JP2010127815A - Defect detection selection correction system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a defect detection selection correction system including a means for selecting beforehand defects required to be corrected, because defects required actually to be corrected are few, though a considerably large number of defects on a color filter substrate are detected by an automatic defect inspection device. <P>SOLUTION: This defect detection selection correction system includes at least the automatic defect inspection device, a defect correction device, and an automatic statistical-correction system. The automatic statistical-correction system automatically determines whether each defect is to be actually corrected or not in consideration of a past correction history with respect to a defect positioned on the same portion and having a similar appearance characteristic. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a defect correction technique on a color filter substrate, and in particular, collective management of all information such as the position of defects in a color filter layer and appearance information characterizing the defects and information related to correction of the defects, and these The present invention relates to a defect detection selection correction system that realizes efficient defect correction using information.

  Color filter layers, overcoat layers, photo spacers, and alignment control protrusions formed on the substrate using photolithography and ink jet methods have pinholes (white spots), chippings, protrusions, and bridges of about 5 to 500 μmφ. Various local defects such as double coating and density unevenness occur. Whether or not there is a defect is determined by observing the entire color filter substrate using an automatic defect inspection device. If a defect is detected, the defect correction device is used thereafter to detect the defect. The defect is corrected to produce a color filter substrate having no defect (see, for example, Patent Documents 1 and 2).

  The automatic defect inspection device picks up the color filter substrate transported from the substrate transport unit with a CCD line sensor, detects the defective portion while enlarging the acquired image with the image processing device, and simultaneously detects the defect size and density. Appearance information such as shape is acquired. Specifically, an image of a certain size is divided in units of cells (for example, a set of red, green, and blue pixels) determined by the size of the color filter, and surrounds the image of the cell of interest and the cell. Density comparison processing is performed with the upper, lower, left, and right cells, binarization is performed with a set threshold value, a defective portion is specified, and defect information is stored as a data file (see, for example, Patent Document 2). This defect search is performed mechanically automatically, and it is considered that the process of judging the necessity of correction by visually observing the acquired image of each defect is rather inefficient.

On the other hand, the defect correction apparatus places a color filter substrate on a regular board, moves a correction unit including a microscope in a gantry shape by two-axis drive, detects a defect, and performs correction processing by appropriate means. (For example, refer to Patent Document 3). Normally, defect correction is performed by loading the substrate after the automatic defect inspection into the defect correction device through the substrate transfer unit, reading a data file related to the defect from the automatic defect inspection device using a LAN (local area network), The microscope unit equipped with the sensor is moved to the defect site within the range of the CCD field of view, and the defect is imaged and visually confirmed in almost the same procedure as the automatic defect inspection apparatus. Perform defect correction processing. The defect correction apparatus also includes a defect data file that holds information about defects.
JP-A-3-274504 JP 2007-327671 A JP 2001-166129 A

By the way, the number of defects of the color filter layer detected by the automatic defect inspection apparatus amounts to a considerable amount. However, defects that are actually required to be corrected are about 5% of defects detected empirically. In the current system, once it is inserted into the defect correction device, it is necessary to re-align with the CCD sensor for each detected defect, and determine whether or not the defect can be corrected while visually checking the defect image. Therefore, there is a problem that it takes a very long time to correct the defect.
Therefore, according to the present invention, not all defects detected by the automatic defect inspection apparatus are visually confirmed by the defect correction apparatus, but all the defects should be aligned by the defect correction apparatus and confirmed by an image. An object of the present invention is to provide a defect detection / selection / correction system including means for extracting defects in advance.

