JP6936577B2 - Misalignment amount acquisition device, inspection device, misalignment amount acquisition method and inspection method - Google Patents

Description

The present invention relates to a misalignment amount acquisition device, an inspection device, a misalignment amount acquisition method, and an inspection method.

Conventionally, in the visual inspection of a printed circuit board, the inspection is performed based on design data such as CAM (Computer Aided Manufacturing) data. For example, the image of the design pattern shown by the design data is compared with the image showing the wiring pattern of the printed circuit board, and the place where there is a certain difference or more is extracted as a defect. In such a comparative inspection, the two images are aligned. In the alignment, the amount of misalignment between the two images is obtained.

In addition, Patent Document 1 discloses an inspection device that detects a defect by comparing an image to be inspected and a reference image with each other for each evaluation section divided into regions of a predetermined size. In the inspection device, the amount of misalignment of both images in each evaluation section is obtained, and the frequency distribution with respect to the amount of misalignment can be obtained. The position of the center of gravity of the distribution in the frequency distribution is acquired as the amount of misalignment of the entire image, and the distance between the point farthest from the center of gravity and the center of gravity is acquired as the maximum value of internal distortion in the pattern shown by the image to be inspected.

Japanese Unexamined Patent Publication No. 6-300703

By the way, when the alignment of two images is performed for each divided region of a predetermined size, the pattern of one divided region may be composed of, for example, only linear pattern elements extending in one direction. In this case, the amount of misalignment in the direction perpendicular to the one direction can be accurately obtained, but it is difficult to accurately determine the amount of misalignment in the one direction.

The present invention has been made in view of the above problems, and an object of the present invention is to accurately obtain the amount of misalignment of each divided region in two images.

The invention according to claim 1 is a misalignment amount acquisition device, which is an image dividing unit that divides one of two images to acquire a divided region group, and a misalignment amount between the two images. By obtaining the suitability of each divided region for the shift amount acquisition process to acquire the A position shift amount acquisition unit that acquires the position shift amount of each appropriate region by performing the shift amount acquisition process between the two images, and a non-position that is not included in the plurality of suitability regions of the divided region group. A misalignment amount calculation unit for obtaining a misalignment amount based on the misalignment amount of the plurality of aptitude regions and the positional relationship between the unsuitable region and the plurality of aptitude regions with respect to the aptitude region is provided . The aptitude area specifying part obtains the aptitude by detecting the feature points in the pattern indicated by each of the divided areas, and the aptitude area specifying part is a pixel that is not included in the pattern area in each of the divided areas. The distance from the first target pixel to the edge of the pattern region is a predetermined distance in seven of the eight directions with respect to each of the eight directions set at intervals of 45 degrees among the first target pixels. The distance to the edge of the pattern region among the recessed pixel that is less than and equal to or greater than the predetermined distance in one of the eight directions and the second target pixel included in the pattern region in each of the divided regions. However, one or both of the convex pixel, which is less than a predetermined distance in 7 of the 8 directions and greater than or equal to the predetermined distance in one of the 8 directions, is detected as the feature point. a plurality of the feature points, and the divided region including a predetermined number or more, it determined that proper region.

The invention according to claim 2 is the position shift amount acquisition device according to claim 1, wherein one of the images is an image derived from design data showing a design pattern, and the other of the two images. The image is an image acquired by imaging an object on which an actual pattern is formed based on the design data.

The invention according to claim 3 is the position shift amount acquisition device according to claim 1 or 2 , and in the shift amount acquisition process, while moving each suitable region in the one image up, down, left and right. By obtaining the difference between each of the aptitude regions and the other image of the two images, the amount of misalignment of each of the aptitude regions can be obtained.

The invention according to claim 4 is the position shift amount acquisition device according to any one of claims 1 to 3 , wherein an equation for the position shift amount with the position of each division region as a parameter is preset. The misalignment amount calculation unit obtains the coefficient of the above equation based on the misalignment amount of the plurality of aptitude regions and the positions of the plurality of aptitude regions.

The invention according to claim 5 is an inspection device, wherein the misalignment amount acquisition device according to any one of claims 1 to 4 and each of the divided regions acquired by the misalignment amount acquisition device. It is provided with an inspection unit that acquires inspection results for the images to be inspected included in the two images while aligning the two images using the amount of misalignment.

