JP2012042483A - Image generation device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a connection image of good precision even with a monotonous pattern using design data as a restriction condition, and to generate a connection image of good precision at a high speed by roughly finding a reference position by matching between the design data and image data, and performing matching between adjacent images based upon an amount of deviation from the design data as a retrieval range.SOLUTION: An image generation method is adapted to inspect an electronic device pattern using a scanning electron microscope, and comprises a design data file in which design data describing layout information on the electronic device pattern is input and stored, a plurality of divided image data obtained by imaging the electronic device pattern at different imaging positions, and image connection means of connecting the plurality of divided image data into one image using the plurality of divided image data and the design data of the design data file.

Description

  The present invention relates to an image generation method and apparatus for imaging an arbitrary position on a sample by imaging an electronic device pattern with a critical-dimension scanning electron microscope (CD-SEM) or the like. is there.

  Conventionally, a critical dimension scanning electron microscope (CD-SEM) or the like has been widely used for measuring and inspecting the result of a highly accurate wiring pattern formed on a semiconductor wafer. Currently, semiconductor devices are being miniaturized and mass production of 65 nm is being performed. Defects to be inspected also become smaller as miniaturization progresses, and the magnification at the time of imaging must be high. On the other hand, when performing a wiring pattern layout simulation, there is a need to perform over a wide range. In the future, in order to carry out detailed simulations, it would be desirable to acquire images at a high magnification. However, if the image is taken at a high magnification, the portion that enters the screen becomes small, so it cannot be seen in a wide range.

  [Patent Document 1] proposes a method of acquiring a whole image at a low magnification and then acquiring a divided image at a high magnification. In this method, even if the pattern is monotonous, the pasting position can be specified by referring to the entire image. However, it is necessary to change the magnification and read the entire image. Although there is no problem when imaging the mask, there is a charging problem when imaging the wafer, and the image may be disturbed once read. Therefore, it is preferable that the entire image is not read if possible.

JP 2001-128106 A JP 2003-098112 A JP 2004-333446 A

  As a method for solving this problem, as in [Patent Document 2], there is a method of obtaining a wide range of high-magnification detailed images by dividing an image into a plurality of images based on the position information of the stage and capturing and pasting them together. Proposed. However, although the amount of movement of the stage is quite accurate, it is misaligned when viewed in nm units. When images taken at a high magnification are combined, they are displaced, and the line is cut at the bonding position. This method cannot be used because a displacement of one pixel is not preferable for performing a detailed simulation. [Patent Document 3] captures a plurality of regions so as to overlap with images of adjacent regions, performs matching processing on the overlapping regions, and can obtain a wide range of matched high-resolution detailed images. . However, since there are many monotonous images of the wiring pattern, when matching processing is performed in an overlapping region, a plurality of candidate points for bonding with a high matching rate may appear. In this case, there is a problem in that it is not known which position is actually aligned, and if two or more images are connected, somewhere in the entire image is not always matched.

  An object of the present invention is to obtain a highly accurate connected image even when measuring a monotonous pattern such as a wiring pattern when measuring and inspecting a highly accurate wiring pattern formed on a semiconductor wafer. It is intended.

  In order to achieve the above object, the present invention provides an electronic device pattern image generation method using an electron microscope, storing design data in which electronic device pattern layout information is described, and changing the imaging position to change the electronic device pattern. Storing a plurality of pieces of divided image data obtained by imaging the plurality of pieces of divided image data and connecting the plurality of pieces of divided image data into a single image using the design data of the design data file. It is characterized by doing.

  Further, in the image generation method of the present invention, when performing the image linking, the plurality of divided image data obtained by dividing and imaging the electronic device pattern by providing an overlapping region where adjacent image regions overlap, and A first matching process that performs matching with design data including a pattern of captured divided images and a second matching process that performs matching using image data of an overlapping area between the divided images are performed. Is.

  In addition, the image generation method of the present invention is characterized in that, when performing the image connection, the parameter used in the second matching process is obtained in the first matching process.

  The image generation method of the present invention is characterized in that, when performing the image connection, the parameter of the second matching process obtained in the first matching process is information on the reference position and the search range.

  In order to achieve the above object, the present invention provides an electronic device pattern image generation apparatus using an electron microscope, the design data storage means in which the layout information of the electronic device pattern is described, and the imaging position is changed. Using the plurality of divided image data storage means obtained by imaging the electronic device pattern, the plurality of divided image data, and the design data of the design data file, the plurality of divided image data is converted into one image. And image linking means for linking to.

