is established, and the precise control of the radiation irradiation parameters is realized, and the defects and unstable factors caused by the irradiation are reduced; first-principles calculation and molecular dynamics simulation are used to deeply study the influence of radiation-induced defects on the barrier characteristics, and provide a theoretical basis for inhibiting the formation of defects; real-time monitoring and intelligent control of the internal barrier parameters of the device during the irradiation process are realized, and the stability and reliability of the device performance are improved; by screening new radiation-resistant materials and optimizing the device structure, the device performance and reliability are coordinated to be optimized, and the adaptability of the device in complex environments is enhanced; orthogonal experiments and big data analysis methods are used to establish a multi-physical field coupling model, analyze the influence of different stress factors on the device performance, and provide a scientific basis for device design and manufacturing; on the whole, the invention provides new impetus for the development of flexible electronic devices, promotes the rapid development of high-performance electronic devices, and has important technical application value and market prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a structural block diagram of an intelligent control method for electric energy barrier control of a flexible electronic device in an embodiment of the present invention;

[0023] Figure 2 This is a flowchart of the steps of an intelligent control method for electric energy barrier control of a flexible electronic device in an embodiment of the present invention;

[0024] Figure 3 This is a structural block diagram of step S7 of a method for intelligently controlling electric energy barrier of a flexible electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0026] The preferred implementation methods of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these implementation methods are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0027] See also Figure 1-Figure 3 As shown, a method for intelligently controlling electric energy barrier control of a flexible electronic device comprises:

[0028] S1. According to the material properties and device structure of the flexible electronic device, a mathematical model of barrier regulation is established, and the barrier region is simulated in detail using the finite element analysis method to obtain a quantitative relationship model between the device structure parameters and the device performance. The device structure parameters include barrier height and barrier width. The quantitative relationship model is trained using a vector machine algorithm to obtain the optimized barrier height and barrier width.

[0029] S2. Using mathematical models and simulation results, the optimized barrier height and barrier width obtained in step S1 are mined and analyzed through machine learning algorithms, and a correlation model is established with inputs of radiation energy, dose, and incident angle irradiation parameters and outputs of quantitative indicators of device performance degradation, so as to evaluate the impact of radiation irradiation on the quantitative indicators of device performance degradation;

[0030] S3. Based on the correlation model of step S2, the first principle calculation and molecular dynamics simulation methods are used to obtain the position and concentration distribution of the defect state in the bandgap and the influence on the height and width of the barrier in view of the changes in the material structure and chemical composition caused by the radiation.

[0031] S4. According to the calculation and simulation in step S3, during the radiation irradiation process, the internal barrier parameters of the device are detected online by means of capacitance-voltage measurement and deep level transient spectrum detection, which are used for the control of the radiation irradiation process and the real-time feedback adjustment of the barrier parameters;

[0032] S5. According to the calculation and simulation of step S3, a barrier control material system is used, barrier layer materials are screened by first-principles calculation, and the composition and thickness of the barrier layer are controlled by atomic layer deposition and molecular beam epitaxy technology to obtain a screened device structure, where the screened device structure is used to suppress the formation of radiation-induced defects;

[0033] S6. Test the screened device structure, regulate the technical parameters of irradiation dose, irradiation time and annealing temperature through orthogonal test, and establish a mapping relationship between barrier characteristics and regulation parameters for the regulation of barrier parameters and coordinated optimization of device performance;

[0034] S7. According to the calculation and simulation of step S3, a device failure mechanism model is established. By establishing a multi-physical field coupling model, the influence weights of different stress factors on device performance are analyzed. Accelerated life testing and big data analysis methods are used to record the performance degradation data of the device, including leakage current and breakdown voltage.

[0035] Furthermore, in step S1, a mathematical model of barrier regulation is first established according to the material properties and device structure parameters of the flexible electronic device;

[0036] The potential barrier height ranges from 0.1eV to 1.0eV, and the potential barrier width ranges from 1nm to 10nm. Then, the finite element analysis method is used to perform a refined simulation of the potential barrier region, and the mesh size is set to 0.1nm. Through the simulation, the current-voltage characteristic curve, threshold voltage, and switching ratio performance parameters of the device under different potential barrier heights and barrier widths are obtained.

[0037] According to the device performance parameters obtained by simulation, a quantitative relationship model between barrier height, barrier width and device performance is established by using polynomial fitting method;

[0038] Then, the quantitative relationship model was trained and optimized using the support vector machine algorithm. The training data set contained 100 sets of device performance data under different barrier parameters. After optimization, the device performance was optimal when the barrier height was 0.6 eV and the barrier width was 5 nm.

[0039] According to the optimized barrier parameters, the mathematical model is modified to obtain a modified barrier control mathematical model; the modified mathematical model is applied to the design and manufacturing process of electronic devices, and the switching ratio of the device is increased by 20% and the threshold voltage is reduced by 0.2V by adjusting the barrier height and barrier width;

[0040] Finally, a genetic algorithm was used to optimize the doping concentration, gate dielectric layer thickness material properties and device structure parameters of the electronic device with the on / off ratio and threshold voltage of the device as the optimization targets. After optimization, the doping concentration of 1×10 18 cm -3 , the device performance is optimal when the gate dielectric layer thickness is 3nm; the optimized material properties and device structure parameters are substituted into the modified barrier control mathematical model, and the optimization results are verified by simulation. The device switching ratio is increased to 1000, and the threshold voltage is reduced to 0.3V, which meets the device performance requirements;

[0041] By comprehensively optimizing material properties, device structural parameters and barrier control mathematical models, we can obtain the electronic device design with the best performance.

[0042] Specifically include:

[0043] Establish a mathematical model for potential barrier regulation: Assume that the potential barrier height of the electronic device is V 0 , the potential barrier width is W, then the mathematical model can be expressed as: M(V 0 ,W) = f(V 0 ,W), where f is a function that represents the relationship between barrier height and width and device performance;

[0044] The value range of the potential barrier parameter: the potential barrier height V 0The value range of is 0.1eV≤V0≤1.0eV, and the value range of barrier width W is 1nm≤W≤10nm; finite element analysis:

[0045] The barrier region is simulated using the finite element method. The mesh size is set to d = 0.1 nm. For each barrier parameter combination (V 0 , W), and simulate to obtain performance parameters, such as current-voltage characteristic curve I(V), threshold voltage V th , switching ratio R; polynomial fitting:

[0046] It is assumed that the data obtained through simulation can be expressed as: , where a, b, c, d are polynomial coefficients obtained by least squares fitting;

[0047] Support vector machine algorithm: Use the support vector machine (SVM) algorithm to train and optimize the quantitative relationship model. Suppose the training data set is ;

[0048] The most advantageous barrier parameters are obtained by optimization ;