The invention according to claim 6 is a method for acquiring a displacement amount, which includes a) a step of dividing one of the two images to acquire a divided region group, and b) a position between the two images. A step of identifying a plurality of aptitude regions suitable for the deviation amount acquisition process from the division area group by obtaining the suitability of each divided region for the deviation amount acquisition process for acquiring the deviation amount, and c) in each aptitude region. , The step of acquiring the displacement amount of each appropriate region by performing the displacement amount acquisition process between the two images, and d) the inadequacy not included in the plurality of appropriate regions in the divided region group. The step b) includes a step of determining the amount of misalignment with respect to the region based on the amount of misalignment of the plurality of suitable regions and the positional relationship between the unsuitable region and the plurality of suitable regions. The aptitude is obtained by detecting the feature points in the pattern indicated by each of the divided regions, and the first target pixels, which are pixels not included in the pattern region in each of the divided regions, are spaced at 45 degree intervals. The distance from the first target pixel to the edge of the pattern region is less than a predetermined distance in 7 of the 8 directions and in 1 of the 8 directions with respect to each of the 8 directions set. Of the recessed pixels that are equal to or greater than the predetermined distance and the second target pixel included in the pattern region in each of the divided regions, the distance to the edge of the pattern region is in 7 of the 8 directions. One or both of the convex pixel, which is less than a predetermined distance and is greater than or equal to the predetermined distance in one of the eight directions, is detected as the feature point, and the division includes a plurality of the feature points and a predetermined number or more. region, Ru is determined that proper region.

The invention according to claim 7 is the method for acquiring the amount of misalignment according to claim 6 , wherein one of the images is an image derived from design data showing a design pattern, and the other of the two images. The image is an image acquired by imaging an object on which an actual pattern is formed based on the design data.

The invention according to claim 8 is the position shift amount acquisition method according to claim 6 or 7 , wherein in the shift amount acquisition process, each suitable region in the one image is moved up, down, left and right. By obtaining the difference between each of the aptitude regions and the other image of the two images, the amount of misalignment of each of the aptitude regions can be obtained.

The invention according to claim 9 is the method for acquiring the amount of misalignment according to any one of claims 6 to 8 , wherein an equation for the amount of misalignment with the position of each divided region as a parameter is preset. The step d) includes a step of obtaining the coefficient of the above formula based on the amount of misalignment of the plurality of aptitude regions and the positions of the plurality of aptitude regions.

The invention according to claim 10 is an inspection method, wherein the misalignment amount acquisition method according to any one of claims 6 to 9 and each of the divided regions acquired by the misalignment amount acquisition method. A step of acquiring an inspection result for an image to be inspected included in the two images while aligning the two images by using the amount of misalignment is provided.

According to the present invention, the amount of misalignment of each divided region in the two images can be accurately obtained.

It is a figure which shows the structure of the pattern inspection apparatus. It is a figure which shows the structure of a computer. It is a block diagram which shows the functional structure in a pattern inspection apparatus. It is a figure which shows the flow of the process of inspecting a substrate. It is a figure which shows the division area group. It is a figure which shows a part of the division area. It is a figure which shows a part of the division area. It is a figure which shows the aptitude region. It is a figure which shows the unsuitable region. It is a figure which shows a plurality of aptitude regions in a division region group. It is a figure for demonstrating the deviation amount acquisition process. It is a figure which shows the pixel array of the division area to be inspected. It is a figure which shows the pixel arrangement of the division area to be inspected corresponding to an unsuitable area.

FIG. 1 is a diagram showing a configuration of a pattern inspection device 1 according to an embodiment of the present invention. The pattern inspection device 1 is, for example, a device that inspects the appearance of a printed circuit board (also referred to as a printed wiring board) before electronic components are mounted.

Here, the printed circuit board has a wiring pattern formed on the surface of a resin substrate 9 (hereinafter, simply referred to as “substrate 9”) with a conductive material such as copper. In the production of a printed circuit board, a film (conductive film) of a conductive material is provided on the surface of the substrate 9. A resist film, which is a photosensitive material, is formed on the conductive film, and an image of a pattern based on design data is directly drawn on the resist film by a drawing device (direct drawing device). The substrate 9 on which the pattern is drawn is subjected to a development process, an etching process, a resist peeling process, and the like. As a result, a wiring pattern is formed on the substrate 9. The etching process for the substrate 9 is, for example, wet etching performed by applying an etching solution to the substrate 9. As the etching process for the substrate 9, for example, dry etching using plasma or the like may be performed. Further, the pattern may be formed (exposed) on the resist film by using a photomask showing the design pattern.

The pattern inspection device 1 includes a device main body 2 that captures an image of the substrate 9, and a computer 3 that controls the overall operation of the pattern inspection device 1 and realizes a calculation unit and the like described later. The apparatus main body 2 has an imaging device 21 that images a substrate 9 to acquire (data) a multi-gradation image, a stage 22 that holds the substrate 9, and a stage 22 relative to the imaging device 21. It has a moving stage drive unit 23. The imaging device 21 electrifies an image of the illumination unit 211 that emits illumination light, the optical system 212 that guides the illumination light to the substrate 9 and the light from the substrate 9 incident on, and the substrate 9 formed by the optical system 212. It has an imaging unit 213 that converts a signal. The stage drive unit 23 is composed of a ball screw, a guide rail, a motor, and the like. When the computer 3 controls the stage driving unit 23 and the imaging device 21, a predetermined region on the substrate 9 is imaged.

