METHOD AND APPARATUS FOR JUDGING DEFECT QUALITY

Technical Field

[0001]

The present invention relates to a method and an apparatus for judging defect quality to judge defect quality by measuring defect height information, which occurs in an inspection subject such as a semiconductor wafer which multilayer thin films are used and a thin film transistor display device in manufacturing by means of an optical image means, and specifying height information of a defect occurring position and a defect occurring layer.

Background Art

[0002]

In manufacturing the semiconductor wafer which the multilayer thin films are used, the thin film transistor display device and the like, a fine pattern is formed by using photolithography. In these manufacturing processes, the defect such as a pattern abnormality and a pinhole is occurred by various factors, and it causes a decrease in yield. In order to improve a manufacturing efficiency by managing these manufacturing processes and removing a factor of the yield decrease, working which inspects the occurring defect and specify the factor is performed. [0003]

In processes of successively forming the plural layers in such the manufacturing processes, a case that the pattern abnormality, the pinhole and the like cannot be inspected, is existed. Thus, as the only means, it is necessary to inspect the defect after forming a final layer and specify the process, which the defect is occurred, from the defect height position information. For example, in a process of sealing a flexible organic electroluminescence

(EL) , technology that prevents from entering oxygen and moisture of atmosphere to an apparatus by multi-laminating inorganic thin films such as silicon nitride and organic thin films such as polyimide, is used. However, existence of fine pinholes occurring in each of the layers is crucial for a lifetime of the apparatus. Especially, it is necessary to accurately measure positions that the pinholes, which are occurred in different layers, exist proximately, and judge quality of a product. However, a formation of these layers needs to perform a short time in a vacuum or in nitrogen atmosphere, and the subject cannot be inspected by stopping in the middle of the processes.

[0004]

FIG.l shows a cross-sectional structure of the typical flexible organic EL display device, and the organic EL emits light by an electric circuit pattern 1. The electric circuit pattern 1 is formed on a base material, which comprises a first base material 5 and a second base material 6, and is sealed by a transparent film which comprises a first sealing layer (inorganic film) 2, a second sealing layer (organic film) 3 and a third sealing layer

(inorganic film) 4. Normally, the first sealing layer 2 and the third sealing layer 4 are a silicon nitride film which is an inorganic material, and formed by a chemical vapor deposition (CVD) . Polyimide which is an organic material is used to the second sealing layer 3, and the second sealing layer 3 is formed by using, for example, an inkjet printing apparatus. The first base material 5 is a transparent substrate. For example, a resin substrate such as

polyethylene-terephthalate (PET) or polycarbonate (PC) is used to the first base material 5.

[0005]

FIG.2 shows are a state that two pinholes 7A and 7B are occurred in the sealing layer 4 in the process of forming the sealing layer (for example, a sealing transparent film which is an organic material or an inorganic material) . The pinholes 7A and 7B are existed in the third sealing layer 4 of the silicon nitride film. Even if oxygen (O2) and water (H2O) in the air permeate the polyimide film, the oxygen (O2) and the water (H2O) in the air are intercepted by the first sealing layer 2 (silicon nitride film) . Then, the EL device which is right under the pinholes 7A and 7B is not immediately broken down, and the pinholes 7A and 7B are not crucial defects to the organic EL display device.

[0006]

On the other hand, FIG.3 shows a state that the pinholes 8A and 8B, which are the defects, are existed in a proximate portion (as an axis in a vertical direction) of respective silicon nitride films of the first sealing layer 2 and the third sealing layer 4, in the forming process of the flexible organic EL display device, similarly. In this case, the oxygen and the water enter from the pinhole 8A of the sealing layer 4 that is contact to the air, and permeate the polyimide film (the sealing layer 3) as time proceeds.

Finally, the oxygen and the water arrive at the pinhole 8B of the sealing layer 2, and the EL device right under the pinhole 8B is broken down. Being shortened a lifetime of the organic EL display device by breaking down the EL device in this way is crucial for the display device.

[0007]

As described with the organic EL example, the defects on the sealing layer of the display device (for example, pinholes and foreign matters) need to be judged whether the defects are crucial for the display device as shown in FIG.3 or not. However, since a thickness of the sealing layer is a few μπι, it is necessary to measure the heights of the defects by a resolution of sub-μπι order. As technology that

satisfies such a precise measuring precision, for example, a distance measuring apparatus by a triangulation method using laser, and a white light interferometer (Japanese Unexamined Patent Publication No .2013-19767 A (Patent Document 1)), a confocal microscope (Japanese Unexamined Patent Publication No .2012-237647 A (Patent Document 2)) and the like are known.

[0008]

In Japanese Unexamined Patent Publication No.2014- 148735 A (Patent Document 3), a multifocal confocal Raman spectroscopy microscope, which Raman scattering light from a sample using a laser observation optical system is observed, is disclosed. Further, in making a three dimension profile map, technology, which generates a three dimension profile map (height map) of a sample by obtaining plural images to the stereoscopic sample depending on a focal length by means of a camera apparatus such as an optical microscope,

combining these images, and evaluating a focus matching degree which light strength and brightness contrast are maximum, is disclosed in Japanese Unexamined Patent

Publication No .2012-220490 A (Patent Document 4) .

Furthermore, in a scanning confocal microscope, technology, which generates luminance and height information (three dimension information) about a sample based on the maximum strength point of confocal images by moving a Z-revolver in a moving pitch ΔΖ and obtaining the images in each of the Z- relative positions, is disclosed in Japanese Unexamined Patent Publication No .2005-172805 A (Patent Document 5) .

The List of Prior Documents

Patent Documents

[0009]

Patent Document 1 : Japanese Unexamined Patent Publication

No.2013-19767 A

Patent Document 2 : Japanese Unexamined Patent Publication

No.2012-237647 A

Patent Document 3 : Japanese Unexamined Patent Publication

No.2014-148735 A

Patent Document 4 : Japanese Unexamined Patent Publication

No.2012-220490 A

Patent Document 5 : Japanese Unexamined Patent Publication

No.2005-172805 A

Patent Document 6 : Japanese Unexamined Patent Publication

No .Hll-337313 A

Summary of the Invention

Problems to be Solved by the Invention

[0010]

Height measurement technology such as the distance measuring apparatus by the triangulation method using the laser, the white light interferometer and the confocal microscope can be measured a flat pattern defect height having 10 μπι or more. However, the above technology cannot accurately be measured the height information of the defect having a few μπι (for example, a pinhole and a foreign matter) and the foreign matter which is not a flat shape.

Especially, the defect having Ιμπι or less cannot be measured by the above technology.

[0011]

A method of measuring interference between reflected light from the subject and reference light or a detecting method of condensing reflected light from the subject by using the confocal optical system cannot be detected the defect having Ιμπι or less. In cases that the size of the subject is Ιμπι or less and a body which illumination light for using a measurement is not a specular reflection and is scattered due to an unevenness of a surface, since the reflection light cannot be observed, the apparatus cannot detect the defect. Then, in conventional height measuring apparatuses, the defect having Ιμπι or less cannot be

detected .

[0012]

The laser observation optical system disclosed in Patent Document 3 does not have a function that inspects the defect of the subject such as the semiconductor wafer, the thin film transistor and especially the sealing layer of the flexible organic EL display device (the sealing transparent film), and calculates the height information of the defect.

[0013]

Further, a depth from defocus (DFD) or a depth from focus (DFF) , which acquires the images in changing a focus position and extracts the height information of the subject based on a luminance changing amount in a portion where the luminance is sharply changed, is known (Japanese Unexamined Patent Publication No . Hll-337313 A (Patent Document 6)).

However, in a conventional DFD process as shown in Patent Document 6, the subject, which the imaging pixel is almost one pixel such as a pattern defect having Ιμπι or less, and a fine defect on the layer (the pinhole and the foreign matter) , cannot accurately be measured the height information due to a positional displacement or an error of a pixel noise of the image. In measuring the height information by using DFD or DFF, the absolute position in the height direction depends on a precision of a machine that connects between the substrate and the microscope.

[0014]

Currently, in the display device using the EL device, in order to improve the efficiency of manufacturing, the glass substrate of a G6-size (1500mm x 1850mm) is used.

Because a stage on which the substrate provided within an inspection apparatus is placed is larger in accordance with the G6-size, it is not realistic to adjust the absolute position in the height direction of the substrate surface within Ιμπι. Thus, it is very difficult to measure the defect height information (the absolute position of the height direction) , which occurs in the pattern formed on the glass substrate, in Ιμπι.

