Inspection system, inspection device, inspection method, and inspection program

Abstract

To more accurately detect a defect of an object.SOLUTION: An inspection device comprises: an acquisition part which acquires measurement results of a distribution of physical properties in a plane of an object having a plane part; a frequency distribution generation part which calculates standard deviations or variation coefficients of the physical properties in a plurality of first unit regions in the plane based upon the measurement results acquired by the acquisition part, and generates a frequency distribution of the standard deviations or the variation coefficients; and an extraction part which extracts some grades of the frequency distribution generated by the frequency distribution generation part.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to an inspection system, an inspection device, an inspection method, and an inspection program.

Background Art

[0002] Conventionally, technologies for inspecting objects such as films have been developed. For example, Patent Document 1 (Japanese Unexamined Patent Application Publication No. 2023-5503) discloses the following inspection method. That is, the inspection method is an inspection method for detecting the presence of defects in a long film that is continuously conveyed, and the physical properties that vary corresponding to the defects in the film are measured at a plurality of measurement points Po over a certain direction L in the plane of the film and a direction W in the plane different from the direction L, and a step of obtaining a data set composed of a plurality of sets of measurement data related to the plurality of measurement points Po, wherein each set of the measurement data includes coordinate information along the direction L and the direction W and information on the measured value of the physical property related to each of the measurement points Po (step (I)), a step of classifying the data of the data set corresponding to a plurality of regions R of the film and constructing a subset of the data for each of the regions R (step (II)), for each of the subsets, obtaining an arithmetic mean value AvC of the measured values at a measurement point group PoC that is aligned along a straight line in a certain direction C in the plane of the film among the plurality of measurement points Po, and constructing a profile E of the relationship between the distance DsD from a certain reference position along a direction D in the plane different from the direction C to the measurement point group PoC and the arithmetic mean value AvC (step (III)), and a step of detecting the presence of the defect based on the profile E (step (IV)).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Beyond the technology described in Patent Document 1, a technology capable of more accurately detecting defects in an object is desired.

[0005] The present invention has been made to solve the above-described problems, and an object thereof is to provide an inspection system, an inspection apparatus, an inspection method, and an inspection program capable of more accurately detecting defects in an object.

Means for Solving the Problems

[0006] (1) In order to solve the above problems, an inspection system according to an aspect of the present invention includes an acquisition unit that acquires a measurement result of the distribution of physical properties in a plane of an object having a planar portion, and based on the measurement result acquired by the acquisition unit, calculates a standard deviation or a coefficient of variation of the physical properties in a plurality of first unit regions in the plane, and a frequency distribution creation unit that creates a frequency distribution of the standard deviation or a frequency distribution of the coefficient of variation, and an extraction unit that extracts a part of classes in the frequency distribution created by the frequency distribution creation unit.

[0007] In this way, by creating a frequency distribution of the standard deviation or the coefficient of variation of the physical properties in a plurality of first unit regions in the plane of the object and extracting a part of classes in the frequency distribution, based on the created frequency distribution, for example, excluding the standard deviation or the coefficient of variation of normal physical properties, the standard deviation or the coefficient of variation of the physical properties derived from defects is extracted, and based on the extracted standard deviation or coefficient of variation, defects can be accurately detected. Therefore, defects in the object can be detected more accurately.

[0008] (2) In the above (1), the inspection system may further include a calculation unit that calculates a statistical value of the standard deviation or a statistical value of the coefficient of variation of the part of classes extracted by the extraction unit.

[0009] With such a configuration, the occurrence state of defects in the object can be normalized using the statistical value, so that the occurrence state of defects in the object can be easily evaluated using the statistical value.

[0010] (3) In the above (1) or (2), the inspection system may further include an image display unit that creates an image showing a two-dimensional distribution of the standard deviation or the coefficient of variation of the partial class in the plane based on the partial class extracted by the extraction unit, and performs a process of displaying the created image.

[0011] With such a configuration, it is possible to easily make the user recognize the location where defects exist in the plane of the object.

[0012] (4) In any of the above (1) to (3), the extraction unit may extract a plurality of different partial classes.

[0013] With such a configuration, it is possible to confirm the validity of the extraction range of the class in the frequency distribution based on the two-dimensional distribution of the standard deviation or the coefficient of variation corresponding to each of the plurality of extracted partial classes.

[0014] (5) In any of the above (1) to (4), the inspection system may further include a reception unit that receives the setting of the size of the first unit region, and the frequency distribution creation unit may calculate the standard deviation or the coefficient of variation of the physical property in the first unit region having the size received by the reception unit.

[0015] With such a configuration, it is possible to calculate the standard deviation or the coefficient of variation of the physical property in the first unit region of an appropriate size according to the type of defect that may occur in the object.

[0016] (6) In any of the above (1) to (5), the frequency distribution creation unit may create a plurality of frequency distributions corresponding to a plurality of the first unit regions having different sizes.

[0017] With such a configuration, based on the classes extracted from the plurality of frequency distributions created, it is possible to detect a plurality of types of defects that can occur in the object.

[0018] (7) In any one of (1) to (6) above, the inspection system may further include a noise processing unit that converts the distribution of the physical property into a distribution of the median or average value of the physical property in a plurality of second unit regions within the plane, and the frequency distribution creation unit may calculate the standard deviation or the coefficient of variation based on the distribution of the median or the average value converted by the noise processing unit, and the size of the second unit region may be equal to or smaller than the size of the first unit region.

[0019] With such a configuration, the noise of the physical property of the object can be reduced, so that the defects of the object can be detected more accurately.

[0020] (8) In any one of (1) to (7) above, the frequency distribution creation unit may create a histogram showing the frequency distribution, and the inspection system may further include a histogram display unit that performs a process of displaying the histogram created by the frequency distribution creation unit.

[0021] With such a configuration, the variation in the standard deviation or the coefficient of variation of the physical property within the plane of the object can be visually recognized by the user.

