Method for evaluating wide spectrum proton non-ionizing energy damage of silicon carbide power devices

Abstract

The application discloses a method for evaluating non-ionizing energy damage of a wide-energy-spectrum proton of a silicon carbide power device, relates to the technical field of space radiation effects and power semiconductor reliability evaluation, and comprises the following steps: reverse analysis modeling and obtaining a proton energy threshold value for making the proton range reach the upper surface of an epitaxial layer, wherein damage is not considered below the threshold value; the threshold value is greater than or equal to the threshold value, then the interaction between the proton and the material is simulated by using a Monte Carlo method to obtain the characteristics of the secondary heavy ions of the epitaxial layer; the non-ionizing energy damage of the epitaxial layer is calculated at multiple energy points to establish an energy-damage correspondence; the orbit proton energy spectrum is divided into time periods, the orbit parameters are dynamically updated, and integral accumulation is performed to obtain in-orbit cumulative damage; reverse analysis modeling is performed, and the proton energy threshold value is set to eliminate low-energy contributions that cannot reach the epitaxial layer; the characteristics of the secondary heavy ions are extracted by using the Monte Carlo method above the threshold value, and the energy-damage correspondence is established; the energy spectrum is dynamically updated according to the orbit parameters, and integral accumulation is performed in segments, so that the precision and efficiency are considered.

Description

Technical Field

[0001] This invention relates to the field of space radiation effects and power semiconductor reliability assessment technology, and more specifically, to a method for assessing the wide-spectrum proton non-ionization energy damage of silicon carbide power devices. Background Technology

[0002] With the increasing demand for high efficiency, high voltage withstand capability, and high temperature operation in aerospace power supplies and electric propulsion, silicon carbide power devices, due to their large bandgap, high breakdown field strength, and low switching losses, are increasingly being used in critical areas such as satellite power regulation, power conversion, and high-voltage power distribution. In the space radiation environment, proton flux often covers a wide energy spectrum from low to high energies and varies over time with factors such as orbital altitude, inclination, local time, and passage through radiation anomaly regions. Besides ionization, protons in device materials introduce lattice displacement defects through processes such as elastic scattering and inelastic nuclear reactions, forming displacement damage effects characterized by non-ionization energy damage. These defects can lead to carrier lifetime decay, conduction characteristic drift, increased leakage current, or changes in voltage withstand margin, thereby affecting the long-term reliability of the device and the accuracy of mission lifetime assessment.

[0003] Existing engineering methods for assessing displacement damage typically fall into the following categories: First, based on the concept of non-ionizing energy damage or equivalent dose of displacement damage at the material level, the environmental proton energy spectrum is combined with the material response function, and damage indices are obtained through weighted integration of the energy spectrum. Second, the device is simplified to a material layer of equivalent thickness or an equivalent target, and the damage contribution in critical regions is estimated using particle range and energy deposition laws. Third, the Monte Carlo particle transport method is used to simulate the energy loss and reaction process of protons in the device material to obtain an energy deposition distribution or damage-related quantities that more closely approximate the physical mechanism. These methods have certain applicability in different scenarios and provide a foundation for engineering applications. However, when the assessment target specifically points to the epitaxial layer of a silicon carbide power device, and the irradiation source is a broad-spectrum proton with superimposed on-orbit parameters changing over time, existing methods still have room for further improvement in terms of refinement, consistency, and engineering reusability.

[0004] On the one hand, silicon carbide power devices typically have a multi-layer structure including a metal layer, a dielectric layer, an epitaxial layer, and a substrate. The epitaxial layer, as the critical operating region, has a decisive influence on whether protons can reach it and where effective displacement damage occurs. If the energy conditions for protons to reach the upper surface of the epitaxial layer are not explicitly introduced in the evaluation, and all low-energy protons are included in the epitaxial layer damage contribution, it may lead to over-inclusion of epitaxial layer damage or inaccurate contribution classification. On the other hand, when protons undergo nuclear reactions in the material, they produce various secondary particles. Among these, secondary heavy ions have high nuclear energy deposition and displacement defect generation efficiency within the epitaxial layer, and their contribution to non-ionizing energy damage of the epitaxial layer has significant mechanistic implications. If the evaluation process lacks statistical characterization and traceable aggregated description of the types, energies, and quantities of secondary heavy ions, it is difficult to accurately reflect the differences in the sources of epitaxial layer damage at different incident energy points under broad energy spectrum conditions. On the other hand, the on-orbit proton energy spectrum often changes dynamically with orbital evolution and spatial position. If a fixed energy spectrum is used for long-term accumulation and integration, or if there is a lack of an energy spectrum update mechanism triggered by changes in orbital parameters, the cumulative damage results may not correspond sufficiently to the actual orbital environment. At the same time, if a full-spectrum, full-process Monte Carlo simulation is performed from scratch for each on-orbit evaluation, although the physical consistency is strong, the computational cost is high, which is not conducive to mission scheme iteration, rapid assessment of orbital changes, and engineering deployment. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a method for assessing the non-ionizing proton energy damage of silicon carbide power devices with a wide energy spectrum.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for assessing the non-ionizing proton energy damage of silicon carbide power devices with a wide energy spectrum, executed on an electronic computing device, includes the following steps: Step 1: Perform reverse analysis on the silicon carbide power device and establish a silicon carbide power device model based on the reverse analysis results; determine the proton energy threshold; when the incident proton energy is lower than the proton energy threshold, the non-ionization energy damage to the epitaxial layer caused by the incident proton is not included; Step 2: When the incident proton energy is not lower than the proton energy threshold, the Monte Carlo method is used to simulate the interaction between the incident proton and the material in the silicon carbide power device model to obtain the feature set of secondary heavy ions generated by the incident proton in the epitaxial layer of the silicon carbide power device. Step 3: For multiple incident proton energy points where the incident proton energy is not lower than the proton energy threshold, calculate the non-ionization energy damage of the epitaxial layer of the silicon carbide power device based on the secondary heavy ion feature set, and establish the correspondence between the incident proton energy and the non-ionization energy damage of the epitaxial layer. Step 4: Obtain the proton energy spectrum under any space orbit and divide the on-orbit accumulation time of the silicon carbide power device into multiple time periods; for each time period, dynamically update the proton energy spectrum according to the changes in orbital parameters. When the changes in orbital parameters exceed the preset orbital parameter change threshold, the dynamic update of the proton energy spectrum is triggered; based on the dynamically updated proton energy spectrum and according to the correspondence established in Step 3, calculate the epitaxial non-ionization energy damage suffered by the silicon carbide power device during on-orbit accumulation.