FIG. 2 is a diagram showing the configuration of the computer 3. The computer 3 has a general computer system configuration including a CPU 31 that performs various arithmetic processes, a ROM 32 that stores basic programs, and a RAM 33 that stores various information. The computer 3 can be read by a computer such as a fixed disk 34 for storing information, a display 35 for displaying various information such as images, a keyboard 36a and a mouse 36b for receiving input from an operator, an optical disk, a magnetic disk, and a magneto-optical disk. Further includes a reading device 37 for reading information from the recording medium 8 and a communication unit 38 for transmitting and receiving signals to and from other configurations of the pattern inspection device 1.

In the computer 3, the program 80 is read in advance from the recording medium 8 via the reading device 37 and stored in the fixed disk 34. The CPU 31 executes arithmetic processing while using the RAM 33 and the fixed disk 34 according to the program 80.

FIG. 3 is a block diagram showing a functional configuration in the pattern inspection device 1, and in FIG. 3, a broken line rectangle with reference numeral 3 indicates a functional configuration realized by the CPU 31, ROM 32, RAM 33, fixed disk 34, etc. of the computer 3. It is surrounded by. The computer 3 has a misalignment amount acquisition device 41, an inspection unit 42, and a storage unit 49. The misalignment amount acquisition device 41 includes an image division unit 411, an appropriate region identification unit 412, a misalignment amount acquisition unit 413, and a misalignment amount calculation unit 414. The storage unit 49 stores design data 48 such as CAM data (or CAD data). Details of the functions realized by these configurations will be described later. It should be noted that these functions may be constructed by a dedicated electric circuit, or a partially dedicated electric circuit may be used.

FIG. 4 is a diagram showing a flow of processing in which the pattern inspection device 1 inspects the substrate 9. Here, it is assumed that a plurality of substrates 9 manufactured as one lot (hereinafter, referred to as “target lot”) are to be inspected. The plurality of substrates 9 included in the target lot are used for manufacturing the same product, and the same wiring pattern is formed.

In the pattern inspection device 1, first, the design data 48 used when forming the wiring pattern on the surface of the substrate 9 (hereinafter, the wiring pattern on the actual substrate 9 is referred to as “actual pattern”) is stored in the storage unit 49. It is input and prepared (step S10). The design data 48 in the present embodiment is vector data indicating a design pattern. The design data 48 may be raster data. The actual patterns on the plurality of substrates 9 included in the target lot are formed by etching or the like based on the design data 48.

The image dividing unit 411 generates a binary image (hereinafter, referred to as “reference image”) showing a design pattern. In the reference image, a plurality of pixels are arranged in the row direction and the column direction, one value (for example, 1) is given to each pixel included in the pattern area of the design pattern, and each pixel included in the background area is given another value. A value (eg, 0) is assigned. As shown in FIG. 5, the reference image 6 is divided into a plurality of regions 60 (hereinafter, referred to as “divided region 60”) (step S11). For example, the reference image 6 is equally divided in the row direction and the column direction, and the plurality of divided areas 60 have the same size. The plurality of divided regions 60 may have different sizes. In the following description, a set of a plurality of divided regions 60 is referred to as a "divided region group".

6 and 7 are views showing a part of the divided region 60. When the divided area group is acquired by the image dividing unit 411, the aptitude area specifying unit 412 sets each pixel not included in the pattern area 61 in the reference image 6, that is, each pixel included in the background area 62 as a target pixel. , The distance from the target pixel to the edge of the pattern region 61 (hereinafter, referred to as “background measurement distance”) is measured in each of the eight directions set at intervals of 45 degrees. In FIG. 6, the background measurement distances in eight directions when the pixel 82 is the target pixel are indicated by arrows 81 (one arrow is designated by reference numeral 81a).

In the target pixel, when only the background measurement distance in one direction is equal to or longer than the predetermined determination distance and the background measurement distance in the other seven directions is less than the determination distance, the target pixel is detected as a concave pixel. In the pixel 82 in FIG. 6, since only the background measurement distance indicated by the arrow 81a is equal to or greater than the determination distance, the pixel 82 is a concave pixel. The recess pixel indicates the presence of the recess 63 (hereinafter, referred to as “pattern recess 63”) in the pattern region 61. The upper limit of the background measurement distance is predetermined, and the arrow 81a in FIG. 6 is the upper limit background measurement distance. The search for the concave pixel may be performed for each predetermined number of pixels in the background region 62 (similar to the search for the convex pixel).

In the aptitude region specifying unit 412, further, each pixel included in the pattern region 61 is set as a target pixel, and the distance from the target pixel to the edge of the pattern region 61 in each of the eight directions set at intervals of 45 degrees (hereinafter, "Pattern measurement distance") is measured. In FIG. 7, the pattern measurement distances in eight directions when the pixels 72 and 72A are the target pixels are indicated by arrows 71 (three arrows are designated by reference numerals 71a, 71b and 71c).

