JP3480641B2 - Pattern inspection method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an inspection method for inspecting a pattern formed on a green sheet, a film carrier or the like, and more particularly to an inspection method for inspecting a pattern disconnection.

[0002]

2. Description of the Related Art Conventionally, PGA (Pin Grid Array) has been known as a mounting technique suitable for a demand for increasing the number of pins of ICs and LSIs. In PGA, a ceramic substrate is used as a base of a package to which a chip is attached, and wiring is performed up to a lead wire extraction position. To make this ceramic substrate, a so-called green sheet is used in which alumina powder is kneaded with a liquid binder to form a sheet, and a paste containing a metal having a high melting point is screen-printed on the green sheet. Then, by stacking a required number of such sheets and firing them, so-called simultaneous firing is performed, in which the green sheets are sintered and the paste is metallized.

As another mounting technique, TAB is used.
(Tape Automated Bonding) is known. The TAB method is a polyimide film carrier (TAB tape).
The copper foil pattern formed above is bonded to the electrodes of the IC chip to form external leads. The copper foil pattern is formed by attaching a copper foil to a film with an adhesive and etching this.

In such a green sheet or film carrier, the pattern is visually inspected by a person using a microscope after the pattern is formed. However, in order to visually inspect a fine pattern, there is a problem that skill is required and eyes are seriously used. Therefore, as an alternative to the visual inspection, a technique has been proposed in which a pattern formed on a film carrier or the like is imaged by a TV camera and automatically inspected (for example, JP-A-6-27313).
No. 2).

8 and 9 are views for explaining a conventional inspection method disclosed in Japanese Patent Laid-Open No. 6-273132. The master pattern created by imaging the measured pattern determined to be non-defective is registered as a set of straight lines indicating pattern edges. Further, the measured pattern is input as a set of edge data (edge coordinates) indicating a pattern edge extracted from a grayscale image obtained by capturing the pattern. Then, the extracted edge data n1, n2, n3 ... Of the measured pattern and the straight line of the master pattern are associated with each other. In order to make this correspondence, as shown in FIG. 8, bisectors A2 ′, A3 ′, which bisect the angles formed by the continuous straight lines A1 and A2, A2 and A3, ... · Ask for.

The circumference of the straight lines A1, A2, A3, ... Of the master pattern is divided by the bisectors A2 ', A3' ,. As a result, the edge data n1, n2, n3 ... Of the measured pattern existing in each area are associated with the straight lines A1, A2, A3 ... Of the master pattern to which the area belongs. . For example, in FIG. 8, the edge data n1 to n3 are associated with the straight line A1, and the data n4 to n6 are associated with the straight line A2. Next, the edge data of the measured pattern is compared with the master pattern to inspect whether the measured pattern is broken.

This inspection is realized by a labeling process for tracing the pattern edges by tracing the linked edge data n1 to n9 of the pattern to be measured.
At this time, due to the disconnection at the tip of the measured pattern,
Since the edge data is not connected at this disconnection portion, there is no edge data corresponding to the straight lines A3 to A5 of the master pattern. Thus, the disconnection of the measured pattern can be detected.

[0008]

As described above, in the conventional inspection method, the edge data of the measured pattern and the straight line of the master pattern are associated with each other to inspect the disconnection of the measured pattern. There is a problem in that it takes time to inspect the disconnection because it takes time to perform the attaching process. The present invention has been made to solve the above problems, and an object of the present invention is to provide an inspection method capable of inspecting a disconnection of a pattern at high speed.

[0009]

According to the present invention, an inspection range of a predetermined size is set in an image of a master pattern which serves as a reference for comparison with a measured pattern, and a pattern edge of the master pattern within this inspection range is set. And sequentially assigning numbers to the patterns intersecting each side of the inspection range, and assigning the same number to one pattern composed of connected pattern edges, the pattern of the master pattern is determined from the image of the measured pattern. The inspection range of the measured pattern corresponding to the inspection range is taken out, the pattern edges of the measured pattern in this inspection range are traced, the patterns intersecting each side of the inspection range are sequentially numbered, and the connected pattern edges The same number is assigned to one pattern consisting of, and the number assigned to the master pattern and the number to be measured are assigned. Compared were a number, in which so as to inspect a break in the pattern to be measured. In this way, in each of the master pattern and the pattern to be measured, the pattern intersecting each side of the inspection range is sequentially numbered, and the same number is assigned to one pattern including the connected pattern edges. When a disconnection exists in the measured pattern, the pattern edges are not connected at this disconnection portion, and different numbers are given to the two patterns that should be originally the same and are separated by the disconnection portion. Therefore, since there is a difference between the number assigned to the master pattern and the number, the disconnection of the measured pattern can be detected.

