Method for measuring valence band maximum of material based on ultraviolet photoelectron spectroscopy

Abstract

The application provides a method for measuring material valence band top based on ultraviolet photoelectron spectroscopy, and relates to the technical field of material analysis, and comprises the following steps: determining an ultraviolet photoelectron binding energy spectrum of a sample to be measured; performing wavelet decomposition on the ultraviolet photoelectron binding energy spectrum to obtain multiple levels, reconstructing a signal after removing components with a frequency higher than a threshold to obtain a denoised ultraviolet photoelectron binding energy spectrum; selecting a valence band top region from the denoised ultraviolet photoelectron binding energy spectrum, and performing fitting on the valence band top region to obtain a fitting result; and determining the value of the material valence band top of the sample to be measured according to the fitting result. The application can overcome the hardware limitation of a measuring instrument, avoid interference in a forbidden band region, and significantly improve the measurement accuracy and reliability in the study of material electronic structure.

Description

Technical Field

[0001] This invention relates to the field of materials analysis technology, specifically to a method for measuring the valence band top of materials based on ultraviolet photoelectron spectroscopy. Background Technology

[0002] In the electronic structure analysis of materials, ultraviolet photoelectron spectroscopy (UPS) is a commonly used characterization technique that is widely used to measure the band structure of materials, especially in determining the location of the valence band maximum (VBM).

[0003] However, in high-precision measurement scenarios, related technologies are often limited by signal noise and instrument accuracy, easily introducing errors when measuring the valence band top position of materials. These errors may stem from hardware limitations of the instrument, leading to unreliable measurement results. Simultaneously, environmental noise, internal instrument interference, and systematic errors often affect the signal during experiments, resulting in a higher noise content in the acquired signal, further impacting the accurate determination of the valence band top. Furthermore, the distinction between the valence band top and the band gap region is often ambiguous; selecting the valence band top region based on this ambiguity can easily lead to selecting too many band gap regions, resulting in erroneous measurement results. Summary of the Invention

[0004] This invention provides a method for measuring the valence band top of materials based on ultraviolet photoelectron spectroscopy, aiming to solve the problems existing in the background art. To solve the above-mentioned technical problems, this invention is implemented as follows: In a first aspect, embodiments of the present invention provide a method for measuring the valence band top of a material based on ultraviolet photoelectron spectroscopy, comprising: The ultraviolet photoelectron binding energy spectrum of the sample to be tested is determined, wherein the abscissa of the ultraviolet photoelectron binding energy spectrum is the photoelectron binding energy and the ordinate is the energy spectrum signal intensity. Wavelet decomposition is performed on the ultraviolet photoelectron binding energy spectrum to obtain multiple levels, and levels above a preset threshold are removed. The signal is then reconstructed to obtain a denoised ultraviolet photoelectron binding energy spectrum. The valence band top region is selected from the denoised ultraviolet photoelectron binding energy spectrum, and the valence band top region is fitted to obtain the fitting result. Based on the fitting results, the value of the material valence band top of the sample to be tested is determined.

[0005] Optionally, wavelet decomposition is performed on the ultraviolet photoelectron binding energy spectrum to obtain multiple levels, and levels exceeding a preset threshold are removed. The signal is then reconstructed to obtain a denoised ultraviolet photoelectron binding energy spectrum, including: Using the Daubechies 8 wavelet basis as the basis function, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is decomposed to obtain the multiple levels and the detail coefficients of each level. The number of multiple levels is determined according to the number of data points of the ultraviolet photoelectron binding energy spectrum, and the detail coefficients characterize the high-frequency features of the energy spectrum signal. The energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is processed according to the detail coefficients of each of the multiple levels to obtain the denoised ultraviolet photoelectron binding energy spectrum.

[0006] Optionally, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is processed according to the detail coefficients of each of the multiple levels to obtain the denoised ultraviolet photoelectron binding energy spectrum, including: Calculate the energy percentage of the detail coefficients for each of the multiple levels; Identify all levels in the high-frequency region of the ultraviolet photoelectron binding energy spectrum; Calculate the median of the energy percentage of all levels located in the high-frequency region, and determine the noise threshold based on the median; Based on the noise threshold, the detail coefficients of each level in the plurality of levels are shrunk to obtain the shrunk detail coefficients of each level. Based on the detail coefficients after shrinkage at each level, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is reconstructed to obtain the denoised ultraviolet photoelectron binding energy spectrum.

[0007] Optionally, the median energy percentage of all levels located in the high-frequency region is calculated, and a noise threshold is determined based on the median, including: The regions containing the n levels with the highest energy percentage of detail coefficients among the multiple levels are defined as high-frequency regions, where n is an integer greater than 1; Calculate the median energy percentage of all levels within the high-frequency region; The noise threshold is determined to be a preset multiple of the median.

[0008] Optionally, a valence band top region is selected from the denoised ultraviolet photoelectron binding spectrum, and the valence band top region is fitted to obtain a fitting result, including: Based on the maximum and minimum values ​​of the energy spectrum signal intensity in the denoised ultraviolet photoelectron binding energy spectrum, the denoised ultraviolet photoelectron binding energy spectrum is normalized to obtain the normalized ultraviolet photoelectron binding energy spectrum. On the normalized ultraviolet photoelectron binding energy spectrum, the region containing the valence band top feature is defined as the valence band top region; A continuous bisegmental linear model is constructed, and the bisegmental linear model is used to fit the valence band top region to obtain the fitting result. The continuous bisegmental linear model defines two straight lines with different slopes on the left and right sides, with the segmentation point as the boundary.

[0009] Optionally, the fitting result is obtained by fitting the valence band top region using the bisegmental linear model, including: Define a population containing a preset number of individuals, and encode each individual as a corresponding three-dimensional parameter vector, wherein the three-dimensional parameter vector includes the segment point, the slope of the left segment line, and the slope of the right segment line; The fitness of each individual is calculated based on the Huber loss function, wherein the threshold of the Huber loss function is determined based on the fitting residuals; In each iteration, a preset number of individuals are randomly selected, and the individuals with the highest fitness among the preset number of individuals are retained to the next generation. In addition, among the individuals in the same generation, the individuals with the highest fitness ranked from high to low are retained to the next generation. With a first preset probability, the encoding bits of the three-dimensional parameter vector corresponding to each individual are randomly swapped, and with a second preset probability, Gaussian perturbation is applied to the parameter values ​​of the three-dimensional parameter vector corresponding to each individual. Under the preset iteration conditions, a fitting result containing the optimal segmentation point, the optimal slope of the left segment, and the optimal slope of the right segment is obtained.

[0010] Optionally, based on the fitting results, determining the value of the material valence band top of the sample to be tested includes: Divide the energy spectrum signal intensity at the optimal segment point by the slope of the optimal left segment line to obtain the energy compensation amount in the direction of the slope of the optimal left segment line. The energy compensation amount is subtracted from the energy spectrum signal intensity corresponding to the optimal segmentation point to obtain the value of the material valence band top of the sample to be tested.