[0022] (9) In order to solve the above problems, an inspection system according to an aspect of the present invention includes an acquisition unit that acquires a measurement result of the distribution of a physical property within the plane of an object having a planar portion, and based on the measurement result acquired by the acquisition unit, generates an image showing a two-dimensional distribution of the standard deviation or the coefficient of variation within a partial numerical range of the standard deviation or the coefficient of variation of the physical property in a plurality of first unit regions within the plane, and an image display unit that performs a process of displaying the generated image.

[0023] In this way, by generating and displaying an image showing the two-dimensional distribution of the standard deviation or coefficient of variation within a partial numerical range among the standard deviations or coefficients of variation of physical properties in a plurality of first unit regions in the plane of the object, for example, an image with defects emphasized, which shows the two-dimensional distribution of some standard deviations or coefficients of variation resulting from defects, can be displayed, and thus defects can be detected based on the image. Therefore, defects in the object can be detected more accurately.

[0024] (10) To solve the above problems, an inspection method according to an aspect of the present invention is an inspection method in an inspection system, including steps of: obtaining a measurement result of the distribution of physical properties in the plane of an object having a planar portion; calculating the standard deviation or coefficient of variation of the physical properties in a plurality of first unit regions in the plane based on the obtained measurement result, and creating a frequency distribution of the standard deviation or a frequency distribution of the coefficient of variation; and extracting a part of classes in the created frequency distribution.

[0025] In this way, by creating a frequency distribution of the standard deviation or coefficient of variation of physical properties in a plurality of first unit regions in the plane of the object and extracting a part of classes in the frequency distribution, based on the created frequency distribution, for example, the standard deviation or coefficient of variation of physical properties resulting from defects excluding the standard deviation or coefficient of variation of normal physical properties can be extracted, and defects can be accurately detected based on the extracted standard deviation or coefficient of variation. Therefore, defects in the object can be detected more accurately.

[0026] (11) To solve the above problems, an inspection program according to an aspect of the present invention is an inspection program used in an inspection apparatus, which causes a computer to function as an acquisition unit that acquires measurement results of the distribution of physical properties in the plane of an object having a flat portion, a frequency distribution creation unit that calculates the standard deviation or coefficient of variation of the physical properties in a plurality of first unit regions in the plane based on the measurement results acquired by the acquisition unit and creates a frequency distribution of the standard deviation or a frequency distribution of the coefficient of variation, and an extraction unit that extracts a part of the classes in the frequency distribution created by the frequency distribution creation unit.

[0027] In this way, by creating a frequency distribution of the standard deviation or coefficient of variation of the physical properties in a plurality of first unit regions in the plane of the object and extracting a part of the classes in the frequency distribution, based on the created frequency distribution, for example, excluding the standard deviation or coefficient of variation of normal physical properties, the standard deviation or coefficient of variation of the physical properties derived from defects can be extracted, and based on the extracted standard deviation or coefficient of variation, defects can be accurately detected. Therefore, defects of the object can be detected more accurately.

Advantages of the Invention

[0028] According to the present invention, defects of an object can be detected more accurately.

Brief Description of the Drawings

[0029]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Embodiments for Carrying Out the Invention

[0030] Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals and their description will not be repeated. Also, at least a part of the embodiments described below may be arbitrarily combined.

[0031] [Configuration and Basic Operation] FIG. 1 is a diagram showing the configuration of an inspection system according to an embodiment of the present invention. Referring to FIG. 1, the inspection system 201 includes an inspection apparatus 101, a measurement apparatus 111, and a display apparatus 121. The inspection system 201 is a system for inspecting an object F having a planar portion Sf. For example, the shape of the object F is film-like or plate-like. As an example, the object F is an optical film. Note that the object F only needs to have a planar portion Sf, and in addition to the planar portion Sf, it may have uneven portions provided with wiring patterns or the like.

[0032] The measurement apparatus 111 measures the distribution of physical properties in the plane of the object F. For example, the measurement apparatus 111 measures the distribution of the film thickness tc in the plane of the planar portion Sf as the distribution of physical properties in the plane of the object F. For example, the measurement apparatus 111 is a non-contact film thickness measurement apparatus such as a spectroscopic film thickness measurement apparatus. Note that the measurement apparatus 111 may be a contact-type film thickness measurement apparatus.

[0033] FIG. 2 is a diagram showing an example of measuring the film thickness of an object by the measurement apparatus according to the embodiment of the present invention. The broken line in FIG. 2 is a virtual line indicating the measurement point P in the plane of the planar portion Sf. Referring to FIG. 2, the measurement apparatus 111 measures the film thickness tc at a plurality of matrix-shaped measurement points P in the plane of the planar portion Sf.

[0034] FIG. 3 is a diagram schematically showing a film thickness distribution image created by the measuring apparatus according to the embodiment of the present invention. When the measuring apparatus 111 measures the distribution of the film thickness tc in the plane of the planar portion Sf, it creates a film thickness distribution image G1 showing the correspondence between the position of the measurement point P in the plane of the planar portion Sf and the film thickness tc at the measurement point P. In the film thickness distribution image G1, the position of the measurement point P is represented by an X coordinate indicating the position in the X direction in the plane of the object F and a Y coordinate indicating the position in the Y direction in the plane of the object F. Here, the Y direction is a direction orthogonal to the X direction. In the film thickness distribution image G1, the film thickness tc at the measurement point P is represented by shading. The measuring apparatus 111 transmits the created film thickness distribution image G1 to the inspection apparatus 101.

[0035] FIG. 4 is a diagram showing the configuration of the inspection apparatus according to the embodiment of the present invention. Referring to FIG. 4, the inspection apparatus 101 includes an acquisition unit 11, a noise processing unit 12, a creation unit 13, an extraction unit 14, a display processing unit 15, and a calculation unit 16. The creation unit 13 is an example of a frequency distribution creation unit and also an example of a reception unit. The display processing unit 15 is an example of an image display unit and also an example of a histogram display unit. A part or all of the acquisition unit 11, the noise processing unit 12, the creation unit 13, the extraction unit 14, the display processing unit 15, and the calculation unit 16 are realized by, for example, a processing circuit (Circuitry) including one or a plurality of processors.

[0036] The acquisition unit 11 acquires the measurement result of the distribution of physical properties in the plane of the object F. More specifically, the acquisition unit 11 receives from the measuring apparatus 111 a film thickness distribution image G1 showing the measurement result of the distribution of the film thickness tc. The acquisition unit 11 outputs the received film thickness distribution image G1 to the noise processing unit 12.

