JP2007333584A - Tft array inspecting device and tft array inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce calculation load in the smoothing processing of two-dimensional arrangement data, and to have the throughput of the inspection of a TFT array substrate improved by reducing the calculation load. <P>SOLUTION: This TFT array inspection method, for inspecting a TFT array by irradiating the TFT array substrate with an electron beam, to detect secondary electrons produced from the pixel of the TFT array substrate by the irradiation with the electron beam, includes a background calculation process for calculating a background from the two-dimensional arrangement data of a process target region, and a data processing process for calculating the flaw data of the TFT substrate using the background. The background calculation process includes a first calculation process for calculating the first background of a block which divides the processing target region from the two-dimensional arrangement data contained in the block, a second calculation process for calculating the second background of the other point in the block from the calculated background of the block, and a third calculation process for calculating the background of the processing target region as a whole. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a TFT array inspection apparatus and a TFT array inspection method, and more particularly to data processing for inspecting a defective pixel using measurement data.

  As a configuration in which TFTs (thin film transistors) are arranged in an array, for example, there is a liquid crystal substrate, which is used for a flat panel display (FPD) such as a liquid crystal display.

  Conventionally, a TFT array substrate inspection apparatus irradiates a TFT substrate with an electron beam or light, measures the potential state of the TFT substrate, and detects defects in the TFT array by detecting pixels having an abnormal potential on the TFT substrate. .

  In the TFT array substrate device, two-dimensional measurement data (two-dimensional array data) corresponding to the substrate pixel array of the TFT substrate to be inspected is acquired, and defective pixels are extracted by processing this two-dimensional array data. Then, defect inspection is performed with respect to the coordinates of defective pixels, classification of defect types, and the like (for example, see Patent Document 1).

  For this data processing, an algorithm of low-pass filter processing is used. FIG. 6 is a schematic diagram for explaining an algorithm of low-pass filter processing.

  In FIG. 6, the low-pass filter processing is performed by, for example, obtaining a background 203 by removing high-frequency components due to defects and noise from the two-dimensional array data 201 by the background calculation unit 202, and using this background from the original two-dimensional array data 201. Defects and noise are extracted by subtracting 203, and this output is normalized with reference to 100 by a threshold value in the defective pixel inspection 204, thereby performing defect detection to obtain inspection output information 205.

  The functional role of this low-pass filter process is to smooth the two-dimensional array data and calculate the background. In order to obtain sufficient smoothing accuracy in this smoothing process, it is necessary to use a region of interest (ROI: Region of Intention) for a sufficiently wide region. Performing the smoothing process on a wide area of interest means performing a process on a large number of data, which means that the calculation burden is heavy. That is, when the entire region to be processed is set as the region of interest and the two-dimensional array data 201 of this region to be processed is smoothed by normal two-dimensional image image data processing, the amount of calculation becomes enormous if the entire region is processed at once. This is a heavy load on the computer.

  As a method of reducing the amount of calculation required to calculate the background, a calculation area of a small two-dimensional array is set in the processing target area, the background is calculated as the attention area, and the attention area is sequentially moved. In some cases, the background in the processing target area is smoothed.

  FIG. 7 is an explanatory diagram for explaining the calculation of the background. The processing target area 1 has a plurality of data points 2 arranged two-dimensionally (FIG. 7A). For each of the data points 2, a calculation area 30a composed of a plurality of data points including the data point 2 is set (FIG. 7B). For the plurality of data in the calculation area 30, for example, median The data value of the center value is obtained by the processing, and the obtained data value is set by replacing the original data value of the data point 31a in the calculation area 30a (FIG. 7C). The process of setting the calculation area and replacing the data value calculated by the median process is sequentially repeated for each of the data points 2 (FIGS. 7D to 7G). Thus, the background of the processing target area 1 is calculated (FIG. 7 (h)).

  In this case, although the amount of calculation is smaller than when all the areas are processed at once, it is necessary to perform median processing for all the data points 2 in the processing target area 1, so that a considerable amount of calculation is still required. . For example, if the total number of data points in the processing target area 1 is n, it is necessary to perform n median processes. As a method for removing the noise component included in the two-dimensional array data, median filter processing is used, for example, in Patent Document 2.