[0011] Optionally, before determining the ultraviolet photoelectron binding energy spectrum of the sample to be tested, the following steps are also included: Measure the ultraviolet photoelectron spectrum of standard samples; The zero point of the Fermi level and the position of the secondary electron cutoff edge were determined from the ultraviolet photoelectron spectrum of the standard sample. The work function of the standard sample is calculated based on the photon energy of the ultraviolet light source irradiating the standard sample, the zero position of the Fermi level, and the position of the secondary electron cutoff edge. Based on the photon energy of the ultraviolet light source irradiating the sample to be tested and the work function of the standard sample, the photoelectron kinetic energy of the sample to be tested is converted into the corresponding binding energy; Using the binding energy as the abscissa and the energy spectrum signal intensity as the ordinate, the ultraviolet photoelectron binding energy spectrum of the sample to be tested is obtained.

[0012] Optionally, after determining the ultraviolet photoelectron binding energy spectrum of the sample to be tested, the method further includes: The baseline noise value is calculated based on the signal intensity within the preset range of the lowest binding energy in the ultraviolet photoelectron binding energy spectrum. The baseline noise value is subtracted from the energy spectrum signal intensity of each data point in the ultraviolet photoelectron binding energy spectrum to obtain the ultraviolet photoelectron binding energy spectrum after dark noise correction.

[0013] Secondly, embodiments of the present invention provide a device for measuring the valence band top of a material based on ultraviolet photoelectron spectroscopy, comprising: The conversion module is used to determine the ultraviolet photoelectron binding energy spectrum of the sample to be tested, wherein the horizontal axis of the ultraviolet photoelectron binding energy spectrum is the photoelectron binding energy and the vertical axis is the energy spectrum signal intensity. The decomposition module is used to perform wavelet decomposition on the ultraviolet photoelectron binding energy spectrum to obtain multiple levels, and remove the levels that are higher than a preset threshold from the multiple levels, and reconstruct the signal to obtain a denoised ultraviolet photoelectron binding energy spectrum. The fitting module is used to select the valence band top region from the denoised ultraviolet photoelectron binding energy spectrum and fit the valence band top region to obtain the fitting result. The determination module is used to determine the value of the material valence band top of the sample to be tested based on the fitting results.

[0014] Optionally, the decomposition module includes: The first decomposition submodule is used to decompose the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum using the Daubechies 8 wavelet basis as the basis function, to obtain the multiple levels and the detail coefficients of each of the multiple levels. The number of the multiple levels is determined according to the number of data points of the ultraviolet photoelectron binding energy spectrum, and the detail coefficients characterize the high-frequency features of the energy spectrum signal. The second decomposition submodule is used to process the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum according to the detail coefficients of each of the multiple levels, so as to obtain the denoised ultraviolet photoelectron binding energy spectrum.

[0015] Optionally, the second decomposition submodule includes: The first decomposition unit is used to calculate the energy percentage of the detail coefficients of each of the multiple levels; The second decomposition unit is used to determine all levels in the high-frequency region of the ultraviolet photoelectron binding energy spectrum; The third decomposition unit is used to calculate the median of the energy proportion of all levels located in the high-frequency region, and to determine the noise threshold based on the median. The fourth decomposition unit is used to shrink the detail coefficients of each of the multiple levels based on the noise threshold, so as to obtain the shrunken detail coefficients of each level. The fifth decomposition unit is used to reconstruct the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum based on the detail coefficients after shrinkage of each level, so as to obtain the denoised ultraviolet photoelectron binding energy spectrum.

[0016] Optionally, the third decomposition unit includes: The first decomposition subunit is used to determine the region where the top n levels with the highest energy proportion of detail coefficients are located in the multiple levels as the high-frequency region, where n is an integer greater than 1; The second decomposition subunit is used to calculate the median of the energy proportion of all levels within the high-frequency region; The third decomposition subunit is used to determine the noise threshold as a preset multiple of the median.

[0017] Optionally, the fitting module includes: The first fitting module is used to normalize the denoised ultraviolet photoelectron binding energy spectrum based on the maximum and minimum values ​​of the energy spectrum signal intensity in the denoised ultraviolet photoelectron binding energy spectrum, so as to obtain the normalized ultraviolet photoelectron binding energy spectrum. The second fitting module is used to determine the region containing the valence band top feature as the valence band top region on the normalized ultraviolet photoelectron binding energy spectrum. The third fitting module is used to construct a continuous bisegment linear model and use the bisegment linear model to fit the valence band top region to obtain the fitting result. The continuous bisegment linear model defines two straight lines with different slopes on the left and right sides, with the segmentation point as the boundary.

[0018] Optionally, the third fitting module includes: The first fitting unit is used to define a population containing a preset number of individuals and encode each individual into a corresponding three-dimensional parameter vector, wherein the three-dimensional parameter vector includes segmentation points, the slope of the left segment of the line and the slope of the right segment of the line; The second fitting unit is used to calculate the fitness of each individual based on the Huber loss function, wherein the threshold of the Huber loss function is determined based on the fitting residuals; The third fitting unit is used to randomly select a preset number of individuals in each iteration, retain the individual with the highest fitness among the preset number of individuals to the next generation, and retain the individuals in the same generation whose fitness ranks from high to low, representing the top preset percentage, to the next generation. The fourth fitting unit is used to randomly swap the encoding bits of the three-dimensional parameter vector corresponding to each individual with a first preset probability, and to apply Gaussian perturbation to the parameter values ​​of the three-dimensional parameter vector corresponding to each individual with a second preset probability. The fifth fitting unit is used to obtain fitting results containing the optimal segment point, the optimal slope of the left segment, and the optimal slope of the right segment, under the condition of meeting the preset iteration conditions.

[0019] Optionally, the determining module includes: The first determining submodule is used to divide the energy spectrum signal intensity at the optimal segmentation point by the slope of the optimal left segment line to obtain the energy compensation amount in the direction of the slope of the optimal left segment line. The second determining submodule is used to subtract the energy compensation amount from the energy spectrum signal intensity corresponding to the optimal segmentation point to obtain the value of the material valence band top of the sample to be tested.

[0020] Optionally, the device further includes: The standard sample measurement module is used to measure the ultraviolet photoelectron spectrum of standard samples; The standard sample determination module is used to determine the position of the Fermi level zero and the position of the secondary electron cutoff edge from the ultraviolet photoelectron spectrum of the standard sample. The calculation module is used to calculate the work function of the standard sample based on the photon energy of the ultraviolet light source irradiating the standard sample, the zero position of the Fermi level, and the position of the secondary electron cutoff edge. The first conversion module is used to convert the photoelectron kinetic energy of the test sample into the corresponding binding energy based on the photon energy of the ultraviolet light source irradiating the test sample and the work function of the standard sample. The second conversion module is used to obtain the ultraviolet photoelectron binding energy spectrum of the sample under test by using the binding energy as the abscissa and the energy spectrum signal intensity as the ordinate.

