Evaluation method, search method, and search system

Abstract

The invention provides an evaluation method, a search method, and a search system. The purpose of the present invention is to provide a technique for evaluating a shape appearing in a cross-sectional image. A representative evaluation method of the present invention evaluates the difference between a target shape and a cross-sectional shape of an electron microscope image, the evaluation method being characterized by comprising: a first step of measuring a characteristic dimension of the cross-sectional shape; a second step of, after the first step, creating a template of the target shape among the cross-sectional shapes obtained from the dimensions; and a third step of comparing the difference between the template and the cross-sectional shape by means of the normalized index.

Description

Technical Field

[0001] The present invention relates to an evaluation method, a search method and a search system. Background Art

[0002] In general, in manufacturing systems, data related to manufacturing is very important and valuable. In addition, with the advancement of data analysis tools and algorithms, it is important to always seek refined and simple methods to reduce our workload in order to actively grow in a highly competitive business environment.

[0003] As more information is collected, the amount of data is likely to increase rapidly, requiring more resources and time for analysis, and process development tends to stagnate.

[0004] This situation also applies to the semiconductor industry. For example, in order to evaluate the results of etching tests, a large amount of analysis using microscope images is required.

[0005] In the past, a method for analyzing microscope images has been proposed. For example, Patent Document 1 discloses a dimension measuring device that shortens the time required for dimension measurement and eliminates errors caused by operators. The dimension measuring device in Patent Document 1 uses a first image recognition model that extracts the boundary line between the processing structure and the background and / or the boundary line of the interface between different materials throughout the entire cross-sectional image, and outputs a second image recognition model that is used to distinguish the information of the boundary line throughout the entire cross-sectional image obtained from the first image recognition model according to each unit pattern constituting a repeated pattern, to obtain the coordinates of a plurality of feature points predefined for each unit pattern, and to measure the dimension defined as the distance between two given points among the plurality of feature points.

[0006] Prior art literature

[0007] Patent Literature

[0008] Patent Document 1: International Publication No. 2020 / 121564

[0009] In Patent Document 1, a plurality of feature points for evaluating the size of a repetitive pattern are extracted from a microscope image. However, there is still room for improvement in the evaluation method of the shape of the pattern itself. Summary of the invention

[0010] Therefore, an object of the present invention is to provide a technique for evaluating a shape appearing in a cross-sectional image.

[0011] In order to solve the above-mentioned problems, one of the representative evaluation methods of the present invention evaluates the difference between a target shape and a cross-sectional shape of an electron microscope image, and the evaluation method is characterized in that it comprises: a first step of measuring a characteristic dimension of the cross-sectional shape; a second step of, after the first step, making a template of the target shape in the cross-sectional shape obtained from the dimension; and a third step of comparing the difference between the template and the cross-sectional shape by a normalized index.

[0012] Effects of the Invention

[0013] According to the present invention, the shape appearing in the cross-sectional image can be evaluated.

[0014] Problems, structures, and effects other than those described above will become apparent from the following description of embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a diagram showing an example of a cross section of a semiconductor processing layer imaged by SEM.

[0016] Figure 2 It is a diagram schematically showing the profile of an ideal rectangle.

[0017] Figure 3 This is a diagram schematically showing a more realistic silhouette found during the etching process.

[0018] Figure 4 Yes Figure 3 The diagram shows an example of the case where the area method is applied to the groove 4b shown.

[0019] Figure 5 Yes Figure 3 The diagram shows an example of the case where the distance method is applied to the groove 4b shown.

[0020] Figure 6 This is a diagram showing the process of data preparation.

[0021] Figure 7 This is a diagram showing an example of a silhouette diagram.

[0022] Figure 8 This is a diagram showing an example of data obtained through the data preparation process.

[0023] Fig. 9 This means that the block area representing the ideal shape is set to include a region having an x ​​coordinate value x max A diagram of the situation in the shoulder area above the point.

[0024] Fig.10 This is a diagram showing the process of calculating the area number in method 1.

[0025] Fig.11 is the value x at the x coordinate max FIG. 3 is a diagram showing a case where a block region 34 representing an ideal shape is set within a range of points.

[0026] Fig.12 This is a diagram showing the process of calculating the area number in method 2.

[0027] Fig.13 This is a diagram showing an example of a case where a silhouette image is scaled.