[0049] Modify the mathematical model: according to the optimal parameters Modify the mathematical model to obtain the modified model ;

[0050] Adjust the potential barrier parameters: According to the modified model, adjust V 0 and W, which increases the switching ratio by 20% and reduces the threshold voltage by 0.2V;

[0051] Genetic algorithm optimization: Assume the doping concentration of the device is N d , the gate dielectric layer thickness is T, and the genetic algorithm is used to calculate the switching ratio R and the threshold voltage V th Optimize for the target and get the optimal parameters ;

[0052] Verify the optimization results: Substituting the modified mathematical model M', the simulation results show that the switching ratio R = 1000 and the threshold voltage V th = 0.3V; get the optimal design solution:

[0053] Combining all optimized parameters, we can get the electronic device design with the best performance;

[0054] In the above calculation, the characters involved and their meanings are as follows:

[0055] V 0 : Barrier height, which indicates the height of the potential energy barrier that prevents electrons from passing through the flexible electronic device, in electron volts (eV);

[0056] W: Barrier width, which indicates the width of the potential barrier of flexible electronic devices, in nanometers (nm);

[0057] M(V 0 ,W): The mathematical model of barrier regulation, which is a function that describes the relationship between barrier height and width and device performance;

[0058] d: mesh size, used for meshing in finite element analysis, in nanometers (nm);

[0059] I(V): Current-voltage characteristic curve, indicating the current value of the device at different voltages;

[0060] V th : Threshold voltage, which refers to the minimum voltage required for the device to turn on;

[0061] R: On / Off ratio, which indicates the current ratio between the on state and the off state of the device, and is an important parameter for measuring device performance;

[0062] a, b, c, d: coefficients in the polynomial fit, used to express the quantitative relationship between barrier height, width and device performance;

[0063] D: training data set, including device performance data under different barrier parameters;

[0064] : The most advantageous barrier parameters obtained by optimizing the support vector machine algorithm;

[0065] : The revised barrier regulation mathematical model was adjusted according to the optimized parameters;

[0066] N d : Doping concentration, which indicates the concentration of doping atoms in flexible electronic devices, in units of atoms per cubic centimeter (cm -3 );

[0067] T: gate dielectric layer thickness, indicating the physical thickness of the gate dielectric layer in the device, in nanometers (nm);

[0068] : The optimal doping concentration and gate dielectric layer thickness obtained by genetic algorithm optimization.

[0069] Furthermore, in step S2, obtaining test data of the radiation-irradiated device is the basis for studying the performance change of the device under the radiation environment. For example, the preset research object is a semiconductor device for flexible electronics, and the test data includes the device's current, voltage, response time performance parameters, and radiation energy (such as 1MeV, 10MeV), irradiation dose (such as 1krad, 10krad), incident angle (such as 0°, 45°) radiation irradiation parameters;

[0070] Preprocessing the acquired test data is a key step to ensure data quality and model accuracy;

[0071] Data cleaning includes removing outliers and filling missing values, including data anomalies caused by equipment failure in a certain test, which need to be identified and eliminated;

[0072] Normalization is to unify data of different dimensions to the same scale for the convenience of subsequent analysis, including normalizing both current value and radiation dose to between 0 and 1;

[0073] The key characteristic parameters extracted include the average response time and maximum current change rate of the device;

[0074] Using the extracted feature parameters, the correlation model between device performance degradation and radiation irradiation parameters is established through the support vector machine (SVM) algorithm. SVM is a powerful machine learning algorithm suitable for classification and regression problems of small samples and high-dimensional data.

[0075] The model input is the radiation irradiation parameters, including energy, dose and angle, and the output is the quantitative index of device performance degradation, including the performance degradation percentage. The specific input parameters are 10MeV energy, 5krad dose, 30° angle, and the model output is 15% performance degradation.

[0076] Using the cross-validation method to evaluate the established association model can effectively avoid overfitting and improve the generalization ability of the model. For example, the data set is divided into five parts, and a five-fold cross-validation is performed. Four parts of the data are used to train the model each time, and one part of the data is used to test the model. Finally, the average performance of the five tests is taken as the model evaluation result.

[0077] The optimized correlation model is used to evaluate the reliability level of the device under different radiation irradiation parameters.

[0078] The specific operation is to input different radiation irradiation parameters into the model to obtain the quantitative index of device performance degradation, and then judge the reliability level of the device. For example, different combinations of radiation energy, dose and angle are input, and the model outputs the corresponding performance degradation index. If the degradation index exceeds the preset threshold (such as 20%), the device is considered unreliable under the irradiation condition.

[0079] Based on the assessed reliability level of the device, determine the impact of radiation exposure on the long-term reliability of the device;

[0080] If radiation exposure causes serious degradation of device performance and reduces its reliability to an unacceptable level, then radiation exposure is considered to have a significant impact on the long-term reliability of the device. For example, through model prediction, it is found that under the combination of high-energy radiation (such as 10MeV) and high dose (such as 10krad), the device performance degradation exceeds 30%, which seriously affects its long-term reliability.

[0081] In view of the significant impact of radiation on device reliability, device reinforcement measures are proposed, including shielding and redundant design;

[0082] Shielding measures include adding a physical shielding layer, including using lead plates or other high atomic number materials to absorb radiation energy and reduce direct damage to the device. For example, adding a 1mm thick lead shielding layer to the outside of the device can effectively reduce radiation energy and performance degradation.

[0083] Redundant design improves the overall reliability of the system by adding backup devices or modules. For example, a dual redundant system can be designed so that when the main device fails due to radiation exposure, the backup device immediately takes over to ensure the continuity of system functions. These specific measures can not only effectively improve the reliability of devices in radiation environments, but also provide important reference for subsequent device design and application.

[0084] Specifically, it also includes: Obtaining test data: Let D = {(I 1 ,V 1 ,t 1 ,E 1 ,D 1 ,θ 1 ),...,(I n ,V n ,t n ,E n ,D n ,θ n )}where I is current, V is voltage, t is response time, E is ray energy, D is irradiation dose, and θ is the incident angle;

[0085] Data preprocessing: Data cleaning:

[0086] Remove outliers: Identify and remove test data that does not conform to physical laws, such as I outlier ,V outlier ,t outlier ;

[0087] Fill missing values: For missing data points, use interpolation or other methods to fill in the missing values. missing ,V missing ,tmissing ;

[0088] Normalization:

[0089] Current normalization:

[0090] Radiation dose normalization:

[0091] Extract key feature parameters:

[0092] Average response time:

[0093] Maximum current change rate:

[0094] Establish an association model: Use the support vector machine (SVM) algorithm to establish the model f:R 3 →R, where the input is (E, D, θ) and the output is the performance degradation percentage P deg ;

[0095] Model training and evaluation: Training model: Use the preprocessed data D′ to train the SVM model, cross validation: Perform five-fold cross validation and calculate the P value for each validation. deg and average performance;

[0096] Evaluate device reliability: Input different radiation irradiation parameters (E j ,D k ,θ l ), and get the output P degjkl , judgment criteria: If P degjkl >Threshold, the device is unreliable;

[0097] Determine the impact of radiation exposure: Analyze the model output. If Pdeg>30% at high energy and high dose, the impact is considered significant.