[0007] Further reverse analysis includes: preparing cross-sectional samples of silicon carbide power devices to obtain the layer structure sequence, measuring layer-by-layer thickness using microscopic imaging, obtaining the elemental composition of each layer using energy dispersive spectroscopy, and determining the material density and doping parameters in conjunction with device process information; and generating a silicon carbide power device model based on layer-by-layer thickness, elemental composition, material density, and doping parameters.

[0008] Furthermore, the proton energy threshold is the minimum incident proton energy required for the incident proton range to reach the upper surface of the epitaxial layer of the silicon carbide power device. The determination of the proton energy threshold includes: simulating the range of different incident proton energies in the silicon carbide power device model using the Monte Carlo method, and determining the proton energy threshold based on the condition that the incident proton range reaches the upper surface of the epitaxial layer of the silicon carbide power device.

[0009] Furthermore, the set of secondary heavy ion characteristics includes: the types of secondary heavy ions produced by the incident protons, the energy of the secondary heavy ions, and the number of secondary heavy ions.

[0010] Furthermore, the Monte Carlo method was used to simulate the interaction between incident protons and materials in the silicon carbide power device model and obtain a secondary heavy ion feature set. This included: setting initial conditions for incident protons based on the silicon carbide power device model, including at least incident proton energy, incident direction, and incident position; randomly sampling particle transport, nuclear reaction, and scattering processes of incident protons in the silicon carbide power device model and tracking the secondary heavy ions generated by incident protons in the silicon carbide power device model; setting the epitaxial layer as a scoring region, performing event statistics on secondary heavy ions entering or generated within the scoring region, and recording the feature parameters corresponding to the secondary heavy ion feature set; and generating the secondary heavy ion feature set based on the event statistics results and the recorded feature parameters.

[0011] Furthermore, step three calculates the epitaxial layer non-ionizing energy damage based on the secondary heavy ion feature set, including: grouping the secondary heavy ion feature set according to the type and energy of secondary heavy ions to obtain the number and energy representative value of secondary heavy ions in each group; calling a pre-established nuclear damage mapping function for each group to convert the energy representative value into a nuclear damage energy share; weighting and summing the nuclear damage energy shares of each group according to the number of secondary heavy ions in each group to obtain the total nuclear damage energy of the epitaxial layer; and normalizing the total nuclear damage energy of the epitaxial layer according to the epitaxial layer mass to obtain the non-ionizing energy damage of the epitaxial layer.

[0012] Furthermore, the nuclear damage mapping function is constrained by a preset displacement threshold parameter of the epitaxial layer material, and a smooth transition is adopted within the energy transition interval centered on the preset displacement threshold parameter and determined by a preset energy bandwidth; when weighting and accumulating the nuclear damage energy share, a robust weight based on the probability of group occurrence is introduced to suppress the influence of low-probability grouping on the non-ionization energy damage of the epitaxial layer.

[0013] Furthermore, establishing the correspondence between incident proton energy and epitaxial layer non-ionization energy damage includes: determining the incident proton energy point set, which covers the target incident proton energy range above the proton energy threshold and satisfies a preset energy step size or preset energy grid; for each incident proton energy point in the incident proton energy point set, performing step two to obtain the secondary heavy ion feature set and performing step three to obtain the corresponding epitaxial layer non-ionization energy damage value; forming a discrete correspondence dataset by combining the incident proton energy point set and the epitaxial layer non-ionization energy damage value set; and generating a mapping model based on the discrete correspondence dataset and storing it as a correspondence.

[0014] Furthermore, the mapping model can be any one of lookup table mapping, piecewise interpolation mapping, or regression fitting mapping, so that the non-ionization energy damage value of the epitaxial layer can be obtained through the correspondence for any incident proton energy within the target incident proton energy range.

[0015] Furthermore, in step four, the proton energy spectrum components with incident proton energies not lower than the proton energy threshold are integrated according to the correspondence established in step three to obtain the epitaxial layer non-ionization energy damage for the corresponding time period. The epitaxial layer non-ionization energy damage for each time period is then accumulated to obtain the epitaxial layer non-ionization energy damage accumulated on-orbit by the silicon carbide power device.

[0016] Compared with the prior art, the present invention has the following beneficial effects: This invention performs reverse analysis on silicon carbide power devices and establishes a silicon carbide power device model consistent with the layer structure and material parameters. At the same time, it determines the proton energy threshold that allows the incident proton range to reach the upper surface of the epitaxial layer, and excludes the energy spectrum components of the incident proton below the proton energy threshold from the epitaxial layer non-ionization energy damage. This achieves physical constraint and screening of invalid contributions at the evaluation source, avoids miscounting low-energy protons that cannot reach the epitaxial layer as epitaxial layer displacement damage, and improves the accuracy and interpretability of epitaxial layer damage contribution classification. When the incident proton energy is not lower than the proton energy threshold, the Monte Carlo method is used to simulate the interaction between the incident proton and the device material, obtain the characteristic set of secondary heavy ions generated by the incident proton in the epitaxial layer, and further calculate the non-ionization energy damage of the epitaxial layer at multiple incident proton energy points, establish the correspondence between the incident proton energy and the non-ionization energy damage of the epitaxial layer, so that the evaluation under the wide energy spectrum can reflect the key contribution of nuclear reaction products to displacement damage by statistical characterization of the type, energy and quantity of secondary heavy ions, while solidifying the high-overhead energy-point physical calculation into a reusable correspondence, taking into account both mechanism consistency and engineering efficiency; By acquiring the proton energy spectrum under arbitrary space orbits and dividing the on-orbit accumulation time of silicon carbide power devices into multiple time periods, the proton energy spectrum is dynamically updated according to changes in orbital parameters. When the changes in orbital parameters exceed a preset threshold, the energy spectrum is updated. Then, based on the dynamically updated proton energy spectrum and calling the corresponding relationship, the energy spectrum integration and segmented accumulation are performed. This ensures that the on-orbit accumulated epitaxial layer non-ionization energy damage results are consistent with orbital evolution. This avoids environmental mismatch caused by using a fixed energy spectrum for a long time and avoids the computational burden caused by repeatedly performing full-process Monte Carlo calculations for each accumulation evaluation, thereby improving the adaptability, scalability and evaluation efficiency of on-orbit applications. Attached Figure Description