[0028] Fig.14 This is a diagram showing the process of calculating the distance number in the distance method.

[0029] Fig.15 This is a diagram showing an example of the configuration of an evaluation system that performs the evaluation method disclosed in the present disclosure.

[0030] Fig.16 It is a figure which shows the result of evaluating a semiconductor structure.

[0031] Description of Reference Numerals

[0032] 1, 1a, 1b: semiconductor substrate, 2, 2a, 2b, 2c: processing layer, 3, 3a, 3c: photoresist layer, 4, 4a, 4b: groove, 10: bottom, 11: frame area, 12: shoulder area, 17: object range, 20: contour line, 21: lowermost point, 22: sidemost point, 23: top point, 24, 34: block area, 40: evaluation system, 41: input unit, 42: control unit, 43: storage unit, 44: display unit, 44: ideal shape, 45: closest point. DETAILED DESCRIPTION

[0033] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In addition, the present invention is not limited by the embodiments. In addition, in the description of the accompanying drawings, the same reference numerals are used to indicate the same parts.

[0034] When there are a plurality of components having the same or similar functions, different subscripts may be added to the same reference numerals for description. In addition, when there is no need to distinguish between these plurality of components, the subscripts may be omitted for description.

[0035] In order to facilitate understanding of the invention, the position, size, shape, range, etc. of each component shown in the drawings may not represent the actual position, size, shape, range, etc. Therefore, the present invention is not necessarily limited to the position, size, shape, range, etc. disclosed in the drawings.

[0036] In this disclosure, when coordinates are shown, conditions may be described in subscripts. For example, "xy＝ymax ” means “the value of y is y max The value of the x-coordinate of the case".

[0037] In addition, the so-called silhouette refers to the outline presented in the cross section of the semiconductor. For example, in the case of a "silhouette view" of a semiconductor structure, it refers to a view showing the outline formed by the boundary between the semiconductor structure part and the space part presented when observing the cross section of the semiconductor substrate.

[0038] Furthermore, the width of the silhouette refers to the length in a direction parallel to the main surface of the semiconductor substrate, and the height of the silhouette refers to the length in a direction perpendicular to the main surface of the semiconductor substrate.

[0039] [Problems of the prior art]

[0040] Generally, accurate evaluation can be performed by collecting a large amount of data, but there are problems in analyzing a large amount of data.

[0041] In order to maximize the information from expensive SEM (Scanning Electron Microscope), STEM (Scanning Transmission Electron Microscope), and TEM (Transmission Electron Microscopy) images while offsetting the increase in data collection, it is often useful to characterize important structures with as few data points as possible and to quickly analyze and report the results of a given experiment. However, the shapes and behaviors that become clear in the microscope images of the semiconductor deposits under test are often difficult to quantify regardless of their relevance.

[0042] Attempts to quantify such shapes are often done qualitatively using a single value using SEM images or by manual grading after other inspections. However, this approach depends on who grades the silhouette and thus involves subjective measurements.

[0043] Due to its ambiguity, this method is not ideal as a machine learning method and also lacks precision.

[0044] For example, it is useful to use edge detection software to describe the squareness of a silhouette by taking the edge as the x- and y-coordinates. However, the more accurate the shape description, the more sampling points are used, which is difficult to manage manually. In addition, it is also troublesome to have a large number of asymmetric coordinates that are not ideal for machine learning analysis.

[0045] Therefore, the present disclosure proposes two methods for quantifying the overall shape of a certain ambiguous shape by comparing it with an ideal shape using a single dimensionless quantity index.

[0046] [Implementation Method]

[0047] (Example of Microscope Image and Shape of Evaluation Object)

[0048] refer to Figures 1 to 3 2A and 2B illustrate examples of microscope images of semiconductor processing substrates. Figures 1 to 3 This is a perspective view of a portion of a cross section including a semiconductor processing substrate. Figure 1 This is a diagram showing an example of a cross section of a semiconductor processing layer imaged by SEM. Figure 2 It is a diagram schematically showing the silhouette of an ideal rectangle. Figure 3 This is a diagram schematically showing a more realistic silhouette found during the etching process.

[0049] Figure 1 This is a diagram obtained by photographing a cross-sectional view in a direction perpendicular to the main surface of a semiconductor processing substrate. It shows the following state: after a processing layer 2 including a TiN / Ti layer, an Al-Si-Cu layer, and a TiN layer and a photoresist layer 3 having a given pattern are formed on a semiconductor substrate 1, etching is performed to form a groove 4 following a given pattern.