Then, in the target pixel, when only the pattern measurement distance in one direction is equal to or longer than the predetermined determination distance and the pattern measurement distance in the other seven directions is less than the determination distance, the target pixel is detected as a convex pixel. NS. In the pixel 72 in FIG. 7, since only the pattern measurement distance indicated by the arrow 71a is equal to or longer than the determination distance, the pixel 72 is a convex pixel (in FIG. 7, a convex pixel indicating the tip of the pad). Further, in the pixel 72A, since the two pattern measurement distances indicated by the arrows 71b and 71c are equal to or longer than the determination distance, the pixel 72A is not a convex pixel. The convex pixel indicates the presence of the convex 64 (hereinafter, referred to as “pattern convex 64”) of the pattern region 61. The upper limit of the pattern measurement distance is predetermined, and the arrows 71a and 71b in FIG. 7 are the upper limit pattern measurement distances.

In the aptitude area specifying unit 412, the number of convex pixel and concave pixel included in each divided area 60 is obtained, and the divided area 60 in which the number is a predetermined number or more is specified as an aptitude area (step S12). In FIG. 8, three divided regions 60 specified as suitable regions are shown vertically side by side. On the other hand, the divided region 60 in which the number of convex pixels and concave pixels is less than a predetermined number is specified as an unsuitable region. In FIG. 9, three divided regions 60 identified as unsuitable regions are shown vertically side by side. In FIG. 10, parallel diagonal lines are attached to the divided regions 60 specified as suitable regions in the divided region group of the reference image 6.

As will be described later, in the inspection of the substrate 9, the captured image of the substrate 9 and the reference image 6 are aligned. In the alignment, the amount of misalignment between the pattern shown by each divided region 60 of the reference image 6 and the pattern in the captured image that matches (or substantially matches) the pattern is obtained. In order to accurately determine the amount of misalignment with respect to each divided region 60, it is preferable that the pattern shown by the divided region 60 includes the pattern concave portion 63 and the pattern convex portion 64 to some extent (the reason will be described later). Therefore, in the aptitude region specifying unit 412, among the divided region groups, the divided region 60 in which the number of concave pixels and convex pixels is a predetermined number or more is a process of acquiring the amount of misalignment (process by the misaligned amount acquisition unit 413). Therefore, it is treated as an aptitude region suitable for the "deviation amount acquisition process"). In other words, among the divided region groups, the divided region 60 in which the number of concave pixels and convex pixels is less than a predetermined number is treated as an unsuitable region that is not suitable for the deviation amount acquisition process. The unsuitable region is a divided region 60 that is not included in a plurality of suitable regions in the divided region group.

Subsequently, in the apparatus main body 2, the first substrate 9 among the plurality of substrates 9 included in the target lot is placed on the stage 22 (see FIG. 1), and a predetermined region on the substrate 9 is provided by the stage drive unit 23. Is arranged in the imaging region by the imaging unit 213. Then, the imaging unit 213 acquires an captured image showing the actual pattern (hereinafter, referred to as “actual pattern image”) and outputs the image to the misalignment amount acquisition device 41 (step S13). In the actual pattern image, a plurality of pixels are arranged in the row direction and the column direction as in the reference image 6. The resolution of the actual pattern image is the same as the resolution of the reference image 6. In other words, the reference image 6 is generated from the design data 48 according to the resolution of the image captured by the imaging unit 213. The size of the region on the substrate 9 indicated by one pixel of the reference image 6 is the same as the size of the region on the substrate 9 indicated by one pixel of the actual pattern image.

In the misalignment amount acquisition unit 413, the actual pattern image is binarized at a predetermined threshold value, and a binary image showing the actual pattern (hereinafter, referred to as “inspected image”) is generated. Then, the deviation amount acquisition process between the image to be inspected and the reference image 6 is performed in each divided region 60 which is an appropriate region, and the displacement amount is acquired for each appropriate region (step S14).

FIG. 11 is a diagram for explaining a deviation amount acquisition process. In FIG. 11, the division area 60 (indicated by a solid line rectangle with parallel diagonal lines) is arranged vertically and horizontally with respect to the inspected division area 50 (indicated by a broken line rectangle), which will be described later. Shown. In this processing example, as in the case of the divided area group of the reference image 6, by dividing the image to be inspected in the row direction and the column direction, the area 50 corresponding to each divided area 60 of the reference image 6 (in the image to be inspected) ( It is a region indicating the same position on the substrate 9 as the division region 60, and hereinafter, “division region 50 to be inspected”) is specified. In the deviation amount acquisition process, as shown in the center of FIG. 11, each pixel of the divided area 60 is overlapped with the entire divided area 50 to be inspected corresponding to the divided area 60. The difference (absolute value) between the value and the value of the pixel of the image to be inspected that overlaps with the pixel is obtained. Then, the sum of the values of the differences obtained for all the pixels of the divided region 60 is acquired as an evaluation value.