[0021] The baseline noise calculation module is used to calculate the baseline noise value based on the signal intensity of the preset range of the lowest binding energy in the ultraviolet photoelectron binding energy spectrum. The noise correction module is used to subtract the baseline noise value from the energy spectrum signal intensity of each data point in the ultraviolet photoelectron binding energy spectrum to obtain the ultraviolet photoelectron binding energy spectrum after dark noise correction.

[0022] The technical solutions provided by the embodiments of the present invention bring at least the following beneficial effects: This invention converts ultraviolet photoelectron spectroscopy into ultraviolet photoelectron binding spectroscopy, and combines this with wavelet decomposition and filtering to effectively reduce noise components in the signal, eliminate the influence of signal noise on the measurement results, and significantly improve the measurement accuracy of the valence band top position. The denoising process ensures that the extracted signal is clearer, thus allowing for more accurate localization of the material's valence band top. Furthermore, by fitting the denoised ultraviolet photoelectron binding spectroscopy, this invention can accurately select the valence band top region, effectively avoiding the problem of misselecting bandgap regions, thereby ensuring the accurate determination of the valence band top position. In summary, this invention overcomes the hardware limitations of measuring instruments while avoiding bandgap interference, significantly improving the measurement accuracy and reliability in the study of material electronic structure. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments of the present invention will be briefly introduced below.

[0024] Figure 1 This is a schematic diagram of the steps of a method for measuring the valence band top of a material based on ultraviolet photoelectron spectroscopy, provided in an embodiment of the present invention; Figure 2 This is an ultraviolet photoelectron spectrum of a standard gold sample in one embodiment of the present invention; Figure 3 This is a hafnium oxide ultraviolet photoelectron spectrum converted to a binding energy spectrum in one embodiment of the present invention; Figure 4 This is a hafnium oxide ultraviolet photoelectron spectrum after noise reduction using wavelet transform in one embodiment of the present invention; Figure 5 This is a schematic diagram of the valence band top region selected in one embodiment of the present invention; Figure 6 This is a schematic diagram of calculating the valence band top using two continuous straight lines based on a genetic algorithm in one embodiment of the present invention; Figure 7 This is a structural block diagram of a device for measuring the valence band top of a material based on ultraviolet photoelectron spectroscopy, provided in one embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] In materials electronic structure analysis, ultraviolet photoelectron spectroscopy (UPS) is an important characterization technique widely used to determine the band structure of materials, especially in identifying the valence band top (VBM) position. However, UPS measurements are often affected by signal noise. Environmental noise, instrumental errors, and systematic interference often lead to poor signal quality, thus affecting the accurate determination of the valence band top position. Furthermore, due to limitations in experimental equipment precision and insufficient data processing algorithms, the division between the measured valence band top and the band gap is often unclear, increasing measurement errors and potentially leading to incorrect selection of the valence band top region, ultimately impacting the assessment and study of the material's electronic properties.

[0027] To address the aforementioned issues, the core idea of ​​this invention is to process the ultraviolet photoelectron binding energy spectrum using wavelet denoising technology, thereby removing noise signals and enhancing data reliability. Simultaneously, a fitting algorithm is used to accurately fit the denoised ultraviolet photoelectron binding energy spectrum to clearly distinguish the valence band top region and the band gap region, avoiding the problem of blurred partitioning. Finally, by fitting the denoised data, the position of the valence band top of the material is accurately determined.

[0028] Figure 1 This is a schematic diagram illustrating the steps of a method for measuring the valence band top of a material based on ultraviolet photoelectron spectroscopy, according to an embodiment of the present invention, including: Step S11: Determine the ultraviolet photoelectron binding energy spectrum of the sample to be tested. The horizontal axis of the ultraviolet photoelectron binding energy spectrum is the photoelectron binding energy, and the vertical axis is the energy spectrum signal intensity.

[0029] In one alternative implementation, before determining the ultraviolet photoelectron binding energy spectrum of the sample to be tested, the method further includes: Step S21: Measure the ultraviolet photoelectron spectrum of the standard sample.

[0030] Considering that gold is an ideal conductor and its Fermi level is strictly aligned with the instrument vacuum level, providing an absolute energy reference, this embodiment of the invention selects elemental gold (Au) with a purity ≥ 99.999% as the standard sample.

[0031] Figure 2 This is an ultraviolet photoelectron spectrum of a standard gold sample in one embodiment of the present invention. The horizontal axis represents the photoelectron kinetic energy (unit: electron volts), indicating the kinetic energy distribution of the photoelectrons. The vertical axis represents the energy spectrum signal intensity (unit: electrons), indicating the number of electrons detected at a specific kinetic energy. Figure 2 The ultraviolet photoelectron spectrum of a standard gold sample (Au) is displayed, including the Fermi edge and the secondary electron cutoff edge. This serves as a reference spectral line, providing a calibration basis for the subsequent calculation of the binding energy of samples to be tested. The Fermi level of the gold sample is defined as 0 eV, used to determine the energy reference point for the entire spectrum.

[0032] Step S22: Determine the zero point of the Fermi level and the position of the secondary electron cutoff edge from the ultraviolet photoelectron spectrum of the standard sample.

[0033] The zero-point position and secondary electron cutoff edge position of the Fermi level were extracted from the ultraviolet photoelectron spectrum of the standard sample. The Fermi level is the midpoint of the edge of the abrupt drop in signal intensity in the high kinetic energy region of the spectral line (Fermi step). The Fermi step was fitted with three straight lines, and the energy value corresponding to the midpoint of the second line was taken as the Fermi level position.

[0034] The secondary electron cutoff edge refers to the region in ultraviolet photoelectron spectroscopy where the spectral signal intensity gradually decreases and approaches zero as the photoelectron binding energy increases. Located on the low-kinetic-energy side of the ultraviolet photoelectron spectrum, the secondary electron cutoff edge represents the lowest energy at which electrons can be released from the surface of a standard sample. It can be selected as the intersection of the linearly extrapolated edge of the abrupt drop in spectral intensity on the low-kinetic-energy side and the noise baseline.

[0035] Step S23: Calculate the work function of the standard sample based on the photon energy of the ultraviolet light source irradiating the standard sample, the zero position of the Fermi level, and the position of the secondary electron cutoff edge.

[0036] The work function of the standard sample is calculated by combining the photon energy of the ultraviolet light source irradiating the sample, the zero position of the Fermi level, and the position of the secondary electron cutoff edge. The work function is the minimum energy required for an electron to escape from the sample surface, i.e., the potential barrier that an electron must overcome to escape from the valence band at the sample surface. The photon energy is determined by the wavelength of the ultraviolet light source. In this embodiment, the photon energy (in eV) of the ultraviolet light source irradiating the standard sample is obtained beforehand.

[0037] Work function ( It can be calculated using the following formula:

[0038] In the formula, The photon energy of an ultraviolet light source; This represents the zero point of the Fermi level of the standard sample. This refers to the secondary electron cutoff edge position of the standard sample.

[0039] The work function of the instrument can be calculated by substituting the photon energy of the ultraviolet light source, the position of the Fermi level zero, and the position of the secondary electron cutoff edge into the above formula.