[0050] Figure 2 : represents the case where etching is performed ideally. Under ideal conditions, etching progresses in a direction perpendicular to the main surface of the semiconductor substrate 1a, and does not progress in a direction parallel to the main surface of the semiconductor substrate 1a. Since the etching amount is the same in the processing layer 2a and the photoresist layer 3a, the processing layer 2a and the photoresist layer 3 have the same surface relative to the groove 4a. Figure 2 As shown, the ideal groove 4a has a rectangular shape extending in a direction perpendicular to the main surface of the semiconductor substrate 1a. However, in reality, it is difficult to achieve an ideal shape because the etching amount varies depending on the composition contained in the photoresist layer 3a and the handle layer 2a.

[0051] Figure 3 This is a diagram that makes ideal etching closer to reality. Etching progresses in a direction perpendicular to the main surface of the semiconductor substrate 1b, and does not progress in a direction parallel to the main surface of the semiconductor substrate 1b. In addition, the groove 4b has an arc shape in the semiconductor substrate 1b. Figure 3 The shape shown is worth evaluating when considering the degree of etching treatment. Figure 3The degree of shape shown can also be used as an indicator for evaluating etching processing.

[0052] (Gist of the Invention)

[0053] refer to Figure 4 as well as Figure 5 In the present disclosure, two methods (area method and distance method) are defined for determining how close a shape is to a rectangular shape. The evaluation method in the present disclosure is an evaluation method for evaluating the difference between a target shape and a cross-sectional shape of an electron microscope image. Figure 4 Yes Figure 3 The diagram shows an example of the case where the area method is applied to the groove 4b shown. Figure 5 Yes Figure 3 The diagram shows an example of the case where the distance method is applied to the groove 4b shown.

[0054] After the shape is numerically evaluated using boundary line detection software, a template of an ideal shape determined by inherent extreme values ​​is defined.

[0055] Then, a scaling is performed in order to be able to compare the different shapes meaningfully with one another.

[0056] In the planimetric method, the area of ​​the silhouette is subtracted from the area of ​​the ideal template and then divided by the area of ​​the ideal template to derive the measurement of the dimensionless quantity.

[0057] In more detail, Figure 4 is a diagram showing the outline of the planimetric method. Figure 4 In (a), the maximum width w of the trench 4b is determined by the length in the direction parallel to the main surface of the semiconductor substrate 1b. max A rectangular region consisting of sides of the same length, sides including the bottom 10 of the groove 4b, and sides of the groove 4b perpendicular to the main surface of the semiconductor substrate 1b is defined as a box region (region indicated by a dotted line) 11. In addition, the portion of the box region 11 excluding the groove 4b ( Figure 4 The portion other than the portion of the checkered pattern in (a) is referred to as the shoulder region 12. Figure 4 As shown in (b) of FIG. 1 , the area number which is an index calculated by the area method is derived by equation (1).

[0058]

Mathematical formula 1

[0059]

[0060] In the distance method, after scaling the data by the local extremum, the point of the silhouette closest to the corner of the ideal template is evaluated to derive the evaluation of a dimensionless quantity.

[0061] In more detail, Figure 5is a diagram showing an overview of the distance method. Figure 5 In (b), the box area 11 and Figure 4 Here, the y coordinate of the bottom 10 is set to y 0 In addition, the maximum width w will be taken max The x-coordinate value at the time is set to x max The corners (x max ,y 0 ) is the nearest point (x, y). Figure 5 As shown in (b) of , the distance number, which is an index derived by the distance method, is derived by equation (2). In addition, α and β are constants used for normalization.

[0062]

Mathematical formula 2

[0063]

[0064] α ≥ x max , β ≥ y max

[0065] By normalizing these values ​​using the intrinsic extreme values ​​of each shape as described above, a single dimensionless value is obtained, and this is used for comparison with the same shape in other images and for evaluating the rectangularity of the sample.

[0066] These normalized values ​​can be used to better judge the results of the experiment and to quantify the ambiguous but important shape information concisely for use in statistical analysis.

[0067] (Installation method)

[0068] Here, two quantitative methods are defined for describing the compatibility with the characteristics of an ideal shape using a quantitative index of a single dimensionless quantity.

