A method and a system for detecting a sample structure based on a FIB

Abstract

The application provides a sample structure detection method based on a FIB device and a FIB detection system. The sample structure detection method comprises the following steps: obtaining a target SEM image and a target energy spectrum image corresponding to a sample to be detected; segmenting the target SEM image to obtain a first SEM segmentation image, and segmenting the target energy spectrum image to obtain an energy spectrum segmentation image; when a film layer segmentation result in the first SEM segmentation image is different from a film layer segmentation result in the energy spectrum segmentation image, determining sample structure information of the sample to be detected based on the first SEM segmentation image and the energy spectrum segmentation image. According to the application, the sample structure information of the sample to be detected is determined by combining the target SEM image and the target energy spectrum image, so that the accuracy of the film layer segmentation result is ensured, and the accuracy of the sample structure information is ensured.

Description

Technical Field

[0001] This application relates to the field of semiconductor manufacturing technology, and in particular to a sample structure detection method and FIB detection system based on FIB. Background Technology

[0002] The Focused Ion Beam Dual-Beam Microscope (FIB-SEM) is an advanced micro-nano fabrication and analysis instrument that integrates a focused ion beam (FIB) and a scanning electron microscope (SEM). Its core function is to combine precise etching with the ion beam with high-resolution imaging / analysis with the electron beam to fabricate and characterize the micro-nano scale structure of a sample.

[0003] One application of this equipment is to perform micro- and nano-scale cutting of samples at designated locations using a focused ion beam, followed by compositional analysis and dimensional measurement of the resulting multi-layered structure profiles. During dimensional measurement, subjective factors by engineers can introduce significant errors at the micro- and nano-scale, and these errors are more pronounced at the blurred boundaries between different structural layers. Using a single image processing method to identify multi-layered non-conductive structures (such as photoresist layers, silicon oxide layers, and silicon nitride layers stacked together) can lead to problems such as charge accumulation causing abnormal brightness in certain areas of the profile, and the inability to distinguish the contrast between non-conductive layers.

[0004] In view of this, it is necessary to improve the existing multi-layer structure profile identification methods to solve the above problems.

[0005] It should be noted that the above description of the background technology is only for the purpose of providing a clear and complete explanation of the technical solutions of this application and facilitating understanding by those skilled in the art. It should not be assumed that these technical solutions are known to those skilled in the art simply because they have been described in the background technology section of this application.

[0006] Application content

[0007] The purpose of this application is to solve the problems that existing methods for analyzing the composition and measuring the dimensions of multi-layered structures are prone to errors due to manual measurement, and that using a single image processing method for identification results in abnormal brightness in certain areas of the cross-section and inability to distinguish the contrast between non-conductive layers.

[0008] To achieve the above objectives, this application provides a sample structure detection method based on FIB equipment, comprising:

[0009] Acquire the target SEM image and target energy spectrum image of the sample to be tested, wherein the sample to be tested comprises at least two stacked film layers;

[0010] The target SEM image is segmented to obtain a first SEM segmentation image, and the target energy spectrum image is segmented to obtain an energy spectrum segmentation image;

[0011] When the membrane segmentation results in the first SEM segmentation image and the membrane segmentation results in the energy dispersive spectroscopy (EDS) segmentation image are different, the sample structure information of the sample to be detected is determined based on the first SEM segmentation image and the EDS segmentation image.

[0012] The sample structure information includes at least the size information of each of the membrane layers.

[0013] As a further improvement of this application, the target SEM image and the target energy spectrum image are of the same size.

[0014] As a further improvement to this application, the step of determining the sample structure information of the sample to be detected based on the first SEM segmentation image and the energy dispersive spectroscopy (EDS) segmentation image includes:

[0015] When the total number of film layers corresponding to the first SEM segmentation image is greater than the total number of film layers corresponding to the energy dispersive spectroscopy (EDS) segmentation image at the same detection location of the sample to be tested, the film layer segmentation result in the first SEM segmentation image is adjusted according to the film layer segmentation result in the EDS segmentation image to obtain the second SEM segmentation image.

[0016] Based on the pixel ratio of each membrane layer in the second SEM segmentation image and the actual size of the sample to be detected, the size information of each membrane layer is calculated.

[0017] The total number of film layer segments in the second SEM segmentation image is the same as the total number of film layer segments in the energy spectrum segmentation image.

[0018] As a further improvement to this application, the step of determining the sample structure information of the sample to be detected based on the first SEM segmentation image and the energy dispersive spectroscopy (EDS) segmentation image further includes:

[0019] When the total number of film layers corresponding to the first SEM segmentation image is less than the total number of film layers corresponding to the energy dispersive spectroscopy segmentation image at the same detection location of the sample to be tested, a new target SEM image corresponding to the sample to be tested is re-acquired, and the new target SEM image is segmented to obtain a second SEM segmentation image.

[0020] When the total number of film layers corresponding to the second SEM segmentation image is equal to the total number of film layers corresponding to the energy dispersive spectroscopy segmentation image at the same detection location of the sample to be tested, the size information of the film layer is calculated based on the pixel ratio of each film layer in the second SEM segmentation image and the actual size of the sample to be tested.

[0021] When the total number of film layers corresponding to the second SEM segmentation image is less than the total number of film layers corresponding to the energy spectrum segmentation image at the same detection location of the sample to be tested, the size information of the film layers is calculated based on the pixel ratio of each film layer to be distinguished in the energy spectrum segmentation image and the actual size of the sample to be tested, and the size information of the remaining film layers is calculated based on the pixel ratio of each remaining film layer in the second SEM segmentation image and the actual size of the sample to be tested.

[0022] Wherein, the film layer to be distinguished is film layer i in the second SEM segmentation image and film layers j to j+n in the energy spectrum segmentation image, and the remaining film layers are film layers other than the film layer to be distinguished; 1≤i≤k1<k2, 1≤j≤k2, n≥1, k1 is the total number of film layers corresponding to the second SEM segmentation image, and k2 is the total number of film layers corresponding to the energy spectrum segmentation image.

[0023] As a further improvement to this application, the sample structure detection method further includes:

[0024] When the membrane segmentation results in the first SEM segmentation image and the membrane segmentation results in the energy dispersive spectroscopy (EDS) segmentation image are the same, the size information of the membrane is calculated based on the pixel ratio of each membrane in the first SEM segmentation image and the actual size of the sample to be tested.

[0025] As a further improvement of this application, the first SEM segmentation image includes the segmentation results of at least two SEM image sub-blocks;

[0026] The step of segmenting the target SEM image to obtain a first SEM segmentation image includes:

[0027] The target SEM image is divided into blocks to obtain at least two SEM image sub-blocks;

[0028] The target SEM segmentation threshold of each SEM image sub-block is determined based on the grayscale value corresponding to the pixel in each SEM image sub-block.

