SiC MOS dynamic characteristics prediction method, system and medium

Abstract

This invention discloses a method, system, and medium for predicting the dynamic characteristics of SiC MOS devices, relating to the field of characteristic prediction technology. The method includes: collecting multi-dimensional feature information of SiC MOS devices and constructing a SiC MOS coupling model based on the multi-dimensional feature information; obtaining simulation records through the SiC MOS coupling model and introducing a predetermined verification strategy to verify and analyze the simulation records to obtain simulation verification results; when the simulation verification results meet predetermined simulation constraints, retrieving the SiC MOS coupling model to simulate and obtain the predicted dynamic characteristics of the SiC MOS device. This invention solves the technical problem that existing technologies cannot effectively predict the dynamic characteristics of SiC MOS devices before switching, achieving accurate prediction of dynamic characteristics by constructing a coupling model and performing simulation verification analysis.

Description

Technical Field

[0001] This invention relates to the field of characteristic prediction technology, specifically to a method, system, and dielectric for predicting the dynamic characteristics of SiC MOS. Background Technology

[0002] SiC MOS devices exhibit excellent performance in high-speed and high-frequency applications, and their dynamic characteristics directly affect system efficiency and device reliability. In practical applications, the switching process is influenced by a combination of factors, including drive parameters, device structure, and operating environment, resulting in complex and variable dynamic behavior. Due to the lack of accurate means to predict device behavior before switching, drivers often rely on experience to adjust drive parameters, which is difficult to adjust in a timely manner to adapt to different operating conditions. This leads to increased switching losses, device instability, and even shortened lifespan. Summary of the Invention

[0003] This application provides a method, system, and dielectric for predicting the dynamic characteristics of SiC MOS devices, which addresses the technical problem that existing technologies cannot effectively predict the dynamic characteristics of SiC MOS devices before switching.

[0004] In view of the above problems, this application provides a method, system and dielectric for predicting the dynamic characteristics of SiC MOS.

[0005] The first aspect of this application provides a method for predicting the dynamic characteristics of SiC MOS, the method comprising:

[0006] Multidimensional feature information of SiC MOS devices is collected, and a SiC MOS coupling model is constructed based on the multidimensional feature information. Simulation records are obtained through the SiC MOS coupling model, and a predetermined verification strategy is introduced to verify and analyze the simulation records to obtain simulation verification results. When the simulation verification results meet predetermined simulation constraints, the SiC MOS coupling model is invoked to simulate the predicted dynamic characteristics of the SiC MOS device. The simulation records include gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data. The MOS coupling model generates a simulation record, and a predetermined verification strategy is introduced to verify and analyze the simulation record to obtain simulation verification results. This includes: extracting a static verification plan from the predetermined verification strategy; measuring static verification information according to the static verification plan, wherein the static verification information includes multiple sets of charging characteristics with temperature indicators, and each set of charging characteristics with temperature indicators includes transfer characteristics and output characteristics; extracting a dynamic verification plan from the predetermined verification strategy; measuring dynamic verification information according to the dynamic verification plan, wherein the dynamic verification information includes switching characteristics and reverse recovery characteristics; and analyzing the transfer characteristics and output characteristics, the switching characteristics and the reverse recovery characteristics to obtain the simulation verification results.

[0007] In one possible implementation, multidimensional feature information of a SiC MOS device is collected, and a SiC MOS coupling model is constructed based on the multidimensional feature information. This includes: constructing a structural parameter model based on the geometric parameters in the multidimensional feature information; constructing a gate oxide characteristic model based on oxide traps in the multidimensional feature information; and constructing a temperature-dependent model based on temperature characteristic parameters in the multidimensional feature information. The geometric parameters include trench depth, trench width, and trench spacing; the oxide traps include interface electronic defects and internal electronic traps in the oxide layer; and the temperature characteristic parameters include mobility temperature coefficient, carrier lifetime temperature coefficient, and saturation velocity temperature coefficient. The SiC MOS coupling model is obtained by combining the structural parameter model, the gate oxide characteristic model, and the temperature-dependent model.

[0008] In one possible implementation, after analyzing the transfer characteristics, output characteristics, switching characteristics, and reverse recovery characteristics to obtain the simulation verification results, the method further includes: extracting a temperature verification plan from the predetermined verification strategy; obtaining an arbitrary temperature according to the temperature verification plan, and matching the arbitrary charging characteristic corresponding to the arbitrary temperature among the multiple sets of charging characteristics with temperature identifiers; matching the arbitrary charging simulation data corresponding to the arbitrary temperature in the gate charging simulation data; comparing the arbitrary charging characteristic with the arbitrary charging simulation data to obtain a comparison result; taking the mean of the comparison result and performing normalization processing to obtain a temperature characteristic coefficient; and calibrating and adjusting the simulation verification results using the temperature characteristic coefficient as a weight.

[0009] In one possible implementation, comparing the arbitrary charging characteristic with the arbitrary charging simulation data to obtain a comparison result includes: comparing the arbitrary transfer characteristic in the arbitrary charging characteristic with the arbitrary simulated transfer characteristic in the arbitrary charging simulation data, and obtaining the arbitrary transfer characteristic deviation; comparing the arbitrary output characteristic in the arbitrary charging characteristic with the arbitrary simulated output characteristic in the arbitrary charging simulation data, and obtaining the arbitrary output characteristic deviation; performing a weighted calculation on the standardized arbitrary transfer characteristic deviation and the arbitrary output characteristic deviation to obtain an arbitrary deviation index; and using the arbitrary deviation index as the comparison result.

