JP2003031630A - Method and apparatus for determining crystal boundary - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable to determine a change in state from amorphous silicon to polysilicon easily and stably. SOLUTION: An LCD substrate 20 is imaged by a CCD camera 10, and brightness data (Gray values) is created from the brightness level of each pixel. While determining a local minimum point on a certain line, such a process is conducted that extracts valleys or peaks on a crystal boundary as one curved line. In a local area on one line, a minimum point is determined from the relationship with an adjacent pixel. This minimum point is compared with a Gray value of a next line. When there is an adjacent minimum point, it is determined that there is a continuous curved line. When there is no adjacent minimum point, it is determined that a curved line is broken. Among the thus obtained curved lines, those which are continuous like a closed loop within a reference inspection area are determined as boundaries of a crystal which could be detected effectively.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for determining a crystal boundary used in a semiconductor process that causes a state change from amorphous silicon to polysilicon.

[0002]

2. Description of the Related Art Conventionally, in a manufacturing process of, for example, a TFT liquid crystal display device, a step of changing a state of an amorphous silicon layer provided on a semiconductor substrate into a polysilicon layer has been used. And in such a process, it is necessary to confirm that the state of amorphous silicon has changed to polysilicon.
For example, the method of observing with SEM (scanning electron microscope) etc. was used. Specifically, after the operator directly observed the SEM image, it was photographed, marked and cut into a mesh to calculate the area.

[0003]

However, the method of manually observing the state change by the operator as described above is extremely complicated, and requires a long inspection time and patience, which requires a great deal of labor for the operator. There was a problem of giving. Further, there is a problem that it is difficult to obtain a uniform determination result because the visual determination has a large individual difference and requires high skill. Further, there is a problem in that it is difficult to visually determine a crystal boundary, and it is easy to miss a delicate boundary line. Further, there is a problem that the brightness level cannot be judged visually because it is dark.

Therefore, an object of the present invention is to determine the state change from amorphous silicon to polysilicon.
It is an object of the present invention to provide a crystal boundary discrimination method and device capable of performing a simple and stable determination work.

[0005]

In order to achieve the above-mentioned object, the present invention provides a method for determining the size of a polysilicon crystal group in a semiconductor process in which amorphous silicon on a substrate causes a state change in polysilicon. A crystal boundary determination method, which includes an imaging step of capturing an image of a predetermined inspection region on the substrate, and an image data creation step of creating luminance data that digitizes the brightness of each point of the captured image. A boundary line detecting step of detecting a closed crystal boundary line in the inspection region by image-processing the digitized luminance data. Further, the present invention is a crystal boundary discriminating apparatus for discriminating a size of a crystal group of the polysilicon in a semiconductor process in which amorphous silicon on the substrate causes a state change in the polysilicon, and a predetermined device on the substrate. In the inspection area, image pickup means for picking up an image of the inspection area and luminance data in which the brightness of each point of the picked-up image is digitized are created, and the digitized luminance data is image-processed. And image processing means for detecting the closed crystal boundary line.

In the crystal boundary discriminating method of the present invention, an image of a predetermined inspection region on the substrate that causes a state change from amorphous silicon to polysilicon is taken, and the brightness of each point of the taken image is digitized. Luminance data is created and the luminance data is image-processed to detect a closed crystal boundary line in the inspection region. Therefore, when judging the state change from amorphous silicon to polysilicon, compared with the judgment method by the operator's visual inspection,
It is possible to quickly obtain a uniform determination result without requiring the skill and labor of an operator, and it is possible to perform a simple and stable determination operation.

Further, in the crystal boundary discriminating apparatus of the present invention, an image of a predetermined inspection area on the substrate where a state change occurs from amorphous silicon to polysilicon, and the brightness of each point of this image is numerically represented. The closed crystal boundary line in the inspection area is detected by creating the converted brightness data and image-processing the brightness data. Therefore, when judging the state change from amorphous silicon to polysilicon, a uniform judgment result can be quickly obtained without requiring the skill and labor of the operator, as compared with the judgment method by the operator's visual observation. Therefore, it is possible to perform a simple and stable determination work.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of a crystal boundary discriminating method and apparatus according to the present invention will be described below. In addition,
The embodiment described below is a preferred specific example of the present invention, and is provided with various technically preferable limitations,
In the following description, the scope of the present invention is not limited to these embodiments unless otherwise specified.

In this embodiment, a thin film transistor (TFT) is used.
In the manufacturing process of liquid crystal substrates, when the growth process of polysilicon crystals by laser annealing is checked, the boundaries of dense clusters of crystal grains of several microns are distinguished and image processing is applied to calculate the area. The present invention provides a boundary detecting means (boundary detector). According to the images of high-magnification microscope observation and SEM observation, the cluster of crystal grains rises or sinks at the interface with the adjacent crystal during the crystal growth process when the crystal surface is viewed from above. There is. Although the degree is a very complicated mottled pattern, the boundary part is captured by a camera with an image processing device and the change in brightness is analyzed to determine the continuity of the boundary part and the size of the crystal grain surrounded by the closed loop ( Grai
n) can be expressed. Therefore, the present embodiment uses an image processing system that determines a crystal boundary from an image obtained by an imaging device such as various microscopes, not a method in which an operator directly visually observes the crystal at a high magnification.