[0017] Figure 1 A schematic diagram of the overall process for assessing the broadband proton non-ionization energy damage of silicon carbide power devices; Figure 2 A schematic diagram showing the relationship between the cross-section of the silicon carbide power device layer structure and the proton energy threshold determination; Figure 3 This is a schematic diagram of obtaining the epitaxial hierarchical heavy ion feature set based on Monte Carlo simulation; Figure 4 A schematic diagram illustrating the on-orbit accumulation of non-ionization energy damage in the epitaxial layer by dynamically updating the proton energy spectrum in segments and integrating the data under any space orbit. Detailed Implementation

[0018] In this specification, "non-ionizing energy damage" refers to the nuclear damage energy corresponding to a unit mass of epitaxial layer material, denoted as the epitaxial layer non-ionizing energy damage value, the unit of which can be energy / mass; if the energy unit is converted to joules, the unit is joules per kilogram; if the energy unit is converted to electron volts, the unit is electron volts per gram or its equivalent unit.

[0019] "Nuclear damage energy share" refers to the energy portion corresponding to the lattice displacement and displacement cascade caused by secondary charged particles in the epitaxial layer material. It is a function of the energy of the secondary charged particles, and its value ranges from zero to the energy of the secondary charged particles.

[0020] "Secondary heavy ions" refer to charged secondary particles generated by incident protons in device materials through nuclear reactions or scattering processes that meet a preset atomic number condition. In one embodiment, secondary heavy ions are charged ion nuclear reaction products with an atomic number not less than a preset threshold. In another embodiment, alpha particles are counted as a separate category or included in the above-mentioned secondary heavy ion statistics, specifically based on the particle nuclide identification at the time of event recording.

[0021] The “scoring region extension layer” refers to the statistical region set in the Monte Carlo transport simulation. Secondary charged particles entering or generated within this statistical region are counted and their energy is recorded according to a preset caliber. The event caliber includes at least the particle nuclide identifier, the kinetic energy when entering the scoring region, and the event count contribution.

[0022] Reference Figures 1-4 A method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices, including... Step 1: Perform reverse engineering on the silicon carbide power device and establish a silicon carbide power device model based on the reverse engineering results; determine the proton energy threshold; when the incident proton energy is lower than the proton energy threshold, the non-ionizing energy damage to the epitaxial layer caused by the incident proton is not included; obtain the device structure and material basis required for evaluation, and establish a judgment boundary in the energy dimension for "whether it enters the epitaxial layer and causes effective non-ionizing energy damage". By performing reverse engineering and establishing a silicon carbide power device model, the subsequent Monte Carlo simulation has a realistic layer structure, material composition, and geometric boundaries, thus ensuring that the transport, reaction, and energy deposition processes of incident protons inside the device have a traceable physical carrier. By determining the proton energy threshold, incident protons are divided into two categories according to energy: "potentially contributing to damage to the epitaxial layer" and "contributing negligiblely to damage to the epitaxial layer". This allows the evaluation to focus on incident protons that can reach and act on the epitaxial layer, avoiding the introduction of invalid contributions from low-energy incident protons into the non-ionizing energy damage results of the epitaxial layer, thereby improving computational efficiency and enhancing the engineering usability and consistency of the results.

[0023] In one specific embodiment, to determine whether the non-ionization energy damage to the epitaxial layer caused by an incident proton with an energy lower than the proton energy threshold is included, this embodiment performs the following steps to form a repeatable determination process: Reverse engineering is performed on silicon carbide power devices to establish silicon carbide power device models. Specifically, reverse engineering involves preparing cross-sectional samples of the silicon carbide power device to obtain the layer structure sequence, measuring layer-by-layer thickness using microscopic imaging, obtaining the elemental composition of each layer using energy dispersive spectroscopy (EDS), and determining material density and doping parameters based on device process information. After obtaining the layer-by-layer thickness, elemental composition, material density, and doping parameters, the layer structure sequence is restored according to spatial position and input into an electronic computing device to generate a silicon carbide power device model. For example, after preparing a cross-sectional sample of a silicon carbide power device with a rated withstand voltage of 1200 volts, the layer structure sequence of the metal layer, dielectric layer, epitaxial layer, and substrate layer can be obtained. Microscopic imaging shows that the epitaxial layer thickness is 10 micrometers. EDS analysis shows that the main elemental composition of the epitaxial layer is silicon and carbon, accompanied by doping element peaks. Process information is used to determine the material density and doping parameters of the epitaxial layer, thus completing the parameterization of the silicon carbide power device model. Candidate ranges for proton energy thresholds were determined, and the Monte Carlo method was used to simulate the range of different incident proton energies in a silicon carbide power device model. During the simulation, the incident direction and incident position were kept in accordance with the evaluation boundary conditions. Multiple incident proton energies were calculated point by point, and the relative positional relationship between the incident proton range and the upper surface of the epitaxial layer of the silicon carbide power device was recorded. To improve the determination accuracy, the candidate range could be covered with a larger energy step size first, and then the energy points could be refined near the critical point with a smaller energy step size. For example, the energy step size could be 0.5 MeV for scanning first, and then the range could be refined with an energy step size of 0.1 MeV in the interval where the range is just close to the upper surface of the epitaxial layer of the silicon carbide power device. The proton energy threshold is determined based on the condition that the incident proton range reaches the upper surface of the epitaxial layer of the silicon carbide power device. This proton energy threshold is then fixed as the criterion for subsequent evaluation. When the incident proton energy is lower than the proton energy threshold, the non-ionization energy damage to the epitaxial layer caused by the incident proton is not included. This eliminates incident protons that cannot reach the upper surface of the epitaxial layer and contribute to effective displacement damage at the source, ensuring that the evaluation of non-ionization energy damage to the epitaxial layer is contributed only by the incident proton energy range that can meet the geometric arrival conditions. For example, if the Monte Carlo method indicates that the incident proton range still does not reach the upper surface of the epitaxial layer of the silicon carbide power device when the incident proton energy is 3 MELV, but reaches the upper surface of the epitaxial layer of the silicon carbide power device when the incident proton energy is 3.4 MELV, then the proton energy threshold is determined to be 3.4 MELV. Based on this, in subsequent calculations, the non-ionization energy damage to the epitaxial layer is not included in cases where the incident proton energy is lower than 3.4 MELV.