[0098] Propose reinforcement measures: Shielding: Calculate the shielding effect, such as , where μ is the linear attenuation coefficient and x is the thickness of the shield. Redundancy design: Design redundant systems and calculate system reliability improvements, such as using parallel or series redundancy;

[0099] In the above calculation, the characters involved and their meanings are as follows:

[0100] D: test data set, containing data points of multiple test samples;

[0101] I: Current, indicating the current value of the device under test;

[0102] V: voltage, indicating the voltage value of the device under test;

[0103] t: response time, which indicates the time required for the device to respond to external stimuli;

[0104] E: Radiation energy, which indicates the radiation energy level of the irradiated device, in mega-electron volts (MeV);

[0105] D: Radiation dose, which indicates the total amount of radiation received by the device, in krad;

[0106] θ: incident angle, which indicates the angle at which the ray hits the device surface;

[0107] I outlier ,V outlier ,t outlier : Outliers, which indicate values ​​in the test data that do not conform to the normal range;

[0108] I missing ,V missing ,t missing : Missing values, indicating missing data points in the test data;

[0109] I norm , Dnorm : Normalized current and irradiation dose, converting the raw data to values ​​between 0 and 1;

[0110] : Average response time, which means the average response time of all test samples;

[0111] : Maximum current change rate, which indicates the maximum rate at which current changes over time;

[0112] f: Support vector machine (SVM) model, used to establish the association between radiation irradiation parameters and device performance degradation;

[0113] P deg : Performance degradation percentage, which is a quantitative indicator of device performance degradation;

[0114] Threshold: The preset threshold used to determine whether the device is reliable;

[0115] E shielded : The energy of the ray after shielding, which indicates the energy level of the ray after passing through the shielding layer;

[0116] E original : Original ray energy, which indicates the energy level of the ray before passing through the shielding layer;

[0117] μ: Linear attenuation coefficient, which indicates the attenuation rate of the ray when it propagates in the material;

[0118] x: Shielding layer thickness, indicating the thickness of the shielding material.

[0119] Furthermore, in step S3, according to the first principle calculation method, the influence of radiation irradiation on the material structure and chemical composition is simulated. First, a typical semiconductor material, such as silicon (Si) or germanium (Ge), needs to be selected;

[0120] Through density functional theory (DFT) calculations, we can obtain the types of defects generated in the material after irradiation, including vacancies (V), interstitial atoms (I), as well as their concentration and size distribution. For example, the calculation results show that under 1MeV electron irradiation, the concentration of vacancy defects in silicon material is about 10 15 cm -3 , the interstitial atomic size distribution is mainly concentrated in the range of 0.5-1.0 nm;

[0121] Using the obtained microstructural parameters, molecular dynamics simulations were performed to study the formation mechanism and evolution behavior of irradiation-induced defects. The simulation time was preset to 1 nanosecond. It can be observed that vacancy defects are formed rapidly in the early stage, and then gradually decrease through recombination with interstitial atoms, while interstitial atoms tend to aggregate to form larger defect clusters.

[0122] By analyzing the defect state density and energy level distribution, it is found that vacancy defects introduce new energy levels at 0.2 eV below the conduction band, while interstitial atomic clusters form defect states at 0.3 eV above the valence band.

[0123] Based on the results of molecular dynamics simulation, the first-principles calculations were used to analyze the effects of defect state density and energy level distribution on the distribution of electronic state density in the material bandgap.

[0124] Calculations show that the energy levels introduced by vacancy defects lead to an increase of about 5% in the electronic state density in the band gap, while interstitial atomic clusters significantly increase the electronic state density at 0.3 eV above the valence band, affecting the electrical conductivity of the material.

[0125] Using semiconductor device simulation software, including SilvacoTCAD, the changes in material barrier height and width were calculated based on the defect state density and energy level distribution. The initial material barrier height was preset to be 1.1 eV and the width was 10 nanometers. The calculation results showed that as the defect state density increased, the barrier height decreased to 0.9 eV and the width narrowed to 8 nanometers, indicating that the material barrier characteristics were significantly degraded.

[0126] Based on the established radiation-induced defect correlation model, the change of the material barrier characteristics is predicted when the defect state density exceeds the intrinsic defect state density of the material. The intrinsic defect state density is preset to be 10 14 cm -3 , when the defect state density reaches 10 15 cm -3When the barrier height is reduced by 20%, the width is reduced by 15%, and the conductive properties of the material are significantly reduced;

[0127] To address the problem of material barrier property degradation caused by irradiation, molecular dynamics simulation was used to adjust the chemical composition of the material, including the doping concentration, and control the defect state density distribution. For example, the doping concentration was increased from 10 16 cm -3 Increase to 10 17 cm -3 The simulation results show that the defect state density is reduced by 30%, the barrier height is restored to 1.0eV, and the width is increased to 9 nanometers, and the material barrier properties are effectively improved;

[0128] The optimized material structure and chemical composition parameters are input into the established radiation-induced defect correlation model to predict the performance changes of the material under the radiation environment. The optimized doping concentration is preset to 10 17 cm -3 ,The calculation results show that under the same irradiation conditions, the material barrier height is only reduced by 5% and the width is reduced by 3%, indicating that the optimized material has better radiation resistance;

[0129] Through the above analysis and examples, it can be seen that the combination of first-principles calculations and molecular dynamics simulations can effectively reveal the impact of radiation irradiation on the microstructure and macroscopic properties of materials, and improve the radiation resistance of materials by optimizing the material structure and chemical composition, providing theoretical guidance for practical applications.