In order to achieve the above object, the invention according to claim 1 provides:
At least the coating film formed on the substrate is inspected, and information on the position of the detected defect on the substrate, information on the characteristics of the defect, instruction information on whether to correct the defect, etc. An automatic defect inspection apparatus that holds the first defect data file through a LAN as a defect data file;
After detecting the defect with the correction instruction on the substrate based on the information of the first defect data file, the defect is corrected if necessary, and the information regarding the position of the defect on the substrate and the actual correction are performed. A defect correction device that holds information as to whether or not the second defect data file is stored, and inputs and outputs the second defect data file through the LAN;
A third defect data file configured by categorizing the information related to the position included in the first defect data file and the information related to the characteristics of the defect, etc., by layer is stored, and the third defect data file The information on the first defect data file and the information on whether or not the second defect data file has been corrected for each inspection board obtained through the LAN is accumulated in the corresponding layer, and acquired using the accumulated information. An automatic statistics / correction system for setting instruction information as to whether or not to correct a defect in the first defect data file and outputting the instruction information to the automatic defect inspection apparatus, It is what.

  Such a system includes a third defect data file in addition to two defect data files related to the automatic defect inspection apparatus and the defect correction apparatus, both of which are connected via a LAN. The file includes a pocket (hereinafter referred to as a layer) finely divided into a range that can be taken by a factor that characterizes the defect, a defect data file from the automatic defect inspection device of the detected defect, and a defect correction device from the defect correction device. By combining defect data files, it is possible to know what defects were actually corrected and at what rate. With this correction rate, it is possible to set a threshold value for defect recognition in the automatic defect inspection apparatus and to set instruction information for the defect correction apparatus.

The invention according to claim 2 is the defect detection selection correction system according to claim 1, wherein the first defect data file includes at least the following information related to the defect.
Information related to position: Defect occurrence coordinate X [mm],
Defect occurrence coordinate Y [mm]
Information on characteristics: number of defective pixels [pix],
Defect long side size [pix],
Defect short side size [pix],
Defective pixel average density [0-255],
Defective pixel maximum density [0-255],
Defective pixel minimum density [0-255],
Others include board serial number and correction instruction information.

The invention according to claim 3 is the defect detection selection correction system according to claim 1, wherein the second defect data file includes at least the following information related to the defect.
Information related to position: Defect occurrence coordinate X [mm],
Defect occurrence coordinate Y [mm],
Other information, such as whether or not it has been corrected. Here, pix indicates the number of pixels.

  In the invention according to claim 4, the instruction information as to whether or not to detect the defect detected on the substrate is selected from one or more attributes defining the defect, and then from the third defect data file. The correction unnecessary rate of the layer to which the defect in the attribute belongs is calculated as the attribute correction unnecessary rate, and further, the fixed correction unnecessary rate obtained by multiplying the attribute correction unnecessary rates of all the selected attributes is set in advance. The defect detection according to any one of claims 1 to 3, further comprising: an algorithm for automatically comparing with a correction unnecessary threshold and determining whether or not the correction unnecessary threshold is exceeded. This is a sorting and correction system.

As described above, according to the present invention, by adding the automatic statistical sorting system to the automatic defect inspection apparatus and the defect correction apparatus, first of all, the number of processed color filter substrates belonging to the same lot is small and the defect data file When there is little defect correction data, if necessary, the color filter substrate that is put into the defect correction device without changing the software by forcibly turning on the correction instruction flag in the automatic statistical sorting system It is possible to inspect the detected defects and collect and accumulate data on whether or not they have been corrected in an automatic statistical sorting system.
Next, after a certain amount of correction data has been accumulated, the necessity of correction is automatically determined based on the accumulated correction history information for each detection defect of the color filter substrate circulating in the inspection. Thus, the correction instruction flag of the defect data file can be corrected and the number of defects that need to be corrected can be narrowed down. As a result, since only the narrowed defects are checked again and corrected, the time required for correction can be greatly reduced.
Further, for the automatic defect inspection apparatus, the number of defects can be narrowed down by resetting a threshold value in consideration of the correction rate. In any case, it is possible to provide a defect detection / selection / correction system that has the effect of improving the productivity of the defect correction apparatus and eliminating the need to increase the number of automatic correction apparatuses.

  Hereinafter, the function of the defect detection selection correction system according to the present invention will be described in detail.