[0029] Based on the target SEM segmentation threshold of the SEM image sub-block, the SEM image sub-block is segmented to obtain the segmentation result of the SEM image sub-block.

[0030] As a further improvement to this application, determining the target SEM segmentation threshold of the SEM image sub-block based on the grayscale value of each SEM image sub-block includes:

[0031] Based on the grayscale values ​​corresponding to the pixels in each SEM image sub-block, an initial SEM segmentation threshold for the SEM image sub-block is determined, and the initial SEM segmentation threshold is used as the current SEM segmentation threshold.

[0032] The SEM image sub-blocks are segmented based on the current SEM segmentation threshold to obtain the current SEM sub-block segmentation result. Based on the current SEM sub-block segmentation result, a new SEM segmentation threshold is calculated.

[0033] When the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is less than a preset threshold, the new SEM segmentation threshold is used as the target SEM segmentation threshold.

[0034] When the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is greater than or equal to the preset threshold, the new SEM segmentation threshold is used as the current SEM segmentation threshold, and the step of segmenting the SEM image sub-block based on the current SEM segmentation threshold to obtain the current SEM sub-block segmentation result is returned, until the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is less than the preset threshold, and the new SEM segmentation threshold is used as the target SEM segmentation threshold.

[0035] As a further improvement of this application, the energy spectrum segmentation image includes the segmentation results of at least two energy spectrum image sub-blocks;

[0036] The step of segmenting the target energy spectrum image to obtain an energy spectrum segmentation image includes:

[0037] The target energy spectrum image is processed to obtain elemental peak information; wherein, the elemental peak information includes the position, intensity and type of each element peak;

[0038] The effective element peaks among all the element peaks are determined based on a set peak-to-valley ratio threshold.

[0039] The target energy spectrum image is divided into blocks to obtain at least two energy spectrum image sub-blocks;

[0040] The target energy spectrum segmentation threshold of the energy spectrum image sub-block is determined based on the effective element peaks and the signal intensities corresponding to the pixels in the energy spectrum image sub-block.

[0041] Based on the target energy spectrum segmentation threshold of the energy spectrum image sub-block, the energy spectrum image sub-block is segmented to obtain the segmentation result of the energy spectrum image sub-block.

[0042] As a further improvement to this application, the processing of the target energy spectrum image to obtain elemental peak information includes:

[0043] Peak identification processing is performed on the target energy spectrum image to determine the intensity and position of the peaks;

[0044] Obtain an element energy database, which includes multiple element peaks and the corresponding energy values ​​of the element peaks;

[0045] When the intensity of the peak is within a preset range based on the energy value in the elemental energy database, the peak is identified as the elemental peak corresponding to the energy value, so as to obtain the type of each elemental peak;

[0046] And / or,

[0047] The step of determining the target energy spectrum segmentation threshold of the energy spectrum image sub-block based on the signal intensity corresponding to the effective element peak and the pixel point in the energy spectrum image sub-block includes:

[0048] The initial energy spectrum segmentation threshold is determined based on the signal intensity corresponding to the effective element peaks in the energy spectrum image sub-block;

[0049] Based on the initial energy spectrum segmentation threshold, an energy spectrum threshold range is determined; wherein, the energy spectrum threshold range includes at least one optional energy spectrum segmentation threshold, and the at least one optional energy spectrum segmentation threshold includes the initial energy spectrum segmentation threshold;

[0050] The energy spectrum image sub-block is segmented based on each of the optional energy spectrum segmentation thresholds to obtain the corresponding current energy spectrum sub-block segmentation result and calculate the corresponding inter-class variance.

[0051] The optional energy spectrum threshold at which the inter-class variance is maximized is determined as the target energy spectrum segmentation threshold.

[0052] As a further improvement to this application, the step of determining the effective element peaks among the element peaks based on a set peak-to-valley ratio threshold includes:

[0053] Calculate the ratio of the peak height of the element peak to the adjacent valley, and determine the element peak as the effective element peak when the ratio is greater than the set peak-valley ratio threshold;

[0054] The adjacent valley is the valley adjacent to the element peak.

[0055] Based on the same inventive concept, this application also discloses a FIB detection system, comprising: a FIB device and a detection device coupled to the FIB device, wherein the FIB device is used to detect the sample to be detected, obtaining a target SEM image and a target energy dispersive spectroscopy image; the detection device includes:

[0056] The acquisition module acquires the target SEM image and target energy spectrum image corresponding to the sample to be detected, wherein the sample to be detected includes at least two stacked film layers;

[0057] The segmentation module segments the target SEM image to obtain a first SEM segmentation image and segments the target energy spectrum image to obtain an energy spectrum segmentation image.

[0058] The determination module determines the sample structure information of the sample to be detected based on the first SEM segmentation image and the energy spectrum segmentation image when the film segmentation result in the energy spectrum segmentation image is different from the film segmentation result in the first SEM segmentation image.

[0059] The sample structure information includes at least the size information of each of the membrane layers.

[0060] Compared with existing technologies, the beneficial effects of this application are as follows: In this application, a target SEM image and a target energy dispersive spectroscopy (EDS) image corresponding to the sample to be tested are acquired; the target SEM image is segmented to obtain a first SEM segmentation image, and the target EDS image is segmented to obtain an EDS segmentation image; when the film segmentation results in the first SEM segmentation image and the film segmentation results in the EDS segmentation image are different, the sample structure information of the sample to be tested is determined based on the first SEM segmentation image and the EDS segmentation image. Compared with the scheme that only uses SEM images to determine sample structure information, this scheme combines the target SEM image and the target EDS image to determine the sample structure information of the sample to be tested, avoiding the error problem caused by relying solely on the target SEM image (such as the charge accumulation due to photoresist in the target SEM image, or the unrecognizable situation due to the superposition of silicon nitride and silicon oxide in the target SEM image). The first SEM segmentation image can be verified by the EDS segmentation image to further ensure the accuracy of the film segmentation results, and thus ultimately ensure the accuracy of the sample structure information of the sample to be tested. Attached Figure Description

[0061] Figure 1 This is a schematic diagram illustrating the steps of a sample structure detection method based on a FIB device as shown in this application.

