JP2010085210A - Defect inspecting device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an inspecting device to reduce a measurement time in a defect inspection greatly. <P>SOLUTION: The defect inspecting device to inspect a defect on the pattern of an object to be measured with a regular pattern arranged on a whole surface thereof, includes two imaging means to image the object to be measured, with a constant distance between the imaging means, a moving means to relatively move the imaging means and the object to be measured, a first operating means to extract the minimum means of the pattern on each of two images taken by the two imaging means and then measure the central position of the minimum means, and a second operating means to detect the presence of the pattern by calculating the variation of the difference value between two coordinates measured by the first operating means and determining the variation with a predetermined threshold. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an inspection apparatus capable of greatly shortening a measurement time when inspecting a defect due to a minute change in the pattern on a substrate on which a regular pattern such as a liquid crystal color filter is arranged.

  In a color filter for a liquid crystal panel, a black matrix (hereinafter referred to as BM) is usually formed with high accuracy with a regular pattern, but when a defect called a dimple defect occurs, it is highly accurate and regular. A slight change in the pattern may occur, resulting in a defective color filter.

The dimple defect is a defect caused by a minute change in a pattern caused by a foreign matter mixed in the exposure apparatus. (See Figure 2)
The cause of the dimple defect is, as shown in the sectional view of the inside of the exposure apparatus in FIG. 2 (a), in a state in which the foreign matter 204 is caught between the exposure table 201 and the substrate 202, and the ultraviolet rays are passed through the exposure mask 203. When the substrate 202 is irradiated with 205, the irradiation direction is not perpendicular to the substrate in the region where the foreign matter 204 exists, that is, the region where the substrate 202 is curved with respect to the exposure table 201, so that there is a slight change in the regular pattern. Will occur.

  The minute change of the pattern is due to a change in the size of a unit (hereinafter referred to as a cell), which is the initial unit for forming a pattern, from a design value, and as shown in FIG. In the BM formation substrate 206 where the defect exists, due to the irradiation direction to the substrate, the size of the cell 207 which is the normal part, such as the thick line of the cell 208 and 209 which is a dimple defect part, or the dotted line of the cell 208 and 209 It is a minute change caused by contraction or expansion rather than size.

In general, as with inspection machines used for defect inspection such as white pins and black pins, the size of adjacent cells is measured to calculate the difference value, and the obtained difference value is used to inspect the defect with a predetermined threshold. The method cannot detect dimple defects. A dimple defect is a phenomenon in which the cell size changes due to shrinkage or expansion, but the amount of change is so small that the difference in size between adjacent cells falls within the variation in measurement accuracy of the inspection equipment. is there. Therefore, a general inspection method using a difference value between adjacent cells cannot be adopted for a dimple defect.
Conventionally, measurement using a length measuring instrument has been used in the inspection of a dimple defect generated on a BM forming substrate of a liquid crystal color filter.

  The inspection method by the length measuring instrument measures the absolute distance from the reference cell on the substrate to the measurement position, obtains the difference between the actual distance and the measurement distance, and determines the quality of the dimple defect according to a predetermined threshold value. I was going. Specifically, the area sensor camera head moves to the measurement cell coordinate position from a center coordinate of a certain reference cell to a cell that is a predetermined distance away, and stops at the measurement cell position by a highly accurate linear scale. After that, the area sensor camera captures the BM pattern of the substrate by applying autofocus in order to capture a clear image. The center coordinate of the measurement cell in the image is measured from the obtained image, and the obtained measurement result and the absolute distance from the center coordinate of a certain reference cell to the center coordinate of the measurement cell are calculated. When the obtained difference between the absolute distance and the actual distance exceeds a predetermined threshold, it is detected as a dimple defect.

  In the inspection method, since the absolute distance is used as a comparison factor, the minute change amount of one cell of the concave defect portion, which is a characteristic of the concave defect, is increased by integration, and an inspection that can be recognized and detected as a concave defect for the first time there. Is the method. Furthermore, in the dimple defect inspection by the length measuring machine, it can be said that it is suitable to set the absolute distance as small as possible in order to detect a minute change in the pattern without overlooking it.

However, with the recent increase in size of liquid crystal panels, the color filter substrate has become larger, so that the number of measurements per substrate has to be increased.
Furthermore, the conventional dimple defect inspection method has a problem that it takes time for one measurement because it stops at the measurement position and performs autofocus every measurement.
As described above, there is a problem that the measurement time is significantly increased due to a large number of measurements and a long time required for one measurement.