[0015]

The technology disclosed in Patent Document 4 and Patent Document 5 generates the three dimension profile map (height map) of the sample, and does not have a function that calculates the defect height information of the sealing transparent film such as the flexible organic EL display device with a resolution of Ιμπι or less.

[0016]

The present invention has been developed in view of the above-described circumstances, and an object of the present invention is to provide the method and the apparatus for judging defect quality to accurately judge the quality of the occurred defect by accurately measuring the occurred position in the height direction of the three dimension structure even against the defect such as the pattern defect and the fine pinhole of the film (layer), which the size is less than Ιμπι, and judging a relative height to a background pattern . Means for Solving the Problems

[0017]

The above-described object of the present invention is achieved by that comprising the steps of: acquiring plural images with a predetermined step in a height direction by an optical image means to an inspection subject which is used to multilayer transparent thin films; calculating sharpness of partial images from luminance differences to adjacent pixels against each pixel of the plural images; calculating height information of the partial images from an image number which a calculating result of the sharpness at a same pixel position is maximum in all images of the plural images;

obtaining three dimension information of the all images from calculating the height information; and judging defect quality of the inspection subject based on the three

dimension information.

[0018]

The above-described object of the present invention is efficiently achieved by that further comprising the steps of: detecting a pattern defect of the image which the sharpness is highest; extracting an image having a maximum density of partial images which have the high sharpness in the plural images; setting the image as a reference position 1 in a height direction of a three dimension pattern structure; and measuring a height in the three dimension pattern structure of the occurred pattern defect from a relationship between the height information of the pattern defect and the reference position 1, or further comprising the steps of: detecting a pattern defect of the image which the sharpness is highest; extracting an image that an interference image of interference fringes which generates at edge portions of the transparent thin film has highest sharpness in the plural images; setting the image as a reference position 2 in a height direction of a three dimension pattern structure; and measuring a height in the three dimension pattern structure of the occurred pattern defect from a relationship between the height information of the pattern defect and the reference position 2, or further comprising the step of: repairing the pattern defect by using the height information of the pattern defect.

[0019]

The above-described object of the present invention is achieved by that comprising: an imaging means that acquires plural image data of an inspection subject, which has multilayer transparent thin films with image numbers, by an optical imaging means which moves upwardly and downwardly with a predetermined step; an extracting section to extract a feature of the image data;

an evaluating value calculating section to calculate an evaluating value based on the feature; an evaluating value comparing section to compare the evaluating value with a previous evaluating value consistent with a position

relationship to the evaluating value, and generate a

comparing result; an evaluating value storing section to store the evaluating value based on the comparing result; an image number storing section to store the image number based on the comparing result; a three dimension information extracting section to extract three dimension information of the inspection subject based on the image number which is stored in the image number storing section; a three dimension information extracting section to extract height information of a defect which exists in the inspection subject based on the three dimension information; and a quality judging section to judge quality of the inspection subject based on differences of the height information among the defects in a case that the plural defects are existed.

[0020]

The above-described object of the present invention is efficiently achieved by that wherein the three dimension information extracting section extracts the three dimension information based on the image number which the evaluating value is highest; or wherein the evaluating value is

calculated based on a luminance difference between an interest pixel and an adjacent pixel which is adjacent to the interest pixel; or wherein the three dimension information extracting section determines a reference of the height information based on electrode patterns of the inspection subject, and the evaluating values of the image data which interference fringes of a sealing layer of the inspection subject is imaged; or wherein the defect is a pattern defect, a pinhole or a foreign matter; or wherein the evaluating value is sharpness which is calculated based on a difference of a luminance value of the interest pixel to a luminance value of the adjacent pixel; or wherein the inspection subject is an organic electroluminescence (EL) display device; or wherein the inspection subject is a flexible organic electroluminescence (EL) display device which is formed on a flexible substrate; or further comprising at least one function to repair the defect based on the height information which the three dimension information extracting section calculates; or further comprising a function to select the at least one function depending on the height information .

Effects of the Invention

[0021]

In accordance with the present invention, the height of the fine pattern and the foreign matter is observed by decreasing the observation light by passing the edge portions of the patterns and neighborhood of the foreign matter, not be limited to the reflection light from the fine pattern and the foreign matter. Thereby, it is possible to accurately measure the height information such as the pinhole and the foreign matter having a diameter of about Ιμπι which cannot be measured by the conventional technology.

[0022]

In the present invention, in a process of acquiring and processing the plural images, even if a position error in the horizontal direction among the acquired images is occurred by vibration of the apparatus itself or a floor, it is possible to accurately measure the height information, which the fine defect is occurred, without the error by considering an amplitude of the horizontal direction

vibration among the images in comparing the image sharpness of the pattern edge and the foreign matter among the images, comparing the values of sharpness among the neighboring pixel positions of the interest pixel position, and replacing the image numbers which are stored in the neighborhood with the image number having the highest sharpness. Then, vibration resistance characteristic of the overall apparatus can be expanded, and the above described process is largely

contributed to a cost reduction of the inspection apparatus .

[0023]

According to the present invention, in manufacturing the devices of thin film multilayers, since the apparatus can specify the layer which is occurred in the pinhole having the diameter of about Ιμπι which cannot be measured by the conventional technology, it is possible to select an optimal repairing means in accordance with a material of the film which form the layer.

Brief Description of the Drawings

[0024]

In the accompanying drawings :

FIG.l is a cross-sectional view illustrating a structure example of a general flexible organic

electroluminescence (EL) display device;

FIG.2 is a cross-sectional view illustrating a state which pinholes are formed in the same layer in sealing layers of the flexible EL display device;

FIG.3 is a cross-sectional view illustrating a state which pinholes are formed in a first sealing layer and a third sealing layer in the flexible organic EL display device ;

FIG.4 is a block diagram illustrating a configuration example of a defect quality judging apparatus according to the present invention;

FIG.5 is a flowchart illustrating an operation example of the present invention; FIG.6 is a perspective view illustrating a state which two fine foreign matters are occurred in one pixel of the flexible organic EL display device;

FIG.7 is a cross-sectional view illustrating a shape of a light emitting portion of the flexible organic EL display device;

FIG.8 is a plan view illustrating a state which is shown in the flexible organic EL display device from a vertical direction of the flexible organic EL display device;

FIG.9 is a flowchart illustrating a detail operation example of imaging;

FIG.10 is a flowchart illustrating a detail operation example of extracting pattern edges of images;

FIG.11 is a flowchart illustrating a detail operation example of extracting a fine foreign matter image of the images ;

FIG.12 is a flowchart illustrating a detail operation example of extracting three dimension information of the images ;

FIG.13 is an exploded perspective view illustrating a state which first image, tenth image, thirtieth image, and fortieth image are disposed in a height direction in acquired forty images;

FIG.14 is a diagram illustrating array data of luminance values (gray values) of the first image; FIG.15 is a diagram illustrating the array data of the luminance values (gray values) of the tenth image;

FIG.16 is a diagram illustrating the array data of the luminance values (gray values) of the thirtieth image;

FIG.17 is a diagram illustrating the array data of the luminance values (gray values) of the fortieth image;

FIG.18 is a diagram illustrating edge evaluating value E(i, j) of the first image;

FIG.19 is a diagram illustrating the edge evaluating value E(i, j) of the tenth image;

FIG.20 is a diagram illustrating the edge evaluating value E(i, j) of the thirtieth image;

FIG.21 is a diagram illustrating the edge evaluating value E(i, j) of the fortieth image;

FIG.22 is a diagram illustrating a state which an edge evaluating storing value EM(i, j) is updated by using the edge evaluating value E(i, j) of the first image;

FIG.23 is a diagram illustrating a state which the edge evaluating storing value EM(i, j) is updated by using the edge evaluating value E(i, j) of the tenth image;

FIG.24 is a diagram illustrating a state which the edge evaluating storing value EM(i, j) is updated by using the edge evaluating value E(i, j) of the thirtieth image;

FIG.25 is a diagram illustrating a state which the edge evaluating storing value EM(i, j) is updated by using the edge evaluating value E(i, j) of the fortieth image;

FIG.26 is a diagram illustrating a state which a foreign matter evaluating storing value FM(i, j) is updated by using foreign matter evaluating value F(i, j) of the thirtieth image;