[0130] Specifically, it also includes: material selection: selecting semiconductor material M (such as silicon Si or germanium Ge);

[0131] First principles calculations (DFT):

[0132] Irradiation conditions: 1MeV electron irradiation; Calculation of defect type and concentration: vacancy defect concentration [V] = 10 15 cm -3 , the interstitial atom size distribution range is 0.5-1.0nm;

[0133] Molecular dynamics simulation:

[0134] Simulation time: t = 1ns; observe defect formation and evolution: vacancy defect formation rate r_V and interstitial atomic cluster formation rate r_I;

[0135] Analysis of defect state density and energy level distribution: vacancy defects introduce new energy levels: E_V = 0.2eV (below the conduction band); interstitial atomic clusters form defect states: E_I = 0.3eV (above the valence band);

[0136] Calculation of the effect of defects on the electronic state density: the increase in electronic state density caused by vacancy defects: Δρ_V = 5%; the increase in electronic state density caused by interstitial atomic clusters: Δρ_I (at 0.3 eV);

[0137] Calculation of material barrier characteristics: initial barrier height: E_B0=1.1eV, width: W_B0=10nm; barrier height after defect state density increases: E_B = 0.9eV, width: W_B=8nm;

[0138] Radiation-induced defect correlation model prediction: Intrinsic defect state density: [D_0]=10 14 cm -3 ; Defect state density reaches [D] = 10 15 cm -3 When , the barrier height decreases: ΔE_B=20%, and the width decreases: ΔW_B=15%;

[0139] Material optimization: Adjust doping concentration: from [D_1] = 10 16 cm -3 , increase to [D_2]=10 17 cm -3 ;

[0140] The defect state density decreases: Δρ_D=30%, the barrier height is restored: E_B'=1.0eV, and the width increases: W_B'=9nm;

[0141] Prediction of optimized material properties: Optimized doping concentration: [D_2]=10 17 cm -3 ;

[0142] Under the same irradiation conditions, the barrier height is reduced by ΔE_B'=5% and the width is reduced by ΔW_B'=3%;

[0143] Through these steps, the entire content logic shows from material selection and radiation defect calculation, to molecular dynamics simulation and defect state density analysis, to the calculation and optimization of material barrier characteristics, and finally to the prediction of optimized material properties through the model. This process shows how to use the combination of first-principles calculations and molecular dynamics simulations to reveal the influence of radiation on the microstructure and macroscopic properties of materials, and to improve the radiation resistance of materials by optimizing the material structure and chemical composition, providing theoretical guidance for practical applications.

[0144] Further, in step S4, according to preset irradiation parameters, the irradiation source is controlled to irradiate the device, for example, the acceleration voltage of the X-ray source is set to 50 kilovolts, the beam intensity is set to 10 microamperes, and the semiconductor device is uniformly irradiated. At the same time, high-precision electrical measurement equipment, including a four-probe tester, is used to collect electrical characteristic data of the device in real time, including current-voltage (IV) curves and capacitance-voltage (CV) curves;

[0145] The collected electrical characteristic data is transmitted to a data processing unit, which has a built-in capacitance-voltage measurement algorithm and a deep energy level transient spectrum detection algorithm;

[0146] The capacitance-voltage measurement algorithm calculates the device's internal barrier height and width parameters by analyzing the CV curve. For example, through the slope and intercept of the CV curve, it can be estimated that the barrier height is about 0.8 electron volts and the barrier width is about 10 nanometers.

[0147] The deep level transient spectrum detection algorithm is used to identify the defect state density and energy level distribution inside the device, including the detection of the presence of a concentration of 10 12 Deep level defects per cubic centimeter;

[0148] The calculated barrier parameter data is compared with the preset standard parameter range, where the preset standard range is 0.7-0.9 electron volts for barrier height and 8-12 nanometers for barrier width;

[0149] If the actual measured potential barrier height is only 0.6 electron volts, which is beyond the standard range, it is judged that the radiation intensity is too high. At this time, the radiation source control unit is triggered to reduce the radiation intensity by adjusting the acceleration voltage of the radiation source to 40 kilovolts and the beam intensity to 5 microamperes;

[0150] A decision tree algorithm is used to comprehensively analyze the collected device electrical characteristic data and barrier parameter data;

[0151] The decision tree algorithm classifies data and identifies key factors that affect device performance through multiple nodes and branches. For example, the decision tree identifies that the significant increase in defect state density under high irradiation intensity is the main factor leading to the degradation of barrier characteristics.

[0152] Based on this, a data-driven device performance statistical evaluation model is constructed. This model can predict the changing trend of device performance under different irradiation conditions, and feed back the analysis results of the evaluation model to the radiation source control unit and the data processing unit to dynamically optimize the radiation irradiation parameters and barrier parameter calculation algorithms. For example, the evaluation model recommends adjusting the fitting parameters in the capacitance-voltage measurement algorithm under specific irradiation conditions to improve the calculation accuracy.

[0153] The radiation source control unit adjusts the irradiation parameters according to the optimization suggestions, such as adjusting the acceleration voltage to 45 kilovolts and the beam intensity to 7 microamperes;

[0154] According to the optimized irradiation parameters and barrier parameter calculation algorithm, the next round of irradiation and device parameter detection is carried out;

[0155] Repeat the above steps to form an intelligent closed-loop control of the irradiation process. For example, after the second round of irradiation, the electrical characteristic data is collected again and the potential barrier parameters are calculated. It is found that the potential barrier height is restored to 0.8 electron volts, which is within the standard range, verifying the effectiveness of the optimization measures.

[0156] Comparing the device performance evaluation result with a preset performance lower limit threshold, where the preset performance lower limit threshold is that the barrier height is not less than 0.7 electron volts;

[0157] If the evaluation result shows that the barrier height is only 0.6 electron volts, which is lower than the threshold, an alarm signal will be issued through an audible and visual alarm to indicate that the radiation intensity is too high or the device performance is degraded. At this time, the operator needs to stop the radiation, reset the radiation parameters, or replace and maintain the device.

[0158] Through the above steps, precise control of the irradiation process and real-time monitoring of device performance are achieved;

[0159] Controlling radiation irradiation parameters can avoid device damage caused by excessive irradiation, optimizing the barrier parameter calculation algorithm can improve the accuracy of parameter measurement, and the application of decision tree algorithm can enhance the intelligence level of data analysis.

[0160] Specifically, it also includes: setting of irradiation parameters: setting the acceleration voltage of the X-ray source to Vacceleration = 50 kV and the beam intensity to Ibeam = 10 μA to uniformly irradiate the semiconductor device;

[0161] Electrical characteristic data acquisition: Use a four-probe tester to collect the device's current-voltage (IV) curve and capacitance-voltage (CV) curve in real time;

[0162] Data processing: The collected electrical characteristic data is transmitted to the data processing unit, which has the following built-in algorithms:

[0163] Capacitance-voltage measurement algorithm: Calculate the potential barrier parameters through CV curve analysis. Assuming the slope of the CV curve is m and the intercept is b, the potential barrier height Vbarrier and width Wbarrier can be estimated as: Vbarrier ≈ 0.8 eV (calculated by the slope m) Wbarrier ≈ 10nm (calculated by the intercept b);

[0164] Deep level transient spectrum detection algorithm: Identify defect state density and energy level distribution, assuming that the defect state density is Ndefect=10 12 cm -3 ;

[0165] Barrier parameter comparison: Compare the calculated barrier parameters with the preset standard range: 0.7 eV≤V barrier≤0.9eV8nm≤W barrier≤12nm. If V barrier = 0.6 eV, the irradiation intensity is judged to be too high.