[0010] In one possible implementation, when the simulation verification result meets predetermined simulation constraints, the SiC MOS coupling model is invoked to simulate the predicted dynamic characteristics of the SiC MOS device. This includes: collecting operating condition information of the SiC MOS device; using the operating condition information as simulation constraints, simulating the predicted dynamic characteristics through the SiC MOS coupling model; wherein the operating condition information includes electrical conditions, environmental conditions, and circuit topology conditions; the electrical conditions include the gate drive voltage range, operating frequency range, and load current range; the environmental conditions include the operating temperature range, heat dissipation conditions, and ambient humidity; and the circuit topology conditions include the application circuit topology, parasitic parameters, and drive circuit parameters.

[0011] In one possible implementation, the method further includes: performing a timing analysis on the predicted dynamic characteristics to obtain a predicted timing; performing a polynomial regression fitting analysis on the predicted timing to obtain a predicted fitting formula; reading a predetermined dynamic characteristic threshold and combining it with the predicted fitting formula to obtain the dynamic characteristic lifetime of the SiC MOS device, denoted as the predicted lifetime; wherein the predicted lifetime is used to characterize the predicted lifetime value of the SiC MOS device.

[0012] In one possible implementation, performing a multinomial regression fitting analysis on the predicted time series to obtain a predicted fitting formula includes: obtaining the prediction time corresponding to the predicted dynamic characteristics; forming the predicted time series based on the correspondence between the predicted dynamic characteristics and the prediction time, and generating a predicted scatter plot; generating a prediction curve of the predicted scatter plot based on the principle of random sampling consistency, and performing a multinomial regression fitting analysis on the prediction curve to obtain the predicted fitting formula.

[0013] A second aspect of this application provides a SiC MOS dynamic characteristic prediction system, the system comprising:

[0014] The information collection module is used to collect multi-dimensional feature information of the SiC MOS device and construct a SiC MOS coupling model based on the multi-dimensional feature information; the verification and analysis module is used to obtain simulation records through the SiC MOS coupling model and to perform verification and analysis on the simulation records by introducing a predetermined verification strategy to obtain simulation verification results; the prediction module is used to retrieve the SiC MOS coupling model to simulate and obtain the predicted dynamic characteristics of the SiC MOS device when the simulation verification results meet the predetermined simulation constraints.

[0015] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the SiC MOS dynamic characteristic prediction method provided in this application.

[0016] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0017] This application collects multidimensional feature information of SiC MOS devices and constructs a SiCMOS coupling model based on this information. Simulation records are obtained through the SiC MOS coupling model, and a predetermined verification strategy is introduced to verify and analyze these records, yielding simulation verification results. When the simulation verification results meet predetermined simulation constraints, the SiC MOS coupling model is invoked to simulate and obtain the predicted dynamic characteristics of the SiC MOS device. This invention solves the technical problem that existing technologies cannot effectively predict the dynamic characteristics of SiC MOS devices before switching. By constructing a coupling model and performing simulation verification analysis, it achieves the technical effect of accurately predicting dynamic characteristics. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart of the SiC MOS dynamic characteristic prediction method provided in the embodiments of this application;

[0020] Figure 2 This is a schematic diagram of the SiC MOS dynamic characteristic prediction system provided in an embodiment of this application.

[0021] Explanation of reference numerals in the attached diagram: Information collection module 11, verification and analysis module 12, prediction module 13. Detailed Implementation

[0022] This application provides a method, system, and dielectric for predicting the dynamic characteristics of SiC MOS devices. It addresses the technical problem that existing technologies cannot effectively predict the dynamic characteristics of SiC MOS devices before switching. By constructing a coupled model and conducting simulation verification analysis, it achieves the technical effect of accurately predicting dynamic characteristics.

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0024] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0025] Example 1, as Figure 1 As shown, this application provides a method for predicting the dynamic characteristics of SiC MOS, the method comprising:

[0026] Step S100: Collect multidimensional feature information of SiC MOS device and construct SiC MOS coupling model based on the multidimensional feature information.

[0027] In this embodiment, multidimensional feature information of SiC MOS devices is retrieved from a pre-set database. This information is based on a large amount of device measurement and simulation results and covers the core physical parameters that affect the dynamic behavior of the devices. The multidimensional feature information includes geometric parameters reflecting the device geometry, such as trench depth, trench width, and trench spacing; oxide trap data describing gate oxide characteristics and defect distribution, including interface electronic defects and internal electronic traps in the oxide layer; and temperature characteristic parameters related to temperature response, such as mobility temperature coefficient, carrier lifetime temperature coefficient, and saturation velocity temperature coefficient.

[0028] Based on the above multidimensional feature information, structural parameter models, gate oxide layer characteristic models, and temperature-related models are constructed respectively. Through collaborative modeling, the three are integrated into a unified SiC MOS coupling model.

[0029] Furthermore, the method provided in the application embodiments, which involves collecting multi-dimensional feature information of the SiC MOS device and constructing a SiC MOS coupling model based on the multi-dimensional feature information, further includes:

[0030] A structural parameter model is constructed based on the geometric parameters in the multidimensional feature information; a gate oxide layer characteristic model is constructed based on the oxide layer traps in the multidimensional feature information; a temperature-dependent model is constructed based on the temperature characteristic parameters in the multidimensional feature information; wherein, the geometric parameters include trench depth, trench width, and trench spacing; the oxide layer traps include interface electronic defects and internal electronic traps in the oxide layer; and the temperature characteristic parameters include mobility temperature coefficient, carrier lifetime temperature coefficient, and saturation velocity temperature coefficient; the SiC MOS coupling model is obtained by combining the structural parameter model, the gate oxide layer characteristic model, and the temperature-dependent model.

[0031] In this embodiment, geometric parameter data, including trench depth, trench width, and trench spacing, are first extracted from a pre-defined database. Trench depth represents the longitudinal dimension of the trench structure in the device, affecting the electric field distribution and the length of the conductive channel. Trench width refers to the physical width of the trench in the lateral direction, determining the effective area of ​​the carrier channel; trench spacing reflects the distance between adjacent trenches, affecting the overall layout density and capacitance characteristics of the device. Based on this geometric data, a structural model of the device is constructed using numerical modeling methods. A three-dimensional mesh and boundary conditions consistent with the actual structure are generated through a simulation platform, thereby obtaining a device structural parameter model that can be used for electric field simulation.