  In order to reduce the amount of calculation processing, the applicant of the present invention selects array data at a specific array position from among the array data included in the processing target region, and calculates the background using the selected array data. A method has been proposed (Patent Document 3).

  FIG. 8 is an explanatory diagram for describing background calculation using array data at a specific array position. FIG. 8 shows an example in which the background is calculated using, for example, cross-shaped array data as the calculation area 40. For example, for a plurality of data in the calculation area 40a (FIG. 8B), the data value of the center value is obtained by median processing, and the obtained data value is used as the original data value of the data point 41a in the calculation area 40a. (Fig. 8 (c)), and processing for setting the calculation area and replacing the data value calculated by the median processing in order for each of the data points 2 is performed (Figs. 8 (d) to (d)). g)), thereby calculating the background of the processing target area 1 (FIG. 8H).

JP-A-8-86633 (paragraph numbers 0033 to 0036) JP-A-10-160632 (paragraph number 0009) JP 2005-221338 A

  As described above, the amount of calculation processing required to calculate the background is reduced, but the TFT array substrate to be inspected is becoming larger and the resolution is increased, and an increase in throughput is required. Furthermore, it is required to reduce the time required for the background.

  Therefore, it is required to further reduce the background calculation time in the TFT array inspection.

  In view of the above, an object of the present invention is to solve the above-described problems and reduce the calculation load in the process of calculating the background from the two-dimensional array data. Another object of the present invention is to improve the inspection throughput of the TFT array substrate by reducing the calculation load.

  In all of the background calculation calculations using the conventional median process, the median process calculation is performed for all data points included in the processing target region, so that the factor of increasing the size of the TFT array substrate is greatly affected.

  Therefore, in the present invention, in the calculation calculation of the background using the median processing, the median processing is performed on the data points selected from all the data points included in the processing target area, and some array positions are obtained. The background is obtained, and the background of the remaining array positions is obtained by performing an interpolation process using the previously obtained background.

  As a result, the number of computations of the median process can be reduced, so that the total computation time for obtaining the background in the processing target area can be shortened.

  In order to solve the above object, the TFT array inspection apparatus of the present invention inspects the TFT array by irradiating the TFT substrate with an electron beam and detecting secondary electrons generated from the pixels of the TFT substrate by this electron beam irradiation. The TFT array inspection apparatus includes a data processing means for obtaining a background from two-dimensional array data obtained by detecting secondary electrons in a processing target region and obtaining defect information of the TFT substrate using the obtained background.

  The data processing means divides the processing target area into a plurality of blocks, and first calculation means for calculating a first background of the block from the two-dimensional array data included in each divided block; Second calculating means for calculating a second background of another point in the block from the background of the block calculated by the calculating means.

  By combining the background calculated by the first calculation unit and the background calculated by the second calculation unit, the background of the entire processing target region can be calculated.

  The first calculation means possessed by the present invention calculates, in each block, an intermediate value of the two-dimensional array data included in the block, and the calculated intermediate value is a first background at a representative point defined in the block; To do.

  Further, the second calculation means of the present invention performs interpolation using the first background of each block, calculates the background of two-dimensional array points other than the representative points in the block, The background.

  The TFT array inspection method of the present invention is a TFT array inspection method for inspecting a TFT array by irradiating an electron beam to the TFT substrate and detecting secondary electrons generated from the pixels of the TFT substrate by the electron beam irradiation. And a background calculation step for obtaining a background from two-dimensional array data obtained by detecting secondary electrons in the processing target region, and a data processing step for obtaining defect information of the TFT substrate using the background.

  The background calculation step divides the processing target area into a plurality of blocks, and calculates a first background of the block from the two-dimensional array data included in each block; A second calculating step of calculating a second background of another point in the block from the background; a first calculating step of calculating a background of the entire processing target region from the first background and the second background; 3 calculation steps.