[0062] Figure 2 for Figure 1Schematic diagram of the steps for segmenting the target SEM image to obtain the first SEM segmentation image;

[0063] Figure 3 For Figure 2 Specific flowchart for determining the target SEM segmentation threshold of the SEM image sub-block according to the gray values respectively corresponding to the pixel points in each SEM image sub-block;

[0064] Figure 4 For Figure 1 Schematic diagram of the steps for segmenting the target energy spectrum image to obtain the energy spectrum segmentation image;

[0065] Figure 5 For Figure 4 Schematic diagram of the steps for determining the target energy spectrum segmentation threshold of the energy spectrum image sub-block according to the effective element peaks and the signal intensities corresponding to the pixel points in the energy spectrum image sub-block;

[0066] Figure 6 Topological diagram of a FIB detection system shown in the present application. Detailed implementation manners

[0067] The present application will be described in detail below in conjunction with the embodiments shown in the accompanying drawings. It should be noted, however, that these embodiments are not limitations to the present application, and any equivalent transformation or substitution in terms of function, method, or structure made by those of ordinary skill in the art based on these embodiments shall fall within the protection scope of the present application. <关于本申请的详细说明,结合附图所示的各实施方式进行,但这些实施方式并非对本申请的限制,普通技术人员基于这些实施方式所作的功能、方法、或者结构上的等效变换或替代,均属于本申请的保护范围之内。>

[0068] Please refer Figures 1 to 5 As shown, a detailed implementation manner of a sample structure detection method based on a FIB device is shown in the present application. This sample structure detection method is used for performing composition analysis and dimension measurement on the structure profile of the sample to be detected, and specifically combines the SEM image and the energy spectrum image collected by the FIB device for the sample to ensure the accuracy of the composition analysis and dimension measurement.

[0069] Refer Figure 1 As shown, a sample structure detection method based on a FIB device includes the following steps S1 to S3.

[0070] Step S1, obtain the target SEM image and the target energy spectrum image corresponding to the sample to be detected.

[0071] The sample to be detected is cut and processed by the ion beam in the FIB device, and then the surface of the sample to be detected is scanned by the electron beam in the FIB device to collect the SEM image and the energy spectrum image corresponding to the sample to be detected. Among them, the FIB device can be, for example, a FIB-SEM device or a FIB-TEM device, etc., which can collect the SEM image and the energy spectrum image corresponding to the sample to be detected. This embodiment does not make specific limitations thereto.

[0072] More specifically, the FIB-SEM device is used as an example for illustration. The electron beam in the FIB-SEM device scans the surface of the sample to be tested, detecting secondary electrons (SE) or backscattered electrons (BSE) to generate a SEM image. The electron beam in the FIB-SEM device scans the surface of the sample to be tested, exciting characteristic X-rays, which are then received by an EDS detector to generate an energy dispersive spectroscopy (EDS) image.

[0073] It should be noted that the sample to be tested can be, for example, a chip. Specific locations on the chip (such as transistors, interconnects, and defect points) are etched at the nanoscale using an ion beam in a FIB device to create cross-sections or extract thin films. Simultaneously, the sample to be tested comprises at least two stacked film layers. SEM images are generated by the interaction between the electron beam and the sample to be tested, producing signals (such as secondary electrons or backscattered electrons), and primarily reflect the surface morphology, microstructure, or compositional contrast of the sample. Energy dispersive X-ray spectroscopy (EDS) images are generated by detecting characteristic X-rays excited by an electron beam using an EDS detector, producing images of the types, amounts, and spatial distribution of elements in the sample to be tested.

[0074] Since SEM images and energy dispersive spectroscopy (EDS) images may contain noise and background interference, especially since EDS images are used to display the types and abundance distribution of elements, their pixel values ​​often represent the intensity information of the elements. Different elements have different characteristic peak intensities, resulting in different grayscale distributions in the EDS image. Due to factors such as noise and element overlap, EDS images may exhibit background fluctuations and peak broadening. Therefore, after acquiring the SEM image, it is preprocessed (e.g., smoothing) to remove noise interference and obtain the target SEM image. Similarly, after acquiring the EDS image, it is preprocessed (e.g., smoothing) to remove noise and background interference and highlight element peaks, obtaining the target EDS image. This improves the accuracy of subsequent threshold segmentation.

[0075] It is worth noting that both the target SEM image and the target energy spectrum image are grayscale images; if the obtained SEM image and energy spectrum image are color images, they are converted to grayscale to obtain the corresponding grayscale images.

[0076] Step S2: Segment the target SEM image to obtain the first SEM segmentation image, and segment the target energy spectrum image to obtain the energy spectrum segmentation image.

[0077] In one implementation, the first SEM segmentation image includes the segmentation results of at least two SEM image sub-blocks. (See reference...) Figure 2 As shown, the first SEM segmentation image is obtained by segmenting the target SEM image, including the following steps S21 to S23.

[0078] Step S21: Divide the target SEM image into blocks to obtain at least two SEM image sub-blocks.

[0079] Based on the resolution and structural features of the target SEM image, the target SEM image is divided into at least two SEM image sub-blocks of equal size. If the target SEM image has a high resolution and a complex structure, the divided SEM image sub-blocks are smaller; if the target SEM image has a low resolution and a simple structure, the divided SEM image sub-blocks are larger. This embodiment does not limit the specific size of the sub-blocks.

[0080] Step S22: Determine the target SEM segmentation threshold for each SEM image sub-block based on the grayscale values ​​corresponding to the pixels in each SEM image sub-block.

[0081] It should be noted that the target SEM segmentation threshold in this application is first determined based on the structural complexity of the sample to be tested and prior knowledge, by determining the appropriate number of target SEM segmentation thresholds. For samples with simple structures, two target SEM segmentation thresholds may be sufficient; for samples with complex structures, three or more target SEM segmentation thresholds may be required.

[0082] Furthermore, the threshold value can be determined based on the structure of the sample to be tested. For example, if the observation area is a sandwich area, the gray value of the middle layer structure is used as the background value. The other two layers have a certain contrast difference with the middle layer, so two threshold values ​​are selected.

[0083] In one embodiment, this embodiment is illustrated using an example where the number of target SEM segmentation thresholds for a SEM image sub-block is one. Figure 3 As shown, step S22 specifically includes the following steps S221 to S224.

[0084] Step S221: Based on the grayscale value corresponding to the pixel in each SEM image sub-block, determine the initial SEM segmentation threshold of the SEM image sub-block, and use the initial SEM segmentation threshold as the current SEM segmentation threshold.

[0085] For each SEM image sub-block, the mean (denoted as gray-level mean) and variance (denoted as gray-level variance) of the gray values ​​of all pixels in the SEM image sub-block are calculated. Corresponding weight coefficients are assigned to the gray-level mean and gray-level variance, and then an offset is added to calculate the initial SEM segmentation threshold. The weight coefficients and offsets corresponding to the gray-level mean and gray-level variance are determined manually and can be pre-adjusted through experiments.

