JP2015018577A - Pattern matching device and computer program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a pattern matching device and a computer program that are capable of properly selecting an area not to be subjected to pattern matching processing.SOLUTION: In order to achieve the purpose, a pattern matching device that executes pattern matching on an image using a template formed based on design data and that acquires an image in an area to be subjected to matching, extracts an area not to be subjected to matching from the acquired image, and executes pattern matching with the area being a non-matching area, or a computer program is proposed.

Description

  The present invention relates to a pattern matching apparatus and a computer program that causes a computer to execute pattern matching, and in particular, an apparatus that performs alignment between an image obtained based on design data and an image obtained based on imaging, and Regarding computer programs.

  A template matching method is known as a method for searching for a specific pattern existing on a target sample. In particular, with the miniaturization of semiconductor devices, higher accuracy is required for a template matching method for aligning images acquired by a scanning electron microscope (SEM) or the like for semiconductor device evaluation. . Patent Document 1 describes a pattern matching method in which a mask is applied to an area that is not subject to pattern matching processing.

JP 2003-109002 A

  According to the pattern matching method disclosed in Patent Document 1, since a region that is not a pattern matching target can be excluded from the matching target, there is a certain effect in delaying processing time and the like. On the other hand, for example, in the ion implantation process in the semiconductor manufacturing process, a mask region (mask layer) is formed in the non-implanted region before the ion implantation process in order to set a implanted region and a non-implanted region and enable partial implantation. Processing to be performed is performed. When a sample including such a mask layer is to be measured or inspected by an electron microscope or the like, the pattern matching is performed in order to position the field of view (FOV) of the electron microscope or the like in a desired region. However, the mask layer formation accuracy may be lower than the pattern formation accuracy. That is, there is a possibility that a geometrical deviation from a previously registered template for matching may occur due to the displacement of the mask layer. If there is a geometric divergence between the template and the actual pattern formed on the sample, the success rate of matching may decrease. Although it is conceivable to apply a mask as described in Patent Document 1 to the mask layer for such a factor of decreasing the matching success rate, the mask layer position is not stable, so the mask application area is determined. It is difficult to do. If a large mask is applied in consideration of the variation of the mask layer, the region used for matching becomes narrow, which may be a factor for reducing the success rate of matching.

  In the following, a pattern matching apparatus for the purpose of appropriately selecting a region that is not subject to pattern matching processing and a computer program that causes a computer to execute the pattern matching processing will be described.

  As one aspect for achieving the above object, there is provided a computer program or a pattern matching device for causing a computer to perform pattern matching on an image using a template formed based on design data as follows. A pattern matching device that extracts a region that is not subject to matching from the acquired image, sets the region as a non-matching region, and performs pattern matching, or performs the pattern matching on a computer We propose a computer program to be executed.

  In particular, the mask layer or the like formed on the upper layer of the sample is not arranged discretely, but is arranged in a partially assembled state in the FOV, and therefore from the side where the upper layer in the FOV is located. A search is performed to find the upper edge portion. Further, the edge portion of the mask is detected by detecting the signal on the side where the mask layer in the FOV exists.

  According to the above configuration, it is possible to execute the matching process while suppressing a decrease in matching accuracy due to the presence of the upper layer.

  In particular, by performing a search for finding the edge portion of the upper layer from the side where the upper layer is located in the FOV, or by detecting a signal existing on the side where the mask layer or the like is present, it can be mistaken for the edge or the like of the lower layer pattern. The edge portion of the mask can be captured, and as a result, the region necessary for masking the upper layer can be selectively set as a non-matching region.

The schematic block diagram of a scanning electron microscope. The cross-sectional structure (vertical structure) figure and top view (Top Down SEM Image) of the mask layer given in the case of the ion implantation process of a semiconductor device. The flowchart which shows the process of performing pattern matching on the image which set the pattern matching prohibition area | region extracted on the SEM image. The figure which shows the outline | summary of a pattern matching process. The figure which shows the outline | summary of the semiconductor measurement or inspection system containing SEM. The figure which shows an example of the extraction method of the edge part of an upper layer. The figure which shows an example of a peak position detection filter. The block diagram which shows the process of the peak selection using a peak position detection filter. The flowchart which shows the process of selecting a non-search area | region on a SEM image. The figure which shows transition of the coincidence degree of a peak position detection filter and an image when detecting an edge position using a peak position detection filter.