[0166] Radiation source control: adjust the radiation source parameters to reduce the irradiation intensity: Vacceleration = 40kV I beam current = 5μA;

[0167] Decision tree algorithm analysis: Comprehensively analyze the electrical characteristic data and barrier parameters to identify key factors: Assume that the decision tree algorithm identifies that the increase in defect state density is the main factor;

[0168] Performance statistical evaluation model: Build a model to predict the performance trend of the device and provide optimization suggestions: adjust the fitting parameters in the capacitance-voltage measurement algorithm;

[0169] Radiation source parameter adjustment: adjust irradiation parameters according to optimization suggestions: V acceleration optimization = 45 kVI beam current optimization = 7μA

[0170] Closed-loop control: Repeat the irradiation and detection steps to form an intelligent closed-loop control: If the V barrier = 0.8eV after the second round of irradiation, the effectiveness of the optimization measures is verified;

[0171] Performance evaluation and alarm: Compare the device performance evaluation results with the performance lower limit threshold: If V barrier < 0.7eV, an audible and visual alarm is triggered.

[0172] Furthermore, in step S5, according to the performance requirements of the device structure, it is first necessary to determine the target performance indicators of the barrier layer material, mainly including stability and resistance to radiation damage. For example, when preparing flexible electronic devices, the barrier layer material needs to have high thermal stability and resistance to radiation damage to ensure that the device can still maintain good performance under high temperature and radiation environment;

[0173] Through first-principles calculation software such as VASP or QuantumESPRESSO, candidate materials such as zinc oxide (ZnO), titanium dioxide (TiO 2 ) Conduct theoretical simulation and performance prediction;

[0174] During the simulation, the band structure and key parameters of the electronic state density of the material were calculated to predict its performance under irradiation. It was found that TiO 2The material system that exhibits higher structural stability and lower defect generation rate under irradiation will be selected as the optimal material system;

[0175] Using molecular beam epitaxy technology, the screened TiO 2 Epitaxial growth of functional layer materials on the surface of materials, such as perovskite (CH 3 NH 3 PbI 3 );

[0176] Molecular beam epitaxy technology can accurately control the temperature and pressure parameters during the growth process to ensure the interface quality between the functional layer material and the barrier layer material. For example, the growth temperature is set to 500°C and the vacuum degree is set to 10⁻ 8 Torr, by precisely controlling the molecular beam flow rate and growth time, a perovskite film with uniform thickness and high crystal quality is obtained;

[0177] The barrier layer material is precisely grown at the specified location of the flexible electronic device structure by atomic layer deposition technology, with TiO 2 For example, the deposition temperature is controlled at 150°C, and the oxygen and precursor gas flow rates are 50 sccm and 10 sccm, respectively, to ensure uniformity and compactness of the deposition process;

[0178] The advantage of atomic layer deposition technology lies in its layer-by-layer growth characteristics, which can accurately control the film thickness and composition at the nanoscale to obtain high-quality barrier layer films;

[0179] Conduct radiation damage tests on the grown flexible electronic device structure to evaluate the radiation resistance of the barrier layer material. For example, place the device under a 60Co radiation source and conduct irradiation experiments with different doses (such as 1kGy, 10kGy, and 100kGy). Through electrical testing and microscopic observation, evaluate the changes in device performance and damage to the material structure before and after irradiation.

[0180] The preset test results show that at a dose of 100 kGy, TiO 2 The electrical properties of the barrier layer remain stable, indicating that it has good radiation resistance;

[0181] According to the test results, the performance of flexible electronic devices can be further optimized by adjusting the composition and thickness of the barrier layer. For example, by doping Al elements to form Al-doped TiO 2 (Al-TiO 2 ), improve its electrical conductivity and radiation resistance;

[0182] Using first-principles calculations, the electrical and optical properties of the optimized device structure were simulated, and the material parameters of the barrier layer and functional layer, such as the band structure and carrier concentration, were adjusted. The preset calculation results showed that Al-TiO2 The band structure of the perovskite is better matched with the perovskite functional layer, and the carrier concentration is increased by 20%, so this optimization scheme is adopted;

[0183] The device reliability test platform is used to evaluate the long-term working stability of the optimized device structure under irradiation environment. For example, the device is placed in an irradiation chamber simulating the space environment for a 1,000-hour continuous irradiation experiment to monitor the changes in the device's current-voltage characteristics and key parameters of photoelectric conversion efficiency.

[0184] The results of the preset experiments show that the optimized device has stable performance under irradiation environment, and the photoelectric conversion efficiency only decreases by 5%, which verifies the effectiveness of the barrier control technology;

[0185] The target performance indicators of the barrier layer materials are determined to ensure the reliability and stability of the device under specific conditions. The first-principles calculation and material screening can predict the performance of the materials at the atomic scale and improve R&D efficiency.

[0186] Molecular beam epitaxy and atomic layer deposition technologies can precisely control the growth of thin films to ensure the uniformity and density of device structures;

[0187] Radiation damage testing and reliability assessment verify the effectiveness of materials and technical solutions to ensure the stable performance of devices in actual applications;

[0188] Through the comprehensive application of these steps, the device performance can be optimized and improved to meet the application requirements of high-performance devices.

[0189] Specifically, it also includes: determining the target performance index: setting the target performance index of the barrier layer material to stability S and radiation damage resistance R, for example, S and R need to meet certain thresholds S_min and R_min;

[0190] First principles calculations: Use VASP or QuantumESPRESSO software to simulate candidate materials, calculate the band structure E and electronic state density D, and predict the performance under irradiation, for example, compare ZnO and TiO 2 E and D, preset TiO 2 The stability of S_TiO2 and R_TiO2 is higher than that of ZnO;

[0191] Material screening: TiO 2 The performance is better than ZnO, that is, S_TiO2>S_ZnO and R_TiO2>R_ZnO, so TiO is selected 2 As a barrier layer material;

[0192] Molecular beam epitaxy technology: Growth parameters are set to temperature T = 500°C and vacuum P = 10⁻8 Torr, by controlling the molecular beam flow rate F and the growth time t, a uniform perovskite film thickness d is obtained. For example, F and t are adjusted to the optimal values ​​to obtain d = 100 nm;

[0193] Atomic layer deposition technology: The deposition parameters are controlled as temperature T = 150 ° C, oxygen flow F_O2 = 50 sccm and precursor gas flow F_precursor = 10 sccm to obtain uniform and dense TiO 2 film;

[0194] Radiation damage test: set different radiation doses D = {1kGy, 10kGy, 100kGy}, and evaluate the performance changes before and after irradiation through electrical testing and microscopic observation. For example, at a dose of 100kGy, the performance remains stable, that is, the performance degradation rate P_degradation <10%;

[0195] Material optimization: Formation of Al-TiO by doping Al 2 , improve the conductivity C and radiation resistance R, for example, after doping, C_Al-TiO2>C_TiO2 and R_Al-TiO2>R_TiO2;

[0196] Performance simulation: Use first-principles calculations to simulate the optimized structural performance, such as band structure E' and carrier concentration n', and preset Al-TiO 2 The n' of Al-TiO2 increases by 20%, that is, n'_Al-TiO2=n_TiO2 + 0.2n_TiO2;

[0197] Reliability test: Conduct a 1000-hour continuous irradiation experiment to monitor the current-voltage characteristic IV and the photoelectric conversion efficiency η. The preset experimental results show that η only decreases by 5%, that is, η_final=η_initial- 0.05η_initial;

[0198] Through these steps, the entire content logic shows that from determining the target performance indicators to verifying the effectiveness of the optimization plan through experiments, the device performance is finally optimized and improved to meet the application requirements of high-performance devices.