[0037] The noise processing unit 12 receives the film thickness distribution image G1 from the acquisition unit 11 and performs noise removal processing on the received film thickness distribution image G1.

[0038] FIG. 5 is a diagram showing an example of noise removal processing by a noise processing unit in an inspection apparatus according to an embodiment of the present invention. Referring to FIG. 5, in the noise removal processing, the noise processing unit 12 converts the distribution of the film thickness tc shown in the film thickness distribution image G1 into the median value M of the film thickness tc in a plurality of unit regions R2 in the plane of the object F. The unit region R2 is an example of the second unit region.

[0039] More specifically, the noise processing unit 12 performs a convolution process on the film thickness distribution image G1 using a median filter Mx corresponding to the unit region R2, thereby creating a film thickness distribution image G2 showing the correspondence between the position of the measurement point P in the plane of the object F and the median value M at the measurement point P. For example, the noise processing unit 12 performs a padding process of adding data of the film thickness tc around the film thickness distribution image G1, and performs a convolution process on the film thickness distribution image G1 after the padding process to create the film thickness distribution image G2. Thereby, a film thickness distribution image G2 having the same size as the film thickness distribution image G1 can be created. The noise processing unit 12 creates a plurality of film thickness distribution images G2 respectively corresponding to the size settings of a plurality of different unit regions R2 according to the types of defects to be detected.

[0040] FIGS. 6 and 7 are diagrams schematically showing film thickness distribution images created by a noise processing unit in an inspection apparatus according to an embodiment of the present invention.

[0041] Referring to FIG. 6, the noise processing unit 12 performs a convolution process on the film thickness distribution image G1 using a median filter M1 corresponding to the unit region R2a which is the unit region R2, thereby creating a film thickness distribution image G2A which is a film thickness distribution image G2 showing the correspondence between the position of the measurement point P in the plane of the object F and the median value Ma at the measurement point P. In the film thickness distribution image G2A, the median value Ma at the measurement point P is represented by shading. The median value Ma is an example of the median value M. The unit region R2a has a size corresponding to (m×n) measurement points P of m rows and n columns. m and n are positive integers. m and n may be the same value or different values.

[0042] m and n are preferably odd numbers, for example, "3". In this case, the noise processing unit 12 creates a film thickness distribution image G2A by performing convolution processing on the film thickness distribution image G1 using a 3×3 median filter M1.

[0043] Referring to FIG. 7, the noise processing unit 12 creates a film thickness distribution image G2B, which is a film thickness distribution image G2 showing the correspondence between the position of the measurement point P in the plane of the object F and the median value Mb at the measurement point P, by performing convolution processing on the film thickness distribution image G1 using a median filter M2 corresponding to the unit region R2b, which is the unit region R2. In the film thickness distribution image G2B, the median value Mb at the measurement point P is represented by shading. The median value Mb is an example of the median value M. The unit region R2b has a size corresponding to (p×q) measurement points P of p rows and q columns. p and q may be the same value or different values.

[0044] p and q are preferably odd numbers, for example, "5". In this case, the noise processing unit 12 creates a film thickness distribution image G2A by performing convolution processing on the film thickness distribution image G1 using a 5×5 median filter M2.

[0045] When the noise processing unit 12 creates the film thickness distribution images G2A and G2B, it outputs the created film thickness distribution images G2A and G2B to the creation unit 13.

[0046] The creation unit 13 calculates the standard deviation Sd of the film thickness tc in a plurality of unit regions R1 in the plane of the object F based on the measurement results obtained by the acquisition unit 11. The unit region R1 is an example of the first unit region. For example, the creation unit 13 receives the setting of the size of the unit region R1 from the user of the inspection device 101. The creation unit 13 calculates the standard deviation Sd of the film thickness tc in the unit region R1 of the received size. For example, the size of the unit region R1 is equal to or greater than the size of the unit region R2. Note that the size of the unit region R1 may be less than the size of the unit region R2. Hereinafter, it is assumed that the sizes of the unit regions R1 and R2 are the same.

[0047] For example, the creation unit 13 calculates the standard deviation Sd based on the distribution of the median M of the film thickness tc converted by the noise processing unit 12. More specifically, the creation unit 13 receives the film thickness distribution images G2A and G2B from the noise processing unit 12 and performs unevenness extraction processing on the received film thickness distribution images G2A and G2B.

[0048] FIG. 8 is a diagram showing an example of unevenness extraction processing by the creation unit in the inspection apparatus according to the embodiment of the present invention. Referring to FIG. 8, in the unevenness extraction processing, the creation unit 13 calculates the standard deviation SdA, which is the standard deviation of the median Ma in a plurality of unit regions R1 in the plane of the object F, based on the median Ma indicated by the film thickness distribution image G2A. Further, in the unevenness extraction processing, the creation unit 13 calculates the standard deviation SdB, which is the standard deviation of the median Mb in a plurality of unit regions R1 in the plane of the object F, based on the median Mb indicated by the film thickness distribution image G2B.

[0049] More specifically, the creation unit 13 performs a convolution process on the film thickness distribution image G2 using a standard deviation filter Fx corresponding to the unit region R1, thereby creating a film thickness distribution image G3 showing the correspondence between the position of the measurement point P in the plane of the object F and the standard deviation Sd at the measurement point P. For example, the creation unit 13 performs a padding process of adding data of the median M around the film thickness distribution image G2, and creates the film thickness distribution image G3 by performing a convolution process on the padded film thickness distribution image G2. Thereby, a film thickness distribution image G3 having the same size as the film thickness distribution image G2 can be created.

[0050] FIGS. 9 and 10 are diagrams schematically showing the film thickness distribution images created by the creation unit in the inspection apparatus according to the embodiment of the present invention.