[0040] Step S24: Based on the photon energy of the ultraviolet light source irradiating the sample to be tested and the work function of the standard sample, the photoelectron kinetic energy of the sample to be tested is converted into the corresponding binding energy.

[0041] Based on the law of conservation of energy in the photoelectric emission process, and combining energy... With kinetic energy The conversion formula is:

[0042] In the formula, Photon energy of ultraviolet light source; This is the work function of the standard sample.

[0043] As an energy reference correction term, it eliminates the systematic error introduced by instrument vacuum level drift, and the converted... Directly characterizing the binding energy of photoelectrons in the valence band of materials enables the comparability of data from different instruments.

[0044] Step S25: Using the binding energy as the abscissa and the energy spectrum signal intensity as the ordinate, the ultraviolet photoelectron binding energy spectrum of the sample to be tested is obtained.

[0045] In the ultraviolet photoelectron binding energy spectrum of the sample, the horizontal axis is reset to the binding energy. The vertical axis remains unchanged, still representing the energy spectrum signal intensity. The horizontal axis is then adjusted according to binding energy. The values ​​are arranged in descending order from left to right to obtain the ultraviolet photoelectron binding energy spectrum of the sample to be tested.

[0046] Figure 3 This is a hafnium oxide ultraviolet photoelectron spectrum converted to a binding energy spectrum in one embodiment of the present invention. Taking hafnium oxide as the sample to be tested, the horizontal axis represents the binding energy (unit: eV), indicating the electron binding energy (relative to the Fermi level). The vertical axis represents the energy spectrum signal intensity (unit: counts), indicating the number of electrons detected at a specific binding energy. Figure 3 The original binding energy spectrum of the hafnium oxide sample after pretreatment is shown. Figure 3 The valence band top is located in the low binding energy region, while significant dark noise fluctuations exist in the high binding energy region. It is evident that the spectral signal-to-noise ratio is low, the spectral partitioning is ambiguous (the boundary between the valence band top and the band gap is unclear), and baseline drift exists (such as non-zero noise in the high binding energy region), highlighting the measurement shortcomings of traditional methods.

[0047] Step S12: Perform wavelet decomposition on the ultraviolet photoelectron binding energy spectrum to obtain multiple levels, remove the levels that are higher than a preset threshold, and reconstruct the signal to obtain a denoised ultraviolet photoelectron binding energy spectrum.

[0048] Because related technologies suffer from the problem of valence band top features being submerged in high-noise environments, this invention proposes the following: First, wavelet transform is used to decompose the original energy spectrum signal into multiple levels of different frequencies, where high-frequency levels typically correspond to noise components, while low-frequency levels retain the main structural features of the signal. By setting an adaptive threshold, high-frequency noise levels above the threshold are identified and removed, while retaining low-frequency and mid-frequency components containing effective information. Finally, the signal is reconstructed based on the retained levels to obtain the denoised ultraviolet photoelectron binding energy spectrum, thus overcoming the problems of large noise interference and blurred partitioning in traditional methods. It is understood that this invention automatically optimizes the decomposition depth according to the data scale, avoiding over-decomposition leading to feature loss, and avoids subjective bias from manually setting thresholds based on the statistical characteristics of the highest frequency noise. Simultaneously, it protects the low-frequency band structure and solves the problem of blurred valence band top partitioning.

[0049] In an optional implementation, step S12 specifically includes steps S121 to S122: Step S121: Using the Daubechies 8 wavelet basis as the basis function, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is decomposed to obtain the multiple levels and the detail coefficients of each level. The number of multiple levels is determined according to the number of data points of the ultraviolet photoelectron binding energy spectrum, and the detail coefficients characterize the high-frequency features of the energy spectrum signal.

[0050] The Daubechies 8 (db8) wavelet basis is used as the fundamental function. Due to its tight support (finite non-zero interval) and high-order vanishing moment characteristics, it can accurately capture the local slope change characteristics of the valence band top.

[0051] Let the number of energy spectrum data points of the ultraviolet photoelectron binding energy spectrum of the sample to be tested be... N Then the number of decomposition layers L Determined according to the following rules:

[0052] In the formula, This is the floor operator.

[0053] The number of levels is dynamically determined to ensure that data at different resolutions is adequately decomposed (e.g., ...). N =512 hours L =6), the minimum number of layers can be set to 5 to prevent loss of low-frequency energy spectrum profile (low-frequency characteristics).

[0054] Perform the following on the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum (the horizontal axis is the binding energy, and the vertical axis is the energy spectrum signal intensity) L Layer-wise wavelet decomposition yields the detail coefficients for each layer, and approximate coefficients are obtained. For the th layer... jEach layer has a detail coefficient that characterizes the high-frequency features (noise and valence band abrupt changes) of that layer. The approximation coefficient characterizes the low-frequency profile of the final layer (reflecting the overall structure of the ultraviolet photoelectron binding spectrum).

[0055] After wavelet decomposition, high-frequency noise is isolated to the shallow layer (the side with lower layer level), while the effective signal is enriched in the deep layer (the side with higher layer level).

[0056] Step S122: Based on the detail coefficients of each of the multiple levels, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is processed to obtain the denoised ultraviolet photoelectron binding energy spectrum.

[0057] This invention addresses the problem of blurred valence band top partitioning caused by low signal-to-noise ratio by employing adaptive filtering to suppress noise while preserving key features at the valence band top. First, it analyzes the energy proportion, i.e., quantifies the signal contribution of each level, distinguishing noise from effective features. Next, it calibrates the high-frequency noise threshold and adaptively sets the filtering intensity based on the statistical results of the energy proportion at each level. Finally, it preserves the effective energy spectrum signal and eliminates dark noise, outputting a denoised ultraviolet photoelectron binding spectrum to support accurate fitting by the subsequent genetic algorithm.

[0058] Figure 4 This is a hafnium oxide ultraviolet photoelectron spectrum after wavelet transform denoising in one embodiment of the present invention. Please refer to [link to relevant documentation]. Figure 4 The energy spectrum of hafnium oxide after wavelet denoising. Compared to Figure 3 High-frequency noise is significantly suppressed, and the boundary between the characteristic peak of the valence band top and the flat region of the band gap becomes clearer, providing high signal-to-noise ratio data for subsequent selection of the valence band top region.

[0059] In an optional implementation, step S122 specifically includes steps S1221 to S1225: Step S1221: Calculate the energy percentage of the detail coefficients for each of the multiple levels.

[0060] This step aims to quantify the signal energy distribution of the detail coefficients at each level after wavelet decomposition, defining the energy proportion formula as follows:

[0061] In the formula, For the first j Layer detail coefficient vector; L This represents the total number of wavelet decomposition layers. For the first The sum of squares of the energy of the layer detail coefficients reflects the signal strength of that layer; The sum of squared energy for all level detail coefficients; energy_ratio j For the first jLayer energy percentage is used to distinguish between noise and effective signal layers.

[0062] Step S1222: Determine all levels in the high-frequency region of the ultraviolet photoelectron binding energy spectrum.

