Image processing apparatus, optical apparatus, image processing method, and method for using the optical apparatus

Abstract

To provide an image processing device for taking appropriate measures to a difference in sensitivity caused by a difference in an imaging direction of a pattern, an optical device, an image processing method, and a method for using the optical device.SOLUTION: In an inspection device 1 having an imaging optical system 20, an image processing device 40 comprises: a setting part for setting a first parameter for a first pattern on the basis of a first sample image area being an area including the first pattern extending in an X axis direction in a reference image plane, and setting a second parameter in a second pattern on the basis of a second sample image area being an area including the second pattern extending in a Y axis direction orthogonal to the X axis direction; an acquisition part for acquiring a sample image being a picked-up image obtained by imaging a sample having a third pattern extending in the X axis direction and a fourth pattern extending in the Y axis direction by the prescribed imaging optical system 20; and a processing part for performing prescribed information processing based on the first parameter and the second parameter to the sample image.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This disclosure relates to an image processing apparatus, an optical apparatus, an image processing method, and a method for using an optical apparatus. [Background technology]

[0002] Patent Document 1 describes a technique for inspecting defects in a photomask by irradiating it with EUV (Extreme Ultra Violet) light. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Patent No. 6249513 [Non-patent literature]

[0004] [Non-Patent Document 1] Masayuki Hayashi, "CVML Expert Guide," [online], [Accessed November 6, 2024], Internet<URL:https: / / cvml-expertguide.net / 2021 / 07 / 12 / spatial-filtering / > [Overview of the project] [Problems that the invention aims to solve]

[0005] Depending on the design and settings of the optical device, the appearance of patterns and defects in inspection images may differ depending on the angle between the direction in which the photomask pattern extends and the direction of incidence of the illumination light, or depending on the numerical aperture (NA) and magnification. For example, even with a pattern surface of the same design, the appearance of patterns and defects in inspection images may differ depending on whether the light is incident from the direction in which the pattern extends or from a direction perpendicular to the direction in which the pattern extends, or depending on the numerical aperture (NA) and magnification. Furthermore, even with a pattern surface of the same design and the same defect, the appearance of patterns and defects may differ depending on the imaging direction, depending on the differences in NA and magnification in the direction corresponding to each direction on the captured image. Therefore, inspection may be required that takes into account differences in sensitivity and appearance caused by the imaging direction of the pattern and the sample.

[0006] This disclosure was made in view of the above problems and provides an image processing apparatus, an optical apparatus, an image processing method, and a method for using the optical apparatus that can appropriately address differences in sensitivity and image quality caused by the imaging direction of the pattern and the imaging direction of the sample. [Means for solving the problem]

[0007] An image processing apparatus according to one aspect of this embodiment includes: a setting unit that sets a first parameter for a first pattern based on a first sample image region which is a region including a first pattern extending in a first direction in a reference image plane, and sets a second parameter for a second pattern based on a second sample image region which is a region including a second pattern extending in a second direction orthogonal to the first direction in the reference image plane; an acquisition unit that acquires a sample image which is an image of a sample having a third pattern extending in a first direction and a fourth pattern extending in a second direction in the captured image when the plane of the captured image is the reference image plane, captured by a predetermined imaging optical system; and a processing unit that performs predetermined information processing on the sample image based on the first parameter and the second parameter.

[0008] In the image processing apparatus described above, the setting unit may set the first parameter to include a threshold used for determining whether or not there are defects in the third pattern included in the sample image, and the second parameter to include a threshold used for determining whether or not there are defects in the fourth pattern included in the sample image, and the processing unit may perform the predetermined information processing, including determining whether or not there are defects in the third and fourth patterns included in the sample image.

[0009] In the image processing apparatus described above, the sample image includes both the third pattern and the fourth pattern, and the processing unit may perform the predetermined information processing, including determining whether or not there are defects in the third pattern and the fourth pattern included in the sample image.

[0010] In the image processing apparatus described above, the setting unit may set the threshold for multiple types of physical quantities relating to the information of the sample image.

[0011] In the image processing apparatus described above, the setting unit may set at least one of the type of physical quantity for which the threshold value is determined and the number of physical quantities as the first parameter to be different from the physical quantity for which the threshold value is determined as the second parameter.

[0012] In the image processing apparatus described above, at least one of the first sample image region and the second sample image region may be a region included in an image generated based on the design data of the sample or the design data of the pattern formed on the sample.

[0013] In the image processing apparatus described above, the first sample image region may also be a region included in the image captured by the predetermined imaging optical system of an object having the first pattern.

[0014] In the image processing apparatus described above, the first sample image region and the second sample image region may also be regions included in the image captured by the predetermined imaging optical system of an object having the first pattern and the second pattern.

[0015] In the above-described image processing apparatus, the first sample image region and the second sample image region are regions included in the captured image of the object having the first pattern by the predetermined imaging optical system. The orientation of the object with respect to the predetermined imaging optical system when capturing the captured image including the first sample image region and the orientation of the object with respect to the predetermined imaging optical system when capturing the captured image including the second sample image region may differ by 90° in the principal plane of the object.

[0016] In the above-described image processing apparatus, the predetermined imaging optical system captures an image by oblique incident illumination on the object, and the first sample image region may be included in the captured image when the angle formed by the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the object and the first direction is a predetermined angle.

[0017] In the above-described image processing apparatus, the predetermined imaging optical system captures an image by oblique incident illumination on the object, and the first sample image region and the second sample image region may be included in the captured image when the angle formed by the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the object and the first direction is a predetermined angle.

[0018] In the above-described image processing apparatus, the predetermined imaging optical system captures an image by oblique incident illumination on the sample, and the sample image may be a captured image by the predetermined imaging optical system when the angle formed by the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the sample and the first direction is a predetermined angle.

[0019] In the above-described image processing apparatus, the sample may include an anamorphic mask.

[0020] In the above-described image processing apparatus, the acquisition unit may acquire sample images in which the number of pixels binned in the first direction and the second direction is different.

[0021] In the above-described image processing apparatus, the predetermined imaging optical system may have different NAs in the first direction and the second direction.

[0022] In the above-described image processing apparatus, the setting unit sets the first parameter to include a correction value of the luminance of pixels belonging to the third pattern included in the sample image, and sets the second parameter to include a correction value of the luminance of pixels belonging to the fourth pattern included in the sample image, and the processing unit may perform the predetermined information processing including image correction of the sample image.

[0023] An optical apparatus according to one aspect of the present embodiment includes an illumination optical system that illuminates the sample, the predetermined imaging optical system that images the illuminated sample, and the image processing apparatus described above.

[0024] An image processing method according to one aspect of the present embodiment includes a step of setting a first parameter for the first pattern based on a first sample image region that is a region including a first pattern extending in a first direction in a reference image plane, a step of setting a second parameter for the second pattern based on a second sample image region that is a region including a second pattern extending in a second direction orthogonal to the first direction in the reference image plane, a step of obtaining a sample image that is the captured image of a sample having a third pattern extending in the first direction and a fourth pattern extending in the second direction in the captured image when the plane of the captured image is the reference image plane by a predetermined imaging optical system, and a step of performing predetermined information processing on the sample image based on the first parameter and the second parameter.

[0025] A method of using an optical apparatus according to one aspect of the present embodiment includes a step of illuminating the sample, a step of imaging the illuminated sample, and the image processing method described above.

Advantages of the Invention

[0026] According to this disclosure, it is possible to provide an image processing device, an optical device, an image processing method, and a method of using the optical device that can appropriately address differences in sensitivity caused by the imaging direction of the pattern. [Brief explanation of the drawing]

[0027] [Figure 1] This is a diagram illustrating an inspection device according to Embodiment 1. [Figure 2] This figure illustrates the mask pattern of a sample according to Embodiment 1. [Figure 3] This figure illustrates an image captured such that the mask pattern of the sample according to Embodiment 1 extends in a first direction. [Figure 4] This figure illustrates an image captured such that the mask pattern of the sample according to Embodiment 1 extends in a second direction. [Figure 5] This is a block diagram illustrating an image processing device according to Embodiment 1. [Figure 6] This figure illustrates a sample mask pattern for a sample image according to Embodiment 1. [Figure 7] This figure illustrates a sample mask pattern for a sample image according to Embodiment 1. [Figure 8] This figure illustrates a sample mask pattern for a sample image according to Embodiment 1. [Figure 9] This figure illustrates a first sample image acquired by the acquisition unit in the image processing apparatus according to Embodiment 1. [Figure 10] This figure illustrates a second sample image acquired by the acquisition unit in the image processing apparatus according to Embodiment 1. [Figure 11] This figure illustrates an image captured by the acquisition unit in the image processing apparatus according to Embodiment 1. [Figure 12] In the image processing apparatus according to Embodiment 1, this graph illustrates the relationship between the difference signal and the probability of occurrence between predetermined pixels in an image captured by the acquisition unit, with the horizontal axis representing the difference signal and the vertical axis representing the probability of occurrence. [Figure 13]In the image processing apparatus according to Embodiment 1, this graph illustrates the relationship between the difference signal and the probability of occurrence between predetermined pixels in an image captured by the acquisition unit, with the horizontal axis representing the difference signal and the vertical axis representing the probability of occurrence. [Figure 14] This flowchart illustrates an image processing method using the image processing apparatus according to Embodiment 1. [Figure 15] This flowchart illustrates an inspection method using the inspection device according to Embodiment 1. [Figure 16] This figure illustrates an image processing apparatus according to Embodiment 2, including a first pattern extending in a first direction, a luminance distribution along a second direction in the first pattern, and a one-dimensional kernel applied to the luminance distribution. [Figure 17] This figure illustrates an image processing apparatus according to Embodiment 2, including a second pattern extending in a second direction, a luminance distribution along the first direction in the second pattern, and a one-dimensional kernel applied to the luminance distribution. [Figure 18] This figure illustrates the luminance distribution of a reference shape along a direction perpendicular to the direction in which the pattern extends, in the image processing apparatus according to Embodiment 2. [Figure 19] This is a schematic diagram illustrating the information processing performed by the processing unit on a sample image or an image based on a sample image in the image processing apparatus according to Embodiment 2. [Figure 20] This figure illustrates a difference image in an image processing apparatus according to a modified example 1 of Embodiment 2. [Figure 21] This flowchart illustrates an image processing method using the image processing apparatus according to Embodiment 2. [Modes for carrying out the invention]

[0028] Embodiments of the present disclosure will be described below with reference to the drawings. The following description illustrates preferred embodiments of the present disclosure and does not limit the scope of the present disclosure to the following embodiments. In the following description, the same reference numerals indicate substantially the same thing.

[0029] (Embodiment 1) This document describes an image processing apparatus, optical apparatus, image processing method, and method of using the optical apparatus according to Embodiment 1. First, in <Inspection Apparatus>, an inspection apparatus will be described as an example of an optical apparatus. Next, in <Pattern>, a pattern of a sample that may be the object of inspection will be described. Next, in <Image Processing Apparatus>, an image processing apparatus provided in the optical apparatus will be described. Next, in <Image Processing Method> and <Inspection Method>, an inspection method will be described as an example of an image processing method using the image processing apparatus and a method of using the optical apparatus.

[0030] The image processing apparatus and image processing method described in this disclosure may be used for inspection equipment, as described in the following embodiments, but is not limited to that. For example, the image processing apparatus and image processing method described in this disclosure may be used as a device (review device) that displays an image (captured image) obtained as a result of illuminating a sample on a display or the like.