[0051] Referring to FIG. 9, for example, in the non-uniformity extraction process, the creation unit 13 performs a convolution process on the film thickness distribution image G2A received from the noise processing unit 12 using the standard deviation filter F1 corresponding to the unit region R1a which is the unit region R1, thereby creating a film thickness distribution image G3A which is a film thickness distribution image G3 showing the correspondence between the position of the measurement point P in the plane of the object F and the standard deviation SdA at the measurement point P. In the film thickness distribution image G3A, the standard deviation SdA at the measurement point P is represented by shading. The unit region R1a has a size corresponding to (m×n) measurement points P in m rows and n columns, similar to the unit region R2a.

[0052] Referring to FIG. 10, for example, in the non-uniformity extraction process, the creation unit 13 performs a convolution process on the film thickness distribution image G2B received from the noise processing unit 12 using the standard deviation filter F2 corresponding to the unit region R1b which is the unit region R1, thereby creating a film thickness distribution image G3B which is a film thickness distribution image G3 showing the correspondence between the position of the measurement point P in the plane of the object F and the standard deviation SdB at the measurement point P. In the film thickness distribution image G3B, the standard deviation SdB at the measurement point P is represented by shading. The unit region R1b has a size corresponding to (p×q) measurement points P in p rows and q columns, similar to the unit region R2b.

[0053] The creation unit 13 creates a frequency distribution Fd of the standard deviation Sd. For example, the creation unit 13 creates a plurality of frequency distributions Fd respectively corresponding to a plurality of unit regions R1 having different sizes. More specifically, the creation unit 13 creates a frequency distribution FdA which is a frequency distribution in which the standard deviation SdA of each measurement point P shown in the film thickness distribution image G3A is classified into a predetermined number of classes CsA, and a frequency distribution FdB which is a frequency distribution in which the standard deviation SdB of each measurement point P shown in the film thickness distribution image G3B is classified into a predetermined number of classes CsB.

[0054] FIGS. 11 and 12 are diagrams showing an example of a histogram created by the creation unit in the inspection apparatus according to the embodiment of the present invention.

[0055] Referring to FIGS. 11 and 12, the creation unit 13 creates a histogram HgA showing the frequency distribution FdA of the standard deviation SdA and a histogram HgB showing the frequency distribution FdB of the standard deviation SdB. Here, when there are no defects in the object F, the histograms HgA and HgB created by the creation unit 13 have a distribution close to a normal distribution. Further, when there are no defects in the object F and the film thickness tc at each measurement point P is uniform, the standard deviations SdA and SdB at each measurement point P become zero. In this case, the histograms HgA and HgB created by the creation unit 13 consist of a single class and are in a state where there is so-called no distribution.

[0056] On the other hand, when there are defects in the object F, the histograms HgA and HgB created by the creation unit 13 have a distorted distribution, which is different from the histograms HgA and HgB created by the creation unit 13 when there are no defects. The creation unit 13 outputs the created histograms HgA and HgB to the display processing unit 15.

[0057] The display processing unit 15 performs a process of displaying the histograms HgA and HgB created by the creation unit 13. More specifically, the display processing unit 15 performs a process of displaying the histograms HgA and HgB received from the creation unit 13 on the display device 121.

[0058] Further, the creation unit 13 outputs the created frequency distributions FdA and FdB and histograms HgA and HgB to the extraction unit 14 and the calculation unit 16.

[0059] The extraction unit 14 extracts some classes CsA and CsB in the frequency distributions FdA and FdB created by the creation unit 13.

[0060] For example, the extraction unit 14 identifies the maximum value VAmax among the standard deviations SdA of each measurement point P and the interquartile range QrA of the standard deviation SdA. Here, the interquartile range QrA is a value obtained by subtracting the first quartile QA2 of the standard deviation SdA from the third quartile QA1 of the standard deviation SdA. The extraction unit 14 calculates a value VAq by adding a value obtained by multiplying the interquartile range QrA by a predetermined coefficient to the third quartile QA1. Then, the extraction unit 14 determines the presence or absence of a defect in the object F based on the comparison result between the value VAmax, the value VAq, and a predetermined threshold value VAth. When the extraction unit 14 determines that a defect exists in the object F, it extracts some classes CsA in the frequency distribution FdA.

[0061] More specifically, as shown in FIG. 11, the extraction unit 14 determines that a defect exists in the object F when the values VAmax, VAq, and the threshold value VAth satisfy the following formula (1). VAth ≦ VAq < VAmax ··· (1) In this case, the extraction unit 14 extracts one or more classes CsA in the range RA from the class CsA1, which is the class CsA to which the value VAq belongs, to the class CsA2, which is the class CsA to which the value VAmax belongs, among the plurality of classes CsA in the frequency distribution FdA.

[0062] Also, the extraction unit 14 determines that a defect exists in the object F when the values VAmax, VAq, and the threshold value VAth satisfy the following formula (2) or (3). VAth ≦ VAmax < VAq ··· (2) VAq < VAth ≦ VAmax ··· (3) In this case, the extraction unit 14 extracts one or more classes CsA in the range RA from the class CsA3, which is the class CsA to which the threshold value VAth belongs, to the class CsA2, which is the class CsA to which the value VAmax belongs, among the plurality of classes CsA in the frequency distribution FdA.

[0063] On the other hand, the extraction unit 14 determines that no defect exists in the object F when the values VAmax, VAq, and the threshold value VAth satisfy the following formula (4), (5), or (6). VAmax < VAth ≤ VAq ··· (4) VAq < VAmax ≤ VAth ··· (5) VAmax < VAq ≤ VAth ··· (6) In this case, the extraction unit 14 does not perform the extraction of the class CsA.

[0064] Also, for example, the extraction unit 14 identifies the maximum value VBmax among the standard deviations SdB of each measurement point P and the third quartile QB1 of the standard deviation SdA. Then, the extraction unit 14 determines the presence or absence of a defect in the object F based on the comparison result between the value VBmax, the third quartile QB1, and a predetermined threshold value VBth. When the extraction unit 14 determines that a defect exists in the object F, it extracts some of the classes CsB in the frequency distribution FdB.

