JP4629086B2 - Image defect inspection method and image defect inspection apparatus - Google Patents

Description

  The present invention relates to an image defect inspection apparatus and a defect inspection method, for example, a defect inspection method for inspecting a fine pattern defect formed on a reticle used in semiconductor manufacturing, and an apparatus therefor.

  In general, the production of LSIs requires a great deal of cost, and thus improvement in yield is indispensable. One factor that reduces the yield is a pattern defect of a reticle used when a fine pattern is exposed and transferred by a lithography technique on a semiconductor wafer. In recent years, with the miniaturization of LSI pattern dimensions, the minimum dimensions of defects that must be detected have also been miniaturized. Therefore, it is necessary to increase the accuracy of the pattern inspection apparatus that inspects the defects of the reticle.

  Methods for inspecting the presence or absence of pattern defects are roughly classified into die-to-die comparison (Die to Die comparison) and die and database comparison (Die-to-Database comparison). The Die to Die comparison (DD comparison) is a method of detecting defects by comparing two dies on a reticle, and the Die to Database comparison (DB comparison) is a database generated from CAD data for die and LSI design. This is a method for detecting defects by comparing them. In the following description, of the two dies in the DD comparison, an image to be subjected to defect inspection is referred to as an optical image, and another reference image is referred to as a reference image. Further, the die image in the DB comparison is called an optical image, and the database is called a reference image.

  In the conventional DD comparison method or DB comparison method, when an optical image and a reference image are simply compared, pixel position shift, image expansion / contraction, sensing noise, focus shift during image acquisition, reference image generation model in DB inspection Since there is a difference between the optical image and the reference image due to the limitation of the above, if the inspection is directly compared as it is, it is buried in the difference and the defect inspection performance is deteriorated.

  In order to detect such fine defects, image correction is first performed to reduce the difference between the optical image and the reference image, and defect inspection is performed using the obtained corrected image (patent) Reference 1).

  However, this conventional technique has the following problems. That is, generally, there are many local variations in the difference between the optical image and the reference image, and it has been difficult to correct these conventionally. In addition, since the conventional technique is a method for reducing the difference as a whole image, the difference in the local area required when detecting the defect is not necessarily reduced.

JP 2006-266860 A

  The present invention was made to solve the above-mentioned problems of the prior art, and an image defect that enables highly accurate image defect determination even when a difference occurs in a part of the area, which was difficult with the prior art. An object is to provide an inspection apparatus and an image defect inspection method.

An image defect inspection method according to the present invention is a pattern defect inspection method for comparing and inspecting an image signal photoelectrically converted from a certain inspection object and a reference image, and obtaining a first corrected image by correcting the reference image. 1 image correcting step, a first defect detecting step for detecting a defect by using the image signal and the first corrected image to obtain a defect coordinate, the image signal, the first corrected image, and the defect coordinate. A weight image calculation step for obtaining a weight image from the second correction step, a second image correction step for re-correcting the first correction image using the first correction image and the weight image to obtain a second correction image, A second defect detection step of detecting a defect using the image signal and the second corrected image is provided.

An image defect inspection apparatus according to the present invention is a pattern defect inspection apparatus that compares and inspects an image signal photoelectrically converted from a certain inspection object and a reference image, and obtains a first corrected image by correcting the reference image. 1 image correcting means, first defect detecting means for detecting a defect by using the image signal and the first corrected image to obtain a defect coordinate, the image signal, the first corrected image, and the defect coordinate. Weight image calculation means for obtaining a weight image from the image, second image correction means for re-correcting the first correction image using the weight image to obtain a second correction image, the image signal and the second image A second defect detecting means for detecting a defect using the corrected image is provided.

  According to the image defect inspection apparatus and the image defect inspection method according to the present invention, when there is a difference in a line of a part of the region or when only a part of the region does not match the shape, the difference is also corrected. Since the defect inspection is performed, it is possible to correct the image according to the local pattern mode, and the defect inspection accuracy is improved.