FIG.27 is a diagram illustrating a state which the foreign matter evaluating storing value FM(i, j) is updated by using the foreign matter evaluating value F(i, j) of the fortieth image;

FIG.28 is a diagram illustrating a state which the image numbers in an edge image number storing section are updated after completing an evaluating value update in the first image;

FIG.29 is a diagram illustrating a state which the image numbers in the edge image number storing section are updated after completing the evaluating value update in the tenth image;

FIG.30 is a diagram illustrating a state which the image numbers in the edge image number storing section are updated after completing the evaluating value update in the thirtieth image;

FIG.31 is a diagram illustrating a state which the image numbers in the edge image number storing section are updated after completing the evaluating value update in the fortieth image; FIG.32 is a diagram illustrating a state which the image numbers in a foreign matter image number storing section are updated after completing an evaluating value update in the thirtieth image;

FIG.33 is a diagram illustrating a state which the image numbers in the foreign matter image number storing section are updated after completing the evaluating value update in the fortieth image; and

FIG.34 is a diagram illustrating three dimension information by a contour display of a sample (the flexible organic EL device);

FIG. 35 is a schematic side view of apparatus for mapping of features in a sample, in accordance with an embodiment of the invention;

FIG. 36 is a plot that schematically illustrates vibrations measured in the apparatus of FIG. 35, in

accordance with an embodiment of the invention;

FIG. 37 is a plot that schematically illustrates measurements of focal quality made by the apparatus of FIG. 35, in accordance with an embodiment of the invention; and

FIG. 38 is a flow chart that schematically illustrates a method for mapping of features in a sample, in accordance with an embodiment of the invention. Mode for Carrying Out the Invention

[0025]

In accordance with the present invention, in

manufacturing a semiconductor, a display device and the like that multilayer transparent thin films are used, height information of a fine pattern defect, which is occurred in film formation, is measured by an optical inspection means. The height information of a defect occurring position and a defect occurring film kind are specified, and defect quality is judged. Particularly, a microscope image apparatus having a mechanism which a focus position mechanically scans in a height direction, continuously images and stores plural images by scanning in the height direction, and calculates contrast differences among adjacent pixels of image

information as evaluating values of numerical values. By comparing the obtained evaluating values in magnitude among each of the pixels in the images, the apparatus selects an image number that has the highest sharpness in pattern edge images, and measures as the height of a vertical direction of the image portion by converting the image number to a height position of the vertical direction which the image is acquired. A reference position of the height is obtained by a density of a maximum evaluating value of contrast or an interference fringe image which occurs at edge portions of a transparent film. With reference to the images of defect points which are extracted as a fine image point such as a pattern abnormality, a pinhole and a foreign matter, the evaluating values are calculated by a similar method. The heights of the defect points are measured by a relative position relationship to the reference height. The layer which is occurred in the defect points is specified from the heights of the vertical direction which are occurred in the defects and the defect quality is judged.

[0026]

An embodiment according to the present invention will be described with reference to the drawings.

[0027]

At first, a configuration example of the embodiment according to the present invention is described with

reference to FIG.4.

[0028]

In the present embodiment, a flexible organic electroluminescence (EL) display device 10 is an inspection subject. The display device 10 is mounted on a predetermined stage (not shown), and is disposed beneath a microscope 20. An objective lens is attached at an inspection subject side of a lens barrel section of the microscope 20, and an image camera 21 is attached at the opposite side. A sequence control section 30 controls a height direction driving motor 23 and an image acquiring section 31. The height direction driving motor 23 is coupled to the microscope 20 via a rack and a pinion or the like. The microscope 20 moves upwardly and downwardly by driving the motor 23 in the sequence control section 30. The camera 22 continuously images the flexible organic EL display device 10, and the image

acquiring section 31 acquires image data from the camera 22 in response to instructions of the sequence control section 30. An image memory 32 stores the image transmitted from the image acquiring section 31. The sequence control section 30 controls the height of the microscope 20 with a predetermined step (up amount or down amount) via the height direction driving motor 23, adjusts a focus position of the objective lens 21, and can image the display device 10 which is the inspection subject. Since the predetermined step is a resolution of the height direction, the smaller the

predetermined step is, the more images can be acquired in the height direction in a measuring range. Conversely, the larger the predetermined step is, the fewer images can be acquired in the height direction in the measuring range. By adjusting the predetermined step, the resolution of the height direction can be adjusted. Then, when the microscope 20 moves to the height of the measuring range, all of the images corresponding to the predetermined step are stored in the image memory 32. In a case that an optical system of the microscope 20 comprises an infinity-corrected optical system, the motor 23 may drive only the objective lens 21 upwardly and downwardly instead of driving the microscope 20 upwardly and downwardly by using the motor 23.

[0029]

The data which is stored in the image memory 32 is processed by a judging process section as described below. The judging process section comprises an edge process section 40 to extract the pattern edges, a foreign matter process section 50 to extract and process the fine foreign matter and a quality judging section 60 to judge the quality of the defects based on three dimension information ED from the edge process section 40 and foreign matter three dimension information FM from the foreign matter process section 50.

[0030]

The edge process section 40 comprises a pattern edge extracting section 41 to extract the edges of the pattern, an edge evaluating value calculating section 42 to calculate edge evaluating values, an edge evaluating value comparing section 43 to compare the edge evaluating values, an edge evaluating value storing section 44 to store the edge evaluating values, an edge image number storing section 45 to store numbers of the edge images (including a symbol and the like) and an edge three dimension information extracting section 46 to extract the edge three dimension information ED based on information of the edge evaluating value storing section 44 and the edge image number storing section 45. The foreign matter process section 50 comprises a fine foreign matter extracting section 51, a foreign matter evaluating value calculating section 52 to calculate the foreign matter evaluating values, a foreign matter evaluating value

comparing section 53 to compare the foreign matter evaluating values, a foreign matter evaluating value storing section 54 to store the foreign matter evaluating values, a foreign matter image number storing section 55 to store numbers of the foreign matter images (including a symbol and the like) and a foreign matter three dimension information extracting section 56 to extract the foreign matter three dimension information FM based on information of the foreign matter evaluating value storing section 54 and the foreign matter image number storing section 55.

[0031]

In such a configuration, the operation example is shown in a flowchart of FIG.5. At first, the images of the flexible organic EL display device 10, which is the

inspection subject, are imaged by using the microscope 20 by means of driving the sequence control section 30 (Step S100) . Next, the pattern edges of the images are extracted at the edge process section 40 (Step S200) and the fine foreign matters of the images are extracted at the foreign matter process section 50 (Step S300) . The order of an edge extracting process and a foreign matter extracting process may be changeable. The edge three dimension information extracting section 46 in the edge process section 40 and the foreign matter three dimension information extracting section 56 in the foreign matter process section 50 perform the extracting process of the three dimension information (Step S400) . The three dimension information ED from the edge three dimension information extracting section 46 and the foreign matter three dimension information FM from the foreign matter three dimension information extracting section 56 are inputted into the quality judging section 60 and the quality of the defects are judged (Step S500) .

[0032]

At first, the method of extracting the edge pixel such as the pattern and the defects with reference to the images acquired in the measuring range will be described.

[0033]

In principle, in the image information, the

evaluating value, which the interest pixel is evaluated, is calculated based on a luminance difference between the interest pixel and the adjacent pixel which is adjacent to the interest pixel. Comparing the evaluating value with a predetermined reference threshold, a degree of defocus with reference to a partial image around the interest pixel is judged. Based on the judging result of the degree of the defocus, the apparatus judges whether the interest pixel is the edge pixel such as the pattern and the defect or not. In a case that the partial image is not defocus, the partial image has high sharpness. In a case that the partial image is defocus, the partial image has low sharpness. As

described below, even in a case of extracting the fine foreign matter pixel, the degree of the defocus in the pixel, which the possibility of the foreign matter is existed, is judged by using a similar method.

[0034]

As an extracting subject such as the pattern and the defect that exists in the image as image information, for example, the electrode pattern of the flexible organic EL display device 10, patterns of an organic film, an inorganic film and the like and the defects such as the pinholes and the like can be considered. In order to evaluate the

interest pixel, variables that represent the pixel are defined as follows. That is, when the positon of each of the pixels, which constitute the image, is assumed as a lateral position i and a longitudinal position j, a gray value (the luminance value) of any pixel in the image is represented by G(i, j ) .