  A pattern matching method is known as one of methods for finding a desired pattern or the like from an image acquired by a semiconductor measurement device including an image acquisition device such as an SEM. An image recognition method typified by such a pattern matching method is a comparison between images of the same type, such as an SEM image and an SEM image, and an OM (Optical Microscope) image and an OM image. Recently, however, image recognition technology using design data of semiconductor devices has appeared. However, the design data shows an ideal shape that does not include shape variation due to process variation. When image recognition is executed based on such an ideal shape, there is a possibility that image recognition may fail due to a deviation from the actual sample state. In an ion implantation process which is one of semiconductor processes, a mask (hereinafter referred to as Top Layer) is set in a non-implanted region in order to set a region where ions are implanted (implanted region) and a region where ions are not implanted (non-implanted region). (Sometimes referred to as an (upper layer) pattern or an Ion Implant pattern). However, it has become clear that the relative position between the Top Layer pattern and the underlying layer (lower layer) is not stable, and pattern recognition may be difficult due to the variation in the position of the edge portion of the Top Layer pattern. . Hereinafter, an image recognition method, apparatus, and computer program suitable for performing pattern matching in a region where an upper layer pattern exists as in the ion implantation process will be described.

  In the following, an image recognition method between design data and an actual SEM image, in which even if the overlay of the top layer of the background of the SEM image greatly deviates, a method capable of image recognition corresponding to the deviation will be described. . In particular, a description will be given of a method that can easily detect a target pattern even in a sample such as Ion Implant Layer where the overlay of a ground layer and a Top Layer pattern may be greatly shifted.

  In order to provide the above method, it is necessary to perform image recognition by removing the Top Layer pattern having a large variation factor.

  For this purpose, a matching algorithm having the following two features is proposed.

  First, when creating a template from design data, the base layer is created by distinguishing Top Layer image recognition templates. As an image recognition procedure, first, since Ion Implant Layer (Top Layer) has a certain shape and directionality, image recognition is performed using the features, and the design data of the background is used for the remaining pattern. Then, image recognition is performed.

  Secondly, in view of the fact that the Top Layers are not discretely arranged in the image, but are partially formed integrally and spread out of the image. Apply a method to selectively exclude partial regions with special features. In this way, by deleting the Ion Implant Layer pattern having a certain shape and directionality, by performing image recognition using only the remaining background, the background is in the presence of misalignment of Top Layer. Regardless, high-precision matching is possible.

  According to the matching method as described above, improvement in the accuracy of image recognition using design data can be expected.

  Hereinafter, a matching algorithm that can stably and highly accurately recognize an image regardless of the position accuracy of the upper layer will be described with reference to the drawings. FIG. 1 is a schematic configuration diagram of an SEM. The primary electron beam 104 applied to the cathode 101 and the first anode 102 is accelerated by the voltage Vacc applied to the first anode 102 and the second anode 103 and proceeds to the subsequent lens system. The primary electron beam 104 is focused as a minute spot on a wafer (sample) 107 by a focusing lens 105 and an objective lens 106 controlled by a lens control power source 114, and two-dimensionally on the wafer (sample) 107 by a two-stage deflection coil 108. Scanned. The scanning signal of the deflection coil 108 is controlled by the deflection controller 109 according to the observation magnification.

  The secondary electrons 110 generated from the sample by the primary electron beam 104 scanned on the wafer (sample) 107 are detected by the secondary electron detector 111. Secondary electron information detected by the secondary electron detector 111 is amplified by the amplifier 112 and displayed on the CRT 113. In the present embodiment, automatic measurement of the pattern is performed by using the sample shape information displayed on the CRT 113.

  FIG. 5 is a diagram showing the measurement and inspection system including the SEM in more detail. The system includes an SEM main body 501, a control device 503 of the SEM main body, and an arithmetic processing unit 505. The arithmetic processing unit 505 is obtained by arranging a recipe execution unit 506 that supplies a predetermined control signal to the control unit 503 and a detection signal obtained by the detector 503 in synchronization with the scanning of the scanning deflector 502. An image processing unit 507 that performs image processing of an image, and a memory 508 that stores acquired image information and recipe information executed by the recipe execution unit 506 are incorporated.

  Electrons emitted from the sample are captured by the detector 503 and converted into a digital signal by an A / D converter built in the control device 504. Image processing according to the purpose is performed by image processing hardware such as a CPU, ASIC, and FPGA incorporated in the image processing unit 507. The image processing unit 507 also has a function of creating a line profile based on the detection signal and measuring a dimension between peaks of the profile.

