Quantitative characterization method for trap characteristics of composite dielectric microcell interface

Abstract

The invention relates to a quantitative characterization method for trap characteristics of a composite dielectric micro-area interface, which comprises the following steps of: under an electrostatic force microscope platform, acquiring micro-topography characteristics of a composite dielectric by utilizing a tapping mode, and carrying out polarization treatment on a sample; setting a measurement area, measuring potential characteristics in the depolarization process of the interface micro-area by adopting a lifting mode, and obtaining an isothermal potential attenuation curve of the interface micro-area through the correspondence of a probe path and measurement time due to the fact that the measured potential value is in an attenuation trend along with the movement of the nano-probe; obtaining an optimal clustering effect according to the measured surface potential distribution characteristics of the sample; analyzing according to the position corresponding to the optimal clustering result, so as to divide the areas between the filler and the interface and between the filler and the mixed surface, and deducing the distribution characteristics of the sample interface; and calculating trap distribution characteristics of corresponding areas by combining an isothermal potential attenuation theory. According to the invention, the micro-region trap characteristics of the composite dielectric can be represented more accurately.

Description

Technical Field

[0001] This invention belongs to the field of solid dielectric performance testing technology, and relates to a method for characterizing the trap properties of micro-nano interfaces in composite dielectrics. Background Technology

[0002] Dielectric materials, due to their excellent insulation and other properties, are widely used in high-voltage power equipment such as cables, insulators, and transformers. The performance of dielectrics plays a crucial role in the service life of power equipment and the stability of power systems. Doping with nanoparticles can enhance the electrical properties of composite dielectrics, such as high dielectric constant, low loss, corona resistance, high and low temperature resistance, high breakdown field strength, and resistance to electrical dendrite formation. The interface, the transition region between the nanofiller and the polymer matrix, is a key factor affecting the electrical and mechanical properties of composite dielectrics. Therefore, studying the microstructure of the interface region has a significant impact on improving the performance of nanodielectrics and accelerating their widespread application.

[0003] Characterizing the charge properties of nanoscale micro-regions is an effective means of analyzing the influence of the interface on the electrical performance of composite dielectrics. Currently, although various methods exist for measuring trap parameters in nanocomposite dielectrics, such as the thermally stimulated current method, the photo-stimulated discharge method, and space charge, these methods still have limitations in quantitatively analyzing the distribution information of traps at different energy levels within the dielectric. Both the conductivity current and thermally stimulated current methods indirectly describe the overall carrier generation and transport characteristics of the dielectric. Furthermore, due to insufficient spatial resolution, it is difficult to separate the individual effects of filler and interface properties on carrier transport behavior, failing to meet the needs of studying the micro-region characteristics of composite dielectric interfaces. Therefore, the lack of measurement methods has resulted in the absence of a complete theoretical framework to explain interface characteristics, thus hindering the targeted design of composite dielectric interface structures to improve insulation performance.

[0004] With the rapid development of microscopic measurement technology, the characterization of the microscopic charge properties of composite dielectrics has become a research hotspot. By applying a voltage to a scanning probe, the electrostatic characteristics of the dielectric at the nanoscale during polarization can be obtained. Peng Jinping et al. studied the dielectric response behavior and temperature characteristics of the micro-interface of nanocomposite epoxy resin dielectrics. Peng et al. obtained the nanoscale space charge morphology distribution inside the composite dielectric based on an improved KPFM technique. The applicant and collaborators Zhang Yunxiao et al. used scanning probe microscopy, based on the potential characteristic measurement results, to obtain the charge morphology distribution through inversion, and further deduced the trap distribution based on the charge inversion. The above methods have promoted the study of the relationship between the microstructure and macroscopic properties of dielectrics. However, existing methods compare the results with different macroscopic measurement methods and microscopic calculations, lacking consistent theoretical support for the measurement results. On the other hand, although the above methods have studied the overall trap characteristics of the interfacial micro-regions, they have not distinguished the individual trap characteristics of the filler, matrix, and interfacial regions. Summary of the Invention

