JP5596812B2 - Pattern generation apparatus and pattern shape evaluation apparatus - Google Patents

Description

  The present invention relates to a pattern generation apparatus and a pattern shape evaluation apparatus for evaluating the shape of a circuit pattern using a captured image of a circuit pattern of an electronic device manufactured on a wafer, a reticle, or the like.

  In recent years, high-density integration of semiconductor devices for the purpose of improving the performance of semiconductor devices and reducing manufacturing costs has been progressing. In order to realize high-density integration of semiconductor devices, progress in lithography technology for forming fine circuit patterns on a silicon wafer is necessary. Lithography is a process that creates a mask to be an original circuit pattern and transfers the circuit pattern of the mask onto a photosensitive light-receiving resin (hereinafter referred to as a resist) applied on a silicon wafer by an exposure apparatus. The trend of miniaturization has been maintained by various technological innovations such as, modified illumination technology, scan exposure method and chemically amplified resist material. However, since the minimum dimension of the circuit pattern falls below the wavelength of the exposure light source, there is a problem that the allowable range (hereinafter referred to as the process window) when the lithography process conditions change from the optimum state becomes extremely small. It has occurred.

  There are various reasons for the minimization of the process window, such as non-uniform illumination due to shortening of the exposure light source, non-uniformity of processes such as antireflection, baking and development, and variations in mask dimensions. Factors can be classified into an effective dose (hereinafter referred to as “dose”) that behaves similarly to the change in exposure light amount and an effective focus (hereinafter referred to as “focus”) that behaves similarly to the focus change. For this reason, a semiconductor manufacturer measures a test pattern manufactured on a silicon wafer by changing a dose and a focus step by step in a semiconductor development stage, and measures a dose range for manufacturing a normal pattern. The process window is derived by repeatedly performing an operation for obtaining a process window in the focus range (hereinafter referred to as “conditioning operation”, disclosed in Patent Document 1) and a dose and focus variation factor analysis operation by optical simulation to derive optimum lithography conditions. Is expanding.

  However, in the process after 65 nm, it is difficult to obtain an optimum process window for all pattern shapes and pattern arrangements in the development stage due to the high density of patterns and the complexity of pattern shapes. Therefore, it is important to monitor the deformation of the pattern shape due to process variations at the mass production stage.

  In order to monitor the deformation of the pattern shape at the stage of mass production, the pattern shape in the chip (hereinafter referred to as the reference pattern) that does not deform the pattern shape due to process variations is compared with the pattern shape in the chip. Thus, a method for obtaining the pattern deformation amount is effective. As such pattern shape evaluation methods, inventions that evaluate pattern shapes using electronic device design data as reference patterns (Patent Documents 2 to 5), and inventions that perform pattern shape evaluation using non-defective patterns as reference patterns (patents) Document 3) is disclosed.

JP-A-11-288879 Japanese Patent Application No. 6-49264 JP 10-31461 A JP 2002-6479 A JP 2001-338304 A

  However, in the inventions disclosed in Patent Documents 2 to 5, a circuit pattern having a shape completely equivalent to the shape of the circuit pattern in the design data is obtained due to the performance problem of the circuit pattern manufacturing apparatus such as a mask manufacturing apparatus and an exposure apparatus. It is difficult to form on the wafer, and the shape evaluation method based on the circuit pattern of the design data has a problem that incorrect shape evaluation is performed due to the difference in shape. In order to deal with such a problem, Patent Document 4 discloses a method for evaluating the shape of a pattern by using a pattern in which the shape of design data is deformed in advance as a shape formed on a silicon wafer as a reference pattern. Although disclosed, it is difficult to completely predict the shape of the manufacturing pattern, and there is a problem that incorrect shape evaluation is performed due to a mismatch between the predicted pattern and the manufacturing pattern.

  From the above problems, the shape evaluation method based on the best pattern (captured image of a good circuit pattern) disclosed in Patent Document 3 is effective, but on the other hand, There is a problem that an erroneous shape evaluation is performed due to the influence of a local deformation portion of the circuit pattern such as edge roughness generated in each circuit pattern in the manufacturing process. Furthermore, in order to evaluate the shape of a fine circuit pattern, an SEM is generally used as a means for photographing the circuit pattern. However, due to the structure of the SEM, the photographed image includes a lot of noise and luminance unevenness. It is. These noises and luminance irregularities are harmful to the image processing for extracting the pattern shape from the SEM image, and the accurate pattern shape of the non-defective pattern cannot be extracted from the image. There is a problem that it can not be done. Furthermore, there is no guarantee that a good product pattern suitable as a reference pattern, that is, a good product pattern that does not undergo process fluctuations, is necessarily manufactured on the wafer.

  In order to solve the above-described problems, the present invention is an apparatus for generating a reference pattern used for shape evaluation of a circuit pattern of an electronic device, and is used for pattern inspection from contour data of at least two circuit patterns. Reference pattern generating means for generating a reference pattern is provided, the reference pattern generating means performs a filling process on an area sandwiched between the at least two or more contour data, and performs a thinning process on the filled area The present invention proposes a pattern generation apparatus characterized by generating the reference pattern.

  Further, a reference pattern generation means for generating a reference pattern used for shape evaluation of a circuit pattern of an electronic device, the reference pattern generating means generating a reference pattern used for pattern inspection from contour data of at least two circuit patterns. And the reference pattern generation unit generates the contour composite image by superimposing the at least two or more contour data, and generates the reference pattern based on the identification of the overlapping portion of the plurality of contour data. A featured pattern generator is proposed.

  In the pattern generation device, the present invention further includes means for performing pattern matching between the contour data and other contour data to specify an overlapping position between the contour data, and the contour distribution data generating means includes the contour data It is characterized in that contour distribution data is generated based on the overlapping position between them.

  In the pattern generation device, the present invention further includes means for performing pattern matching between the design data and the contour data to identify the overlapping position of the contour data, and the contour distribution data generating means includes the design data and Contour distribution data is generated based on the overlapping position between the contour data.

  In the pattern generation apparatus according to the present invention, the contour distribution data generation unit compares only the contour data satisfying the prescribed value of the circuit pattern shape by comparing the prescribed value of the shape of the circuit pattern with the shape of the contour data. It is used to generate contour distribution data.

  Further, the present invention is the pattern generation device, wherein the reference pattern generation unit specifies a distribution of the contour from the contour distribution data, and sets a reference pattern within the distribution range of the contour. is there.

  In the pattern generation apparatus according to the present invention, the reference pattern generation means generates the reference pattern from the contour distribution data with reference to the shape of circuit pattern design data.