[0032] Subsequently, data related to oxide traps were retrieved, including interface electronic defects and internal electronic traps in the oxide layer. Interface electronic defects describe unstable charge states at the interface between the gate oxide layer and the semiconductor material, which affect the device's controllability and stability. Internal electronic traps refer to localized charge trapping centers within the oxide layer material, affecting the device's leakage current and long-term reliability. By incorporating this defect data, a physical model reflecting defect behavior was constructed to simulate the impact of defects on the device's electrical characteristics, especially threshold voltage changes, ultimately yielding a gate oxide characteristic model describing the device's defect response characteristics.

[0033] Next, temperature characteristic parameters are obtained, including the mobility temperature coefficient, carrier lifetime temperature coefficient, and saturation velocity temperature coefficient. The mobility temperature coefficient reflects the rate at which the mobility of charge carriers in the material decreases with increasing temperature; the carrier lifetime temperature coefficient represents the change in the average residence time of charge carriers at different temperatures, used to assess the impact of temperature on carrier recombination behavior; the saturation velocity temperature coefficient represents the degree of influence of temperature when charge carriers reach their maximum drift velocity under high electric field conditions. Based on these parameters, a temperature response model is constructed to characterize the conductivity and response time of the device at different temperatures, ultimately forming a temperature-dependent model.

[0034] Finally, the structural parameter model, gate oxide characteristic model, and temperature-related model are integrated and a physical coupling relationship is established through a unified modeling platform, so that the models can transmit influence to each other during the simulation process, resulting in a SiC MOS coupled model that comprehensively considers structural, electric field, defect, and temperature factors.

[0035] Step S200: Obtain the simulation record through the SiC MOS coupling model, and introduce a predetermined verification strategy to verify and analyze the simulation record to obtain the simulation verification result.

[0036] In this embodiment, simulation records are acquired under pre-defined simulation conditions using a SiC MOS coupling model. These records include gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data. To verify the accuracy of the simulation records, a predetermined verification strategy is introduced for analysis. This strategy includes a static verification plan and a dynamic verification plan. The static verification plan acquires multiple sets of charging characteristics with temperature indicators, each set including transfer and output characteristics. The dynamic verification plan acquires switching and reverse recovery characteristics. Finally, by analyzing the transfer, output, switching, and reverse recovery characteristics, and comparing the simulation data with actual measurement results, the simulation verification results are obtained.

[0037] Furthermore, in the method provided in the application embodiments, the simulation record includes gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data. The simulation record is obtained through the SiC MOS coupling model, and a predetermined verification strategy is introduced to verify and analyze the simulation record to obtain simulation verification results. The method further includes:

[0038] Extract the static verification plan from the predetermined verification strategy; measure static verification information according to the static verification plan, wherein the static verification information includes multiple sets of charging characteristics with temperature labels, and each set of charging characteristics with temperature labels includes transfer characteristics and output characteristics; extract the dynamic verification plan from the predetermined verification strategy; measure dynamic verification information according to the dynamic verification plan, wherein the dynamic verification information includes switching characteristics and reverse recovery characteristics; analyze the transfer characteristics and the output characteristics, the switching characteristics and the reverse recovery characteristics to obtain the simulation verification results.

[0039] In this embodiment, after generating gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data using the SiC MOS coupling model, a static verification plan is first extracted from the predetermined verification strategy to verify the accuracy of the simulation records. The static verification plan is used to set the parameter boundaries during the static simulation process, including temperature point settings, voltage scanning methods, and output variable extraction methods. According to the static verification plan, multiple simulation temperature points are set in the SiC MOS coupling model. Static verification information is obtained by measuring the gate-source voltage scan and drain-source voltage scan, respectively. The static verification information is a set of data output from the SiC MOS coupling model at each temperature point, including multiple sets of charging characteristics with temperature labels. Each set of charging characteristics with temperature labels corresponds to a specific temperature value, and each set of charging characteristics includes transfer characteristics and output characteristics. The transfer characteristics are the drain current response curve caused by changes in gate-source voltage, used to reflect the device's turn-on threshold and gate control efficiency; the output characteristics are the drain current output curve caused by changes in drain-source voltage, used to reflect the device's conduction capability and current carrying capacity. This step ultimately yields static verification information, including transfer and output characteristics, under multiple temperature conditions.

[0040] Subsequently, the dynamic verification plan is extracted from the predetermined verification strategy. The dynamic verification plan is used to set simulation parameters such as pulse excitation, load model, topology configuration, and time nodes required for time-domain simulation under dynamic operating conditions. Based on the dynamic verification plan, a switching action simulation scenario is constructed in the SiC MOS coupled model, and a driving signal is applied for timing simulation to obtain dynamic verification information. The dynamic verification information consists of the response data output by the SiC MOS coupled model under dynamic conditions, including switching characteristics and reverse recovery characteristics. Switching characteristics are the changes in drain voltage and current over time output by the SiC MOS coupled model during conduction and turn-off, used to extract indicators such as turn-on time, turn-off time, voltage rise time, and current fall time; reverse recovery characteristics are the conduction and cut-off processes of the parasitic diode or body diode after the current direction reverses, mainly extracting reverse recovery time and reverse recovery charge. Through this step, dynamic verification information containing switching characteristics and reverse recovery characteristics is obtained.