[0035]

Next, a method of calculating the edge evaluating value E(i, j) by using functions will be described. Used functions are a function MAX (X, Y) that compares X with Y and outputs a larger value, and a function ABS (X) that outputs an absolute value of X. Using these functions, the edge

evaluating value E(i, j) can be calculated. Then, comparing the edge evaluating value E(i, j) with the edge threshold value, the edge pixel of the pattern (an edge partial image) can be extracted. When the calculated edge evaluating value E(i, j) is larger than the edge threshold value, it is considered that the pixel having G(i, j) is the edge pixel. When the calculated edge evaluating value E(i, j) is not larger than the edge threshold value, it is considered that the pixel having G(i, j) is not the edge pixel.

Particularly, the edge evaluating value E(i, j) is given by Equation 1. Using Equation 2, it is determined whether the pixel having G(i, j) is the edge pixel or not.

[Equation 1]

Edge evaluating value E(i, j)

= MAX (MAX (ABS (G (1 _r j) -G(i+1, j)), ABS (G (i , j) - G(i,

3+1))),

MAX (ABS (G (1 , j) - G(l-l _r j)), ABS (G (1 , j) - G(i, j-1))))

[Equation 2]

Edge evaluating value E(i, j) > edge threshold value

The process of the edge process section 40 that extracts the edge pixel such as the pattern and the defect is performed as follows .

[0036]

At first, the pattern edge extracting section 41 acquires the gray value of any pixel in the image G(i, j), the gray values being adjacent to any pixel in the lateral direction in the image G(i-1, j) and G(i+1, j), and the gray values being adjacent to any pixel in the longitudinal direction in the image G(i, j-1) and G(i, j+1) from the image memory 32. The edge evaluating value calculating section 42 calculates the edge evaluating value E(i, j) by using the calculating method as shown in Equation 1, and compares the edge evaluating value E(i, j) with the edge threshold value. As a result of comparison, when the edge evaluating value E(i, j) is larger than the edge threshold value, it is considered that the pixel having G(i, j) is the edge pixel. When the edge evaluating value E(i, j) is not larger than the edge threshold value, it is not considered that the pixel having G(i, j) is the edge pixel.

[0037]

The edge evaluating value comparing section 43 compares the subsequently acquired edge evaluating value E(i, j) with the edge evaluating storing value EM(i, j) of the corresponding position (the lateral position i, the

longitudinal position j ) which is stored in the edge

evaluating value storing section 44. The edge evaluating storing value EM(i, j) is the previously acquired edge evaluating value E(i, j) prior to the subsequently acquired edge evaluating value E(i, j) . As a result of comparison, when the edge evaluating value E(i, j) is larger than the edge evaluating storing value EM(i, j), the edge evaluating value storing section 44 rewrites the edge evaluating storing value EM(i, j) to the edge evaluating value E(i, j) . In a case of rewriting the edge evaluating storing value EM(i, j), the edge image number storing section 45 updates the edge image number EN(i, j), which is an element of the

corresponding position, to the image number in currently processing, and associates the edge evaluating value E(i, j) with the image number. The number is not necessary to the numerical number, and may be a symbol which can distinguish others. Thus, the apparatus judges whether all of the pixels which have the image number in currently processing are the edge pixel or not, sequentially. Depending on the judging, the edge evaluating storing value EM(i, j) and the edge image number EN(i, j) is rewritten to the edge evaluating value E(i, j) and the image number N in currently processing.

[0038]

Completing the above process to all of the pixels which have the image number N in currently processing, a similar process is performed to the image having the next image number. Completing the above process to all image numbers of the images, the edge three dimension information extracting section 46 generates the height information of the edge images such as the pattern and the defect, based on the image number EN(i, j) stored in the edge image number storing section 45.

[0039]

Next, the foreign matter process section 50 that extracts the fine foreign matter and extracts the three dimension information of the fine foreign matter will be described. The process of the image of the fine foreign matter may be performed after the process of generating the pattern edge height information, may be performed before the process of generating the pattern edge height information, or may be performed in parallel with the process of generating the pattern edge height information.

[0040]

Similar to the process of extracting the edge pixels such as the pattern and the defect, when the positon of each of the pixels, which constitute the image, is assumed as a lateral position i and a longitudinal position j , the gray value (the luminance value) of any pixel in the image is represented by G(i, j) . A function, which is used in

calculating the foreign matter evaluating value F(i, j), is a function MIN(X, Y) that compares X with Y and outputs a smaller value. Using this function, the foreign matter evaluating value F(i, j) can be calculated.

[0041]

Comparing the foreign matter evaluating value F(i, j) with the fine foreign matter threshold value, the fine foreign matter pixel of the pattern (a fine foreign matter partial image) can be extracted. If the calculated foreign matter evaluating value F(i, j) is larger than the fine foreign matter threshold value, it is considered that the pixel having G(i, j) is the fine foreign matter pixel. If the calculated foreign matter evaluating value F(i, j) is not larger than the fine foreign matter threshold value, it is considered that the pixel having G(i, j) is not the fine foreign matter pixel. Particularly, the foreign matter evaluating value F(i, j) is given by Equation 3. Using

Equation 4, it is determined whether the pixel having G(i, j) is the fine foreign matter pixel or not.

[Equation 3]

Foreign matter evaluating value F(i, j)

= MIN(G(i-l, j) + G(i+1, j) - G(i, j)*2 _r G(i, j-1) +

G(i, - G(i, j) *2)

[Equation 4]

Foreign matter evaluating value F(i, j) > fine foreign matter threshold value

By using Equation 4, it is possible to extract a pixel that has the lower luminance pixel than adjacent pixels in both the lateral direction and the longitudinal direction, that is, a scotoma having the size of about one pixel.

[0042]

The process of the foreign matter process section 50 that extracts the fine foreign matter pixel is performed as follows .

[0043]

At first, the fine foreign matter extracting section 51 acquires the gray value of any pixel in the image G(i, j), the gray values being adjacent to any pixel in the lateral direction in the image G(i-1, j) and G(i+1, j), and the gray values being adjacent to any pixel in the longitudinal direction in the image G(i, j-1) and G(i, j+1) from the image memory 32. The foreign matter evaluating value calculating section 52 calculates the foreign matter evaluating value F(i, j) by using the calculating method as shown in Equation 3, and compares the foreign matter evaluating value F(i, j) with the fine foreign matter threshold value. As a result of comparison, if the foreign matter evaluating value F(i, j) is larger than the fine foreign matter threshold value, it is considered that the pixel having G(i, j) is the fine foreign matter pixel. If the foreign matter evaluating value F(i, j) is not larger than the fine foreign matter threshold value, it is not considered that the pixel having G(i, j) is the fine foreign matter pixel.

[0044]

The foreign matter evaluating value comparing section 53 compares the subsequently acquired foreign matter

evaluating value F(i, j) with the foreign matter evaluating storing value FM(i, j) of the corresponding position (the lateral position i, the longitudinal position j) which is stored in the foreign matter evaluating value storing section 54. The foreign matter evaluating storing value FM(i, j) is the previously acquired foreign matter evaluating value F(i, ) prior to the subsequently acquired foreign matter

evaluating value F(i, j) . As a result of comparison, when the foreign matter evaluating value F(i, j) is larger than the foreign matter evaluating storing value FM(i, j), the foreign matter evaluating value storing section 54 rewrites the foreign matter evaluating storing value FM(i, j) to the foreign matter evaluating value F(i, j) . In a case of rewriting the foreign matter evaluating storing value FM(i, j ) , the foreign matter image number storing section 55 updates the foreign matter image number FN(i, j), which is an element of the corresponding position, to the image number N in currently processing, and associates the foreign matter evaluating value F(i, j) with the image number N.

[0045]

The judging whether the pixel is the foreign matter pixel or not, is sequentially performed to all of the pixels of the image number N in currently processing. Depending on the judging, the foreign matter evaluating storing value FM(i, j) and the foreign matter image number FN(i, j) are rewritten to the foreign matter evaluating value F(i, j) and the image number N in currently processing, respectively. Completing the above process to all of the pixels which have the image number N in currently processing, a similar process is performed to the image having the next image number (N+l) .

[0046]

Completing the above process to all image numbers of the images, the foreign matter three dimension information extracting section 56 generates the height information of the fine foreign matter, based on the foreign matter image number FN(i, j) stored in the foreign matter image number storing section 55.