Further, an inspection range of a predetermined size is set in the image of the master pattern which is a reference for comparison with the measured pattern, the pattern edges of the master pattern within this inspection range are traced, and the sides of the inspection range Detects the tip of the pattern that does not intersect with, and uses the coordinates of this tip as the tip point, and extracts the inspection range of the measured pattern that corresponds to the inspection range of the master pattern from the image of the measured pattern, Is traced, and the disconnection of the measured pattern is inspected depending on whether or not the connected pattern edge passes within a predetermined range centered on the tip point. In this way, the pattern tip that does not intersect the side of the inspection range in the master pattern is detected, the coordinates of this tip are used as the tip points, and the connected pattern edges of the measured pattern are within a predetermined range centered on the tip points. Determine whether to pass. If a disconnection exists in the measured pattern, the pattern edges are not connected at this disconnection portion, so that the pattern edge does not pass within a predetermined range that should exist before the disconnection portion when tracing the connected pattern edges. Thereby, the disconnection of the measured pattern can be detected.

Further, the detection of the tip of the pattern is made such that one intersection of the side of the inspection range and the pattern edge is used as a reference point, and the pattern edge whose distance from the reference point shows a maximum value is used as the pattern tip. is there. In this way, if one intersection of the side of the inspection range and the pattern edge is set as the reference point, the distance from the reference point, which has been increasing, will decrease at the tip of the pattern. Thereby, the tip of the pattern can be detected. Further, among the pattern edges whose distance from the reference point has the maximum value when the disconnection of the measured pattern is detected, the one not registered as the tip point indicates the disconnection portion.

[0012]

1 is a flow chart showing an inspection method according to a first embodiment of the present invention. First, registration of the master pattern will be described. First, a pattern that has been determined to be non-defective is captured by a camera, a grayscale image output from the camera is digitized by an A / D converter, and is temporarily stored in an image memory (step 101).

Subsequently, the edge data (edge coordinates) of the master pattern is extracted from the multi-gradation grayscale image data stored in the image memory as follows (step 10).
2). FIG. 2 is a diagram for explaining a method of extracting edge data. Here, the pattern is represented by white circles as pixels with high density, and the base material (green sheet or film carrier) is represented by black circles as pixels with low density.

Although the grayscale image data includes a pattern and a background (base material) other than the pattern, since there is generally a density difference between the pattern and the background, a density histogram showing the frequency of the density of the image data. As is well known, the histogram shows a bimodal property having two maximum values, that is, the frequency corresponding to the background and the frequency corresponding to the pattern. In order to binarize the grayscale image data, the valley point between the two peaks may be set as the threshold value S.

If the threshold S determined in this way indicates a pattern edge in a grayscale image, X and Y of the threshold S from the dark and light pixels on both sides of the threshold S as boundaries.
By obtaining the coordinates, the edge coordinates of the master pattern can be obtained. For example, when the edge coordinates are obtained at the position C in FIG. 2A, the pixels of the coordinates X2 and Y1 and the coordinates X
The density gradient as shown in FIG. 2B is obtained from the pixels of 1 and Y1, and the coordinate X of the point where this density gradient intersects the threshold value S.
Should be calculated (Y coordinate is Y1). In this way, the edge data of the master pattern can be extracted.

Next, an inspection range of a predetermined size (hereinafter referred to as a window) is set in the master pattern, and an image of the master pattern is cut out by this window (step 103). Then, a number is given to the master pattern that intersects with each side of the window by labeling the edge data in which the master pattern cut out by the window is connected (step 104).

FIG. 3 is a diagram for explaining this numbering process, where W1 is a window and P1 to P4 are patterns. Although the master pattern is a set of points (edge data) indicating pattern edges, the pattern edges are represented by straight lines here. In the numbering process, the upper left of the window W1 is used as a starting point, and the master pattern is scanned clockwise as shown in FIG. The numbers given below to each pattern are also given in this order.