  In order to solve the above problem, the present invention provides a pattern shape evaluation apparatus, wherein the contour distribution data generation unit, the reference pattern generation unit, and the evaluation target for the reference pattern by comparing the reference pattern and the evaluation target pattern It has a shape evaluation means for generating a pattern shape evaluation value.

  In addition, the present invention is characterized in that the pattern shape evaluation apparatus includes a quality determination unit that performs quality determination of the evaluation target pattern using the shape evaluation value.

  In the pattern shape evaluation apparatus according to the present invention, the shape evaluation unit may measure the interval between the reference pattern and the evaluation target pattern, or may provide an average of the intervals, a dispersion of the intervals, or a reference pattern. One of the inside and outside of the evaluation target pattern with respect to the region for allowing deformation of the pattern shape, or the area of the reference pattern and the evaluation target pattern is calculated as the shape evaluation value of the evaluation target pattern. The pass / fail judgment means performs the pass / fail judgment of the pattern by comparing the shape evaluation value with a predetermined value for judging pass / fail of the pattern to be evaluated.

  In the pattern shape evaluation apparatus according to the present invention, all or a part of the contour distribution data generation unit, the reference pattern generation unit, the shape evaluation unit, the pass / fail determination unit, and the evaluation target pattern Image capturing means for acquiring captured images and captured images of the at least two or more circuit patterns for generating a reference pattern; means for extracting contour data of the circuit patterns from the captured images; and the pattern shape evaluation And a data storage means for storing the captured image, the contour data, the contour distribution data, the reference pattern, the shape evaluation value, and the quality determination result data.

  Also, the present invention provides a pattern shape evaluation apparatus, wherein the pattern shape evaluation apparatus includes means for generating a recipe for acquiring a photographed image of the circuit pattern from the circuit pattern design data by the image photographing means. It is a feature.

  The pattern shape evaluation apparatus according to the present invention is characterized in that the pattern shape evaluation apparatus includes means for displaying the contour distribution data, the reference pattern, and data of the shape evaluation result.

  In the pattern shape evaluation apparatus according to the present invention, the pattern shape evaluation apparatus further includes a data input unit for reflecting an instruction from a user to the reference pattern generation unit, the shape evaluation unit, and the non-defective product determination unit. The reference pattern generation unit, the shape evaluation unit, and the non-defective product determination unit perform generation of a reference pattern and shape evaluation based on instruction data from the data input unit.

  In order to solve the above-mentioned problem, the present invention provides a pattern generation apparatus comprising data input means for reflecting an instruction from a user to the reference pattern generation means, wherein the reference pattern generation means is the data input A reference pattern is generated based on instruction data from the means.

  In order to solve the above problems, the present invention provides a pattern shape evaluation apparatus that generates an image showing a wafer map showing the shape evaluation value and the non-defective product determination result, and displays the wafer map image on the data display means. It is characterized by doing.

  According to the present invention, a reference pattern used for evaluating the shape of a circuit pattern of an electronic device is generated from captured images of a plurality of circuit patterns, so that it conforms to the manufacturing conditions of the circuit pattern and suppresses distortion of each circuit pattern. A reference pattern can be generated. By comparing the reference pattern with the evaluation target pattern, the shape evaluation of the shape evaluation pattern can be performed with high accuracy.

3 is a flowchart illustrating a pattern shape evaluation method according to the present invention described in Example 1; It is a structure schematic diagram of the pattern shape evaluation apparatus of this invention. It is the figure which showed the pattern which exists between wafers, shots, between chips, and the same chip. It is the figure which showed the contour data of the pattern extracted from the picked-up image of the pattern, and the picked-up image. It is the figure which showed the procedure which superimposes contour data. It is a flowchart which shows the procedure which produces | generates outline synthetic data. It is the figure which showed the procedure which superimposes the outline data in the same FOV. It is a flowchart which shows the procedure which produces | generates outline synthetic data from the outline data in the same FOV. It is the figure which showed the method of producing | generating a reference pattern from outline synthetic data on the basis of design data. It is the flowchart which showed one method which produces | generates a reference pattern from outline synthetic data. It is the flowchart which showed one method which produces | generates a reference pattern from outline synthetic data. It is the figure which showed the method of specifying the point of a reference pattern from outline distribution data. It is the flowchart which showed the process sequence which performs the shape comparison of a reference pattern and an evaluation object pattern, and produces | generates the shape evaluation value of an evaluation object pattern. It is the figure which showed the example of shape comparison of a reference pattern and an evaluation object pattern. It is the figure which showed the table data which stores the result of a shape evaluation value. It is the flowchart which showed the quality pass / fail judgment procedure. It is the figure which showed the example of measurement of a pattern space | interval. It is the figure which showed the image data for providing a user with the quality determination result. It is the flowchart which showed the procedure for performing the superimposition of outline data with high precision. It is the figure which showed the rotation of the outline data, position shift, and its correction result. It is the figure which showed the example of a synthesis | combination of outline data. It is the flowchart which showed the imaging | photography sequence of SEM. It is the figure which showed the imaging | photography sequence layout of SEM. It is the figure which showed the example which specifies a reference pattern from outline distribution data. It is the figure which showed the relationship between the design data of a circuit pattern, and the centerline of the pattern of design data. It is the figure which showed the example which produces | generates a reference pattern from outline distribution data. It is the flowchart which showed the procedure which excludes the pattern deform | transformed greatly from the synthetic | combination object of an outline, and produces | generates outline synthetic data.

  Embodiments of the present invention will be described with reference to the drawings.

  FIG. 2 shows a schematic configuration block of a scanning electron microscope (hereinafter referred to as SEM) that acquires an image of a circuit pattern according to the present invention. The electron optical system 202 includes an electron gun 203 that generates an electron beam (primary electron) 204, a condenser lens 205 that converges the electron beam 204 generated from the electron gun 203, and a deflector that deflects the converged electron beam 204. 206, an ExB deflector 207 for detecting secondary electrons, and an objective lens 208 that forms an image of the converged electron beam on the semiconductor wafer 201. The wafer 201 is placed on the XY stage 217. As a result, the deflector 206 and the objective lens 208 set the electron beam irradiation position and the aperture so that the electron beam is focused and irradiated at an arbitrary position on the wafer 201 placed on the stage 217. Control. By the way, the XY stage 217 moves the wafer 201 and enables imaging of an arbitrary position of the wafer 201. Therefore, changing the observation position by the XY stage 217 is called stage shift, and changing the observation position by deflecting the electron beam by the deflector 206 is called beam shift. On the other hand, secondary electrons and reflected electrons are emitted from the wafer 201 irradiated with the electron beam, and the secondary electrons are detected by the secondary electron detector 209. On the other hand, the reflected electrons are detected by the reflected electron detectors 210 and 211. The backscattered electron detectors 210 and 211 are installed at different positions. The secondary electrons and reflected electrons detected by the secondary electron detector 209 and the reflected electron detectors 210 and 211 are converted into digital signals by the A / D converters 212, 213 and 214 and input to the processing control unit 215. The shape of the circuit pattern is evaluated by performing image processing according to the purpose by the CPU 251 or the LSI 253 which is image processing hardware, which is stored in the image memory 252. That is, the process control unit 215 uses the stage controller 219 and the deflection control unit 220 to image the shape evaluation point based on the imaging recipe that is created by the imaging recipe generation unit 225 described later and shows the pattern shape evaluation procedure. A control signal is sent to the observation image on the wafer 201, and processing and control such as various image processing are performed to evaluate the shape of the circuit pattern.