Further, on the image to be inspected, the evaluation value is similarly acquired while moving the division area 60 one pixel at a time in the vertical and horizontal directions (row direction and column direction) from the position of the division area 50 to be inspected. In this way, each pixel of the pixel array of M rows and N columns (M and N are arbitrary integers) centered on the central pixel of the divided region 50 to be inspected overlaps with the central pixel of the divided region 60. , The evaluation value is acquired while arranging the division area 60. Then, in the divided area 50 to be inspected, among a plurality of pixels in which the central pixels of the divided area 60 are overlapped, a pixel having an evaluation value equal to or less than a predetermined threshold value is specified as a specific pixel. When the central pixel of the divided region 60 is arranged at a position where it overlaps with the specific pixel, the difference between the divided region 60 and the image to be inspected is relatively small.

FIG. 12 is a diagram showing an array 510 (that is, pixel array 510) of pixels 51 having 5 rows and 5 columns centered on the center pixel 51a of the divided area 50 to be inspected, and the specific pixels 51 are shaded in parallel. There is. In the shift amount acquisition process in the divided region 60, which is an aptitude region, a plurality of specific pixels 51 are gathered together as a specific pixel group. Then, a vector from the central pixel 51a of the divided region 50 to be inspected toward the position of the center of gravity 52 in the specific pixel group is acquired as the amount of misalignment of the divided region 60. The misalignment amount includes a row direction component and a column direction component, and in the following description, these components are referred to as "row direction misalignment amount" and "column direction misalignment amount". As described above, the misalignment amount acquisition unit 413 moves each aptitude region in the reference image 6 up / down / left / right and obtains the difference between the aptitude region and the image to be inspected to obtain the misalignment amount of the aptitude region. Is obtained.

In addition, even if the division area 60 is arranged according to any pixel 51 of the pixel array 510 of M rows and N columns, if the evaluation value below the threshold value cannot be obtained (the specific pixel 51 is not acquired), the above The same processing as described above is performed on the new pixel array adjacent to the pixel array 510. In the deviation amount acquisition process, the above process is repeated until the specific pixel 51 is acquired or a predetermined number of new pixel arrays are set. If the specific pixel 51 is not acquired even if a predetermined number of new pixel arrays are set, the repetition of the above processing is stopped, and for example, the divided region 60 is changed from an appropriate region to an unsuitable region.

In the pattern inspection device 1, an equation (hereinafter, referred to as “displacement amount calculation formula”) for the amount of misalignment with the position of each division region 60 as a parameter is set in advance. For example, in the deviation amount calculation formula of Equation 1, the position deviation amount ΔX in the row direction and the position deviation in the column direction are used by using the position Xb in the row direction and the position Yb in the column direction of the central pixel of the division region 60 shown in FIG. The quantity ΔY is represented. In Equation 1, αx, βx, γx, αy, βy, and γy are coefficients.

In the misalignment amount calculation unit 414, the positions of the plurality of aptitude regions (positions of the divided regions 60 having parallel diagonal lines in FIG. 10) and the misalignment of the plurality of aptitude regions acquired by the misalignment amount acquisition unit 413. Based on the quantity, the coefficients αx, βx, γx, αy, βy, and γy (values) of the deviation amount calculation formula are obtained (step S15). For example, these coefficients are obtained by the method of least squares. In the least squares method, in the number 2 derived from the number 1, the position Xb in the row direction and the position Yb in the column direction of the central pixel in a plurality of aptitude regions, and the position shift amount ΔX in the row direction and the position shift amount in the column direction ΔY is substituted. Then, the coefficients αx, βx, γx, αy, βy, and γy that minimize the Ex and Ey of Equation 2 are obtained.

When the coefficient of the deviation amount calculation formula is obtained, the row direction position Xb and the column direction position Yb of the center pixel of each inappropriate region (white division region 60 in FIG. 10) are substituted into the deviation amount calculation formula. .. As a result, the misalignment amount ΔX in the row direction and the misalignment amount ΔY in the column direction can be obtained for the unsuitable region (step S16). As described above, the coefficients αx, βx, γx, αy, βy, and γy of the deviation amount calculation formula are the positional deviation amounts of the plurality of aptitude regions and the positions of the plurality of aptitude regions in the row direction and the column direction. Obtained based on. Further, the misalignment amount of the unsuitable region is obtained by a misalignment amount calculation formula using the positions of the inappropriate region in the row direction and the column direction. Therefore, it can be said that the amount of misalignment of the unsuitable region is substantially determined based on the amount of misalignment of the plurality of aptitude regions and the positional relationship between the unsuitable region and the plurality of aptitude regions.

In the inspection unit 42, for example, the central pixel of each division region 60 of the reference image 6 is arranged at a position moved from the center pixel 51a of the corresponding division region 50 by the amount of misalignment of the division region 60. A difference region image showing the difference between the divided region 60 and the image to be inspected is acquired. The difference area image is a binary value indicating the exclusive OR of the value of each pixel of the divided area 60 and the value of the pixel of the image to be inspected that overlaps the pixel in the aligned reference image 6 and the image to be inspected. It is an image. The difference region image shows defect candidate pixels. When the size of a set of defect candidate pixels adjacent to each other is larger than a predetermined area threshold value, the set of defect candidate pixels is detected as a defect region.

