Etching Shape Evaluation System and Method

Abstract

The purpose of this invention is to provide a technique to improve the accuracy and efficiency of etched shape evaluation. An etched shape evaluation system includes a computer device comprising a processing unit and a data storage unit; the data storage unit stores a basic model image; the processing unit acquires one or more fabrication cross-sectional observation images taken with a scanning electron microscope; a grid pattern is applied to the fabrication cross-sectional observation image; the width of the cross-sectional outline in the region of the grid pattern corresponding to each depth position is defined by the acquired source image; the source images are captured and combined to create a montage image; the data storage unit stores the montage image as a basic model image. The processing unit acquires one or more evaluation cross-sectional observation images and determines the similarity between the basic model image and the evaluation cross-sectional observation image.

Description

[Technical Field]

[0001] This invention relates to an etching shape evaluation system and method. [Previous Technology]

[0002] In semiconductor manufacturing engineering, the semiconductor wafer processing process, also known as the front-end process, involves forming electronic circuits or integrated circuits containing transistors on the surface of a silicon wafer. A fundamental part of the process for forming these integrated circuits is the etching process, which uses photoresist or similar materials as a mask after development, and then uses etching equipment to shape the wiring and other components. These processes are repeated through precision machining based on film deposition, pattern transfer, and etching.

[0003] The performance improvement of integrated circuits has been achieved through the development of miniaturization processing technology to date, which has enabled the miniaturization of components and increased integration density. However, in recent years, the miniaturization process of wiring has stagnated, making new integration technologies necessary. As new integration technologies, 2.5D packaging technology and 3D packaging technology have attracted much attention. These methods differ from previous miniaturization methods that increased integration density in planar structures; instead, they utilize three-dimensional structures to increase integration density.

[0004] The etching equipment used in wiring processing is used to etch (remove) the layered film formed on the wafer to form the target structure. It is an essential component of the manufacturing process, required to create patterns according to the designed dimensions and shapes, and to precisely carve out the narrow openings (deep holes) and trenches (deep trenches) needed for the wiring contained within them. In etching, repeated processing trials are necessary to create patterns according to the designed dimensions and shapes, while adjusting processing conditions. Since higher processing precision is required in the finished shape, methods to quantitatively confirm the amount of change, such as the actual cross-sectional shape and the dimensions of the cross-sectional shape, are frequently used as indicators of the processed shape.

[0005] Patent Document 1 discloses a technique: capturing a scanning electron microscope (SEM) image of a pattern on a substrate, then performing an arrangement pattern detection process to detect the regular arrangement of patterns within the image area; performing template matching to detect the position of patterns within the image area, integrating the results of arrangement pattern detection and template matching, setting a measurement vernier for the position of each pattern, and measuring the pattern size of each measurement vernier. Furthermore, Patent Document 2 discloses a method for estimating the cross-sectional shape of an inspected sample, comprising: a matching process for matching the cross-sectional shape data of the inspected sample using a complex number of shape models; and a selection process in which, based on the accuracy index, i.e., the error function value, of the matched model, at least one shape model is selected from the complex number of shape models as the most suitable model. [Prior Art Documents] [Patent Documents]

[0006] [Patent Document 1] Japanese Patent Application Publication No. 2015-21942 [Patent Document 2] Japanese Patent Application Publication No. 2013-73776 [Summary of the Invention]

[0007] [Problem to be Solved by the Invention] However, when observing semiconductor circuit patterns with deep holes or trenches on silicon wafers having three-dimensional structures such as micro-cylinders, cones, and N-prisms using SEM from the top-view of the silicon wafer in a planar state, the cross-sectional shape of the three-dimensional structure of the deep holes or trenches is difficult to evaluate. Therefore, in order to confirm the cross-sectional shape, the method of cleaving the sample to be etched and using SEM to confirm the cross-sectional shape has been used in the past. At this time, in device manufacturing, it is important to shift the manufacturing process to a fast start-up, high-yield mass production system early on. Therefore, it is important to shorten the time for observing and evaluating the cross-sectional shape when evaluating the performance of etching processes in repeated processing tests.

[0008] However, when the micro-pattern is a pattern with a circular (Hole) deep hole (also called a hole pattern), it is necessary to process the sample by cleaving along the vertical direction in order to confirm its cross-sectional shape (cleavage: if a small scratch is made along the surface using a tool such as a diamond cutter, and then force is applied to cut it, a smooth cleavage surface will be formed on the surface). In this case, the smaller the size of the circle in the pattern, the longer it takes to measure the roundness of each one, and it is difficult to manually cleave while maintaining the positional relationship of the center point, using a CD-SEM (Critical Dimension-SEM) or similar tool to locate the pattern position from the top surface and aim at the same position. Furthermore, it is even more difficult to cleave in a symmetrical state with the center of the circular pattern as the target surface, and to make it equivalent to the normal section (i.e., all positions from the top (face) to the bottom (face) are equivalent to the center position of the pattern) and confirm the above.