  The process control unit 215 observes the global alignment mark on the wafer 201 with an optical microscope (not shown) or the like, thereby including the global alignment control for correcting the deviation of the origin of the wafer 201 and the rotation of the wafer. A stage controller 219 that controls the position and movement, a deflection controller 220 that controls the deflector 206 to control the beam shift (beam deflection) of the electron beam, and a focus controller 221 that controls the focus by controlling the objective lens 208. And connected to. Further, the processing control unit 215 has a function such as a GUI (Graphical User Interface) that is connected to a display 216 having an input unit and displays an image, a shape evaluation result, and the like to the user. In addition, although the Example provided with the detector of two reflected electron images was shown, it is also possible to increase or decrease the number of the detectors of the said reflected electron image. In addition, part or all of the control in the processing control unit 215 can be assigned to a CPU or an electronic computer equipped with a memory capable of storing images, and can be processed and controlled.

  The processing control unit 215 further creates an imaging recipe that includes information about the coordinates of the circuit pattern, a design data template for positioning corresponding to the coordinates, and imaging conditions (including imaging magnification and image quality) for SEM observation. It is connected to the recipe generation unit 225 via a network or a bus. The imaging recipe generation unit 225 is connected to a design system 230 such as an EDA (Electronic Design Automation) tool via a network or the like in order to acquire design data. The imaging recipe generation unit 225 creates an imaging recipe using design data from information on imaging points on the wafer whose shape is to be evaluated. For example, an imaging recipe creation disclosed in JP-A-2006-3517146 is disclosed. The device corresponds to this. The concept of creating a shooting recipe from design data has been proposed for a long time, and does not limit the method and apparatus for generating a shooting recipe from design data. The creation of a photographing recipe is generally executed by software processing of an electronic computer equipped with a CPU, memory, etc., and hardware processing equipped with a CPU, ASIC, FPGA, memory, etc.

  Next, an imaging sequence for observing an arbitrary shape evaluation point (hereinafter referred to as EP) on the wafer will be described with reference to FIG. Further, FIG. 23 shows an addressing point (hereinafter referred to as AP) 2303, an auto focus point (hereinafter referred to as FP) 2302, an auto stigma point (hereinafter referred to as SP) 2306 as a stigma correction point, brightness, and contrast adjustment for EP 2305 on the design layout 2301. FIG. 6 is a diagram showing an example of setting of brightness and contrast point (hereinafter referred to as BP) 2304 as points. The imaging location and imaging conditions (including imaging magnification and image quality) in the imaging sequence, and the shape evaluation conditions in the EP are created by the imaging recipe generation unit 225 as an imaging recipe based on the design data and information on the shape evaluation points, for example, stored. It is stored and managed in the device 223.

  First, the wafer 201 is mounted on the stage 217 (2201). Next, by observing the global alignment mark on the sample with an optical microscope (not shown) or the like, the processing control unit 215 calculates the origin deviation and rotation deviation of the sample, and the stage controller 219 is controlled based on these deviation amounts. Then, correction is performed by controlling the stage 217 (step 2202). Next, the process control unit 215 moves the stage 217, moves the imaging position to the AP according to the coordinates and imaging conditions of the imaging point created by the imaging recipe generation unit 225, and performs imaging at a lower magnification than that during EP imaging. Imaging is performed under conditions (2203). Here, a description of the AP will be added. In order to solve the problem that the observation location is deviated from the field of view of the SEM due to reasons such as positioning accuracy of the stage when directly observing the EP, the imaging recipe generation unit 225 once creates the image for the purpose of positioning, and the storage device 223. The processing control unit 215 temporarily observes the AP whose coordinates registered in the image are known, and the process control unit 215 pre-creates the imaging recipe generation unit 225 and registers it in the storage device 223 and the SEM image of the observed AP. Is used to detect a deviation vector between the center coordinates of the design data template and the center coordinates when the AP is actually observed. Next, the processing control unit 215 controls the deflector 206 via the deflection control unit 220 by the amount obtained by subtracting the detected deviation vector from the relative vector between the coordinates of the design data template and the coordinates of the EP, and the beam. By shifting (inclining the incident direction of the beam and changing the irradiation position), moving the imaging position and observing the EP, the EP can be imaged with high coordinate accuracy (generally beam shift). The positioning accuracy is higher than the positioning accuracy of the stage). Next, based on the control and processing of the processing control unit 215, the imaging position is moved to the FP by beam shift and imaging is performed to obtain autofocus parameters, and autofocus is performed based on the obtained parameters (2204). .

  Next, based on the control and processing of the processing control unit 215, the imaging position is moved to the SP by beam shift and imaging is performed to obtain astigmatism correction parameters, and automatic astigmatism is calculated based on the obtained parameters. Correction (auto stigma correction) is performed (2205).

  Next, based on the control and processing of the processing control unit 215, the imaging position is moved to the BP by beam shift and imaging is performed to obtain parameters for brightness & contrast adjustment. Based on the obtained parameters, automatic brightness & contrast is obtained. Adjustment is performed (2206). Note that the addressing, autofocus, autostigma, autobrightness & contrast in steps 2203, 2204, 2205, and 2206 described above may be partially or completely omitted depending on the case, or the order of 2203, 2204, 2205, and 2206 is arbitrary. There are variations such that the coordinates of AP, FP, SP, and BP overlap (for example, autofocus and autostigma are performed at the same location). Finally, based on the control and processing of the processing control unit 215, the imaging position is moved to the EP by beam shift, and imaging is performed. The design data template in the EP registered in the storage device 223 and the SEM image of the observed EP are Matching is performed to calculate the shift amount of the shape evaluation target point in the SEM image (2207).