[0031] <Inspection equipment> Figure 1 is a diagram illustrating an inspection apparatus 1 according to Embodiment 1. The inspection apparatus 1 comprises an illumination optical system 10, an imaging optical system 20, and an image processing device 40. The illumination optical system 10 illuminates the sample 50 using illumination light L11. The illumination optical system 10 may also illuminate objects used for acquiring the first sample image region and the second sample image region, described later, using illumination light L11, similar to the sample 50. The illumination optical system 10 includes, for example, a light source 11, an ellipsoidal mirror 12, an ellipsoidal mirror 13, and a recessed mirror 14. The imaging optical system 20 captures an image of the sample 50 illuminated by the illumination light L11. The image of the sample 50 is sometimes referred to as the sample image. The imaging optical system 20 includes, for example, a perforated concave mirror 21, a convex mirror 22, and a detector 23. The perforated concave mirror 21 and the convex mirror 22 constitute a Schwarzschild magnifying optical system. The illumination optical system 10 and the imaging optical system 20 may further include optical components other than those described above, or any of the above optical components may be omitted.

[0032] Here, for the sake of convenience in explaining the inspection device 1, we will introduce the XYZ Cartesian coordinate system. For example, the plane parallel to the stage surface of the stage 52 on which the sample 50 is placed will be called the XY plane, and the direction perpendicular to the stage surface will be called the Z axis direction. For convenience, the +Z axis direction will be called upward, and the -Z axis direction will be called downward. Note that upward and downward are for the sake of convenience in explaining the inspection device 1 and do not indicate the actual direction in which the inspection device 1 is placed.

[0033] The light source 11 generates illumination light L11. The illumination light L11 contains, for example, EUV light of 13.5 nm, the same wavelength as the exposure wavelength of the EUV mask, which is the sample 50. The illumination light L11 generated from the light source 11 is reflected by the ellipsoidal mirror 12. The illumination light L11 reflected by the ellipsoidal mirror 12 travels while being focused and is concentrated at the focal point IF1. The focal point IF1 is located in a position conjugate to the upper surface 51 of the sample 50.

[0034] The illumination light L11, after passing through the focal point IF1, spreads out as it travels and enters a reflecting mirror such as the ellipsoidal mirror 13. The illumination light L11 that enters the ellipsoidal mirror 13 is reflected by the ellipsoidal mirror 13, and as it travels, it is focused and enters the recessed mirror 14. In other words, the ellipsoidal mirror 13 causes the illumination light L11 to enter the recessed mirror 14 as a focused beam. The recessed mirror 14 is positioned above the sample 50. The illumination light L11 that enters the recessed mirror 14 and is reflected enters the sample 50. That is, the recessed mirror 14 causes the illumination light L11 to enter the sample 50.

[0035] The ellipsoidal mirror 13 focuses the illumination light L11 onto the sample 50. The illumination optical system 10 is positioned so that when the illumination light L11 illuminates the sample 50, the image of the light source 11 is projected onto the upper surface 51 of the sample 50. Therefore, the illumination optical system 10 provides critical illumination. In this way, the illumination optical system 10 illuminates the sample 50 using critical illumination provided by the illumination light L11 generated by the light source 11.

[0036] The sample 50 is placed on the stage 52. The illumination light L11 is incident on the sample 50 from a direction inclined from the Z-axis direction. That is, the illumination light L11 is incident on the sample 50 at an oblique angle as oblique incidence illumination. The illumination light L11 may illuminate the sample 50 as such oblique incidence illumination.

[0037] Stage 52 is an XYZ drive stage. By moving Stage 52 in the X-axis and Y-axis directions, a desired area of ​​the sample 50 can be illuminated. Furthermore, by moving Stage 52 in the Z-axis direction, focus adjustment can be performed. Stage 52 may also be rotated using the X-axis, Y-axis, and Z-axis as rotation axes. Alternatively, instead of moving and rotating Stage 52 along the X-axis, Y-axis, and Z-axis directions, the illumination optical system 10 and the imaging optical system 20 may be moved and rotated.

[0038] Illumination light L11 from light source 11 illuminates the inspection area of ​​sample 50. Reflected light L12, which is incident from a direction inclined with respect to the Z axis and reflected by sample 50, enters the perforated concave mirror 21. A hole 21a is provided in the center of the perforated concave mirror 21.

[0039] The reflected light L12 reflected by the perforated concave mirror 21 is incident on the convex mirror 22. The convex mirror 22 reflects the reflected light L12 incident from the perforated concave mirror 21 toward the hole 21a of the perforated concave mirror 21. The reflected light L12 that has passed through the hole 21a is detected by the detector 23. The detector 23 may be a detector 23 that includes a TDI (Time Delay Integration) sensor. The detector 23 acquires image data of the sample 50. The detector 23 includes multiple image sensors arranged in a line in one direction. Linear image data captured by multiple image sensors arranged in a line is called one-dimensional image data, or one frame. The detector 23 acquires multiple one-dimensional image data by scanning in a direction orthogonal to one direction. The image sensor is, for example, a CCD (Charge Coupled Device). However, the image sensor is not limited to a CCD.

[0040] In this manner, the imaging optical system 20 collects the reflected light L12 from the sample 50 illuminated by the illumination light L11, and detects the collected reflected light L12 with the detector 23 to acquire image data of the sample 50. The image data is, for example, one-dimensional image data. Alternatively, the imaging optical system 20 may collect the reflected light L12 from an object illuminated by the illumination light L11, and detect the collected reflected light L12 with the detector 23 to acquire image data of the object.

[0041] The reflected light L12 contains information about defects in the sample 50. The specularly reflected light of the illumination light L11 incident on the sample 50 from a direction tilted with respect to the Z axis is detected by the imaging optical system 20. If defects exist in the sample 50, the defects are observed as dark images. This observation method is called bright-field observation. Multiple one-dimensional image data of the sample 50 acquired by the detector 23 are output to the image processing device 40 and processed into two-dimensional image data.

[0042] The image processing device 40 is connected to the imaging optical system 20 by a signal line or wirelessly. The image processing device 40 receives image data of the sample 50 from the detector 23 in the imaging optical system 20. The image processing device 40 processes the image data of the sample 50 received from the detector 23 as a two-dimensional captured image. The image processing device 40 inspects the sample 50 using the image-processed captured image.

[0043] With this configuration, the inspection device 1 inspects the sample 50 for defects and contamination. The sample 50 is, for example, an EUV mask that corresponds to EUV light. However, the sample 50 is not limited to an EUV mask. The sample 50 may be a photomask that corresponds to illumination light L11 having a different wavelength, or a semiconductor substrate. The photomask has a pattern.

[0044] <Pattern> Next, the pattern formed on sample 50 will be described. As an example, the mask pattern of a photomask will be described. Note that the pattern is not limited to a mask pattern, but may also be a pattern formed on a semiconductor substrate, etc. In the following, sample 50, which is the subject of predetermined information processing such as the determination of the presence or absence of defects, will be described, but the same applies to the object used to acquire the first sample image region and the second sample image region described later. Here, the object used to acquire the first sample image region and the second sample image region may be the same object as sample 50, which is the subject of predetermined information processing such as the determination of the presence or absence of defects, or it may be a different object.

[0045] Figure 2 is an example of the mask pattern MP of sample 50 according to Embodiment 1. As shown in Figure 2, sample 50 has a mask pattern MP. The mask pattern MP may include manufacturing tolerances called line edge roughness (hereinafter referred to as LER) at the edges.

[0046] Figure 3 is an example of an image taken such that the mask pattern MP of sample 50 according to Embodiment 1 extends in a first direction. Figure 4 is an example of an image taken such that the mask pattern MP of sample 50 according to Embodiment 1 extends in a second direction. As shown in Figure 3, in image G01, the mask pattern MP extends in a first direction. As shown in Figure 4, in image G02, the mask pattern MP extends in a second direction perpendicular to the first direction.

[0047] Here, we introduce an αβ orthogonal coordinate system within the plane of the image. For example, the image plane is sometimes called the reference image plane. Within the plane of the image, the vertical direction is defined as the α-axis direction, and the horizontal direction as the β-axis direction. The α-axis direction is called the first direction, and the β-axis direction is called the second direction. In this case, the mask pattern MP in image G01 extends in the α-axis direction, and the mask pattern MP in image G02 extends in the β-axis direction.

[0048] The α-axis and β-axis directions may or may not correspond to the X-axis and Y-axis directions of the stage 52. For example, the α-axis and β-axis directions in the reference image plane may be made to correspond to the X-axis and Y-axis directions of the stage 52. Therefore, when the mask pattern MP shown in Figure 2 is positioned along the X-axis direction of the stage 52, imaging is performed along the α-axis direction, as shown in image G01 in Figure 3. On the other hand, when the mask pattern MP shown in Figure 2 is positioned along the Y-axis direction of the stage 52, imaging is performed along the β-axis direction, as shown in image G02 in Figure 4. In this way, by positioning the same mask pattern MP along the X-axis or Y-axis direction of the stage 52, imaging is performed along the α-axis or β-axis direction.

[0049] For the sake of explanation, in the following, we will assume that when the mask pattern MP of sample 50 is positioned to extend in the X-axis direction of stage 52, the mask pattern MP will be imaged to extend in the α-axis direction. Also, when the mask pattern MP is positioned to extend in the Y-axis direction of stage 52, we will assume that the mask pattern MP will be imaged to extend in the β-axis direction.

[0050] Furthermore, depending on the number and type of optical elements constituting the imaging optical system 20, if the mask pattern MP is positioned to extend in the X-axis direction of the stage 52, the mask pattern MP may be imaged to extend in the β-axis direction, and if the mask pattern MP is positioned to extend in the Y-axis direction of the stage 52, the mask pattern MP may be imaged to extend in the α-axis direction.

[0051] When the sample 50 is imaged, the mask pattern MP that is imaged extending in the α-axis direction is called the first pattern 71 or the third pattern 73. For example, among the mask pattern MPs that extend in the α-axis direction which is the first direction in the reference image plane, the mask pattern MP in the first sample image that includes the first sample image region described later is the first pattern 71. Also, for example, among the mask pattern MPs that extend in the α-axis direction which is the first direction in the reference image plane, the mask pattern MP in the sample image of the object to which predetermined information processing, such as the determination of the presence or absence of defects described later, is called the third pattern 73.

[0052] When sample 50 is imaged, the mask pattern MP that is imaged extending in the β-axis direction is the second pattern 72, or the fourth pattern. 74 This is referred to as the second pattern 72. For example, among the mask patterns MP extending in the β-axis direction, which is the second direction orthogonal to the first direction in the reference image plane, the mask pattern MP in the second sample image that includes the second sample image region described later is the second pattern 72. Also, for example, among the mask patterns MP extending in the β-axis direction, which is the second direction orthogonal to the first direction in the reference image plane, the mask pattern MP in the sample image of the object to which predetermined information processing, such as the determination of the presence or absence of defects described later, is taken is the fourth pattern 74.

[0053] For example, images G01 and G02 may be images of the sample 50 taken with an imaging optical system 20 having different numerical apertures (hereinafter referred to as NA) in the X-axis direction and the Y-axis direction. Specifically, image G01 may be an image taken when the sample 50 having a mask pattern MP is placed on the stage 52 such that the mask pattern MP extends in the X-axis direction. In other words, image G01 may be an image taken when the sample 50 having a mask pattern MP is placed on the stage 52 such that it has a first pattern 71. Image G02 may be an image taken when the sample 50 having a mask pattern MP is placed on the stage 52 such that the mask pattern MP extends in the Y-axis direction. In other words, image G02 may be an image taken when the sample 50 having a mask pattern MP is placed on the stage 52 such that it has a second pattern 72. The imaging optical system 20 may be an imaging optical system in which the NA differs in a first direction when the captured image is used as a reference image plane and in a second direction orthogonal to the first direction. Furthermore, the imaging optical system may have different imaging magnifications in a first direction when the captured image is considered as a reference image plane, and in a second direction perpendicular to the first direction.