[0041] Finally, the transfer characteristics, output characteristics, switching characteristics, and reverse recovery characteristics were analyzed. The characteristic curves from the static and dynamic verification information were compared one by one with the gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data generated by the SiC MOS coupling model. During the comparison, a key point comparison method was used to extract key physical parameters such as the threshold voltage in the transfer characteristics, the conduction region slope in the output characteristics, the peak current in the switching characteristics, and the inflection point voltage in the reverse recovery characteristics, and then performed numerical comparisons. For example, in the transfer characteristics, if the threshold voltage output by the SiC MOS coupling model is 2.5V, and the corresponding value in the static verification information is 2.55V, then the deviation is 0.05V. If the tolerance range is set to ±0.1V, then the result is considered to meet the requirements. In the switching characteristics, if the peak current output by the SiC MOS coupling model is 19.1A, and the corresponding value in the dynamic verification information is 19.3A, then the deviation is 0.2A, which is also within the ±0.5A tolerance range. This method is used to perform parameter alignment and difference evaluation on various transfer characteristics, output characteristics, switching characteristics, and reverse recovery characteristics, resulting in simulation verification results. The simulation verification results are verification outputs containing the deviation values ​​of each key physical parameter, used to characterize the prediction accuracy and consistency of the SiCMOS coupling model for device behavior under various simulation conditions.

[0042] Furthermore, in the method provided in the application embodiments, after analyzing the transfer characteristics, the output characteristics, the switching characteristics, and the reverse recovery characteristics to obtain the simulation verification results, the method further includes:

[0043] Extract the temperature verification plan from the predetermined verification strategy; according to the temperature verification plan, obtain an arbitrary temperature and match the arbitrary charging characteristic corresponding to the arbitrary temperature in the multiple sets of charging characteristics with temperature labels; match the arbitrary charging simulation data corresponding to the arbitrary temperature in the gate charging simulation data; compare the arbitrary charging characteristic with the arbitrary charging simulation data to obtain a comparison result; take the mean of the comparison result and perform normalization processing to obtain the temperature characteristic coefficient; use the temperature characteristic coefficient as a weight to calibrate and adjust the simulation verification result.

[0044] In this embodiment, a temperature verification plan is first extracted from the predetermined verification strategy. This temperature verification plan sets out a temperature-related calibration process, including a temperature point extraction method, a data matching method, difference quantification rules, and a correction process.

[0045] Subsequently, according to the temperature verification plan, an arbitrary temperature is obtained, and the arbitrary charging characteristic corresponding to the arbitrary temperature is matched among multiple sets of charging characteristics with temperature labels. The arbitrary temperature can be set by the user or determined by the simulation environment, for example, set to 80℃. The characteristic data set closest to this temperature is searched among the multiple sets of charging characteristics with temperature labels. If two temperature points, 75℃ and 85℃, exist, linear interpolation can be performed to generate the transfer and output characteristics corresponding to 80℃, thus forming the arbitrary charging characteristic.

[0046] Next, arbitrary charging simulation data corresponding to any temperature is matched in the gate charging simulation data. This matching is based on temperature indexing, extracting the simulation results corresponding to 80℃ from the gate charging simulation data of the SiC MOS coupled model, including the numerical output of transfer characteristics and output characteristics under simulation conditions, to ensure that the simulation data and verification data are aligned in the temperature dimension.

[0047] Next, the arbitrary charging characteristics are compared with arbitrary charging simulation data. In this process, the transfer characteristics in the arbitrary charging characteristics are first compared with the simulated transfer characteristics in the arbitrary charging simulation data to obtain the transfer characteristic deviation. Then, the output characteristics are processed in the same way to obtain the output characteristic deviation. The two deviation values ​​are standardized and weighted for a weighted calculation, forming a comprehensive value as the comparison result.

[0048] Next, the mean of the comparison results is taken and normalized to obtain the temperature characteristic coefficient. The purpose of normalization is to convert the deviation value into a standard scale, reflecting its position within the maximum allowable deviation range. For example, if the mean of the comparison results is 0.05, and the maximum acceptable error is set to 0.2, then the normalized temperature characteristic coefficient is 1 - (0.05 / 0.2) = 0.75. This temperature characteristic coefficient is a dimensionless value used to quantify the simulation fit of the model at that temperature point; the closer it is to 1, the better the simulation effect.

[0049] Finally, the simulation verification results are calibrated and adjusted using the temperature characteristic coefficient as a weight. This is achieved by multiplying the deviation values ​​in the original simulation verification results by the temperature characteristic coefficient, thus correcting the overall verification error for temperature. For example, if the original deviation is ±6% and the temperature characteristic coefficient is 0.75, the corrected deviation will be ±4.5%.

[0050] Furthermore, in the method provided in the application embodiments, comparing the arbitrary charging characteristics with the arbitrary charging simulation data to obtain a comparison result further includes:

[0051] The arbitrary transfer characteristic in the arbitrary charging characteristic is compared with the arbitrary simulated transfer characteristic in the arbitrary charging simulation data to obtain the arbitrary transfer characteristic deviation; the arbitrary output characteristic in the arbitrary charging characteristic is compared with the arbitrary simulated output characteristic in the arbitrary charging simulation data to obtain the arbitrary output characteristic deviation; the arbitrary transfer characteristic deviation and the arbitrary output characteristic deviation after standardization are weighted and calculated to obtain the arbitrary deviation index; the arbitrary deviation index is used as the comparison result.

[0052] In this embodiment, the arbitrary transfer characteristic in the arbitrary charging characteristic is first compared with the arbitrary simulated transfer characteristic in the arbitrary charging simulation data to obtain the arbitrary transfer characteristic deviation. This step uses the threshold voltage comparison method, that is, extracting the voltage value at which the gate voltage starts to conduct in both the arbitrary charging characteristic and the arbitrary charging simulation data, and calculating the difference as the transfer characteristic deviation. For example, if the threshold voltage in the verification data is 2.5V and the simulation result is 2.2V, then the transfer characteristic deviation is 0.3V.

[0053] Next, the arbitrary output characteristic from the arbitrary charging characteristics is compared with the arbitrary simulated output characteristic from the arbitrary charging simulation data to obtain the arbitrary output characteristic deviation. This step uses the saturation current deviation method, which extracts the maximum drain current value of the two sets of output characteristic curves when the drain voltage tends to stabilize, and calculates the difference as the output characteristic deviation. For example, if the measured maximum drain current is 18A and the simulation result is 16.5A, then the output characteristic deviation is 1.5A.