  Further, in the image generating apparatus of the present invention, the image connecting means includes the plurality of pieces of divided image data obtained by dividing and imaging the electronic device pattern by providing an overlapping region where adjacent image regions overlap. A first matching unit that performs matching with design data including a pattern of the divided images, and a second matching unit that performs matching using image data of an overlapping area between the divided images. Is.

  In the image generating apparatus of the present invention, the image connecting means obtains parameters used in the second matching means by the first matching means.

  In the image generating apparatus of the present invention, the parameter of the second matching means obtained by the first matching means of the image connecting means is information relating to a reference position and a search range.

  Further, in order to achieve the above object, in an image generation apparatus for an electronic device pattern using an electron microscope, an imaging unit for imaging an electronic device pattern and an imaging position for moving the electronic device imaged by the imaging unit to an imaging position Control means, design data storage means in which layout information of the electronic device pattern is described, and divided image data storage means for storing a plurality of pieces of divided image data obtained by imaging the electronic device pattern at different imaging positions; Image connecting means for connecting the plurality of divided image data to one image using the plurality of divided image data stored in the divided image data storage means and the design data of the design data storage means; And the imaging position control means changes imaging conditions based on the design data.

  In the image generation apparatus of the present invention, the imaging position control means obtains and determines a feature amount of an overlapping area using design data including a pattern of an image to be captured in advance when the imaging condition is changed. When the feature amount is smaller than a specific threshold value, the size of the overlapping region where the feature amount is larger is obtained, and imaging is performed by changing the imaging conditions so that the obtained overlapping region size is obtained. It is.

  In the image generating apparatus of the present invention, the feature amount obtained by the imaging position control means is a pattern shape or the number of patterns.

  In the image generation apparatus of the present invention, the imaging position control means detects a characteristic pattern based on design data including a pattern of an image to be captured, and sets an imaging condition so as to be an overlapping region between images. It is characterized by changing.

  Further, in the image generating apparatus of the present invention, the imaging position control means is configured so that a portion to be focused on is not an overlapping area between images but near the center of the image based on design data including a pattern of an image to be captured. The image pickup position is changed.

  Further, in order to achieve the above object, in an image generation apparatus for an electronic device pattern using an electron microscope, an imaging unit for imaging an electronic device pattern and an imaging position for moving the electronic device imaged by the imaging unit to an imaging position A control means; a divided image data storage means for storing divided image data for storing a plurality of pieces of divided image data obtained by imaging the electronic device pattern; and a plurality of sheets stored in the divided image data storage means. An image linking unit for linking to one piece of image data using the divided image data, and the imaging position control unit is required to perform an image linking process based on information on a range to be inspected and an imaging resolution of the imaging unit. This is characterized in that the image pickup position is set.

  In the image generation apparatus of the present invention, when the image connection process is performed, the user is notified that the image connection process has been performed.

  According to the present invention, an accurate linked image can be obtained even with a monotonous pattern such as a wiring pattern.

The block diagram which shows the apparatus structure of this invention. The figure which shows the problem of the connection process of a wiring pattern. The figure which shows the outline | summary of the image connection process of this invention. The figure which shows the processing flow of the reference | standard alignment stage of this invention. The figure which shows the outline | summary which converts design data into an image. The figure which shows the shift | offset | difference of the pattern of an overlap area. The figure which shows the processing flow of the image position shift correction | amendment stage of this invention. The figure which shows the structural example of the image connection means of this invention. The block diagram which shows the apparatus structure of this invention. The figure which shows the outline | summary of a feature pattern. The figure which shows the processing flow of the imaging position change means of this invention. The block diagram of the apparatus structure of this invention. The block diagram of the apparatus structure of this invention. The block diagram of the apparatus structure of this invention. The block diagram of the apparatus structure of this invention. The block diagram of the apparatus structure of this invention. The figure which shows the duplication area | region between images. The figure which shows the thinning and thickening of a pattern.

  Embodiments of an image generation method and an image generation apparatus using a scanning electron microscope such as a SEM apparatus (length measurement scanning electron microscope: CD-SEM) according to the present invention will be described.

  The apparatus configuration of the present invention will be described below with reference to FIG.