[0054] Furthermore, for example, images G01 and G02 may be images taken by changing the angle between the principal axis of oblique incidence illumination and the direction in which the mask pattern MP extends. For example, image G01 may be an image taken such that the angle between the direction in which the principal axis of oblique incidence illumination is projected onto the upper surface of the sample 50 and the X-axis direction in which the mask pattern MP extends is a predetermined angle. Image G02 may have the principal axis of oblique incidence illumination in the same direction as image G01. In this case, the mask pattern MP extends in the Y-axis direction, and the angle between the direction in which the principal axis of oblique incidence illumination is projected onto the upper surface of the sample 50 and the Y-axis direction in which the mask pattern MP extends is taken such that the angle differs by approximately 90° from the predetermined angle. Here, approximately 0° and approximately 90° mean not only 0° and 90° but also within a range that includes measurement and instrument errors. Note that angles such as 0° and 90° shown below may be used with the meaning including the abbreviation even when the abbreviation is omitted.

[0055] Furthermore, for example, images G01 and G02 may be images of each mask pattern MP of a sample having both a mask pattern MP extending in the X-axis direction and a mask pattern MP extending in the Y-axis direction (for example, Figure 6 described later) captured by oblique incidence illumination.

[0056] As shown in Figures 3 and 4, the magnitude of the LER's influence on the image changes depending on the orientation of the mask pattern MP during imaging. For example, if the NA of the imaging optical system 20 used in a photomask inspection device using EUV light is different in the X-axis and Y-axis directions, the resolution of images G01 and G02 may change in the α-axis and β-axis directions. For example, the appearance of the mask pattern MP differs depending on whether it is viewed along the α-axis direction (e.g., approximately 0° direction), as in the first pattern 71 shown in Figure 3, or along the β-axis direction (e.g., 90° direction), as in the second pattern 72 shown in Figure 4. As a result, even for defects of the same shape, the detection sensitivity changes depending on whether the defect is on the edge of the pattern in the α-axis direction or the edge of the pattern in the β-axis direction.

[0057] Therefore, the image processing device 40 of this embodiment detects the mask pattern MP in the α-axis direction and the β-axis direction, and allows for independent adjustment of the sensitivity in each direction. Specifically, as will be described later, the setting unit 42 of the image processing device 40 independently sets parameters used for predetermined information processing, including the determination of the presence or absence of defects in each direction, for each direction. This makes it possible to set different algorithms for each direction, thereby reducing the sensitivity difference between directions.

[0058] <Image Processing Device> Next, the image processing apparatus 40 will be described. Figure 5 is a block diagram illustrating the image processing apparatus 40 according to Embodiment 1. As shown in Figure 5, the image processing apparatus 40 comprises an acquisition unit 41, a setting unit 42, and a processing unit 43. The acquisition unit 41, the setting unit 42, and the processing unit 43 have the functions of acquisition means, setting means, and processing means. The image processing apparatus 40 may further include units having other functions, such as a storage unit that serves as a storage means. The processing unit 43 may be at least one of the determination unit 43a and the image correction unit 43b, which will be described later. The determination unit 43a and the image correction unit 43b have the functions of determination means and image correction means.

[0059] The acquisition unit 41 acquires a first sample image region and a second sample image region. The first sample image region and the second sample image region may be acquired from images of objects actually prepared for acquisition of the first and second sample image regions, or from images formed from design data. The images used to acquire the first and second sample image regions are called sample images. Therefore, the sample images may be images captured by the imaging optical system 20, or images based on design data.

[0060] Figures 6-8 illustrate the mask patterns MP1 and MP2 of samples 50a-50c for sample images according to Embodiment 1. As shown in Figures 6-8, samples 50a-50c for sample images have at least one of mask patterns MP1 and MP2. Samples 50a-50c for sample images may be referred to as "objects" to distinguish them from sample 50 for sample images. As shown in Figure 6, sample 50a for sample images has mask patterns MP1 and MP2. As shown in Figure 7, sample 50b for sample images has mask pattern MP1. As shown in Figure 8, sample 50c for sample images has mask pattern MP2.

[0061] Figure 9 is a diagram illustrating a first sample image G11 acquired by the acquisition unit 41 in the image processing apparatus 40 according to Embodiment 1. As shown in Figure 9, the first sample image G11 includes a first pattern 71 that extends in the α-axis direction within the plane of the image surface. The region including the first pattern 71 is called the first sample image region. Therefore, the first sample image G11 includes the first sample image region. The acquisition unit 41 may acquire the first sample image G11 and the first sample image region from an image of a sample 50a for the sample image, or it may acquire the first sample image G11 and the first sample image region from an image of a sample 50b for the sample image. In other words, the first sample image region may be a region included in the image captured by a predetermined imaging optical system 20 of an object having the first pattern 71.

[0062] Figure 10 is an example of a second sample image G12 acquired by the acquisition unit 41 in the image processing apparatus 40 according to Embodiment 1. As shown in Figure 10, the second sample image G12 includes a second pattern 72 that extends in the β-axis direction within the plane of the image. The region including the second pattern 72 is called the second sample image region. Therefore, the second sample image G12 includes the second sample image region. The acquisition unit 41 may acquire the second sample image G12 and the second sample image region from an image of a sample 50a for the sample image, or it may acquire the second sample image G12 and the second sample image region from an image of a sample 50c for the sample image. In other words, the second sample image region may be a region included in the image captured by a predetermined imaging optical system 20 of an object having the second pattern 72.

[0063] The acquisition unit 41 may acquire the first sample image G11 and the first sample image region on the stage 52 by oblique incident illumination onto a sample 50a or sample 50b having the first pattern 71, such that the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the sample 50a and the X-axis direction (i.e., the first direction or α-axis direction) to which the first pattern 71 extends is a predetermined angle. In other words, the predetermined imaging optical system 20 captures an image of the object by oblique incident illumination. The first sample image region is included in the captured image when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the object and the first direction is a predetermined angle.

[0064] Furthermore, when the acquisition unit 41 acquires the second sample image G12 and the second sample image region by oblique incident illumination onto a sample 50a or sample 50c having the second pattern 72, the second sample image G12 may be acquired such that the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the sample 50a and the Y-axis direction (i.e., the second direction and the β-axis direction) to which the second pattern 72 extends is 90° different from a predetermined angle. Also, the predetermined imaging optical system 20 captures an image of the object by oblique incident illumination. The first sample image region and the second sample image region may be included in the captured image when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the object and the first direction is a predetermined angle.

[0065] Furthermore, the acquisition unit 41 may acquire the second sample image G12 and the second sample image region from an image of sample 50b. That is, sample 50b is rotated 90° around the Z-axis on the stage 52. This allows the first pattern 71 to be imaged as the second pattern. Similarly, the acquisition unit 41 may acquire the first sample image G11 and the first sample image region from an image of sample 50c. That is, sample 50c is rotated 90° around the Z-axis on the stage 52. This allows the second pattern 72 to be imaged as the first pattern. Thus, the first sample image region and the second sample image region may be regions included in the image of the object having the first pattern captured by the predetermined imaging optical system 20. The orientation of the object relative to the predetermined imaging optical system 20 when capturing the image including the first sample image region and the orientation of the object relative to the predetermined imaging optical system 20 when capturing the image including the second sample image region differ by 90° in the main plane of the object.

[0066] Furthermore, when the acquisition unit 41 acquires the first sample image G11, the first sample image region, the second sample image G12, and the second sample image region on the stage 52 by oblique incident illumination onto a sample 50a having the first pattern 71 and the second pattern 72, it may acquire the first sample image G11, the first sample image region, the second sample image G12, and the second sample image region by setting the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the sample 50a and the X-axis direction (i.e., the first direction or the α-axis direction) to a predetermined angle. In other words, the first sample image region and the second sample image region may also be regions included in the image captured by the predetermined imaging optical system 20 of the object having the first pattern 71 and the second pattern 72.

[0067] When the first sample image G11, the first sample image region, the second sample image G12, and the second sample image region are acquired by oblique incidence illumination on a sample 50a having the first pattern 71 and the second pattern 72, the orientation of the object relative to the predetermined imaging optical system 20 when acquiring the image including the first sample image region and the orientation of the object relative to the predetermined imaging optical system 20 when acquiring the image including the second sample image region may be substantially the same. Alternatively, the image including the first sample image region and the image including the second sample image region may be the same image.

[0068] The acquisition unit 41 may acquire at least one of the first sample image G11 and the second sample image G12 based on the design data of the sample 50 or the design data of the pattern formed on the sample 50.

[0069] Figure 11 is a diagram illustrating an image G13 acquired by an acquisition unit 41 in an image processing apparatus 40 according to Embodiment 1. As shown in Figure 11, the acquisition unit 41 acquires the image G13. The image G13 includes a third pattern 73 extending in the α-axis direction and a fourth pattern 74 extending in the β-axis direction within the plane of the image surface, when the plane of the image G13 is used as the reference image plane. The image G13 is a sample image obtained by imaging a sample having the third pattern 73 and the fourth pattern 74. The third pattern 73 is a mask pattern MP1 that is imaged to extend in the α-axis direction when the sample is imaged, similar to the first pattern 71. The fourth pattern 74 is a mask pattern MP2 that is imaged to extend in the β-axis direction when the sample is imaged, similar to the second pattern 72.

[0070] If the predetermined information processing using the first and second parameters described later can be performed, the pattern width and pitch spacing of the first pattern 71 and the third pattern 73 may be the same or different. Similarly, the pattern width and pitch spacing of the second pattern 72 and the fourth pattern 74 may be the same or different. Note that the same or different pattern widths and pitch spacings here may refer to a comparison of the pattern widths and pitch spacings of patterns physically formed on the sample, or to a comparison of the pattern widths and pitch spacings of patterns visible in the sample image or patterns drawn as images in the first sample image region and the second sample image region.

[0071] The captured image G13 is, for example, an image of a sample on which both a pattern extending in one direction and a pattern extending in a direction approximately 90° different are formed on the sample surface. The pattern extending in one direction and the pattern extending in a direction 90° different from it may be formed with physically approximately identical pattern widths and pitches, even though their directions of extension are different. Alternatively, like an anamorphic mask, they may be formed so that their directions of extension and physical pattern widths and pitches are different, but appear approximately identical on the image plane. The captured image G13 may be, for example, an image of the sample 50a used for the sample image shown in Figure 6. In that case, the sample 50a is an object that is the subject of predetermined information processing such as defect judgment, and is also the object for acquiring the first sample image region and the second sample image region. The third pattern 73 is the first pattern, and the fourth pattern 74 is the second pattern. In the following explanation, image G13 will be described as a sample image taken of sample 50a.

[0072] The captured image G13 may include both the third pattern 73 and the fourth pattern 74. The acquisition unit 41 acquires the captured image G13 of the sample 50a including the third pattern 73 and the fourth pattern 74. The acquisition unit 41 may also acquire the captured image G13 captured by the imaging optical system 20 with different NAs in the X-axis direction and the Y-axis direction.

[0073] The acquisition unit 41 may acquire the captured image G13 by oblique incident illumination onto a sample 50a having a third pattern 73, such that the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the sample 50a and the X-axis direction (i.e., the first direction and the α-axis direction) is a predetermined angle. In other words, the predetermined imaging optical system 20 captures an image of the sample 50a by oblique incident illumination. The sample image is an image captured by the predetermined imaging optical system 20 when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the sample and the first direction is a predetermined angle.

[0074] Figures 12 and 13 are graphs illustrating the relationship between the difference signal between predetermined pixels and the probability of occurrence in the captured image G13 acquired by the acquisition unit 41 in the image processing apparatus 40 according to Embodiment 1. The horizontal axis represents the difference signal, and the vertical axis represents the probability of occurrence. Figure 12 shows the difference signal between predetermined pixels of the third pattern 73 in the captured image G13, and Figure 13 shows the difference signal of the fourth pattern 74 in the captured image G13.