  Next, the pattern shape of the present invention is evaluated using the SEM image of the EP and the shift amount by EP matching. In pattern shape evaluation, first from the captured image of a plurality of patterns with different production points or wafers on the wafer, pattern deformation due to process fluctuations, distortion of the pattern shape due to the influence of noise included in the image, A reference pattern in which distortion such as roughness of the pattern is suppressed is generated. Next, the shape of the evaluation target pattern with respect to the shape of the reference pattern is generated by comparing the shape of the reference pattern with the shape of the evaluation target pattern. The plurality of patterns indicate a plurality of patterns whose shapes are equivalent on the design data.

  The plurality of patterns are, for example, patterns having conditions as shown in (A) to (D) below. FIG. 3 is a diagram showing the relationship between these A to D. The relationship between the two wafers 301 and 305, the shot area 302 in the wafer 301, the chip 303 in the shot area, and the pattern 304 in the chip is shown. FIG. A shot is an exposure area that can be transferred at once, and a plurality of chips exist in this shot.

(A) A plurality of patterns with different wafers (Example 306).
(B) A plurality of patterns with different shots (Example 307).
(C) A plurality of patterns having different chips in the same shot (example 308).
(D) A plurality of patterns having different coordinates in the same chip (example 309).

  The above condition is merely an example, and a reference pattern may be generated using a plurality of patterns in which these A to D are mixed.

  Hereinafter, the pattern shape evaluation method of the present invention will be described in detail.

  FIG. 1 shows a flowchart of a pattern shape evaluation method according to the present invention, which is executed by software processing using the CPU 251 and the image memory 252 of the processing control unit 215. However, the CPU, memory, etc. of an electronic computer that can input an image from the electron optical system 202 and a design data template from the photographing recipe generation unit 225 via a LAN or bus, or a storage medium such as a portable memory or a hard disk. It can also be executed by using software processing. Details of each step will be described below.

  First, a captured image of a circuit pattern is read (101). An image used for pattern shape evaluation is a recipe for capturing images of a plurality of patterns for the purpose of generating a reference pattern under the conditions A to D and an evaluation target pattern in the shooting recipe generation unit 225. Is obtained by controlling the electron optical system 202 with the recipe.

  Next, the outline of the circuit pattern is extracted from the image (102). Contour extraction has been proposed in various ways. For example, the method disclosed in Japanese Patent Application Laid-Open No. 2006-66478, “R. Matsuoka, New method of Contour based mask shape compiler, SPIE Proc 6730-21, 2007.9. 21 ”can be applied. When a pattern is photographed with the SEM, as shown in FIG. 4A, an inclined portion and a protruding portion of the pattern are imaged as a white belt-like image. By applying the method disclosed above, the white belt image 401 can be extracted as outline data 403 of a line drawing as shown in FIG. Image reading (101) and contour extraction (102) are executed for all images of the reference pattern generation image and the evaluation target pattern (103).

Next, a reference pattern is generated using a plurality of contour data extracted from the reference pattern generation image. For the generation of the reference pattern, in order to suppress the deformation of the pattern due to process fluctuations, the distortion of the pattern caused by the roughness of each pattern, and the distortion of the contour due to the noise included in the image, more patterns are used. It is desirable to use it.
[Outline composition]
As a preliminary preparation for generating a reference pattern, a plurality of contour data 2101 and 2102 extracted from an image are superimposed as shown in FIG. 21A to generate a contour composite image (104). As described above, when the contour position of the pattern in the image differs for each contour data due to problems such as the positioning accuracy of the stage, the contour data is superimposed in consideration of the contour position in the image. The overlapping position can be automatically specified by using the EP matching result used for specifying the shape evaluation point. A method for automatically specifying the overlapping position will be described with reference to FIG. FIGS. 5A to 5C show three pieces of contour data used for generating a reference pattern. These are problems such as stage position accuracy, and the contour positions in the image are different. FIG. 5D shows a design data template used for matching. By EP matching, shift amounts 501 to 503 of the images (e), (f), and (g) with respect to the design data template can be detected. With reference to this shift amount, the overlapping position 504 of the three contour data is specified as shown in FIG. Although the shapes of the design data and the contour data are different, various techniques for absorbing the difference between these shapes and obtaining the matching position with high accuracy and detecting the shift amount have been proposed, for example, disclosed in Japanese Patent Application Laid-Open No. 2007-79982. The shift amount can be detected by applying the above method.

  FIG. 6 shows a processing flow for superimposing contour data. First, a plurality of contour data (or an image before contour extraction) is read (601). Next, matching between the contour data (or the image before extracting the contour) and the design data template is performed (602). From the result of matching, the shift amount of each contour data and design data template is calculated (603). Next, an image is formed by superimposing a plurality of contour data on the basis of the shift amount (604). The contour composite image is written into the memory or the like (605).

  In addition, even when multiple images to be superimposed are included in one image, they can be easily superimposed using the shift amount by EP matching and the pattern interval derived from the design data. Can do. An example of synthesizing a plurality of contour data in the same image will be described using the processing outline diagram of FIG. 7 and the flowchart of FIG. First, contour data is read (801). Next, pattern matching is performed between the design data template and contour data (or an image from which the contour is extracted) (802, equivalent to EP matching). From the pattern matching result, the shift amount (FIG. 7A) 701 is calculated (803). Next, contour synthesis points 702 to 704 on the contour data and synthesis regions 705 to 707 are determined from the shift amount 701 and the pattern arrangement interval on the design data. Next, a contour composite image is generated 805 based on the contour composite points 702 to 704 and the composite regions 705 to 707. FIG. 7B shows an image in which the contour synthesis points 702 to 704 are made to correspond to the coordinates 708 of the contour synthesis image and the contour data of the synthesis regions 705 to 707 are superimposed. Finally, the contour composite image is written in the memory (806). Note that it is also possible for the user to specify a contour synthesis point and a synthesis area via input means connected to the display 216. In this case, the contour data is displayed on the display 216, and the contour synthesis point and the synthesis area designated by the input means are stored in the memory. Based on the contour synthesis point and the synthesis area, the contour data is synthesized and written into the memory.

An example of the contour composite image is shown in FIG. When the evaluation target patterns have different shapes due to process fluctuations, individual pattern roughness, and the like, a contour composite image having a pattern shape distribution (hereinafter referred to as a contour distribution) as shown in FIG. Is obtained. A reference pattern is generated from such a contour composite image (105).
[Reference pattern generation]
A reference pattern is determined based on a predetermined rule from the contour distribution state of the contour composite image. The predetermined rule is to be changed according to the purpose of shape evaluation, and is not limited thereto. Two examples of generating a reference pattern from a contour composite image will be described below.