  Here, the first calculation step includes, in each block, a step of calculating an intermediate value of the two-dimensional array data included in the block, and a first background at a representative point that defines the intermediate value in the block. The process of carrying out. Further, in the second calculation step, interpolation processing is performed using the first background of each block obtained in the first calculation step, and the background of the two-dimensional array points other than the representative points in this block is obtained. A step of calculating.

  The two-dimensional array data processed by the present invention can be acquired in one pass when the processing target region is scanned by a plurality of passes.

  According to the present invention, the calculation load can be reduced in the smoothing process of two-dimensional array data. In addition, the throughput of inspection of the TFT array substrate can be improved by reducing the calculation load.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 is an explanatory diagram for explaining a configuration example of a TFT array inspection apparatus according to the present invention, taking a TFT array inspection apparatus using a voltage contrast technique as an example.

  In this inspection apparatus, a TFT substrate to be inspected is transported into a high vacuum chamber and inspected in a state of being placed on a stage.

  In FIG. 1, the TFT substrate inspection apparatus 20 includes an electron beam generation source 21, a secondary electron detector 22, and data processing means 23. The electron beam generation source 21 irradiates each pixel 51 of the TFT substrate 50 with the electron beam 24. The secondary electron detector 22 detects secondary electrons 25 generated by irradiating each pixel 51 of the TFT substrate 50 with the electron beam 24. The secondary electron detector 22 outputs a signal representing a waveform corresponding to the voltage waveform of the pixel 51 to the data processing means (computer system or the like) 23 based on the detected amount of the secondary electrons 25. The data processing means 23 analyzes the output signal of the secondary electron detector 22 and inspects the state of the pixel, in particular, the presence or absence of the pixel defect and the content of the defect.

  The data processing means 23 can also include a drive signal supply means, and outputs a drive signal for driving each pixel 51 of the TFT substrate 50 via the line 26. This drive signal is supplied in synchronization with the scanning control of the electron beam 24 that scans the TFT substrate 50, whereby the detection position can be specified.

  Here, the data processing means 23 is a two-dimensional array data obtained by detecting the secondary electrons 25 obtained by scanning the electron beam 24 over the region to be processed on the TFT substrate 50 with the secondary electron detector 22. The background is obtained from the above, and the defect information of the TFT substrate is obtained using the obtained background.

  FIG. 2 is an explanatory diagram for explaining an algorithm for defect detection performed by the data processing means of the present invention.

  In FIG. 2, for the processing target data 101, a background calculation process 102 obtains a background 103 from the processing target data 101 and subtracts the background 103 from the original processing target data 101 to extract defects and noise. This output is subjected to defect detection by a defect extraction process 104 to obtain an inspection output 105.

  Here, the background calculation processing 102 divides the processing target region into a plurality of blocks, and calculates the first background of the block from the two-dimensional array data included in each block, A second background is calculated by using the background of each block calculated in the first calculation process 102a to calculate a second background of other points in the block and calculating a background 103 for all data in the processing target area. Calculation processing 102b.

  FIG. 3 is a diagram for explaining blocks used for calculating the first background in the first calculation process 102a.

  FIG. 3 is a configuration example of a plurality of blocks that divide the processing target area. In FIG. 3, the processing target area 1 corresponds to the scanning area of the electron beam 24 on the TFT substrate 50. For example, the electron beam 24 of the electron beam generation source 21 scans the substrate 50 by a plurality of passes. Can correspond to an area scanned by one pass.

  When this processing target area 1 is made to correspond to a scanning area of one pass, data processing can be performed in order while acquiring data as the path is scanned.

  Each point 2 shown on the processing target area 1 corresponds to a detection point detected by scanning with the electron beam 24, and two-dimensional array data can be represented by corresponding data obtained by detecting each point. . In the following, point 2 is represented as data point 2.

  Here, in the first calculation process of the background of the present invention, the calculation process for obtaining the background is not performed for all the data points 2 on the processing target area 1, but one selected on the processing target area 1 is selected. By performing the calculation process for obtaining the background on the data point 2 of the part, the calculation process time is shortened and the burden on the calculation process is reduced.