[0075] As shown in Figures 12 and 13, the probability of the difference signal appearing in the third pattern 73 is wider than the probability of the difference signal appearing in the fourth pattern 74. In other words, the probability of the third pattern 73 appearing is wider than the probability of the fourth pattern 74 appearing. 74 This indicates that the influence of noise and other factors is greater than that of the previous pattern. In this way, the probability distribution of occurrence may be created from the histogram of the difference signal obtained from the patterns in the α-axis and β-axis directions of the captured image G13. Then, the influence of noise in each direction may be predicted from the probability distribution of occurrence of the difference signal created. This makes it possible to optimize the defect detection sensitivity in each direction. For example, the third pattern 73 is more likely to be affected than the fourth pattern 74 If the impact of noise etc. is greater, then the third pattern 73 The threshold for defect detection is the fourth pattern 74 It will be milder than that.

[0076] The setting unit 42 sets a first parameter for the first pattern 71 based on the first sample image region. The first parameter may include a parameter for detecting the first pattern 71, or a parameter for detecting defects in the first pattern 71. The setting unit 42 may set the first parameter for the first pattern 71 based on the line edge roughness (LER) distribution in the first sample image region. However, the setting unit 42 may set the first parameter for the first pattern 71 not only based on the LER distribution in the first sample image region, but also based on the central brightness, average brightness, line width, etc., of the first pattern 71 in the first sample image region.

[0077] The setting unit 42 may set the first parameter to include a threshold used to determine the presence or absence of defects in the third pattern 73 contained in the captured image G13 (sample image as described above, hereinafter the same). Specifically, the setting unit 42 may set the first parameter to include a threshold used for defect inspection of the third pattern 73. Alternatively, the setting unit 42 may set the first parameter to include a threshold used for defect inspection based on the brightness difference of the third pattern 73. The brightness difference includes the difference between the brightness of the first pattern 71 in the first sample image G11 and the brightness of the third pattern 73 in the captured image G13.

[0078] Furthermore, the setting unit 42 sets a second parameter for the second pattern 72 based on the second sample image region. The second parameter may include a parameter for detecting the second pattern 72, or it may include a parameter for detecting defects in the second pattern 72. The setting unit 42 may set the second parameter for the second pattern 72 based on the LER distribution in the second sample image region. Note that the setting unit 42 may set the second parameter for the second pattern 72 not only based on the LER distribution in the second sample image region, but also based on the central brightness, average brightness, line width, etc., of the second pattern 72 in the second sample image region.

[0079] The setting unit 42 may set the second parameter to include a threshold used to determine the presence or absence of defects in the fourth pattern 74 contained in the captured image G13 (sample image as described above, hereinafter the same). Specifically, the setting unit 42 may set the second parameter to include a threshold used for defect inspection of the fourth pattern 74. Alternatively, the setting unit 42 may set the second parameter to include a threshold used for defect inspection based on the brightness difference of the fourth pattern 74. The brightness difference includes the difference between the brightness of the second pattern 72 in the second sample image G12 and the brightness of the fourth pattern 74 in the captured image G13.

[0080] The setting unit 42 may set thresholds for multiple types of physical quantities related to the information of the captured image G13. The setting unit 42 may set at least one of the types and number of physical quantities for which thresholds are defined as first parameters to be different from the physical quantities for which thresholds are defined as second parameters. As described above, for example, the third pattern 73 is better than the fourth pattern 74 If the impact of noise etc. is greater, then the third pattern 73 The threshold for defect detection is the fourth pattern 74 In some cases, it is preferable to relax the setting. To that end, the setting unit 42 may set the number of physical quantities used as thresholds when the second parameter is set as a threshold used to determine whether or not there is a defect in the fourth pattern 74 to be greater than the number of physical quantities used as thresholds when the first parameter is set as a threshold used to determine whether or not there is a defect in the third pattern 73. The physical quantities include the LER of the pattern, the central brightness of the pattern, the average brightness of the pattern, the line width of the pattern, the brightness difference, etc., and may be selected from this group.

[0081] The processing unit 43 performs predetermined information processing on the captured image G13 based on the first and second parameters. The predetermined information processing may include performing predetermined judgments and predetermined image corrections. The predetermined judgments may include determining whether or not there are defects. The processing unit 43 performs predetermined information processing, including determining whether or not there are defects in the third pattern 73 and the fourth pattern 74 included in the captured image G13. In this case, the processing unit 43 may be referred to as the judgment unit 43a instead of the processing unit 43. The processing unit 43 (judgment unit 43a) may perform predetermined information processing, including determining whether or not there are defects in at least one of the third pattern 73 and the fourth pattern 74 included in the captured image G13.

[0082] Furthermore, the processing unit 43 may perform predetermined information processing, including image correction of the captured image G13, to correct or mitigate the expected difference in sensitivity between the third pattern 73 and the fourth pattern 74 included in the captured image G13. In this case, the processing unit 43 may be referred to as the image correction unit 43b instead of the processing unit 43. In this case, for example, the setting unit 42 sets the first parameter as a brightness correction value for pixels belonging to the third pattern 73 included in the captured image G13, and sets the second parameter as a brightness correction value for pixels belonging to the fourth pattern 74 included in the captured image G13. Based on these first and second parameters, the processing unit 43 (image correction unit 43b) performs predetermined information processing, including image correction, to correct the brightness of at least some of the pixels belonging to the third pattern 73 or the fourth pattern 74 included in the captured image G13. The setting unit 42 may also set the first and second parameters by, for example, specifying the brightness correction values ​​for pixels belonging to the first pattern 71 and the second pattern 72 so that the LER distribution of the first pattern 71 and the LER distribution of the second pattern 72 have similar shapes, thereby setting the brightness correction values ​​for pixels belonging to the third pattern 73 and the fourth pattern 74.

[0083] For example, if the LER value of the third pattern 73 in the captured image G13 is greater than the threshold value of the first parameter set in advance, the determination unit 43a determines that there is a defect in the third pattern 73. Also, if the LER value of the fourth pattern 74 in the captured image G13 is greater than the threshold value of the second parameter set in advance, the determination unit 43a determines that there is a defect in the fourth pattern 74.

[0084] For example, as shown in Figures 2 to 4, even when the same mask pattern MP is imaged, the sensitivity may differ depending on the orientation of the imaged mask pattern MP on the image plane. Depending on such differences in sensitivity, for example, the threshold of LER as the first parameter for a pattern extending in a first direction may be different from the threshold of LER as the second parameter for a pattern extending in a second direction. The threshold for a coarsely imaged edge, as shown in Figure 3, will be coarser than the threshold for a smoothly imaged edge, as shown in Figure 4. By inspecting the sample based on a preset threshold for each direction in which the pattern extends using the setting unit 42, it is possible to suppress misjudgments such as identifying something that appears to be a rough edge but is not a defect as a defect, and overlooking defects that appear to be not defects because they appear to be smooth edges.

[0085] In sample 50, if a pattern extending in a first direction and a pattern extending in a second direction 90° different from the first are formed with physically similar pattern widths and pitch spacings, the setting unit 42 sets a first parameter and a second parameter (for example, a defect detection threshold) for each direction in which the pattern extends. This makes it possible to make a judgment that takes into account differences in resolution due to differences in NA, for example. Furthermore, if sample 50 is an anamorphic mask, where the physical pattern widths and pitch spacings of a pattern extending in a first direction and a pattern extending in a second direction 90° different from the first are different, but the pattern widths and pitch spacings appear to be approximately the same on the captured image plane, the setting unit 42 sets a first parameter and a second parameter (for example, a defect detection threshold) for each direction in which the pattern extends. This makes it possible to make a judgment that takes into account differences in resolution due to differences in pattern orientation, for example.

[0086] <Image Processing Method> Next, an image processing method using the image processing apparatus according to this embodiment will be described. Figure 14 is a flowchart illustrating an image processing method using the image processing apparatus according to Embodiment 1. As shown in step S11 of Figure 14, the first and second parameters are set. Specifically, the setting unit 42 sets the first parameter for the first pattern 71 based on the first sample image area. The setting unit 42 also sets the second parameter for the second pattern 72 based on the second sample image area.

[0087] Here, the setting unit 42 may set the first and second parameters based on the first and second sample image regions included in the image obtained by setting the angle of oblique incident illumination on the object having the first pattern to 0° and 90° with respect to the direction in which the first pattern extends, as described above. The setting unit 42 may also set the first and second parameters based on the first and second sample image regions included in the image obtained by not changing the angle of oblique incident illumination on the object having the first and second patterns with respect to the direction in which the first pattern extends. Alternatively, the setting unit 42 may set the first and second parameters based on the first and second sample image regions generated based on the design data of the object having the first pattern or the object having the first and second patterns.

[0088] When setting the first parameter, the setting unit 42 may set the first parameter to include a threshold used to determine whether or not there is a defect in the third pattern included in the captured image G13. Furthermore, when setting the second parameter, the setting unit 42 may set the second parameter to include a threshold used to determine whether or not there is a defect in the fourth pattern included in the captured image G13.

[0089] Next, as shown in step S12, an image G13 is acquired. Specifically, the acquisition unit 41 acquires an image G13 of the sample 50a including the third pattern 73 and the fourth pattern 74. The image G13 may include both the third pattern 73 and the fourth pattern 74, or it may include at least one of the third pattern 73 and the fourth pattern 74.

[0090] Next, as shown in step S13, predetermined information processing is performed. Specifically, the processing unit 43 performs predetermined information processing on the captured image G13 based on the first and second parameters. When performing predetermined information processing, the processing unit 43 may perform predetermined information processing including determining whether or not there are defects in the third pattern 73 and fourth pattern 74 included in the captured image G13. Alternatively, when performing predetermined information processing, the processing unit 43 may perform predetermined information processing including image correction of the captured image G13.

[0091] <Testing Method> Next, the inspection method will be described as the operation of the inspection device 1 of this embodiment. 15 This is a flowchart illustrating an inspection method using the inspection apparatus 1 according to Embodiment 1. As shown in step S101 of Figure 15, first, the sample 50 is illuminated with illumination light L11. For example, when illuminating the sample 50 with illumination light L11 generated by the light source 11, the illumination optical system 10 is arranged to provide critical illumination. Then, the sample 50 is illuminated with illumination light L11.

[0092] Next, as shown in step S102, the sample 50 is imaged. For example, the imaging optical system 20 images the sample 50 illuminated by the illumination light L11. The imaging optical system 20 then outputs the imaged image G13 of the sample 50 to the image processing device 40 as the detection result of the detector 23.

[0093] Next, as shown in step S103, the image processing device 40 processes the captured image G13. Specifically, the image processing device 40 sets the first parameter, the second parameter, etc., as in the image processing method described above. Then, the image processing device 40 performs predetermined information processing on the captured image acquired from the imaging optical system 20.

[0094] Next, the effects of this embodiment will be described. The image processing device 40 of this embodiment sets a first parameter for the first pattern 71 and a second parameter for the second pattern 72. This allows the parameters to be adjusted independently in the α-axis direction and the β-axis direction. Therefore, appropriate measures can be taken to address differences in sensitivity caused by the imaging direction of the pattern.

[0095] Furthermore, since parameters are set in advance for the first pattern 71 and the second pattern 72, the effects of differences in resolution in each direction can be predicted beforehand.

[0096] For example, based on the first sample image G11 and the second sample image G12, thresholds for defect detection may be set for parameters such as LER, brightness, and line width. This allows for appropriate defect detection in response to differences in sensitivity caused by the imaging direction of the pattern. For example, by setting separate thresholds for LER defect detection in the α-axis and β-axis directions for the first pattern 71 and the second pattern 72, the influence of LER and other parameters can be predicted, and the defect detection sensitivity can be optimized.