  Further, the arithmetic processing device 505 is connected to an input device 518 having an input means, and displays a graphic user interface (GUI) for displaying images, inspection results, and the like to the operator on a display device provided in the input device 518. Etc.

  Note that some or all of the control and processing in the arithmetic processing unit 507 can be assigned to a CPU or an electronic computer equipped with a memory capable of storing images and processed and controlled. Also, the input device 518 stores an imaging recipe including the coordinates of the electronic device required for inspection, a pattern matching template used for positioning, imaging conditions, etc. manually or in the design data storage medium 517 of the electronic device. It also functions as an imaging recipe creation device that creates by utilizing the designed data.

  The input device 518 includes a template creation unit that cuts out a part of a diagram image formed based on design data and uses it as a template. The created template is a matching processing unit built in the image processing unit 507. The template matching template 509 is registered in the memory 508. Template matching is a technique for identifying a location where a captured image to be aligned and a template match based on matching degree determination using a normalized correlation method or the like, and the matching processing unit 509 performs matching degree determination. Based on the above, a desired position of the captured image is specified.

  The embodiment described below relates to pattern matching between edge information obtained mainly based on design data and captured images taken by SEM or the like, and edge information obtained based on design data. Is line segment image information indicating the ideal shape of the pattern formed based on the design data, or line segment image information that has been subjected to deformation processing close to the actual pattern by the simulator. The design data is expressed in, for example, the GDS format or the OASIS format, and is stored in a predetermined format. The design data can be of any type as long as the software that displays the design data can display the format and can handle the data as graphic data.

  In the embodiment described below, an example in which matching processing is executed by a control device mounted on the SEM or an arithmetic processing device 505 connected to the SEM via a communication line or the like will be described. The present invention is not limited, and a process as described later may be performed using a general-purpose arithmetic device that executes image processing by a computer program. Furthermore, the method described later can be applied to other charged particle beam apparatuses such as a focused ion beam (FIB) apparatus.

  The embodiment described below relates to an apparatus that performs pattern matching, a program that causes a computer to execute pattern matching, and a storage medium that stores the program.

  FIG. 2 shows a cross-sectional structure (vertical structure) and a top view (Top Down SEM Image) of the mask layer applied in the ion implantation process of the semiconductor device. FIG. 2A is a cross-sectional view and a top view of the mask layer and the underlayer during the ion implantation process. A pattern 202 that is a base pattern is formed on the wafer 201. A mask layer 203a is deposited on the sample so that a part of the pattern 202 serves as an ion implantation region 205 and another part of the pattern 202 serves as a non-implantation region 206, and ions 204 are implanted thereon.

  As shown in the upper diagram of FIG. 2, the mask layer 203a may be properly formed, and as shown in the lower diagram of FIG. 2, the mask layer 203b may be formed in a shifted position. In this way, when the position of the mask layer changes for each sample, in the case of matching using a template, the matching degree between the template and the image at the correct matching position changes depending on the image (sample), and the matching position is incorrect. It is possible to end up. For example, the line end of the pattern 202 may be recognized as the edge portion of the mask layer 203b. Or, since the degree of coincidence is low, there is a possibility that a part that is completely unrelated is recognized as a matching position.

  In a region where the mask layer 203 exists, one of the factors that may cause a matching error is a change in the positional accuracy of the deposition of the mask layer 203 and the inclination of the pattern. Like the mask layer 203a and the mask layer 203b, the Top Down SEM image may change with process changes, whereas the positional relationship between the mask layer 203 and the pattern 202 in the design data is constant. For this reason, in the ion implantation process in which the process variation is severe, it may be difficult to correctly detect the pattern by image recognition using design data.

  In the present embodiment, particularly in the measurement and inspection of a process in which the relative positional relationship of the pattern is remarkably changed due to the process change such as the ion implantation process, it is illustrated in FIG. 3 in order to execute an appropriate matching process. We propose a method for executing the matching process through these steps.

  In step 301, an image recognition template is created by distinguishing the design data into a top layer (Ion Implant Layer) and a base layer as in steps 304 and 305. The pattern directionality is detected using only the top layer (Ion Implant layer) created in step 303 (step 305). Using the directionality of the detected Ion Implant pattern, an Ion Implant pattern is detected from the Top down SEM image (Step 306) (Step 307). When the Ion Implant pattern is detected in the Top down SEM image, the Ion Implant pattern is deleted by image processing (Step 308), or set as a prohibited Area at the time of image recognition using the background in Step 309 (Step 308). In the final image recognition in step 309, the Ion Implant pattern is omitted and image recognition is performed only on the background, thereby improving the image recognition success rate between the design data and the actual Top down SEM image.