[0054] Then, the arbitrary transfer characteristic deviations and arbitrary output characteristic deviations are standardized. This step uses the normalized maximum method, which divides each deviation value by its set maximum tolerance. For example, if the transfer characteristic tolerance is 0.5V and the output characteristic tolerance is 3A, then the above deviations are normalized to 0.6 and 0.5, respectively. This step ultimately yields standardized transfer characteristic deviations and standardized output characteristic deviations, used to unify dimensions and scales.

[0055] Subsequently, the arbitrary transfer characteristic deviation and arbitrary output characteristic deviation after standardization are weighted and calculated to obtain the arbitrary deviation index. This step uses a fixed-weighting method; for example, the weight of the transfer characteristic is set to 0.6 and the weight of the output characteristic is set to 0.4. The weighted sum is calculated using a linear weighting method: 0.6 × 0.6 + 0.4 × 0.5 = 0.56. This step ultimately yields the arbitrary deviation index, which reflects the overall degree of difference between the two characteristics.

[0056] Finally, the arbitrary deviation index is used as the comparison result.

[0057] Step S300: When the simulation verification result meets the predetermined simulation constraints, the SiC MOS coupling model is retrieved to simulate and obtain the predicted dynamic characteristics of the SiC MOS device.

[0058] In this embodiment, before making a prediction, the simulation verification results are first judged according to predetermined simulation constraints. The predetermined simulation constraints are a set of criteria for determining whether the model accuracy meets the standards, including indicators such as static verification deviation threshold, dynamic verification deviation threshold, and lower limit of temperature characteristic coefficient. When the static verification deviation and dynamic verification deviation in the simulation verification results are both lower than the corresponding set thresholds, and the temperature characteristic coefficient is higher than the set reliability lower limit, it is determined that the predetermined simulation constraints are met.

[0059] After meeting the above criteria, the SiC MOS coupling model is retrieved for predictive simulation. This process first collects the operating condition information of the SiC MOS device as input constraints for the model simulation. The operating condition information includes electrical conditions, environmental conditions, and circuit topology conditions. The electrical conditions cover the gate drive voltage range, operating frequency range, and load current range. The environmental conditions involve the operating temperature range, heat dissipation conditions, and ambient humidity. The circuit topology conditions include the application circuit topology, parasitic parameters, and drive circuit parameters.

[0060] After inputting the above operating condition information into the SiC MOS coupling model, a joint simulation is performed based on the integrated structural parameter model, gate oxide characteristic model, and temperature-related model. The final output is the predicted dynamic characteristics of the SiC MOS device under the target operating conditions. These predicted dynamic characteristics include gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data, used to reflect the dynamic response behavior of the device in specific application scenarios.

[0061] Furthermore, in the method provided in the application embodiments, when the simulation verification result meets the predetermined simulation constraints, the SiC MOS coupling model is invoked to simulate and obtain the predicted dynamic characteristics of the SiC MOS device, which further includes:

[0062] The operating condition information of the SiC MOS device is collected; using the operating condition information as simulation constraints, the predicted dynamic characteristics are simulated through the SiC MOS coupling model; wherein, the operating condition information includes electrical conditions, environmental conditions, and circuit topology conditions; the electrical conditions include the gate drive voltage range, operating frequency range, and load current range; the environmental conditions include the operating temperature range, heat dissipation conditions, and ambient humidity; and the circuit topology conditions include the application circuit topology, parasitic parameters, and drive circuit parameters.

[0063] In this embodiment, after the SiC MOS coupling model has been verified and meets predetermined simulation constraints, operating condition information of the SiC MOS device is collected to achieve simulation calculations for predicting dynamic characteristics. The operating condition information is a pre-defined set of parameters configured according to typical application scenarios, device design specifications, or system-level requirements. This information is directly derived from design specification documents, reference circuit manuals, or simulation templates. The operating condition information includes three types of parameters: electrical conditions, environmental conditions, and circuit topology conditions. Electrical conditions are set as a set of typical electrical excitation parameters, such as a gate drive voltage range of +15V to +20V, an operating frequency range of 20kHz to 250kHz, and a load current range of 5A to 100A, to cover response conditions under different workloads and switching frequencies. Environmental conditions include the operating temperature range (e.g., –40°C to 150°C), heat dissipation conditions (e.g., natural convection, air cooling, or liquid cooling), and ambient humidity (e.g., 85%RH industrial humid heat conditions), used to simulate the electrical changes of the device under thermal and climatic stresses. The circuit topology conditions are determined by the preset application circuit architecture, such as half-bridge, full-bridge, synchronous Buck or three-phase inverter topology, and also include the parasitic inductance (e.g. 10nH), parasitic capacitance (e.g. 100pF) and drive circuit parameters (e.g. rise time 50ns, drive impedance 5Ω, etc.) set in the simulation, so as to ensure that the modeling environment is close to the actual application system.

[0064] The aforementioned preset operating conditions are used as simulation constraints input to the SiC MOS coupled model. Simulations are then performed using integrated structural parameter models, gate oxide characteristic models, and temperature-dependent models. The final output includes predicted dynamic characteristics, such as gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data. The gate charging simulation data reflects the relationship between gate voltage and charge response. The switching characteristic simulation data reflects the current and voltage changes during device turn-on and turn-off. The reverse recovery simulation data reflects the recovery behavior of the internal body diode or freewheeling path when the current direction reverses.

[0065] Furthermore, the method provided in the application embodiments also includes:

[0066] A time-series analysis is performed on the predicted dynamic characteristics to obtain the predicted time series; a polynomial regression fitting analysis is performed on the predicted time series to obtain the predicted fitting formula; a predetermined dynamic characteristic threshold is read, and the dynamic characteristic lifetime of the SiC MOS device is obtained by combining the predicted fitting formula, which is denoted as the predicted lifetime; wherein, the predicted lifetime is used to characterize the predicted lifetime value of the SiC MOS device.