[0065] More specifically, as shown in FIG. 12, when the value VBmax, the third quartile QB1, and the threshold value VBth satisfy the following formula (7), the extraction unit 14 determines that a defect exists in the object F. VBth ≤ QB1 < VBmax ··· (7) In this case, the extraction unit 14 extracts one or more classes CsB in the range RB from the class CsB1, which is the class CsA to which the third quartile QB1 belongs, to the class CsB2, which is the class CsB to which the value VBmax belongs, among the plurality of classes CsB in the frequency distribution FdB.

[0066] Also, when the value VBmax, the third quartile QB1, and the threshold value VBth satisfy the following formula (8), the extraction unit 14 determines that a defect exists in the object F. QB1 < VBth ≤ VBmax ··· (8) In this case, the extraction unit 14 extracts one or more classes CsB in the range RB from the class CsB3, which is the class CsA to which the threshold value VBth belongs, to the class CsB2, which is the class CsB to which the value VBmax belongs, among the plurality of classes CsB in the frequency distribution FdB.

[0067] On the other hand, when the value VBmax, the third quartile QB1, and the threshold value VBth satisfy the following formula (9), the extraction unit 14 determines that there is no defect in the object F. QB1 < VBmax ≦ VBth ··· (9) In this case, the extraction unit 14 does not perform the extraction of the class CsB.

[0068] Note that the extraction unit 14 may extract a plurality of different partial classes CsA. Also, the extraction unit 14 may extract classes CsA in a plurality of non-overlapping ranges. Similarly, the extraction unit 14 may extract a plurality of different partial classes CsB. Further, the extraction unit 14 may determine the presence or absence of a defect in the object F using other values such as the 74th percentile, 76th percentile, and 70th percentile instead of the third quartile QB1, or the value used instead of the third quartile QB1 may be settable. Also, the extraction unit 14 may determine the presence or absence of a defect in the object F using other values such as the 74th percentile, 76th percentile, and 70th percentile instead of the third quartile QA1, or the value used instead of the third quartile QA1 may be settable.

[0069] When the extraction unit 14 determines that there is a defect in the object F and performs the extraction of the class CsA, the extracted class CsA is output to the display processing unit 15 and the calculation unit 16. Also, when the extraction unit 14 determines that there is a defect in the object F and performs the extraction of the class CsB, the extracted class CsB is output to the display processing unit 15 and the calculation unit 16.

[0070] Based on some of the classes CsA and CsB extracted by the extraction unit 14, the display processing unit 15 creates an image showing the two-dimensional distribution of the standard deviations SdA and SdB of these some classes CsA and CsB in the plane of the object F.

[0071] FIG. 13 and FIG. 14 are diagrams schematically showing images created by the display processing unit in the inspection apparatus according to the embodiment of the present invention.

[0072] Referring to FIG. 13, based on the class CsA received from the extraction unit 14, the display processing unit 15 creates a film thickness distribution image G4A showing the correspondence between the position of the measurement point P in the plane of the object F and the standard deviation SdA of the class CsA at the measurement point P. In the film thickness distribution image G4A, the standard deviation SdA at the measurement point P is represented by shading. In the film thickness distribution image G4A, defects such as the point defect D1 where the inclination of the film thickness tc in the plane of the object F becomes a predetermined value or more are emphasized. The display processing unit 15 performs a process of displaying the generated film thickness distribution image G4A. For example, as shown in FIG. 13, the display processing unit 15 performs a process of displaying the film thickness distribution image G4A without performing gain adjustment such as adjusting the shading of the generated film thickness distribution image G4A.

[0073] Referring to FIG. 14, based on the class CsB received from the extraction unit 14, the display processing unit 15 creates a film thickness distribution image G4B showing the correspondence between the position of the measurement point P in the plane of the object F and the standard deviation SdB of the class CsB at the measurement point P. In the film thickness distribution image G4B, the standard deviation SdB at the measurement point P is represented by shading. In the film thickness distribution image G4B, defects such as the strip-shaped defect D2 and the island-shaped defect where the inclination of the film thickness tc in the plane of the object F is less than a predetermined value are emphasized. The display processing unit 15 performs a process of displaying the generated film thickness distribution image G4B. For example, as shown in FIG. 14, the display processing unit 15 performs a process of displaying the film thickness distribution image G4B without performing gain adjustment such as adjusting the shading of the generated film thickness distribution image G4B.

[0074] The calculation unit 16 calculates statistical values SvA and SvB of the standard deviations SdA and SdB of some of the classes CsA and CsB extracted by the extraction unit 14. For example, based on the class CsA received from the extraction unit 14, the calculation unit 16 calculates, as the statistical value SvA, the median MeDA of the standard deviation SdA of the received class CsA. Also, for example, based on the class CsB received from the extraction unit 14, the calculation unit 16 calculates, as the statistical value SvB, the median MeDB of the standard deviation SdB of the received class CsB. Note that the calculation unit 16 may calculate the average value of the standard deviation SdA of the class CsA received from the extraction unit 14 instead of the median MeDA as the statistical value SvA. Also, the calculation unit 16 may calculate the average value of the standard deviation SdB of the class CsB received from the extraction unit 14 instead of the median MeDB as the statistical value SvB.

[0075] Figures 15 and 16 are diagrams showing an example of the calculation result of the median by the calculation unit in the inspection apparatus according to the embodiment of the present invention. In Figures 15 and 16, the horizontal axis is the sample number of the object F, and the vertical axis is the median.

[0076] Referring to Figure 15, the calculation unit 16 calculates the median MeDA for each sample of the object F and creates a graph GRA showing the correspondence between the samples of the object F and the median MeDA. The median MeDA shown by the graph GRA can be used as an index representing the occurrence state of defects such as the dot defect D1 in the samples of the object F. The calculation unit 16 performs a process of displaying the created graph GRA.

[0077] Referring to Figure 16, the calculation unit 16 calculates the median MeDB for each sample of the object F and creates a graph GRB showing the correspondence between the samples of the object F and the median MeDB. The median MeDB shown by the graph GRB can be used as an index representing the occurrence state of defects such as the strip defect D2 and the island-like defect in the samples of the object F. The calculation unit 16 performs a process of displaying the created graph GRB.

[0078] Further, the calculation unit 16 receives the frequency distributions FdA, FdB and the histograms HgA, HgB from the creation unit 13, and calculates the feature amounts FA, FB indicating the normality of the frequency distributions FdA, FdB and the histograms HgA, HgB.