[0047]

The quality judging section 60 performs a quality judging about the flexible organic EL display device 10 which is a sample, based on relative heights of the edge three dimension information ED (height information of the edge pixel of the pattern) from the edge three dimension

information extracting section 46 and the foreign matter three dimension information FM (height information of the fine foreign matter) from the foreign matter three dimension information extracting section 56, and outputs the judging result .

[0048]

At first, as an example of the inspection subject (the sample), a state, which two fine foreign matters 101 and 102 are occurred in one pixel of the flexible organic EL display device 10, is shown in FIG.6. As described below, a first fine foreign matter 101 is disposed in a plane 10-30 which the thirtieth image is the focus position and a second fine foreign matter 102 is disposed in a plane 10-40 which the fortieth image is the focus position, respectively. A reference plane of the height is the electrode pattern 103. Since the images are acquired per 0. Ιμπι in the present embodiment, it is detected that the fine foreign matter 101 is disposed on a position which is 2. Ομπι above the pattern and the fine foreign matter 102 is disposed on a position which is 3. Ομπι above the pattern.

[0049]

Next, the reference of the height of the three dimension information (structure) will be described. As the structure which is difficult to detect the pattern edge of the circuit of the flexible organic EL, a case that a shape of the light emitting layer 107 is a rectangular shape

(window) formed by the organic film 105 which covers the cathode electrode 104, as shown in FIG.7, is existed. As described below, a portion of light emitting is described as the window. Illumination light is reflected on an

interfacial surface between the organic film 105 and a transparent film 106. By the illumination light using coaxial vertical illumination, the interference fringes having a ring shape are generated around the window. The interference fringes can be observed by using an image processing. The interference fringe 107A which is existed in an innermost position is an edge portion (edge) of the window 107 of the organic film 105. Consequently, the edge

evaluating value E(i, j), which is detected at the region which is existed in this edge, is suitable to the reference of the height. The pattern edge of the circuit of the flexible organic EL substrate is difficult to detect due to the structure. However, the edge portion (edge) of the window 107 of the organic film 105 is suitable to the reference of the height of the three dimension information (structure) . Further, when observing the flexible organic EL substrate from the vertical direction of the flexible organic EL substrate (for example, by using the microscope), the cathode electrode 104 acts like a mirror surface, and the strong interference fringes having a ring shape are observed around the window by the illumination light which is

reflected in an inverse direction at the interfacial surface between the organic film 105, which forms the window, and the transparent film 106. The above described state is shown in FIG.8. In FIG.8, an edge image 104A of the cathode electrode 104, an edge image 107A of the organic EL light emitting layer 107 and the interference image 107B which generates at the edge portion of the organic film 105, which forms the window, are illustrated.

[0050]

Details of respective operations as described above will be described with reference to flowcharts.

[0051]

A detail operation example of the imaging (Step S100 in FIG.5) will be described with reference to the flowchart of FIG.9. At first, the sequence control section 30

initializes the image number (Step S101) and adjusts the height of the microscope 20 to a measuring starting point by driving the height direction driving motor 23 (Step S102) . In this state, the objective lens 10 and the camera 22 are a relationship of the focus position, the image of the display device 10 is imaged via the image reading section 31 (Step S103) and the image data is stored in the image memory 32 (Step S104) . Next, the microscope 20 moves in the

predetermined step in the height direction (Step S105), and the apparatus judges whether the height of the microscope 20 is within the measuring range or not (Step S106) . In a case that the height of the microscope 20 is within the measuring range, the image of the display device 10 is imaged repeatedly, and the image data is stored sequentially. When the height of the microscope 20 is out of the measuring range, the image number N in currently processing is set as an image number maximum value Nmax and is stored in the image memory 32 (Step S107), and the imaging is completed.

[0052]

Next, a detail of the pattern edge extracting operation of the display device 10 (Step S200 in FIG.5) will be described with reference to the flowchart of FIG.10. At first, values, which indicate the image number N, the height of a lens barrel section of the microscope 20, the lateral position i, the longitudinal position j and an edge pixel number EC(N) , are initialized (Step S201) . Next, the edge evaluating value calculating section 42 acquires the gray values G(i, j), G(i+1, j), G(i, j+1), G(i-1, j) and G(i, j-1) of the image number N in currently processing from the image memory 16 (Step S202), and calculates the edge evaluating value E(i, j) from the gray values, based on Equation 1 (Step S203) . The edge evaluating value comparing section 43 compares the edge evaluating value E(i, j) with the edge threshold value (Step S204) . In a case that the edge

evaluating value E(i, j) is larger than the edge threshold value, the edge evaluating value comparing section 43 rewrites the edge evaluating storing value EM(i, j) of the edge evaluating value storing section 44 to the calculated edge evaluating value E(i, j) (Step S205), and rewrites the edge image number EN(i, j) of the edge image number storing section 45 to the image number N in currently processing (Step S206) . Further, the edge pixel number EC (N) , which indicates the pixel number which the pixels of the image number N are judged to the edge pixel, is incremented by one ("+1") (Step S207), and the lateral position i of the pixel is incremented by one ("+1") (Step S208) . In the above Step S204, in a case that the edge evaluating value E(i, j) is the threshold value or less, only the lateral position i of the pixel is incremented by one ("+1") (Step S208) .

[0053]

Then, comparing the lateral position i of the pixel with the maximum lateral position imax (the lateral position is a positon of the edge of the image) (Step S209) , in a case that the lateral position i of the pixel is the maximum lateral position imax or more, the longitudinal position j is incremented by one ("+1") (Step S210), and the lateral position i is initialized (Step S210) . In a case that the lateral position i of the pixel is less than the maximum lateral position, the process is returned to the above Step S202, the gray values of the pixel G(i, j), G(i+1, j), G(i, j+1), G(i-1, j) and G(i, j-1) are acquired again, and the process of extracting the edge pixels of the pattern is performed .

[0054]

Next, comparing the longitudinal position j of the pixel with the maximum longitudinal position jmax (Step S211), in a case that the longitudinal position j of the pixel is the maximum longitudinal position jmax or more, the image number N is incremented by one ("+1") , and the

longitudinal position j is initialized (Step S212) . In a case that the longitudinal position j of the pixel is less than the maximum longitudinal position jmax, the process is returned to the above Step S202, the gray values of the pixel G(i, j), G(i+1, j), G(i, j+1), G(i-1, j) and G(i, j-1) are acquired similarly, and the process of extracting the edge pixels of the pattern is performed.

[0055]

Then, comparing the image number N with the image number maximum value Nmax (Step S213), in a case that the image number N is the image number maximum value Nmax or more, the process of extracting the edge pixels of the pattern is completed. In a case that the image number N is less than the image number maximum value Nmax, the process is returned to the above Step S202, and the process of

extracting the edge pixels of the pattern is performed.

[0056]

Next, a detail of the fine foreign matter image extracting operation of the display device 10 (Step S300 in FIG.5) will be described with reference to the flowchart of FIG.11. At first, values, which indicate the image number N, the height of the microscope, the lateral position i, the longitudinal position j and a foreign matter pixel number FC (N) , are initialized (Step S301) . Next, the foreign matter evaluating value calculating section 52 acquires the gray values of the image G(i, j), G(i+1, j), G(i, j+1), G(i-1, j) and G(i, j-1) (Step S302), and calculates the foreign matter evaluating value F(i, j) in accordance with Equation 3 (Step S303) . The foreign matter evaluating value comparing section 53 compares the foreign matter evaluating value F(i, j) with the fine foreign matter threshold value (Step S304) . In a case that the foreign matter evaluating value F(i, j) is larger than the fine foreign matter threshold value, the foreign matter value comparing section 53 rewrites the foreign matter evaluating storing value FM(i, j) of the foreign matter evaluating value storing section 54 to the foreign matter evaluating value F(i, j) (Step S305), and rewrites the foreign matter image number FN(i, j) of the foreign matter image number storing section 55 to the image number N in currently processing (Step S306) . The foreign matter number FC(N), which indicates the number which the pixels of the image number N are judged to the foreign matter, is incremented by one ("+1") (Step S307), and the lateral position i of the pixel is incremented by one ("+1") (Step S308) . In a case that the foreign matter evaluating value F(i, j) is the fine foreign matter threshold value or less, only the lateral position i of the pixel is incremented by one P+l") (Step S308) .