[0097] The first sample image G11 and the second sample image G12 may be actual images captured. This allows the parameter accuracy to be matched to the actual captured image G13. Alternatively, the first sample image G11 and the second sample image G12 may be obtained by performing an optical simulation using the mask design pattern and the NA of the imaging optical system as input. This eliminates the need for actual imaging and allows the parameters to approach ideal values.

[0098] This embodiment is more effective against differences in appearance due to illumination direction and characteristics of the imaging optical system 20, such as oblique incidence illumination where EUV light is incident at an angle. Furthermore, the sample 50 may include an anamorphic mask used by the imaging optical system 20, which has different magnifications in the X-axis and Y-axis directions. In a HighNA mask such as an anamorphic mask, the linewidth in the X-axis direction and the linewidth in the Y-axis direction may differ. By applying this embodiment to such a sample 50, the defect detection sensitivity can be further optimized. Also, the acquisition unit 41 may acquire an image G13 with a different number of binned pixels in the X-axis and Y-axis directions. Such an image G13 also has differences in sensitivity due to the imaging direction of the pattern, so by applying this embodiment, appropriate countermeasures can be taken. A single device does not need to have all the components such as the acquisition unit 41, setting unit 42, and processing unit 43. For example, the acquisition unit 41, setting unit 42, and processing unit 43 may be implemented in two or more devices such as an optical device or a server device, and the image processing device may be configured by communicating and coordinating each device. Alternatively, the image processing apparatus may be configured to function as an acquisition unit 41, a setting unit 42, and a processing unit 43 by communicating and coordinating two or more devices, such as optical devices and server devices.

[0099] (Embodiment 2) Next, Embodiment 2 to The image processing device 40 is described below. This embodiment utilizes image conversion parameters such as a kernel. The image processing device 40 of this embodiment also has the same configuration as the image processing device 40 of Embodiment 1, and includes an acquisition unit 41, a setting unit 42, and a processing unit 43.

[0100] The acquisition unit 41 acquires a sample image, which is an image of the sample 50 captured by a predetermined imaging optical system 20.

[0101] The setting unit 42 sets image transformation parameters based on a first sample image region, which is a region containing a first pattern extending in a first direction in the reference image plane, and a second sample image region, which is a region containing a second pattern extending in a second direction in the reference image plane. Here, the image transformation parameters are, for example, Next Includes the original kernel and the two-dimensional kernel, etc. Below, as an example of image conversion parameters, Next The original kernel and the two-dimensional kernel will be used for the explanation. Note that the image transformation parameters are one Next The original kernel and the two-dimensional kernel are not the only options; for example, the first and second parameters mentioned above may also be included. The first sample image region and the second sample image region may be acquired by the acquisition unit 41 by imaging the target object or by imaging the sample 50.

[0102] Figure 16 illustrates an image processing apparatus 40 according to Embodiment 2, showing a first pattern extending in a first direction, a luminance distribution along a second direction in the first pattern, and a one-dimensional kernel K11 applied to the luminance distribution. Figure 17 illustrates an image processing apparatus 40 according to Embodiment 2, showing a second pattern extending in a second direction, a luminance distribution along the first direction in the second pattern, and a one-dimensional kernel K12 applied to the luminance distribution. Figure 18 illustrates an image processing apparatus 40 according to Embodiment 2, showing a luminance distribution of a reference shape along a direction orthogonal to the direction in which the pattern extends.

[0103] As shown in Figure 16, the one-dimensional kernel K11 may correspond to the luminance distribution along the second direction in the first pattern extending in the first direction. The luminance distribution along the second direction in the first pattern is sometimes called the first luminance distribution. The one-dimensional kernel K11 may also include parameters of vector or matrix elements arranged in the second direction. The one-dimensional kernel K11 is an image transformation parameter that transforms the first luminance distribution shown in Figure 16 to resemble the luminance distribution of the reference shape shown in Figure 18.

[0104] For example, the setting unit 42 sets a one-dimensional kernel K11 that includes components of a moving average filter and components of a sharpening filter. Here, the moving average filter is a filter that smooths the luminance distribution. The sharpening filter is a filter that sharpens the luminance distribution. The setting unit 42 sets the parameters of the one-dimensional kernel K11 by adjusting the proportions of the moving average filter components and the sharpening filter components. The setting unit 42 sets the parameters of the one-dimensional kernel K11 so that the first luminance distribution resembles the luminance distribution of the reference shape.

[0105] As shown in Figure 17, the one-dimensional kernel K12 may correspond to the luminance distribution along the first direction in the second pattern extending in the second direction. The luminance distribution along the first direction in the second pattern is sometimes called the second luminance distribution. The one-dimensional kernel K12 may also include parameters of vectors or matrix elements arranged in the first direction. The one-dimensional kernel K12 is an image transformation parameter that transforms the second luminance distribution shown in Figure 17 to resemble the luminance distribution of the reference shape shown in Figure 18. In this case, the luminance distribution of the reference shape shown in Figure 18 is assumed to be rotated by 90° so that it is a luminance distribution along the first direction, but for simplicity, the explanation of the 90° rotation may be omitted in the following explanation.

[0106] For example, the setting unit 42 includes components of the moving average filter and components of the sharpening filter. Next The original kernel K12 is set. The setting unit 42 sets the parameters of the one-dimensional kernel K12 by adjusting the proportion of the moving average filter component and the proportion of the sharpening filter component. The setting unit 42 sets the parameters of the one-dimensional kernel K12 so that the second luminance distribution resembles the luminance distribution of the reference shape. For example, the setting unit 42 may make the sharpening filter component larger than the moving average filter component in order to make the second luminance distribution in Figure 17 resemble the luminance distribution of the reference shape in Figure 18.

[0107] The luminance distribution of the reference shape shown in Figure 18 may be the luminance distribution along the direction perpendicular to the direction in which the pattern extends in the pattern included in the reference image, or it may be the luminance distribution obtained from optical simulation. Furthermore, the luminance distribution of the reference shape may be the first luminance distribution shown in Figure 16, or the second luminance distribution shown in Figure 17.

[0108] The setting unit 42 may also set the one-dimensional kernel K11 and the one-dimensional kernel K12 without using the luminance distribution of the reference shape. For example, the setting unit 42 may set the one-dimensional kernel K11 and the one-dimensional kernel K12 so that the luminance distribution when the one-dimensional kernel K11 is applied to the first luminance distribution shown in Figure 16 is similar to the luminance distribution when the one-dimensional kernel K12 is applied to the second luminance distribution shown in Figure 17. In this case, one of the luminance distributions is rotated by 90°.

[0109] For example, the setting unit 42 changes at least one of the parameters of the one-dimensional kernel K11 and the one-dimensional kernel K12 so that the difference in brightness of each pixel between the brightness distribution obtained when the one-dimensional kernel K11 is applied to the first brightness distribution shown in Figure 16 and the brightness distribution obtained when the one-dimensional kernel K12 is applied to the second brightness distribution shown in Figure 17 is minimized. The setting unit 42 may also set the one-dimensional kernel K11 and the one-dimensional kernel K12 to include the parameters that result in the minimum brightness difference between the two.

[0110] In the embodiments illustrated above, the luminance distribution obtained by applying the one-dimensional kernel K11 set by the setting unit 42 to the luminance distribution along the second direction in the first pattern extending in the first direction, and the luminance distribution obtained by applying the one-dimensional kernel K12 to the luminance distribution along the first direction in the second pattern extending in the second direction, will have similar shapes.

[0111] Applying one-dimensional kernels K11 and K12 to the luminance distribution means performing convolution processing using one-dimensional kernels K11 and K12 on the pixels that make up the luminance distribution. The convolution processing using one-dimensional kernels K11 and K12 may be carried out in accordance with the convolution processing using two-dimensional kernel K21 or kernel W, etc., which will be described later.

[0112] The setting unit 42 sets a two-dimensional kernel K21 having two-dimensional parameters corresponding to the first and second directions. That is, the setting unit 42 sets the two-dimensional kernel K21 as an image transformation parameter. The setting unit 42 sets the two-dimensional kernel K21 based on the one-dimensional kernel K11 and the one-dimensional kernel K12. Specifically, the setting unit 42 may set the two-dimensional kernel K21 from the product of the vector that is the one-dimensional kernel K11 and the vector that is the one-dimensional kernel K12. That is, when the one-dimensional kernel K12 is a column vector (b1, b2, ..., bn) and the one-dimensional kernel K11 is a row vector (a1, a2, ..., an), the two-dimensional kernel K21 may be set by obtaining an n × n matrix by the product of each element. Here, the i-th column and j-th row element Cij of the two-dimensional kernel K21 is bi × aj. The setting unit 42 may also set the two-dimensional kernel K21 from the cross product of the one-dimensional kernel K11 and the one-dimensional kernel K12. In addition, if the setting unit 42 can set the two-dimensional kernel K21 using the one-dimensional kernel K11 and the one-dimensional kernel K12, it may also set the two-dimensional kernel K21 using a calculation method other than the product or cross product.

[0113] The two-dimensional kernel K21 may have an asymmetric distribution of two-dimensional parameters between the first and second directions. Such a two-dimensional kernel K21 can be described as a kernel set considering, for example, the difference in sensitivity between the vertical and horizontal directions. In other words, depending on the optical device, the sensitivity may differ between the first and second directions. Therefore, in order to mitigate the difference between the sensitivity in the first direction and the sensitivity in the second direction, the setting unit 42 sets the parameter distribution to be asymmetric between the first and second directions. two dimensionalThe kernel K21 is set. Therefore, the parameter distribution in the two-dimensional kernel K21 is asymmetric between the first and second directions. The two-dimensional kernel K21 is an example of a kernel set considering the difference in sensitivity between the vertical and horizontal directions. In other words, the two-dimensional kernel K21 is an example of a kernel set to mitigate the difference in sensitivity between the vertical and horizontal directions.

[0114] The processing unit 43 performs predetermined information processing on the sample image or an image based on the sample image by applying image transformation parameters such as the two-dimensional kernel K21. The two-dimensional kernel K21 is sometimes simply referred to as kernel W. As described above, the processing unit 43 performs predetermined information processing by applying kernel W in order to mitigate the difference in the appearance of images captured with different sensitivities in the first and second directions.

[0115] Below, we will describe the process of applying a kernel to a sample image as a predetermined information processing step. Then, we will describe the defect inspection step. Furthermore, in Modification 1 of Embodiment 2, we will describe the kernel parameters based on the difference image. Then, we will describe the process of applying a kernel to a sample image as a predetermined processing step. In addition, in Modification 2 of Embodiment 2, we will describe the process of applying a kernel to the difference image as a predetermined information processing step, and then describe the defect inspection step. Furthermore, in Modification 3 of Embodiment 2, we will describe the process of using a modified sample image and a modified difference image as a predetermined information processing step. Note that the image based on the sample image includes the difference image, the modified sample image, and the modified difference image.

[0116] <Processing that applies a kernel to the sample image> The processing unit 43, as a predetermined information processing step, applies a kernel W having a parameter distribution pattern set by the setting unit 42 to the sample image and outputs a corrected sample image. The kernel W applied to the sample image is sometimes referred to as the first kernel W.

[0117] Figure 19 is a schematic diagram illustrating the information processing performed by the processing unit 43 on a sample image or an image based on a sample image in the image processing apparatus 40 according to Embodiment 2. As shown in Figure 19, the processing unit 43 may perform predetermined information processing on the sample image by applying a kernel W.