[0086] Specifically, the formula for calculating the initial SEM segmentation threshold is:

[0087] T i (0)=α*μ i +β*σ i +γ; where α and β are weighting coefficients, γ is the offset, and μ is the weighting factor. i It is the grayscale mean, σ i It is the grayscale variance, T i (0) It is the initial SEM segmentation threshold.

[0088] Understandably, the gray-scale mean reflects the average brightness of a sub-block in the SEM image, and the gray-scale variance represents the degree of variation in gray-scale values. The calculation schemes for the gray-scale mean and gray-scale variance are the same as those in related technologies, and will not be repeated here.

[0089] Step S222: Segment the SEM image sub-blocks based on the current SEM segmentation threshold to obtain the current SEM sub-block segmentation result. Calculate the new SEM segmentation threshold based on the current SEM sub-block segmentation result.

[0090] Based on the current SEM segmentation threshold, pixels in a SEM image sub-block are divided into foreground (i.e., pixels with gray values ​​greater than the current SEM segmentation threshold) and background (i.e., pixels with gray values ​​less than or equal to the current SEM segmentation threshold). The pixel percentages of the foreground and background pixels relative to all pixels in the SEM image sub-block are calculated, along with the mean and variance of gray values ​​for the foreground and background pixels. Based on this, combined statistics corresponding to the mean and variance of gray values ​​are calculated, and corresponding weighting coefficients are assigned to these combined statistics. An offset is then added to calculate a new SEM segmentation threshold.

[0091] Specifically, the formula for calculating the combined statistic corresponding to the grayscale mean is as follows:

[0092] μ i (k+1) =ρ*μ fg +(1-ρ)*μ bg ;

[0093] The formula for calculating the pooled statistic corresponding to the gray variance is:

[0094] (σ i (k+1) ) 2 =ρ(σ fg 2 +(μ fg -μ i( k +1)) 2 )+(1-ρ)(σ bg 2 +(μ bg -μi (k+1) ) 2 );

[0095] The formula for calculating the new SEM segmentation threshold is:

[0096] T i (k+1) =α*μ i (k+1) +β*σ i (k+1) +γ;

[0097] Where ρ is the proportion of foreground pixels, (1-ρ) is the proportion of background pixels, and μ fg It is the mean gray level of the foreground, μ bg It is the average grayscale value of the background, σ fg 2 It is the foreground variance, σ bg 2 It is the background variance, μ i (k+1) It is the combined statistic corresponding to the grayscale mean, σ i (k+1) It is the pooled statistic corresponding to the gray variance, T i (k+1) It is the new SEM segmentation threshold.

[0098] Step S223: Calculate the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold.

[0099] Step S224: Determine whether the absolute difference is less than the preset threshold.

[0100] The preset threshold is a positive value, and the preset threshold can be set according to specific application needs.

[0101] If yes (i.e., the absolute difference is less than the preset threshold), then proceed to step S225; if no (i.e., the absolute difference is greater than or equal to the preset threshold), then proceed to step S226.

[0102] Step S225: Use the new SEM segmentation threshold as the target SEM segmentation threshold.

[0103] Step S226: Use the new SEM segmentation threshold as the current segmentation threshold.

[0104] When the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is greater than or equal to a preset threshold, the new SEM segmentation threshold is used as the current SEM segmentation threshold, and the process returns to step S222 (i.e., the SEM image sub-blocks are segmented based on the current SEM segmentation threshold to obtain the current SEM sub-block segmentation result), until the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is less than the preset threshold, at which point the new SEM segmentation threshold is used as the target SEM segmentation threshold.

[0105] For example, the current SEM segmentation threshold is A1, and the new SEM segmentation threshold is A2. If the absolute difference between A1 and A2 is less than a preset threshold, then A2 is used as the target SEM segmentation threshold. If the absolute difference between A1 and A2 is greater than or equal to the preset threshold, then A2 is used as the current SEM segmentation threshold, and a new SEM segmentation threshold A3 is calculated based on A2. If the absolute difference between A2 and A3 is less than the preset threshold, then A3 is used as the target SEM segmentation threshold. If the absolute difference between A2 and A3 is greater than or equal to the preset threshold, then A3 is used as the current SEM segmentation threshold, and a new SEM segmentation threshold A4 is calculated based on A3, and so on.

[0106] It should be noted that the initial SEM segmentation threshold is calculated based on the grayscale value of the pixels in each SEM image sub-block. However, since the grayscale values ​​are often similar near the boundary between two layers, the boundary may be unclear. If the initial SEM segmentation threshold is directly used as the target SEM segmentation threshold and the SEM image sub-blocks are segmented based on it, the final SEM image sub-block segmentation result will have a large error. Therefore, in this application, the initial SEM segmentation threshold is first used as the current SEM segmentation threshold. Then, the SEM image is segmented based on the current SEM segmentation threshold to obtain the current SEM sub-block segmentation result. A new SEM segmentation threshold is calculated based on the current SEM sub-block segmentation result. This process is repeated until the threshold converges (i.e., the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is less than a preset threshold), obtaining the optimal SEM segmentation threshold. The obtained optimal SEM segmentation threshold is used as the target SEM segmentation threshold, thereby ensuring the accuracy of subsequent SEM image sub-block segmentation based on the target SEM segmentation threshold.

[0107] Step S23: Based on the target SEM segmentation threshold of the SEM image sub-block, segment the SEM image sub-block to obtain the segmentation result of the SEM image sub-block.

[0108] When the target SEM segmentation threshold is 1, the first SEM segmented image is a binary image. Specifically, based on the target SEM segmentation threshold, the pixels in the SEM image sub-block are divided into two categories: pixels with gray values ​​greater than the target SEM segmentation threshold are classified into one category, and pixels with gray values ​​less than the target SEM segmentation threshold are classified into another category, thus obtaining the segmentation result of the SEM image sub-block.

[0109] In one implementation, after obtaining the segmentation results of SEM image sub-blocks, a first SEM segmentation image is constructed from the segmentation results of all SEM image sub-blocks. Morphological processing is then performed on the first SEM segmentation image to make the structures in the first SEM segmentation image clearer and more complete, thereby further optimizing the segmentation results. Furthermore, contour extraction can be performed on the first SEM segmentation image to obtain the boundary contour of the sample to be detected, and then the sample structural information can be measured.

[0110] Morphological processing includes, but is not limited to, erosion, dilation, opening, and closing operations. Erosion can remove small noise points and burrs from an image; dilation can fill holes and gaps in an image; opening operations, which are performed by first erosion and then dilation, can remove small interfering objects; closing operations, which are performed by first dilation and then erosion, can connect broken objects.