[0024] In one implementation, to avoid overlooking the contribution of secondary charged particles generated by the nuclear reaction of the upper material that can still enter the epitaxial layer, by using only "the incident proton range reaches the upper surface of the epitaxial layer" as the proton energy threshold, after determining the candidate proton energy threshold, sampling calculations in steps two and three can be performed on several incident energy points below the candidate threshold to obtain an upper bound estimate of the total nuclear damage energy of the epitaxial layer; when the upper bound estimate does not exceed a preset proportion threshold (e.g., does not exceed a preset proportion of the total nuclear damage energy of the epitaxial layer accumulated in the energy range above the threshold), the candidate proton energy threshold is solidified as the final proton energy threshold; if it exceeds the preset proportion threshold, the candidate threshold is adjusted upward or "the nuclear damage energy of the epitaxial layer reaches a negligible upper limit" is used as the threshold criterion, thereby ensuring that the threshold screening will not cause a systematic omission of non-ionizing energy damage of the epitaxial layer.

[0025] Step Two: When the incident proton energy is not lower than the proton energy threshold, the Monte Carlo method is used to simulate the interaction between the incident proton and the material in the silicon carbide power device model, obtaining the characteristic set of secondary heavy ions generated by the incident proton in the epitaxial layer of the silicon carbide power device. The "real interaction process after the incident proton enters the device" is transformed into a more directly relevant characterization quantity for epitaxial layer damage, namely, the secondary heavy ion characteristic set. When the incident proton energy is not lower than the proton energy threshold, the incident proton can undergo processes such as particle transport, scattering, and nuclear reactions in the silicon carbide power device model, and may generate various secondary heavy ions. Secondary heavy ions typically contribute more concentratedly and significantly to energy deposition and displacement damage in epitaxial layers. Therefore, statistical simulation of interactions using the Monte Carlo method can obtain information such as the types, energies, and quantities of secondary heavy ions induced by incident protons in epitaxial layers under conditions of randomness and multiple physical processes. This provides a foundation of input data for subsequent calculations of non-ionizing energy damage in epitaxial layers using the "secondary heavy ion-damage" path.

[0026] In one specific embodiment, when the incident proton energy is not lower than the proton energy threshold, in order to obtain the characteristic set of secondary heavy ions generated by the incident proton in the epitaxial layer of the silicon carbide power device, this embodiment uses the Monte Carlo method to simulate the interaction between the incident proton and the material in the silicon carbide power device model, and performs the following steps in a step-by-step manner: The initial conditions for incident protons are set based on a silicon carbide power device model. The initial conditions include at least the incident proton energy, incident direction, and incident position. The incident proton energy is selected as a target energy point not lower than the proton energy threshold. The incident direction is set as vertical or oblique incident according to the irradiation attitude of the device. The incident position is set as a specific position within the projection area of ​​the epitaxial layer or a position sampled according to a uniform distribution, according to the coverage of the irradiated area. This is to enable the simulation to reflect the influence of different incident geometries on the interaction path length and nuclear reaction probability. For example, after the proton energy threshold is determined, incident proton energies such as 5 MeV, 10 MeV, and 30 MeV can be selected. The incident direction is set as either consistent with the normal of the upper surface of the epitaxial layer or deviated from the normal by a certain angle. At the same time, multiple incident positions are selected within the incident window corresponding to the epitaxial layer to cover the difference between the edge and center of the device. In a silicon carbide power device model, random sampling of incident protons is performed on particle transport, nuclear reactions, and scattering processes. The secondary heavy ions generated by the incident protons within the model are tracked. Specifically, this includes random sampling of the path of freedom within the material layer based on interaction cross-sections and energy loss patterns; random sampling of possible nuclear reactions and scattering types; and random sampling of the energy and emission angle of reaction products. Charged heavy ions generated in each reaction or scattering event are used as candidates for secondary heavy ions, and trajectory advancement and energy decay calculations are performed until the secondary heavy ion is reached. Heavy ions stop, escape from the device boundary, or enter the epitaxial layer scoring region. To ensure statistical reliability, the incident proton process simulation can be repeated multiple times for each set of initial conditions, and the relative uncertainty of the event statistical results reaching a preset level can be used as the convergence criterion. This improves the stability of secondary heavy ion statistics while maintaining the consistency of the physical process. For example, for the initial conditions of incident proton energy of 10 MeV and incident direction of oblique incidence, tens of thousands of incident proton processes can be simulated repeatedly to bring the statistical fluctuation of the number of secondary heavy ions in the epitaxial layer to an acceptable range. Monte Carlo physical processes and convergence criteria: In one implementation, the Monte Carlo simulation at least activates the electromagnetic energy loss process, the elastic scattering process, and the inelastic nuclear reaction process, and records the charged secondary particles generated by the nuclear reaction; the number of repeated processes for each set of initial incident conditions is not less than the preset minimum number of processes. The "relative uncertainty" can be obtained by repeated batch statistics or variance estimation: the incident proton processes under the same initial conditions are divided into several statistical batches, and the batch mean and batch standard deviation are calculated for the total number of secondary heavy ions or the total nuclear damage energy of the epitaxial layer in the scoring region, respectively, and the ratio of the batch standard deviation to the batch mean is used as the relative uncertainty. The "preset level" can be set so that the relative uncertainty is not greater than a preset threshold; in one example, the preset threshold can be in the range of 5% to 10%; when the relative uncertainty does not reach the preset level, the number of incident proton processes is increased until the preset level is met, thereby ensuring the repeatability and numerical stability of the secondary heavy ion statistical results and the epitaxial layer non-ionizing energy damage results; The epitaxial layer is set as the scoring region. Event statistics are performed on secondary heavy ions entering or generated within the scoring region, and the characteristic parameters corresponding to the secondary heavy ion feature set are recorded. A secondary heavy ion feature set is generated based on the event statistics and the recorded characteristic parameters. This feature set includes the type, energy, and quantity of secondary heavy ions generated by the incident protons. During recording, each secondary heavy ion entering the scoring region is considered a basic event. At least the type of secondary heavy ion corresponding to its nuclide identifier, the energy of the secondary heavy ion entering the scoring region, and the count contribution of that energy event are recorded. During the statistical summary phase, the event is further categorized by secondary heavy ion type and energy range. Events are merged to obtain the energy and quantity distribution of each type of secondary heavy ion in the epitaxial layer, thus forming a secondary heavy ion feature set that can be directly used for subsequent calculations. For example, multiple secondary heavy ion events can be statistically analyzed within the scoring region of the epitaxial layer. Some events correspond to higher energy ranges and are fewer in number, while others correspond to medium energy ranges and are more in number. Finally, the secondary heavy ion feature set is output with the type of secondary heavy ion as the main index, the energy of secondary heavy ion as the attribute, and the statistical value of the quantity of secondary heavy ions attached. This allows the generation pattern of secondary heavy ions in the epitaxial layer under different incident proton energies to be quantified and compared, and to serve as the input basis for subsequent calculation of non-ionizing energy damage in the epitaxial layer.