[0063] Considering that dark noise and random noise are mainly distributed in the highest frequency band, which covers most of the noise energy, it is necessary to identify the high-frequency level dominated by noise, that is, the interval for calculating the focusing threshold. L Each level is sorted from highest to lowest frequency (serial number) j (The smaller the value, the higher the frequency). Before selection K Each level is designated as a high-frequency zone, among which K The value can be determined according to a preset percentage. For example, if the top 25% of levels are selected, if... L= 8, then K= 2 (Take the first and second layers).

[0064] Step S1223: Calculate the median of the energy proportion of all levels located in the high-frequency region, and determine the noise threshold based on the median.

[0065] To address the quantization and suppression of high-frequency noise, a statistical method is used to adaptively determine the shrinkage threshold of wavelet coefficients. First, the high-frequency region is identified based on the energy proportion of different levels, as it concentrates the main energy of instrument dark noise and random interference. Then, the median of the energy proportion in the high-frequency region is calculated as a statistic of noise intensity, and finally, the noise threshold is set as a multiple of the median. This invention combines the concentrated characteristics of high-frequency noise with statistical outliers to effectively distinguish noise from effective signal levels, avoiding the loss of band structure characteristics due to excessive shrinkage.

[0066] In an optional implementation, step S1223 specifically includes steps S12231 to S12233: Step S12231: Among the multiple levels, select the level with the highest energy percentage for detail coefficients. n The area where each level is located is identified as a high-frequency zone. n It is an integer greater than 1.

[0067] After wavelet decomposition L Each level is arranged in descending order of frequency (index) j (The smaller the value, the higher the frequency). Before selection n Each level is designated as a high-frequency zone. n The value can be set to the top 25% of the total number of floors. For example, if the total number of floors... L=8 ,but n=2 The high-frequency region is j=1,2 The hierarchy.

[0068] Step S12232: Calculate the median of the energy proportion of all levels within the high-frequency region.

[0069] Extract the energy percentage of all levels in the high-frequency region energy_ratio j ,( j ∈[1,n]). Calculate the median of this set. median_ratio Right now:

[0070] In the formula, energy_ratio j For the first j Layer energy percentage.

[0071] Step S12233: The noise threshold is determined to be a preset multiple of the median.

[0072] The threshold formula is as follows:

[0073] In the formula, m It is a natural number greater than 1; threshold This is the noise threshold.

[0074] The preferred value is 3. If the number of effective layers selected according to the noise threshold is less than 3, at least 3 layers shall be forcibly retained to prevent excessive loss of band features.

[0075] Step S1224: Based on the noise threshold, shrink the detail coefficients of each of the multiple levels to obtain the shrunken detail coefficients of each level.

[0076] The energy proportion of each level is compared with the noise threshold: for levels with an energy proportion exceeding the noise threshold, they are considered to be rich in effective energy spectrum characteristic signals, and therefore their detail coefficients are fully preserved, i.e., multiplied by the selection factor "1"; for levels with an energy proportion not exceeding the noise threshold, they are considered to be dominated by noise components, and therefore their detail coefficients are set to zero, i.e., multiplied by the selection factor "0". Through a binary contraction operation of either 0 or 1, the total number of levels participating in reconstruction is kept constant in the programming implementation, while effectively filtering out high-frequency noise levels.

[0077] Step S1225: Based on the detail coefficients after shrinkage of each level, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is reconstructed to obtain the denoised ultraviolet photoelectron binding energy spectrum.

[0078] After the aforementioned steps, the signals at each level obtained from wavelet decomposition are themselves components of the original energy spectrum. Therefore, the reconstruction process does not require complex inverse transform operations; instead, it directly performs a linear superposition of the signal components retained after shrinking all levels. Specifically, the signals at all levels assigned a selection coefficient of "1" are directly added together, while the signals at levels assigned a coefficient of "0" do not contribute because they have been set to zero. Finally, by directly summing the retained effective signal levels in this way, the denoised ultraviolet photoelectron binding energy spectrum is obtained.

[0079] Optionally, adaptive smoothing filtering and Savitzky-Golay filtering are used to process the denoised ultraviolet photoelectron binding energy spectrum signal to eliminate high-frequency noise and Gibbs oscillations, presenting a clear valence band top feature region.

[0080] This invention is based on wavelet transform, which can not only remove noise from spectral lines to a great extent, but also preserve the original information of spectral lines to the maximum extent, reducing the impact of human selection factors on test results.

[0081] Step S13: Select the valence band top region from the denoised ultraviolet photoelectron binding energy spectrum, and fit the valence band top region to obtain the fitting result.

[0082] To address the issues of ambiguous spectral lines and significant noise interference in traditional methods, a normalization preprocessing technique is employed to eliminate dimensional differences. Combined with feature interval selection based on physical constraints, the core region of the valence band top is pinpointed. Finally, the band structure inflection point is analyzed using a continuous bisegmental linear model. This approach overcomes the subjectivity of manual fitting and significantly reduces the positioning error of the valence band top in testing complex materials such as hafnium oxide.

[0083] In an optional implementation, step S13 specifically includes steps S131 to S133: Step S131: Based on the maximum and minimum values ​​of the energy spectrum signal intensity in the denoised ultraviolet photoelectron binding energy spectrum, normalize the denoised ultraviolet photoelectron binding energy spectrum to obtain the normalized ultraviolet photoelectron binding energy spectrum.

[0084] To eliminate the difference in signal intensity dimensions and highlight the valence band top morphology, the denoised ultraviolet photoelectron binding energy spectrum was normalized to obtain the normalized ultraviolet photoelectron binding energy spectrum. The normalization formula is as follows:

[0085] In the formula, min(Y) This is the minimum intensity value across the entire spectrum; max(Y) This represents the maximum intensity value across the entire spectrum.

[0086] Compared to the unnormalized ultraviolet photoelectron binding energy spectrum, the visual distinctiveness of the steep rise of the valence band peak in the normalized ultraviolet photoelectron binding energy spectrum is enhanced.

[0087] Step S132: On the normalized ultraviolet photoelectron binding energy spectrum, the region containing the valence band top feature is determined as the valence band top region.

[0088] Figure 5 This is a schematic diagram of the selected valence band top region in one embodiment of the present invention. Figure 5 The normalized valence band top region is shown. The valence band top is characterized by a distinct peak or a steep rise in the ultraviolet photoelectron binding spectrum. Optionally, the starting point for selecting the interval containing the valence band top feature can be determined as the energy point where the normalized spectral signal intensity first exceeds 0.1, and the ending point can be determined as the energy point where the intensity decays from the peak value to 0.3.

[0089] Step S133: Construct a continuous bisegmental linear model and use the bisegmental linear model to fit the valence band top region to obtain the fitting result. The continuous bisegmental linear model defines two straight lines with different slopes on the left and right sides, with the segmentation point as the boundary.

[0090] To address the issues of high noise and ambiguous partitioning in traditional methods, this invention analyzes the abrupt changes in the valence band top energy band structure using a continuous bisegmental linear model and employs a genetic algorithm for global parameter optimization. In the continuous bisegmental linear model, the slope of the left segment ( The density of electronic states at the top of the valence band rises sharply, and the slope of the right segment ( Characterizes bandgap attenuation; segmentation point ( Locating the band structure transition points. By using a genetic algorithm for fitting, the traditional method of fitting step functions overcomes its reliance on human experience.