  Selection of a data point on which the background is calculated on the processing target area 1 divides the processing target area 1 into a plurality of blocks 3 and 4, and the data points included in the divided blocks 3 and 4 are displayed in the background. This is done by using the target data for calculating.

  Here, the median processing is applied as the calculation for calculating the background. The median process is a process in which a plurality of pieces of target data are arranged in order of size, and data in the middle order is used as representative data. Here, median processing is applied in units of blocks. That is, the data values of the data points included in each block divided in the processing target area 1 are arranged in order of size, and the data value of the data in the middle order is selected as the data value representing the block. Note that it is set in advance for each block which data point in the block corresponds to the selected data value.

  The block example shown in FIG. 3A is an example in which all blocks set on the processing target area 1 have the same shape and the number of data points included in the block is the same, and the block example shown in FIG. This is an example in which the blocks set on the processing target area 1 have different shapes, and the number of data points included in the blocks is allowed to be different.

  FIG. 3A shows an example in which nine data points arranged in a square shape are set as one block, and the processing target area 1 is divided into a plurality of blocks 3a, 3b, to 3n. Note that the number of data points and the shape of one block can be arbitrarily set.

  In addition, the block example illustrated in FIG. 3B is an example including a plurality of types of blocks 4a, 4b,..., 4n having different shapes and different numbers of data points as blocks set on the processing target area 1. is there. This block may include the same shape block or the same number of data points. It should be noted that it is possible to arbitrarily set what kind of block is set in what part on the processing target area 1 or how many data points are set in which part.

  FIG. 4 is an explanatory diagram for explaining the calculation of the background using the block 3 shown in FIG.

  The processing target area 1 has a plurality of data points 2 arranged two-dimensionally (FIG. 4A). The processing target area 1 is divided into a plurality of blocks 3a, 3b, to 3m having the same shape and the same number of data points, and each block includes a plurality of data points 2. According to the calculation of the background according to the present invention, first, one background representing the inside of each block is calculated and assigned to one data point in the block (first background calculation). By using the background assigned to one data point, the background of the remaining data points in the processing target region 1 is calculated by interpolation (calculation of the second background), and by combining these backgrounds, The background values of all data points in the processing target area 1 are calculated.

  FIG. 4B shows one block 3 a in the processing target area 1. A median process is performed on the data points included in the block 3a to calculate one background value. The median process is a process of selecting a value representing this block by arranging values in a plurality of data points in order of magnitude and extracting an intermediate value.

  The background value calculated by the median processing is assigned to any data point included in the block 3a. FIG. 4C shows a state in which the background calculated by the median process is assigned to the data point 10a defined in the block 3a.

  FIG. 4D shows another block 3 b adjacent to the block 3 a in the processing target area 1. One background value is calculated by subjecting the data points included in the block 3b to median processing. The background value calculated by the median processing is assigned to any data point included in the block 3b. FIG. 4E shows a state in which the background calculated by the median process is assigned to the data point 10b defined in the block 3b.

  For the remaining blocks determined in the processing target area 1, the background is calculated in the same manner by median processing, and assigned to any data point included in the block.

  4 (f) and 4 (g), the background is calculated by the median process in the same manner for the last block 3m determined in the processing target area 1, and any data point included in the block is calculated. assign. It is assumed that the processing target area 1 in FIG. 4 is divided into m blocks. Therefore, when the processing target area 1 includes m blocks, m backgrounds are calculated by m median processes and assigned to data points in each block of the processing target area 1.

  FIG. 4H shows a distribution of m backgrounds calculated by the median process in the processing target area 1.

  After this median processing, m backgrounds are calculated at n data points included in the processing target area 1, and the background is obtained for the remaining (n−m) data points. The background value is uncalculated with the data value unchanged.

  Therefore, the present invention obtains the background value at the data point in the data point remaining uncalculated in the processing target region 1 by the calculation of the second background. Here, the background at the remaining data points is calculated by performing an interpolation process using the background value obtained by the calculation of the first background.