  The electron optical system 1 includes an electron gun 100 that generates an electron beam 104, a condenser lens 105 that converges the electron beam 104 generated from the electron gun 100, a deflector 106 that converges and deflects the electron beam 104, and a secondary electron. An ExB deflector 107 for detection and an objective lens 108 that forms an image of the converged electron beam on the sample (mask and wafer) 102 are configured. The sample 102 is placed on the XY stage 103. As a result, the deflector 106 and the objective lens 108 are irradiated with the electron beam irradiation position and the aperture so that the electron beam is focused and irradiated at an arbitrary position on the sample 102 placed on the XY stage 103. To control. By the way, the sample 102 is moved on the XY stage 103, and an image at an arbitrary position of the sample can be acquired.

  On the other hand, secondary electrons and reflected electrons are emitted from the sample 102 irradiated with the electron beam, and the secondary electrons are detected by the secondary electron detector 101. The secondary electron signal detected by the secondary electron detector 101 is converted into a digital signal by the A / D converter 12 and subjected to image processing by the image processing unit 2. In the image processing 2, the signal converted into the digital signal is stored in the image memory 21.

  Here, when acquiring a wide range of pattern images, for example, if the image is divided into nine parts, nine divided pattern images are stored in the image memory 21. Then, in order to connect these nine divided pattern images to one image, the image connecting means 22 uses the divided pattern image data stored in the image memory 21 and the design data from the design data storage means 23. Connect the divided pattern images. When the image is divided into nine, the image is picked up nine times while changing the image pickup position, and the image pickup position and the picked-up divided pattern image are managed in association with each other. Therefore, the rough arrangement of the nine divided pattern images can be known in advance. Then, adjacent divided pattern images at the imaging positions are connected, and finally connected to one pattern image. The pattern images connected to one sheet are stored in the image memory 25 and displayed on the display means 30.

  The problem of wiring pattern connection processing will be briefly described with reference to FIG. Here, a case where nine divided pattern images are connected to one sheet will be described.

  As shown in FIG. 2A, the method of linking by using overlapping regions between images is that when the imaging range is divided into nine regions A to I: nine regions A to I: nine adjacent image regions Images are taken so that they overlap (overlap) at a certain size.

  The wiring pattern to be imaged is often a linear monotonous pattern as shown in FIG. For this reason, when viewed in the overlapping area between images, the number of monotonous patterns increases. For example, when the cut between F and I in FIG. 2B is enlarged, it becomes as shown in FIG. Here, if it is considered that matching is performed between images in the overlapping area of the thick dotted line, the horizontal direction can be specified, but the matching ratio is high at any position in the vertical direction, and the connecting position cannot be specified.

  On the other hand, considering that the pattern images are connected using the design data, there is no need for an overlapping area between the images, and the design data and the entire pattern image are matched. For example, when design data is converted into a pattern image, a pattern image as shown in FIG. 2D is obtained. When each of the captured divided images A to I in FIG. 2B is overlapped with the pattern image in FIG. 2D, a connected image can be generated as shown in FIG. A dotted line is an image pattern obtained from design data, and a solid line is a divided image bonded based on the matching result. In this case, there is no need for an overlapping area between images, and a matching result (bonding position) can be obtained if there is a feature in the image. However, there are places where the detailed pattern image is not connected. For example, FIG. 2F is an enlarged view of the lower right of FIG. 2E (the overlapping region of F and I), and the captured pattern image is affixed at a position that matches the design data. There is a slight gap. The captured pattern was originally created based on the pattern of the design data, but since it has been deformed, it does not completely match even if it is matched. Therefore, we considered combining the method of matching based on the overlapping area between the captured pattern images described above and the method of matching the pattern image created from the design data and the captured pattern image.

  FIG. 3 shows an outline of the image connection process. First, after obtaining the divided image in s01, the pattern image created from the design data is matched with the captured pattern image in the reference position alignment step in s10, and the position is roughly aligned as the reference position. Thereafter, matching is performed based on the overlapping area between the pattern images captured in the image position deviation correction step of s20, and the reference position is finely adjusted.

  Next, the processing flow of the reference alignment stage of the present invention will be described with reference to FIG.

  An image is generated based on the design data in the reference pattern image generation in s11.