[First embodiment: Image defect inspection apparatus]
Hereinafter, the image defect inspection apparatus 1 that can be used in the defect detection method of the present invention will be described in detail. FIG. 1 is a schematic diagram illustrating a configuration of an image defect inspection apparatus 1 according to the first embodiment. The apparatus configuration will be described in detail with reference to the schematic diagram of FIG. The image defect inspection apparatus 1 according to the first embodiment is an apparatus that inspects whether a pattern formed on an inspection object such as a reticle 101 is formed in a predetermined shape, that is, whether an image has a defect. It is. The image defect inspection apparatus 1 includes an optical image acquisition unit 110 and a data processing unit 120.

  As shown in FIG. 1, the optical image acquisition unit 110 reads a pattern drawn on an inspection object on which a pattern such as a reticle and a mask is formed, and obtains an optical image. The optical image 100 is an image signal that is photoelectrically converted after a pattern drawn on the inspection object is read by the optical image acquisition unit 110.

(Optical image acquisition unit)
The hardware configuration of the optical image acquisition unit 110 will be described with reference to FIG.
The optical image acquisition unit 110 acquires an optical image of the reticle 101. A reticle 101 to be inspected is placed on an XYθ table 102. The XYθ table 102 is controlled to move in the horizontal direction and the rotation direction by the motors 150, 151, and 152 of the respective axes of the XYθ table 102 in response to a command from the table control unit 114. Light from the light source 103 is applied to the pattern formed on the reticle 101. The light that has passed through the reticle 101 forms an optical image on the photodiode array 105 via the magnifying optical system 104. The image captured by the photodiode array 105 is processed by the sensor circuit 106 to become optical image data. On the other hand, the laser length measurement system 132 transmits the measured position signal of the XYθ table 102 to the position measurement unit 107.

  The XYθ table 102 is driven by the table control unit 114 under the control of the central processing unit 131. The movement position of the XYθ table 102 is measured by the laser length measurement system 132 and sent to the position measurement unit 107. The reticle 101 on the XYθ table 102 is transported from the autoloader 130 under the control of the autoloader control unit 113. The measurement pattern data of the inspection stripe image output from the sensor circuit 106 is sent to the defect coordinate calculation unit 108 together with the data indicating the position of the reticle 101 on the XYθ table 102 output from the position measurement unit 107. The optical image data and the corrected reference image to be compared are cut out into an area having an appropriate pixel size. For example, it is cut out into an area of 512 × 512 pixels.

  The optical image acquisition unit 110 further includes an autoloader 130 as necessary, so that image inspection can be performed automatically continuously. The optical image obtained by the optical image acquisition unit 110 is image data for each pixel having, for example, an 8-bit gradation.

(Data processing part)
1, the data processing unit 120 includes a reference image generation unit 122, an image correction processing unit 124, a central processing unit 131, an autoloader control unit 113, a table control unit 114, a database 140, and a database creation unit. 141, a defect coordinate calculation unit 108, a weight image inspection unit 109, and a position measurement unit 107. In addition, peripheral devices such as a magnetic disk device, a magnetic tape device, an FD, a CRT, a pattern monitor, and a printer that are installed in a normal computer are provided. These peripheral equipments are equipped with a normal computer and will not be described in detail. All these components are connected by a bus 153 so that they can communicate with each other. The data processing unit 120 can be configured by an electronic circuit, a program, or a combination thereof.

  An optical image acquisition procedure will be described. FIG. 2 is a schematic diagram illustrating acquisition of an inspection stripe image of the image defect inspection apparatus 1 according to the first embodiment. The inspection area of the reticle 101 is virtually divided into a plurality of strip-shaped inspection stripes having a constant scanning width, and each of the divided inspection stripes is continuously scanned. Therefore, the XYθ table 102 moves under the control of the table control unit 114. According to the movement, an optical image of each inspection stripe is acquired by the photodiode array 105. The photodiode array 105 continuously acquires images having a certain scanning width. After acquiring the image of the first inspection stripe A, the photodiode array 105 is in the reverse direction to the acquisition of the image, but the image of the second inspection stripe B is continuously acquired with the scanning width W in the same manner. To get to. The image of the third inspection stripe C is acquired in the direction opposite to the direction in which the image of the second inspection stripe B is acquired, that is, the direction in which the image of the first inspection stripe A is acquired. In this way, it is possible to shorten a useless processing time by continuously acquiring images. Here, for example, the scanning width is 2048 pixels.