[0079] More specifically, the calculation unit 16 calculates, as the feature amount FA, at least any one of the skewness F1A of the histogram HgA, the kurtosis F2A of the histogram HgA, the correlation coefficient F3A between the ideal normal distribution and the frequency distribution FdA, and the index value F4A indicated by the Q-Q (Quantile-Quantile) plot between the ideal normal distribution and the frequency distribution FdA.

[0080] Also, the calculation unit 16 calculates, as the feature amount FB, at least any one of the skewness F1B of the histogram HgB, the kurtosis F2B of the histogram HgB, the correlation coefficient F3B between the ideal normal distribution and the frequency distribution FdB, and the index value F4B indicated by the Q-Q plot between the ideal normal distribution and the frequency distribution FdB.

[0081] For example, the correlation coefficients F3A, F3B are Pearson's correlation coefficients. The index value F4A is the average value or the median of the product DAq of the difference DA between the q-quantile of the ideal normal distribution and the q-quantile of the frequency distribution FdA and the q-quantile of the ideal normal distribution when the real number q is scanned in the predetermined numerical range Rn. The index value F4B is the average value or the median of the product DBq of the difference DB between the q-quantile of the ideal normal distribution and the q-quantile of the frequency distribution FdB and the q-quantile of the ideal normal distribution when the real number q is scanned in the predetermined numerical range Rn. The index values F4A, F4B can be used as values indicating the unevenness amounts of the film thickness distribution images G2A, G2B. Preferably, the numerical range Rn is set to a range in which the q-quantile of the ideal normal distribution becomes a positive value. By setting the numerical range Rn to a range in which the q-quantile of the ideal normal distribution becomes a positive value, the index values F4A, F4B having a high correlation with the unevenness amounts of the film thickness distribution images G2A, G2B can be calculated.

[0082] The calculation unit 16 calculates feature amounts FA and FB for each sample of the object F, and creates a graph GpA showing the correspondence between the samples of the object F and the feature amount FA, and a graph GpB showing the correspondence between the samples of the object F and the feature amount FB. The calculation unit 16 performs a process of displaying the created graphs GpA and GpB. As an example, every time the calculation unit 16 receives the frequency distributions FdA and FdB from the creation unit 13, the calculation unit 16 calculates the feature amounts FA and FB of the frequency distributions FdA and FdB, and performs a process of including the calculated feature amounts FA and FB in the graphs GpA and GpB and displaying them. Thereby, based on the change tendencies of the feature amounts FA and FB, it is possible to predict the appearance of defects in the object F and the like.

[0083] [Operation flow] Each device in the inspection system according to the embodiment of the present invention includes a computer including a memory, and a processor such as a CPU in the computer reads and executes a program including part or all of each step of the following flowchart from the memory. The programs of these multiple devices can be installed from the outside. The programs of these multiple devices are stored in a recording medium or distributed via a communication line.

[0084] FIG. 17 is a flowchart showing an example of the operation when the inspection device according to the embodiment of the present invention inspects an object.

[0085] Referring to FIG. 17, first, the inspection device 101 acquires the measurement result of the distribution of physical properties in the plane of the object F. More specifically, the inspection device 101 receives a film thickness distribution image G1 showing the measurement result of the distribution of the film thickness tc from the measurement device 111 (step S11).

[0086] Next, when the inspection device 101 determines that there are defects in the object F, for example, the inspection device 101 creates film thickness distribution images G4A and G4B showing the two-dimensional distribution of the standard deviations SdA and SdB within a part of the numerical ranges among the standard deviations SdA and SdB of the film thickness tc in a plurality of unit regions R1 in the plane of the object F (step S12).

[0087] Next, the inspection apparatus 101 performs a process of displaying the film thickness distribution images G4A and G4B (step S13).

[0088] Next, the inspection apparatus 101 calculates the median values MeDA and MeDB as statistical values SvA and SvB of the standard deviations SdA and SdB of the extracted partial classes CsA and CsB (step S14).

[0089] Next, the inspection apparatus 101 creates a graph GRA showing the correspondence between the samples of the object F and the median value MeDA, and a graph GRB showing the correspondence between the samples of the object F and the median value MeDB, and performs a process of displaying the graphs GRA and GRB (step S15).

[0090] FIG. 18 is a flowchart showing an example of the operation when the inspection apparatus according to the embodiment of the present invention generates a film thickness distribution image. FIG. 18 shows the details of step S12 in FIG. 17.

[0091] Referring to FIG. 18, first, the inspection apparatus 101 performs a noise removal process of converting the distribution of the film thickness tc indicated by the film thickness distribution image G1 into the median value M of the film thickness tc in a plurality of unit regions R2 within the plane of the object F. More specifically, the inspection apparatus 101 creates film thickness distribution images G2A and G2B by performing a convolution process on the film thickness distribution image G1 using median filters M1 and M2 (step S21).

[0092] Next, the inspection apparatus 101 performs a non-uniformity extraction process of calculating the standard deviation Sd based on the distribution of the median value M of the film thickness tc indicated by the film thickness distribution images G2A and G2B. More specifically, the inspection apparatus 101 creates a film thickness distribution image G3A by performing a convolution process on the film thickness distribution image G2A using a standard deviation filter F1. Also, the inspection apparatus 101 creates a film thickness distribution image G3B by performing a convolution process on the film thickness distribution image G2B using a standard deviation filter F2 (step S22).

[0093] Next, the inspection apparatus 101 creates a frequency distribution FdA in which the standard deviation SdA of each measurement point P shown in the film thickness distribution image G3A is classified into a predetermined number of classes CsA, and a frequency distribution FdB in which the standard deviation SdB of each measurement point P shown in the film thickness distribution image G3B is classified into a predetermined number of classes CsB (step S23).

[0094] Next, the inspection apparatus 101 creates a histogram HgA showing the frequency distribution FdA of the standard deviation SdA and a histogram HgB showing the frequency distribution FdB of the standard deviation SdB (step S24).

[0095] Next, the inspection apparatus 101 determines the presence or absence of defects in the object F based on the frequency distributions FdA and FdB. If it is determined that there are no defects in the object F (NO in step S25), the process ends.

