JPH0450709A - Position detecting method for scribing line crossing area - Google Patents

Abstract

PURPOSE:To enable position detection for a variety of substrates by regarding a coordinate position, based upon coordinate values of two integral values which are judged to match scribing line widths in an X and a Y direction, as the position of the angle in the crossing area of the scribing lines. CONSTITUTION:An enlarged gradational image of the scribing line crossing area of a substrate W which is picked up by a television camera 6 is inputted to an image processing unit 7. The image which is inputted to the unit 7 is an image which is photographed through a downward lighting microscopy opti cal system, so the gradation value of the edges of the scribing lines on the substrate W is small. Then the edges of the scribing lines slant to the substrate surface because of a step due to the high-low difference between the areas of the scribing lines on the substrate surface and a pattern area. Therefore, light which is projected from an objective 2 and irradiates the substrate W is reflected off the entrance pupil of the lens 2, so that the light does not reach the camera 6.

Description

[Detailed Description of the Invention] <Industrial Application Field> The present invention is used in various characteristic measuring devices such as a film thickness measuring device to position a substrate such as a semiconductor wafer. The present invention relates to a method of automatically detecting the position of an intersection area, and particularly to a method of searching for an intersection area of scribe lines from an image of the scribe lines on the surface of a substrate and detecting the position.

When positioning the board using various measuring devices, the position of the intersection area of the scribe lines is detected to be used as an index for positioning, and by using the scribe lines, positioning marks can be placed on the board. This is to make it unnecessary to form such.

<Prior art> In order to detect the position of the scribe line intersection area from an image of the scribe line on the substrate surface, the first step is to detect the scribe line area from the image, and the second step is to detect the scribe line intersection area. Two steps are required: detecting the position of the area, and conventionally, the so-called 2i method is used as the first step to detect the scribe line.
A method called oxidation processing is known.

This method takes a multi-gradation image taken of the board surface, sets the gradation value set using a method called the mode method as a threshold value,
The scribe line area in the image is detected by converting it into a binary box image.

The conventional method will be explained below with reference to FIGS. 11 and 12.

FIG. 11 is an enlarged image of a part of the substrate surface (the area where the scribe lines intersect), and in the figure, SX is the scribe line in the X direction, and SY is the scribe line in the Y direction.

In the so-called mode method, when the gradation value histogram of a captured image as shown in FIG. 11 shows bimodality as shown in FIG. 2417. It is a process that undergoes chemical processing. Based on the binarized image from which the scribe line has been extracted, the position of the intersection area is detected, which is the subsequent step of the scribe line intersection area position detection method. For the extracted binarized image, an image processing method such as pattern recognition is appropriately used to specify the intersection area of the scribe lines, and its position is detected. Note that from the position of the detected intersection area of the scribe line, a TEG (not shown) located at the center of the intersection area or at a predetermined positional relationship therefrom is determined.
Inspection monitor element group: Te5t EleIIlent
Measurement of film thickness and various electrical characteristics are performed for the following.

<Problems to be Solved by the Invention> However, the method of detecting scribe lines through binarization processing using the mode method as described above has the following problems.

That is, in the binarization process, in the image taken of the board surface, the scribe line area has a gradation value unique to the scribe line, and the area other than the scribe line (mainly the area where the circuit pattern is formed) It is assumed that the pattern area (hereinafter referred to as a "pattern area") has a unique tone value different from that of the scribe line, and is in the state of a tone value histogram exhibiting bimodality as shown in FIG. 12.

However, the surface conditions of substrates vary widely in each process of semiconductor manufacturing, and all substrates have
Rather than meeting this assumption, it is often difficult to detect scribe lines with the method described above.

For example, if only the seed film that will become the pattern area is formed and the intensity of the reflected light is almost the same in the pattern area and the scribe line area, the gradation value histogram will be as shown in Figure 13 (a). In addition, at various stages of semiconductor manufacturing, various miscellaneous films are formed at various locations in the pattern area, and the gradation value histogram thereof is as shown in Figure 13, etc. Indicates three or more peaks.