Further, as a method for inspecting a dimple defect, for example, there is JP-A-2007-24574 (Patent Document 1). Patent Document 1 is an inspection method in which a stripe pattern is imaged and inspected by a line sensor when inspecting a dimple defect generated on a substrate having a stripe pattern. The substrate is always arranged so that the stripe pattern is perpendicular to the transport direction. This is an inspection method for recognizing the direction of the stripe pattern with respect to the substrate and rotating the substrate with a conveying means holding the substrate rotation mechanism.
However, in the dimple defect inspection method described above, the BM pattern of the inspectable substrate is limited to a stripe pattern, and the substrate such as a lattice pattern is not an inspection target.
JP 2007-24574 A

  It is an object of the present invention to provide an inspection apparatus that can significantly reduce the measurement time in the inspection of a dimple defect.

In order to achieve the above object, the present invention provides the following inspection apparatus.
1) An inspection apparatus for inspecting minute defects in a pattern of an object to be measured in which regular patterns are arranged on one surface, and the distance between two imaging means for imaging the object to be measured is constant. Two imaging means, a moving means for relatively moving the imaging means and the object to be measured, and each of the two images taken by the two imaging means at a predetermined position in the image field of view. Extracting a minimum unit part that forms a pattern, and further measuring a central position of the minimum unit part by coordinates within an image field; and two measured by the first calculation unit An inspection apparatus comprising: a second calculation unit that takes a difference between coordinate values, calculates a change amount of the difference value, and detects the presence or absence of a minute defect in the pattern by a predetermined threshold determination.
2) The inspection apparatus according to 1), wherein the imaging unit captures images synchronously.
3) The inspection apparatus according to 1), wherein the moving unit picks up an image while moving when the image pickup unit picks up an image.
4) The inspection apparatus according to 1), wherein the object to be measured has a rectangular shape.
5) The inspection apparatus according to 1) or 4), wherein in the imaging unit, a constant distance between the imaging units is set to a half of one side of the object to be measured.
6) The inspection apparatus according to claim 1 wherein the object to be measured is a color filter.

  The present invention compares the measured values of images taken simultaneously using two imaging means when inspecting a dimple defect on a substrate on which a pattern having regularity as described above is arranged. Even with inexpensive moving means that can offset the movement error of the means, high-accuracy measurement is possible as in the prior art, and the imaging means with a constant distance between the two units can capture images while moving. An inspection apparatus that can greatly reduce the measurement time is provided.

Embodiments of the present invention will be described with reference to the drawings.
Fig. 1 is a schematic diagram of the dimple defect inspection device, Fig. 2 is a schematic diagram of the mechanism of the dimple defect occurrence, Fig. 3 is a color filter in which the dimple defect has occurred, and Fig. 4 is a flowchart of the dimple defect inspection. Fig. 5 shows a flowchart of the processing in the inspection part of the dimple defect inspection, Fig. 6 is an image obtained by simultaneously imaging the substrate with two cameras (camera 1 and camera 2), and (a) two (B) shows an image obtained by simultaneously capturing the normal portion of the two cameras and the camera 2 simultaneously capturing the dimple defect portion.

  Referring to FIG. 1, a table 12 for supporting a substrate 11 is provided. Above the table 12, a CCD camera 14a of a first area sensor and a CCD camera 14b of a second area sensor are two area sensor CCDs. A stage 13 is provided that can move the camera in the X direction with the distance fixed, and can also move in the Y direction.

  An optical fiber 18 is connected to a white light source 17 having a strobe function of a xenon lamp as an illumination means, and 18a and 18b are connected to first and second area sensor CCD cameras 14a and 14b, respectively. ing. Further, lenses 19a and 19b are connected to the lower surfaces of the first and second area CCD sensor cameras, respectively.

  Image data of images taken by the first and second area sensor cameras is transmitted to the inspection unit 16 via the camera cables 15a and 15b. The inspection unit 16 performs image processing on the received two-sheet image data in the first calculation unit, and the obtained result is transmitted to the second calculation unit, and the second calculation unit determines the quality of the dimple defect.