[0069] These methods utilize boundary line detection software in order to quickly obtain numerical images of silhouettes.

[0070] (Data preparation)

[0071] refer to Figure 6 To illustrate the preparation of data used in the calculation of the indicator. Figure 6 This is a diagram showing the process of data preparation.

[0072] The data preparation process is the first step to determine the characteristic dimensions of the cross-sectional shape. First, the dimensions of the shape are collected through the outline of the silhouette. They are optimally installed using boundary line extraction software. In addition, in the following description, the dimensions are sometimes also referred to as CD values ​​(critical dimensions).

[0073] Specifically, first, a SEM image obtained by photographing a cross section of a semiconductor processing substrate is obtained (step S1). Next, a silhouette image is extracted from the image of the shape of the object and the boundary line with the background (step S2). Here, SEM is used as an example for explanation, but the present disclosure is not limited to this. It can also be an image obtained by other so-called microscopes such as STEM and TEM.

[0074] Next, the change in silhouette width is considered as a function of the silhouette height. The y component of the data is transformed so that the bottom of the silhouette is at y=0. Specifically, the y component of the data is transformed so that the bottom of the silhouette is at y=0 (step S3).

[0075] Next, the x component of the data is transformed so that the outermost point of the silhouette is at x = 0 (step S4). Specifically, the x component of the silhouette data is transformed using the maximum width value of the silhouette, and then the point on one side of the width when the silhouette becomes the widest is located at x = 0 (step S4).

[0076] Through the above-described processing, a set of unique extreme values ​​and parameters required for calculating the area number and the distance number can be prepared.

[0077] refer to Figure 7 as well as Figure 8 To illustrate Figure 6 A specific example of data. Figure 7 is a diagram showing an example of a silhouette diagram. Figure 8 This is a diagram showing an example of data obtained through the data preparation process.

[0078] Figure 7 This is a silhouette image obtained by obtaining the boundary line from the SEM image using boundary line detection software. Figure 7 Yes Figure 6 The graph obtained after step S2 in . Figure 7 In (a), the contours of the handle layer 2c and the photoresist layer 3c are extracted from a microscope image including the handle layer 2c and the photoresist layer 3c. Figure 7 (b) shows a case where the object range 17 to be evaluated is extracted from the silhouette image.

[0079] Figure 8 (a) means in progress Figure 6 The figure obtained after step S3 in . Figure 8 (a) shows the boundary line (contour) between the groove and the semiconductor processing layer, namely the contour line 20. The contour line 20 is a set of data that specifies N (N is a positive integer) coordinates, and the boundary line is derived by smoothing. In the silhouette diagram, the lowest point of the contour line 20, namely the lowest point 21, is the bottom of the silhouette and represents the lowest part. The silhouette diagram is transformed so that the lowest point 21 is at the position of y=0. In addition, relative to the paper surface, the left side of the contour line 20 represents the portion equivalent to the groove, and the right side of the contour line 20 represents the portion equivalent to the cross-section of the semiconductor processing layer. In addition, the top point 23, namely the top point 20, is the end of the semiconductor processing layer, which is equivalent to the boundary portion between the semiconductor processing layer and the photoresist layer.

[0080] Figure 8 (b) is in progress Figure 6 The most lateral point 22 is the maximum width w of the groove defined in the contour line 20. max The end of the line segment. Figure 8 In (b), the contour line 20 is reversed so that the most side point 22 is at the position of x=0. In the following description, the value of the y coordinate of the top point 23 is also indicated as y max In addition, since the position x=0 is also the location where the maximum width W is achieved max Therefore, the position of x = 0 is also recorded (x max ). In addition, here, an example of transformation in which the contour line 20 is reversed is shown, but the method of transformation is not limited to this. For example, the present disclosure can also be applied when the contour line 20 is moved in parallel.

[0081] The prepared silhouette data is used in the calculation of the area method and the distance method.

[0082] (Calculation using the area method)

[0083] By preparing the silhouette data as described above and defining the inherent extreme values, the area number can be calculated. Figure 4 As shown in FIG. 1 , a block area representing an ideal rectangular shape is set. The process of setting the block area is the second step of making a template of the target shape in the cross-sectional shape obtained from the size after the process of preparing data (the first step). In addition, the process of calculating the index is the third step of comparing the difference between the template and the cross-sectional shape by the normalized index. Here, several methods are considered in the method of setting the block area and calculating the index. The calculation method of the area number corresponding to the method of setting the block area is described as method 1 and method 2.