When the master pattern is scanned from the upper left of the window W1, the pattern P1 is detected as the first pattern intersecting the side of the window W1. As a result, as shown in FIG. 3, the pattern P1 intersecting the upper side of the window W1.
No. "1" is given to the end of the. Subsequently, with this end portion as the start position, the edges of the pattern P1 are tracked by sequentially searching for edge data connected in a clockwise direction from the start position (that is, as shown by the arrow in FIG. Track the edges).

In the case of the pattern P1, since it collides with the lower side of the window W1 during this tracking, "1" is similarly given to the end of the pattern P1 which intersects with the lower side of the window W1. Furthermore, the edge of the pattern P1 is tracked (that is,
The edge on the left side of the pattern P1 is traced), and the boundary tracing of the pattern P1 is completed by returning to the start position.

Next, when the master pattern is scanned from the upper left of the window W1, the pattern P2 is detected as the next pattern for which boundary tracking has not been performed. Since the pattern P2 is a different pattern that is not connected to the pattern P1,
The next number "2" is given to the end of the pattern P2 intersecting the upper side of the window W1, and the number "2" is similarly given to the end of the pattern P2 intersecting the lower side of the window W1. Such numbering processing is similarly performed for the pattern P3 and the pattern P4, and each is "3".
No., “4” is given.

By repeating the above numbering process for each window (steps 103 to 10).
5), numbers are assigned to the entire range of the master pattern. Next, the inspection of the measured pattern will be described. First, the pattern to be measured is imaged by the camera, and the grayscale image output from the camera is digitized by the A / D converter and temporarily stored in the image memory (step 106). The multi-grayscale image data stored in the image memory is binarized after alignment with the master pattern is performed (step 107).

Subsequently, a window at a position corresponding to the window of the master pattern is cut out from the binary image of the measured pattern, and each window of the window is labeled by tracking connected pixels of the measured pattern cut out by this window. A number is given to the measured pattern that intersects the side (step 108). FIG. 4 is a diagram for explaining the numbering process of the measured pattern.

The numbering process of the measured pattern is almost the same as that of the master pattern. When the measured pattern is scanned from the upper left of the window W2, the pattern corresponding to the above pattern P1 is the first pattern intersecting the side of the window W2. P11 is detected. As a result, FIG.
As shown in, the pattern P intersecting with the upper side of the window W2
The number “1” is given to the end of 11.

Subsequently, the edge of the pattern P11 is traced by successively searching for boundary points connected clockwise from the starting position with this end as the starting position. In the case of the pattern P11, since it collides with the lower side of the window W2 during this tracking, "1" is also given to the end of the pattern P11 that intersects with the lower side of the window W2. Further, by tracing the edge of the pattern P11 and returning to the start position, the boundary tracing of the pattern P11 ends.

In this embodiment, in the binary image of the measured pattern in which the pattern is represented by white pixels and the substrate is represented by black pixels, when scanning from left to right, the image rises from black to white. The white pixel is the boundary point, and the black pixel is the boundary point when falling from white to black.

Next, the pattern corresponding to P2 above is
Both are divided into patterns P12 and P16 due to the disconnection. As a result, since the boundary points are not connected at this disconnection portion, when the pattern P12 is detected as the next pattern for which boundary tracking has not been performed and the boundary point is tracked, as shown by the arrow in FIG. After tracing the generated edge, the left edge is traced back to the starting position.

Therefore, the number given to the pattern P12 is only "2" given to the end portion that intersects with the upper side of the window W2. The pattern corresponding to the above P3 is divided into the pattern P13 and the pattern P15 due to the disconnection.
By the same processing as 12, the number “3” is given to the end portion that intersects with the upper side of the window W2. Then, the pattern P14
Is processed in the same manner as the pattern P11.

Next, when the measured pattern is scanned clockwise from the upper left of the window W2, the pattern P15 is detected as a pattern intersecting the lower side of the window W2. As a result, the pattern P15 that intersects with the lower side of the window W2
The number "5" is given to the end of the. Then, with this end as the starting position, the edges of the pattern P15 are traced by successively searching for boundary points connected clockwise from the starting position (that is, as shown by the arrow in FIG. Track the edges).