An inspection method in the apparatus of FIG. 1 will be described based on the flowchart of FIG.
After the substrate 11 is placed on the stage 12 so as to be stably supported, the measurement is started. In the stage movement of S1 in FIG. 4, the movement stage 13 supporting the area sensor cameras 14a and 14b starts to move in the X direction at a constant speed. Thereafter, in the simultaneous imaging of S2, the area sensor cameras 14a and 14b simultaneously capture images every certain predetermined time. At the moment of the imaging, the white light xenon strobe lamp light source 17 which is the reflected illumination of the coaxial incident light strikes the strobe and takes a clearer image.

The image data of the captured image is transmitted to the inspection unit 16 via the camera cable as indicated by S3, and image processing, measurement, pass / fail determination, etc. are performed in the inspection unit. The processing contents of the inspection unit 16 are as shown in the flowchart of FIG. From the images simultaneously captured by the two area cameras CCD cameras, the corresponding cells in the images are extracted by the first calculation unit of the inspection unit 16, respectively. It is preferable to extract a plurality of corresponding cells in the two images, and in particular, adjacent cells are extracted in order to improve the inspection accuracy. After extracting the corresponding cells in the two images, the coordinate values of the corresponding cells are measured, and the coordinate values of the BM edge positions are measured.
The coordinate values of the two images measured by the first calculation unit are transmitted to the second calculation unit. The second calculation unit calculates the difference between the coordinate values of the corresponding cells in the first and second images. When there are a plurality of extracted corresponding cells in the first calculation unit, the difference is calculated by the number of corresponding cell pairs.

  The second calculation unit measures the amount of change due to the number of measurements of the difference value of each target cell coordinate between the two images measured by the first calculation unit. As shown in FIG. 6 (a), in the case of an image captured in the normal part, the difference value of the X coordinate of each target cell in the two images is zero. This difference value does not change as long as the two area sensor CCD cameras measure a normal pattern. However, the image in FIG. 6 (b) is an image of a dimple defect taken by one of the two area sensor CCD cameras. Since the position of the target cell slightly changes in the dimple defect portion, the difference value is not zero. In the measurement of the normal part of the measurement substrate, the same difference value is always measured, whereas in the difference value when the dimple defect part is measured by one camera, a change appears. A dimple defect can be searched by measuring this amount of change.

The inspection determination unit will be described based on FIG. Assume that attention is paid to a cell having a lattice pattern located at the center of images simultaneously photographed by the camera 1 and the camera 2 in FIG. At this time, regarding the coordinates of the cell of interest, the camera 1 is (i ₁ , j ₁ ), the camera 2 is (i ₂ , j ₂ ), and the gray value of the cell of interest is f (i ₁ , j ₁ ) and f ( i ₂ , j ₂ ). Here, in order to reduce the influence of noise and the like, the calculation was performed with the measurement range up to 8 cells in the vicinity of the cell of interest (shaded cells in FIG. 6B). The difference value S between the two images is as follows.

  If the number of measurements is t and the determination threshold L is the difference value S obtained by the above equation, the determination formula for the dimple defect inspection is as follows.

When the above equation is satisfied, it is determined that the measured substrate has a dimple defect.
The best mode of each means of the invention will be described. In the present invention, in a substrate on which a BM pattern of a liquid crystal color filter is formed, an inspection is performed on a substrate having a defect in which a change from a design value occurs due to continuous shrinkage or expansion of cells in a region where the pattern is present. The present invention relates to an inspection apparatus that detects The regular pattern refers to a lattice pattern, a stripe pattern, or the like.

  In the two image pickup means in the present invention, it is often preferable that the area sensor cameras as the respective image pickup means are fixed to the same camera head and the distance between the two is always constant.

  In addition, the imaging means does not require an accurate reference image that serves as a master image, and includes a plurality of imaging means and compares the captured images. Is preferred. Furthermore, it is preferable to increase the number of imaging means in order to shorten the measurement time.

The two imaging units are preferably provided with telecentric lenses each having a deep depth of field in order to prevent the influence of defocusing for imaging while the imaging unit or the substrate is moving. More preferably, the telecentric lens is a high-magnification lens so that a minute change can be clearly detected in order to detect a defect that is a minute change. The telecentric lens is preferably provided with coaxial epi-illumination and strobe illumination in order to capture a clear image even with a short exposure.
In the area sensor camera, lens, and illumination, it is preferable to increase the number of heads to an even number in order to further shorten the measurement time.

  Further, it is preferable that the moving means can be measured in a form provided on either the camera head or the substrate side, can be stably arranged, and moves at a constant speed and in a constant direction at the time of measurement. The moving means may be an inexpensive moving means if the above conditions are satisfied.