[0096] On the other hand, if the inspection apparatus 101 determines that there are defects in the object F (YES in step S25), it extracts some of the classes CsA and CsB in the frequency distributions FdA and FdB. More specifically, when the inspection apparatus 101 determines that there are defects in the object F, it extracts some of the classes CsA out of the plurality of classes CsA in the frequency distribution FdA. Also, when the inspection apparatus 101 determines that there are defects in the object F, it extracts some of the classes CsB out of the plurality of classes CsB in the frequency distribution FdB (step S26).

[0097] Next, the inspection apparatus 101 creates film thickness distribution images G4A and G4B showing two-dimensional distributions of the standard deviations SdA and SdB of the respective partial classes CsA and CsB in the plane of the object F based on the extracted partial classes CsA and CsB (step S27).

[0098] In the inspection system 201 according to the embodiment of the present invention, the measuring device 111 is configured to measure the distribution of the film thickness tc in the plane of the object F. However, the present invention is not limited to this. The measuring device 111 may be configured to measure the distribution of physical properties other than the film thickness tc, such as the distribution of the phase difference generated by birefringence in the plane of the object F and the distribution of the optical axis, and create a film thickness distribution image G1 showing the correspondence between the position of each measurement point P and the measured physical properties.

[0099] Further, the inspection device 101 according to the embodiment of the present invention is configured to include the noise processing unit 12. However, the present invention is not limited to this. The inspection device 101 may be configured not to include the noise processing unit 12. In this case, the creation unit 13 receives the film thickness distribution image G1 from the acquisition unit 11, and creates the film thickness distribution image G3A by performing convolution processing using the standard deviation filter F1 on the received film thickness distribution image G1. Further, the creation unit 13 creates the film thickness distribution image G3B by performing convolution processing using the standard deviation filter F2 on the film thickness distribution image G1 received from the acquisition unit 11.

[0100] Further, the inspection device 101 according to the embodiment of the present invention is configured to include the display processing unit 15 and the calculation unit 16. However, the present invention is not limited to this. The inspection device 101 may be configured not to include one or both of the display processing unit 15 and the calculation unit 16.

[0101] Also, in the inspection apparatus 101 according to the embodiment of the present invention, although the creation unit 13 is configured to calculate the standard deviation Sd of the film thickness tc in a plurality of unit regions R1 in the plane of the object F, the present invention is not limited to this. The creation unit 13 may be configured to calculate the coefficient of variation of the film thickness tc in a plurality of unit regions R1 in the plane of the object F instead of the standard deviation Sd, and create film thickness distribution images G3A and G3B showing the correspondence between the position of the measurement point P in the plane of the object F and the coefficient of variation at the measurement point P. Here, the coefficient of variation is a value obtained by dividing the standard deviation Sd by the average value or the median value of the film thickness tc. For example, the creation unit 13 calculates the coefficient of variation at the measurement point P using the median value M at the measurement point P calculated by the noise processing unit 12.

[0102] Also, in the inspection apparatus 101 according to the embodiment of the present invention, although the creation unit 13 is configured to create the frequency distributions FdA and FdB, the present invention is not limited to this. The creation unit 13 may be configured not to create either one of the frequency distributions FdA and FdB.

[0103] Also, in the inspection apparatus 101 according to the embodiment of the present invention, although the creation unit 13 is configured to create the frequency distribution Fd, the present invention is not limited to this. The creation unit 13 may be configured not to create the frequency distribution Fd. In this case, for example, the extraction unit 14 extracts the standard deviations SdA and SdB within a predetermined numerical range among the standard deviations SdA and SdB of each measurement point P shown in the film thickness distribution images G3A and G3B created by the creation unit 13, and outputs the extracted standard deviations SdA and SdB to the display processing unit 15 and the calculation unit 16. Then, the display processing unit 15 performs a process of creating and displaying the film thickness distribution images G4A and G4B based on the standard deviations SdA and SdB received from the extraction unit 14.

[0104] In addition, in the inspection apparatus 101 according to the embodiment of the present invention, the extraction unit 14 is configured to determine the presence or absence of a defect in the object F based on the comparison results of the value VAmax, the value VAq, and the threshold value VAth, but the present invention is not limited thereto. The extraction unit 14 may be configured to determine the presence or absence of a defect in the object F based on the normality of the histogram HgA.

[0105] For example, the extraction unit 14 determines the normality of the histogram HgA based on the skewness SkA and kurtosis KuA of the histogram HgA. More specifically, when the absolute value of the skewness SkA is less than a predetermined threshold value Skth and the kurtosis KuA is within a predetermined interval Kuth, the extraction unit 14 determines that the histogram HgA is a normal distribution and determines that there is no defect in the object F. On the other hand, when the absolute value of the skewness SkA is greater than or equal to the threshold value Skth, or when the kurtosis KuA is outside the interval Kuth, the extraction unit 14 determines that the histogram HgA is not a normal distribution and determines that there is a defect in the object F. The threshold value Skth is, for example, 0.5. The interval Kuth is, for example, a value range of 2.5 or more and 3.5 or less.

[0106] Alternatively, the extraction unit 14 determines the normality of the histogram HgA based on the comparison results of the average value, median value, and mode value of the histogram HgA. More specifically, when the difference between the average value, median value, and mode value of the histogram HgA is less than a predetermined value, the extraction unit 14 determines that the histogram HgA is a normal distribution and determines that there is no defect in the object F.

[0107] Alternatively, the extraction unit 14 determines the normality of the histogram HgA based on the Q-Q plot of the histogram HgA. More specifically, when the residual in the linear approximation of the Q-Q plot of the histogram HgA is less than a predetermined value, the extraction unit 14 determines that the histogram HgA is a normal distribution and determines that there is no defect in the object F.

[0108] Alternatively, the extraction unit 14 determines the normality of the histogram HgA by means of the D'Agostino test based on skewness SkA, the D'Agostino test based on kurtosis KuA, the omnibus test based on skewness SkA and kurtosis KuA, the Kolmogorov-Smirnov test, or the Shapiro-Wilk test.