  As shown in FIG. 1, the defect detection sorting / correcting system includes at least an automatic defect inspection apparatus 1, a defect correcting apparatus 2, an automatic statistics / sorting system 3, a transport system (not shown) for transporting a color filter substrate, and each It consists of LANs 5, 6, 7, and 8 for mutual communication of control signals for controlling the device and various information and data necessary for the device.

The information relating to the defects detected and collected by the automatic defect inspection apparatus 1 is the contents of the first defect data file 10 as shown below, with respect to the position, the appearance characteristics of the defects, and other necessary information. . Hereinafter, this information is referred to as a storage factor or a factor or attribute.
Coordinate X: Coordinate X [mm] where a defect is seen from the reference point,
Coordinate Y: Coordinate Y [mm] where a defect is seen from the reference point,
Area: Number of pixels related to defects [pix]
Size x: Defect length in the x direction [pix],
Size y: the length of the defect in the y direction [pix],
Size z: Defect height in the z direction [micron]
Defect average density [0-255]: Average density of defective pixels,
Defect maximum density [0-255]: maximum density of defective pixel,
Defect minimum density [0-255]: Minimum density of defective pixel,
Other: Correction instruction information [ON, OFF], serial number, board ID, total number of defects, etc.

Here, pix is an abbreviation of an English character of a unit pixel for digitizing the image data acquired by the CCD line sensor. Defects are x and y, which are the minimum four-sided pixel numbers surrounded by straight lines parallel to the x-axis and y-axis, including the defect, as viewed in this pixel unit. The coordinates of the center of the quadrilateral are represented by X and Y as the position coordinates of the defect from the reference point. The defect area is a product xy of two side lengths. The density indicates the level of the monochrome image of each pixel related to the defect with 256 gradations. The average density is a weighted average of the density of all pixels related to the defect. Maximum and minimum are literal meanings.

  These data are held in the computer memory 11 attached to the defect detection apparatus, and are accessed from the automatic statistics / sorting system 5 and 6 through the LAN. In addition, the correction instruction flag is set to ON for all detected defects that need to be corrected.

  The correction substrate 4 is conveyed to the defect correction device 2, and defect inspection information of the substrate is transmitted to the defect correction device 2 through the LAN 7 using the substrate ID as a key. According to this defect information correction instruction flag, the defective part is searched, an image around the defect is displayed, and it is determined again by inspection whether correction is necessary. The second defect data file 11 used in the defect correction apparatus can hold almost the same storage factor as the first defect data file 10, but in general, the defect location information and the defect for which the correction instruction has been given are actually used. Only the data indicating whether or not it was corrected is sufficient. The storage factor file is also accessed from the automatic statistics / sorting system 3 through the LAN 8.

  The automatic statistics / sorting system 3 associates the first defect data file 10 with the second defect data file 11 and records the information on whether or not the detected defect is actually corrected. A data file 12 is held. This file has a structure obtained by finely dividing (stratifying) the range in which each factor can take the plurality of storage factors. That is, in the case of the coordinate X, if the length of the color filter to be inspected in the x direction is 700 mm, the color filter is divided into seven layers obtained by dividing the coordinate X in units of 100 mm. Of course, it may be divided into 70 layers of 10 mm units. The number of layers is appropriately set according to the inspection object. If it is a density | concentration, it will be divided | segmented into a maximum of 256 layers. The layer structure may be one-dimensional or multi-dimensional. For example, a combination of a defect area factor and a concentration factor. In this case, the defect area factor can be divided into 10 layers, and each layer can correspond to 256 layers having an average concentration. In this case, a two-dimensional file is formed. When a defect is specified by a plurality of factors, the multi-dimensional file is prepared.