Therefore, according to the conventional method, it is difficult to set the dark value,
Since the binarization process itself cannot be performed, it is difficult to detect the scribe line, and there are problems in that it is often impossible to detect the position of the intersection area of the scribe line.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a method for detecting the position of an intersection area of scribe lines that can be applied to a variety of substrates.

Means for Solving the Problems> In order to achieve the above objects, the present invention has the following configuration.

(1) A method for detecting a scribe line according to the first invention includes X, Y orthogonal directions along the scribe line with respect to multi-tone image data of a scribe line intersection area on a substrate surface that is magnified and imaged by an epi-microscopic optical system. An integral processing process in which gradation values are sequentially integrated for each pixel column in the direction, and two minimum integrated values are detected from among the group of integrated values in the X and Y directions obtained by the above integration process. an interval calculation process that calculates the X-direction interval and the Y-direction interval from the coordinate values corresponding to the two integral values detected in the X and Y directions, respectively; a determination step of determining whether the directional spacing each matches a predetermined scribe line width; Value X1. X
z, Yt, coordinate position (X,,Y,) based on Yt,
(xz, Y,), (Xl, Yt) (Xl, Y,) is the position of the corner of the intersection area of the scribe lines.

(2) The method for detecting a scribe line according to the second invention includes a differential processing process in which the integrated values in the X and Y directions obtained by the above integration process are respectively differentiated, and each in the X and Y directions. An inflection point calculation process of calculating two inflection points in each of the X and Y directions from the obtained differential processing results, and coordinate values corresponding to the two calculated inflection points in each of the X and Y directions. In addition to the interval calculation process of calculating the X-direction interval and the Y-direction interval from X, X1, X2, Y1, and Y2, the same judgment as in the first invention is made regarding the calculated X and Y-direction intervals. The coordinates ('iXl
, xi , y+ , coordinate value lCL based on yz ,
Yt), (L, y,), (x,, y□).

Let (Xl, Yt) be the corner position of the intersection area of the scribe lines.

(3) A method for detecting a scribe line according to a third aspect of the present invention includes a minimum value detection step of detecting the minimum integral value from a group of integral values in the X and Y directions obtained by the above-mentioned integral processing; In relation to the calculated minimum integral value in the X and Y directions,
, a threshold calculation process of calculating two thresholds according to the Y direction, and two integral value groups in the X and Y directions based on the two thresholds.
Based on the binarization process and the groups of integral values binarized in the X and Y directions, the central coordinate values X1.
This invention comprises an interval calculation process for calculating the X-direction interval and Y-direction interval from Xl and YIY2, respectively, and a judgment process for making the same judgment as in the first invention regarding the calculated X- and Y-direction intervals. Yes, for X, Y'jj leap, two center coordinate values X, , x that are determined to match the scribe line width, respectively.
Coordinate value 11f (XY+) based on i, y, yz,
(Xg, Yt), (L, Yt), (XZYt
) is the position of the corner of the intersection area of the scribe lines.

(4) The scribe line detection method according to the fourth invention combines the main components of the first invention and the main components of the second invention, and includes a method for detecting scribe lines obtained by the above-mentioned integral processing. 7 in each pixel column in the X and Y directions, and in turn the floor ff1
an integral processing process that performs an integral process of 4t; a minimum value detection process that detects the two minimum integral values from among the group of integral values in the X and Y directions obtained in the integral processing process; Coordinate values X, , X2 . corresponding to the two integral values detected in the Y direction, respectively. Y,,Y.

and a step of determining whether or not the calculated X and Y direction intervals match respective predetermined scribe line widths. In the first judgment process, which is the same as the main configuration of the first invention, and the first judgment process, if the rain sounds do not match,
An inflection point in which two inflection points are calculated in each of the X and Y directions from the differential processing process in which differential processing is applied to the integral value in the Y direction, and the results of the differential processing obtained in the X and Y directions, respectively. Calculation process; Coordinate values X, , X, , Y corresponding to the two calculated inflection points in each of the X and Y directions.