In another example of processing in the inspection unit 42, the width of the pattern area or the width of the background area is acquired as the pattern width or the background width at each position of the image to be inspected. Further, in the pattern inspection device 1, a width threshold value is set for each position of the reference image 6, and in the aligned reference image 6 and the image to be inspected, the reference image that overlaps each position of the image to be inspected. The threshold value of the position 6 is acquired. Then, by comparing the pattern width or the background width of the position of the image to be inspected with the threshold value, it is determined whether or not the position of the image to be inspected is defective. As described above, the inspection unit 42 acquires the inspection result for the image to be inspected while aligning the reference image 6 and the image to be inspected by using the amount of misalignment of each division region 60 (step S17). ..

When the inspection of the first substrate 9 is completed, the second substrate 9 included in the target lot is imaged by the imaging unit 213, and the image to be inspected is acquired (steps S18 and S13). Subsequently, in the same manner as described above, the displacement amount of the appropriate region is acquired, and the coefficient of the displacement amount calculation formula is obtained (steps S14 and S15). Since the plurality of substrates 9 included in the target lot are based on the same design data 48, the suitability region in the divided region group is the same for the plurality of substrates 9. Then, the amount of misalignment of the inappropriate region is acquired, and the inspection result is acquired from the aligned reference image 6 and the image to be inspected (steps S16 and S17). When the processes of steps S13 to S17 are performed on all the substrates 9 included in the target lot, the processes in the pattern inspection apparatus 1 are completed (step S18).

Here, a pattern inspection device of a comparative example that performs a deviation amount acquisition process for all the divided regions 60 of the divided region group will be described. In the pattern inspection apparatus of the comparative example, for example, the deviation amount acquisition process is also performed on the three divided regions 60 of FIG. The patterns of the upper, middle, and lower division regions 60 of FIG. 9 are composed of only linear pattern elements extending in the row direction, the column direction, and the diagonal direction, respectively. Therefore, as shown in the upper, middle, and lower rows of FIG. 13, in the pixel array 510 of M rows and N columns centered on the central pixel 51a of the divided region 50 to be inspected, the specific pixels 51 with parallel diagonal lines are in the row direction, respectively. , Lined up in a row and diagonally. As a result, it becomes difficult to accurately acquire the amount of misalignment in the direction in which the specific pixels 51 are lined up. Even if a vector from the center pixel 51a of the divided region 50 to be inspected toward the position of the center of gravity in the specific pixel group is acquired as the amount of misalignment, the amount of misalignment is unreliable, and the divided region 42 by the inspection unit 42 In the inspection result in No. 60, many false reports appear in which a portion that is not a defect is detected as a defect. This is the case when the divided region 60 shows another pattern having periodicity in one direction.

On the other hand, the pattern of the divided region 60 including the pattern concave portion 63 or the pattern convex portion 64 is not limited to only linear pattern elements extending in one direction. In other words, in the divided region 60 (see FIG. 8) where the concave pixel or the convex pixel is detected, the concave pixel or the convex pixel becomes a feature point in the pattern indicated by the divided region 60, and the above M rows and N columns. In the pixel array 510 of the above, the specific pixels 51 are arranged in a row. Actually, it can be said that the larger the number of concave pixels and convex pixels included in each divided region 60, that is, the number of feature points, the more the specific pixels 51 are arranged in a row, and the number of concave pixels and convex pixels is , The degree of suitability for the deviation amount acquisition process of the divided region 60 is shown.

In the pattern inspection device 1, a plurality of suitable regions are specified from the divided region group by determining the suitability of each divided region 60 for the shift amount acquisition process, and the displacement amount of each suitable region is acquired by the shift amount acquisition process. NS. Then, the amount of misalignment with respect to the unsuitable region is obtained based on the amount of misalignment of the plurality of aptitude regions and the positional relationship between the unsuitable region and the plurality of aptitude regions. In this way, by using the misalignment amount of the appropriate region suitable for the misalignment amount acquisition process to obtain the misalignment amount of the inappropriate region unsuitable for the misalignment amount acquisition process, each division in the reference image 6 and the image to be inspected. The amount of misalignment of the region 60 can be accurately obtained. Further, since the deviation amount acquisition processing is performed only on a plurality of appropriate regions, the processing amount in the position deviation amount acquisition unit 413 is reduced as compared with the case where the deviation amount acquisition processing is performed on all the divided regions 60. That is, the time required to acquire the amount of misalignment can be shortened.

Further, in the deviation amount acquisition process, the difference between the appropriate region and the image to be inspected is obtained while moving each appropriate region in the reference image 6 up / down / left / right. This makes it possible to easily obtain the amount of misalignment of the appropriate region. Further, a deviation amount calculation formula with the position of each division region 60 as a parameter is set in advance, and the deviation amount calculation formula is based on the positional deviation amount of the plurality of appropriate regions and the positions of the plurality of appropriate regions. The coefficient is calculated. As a result, the amount of misalignment of the inappropriate region can be quickly obtained.