After preparing the above three defect data file files, the defect correction process is performed in the following procedure.
First, the substrate 4 on which some film is formed is carried into the automatic defect inspection apparatus 1. The formed film may be a substrate on which any of black matrix, red, blue, and green colored pixels is formed, a substrate on which a plurality of films are laminated, or a final color filter substrate. . The automatic defect inspection apparatus 1 writes specific numerical values related to the plurality of storage factors to the first defect data file 10 and stores them in a computer memory attached to the automatic defect inspection apparatus with respect to the detected defects, and automatic statistics. Transfer to the sorting system 3 via the LAN.

Then, the automatic statistics / sorting system selects one of the layers belonging to the specific factor including the numerical value based on the numerical data related to the specific factor, and, for example, according to a regular method of the count number of this layer + 1 (In the initial state of the lot, it is set to zero). This is done for all factors. Next, the substrate 4 is transported to the defect correction apparatus 2, but the defect correction apparatus determines whether or not correction is necessary by inspection for a defect whose correction instruction information is ON (for correction). If it is determined that it is necessary and correction processing is performed, the correction flag of the second defect data file 11 is set to ON. If the correction instruction flag is ON but not corrected, the correction flag is turned OFF. The information thus obtained is held in the computer of the defect correction apparatus as the second defect data file 11 and is accessed from the automatic statistics / sorting system 3 through the LAN.

Based on this information, the automatic statistics / sorting system 3 increments the count number of the area in the file holding the number of corrections corresponding to the defect. In this way, whether or not a defect specifically identified by layering has been corrected is recorded, and when this is repeated for a substrate that has been input, correction data will accumulate.
In an automatic defect inspection device, the correction instruction flag is normally set to ON. However, in the automatic statistics / sorting system, the correction instruction flag can be automatically reset by a certain algorithm based on accumulated data. Is possible.
For example, it is a case where the part is a defect but has no harm and no correction is required. In the case where the number of inputs is small, it is conceivable that the automatic resetting mechanism does not work, and all of them are left to the inspection judgment of the defect correcting device to make efforts to accumulate data. These can be executed by software on a computer.

  The first defect data file 10 transmitted from the automatic defect inspection apparatus 1 is sent back to the automatic defect inspection apparatus 1 after the correction instruction flag is reset, and this information and the color filter substrate 4 are transferred to the defect correction apparatus 2. Will be sent to. However, sending to the automatic defect inspection apparatus 1 may be said to be substantially sent to the defect correction apparatus 2.

  The third defect data file can be used as a reference for setting a threshold. For example, if the correction rate of a defect of a specific appearance at a specific position is extremely low, it is considered that this defect does not need to be corrected. Accordingly, the correction instruction information can be automatically set to OFF on the automatic defect inspection apparatus 1 side.

Finally, an algorithm used when writing the instruction information on whether or not to correct the defect Z to the first defect data file using the accumulated third defect data file will be described.
First, an attribute (factor) that defines a defect is determined. Which attribute is used is selected and set on the system by the person to be inspected. It can also be set in advance according to the defect characteristics.
As the attribute A, the defect X coordinate and the defect size are selected and stratified as follows.
Layer 1: X coordinate; 0-9mm Size; 0-9mm □
Layer 2: X coordinate; 0-9 mm Size; 10-19 mm □
Layer 3: X coordinate; 10-19 mm Size; 0-9 mm □
Layer 4: X coordinate; 10-19 mm Size; 10-19 mm □
For each layer, it is easily calculated whether the correction unnecessary rate is stored in the third defect data file. The values are layer 1 → A1%, layer 2 → A2%, layer 3 → A3%, layer 4 → A4%.

As the attribute B, the average defect density and the defect size are selected and stratified as follows.
Layer 1: Average concentration; 0-19 Size; 0-9 mm □
Layer 2: Average concentration; 0-19 Size; 10-19 mm □
Layer 3: Average concentration; 20 to 29 Size; 0 to 9 mm
Layer 4: Average concentration; 20 to 29 Size; 10 to 19 mm □
For each layer, it is easily calculated whether the correction unnecessary rate is stored in the third defect data file. The values are layer 1 → B1%, layer 2 → B2%, layer 3 → B3%, layer 4 → B4%.