  The pattern image formed on the photodiode array 105 by the above optical image acquisition procedure is photoelectrically converted by the photodiode array 105 and further A / D (analog-digital) converted by the sensor circuit 106. The light source 103, the magnifying optical system 104, the photodiode array 105, and the sensor circuit 106 constitute a high-magnification image inspection reading system.

  Hereinafter, the processes leading to image correction and defect determination will be described in more detail. FIG. 3 is a block diagram showing a main part configuration of the image defect inspection apparatus 1 according to the first embodiment. As shown in FIG. 3, the data processing unit 120 includes a reference image generation unit 122, an image correction processing unit 124, a defect coordinate calculation unit 108, and a weight image inspection unit 109 as main components. ing. In FIG. 3, a reference image generation unit 122 reads design data such as CAD data 121 from a magnetic tape device through a central processing unit 131 and converts it into a reference image 123 of image data.

  Further, the image correction processing unit 124 receives the image data of the reference image 123, and corrects the corrected reference image that has been subjected to correction processing such as alignment correction to resemble the optical image 100 by rounding or slightly blurring the figure. 125 is generated. The corrected reference image 125 generated by the image correction processing unit 124 is output to the defect coordinate calculation unit 108. Thereafter, the defect coordinate calculation unit 108 compares the optical image 100 with the corrected reference image 125 and calculates the position coordinates of the pattern defect. Further, the defect coordinate information 126 output from the defect coordinate calculation unit 108 is input to the weight image inspection unit 109, and finally a defect determination result 129 is generated.

  The reference image 123 is an image created so as to be similar to the optical image 100 from the CAD data 121 of the inspection object. The corrected reference image 125 is a corrected image obtained by correcting the reference image 123 so as to remove the shift, distortion, or shift and distortion between the optical image 100 and the reference image. The CAD data 121 is data such as CAD for design that is a source of drawing an image on the inspection object. Hereinafter, the reticle 101 will be described as an inspection object. However, the inspection object may be any object as long as a pattern is formed, and may be a mask, a wafer, or the like. Further, in the case of DB comparison, an optical image and a reference image are compared. However, in the case of performing the above-described DD comparison, as shown in FIG. 3, an image of a reference pattern (optical image A) and an inspection object are compared. A pattern image (optical image B) may be captured and input to the data processing unit 120.

  Hereinafter, the image correction for the reference image will be described in more detail. FIG. 4 is a block diagram illustrating a configuration of the image correction processing unit 124 of the image defect inspection apparatus 1 according to the first embodiment. For example, as illustrated in FIG. 4, the image correction processing unit 124 includes a correction model parameter identification unit 200 and an image correction unit 201. The correction model parameter identification unit 200 uses the optical image 100 obtained by the optical image acquisition unit 110, the feature data 202 indicating the pattern characteristics of the reticle 101, and the reference image 123 created by the reference image generation unit 122. Thus, the correction model parameter 204 is created. The image correction unit 201 operates the correction model parameter 204 created by the correction model parameter identification unit 200 on the reference image 123 and performs arithmetic processing to create a corrected reference image 125.

  Next, the defect coordinate calculation unit 108 inputs the corrected reference image 125 and outputs defect coordinate information 126. The defect coordinate calculation unit 108 obtains a pixel value difference from the input corrected reference image 125 and the optical image 100, and determines that the defect is a defect when the pixel value difference exceeds a predetermined threshold. And the defect coordinate information 126 linked | related with the coordinate value of the area | region determined to be a defect is output. The image correction processing unit 124 can be configured by an electronic circuit, a program, or a combination thereof.