  An example of generating an image based on design data is shown. For example, a pattern 501 as shown in FIG. 5 includes four vertices P1 to P4. In the design data, line segment information representing the contour of the pattern shape represented by the x and y coordinate values of the four vertices P1 to P4 is described. The x and y coordinate values are represented by an arbitrary reference coordinate system 500 and have units such as nanometers. A pattern image is generated using information in the design data. Further, it is conceivable that the line segment is white or black according to the image. It is also possible to add deformation. Here, an image generated based on the design data is used as a reference pattern image.

  The matching process is performed on the reference pattern image and the divided pattern image created based on the design data in the matching process of s12. If the image is divided into nine, matching is performed for each of nine images. If the divided pattern image corresponding to the imaging position is known, the matching position corresponding to the design data can be narrowed down by arranging nine images.

  In the reference position storage of s13, the obtained matching position of each divided pattern image is stored as the reference position.

  In the constraint width calculation of s14, a difference between patterns is obtained and stored between the obtained divided pattern image and the reference pattern image. For example, when a pattern as shown in FIG. 6 is formed in the vicinity of an image break or overlapping area, the pattern of the reference pattern image (dotted line (1)) and the pattern of the divided pattern image (solid line (1) ′) ) Is −A, and the difference between the pattern of the reference pattern image (dotted line (2)) and the pattern of the divided pattern image (solid line (2) ′) is B. The matching of the adjacent divided pattern image is also matched with the reference pattern image, and the pattern of the adjacent divided pattern image is considered to be within the pattern width of the reference pattern image. Therefore, when the difference is −A and B, it is considered that there are also patterns of divided pattern images adjacent to the range of −A to + B from the reference position even when they are deviated. Similarly, the difference between the patterns of the divided pattern image is obtained in each overlapping region. Although the captured image pattern is narrower than the design data pattern here, the pattern may actually be thicker as shown in FIG. When the pattern is thick, the difference between the pattern of the reference pattern image (dotted line (3)) and the pattern of the divided pattern image (solid line (3) ′) is A, the pattern of the reference pattern image (dotted line (4)), The difference from the pattern of the divided pattern image (solid line (4) ′) is −B. In this case, it is considered that there are also patterns of divided pattern images adjacent to the range from A to -B from the reference position. The difference between the patterns is stored as a constraint width.

  Next, FIG. 7 is used to show the processing flow of the image position deviation correction stage of the present invention.

  The matching process is performed in an overlapping area between adjacent images of the pattern image captured in the matching process of s21. The width of the overlapping area may be, for example, the MAX value of the imaging error.

  The reference position to be searched is the reference position obtained in s12. The search range is the value of the constraint width obtained in s13. For example, there is a pattern image pattern that is captured in a pattern surrounded by a dotted line design data pattern in FIG. 2E (when the pattern image pattern is narrower than the design data pattern), or (When the pattern image pattern is thicker than the design data pattern), it is assumed that there is a design data pattern inside the captured pattern image, and matching is performed between them. Further, when matching, for example, when nine images are connected, there are overlapping regions at 12 locations (1) to (12) as shown in FIG. It is conceivable that the evaluation value at this time is the total value of the matching rates at the 12 locations. Moreover, even if a restriction is provided, it takes time to match 12 locations in a round robin manner. As a shortening method, it can be obtained by a multi-point search such as GA. First, matching is performed at locations that have characteristics in the overlapping area, and after confirming the bonding position, only the locations that do not have the characteristics in the overlapping area are subjected to round-robin matching to increase the speed. Can be planned.

  In the connection of the divided images in S22, the images are connected to the matching position obtained in S21 to create one image.

  Next, a configuration example of the image connecting means of the present invention will be shown using FIG. The image connecting hand 17 is mainly composed of a reference position aligning unit 2210 and an image position misalignment correcting unit 2220.

The reference alignment unit 2210 reads design data from the design data storage unit 23 and the image generation unit 221 converts it into a pattern image. Here, the converted pattern image is set as a reference pattern image. The reference pattern image is expanded by the expansion processing unit 222. In the matching process 223, the expanded reference pattern image is matched with the captured divided pattern image. The reason why the expansion processing is performed is to perform matching well. In the constraint width calculation unit 224, the divided pattern image and the reference pattern image before expansion are overlapped at the matched position to obtain a difference in pattern width between overlapping regions. This is described with reference to FIG. The obtained reference position and constraint width are stored in the constraint width storage unit 225. The image misalignment correction unit 2220 reads the division pattern image from the image memory 21 and stores it in the partial memory 226, and the matching processing unit 227 performs matching processing between the divided images. At the time of this matching, parameters of an initial value (start point) of matching and a search range (constraint width) of matching are read from the constraint width storage unit 225. For example, if the constraint width as shown in FIG. 6 is -A, B, -A to + B are set as the search range. Then, based on the matching position of each divided pattern image obtained by the matching processing unit 227, a connected image of each divided pattern image is created by the connected image generating unit.