[0067] In this embodiment, the predicted dynamic characteristics are first analyzed using a time-series analysis method. This method is based on simulated gate charging data, switching characteristic data, and reverse recovery data from the simulation output. Through signal extraction and key node identification operations, it obtains time-domain feature values ​​in each dynamic behavior. For example, it extracts the turn-on delay time from the rising edge of the gate-source voltage, the turn-off time from the falling segment of the drain current, and the reverse recovery time and recovery current peak from the reverse current waveform. Each data point corresponds to a specific timestamp, forming a complete response curve, ultimately generating a predicted timing sequence with a time-series structure.

[0068] Next, a multinomial regression fitting analysis is performed on the predicted time series. In this process, firstly, the predicted time corresponding to the predicted dynamic characteristics is obtained, and a correspondence between this time series and the response data is established to form the predicted time series, which is then used to generate a predicted scatter plot. Subsequently, a prediction curve is constructed based on the principle of random sampling consistency, and this curve is used for multinomial regression fitting analysis to obtain the predicted fitting formula.

[0069] Subsequently, predetermined dynamic characteristic thresholds are read. These thresholds are performance boundary parameters set to evaluate device reliability, such as maximum allowable turn-on delay time, maximum reverse recovery current, and maximum gate charge. The predetermined dynamic characteristic thresholds are then substituted one by one into the prediction fitting formula, and the time point or number of iterations at which the output value of the prediction fitting formula first reaches or exceeds the corresponding threshold is calculated in reverse. This calculation process is implemented using numerical methods, such as using an iterative method to gradually approximate the point where the output result of the prediction fitting formula equals the threshold. The obtained time point represents the expected running time for the dynamic characteristic to reach the preset performance upper limit.

[0070] Ultimately, this point in time or number of cycles is defined as the dynamic characteristic lifetime of the SiC MOS device, denoted as the predicted lifetime. The predicted lifetime is a numerical result, measured in "h" (hours) or "cycles" (number of cycles), used to characterize the available lifespan of the device to maintain stable performance under current operating conditions.

[0071] Furthermore, in the method provided in the application embodiments, performing multinomial regression fitting analysis on the predicted time series to obtain the predicted fitting formula further includes:

[0072] Obtain the prediction time corresponding to the predicted dynamic characteristics; form the prediction time series according to the correspondence between the predicted dynamic characteristics and the prediction time, and generate a prediction scatter plot; generate the prediction curve of the prediction scatter plot based on the principle of random sampling consistency, and perform multinomial regression fitting analysis on the prediction curve to obtain the prediction fitting formula.

[0073] In this embodiment, the prediction time corresponding to the predicted dynamic characteristics is first obtained. The predicted dynamic characteristics include gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data, all of which are generated by simulation using a SiC MOS coupling model, and each data point is accompanied by a time tag. By reading the time coordinates corresponding to key points with characteristic significance (such as peaks, inflection points, zero crossings, etc.) in these simulation data, the prediction time reflecting the changes in dynamic behavior is extracted, thereby obtaining the prediction time corresponding to the predicted dynamic characteristics.

[0074] Subsequently, based on the correspondence between predicted dynamic characteristics and predicted time, a predicted time series is formed. The predicted time series consists of multiple data pairs with time as the independent variable and response value as the dependent variable, covering the entire process of dynamic behavior. Based on this, a two-dimensional predicted scatter plot is plotted with predicted time as the x-axis and response value as the y-axis, visually displaying the changing trajectories of dynamic parameters such as turn-on delay time, turn-off time, and reverse recovery time during the simulation process. The predicted scatter plot provides the original data foundation for subsequent characteristic trend modeling.

[0075] After obtaining the predicted scatter plot, data filtering and prediction curve extraction are performed based on the principle of random sampling consistency. Specifically, three non-collinear data points are randomly selected from the predicted scatter plot, and an initial curve model is fitted to generate it. Then, using this model as a benchmark, the vertical distance from all data points in the plot to the curve is calculated, and an error threshold is set. For example, points with an error less than 1% of the maximum response value are considered "interior points." The number of interior points that meet this condition is counted. When this number exceeds 80% of the total number of points, the model is considered valid, and the set of interior points and its corresponding model are recorded. This process is repeated 300 times, and each time the set of data points with the most error interior points is retained as the final predicted curve point set, thereby generating a predicted curve that reflects the main trend of dynamic characteristics.

[0076] Finally, a polynomial regression fitting analysis was performed on the above-mentioned prediction curves. Using the least squares method, the extracted prediction curve data points were numerically processed. A third- or fourth-order polynomial model was selected as the fitting function. Based on the principle of minimizing the sum of squared residuals, a continuous and differentiable mathematical expression was fitted, forming the prediction fitting formula.

[0077] In summary, the embodiments of this application have at least the following technical effects:

[0078] This application collects multidimensional feature information of SiC MOS devices and constructs a SiCMOS coupling model based on this information. Simulation records are obtained through the SiC MOS coupling model, and a predetermined verification strategy is introduced to verify and analyze these records, yielding simulation verification results. When the simulation verification results meet predetermined simulation constraints, the SiC MOS coupling model is invoked to simulate and obtain the predicted dynamic characteristics of the SiC MOS device. This invention solves the technical problem that existing technologies cannot effectively predict the dynamic characteristics of SiC MOS devices before switching. By constructing a coupling model and performing simulation verification analysis, it achieves the technical effect of accurately predicting dynamic characteristics.