an interval calculation process of calculating an interval in the X direction and an interval in the Y direction from Yz'; and a process of determining whether the calculated intervals in the X and Y directions match respective predetermined scribe line widths. The second invention is the same as the main structure of the second invention.
determining a coordinate position (XI, Yl), (L, Y,)(x,, Yl
), (x,, Y,) or the coordinate values χ+, L of the two inflection points that were determined to match the scribe line width in the X and Y directions, respectively, in the second determination process.
, Yl, Y% coordinate position (L, Yl'
), (XtYl'), (x+, yz'), (
x, , yz') are the corner positions of the intersection areas of the scribe lines.

<Effect> (1) The effects of the first invention are as follows.

Various films are formed on the surface of the substrate to form a semiconductor circuit.
Since the scribe line has one or more layers, the edge of the scribe line is inclined on the substrate surface due to the difference in height from the pattern area.

For this reason, in images taken of the board using an epi-illumination microscope optical system,
The substrate is illuminated only by the light emitted from the objective lens, and when the illumination light is irradiated onto the edge of the scribe line, the entrance pupil of the epi-illumination microscope optical system is illuminated because the edge of the scribe line is inclined. Therefore, in this image, the gradation (brightness, light intensity) value of the edge part of the scribe line is small. Therefore, the multi-gradation image data of the scribe line intersection area is When integrating the gradation values in sequence for the pixel row, we get
The integral value of the pixel column corresponding to the edge of the scribe line takes a smaller value compared to other pixel columns. In other words, 1
．． Two edges of the scribe line are detected by detecting the two minimum integral values. Then, the X-direction interval and the Y-direction interval are calculated from the coordinate values corresponding to the two integral values detected in the X and Y directions, respectively, and if they match the predetermined scribe line width, Coordinate values χ1. of the two integral values in the Y direction, respectively. X2Yl, Y! is, X, Y
The coordinate values of the edges of each scribe line in the direction are confirmed. The corner of the intersection area of the scribe lines is
Since this is the intersection of the edge of the scribe line in the direction and the Y direction, the coordinates of the intersection of the edges are (XI, Yl), (X2
゜Yl), (x,, Yt), (xi, YZ)
is detected as the position of the corner of the intersection area of the scribe lines.

(2) The effects of the second invention are as follows.

According to the second invention, a differential process is applied to the XY force direction integral values obtained by the integral process, and two inflection points are calculated for each of the X and Y directions from the group of differential values. do. These inflection points correspond to the edges of the scribe line, so when the coordinate values XXz, Yl, and Yz corresponding to the two inflection points in the X and Y directions are calculated, similarly to the first invention, Coordinates of the intersection of the edges of the scribe lines in the X and Y directions (X,, Y,) (xz
, Yl ), (X,, ¥2 ), (XZ, Yz
) becomes clear, and its coordinates are detected as the position of the corner of the intersection area of the scribe lines.

(3) The effects of the third invention are as follows.

According to the third invention, X obtained by integral processing.

The minimum integral values are respectively detected from a group of integral values in the Y direction, and two threshold values corresponding to the X and Y directions are set in relation to these. Based on these two thresholds, based on the disulfide integral value group obtained by binarizing the integral value group in the X and Y directions, the central coordinate value , ,X, ,Y, ,Yi are determined. This center coordinate value corresponds to the coordinate of the edge of the scribe line. These center coordinate values X+, Xi, Yl, Yi
, and the coordinates of the intersection of the scribe lines in the X and Y directions (
X,,Y,), (χ2.Y,)(XI, Yz)
, (Xz, Yz), and its coordinates are detected as the position of the corner of the intersection area of the scribe lines.

(4) The effects of the fourth invention are as follows.