[0118] When a reference image plane is placed in the sample image, the sample image is an image composed of multiple pixel rows arranged in a first direction and multiple pixel rows arranged in a second direction. Here, the sample image is assumed to be an image in which M pixels p are arranged in the first direction and N pixels p are arranged in the second direction.

[0119] The first kernel W, which has two-dimensional parameters set by the setting unit 42, is, for example, a kernel with 3 rows and 3 columns, i.e., 3 × 3 = 9 cells. The size of each element (sometimes called a cell) of the first kernel W may virtually correspond to the size of the pixel p of the sample image.

[0120] The predetermined information processing may be, for example, the convolution process described in Non-Patent Document 1. That is, the processing unit 43 may perform a convolution process on the sample image by applying the first kernel W. Specifically, at the pixel p corresponding to the center of the first kernel W, a convolution process is performed between the surrounding 3x3 pixel values ​​(e.g., luminance) and the parameters of the first kernel W to filter each pixel in the sample image. For example, each cell of the first kernel W is represented by (i, j) using local coordinate axes including the x and y axes. If J(p) is the local operator that performs image filtering in the 3x3 neighborhood of each pixel p, then J(p) can be defined as the convolution between the luminance value I(x+i, y+j) around the pixel p=(x,y) on the sample image and W'(p), as shown in equation (1) below (the convolution is a two-dimensional discrete approximation).

[0121]

number

[0122] Here, W'(p) is the kernel obtained by inverting the original first kernel W(p) with respect to both the x and y axes of the local coordinate system. i、j is the weight coefficient at each position (i, j) of the inverted kernel W'(p), and S>0 is the coefficient for normalization and scaling. If i and j extend beyond the boundary range of the sample image, a padding process may be performed beforehand to fill the periphery with pixel values ​​I(x+i, y+j). In this way, the processing unit 43 performs a convolution process on each pixel of the sample image and outputs the processed pixel value. The processing unit 43 can sweep the entire sample image with the first kernel W by sequentially changing the pixels of the sample image where the central cell of the first kernel W is located. As a result, the processing unit 43 obtains a modified sample image to which the first kernel W has been applied to the entire sample image.

[0123] The corrected sample image is an image in which M pixels are arranged in the first direction and N pixels are arranged in the second direction, similar to the sample image. Each pixel (p') of the corrected sample image m、n Brightness I' m、n By applying the first kernel W, each pixel (p) of the sample image is processed. m、n ) Brightness I m、n The modified sample image is an image output by applying a first kernel W, which has different parameter distributions in the first and second directions, to the sample image. Therefore, the modified sample image is an image in which the difference in sensitivity between the first and second directions is mitigated compared to the sample image.

[0124] <Defect Inspection> The processing unit 43 detects defects based on the difference image between the reference image and the corrected sample image. The reference image may be a good product image representing an image without abnormalities or an image without critical abnormalities. The difference image may include the difference in brightness of pixels in the reference image and the corrected sample image that are in the same relative position as a parameter for each pixel. The difference image may also include the ratio of the brightness of pixels in the reference image and the corrected sample image that are in the same relative position as a parameter for each pixel. Hereafter, the pixel parameter of the difference image will simply be called brightness. The processing unit 43 may detect defects based on whether the brightness of the pixels included in the difference image exceeds a predetermined threshold.

[0125] <Modification 1 of Embodiment 2> <Kernel parameter settings based on differential images> The setting unit 42 may set image conversion parameters based on an image obtained by comparing the captured image of the object with a reference image of the captured image, or an image obtained by comparing the sample image, which is an captured image of the sample, with a reference image of the sample.

[0126] In the following description, the setting unit 42 will be described as setting image conversion parameters based on an image obtained by comparing the captured image of the object with a reference image of the captured image. The image obtained by comparing the captured image of the object with a reference image of the captured image is, for example, a difference image between the captured image and the reference image. The reference image may be a good image representing an image without abnormalities, or an image without critical abnormalities. The difference image may include the difference in brightness of pixels in the reference image and the captured image of the object that are in the same relative position as a parameter for each pixel. The difference image may also include the ratio of the brightness of pixels in the reference image and the captured image that are in the same relative position as a parameter for each pixel. Hereafter, the pixel parameter of the difference image will simply be referred to as brightness.

[0127] As will be described later, in Modification 1 of Embodiment 2, it is not essential that the object or sample 50 has a pattern extending in one direction. In other words, the object or sample 50 does not need to have a pattern extending in one direction. Due to the increasing complexity of semiconductor structures in recent years, various patterns can be formed on the sample 50, such as a photomask or wafer, including patterns extending in one direction, patterns extending perpendicular to one direction, curved patterns, diagonally extending patterns, and contact holes. In this case, according to Modification 1 of Embodiment 2, regardless of the shape or type of pattern formed on the object or sample 50, the setting unit 42 can set image conversion parameters that take into account the difference in NA and sensitivity between the vertical and horizontal directions.

[0128] Figure 20 is a diagram illustrating a difference image in an image processing apparatus 40 according to a modified example 1 of Embodiment 2. Difference image (a) in Figure 20 shows a luminance distribution along a first direction, and difference image (b) shows a luminance distribution along a second direction. As shown in Figure 20, the setting unit 42 sets a one-dimensional kernel K12 in the difference image (a) between the captured image and a reference image of the captured image, when the relative orientation between the object and the imaging optical system is set to a predetermined direction. A region having luminance exceeding a predetermined threshold is called a defect region. The luminance distribution when the one-dimensional kernel K12 is applied to the luminance distribution along the first direction, when the plane of the difference image (a) containing the defect region is used as the reference image plane, is called the third luminance distribution. The imaging optical system has different numerical apertures (NA) in the first direction and in the second direction orthogonal to the first direction, when the captured image is used as the reference image plane. The imaging optical system may, in accordance with or instead of this, have different imaging magnifications in the first direction and in the second direction orthogonal to the first direction, when the captured image is used as the reference image plane.

[0129] The setting unit 42 sets a one-dimensional kernel K11 in the difference image (b) between the captured image and a reference image of the captured image, when the relative orientation of the object and the imaging optical system is 90° different from a predetermined orientation. The region corresponding to the defect region in the difference image (b) is called the rotation defect region. The luminance distribution obtained when the one-dimensional kernel K11 is applied to the luminance distribution along the second direction, with the plane of the difference image (b) containing the rotation defect region being the reference image plane, is called the fourth luminance distribution. The second direction is, for example, a direction 90° different from the first direction.

[0130] The setting unit 42 sets the one-dimensional kernel K11 and the one-dimensional kernel K12 so that the third luminance distribution and the fourth luminance distribution are similar to each other.

[0131] In this way, the setting unit 42 sets the one-dimensional kernel K11 and the one-dimensional kernel K12 so that the brightness distribution obtained when applying the one-dimensional kernel K12 to the brightness distribution along the first direction when the plane of the difference image is the reference image plane, for regions having brightness exceeding a predetermined threshold (hereinafter referred to as "defect regions") in the difference image between the captured image and the reference image of the captured image when the relative orientation of the object and the imaging optical system is set to a predetermined direction, and the brightness distribution obtained when applying the one-dimensional kernel K11 to the brightness distribution along the second direction, which is 90° different from the first direction when the plane of the difference image is the reference image plane, for the rotational defect regions, which are regions corresponding to the defect regions in the difference image between the captured image and the reference image of the captured image when the relative orientation of the object and the imaging optical system is set to a direction 90° different from the predetermined direction, are similar to each other. The defect regions may be referred to as "regions of interest" as appropriate.

[0132] Here, the difference in the brightness distribution of the defect region in the first and second directions in the difference images (a) and (b) can be attributed to differences in numerical aperture (NA) in the vertical and horizontal directions, differences in magnification and sensitivity, and the shape of the defect region (in Figure 20, it is elongated vertically in the α direction). Therefore, the setting unit 42 sets the image conversion parameters based on the brightness distribution in the first and second directions of the defect regions in the difference images, where the relative orientation of the object and the imaging optical system differs by 90°. This minimizes the difference in brightness distribution caused by the shape of the defect region and allows for the setting of appropriate image conversion parameters.

[0133] The setting unit 42 may set the one-dimensional kernels K11 and K12 such that the luminance distribution when the one-dimensional kernel K12 is applied to the luminance distribution along the first direction of the defect region, and the luminance distribution when the one-dimensional kernel K11 is applied to the luminance distribution along the second direction of the rotation defect region, are both similar to the luminance distribution of the reference shape. Alternatively, the setting unit 42 may set the one-dimensional kernels K11 and K12 without using the luminance distribution of the reference shape; for example, the one-dimensional kernels K11 and K12 may be set so that one is similar to the luminance distribution of the other. Or, the setting unit 42 may set the one-dimensional kernels K11 and K12 by changing at least one parameter of the one-dimensional kernels K11 and K12 such that the difference in luminance between each pixel in both luminance distributions is minimized when the corresponding one-dimensional kernels are applied to both luminance distributions.

[0134] The setting unit 42 may set the two-dimensional kernel K21 based on the one-dimensional kernels K11 and K12. Specifically, the setting unit 42 may set the two-dimensional kernel K21 from the product of the vector that is the one-dimensional kernel K11 and the vector that is the one-dimensional kernel K12. Alternatively, the setting unit 42 may set the two-dimensional kernel K21 from the cross product of the one-dimensional kernel K11 and the one-dimensional kernel K12. Furthermore, if the setting unit 42 can set the two-dimensional kernel K21 using the one-dimensional kernels K11 and K12, it may set the two-dimensional kernel K21 using calculation methods other than product or cross product.

[0135] In the setting section 42, the two-dimensional kernel K21, which is set based on the difference image, can be said to be an example of a kernel set considering the differences in NA, magnification, and sensitivity between the vertical and horizontal directions. In other words, it is an example of a kernel set to mitigate the differences in NA, magnification, and sensitivity between the vertical and horizontal directions.

[0136] <Processing that applies a kernel to the sample image> The processing unit 43 outputs a corrected sample image by applying a kernel W having a parameter distribution pattern set by the setting unit 42 to the sample image as a predetermined information processing step. That is, the processing unit 43 may output a corrected sample image by applying a two-dimensional kernel K21, which is an image conversion parameter set by the setting unit 42 based on an image obtained by comparing the captured image of the object with a reference image of the captured image, to the sample image. In this case, the two-dimensional kernel K21 is an example of a first kernel W, as it is the kernel W applied to the sample image.

[0137] As mentioned above, the sample 50, which is the subject of predetermined information processing such as the determination of whether or not there are defects, and the object may be the same object or different objects.

[0138] Furthermore, in this modified example 1, the object or sample 50 does not necessarily have a pattern extending in one direction. In this modified example 1, the image transformation parameters are set based on the defect region in the difference image, and here, the defect region in the difference image can occur regardless of the shape or type of the pattern.

[0139] <Modification 2 of Embodiment 2> <Processing that applies a kernel to the difference image> The processing unit 43 may, as a predetermined information processing step, apply a kernel having a distribution of parameters set by the setting unit 42 to the difference image and output a corrected difference image. The reference image may be a good image representing an image without abnormalities, or an image without critical abnormalities. The difference image may include the difference in brightness of pixels in the reference image and the corrected sample image that are at the same relative position as a parameter for each pixel. The difference image may also include the ratio of brightness of pixels in the reference image and the corrected sample image that are at the same relative position as a parameter for each pixel. Alternatively, the difference image may include the difference in brightness of pixels in the reference image and the sample image that are at the same relative position as a parameter for each pixel. The difference image may also include the ratio of brightness of pixels in the reference image and the sample image that are at the same relative position as a parameter for each pixel.