At this time, as in the case of the pattern P12, the boundary points on the side of the pattern P13 are not regarded as the connected boundary points, so as shown by the arrow in FIG. Follow the right edge and return to the starting position. Therefore, the number given to the pattern P15 is only "5" given to the end portion that intersects with the lower side of the window W2. For pattern P16,
By the same process as the pattern P15, the number "6" is given to the end portion that intersects with the lower side of the window W2.

After performing the numbering process as described above, the number given to the master pattern and the number given to the measured pattern are compared for each side of the window (step 109).

[0031]

[Table 1]

There is no difference in the numbers of the master pattern and the measured pattern provided on the upper side of the window, and all are judged to be normal (◯ in Table 1). On the other hand, there is a discrepancy between the numbers of the master pattern and the measured pattern provided on the lower side of the window, so it is judged that there is a disconnection in this portion (X in Table 1). By repeating the above processing for all windows (steps 108 to 11)
0), it is possible to inspect all the disconnections of the measured pattern.

As described above, in the present embodiment, it is not necessary to associate the edge data of the pattern to be measured with the straight line of the master pattern, so that the inspection time (the time required for the processing after step 106) can be shortened. You can

In the first embodiment, the disconnection can be inspected when both ends of the pattern abut the side of the window, but the inspection can be performed when one end of the pattern does not abut the side of the window. In such a case, another inspection method is necessary because it is not possible. FIG. 5 is a flow chart showing the inspection method according to the second embodiment of the present invention.

First, steps 201 to 203 are the same as steps 101 to 103 in FIG. Then, by the labeling process that tracks the edge data that is the concatenation of the master pattern cut out by the window,
Detects the tip of the pattern that does not intersect the edge of the window,
This is stored (step 204). FIG. 6 is a diagram for explaining this marking process.

In the marking process, the master pattern is scanned clockwise from the upper left of the window W3. As a result, the point O is detected as the first pattern edge that intersects the side of the window W3, and the coordinates of this reference point O are stored. Then, with the reference point O as the starting position, the pattern edges connected clockwise from the starting position are traced as indicated by the arrows in FIG.

At this time, the pattern edge where the distance from the reference point O shows the maximum value is defined as the pattern tip. For example, at the point M1, the distance from the reference point O that has been increasing is decreasing. As a result, this point M1 is detected as the tip of the pattern P7, and its coordinates are stored. Similarly, the point M2
However, the distance from the reference point O, which had been increasing again, is going to decrease. As a result, the coordinates of this point M2 are stored as the tip of the pattern P7. The same applies to the other points M3 and M4.

By repeating the above marking process for each window (steps 203 to 20)
5), the tip of the pattern is marked in the entire range of the master pattern. Next, the inspection of the measured pattern will be described. Steps 206 and 207 are the same as steps 106 and 107 in FIG.

Then, a window at a position corresponding to the window of the master pattern is cut out from the binary image of the measured pattern, and the connected pattern is traced by the connected pixels of the measured pattern cut out by this window. It is determined whether the edge passes within a predetermined range centered on the stored tip point (step 2
08). FIG. 7 is a diagram for explaining this passage determination processing.

In the passage determination processing, the pattern to be measured is scanned from the upper left of the window W4 in the clockwise direction and the connected boundary points are searched one after another to trace the edge of the pattern P17 as shown by the arrow in FIG. At the tip of the first pattern, since the pattern edge passes within a predetermined range R1 centered on the tip point M1 marked by the master pattern,
It is judged to be normal. The same applies to the tip of the second pattern.

Regarding the third tip, due to the disconnection, a predetermined range R3 centered on the marked tip point M3
Is divided into a tip and a part in the middle. As a result, the boundary points are no longer connected at the disconnection portion, and the edge caused by the disconnection is traced as shown by the arrow in FIG. 7, so that it passes through the predetermined range R3 that should exist before the disconnection portion. do not do. Thus, the disconnection of the measured pattern can be detected.