[0005] To overcome the shortcomings of existing technologies, this invention provides a quantitative characterization method for the micro-region interface trap properties of composite dielectrics, enabling more precise characterization of these properties. It establishes a novel method for characterizing the relationship between nano-micro-region interfaces and trap properties, based on the isothermal potential decay theory for measuring micro-region trap properties. By introducing an intelligent clustering algorithm based on potential data from different times and locations within the measurement region, trap properties of filler, matrix, and interface regions are extracted and differentiated, enabling systematic and accurate analysis of the trap properties of micro-regions at the composite dielectric interface. This work contributes to revealing the microscopic properties of dielectrics and provides technical and data support for optimizing the performance of composite dielectrics.

[0006] The technical solution adopted by this invention to solve its technical problem is:

[0007] A quantitative characterization method for the interface trap properties of micro-regions in composite dielectrics includes the following steps:

[0008] 1) Under the electrostatic force microscope platform, the microstructure characteristics of the composite dielectric were obtained by tapping mode, and the sample was polarized.

[0009] 2) After polarization is completed, the measurement area is set, and the potential characteristics during the depolarization process of the interface micro-region are measured using the elevation mode. Since the measured potential value decreases as the nanoprobe moves, the isothermal potential decay curve of the interface micro-region can be obtained by corresponding the probe path and measurement time.

[0010] 3) Based on the sample surface potential distribution characteristics measured in step 2), combined with the scanned points, a dataset containing three features of potential, time, and location can be formed; visualize it as a three-dimensional graphic, use the k-means clustering algorithm to build a model for training, and continuously adjust the number of clusters and the initial cluster center points to obtain the optimal clustering effect.

[0011] 4) Based on the optimal clustering results obtained in step 3), we can analyze them according to their corresponding positions to divide the regions between the filler and the interface and the mixing surface, and infer the distribution characteristics of the sample interface.

[0012] 5) Based on the potential decay characteristics of different regions of the composite dielectric measured in step 4), and combined with the isothermal potential decay theory, calculate the trap distribution characteristics of the corresponding regions.

[0013] Furthermore, the method also includes the following steps:

[0014] 6) By combining the sample surface morphology, interface distribution and corresponding trap distribution in the region, the precise trap distribution characteristics of the nano-micro area interface are obtained.

[0015] The beneficial effects of this invention are mainly reflected in the following aspects: it can use the isothermal potential decay theory to carry out the comparison and analysis of micro and macro trap parameters, and at the same time, it can use clustering algorithms to divide the region between the filler and the interface and the mixing surface, which can more accurately characterize the micro-area trap characteristics of composite dielectrics. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the method for quantitatively characterizing the interface structure and trap properties of nano-dielectric microregions according to the present invention.

[0017] Figure 2 This is a schematic diagram of the charge distribution after polarization.

[0018] Figure 3 This is a schematic diagram of a method for measuring the isothermal potential decay of nanoscale interfaces based on scanning probes, where 1 represents the substrate and 2 represents the nanoparticles.

[0019] Figure 4 This is a schematic diagram of the principle of testing the dielectric properties of nano-interface micro-regions based on scanning probes.

[0020] Figure 5 This is a flowchart of a clustering algorithm. Detailed Implementation

[0021] The present invention will now be further described with reference to the accompanying drawings.

[0022] Reference Figures 1-5 A quantitative characterization method for the interface trap properties of micro-regions in composite dielectrics includes the following steps:

[0023] 1) Under the electrostatic force microscope platform, the microstructure characteristics of the composite dielectric were obtained by tapping mode, and the sample was polarized.

[0024] 2) After polarization is completed, the measurement area is set, and the potential characteristics during the depolarization process of the interface micro-region are measured using the elevation mode. Since the measured potential value decreases as the nanoprobe moves, the isothermal potential decay curve of the interface micro-region can be obtained by corresponding the probe path and measurement time.

[0025] 3) Based on the sample surface potential distribution characteristics measured in step 2), combined with the scanned points, a dataset containing three features of potential, time, and location can be formed; visualize it as a three-dimensional graphic, use the k-means clustering algorithm to build a model for training, and continuously adjust the number of clusters and the initial cluster center points to obtain the optimal clustering effect.