[0084] (Method 1)

[0085] The ideal shape is determined by the inherent extreme value. In the area method, the dimension is the maximum value of the width of the cross-sectional shape and the maximum value of the depth of the cross-sectional shape, the template is a rectangle formed by the maximum value of the width and the maximum value of the depth, and the indicator is the area difference between the rectangle (template) and the cross-sectional shape.

[0086] Generally speaking, a box created to assume an ideal shape (template) is defined by the maximum width (maximum value of width) and the maximum height (maximum value of depth) observed in the silhouette to be evaluated.

[0087] The x-coordinate value of the widest part of the silhouette max The y coordinate value of the vertex that is not in the silhouette max In this case, the method will be located at a ratio x max Any shoulder area on the upper side is treated in the same manner as the shoulder area on the bottom side of the silhouette.

[0088] Therefore, method 1 can be used to describe the conformity of the optimal overall silhouette to an ideal shape having the same inherent extreme value in order to determine the rectangularity of the silhouette, and the ideal value is 1, which is the same as the case where there is no shoulder area to be measured.

[0089] Fig. 9 This means that the block area 24 representing the ideal shape is set to include a region having an x ​​coordinate value x max The figure shows the situation of the shoulder area above the point. Fig. 9 As shown in (a), in the method 1 of the area method, the block area 24 representing the ideal shape includes: the bottom (y=0) or the lowest point 21; max The end of the case and located at x = x max The point of the position of the groove is the most side point 22; and represents the maximum height of the groove (y = y max ) of the top point 23. The area enclosed by the contour line 20, the coordinate axis of x=0 and the coordinate axis of y=0 corresponds to the shoulder area.

[0090] like Fig. 9 The area of ​​the shoulder region can be calculated by dividing the shoulder region into rectangles having long sides in a direction parallel to the y-axis direction and taking the sum of the areas of the rectangles as shown in (b) of FIG. 1. The rectangle has a side of length Δy in the y-axis direction and a side of length to a point set on the contour line 20.

[0091] Fig.10: is a diagram showing the process of calculating the area number in method 1. First, the length Δy of the side in the y-axis direction of the rectangle dividing the shoulder area is calculated by the following formula (3) (step S11). Here, N is the number of coordinates on the contour line 20 set by the boundary line detection software. Therefore, Δy is determined based on the resolution of the boundary line detection software.

[0092]

Mathematical formula 3

[0093] Δy=y max / N...(3)

[0094] Next, the area of ​​the ideal shape is calculated (step S12). The area S of the block region 24 representing the ideal shape is Ideal The calculation is as follows: Formula (4).

[0095]

Mathematical formula 4

[0096] S Ideal =x y＝0 ×y max =x y＝0 ×Δy×N...(4)

[0097] Next, the area of ​​the shoulder region is calculated (step S13). Shoulde r is approximated by the area of ​​N rectangles and calculated as the following equation (5). i (i is an integer not less than 1 and not more than N) represents the value of the x coordinate of the N coordinates set in the contour line 20.

[0098]

Mathematical formula 5

[0099]

[0100] Finally, using equations (3) to (5), the area value is calculated as in the following equation (6) (step S14).

[0101]

Mathematical formula 6

[0102]

[0103] (Method 2)

[0104] To check and compare the silhouette, replace the y coordinate value y with respect to the height of the ideal shape. max It is effective to define the height by using the y-coordinate value of the point where the silhouette has the widest width as the height.

[0105] Therefore, the area value for x y＝ymax <x max Such a silhouette of the adverse situation y x = xmax <ymax Such arbitrary shapes have extremely high sensitivity. Such conditions occur, for example, in a situation encountered in semiconductor process engineering called "side etch".

[0106] In the perfect rectangular silhouette, the x-coordinate value is close to x max If there are many points with wide silhouette widths such as the value of , defining the widest point as the vertex of the ideal shape will reduce the sensitivity to the rectangular value. On the other hand, if the narrowest point is defined as the vertex of the ideal shape, the value of rectangularity will be very sensitive when describing the exact bottom of the silhouette, but there is a possibility that the value of rectangularity will be misjudged.