[0111] In one embodiment, the energy spectrum segmentation image includes the segmentation results of at least two energy spectrum image sub-blocks. This embodiment is illustrated using a segmentation threshold of 1 as an example. Figure 4 As shown, the target energy spectrum image is segmented to obtain an energy spectrum segmentation image, including the following steps S24 to S28.

[0112] Step S24: Process the target energy spectrum image to obtain elemental peak information.

[0113] Elemental peak information includes the position, intensity, and type of each element peak.

[0114] In one embodiment, step S24 specifically includes steps S241 to S243.

[0115] S241. Perform peak identification processing on the target energy spectrum image to determine the intensity and position of the peaks.

[0116] Peak identification processing of the target energy spectrum image can be achieved through the derivative method. The first derivative is used to determine the position of the peak, and the second derivative is used to distinguish the peak from noise, thereby determining the intensity and position of the peak.

[0117] S242, Obtain the elemental energy database.

[0118] The elemental energy database includes various elemental peaks and their corresponding energy values. For example, the Kα and Lα lines in EDS, the binding energy in XPS, etc. The energy value of the Kα line for carbon is 277 eV, and the energy value of the Kα line for iron is 6.4 keV.

[0119] S243. When the intensity of a peak is within a preset range based on the energy value in the element energy database, the peak is identified as the element peak corresponding to the energy value, and the type of each element peak is obtained.

[0120] A preset range, such as ±0.1 keV, is formed based on the energy value. Taking the Kα line of iron as an example, the energy value of the Kα line of iron is 6.4 keV, so the corresponding preset range is 6.43 keV-6.5 keV. If the intensity of the peak is between 6.43 keV and 6.5 keV, then the peak is identified as an iron element peak, and the element peak type is determined to be iron.

[0121] It should be noted that if overlapping peaks appear in the target energy spectrum, the overlapping peaks are fitted and separated (e.g., using Gaussian or Lorentz functions), the error of the EDS detector is calibrated, and the intensity of the overlapping peaks is adjusted using the determined element peaks. Any method in the prior art can be used, and this embodiment will not be elaborated here.

[0122] Step S25: Determine the effective element peaks among all element peaks based on the set peak-to-valley ratio threshold.

[0123] The ratio of the peak height of an element peak to its adjacent valley is calculated. If the ratio is greater than a set peak-to-valley ratio threshold, the element peak is considered a true distribution and is identified as a valid element peak. If the ratio is less than or equal to the set peak-to-valley ratio threshold, the element peak is considered noise. Adjacent valleys are those adjacent to the element peak.

[0124] It should be noted that the peak-to-valley ratio threshold is set manually and can be obtained through Monte Carlo simulation or statistical analysis of actual data. For example, a peak-to-valley ratio threshold of PVRT = 1.5-2.5 can balance sensitivity and specificity in most scenarios; PVR = 1.5 (i.e., indicating that the peak intensity is 50% higher than the background) is suitable for high signal-to-noise ratio and narrow peaks (such as XPS main peak); PVR = 2.5 (i.e., indicating that the peak intensity is 150% higher than the background) is suitable for low signal-to-noise ratio and wide peaks (such as EDS overlapping peaks).

[0125] Step S26: Divide the target energy spectrum image into blocks to obtain at least two energy spectrum image sub-blocks.

[0126] Similar to the aforementioned segmentation of the target SEM image, the target energy spectrum image is divided into at least two energy spectrum image sub-blocks of equal size based on the structural characteristics of the target energy spectrum image. This embodiment will not be elaborated further here.

[0127] Step S27: Determine the target energy spectrum segmentation threshold of the energy spectrum image sub-block based on the signal intensity corresponding to the effective element peaks and the pixels in the energy spectrum image sub-block.

[0128] Each energy spectrum image sub-block corresponds one-to-one with the target energy spectrum segmentation threshold.

[0129] In one implementation, the reference Figure 5 As shown, step S27 includes steps S271 to S274.

[0130] Step S271: Determine the initial energy spectrum segmentation threshold based on the signal intensity corresponding to the effective element peaks in the energy spectrum image sub-blocks.

[0131] Step S272: Determine the energy spectrum threshold range based on the initial energy spectrum segmentation threshold.

[0132] The energy spectrum threshold range includes at least one optional energy spectrum segmentation threshold, and the at least one optional energy spectrum segmentation threshold includes an initial energy spectrum segmentation threshold.

[0133] In one implementation, for each effective element peak in a sub-block of the energy spectrum image, the value of the signal intensity corresponding to that effective element peak within a certain range is used as an optional energy spectrum segmentation threshold. For example, if the signal intensity corresponding to the effective element peak is D, all values ​​within [D-D1, D+D2] are used as optional energy spectrum segmentation thresholds (the difference between two adjacent optional energy spectrum segmentation thresholds is d). Subsequently, the inter-class variance corresponding to all optional energy spectrum segmentation thresholds is calculated, and the optional energy spectrum segmentation threshold with the largest inter-class variance is determined as the target energy spectrum segmentation threshold. Here, D1 < D, and the specific values ​​of D, D1, and D2 can be set according to application needs; this embodiment does not limit this.

[0134] In another implementation, the signal intensities corresponding to all valid element peaks in a single energy spectrum image sub-block are used as optional energy spectrum segmentation thresholds. Subsequently, the inter-class variances corresponding to all optional energy spectrum segmentation thresholds are calculated, and the optional energy spectrum segmentation threshold with the largest inter-class variance is determined as the target energy spectrum segmentation threshold.

[0135] Understandably, if a single energy spectrum image sub-block does not contain any valid element peaks, then the pixels in that energy spectrum image sub-block are classified into one category.

[0136] Step S273: Segment the energy spectrum image sub-blocks based on each selectable energy spectrum segmentation threshold, obtain the corresponding current energy spectrum sub-block segmentation result, and calculate the corresponding inter-class variance.

[0137] The pixels in the energy spectrum image sub-block are divided into foreground (i.e., pixels in the energy spectrum image sub-block with gray values ​​greater than the optional energy spectrum segmentation threshold) and background (i.e., pixels in the energy spectrum image sub-block with gray values ​​less than or equal to the optional energy spectrum segmentation threshold) according to the optional energy spectrum segmentation threshold. The pixel proportions of the pixels in the foreground and the pixels in the background relative to the pixels in the energy spectrum image sub-block are calculated, and the mean gray value of the gray values ​​corresponding to the pixels in the foreground and the mean gray value of the gray values ​​corresponding to the pixels in the background are calculated.