  Note that the feature data 202 used in FIG. 4 designates a specific pattern of the image of the reticle 101 and indicates a feature portion of the image of the reticle 101. The feature data 202 is created, for example, at the stage of designing the image of the reticle 101, corresponds to the pattern position of the reticle 101, and is pattern identification data that designates a pattern. The feature data 202 is represented by an image corresponding to the image of the reticle 101, for example, and includes pixel value data. The feature data 202 can give a weight to the image pattern of the reticle 101. The feature data 202 indicates a pattern, auxiliary pattern, dummy pattern, or the like that requires drawing accuracy. Note that the pattern used in this embodiment may have any shape, for example, an independent pattern, a combined pattern of independent patterns, a pattern of a part (part) of an independent pattern, or The pattern of the part (part) of the combined pattern may be used.

  Hereinafter, the process of obtaining the defect determination result 129 from the defect coordinate information 126 will be described in detail. FIG. 5 is a block diagram illustrating an internal configuration of the weight image inspection unit 109 of the image defect inspection apparatus 1 according to the first embodiment. The weight image inspection unit 109 generates a defect determination result 129 from the input defect coordinate information 126, and includes a weight image calculation unit 133, a weight image correction unit 134, and a defect determination unit 135. .

The weighted image calculation unit 133 creates a weighted image 127 from the defect coordinates (px, py) included in the defect coordinate information 126. That is, the SSD value of 15 × 15 pixels centered on the defect coordinates (px, py) and 15 × 15 pixels centered on (x, y) is obtained and set as SSD (x, y). Then, A · exp ( −SSD (x, y) / T) is set as the gradation value of the coordinates (x, y) of the weighted image. However, if the city block distance between (x, y) and (px, py) is equal to or smaller than a predetermined value, the gradation value of the coordinate (x, y) of the weighted image is set to 0. In accordance with the above rules, the weighted image 127

  The weighted image correction unit 134 creates a weighted reference image 128 from the weighted image 127 and the corrected reference image 125. In the first embodiment, the weighted reference image 128 is obtained by correcting the corrected reference image 125 using the conventional technique (here, the technique disclosed in JP-A-2006-266860 is used).

  Next, the defect determination unit 135 inputs the weighted reference image 128 and the optical image 100 created as described above, and performs defect determination. A defect determination result 129 is obtained by performing the defect determination. The defect determination unit 135 receives the optical image 100 and the weighted reference image 128, calculates the difference between the pixel values of the two images, and determines that the pixel is defective if the absolute value of the difference exceeds a predetermined value. It has a function. The finally obtained defect determination result 129 becomes the inspection result of the image defect inspection apparatus according to the first embodiment.

  As described above, according to the first embodiment, the defect coordinates are calculated from the optical image and the corrected reference image, and the weighted reference image is generated from the weighted image generated from the defect coordinate value and the corrected reference image. By performing defect determination again, there is provided an image inspection apparatus capable of performing high-accuracy image defect determination even when a difference occurs in a part of a region that has been difficult with the prior art.

[Second Embodiment: Image Defect Inspection Method]
Details of the image defect inspection method according to the second embodiment will be described below. The second embodiment is different from the first embodiment in an image defect inspection method for inspecting an optical image acquired by a predetermined optical image acquisition unit. A processing procedure for inspecting an optical image acquired by the optical image acquisition unit 110 of the image defect inspection apparatus according to the first embodiment will be described in detail. The same reference numerals denote the same components, and detailed description thereof is omitted. FIG. 6 is a flowchart illustrating a processing procedure of an image defect inspection method according to the second embodiment.

  In the image defect inspection method according to the second embodiment, as shown in FIG. 6, the optical image 100 and the reference image 123 are input, and the first image correction step (S41), the first defect detection step ( The image defect is inspected by processing in S42), a weight image calculation step (S43), a second image correction step (S44), and a second defect detection step (S45). Hereinafter, each step of inspecting the optical image with respect to the reference image will be described in detail. Hereinafter, the reference image is described in correspondence with the inspection standard pattern image, and the optical image is described in correspondence with the inspection pattern image.