1) Average shape of contour distribution An average point of the contour distribution is converted into a reference pattern. By comparing the shape of the reference pattern with the shape of the evaluation target pattern, the shape of the evaluation target pattern can be evaluated with respect to the average pattern shape. FIG. 24A is a diagram illustrating the contour distribution, and is a diagram illustrating an example in which an average point of the contour distribution is converted into a reference pattern (broken line). FIG. 24C is a graph showing the number of contours on a straight line of the contour distribution QP, and shows an average reference pattern point 2401 in the contour distribution between QP.

  In order to specify the average point of the contour distribution, the number of overlapping contours when the contour composite image is generated is stored as the pixel value of the contour composite image. Further, by applying a moving average filter or the like in the P11 smoothing and noise removal section of computer image processing (hereinafter referred to as Reference Document 1) by Hideyuki Tamura to this contour composite image, an average of the contour distribution is obtained. A contour composite image having a peak at a point can be generated. Finally, a continuous peak position in the contour distribution is specified, and the peak position is set as a reference pattern (broken line in FIG. 23A).

2) Center shape of the contour distribution frame The center of the range frame of the contour distribution is converted into a reference pattern. By comparing the shapes of the reference pattern and the evaluation target pattern, it is possible to evaluate the shape of the evaluation target pattern with respect to the center of the shape deformation range of the evaluation target pattern. FIG. 24B is a diagram showing an example in which the same contour distribution and the center position of the contour distribution frame as in FIG. 24A are converted to a reference pattern (solid line), and FIG. FIG. 6 is a graph showing the number of contour lines on the straight line, and shows the contour distribution frames 2403 and 2404 between Q and P, and the reference pattern points 2402 corresponding to the center positions of the contour distribution frames 2403 and 2404. Yes.

  FIG. 10 shows a flowchart of a method for detecting the center of the contour distribution frame from the contour composite image.

  First, a contour composite image is read (1001). Next, a contour is filled 1002. FIG. 12B is an enlarged view of the contour distribution of the contour composite image as shown in FIG. Thus, in the contour distribution, there are a point where a contour exists and a point where a contour does not exist. Filling is a process of writing the same value as the outline to the pixels in the area between the outlines. For example, when the contour composite image is binary data, the pixel value of the point where the contour is present is “1”, and the point where the contour does not exist is “0”, the term of P154 “Shrinkage and expansion” in Reference 1 By applying an image processing method using a certain morphology filter (expansion processing → contraction processing), all the pixel values in the region sandwiched between the contours are changed to “1” as shown in FIG. be able to.

  Next, the center position 1203 of the contour distribution is detected 1103 using the contour filling result. The center position can be detected, for example, by applying a thinning process described in “Thinning” in P158 of Reference 1. The thinning process is a process aiming to specify the center line of a pattern distributed over a wide range. By thinning the filled outline distribution, the center position of the outline distribution as shown in FIG. 1203 can be determined. The center position of the contour distribution is set as a reference pattern and written in a memory or the like 1004.

  Further, the reference pattern position is not the center position of the contour distribution frame, but a point 1213 shifted by several pixels from the center position 1203 of the contour distribution toward the outer frame 1202 or the inner frame 1201 as shown in FIG. It is also possible to use 1214. FIG. 11 shows a processing flow of a method for generating such a reference pattern. Note that 1101 to 1103 are equivalent in processing to the processing flows 1001 to 1003 shown in FIG.

  After specifying the center position of the contour distribution, an inner frame 1201 and an outer frame 1202 as shown in FIG. In the case of a contour distribution image in which the pixel value of the contour after filling is “1” and the other region is “0”, a portion that changes from the pixel value “0” to “1” with respect to the center line 1212 1201, a part that changes from “1” to “0” with respect to the center line 1212 of the pattern shape is an outer frame 1202. The position of the center line of the pattern shape in the contour distribution image is the design data 1211 and the shift amount of the contour data (any one of a plurality of patterns) by the EP matching as shown in FIG. The shape can be specified by using the center line 1212 of the shape. A method for obtaining the center line position of the pattern shape from the design data will be described with reference to FIG.

  As shown in FIG. 25A, the design data is provided from the design system 230 or the like as a set of a plurality of vertex coordinate data 2501 constituting the closed figure of the pattern. A pattern of design data is obtained by connecting the vertex coordinates with a straight line. A pattern composed of the straight lines is drawn, and an image in which the inside of the pattern is filled as shown in FIG. 25B is created. For example, the pixel value of the straight line and the painted area of the design data is “1”, and the other areas are “0”.

  Next, the center line position of the pattern shape of the design data as shown in FIG. 25C can be specified by applying the thinning process described above to the image in which the design data pattern is filled.

  The center line 1212 in the contour distribution image can be specified from the center line position of the pattern shape of the design data specified in this way and the EP matching shift amount, and depending on the positional relationship between the center line 1212 and the switching position of the pixel value. The inner frame 1201 and the outer frame 1202 can be specified.

  After specifying the inner frame 1201 and the outer frame 1202 using the center position of the pattern shape, the position of the reference pattern is specified 1105 based on the center position 1203 of the contour distribution. An example of specifying a reference pattern based on a rule such as “points of L pixels in the outer frame direction with respect to the center position of the contour distribution” will be described with reference to FIG.

  First, the morphology filter (expansion process) described above is applied to the pattern at the center position 1203 of the contour distribution. The expansion process is a process for expanding the width of the pattern. The pattern of the center position of the contour distribution is 1 pixel wide. By performing dilation processing once on this pattern, it is possible to generate a pattern with a width of 3 pixels centered on the pixel at the center position of the contour distribution. FIG. 26 shows an example of the expansion process. FIG. 26 is a diagram showing a center pattern 2602 and an inner frame 2601 and an outer frame 2603 of the contour distribution. When the expansion process is performed once on the center pattern 2602, both sides of the center pattern 2602 are expanded by one pixel as in the enlarged window of FIG. The pixel newly added by the expansion process is a pixel existing at a point one pixel away from the center pattern 2602. Therefore, the distance value “1” from the center pattern is written in the pixel newly added in the first expansion process. By repeating this expansion process and writing of the distance value, the distance value from the center pattern is set to the pixel existing between the inner frame 2601 of the contour distribution and the outer frame 2603 of the contour distribution as shown in the enlarged window FIG. Can write. By using the distance information from the center of the contour distribution obtained in this way and the relationship between the inner frame 2601 and the outer frame 2603 of the contour distribution, it is easy to point the L pixel from the center of the contour distribution toward the outer frame. Can be specified.