[0140] Here, the kernel to be applied to the difference image may be a kernel set by the setting unit 42 based on a first sample image region, which is a region including a first pattern extending in a first direction in the reference image plane, and a second sample image region, which is a region including a second pattern extending in a second direction orthogonal to the first direction in the reference image plane, as shown in Embodiment 2. Alternatively, the kernel to be applied to the difference image may be a kernel set by the setting unit 42 based on an image based on a comparison between a captured image of an object (or a sample) and a reference image of the captured image, as shown in Modified Example 1 of Embodiment 2. In this case, the object or the sample 50 may not have a pattern extending in one direction. The kernel to be applied to the difference image may be referred to as the second kernel W2.

[0141] When the reference image plane is placed in the difference image, the difference image may be an image composed of pixel columns arranged in a plurality in the first direction and pixel columns arranged in a plurality in the second direction. Here, it is assumed that the difference image is an image in which M pixels are arranged in the first direction and N pixels are arranged in the second direction. Then, the size of each cell of the second kernel W2 virtually corresponds to the size of the pixels of the difference image. Then, for the entire sample image described above, by the same process as when the first kernel W is applied, the processing unit 43 can acquire a corrected difference image by applying the second kernel W2 to the entire difference image.

[0142] The corrected difference image is, like the difference image, an image in which M pixels are arranged in the first direction and N pixels are arranged in the second direction. The luminance I''' m、n at each pixel (p''') m、n of the corrected difference image has been changed from the luminance I'' m、n at each pixel (p'') m、n of the difference image by the application of the second kernel W2. The corrected difference image is an image output by applying the second kernel W2, which has different parameter distributions in the first direction and the second direction, to the difference image. Therefore, the corrected difference image is an image in which the difference in sensitivity in the first direction and the second direction is alleviated compared to the difference image.

[0143] <Defect Inspection> The processing unit 43 detects defects based on the corrected difference image. The processing unit 43 may detect defects based on whether or not the brightness of the pixels included in the corrected difference image exceeds a predetermined threshold.

[0144] <Modification 3 of Embodiment 2> The processing unit 43 may, as a predetermined information processing step, acquire a corrected sample image by applying a first kernel W to the sample image, and also acquire a corrected difference image by applying a second kernel W2 to the difference image between the corrected sample image and a reference image (which may be a good product image corresponding to the sample image, or an image obtained by applying the first kernel W to the good product image). Then, the processing unit 43 may detect defects based on the corrected difference image.

[0145] Here, the processing unit 43 may obtain a modified difference image by applying the second kernel W2 to the difference image if the difference image between the modified sample image and the reference image contains a specific defect image. That is, if the processing unit 43 obtains a modified difference image by applying the second kernel W2 to the difference image if the difference image between the modified sample image and the reference image contains a specific defect image, it may evaluate the sample 50 based on the difference image if the difference image between the modified sample image and the reference image does not contain a specific defect image. The case in which a specific defect image is included may include, for example, the case in which pixels with a brightness exceeding a predetermined threshold exist over a predetermined number of regions in the difference image.

[0146] Here, the setting unit 42 may make the first kernel W and the second kernel W2 the same. The setting unit 42 may make the first kernel W and the second kernel W2 different. The setting unit 42 may set the first kernel W to have a parameter distribution that is asymmetric in the first direction and the second direction, based on a first sample image region which is a region containing a first pattern extending in a first direction in the reference image plane and a second sample image region which is a region containing a second pattern extending in a second direction orthogonal to the first direction in the reference image plane, as described above (Embodiment 2). Alternatively, the setting unit 42 may set the first kernel W to have a parameter distribution that is asymmetric in the first direction and the second direction, based on the defect region in the difference image between the captured image of the object (or sample 50) and the reference image (Modification 1 of Embodiment 2). On the other hand, the setting unit 42 may set the second kernel W2 by a method based on a different concept or purpose, as described later. In other words, the first kernel W is a kernel that is set considering the difference in sensitivity in the vertical and horizontal directions, while the second kernel W2 may be a kernel that is set for a different effect.

[0147] For example, the setting unit 42 may set the parameters of the second kernel W2 based on the transferability of defects in the sample, as follows: That is, the second kernel W2 may be set considering the transferability of foreign substances such as defects when exposed.

[0148] <How to configure the second kernel> For example, a sample image and an image of a substrate such as a wafer patterned using the sample as a patterning device (referred to as a substrate image) are prepared. The setting unit 42 then sets the parameters of kernel W so that the converted sample image obtained by applying kernel W to the sample image resembles the substrate image. When setting the parameters of kernel W, the setting unit 42 may set the parameters of kernel W based on a learning model that has been trained using multiple pairs of sample images and substrate images as training data. Alternatively, the setting unit 42 may use optical simulation to pre-derive the substrate image from the sample image and set kernel W such that equivalent results can be obtained. The setting unit 42 may also set the parameters of a second kernel W2 to be applied to the difference image based on the parameters of kernel W derived from the sample image in the above example. Such a kernel can be said to be an example of a kernel set considering the transferability when foreign matter is exposed.

[0149] <Prescribed information processing> The processing unit 43 performs a predetermined information processing, which involves adding a first kernel to the sample image. W A modified sample image to which the first kernel W has been applied may be obtained, and a modified difference image may be obtained by applying the second kernel W2 to the difference image between the reference image and the modified sample image. Then, the processing unit 43 may detect defects based on the modified difference image. Here, the first kernel W is a kernel set considering the difference in sensitivity in the vertical and horizontal directions, as explained above. Also, the second kernel W 2 As mentioned above, this is the second kernel W2, which is set considering the transferability of foreign substances during exposure.

[0150] Thus, when analyzing sample images, by applying both a kernel (first kernel W) that takes into account the difference in sensitivity in the vertical and horizontal directions, and a kernel (second kernel W2) that takes into account the transferability of foreign objects during exposure, it is possible to achieve appropriate inspection that takes into account the difference in sensitivity in the vertical and horizontal directions, as well as the transferability of foreign objects during exposure.

[0151] In the above explanation, the first kernel W applied to the sample image was set considering the difference in sensitivity in the vertical and horizontal directions, and the second kernel W2 applied to the difference image was set considering the transferability of foreign matter during exposure. However, this is not the only explanation. difference The second kernel W2 applied to the image may be set considering the difference in sensitivity in the vertical and horizontal directions, and the first kernel W applied to the sample image may be set considering the transferability of foreign matter during exposure.

[0152] Figure 21 is a flowchart illustrating an image processing method using the image processing device 40 according to Embodiment 2. As shown in Figure 21, the image processing method of this embodiment includes a step S21 for setting image conversion parameters, a step S22 for acquiring an image, and a step S23 for performing predetermined information processing. Steps S21 and S22 may be performed in reverse order or in parallel.

[0153] In step S21, the setting unit 42 sets image transformation parameters based on a first sample image region, which is a region containing a first pattern extending in a first direction in the reference image plane, and a second sample image region, which is a region containing a second pattern extending in a second direction orthogonal to the first direction in the reference image plane. The image transformation parameters are kernels having two-dimensional parameters corresponding to the first and second directions, and may include kernels in which the distribution of parameters is asymmetric between the first and second directions. In step S21, the setting unit 42 may also set image transformation parameters based on the transferability of defects in the sample.

[0154] Furthermore, the setting unit 42 may set image transformation parameters based on the following: the brightness distribution along a first direction when the plane of the comparison-based image is the reference image plane, in an image based on a comparison between an image of an object and a reference image corresponding to the image of the object, when the relative orientation of the object and the predetermined imaging optical system is a predetermined orientation; and the brightness distribution along a second direction when the plane of the comparison-based image is the reference image plane, in an image based on a comparison between an image of an object and a reference image corresponding to the image of the object, when the relative orientation of the object and the predetermined imaging optical system is approximately 90° different from the predetermined orientation, in a rotating area of ​​interest which is the area of ​​interest, in an image based on a comparison-based image which is the area of ​​interest.

[0155] In step S22, the acquisition unit 41 acquires an image. The acquisition unit 41 acquires, for example, a sample image, which is an image of the sample taken with a predetermined imaging optical system. The acquisition unit 41 may also acquire an image of the object taken with a predetermined imaging optical system in which the NA differs in a first direction and in a second direction perpendicular to the first direction in the reference image plane. The acquisition unit 41 may also acquire an image of the object taken with a predetermined imaging optical system in which the imaging magnification differs in a first direction and in a second direction perpendicular to the first direction in the reference image plane.

[0156] In step S23, the processing unit 43 performs predetermined information processing on the sample image or an image based on the sample image by applying image conversion parameters. The predetermined information processing may be a process of applying a kernel to the sample image, or a process of applying a kernel to the difference image. Alternatively, the predetermined information processing may be the above-described process using the modified sample image and the modified difference image.

[0157] While embodiments of this disclosure have been described above, this disclosure includes appropriate modifications that do not impair its purpose and advantages, and is not limited by the above embodiments. Furthermore, appropriate omissions and combinations of the components of Embodiment 1 are also within the scope of the technical concept of this disclosure. In addition, the following configurations are also within the scope of the technical concept of the embodiments.

[0158] (Note 1) A step of setting a first parameter for the first pattern based on a first sample image region which is a region containing a first pattern extending in a first direction in a reference image plane, A step of setting a second parameter for the second pattern based on a second sample image region which is a region in the reference image plane that includes a second pattern extending in a second direction orthogonal to the first direction, A step of acquiring a sample image, which is an image of a sample having a third pattern extending in the first direction and a fourth pattern extending in the second direction in the captured image, with the plane of the captured image being the reference image plane, using a predetermined imaging optical system; The steps include performing predetermined information processing on the sample image based on the first and second parameters, An image processing method equipped with [a specific feature]. (Note 2) In the step of setting the above first parameter, The first parameter is set to include a threshold used to determine the presence or absence of defects in the third pattern included in the sample image. In the step of setting the second parameter, The second parameter is set to include a threshold used to determine the presence or absence of defects in the fourth pattern included in the sample image, In the step of performing the predetermined information processing, The predetermined information processing includes determining whether or not there are defects in the third pattern and the fourth pattern included in the sample image. The image processing method described in Appendix 1. (Note 3) The sample image includes both the third pattern and the fourth pattern. In the step of performing the predetermined information processing, The predetermined information processing includes determining whether or not there are defects in the third pattern and the fourth pattern included in the sample image. The image processing method described in Appendix 2. (Note 4) In the steps of setting the first parameter and setting the second parameter, The threshold is set for multiple types of physical quantities relating to the information of the sample image. The image processing method described in Appendix 2. (Note 5) In the steps of setting the first parameter and setting the second parameter, The first parameter is set to be different from the second parameter, which is set to be different from the second parameter, which is set to be the type of physical quantity for which the threshold is determined and the number of physical quantities for which the threshold is determined. The image processing method described in Appendix 4. (Note 6) The step further comprises acquiring the first sample image region and the second sample image region, In the step of acquiring the first sample image region and the second sample image region, At least one of the first sample image region and the second sample image region is a region included in an image generated based on the design data of the sample or the design data of the pattern formed on the sample. The image processing method described in Appendix 1. (Note 7) The first sample image region is a region included in the image captured by the predetermined imaging optical system of an object having the first pattern. The image processing method described in Appendix 1. (Note 8) The first sample image region and the second sample image region are regions included in the image captured by the predetermined imaging optical system of an object having the first pattern and the second pattern. The image processing method described in Appendix 7. (Note 9) The first sample image region and the second sample image region are regions included in the image captured by the predetermined imaging optical system of an object having the first pattern, The orientation of the object with respect to the predetermined imaging optical system when capturing an image including the first sample image region, and the orientation of the object with respect to the predetermined imaging optical system when capturing an image including the second sample image region, differ by 90° in the main plane of the object. The image processing method described in Appendix 7. (Note 10) The predetermined imaging optical system captures an image of the object by oblique incident illumination, The first sample image region is included in the captured image when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the object and the first direction is a predetermined angle. The image processing method described in Appendix 7. (Note 11) The predetermined imaging optical system captures an image of the object by oblique incident illumination, The first sample image region and the second sample image region are included in the captured image when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the object and the first direction is a predetermined angle. The image processing method described in Appendix 9. (Note 12) The predetermined imaging optical system captures an image of the sample by oblique incident illumination, The sample image is an image captured by the predetermined imaging optical system when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the sample and the first direction is a predetermined angle. The image processing method described in Appendix 1. (Note 13) The aforementioned sample includes an anamorphic mask. The image processing method described in Appendix 12. (Note 14) In the step of acquiring the captured image, The sample image is obtained in which the number of pixels binned differs between the first direction and the second direction. The image processing method described in Appendix 1. (Note 15) In the step of acquiring the captured image, The captured image is obtained by an imaging optical system in which the NA differs between the first and second directions. The image processing method described in Appendix 1. (Note 16) In the step of setting the above first parameter, The first parameter is set to include a brightness correction value for pixels belonging to the third pattern included in the sample image, In the step of setting the second parameter, The second parameter is set to include a brightness correction value for pixels belonging to the fourth pattern included in the sample image, In the step of performing the predetermined information processing, The predetermined information processing, including image correction of the sample image, is performed. The image processing method described in Appendix 1. (Note 17) The steps include illuminating the aforementioned sample, The steps include: imaging the illuminated sample, The image processing method described in any one of the appendices 1 to 16, How to use an optical device equipped with this device. [Explanation of symbols]