[0107] In x max Located in max In other words, when the point where the silhouette has the widest width is the most side point and the top point, the ideal shape is the same as defined in method 1.

[0108] Fig.11 is the value x at the x coordinate max FIG. 34 is a diagram showing a situation where a block region 34 representing an ideal shape is set within a range of points. Fig.11 As shown in (a), in the area method 2, the block area 34 representing the ideal shape includes the bottom (y=0), that is, the lowest point 21; max The end of the case and x = x max The height up to the most lateral point 22 (the y coordinate of the most lateral point 22 ) is equal to the height of the block area 34 .

[0109] like Fig.11 As shown in (b), similar to the case of method 1, the shoulder region is divided into rectangles having long sides in a direction parallel to the y-axis direction, and the area of ​​the shoulder region is calculated by taking the sum of the areas of the rectangles.

[0110] Fig.12 : is a diagram showing the process of calculating the area number in method 2. First, the number P (P is an integer less than N) of rectangles dividing the shoulder area is calculated by the following equation (7) (step S21). Δy is derived from equation (3).

[0111]

Mathematical formula 7

[0112] P=y x＝0 / Δy…(7)

[0113] Next, the area of ​​the ideal shape is calculated (step S22). The area S of the block region 34 representing the ideal shape is Ideal The calculation is as follows: Equation (8).

[0114]

Mathematical formula 8

[0115] S Ideal =x y＝0 ×y y＝0 =x y＝0 ×Δy×P...(8)

[0116] Next, the area of ​​the shoulder region is calculated (step S23). Shoulder The area of ​​P rectangles is used for approximation and the calculation is performed as the following formula (9).

[0117]

Mathematical formula 9

[0118]

[0119] Finally, using equations (7) to (9), the area value is calculated as in the following equation (10) (step S24).

[0120]

Mathematical formula 10

[0121]

[0122] (Calculation of distance method)

[0123] The calculation of the distance method is explained. In the distance method, the size and template can be set in the same way as in the area method, and the indicator is the shortest distance from the corner of the rectangle (template) to the cross-sectional shape. Figure 6 and related descriptions to create silhouette data, and calculate distance values ​​after defining a unique extreme value.

[0124] (Zoom)

[0125] In order to achieve the dimensionless index and to be able to compare different shapes with good discrimination, the silhouette data is scaled by the unique extremum calculated in advance. The scaling method can be selected from the following methods.

[0126] To maintain the aspect ratio, both x and y are scaled by the same constant α.

[0127] Furthermore, if we want to ensure that the distance method index must be less than 1, we need to ensure that α and x are max and max equal to or greater than it. Since in most cases, x max and max is a relative size, and the extreme values ​​on both sides may be greatly different for each sample. Therefore, the above situation is difficult to achieve in practice.

[0128] Therefore, when a single parameter is used to scale both the x-coordinate data and the y-coordinate data, there is no guarantee that the distance value will always be less than 1 or that a meaningful comparison can be made with other silhouettes outside the data set, but even so, it can be useful for analysis.

[0129] We can also import other scaling parameters β = y max , choose α = x max , x and y are scaled by α and β respectively, so that the value of the distance method is always less than 0. At the cost of slightly abstracting the original shape of the silhouette due to deformation of the aspect ratio, a general and meaningful comparison can be made.

[0130] As with the area method, we can also use β = y max Substitute β = y x＝xmax , which is also unfavorable for any "side-etched" samples, but can lead to the values ​​of the anomalies observed in the distance method.

[0131] The region of interest is calculated and then, after choosing a scaling method, the distances of all data points within the data set are determined to find the point with the smallest length and thus the closest point.

[0132] This is the definition of the distance method value where the ideal value for a silhouette of a completely rectangular shape is 0.

[0133] Fig.13 This is a diagram showing an example of a case where a silhouette image is scaled. Fig.13 (a) shows the case of scaling. Fig.13 (b) will Fig.13 The y coordinate data is divided by y x＝xmax The y coordinate value of the (most side point 22) is obtained by dividing the x coordinate data by the x coordinate value of the lowest point 21. The scaling method is not limited to Fig.13 The method shown can be appropriately set according to the type of etching shape and the evaluation method.

[0134] like Fig.13 As shown in (b), an ideal shape 44 is set. The ideal shape 44 is a rectangle with (0, 0), the side point 22 (0, 1), the bottom point 21 (1, 0), and (1, 1) as vertices. In addition, (0, 0) of the ideal shape 44 and the point closest to (0, 0) (x 45) are connected. c ,y c ) is calculated as the distance number.