  Further, based on the pattern shape of the design data, the center position of the contour distribution frame can be converted into a reference pattern. FIG. 9 is a diagram in which the design data 900 and the contour distribution inner frame 905 and outer frame 903 are superimposed based on the shift amount by EP matching. The center position 908 between the contour distribution frames is an intersection 906 between the inner frame 905 and the straight line on the straight line drawn in the normal direction (center line end points 909 and 910 are fan-shaped) with respect to the center line 901 of the design data 900. What is necessary is just to obtain | require as a midpoint of the intersection 907 of a straight line with the outer frame 903. Finally, an approximate line and an approximate curve are obtained from the distribution of the midpoints, and a reference pattern is generated.

  In addition, as shown in FIG. 12E, the user can determine the position of the reference pattern via the input means connected to the display 216 by operating the cursor 1210. In such a case, by displaying the contour distribution image on the display 216, the user can determine the reference pattern while confirming the contour distribution image.

[Shape comparison inspection]
Next, the reference pattern generated from the contour composite image is compared with the shape of the evaluation target pattern, and a shape evaluation value of the evaluation target pattern with respect to the shape of the reference pattern is generated 106. The shape evaluation value is data used for determining whether the evaluation target pattern described later is acceptable.

  The processing flow is shown in FIG. First, a reference pattern is read 1301. Subsequently, the contour data of the evaluation target pattern is read 1302. Next, using the reference pattern, the evaluation value of the shape of the evaluation target pattern with respect to the shape of the reference pattern is obtained 1303 by a technique illustrated below, and the shape evaluation value is written in the memory 1304. Hereinafter, three methods for generating a shape evaluation value will be described using FIG. 14, but any shape evaluation value generation method that can determine the quality of an evaluation target pattern may be used, and the method is limited to this generation method. It is not a thing.

(1) Pattern Interval FIG. 14A is an image in which the reference pattern 1401 and the evaluation target pattern 1402 are superimposed based on the EP matching shift amount.

  FIG. 14B is a diagram showing an example in which the interval between the reference pattern 1401 and the evaluation target pattern 1402 having the relationship shown in FIG. 14A is measured and the interval value is used as the shape evaluation value of the evaluation target pattern. is there. The interval can be measured, for example, by measuring the point 1410 of the evaluation target pattern existing in the normal direction of the reference pattern at the point 1409 on the reference pattern and the interval 1403 thereof. For example, by performing such an interval measurement for every other pixel on the reference pattern, it is possible to quantify how much the shape of the evaluation target pattern is deformed with respect to the reference pattern. Further, the shape evaluation value can be obtained by obtaining the average or variance of the pattern intervals. For example, FIG. 17A is a diagram illustrating an example in which the distance L (n) between the reference pattern 1702 and the pattern 1703 (n: the number of pixels constituting the pattern 1703) is measured. In the interval measurement, interval values are generated in units of pixels or sub-pixels constituting the pattern 1703, and thus the interval value L (n) in the contour data region 1701 becomes enormous. Therefore, from the interval value L (n), the average of the intervals (ΣL (n) / n)) and the variance of the intervals (Σ (L (n) −ΣL (n) / n)) ^ 2 / n) are obtained. Thus, the shape evaluation value is simplified. Thereby, for example, the pattern 1704 as shown in FIG. 17B is expanded and contracted as a whole with respect to the shape of the reference pattern 1702, and therefore, the interval dispersion value tends to be small and the interval average value tends to be large. Appears in the shape evaluation value. In addition, since the pattern 1705 as shown in FIG. 17C is locally distorted with respect to the shape of the reference pattern 1702, a tendency that the interval dispersion value becomes large appears in the shape evaluation value. The shape evaluation value based on such interval measurement is written in a memory or the like as table information as shown in FIG.

(2) Shape Deformation Allowable Range FIG. 14C shows a band inside / outside determination result of the evaluation target pattern 1402 by setting a band 1404 indicating the shape deformation allowable range in the reference pattern 1401 having the relationship as shown in FIG. It is the figure which showed the example which uses as a shape evaluation value. For example, the expansion process using the morphology filter described above is applied to the reference pattern, and the pattern is expanded to a range where shape deformation can be allowed. Thereby, a band 1404 indicating a shape deformation allowable range with respect to the reference pattern 1401 can be generated. The band 1404 and the evaluation target pattern 1402 are overlapped based on the EP matching shift amount, and it is determined whether each point on the evaluation target pattern 1402 exists inside the band 1404 or outside. This result is written as a shape evaluation value in a memory or the like as table information as shown in FIG.

(3) Pattern Area FIG. 14D is a diagram showing a region 1407 surrounded by the reference pattern 1401 and a region 1408 surrounded by the evaluation target pattern 1402. The area of these regions can be obtained by counting the number of pixels surrounded by the pattern. The area ratio of the evaluation target pattern 1402 to the reference pattern 1401 is written in a memory or the like as a shape evaluation value.
[Pass / fail judgment]
Using the shape evaluation value obtained as described above, the quality of the pattern is determined 108. FIG. 16 shows a processing flow for pass / fail judgment. First, 1601 is read in the shape evaluation value of each evaluation target pattern. Next, the shape evaluation value is compared with data defining the shape information of the non-defective product (specified value), and the quality of the evaluation target pattern is determined 1602. The determination is made for all the evaluation target patterns 1605, and finally the evaluation result is written in a memory or the like. Further, in order to provide the user with easy-to-understand information on the pattern determined to be non-defective and the point on the wafer determined to be abnormal, as shown in FIG. 18, the position of the pattern on the wafer 1802 and the area where the non-defective product exists. A wafer map image 1801 is generated and displayed on the display. With such a wafer map, the user can grasp a point on the wafer where the deformation of the pattern shape increases due to process variations.

  The method for determining the shape evaluation value differs depending on the generated shape evaluation value. A determination method for the shape comparison values (1) to (3) described above will be described below.

(1) Pattern interval The reference pattern and evaluation target pattern interval values obtained by shape evaluation are compared with the non-defective pattern interval value relative to the reference pattern, and the interval between the evaluation target pattern and the non-defective pattern interval is greater than the non-defective pattern interval. Is regarded as an abnormal part of the pattern, and the evaluation target pattern in which the abnormality determination part exists is determined to be abnormal. Similarly, when the shape evaluation value is the dispersion or average of pattern intervals, pattern abnormality is determined by threshold processing.

(2) Shape Deformation Allowable Range If there is a pattern portion outside the pattern deformation allowance range in the evaluation target pattern, the evaluation target pattern is determined as an abnormal pattern. However, as described above, the contour data extracted from the image includes noise information, and an erroneous determination may occur due to the deformation of the contour due to noise. For this reason, it is possible to reduce erroneous detection of an abnormal pattern due to the influence of noise included in an image by performing threshold determination such that an abnormal pattern is determined when a pattern part that exceeds the allowable range of shape deformation of the pattern is greater than a certain value. it can.