  However, there is a possibility that a size error corresponding to the constraint width will occur. If there is a characteristic pattern in the overlapping area, matching can be determined. Therefore, it is conceivable to change the imaging position so that a characteristic pattern enters the overlapping area.

  Next, an apparatus configuration for changing the imaging position of the present invention will be described with reference to FIG.

  The description of the electron optical system 1 has been described with reference to FIG. The design data including the pattern to be imaged is captured from the design data storage unit 23 to the imaging position changing unit 24, converted into an image, and an overlap region of the converted image is viewed to determine whether there is a feature or not. Search for more overlapping areas. If there is an overlapping area with more features, the stage controller 11 and / or the deflection control means 10 are instructed to obtain an imaging position corresponding to the obtained overlapping area. Further, when connecting the captured divided images, the size of the overlapping area is notified to the image connecting means 22 for each divided image.

  Next, a characteristic pattern for determining whether or not there is a feature will be described with reference to FIG. For example, in FIG. 10A, although there are four line segment patterns, all of them are in the same vertical direction, when matching processing is performed, the matching points in the horizontal direction can be specified, but the vertical direction cannot be specified. FIG. 10B shows a pattern of two line segments, but since the directions are different, the vertical direction and the horizontal direction can be specified. FIG. 10 (c) can be specified in an oblique direction with diagonal lines, but cannot be specified in both vertical and horizontal directions. Although FIG. 10D is connected by one line, since there are line segments in two directions, both vertical and horizontal can be specified. Therefore, it is conceivable that the criterion for the presence / absence of a feature is determined by, for example, whether there are two or more different lines. Therefore, if the design data information is the vertex coordinates of a closed figure, the direction of the line connecting the vertices can be easily obtained. If the positions to be imaged are determined in advance at the time of generating the pattern image, it can be determined whether or not the pattern enters the overlapping area, and therefore, it can be determined how many lines connecting the vertices entering the overlapping area. That is, the number of lines existing in the overlapping region and the respective directions and angles can be easily understood. Therefore, it is determined whether or not there is a feature pattern based on whether or not there are two or more line segments with different directions in the overlapping region. Moreover, although it was set as the line segment here, the shape of a pattern may be sufficient. For example, it is possible to determine the direction of the pattern by obtaining the edge amount for each direction in the vertical, horizontal, and diagonal directions. Here, the edge amount is an output of filter processing. For example, a value obtained based on the output of Laplacian filter processing or vertical, horizontal, and diagonal soble filter processing can be considered.

  The processing flow of the imaging position changing means of the present invention will be described with reference to FIG. Using the design data, the feature amount of the pattern included in the overlapping region is calculated in S31. The feature amount here is the number of line segments with different directions existing in the overlapping region. In S32, it is determined whether or not the feature amount is equal to or greater than the threshold value T1, and if it is equal to or greater than the threshold value T1, the width of the current overlap region is stored in s35, and an imaging position corresponding to the overlap region is obtained. Here, the threshold T1 is the number of line segments in different directions, and the value is 2. If it is not equal to or greater than the threshold T1, the size of the overlapping area is increased by + d in S33. In S34, it is confirmed that the size of the current overlap area is not equal to or greater than the upper limit value L1, and if it is greater than or equal to the upper limit value L1, the overlap area is set as L1 and the width of the current overlap area is stored in S35. An imaging position corresponding to the area is obtained.

  If it is not the upper limit value, the feature amount of the pattern included in the overlapping region is detected in S31, and the above processing is repeated.

  In addition, the part that is desired to be seen in the simulation is considered to be the part where the problem is most likely to occur, and not all of them are necessarily required. For example, when there is a wiring pattern shown in FIG. 12A, it is conceivable that an area of interest is imaged at a position where the image is captured so as not to follow an overlapping area (cut) between images. For example, when it is desired to pay attention to a portion where the angle of two line segments is 90 °, such as a rectangular corner, the rectangular corner is placed in one image divided as shown in FIG. It is conceivable to take an image. Then, the portion that is not noticed so much may be located anywhere in the image. For example, the imaging position of the other portion is shifted so that the portion to be noticed is in the center of the image, or FIG. It is also conceivable that after the divided images are acquired uniformly as shown in FIG. In this way, the effect of matching accuracy can be prevented by avoiding the point of interest to be a break (overlapping region).