According to the fourth invention, similar to the first invention, the scribe line is initially detected based on detecting the minimum integral value, and when the scribe line is not detected, the sensitivity is further increased. By utilizing the good second invention, the inflection point is determined by differentiating the integral value. That is, in the first judgment process similar to the main part of the first invention, X,
In the Y direction, coordinate values Xl and Xx of two integral values determined to match the scribe line width, respectively.
, Yl, Yz, or coordinate values XI, XtYl, Y of two inflection points determined to match the scribe line width in the X and Y directions, respectively, in the second determination process similar to the second invention. ! ', the coordinate values of the edge of the scribe line in each of the X and Y directions become clear.

Since the corner of the intersection area of the scribe lines is the intersection of the edges of the scribe lines in the X and Y directions, the coordinates of the intersection of the edges determined in the first judgment process (x+, Yl),
(xi, Yl), (x,, Yi) (XI, Yi
) or the coordinates of the intersection of the edges found in the second judgment process (XI, Yl'), CX2Y+'), (
L, Yz') and (Xz, Y2') are detected as the corner positions of the intersection areas of the scribe lines.

<Example> Hereinafter, an example of the present invention will be described in detail based on the drawings.

1. Figure 1 is a block diagram showing a schematic configuration of a film thickness measuring device used in this embodiment and each embodiment to be described later. Here, a film thickness measuring device is cited as an example of a device to which the method of the present invention is applied, but the method of the present invention can also be applied to, for example, various characteristic measuring devices equipped with a positioning mechanism that utilizes scribe line detection. can do.

In the first article, the symbol W is a substrate such as a semiconductor wafer on which a scribe line is formed.

The substrate W is placed on an XY stage 1 that moves horizontally in two orthogonal directions (X and Y directions). Code 2 is the substrate W
The illumination light emitted from the light source 3 is reflected onto the substrate W via the half mirror 4 and the objective lens 2.

The reflected light from the substrate W enters the half mirror 5 via the objective lens 2 and the half mirror 4, and a part of it is reflected and enters the television camera 6, where the surface of the substrate W is imaged in an enlarged manner. A series of optical means consisting of the light source 3, half mirror 4, objective lens 2, and half mirror 5 constitute an epi-illumination microscope optical system.

Note that the optical system required to carry out the method according to the present invention is not limited to the structure shown in the first diagram, but may be any system as long as it constitutes an epi-illumination microscope optical system.

The multi-gradation image of the surface of the substrate W is provided to the CRT 8 via an image processing unit 7 that performs alignment processing to be described later, and an enlarged image of the surface of the base Fiw is displayed. ! ii
After detecting the center position of the intersection area of the scribe lines of the substrate W, the image processing unit 7 detects the film thickness measurement spot superimposed on the image of the substrate W on the CRTB and the center position of the intersection area. The XY stage 1 is moved via the XY stage drive circuit 9 by the amount of deviation, and the film thickness measurement spot is moved to the center of the intersection area of the scribe lines (
center of gravity).

After being aligned in this way, the reflected light from the film thickness measurement spot on the substrate W is transmitted to the objective lens 2, half mirror 4,
5 and enters the spectrometer 10 to measure the film thickness.

Hereinafter, the position detection operation of the scribe line intersection area in this embodiment will be explained with reference to the flowchart of FIG.

Step Sl: First, an enlarged multi-tone image of the scribe line crossing area of the substrate W captured by the television camera 6 is taken into the image processing unit 7. The image taken into this image processing unit 7 is an image taken through the epi-illumination U4@optical system as was done previously, so the gradation value of the edge of the scribe line on the substrate W is small.

This is because the edge of the scribe line is inclined with respect to the substrate surface due to the difference in height between the scribe line area and the pattern area on the substrate surface. The light that illuminates
This is because at the edges of the scribe line, the light is reflected out of the entrance pupil of the objective lens 2 and does not reach the television camera 6.

Note that when observing the board surface with a magnifying glass or photographing the board surface under natural light, the board is illuminated with light from a wide range of directions, so only the edges of the scribe lines are dark (with small gradation values). ) In the present invention, images with small gradation values at the edge of the scribe line are photographed as described above. This is because you are photographing.