[0009] On the other hand, observation and measurement are easily affected by the cleavage fracture mode. In experiments such as etching processes where processing conditions are repeatedly adjusted to achieve the desired shape, the acquisition of the cross-sectional shape, which can be used as an indicator of the processed shape, can easily affect the accuracy of dimensional measurement results and shape evaluation. In addition, observation and measurement are also easily affected by human experience (judgment) and technique (cutting method, structure confirmation method, etc.). Although focused ion beam (FIB) processing can produce samples with high-precision cross-sectional observation, it requires learning the processing technique and sample preparation takes several hours, which can take even longer depending on the material. The understanding of issues such as accuracy and efficiency when evaluating the shape by cleaving the sample and observing the cross-sectional shape is not fully disclosed in Patent Documents 1 and 2.

[0010] Therefore, the object of the present invention is to provide a technique that, by observing the cleavage profile of a sample being processed, makes it possible to improve the accuracy and efficiency of etch shape evaluation. [Means for solving the problem]

[0011] To solve the above-mentioned problems, the present invention provides a representative etching shape evaluation system, comprising a computer device, wherein the computer device has a processing unit and a data storage unit; the data storage unit stores a basic model image; the processing unit acquires one or more cross-sectional observation images for fabrication taken by a scanning electron microscope; a grid-like template is applied to the cross-sectional observation image for fabrication; images (material images) within the regions of the grid-like template corresponding to each depth position, wherein the width of the cross-sectional outline is maximized; the material images are captured and combined to create a montage image; the data storage unit stores the montage image as a basic model image. [Effects of the Invention]

[0012] According to the present invention, by observing the cleavage profile of the sample being processed, it becomes possible to improve the accuracy and efficiency of the etching shape evaluation. Other issues, structures, and effects not described above are disclosed in the embodiments.

Implementation Method

[0013] Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the present invention is not limited to these embodiments. Furthermore, in the drawings, the same symbols are used to denote the same parts.

[0014] [First Embodiment] (Creation of a Montage Image) In the first embodiment of the present invention, the main focus is on creating a montage image as an evaluation index for etching shape from a cross-sectional image of the etched sample (hereinafter referred to as the sample) after cleavage. FIG1 is a flowchart showing the steps of creating the montage image in the first embodiment. Hereinafter, steps 101 to 109 of FIG1 will be described.

[0015] (Step 101) First, an etching process test is performed on the sample to be processed. Figure 2 is a schematic diagram of the sample to be processed before and after etching in the first embodiment. In this embodiment, a scanning electron microscope (SEM) is used as the imaging device to photograph the sample to be processed. However, other than an SEM may be used as the imaging device.

[0016] The processing sample 201 is composed of at least an etched layer 203 and a masking layer 202 deposited on a silicon wafer. As shown in the schematic diagram (pattern top view) of the masking layer of the processing sample viewed from the top surface of FIG2(a) and the schematic diagram viewed from the cross section of FIG2(b), a masking pattern of regularly arranged circular concave shapes is patterned on the surface of the masking layer 202. When the concave masking pattern is used and etching is performed by an etching apparatus, a deep hole pattern with a three-dimensional structure of voids such as cylinders, cones or N-shaped prisms is formed on the etched layer 203 that is removed due to the processing. FIG3 is a schematic diagram of several examples of masking pattern shapes and deep hole pattern shapes. FIG3(a) is a pattern top view of a circular masking pattern and a three-dimensional view of a cylindrical deep hole pattern. Figure 3(b) shows the top view of the circular masking pattern and the three-dimensional view of the conical deep hole pattern; Figure 3(c) shows the top view of the triangular masking pattern and the three-dimensional view of the triangular prism deep hole pattern; Figure 3(d) shows the top view of the square masking pattern and the three-dimensional view of the square prism deep hole pattern.

[0017] On the other hand, if the mask pattern is convex, the etched shape will be formed by removing the etched layer other than the mask layer, resulting in a columnar (pillar) three-dimensional structure. When regularly arranged convex patterns are closely arranged, deep grooves will be formed between the patterns. In this embodiment, although a concave mask pattern is used to achieve a deep hole, it can also be applied to deep grooves.

[0018] (Step 102) Next, the sample to be processed is cleaved, and a SEM image of the cross-section of the sample to be measured is obtained (also called a "cross-sectional observation image"). (The cross-sectional observation image obtained for creating a montage image is also called a "processing cross-sectional observation image"). The shape of the deep hole pattern after etching is important for judging the etched shape. At this time, it is ideal to observe the cross-sectional shape of the cleaved sample, i.e., the three-dimensional structure of the deep hole pattern, in a state where the three-dimensional structure of the deep hole pattern is symmetrical through the ideal cleaving position A*-B* at the center of the circular mask pattern (see Figure 2(a)(b)). In this embodiment, a method is to insert a small scratch into the surface of the silicon wafer using a tool such as a diamond cutter at any position along the ideal cleaving position A*-B* that contains the three-dimensional structure of the deep hole pattern as much as possible and parallel to the Y-direction plane, and then apply force to cut it. According to this method, it will be possible to form a relatively smooth cleaving surface and observe it in cross-section. However, the masking patterns are very fine, and most of the cleaving is done manually, making it difficult to always cleave the sample at the ideal cleavage position A*-B*. While there are methods using specialized scribing devices with microscopes to achieve more precise cleavage positions, it is difficult to achieve the high precision of focused ion beam (FIB) processing, and it is impossible to aim at the center of every circle of the masking pattern on the tiny circles observed from the top surface.