[0138] The inter-class variance is calculated based on the pixel proportion and mean grayscale value of the foreground and the pixel proportion and mean grayscale value of the background. Specifically, the formula for calculating the inter-class variance is:

[0139] σ 2 (k)=ω0(k)ω1(k)[μ0(k)-μ1(k)] 2 ;

[0140] Where k is the optional energy spectrum segmentation threshold, ω0(k) is the proportion of foreground pixels, ω1(k) is the proportion of background pixels, μ0(k) is the average gray level of the foreground, μ1(k) is the average gray level of the background, and σ 2 (k) is the between-class variance.

[0141] Step S274: Determine the selectable energy spectrum segmentation threshold when the inter-class variance is maximized as the target energy spectrum segmentation threshold.

[0142] Based on the above steps S271 to S273, the inter-class variance corresponding to each effective element peak is calculated, and the intensity corresponding to the effective element peak with the largest inter-class variance is used as the target energy spectrum segmentation threshold.

[0143] Specifically, the formula for calculating the target energy spectrum segmentation threshold is as follows:

[0144] k* = argmax k σ 2 (k).

[0145] Where k* is the target energy spectrum segmentation threshold.

[0146] Step S28: Based on the target energy spectrum segmentation threshold of the energy spectrum image sub-block, segment the energy spectrum image sub-block to obtain the segmentation result of the energy spectrum image sub-block.

[0147] Based on the target energy spectrum segmentation threshold, the pixels in the energy spectrum image sub-block are divided into two categories: pixels with signal intensity greater than the target energy spectrum segmentation threshold are classified into one category, and pixels with signal intensity less than the target energy spectrum segmentation threshold are classified into another category, thus obtaining the segmentation result of the energy spectrum image sub-block.

[0148] It should be noted that for multi-element target energy spectrum images, the above threshold segmentation method may need to be applied multiple times, setting thresholds for each element peak to accurately segment different element regions.

[0149] Step S3: When the membrane segmentation results in the first SEM segmentation image and the membrane segmentation results in the energy dispersive spectroscopy (EDS) segmentation image are different, determine the sample structure information of the sample to be tested based on the first SEM segmentation image and the EDS segmentation image.

[0150] In one embodiment, when the total number of film layers corresponding to the first SEM segmentation image is greater than the total number of film layers corresponding to the energy dispersive spectroscopy (EDS) segmentation image at the same detection location of the sample to be tested, the film layer segmentation results in the first SEM segmentation image are adjusted according to the film layer segmentation results in the EDS segmentation image to obtain a second SEM segmentation image; based on the pixel ratio of each film layer in the second SEM segmentation image and the actual size of the sample to be tested, the size information of each film layer is calculated. The total number of film layer segments in the second SEM segmentation image is the same as the total number of film layer segments in the EDS segmentation image.

[0151] It should be noted that when a non-conductive layer structure such as photoresist is present in the sample to be tested, charge accumulation occurs. This layer appears abnormally bright in the target SEM image. Therefore, when segmenting the target SEM image, it is processed into two different structures, while the target energy dispersive spectroscopy (EDS) image only shows one structure after segmentation. Based on this, the film layer segmentation result in the first SEM segmentation image is adjusted to obtain the second SEM segmentation image. That is, the two-layer structure in the first SEM segmentation image is adjusted to a single structure to correct the first SEM segmentation image and obtain the second SEM segmentation image. This ensures the accuracy of the film layer segmentation result and the accuracy of the final calculation of the size information of each film layer.

[0152] In another implementation, the following approach can be adopted:

[0153] (a) When the total number of film layers corresponding to the first SEM segmentation image is less than the total number of film layers corresponding to the energy spectrum segmentation image at the same detection location of the sample to be tested, a new target SEM image corresponding to the sample to be tested is re-acquired, and the new target SEM image is segmented to obtain the second SEM segmentation image.

[0154] (b) When the total number of film layers in the second SEM segmentation image is equal to the total number of film layer segments in the energy spectrum segmentation image at the same detection location of the sample to be tested, the size information of the film layer is calculated based on the pixel ratio of each film layer in the second SEM segmentation image and the actual size of the sample to be tested.

[0155] (c) When the total number of film layers corresponding to the second SEM segmentation image is less than the total number of film layers corresponding to the energy spectrum segmentation image at the same detection location of the sample to be detected, the size information of the film layer is calculated based on the pixel ratio of each film layer to be distinguished in the energy spectrum segmentation image and the actual size of the sample to be detected, and the size information of the remaining film layer is calculated based on the pixel ratio of each remaining film layer in the second SEM segmentation image and the actual size of the sample to be detected.

[0156] The film layer to be distinguished is film layer i in the second SEM segmentation image and film layers j to j+n in the energy spectrum segmentation image. The remaining film layers are the film layers other than the film layer to be distinguished. 1≤i≤k1<k2, 1≤j≤k2, n≥1, k1 is the total number of film layers corresponding to the second SEM segmentation image, and k2 is the total number of film layers corresponding to the energy spectrum segmentation image.

[0157] Understandably, since the target SEM image misidentifies a multilayer structure (i.e., a higher-level concept of the aforementioned two-layer structure) as a single-layer structure, the film layer to be distinguished is a single-layer structure (i.e., film layer i) in the second SEM segmentation image, while it is a multilayer structure (i.e., film layers j to j+n) in the energy spectrum segmentation image.

[0158] It should be noted that when the sample contains two non-conductive layers, silicon nitride and silicon oxide, stacked together, the normal target SEM image cannot distinguish them. Therefore, during the segmentation of the target SEM image, this is treated as a single-layer structure, while the target energy dispersive spectroscopy (EDS) image reveals two layers. Based on this, the operator readjusts the imaging conditions for this part of the structure (e.g., changing parameters such as current or voltage) to enable the identification of the two-layer structure in the new target SEM image. The new target SEM image is then segmented to obtain a second SEM segmentation image. If changing the imaging conditions still fails to identify the two-layer structure in the new target SEM image, then the size information of the film layer is calculated based on the EDS segmentation image for this part of the structure, and the size information of the film layer for the remaining part of the structure is calculated based on the second SEM segmentation image. This ensures the accuracy of the film layer segmentation results and guarantees the accuracy of the final calculated size information for each film layer.

[0159] For example, suppose the sample to be tested consists of five stacked film layers, with the second film layer being a silicon nitride layer and the third film layer being a silicon oxide layer. The total number of film layers corresponding to the second SEM segmentation image is 4 (the second and third film layers are identified as one layer, denoted as film layers A1 to A4, respectively), and the total number of film layers corresponding to the energy dispersive spectroscopy (EDS) segmentation image is 5 (denoted as film layers B1 to B5, respectively). In this case, the film layers to be distinguished are film layers B2 and B3, and their size information is calculated based on the EDS segmentation image; the remaining film layers are film layers A1, A3, and A4, and their size information is calculated based on the second SEM segmentation image.