(a) of FIG. 3 shows that the C
Multi-story 11 projected on RT8! showing the image. At this time, the scribe lines SX and SY are
The positioning of the XY stage 1 is performed in advance so as to substantially match the Y force direction (displayed with a cursor on the CRT screen).

Step Sl Here, the position of the scribe line SY is detected by processing the multi-tone image taken into the image processing unit 7. The specific process of step S2 is as follows.

Step S2. : As shown in Fig. 3(a), if the upper left corner of this figure is the origin (0, 0), each pixel L XO+ LXl+ .
Starting from Lxxss, Y-0 to Y-23 respectively
The gradation values are integrated (integrated) for each of the 256 pixel columns parallel to the Y direction extending up to 9. Figure 3 (b) is a distribution diagram of the Y-direction gradation integral values obtained in this way. It is. In the figure, the joint axis shows the integral value converted by replacing the maximum integral value with rl and OJ.

Step S2t Next step S2. In this step S2., the group of integral values obtained in Step S2 is sorted in order from the integral value indicating the minimum value along with its coordinates.
corresponds to the minimum value detection process in the present invention. In addition,
FIG. 4 shows an example of sorted integral value data. In FIG. 4, the integral value is calculated in step 32. Conversion (replace the maximum value with rl and OJ and convert)
Shows raw data that has not been performed. As mentioned above, when the surface of the substrate W is imaged by epi-illumination, the edges and edges of the scribe line (LX (+
The line along Lllj) becomes dark, so the line along Lllj
The integral value near the line i+LXj is smaller than the integral value of other lines.

Step S2. ; Based on the result of sorting in step S2□, as shown in the following formula
l 41J ), the coordinate values having the next smallest integral value are subtracted in order to calculate the distance W from the coordinate XMIN.

Wa ” l XIIIIN X-l ・
...■ Here, n-1, 2, 3, . . .

FIG. 4 shows the interval W4 determined in this way.

Step S24 Next step S2. The interval W calculated by
, matches the predetermined width of the scribe line SY. Here, it is assumed that there is a match within a range of ±5% with respect to the step SYO width, and if there are multiple matches, the closest match is selected. If the interval W matching the width of the scribe line SY is not detected, an error is displayed.

Step S2. When an interval matching the width of the varnish line SY is selected, the corresponding X coordinate XI (
-X, IN) and Xi are stored. The above locus mx,,
χ8 corresponds to the position coordinates of the edge of the scribe line SY.

Step S3 When the position coordinates of the scribe line SY are determined in step S2 as described above, the position coordinates of the scribe line SX are similarly determined. (C
) is each pixel lined up on a straight line in the Y direction at x=0. +
The distribution of each integral value obtained by integrating the gradation value for each of 240 pixel columns parallel to the X direction extending from X-0 to ,In the figure,Y,,Y.

is the scribe line SX obtained in step S3
is the position coordinate of the edge of

Step S4 2 Step S2. Coordinate values x, , x, , Y of the edges of the scribe lines sy and sx found in S3
,, from Yz, the coordinates of the intersection of the edges of the scribe lines in the X and Y directions (XYl), (x,, Yl)
, (Xl, Yt'')(Xl, Yz) are determined, and these four coordinate positions are detected as the positions of the four corners of the intersection area of the scribe lines, and then the following equation
, ■, the coordinates Gx, Gy of the center G of the intersection area of the scribe lines SY, Sχ are calculated.

Yl +Y.

GV-...■ Step S5 When the center coordinates C,,C, of the intersection area of the varnish line lines SY and SX are calculated, the coordinates between the measurement spot MP shown on C and RT8 and the center G are calculated. After determining the amount of deviation, the XY stage 1 is moved via the XY stage drive circuit 9 so that MP coincides with the center C.

By the above processing, the alignment of the measurement spot MP is completed, and the film thickness measurement at that point is performed.

12 旌± The feature of this embodiment is step S6 of the flowchart shown in FIG. S7, and other image import processing (
step S1), intersection area position detection processing (step S1),
4) The process of moving the measurement spot MP (step 35) is the same as in the first embodiment, so the explanation here will be omitted.