[0109] Also, in the inspection apparatus 101 according to the embodiment of the present invention, the extraction unit 14 is configured to determine the presence or absence of a defect in the object F based on the comparison result between the value VBmax, the third quartile QB1, and the threshold value VBth, but is not limited thereto. The extraction unit 14 may be configured to determine the presence or absence of a defect in the object F based on the normality of the histogram HgB. More specifically, the extraction unit 14 determines the presence or absence of a defect in the object F by determining the normality of the histogram HgB in the same manner as the histogram HgA.

[0110] Also, in the inspection apparatus 101 according to the embodiment of the present invention, the calculation unit 16 is configured to calculate the feature amounts FA and FB and perform the process of displaying the graphs GpA and GpB, but is not limited thereto. The calculation unit 16 may be configured not to perform the process of displaying the graphs GpA and GpB while calculating the feature amounts FA and FB. Further, the calculation unit 16 may be configured not to calculate the feature amounts FA and FB.

[0111] Also, in the inspection system 201 according to the embodiment of the present invention, the inspection apparatus 101 is configured to include the noise processing unit 12, the creation unit 13, the extraction unit 14, the display processing unit 15, and the calculation unit 16, but is not limited thereto. A part or all of the noise processing unit 12, the creation unit 13, the extraction unit 14, the display processing unit 15, and the calculation unit 16 may be provided in the measuring apparatus 111, or may be provided in another apparatus other than the measuring apparatus 111 and the inspection apparatus 101. By providing a part or all of the noise processing unit 12, the creation unit 13, the extraction unit 14, the display processing unit 15, and the calculation unit 16 in the measuring apparatus 111, the processing necessary for the inspection of the object F can be speeded up.

[0112] The above embodiments should be considered illustrative in all respects and not restrictive. The scope of the present invention is indicated by the scope of the claims rather than the above description, and it is intended that all modifications within the meaning and scope equivalent to the scope of the claims be included.

Explanation of Signs

[0113] 11 Acquisition unit 12 Noise processing unit 13 Creation unit 14 Extraction unit 15 Display processing unit 16 Calculation unit 101 Inspection device 111 Measurement device 201 Inspection system F Object P Measurement point G1, G2A, G2B, G3A, G3B, G4A, G4B Film thickness distribution image R1, R2 Unit area HgA, HgB Histogram QA1, QB1 Third quartile QA2 First quartile QrA Interquartile range VAmax, VAq, VBmax Value RA, RB Range CsA1, CsA2, CsB1, CsB2 Class D1 Point defect D2 Strip defect GRA, GRB Graph

Claims

1. An acquisition unit that acquires measurement results of the distribution of physical properties within the plane of an object having a planar portion; Based on the measurement results acquired by the acquisition unit, a standard deviation or coefficient of variation of the physical properties in a plurality of first unit regions within the plane is calculated, and a frequency distribution creation unit that creates a frequency distribution of the standard deviation or a frequency distribution of the coefficient of variation; An inspection system comprising: an extraction unit that extracts a part of the classes in the frequency distribution created by the frequency distribution creation unit.

2. The inspection system further comprises: A calculation unit that calculates a statistical value of the standard deviation or a statistical value of the coefficient of variation of the part of the classes extracted by the extraction unit, according to the inspection system described in claim 1.

3. The inspection system further comprises: Based on the part of the classes extracted by the extraction unit, an image display unit that creates an image showing a two-dimensional distribution of the standard deviation or the coefficient of variation of the part of the classes within the plane, and performs a process of displaying the created image, according to the inspection system described in claim 1 or claim 2.

4. The extraction unit extracts a plurality of different parts of the classes, according to the inspection system described in claim 1 or claim 2.

5. The inspection system further comprises: A reception unit that receives a setting of the size of the first unit region, The frequency distribution creation unit calculates the standard deviation or the coefficient of variation of the physical properties in the first unit region having the size received by the reception unit, according to the inspection system described in claim 1 or claim 2.

6. The frequency distribution creation unit creates a plurality of frequency distributions respectively corresponding to a plurality of first unit regions having different sizes, according to the inspection system described in claim 1 or claim 2.

7. The inspection system further comprises: A noise processing unit that converts the distribution of the physical property into a distribution of the median or average value of the physical property in a plurality of second unit regions in the plane is provided. The frequency distribution creation unit calculates the standard deviation or the coefficient of variation based on the distribution of the median or the average value converted by the noise processing unit. The inspection system according to claim 1 or claim 2, wherein the size of the second unit region is equal to or smaller than the size of the first unit region.

8. The frequency distribution creation unit creates a histogram showing the frequency distribution. The inspection system further includes A histogram display unit that performs a process of displaying the histogram created by the frequency distribution creation unit, the inspection system according to claim 1 or claim 2.

9. An acquisition unit that acquires a measurement result of the distribution of a physical property in a plane of an object having a planar portion, An image display unit that generates an image showing a two-dimensional distribution of the standard deviation or the coefficient of variation within a partial numerical range of the standard deviation or the coefficient of variation of the physical property in a plurality of first unit regions in the plane based on the measurement result acquired by the acquisition unit, and performs a process of displaying the generated image, an inspection apparatus.

10. An inspection method in an inspection system, comprising: A step of acquiring a measurement result of the distribution of a physical property in a plane of an object having a planar portion, A step of calculating the standard deviation or the coefficient of variation of the physical property in a plurality of first unit regions in the plane based on the acquired measurement result, and creating a frequency distribution of the standard deviation or a frequency distribution of the coefficient of variation, A step of extracting a part of classes in the created frequency distribution, an inspection method.

11. An inspection program used in an inspection apparatus, comprising: A computer, An acquisition unit that acquires measurement results of the distribution of physical properties in the plane of an object having a planar portion; Based on the measurement results acquired by the acquisition unit, calculate the standard deviation or coefficient of variation of the physical properties in a plurality of first unit regions in the plane, and create a frequency distribution of the standard deviation or a frequency distribution of the coefficient of variation; a frequency distribution creation unit; An extraction unit that extracts some classes in the frequency distribution created by the frequency distribution creation unit; An inspection program for causing it to function as such.

Images