[0019] Therefore, in practice, the cross-sectional image of the cleaved sample is observed at the actual cleavage position AB. Figure 2(c) is a schematic diagram of the cross-sectional image of the cleaved sample after etching. If the sample is cleaved after etching, the cross-section of the three-dimensional structure (void) formed by the combination of the three-dimensional structure of the masking pattern of the masking layer 202 and the three-dimensional structure of the deep hole pattern of the etched layer 203 will be presented as a cross-sectional outline 204. The cross-sectional outline 204, as shown in Figure 2(c), can be of various shapes, and the dimension width (CD: Critical Dimension) 205 at any depth 206 can also be of various values. Therefore, it is possible to measure the shape and size of the cross-sectional outline 204 and obtain data such as the changes and differences between the shape and size of the target, i.e., the processed cross-sectional outline. Furthermore, in the following description, the horizontal line along the actual cleavage position AB on the top surface of the masking layer is defined as the X direction, and the vertical direction of the top surface of the masking layer is defined as the Y direction.

[0020] The magnification when obtaining the cross-sectional observation image is adjusted such that the depth 206 at any position (Y direction) from the top of the mask layer 202 to the bottom of the etched layer 203 for the cross-sectional outline 204 can be measured; and the change in the width 205 (X direction) at each depth position can be quantitatively measured using a measuring tool or the like on the screen of the computer device; and the observation area is included within the standard image size range.

[0021] In this embodiment, a scanning electron microscope (SEM) is used for observation, displaying at 200K magnification, and multiple cross-sectional observation images are obtained from any observation position within the sample. Figure 4 shows examples of the multiple cross-sectional observation images (A) to (C) that have been obtained. Furthermore, the image data may differ depending on the basic performance of the scanning electron microscope (SEM) used and the settings of its parameters, such as screen size or resolution. Additionally, when the multiple cross-sectional observation image data after etching shown in Figure 4 are loaded into a computer device, it is best to verify that the image data is a regular arrangement of deep hole patterns with the same three-dimensional structure (e.g., cylindrical), and that the image data size, resolution, observation range, and magnification are consistent standard images.

[0022] (Step 103) Next, a grid pattern is applied to the cross-sectional view image. Figure 5 is a schematic diagram showing the state of applying a grid pattern to the cross-sectional view image. The grid pattern A501 is composed of checkerboard-shaped regions (also referred to as "grid regions" or simply "regions"). By applying the grid pattern to the cross-sectional view image, the shape of each cross-sectional outline 204 of the image cut out in each region (also referred to as "image within the region") can be measured. The shape of the grid region is not necessarily limited to a square, as long as the shape of each plurality of cross-sectional outlines 204 of the image within the region can be measured and a montage image can be created.

[0023] Regarding the horizontally arranged areas of the grid-like template A501, since the masking pattern of the masking layer 202 is arranged based on a consistent spacing rule (see Figure 2(a)), and the deep hole pattern after etching is also arranged according to the masking pattern rule, the masking pattern and the deep hole pattern are arranged at the same spacing. It is preferable that the width of each cross-sectional outline 204 is set to be able to be accommodated within the size of each area in the horizontal direction (area X size). Regarding the overall horizontal size of the grid-like template (template X size), although there is no particular limitation, it is preferable to set it appropriately to accommodate the multiple cross-sectional outlines contained in the cross-sectional observation image.

[0024] On the other hand, the areas arranged vertically in the grid pattern A501 can be appropriately adjusted according to the accuracy or efficiency of the etching shape evaluation. If the size of each area in the vertical direction (area Y size) becomes more refined, more detailed measurement data can be obtained for the profile outline and the evaluation accuracy will increase, but the evaluation time and cost will also increase. Therefore, it is best to make an adjustment that can take both into account.

[0025] Regarding the overall vertical size of the grid-like template (template Y size), while there are no particular limitations, it is preferable to set it appropriately to accommodate the multiple cross-sectional outlines contained in the cross-sectional observation image. However, when measuring depth in the vertical direction, the starting point of the depth, i.e., the reference position, is also designated as the uppermost or lowermost position. For example, there is the case where the uppermost position of the depth starting point is configured at position C1 on the top surface of the mask layer 202, and the case where the uppermost position of the depth starting point is configured at position C2 on the boundary surface between the mask layer 202 and the etched layer 203. In this case, the template Y size, as illustrated in Figure 5, is preferably adjusted to accommodate the uppermost position (C2) and the lowermost position (D) designated as the reference position.

[0026] Figure 6 shows a schematic diagram of applying a different grid pattern to a cross-sectional observation image. Compared with grid pattern A501, grid pattern B601 has an increased number of regions (segments) in the Y direction. Therefore, the number of regions in the vertical direction (Y direction) of the grid pattern or the Y-size of these regions can be flexibly changed according to the observation magnification, the depth range of observation, or the accuracy of performance evaluation.