  In the interpolation process for calculating the second background, it is possible to arbitrarily determine which positional value to use for the data point to be calculated. For example, in addition to performing interpolation using the background values in the x and y directions of the data point to be calculated, interpolation using the background value in the x direction or the background value in the y direction is used. Interpolation may be performed. Also, interpolation may be performed using a background in an oblique direction.

  At this time, by interpolating using a background in a specific direction, such as a background in the x direction or a background in the y direction, a background that further represents the characteristics of the defect can be obtained.

  FIG. 4 (i) shows the background for all data points in the processing target region 1 by combining the backgrounds obtained by the first background calculation by the median process and the second background calculation by the interpolation process. It shows the state that has been obtained.

  FIG. 5 is an explanatory diagram for explaining the calculation of the background using the block 3 shown in FIG.

  The processing target area 1 has a plurality of data points 2 arranged two-dimensionally (FIG. 5A). The processing target area 1 is divided into a plurality of blocks 4a, 4b, to 4m having different shapes and data points, and a plurality of data points 2 are included in each block. In the background calculation according to the present invention, similarly to the case shown in FIG. 4, first, one background representing the inside of each block is calculated and assigned to one data point in the block (the first background). Next, the background of the remaining data points in the processing target region 1 is calculated by interpolation using the background assigned to one data point for each block (calculation of the second background). The background values of all data points in the processing target area 1 are calculated by combining these backgrounds.

  FIG. 5B shows one block 4 a in the processing target area 1. One background value is calculated by performing median processing on the values of the data points included in the block 4a.

  The background value calculated by the median processing is assigned to any data point included in the block 4a. FIG. 5C shows a state in which the background calculated by the median process is assigned to the data point 11a defined in the block 4a.

  FIG. 5D shows another block 4 b adjacent to the block 4 a in the processing target area 1. A median process is performed on the values of the data points included in the block 4b to calculate one background value. The background value calculated by the median processing is assigned to any data point included in the block 4b. FIG. 5E shows a state in which the background calculated by the median processing is assigned to the data point 11b defined in the block 4b.

  For the remaining blocks determined in the processing target area 1, the background is calculated in the same manner by median processing, and assigned to any data point included in the block.

  5 (f) and 5 (g) show the background calculated by the median process for the last block 4m determined in the processing target area 1 in the same way, and the data points included in the block are calculated. assign. The processing target area 1 in FIG. 5 is assumed to be divided into m blocks. However, the number of blocks, the number of data points included in each block, and the shape of the block can be arbitrarily determined. it can. When the processing target area 1 includes m blocks, m backgrounds are calculated by m times of median processing and assigned to data points in each block of the processing target area 1.

  FIG. 5H shows the distribution in the processing target area 1 of m backgrounds calculated by the median processing.

  After this median processing, m backgrounds are calculated at n data points included in the processing target area 1, and the background is obtained for the remaining (n−m) data points. The background value is uncalculated with the data value unchanged.

  Therefore, the present invention obtains the background value at the data point in the data point remaining uncalculated in the processing target region 1 by the calculation of the second background. The calculation of the background at the remaining data points can be the same as in the case of FIG. 4 described above, and thus the description thereof is omitted here.

  FIG. 5 (i) shows the background for all data points in the processing target region 1 by combining the backgrounds obtained by the first background calculation by the median process and the second background calculation by the interpolation process. It shows the state that has been obtained.

  The number of median processing operations performed in the first background calculation of the present invention corresponds to the number of blocks m included in the processing target region, and the number of interpolation processing operations performed in the second background calculation is calculated in the processing target region. Subtracting the number of blocks m from the number of included data points (n−m) times, the total number of operations itself is m + (n−m), and median processing is performed on all data points included in the processing target area. This is the same as the number of operations n in the case of performing. However, in general, the number of data points handled by the median processing is larger than the number of data points handled by the interpolation processing, so that the computation processing time required for one median processing is T, and the computation processing time required for one interpolation processing is Assuming t, the processing time required to calculate the background according to the present invention is represented by mT + (n−m) t, and the processing time when median processing is performed on all data points is represented by nT. In general, since the number of data points n included in the processing target area 1 is sufficiently larger than the number of blocks m, the computation processing time mT + (n−m) t required for background calculation according to the present invention medians all data points. Is sufficiently shorter than the computation processing time nT required for.