(S41: First image correction step)
This step is a process for recovering the deterioration of the inspection pattern image or the inspection reference pattern image such as the alignment of the inspection reference pattern image or the expansion / contraction of the image, waviness, sensing noise, defocus, etc. The method disclosed in Japanese Patent No. 266860 can be employed. The method will be briefly described below. As described below, this method includes a two-dimensional linear prediction model setting stage, simultaneous equation solving stage, and model image generation stage.

(1) Setting of two-dimensional linear prediction model (simultaneous equation generation step)
First, a method of setting a two-dimensional linear prediction model (two-dimensional input / output linear prediction model) by regarding the inspection reference pattern image as two-dimensional input data and the inspection pattern image as two-dimensional output data will be described. Here, a 5 × 5 two-dimensional linear prediction model using a 5 × 5 pixel region is taken as an example. Table 1 shows the suffixes (corresponding to the positions of 5 × 5 pixels) used in this model. FIG. 7 is a diagram illustrating comparison of pattern images in the image defect inspection method according to the second embodiment. In FIG. 7, the left figure is an inspection reference pattern image, and the right figure is an inspection pattern image.

  The two-dimensional input data and the two-dimensional output data are u (i, j) and y (i, j), respectively. The suffixes of the pixel of interest are i and j, and the suffixes of a total of 25 pixels around 2 rows and 2 columns surrounding this pixel are set as shown in Table 1. A relational expression such as Expression (1) is set for a certain set of 5 × 5 region pixel data. Coefficients b00 to b44 of each input data u (i, j) in Expression (1) are model parameters to be identified.

  The expression (1) means that one pixel data yk = y (i, j) of the pattern image to be inspected is a linear combination of 5 × 5 pixel data surrounding one pixel of the corresponding inspection reference pattern image. It can be expressed as

(2) Simultaneous equation solving step (identification of model parameters)
When Expression (1) is expressed as a vector, Expression (2) is obtained. Here, the unknown parameter vector α is α = [b00, b01,..., B44] T, and the data vector xk is xk = [u (i−2, j−2), u (i− 2, j−1),..., U (i + 2, j + 2)] T.

  If the coordinates i and j of the inspection reference pattern image and the pattern image to be inspected are scanned and 25 sets of data are combined, the model parameters can be identified. Actually, from a statistical point of view, n (> 25) sets of data are prepared as in Equation (3), and a 25-dimensional simultaneous equation is solved based on the following least-squares method to identify α. . Here, A = [x1, x2,..., Xn] T, y = [y1, y2,..., Yn] T, xkTα = yk, and k = 1, 2,. .., n. As a method of solving these equations, there is a maximum likelihood estimation method in addition to the least square method, and any method may be used.

For example, if the inspection reference pattern image and the pattern image to be inspected are each 512 × 512 pixels, the periphery of the image is reduced by 2 pixels by scanning the 5 × 5 model, so the number of equations is expressed by Equation (4) Thus, 258064 sets of data are obtained. Thereby, it is possible to ensure a sufficient number from a statistical viewpoint.

(3) Generation of Model Image By substituting the identified model parameter α and the input / output image data used for identification into Equation (1), and performing a simulation calculation that scans the coordinates i and j of the pixel, an estimated model Generate an image. This estimated model image is a target first corrected image. In this estimated model image, a pixel position shift of less than one pixel, expansion / contraction / swell noise, resizing processing, and reduction of sensing noise are realized by fitting based on the least square method. Here, the data used for the simulation naturally includes defective pixels, but since it is a very small number compared to the total number of data used for identification, it is not fitted by the least square method, and the estimated model image It does not appear. Further, since the surrounding S / N ratio is improved, there is an effect that the defective pixel is emphasized.

(S42: First defect detection step)
In this step, the conventional technique is used as it is, the difference between the pattern image to be inspected and the first correction image is obtained, the area where the difference exceeds a predetermined threshold is determined as a defect, and the position coordinates of the defect are obtained. The obtained position coordinates of a series of defects are called coordinates (px, py) by the first defect detection.