[0027] Step 3: For multiple incident proton energy points with incident proton energies not lower than the proton energy threshold, calculate the non-ionizing energy damage of the epitaxial layer of the silicon carbide power device based on the secondary heavy ion feature set, and establish the correspondence between incident proton energy and epitaxial layer non-ionizing energy damage. Map the discrete "incident proton energy points" to the "epitaxy layer non-ionizing energy damage" results that can be used for engineering evaluation, and form a reusable energy-damage correspondence. By selecting multiple incident proton energy points and performing calculations point by point, the variation law of the contribution of different energies to epitaxial layer damage within the target incident proton energy range can be covered. Calculating the epitaxial layer non-ionizing energy damage based on the secondary heavy ion feature set obtained in Step 2 establishes a link between damage assessment and the actual secondary particle processes induced by incident protons inside the device, thereby enabling the quantitative expression of the variation law of epitaxial layer non-ionizing energy damage with incident proton energy. Further establishing the correspondence between incident proton energy and epitaxial layer non-ionization energy damage allows for rapid calculation by directly calling the correspondence instead of repeatedly performing complex Monte Carlo simulations and event-by-event statistics during subsequent on-orbit proton energy spectrum integration. This enables an executable evaluation process with "wide energy spectrum and long-term accumulation".

[0028] In one specific implementation, to calculate the non-ionization energy damage of the epitaxial layer of a silicon carbide power device based on the secondary heavy ion feature set for multiple incident proton energy points with incident proton energies not lower than the proton energy threshold, and to establish the correspondence between incident proton energies and the non-ionization energy damage of the epitaxial layer, this implementation performs the following steps in a step-by-step manner and forms a storable and reusable correspondence: A set of incident proton energy points is determined, covering the target incident proton energy range above the proton energy threshold and satisfying a preset energy step size or preset energy grid. The preset energy step size or preset energy grid is used to balance computational overhead and energy resolution. Typically, when the energy range is wide, a segmented energy grid is used, i.e., the energy points are denser in the range where energy changes are sensitive to damage to the non-ionizing energy of the epitaxial layer, and sparse in the range where the changes are gradual. For example, in the range from above the proton energy threshold to higher incident proton energies, a coarse energy point set is first generated with a larger energy interval, and then supplementary energy points are inserted in the energy range where the non-ionizing energy damage of the epitaxial layer changes significantly with the slope of the incident proton energy. This ensures that the subsequently generated discrete corresponding dataset has both coverage and representativeness. For each incident proton energy point in the incident proton energy point set, step two is performed to obtain the secondary heavy ion feature set and a corresponding secondary heavy ion feature set record is established for each incident proton energy point. The record content includes at least the type, energy, and quantity of secondary heavy ions generated by the incident proton, so as to ensure that the subsequent calculation of non-ionization energy damage to the epitaxial layer can be traced back to the statistical results of secondary heavy ions generated by the incident proton in the epitaxial layer on an energy point-by-energy basis. For example, secondary heavy ion feature sets can be formed for the same incident proton energy point under different incident direction conditions. In this embodiment, the same incident direction and incident position constraints are selected according to the preset evaluation boundary conditions so that the differences between different incident proton energy points mainly reflect the changes in the secondary heavy ion generation pattern caused by the change in incident proton energy. For each incident proton energy point, the secondary heavy ion feature set is grouped according to the type and energy of the secondary heavy ion. This yields the number of secondary heavy ions and the energy representative value for each group. The energy representative value can be the statistical representative quantity of the secondary heavy ion energy within the group to maintain the consistency and comparability of the grouping. A fixed division rule is used for the boundary of the energy interval during grouping to achieve isomorphic grouping between different incident proton energy points. For example, secondary heavy ions of the same type can be divided into multiple groups according to several continuous energy intervals. The number of secondary heavy ions in each group is counted, and the representative value of the energy interval of that group is used as the input for the subsequent nuclear damage mapping function, thereby converting the event-level secondary heavy ion statistical results into a group-level computable input. For each group, a pre-established nuclear damage mapping function is invoked to convert the energy representative value into a nuclear damage energy share. The nuclear damage energy shares of each group are then weighted and accumulated based on the number of secondary heavy ions in that group to obtain the total nuclear damage energy of the epitaxial layer. Finally, the total nuclear damage energy of the epitaxial layer is normalized according to the epitaxial layer mass to obtain the non-ionizing energy damage of the epitaxial layer. The nuclear damage mapping function is constrained by a preset displacement threshold parameter of the epitaxial layer material and employs a smooth transition within the energy transition range centered on the preset displacement threshold parameter and determined by a preset energy bandwidth, thereby avoiding discontinuities or non-physical abrupt changes near the critical energy. The preset displacement threshold parameter and the preset energy bandwidth... The setting rules for the wide range of displacement thresholds are explained in conjunction with existing technologies as follows: In existing technologies, displacement damage assessment of semiconductor materials typically uses the minimum energy required for a material atom to be ejected from its lattice position as the physical basis for the displacement threshold. This is combined with the material's crystal structure and the differences in displacement thresholds between different atomic sublattices. In engineering calculations, an equivalent displacement threshold parameter matching the material is used to constrain the starting conditions for the nuclear damage energy share. For the epitaxial layer of silicon carbide power devices, the preset displacement threshold parameter can be selected based on the range of values ​​for the silicon carbide material displacement threshold in published literature and general irradiation damage modeling methods. This is combined with the epitaxial layer doping and density obtained from reverse analysis. Consistency verification of the degree information ensures that the nuclear damage mapping function's suppression of displacement damage in the low-energy region and its response to displacement damage in the high-energy region conform to the material's physical properties. The preset energy bandwidth defines the width of the energy transition range. Existing technologies often improve numerical stability by smoothing the probability of energy deposition and displacement near the critical point. In this embodiment, the preset energy bandwidth is determined jointly by the energy group width and the statistical dispersion of secondary heavy ion energy, ensuring it is not less than the energy span of a single energy group and not greater than the range allowed by the statistical resolution near the critical point. This suppresses grouping without masking the true energy dependence. Boundary effect; simultaneously, when weighting and accumulating the nuclear damage energy share, a robust weight based on the group occurrence probability is introduced to suppress the influence of low occurrence probability groups on the non-ionization energy damage of the epitaxial layer. The robust weight can change monotonically with the group occurrence probability, so that the contribution of groups with low occurrence probability and large statistical fluctuations is moderately converged without changing the overall trend. For example, for a certain incident proton energy point, if the number of secondary heavy ions in some high-energy groups is extremely small and fluctuates greatly, the robust weight reduces its amplification effect on the total nuclear damage energy of the epitaxial layer, thereby keeping the epitaxial layer non-ionization energy damage results stable under different repeated simulation conditions. In one implementation, the nuclear damage mapping function is used to convert the energy representative value of a group into the nuclear damage energy share of that group. Let the energy representative value of the group be... The equivalent displacement threshold parameter of the epitaxial layer material is The preset energy bandwidth is The nuclear damage energy fraction is defined as... ;in, Let be the proportion function of nuclear damage energy share, satisfying ; For threshold constraints and transition smoothing terms, defined as ; Therefore, we can conclude that: when At that time, the nuclear damage energy share is zero; when At that time, the energy share of nuclear damage is taken ;when To achieve a smooth transition, a smoothing function is introduced. This makes the nuclear damage energy share of ,in It monotonically transitions from zero to one within the transition interval, and satisfies at least first-order continuity at the endpoints of the interval.