[0027] By subdividing the size of the grid-like region, as described later, the number of source images for the montage composite image formed by combining images from regions at each depth position will increase, resulting in a more detailed montage composite image. When detailed measurements are desired within a specific depth range, techniques such as subdividing the size of the grid-like region only within that range can be used, allowing for arbitrary setting of the grid-like region size. Figure 7 is a schematic diagram showing the application of grid-like templates A and B of different sizes on a multi-section observation image.

[0028] (Step 104) Next, the profile outline 204 is measured in the entire area of ​​the grid pattern A501. Figure 8 is a schematic diagram showing the application of the grid pattern on the cross-sectional view image and the measurement of the profile outline in the entire area. As shown in Figure 8, the entire area of ​​the grid pattern A501 is defined by numbers in the X and Y directions. In the example of Figure 8, numbers 1 to 10 are defined in the X direction corresponding to the hole positions of the circular mask pattern, and numbers a to e are defined in the Y direction. Therefore, the entire area can be defined by area numbers such as 1-a, 1-b, ~10-d, 10-e. Next, the depth 206 of the profile outline 204 and the width 205 at that depth are measured in the entire area to obtain numerical data. In addition, the position for measuring the depth 206 in each area is not limited, and the centroid position of the area can also be used.

[0029] However, depending on the location of the cleavage or the state of its fracture profile, the three-dimensional structure of the deep hole pattern may deviate significantly from the ideal (see the dashed box in Figure 2), and the profile outline 204 may disappear or become unclear. In such cases, the measurement of the width 205 and depth 206 within the area may be omitted (see the X mark in Figure 8).

[0030] When the obtained cross-sectional observation image data is multiple (see Figure 7), it is preferable to load the image data containing the same image size or magnification into the computer device. Then, using these same-sized grid templates, the measured numerical data in each region of the image are obtained. Regarding the measurement method, while an application with applicable measurement functions can be used in the computer device, it is not limited to one method if a quantitative change can be measured. However, the measurement conditions should ideally yield standard and stable output values.

[0031] In one or more cross-sectional observation images, the measured dimensions 205 and 206 within the entire grid area of ​​the applied grid pattern are input into a list using a spreadsheet application on a computer device. The list should also include the measurement data for the width or depth, along with the area number (1-a, etc.) of the area being measured. The width or depth can be measured using any method. It can be done using publicly available image recognition technology to automatically identify and measure the two ends of the cross-sectional outline at a predetermined depth, or it can be done using a measurement tool on the screen of a computer device.

[0032] In the overview list, it is desirable to record the maximum width of the areas marked with the same depth location area numbers (a~e). Furthermore, by calculating and recording the minimum, average, or median values ​​other than the maximum value, it is possible to verify whether the obtained maximum value information deviates from the expected design size value, which will also help improve the accuracy of the montage image mentioned below. These calculations can be easily performed using calculation functions such as table calculation applications on all obtained measurement results. In addition, the overview list can be edited arbitrarily. By highlighting the areas with the maximum or minimum values ​​in the grid table with color, the positional relationship of the areas with the maximum values ​​in the subsequent grid template will be easily grasped.

[0033] (Step 105) Next, in the image of the area of ​​the grid pattern corresponding to each depth position, select the value with the largest dimension width of the profile outline. To do this, firstly, we will explain the situation where the ideal cleavage position deviates from the actual cleavage position when cleaving the sample for processing. Figure 9 is a schematic diagram of the cleavage pattern of a cylindrical deep hole pattern and an example of the cleaved profile shape. The interior of the deep hole pattern is an etched cavity, and the top and bottom surfaces of the cylindrical pattern's three-dimensional structure are circular. Figure 9(a) shows the cleavage along the direction Pa, which is perpendicular to the top surface and passes through the center position A-A' (ideal cleavage position) of the circle (refer to (a-1)). Qa in the profile is equivalent to the dimension width of the profile outline (refer to (a-2)). In this case, Qa represents the maximum value that the dimension width of the profile outline can be obtained at any depth position in the vertical direction, and in this state, the center of the circle, i.e., the plane of symmetry, becomes the cleaved state.

[0034] Figure 9(b) shows the cleavage along direction Pb, perpendicular to the top surface and passing through B-B' slightly off-center from the circle's center position A-A' (refer to (b-1)). Qb in the section is equivalent to the width of the section profile (refer to (b-2)). The width of the section profile Qb is reduced compared to the maximum value Qa, suggesting that the cleavage location is the state where the asymmetric surface slightly off-center from the center position A-A' is cleaved.

[0035] Similarly, Figure 9(c) shows the cleavage along direction Pc, which is perpendicular to the top surface and passes through C-C', which is significantly deviated from the center position A-A' of the circle (refer to (c-1)). Qc in the section is equivalent to the width of the section profile (refer to (c-2)). The width Qc of the section profile is significantly reduced compared to the maximum value Qa, which suggests that the cleavage location is the state where the asymmetric surface significantly deviated from the center position A-A' is cleaved.