(S43: Weighted image calculation step)
For the first corrected image, the following processing is performed on all pixels (x, y).
That is, SSD values of 15 × 15 pixels centered on (px, py) and 15 × 15 pixels centered on (x, y) are obtained and set as SSD (x, y). Then, A · exp (−ssd (x, y) / T) is set as the gradation value of the coordinates (x, y) of the weighted image. However, if the city block distance between (x, y) and (px, py) is less than or equal to d, the gradation value of the coordinate (x, y) of the weighted image is set to 0. Here, A, T, and d are constants. For example, d = 15, A = 100, and T = 1000000.

(S44: Second image correction step)
Using the conventional technique (here, the technique disclosed in JP-A-2006-266860 is used), the first corrected image is corrected to obtain a second corrected image. However, the criteria for the error to be minimized are p (x, y) for the gradation value of the pattern image to be inspected, q (x, y) for the second corrected image, and W (x, y) for the weighted image. y) is taken as shown in equation (5).

  That is, when w = [W (1), W (2),..., W (n)] ^ t in equation (3) of Japanese Patent Laid-Open No. 2006-266860, wAα = wy instead of Aα = y. Then, the parameter α of the mathematical model may be obtained.

(S45: Second defect detection step)
In this step, the conventional technique is used as it is, and the second image defect detection is executed using the pattern image to be inspected and the second corrected image. As a conventional technique, for example, a difference between a pattern image to be inspected and a second corrected image is taken, and an area where the difference exceeds a predetermined value can be determined as a defect. The position coordinates of the defect thus obtained are defined as (rx, ry).
As described above, according to the second embodiment, the defect coordinates are calculated from the optical image and the corrected reference image, and the weighted reference image is generated from the weighted image generated from the defect coordinate value and the corrected reference image. By performing defect determination again, there is provided an image inspection method that enables highly accurate image defect determination even when a difference occurs in a part of a region that has been difficult with the prior art.

  Hereinafter, the present invention will be described based on examples.

  Hereinafter, an example in which a defect is detected using an inspection pattern image obtained by the optical image acquisition unit 110 according to the first embodiment and an inspection reference pattern image from CAD data will be described. FIG. 8 is a diagram for comparing images when the image defect inspection method according to the prior art is executed. FIG. 8 shows a comparison between the pattern image to be inspected and the first corrected image obtained by fitting only the alignment. 8A is a pattern image to be inspected, and FIG. 8B is a first corrected image that has been subjected to the first image correction step (S41) from the reference image 123. In FIG. FIG. 8C shows a difference image between the pattern image to be inspected and the first corrected image. An area having a large contrast between light and dark is recognized in the central portion of FIG. 8C, but this area is an area where the difference between the pattern image to be inspected and the first corrected image is large, and may be a defect.

  Using these pattern images to be inspected and the first corrected image, images obtained by performing the image defect inspection method according to the second embodiment were compared by the method described in Japanese Patent Application Laid-Open No. 2006-266860. . The result is shown in FIG. As can be seen from the difference image in FIG. 8C, as described above, the difference slightly decreases as compared with the image simply subjected to the fitting process, but it can be seen that the difference is still generated in the corner portion. When this conventional technique is applied to the data of the embodiment, there arises a problem that a difference in corner portion is easily detected as a defect.

  By the way, this corner difference is not really a defect. That is, the same difference occurs in all upper right corners due to the limit of the generation model of the first corrected image. Such a pattern-dependent difference is difficult to correct by the conventional technique, and there is a high possibility that this noise is mistaken as a defect.

  As shown in FIG. 8, the image obtained by the conventional method is subjected to the weight image calculation step and the second image correction step of the image defect inspection method according to the second embodiment, and the second defect Detection was performed. In the above formula (5), A, T, and d are constants, and d = 15, A = 100, and T = 1000000.