  In this case, as shown in FIG. 13, it is conceivable to specify the coordinates of the region to which the user pays attention. Here, when the user designates coordinates with the coordinate designating unit 40, the imaging position changing means 24 obtains the imaging position so that the area of the designated coordinates is at the center of the divided image to be imaged, and the deflection control means 10 and the stage controller. 11 changes the imaging position 24. In this case, the user may specify the coordinates while viewing an image based on the design data on a display screen or the like. The imaging position obtained by the imaging position changing unit 24 is an imaging position where the area designated by the coordinate designating unit 40 is located at the center of the captured image.

  Here, it is described that the user gives an instruction, but the pattern to be noticed is often a pattern specified to some extent. Therefore, as shown in FIG. 14, a pattern of interest is registered in advance in the pattern registration unit 26, and it is determined whether or not the pattern is within the range of the image to be acquired. It is conceivable to change the imaging position so as to be in the center.

  The determination of whether there is a pattern can be realized by a general matching process. If the pattern to be focused on and the pattern image converted from the design data including the pattern to be acquired into a pattern image are matched and the matching value is high, it is determined that there is a pattern to be focused on at that position, and the pattern to be focused on at that position is Change or set to an imaging position that becomes the center when imaging.

  Here, in FIG. 13 and FIG. 14, a pattern to be noticed is specified by a user instruction or matching with a registered pattern, and the image pickup position is changed to an imaging position in which the pattern to be noticed is centered. It is also conceivable that a characteristic pattern is used as an image break (overlapping region). That is, it is conceivable that a characteristic pattern is specified by an instruction from the user or matching with a registered pattern, and the characteristic pattern is changed to an imaging position where an image break (overlapping region) occurs. The characteristic pattern may be considered as the pattern described with reference to FIG.

  In addition, it is troublesome for the user to set in advance how many divided images are required in the image connection process. However, whether or not image connection processing is necessary, or how many divided images are necessary, can be determined by one image range at the magnification for imaging and the image range that the user wants to acquire.

  If the image range that the user wants to acquire is larger than one image range at the magnification at which the image is captured, it is understood that the image connection processing is necessary. For example, when considering only the horizontal direction, the number of divided images can be determined from the image width that the user wants to acquire / one image width at the magnification for imaging-the overlapping region width. For example, an acquired image range instruction unit 50 may be provided as shown in FIG. 15 and the acquired image range may be set by the user. One image range at the imaging magnification may also be set by the user, or the information may be stored in the system information storage unit 27. The imaging magnification may be the limit performance of the imaging device. Then, the system information storage unit becomes a fixed value.

  In addition, when images are automatically connected, if they are not connected well, it is difficult to determine whether they are actually connected due to a defect or because of the accuracy of matching in the image connection process. For this reason, as shown in FIG. 16, the image connection process notification means 60 is provided, and when the image connection process is performed, the fact that the image connection process has been performed is notified. For example, the LED illumination may be turned on, or it may be written in a file. In addition, when displaying a connected image to be displayed on the display device 30, it may be considered that the connected image has been subjected to the image connecting process.

  In addition, by displaying the first or last pixel or multiple pixels of connected images in different colors or brightness, or displaying them with lines, markers, etc., the accuracy of the connected positions can be confirmed visually. it can.

  Further, it is conceivable that the imaging positions of a series of divided images obtained based on the design data are obtained and stored in advance depending on the imaging range. If it does so, what is necessary is just to read and memorize | store the imaging position memorize | stored at the time of imaging.

  In the above embodiment, an example using an image obtained by a secondary electron detector is shown. However, it is also conceivable to use an image generated using secondary reflected electrons when the sample 102 is irradiated with an electron beam. It is done.

  In the above embodiment, the image processing unit 2 of the present invention may be performed by software processing. At that time, software processing may be performed by a personal computer, or it may be incorporated into an LSI and performed by hardware processing.

  As described above, in the present invention, since there are many monotonous images in the wiring pattern, there are cases where the bonding position cannot be specified only by image data. On the other hand, when inspecting the imaged wiring pattern, the inspection is performed using design data in which layout information of the electronic device pattern is described. The design data can be used to narrow down and specify the position where the divided images are to be combined. If it is design data, it can obtain and refer to the pattern information around the point of interest without imaging.