The process of step S6.37 will be described in detail below.

Step S6. :Similar to step S2+ of the first embodiment, integration processing is performed in the Y direction for each pixel column arranged in the X direction. FIG. 6(a) is an enlarged upper image of the scribe line intersection area of the substrate W. , FIG. 3B is a distribution diagram of Y-direction integral values of each pixel column.

Step S6! Ni step 36. Differential processing is performed on each of the integral values obtained in . Here, the efficiency of the subsequent inflection point detection process is improved by calculating the average DA of the absolute values of each differential value and ignoring differential values less than ±DA. FIG. 6(d) is a differential value distribution diagram showing the result of differential processing of the integral value shown in FIG. 6(b). Further, FIG. 7 shows an example of differential value data obtained by such processing and coordinate values corresponding thereto.

Step S6. Based on the result of the differential processing obtained in the second step S62, the maximum positive and negative differential values between which the number of pixels is within the number of pixels (for example, 10 pixels) corresponding to the edge width of the scribe line are determined by a pair of inflections. Turn it on and adopt it.

Two sets of such inflection points are detected, and the midpoint of the coordinate values of these positive and negative inflection points is set as the coordinate XlX2 of the edge of the scribe line SY (see FIG. 7).

Ste, bu S6. ~S6. :Step 32f of the first embodiment
Similar to l-32s, calculation of distance between edges, judgment of match of line width, X3. Stores the Xi coordinates.

Step S7: In the same manner as step S6, the position coordinates Y, Y, of the scribe line SX are determined. Figure 6 (C
) is a distribution diagram of the X-direction integral values of each pixel row arranged in the X-direction,
Figure (e) is a distribution diagram of the differential values.

Thereafter, the scribe line SY is determined in step S4.

The center coordinate value of the intersection area of SX is calculated, and the measurement sub-board-MP is moved (positioned) in step S5.

According to this embodiment, step S6. The integral value obtained in is further differentiated to find an inflection point, and the position of the intersection area of the scribe lines sy and sx is found from the coordinate values of the inflection point, so the number of processes is lower than in the first embodiment. Although the number increases, the scribe line can be detected with high accuracy.

Star 11 Blood■ The feature of this embodiment is step S8 of the flowchart shown in FIG. There is S94.

Hereinafter, step S8. The process of S9 will be explained in detail.

Step S81; Similar to the stencil S2 of the first embodiment, the scribe line intersection area of the 9111(a) ii group 7 anti-W is subjected to integration processing in the X direction for each pixel column arranged in the X direction. FIG. 3B is an enlarged close-up image of FIG.

Step S8? : Ste Nbu S8. The minimum integral value I NIN is detected from the integral value group obtained in .

Step S8. : From the relative ratio between this minimum integral value 11418 and the maximum value (here, set to rx o o), the threshold value TH is calculated using the following formula (2).

TH-111111+, r■■σ-01T; low (%
) y-----■ That is, TH for the maximum integral value
A threshold value THv is set at %.

Step S84: Threshold T determined as described above
Using Hv, step S8. The integral value obtained by 2
Value processing. FIG. 9(d) shows the result of binarizing the integral value shown in FIG. 9 (bl) using the threshold value TH.

Step S8. : In the binarized waveform shown in Fig. 9(d), find the center coordinates X, , The interval W6 is calculated from the values L and X2.

Step SB6.38. :Step S2i of the first embodiment
, S2s, the determination of matching of line widths and the coordinate values X+ and Xt are stored.

Step S1 Similarly to step S8, the position coordinates Y+ and Yt of the scribe line Sx are determined. Figure 9 (C)
is a distribution diagram of integral values of pixel rows arranged in the X direction, (e)
is a distribution diagram of binarized data obtained by binarizing the integral value of FIG.

Thereafter, the scribe line SY is determined in step S4.

The position of the intersection area of SX is detected, and in step S5, movement (alignment) of the measurement subframe MP is performed.