  The processing for calculating the background according to the present invention is not limited to the TFT array inspection, but can be applied to general low-pass filter processing for calculating the background from two-dimensional array data.

It is explanatory drawing for demonstrating the example of 1 structure of the TFT array test | inspection apparatus of this invention. It is explanatory drawing for demonstrating the algorithm of the defect detection which the data processing means of this invention performs. It is an example of 1 structure of the some block which divides | segments the process target area | region of this invention. It is explanatory drawing for demonstrating calculation of the background using the block of this invention. It is explanatory drawing for demonstrating calculation of the other background using the block of this invention. It is the schematic for demonstrating the algorithm of the conventional low-pass filter process. It is explanatory drawing for demonstrating the calculation of the conventional background. It is explanatory drawing for demonstrating the background calculation using the arrangement | sequence data in the conventional specific arrangement | positioning position.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Processing object area | region, 2 ... Data point, 3, 3a-3n ... Block, 4, 4a-4n ... Block, 10a-10m ... Data point, 11a-11m ... Data point, 20 ... TFT substrate inspection apparatus, 21 ... Electron beam generation source, 22 ... secondary electron detector, 23 ... data processing means, 24 ... electron beam, 25 ... secondary electron, 26 ... line, 30a, 30b-30n ... calculation area, 31a, 31b-31n ... data Point, 40a, 40b to 40n ... operation area, 41a, 41b to 41n ... data point, 50 ... TFT substrate, 51 ... pixel,

Claims (6)

In a TFT array inspection apparatus for inspecting a TFT array by irradiating an electron beam to the TFT substrate and detecting secondary electrons generated from the pixels of the TFT substrate by the electron beam irradiation,
A data processing means for obtaining a background from the two-dimensional array data obtained by detecting the secondary electrons in the processing target region, and obtaining defect information of the TFT substrate using the background,
The data processing means divides the processing target area into a plurality of blocks, and first calculation means for calculating a first background of the block from two-dimensional array data included in each block;
Second calculating means for calculating a second background of other points in the block from the background of the block;
A TFT array inspection apparatus, wherein a background of an entire processing target region is calculated from the first background and the second background. The first calculation means calculates an intermediate value of two-dimensional array data included in the block in each block, and sets the intermediate value as a first background at a representative point defined in the block;
The second calculation means performs an interpolation process using the first background of each block, calculates a background of a two-dimensional array point other than the representative point in the block, The TFT array inspection apparatus according to claim 1, wherein:   3. The TFT array inspection apparatus according to claim 1, wherein the two-dimensional array data is data acquired in one pass when the processing target region is scanned in a plurality of passes. In the TFT array inspection method for inspecting the TFT array by irradiating the TFT substrate with an electron beam and detecting secondary electrons generated from the pixel of the TFT substrate by the electron beam irradiation,
A background calculation step for obtaining a background from two-dimensional array data obtained by detecting the secondary electrons in the processing target region;
And a data processing step for obtaining defect information of the TFT substrate using the background.
The background calculation step includes
A first calculation step of dividing the processing target region into a plurality of blocks and calculating a first background of the blocks from the two-dimensional array data included in each block;
A second calculating step of calculating a second background of other points in the block from the background of the block;
A TFT array inspection method comprising: a third calculation step of calculating a background of the entire processing target region from the first background and the second background. In the first calculation step, in each block, a step of calculating an intermediate value of the two-dimensional array data included in the block;
And having the intermediate value as a first background at a representative point defined in the block,
In the second calculation step, interpolation processing is performed using the first background of each block obtained in the first calculation step, and the background of two-dimensional array points other than the representative points in the block is obtained. 5. The TFT array inspection method according to claim 4, further comprising a calculating step.   6. The TFT array inspection method according to claim 4, wherein the two-dimensional array data is acquired in one pass when the processing target region is scanned in a plurality of passes.

Images