[0135] Fig.141 is a diagram showing the process of calculating the distance number in the distance method. First, an ideal shape is set in the silhouette image (step S31). The ideal shape has vertices (0, 0), (0, y x＝0 )、(x y＝0 ,y x＝0 )、(x y＝0 , 0). Here, y x＝0 represents the value of the y coordinate of the point at the end of the groove corresponding to the case where the maximum width of the groove is taken. y＝0 Indicates the value of the x coordinate of the point corresponding to the bottom of the groove.

[0136] Next, in the silhouette image, the point closest to (0, 0) is extracted to calculate the distance number (step S32). The distance number is derived by the following equation (11).

[0137]

Mathematical formula 11

[0138]

[0139] In addition, the distance between the vertex (0, 0) of the ideal shape 44 and the closest point is calculated, but depending on the method of setting the ideal shape, the vertex of the ideal shape may not overlap with the coordinates (0, 0). In this case, the distance between the appropriate vertex of the ideal shape and the closest point is calculated as the distance number.

[0140] (System Structure)

[0141] Fig.15 4 is a diagram showing an example of the structure of an evaluation system for performing the evaluation method disclosed in the present disclosure. The evaluation system 40 includes an input unit 41, a control unit 42, a storage unit 43, and a display unit 44. The evaluation system 40 is a search system for searching for etching conditions that make the processing result of the plasma etching device a target shape using machine learning. In addition, in the evaluation system 40, a search method for searching for etching conditions that make the processing result of the plasma etching device a target shape using machine learning is performed.

[0142] The input unit 41 functions as an interface for inputting data between devices and the like into the evaluation system 40, and for inputting a user's request of the evaluation system 40. The input unit 41 acquires a microscope image generated by a microscope device such as an SEM. In addition, the input unit 41 can also acquire data on processing conditions of a semiconductor processing substrate associated with the microscope image.

[0143] The control unit 42 performs processing for image evaluation. For example, the control unit 42 executes boundary line detection software, such as Figure 6 The image is prepared as shown in FIG. 4 and a silhouette image is generated from the microscope image. Fig.10 as well as Fig.12The area method shown, and Fig.14 Processing of the distance method shown.

[0144] The storage unit 43 stores boundary line detection software. In addition, the storage unit 43 stores a program for performing an area method and a program for performing a distance method. In addition, the storage unit 43 stores the area number and the distance number calculated by the control unit 42 in association with the processing conditions of the semiconductor processing substrate.

[0145] The display unit 44 presents the image evaluation result including the area number and the distance number calculated by the control unit 42 to the user.

[0146] Furthermore, the control unit 42 can also perform machine learning using the image evaluation result as one of the target variables to perform processing condition optimization for searching for the optimal processing condition.

[0147] (Evaluation example)

[0148] refer to Fig.16 An example of applying the area method and the distance method to the evaluation of a semiconductor structure is shown below. Fig.16 is a diagram showing the results of evaluating the semiconductor structure. Fig.16 (a) and Fig.16 As shown in (b), two types of semiconductor structures are evaluated. Fig.16 (a) and Fig.16 In (b), Figure (1) is a SEM image, Figure (2) is a silhouette image obtained by performing boundary line detection processing, and Figure (3) is a silhouette image with the color of the silhouette changed for emphasis.

[0149] exist Fig.16 In the structure shown in (a), the area number is 0.846351 and the distance number is 0.385486. Fig.16 In the structure shown in (b), the area number is 0.980214 and the distance number is 0.065626. The ideal shape of the groove formed by etching is a rectangle, so Fig.16 The area of ​​the structure of (b) is greater than Fig.16 Therefore, it can be inferred that Fig.16 The shape of the groove shown in (b) has a space that is closer to a rectangle.

[0150] In addition, regarding the distance number, in particular, the shape of the bottom of the groove shows a difference from the ideal shape, and thus, Fig.16 The distance number of the construction of (b) is less than Fig.16 Therefore, it can be inferred that Fig.16 The shape of the groove shown in (b) particularly ensures the flatness of the bottom.