[0036] Next, regarding the case where the three-dimensional structure of the deep hole pattern after etching is conical (refer to Figure 3(b)), the pattern where the cleavage position deviates from the ideal position will be explained. In this case, the width of the profile of the deep hole pattern decreases continuously at each depth from the top surface to the bottom surface. Figure 10 is a schematic diagram of the cleavage pattern of the conical deep hole pattern and an example of the cross-sectional shape after cleavage. The interior of the deep hole pattern is the cavity after etching, and the top surface of the three-dimensional structure of the conical pattern is circular. Figure 10(a) shows the cleavage along the direction Pa, which is perpendicular to the top surface and passes through the center position A-A' (ideal cleavage position) of the circle (refer to (a-1)). Qa in the cross-section is equivalent to the width of the profile (refer to (a-2)). In this case, the maximum value of the width of the profile line can be obtained at any depth position in the vertical direction of the Qa system, and the depth Ra in the vertical direction of the profile line is also the maximum value at any position in the horizontal direction. In this state, the center of the circle, i.e. the plane of symmetry, will become the state to be cleaved.

[0037] Figure 10(b) shows the cleavage along direction Pb, perpendicular to the top surface and passing through B-B' off the center position A-A' of the circle (refer to (b-1)). Qb in the section is equivalent to the width of the section profile (refer to (b-2)). The width of the section profile Qb is reduced by the maximum value Qa, and the depth Rb in the vertical direction is also reduced by the maximum value Ra. It can be inferred that the cleavage position is the state of the asymmetric plane slightly off the center position A-A' being cleaved.

[0038] Next, the case where the three-dimensional structure of the deep hole pattern is deformed or bent in the vertical direction will be described. The regularly arranged circular masking pattern shown in Figure 2(a) and the N-sided masking pattern shown in Figure 3, when the design size is more refined, the size of the etched pattern is more likely to deviate from the design size. Figure 11 is a schematic diagram showing the deviation of the masking pattern shape before and after etching. Ideally, the designed ideal circular masking pattern 1101 and the actual masking patterns 1102-1104 observed from the top surface of the etched deep hole pattern should have the same shape and size. However, in reality, as shown in Figure 11, there will generally be slight differences, and deviations are expected, including portions where the design size of the masking pattern 1101 exceeds or falls short of the design size.

[0039] Furthermore, in the three-dimensional structure of a deep hole pattern formed by etching a circular mask pattern in the vertical direction, the ideal shape deviates from the actual shape. Figure 12 is a schematic diagram of an example of the three-dimensional structure of an actual deep hole pattern. Figure 12(a) shows the three-dimensional structure of a deep hole pattern with an ideal shape, and also shows the top surface 1201 and bottom surface 1202 of the deep hole pattern (both circular) viewed from the top. In contrast, Figure 12(b) shows the three-dimensional structure and bottom surface 1203 where the hole axis of the deep hole pattern is bent; Figure 12(c) shows the three-dimensional structure and bottom surface 1204 where the width of the deep hole pattern is reduced; Figure 12(d) shows the three-dimensional structure and bottom surface 1205 where the deep hole pattern is tilted and the etching process is insufficient; and Figure 12(e) shows the three-dimensional structure and bottom surface 1206 where the deep hole pattern is tilted and deviates from the center line. Thus, the actual three-dimensional structure of the deep hole pattern deviates from the ideal shape in various ways.

[0040] Thus, when the three-dimensional structure of the deep hole pattern deforms or bends in the vertical direction, it is considered that a cleavage similar to that of a conical three-dimensional structure as shown in Figure 10 will occur. Assuming that cleavage can be performed along the direction Pa passing through the center of the circle A-A', the dimension width of the profile outline Qb and the depth of Rb can be expected to be represented by the shape state deviation in the depth direction. In this case, it can be expected that the dimension width at each depth position will decrease from the maximum value of the face symmetry, and the depth will also show a decrease in the measured value.

[0041] As described above, when comparing the cross-sectional outline obtained from the cleavage of the three-dimensional structure of the deep hole pattern with the width of the dimensions at each depth position, the larger the value, the more likely the cleavage state can be inferred through the plane of symmetry (positive section) close to the center of the masking pattern. Therefore, when it is necessary to infer the cleavage state of the plane of symmetry, it is better to select the image with the maximum value. Furthermore, the greater the depth of the cross-sectional outline in the vertical direction, the less bending and deformation of the processed deep hole pattern can be inferred.

[0042] (Step 106) Next, the data of the selected maximum value is compared with the position information of the grid area of ​​the grid pattern to define the grid area where the maximum value is obtained. Figure 13 is a schematic diagram showing the shape of the grid area where the maximum value is obtained. As described in step 104, the grid pattern A501 is applied to the cross-sectional view image (A), and the data of the width 205 of the cross-sectional outline 204 is obtained in all grid areas (1-a to 10-e) (Figure 13(a)). However, the measurement data of the areas where the cross-sectional outline is not clear (see the X mark in Figure 13) does not need to be obtained.