  Hereinafter, the steps illustrated in FIG. 3 will be described in more detail with reference to FIGS. 4 and 5. 4A is an actual Top down SEM image, 401 is a ground pattern, and 402 is a Top Layer (Ion Implant Layer) pattern. 4B is a design template of Ion Implant Layer, and C of FIG. 4 is a base design template. Using B in FIG. 4 and C in FIG. 4, the base pattern hidden in the Ion Implant pattern is deleted as shown in FIG. 4D, and a base design template like E in FIG. 4 is created.

  Next, the Ion Implant pattern is detected from the Top down SEM image of FIG. 4A using the Ion Implant design template of FIG. 4B. The detection method of the Ion Implant pattern in the top down SEM image is to detect the edge of the pattern by searching from the right side (in FIG. 4), as indicated by 403 in FIG. Thus, it is possible to detect the Ion Implant pattern without the influence of the background. After detecting the Ion Implant pattern in the Top down SEM image, the Area in which the Ion Implant design of B in FIG. 4 exists is set as an image recognition prohibition Area (404), and E in FIG. The Ion Implant pattern is prevented from interfering with the background image recognition by the image recognition of the background design data) and the F image of FIG.

  In step 301, design data is read from the design data storage medium 301, and in step 302, templates of desired positions (areas) are created for each layer. This template area is an area that includes both the edge portion of the mask layer and the underlying pattern in the design data. The template area selection unit 513 performs area selection based on coordinate information on design data, arbitrarily selected area information, or coordinates of a captured image, magnification information (field of view (FOV) size information), and the like. The layout data of the pattern of the area is selected. The selected area includes the layout data 303 of the mask layer (Top Layer) and the layout data of the base pattern (Low Layer), and a matching template is created based on the layout data of the base pattern (Step 304). At this time, a process for bringing the layout data closer to the SEM image may be performed by performing a smoothing process or the like. Such processing is performed by the image processing unit 512, for example.

  On the other hand, for the mask layer, the position (direction) of the mask layer is detected (step 305). Specifically, it is determined on which side the mask layer exists in the FOV, and is set as a search start point for searching for an edge of the mask layer. In the case of the mask layer as illustrated in FIG. 2, the edge of the mask layer exists on the right side of the SEM image, and the edge of the base pattern exists on the left side relative to the edge of the mask layer. That is, if the right side of the FOV is set as the search start point and the search is started, the first appearing should be the edge of the mask layer. In order to enable such a search, the search direction / start point setting unit 516 determines, for each piece of FOV set on the layout data, to which position on the layout data the piece belongs. When it can be determined that all the pieces belong to the mask layer, a setting may be made such that the piece is set as a search start point. The search start point may be arbitrarily set via the input device 518.

  Further, it is conceivable to use a template having a pixel array as illustrated in FIG. 7 for the search. The template illustrated in FIG. 7 includes pixels of M × N (5 × 5 in FIG. 7), and the center column (3 columns) has the brightest pixel value, and goes to the end (goes to 0 column and 4 columns). And a luminance distribution in which the luminance value gradually decreases. When such a template is searched from the right to the left of the image, the search position where the degree of coincidence between the template and the image is the highest is the brightest part of the edge of the mask layer. And the position of the central row of the template match. In order to detect the edge position of the mask layer based on such coincidence determination, it is preferable to create a template as illustrated in FIG.

  As described above, as the edge position detection method, the start point and direction of the search position may be determined, and the first appearing edge may be detected, or from the right or left using the template. By performing a search to the right, the position with the highest degree of matching obtained may be determined as the edge position. For example, as illustrated in FIG. 10, the degree of coincidence (Degree of Correspondence) between the template and the image with respect to the search position (Position) is obtained, and the position where the degree of coincidence exceeds a predetermined value may be determined as the edge position. good. In particular, this is an effective technique when there is a clear difference in edge width and brightness between the edge portion of the mask layer and the edge portion of the base pattern.

  Further, when there is a large difference between the width and the luminance value of the edge portion of the base pattern and the edge portion of the mask layer, the luminance distribution is acquired in the X direction from the acquired captured image as illustrated in FIG. However, when the peak position is greater than or equal to a predetermined brightness, or when the peak width is greater than or equal to a predetermined value, the position on the image corresponding to the peak position may be determined as the edge position 601.