As the attribute C, the Y coordinate of the defect and the defect size are selected and stratified as follows.
Layer 1: Y coordinate; 0-9mm Size; 0-9mm □
Layer 2: Y coordinate; 0-9 mm Size; 10-19 mm □
Layer 3: Y coordinate; 10-19 mm Size; 0-9 mm □
Layer 4: Y coordinate; 10-19 mm Size; 10-19 mm □
For each layer, it is easily calculated whether the correction unnecessary rate is stored in the third defect data file. The values are layer 1 → C1%, layer 2 → C2%, layer 3 → C3%, and layer 4 → C4%.

If the layer corresponding to the attribute A on the third defect data file of the defect Z is 1, the attribute correction unnecessary rate is A1. Similarly, in attribute B, the attribute correction unnecessary rate is B3, and in attribute C, the attribute correction unnecessary rate is C4. As a product of the three attribute correction unnecessary rates, a fixed correction unnecessary rate K is calculated as K = A1 × B3 × C4.
This fixed correction unnecessary rate is compared with a correction unnecessary threshold that can be set empirically in advance, and it is determined whether or not to correct, and instruction information on the first defect data file is set. Of course, it is possible to automatically compare and set.

It is drawing explaining the concept of the defect detection selection correction system which becomes this invention.

Explanation of symbols

1. Automatic defect inspection device 2, defect correction device 3, automatic statistics / sorting system 4, inspection board 5, 6, 7, 8, data transfer LAN
10, 11, 12, defect data file

Claims (4)

At least the coating film formed on the substrate is inspected, and information on the position of the detected defect on the substrate, information on the characteristics of the defect, instruction information on whether to correct the defect, etc. An automatic defect inspection apparatus that holds the first defect data file through a LAN as a defect data file;
After detecting the defect with the correction instruction on the substrate based on the information of the first defect data file, the defect is corrected if necessary, and the information regarding the position of the defect on the substrate and the actual correction are performed. A defect correction device that holds information as to whether or not the second defect data file is stored, and inputs and outputs the second defect data file through the LAN;
A third defect data file configured by dividing information relating to the position included in the first defect data file and information relating to the characteristics of the defect, etc., into respective layers is held, and the third defect data file The information on the first defect data file and the information on whether or not the second defect data file has been corrected for each inspection board obtained through the LAN is accumulated in the corresponding layer, and acquired using the accumulated information. An automatic statistics / correction system for setting instruction information as to whether or not to correct a defect in the first defect data file and outputting the instruction information to the automatic defect inspection apparatus, . 2. The defect detection / selection / correction system according to claim 1, wherein the first defect data file includes at least the following information related to the defect.
Information related to position: Defect occurrence coordinate X [mm],
Defect occurrence coordinate Y [mm]
Information on characteristics: number of defective pixels [pix],
Defect long side size [pix],
Defect short side size [pix],
Defective pixel average density [0-255],
Defective pixel maximum density [0-255],
Defective pixel minimum density [0-255],
Others include board serial number, correction instruction information, total number of defects, etc. 2. The defect detection / selection / correction system according to claim 1, wherein the second defect data file includes at least the following information relating to the defect.
Information related to position: Defect occurrence coordinate X [mm],
Defect occurrence coordinate Y [mm],
Other information, such as whether or not the defect has been corrected. Here, [pix] indicates that the unit is a pixel.   The instruction information on whether or not to correct the defect detected on the substrate is selected from one or more attributes that define the defect, and the defect in the attribute belongs from the third defect data file. The layer correction unnecessary rate is calculated as the attribute correction unnecessary rate, and further, the value obtained by multiplying the attribute correction unnecessary rate of all the selected attributes is set as the fixed correction unnecessary rate, and this is set with the preset correction unnecessary threshold. The defect detection selection correction system according to any one of claims 1 to 3, further comprising an algorithm that automatically compares and determines whether or not a correction unnecessary threshold has been exceeded.

Images