[0091] In an optional implementation, step S133 specifically includes steps S1331 to S1335: Step S1331: Define a population containing a preset number of individuals, and encode each individual as a corresponding three-dimensional parameter vector, wherein the three-dimensional parameter vector includes the segment point, the slope of the left segment line, and the slope of the right segment line.

[0092] Construct the parameter search space for the genetic algorithm and define a population containing a preset number of individuals, for example, 100 individuals. Each individual is encoded as a three-dimensional parameter vector. .in, The segmentation point (located in the characteristic interval at the top of the price band) Inside), The slope of the left segment (constrained to be negative). The slope of the right segment (constrained to be near zero).

[0093] In the three-dimensional parameter vector corresponding to each individual code, each parameter is uniformly and randomly generated within the domain.

[0094] Step S1332: Calculate the fitness of each individual based on the Huber loss function, wherein the threshold of the Huber loss function is determined based on the fitting residuals.

[0095] No. The fitting residuals of each data point Calculate according to the following formula:

[0096] In the formula, This is the output value of the bisegmental linear model under the current parameters, i.e., the predicted value; The measured value is the normalized ultraviolet photoelectron binding energy spectrum.

[0097] Huber loss function The threshold is dynamically determined based on the fitted residuals, as shown in the following formula:

[0098] In the formula,

[0099] .

[0100] The fitness of each individual is calculated based on the Huber loss function. Higher fitness indicates better fit quality.

[0101] Step S1333: In each iteration, a preset number of individuals are randomly selected, and the individuals with the highest fitness among the preset number of individuals are retained to the next generation. Also, among the individuals in the same generation, the individuals with the highest fitness ranked from high to low are retained to the next generation.

[0102] In each iteration of the fitting process, a preset number (e.g., 3) of individuals are randomly selected to compete, and the one with the highest fitness is retained to enter the next generation. This process is repeated until the population size requirement (e.g., 100 individuals) is met. In addition, the top 10% of the individuals with the highest fitness in each generation are directly copied to the next generation to prevent the loss of high-quality genes and accelerate the convergence process.

[0103] Step S1334: Randomly swap the encoding bits of the three-dimensional parameter vector corresponding to each individual with a first preset probability, and apply Gaussian perturbation to the parameter values ​​of the three-dimensional parameter vector corresponding to each individual with a second preset probability.

[0104] According to a first preset probability (e.g., 80%), the parent individuals exchange coded bits using a randomly generated binary mask. The coded bits are... For example: Parents

[0105] mask offspring

[0106] According to a second preset probability (e.g., 20%), a Gaussian perturbation is applied to the parameter values ​​of the three-dimensional parameter vector corresponding to each individual, with the perturbation intensity... It can be set to .

[0107] Step S1335: Under the condition of meeting the preset iteration conditions, obtain the fitting result containing the optimal segment point, the optimal slope of the left segment line, and the optimal slope of the right segment line.

[0108] The iteration termination condition (sufficient for any one of the following) is: First, the rate of change of the optimal fitness is <0.1% for 20 consecutive generations, that is:

[0109] Second, the total number of iterations is ≥200 generations.

[0110] The parameters of the individual with the highest fitness in the current generation are taken as the fitting result, that is, as shown in the figure. If they are the most fit individuals of our time, then These represent the optimal segmentation point, the optimal slope of the left segment, and the optimal slope of the right segment, respectively.

[0111] Step S14: Determine the value of the material valence band top of the sample to be tested based on the fitting results.

[0112] Based on the parameters of the bisegmental linear model optimized by a genetic algorithm, the valence band top energy is calculated by extrapolation derived from physical derivation.

[0113] Figure 6 This is a schematic diagram illustrating the calculation of the valence band top using two continuous straight lines based on a genetic algorithm in one embodiment of the present invention. The left fitted straight line corresponds to an increase in the electronic state density at the valence band top; the right fitted straight line represents bandgap decay; the optimal segmentation point... The left endpoint of the extended dashed line; the energy at the top of the price band. eV represents the intersection of the extended dashed line and the zero binding energy baseline. It can be seen that the fitting residuals based on the bisegmental linear model and genetic algorithm are uniformly distributed, and the fitting results are accurate.

[0114] In an optional implementation, step S14 specifically includes steps 141 to S142: Step S141: Divide the energy spectrum signal intensity at the optimal segment point by the slope of the optimal left segment line to obtain the energy compensation amount in the direction of the slope of the optimal left segment line.

[0115] The energy offset of the valence band apex along the left-side straight line direction is calculated using the following formula:

[0116] In the formula, The optimal segmentation point The fitting strength value at that location; The optimal slope of the left segment of the line; The energy compensation amount represents the length of the left straight line from... The energy shift required to extrapolate to the zero intensity point.

[0117] Step 142: Subtract the energy compensation amount from the energy spectrum signal intensity corresponding to the optimal segmentation point to obtain the value of the material valence band top of the sample to be tested.

[0118] The value of the material valence band peak of the sample to be tested is determined according to the following formula. :

[0119] In one alternative implementation, after determining the ultraviolet photoelectron binding energy spectrum of the sample to be tested, the method further includes: S31, Calculate the baseline noise value based on the signal intensity of the preset range with the lowest binding energy in the ultraviolet photoelectron binding energy spectrum.

[0120] To address the high dark noise problem inherent in traditional methods, this invention quantifies the instrument's background noise through a physical-driven approach. Specifically, it selects the energy range (i.e., the high kinetic energy region) of the lowest binding energy in the ultraviolet photoelectron binding spectrum (e.g., 10%). This region is far from the valence band peak and mainly contains the instrument's dark current and electron multiplier thermal noise, purely reflecting the instrument's background noise characteristics.

[0121] Calculate the arithmetic mean of the energy spectrum signal intensity of all data points within the selected interval, and define it as the baseline noise value.

[0122] S32, subtract the baseline noise value from the energy spectrum signal intensity of each data point in the ultraviolet photoelectron binding energy spectrum to obtain the ultraviolet photoelectron binding energy spectrum after dark noise correction.

[0123] Global signal shifting eliminates inherent instrument noise, providing a clean data foundation for subsequent analysis. Specifically, the baseline noise value calculated in step S31 is subtracted from the energy spectrum signal intensity of each data point in the ultraviolet photoelectron binding spectrum to generate a dark noise-corrected ultraviolet photoelectron binding spectrum.

[0124] This invention converts ultraviolet photoelectron spectroscopy into ultraviolet photoelectron binding spectroscopy, and combines this with wavelet decomposition and filtering to effectively reduce noise components in the signal, eliminate the influence of signal noise on the measurement results, and significantly improve the measurement accuracy of the valence band top position. The denoising process ensures that the extracted signal is clearer, thus allowing for more accurate localization of the material's valence band top. Furthermore, by fitting the denoised ultraviolet photoelectron binding spectroscopy, this invention can accurately select the valence band top region, effectively avoiding the problem of misselecting bandgap regions, thereby ensuring the accurate determination of the valence band top position. In summary, this invention overcomes the hardware limitations of measuring instruments while avoiding bandgap interference, significantly improving the measurement accuracy and reliability in the study of material electronic structure.