FIG. 1 is a block diagram showing the structure of a crystal boundary discriminating apparatus used in this embodiment. This crystal boundary discriminating device has a CCD camera 10, an optical lens 12, a coaxial illuminating device 14, an observation monitor 16 and a personal computer 18. The CCD camera 10 images the substrate surface of the LCD substrate 20 through the optical lens 12 and the coaxial illumination device 14, and transfers the imaged data to the personal computer 18.
Further, the optical lens 12 and the coaxial illuminator 14 constitute an electron microscope, and the coaxial illuminator 14 to LC
The image formed by the reflected light of the light applied to the substrate surface of the D substrate 20 is enlarged by the optical lens 12 and supplied to the CCD camera 10.

The personal computer 18 controls the entire crystal boundary discriminating apparatus of this example, and particularly, the image processing system 18 for realizing the crystal boundary discriminating method according to the present embodiment.
It is equipped with A. The personal computer 18 is connected to a network 22.
Through, for example, communication is performed with a host system at a remote place, the determination result data is provided to the host system, or work is performed in response to an instruction from the host system. Further, the observation monitor 16 is for displaying a picked-up image and the like for visual confirmation by a work manager, but in the apparatus of this example, the boundary discrimination processing itself is automatically advanced by the personal computer 18. Because there is
The observation monitor 16 is used by a work manager in an auxiliary manner.

Next, the crystal boundary discrimination method in this example will be described. First, FIGS. 2A and 2B are views showing an example of a crystal surface by SEM, FIG. 2A shows an observation image from a vertical direction, and FIG. 2B shows an observation from an oblique direction. The image (tilt screen) is shown. It is assumed that a mottled image as shown in FIG. 2A is obtained by capturing an image of a crystal surface having irregularities due to crystal growth as shown in FIG. 2B from above. It should be noted that these images are examples, and there are actually infinite patterns. Therefore, FIG.
Pay attention to the image of (A). Here, the image pickup signal in the predetermined reference inspection area is regarded as an image of x pixels × y lines, and xx
The image data values of the y pixels are digitized into brightness level data from 0 to 255. The digitized data is called a Gray value. The brightness of the crystal surface pattern captured in the memory as a set of Gray values is different from that of the surroundings at the crystal boundary, and is a pattern in which a white background and a black character intersect.

By the way, a method of setting a certain threshold value for the digitized data of such an inspection region and detecting the area, position, etc. from the image obtained by binarization is also conceivable. Crystal grains are not always visible within the range of equivalent brightness, and when they are discarded at a certain threshold,
There is a risk that crystal grains below or above the threshold may be missed. Therefore, in this example, in order to avoid this, a quadratic polynomial parameter of x and y for each point in the image is determined from the luminance change on y = n in FIG. While determining the local minimum point in the quadratic differential that is perpendicular to the direction, processing is performed such that the valley or peak of the crystal boundary is extracted with a single curve. That is, first, a pixel having a local minimum value (that is, a minimum value) is extracted from each pixel on one line. Here, since one line usually includes a plurality of boundary lines, there are a plurality of local minimum values, and therefore a plurality of minimum points are determined from the relationship with adjacent pixels in a local region. To seek. Then, the minimum point obtained in this way and y = n + 1 or y =
Compare with the Gray value of the next line corresponding to n-1,
If there are adjacent minimum points on continuous adjacent lines, it is determined that continuous curves exist. If there is no adjacent minimum point, it is determined that the curve is interrupted.

Then, it is judged that the curve obtained in this way that is continuous in a closed loop in the reference inspection area is the boundary line of the crystal that can be effectively detected. For example,
As shown in FIG. 3, closed curves A and B in the reference inspection area
When it is possible to detect, the closed curves A and B are judged to be the boundaries of the crystals that can be effectively detected. Then, the location, the number, the area, etc. of the regions delimited by the boundary line of such a closed loop are obtained by the internal function of the image processing system, and a process for determining the size of the crystal is performed if necessary. In a normal image processing system, the area and the like of a predetermined area can be easily obtained by using coordinate calculation for the reference inspection area.

The same Gray which is judged to be the minimum value
The processing when the values are adjacent to each other will be described. FIG. 4 is an explanatory diagram showing a specific example of the brightness data (Gray value). In the illustrated example, in the uppermost line, the local minimum value of 10 is laterally continuous with 3 pixels, and in the next line, the local minimum value of 5 is continuous to the right. , Connect these points and judge as a continuous curve. Further, in the lines below it, 10 are local minimum values, and these are judged as continuous curves. In the determination of the uppermost line, it may be determined that one curve extends to the right side and then extends downward, or it may be defined that a part of the line width has widened. Such a definition can be set in advance as an analysis condition. That is,
As for an algorithm for determining the direction of such a curve, a plurality of methods are conceivable and can be appropriately selected and used. For example, when the same Gray values are arranged side by side, the continuity is interrupted. It may be defined, or it may be interpreted that the line width of the curve is expanded as described above. Further, such a judgment condition may be arbitrarily decided by the worker or the administrator. Information such as the number of crystal grains and the size location obtained as described above is
It can be effectively used for determining the optimum condition of laser annealing energy in the LCD manufacturing process. It is also effective as a tool for process analysis.