[0029] In one example, the smoothing function can take the cubic smoothing form: Let ;but ; proportional function It can be obtained in any of the following ways: First, by using an analytical or semi-analytical model based on the theory of nuclear energy separation. Secondly, a lookup table of "energy-nuclear damage fraction ratio" for epitaxial layer materials was established in advance, and the data was analyzed... Interpolation is performed to obtain .

[0030] In one implementation, the equivalent displacement threshold parameter The displacement threshold is determined based on the physical meaning of the lattice displacement of the epitaxial layer material and selected in conjunction with the material displacement threshold range given in published literature. When the epitaxial layer material contains different sublattices or displacement threshold differences in different directions, a conservative selection principle (taking the larger value) or a weighted equivalence principle (weighted according to the sublattice ratio or according to the sensitivity to displacement damage) can be used to obtain the equivalent displacement threshold parameter. To ensure reproducibility, the selected parameters are given in the examples. The specific numerical value or range of values ​​and the basis for their source.

[0031] In one implementation, a preset energy bandwidth is provided. Not less than the width of the energy group And it is related to the energy dispersion within the group; for example, take ; in, The standard deviation of the secondary heavy ion energy for this group. For preset coefficients and By clarifying The relationship between group width and dispersion makes the determination of the "energy transition range" feasible.

[0032] One way to implement robust weighting is to set the weighted sum of the nuclear damage energy shares of each group, assuming the first... The probability of grouping is ;in This represents the number of secondary heavy ions in this group. Robust weighting. It can be taken as a monotonic function, making When the value is small, the weight tends to be small to suppress the amplification of statistical fluctuations; for example, it can be taken as... in The lower bound threshold and , For preset index and .when Time can be ordered .

[0033] A discrete correspondence dataset is formed by combining the set of incident proton energy points with the set of epitaxial layer non-ionizing energy damage values. A mapping model is generated based on the discrete correspondence dataset and stored as a correspondence relationship. The mapping model can be any of the following: lookup table mapping, piecewise interpolation mapping, or regression fitting mapping. This ensures that the epitaxial layer non-ionizing energy damage value can be obtained through the correspondence relationship for any incident proton energy within the target incident proton energy range. For example, when the energy point density of the discrete correspondence dataset is high, lookup table mapping can be used for fast retrieval. When the discrete correspondence dataset requires higher continuity in local energy bands, piecewise interpolation mapping can be used to ensure monotonicity and smoothness. When it is necessary to express the overall trend with less storage over a wider energy range, regression fitting mapping can be used, and the accuracy can be constrained by the fitting residual. This allows subsequent integral calculations for a wide-spectrum proton environment to directly call this correspondence relationship and obtain an engineering-reusable basis for epitaxial layer non-ionizing energy damage assessment while maintaining sufficient public access.

[0034] Errors and constraints of the correspondence mapping model: In one implementation, a preset accuracy index is set when generating the mapping model: cross-validation or leave-one-out validation is performed on the discrete correspondence dataset, the relative error between the output of the mapping model and the discrete calculated value is calculated, and the relative error not exceeding a preset error threshold is used as the criterion for model usability; when the preset error threshold is not met, the incident proton energy points are densified in the energy range with large errors or piecewise interpolation is used to improve local accuracy. To avoid non-physical oscillations, monotonicity or non-negativity constraints can be applied when using interpolation or regression fitting to keep the non-ionization energy damage value of the epitaxial layer non-negative within the target energy range and to meet the preset smoothness requirements in the change of adjacent energy bands, thereby ensuring the stability of the correspondence when calling the energy spectrum integral.

[0035] In one embodiment, the equivalent displacement threshold parameter can be selected by combining the threshold displacement energy levels of different sublattices of silicon carbide in published literature, and the equivalent displacement threshold parameter of the epitaxial layer material can be obtained by using a conservative selection or weighted equivalent method. For example, for 4H-SiC material, the literature reports that the threshold displacement energy varies for different atoms and directions, and the range of values ​​for the equivalent displacement threshold parameter can be given accordingly.