  When the edge position is detected using the template as described above, the processing conditions for specifying the edge position are registered by the template setting unit 515 or the search direction / start point setting unit 516, and based on the luminance information, When specifying the edge position, conditions such as a threshold value are registered. When creating a template as illustrated in FIG. 7, for example, an edge width information setting unit 801 included in the template setting unit 515 (see FIG. 8) inputs conditions such as edge width and luminance distribution, Create a template.

With the condition set as described above, an SEM image is acquired (step 306), and the edge of the mask layer (Ion Implant pattern) is detected by the method as described above (step 307). This processing is performed by the edge extraction unit 514. The mask setting unit 519 sets the mask layer area (area of the Ion Implant pattern) as an image recognition exclusion area (image recognition prohibited area) (step 308). In this process, a matching exclusion region is set in the mask layer using the edge position detected in step 307 as a reference. This setting includes, for example, a piece that can be determined that all one piece of the FOV set in the layout data belongs to the mask layer, a detected edge position, or a position that is a predetermined value away from the one piece. It is possible to realize this by setting a matching exclusion area.

  A matching process is executed between the SEM image in which the matching exclusion region is set in this way and the template created in step 304 (step 309).

  FIG. 9 is a flowchart illustrating a process of setting a non-search area (matching exclusion area) using the template illustrated in FIG. After acquiring the SEM image (step 901), the peak position is detected using the template (peak position detection filter 803) illustrated in FIG. 7 (step 902). For example, the peak position detection may include a step (step 903) of selectively detecting a peak having a predetermined condition by the peak detection unit 804. Through these steps, a non-search area is set (step 904).

  According to the above-described embodiment, matching processing can be performed after appropriately setting a matching exclusion region according to the obtained SEM image, so that stable matching is performed regardless of process variations. Is possible. In other words, by using the above-described algorithm, it is possible to improve the accuracy and success rate of image recognition using design data even for a pattern such as Ion Implant that has a large process variation.

101 Cathode 102 First Anode 103 Second Anode 104 Primary Electron Beam 105 Focusing Lens 106 Objective Lens 107 Wafer 108 Deflection Coil 109 Deflection Controller 110 Secondary Electron 111 Secondary Electron Detector 112 Amplifier 113 CRT

Claims (10)

In a pattern matching device including an image processing unit that performs pattern matching on an image using a template formed based on design data,
The image processing unit acquires an image of a region to be matched, extracts a region that is not a matching target from the acquired image, and on the image excluding the region that is not a matching target, A pattern matching apparatus that performs pattern matching using the template, and the region that is not the target of matching is a mask layer that blocks ions from reaching the sample by ion implantation. In claim 1,
The image processing unit extracts a region that is not subject to matching based on detection of an edge position of the mask layer. In claim 2,
The pattern processing apparatus, wherein the image processing unit extracts a region that is not subject to matching based on information on a region where the mask layer exists and detection position information on the edge position. In a pattern matching device including an image processing unit that performs pattern matching on an image using a template formed based on design data,
The image processing unit acquires an image of a region to be matched, detects a portion that becomes an upper layer edge from the acquired image by searching the image, and based on the detection of the edge, A pattern matching apparatus that performs pattern matching using the template on the image from which edges are excluded so as to be excluded from the matching target. In claim 4,
The pattern processing apparatus, wherein the image processing unit extracts a region that is not a target of matching based on information on a region where the upper layer exists and detection position information on the edge position. In a computer program that causes a computer to perform pattern matching on an image using a template formed based on design data,
The program causes the computer to acquire an image of a region to be matched, to extract a region that is not a matching target from the acquired image, and on the image excluding the region that is not a matching target. A computer program characterized by executing pattern matching using the template,
The computer program characterized in that the region that is not subject to matching is a mask layer that blocks ions from reaching the sample by ion implantation. In claim 6,
The computer program causes the computer to extract a region that is not subject to matching based on detection of an edge position of the mask layer. In claim 7,
The computer program causes the computer to extract a region that is not subject to matching based on information on a region where the mask layer is present and detection position information on the edge position. In a computer program that causes a computer to perform pattern matching on an image using a template formed based on design data,
The program causes the computer to acquire an image of a region to be matched, causes the upper edge of the acquired image to be detected by searching the image, and based on the detection of the edge, A computer program for executing pattern matching using the template on the image from which the edge is excluded from the matching target. In claim 9,
The computer program causes the computer to extract a region that is not subject to matching based on information on a region in which the upper layer exists and detection position information on the edge position.