The above-described crystal boundary discrimination method can be expected to obtain the following various effects. 1. Boundary detection can be performed in a short time, and effective data can be obtained. 2. There is no individual difference, and the same result can be obtained by any worker. 3. For judgment by parameter setting on image processing,
The analysis conditions can be set flexibly. 4. For judgment by parameter setting on image processing,
Preprocessing such as noise removal is possible. 5. It is possible to deal with patterns with uneven brightness. 6. By loading the determination result into the computer, it is possible to feed back to the related device. 7. Conversion to a visual pattern is possible by processing such as coloring after boundary detection. 8. The system configuration can be realized at low cost. 9. By sending the data to the network, it can be used widely. 10. The speed can be increased depending on the ability of the personal computer that performs the processing.

The above-described treatment can be applied to both the case where the boundary is raised and the case where the boundary is dug down during the crystal growth process. In this case, as the brightness data, it is possible to use a local maximum value for the boundary detection instead of performing the boundary detection with the local minimum value. Further, although the liquid crystal manufacturing process has been described above as an example, it may be used in another semiconductor process. Further, as a method of determining the boundary portion by using image processing, it can be applied to analysis of the locus of capillaries in the medical field, analysis of the number of platelets in blood, and the like.

[0018]

As described above, according to the crystal boundary discriminating method of the present invention, an image of a predetermined inspection region on a substrate that causes a state change from amorphous silicon to polysilicon is captured, and the captured image Luminance data that digitizes the brightness of each point is created, and the image processing of this luminance data detects closed crystal boundaries in the inspection area. As a result, it is possible to quickly obtain a uniform determination result without requiring the skill and labor of the operator, and it is possible to perform a simple and stable determination operation.

Further, according to the crystal boundary discriminating apparatus of the present invention, an image of a predetermined inspection region on the substrate which causes a state change from amorphous silicon to polysilicon is picked up, and the brightness of each point of the picked-up image is measured. By creating the brightness data that digitized, and performing image processing of this brightness data,
Since the closed crystal boundary line in the inspection area is detected, it is possible to quickly obtain a uniform judgment result without requiring the skill and labor of the operator, as compared with the visual judgment method of the operator. Therefore, there is an effect that a simple and stable determination work can be performed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a crystal boundary discriminating apparatus used in an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a crystal surface by SEM,
(A) shows an observation image from a vertical direction, and (B) shows an observation image from an oblique direction.

FIG. 3 is an explanatory diagram showing a specific example of a crystal boundary line.

FIG. 4 is an explanatory diagram showing a specific example of luminance data (Gray value).

[Explanation of symbols] 10 ... CCD camera, 12 ... Optical lens, 14 ...
Coaxial illumination device, 16 ... Observation monitor, 18 ... Personal computer, 18A ... Image processing system, 20 ... LCD substrate, 22 ... Network.

Claims (6)

[Claims] 1. A crystal boundary determination method for determining the size of a crystal group of the polysilicon in a semiconductor process in which amorphous silicon on the substrate causes a state change in the polysilicon. By an image pickup step of picking up an image of an inspection region, an image data creating step of creating luminance data that digitizes the brightness of each point of the imaged image, and image processing the digitized luminance data. And a boundary line detecting step of detecting a closed crystal boundary line in the inspection region, the crystal boundary discriminating method. 2. In the boundary line detecting step, after detecting a first point having a minimum value or a maximum value of local luminance data of a certain line in the inspection area, a line adjacent to the certain line is detected. A second value having a minimum value or a maximum value of the luminance data from points close to the first point,
2. The crystal boundary discrimination method according to claim 1, wherein one curve is sequentially extracted by detecting the point. 3. The crystal boundary determining method according to claim 2, wherein, in the boundary line detecting step, the closed curve extracted in the inspection region is determined as a crystal boundary line. 4. The crystal boundary determining method according to claim 2, wherein, in the boundary line detecting step, when the second point cannot be detected, it is determined that the curve is a curve that is interrupted. 5. The crystal boundary determination method according to claim 3, wherein the crystal size is calculated from the closed curve. 6. A crystal boundary discriminating apparatus for discriminating the size of a crystal group of polysilicon in a semiconductor process in which amorphous silicon on a substrate causes a state change in polysilicon. In the inspection area, image pickup means for picking up an image of the inspection area and luminance data in which the brightness of each point of the picked-up image is digitized are created, and the digitized luminance data is image-processed. An image processing unit for detecting a closed crystal boundary line in, and a crystal boundary discriminating apparatus.