Further, as a result of tracing the pattern edges, when a disconnection is detected, among the pattern edges whose distance from the reference point O has the maximum value, the one not registered as the tip point indicates the disconnection portion. The position of the disconnection can be accurately known. That is, when the edge caused by the disconnection is traced, the distance from the reference point O has a maximum value, but since this position is not one of the stored tip points M1 to M4, this is recognized as the disconnection portion. be able to. By repeating the above processing for all windows (steps 208 and 209), all disconnections of the measured pattern can be inspected.

[0043]

According to the present invention, in each of the master pattern and the pattern to be measured, the pattern intersecting each side of the inspection range is sequentially numbered, and one pattern consisting of connected pattern edges is assigned. When the number to be given is the same, if a disconnection exists in the measured pattern, a difference occurs between the number and the number given to the master pattern, so that the disconnection of the measured pattern can be detected. Therefore, it is not necessary to associate the edge data of the measured pattern with the straight line of the master pattern as in the conventional case, so that the disconnection of the measured pattern can be inspected at high speed.

In the master pattern, the tip of the pattern which does not intersect with the side of the inspection range is detected, the coordinates of the tip are used as the tip point, and the pattern edge to which the measured pattern is connected is within a predetermined range with the tip point as the center. If there is a disconnection in the measured pattern by determining whether it passes through, the pattern edge does not pass within the predetermined range that should exist at the tip of the disconnection part, so it is possible to detect the disconnection of the measured pattern. You can Therefore, it is not necessary to associate the edge data of the measured pattern with the straight line of the master pattern as in the conventional case, so that the disconnection of the measured pattern can be inspected at high speed.

Further, by setting one intersection of the side of the inspection range and the pattern edge as a reference point, the tip of the pattern can be easily detected and the position of the broken portion can be accurately known.

[Brief description of drawings]

FIG. 1 is a flow chart diagram showing an inspection method according to a first embodiment of the present invention.

FIG. 2 is a diagram for explaining a method of extracting edge data of a master pattern.

FIG. 3 is a diagram for explaining a master pattern numbering process.

FIG. 4 is a diagram for explaining a numbering process of a measured pattern.

FIG. 5 is a flowchart showing an inspection method according to a second embodiment of the present invention.

FIG. 6 is a diagram illustrating a master pattern marking process.

FIG. 7 is a diagram for explaining a passage determination process of a measured pattern.

FIG. 8 is a diagram for explaining a conventional inspection method.

FIG. 9 is a diagram for explaining a conventional inspection method.

[Explanation of symbols]

P1 to P4, P7, P11 to P17 ... Pattern, W1
W4 ... Wind, M1 to M4 ... Tip point.

Front page continued (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06T ⁷ /00-7/60 G06T 1/00 G01B 11/00-11/30 G01N 21/84-21/958 H01L 21 / 64-21/66

Claims (3)

(57) [Claims] 1. An inspection range of a predetermined size is set in an image of a master pattern, which serves as a reference for comparison with a measured pattern, and a pattern edge of the master pattern within this inspection range is traced to check the inspection range. Numbers are sequentially assigned to the patterns intersecting with each side, and the same number is assigned to one pattern composed of connected pattern edges, and the measured pattern corresponding to the inspection range of the master pattern is measured from the image of the measured pattern. The inspection range of the pattern is taken out, the pattern edges of the measured pattern in this inspection range are traced, the patterns intersecting each side of the inspection range are sequentially numbered, and one pattern consisting of the connected pattern edges is The numbers given to the master pattern and the number given to the pattern to be measured are the same. And compare the inspection method of a pattern, characterized by checking the breakage of the pattern to be measured. 2. A side of the inspection range is set by setting an inspection range of a predetermined size in an image of a master pattern which is a reference for comparison with a measured pattern, and tracing a pattern edge of the master pattern within the inspection range. The tip of the pattern that does not intersect is detected, the coordinates of this tip are used as the tip point, and the inspection range of the measured pattern corresponding to the inspection range of the master pattern is extracted from the image of the measured pattern. Pattern disconnection of the pattern to be measured is inspected by tracing the pattern edges connected to the pattern and checking whether the connected pattern edges pass within a predetermined range centered on the tip point. Inspection method. 3. The pattern inspection method according to claim 2, wherein the detection of the pattern tip uses a single intersection of a side of the inspection range and a pattern edge as a reference point, and a distance from the reference point indicates a maximum value. A pattern inspection method characterized in that a pattern edge is used as a pattern tip.