[0043] The comparison between the data of the maximum value and the position information of the grid-like area is as shown in the example of Figure 13(b), and is defined by area symbols such as 1-a, 8-b, 3-c, 8-d, and 2-e according to the depth order. At this time, in order to make the selection of the image of the maximum value easier, the display in its area can be emphasized by changing the thickness of the line and the color, or the images in other unsuitable areas can be deleted or hidden, etc., to make the selection easier.

[0044] Furthermore, as shown in Figure 13, when there is only one cross-sectional image, the region of maximum value is identified from within that image. However, as shown in Figure 7, when there are multiple image data, all regions in all image data are measured, and the maximum value is selected from each depth position in the entire image and the entire region. Then, the region where the maximum value is obtained is identified. However, depending on the amount of information displayed (resolution, pattern size, and magnification) and the size of the region that can be changed, the measurement data (width and depth) will have different levels of accuracy.

[0045] (Step 107) Next, the image within the grid-like region where the maximum value is obtained is captured as a source image. Figure 14 is a schematic diagram showing the capture of the image within the grid-like region where the maximum value is obtained as a source image. From the position of the grid-like region where the maximum value is obtained as defined in step 106 (Figure 14(a)), as shown in Figure 14(b), a method of maintaining the region size and segmenting and capturing the image data is adopted. (The image segmented is called the "source image"). In addition, when the same maximum value exists, one of the following can be used: for example, the conformity of the design of the membrane structure and the boundary line of the laminated membrane, the clarity of the cross-sectional outline, whether there is deformation or foreign matter adhering, etc.

[0046] (Step 108) Next, the cut material image data is joined and synthesized into a montage image. Figure 15 shows a schematic diagram of joining the cut material image data and synthesizing a montage image. The material images 1-a, 8-b, 3-c, 8-d, and 2-e (Figure 15(a)) extracted in step 107 are joined and synthesized into a montage image 1501 (Figure 15(b)). At this time, in order to avoid a large deviation between the cross-sectional outline of the montage image and the seam of the adjacent material images, it is desirable to join the material images while maintaining a continuous intermediate position as much as possible. Furthermore, by using a size smaller than the area Y size of the applied grid pattern, the seam between the upper and lower images can be further smoothed when synthesizing the montage image, so as to form a montage composite image with higher accuracy. However, if the number of vertical regions (segments) of the grid pattern is too large and the region Y size is too small, the measurement data will increase significantly. In addition, the selection and capture of source images and the method of combining source images to synthesize montage images can be expected to take a lot of time. Therefore, a balance with accuracy is very important.

[0047] (Step 109) The montage image thus created is an image within the region with the maximum size and width at each depth position of the source image. Therefore, the ideal cross-sectional outline on the plane of symmetry (front section) cleaved at the center position of the mask pattern shape as a whole can be deduced. Therefore, the montage image can be used as an image for evaluating etching shape indicators (the image for evaluating indicators is called the "basic model image"). Here, the montage image is stored in the data memory as a basic model image for evaluating the performance of the etching process. It is then used as a performance evaluation indicator for the etching process.

[0048] (Hardware Configuration) This embodiment can be configured as a system with a computer device comprising a processing unit, memory, input I / F, indicator input unit, data storage unit, communication I / F, output I / F, and display unit as hardware. The processing unit is a processor (CPU) that executes commands according to the program stored in memory. Memory is a medium for storing data or programs and may include random access semiconductor memory, storage devices, or storage media (volatile or non-volatile). Input I / F is an interface such as a keyboard or mouse that connects to the indicator input unit for inputting user commands. Data storage unit stores data such as basic model images or grid patterns containing montage images and is composed of known storage devices such as hard disk drives (HDDs) or solid-state drives (SSDs). Communication I / F is an interface that connects to a scanning electron microscope (SEM) or etching apparatus and transmits and receives signals to other machines for image capture or etching conditions. The output I / F is an interface connected to a display unit such as a monitor, and outputs various display information. The computer device can be any electronic machine such as a desktop computer or a laptop computer. In this embodiment, for example, steps 103 (application of the grid pattern), 104 (measurement of the profile outline), 105 (selection of the maximum value), 106 (identification of the area where the maximum value is obtained), 107 (capturing the area where the maximum value is obtained), and 108 (creation of the montage image) can be automatically processed in the processing unit of the computer device.

[0049] Thus, montage images can be inferred as ideal cross-sectional contours, which can serve as an indicator contributing to the accuracy of shape evaluation. In addition, there is the advantage that cross-sectional observation images can be produced more efficiently using computer devices.

[0050] [Second Embodiment] (Performance Evaluation of Etching Process) The second embodiment of the present invention uses a basic model image comprising a montage image created in the first embodiment to perform performance evaluation of the etching process. FIG16 is a flowchart showing the steps of performing performance evaluation of the etching process in the second embodiment by using image recognition sequence processing based on the basic model image. Hereinafter, steps 1601 to 1609 of FIG16 will be described. The hardware configuration is also the same as in the first embodiment, so the differences will be mainly described.