■Douriboat team As shown in Fig. 10, the features of this embodiment are the scribe line detection method based on the minimum value detection of the integral value shown in the first embodiment, and the integral detection method shown in the second embodiment. This method combines a scribe line detection method based on value inflection point detection.

That is, in the detection of the scribe line based on the detection of the minimum value of the integral value, the calculated X9. χ.

If the interval W4 does not match the predetermined line width of the scribe line SY, steps 362 to S6. from step S2° to step S6' are performed. Next, the position of the scribe line SY is detected using a method based on detecting the inflection point of the integral value, that is, a scribe line detection method that is more accurate than the former method.The scribe line SY is not detected even with this method. An error message will be displayed if the error occurs. Similarly, if the position of the scribe line SX cannot be detected based on the detection of the minimum value of the integral value, step S3. The process then proceeds to step S7', where the position of the scribe line SX is detected based on the detection of the inflection point of the integral value, similar to step S6'.

Using any of the above methods, the edge coordinates X1. of the scribe line 5YSX. If Xi, Yl+ Y- is detected, step S4. Proceed to $5, scribe line s
Position detection and measurement spot MP of intersection area of y, sx
Positioning is performed by moving the.

According to this embodiment, only when the position of the scribe line cannot be detected based on the detection of the minimum value of the integral value, the position of the scribe line is detected based on the highly accurate detection of the inflection point, although the processing process is somewhat complicated. Because I try to do it,
More practical scribe line detection can be performed.

Note that step 32 of the first embodiment. , step S68 of the second embodiment, step S81 of the third embodiment,
Step S2 of the fourth embodiment. Moving average processing may be performed on each integral value of the integral value group obtained in . By doing so, small changes between integral values that are not related to the detection of scribe line intersection areas can be removed. to highlight major changes.

In addition, in each of the above embodiments, since the film thickness measuring device using the method according to the present invention measures the center position G of the scribe line intersection area, the scribe line Although the position G of the center of the intersection area has been calculated, for example, if the measurement target is at a position different from such position G, instead of the above formulas It may be calculated based on the position of the scribe line intersection area detected by the method according to the present invention using another measurement formula that defines the positional relationship with the area.

As described above, the present invention is not limited to detecting the center of the scribe line intersection area, but is capable of detecting the position of the scribe line intersection area from which a required position such as the center position is calculated. be.

In each of the embodiments described above, the minimum value of the gradation value is detected when, for example, brightness, that is, the light intensity caught at the imaging position is used as the gradation value.
It is obvious that if a density value is used as a gradation value, the maximum value should be detected instead of the minimum value.
Since both are technically equivalent, the latter case is also included in the present invention.

<Effects of the Invention> As is clear from the above description, according to the present invention, since the substrate on which the scribe line is formed is photographed using an epi-illumination microscope optical system, the edge portion of the scribe line in the photographed image is The first invention detects the minimum value by integrating the multi-gradation image data of the board photographed in the X and Y directions as the tone value becomes smaller, and by using the integrated values. Based on this, the second invention is based on detecting an inflection point of the integral value, and the third invention is based on determining a threshold value in relation to the minimum value of the integral value to obtain a group of binarized integral values. Therefore, in the fourth invention, when the scribe line cannot be detected by detecting the minimum value of the integral value, the position of each intersection area of the scribe line is detected by detecting the inflection point of the integral value. , if the position of the intersection area of the scribe line becomes impossible to detect because the scribe line itself cannot be detected as in the method based on binarization processing using the conventional modal method (no 1 occurs, the surface state The position of the intersection area of scribe lines can be detected with high accuracy for various substrates.