The inspection unit 42 acquires the inspection result for the image to be inspected while aligning the reference image 6 and the image to be inspected by using the amount of misalignment of each divided region 60. At this time, by aligning the reference image 6 and the image to be inspected for each divided region 60, it is possible to accurately acquire the inspection result of the image to be inspected (reduce false information).

The misalignment amount acquisition device 41 and the pattern inspection device 1 can be modified in various ways.

The aptitude region specifying unit 412 facilitates the suitability of the divided region 60 for the deviation amount acquisition process by detecting the feature points included in the pattern indicated by each divided region 60 with the concave pixel and the convex pixel as the feature points. However, the feature points may be other than the concave pixels and the convex pixels. For example, in the pattern shown by each division region 60, the vertices of the polygonal figure which is the pattern element and the curved portion (excluding the straight line) at the edge of the pattern element may be treated as feature points. Further, the suitability may be determined by a method other than the detection of feature points.

In the image division unit 411, the division area group of the image to be inspected may be acquired. In this case, a plurality of suitable regions are specified from the divided region group, and a deviation amount acquisition process for obtaining the difference between each suitable region and the reference image 6 is performed. As described above, in the pattern inspection apparatus 1, the divided region group is acquired by dividing one of the two images of the image to be inspected and the reference image 6, and a plurality of suitable regions suitable for the deviation amount acquisition process are obtained. , Specified from the divided area group. Further, in the shift amount acquisition process, the position shift of the appropriate region is obtained by finding the difference between the appropriate region and the other image of the two images while moving each suitable region in the one image up / down / left / right. The quantity is obtained.

On the other hand, in the image to be inspected obtained by imaging the substrate 9, when defects or the like are included, the suitability of the divided region cannot be accurately determined. Therefore, from the viewpoint of accurately obtaining the appropriateness of each divided region, it is preferable to acquire the divided region group in the reference image 6 derived from the design data 48 showing the design pattern. Further, when an actual pattern is formed on the plurality of substrates 9 based on the same design data 48, the divided region group is acquired in the reference image 6 derived from the design data 48 to cover the plurality of substrates 9. It is possible to efficiently inspect a plurality of substrates 9 using the same aptitude region for an inspection image.

As the displacement amount acquisition process for acquiring the displacement amount between the two images, a method other than the method described with reference to FIGS. 11 and 12 may be used. In the deviation amount acquisition process, it is possible to adopt various methods related to pattern matching (for example, the normalization correlation method).

In the above embodiment, the amount of positional deviation between the image to be inspected obtained by binarizing the actual pattern image and the binarized reference image 6 is acquired, but in the deviation amount acquisition process, the actual pattern of multiple gradations is acquired. The image may be used as the image to be inspected. In this case, for example, an image of a design pattern in which the average value of the pattern area and the average value of the background area in the image to be inspected are added to the pattern area 61 and the background area 62 is generated as a reference image. Then, the above-mentioned processing is performed by the position shift amount acquisition device 41 in the multi-gradation image to be inspected and the reference image. Similarly, the inspection unit 42 may acquire the inspection result for the multi-gradation image to be inspected.

The deviation amount calculation formula of Equation 1 is only an example, and a quadratic or higher-order formula may be used. Further, the misalignment amount calculation unit 414 may obtain the misalignment amount in the inappropriate region without using the misalignment amount calculation formula. For example, by interpolation calculation using the amount of misalignment of several aptitude regions located around each aptitude region and the positional relationship between these aptitude regions and the aptitude region, the misalignment of the aptitude region is performed. The amount may be determined.

The substrate to be inspected by the pattern inspection apparatus 1 may be a semiconductor substrate, a glass substrate, or the like. Further, in the pattern inspection device 1, in addition to the substrate, a film-like object, a three-dimensional object, or the like may be inspected.

The position shift amount acquisition device 41 for obtaining the position shift amount of the two images may be used for a drawing device or the like, and can be used for acquiring the position shift amount in various types of images.

The above-described embodiment and the configurations in each modification may be appropriately combined as long as they do not conflict with each other.

1 Pattern inspection device 6 Reference image 9 Board 41 Position deviation amount acquisition device 42 Inspection unit 48 Design data 60 Division area 411 Image division area 412 Appropriate area identification unit 413 Position deviation amount acquisition unit 414 Position deviation amount calculation unit S10 to S18 Step

Claims (10)