[0199] Furthermore, in step S6, the material, size and electrode arrangement parameters of the screened device structure are obtained as basic data for barrier control. For example, a semiconductor device based on gallium nitride (GaN) is preset, the barrier layer material of the device is aluminum oxide (Al2O3), and the functional layer material is indium gallium nitride (InGaN);

[0200] The device's dimensional parameters include a barrier layer thickness of 10 nanometers, a functional layer thickness of 20 nanometers, and an electrode arrangement of top and bottom electrodes with a spacing of 50 microns. These basic data provide initial conditions for subsequent barrier regulation;

[0201] According to the obtained device structure parameters, the device is acted on and used to change the device barrier characteristics through the irradiation barrier control technology. For example, by using X-ray irradiation technology, by controlling the irradiation dose (such as 10Gy, 20Gy, 30Gy) and irradiation time (such as 10 minutes, 20 minutes, 30 minutes), the electronic structure and defect state density of the barrier layer can be changed, thereby adjusting the barrier height and conductive performance of the device;

[0202] By using the orthogonal test method and comprehensively considering various barrier control technologies, multiple groups of barrier control parameter combinations are generated. Orthogonal test design can effectively reduce the number of experiments and comprehensively examine the influence of various factors on device performance. For example, three factors are set: irradiation dose (10Gy, 20Gy, 30Gy), irradiation time (10 minutes, 20 minutes, 30 minutes) and annealing temperature (300℃, 400℃, 500℃). 2 ) Generate 9 different control parameter combinations, each parameter combination corresponds to different experimental conditions;

[0203] For each set of barrier control parameter combinations, test the device performance indicators to obtain the corresponding data of barrier characteristics and device performance, for example, test the current-voltage (IV) characteristics, carrier mobility, and key performance indicators of photoelectric conversion efficiency of the device under each set of parameter combinations;

[0204] Under a certain set of parameter combinations (20 Gy, 20 minutes, 400 ° C), the current density of the device reaches 100 A / cm² and the photoelectric conversion efficiency is 20%. These data are recorded for subsequent analysis. Using the obtained barrier characteristics and device performance data, the support vector regression (SVR) algorithm is used to establish a mapping relationship model between barrier characteristics and comprehensive control parameters.

[0205] SVR is a powerful machine learning algorithm that can build a high-precision prediction model with limited data samples. For example, 9 sets of test data are input into the SVR model. Through training and learning, the model can capture the nonlinear relationship between irradiation dose, irradiation time, annealing temperature and device performance.

[0206] Input the target device performance, use the established SVR mapping model, and solve the optimization problem to obtain the corresponding optimal barrier control parameter combination. For example, set the target device performance to a current density of 120A / cm² and a photoelectric conversion efficiency of 25%. Input this target into the SVR model, and the model searches for the optimal parameter combination through an optimization algorithm (such as a genetic algorithm or a particle swarm algorithm). The preset optimal parameter combination is (25Gy, 25 minutes, 450°C).

[0207] The optimal barrier control parameter combination obtained by solving the problem is applied to the preparation process of the screened device structure, and the device barrier characteristics are precisely controlled by controlling the radiation dose and annealing temperature conditions. For example, in the actual preparation process, X-ray irradiation and annealing treatment are performed according to the optimal parameter combination (25Gy, 25 minutes, 450°C) to ensure that the barrier layer of the device reaches the expected electronic structure and defect state density, thereby achieving high-performance device characteristics;

[0208] In this way, the barrier modulation technology not only improves the performance of the device, but also enhances the device's radiation resistance and long-term working stability;

[0209] Precise parameter control and model prediction make the device preparation process more efficient and reliable, and ultimately achieve excellent performance of the screened device structure in practical applications. This systematic regulation method provides strong technical support for the design and optimization of semiconductor devices and promotes the rapid development of high-performance devices.

[0210] Specifically, it also includes: Basic data collection: Setting the parameters of semiconductor devices based on gallium nitride (GaN): barrier layer thickness d = 10 nm, functional layer thickness = 20 nm, electrode spacing L = 50μm;

[0211] Application of barrier control technology: Using X-ray irradiation technology, set the irradiation dose D = {10Gy, 20Gy, 30Gy} and irradiation time T = {10 minutes, 20 minutes, 30 minutes} to change the electronic structure and defect state density of the barrier layer;

[0212] Orthogonal experimental design: set three factors: irradiation dose, irradiation time and annealing temperature, respectively D = {10Gy, 20Gy, 30Gy}, T = {10min, 20min, 30min}, annealing temperature A = {300℃, 400℃, 500℃}, use orthogonal table L9(3 2 ) Design 9 groups of experiments;

[0213] Performance test: Test the device performance for each parameter combination and record the data, for example, test the data of current density J=100A / cm² and photoelectric conversion efficiency η=20% under the parameter combination (20Gy, 20 minutes, 400℃);

[0214] Establishing a mapping relationship model: Using the support vector regression (SVR) algorithm, 9 sets of test data were input into the model to establish the relationship between the barrier characteristics and the control parameters;

[0215] Optimize parameter combination: set the target performance J target = 120A / cm², η target = 25%, use the SVR model and optimization algorithm (such as genetic algorithm, particle swarm algorithm) to search for the optimal parameter combination, and the preset optimal parameter combination is (25Gy, 25 minutes, 450℃);

[0216] Apply optimal parameters: In the actual preparation process, X-ray irradiation and annealing are performed according to the optimal parameter combination to ensure that the device achieves the expected performance;

[0217] Reliability Analysis and Optimization:

[0218] Use principal component analysis (PCA) to determine the influence weight of each stress factor on device performance and obtain the influence weight ranking;

[0219] According to the weight ranking, the key stress factors are selected for accelerated life test. For example, the γ-ray dose and electrical stress are selected, and the stress level D = {20Gy, 30Gy, 40Gy}, V = {200V, 300V, 400V} are set. 2 ) Design of experiments;

[0220] Use support vector machine regression algorithm to establish device performance degradation model and predict the long-term reliability of devices;

[0221] If the prediction results do not meet the reliability index, adjust the test conditions and re-model the prediction;

[0222] Parameter optimization and device preparation:

[0223] Based on the prediction of the long-term reliability of the device, the particle swarm optimization (PSO) algorithm is used to optimize the structural parameters and process parameters of the device;

[0224] The preset optimization results show that the optimal parameters are barrier layer thickness of 15nm, electrode material silver, and doping concentration of 1×10 18 cm -3 ;

[0225] The optimized parameters are applied to the device design and manufacturing process to ensure that the parameters reach the expected values.