[0160] In one embodiment, the above-described sample structure detection method further includes the following step S4.

[0161] Step S4: When the membrane segmentation results in the first SEM segmentation image and the membrane segmentation results in the energy dispersive spectroscopy (EDS) segmentation image are the same, calculate the size information of the membrane layer based on the pixel ratio of each membrane layer in the first SEM segmentation image and the actual size of the sample to be detected.

[0162] In one implementation, the formula for calculating the size information of the film layer is:

[0163] D i =L*P i ;

[0164] Where L is the actual size of the sample to be tested, and P i It refers to the pixel ratio of the film layer.

[0165] Specifically, assuming the length of the sample to be tested is L1, and the proportion of film layer i in the length direction is P. i1 Then the actual length of film layer i is L1*P i1 Assume the width of the sample to be tested is L2, and the proportion of film layer i in the width direction is P. i2 Then the actual width of film layer i is L2*P i2 .

[0166] In summary, this application first preprocesses the detected SEM image and energy spectrum image to obtain the target SEM image and target energy spectrum image corresponding to the sample to be detected. The grayscale histogram corresponding to the target SEM image is calculated, and the target SEM image is segmented using a multi-threshold segmentation algorithm to obtain a first SEM segmentation image. Contour extraction is performed on the first SEM segmentation image, and threshold segmentation is performed multiple times using an adaptive threshold segmentation algorithm (e.g., local Otsu's algorithm) to segment the target energy spectrum image into distribution regions of different elements, resulting in an energy spectrum segmentation image. Based on the element types and distributions determined in the target energy spectrum image, the first SEM segmentation image is combined with the first SEM segmentation image of the target SEM image to verify and optimize the first SEM segmentation image. In multiphase materials, the boundaries and composition of different phases can be determined by elemental distribution, correcting potential misjudgment areas in the first SEM segmentation image. Meanwhile, if abnormal brightness is caused by charge accumulation in a cross-sectional area when the sample to be tested has a multilayer non-conductive structure (e.g., photoresist layer, silicon oxide layer, and silicon nitride layer stacked together), the first SEM segmentation image and energy dispersive spectroscopy (EDS) segmentation image are optimized for verification. Based on the processed and verified grayscale binary image, intelligent size measurement of the sample to be tested is performed.

[0167] Specifically, when the film segmentation results in the first SEM segmentation image and the film segmentation results in the energy dispersive spectroscopy (EDS) segmentation image differ, the sample structure information of the sample to be tested is determined based on both the first SEM segmentation image and the EDS segmentation image. By combining the target SEM image and the target EDS image, the sample structure information of the sample to be tested is determined. This avoids errors caused by relying solely on the target SEM image due to certain special circumstances (such as charge accumulation due to photoresist in the target SEM image, or unrecognizable situations due to the superposition of silicon nitride and silicon oxide in the target SEM image). The first SEM segmentation image can be verified through the EDS segmentation image to further ensure the accuracy of the film segmentation results, and ultimately ensure the accuracy of the sample structure information of the sample to be tested.

[0168] Based on the same inventive concept, this application also discloses a FIB detection system 100. (See reference...) Figure 6 As shown, the FIB detection system 100 includes: a FIB device 10 and a detection device 20 coupled to the FIB device 10. The FIB device 10 is used to detect the sample to be tested, obtaining a target SEM image and a target energy dispersive spectroscopy (EDS) image. The detection device 20 includes an acquisition module 21, a segmentation module 22, and a determination module 23. The acquisition module 21 acquires the target SEM image and the target EDS image corresponding to the sample to be tested, wherein the sample to be tested includes at least two stacked film layers; the segmentation module 22 segments the target SEM image to obtain a first SEM segmentation image, and segments the target EDS image to obtain an EDS segmentation image; when the film layer segmentation result in the EDS segmentation image differs from the film layer segmentation result in the first SEM segmentation image, the determination module 23 determines the sample structure information of the sample to be tested based on the first SEM segmentation image and the EDS segmentation image; wherein the sample structure information includes at least the size information of each film layer.

[0169] The FIB device can be, for example, a FIB-SEM device or a FIB-TEM device, which can acquire SEM images and energy spectrum images of the sample to be tested. This embodiment does not specifically limit this.

[0170] It should be noted that step S1 in the sample structure detection method based on FIB equipment is implemented by the acquisition module 21 in the FIB detection system 100, step S2 in the sample structure detection method based on FIB equipment is implemented by the segmentation module 22 in the FIB detection system 100, and step S3 in the sample structure detection method based on FIB equipment is implemented by the determination module 23 in the FIB detection system 100. The specific technical solutions can be found above and will not be repeated here.

[0171] The detailed descriptions listed above are merely specific descriptions of feasible implementation methods of this application and are not intended to limit the scope of protection of this application. All equivalent implementation methods or modifications made without departing from the spirit of the art of this application should be included within the scope of protection of this application.

[0172] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from the spirit or essential characteristics of this application. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of this application is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within this application. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0173] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A sample structure detection method based on FIB equipment, characterized in that, include: Acquire the target SEM image and target energy spectrum image of the sample to be tested, wherein the sample to be tested comprises at least two stacked film layers; The target SEM image is segmented to obtain a first SEM segmentation image, and the target energy spectrum image is segmented to obtain an energy spectrum segmentation image; When the membrane segmentation results in the first SEM segmentation image and the membrane segmentation results in the energy dispersive spectroscopy (EDS) segmentation image are the same, the size information of the membrane is calculated based on the pixel ratio of each membrane in the first SEM segmentation image and the actual size of the sample to be detected. When the membrane segmentation results in the first SEM segmentation image and the membrane segmentation results in the energy dispersive spectroscopy (EDS) segmentation image are different, the sample structure information of the sample to be detected is determined based on the first SEM segmentation image and the EDS segmentation image. The sample structure information includes at least the size information of each of the membrane layers; determining the sample structure information of the sample to be tested based on the first SEM segmentation image and the energy dispersive spectroscopy (EDS) segmentation image includes: When the total number of film layers corresponding to the first SEM segmentation image is greater than the total number of film layers corresponding to the energy dispersive spectroscopy (EDS) segmentation image at the same detection location of the sample to be tested, the film layer segmentation results in the first SEM segmentation image are adjusted according to the film layer segmentation results in the EDS segmentation image to obtain a second SEM segmentation image. Based on the pixel ratio of each film layer in the second SEM segmentation image and the actual size of the sample to be tested, the size information of each film layer is calculated; the total number of film layer segments in the second SEM segmentation image is the same as the total number of film layer segments in the EDS segmentation image. When the total number of film layers corresponding to the first SEM segmentation image is less than the total number of film layers corresponding to the energy dispersive spectroscopy segmentation image at the same detection location of the sample to be tested, a new target SEM image corresponding to the sample to be tested is re-acquired, and the new target SEM image is segmented to obtain a second SEM segmentation image. When the total number of film layers corresponding to the second SEM segmentation image is equal to the total number of film layers corresponding to the energy dispersive spectroscopy segmentation image at the same detection location of the sample to be tested, the size information of the film layer is calculated based on the pixel ratio of each film layer in the second SEM segmentation image and the actual size of the sample to be tested. When the total number of film layers corresponding to the second SEM segmentation image is less than the total number of film layers corresponding to the energy dispersive spectroscopy (EDS) segmentation image at the same detection location of the sample to be tested, the size information of the film layers is calculated based on the pixel ratio of each film layer to be distinguished in the EDS segmentation image and the actual size of the sample to be tested. The size information of the remaining film layers is also calculated based on the pixel ratio of each remaining film layer in the second SEM segmentation image and the actual size of the sample to be tested. The film layer to be distinguished is film layer i in the second SEM segmentation image and film layers j to j+n in the EDS segmentation image. The remaining film layers are film layers other than the film layer to be distinguished. 1≤i≤k1<k2, 1≤j≤k2, n≥1, k1 is the total number of film layers corresponding to the second SEM segmentation image, and k2 is the total number of film layers corresponding to the EDS segmentation image.