[0036] Step 4: Obtain the proton energy spectrum under arbitrary space orbits and divide the on-orbit accumulation time of the silicon carbide power device into multiple time periods. For each time period, dynamically update the proton energy spectrum according to the changes in orbital parameters. When the changes in orbital parameters exceed a preset orbital parameter change threshold, the dynamic update of the proton energy spectrum is triggered. Based on the dynamically updated proton energy spectrum and according to the correspondence established in Step 3, calculate the epitaxial layer non-ionization energy damage accumulated by the silicon carbide power device in orbit. By combining the device-level correspondence of "incident proton energy - epitaxial layer non-ionization energy damage" with the mission-level "proton energy spectrum under arbitrary space orbits," a cumulative damage assessment oriented towards the actual on-orbit environment can be achieved. By dividing the on-orbit accumulation time into multiple time periods, the changes in proton irradiation conditions caused by the changes in orbital environment over time can be expressed in segments, avoiding the deviation caused by simply equating the time-varying energy spectrum to a single fixed energy spectrum. For each time period, the proton spectrum is dynamically updated based on changes in orbital parameters, with a preset orbital parameter change threshold used as a trigger condition. This allows for control of the update frequency while ensuring assessment accuracy, synchronizing the proton spectrum update with key changes in the orbital environment. Finally, based on the dynamically updated proton spectrum and the correspondence established in step three, calculations are performed to convert the proton spectrum contribution for each time period into the contribution of epitaxial layer non-ionizing energy damage. This accumulated non-ionizing energy damage to the epitaxial layer of silicon carbide power devices over time provides a quantitative basis for lifetime assessment, hardening design, and mission risk assessment.

[0037] In one specific implementation, to obtain the proton energy spectrum under arbitrary space orbits and calculate the on-orbit accumulation of non-ionizing energy damage to the epitaxial layer of silicon carbide power devices, the following steps are performed: The proton energy spectrum is obtained under arbitrary space orbits, and the on-orbit accumulation time of silicon carbide power devices is divided into multiple time periods. For each time period, the proton energy spectrum is dynamically updated according to changes in orbital parameters. When the changes in orbital parameters exceed a preset threshold, the dynamic update of the proton energy spectrum is triggered. The changes in orbital parameters can be characterized by any or a combination of parameters such as orbital altitude, orbital inclination, orbital eccentricity, right ascension of the ascending node, argument of perigee, and local time. The setting rule for the preset threshold of orbital parameter changes can be determined based on the sensitivity of orbital parameter changes to the proton energy spectrum, combined with existing technologies. That is, in the mission design phase, representative orbital parameter perturbations are selected based on commonly used space radiation environment modeling and orbit prediction methods, and the corresponding perturbations are calculated respectively. The proton energy spectrum changes, and then the minimum orbital parameter change that keeps the proton energy spectrum change within an acceptable range over the same time period is selected as the preset orbital parameter change threshold. This ensures that the proton energy spectrum is reflected in a timely manner as the orbit evolves, while avoiding excessive updates that would increase computational costs. For example, in the scenario where the orbital altitude of a low Earth orbit gradually decreases under the influence of aerodynamic drag, the proton energy spectrum at different orbital altitudes can be obtained using the orbital propagation results and its differences can be compared. Then, the orbital altitude increment with significant energy spectrum differences can be mapped to the preset orbital parameter change threshold. When the cumulative change in orbital altitude exceeds this threshold, the dynamic update of the proton energy spectrum is triggered, so that the proton energy spectrum used in each time period is consistent with the orbital parameters in that time period. In one implementation, the requirement that "the proton energy spectrum variation remains within an acceptable range within the same time period" is quantified using a spectral difference metric. Let the proton energy spectrum used in the previous time period be... The candidate updated energy spectrum is Then define the relative difference index of the energy spectrum. for in and This represents the boundary of the target energy range of the proton energy spectrum. When Not exceeding the preset difference threshold When the energy spectrum change is considered to be within an acceptable range, the energy spectrum from the previous time period is used; when Exceeding the preset difference threshold At that time, the proton energy spectrum is dynamically updated.

[0038] Preset difference threshold The threshold can be determined based on a trade-off between the allowable error and computational cost in the task evaluation; in one example, several candidate difference thresholds can be selected first. Sensitivity tests were performed to calculate the variation range of the accumulated non-ionizing energy damage to the epitaxial layer in orbit, and the maximum difference threshold that ensures the variation range does not exceed the preset error target was selected as the threshold. This reduces the update frequency while maintaining accuracy.

[0039] Based on the dynamically updated proton energy spectrum and according to the correspondence established in step three, the on-orbit accumulated epitaxial layer non-ionization energy damage of the silicon carbide power device is calculated. In step four, the proton energy spectrum components with incident proton energies not lower than the proton energy threshold are integrated according to the correspondence established in step three to obtain the epitaxial layer non-ionization energy damage for the corresponding time period. The epitaxial layer non-ionization energy damage for each time period is then accumulated to obtain the on-orbit accumulated epitaxial layer non-ionization energy damage of the silicon carbide power device. Specifically, within each time period, the energy components of the proton energy spectrum are used as the integration object. Energy components with incident proton energies lower than the proton energy threshold are not included in the integration. For energy components with incident proton energies not lower than the proton energy threshold, the correspondence established in step three is called on an energy interval basis to obtain the epitaxial layer non-ionization energy damage contribution corresponding to that energy component, and then... The duration of the time period is used to weight the contribution over time to obtain the integral result for that time period. Finally, the integral results of each time period are summed to obtain the on-orbit cumulative result. For example, the on-orbit cumulative time is divided into multiple time periods based on weeks. If the change in orbital parameters in a certain week does not exceed the preset orbital parameter change threshold, the proton energy spectrum of the previous week is used. If the change in orbital parameters in a certain week exceeds the preset orbital parameter change threshold, the proton energy spectrum corresponding to that week is updated. Only the components of the updated proton energy spectrum where the incident proton energy is not lower than the proton energy threshold are integrated to obtain the epitaxial non-ionization energy damage for that week. Then, the epitaxial non-ionization energy damage of all weeks is summed to obtain the epitaxial non-ionization energy damage of silicon carbide power devices accumulated on-orbit. This allows the cumulative result to simultaneously reflect the effective contribution constraints brought about by the change of orbital environment over time and the screening of incident proton energy threshold.

[0040] Energy spectrum integration and time-weighted approach: In one embodiment, the proton energy spectrum is represented by flux per unit area, per unit time, and per unit energy interval. For each time interval, the flux is multiplied by the duration of that time interval to obtain the differential energy flux for that time interval. Then, for each energy interval where the incident proton energy is not lower than the proton energy threshold, the corresponding relationship is called segment by segment to obtain the contribution of the epitaxial layer to non-ionization energy damage and complete the integration. Energy intervals below the proton energy threshold are not included in the integration to ensure that the energy spectrum integration result is consistent with the geometric reach determination boundary of the epitaxial layer.