[0226] Further, in step S7, according to the failure mechanism model, it is first necessary to obtain data of multiple stress factors such as radiation irradiation, electrical stress and thermal stress. For example, a power device based on silicon carbide (SiC) is preset, and the radiation irradiation adopts gamma rays with a dose range of 10Gy to 50Gy; the electrical stress is obtained by applying different voltages in a range of 100V to 500V; the thermal stress is obtained by temperature changes in a range of 100°C to 500°C. These data provide a basis for subsequent analysis;

[0227] Establish a multi-physics field coupling model through finite element analysis method to calculate the coupling strength of different stress factors. For example, use ANSYS software to establish a three-dimensional model of the device, input the radiation dose, electric field intensity and temperature distribution, and simulate and calculate the distribution of various stress factors inside the device and their interactions;

[0228] The preset simulation results show that under the conditions of γ-ray dose of 30Gy, voltage of 300V and temperature of 300℃, the maximum stress concentration area inside the device is located in the barrier layer near the electrode;

[0229] The principal component analysis method is used to determine the weight coefficient of each stress factor on the device performance and obtain the influence weight ranking. For example, the principal component analysis of the simulation data is performed by SPSS software. The preset results show that the influence weight of the gamma ray dose on the device performance is the highest, followed by electrical stress, and finally thermal stress.

[0230] This ranking helped identify key stress factors;

[0231] According to the ranking results of the influence weights, the key stress factors are selected, and the orthogonal test is adopted to set the stress level and acceleration multiple to carry out the accelerated life test. For example, the γ-ray dose and electrical stress are selected as the key factors, and three stress levels are set: the γ-ray dose is 20Gy, 30Gy, and 40Gy, and the electrical stress is 200V, 300V, and 400V;

[0232] Through the orthogonal table L9 (3 2 ) Generate 9 different test conditions, conduct accelerated life tests for each condition, and record the performance degradation data of the device, including leakage current and breakdown voltage;

[0233] The support vector machine regression algorithm is used to establish a mathematical model between device performance degradation and key stress factors. For example, 9 sets of test data are input into the LibSVM software, and a regression model between leakage current, gamma-ray dose and electrical stress is established through training and learning.

[0234] The preset model shows that the leakage current increases nonlinearly with the increase of γ-ray dose and electrical stress;

[0235] Through model extrapolation, the long-term reliability trend of the device under the use conditions is predicted. For example, the preset model predicts that under the normal use conditions of γ-ray dose of 10Gy and electrical stress of 250V, the mean time between failures of the device is 5000 hours and the failure rate is 0.01% / 1000 hours.

[0236] If the prediction results meet the reliability index requirements, proceed to the next step; otherwise, adjust the accelerated life test conditions, reacquire data and make model predictions;

[0237] On the basis of accurately predicting the long-term reliability of the device, the particle swarm optimization intelligent algorithm is used to optimize the device's structural parameters and process parameters. For example, the PSO algorithm is used to optimize the device's barrier layer thickness, electrode material, doping concentration and other parameters;

[0238] The preset optimization results show that the barrier layer thickness is 15 nanometers, the electrode material is silver, and the doping concentration is 1×10 18 cm -3 When , the reliability of the device is the highest;

[0239] The optimized parameters are applied to the device design and manufacturing process. For example, in the actual preparation process, material selection and process control are carried out according to the optimized parameters to ensure that the barrier layer thickness, electrode material and doping concentration of the device reach the expected values.

[0240] Specifically, it also includes: Data collection: setting the radiation dose range to [D]=[10Gy, 50Gy], the electrical stress range to [V]=[100V, 500V], and the thermal stress range to [T]=[100℃, 500℃];

[0241] Establish a multi-physics field coupling model: Use finite element analysis (such as ANSYS) to establish the model, with input parameters of D=30Gy, V=300V, T=300℃, and calculate the internal stress distribution S(D, V, T) of the device;

[0242] Principal component analysis (PCA): The simulation data were analyzed by SPSS software to obtain the weight coefficients of each stress factor Wγ, WE, WT. The preset result is Wγ>WE>WT, indicating that the radiation exposure has the greatest impact;

[0243] Orthogonal experimental design: Select the key stress factors γ-ray dose and electrical stress, set the stress level to D = {20Gy, 30Gy, 40Gy} and V = {200V, 300V, 400V}, and use the orthogonal table L9 (3 2 ) Design 9 groups of experiments;

[0244] Accelerated life test: record the performance degradation data for each set of test conditions to obtain the data set {Ileak, Vbreak};

[0245] Establish a performance degradation model: Use the support vector machine regression (SVMR) algorithm and LibSVM software to establish a model, such as Ileak = f(D, V). The preset model shows that Ileak increases nonlinearly with the increase of D and V.

[0246] Predicting long-term reliability: Predicting the reliability of devices under normal conditions through model extrapolation, such as predicting the mean time between failures (MTBF) = 5000 hours and the failure rate (λ) = 0.01% / 1000 hours;

[0247] Parameter optimization: If the prediction results meet the reliability index, the particle swarm optimization (PSO) algorithm is used to optimize the device parameters. The optimization goal is to maximize the reliability R and obtain the optimal parameters. ; Apply optimized parameters: Apply optimized parameters during device design and manufacturing to ensure that actual parameters meet expected values.