2. The sample structure detection method according to claim 1, characterized in that, The first SEM segmentation image includes the segmentation results of at least two SEM image sub-blocks; The step of segmenting the target SEM image to obtain a first SEM segmentation image includes: The target SEM image is divided into blocks to obtain at least two SEM image sub-blocks; The target SEM segmentation threshold of each SEM image sub-block is determined based on the grayscale value corresponding to the pixel in each SEM image sub-block. Based on the target SEM segmentation threshold of the SEM image sub-block, the SEM image sub-block is segmented to obtain the segmentation result of the SEM image sub-block.

3. The sample structure detection method according to claim 2, characterized in that, Determining the target SEM segmentation threshold for each SEM image sub-block based on its grayscale value includes: Based on the grayscale values ​​corresponding to the pixels in each SEM image sub-block, an initial SEM segmentation threshold for the SEM image sub-block is determined, and the initial SEM segmentation threshold is used as the current SEM segmentation threshold. The SEM image sub-blocks are segmented based on the current SEM segmentation threshold to obtain the current SEM sub-block segmentation result. Based on the current SEM sub-block segmentation result, a new SEM segmentation threshold is calculated. When the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is less than a preset threshold, the new SEM segmentation threshold is used as the target SEM segmentation threshold. When the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is greater than or equal to the preset threshold, the new SEM segmentation threshold is used as the current SEM segmentation threshold, and the step of segmenting the SEM image sub-block based on the current SEM segmentation threshold to obtain the current SEM sub-block segmentation result is returned, until the absolute difference between the current SEM segmentation threshold and the new SEM segmentation threshold is less than the preset threshold, and the new SEM segmentation threshold is used as the target SEM segmentation threshold.

4. The sample structure detection method according to claim 1, characterized in that, The energy spectrum segmentation image includes the segmentation results of at least two energy spectrum image sub-blocks; The step of segmenting the target energy spectrum image to obtain an energy spectrum segmentation image includes: The target energy spectrum image is processed to obtain elemental peak information; wherein, the elemental peak information includes the position, intensity and type of each element peak; The effective element peaks among all the element peaks are determined based on a set peak-to-valley ratio threshold. The target energy spectrum image is divided into blocks to obtain at least two energy spectrum image sub-blocks; The target energy spectrum segmentation threshold of the energy spectrum image sub-block is determined based on the effective element peaks and the signal intensities corresponding to the pixels in the energy spectrum image sub-block. Based on the target energy spectrum segmentation threshold of the energy spectrum image sub-block, the energy spectrum image sub-block is segmented to obtain the segmentation result of the energy spectrum image sub-block.

5. The sample structure detection method according to claim 4, characterized in that, The process of processing the target energy spectrum image to obtain elemental peak information includes: Peak identification processing is performed on the target energy spectrum image to determine the intensity and position of the peaks; Obtain an element energy database, which includes multiple element peaks and the corresponding energy values ​​of the element peaks; When the intensity of the peak is within a preset range based on the energy value in the elemental energy database, the peak is identified as the elemental peak corresponding to the energy value, so as to obtain the type of each elemental peak; The step of determining the target energy spectrum segmentation threshold of the energy spectrum image sub-block based on the signal intensity corresponding to the effective element peak and the pixel point in the energy spectrum image sub-block includes: The initial energy spectrum segmentation threshold is determined based on the signal intensity corresponding to the effective element peaks in the energy spectrum image sub-block; Based on the initial energy spectrum segmentation threshold, an energy spectrum threshold range is determined; wherein, the energy spectrum threshold range includes at least one optional energy spectrum segmentation threshold, and the at least one optional energy spectrum segmentation threshold includes the initial energy spectrum segmentation threshold; The energy spectrum image sub-block is segmented based on each of the optional energy spectrum segmentation thresholds to obtain the corresponding current energy spectrum sub-block segmentation result and calculate the corresponding inter-class variance. The optional energy spectrum threshold at which the inter-class variance is maximized is determined as the target energy spectrum segmentation threshold.

6. The sample structure detection method according to claim 5, characterized in that, The determination of effective element peaks among the element peaks based on a set peak-to-valley ratio threshold includes: Calculate the ratio of the peak height of the element peak to the adjacent valley, and determine the element peak as the effective element peak when the ratio is greater than the set peak-valley ratio threshold; The adjacent valley is the valley adjacent to the element peak.

7. A FIB detection system, characterized in that, include: FIB device and detection device coupled to the FIB device, the FIB device being used to detect the sample to be tested, to obtain target SEM image and target energy spectrum image; The detection device is applied to the sample structure detection method based on FIB equipment as described in any one of claims 1 to 6, and the detection device comprises: The acquisition module acquires the target SEM image and target energy spectrum image corresponding to the sample to be detected, wherein the sample to be detected includes at least two stacked film layers; The segmentation module segments the target SEM image to obtain a first SEM segmentation image and segments the target energy spectrum image to obtain an energy spectrum segmentation image. The determination module determines the sample structure information of the sample to be detected based on the first SEM segmentation image and the energy spectrum segmentation image when the film segmentation result in the energy spectrum segmentation image is different from the film segmentation result in the first SEM segmentation image. The sample structure information includes at least the size information of each of the membrane layers.

Images