[0051] (Step 1601) An etching test is performed on the processing sample under etching conditions different from those in the first embodiment. However, the processing sample used is the same as that used in the first embodiment when creating the montage image, and the same mask pattern arrangement is used. However, it is acceptable to perform the processing test under the same etching conditions as in the first embodiment.

[0052] (Step 1602) Similar to the first embodiment, a cross-sectional observation image of the object being measured is obtained (the cross-sectional observation image obtained for evaluating the etched shape is called an "evaluation cross-sectional observation image"). The observation environment, such as the observation range and magnification, is the same as that in the first embodiment.

[0053] (Step 1603) The cross-sectional observation image composed of any number of images obtained in step 1602 is stored in the data memory of the computer device.

[0054] (Step 1604) In the processing unit of the computer device, image recognition sequence processing begins.

[0055] (Step 1605) First, the montage image created in the first embodiment and stored as a basic model image (step 109) is read from the data memory unit. Similarly, the cross-sectional observation image saved in step 1603 is read.

[0056] (Step 1606) Next, determine the similarity between the basic model image and the cross-sectional observation image. There are no particular restrictions on the method for determining similarity, and image recognition technology using AI (artificial intelligence) can be used. If there is no similarity, that is, if the image of the cross-sectional outline with the same characteristics is not detected, return to step 1602, re-cleave, and obtain the cross-sectional observation image. Alternatively, return to step 1601 and restart the processing experiment. The reasons for the judgment of no similarity may be as follows: (1) The desired design size cannot be achieved under etching conditions. (2) The cross-sectional observation image is not near the center position (symmetry plane) of the mask pattern, but is taken at a position that is significantly off the center or does not capture the deep hole pattern. (3) The cross-sectional part is contaminated or damaged, making the cross-sectional outline unclear. (4) The deep hole is blocked by deposits or other substances.

[0057] When images with similarity, i.e., those with the same cross-sectional contour lines, are detected, the images consisting of regions containing similar cross-sectional contour lines (referred to as "similar cross-sectional pattern images") will be identified and captured along with their location. Figure 17 is a schematic diagram illustrating the definition and capture of similar cross-sectional pattern images that are similar to the montage image. Multiple similar cross-sectional pattern images may be captured.

[0058] (Step 1608) For the similar cross-sectional pattern image captured in step 1607, the dimensions of the width and depth of the cross-sectional outline in all grid regions are measured, similar to step 104 of the first embodiment.

[0059] (Step 1609) Similar to the montage image, the ideal cross-sectional outline on the symmetry plane (front section) cleaved at the center position of the mask pattern shape can be deduced from the similar cross-sectional pattern image, and therefore can be used as a new basic model image. Here, the similar cross-sectional pattern image is stored together with the measurement data obtained in step 1608 in the data memory unit. The stored similar cross-sectional pattern image is read as a basic model image in subsequent image recognition sequence processing (step 1605).

[0060] Thus, by using montage images as the base model images for cross-sectional observation images obtained in etching processing experiments, and employing image recognition technologies such as AI to determine similarity and evaluate the performance of the etching process, short-term and effective performance evaluation can be achieved, while reducing the impact of human experience bias. Furthermore, montage images, or existing cross-sectional pattern images that are judged to be similar to the base model images, can be used as new base model images for updating AI learning data. Therefore, an increase in the accuracy of similarity judgment using AI image recognition sequence processing can be expected.

[0061] Although the embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above, and various modifications can be made without departing from the spirit of the present invention.

[0062] The possible forms of the present invention are described below, but are not limited thereto. (Form 1) An etching shape evaluation system equipped with a computer device, comprising: the computer device having a processing unit and a data storage unit; the data storage unit storing a basic model image; the processing unit acquiring one or more cross-sectional observation images for manufacturing taken by a scanning electron microscope; a grid pattern applied to the cross-sectional observation image for manufacturing; an image (material image) in the region of the grid pattern corresponding to each depth position, wherein the width of the cross-sectional outline is at its maximum value; capturing the material images and combining them to create a montage image; the data storage unit storing the montage image as the basic model image. (Form 2) The etching shape evaluation system of Form 1, wherein the vertical size of the region of the grid pattern can be changed according to the accuracy of the performance evaluation. (Form 3) In any of Form 1 or 2, the masking layer of the processed sample, which is the object of the etching shape evaluation, is formed as a circular masking pattern. (Form 4) In any of Forms 1 to 3, the aforementioned processing unit acquires one or more evaluation cross-sectional observation images captured by a scanning electron microscope and determines the similarity between the aforementioned basic model image and the aforementioned evaluation cross-sectional observation image. (Form 5) In any of Form 4, the aforementioned basic model image includes a similar cross-sectional pattern image extracted from the aforementioned evaluation cross-sectional observation image that is considered to be similar in the aforementioned determination. (Form 6) In any of Form 4 or 5, the aforementioned determination has the feature of image recognition using artificial intelligence (AI). (Form 7) An etching shape evaluation method, wherein: in the processing unit of a computer device, one or more cross-sectional observation images for fabrication are acquired by scanning electron microscope; a grid pattern is applied to the aforementioned cross-sectional observation images for fabrication; images (material images) in regions of the aforementioned grid pattern corresponding to each depth position, where the width of the cross-sectional outline is at its maximum value, are defined; the aforementioned material images are captured and combined to create a montage image; the aforementioned montage image is saved as a basic model image in the data storage unit of the aforementioned computer device. (Form 8) The etching shape evaluation method of Form 7, wherein: in the aforementioned processing unit, one or more evaluation cross-sectional observation images are acquired by scanning electron microscope; the similarity between the aforementioned basic model image and the aforementioned evaluation cross-sectional observation image is determined. [Simplified Explanation of the Diagram]