[0041] In one implementation, the energy spectrum integration is achieved using a discrete energy grid: the target energy range is divided into several energy intervals, and the energy at the center or boundary of each interval is taken as a representative value. The correspondence established in step three is then used to obtain the epitaxial non-ionization energy damage value of that energy interval. This value is then multiplied by the differential flux of that energy interval (flux multiplied by time and then by the width of the energy interval) and summed over each energy interval to obtain the epitaxial non-ionization energy damage for that time period. In one example, the trapezoidal method or the piecewise constant method can be used to complete the numerical integration, thus making the integration calculation have clear and feasible steps.

[0042] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices, characterized in that, Performed on an electronic computing device, including the following steps: Step 1: Perform reverse analysis on the silicon carbide power device and establish a silicon carbide power device model based on the reverse analysis results; determine the proton energy threshold; when the incident proton energy is lower than the proton energy threshold, the non-ionization energy damage to the epitaxial layer caused by the incident proton is not included; Step 2: When the incident proton energy is not lower than the proton energy threshold, the Monte Carlo method is used to simulate the interaction between the incident proton and the material in the silicon carbide power device model to obtain the feature set of secondary heavy ions generated by the incident proton in the epitaxial layer of the silicon carbide power device. Step 3: For multiple incident proton energy points where the incident proton energy is not lower than the proton energy threshold, calculate the non-ionization energy damage of the epitaxial layer of the silicon carbide power device based on the secondary heavy ion feature set, and establish the correspondence between the incident proton energy and the non-ionization energy damage of the epitaxial layer. Step 4: Obtain the proton energy spectrum under any space orbit and divide the on-orbit accumulation time of the silicon carbide power device into multiple time periods; for each time period, dynamically update the proton energy spectrum according to the changes in orbital parameters. When the changes in orbital parameters exceed the preset orbital parameter change threshold, the dynamic update of the proton energy spectrum is triggered; based on the dynamically updated proton energy spectrum and according to the correspondence established in Step 3, calculate the epitaxial non-ionization energy damage suffered by the silicon carbide power device during on-orbit accumulation.

2. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 1, characterized in that, The reverse analysis includes: preparing cross-sectional samples of silicon carbide power devices to obtain the layer structure sequence, measuring the layer thickness using microscopic imaging, obtaining the elemental composition of each layer using energy dispersive spectroscopy, and determining the material density and doping parameters in conjunction with device process information; and generating a silicon carbide power device model based on the layer thickness, elemental composition, material density, and doping parameters.

3. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 1, characterized in that, The proton energy threshold is the minimum incident proton energy required for the incident proton range to reach the upper surface of the epitaxial layer of the silicon carbide power device. The determination of the proton energy threshold includes: simulating the range of different incident proton energies in the silicon carbide power device model using the Monte Carlo method, and determining the proton energy threshold based on the condition that the incident proton range reaches the upper surface of the epitaxial layer of the silicon carbide power device.

4. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 1, characterized in that, The set of secondary heavy ion characteristics includes: the types of secondary heavy ions produced by incident protons, the energy of secondary heavy ions, and the number of secondary heavy ions.

5. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 1, characterized in that, The Monte Carlo method was used to simulate the interaction between incident protons and materials in a silicon carbide power device model and obtain a secondary heavy ion feature set. This included: setting initial conditions for the incident protons based on the silicon carbide power device model, including at least the incident proton energy, incident direction, and incident position; randomly sampling the particle transport, nuclear reaction, and scattering processes of the incident protons in the silicon carbide power device model and tracking the secondary heavy ions generated by the incident protons in the model; setting the epitaxial layer as a scoring region, statistically analyzing the events of secondary heavy ions entering or generated within the scoring region, and recording the characteristic parameters corresponding to the secondary heavy ion feature set; and generating the secondary heavy ion feature set based on the event statistics and the recorded characteristic parameters.

6. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 1, characterized in that, Step 3 calculates the epitaxial layer non-ionizing energy damage based on the secondary heavy ion feature set, including: grouping the secondary heavy ion feature set according to the type and energy of secondary heavy ions to obtain the number and energy representative value of secondary heavy ions in each group; calling a pre-established nuclear damage mapping function for each group to convert the energy representative value into a nuclear damage energy share; weighting and summing the nuclear damage energy shares of each group based on the number of secondary heavy ions in each group to obtain the total nuclear damage energy of the epitaxial layer; and normalizing the total nuclear damage energy of the epitaxial layer according to the epitaxial layer mass to obtain the non-ionizing energy damage of the epitaxial layer.

7. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 6, characterized in that, The nuclear damage mapping function is constrained by a preset displacement threshold parameter of the epitaxial layer material, and a smooth transition is adopted within the energy transition interval centered on the preset displacement threshold parameter and determined by a preset energy bandwidth. When weighting and accumulating the nuclear damage energy share, a robust weight based on the probability of group occurrence is introduced to suppress the influence of low-probability group occurrences on the non-ionization energy damage of the epitaxial layer.

8. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 1, characterized in that, The establishment of the correspondence between incident proton energy and epitaxial layer non-ionization energy damage includes: determining the incident proton energy point set, which covers the target incident proton energy range above the proton energy threshold and satisfies a preset energy step size or preset energy grid; for each incident proton energy point in the incident proton energy point set, performing step two to obtain the secondary heavy ion feature set and performing step three to obtain the corresponding epitaxial layer non-ionization energy damage value; forming a discrete correspondence dataset by combining the incident proton energy point set and the epitaxial layer non-ionization energy damage value set; and generating a mapping model based on the discrete correspondence dataset and storing it as a correspondence.

9. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 8, characterized in that, The mapping model can be any of the following: lookup table mapping, piecewise interpolation mapping, or regression fitting mapping, so that the non-ionization energy damage value of the epitaxial layer can be obtained through the correspondence for any incident proton energy within the target incident proton energy range.

10. The method for assessing the broad-spectrum proton non-ionization energy damage of silicon carbide power devices according to claim 1, characterized in that, In step four, the proton energy spectrum components with incident proton energy not lower than the proton energy threshold are integrated according to the correspondence established in step three to obtain the epitaxial non-ionization energy damage for the corresponding time period. The epitaxial non-ionization energy damage for each time period is then accumulated to obtain the epitaxial non-ionization energy damage accumulated on-orbit by the silicon carbide power device.

Images