[0026] 4) By using the optimal clustering effect obtained in step 3), the regions between the filler and the interface and the mixing surface can be divided, and the distribution characteristics of the sample interface can be inferred.

[0027] 5) Based on the potential decay characteristics of different regions of the composite dielectric measured in step 4), and combined with the isothermal potential decay theory, calculate the trap distribution characteristics of the corresponding regions.

[0028] 6) By combining the sample surface morphology, interface distribution and corresponding trap distribution in the region, the precise trap distribution characteristics of the nano-micro area interface are obtained.

[0029] The implementation scheme of this embodiment is as follows:

[0030] 1) Prepare composite dielectric thin film samples with precisely controllable thickness, and perform such preparations under a high-precision nanoprobe microscope platform, such as... Figure 2 As shown, the microstructure of the insulating sample was obtained using a tapping mode.

[0031] The sample was placed on an electrostatic force microscope platform and polarized for 30 minutes.

[0032] 2) such as Figure 2 As shown, this illustrates the charge distribution after polarization. Figure 3 , Figure 4 As shown, a measurement area is set, and the potential characteristics of the sample interface micro-region are measured in the lifting mode during the depolarization process. Since the measured potential value decreases as the nanoprobe moves, the isothermal potential decay characteristics of the interface micro-region can be obtained by corresponding the probe path and measurement time.

[0033] 3) Select the data such as isothermal potential decay characteristics obtained in step 2). The prediction of interface structure is affected by many factors. In this embodiment, the factors with the greatest correlation to the prediction will be selected. After comprehensive consideration, time, location and potential will be used as influencing factors.

[0034] The data is normalized using the following formula:

[0035]

[0036] Where x' is the normalized data, x is the original data, and x' is the normalized data. max x represents the maximum value in the original data. min This represents the minimum value in the original data. Normalization is used to process the input data, transforming it to the range [0, 1], thus resolving the issue caused by different units. The final output of the model is also the normalized data. To obtain the true value of the interface distribution, the output value needs to be inversely normalized. The specific formula is:

[0037] p = (p max -p min )·p'+p min (2)

[0038] Where P is the true interface distribution output after inverse normalization, and P' is the normalized interface distribution output by the model. max P is the maximum value in the input. min The minimum value in the input;

[0039] These influencing factors are used as the dataset S to be clustered. Each data object in dataset S has three features: location, time, and potential. A three-dimensional data distribution map is plotted using these three features. A training function is defined based on the data distribution map. The number of clusters to be divided is k, where k is set manually. k data objects are randomly selected from dataset S, and each data object serves as the centroid of the first cluster. The k-means clustering algorithm calculates the distance between two sample objects using Euclidean distance, defined as follows:

[0040]

[0041] Where x i =(x i1 x i2 , ...x ip ),x j =(x j1 x j2 , ...x jp ), representing a data object with two P-dimensional attributes;

[0042] Calculate the distance of each of the remaining data objects to the cluster center according to formula (3), and then assign each data object to the class of the nearest center point;

[0043] The formula for calculating the average distance of all sample points is as follows:

[0044]

[0045] In the formula, n is the total number of sample objects in the dataset, and d(x j x i ) represents the sample point x i and x j The Euclidean distance;

[0046] Recalculate the center of each cluster according to equation (4), adjust the partitioning of all data objects, and compare whether there is any change compared with the previous clustering partitioning;

[0047] The objective function uses the squared error criterion function, defined as follows:

[0048]

[0049] Calculate the value of E according to equation (5). If the value of E converges, the clustering process ends and the clustering result is output. Otherwise, return to equation (4) and continue iterating until the clustering partition no longer changes or E reaches the convergence condition.

[0050] 4) By using the clustering results obtained in step 3), we can know the probability of each sampling point being located in the region corresponding to the filler, interface and mixing surface. Combined with the sampling point location on the sample surface, we can artificially divide the range of filler, matrix and interface, thereby inferring the distribution characteristics of the nano-micro area interface.