[0063] [Fig. 1] Fig. 1 is a flowchart of the montage image creation steps in the first embodiment. [Fig. 2] Fig. 2 is a schematic diagram of the sample before and after etching in the first embodiment. [Fig. 3] Fig. 3 is a schematic diagram of several examples of mask pattern shapes and deep hole pattern shapes. [Fig. 4] Fig. 4 is a diagram of examples of multiple cross-sectional observation images (A) to (C) that have been obtained. [Fig. 5] Fig. 5 is a schematic diagram of the state of applying a grid pattern on the cross-sectional observation image. [Fig. 6] Fig. 6 is a schematic diagram of the state of applying another grid pattern on the cross-sectional observation image. [Fig. 7] Fig. 7 is a schematic diagram of the state of applying grid patterns A and B of different sizes on multiple cross-sectional observation images. [Fig. 8] Fig. 8 is a schematic diagram of applying a grid pattern on the cross-sectional observation image and representing the pattern of measuring the cross-sectional contour line in the entire area. [Figure 9] Figure 9 is a schematic diagram showing the cleavage pattern of a cylindrical deep hole pattern and an example of the cross-sectional shape after cleavage. [Figure 10] Figure 10 is a schematic diagram showing the cleavage pattern of a conical deep hole pattern and an example of the cross-sectional shape after cleavage. [Figure 11] Figure 11 is a schematic diagram showing the deviation of the mask pattern shape before and after etching. [Figure 12] Figure 12 is a schematic diagram showing an example of the three-dimensional structure of an actual deep hole pattern. [Figure 13] Figure 13 is a schematic diagram showing the shape of the grid-like region where the maximum value is obtained. [Figure 14] Figure 14 is a schematic diagram showing the image within the grid-like region where the maximum value is obtained as the source image. [Figure 15] Figure 15 is a schematic diagram showing the combination of the extracted source image data and the synthesis of a montage image. [Figure 16] Figure 16 is a flowchart showing the steps of evaluating the etching performance in the second embodiment by using image recognition sequence processing based on the basic model image. [Figure 17] Figure 17 is a schematic diagram showing the definition and extraction of similar cross-sectional pattern images that are similar to montage images.

Claims

1. An etching shape evaluation system equipped with a computer device, wherein the computer device has a processing unit and a data storage unit; the data storage unit stores a basic model image; the processing unit acquires one or more cross-sectional observation images for fabrication taken by a scanning electron microscope; a grid pattern is applied to the cross-sectional observation image for fabrication; images, i.e., material images, are defined in the regions of the grid pattern corresponding to each depth position where the width of the cross-sectional outline is at its maximum value; the material images are captured and combined to create a montage image; the data storage unit stores the montage image as the basic model image.

2. The etching shape evaluation system as described in claim 1, wherein, The size of the area in the vertical direction of the aforementioned grid-like template can be changed as the accuracy of the etched shape evaluation increases.

3. The etching shape evaluation system as described in claim 1, wherein, The etched shape that becomes the object of the etched shape evaluation is the shape etched using a mask pattern of holes.

4. The etching shape evaluation system as described in claim 1, wherein, The aforementioned processing unit acquires one or more evaluation cross-sectional observation images captured by a scanning electron microscope; and determines the similarity between the aforementioned basic model image and the aforementioned evaluation cross-sectional observation image.

5. The etching shape evaluation system as described in claim 4, wherein, The aforementioned basic model image includes similar cross-sectional pattern images extracted from the aforementioned evaluation cross-sectional observation images that are considered similar from the aforementioned judgment.

6. The etching shape evaluation system as described in claim 4, wherein, The aforementioned judgment is characterized by the use of artificial intelligence (AI) for image recognition.

7. An etched shape evaluation method, wherein: In the processing unit of the computer device, one or more cross-sectional observation images for manufacturing are acquired by scanning electron microscope; a grid pattern is applied to the aforementioned cross-sectional observation images for manufacturing; the images in the regions of the aforementioned grid pattern corresponding to each depth position, where the width of the cross-sectional outline is at its maximum value, i.e., the source images, are defined; the aforementioned source images are captured and combined to create a montage image; the aforementioned montage image is saved as a basic model image in the data memory unit of the aforementioned computer device.

8. The etching shape evaluation method as described in claim 7, wherein, In the aforementioned processing unit, one or more evaluation cross-sectional observation images are acquired by scanning electron microscope; the similarity between the aforementioned basic model image and the aforementioned evaluation cross-sectional observation image is determined.

Images