(3) Area An evaluation target pattern in which the area ratio of the reference pattern region 1407 is N or more or M or less (M <N) is determined as an abnormal pattern.

  As described above, according to the present invention, a reference pattern used for evaluating the shape of a circuit pattern of an electronic device is generated from captured images of a plurality of circuit patterns, so that the circuit pattern conforms to the manufacturing conditions of the circuit pattern. A reference pattern in which distortion of each pattern is suppressed can be generated. By comparing the reference pattern with the evaluation target pattern, the shape evaluation of the shape evaluation pattern can be performed with high accuracy.

  In the first embodiment, the example in which the overlapping position of the contour data is specified based on the matching result between the design data and the contour data (or the image from which the contour is extracted) has been described. It is also possible to improve overlay accuracy by detecting a minute position shift due to a mismatch in the shape of the pattern to be evaluated and a minute position shift due to rotation of the wafer and finely adjusting the overlay position. FIG. 19 shows a processing flow for finely adjusting the overlay position using the contour data and generating a contour composite image.

  First, as described in the first embodiment, the contour data is read 1901, the design data is matched with the evaluation target pattern 1902, and the shift amount is calculated 1903. FIG. 20A shows an example in which two contour data are superimposed based on this shift amount. Although this is a very extreme example, the contour data 2002 is inclined with respect to the contour data 2001. In order to detect such rotation between the contour data and minute positional deviation, matching between the contour data is performed 1904 and a position correction amount is calculated 1905.

  Specifically, among the contour data of the evaluation target pattern, for example, the contour data 2001 is used as a template, and the contour data 2002 is matched with respect to the contour data 2002 of the remaining evaluation target pattern to obtain a position correction amount (rotation angle, position (Shift amount) 2003 is calculated.

  In pattern matching, for example, the size of the pattern to be evaluated is different from the template as in the generalized Hough transform described in “CG-ARTS Association, Digital Image Processing” P215. Except for abnormalities, most of the cases are expanding and contracting while maintaining the pattern shape.) Even when the pattern to be evaluated is rotated relative to the template, the matching point can be searched well. Apply the method. In addition to the generalized Hough transform, various matching methods that are robust against the rotation angle between patterns and pattern deformation have been proposed, and the contour data matching method is not limited to this. Finally, by performing matching between the contour data, the rotation of the evaluation target pattern is corrected based on the detected position correction amount 2003 of the evaluation target pattern, and the contour data is superimposed at a point corresponding to the overlapping position.

  As described above, in the present invention, before superimposing the contour data, the contour data is matched by considering the deformation and rotation between the contour data using one contour data as a template. The position correction amount can be calculated, and the contour data can be superimposed with high accuracy by correcting the position of the contour data based on the position correction amount. As a result, a reference pattern used for pattern shape shape evaluation can be generated more accurately, and the shape of the pattern can be evaluated with high accuracy by comparing the shape of the reference pattern with the pattern to be evaluated.

  Hereinafter, an example in which process window measurement is performed will be described.

  In the shape evaluation method described in the first and second embodiments, the plurality of patterns for generating the reference pattern and the evaluation target pattern are described as separate patterns. However, the reference pattern is generated from the plurality of evaluation target patterns, and the reference pattern is generated. By evaluating the shape of the evaluation target pattern using, for example, it is possible to perform a relative shape evaluation with respect to the shape variation of the evaluation target pattern such as a difference in the shape of the evaluation target pattern with respect to the average shape of the evaluation target pattern. .

  For example, a plurality of patterns with the same shape and different chips on the design data are photographed, contour data is extracted from the photographed image of each chip, and contour distribution data is generated. This contour distribution indicates variations in the pattern shape due to process variations occurring between chips. Next, the average position of the contour distribution is converted into a reference pattern. This reference pattern shows an average point of the pattern shape due to the process variation generated between the chips. By measuring the interval between the reference pattern and the evaluation target pattern as described in the first embodiment, it is evaluated how far the evaluation target pattern is deviated from the average point of pattern deformation caused by process variation. it can.

  Further, the quality of the pattern can be determined by comparing the deviation width with the allowable deformation range of the pattern shape as shown in the first embodiment. This pass / fail judgment is performed for all the evaluation target patterns.

  ADVANTAGE OF THE INVENTION According to this invention, the relative evaluation of the evaluation object pattern with respect to the shape variation of an evaluation object pattern can be performed by using the reference pattern produced | generated from the some evaluation object pattern for the shape evaluation of an evaluation object pattern.

  By using a plurality of pattern images whose pattern deformation due to process variation is relatively small, it is possible to generate a reference pattern excluding small edge deformation due to edge roughness or noise. In recent years, there are increasing cases of using contour data of a pattern manufactured on a wafer for calibration of an OPC (Optical Proximity Correction) model. In the construction of the OPC model, it is particularly important to suppress the deformation of the pattern due to noise included in the image. Thus, a reference pattern is generated from a plurality of patterns having no process variation, and the reference pattern is generated. It is effective to use for calibration.

  The contour synthesis method and the shape evaluation method based on the generated reference pattern can use those of the above-described embodiments. However, an object of the present embodiment is to generate and detect a reference pattern using a pattern whose pattern shape deformation due to process variation is small. When a reference pattern is generated using a pattern whose shape is greatly deformed, the reference pattern There is also a possibility that the shape of will greatly deform. Therefore, for example, in the process of synthesizing the contour data as shown in FIG. 27, it is determined whether the pattern is a normal pattern or an abnormal pattern 2701, and a process for applying only a normal pattern as a composition target of contour is applied. Thereby, it is possible to generate a reference pattern that excludes small edge deformation due to edge roughness or noise and large deformation caused by process fluctuations. Specifically, the interval between representative points of the pattern, the interval between the pattern and the design data, and the like are measured, and the measured value is compared with the interval value (specified value) in the case of a normal pattern.

  As described above, in the present invention, in the stage before contour synthesis described in the first to third embodiments, contour data in which the pattern is greatly distorted is detected by pattern shape determination, and the pattern is subjected to contour synthesis. By adding a process such as excluding the reference pattern, it is possible to generate a reference pattern suitable for OPC model calibration and pattern shape evaluation.

  The technique of the present invention can be widely applied to an apparatus for evaluating the shape of a circuit pattern using a captured image of the circuit pattern of an electronic device manufactured using a wafer, a reticle, or the like.