The first computing means can recognize the shape of the pattern in the image picked up by the image pickup means, and can recognize the minimum unit that forms the pattern and the boundary line formed by the unit. Further, it is preferable that a predetermined coordinate position can be searched in the coordinate system within the image field of view, and the coordinate position of the boundary line forming the searched unit can be measured. Preferably, for the purpose of reducing the influence of noise due to disturbance, the minimum unit is not a single unit but a predetermined minimum unit and a plurality of adjacent units are included in a predetermined target range for processing.
Further, since the first calculation means captures images for the number of cameras in one measurement, the processing in FIG. 6 is processed for all the number of cameras per measurement.

  In the second calculation means, data equivalent to the number of cameras is taken from the first calculation means in the same manner as the first calculation means. In the second calculation means, the difference between the coordinate values calculated by the first calculation means is calculated for each camera between corresponding units in the image field of view. When a predetermined unit is defined as a predetermined area of a plurality of units, a plurality of difference values are also calculated by comparing two images.

  In the above case, it is preferable to average one difference value so that the difference value obtained by comparing two images is one. The second computing means preferably has a function of storing difference value data obtained for each measurement in order to inspect the presence or absence of a dimple defect using a predetermined threshold from the amount of change in the difference value.

  3. The two area sensor cameras according to claim 2, wherein the two area sensor cameras can cancel the movement error of the moving stage by imaging at the same timing. Are preferably synchronized.

Furthermore, it is preferable that the strobe illumination is emitted in synchronization with the photographing timing when the two image pickup means picks up images. The exposure time of the image pickup means and the irradiation time of the strobe light are clearly related to the speed of the moving means. It is preferable to set the optimum setting so that an image can be captured.
In the present invention, the images taken by the two area sensor cameras, if the BM pattern is not captured on the entire surface of the image, the difference value is greatly calculated and erroneous detection occurs. Therefore, in many cases, it is preferable that the measurement object is rectangular so that the two imaging means can image the same pattern.

  An object of the present invention is to reduce the measuring time with an inspection apparatus that detects a dimple defect that is a deviation of a minute change. The distance between the cameras that is constant in the two image pickup means is often preferably half the distance in the longitudinal direction of the substrate to be measured or the river width direction in order to greatly reduce the measurement time.

  In the present invention, an accurate reference image serving as a master image is not required, and therefore, measurement accuracy equivalent to that of the prior art can be obtained without requiring absolute position accuracy as in the prior art. Can do.

  The present invention is a defect inspection apparatus for a liquid crystal color filter, but is not limited to a liquid crystal color filter, and can be applied to a TFT substrate or a PDP (plasma display panel) back substrate.

Example 1
Hereinafter, the present invention will be described with reference to examples. In addition, this invention is not limited to an Example. In the dimple defect inspection of Example 1, the substrate 11 on which the pattern formation of the BM which is the lowermost layer of the liquid crystal color filter was completed was placed on the moving stage 13 and SG SP 26-200 manufactured by Sigma Kogyo Co., Ltd. The moving stage speed was set to 10 mm / sec. Both area sensor CCD cameras used CV-200M manufactured by Keyence Corporation. The area sensor CCD cameras 14a and 14b are equipped with a telecentric lens 6x TV-6F manufactured by Opto Art, respectively, and the telecentric lenses 19a and 19b are respectively 18a and 18b, made by Keyence Corp., forked fiber 93089. Were used, and a coaxial incident xenon strobe lamp 17 and FS-025J60 manufactured by Nisshin Kogyo Co., Ltd. were used.

  The light emission time of the strobe illumination was 2.5 μm with a half width. Image data captured by the two area sensor cameras 14a and 14b are transmitted to the image processing device CV-3500 by the camera cable OP51499 manufactured by Keyence Corporation 15a and 15b, respectively, and are subjected to calculations such as image processing and are calculated. The inspection data 16 was subjected to image processing and pass / fail measurement by the OPTIPLEX GX280 manufactured by Personal Computer DELL. A dimple defect inspection was carried out using the above apparatus.

  In Example 1, the BM pattern of the substrate inspected by the dimple defect inspection is a lattice pattern, and the shape of one cell is shown in FIG. It becomes a shape surrounded by BM around the cell. One cell has a horizontal pitch 71 of 100 μm and a vertical pitch 72 of 300 μm. A substrate in which cells having a BM line width of 20 μm in width 73 and 7 μm in length 74 was arranged with regularity was used.