[0051] 5) Based on the potential decay characteristics of different regions of the composite dielectric obtained in step 4), the trap energy levels at the microscopic interface surface are deduced and calculated, as follows:

[0052] Assuming the charge is uniformly distributed within a distance δ from the surface, and the charge density in the central region is very small and can be ignored, the points on the sample surface can be represented as follows:

[0053]

[0054] Where, ρ + ρ is the positive charge density. - Let ρ be the negative charge density, whose magnitudes are approximately equal.

[0055] Solving the above equation yields the expression for the surface point potential:

[0056]

[0057] The total number of electrons emitted from the upper half of the band gap to the conduction band:

[0058]

[0059] In the formula, f0 is the initial occupancy rate of the trap, assuming f0≈1; N(E T ) is the energy distribution function of the trap, k B Boltzmann's constant is T, and T is the thermodynamic temperature.

[0060] Charge carriers escape from the traps to form a current, which is externally manifested as the decay of the surface potential over time. The relationship between them is expressed as:

[0061]

[0062] Therefore, combining equations (8) and (9), the electron trap density can be obtained from the surface potential decay curve:

[0063]

[0064] In the formula, ε0 is the vacuum permittivity, ε r q is the relative permittivity of the sample. e Where L is the unit charge and L is the sample thickness;

[0065] Considering the different trap energy levels where charge carriers reside, and to better characterize the energy level distribution, traps are classified into two types: deep traps and shallow traps. Due to the different surface potential decay processes corresponding to these traps, a double exponential function is used to fit the measured potential decay curves.

[0066]

[0067] In the formula, a1, a2, b1, and b2 are the fitting parameters;

[0068] Combining equation (11) with the electron trap energy level E T The distribution of the corresponding electron trap energy levels is calculated.

[0069] E T =k B Tln(γ ATE t) (12)

[0070] In the formula γ ATE The frequency at which electrons attempt to escape from the trap;

[0071] 6) Based on the clustering results, the ranges of filler, matrix, and interface are divided and compared with the trap characteristics obtained by inversion calculation in step 5). Combined with the surface morphology of composite dielectric in step 1), the accurate trap distribution characteristics of nano-micro area interface are obtained.

[0072] The scheme in this embodiment utilizes an electrostatic force microscope platform and introduces isothermal potential decay theory and clustering algorithm to accurately obtain the distribution of nano-interface structure, and remove the influence of matrix and nanoparticles, thereby obtaining more accurate trap distribution characteristics of nano-micro-region interfaces. This can provide new ideas and methods for characterizing the micro-region interface structure and trap characteristics of nanocomposite materials.

[0073] The embodiments described in this specification are merely examples of implementations of the inventive concept and are for illustrative purposes only. The scope of protection of this invention should not be considered limited to the specific forms described in these embodiments; rather, it extends to equivalent technical means conceived by those skilled in the art based on the inventive concept.

Claims

1. A quantitative characterization method for the interface trap properties of micro-regions in composite dielectrics, characterized in that, The method includes the following steps: 1) Under the electrostatic force microscope platform, the microstructure characteristics of the composite dielectric were obtained by tapping mode, and the sample was polarized. 2) After polarization is completed, the measurement area is set, and the potential characteristics during the depolarization process of the interface micro-region are measured using the elevation mode. Since the measured potential value decreases as the nanoprobe moves, the isothermal potential decay curve of the interface micro-region can be obtained by corresponding the probe path and measurement time. 3) Based on the sample surface potential distribution characteristics measured in step 2), combined with the scanned points, a dataset containing three features of potential, time, and location can be formed; visualize it as a three-dimensional graphic, use the k-means clustering algorithm to build a model for training, and continuously adjust the number of clusters and the initial cluster center points to obtain the optimal clustering effect. 4) Based on the optimal clustering effect diagram obtained in step 3), data analysis can be performed according to its corresponding position to divide the area between the filler and the interface and the mixing surface, and to infer the distribution characteristics of the sample interface. 5) Based on the potential decay characteristics of different regions of the composite dielectric measured in step 4), and combined with the isothermal potential decay theory, calculate the trap distribution characteristics of the corresponding regions.