[0125] Figure 7 This is a structural block diagram of a device for measuring the valence band top of a material based on ultraviolet photoelectron spectroscopy, according to an embodiment of the present invention. The device includes: The conversion module 401 is used to determine the ultraviolet photoelectron binding energy spectrum of the sample to be tested, wherein the horizontal axis of the ultraviolet photoelectron binding energy spectrum is the photoelectron binding energy and the vertical axis is the energy spectrum signal intensity.

[0126] The decomposition module 402 is used to perform wavelet decomposition on the ultraviolet photoelectron binding energy spectrum to obtain multiple levels, remove the levels that are higher than a preset threshold, and reconstruct the signal to obtain a denoised ultraviolet photoelectron binding energy spectrum.

[0127] The fitting module 403 is used to select the valence band top region from the denoised ultraviolet photoelectron binding energy spectrum and fit the valence band top region to obtain the fitting result.

[0128] The determination module 404 is used to determine the value of the material valence band top of the sample to be tested based on the fitting result.

[0129] Optionally, the decomposition module includes: The first decomposition submodule is used to decompose the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum using the Daubechies 8 wavelet basis as the basis function, to obtain the multiple levels and the detail coefficients of each of the multiple levels. The number of the multiple levels is determined according to the number of data points of the ultraviolet photoelectron binding energy spectrum, and the detail coefficients characterize the high-frequency features of the energy spectrum signal. The second decomposition submodule is used to process the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum according to the detail coefficients of each of the multiple levels, so as to obtain the denoised ultraviolet photoelectron binding energy spectrum.

[0130] Optionally, the second decomposition submodule includes: The first decomposition unit is used to calculate the energy percentage of the detail coefficients of each of the multiple levels; The second decomposition unit is used to determine all levels in the high-frequency region of the ultraviolet photoelectron binding energy spectrum; The third decomposition unit is used to calculate the median of the energy proportion of all levels located in the high-frequency region, and to determine the noise threshold based on the median. The fourth decomposition unit is used to shrink the detail coefficients of each of the multiple levels based on the noise threshold, so as to obtain the shrunken detail coefficients of each level. The fifth decomposition unit is used to reconstruct the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum based on the detail coefficients after shrinkage of each level, so as to obtain the denoised ultraviolet photoelectron binding energy spectrum.

[0131] Optionally, the third decomposition unit includes: The first decomposition subunit is used to determine the region where the n levels with the highest frequency of detail coefficients are located among the multiple levels as the high-frequency region, where n is an integer greater than 1; The second decomposition subunit is used to calculate the median of the energy proportion of all levels within the high-frequency region; The third decomposition subunit is used to determine the noise threshold as a preset multiple of the median.

[0132] Optionally, the fitting module includes: The first fitting module is used to normalize the denoised ultraviolet photoelectron binding energy spectrum based on the maximum and minimum values ​​of the energy spectrum signal intensity in the denoised ultraviolet photoelectron binding energy spectrum, so as to obtain the normalized ultraviolet photoelectron binding energy spectrum. The second fitting module is used to determine the region containing the valence band top feature as the valence band top region on the normalized ultraviolet photoelectron binding energy spectrum. The third fitting module is used to construct a continuous bisegment linear model and use the bisegment linear model to fit the valence band top region to obtain the fitting result. The continuous bisegment linear model defines two straight lines with different slopes on the left and right sides, with the segmentation point as the boundary.

[0133] Optionally, the third fitting module includes: The first fitting unit is used to define a population containing a preset number of individuals and encode each individual into a corresponding three-dimensional parameter vector, wherein the three-dimensional parameter vector includes segmentation points, the slope of the left segment of the line and the slope of the right segment of the line; The second fitting unit is used to calculate the fitness of each individual based on the Huber loss function, wherein the threshold of the Huber loss function is determined based on the fitting residuals; The third fitting unit is used to randomly select a preset number of individuals in each iteration, retain the individual with the highest fitness among the preset number of individuals to the next generation, and retain the individuals in the same generation whose fitness ranks from high to low, representing the top preset percentage, to the next generation. The fourth fitting unit is used to randomly swap the encoding bits of the three-dimensional parameter vector corresponding to each individual with a first preset probability, and to apply Gaussian perturbation to the parameter values ​​of the three-dimensional parameter vector corresponding to each individual with a second preset probability. The fifth fitting unit is used to obtain fitting results containing the optimal segment point, the optimal slope of the left segment, and the optimal slope of the right segment, under the condition of meeting the preset iteration conditions.

[0134] Optionally, the determining module includes: The first determining submodule is used to divide the energy spectrum signal intensity at the optimal segmentation point by the slope of the optimal left segment line to obtain the energy compensation amount in the direction of the slope of the optimal left segment line. The second determining submodule is used to subtract the energy compensation amount from the energy spectrum signal intensity corresponding to the optimal segmentation point to obtain the value of the material valence band top of the sample to be tested.

[0135] Optionally, the device further includes: The standard sample measurement module is used to measure the ultraviolet photoelectron spectrum of standard samples; The standard sample determination module is used to determine the position of the Fermi level zero and the position of the secondary electron cutoff edge from the ultraviolet photoelectron spectrum of the standard sample. The calculation module is used to calculate the work function of the standard sample based on the photon energy of the ultraviolet light source irradiating the standard sample, the zero position of the Fermi level, and the position of the secondary electron cutoff edge. The first conversion module is used to convert the photoelectron kinetic energy of the test sample into the corresponding binding energy based on the photon energy of the ultraviolet light source irradiating the test sample and the work function of the standard sample. The second conversion module is used to obtain the ultraviolet photoelectron binding energy spectrum of the sample under test by using the binding energy as the abscissa and the energy spectrum signal intensity as the ordinate.

[0136] The baseline noise calculation module is used to calculate the baseline noise value based on the signal intensity of the preset range of the lowest binding energy in the ultraviolet photoelectron binding energy spectrum. The noise correction module is used to subtract the baseline noise value from the energy spectrum signal intensity of each data point in the ultraviolet photoelectron binding energy spectrum to obtain the ultraviolet photoelectron binding energy spectrum after dark noise correction.