[0057]

Next, comparing the lateral position i of the pixel with the maximum lateral position imax (the lateral position is a positon of the edge of the image) (Step S309) , in a case that the lateral position i of the pixel is the maximum lateral position imax or more, the longitudinal position j is incremented by one ("+1") and the lateral position i is initialized (Step S310) . In a case that the lateral position i of the pixel is less than the maximum lateral position, the process is returned to the above Step S302, the gray values of the pixel G(i, j), G(i+1, j), G(i, j+1), G(i-1, j) and G(i, j-1) are acquired again, and the process of extracting the fine foreign matter image is performed. Then, comparing the longitudinal position j of the pixel with the maximum longitudinal position jmax (Step S311), in a case that the longitudinal position j of the pixel is the maximum

longitudinal position jmax or more, the image number N is incremented by one ("+1") , and the longitudinal position j is initialized (Step S312) . In a case that the longitudinal position j of the pixel is less than the maximum longitudinal position jmax, the process is returned to the above Step S302, the gray values of the pixel G(i, j), G(i+1, j), G(i, j+1), G(i-1, j) and G(i, j-1) are acquired similarly, and the process of extracting the fine foreign matter image is performed. Comparing the image number N with the image number maximum value Nmax (Step S313), in a case that the image number N is the image number maximum value Nmax or more, the process of extracting the fine foreign matter image is completed. In a case that the image number N is less than the image number maximum value Nmax, the process is returned to the above Step S302, and the process of extracting the fine foreign matter image is performed.

[0058]

Next, a detail of the three dimension information extracting operation in the edge three dimension information extracting section 46 and the foreign matter three dimension information extracting section 56 (Step S400 in FIG.5) will be described with reference to the flowchart of FIG.12.

Here, an example, which the process of the edge is firstly performed and the process of the foreign matter is

subsequently performed, is described. The order of the processes may be changeable. The both processes can be performed in parallel.

[0059]

Firstly, detecting the image number, which the edge pixel number is the largest, is performed. At first, the image number N and the edge pixel number maximum value ECmax are initialized (Step S401), and the edge pixel number EC(N) is acquired (Step S402) . Next, the apparatus judges whether the edge pixel number EC (N) is larger than the edge pixel number maximum value ECmax or not (Step S403) . In a case that the edge pixel number EC (N) is larger than the edge pixel number maximum value ECmax, the apparatus replaces the edge pixel number maximum value image number ECNmax with the image number N (Step S404), and the image number N is

incremented by one ("+1") (Step S405) . In a case that the edge pixel number EC (N) is the edge pixel number maximum value ECmax or less, only the image number N is incremented by one ("+1") (Step S405) . Then, the apparatus judges whether the image number N is the image number maximum value Nmax or more, or not (Step S406) . In a case that the image number N is less than the image number maximum value Nmax, the process is returned to the above Step S403 and the above processes are repeatedly performed. In this way, the

apparatus detects the image number which the edge pixel number is the largest, and detects the image number which the electrode pattern of the display device 10 is in focus.

[0060]

In the next step, detecting the image number N, which the foreign matter pixel number is the largest, is performed. At first, the image number N, a first foreign matter pixel number maximum value FCNlmax and a second foreign matter pixel number maximum value FCN2max are initialized (Step S407), and the foreign matter pixel number FC (N) is acquired (Step S408) . Next, the apparatus judges whether the foreign matter pixel number FC(N) is larger than the first foreign matter pixel number maximum value FCNlmax (Step S409) . In a case that the foreign matter pixel number FC (N) is larger than the first foreign matter pixel number maximum value FCNlmax, the apparatus rewrites the value of the first foreign matter pixel number image number FCNlmax to the value of the second foreign matter pixel number image number

FCN2max, and rewrites the value of the first foreign matter pixel number image number to the image number N (Step S410) . The image number N is incremented by one ("+1") (Step S411) . In a case that it is judged that the foreign matter pixel number FC(N) is the first foreign matter pixel number maximum value FCNlmax or less, only the image number N is incremented by one ("+1") (Step S411) . Then, the apparatus judges whether the image number N is the image number maximum value Nmax or more, or not (Step S412) . In a case that the image number N is less than the image number maximum value Nmax, the process is returned to the above step S409, and the above operation is repeatedly performed. Thus, the image number which has the largest foreign matter pixel number, that is, the first foreign matter pixel number image number FCNlmax is detected, and the image number which the largest foreign matter rewriting pixel number or the image number which has the second largest foreign matter pixel number, that is, the second foreign matter pixel number image number FCN2max is detected. The image number which the foreign matter is in focus is detected.

[0061]

Finally, a first difference between the first foreign matter pixel number image number FCNlmax and the edge pixel number maximum image number ECNmax, and a second difference between the second foreign matter pixel number image number FCN2max and the edge pixel number maximum image number ECNmax are calculated. Then, the height information of the foreign matter is extracted based on the first difference and the second difference (Step S413) .

[0062]

Next, based on the image, which actually the sample of the flexible organic EL (one pixel size) is imaged, the process of extracting the edge of the sample and the three dimension information (height information) of the fine foreign matter is sequentially described.

[0063]

At first, by changing the height of the microscope in the height direction in an equal interval (a predetermined step, for example, 0. Ιμπι) , forty images of the flexible organic EL (one pixel size), which is the inspection subject, are acquired. Then, the first image, the tenth image, the thirtieth image and the fortieth image in the acquired forty images are arranged in the height direction, as shown in FIG.13. As shown in FIG.13, the first image 10-1 is an image that is acquired from the height which is Ιμπι below the electrode pattern. The images are acquired in the upper direction per 0. Ιμπι from the height which is Ιμπι below the electrode pattern. The image 10-10 that the electrode pattern 103 is completely in focus is acquired in the tenth image. The image 10-30 that the first foreign matter 101 is in focus is acquired in the thirtieth image. The image 10-40 that the second foreign matter 102 is in focus is acquired in the fortieth image. A focusing image has a feature that the luminance difference between the interest pixel and the adjacent pixels is large and the sharpness of the image is high. For example, the interference image of the sample, the defect of the film and the like in the focusing image are existed as an image, which thin lines or points which have a large brightness difference (luminance difference) to the surrounding, that is, a partial image which has the high sharpness (not defocus) .

[0064]

Here, the flexible organic EL (one pixel size) is imaged on an area which is arranged with twenty pixels in the longitudinal direction and twenty pixels in the lateral direction. States which are converted to array data of the luminance values are shown in FIG.14 to FIG.17. In FIG.14 to FIG.17, the array data of the luminance value (gray value) of the first image 36, the array data of the luminance value (gray value) of the tenth image 37, the array data of the luminance value (gray value) of the thirtieth image 38 and the array data of the luminance value (gray value) of the fortieth image 39 are shown. In FIG.14 to FIG.17, the position of the lateral direction is corresponding to the lateral position i, and the position of the longitudinal direction is corresponding to the longitudinal position . The position relationship between the position of the image and the array data is also applicable to array data of the luminance value as described below.

[0065]

Next, the edge evaluating value E(i, j) of each of the images is calculated by using Equation 1. In the area that the flexible organic EL (one pixel size) of the

inspection subject is arranged with twenty pixels in the longitudinal direction and twenty pixels in the lateral direction, distributions of the edge evaluating value E(i, j) are shown in FIG.18 to FIG.21. In FIG.18 to FIG.21, the edge evaluating value E(i, j) of the first image 36, the edge evaluating value E(i, j) of the tenth image 37, the edge evaluating value E(i, j) of the thirtieth image 38 and the edge evaluating value E(i, j) of the fortieth image 39 are shown. Similarly, the foreign matter evaluating value F(i, ) of each of the images can be calculated by using Equation 3.

[0066]

After calculating the edge evaluating value E(i, j) corresponding to all of the image number (N = 1 to 40) which are from the first image to the fortieth image, the

evaluating value storing section 44 updates the edge

evaluating storing value EM(i, j), as shown in FIG.22 to FIG.25. In FIG.22 to FIG.25, a state that the edge

evaluating storing value EM(i, j) is updated by using the edge evaluating value E(i, j) of the first image 36, a state that the edge evaluating storing value EM(i, j) is updated by using the edge evaluating value E(i, j) of the tenth image 37, a state that the edge evaluating storing value EM(i, j) is updated by using the edge evaluating value E(i, j) and the foreign matter evaluating value F(i, j) of the thirtieth image 38 and a state that the edge evaluating storing value EM(i, j) is finally updated by using the edge evaluating value E(i, j) and the foreign matter evaluating value F(i, j) of the fortieth image 39, are shown.