[0248] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0249] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and rules of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for intelligently controlling electric energy barrier of flexible electronic devices, characterized in that: The following steps are involved: S1. According to the material properties and device structure of the flexible electronic device, a mathematical model of barrier regulation is established, and the barrier region is simulated in detail using the finite element analysis method to obtain a quantitative relationship model between the device structure parameters and the device performance. The device structure parameters include barrier height and barrier width. The quantitative relationship model is trained using a vector machine algorithm to obtain the optimized barrier height and barrier width. S2. Using mathematical models and simulation results, the optimized barrier height and barrier width obtained in step S1 are mined and analyzed through machine learning algorithms, and a correlation model is established with inputs of radiation energy, dose, and incident angle irradiation parameters and outputs of quantitative indicators of device performance degradation, so as to evaluate the impact of radiation irradiation on the quantitative indicators of device performance degradation; S3. Based on the correlation model of step S2, the first principle calculation and molecular dynamics simulation methods are used to obtain the position and concentration distribution of the defect state in the bandgap and the influence on the height and width of the barrier in view of the changes in the material structure and chemical composition caused by the radiation. S4. According to the calculation and simulation in step S3, during the radiation irradiation process, the internal barrier parameters of the device are detected online by means of capacitance-voltage measurement and deep level transient spectrum detection, which are used for the control of the radiation irradiation process and the real-time feedback adjustment of the barrier parameters; S5. According to the calculation and simulation of step S3, a barrier control material system is used, barrier layer materials are screened by first-principles calculation, and the composition and thickness of the barrier layer are controlled by atomic layer deposition and molecular beam epitaxy technology to obtain a screened device structure, where the screened device structure is used to suppress the formation of radiation-induced defects; S6. Test the screened device structure, regulate the technical parameters of irradiation dose, irradiation time and annealing temperature through orthogonal test, and establish a mapping relationship between barrier characteristics and regulation parameters for the regulation of barrier parameters and coordinated optimization of device performance; S7. According to the calculation and simulation of step S3, a device failure mechanism model is established. By establishing a multi-physical field coupling model, the influence weights of different stress factors on device performance are analyzed. Accelerated life testing and big data analysis methods are used to record the performance degradation data of the device, including leakage current and breakdown voltage.

2. According to claim 1, a flexible electronic device electric energy barrier control intelligent control method is characterized by: Step S1 includes the following steps: Establish a barrier control mathematical model based on the material properties and structural parameters of electronic devices, and set the value range of barrier height and barrier width; Finite element analysis method is used to perform refined simulation of mesh division of the barrier area to obtain current-voltage characteristic curve, threshold voltage and switch ratio performance parameters under different barrier parameters. The mesh division accuracy is 0.1nm. The quantitative relationship model between barrier parameters and device performance was established using the polynomial fitting method, and the optimal barrier height and barrier width were obtained through training and optimization using the support vector machine algorithm.

3. According to claim 1, a method for intelligently controlling electric energy barrier of a flexible electronic device is characterized in that: Step S2 includes the following steps: Collecting test data of the radiation irradiation device, the test data including performance parameters of current, voltage, response time and radiation parameters of radiation energy, dose, and incident angle; Performing data preprocessing on the test data, wherein the data preprocessing includes data cleaning to remove outliers, fill in missing values ​​and normalize the data; The key characteristic parameters in the test data are extracted, wherein the key characteristic parameters include average response time and maximum current change rate, and a correlation model between device performance degradation and radiation irradiation parameters is established using a support vector machine (SVM) algorithm. The correlation model is evaluated through cross-validation to avoid overfitting.

4. The method for intelligently controlling the electric energy barrier of a flexible electronic device according to claim 1, characterized in that: Step S3 includes the following steps: By using the first principles calculation method, semiconductor materials including silicon or germanium are selected to simulate the effect of radiation irradiation on their structure and chemical composition, and density functional theory DFT calculation is used to determine the type of defect state generated by radiation irradiation, as well as the position and concentration distribution of the defect state in the band gap; The formation mechanism and evolution behavior of radiation-induced defects are simulated by molecular dynamics to form a radiation-induced defect correlation model, and the influence of defect state density and energy level distribution on the electronic state density distribution is analyzed. The calculation results show that the increase of defect state density leads to the degradation of material barrier properties and conductivity properties. Based on the irradiation-induced defect correlation model, the change of material barrier characteristics when the defect state density exceeds the intrinsic value is predicted, and the chemical composition, including the doping concentration, is adjusted through molecular dynamics simulation to control the defect state density distribution, and the optimized parameters are input into the model.

5. The method for intelligently controlling the electric energy barrier of a flexible electronic device according to claim 1, characterized in that: Step S4 includes the following steps: According to the preset radiation irradiation parameters, the semiconductor device is irradiated with an X-ray source, and the electrical characteristic data is collected in real time. The data is transmitted to the data processing unit. The built-in capacitance-voltage measurement algorithm and deep energy level transient spectrum detection algorithm respectively calculate the internal barrier parameters and defect state density of the device, and the calculation results are compared with the standard parameter range to determine the radiation intensity; A decision tree algorithm is used to analyze the electrical characteristics and barrier parameter data, identify the key factors affecting device performance, and build a data-driven statistical evaluation model to predict performance change trends. The predicted results of the performance change trends of the statistical evaluation model are fed back to the control unit and the data processing unit. The radiation source control unit adjusts the parameters according to the predicted results and conducts the next round of irradiation and detection. By repeating the optimization process, the effectiveness of the measures is verified, and the evaluation results are compared with the performance lower limit threshold to determine whether to trigger an audible and visual alarm.

6. The method for intelligently controlling electric energy barrier of a flexible electronic device according to claim 1, characterized in that: Step S5 includes the following steps: Determine the target performance indicators of the barrier layer material according to the device performance requirements, including the ability to resist radiation damage, simulate and predict the performance of candidate materials through first-principles calculation software, and select the optimal material system; Molecular beam epitaxy technology is used to epitaxially grow functional layer materials on the surface of the optimal material system, and atomic layer deposition technology is used to accurately grow barrier layer materials at specified positions. Radiation damage tests are carried out, and device performance is optimized based on the test results. First-principles calculations are used to simulate the performance of the optimized device structure.

7. The method for intelligently controlling the electric energy barrier of a flexible electronic device according to claim 1, characterized in that: Step S6 includes the following steps: Obtaining materials, dimensional parameters and electrode arrangement parameters of the screened device structure, wherein the dimensional parameters include layer thickness and electrode spacing; Use radiation irradiation technology to control the device barrier characteristics, change the electronic structure and defect state density by controlling the irradiation dose and time, and use orthogonal experimental methods to generate multiple sets of control parameter combinations; For each set of parameter combinations, the device performance indicators are tested, and data including current density and photoelectric conversion efficiency are recorded. The support vector regression (SVR) algorithm is used to establish a mapping relationship model between barrier characteristics and control parameters. The mapping relationship model is used to capture the nonlinear relationship between parameters and performance. The target performance is input, and the SVR model solves the optimal barrier control parameter combination through an optimization algorithm.

8. The method for intelligently controlling electric energy barrier of a flexible electronic device according to claim 1, characterized in that: Step S7 includes the following steps: Collect data on multiple stress factors such as radiation, electrical stress and thermal stress to provide data input for device failure mechanism analysis. Use finite element analysis to establish a multi-physics field coupling model to determine the maximum stress concentration area. The principal component analysis method is used to determine the weight of each stress factor on device performance. After sorting, orthogonal experiments are carried out, and the support vector machine regression algorithm is used to establish a mathematical model between performance degradation and key stress factors.

Images