[0137] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, electronic devices, and media. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0138] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods and apparatus according to embodiments of the present invention. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0139] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0140] Finally, it should be noted that in this invention, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the term "comprising" or any other variations thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element. The above provides a detailed description of a method for measuring the valence band top of materials based on ultraviolet photoelectron spectroscopy provided by this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and its core ideas; at the same time, for those skilled in the art, based on the ideas of this invention, there will be changes in specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for measuring the valence band top of a material based on ultraviolet photoelectron spectroscopy, characterized in that, include: Determine the ultraviolet photoelectron binding energy spectrum of the sample to be tested, wherein the abscissa of the ultraviolet photoelectron binding energy spectrum is the photoelectron binding energy and the ordinate is the energy spectrum signal intensity; Wavelet decomposition is performed on the ultraviolet photoelectron binding energy spectrum to obtain multiple levels, and levels above a preset threshold are removed. The signal is then reconstructed to obtain a denoised ultraviolet photoelectron binding energy spectrum. The valence band top region is selected from the denoised ultraviolet photoelectron binding energy spectrum, and the valence band top region is fitted to obtain the fitting result. Based on the fitting results, the value of the material valence band top of the sample to be tested is determined.

2. The method according to claim 1, characterized in that, The ultraviolet photoelectron binding energy spectrum is decomposed using wavelet decomposition to obtain multiple levels. Levels exceeding a preset threshold are removed, and the signal is reconstructed to obtain a denoised ultraviolet photoelectron binding energy spectrum, including: Using the Daubechies 8 wavelet basis as the basis function, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is decomposed to obtain the multiple levels and the detail coefficients of each level. The number of multiple levels is determined according to the number of data points of the ultraviolet photoelectron binding energy spectrum, and the detail coefficients characterize the high-frequency features of the energy spectrum signal. The energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is processed according to the detail coefficients of each of the multiple levels to obtain the denoised ultraviolet photoelectron binding energy spectrum.

3. The method according to claim 2, characterized in that, Based on the detail coefficients of each of the multiple levels, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is processed to obtain the denoised ultraviolet photoelectron binding energy spectrum, including: Calculate the energy percentage of the detail coefficients for each of the multiple levels; Identify all levels in the high-frequency region of the ultraviolet photoelectron binding energy spectrum; Calculate the median of the energy percentage of all levels located in the high-frequency region, and determine the noise threshold based on the median; Based on the noise threshold, the detail coefficients of each level in the plurality of levels are shrunk to obtain the shrunk detail coefficients of each level. Based on the detail coefficients after shrinkage at each level, the energy spectrum signal of the ultraviolet photoelectron binding energy spectrum is reconstructed to obtain the denoised ultraviolet photoelectron binding energy spectrum.

4. The method according to claim 3, characterized in that, Calculate the median energy percentage of all levels located in the high-frequency region, and determine the noise threshold based on the median, including: The region containing the n most frequent levels among the multiple levels is defined as the high-frequency region, where n is an integer greater than 1; Calculate the median energy percentage of all levels within the high-frequency region; The noise threshold is determined to be a preset multiple of the median.

5. The method according to claim 4, characterized in that, The valence band top region is selected from the denoised ultraviolet photoelectron binding energy spectrum, and the valence band top region is fitted to obtain the fitting result, including: Based on the maximum and minimum values ​​of the energy spectrum signal intensity in the denoised ultraviolet photoelectron binding energy spectrum, the denoised ultraviolet photoelectron binding energy spectrum is normalized to obtain the normalized ultraviolet photoelectron binding energy spectrum. On the normalized ultraviolet photoelectron binding energy spectrum, the region containing the valence band top feature is defined as the valence band top region; A continuous bisegmental linear model is constructed, and the bisegmental linear model is used to fit the valence band top region to obtain the fitting result. The continuous bisegmental linear model defines two straight lines with different slopes on the left and right sides, with the segmentation point as the boundary.

6. The method according to claim 5, characterized in that, The fitting result obtained by fitting the valence band top region using the bisegmental linear model includes: Define a population containing a preset number of individuals, and encode each individual as a corresponding three-dimensional parameter vector, wherein the three-dimensional parameter vector includes the segment point, the slope of the left segment line, and the slope of the right segment line; The fitness of each individual is calculated based on the Huber loss function, wherein the threshold of the Huber loss function is determined based on the fitting residuals; In each iteration, a preset number of individuals are randomly selected, and the individuals with the highest fitness among the preset number of individuals are retained to the next generation. In addition, among the individuals in the same generation, the individuals with the highest fitness ranked from high to low are retained to the next generation. With a first preset probability, the encoding bits of the three-dimensional parameter vector corresponding to each individual are randomly swapped, and with a second preset probability, Gaussian perturbation is applied to the parameter values ​​of the three-dimensional parameter vector corresponding to each individual. Under the preset iteration conditions, a fitting result containing the optimal segmentation point, the optimal slope of the left segment, and the optimal slope of the right segment is obtained.

7. The method according to claim 6, characterized in that, Based on the fitting results, the value of the material valence band top of the sample to be tested is determined, including: Divide the energy spectrum signal intensity at the optimal segment point by the slope of the optimal left segment line to obtain the energy compensation amount in the direction of the slope of the optimal left segment line. The energy compensation amount is subtracted from the energy spectrum signal intensity corresponding to the optimal segmentation point to obtain the value of the material valence band top of the sample to be tested.

8. The method according to any one of claims 1 to 7, characterized in that, Before determining the ultraviolet photoelectron binding energy spectrum of the sample to be tested, the following steps are also included: Measure the ultraviolet photoelectron spectrum of standard samples; The zero point of the Fermi level and the position of the secondary electron cutoff edge were determined from the ultraviolet photoelectron spectrum of the standard sample. The work function of the standard sample is calculated based on the photon energy of the ultraviolet light source irradiating the standard sample, the zero position of the Fermi level, and the position of the secondary electron cutoff edge. Based on the photon energy of the ultraviolet light source irradiating the sample to be tested and the work function of the standard sample, the photoelectron kinetic energy of the sample to be tested is converted into the corresponding binding energy; Using the binding energy as the abscissa and the energy spectrum signal intensity as the ordinate, the ultraviolet photoelectron binding energy spectrum of the sample to be tested is obtained.

9. The method according to claim 1, characterized in that, After determining the ultraviolet photoelectron binding energy spectrum of the sample to be tested, the following steps are also included: The baseline noise value is calculated based on the signal intensity within the preset range of the lowest binding energy in the ultraviolet photoelectron binding energy spectrum. The baseline noise value is subtracted from the energy spectrum signal intensity of each data point in the ultraviolet photoelectron binding energy spectrum to obtain the ultraviolet photoelectron binding energy spectrum after dark noise correction.

10. A device for measuring the valence band top of a material based on ultraviolet photoelectron spectroscopy, characterized in that, include: The conversion module is used to determine the ultraviolet photoelectron binding energy spectrum of the sample to be tested, wherein the horizontal axis of the ultraviolet photoelectron binding energy spectrum is the photoelectron binding energy and the vertical axis is the energy spectrum signal intensity. The decomposition module is used to perform wavelet decomposition on the ultraviolet photoelectron binding energy spectrum to obtain multiple levels, and remove the levels that are higher than a preset threshold from the multiple levels, and reconstruct the signal to obtain a denoised ultraviolet photoelectron binding energy spectrum. The fitting module is used to select the valence band top region from the denoised ultraviolet photoelectron binding energy spectrum and fit the valence band top region to obtain the fitting result. The determination module is used to determine the value of the material valence band top of the sample to be tested based on the fitting results.

Images