It is a misalignment amount acquisition device
An image division unit that divides one of the two images and acquires a division area group,
By obtaining the suitability of each divided region for the shift amount acquisition process for acquiring the displacement amount between the two images, the suitability region for specifying a plurality of suitable regions suitable for the shift amount acquisition process from the divided region group. With a specific part
In each aptitude region, a position shift amount acquisition unit that acquires the position shift amount of each aptitude region by performing the shift amount acquisition process between the two images, and a position shift amount acquisition unit.
Based on the amount of misalignment of the plurality of aptitude regions and the positional relationship between the aptitude regions and the plurality of aptitude regions with respect to the unsuitable regions not included in the plurality of aptitude regions in the divided region group. And the position shift amount calculation unit that obtains the position shift amount,
Equipped with a,
The aptitude region specifying unit obtains the aptitude by detecting feature points in the pattern indicated by each divided region.
The aptitude region identification part is
Of the first target pixels that are not included in the pattern area in each of the divided regions, the distance from the first target pixel to the edge of the pattern region in each of the eight directions set at 45 degree intervals. However, the recessed pixels are less than a predetermined distance in 7 of the 8 directions and greater than or equal to the predetermined distance in one of the 8 directions.
Of the second target pixels included in the pattern region in each of the divided regions, the distance to the edge of the pattern region is less than a predetermined distance in 7 of the 8 directions and 1 of the 8 directions. Convex pixels that are greater than or equal to the predetermined distance in the direction
One or detects both as the feature point, a plurality of the feature points, and the divided region including a predetermined number or more, positional shift amount obtaining device, characterized that you determined suitability region of. The misalignment amount acquisition device according to claim 1.
One of the images is an image derived from design data showing a design pattern, and the other image of the two images is acquired by imaging an object on which an actual pattern based on the design data is formed. A misalignment amount acquisition device characterized by being an image. The misalignment amount acquisition device according to claim 1 or 2.
In the shift amount acquisition process, the aptitude regions of the one image are moved up, down, left and right, and the difference between the aptitude regions and the other image of the two images is obtained. A misalignment amount acquisition device, characterized in that the misalignment amount is acquired. The misalignment amount acquisition device according to any one of claims 1 to 3.
An expression for the amount of misalignment with the position of each division region as a parameter is set in advance.
A misalignment amount acquisition device, wherein the misalignment amount calculation unit obtains a coefficient of the above formula based on the misalignment amount of the plurality of suitable regions and the positions of the plurality of aptitude regions. It ’s an inspection device,
The misalignment amount acquisition device according to any one of claims 1 to 4,
An inspection that acquires an inspection result for an image to be inspected included in the two images while aligning the two images using the displacement amount of each of the divided regions acquired by the misalignment amount acquisition device. Department and
An inspection device comprising. It is a method of acquiring the amount of misalignment,
a) A step of dividing one of the two images to obtain a divided area group, and
b) By obtaining the suitability of each divided region for the shift amount acquisition process for acquiring the displacement amount between the two images, a plurality of suitable regions suitable for the shift amount acquisition process are specified from the divided region group. Process and
c) In each aptitude region, a step of acquiring the displacement amount of each aptitude region by performing the deviation amount acquisition process between the two images, and
d) The amount of misalignment of the plurality of aptitude regions and the positional relationship between the aptitude regions and the plurality of aptitude regions with respect to the unsuitable regions not included in the plurality of aptitude regions in the divided region group. The process of finding the amount of misalignment based on
Equipped with a,
In the step b), the suitability is determined by detecting the feature points in the pattern indicated by each of the divided regions.
Of the first target pixels that are not included in the pattern area in each of the divided regions, the distance from the first target pixel to the edge of the pattern region in each of the eight directions set at 45 degree intervals. However, the recessed pixels are less than a predetermined distance in 7 of the 8 directions and greater than or equal to the predetermined distance in one of the 8 directions.
Of the second target pixels included in the pattern region in each of the divided regions, the distance to the edge of the pattern region is less than a predetermined distance in 7 of the 8 directions and 1 of the 8 directions. Convex pixels that are greater than or equal to the predetermined distance in the direction
One or both are detected as the feature points of a plurality of the feature points, and the divided region including a predetermined number or more, positional deviation amount acquisition method, wherein Rukoto is determined that proper region. The method for acquiring the amount of misalignment according to claim 6.
One of the images is an image derived from design data showing a design pattern, and the other image of the two images is acquired by imaging an object on which an actual pattern based on the design data is formed. A method for acquiring the amount of misalignment, which is characterized by being an image. The method for acquiring the amount of misalignment according to claim 6 or 7.
In the shift amount acquisition process, the aptitude regions of the one image are moved up, down, left and right, and the difference between the aptitude regions and the other image of the two images is obtained. A method for acquiring a misalignment amount, which comprises acquiring the misalignment amount. The method for acquiring a misalignment amount according to any one of claims 6 to 8.
An expression for the amount of misalignment with the position of each division region as a parameter is set in advance.
The method for obtaining the amount of misalignment, wherein the step d) includes a step of obtaining the coefficient of the above formula based on the amount of misalignment of the plurality of aptitude regions and the positions of the plurality of aptitude regions. It ’s an inspection method,
The method for acquiring the amount of misalignment according to any one of claims 6 to 9,
A step of acquiring an inspection result for an image to be inspected included in the two images while aligning the two images using the displacement amount of each of the divided regions acquired by the misalignment amount acquisition method. When,
An inspection method characterized by comprising.

Images