201 Wafer 202 Electron optical system 203 Electron gun 204 Electron beam 205 Condenser lens 206 Deflector 207 ExB deflector 208 Objective lens 209 Secondary electron detectors 210 and 211 Backscattered electron detectors 212 to 214 A / D converter 215 Processing control unit 216 Display 217 Stage 219 Stage controller 220 Deflection controller 221 Focus controller 223 Storage device 225 Shooting recipe generator 230 Design system 251 CPU
252 Image memory 253 LSI
301, 305 Wafer 302 Shot area 303 Chip 304 In-chip pattern 306 Relationship between patterns with different wafers 307 Relationship between patterns with different shots 308 Relationship between patterns with different chips 309 Pattern within FOV White band 402 Background 403 Pattern outline 501 Shift Amount A
502 Shift amount B
503 Shift amount C
504 Contour superposition position 701 Shift amount 702 to 704 Contour composite points 705 to 707 Composite region 708 Coordinates 900 and 1211 Design data 901 Design data center line 902 Normal lines 903 and 2603 to the center line of design data Outline frame of contour distribution 904 Contour distribution center 905, 2601 Contour distribution inner frame 906, 907, 908 Intersection points of contour distribution center and normal 909, 910 End point 1201 of inner line of design data Inner frame 1202 Outer frame 1203 Center position of contour distribution 1210 Cursor 1212 Center line 1213, 1214 Reference pattern position 1401, 1702, 2103 Reference pattern 1402 Evaluation target pattern 1403 Interval 1404 Band 1405 Abnormal part 1406 of evaluation target pattern Shape deformation allowable range 1407, 1408 Area 1409 Point 1410 on reference pattern Point 1701 Contour data area 1703, 1704, 1705 Pattern 1801 Pass / fail judgment results 2001, 2002 Contour data 2101 Contour data A
2102 Outline data B
2003 Position correction amount 2301 Design layout 2302 Focus point 2303 Addressing point 2304 Brightness, contrast point 2305 Shape evaluation point 2306 Autostigma points 2401, 402 Points 2403, 2404 Contour distribution frame 2501 Vertex coordinate data 2502 Design data center line 2602 Center pattern

Claims (16)

An apparatus for generating a reference pattern used for shape evaluation of a circuit pattern of an electronic device, comprising reference pattern generation means for generating a reference pattern used for pattern inspection from contour data of at least two circuit patterns, the reference pattern generating means, for the region between the at least two contour data, filling processing is executed, with respect to the filled areas, be fine line reference pattern generated on the basis of the thinning process A pattern generator characterized by the above. An apparatus for generating a reference pattern used for shape evaluation of a circuit pattern of an electronic device, comprising reference pattern generation means for generating a reference pattern used for pattern inspection from contour data of at least two circuit patterns, the reference pattern generating means, said generating a contour composite image superimposed at least two contour data, based on the particular site overlapping the said plurality of contour data, that the reference pattern the specific site A featured pattern generator.   The pattern generation apparatus according to claim 1 or 2, wherein pattern matching is performed between contour data and other contour data to identify a superposition position between contour data, and based on the identified superposition position. A pattern generating apparatus comprising contour distribution data generating means for generating contour distribution data.   The pattern generation apparatus according to claim 1, further comprising a unit that performs pattern matching between the design data and the contour data to identify a position where the contour data is overlapped, and overlaps the design data and the contour data. A pattern generation apparatus comprising contour distribution data generation means for generating contour distribution data based on a matching position.   3. The contour distribution according to claim 1, wherein the contour distribution data is generated using only contour data that satisfies a prescribed value of the shape of the circuit pattern by comparing a prescribed value of the shape of the circuit pattern with the shape of the contour data. A pattern generation apparatus comprising a data generation means.   3. The pattern according to claim 1, wherein the reference pattern generation unit specifies a contour distribution from the at least two or more contour data, and sets a reference pattern within the contour distribution range. Generator.   3. The pattern generation apparatus according to claim 1, wherein the reference pattern generation unit generates the reference pattern from the at least two or more contour data with reference to a shape of circuit pattern design data. .   The pattern generation apparatus according to claim 1, wherein a contour distribution data generation unit that generates contour distribution data of a circuit pattern from contour data of at least two or more circuit patterns, a reference pattern generation unit, and the reference pattern And a shape evaluation means for generating a shape evaluation value of the evaluation target pattern with respect to the reference pattern by comparing the evaluation target pattern with the evaluation target pattern.   The pattern shape evaluation apparatus according to claim 8, further comprising: a quality determination unit that performs quality determination of the evaluation target pattern using the shape evaluation value.   The pattern shape evaluation apparatus according to claim 8, wherein an interval between the reference pattern and the evaluation target pattern is measured, an average of the intervals, a dispersion of the intervals, or a pattern shape provided in the reference pattern. One of inside and outside of the evaluation target pattern with respect to a region for allowing deformation, or an area of the reference pattern and the evaluation target pattern is calculated as a shape evaluation value of the evaluation target pattern, and the shape A pattern shape evaluation apparatus comprising: a pass / fail judgment means for judging pass / fail of a pattern by comparing an evaluation value with a prescribed value for judging pass / fail of the pattern to be evaluated.   The overlapped contour data generating means, the reference pattern generating means according to claim 1 or 2, the shape evaluating means according to claim 8, the quality determining means, all means, or a part of the means, and evaluation Image capturing means for acquiring a captured image of a target pattern, a captured image of the at least two circuit patterns for generating a reference pattern, means for extracting contour data of the circuit pattern from the captured image, and Data storage means for storing the photographed image and the contour data by pattern shape evaluation, the at least two or more contour data, the reference pattern, the shape evaluation value, and the data of the pass / fail judgment result. Characteristic pattern shape evaluation apparatus.   The pattern shape evaluation apparatus according to claim 11, further comprising: means for generating a recipe for acquiring a photographed image of the circuit pattern from the design data of the circuit pattern by the image photographing means. Shape evaluation device.   12. The pattern shape evaluation apparatus according to claim 11, further comprising means for displaying the at least two contour data, the reference pattern, and data of the shape evaluation result.   The pattern shape evaluation apparatus according to claim 11, further comprising a data input unit for reflecting an instruction from a user to the reference pattern generation unit, the shape evaluation unit, and the non-defective product determination unit, and the reference pattern The pattern shape evaluation apparatus characterized in that the generation means, the shape evaluation means, and the non-defective product determination means perform reference pattern generation and shape evaluation based on instruction data from the data input means.   The said pattern generation apparatus of Claim 1 or 2 has a data input means for reflecting the instruction | indication from a user on the said reference pattern generation means, The said reference pattern generation means is an instruction | indication from the said data input means A pattern generation apparatus that generates a reference pattern based on data.   The pattern shape evaluation apparatus according to claim 13, wherein the pattern shape evaluation apparatus generates an image showing the shape evaluation value and the non-defective product determination result on a wafer map, and displays the wafer map image on the data display means. Shape evaluation device.

Images