[0159] 1. Inspection device 10 Illumination optical system 11 Light source 12 Ellipsoidal mirrors 13 Ellipsoidal mirror 14. Recessed mirror 20 Imaging optical system 21. Perforated concave mirror 21a hole 22 Convex mirror 23 Detectors 40 Image Processing Devices 41 Acquisition Department 42 Setting section 43 Processing Unit 43a Judgment part 43b Image Correction Unit Samples 50, 50a, and 50b 51 Top side 52 stages 71 Pattern 1 72. Pattern 2 73. Pattern 3 74. Pattern 4 G01, G02 Images G11 First Sample Image G12 2nd sample image G13 Image Capture IF1 Focus point K11, K12 one-dimensional kernels K21 2D kernel L11 Illumination Light L12 reflected light MP, MP1, MP2 Mask Pattern W First kernel W2 2nd kernel

Claims

1. Based on a first sample image region which is a region containing a first pattern extending in a first direction in the reference image plane, a first parameter is set for the first pattern. A setting unit sets a second parameter for the second pattern based on a second sample image region which is a region in the reference image plane that includes a second pattern extending in a second direction perpendicular to the first direction, An acquisition unit acquires a sample image which is an image of a sample having a third pattern extending in the first direction and a fourth pattern extending in the second direction in the captured image, with the plane of the captured image being the reference image plane, and the image being captured by a predetermined imaging optical system in which at least one of the magnification or NA differs between the first direction and the second direction. A processing unit performs predetermined information processing on the sample image based on the first parameter and the second parameter, Equipped with, The predetermined information processing includes an image processing apparatus that reduces the sensitivity difference between the first direction and the second direction, which occurs when at least one of the magnification or NA differs between the first direction and the second direction.

2. The setting unit is, The first parameter is set to include a threshold used to determine whether or not there are defects in the third pattern included in the sample image. The second parameter is set to include a threshold used to determine the presence or absence of defects in the fourth pattern included in the sample image. The processing unit performs the predetermined information processing, including determining whether or not the third pattern and the fourth pattern included in the sample image have defects. The image processing apparatus according to claim 1.

3. The sample image includes both the third pattern and the fourth pattern. The processing unit performs the predetermined information processing, including determining whether or not the third pattern and the fourth pattern included in the sample image have defects. The image processing apparatus according to claim 2.

4. The setting unit sets the threshold for multiple types of physical quantities relating to the information of the sample image. The image processing apparatus according to claim 2.

5. The setting unit sets at least one of the type of physical quantity and the number of physical quantities for which the threshold is determined as the first parameter, so as to be different from the type of physical quantity and the number of physical quantities for which the threshold is determined as the second parameter. The image processing apparatus according to claim 4.

6. The aforementioned physical quantities include line edge roughness, central luminance, average luminance, line width, and luminance difference, and are selected from this group. The image processing apparatus according to claim 5.

7. At least one of the first sample image region and the second sample image region is a region included in an image generated based on the design data of the sample or the design data of the pattern formed on the sample. The image processing apparatus according to claim 1.

8. The first sample image region is a region included in the image captured by the predetermined imaging optical system of an object having the first pattern. The image processing apparatus according to claim 1.

9. The first sample image region and the second sample image region are regions included in the image captured by the predetermined imaging optical system of an object having the first pattern and the second pattern. The image processing apparatus according to claim 8.

10. The first sample image region and the second sample image region are regions included in the image captured by the predetermined imaging optical system of the object having the first pattern, The orientation of the object with respect to the predetermined imaging optical system when capturing the image including the first sample image region, and the orientation of the object with respect to the predetermined imaging optical system when capturing the image including the second sample image region, differ by 90° in the main plane of the object. The image processing apparatus according to claim 8.

11. The predetermined imaging optical system captures an image of the object by oblique incident illumination, The first sample image region is included in the captured image when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the object and the first direction is a predetermined angle. The image processing apparatus according to claim 8.

12. The predetermined imaging optical system captures an image of the object by oblique incident illumination, The first sample image region and the second sample image region are included in the captured image when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the object and the first direction is a predetermined angle. The image processing apparatus according to claim 9.

13. The predetermined imaging optical system captures an image of the sample by oblique incident illumination, The sample image is an image captured by the predetermined imaging optical system when the angle between the direction in which the principal axis of the oblique incident illumination is projected onto the upper surface of the sample and the first direction is a predetermined angle. The image processing apparatus according to claim 1.

14. The aforementioned sample includes an anamorphic mask. The image processing apparatus according to claim 13.

15. The acquisition unit acquires the sample image in which the number of pixels to be binned differs between the first direction and the second direction. The image processing apparatus according to claim 1.

16. The predetermined imaging optical system has different numerical apertures (NA) in the first direction and the second direction. The image processing apparatus according to claim 1.

17. The setting unit is, The first parameter is set to include a brightness correction value for pixels belonging to the third pattern included in the sample image, The second parameter is set to include a brightness correction value for pixels belonging to the fourth pattern included in the sample image, The processing unit performs the predetermined information processing, including image correction of the sample image. The image processing apparatus according to claim 1.

18. A setting unit sets image conversion parameters based on a first sample image region which is a region containing a first pattern extending in a first direction in the reference image plane, and a second sample image region which is a region containing a second pattern extending in a second direction perpendicular to the first direction in the reference image plane. An acquisition unit that acquires a sample image, which is an image of a sample taken with a predetermined imaging optical system in which the magnification or NA differs in at least one of the first direction and the second direction; A processing unit that reduces the sensitivity difference between the first and second directions caused by a difference in magnification or NA between the first and second directions, by performing predetermined information processing on the sample image or an image based on the sample image and applying the image conversion parameters; Equipped with, The image transformation parameters are parameters corresponding to the luminance distributions in the first and second directions, respectively.

19. An acquisition unit that acquires an image of an object captured by a predetermined imaging optical system in which the magnification or NA differs in at least one of the first direction and a second direction perpendicular to the first direction in the reference image plane, A setting unit sets image conversion parameters based on the following: the brightness distribution along the first direction when the plane of the comparison-based image is the reference image plane, in an image based on a comparison between an image of the object and a reference image corresponding to the image when the relative orientation of the object and the predetermined imaging optical system is a predetermined direction; and the brightness distribution along the second direction when the plane of the comparison-based image is the reference image plane, in an image based on a comparison between an image of the object and a reference image corresponding to the image when the relative orientation of the object and the predetermined imaging optical system is approximately 90° different from the predetermined direction, in an image that is a rotating area of ​​interest corresponding to the area of ​​interest, in an image based on a comparison-based image the reference image plane, in a plane of the comparison-based image. A processing unit that performs predetermined information processing by applying the image conversion parameters to a sample image which is an image of a sample captured by the predetermined imaging optical system, or an image based on the sample image, Equipped with an image processing device.

20. The parameter is a kernel having two-dimensional parameters corresponding to the first direction and the second direction. The image processing apparatus according to claim 18.

21. The image transformation parameters include a kernel having two-dimensional parameters corresponding to the first and second directions, wherein the distribution of the parameters is asymmetric between the first and second directions. The image processing apparatus according to claim 19.

22. An illumination optical system for illuminating the aforementioned sample, A predetermined imaging optical system for imaging the illuminated sample, An image processing apparatus according to any one of claims 1 to 19, An optical device equipped with an optical system.

23. A step of setting a first parameter for the first pattern based on a first sample image region which is a region containing a first pattern extending in a first direction in a reference image plane, A step of setting a second parameter for the second pattern based on a second sample image region which is a region in the reference image plane that includes a second pattern extending in a second direction perpendicular to the first direction, Steps include acquiring a sample image, which is an image of a sample having a third pattern extending in the first direction and a fourth pattern extending in the second direction in the captured image, with the plane of the captured image being the reference image plane, using a predetermined imaging optical system in which at least one of the magnification or NA differs between the first and second directions; The steps include: performing predetermined information processing on the sample image based on the first and second parameters; Equipped with, The predetermined information processing is an image processing method that includes processing to reduce the sensitivity difference between the first direction and the second direction, which occurs when at least one of the magnification or NA differs between the first direction and the second direction.

24. The steps of setting image transformation parameters based on a first sample image region which is a region containing a first pattern extending in a first direction in the reference image plane, and a second sample image region which is a region containing a second pattern extending in a second direction perpendicular to the first direction in the reference image plane, The steps include acquiring a sample image, which is an image of the sample taken with a predetermined imaging optical system in which the magnification or NA differs in at least one of the first direction and the second direction, The steps include: applying the image conversion parameters to the sample image or an image based on the sample image to perform predetermined information processing to reduce the sensitivity difference between the first and second directions caused by at least one of the magnification or NA being different in the first and second directions; Equipped with, The image processing method includes an image transformation parameter having parameters corresponding to the luminance distribution in the first and second directions.

25. The steps include acquiring an image of an object by imaging it with a predetermined imaging optical system in which the magnification or NA differs in at least one of the first direction and the second direction perpendicular to the first direction in the reference image plane, A step of setting image transformation parameters based on: the brightness distribution along the first direction when the plane of the comparison-based image is the reference image plane, of the region of interest in the image based on a comparison between the captured image of the object and a reference image corresponding to the captured image, when the relative orientation of the object and the predetermined imaging optical system is a predetermined direction; and the brightness distribution along the second direction when the plane of the comparison-based image is the reference image plane, of the rotating region of interest, which is the region of interest, in the image based on a comparison between the captured image of the object and a reference image corresponding to the captured image, when the relative orientation of the object and the predetermined imaging optical system is approximately 90° different from the predetermined direction; The steps include: applying the image conversion parameters to a sample image, which is an image of a sample captured by the predetermined imaging optical system, or an image based on the sample image, and performing predetermined information processing on the sample image; An image processing method equipped with [a specific feature].

26. The parameter is a kernel having two-dimensional parameters corresponding to the first direction and the second direction. The image processing method according to claim 24.

27. The image transformation parameters include a kernel having two-dimensional parameters corresponding to the first and second directions, wherein the distribution of the parameters is asymmetric between the first and second directions. The image processing method according to claim 25.

28. The steps include illuminating the aforementioned sample, The steps include: imaging the illuminated sample, The image processing method according to any one of claims 23 to 25, How to use an optical device equipped with this device.

Images