[0079] Example 2, based on the same inventive concept as the SiC MOS dynamic characteristic prediction method in the previous examples, such as... Figure 2 As shown, this application provides a SiC MOS dynamic characteristic prediction system. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0080] The information collection module 11 is used to collect multi-dimensional feature information of the SiC MOS device and construct a SiC MOS coupling model based on the multi-dimensional feature information; the verification and analysis module 12 is used to obtain simulation records through the SiC MOS coupling model and to perform verification and analysis on the simulation records by introducing a predetermined verification strategy to obtain simulation verification results; the prediction module 13 is used to retrieve the SiC MOS coupling model to simulate and obtain the predicted dynamic characteristics of the SiC MOS device when the simulation verification results meet the predetermined simulation constraints.

[0081] Furthermore, the system is also used to implement the following functions:

[0082] Construct a structural parameter model based on the geometric parameters in the multidimensional feature information;

[0083] A gate oxide layer characteristic model is constructed based on the oxide layer traps in the multidimensional feature information.

[0084] A temperature-related model is constructed based on the temperature characteristic parameters in the multidimensional feature information;

[0085] The geometric parameters include trench depth, trench width, and trench spacing; the oxide layer traps include interface electronic defects and internal electronic traps of the oxide layer; and the temperature characteristic parameters include mobility temperature coefficient, carrier lifetime temperature coefficient, and saturation velocity temperature coefficient.

[0086] By combining the structural parameter model, the gate oxide characteristic model, and the temperature-related model, the SiC MOS coupling model is obtained.

[0087] Furthermore, the system is also used to implement the following functions:

[0088] Extract the static verification plan from the predetermined verification strategy;

[0089] According to the static verification plan, static verification information is measured and obtained. The static verification information includes multiple sets of charging characteristics with temperature labels, and each set of charging characteristics with temperature labels includes transfer characteristics and output characteristics.

[0090] Extract the dynamic verification plan from the predetermined verification strategy;

[0091] According to the dynamic verification plan, dynamic verification information is measured, wherein the dynamic verification information includes switching characteristics and reverse recovery characteristics;

[0092] The simulation verification results are obtained by analyzing the transfer characteristics, the output characteristics, the switching characteristics, and the reverse recovery characteristics.

[0093] Furthermore, the system is also used to implement the following functions:

[0094] Extract the temperature verification plan from the predetermined verification strategy;

[0095] According to the temperature verification plan, an arbitrary temperature is obtained, and the arbitrary charging characteristic corresponding to the arbitrary temperature is matched among the multiple sets of charging characteristics with temperature labels.

[0096] Match the arbitrary charging simulation data corresponding to the arbitrary temperature in the gate charging simulation data.

[0097] By comparing the arbitrary charging characteristics with the arbitrary charging simulation data, the comparison results are obtained;

[0098] The average value of the comparison results is taken and normalized to obtain the temperature characteristic coefficient;

[0099] The simulation verification results are calibrated and adjusted using the temperature characteristic coefficient as a weight.

[0100] Furthermore, the system is also used to implement the following functions:

[0101] The arbitrary transfer characteristic in the arbitrary charging characteristic is compared with the arbitrary simulated transfer characteristic in the arbitrary charging simulation data, and the deviation of the arbitrary transfer characteristic is calculated.

[0102] Compare any output characteristic in the arbitrary charging characteristics with any simulated output characteristic in the arbitrary charging simulation data, and the deviation of the arbitrary output characteristic;

[0103] The arbitrary transfer characteristic deviation and the arbitrary output characteristic deviation after standardization are weighted and calculated to obtain the arbitrary deviation index;

[0104] The arbitrary deviation index is used as the comparison result.

[0105] Furthermore, the system is also used to implement the following functions:

[0106] Collect the operating condition information of the SiC MOS device;

[0107] Using the operating condition information as simulation constraints, the predicted dynamic characteristics are obtained through simulation using the SiC MOS coupling model;

[0108] The operating condition information includes electrical conditions, environmental conditions, and circuit topology conditions; the electrical conditions include the gate drive voltage range, operating frequency range, and load current range; the environmental conditions include the operating temperature range, heat dissipation conditions, and ambient humidity; and the circuit topology conditions include the application circuit topology, parasitic parameters, and drive circuit parameters.

[0109] Furthermore, the system is also used to implement the following functions:

[0110] A time series analysis is performed on the predicted dynamic characteristics to obtain the predicted time series;

[0111] A polynomial regression fitting analysis was performed on the predicted time series to obtain the predicted fitting formula;

[0112] The predetermined dynamic characteristic threshold is read and combined with the predicted fitting formula to obtain the dynamic characteristic lifetime of the SiC MOS device, which is denoted as the predicted lifetime.

[0113] The predicted lifetime is used to characterize the predicted lifetime of the SiC MOS device.

[0114] Furthermore, the system is also used to implement the following functions:

[0115] Obtain the prediction time corresponding to the predicted dynamic characteristics;

[0116] The prediction time series is formed based on the correspondence between the prediction dynamic characteristics and the prediction time, and a prediction scatter plot is generated.

[0117] The prediction curve of the prediction scatter plot is generated based on the principle of random sampling consistency, and the prediction curve is subjected to polynomial regression fitting analysis to obtain the prediction fitting formula.

[0118] In Example 3, based on the same inventive concept as the SiC MOS dynamic characteristic prediction method in the foregoing embodiments, this application also provides a computer-readable storage medium storing a computer program, which, when executed, implements the steps of any one of the methods in Example 1 above.