  First, matching processing is performed between a plurality of divided image data obtained by dividing and imaging an electronic device pattern by overlapping an adjacent image region and design data corresponding to the position of each captured divided image. The stage to perform is provided and the bonding position is roughly determined. When matching images are arranged on the basis of design data (all design data are connected), there are places where the images are not connected. The captured image is created based on the design data, but cannot be completely matched. For this reason, even if the matching process is performed, there is a slight shift. Therefore, this time, a stage of matching processing is provided using image data of an overlapping area between divided images. At this time, the amount of deviation between each image and the design data is obtained, and the matching is performed within the range of the obtained amount of deviation, with the position of the combination obtained by matching the design data and the image as the origin. Identify the location.

  Even if all lines are visually connected, the connected images may not match the actual images as long as there is a deviation between the design data and the images. Therefore, it is desirable to have a characteristic pattern in the overlapping area between images as much as possible. The feature is expressed as an edge amount or a vector number and / or a direction number for each of vertical, horizontal, and diagonal directions, and a feature pattern presence / absence determining unit that determines whether or not these values are equal to or greater than a specific value is provided. If the result of the determination is that there is a feature pattern, the image is taken at the current imaging position, and if it is determined that there is no feature pattern, the length of the overlap region that is determined to have a feature pattern is obtained, and the overlap region The image is changed to an image pickup position having a length of. By referring to the design data, it is possible to know in advance what kind of pattern is in the imaging position. Based on the design data, there is a characteristic by changing the imaging position by the imaging position control means that moves the stage to the imaging position. The image data of the overlapping area can be used, and the accuracy of the matching process is increased. In addition, for a series of image capturing positions to be captured for simulation, the image capturing positions obtained based on the feature pattern determination unit are stored in advance.

  In addition, the part that is desired to be seen in the simulation is considered to be the part where the problem is most likely to occur, and not all of them are necessarily required. For example, for the point of interest, move to the position where the stage is imaged based on the design data corresponding to the position of the image to be imaged so that it will be in the center of the screen so that it does not follow the overlapping area (cut) between images By changing the image pickup position by the image pickup position control means, the influence of the matching accuracy can be prevented for the point of interest.

  In addition, the imaging position control means can determine whether or not an image connection process is necessary based on the information to be inspected and the imaging resolution of the imaging means, and can set the imaging position, the number of readings, etc. If you set only the inspection range and imaging resolution, you can automatically set the appropriate imaging position, the number of readings, etc. without being conscious of it, and perform image connection processing to obtain a connected image. Can be improved.

DESCRIPTION OF SYMBOLS 1 Electro-optical system 2 Image processing part 10 Deflection control means 11 Stage controller 12 A / D converter 21, 25 Image memory 22 Image connection means 23 Design data storage means 24 Imaging position change means 26 Pattern registration part 27 Pattern information storage part 30 Display means
40 Coordinate instruction unit 50 Acquired image range instruction unit 60 Image connection processing notification unit 100 Electron gun 101 Secondary electron detector 102 Sample 103 XY stage 104 Electron beam 105 Condenser lens 106 Deflector 107 ExB deflector 108 Objective lens 221 Pixel generation unit 222 Expansion processing units 223 and 227 Matching processing unit 224 Constraint width calculation unit 225 Constraint width storage unit 226 Memory 228 Concatenated image generation unit 500 Reference coordinate system 501 Pattern 2201 Reference alignment unit 2202 Image position deviation correction unit

Claims (2)

In an image generation apparatus for an electronic device pattern using an electron microscope,
Imaging means for imaging an electronic device pattern;
Imaging position control means for moving an electronic device imaged by the imaging means to a position for imaging;
Divided image data storage means for storing divided image data for storing a plurality of divided image data obtained by imaging the electronic device pattern;
Image linking means for linking to one piece of image data using a plurality of pieces of divided image data stored in the divided image data storage means;
The image generation apparatus according to claim 1, wherein the imaging position control unit determines whether or not an image connection process is necessary based on information on a range to be inspected and imaging resolution of the imaging unit, and sets an imaging position. The image generation apparatus according to claim 1,
An image generating apparatus that notifies a user that an image connection process has been performed when the image connection process is performed.