[0151] (Function, Effect)

[0152] Thus, according to the present disclosure, an index for evaluating the shape appearing in the cross-sectional image can be obtained. Since the actual shape can be quantitatively evaluated with one index relative to the ideal cross-sectional shape, an absolute quantitative evaluation can be performed compared to the evaluation performed using multiple measurement values ​​in the past. Since one index can be associated with the evaluation result, the burden of condition optimization performed by machine learning or the like can also be reduced.

[0153] As mentioned above, although embodiment of this invention is described, this invention is not limited to the said embodiment, Various changes can be made within the range which does not deviate from the summary of this invention.

[0154] The following describes embodiments that may constitute the content of the present invention, but the present invention is not limited thereto.

[0155] (Method 1)

[0156] An evaluation method for evaluating the difference between a target shape and a cross-sectional shape of an electron microscope image, characterized in that it comprises: a first step of measuring a characteristic dimension of the cross-sectional shape; a second step of, after the first step, creating a template of the target shape in the cross-sectional shape obtained from the dimension; and a third step of comparing the difference between the template and the cross-sectional shape using a normalized index.

[0157] (Method 2)

[0158] In the evaluation method described in aspect 1, the dimension is a maximum value of a width of the cross-sectional shape and a maximum value of a depth of the cross-sectional shape, and the template is a rectangle formed based on the maximum value of the width and the maximum value of the depth.

[0159] (Method 3)

[0160] In the evaluation method described in aspect 1 or 2, a feature is that an area difference between the rectangle and the cross-sectional shape is used as the index.

[0161] (Method 4)

[0162] In the evaluation method described in aspect 1 or 3, the shortest distance from the corner of the rectangle to the cross-sectional shape is used as the index.

[0163] (Method 5)

[0164] A search method uses machine learning to search for etching conditions in which a processing result of a plasma etching device becomes a target shape, the search method being characterized in that an evaluation result of an electron microscope image obtained by executing the following steps is set as one of the target variables: a first step of measuring a characteristic dimension of a cross-sectional shape; a second step of, after the first step, creating a template of the target shape in the cross-sectional shape obtained from the dimension; and a third step of comparing the difference between the template and the cross-sectional shape using a normalized index.

[0165] (Method 6)

[0166] A search system uses machine learning to search for etching conditions in which a processing result of a plasma etching device becomes a target shape. The search system is characterized in that an evaluation result of an electron microscope image obtained by executing the following steps is set as one of the target variables: a first step of measuring a characteristic dimension of a cross-sectional shape; a second step of, after the first step, creating a template of the target shape in the cross-sectional shape obtained from the dimension; and a third step of comparing the difference between the template and the cross-sectional shape using a normalized index.

Claims

1. An evaluation method for evaluating the difference between a target shape and a cross-sectional shape of an electron microscope image, The evaluation method is characterized by having: The first step is to determine the characteristic dimensions of the cross-sectional shape; A second step, after the first step, preparing a template of the target shape in the cross-sectional shape obtained from the size; and The third step is to compare the difference between the template and the cross-sectional shape through a normalized index.

2. The evaluation method according to claim 1, wherein: The dimension is the maximum value of the width of the cross-sectional shape and the maximum value of the depth of the cross-sectional shape, The template is a rectangle formed based on the maximum value of the width and the maximum value of the depth.

3. The evaluation method according to claim 2, wherein: The area difference between the rectangle and the cross-sectional shape is set as the index.

4. The evaluation method according to claim 2, wherein: The shortest distance from the corner of the rectangle to the cross-sectional shape is set as the index.

5. A search method for searching for etching conditions for a plasma etching device in which a processing result becomes a target shape using machine learning, The search method is characterized in that an evaluation result of an electron microscope image obtained by executing the following steps is set as one of the target variables: In the first step, the characteristic dimensions of the cross-sectional shape are determined; A second step, after the first step, of making a template of the target shape in the cross-sectional shape obtained from the size; and The third step is to compare the difference between the template and the cross-sectional shape through a normalized index.

6. A search system that uses machine learning to search for etching conditions that result in a target shape for a plasma etching device, The search system is characterized in that an evaluation result of an electron microscope image obtained by executing the following steps is set as one of the target variables: In the first step, the characteristic dimensions of the cross-sectional shape are determined; A second step, after the first step, of making a template of the target shape in the cross-sectional shape obtained from the size; and The third step is to compare the difference between the template and the cross-sectional shape through a normalized index.

Images