[0119] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0120] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0121] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for predicting the dynamic characteristics of SiC MOS, characterized in that, include: Multidimensional feature information of SiC MOS devices is collected, and a SiC MOS coupling model is constructed based on the multidimensional feature information; The simulation record is obtained through the SiC MOS coupling model, and a predetermined verification strategy is introduced to verify and analyze the simulation record to obtain the simulation verification results. When the simulation verification results meet the predetermined simulation constraints, the SiC MOS coupling model is retrieved to simulate and obtain the predicted dynamic characteristics of the SiC MOS device. The simulation records include gate charging simulation data, switching characteristic simulation data, and reverse recovery simulation data. These simulation records are obtained through the SiC MOS coupling model, and a predetermined verification strategy is introduced to verify and analyze them, yielding simulation verification results, including: Extract the static verification plan from the predetermined verification strategy; According to the static verification plan, static verification information is measured and obtained. The static verification information includes multiple sets of charging characteristics with temperature labels, and each set of charging characteristics with temperature labels includes transfer characteristics and output characteristics. Extract the dynamic verification plan from the predetermined verification strategy; According to the dynamic verification plan, dynamic verification information is measured, wherein the dynamic verification information includes switching characteristics and reverse recovery characteristics; The simulation verification results are obtained by analyzing the transfer characteristics, the output characteristics, the switching characteristics, and the reverse recovery characteristics.

2. The SiC MOS dynamic characteristic prediction method as described in claim 1, characterized in that, The process involves collecting multidimensional feature information of SiC MOS devices and constructing a SiC MOS coupling model based on this information, including: Construct a structural parameter model based on the geometric parameters in the multidimensional feature information; A gate oxide layer characteristic model is constructed based on the oxide layer traps in the multidimensional feature information. A temperature-related model is constructed based on the temperature characteristic parameters in the multidimensional feature information; The geometric parameters include trench depth, trench width, and trench spacing; the oxide layer traps include interface electronic defects and internal electronic traps of the oxide layer; and the temperature characteristic parameters include mobility temperature coefficient, carrier lifetime temperature coefficient, and saturation velocity temperature coefficient. The SiCMOS coupling model is obtained by combining the structural parameter model, the gate oxide layer characteristic model, and the temperature-related model.

3. The SiC MOS dynamic characteristic prediction method as described in claim 1, characterized in that, After analyzing the transfer characteristics, output characteristics, switching characteristics, and reverse recovery characteristics to obtain the simulation verification results, the method further includes: Extract the temperature verification plan from the predetermined verification strategy; According to the temperature verification plan, an arbitrary temperature is obtained, and the arbitrary charging characteristic corresponding to the arbitrary temperature is matched among the multiple sets of charging characteristics with temperature labels. Match the arbitrary charging simulation data corresponding to the arbitrary temperature in the gate charging simulation data. By comparing the arbitrary charging characteristics with the arbitrary charging simulation data, the comparison results are obtained; The average value of the comparison results is taken and normalized to obtain the temperature characteristic coefficient; The simulation verification results are calibrated and adjusted using the temperature characteristic coefficient as a weight.

4. The SiC MOS dynamic characteristic prediction method as described in claim 3, characterized in that, By comparing the arbitrary charging characteristics with the arbitrary charging simulation data, the comparison results are obtained, including: The arbitrary transfer characteristic in the arbitrary charging characteristic is compared with the arbitrary simulated transfer characteristic in the arbitrary charging simulation data, and the deviation of the arbitrary transfer characteristic is calculated. Compare any output characteristic in the arbitrary charging characteristics with any simulated output characteristic in the arbitrary charging simulation data, and the deviation of the arbitrary output characteristic; The arbitrary transfer characteristic deviation and the arbitrary output characteristic deviation after standardization are weighted and calculated to obtain the arbitrary deviation index; The arbitrary deviation index is used as the comparison result.

5. The SiC MOS dynamic characteristic prediction method as described in claim 1, characterized in that, When the simulation verification results meet the predetermined simulation constraints, the SiC MOS coupling model is invoked to simulate the predicted dynamic characteristics of the SiC MOS device, including: Collect the operating condition information of the SiC MOS device; Using the operating condition information as simulation constraints, the predicted dynamic characteristics are obtained through simulation using the SiC MOS coupling model; The operating condition information includes electrical conditions, environmental conditions, and circuit topology conditions; the electrical conditions include the gate drive voltage range, operating frequency range, and load current range; the environmental conditions include the operating temperature range, heat dissipation conditions, and ambient humidity; and the circuit topology conditions include the application circuit topology, parasitic parameters, and drive circuit parameters.

6. The SiC MOS dynamic characteristic prediction method as described in claim 5, characterized in that, Also includes: A time series analysis is performed on the predicted dynamic characteristics to obtain the predicted time series; A polynomial regression fitting analysis was performed on the predicted time series to obtain the predicted fitting formula; The predetermined dynamic characteristic threshold is read and combined with the predicted fitting formula to obtain the dynamic characteristic lifetime of the SiC MOS device, which is denoted as the predicted lifetime. The predicted lifetime is used to characterize the predicted lifetime of the SiC MOS device.

7. The SiC MOS dynamic characteristic prediction method as described in claim 6, characterized in that, A polynomial regression fitting analysis is performed on the predicted time series to obtain the predicted fitting formula, including: Obtain the prediction time corresponding to the predicted dynamic characteristics; The prediction time series is formed based on the correspondence between the prediction dynamic characteristics and the prediction time, and a prediction scatter plot is generated. The prediction curve of the prediction scatter plot is generated based on the principle of random sampling consistency, and the prediction curve is subjected to polynomial regression fitting analysis to obtain the prediction fitting formula.

8. A SiC MOS dynamic characteristic prediction system, characterized in that, The system is used to execute the SiC MOS dynamic characteristic prediction method as described in any one of claims 1-7, and the system comprises: An information collection module is used to collect multi-dimensional feature information of SiC MOS devices and construct a SiC MOS coupling model based on the multi-dimensional feature information; The verification and analysis module is used to obtain the simulation record through the SiC MOS coupling model, and to perform verification and analysis on the simulation record by introducing a predetermined verification strategy to obtain the simulation verification result; The prediction module is used to retrieve the SiC MOS coupling model to simulate and obtain the predicted dynamic characteristics of the SiC MOS device when the simulation verification results meet the predetermined simulation constraints.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the SiC MOS dynamic characteristic prediction method as described in any one of claims 1-7.

Images