[0067] In this way, after completing an update of the edge evaluating storing value EM(i, j) by using the edge

evaluating value E(i, j) (hereinafter referred to as

"evaluating value update"), states that the image number N is updated in the edge image number storing section 45, are shown in FIG.28 to FIG.31.

[0068]

In the process of detecting the foreign matter, a state that the foreign matter evaluating storing value FM(i, ) is updated by using the foreign matter evaluating value F(i, j) of the thirtieth image, is shown in FIG.26. A state that the foreign matter evaluating storing value FM(i, j) is updated by using the foreign matter evaluating value F(i, j) of the fortieth image, is shown in FIG.27. Because the foreign matter evaluating value F(i, j) of the first image and the tenth image is the foreign matter threshold value or less, the foreign matter evaluating value F(i, j) of the first image and the tenth image is set to zero.

[0069]

Next, in the process of detecting the edge, a state that the image number N is updated in the image number storing section 45 after completing the evaluating value update about the first image, similarly, a state that the image number N is updated in the image number storing section 45 after completing the evaluating value update about the tenth image, a state that the image number N is updated in the image number storing section 45 after completing the evaluating value update about the thirtieth image and a state that the image number N is updated in the image number storing section 45 after completing the evaluating value update about the fortieth image, are shown in FIG.28 to FIG.31.

[0070]

Next, in the process of detecting the foreign matter, a state that the image number is updated in the foreign matter image number storing section 55 after completing the evaluating value update about the thirtieth image, is shown in FIG.32. A state that the image number is updated in the foreign matter image number storing section 55 after

completing the evaluating value update about the fortieth image, is shown in FIG.33. The foreign matter three

dimension information extracting section 56 judges that pixels, which the foreign matter is represented, are existed in the position (5, 14) in the image number 30 (thirtieth image) and the position (15, 4) in the image number 40

(fortieth image) . Since the foreign matter evaluating value F(i, j) of the first image and the tenth image is the foreign matter threshold value or less and is set to zero, in the first image and the tenth image, the evaluating value update is not performed and the pixel which the foreign matter is represented is not detected.

[0071]

In this way, after completing the process of updating the image number that the edge and the foreign matter are detected, finally, the apparatus creates a contour graph about the inspection subject, based on the array data of the image number shown in FIG.31. As shown in FIG.34, the three dimension information (height information) of the sample can be analyzed from the contour graph. Based on FIG.34, the three dimension information about the inspection subject is particularly judged.

[0072]

Since the array of the edge evaluating value E(i, j) of the tenth image as shown in FIG.19 has the highest density, it can be judged that the tenth image is an image which is focused on the thin film transistor (TFT) circuit portion existed in the lowest portion of the subject, which is the reference of the three dimension information

(structure) thereof. Thus, the height of the tenth image is set as the height of the reference (Ομπι) . The image number corresponding to the reference height of the three dimension information is ten. It can be set the height of the tenth image as the reference.

[0073]

Because the image number of the image that is focused on the first fine foreign matter 101 is thirty, the apparatus can judge that the first fine foreign matter 101 is existed on the height which is 2. Ομπι above the reference. Since the image number of the image that is focused on the second fine foreign matter 102 is forty, the apparatus can judge that the second fine foreign matter 102 is existed on the height which is 3. Ομπι above the reference. Then, the apparatus can judge that the first fine foreign matter 101 is existed on the height which is 2. Ομπι above the electrode pattern. Further, the apparatus can judge that the second fine foreign matter 102 is existed on the height which is 3. Ομπι above the first foreign matter, that is, 5. Ομπι above the electrode pattern. The heights of the peaks of the contour graph as shown in FIG.34 indicate the three dimension information of the fine foreign matters in the inspection subject.

[0074]

In a case that the defects such as the fine foreign matters or the pinholes are detected, the quality judging section 60 analyzes the three dimension information of the defects. From the analysis result, the apparatus judges whether the defects are existed on the same sealing layer or not. In a case that the defects are existed on the same sealing layer, the subject, which is the organic EL display device is judged as a non-defective product. Then, the apparatus judges whether the plural defects are existed on the different heights (thickness direction) or not. In a case that the plural defects are existed on the different heights (thickness direction), the subject is judged as a defective product. As described above, as time proceeds, the oxygen and the water, which enter from the defect which is existed on the organic film, permeate the organic film.

Further, the oxygen and the water arrive at the defect which is existed below the organic film, and then the measuring subject (for example, the EL display device), which is existed right under the defect, is broken down. As a result, a lifetime of the measuring subject is shortened.

[0075]

Further, a means for repairing the defect may be added to the apparatus for judging the defect quality of the present invention. For example, in manufacturing the devices of the thin film multilayers such as the organic EL display device, the apparatus for judging the defect quality of the present invention can specify the layer that has the defect having the diameter of Ιμπι or less (for example the pinhole or the foreign matter), which cannot be judged by the conventional technology. Then, the optimal repairing means depending on the material of the sealing film that forms the layer, which the defect is existed, can be selected by using the apparatus for judging the defect quality of the present invention. In a case that the defect is the foreign matter of the organic film, the foreign matter is removed by using laser, and the film can be repaired. Further, the optimal repairing can be performed by selecting the wavelength [nm] of the laser light and the energy density [J/cm ² ] . In a case that the laser light cannot be used, the repairing method, which the foreign matter is pushed downward, can be selected. In a case that the defect is the pinhole, the following method can be adopted. The method is that by using a

cartridge, which a fine tip processing tube (a micro

dispenser) is attached, the micro amount film material is applied to the pinhole that is the defect, and then the film is cured by firing or ultraviolet irradiation.

[0076]

In the present embodiment, it is described that one image number which is the largest value of the edge pixel number of the pattern is detected, and the image number which is the largest value of the foreign matter pixel number and the image number which is the second largest value of the foreign matter pixel number (including in a case that the second largest value of the foreign matter pixel number is the same as the largest one) are detected. However, it is not limited to the above embodiment. For example, depending on the organic film, the inorganic film and the electrode pattern which constitute the sample, or depending on the size, the density, the occurring position and the like of the foreign matter and the defect, a modification can appropriately be added.

[0077]

The edge evaluating value E(i, j) and the foreign matter evaluating value F(i, j) can be stored in the edge evaluating value storing value EM(i, j) . Whether the

apparatus prioritizes the edge evaluating value E(i, j) or the foreign matter evaluating value F(i, j) may be determined based on, for example, the size and the sharpness of the partial image of the edge evaluating value E(i, j) and the foreign matter evaluating value F(i, j) .

[0078]

In the method and the apparatus for judging defect quality, in the imaging process, a position error in the horizontal direction among the acquired images is occurred by vibration of the apparatus itself or a floor. By detecting the image number that is the highest sharpness, without affecting the position error of the horizontal direction, the height information, which the fine defect is occurred, is accurately measured without the error. Then, vibration resistance characteristic of the overall apparatus can be expanded, and the above described process is largely

contributed to a cost reduction of the apparatus .

[0079]

In the above described embodiment, an example that the height measuring of the edge pixels of the pattern and the height measuring of the fine foreign matter are performed independently is described. In a case that the edge

evaluating value of the pattern and the fine foreign matter evaluating value are converted to the values in the same order of magnitude, the process from the edge evaluating value comparing section 43 to the edge three dimension information extracting section 46 and the process from the foreign matter evaluating value comparing section 53 to the foreign matter three dimension information extracting section 56 can be treated as the common process. By communalizing such the process sections, the apparatus for judging defect quality of the present invention can be simplified. As such an example, the apparatus comprises a feature extracting section which the pattern edge extracting section 41 and the fine foreign matter extracting section 51 are integrated, an evaluating value calculating section which the edge

evaluating value calculating section 42 and the foreign matter evaluating value calculating section 52 are

integrated, an evaluating value comparing section which the edge evaluating value comparing section 43 and the foreign matter evaluating value comparing section 53 are integrated, the evaluating value storing section which the edge

evaluating value storing section 44 and the foreign matter evaluating value storing section 54 are integrated, an image number storing section which the edge image number storing section 45 and the foreign matter image number storing section 55 are integrated, and a three dimension information extracting section which the edge three dimension information extracting section 46 and the foreign matter three dimension information extracting section 56 are integrated, and then the configuration of the apparatus for judging defect quality of the present invention can be simplified. The software processing can be performed in the edge process section 40, the foreign matter process section 50, and the quality judging section 60 except for the storing sections.