  The results of Example 1 are shown in FIG. In FIG. 8, the horizontal axis represents the position where the camera has moved, and the vertical axis represents the difference value calculated from images captured by two area sensor cameras. Referring to FIG. 8, a square plot is a normal sample, and a triangular plot is a sample including a dimple region. Referring to FIG. 8, signal undulations are severely generated in the area of the dimples, and a defective portion can be identified. In particular, when scanning one defective part with two cameras, the difference between the BM coordinates captured by both cameras when one camera measures the abnormal part first and then the other camera measures the abnormal part. You can see how the value changes twice, plus and minus. When the BM coordinate difference change amount exceeded 1 μm, it was decided to recognize and detect a dimple defect. In the result of the example, it is clear that the difference change amount of the BM coordinate is 1 μm or more, and therefore, it can be detected that the defect is a dimple defect.

  In Example 1 relating to the present invention, the measurement time for one measurement was 0.4 seconds. The size of the substrate used in the example was 360 × 460 mm, and measurement was performed once every 3 mm. The number of times of measurement of the entire surface of the measurement board was 18360, and the time required for measurement of the whole board surface was 7.6 hours. Although the measurement time required for one measurement is 0.4 seconds, since the measurement substrate is measured while moving, the inspection time is substantially equal to the time required for scanning the entire surface of the substrate. Therefore, the measurement time can be greatly shortened by taking measures such as increasing the number of cameras or improving the moving speed of the moving unit.

Comparative Example 1
Using the same sample as in Example 1, the measurement interval, the speed of the moving unit, and the number of measurements were made the same, and the measurement was carried out with a length measuring machine which is a conventional measurement method. The time required for one measurement was 4.7 seconds. The time required to measure the entire surface of the substrate was 24.0 hours.

  Comparing Example 1 and Comparative Example 1, it can be seen that the measurement time is significantly shortened in Example 1. In Comparative Example 1, the measurement time was greatly increased by positioning and focusing until the camera was placed at the measurement position and picked up images. However, in Example 1, two cameras were installed and two images taken. In order to compare each other, it is not necessary to place the camera at an accurate absolute position as in the first comparative example, so that it is possible to reduce the time required for positioning and focusing.

It is a figure which shows the whole structure of a dimple defect inspection apparatus. (a) is a figure explaining the mechanism of the occurrence of a dimple defect, and (b) is a figure showing the change in the size of the cell of the color filter in which the dimple defect has occurred. It is a figure which shows the color filter which the dimple defect generate | occur | produced. It is a figure which shows the process of an indentation defect inspection. It is a figure which shows the process in the test | inspection part of a dimple defect test | inspection. It is a figure which shows the image imaged by a dimple defect inspection. It is a figure which shows the cell shape of the board | substrate used for the Example. FIG. 4 is a diagram showing the results of Example 1.

Explanation of symbols

11,202: Board
12: stand
13: Moving stage
14,14a, 14b: Camera
15,15a, 15b: Camera cable
16: Inspection department
17: Light source
18,18a, 18b: Optical fiber
19a, 19b: Lens
201: Exposure table
203: Exposure mask
204: Foreign object
205: UV
206: BM
207,208,209,41,42: Cell
S1-S7, S31-S38: Step

Claims (6)

An inspection apparatus for inspecting a defect in a pattern of an object to be measured in which a regular pattern is arranged on one surface, wherein two imaging means for imaging the object to be measured have a constant distance between the imaging means The minimum unit part of the pattern is extracted for each of the two images captured by the two imaging units, and the imaging unit, the moving unit that relatively moves the imaging unit and the object to be measured, Calculating a change amount of a difference value between two coordinates measured by the first calculation means for measuring the center position of the minimum unit and the first calculation means, and then setting a predetermined threshold value for the change amount; And a second calculation means for detecting the presence or absence of a defect in the pattern by performing the determination used. The defect inspection apparatus according to claim 1, wherein the two image pickup units pick up images synchronously. The defect inspection apparatus according to claim 1, wherein the image pickup unit picks up an image while moving the image pickup unit when picking up an image. The defect inspection apparatus according to claim 1, wherein the shape of the object to be measured is a rectangle. The defect inspection apparatus according to claim 4, wherein a distance between the two imaging units is equal to half of the length of any side of the object to be measured. 2. The defect inspection apparatus according to claim 1, wherein the object to be measured is a color filter substrate.

Images