2. The quantitative characterization method for the interface trap properties of composite dielectric microregions as described in claim 1, characterized in that, The method further includes the following steps: 6) By combining the sample surface morphology, interface distribution and corresponding trap distribution in the region, the precise trap distribution characteristics of the nano-micro area interface are obtained.

3. A quantitative characterization method for the interface trap properties of a composite dielectric microregion as described in claim 1 or 2, characterized in that, In step 3), time, location, and potential are selected as influencing factors, and the data is normalized using the following formula: Where x' is the normalized data, x is the original data, and x' is the normalized data. max x represents the maximum value in the original data. min The minimum value in the original data is used. Normalization is applied to the input data, transforming it to the range [0, 1], thus resolving the issue caused by different units. The final output of the model is also the normalized data. To obtain the true value of the interface distribution, the output value needs to be inversely normalized. The formula is: p=(p max -p min )·p'+p min (2) Where P is the true interface distribution output after inverse normalization, and P' is the normalized interface distribution output by the model. max P is the maximum value in the input. min The minimum value in the input; These influencing factors are used as the dataset S to be clustered. Each data object in dataset S has three features: location, time, and potential. A three-dimensional data distribution map is plotted using these three features. A training function is defined based on the data distribution map. The number of clusters to be divided is k, where k is set manually. k data objects are randomly selected from dataset S, and each data object serves as the centroid of the first cluster. The k-means clustering algorithm calculates the distance between two sample objects using Euclidean distance, defined as follows: Where x i =(x i1 x i2 , ...x ip ),x j =(x j1 x j2 , ...x jp ), representing a data object with two P-dimensional attributes; Calculate the distance of each of the remaining data objects to the cluster center according to formula (3), and then assign each data object to the class of the nearest center point; The formula for calculating the average distance of all sample points is as follows: In the formula, n is the total number of sample objects in the dataset, and d(x j x i ) represents the sample point x i and x j The Euclidean distance; Recalculate the center of each cluster according to equation (4), adjust the partitioning of all data objects, and compare whether there is any change compared with the previous clustering partitioning; The objective function uses the squared error criterion function, defined as follows: Calculate the value of E according to equation (5). If the value of E converges, the clustering process ends and the clustering result is output. Otherwise, return to equation (4) to continue iterating until the clustering partition no longer changes or E reaches the convergence condition.

4. A quantitative characterization method for the interface trap properties of a composite dielectric microregion as described in claim 1 or 2, characterized in that, The process of step 4) is as follows: Assuming the charge is uniformly distributed within a distance δ from the surface, and the charge density in the central region is very small and can be ignored, the points on the sample surface can be represented as follows: Where, ρ + ρ is the positive charge density. - Let ρ be the negative charge density, and let its magnitude be approximately equal to ρ. Solving the above equation yields the expression for the surface point potential: The total number of electrons emitted from the upper half of the band gap to the conduction band: In the formula, f0 is the initial occupancy rate of the trap, assuming f0≈1; N(E T ) is the energy distribution function of the trap, k B Boltzmann's constant is denoted as T, and T is the thermodynamic temperature. Charge carriers escape from the traps to form a current, which is externally manifested as the decay of the surface potential over time. The relationship between them is expressed as: Therefore, combining equations (8) and (9), the electron trap density can be obtained from the surface potential decay curve: In the formula, ε0 is the vacuum permittivity, ε r q is the relative permittivity of the sample. e Where L is the unit charge and L is the sample thickness; Considering the different trap energy levels where charge carriers reside, and to better characterize the energy level distribution, traps are classified into two types: deep traps and shallow traps. Due to the different surface potential decay processes corresponding to these traps, a double exponential function is used to fit the measured potential decay curves. In the formula, a1, a2, b1, and b2 are the fitting parameters; Combining equation (11) with the electron trap energy level E T The distribution of the corresponding electron trap energy levels is calculated. E T =k B Tln(γ ATE t) (12) In the formula γ ATE The frequency at which electrons attempt to escape from the trap.

Images