  FIG. 9 is a diagram for comparing images when the image defect inspection method according to the second embodiment is executed. The result is shown in FIG. FIG. 9A shows an inspection pattern image, and FIG. 9B shows a second corrected image. As shown in FIG. 9C, the difference image between the pattern image to be inspected and the second corrected image eliminates the image difference at the corner portion that appeared in FIG. 8, and this image difference is not a defect. To be judged. Here, when the second corrected image is obtained for the embodiment, it can be seen that the difference in corners remaining in the first corrected image is eliminated.

  As described above, when the image defect inspection method according to the second embodiment is applied to the data of the example, the difference in the corner portion is eliminated, so that it becomes difficult to detect the defect as a defect. Thus, the possibility of outputting noise as a defect could be reduced. By such inspection methods, it is possible to inspect defects after correcting for differences in "when there are differences in the lines of some areas" and "when some areas do not match the shape", which was difficult with the prior art. It becomes.

1 is a schematic diagram showing a configuration of an image defect inspection apparatus 1 according to a first embodiment. FIG. 3 is a schematic diagram showing acquisition of an inspection stripe image of the image defect inspection apparatus 1 according to the first embodiment. The block diagram which shows the principal part structure of the image defect inspection apparatus 1 concerning 1st Embodiment. FIG. 3 is a block diagram showing a configuration of an image correction processing unit 124 of the image defect inspection apparatus 1 according to the first embodiment. The block diagram which shows the structure inside the weight image inspection part 109 of the image defect inspection apparatus 1 concerning 1st Embodiment. 9 is a flowchart showing a processing procedure of an image defect inspection method according to a second embodiment. The figure which shows the comparison of the pattern image of the image defect inspection method concerning 2nd Embodiment. The figure which compares the image at the time of performing the image defect inspection method by a prior art. The figure which compares the image at the time of performing the image defect inspection method concerning 2nd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Image defect inspection apparatus 100 ... Optical image 101 ... Reticle 102 ... XY (theta) table 103 ... Light source 104 ... Enlargement optical system 105 ... Photodiode array 106 ... Sensor circuit 107 ... Position measuring part 108 ... Defect coordinate calculation part 109 ... Weight image inspection Unit 110 ... Optical image acquisition unit 120 ... Data processing unit 121 ... CAD data 122 ... Reference image generation unit 123 ... Reference image 124 ... Image correction processing unit 125 ... Corrected reference image 126 ... Defect coordinate information 127 ... Weighted image 128 ... Weighting Reference image 129 ... defect determination result 131 ... central processing unit 132 ... laser measurement system 133 ... weight image calculation unit 134 ... weight image correction unit 135 ... defect determination unit 140 ... database 141 ... database creation unit 150 ... X motor 151 ... Y motor 152 ... θ motor 153 ... bus 200 ... complement Positive model parameter identification unit 201 ... image correction unit 202 ... feature data 204 ... correction model parameter

Claims (2)

In a pattern defect inspection method for comparing and inspecting an image signal photoelectrically converted from a certain inspection object and a reference image,
A first image correction step of correcting the reference image to obtain a first corrected image;
A first defect detection step for detecting a defect using the image signal and the first corrected image to obtain a defect coordinate;
A weight image calculation step of obtaining a weight image from the image signal, the first correction image, and the defect coordinates;
A second image correction step of re-correcting the first correction image using the first correction image and the weight image to obtain a second correction image;
An image defect inspection method comprising a second defect detection step of detecting a defect using the image signal and the second corrected image. In a defect inspection apparatus for a pattern for comparing and inspecting an image signal photoelectrically converted from a certain inspection object and a reference image,
First image correcting means for correcting the reference image to obtain a first corrected image;
First defect detection means for detecting a defect using the image signal and the first corrected image to obtain a defect coordinate;
Weight image calculation means for obtaining a weight image from the image signal, the first correction image, and the defect coordinates;
Second image correction means for re-correcting the first correction image using the first correction image and the weight image to obtain a second correction image;
An image defect inspection apparatus comprising: second defect detection means for detecting a defect using the image signal and the second corrected image.