ALTERNATIVE EMBODIMENT

[0080]

FIG. 35 is a schematic side view of inspection

apparatus 500, which maps features in a sample 502, in accordance with an embodiment of the invention. Apparatus 500 operates on principles similar to the embodiments described above, with additions and variations that are explained below. As described in the preceding embodiments and shown, for example, in FIGS. 1-3, sample 502 comprises multiple thin-film layers, typically including transparent layers, which are overlaid on a surface of the sample.

[0081]

Apparatus 500 comprises a video camera 506, which captures electronic images of sample 502 via a lens 508, typically a microscope lens with high magnification, high numerical aperture, and shallow depth of focus. An

illumination source 504 illuminates sample 502 while camera 506 captures the images. In the present embodiment,

illumination source 504 emits light of a single color, i.e., light having a bandwidth no greater than 40 nm (full width at half maximum) . Such single-color illumination is

advantageous in eliminating the effect of chromatic

aberration in the images captured by camera 506. For

enhanced contrast of image features, it is also advantageous that illumination source 504 illuminate sample 506 in a dark field mode. Alternatively, however, illumination source 504 may emit white or other broadband light, and may provide bright field illumination.

[0082]

A motor 510 scans the front focal plane of camera 506 in a direction perpendicular to the surface of sample 502. The scan may be either continuous or step-wise. In the pictured embodiment, motor 510 translates camera 506 and lens 508 upward and downward. Alternatively or additionally, the motor may shift the vertical position of sample 502 or may adjust the focal setting of lens 508 in order to scan the focal plane. In the course of the scan, camera 506 captures a sequence of images of the thin-film layers on sample 502 at different, respective focal depths within the sample.

Consequently, features located at different depths within the sample will be in focus in different images in the sequence, wherein the sharpest focus occurs when the front focal plane of the camera coincides with the position of the feature. For features that extend over a range in the depth dimension

(i.e., the dimension perpendicular to the surface of sample 502), the top end of the feature may be sharply focused in one image and the bottom end sharply focused in another.

[0083]

A processor 512 processes the sequence of images captured by camera 506 over the course of a scan of motor 510 in order to identify features of interest in the images.

Such features may include, for example, defects within the thin film layers, as explained above. Processor 512 typically comprises a general-purpose computer processor, which has suitable interfaces for receiving electronic images from camera 506 and signals from other components of

apparatus 500, and is programmed in software to carry out the functions that are described herein. Alternatively or additionally, at least some of the functions of processor 512 may be implemented in programmable or hard-wired logic. Upon identifying a feature of interest, processor 512 computes the depth of optimal focus of the feature within the sequence of the images, and thus estimates the location of the feature - and specifically the location in the depth (vertical) dimension - within the thin-film layers. For this purpose, as explained above in detail, processor 512 computes a measure of the sharpness of edges of the feature in the images, and finds the depth that maximizes the sharpness.

[0084]

In the present embodiment, apparatus 500 comprises a rangefinder, including a laser 514 and a detector 516, which measures distance between camera 506 and sample 502. The pictured rangefinder operates by sensing shifts in the location of the laser spot that is reflected from sample 502 onto detector 516 as the distance between the camera and sample changes. Alternatively, other sorts of rangefinders that are known in the art can be used, such as ultrasonic or interferometric rangefinders. Processor 512 applies the distance measured by the rangefinder in estimating the locations of features of interest, and particularly in correcting for variations in the location of the front focal plane of camera 506 within the thin film layers on sample 502 that can occur, for example, due to the vibration of the sample. Processor 512 is able to detect this vibration based on periodic changes of the distance measured by the

rangefinder over time, and can then correct the depth measurements in the captured images so as to compensate for this vibration and thus estimate the positions of features of sample 502 with greater accuracy.

[0085]

FIG. 36 is a plot that schematically illustrates vibrations measured in apparatus 500, in accordance with an embodiment of the invention. Data points 520 in the plot indicate the height (in microns) of camera 506 above sample 502 relative to a baseline height, which would be expected in the absence of vibration, as a function of time (in seconds) . Each data point 520 corresponds to a reading made by

rangefinder detector 516. Processor 512 fits a periodic function to data points 520 and thus generates a curve 522, which gives the estimated amplitude of the vibration at any point in time. Camera 506 captures images at times that are indicated by marks 524 on curve 522. At each such time, processor 512 reads the value of curve 522 to give a height correction and adds (or subtracts) this correction value to the nominal depth, given by the scan of motor 510, in order to compute a corrected focal depth. Processor 512 is thus able to compensate for the vibration of sample 502 and more accurately estimate the locations of features that appear in the images .

[0086]

FIG. 37 is a plot that schematically illustrates measurements of focal quality made by apparatus 500, in accordance with an embodiment of the invention. Data points 530 correspond to the focus scores computed for a given feature as a function of the focal depth of camera 506 within the thin film layers on sample 502. The nominal focal depths may be corrected for vibration of the sample as explained above. The focus scores measure the sharpness of the edges of the feature of interest, for example based on the image derivative. The Z locations of data points 530 are corrected for the measured vibration and therefore may be unevenly spaced in the plot. The focus score as a function of the depth of the front focal plane of camera 506 has the form of an inverted parabola. Processor 512 thus fits a suitable curve 532 to data points 530 and finds the peak of curve 532, which indicates the depth of the feature within sample 502.

[0087]

FIG. 38 is a flow chart that schematically illustrates a method for mapping of features in a sample, in accordance with an embodiment of the invention. The method is described hereinbelow, for the sake of convenience and clarity, with reference to the features of apparatus 500 (FIG. 35) .

Alternatively, this method may be applied, mutatis mutandis, using the apparatus of the preceding embodiments or using any other suitable inspection system, as will be apparent to those skilled in the art after reading the present

description .

[0088] Processor 512 measures the distance of camera 506 and lens 508 from sample 502 using a rangefinder, such as laser 514 and sensor 516, at a distance measurement step 540.

Typically, apparatus 500 is constructed so that this distance remains substantially constant (other than small motions due to vibration) while motor 510 scans the depth of the front focal plane of the camera through the thin film layers on sample 502, at a scanning step 542. Alternatively, the rangefinder may measure the shift induced by operation of motor 510 in this step. Processor 512 acquires images of sample 502 from camera 506 as motor 510 scans the focus of the camera in the depth dimension.

[0089]

Based on the rangefinder measurements made at step 540, processor 512 reconstructs the pattern of vibration of sample 502, as illustrated in Fig. 36, for example, at a vibration reconstructions step 544. Processor 512 is then able to correct the nominal focal depths of the images acquired at step 542 to compensate for error induced by the vibration, at a depth correction step 546. Processor 512 identifies one or more features of interest in the images, for example, potential defects, and calculates focus scores for these features as a function of the corrected depth, at a focus scoring step 548. For each such feature, processor 512 fits a curve to the calculated focus scores, and thus finds the coordinates of the feature in three dimensions, at a position calculation step 550.

[0090]

It will be appreciated that the embodiments described above are cited by way of example, and that the present invention is not limited to what has been particularly shown and described hereinabove. Rather, the scope of the present invention includes both combinations and subcombinations of the various features described hereinabove, as well as variations and modifications thereof which would occur to persons skilled in the art upon reading the foregoing description and which are not disclosed in the prior art.

Explanation of Reference Numerals

[0080]

1 electric circuit pattern

2 first sealing layer

3 second sealing layer

4 third sealing layer

5 first base material

6 second base material

10 flexible organic electroluminescence (EL) display device

20 microscope (lens barrel section)

21 objective lens camera

height direction driving motor

sequence control section

image acquiring section

image memory

edge process section

pattern edge extracting section

edge evaluating value calculating section

edge evaluating value comparing section

edge evaluating value storing section

edge image number storing section

edge three dimension information extracting section foreign matter process section

fine foreign matter extracting section

foreign matter value calculating section

foreign matter value comparing section

foreign matter value storing section

foreign matter image number storing section foreign matter three dimension information

extracting section

quality judging section

first fine foreign matter

second fine foreign matter

electrode pattern

cathode electrode 105 organic film

106 transparent film

107 light emitting layer

500 inspection apparatus

502 sample

504 illumination source

506 camera

508 microscope lens

510 focus adjustment motor

512 processor

514 rangefinder laser

516 rangefinder detector