[Brief explanation of the drawing]

1 to 4 relate to the first embodiment G of the present invention, FIG. The operation flowchart of the first embodiment, FIG. 3 is a schematic diagram for explaining the integral processing of multi-tone image data, FIG. 4 is a schematic diagram of sorted integral value data, and FIG. 5 is the operation flowchart of the second embodiment. , No. 6m is a schematic diagram for explaining integral value processing and inflection point detection processing of multi-tone image data, and FIG. 7 is a schematic diagram of differential value data used for inflection point detection. FIG. 8 is an operation flowchart of the third embodiment, and FIG. 9 is a schematic diagram for explaining the integration process of multi-tone image data and its binarization process. FIG. 10 is an operation flowchart of the third embodiment. Figures 11 to 13 relate to the explanation of the conventional method.
Figure 1 shows an enlarged image of the intersection area of scribe lines on the board, Figure 12 shows an example of the tone value histogram of the image in Figure 11, and Figure 13 shows the problem of the conventional example. FIG. 3 is a diagram showing a PM adjustment value histogram for explaining points. W8. , board 1...XY stage 2...
...Objective lens 3...Light source 4.5...Half mirror 6...Television camera 7...Image processing unit 8...CRT9...XY stage drive circuit 10...Spectrometer applicant Dainippon Screen Mfg. Co., Ltd. Representative Patent Attorney Tsutomu Sugitani Figure (a) Figure H Gradation value diagram (b')

Claims (4)

[Claims] (1) For each pixel row in the X and Y directions perpendicular to the scribe line, integrate the gradation values in order for the multi-gradation image data of the scribe line intersection area on the substrate surface, which is magnified and imaged by the epi-microscope optical system. an integral processing process for performing the processing; a minimum value detection process for detecting the two minimum integral values from among the group of integral values in the X and Y directions obtained in the above integration process; From the coordinate values X_1, X_2, Y_1, Y_2 corresponding to the two detected integral values,
an interval calculation step of calculating an interval in the X direction and an interval in the Y direction, and a determination step of determining whether the calculated intervals in the X and Y directions each match a predetermined scribe line width; Coordinate values X_1 of two integral values determined to match the scribe line width in the X and Y directions, respectively;
Coordinate position based on X_2, Y_1, Y_2 (X_1, Y
_1), (X_2, Y_1), (X_1, Y_2), (
A method for detecting the position of a scribe line intersection area, characterized in that X_2, Y_2) is the position of a corner of the scribe line intersection area. (2) For each pixel row in the X and Y directions perpendicular to the scribe line, integrate the gradation values in order for the multi-gradation image data of the scribe line intersection area on the substrate surface, which is magnified and imaged by the epi-illumination microscope optical system. From the integral processing step that performs processing, the differential processing step that performs differential processing on the integral values in the X and Y directions obtained in the above integration processing, and the results of the differential processing obtained respectively in the X and Y directions, An inflection point calculation process that calculates two inflection points in each of the X and Y directions, and an interval in the X direction and an interval in the Y direction from the coordinate values corresponding to the two inflection points in each of the calculated X and Y directions. and a determination step that determines whether the calculated X and Y direction intervals match predetermined scribe line widths, respectively, for each of the X and Y directions. coordinate values X_1 of two inflection points determined to match the scribe line width;
Coordinate position based on X_2, Y_1, Y_2 (X_1, Y
_1), (X_2, Y_1), (X_1, Y_2), (
A method for detecting the position of a scribe line intersection area, characterized in that X_2, Y_2) is the position of a corner of the scribe line intersection area. (3) For each pixel column in the X and Y directions perpendicular to the scribe line, integrate the gradation values in order for the multi-gradation image data of the scribe line intersection area on the substrate surface magnified by the epi-illumination microscope optical system. an integral processing step that performs the processing; a minimum value detection step that detects the minimum integral value from among the group of integral values in the X and Y directions obtained by the integral processing; In relation to the integral value,
, a threshold calculation process of calculating two thresholds according to the Y direction, and two integral value groups in the X and Y directions based on the two thresholds.
A binarization process for digitizing, and central coordinate values X_1, X_2, Y_1, Y_2 in the area corresponding to the edge of the scribe line based on the binarized integral value groups in the X and Y directions. from,
an interval calculation step of calculating an interval in the X direction and an interval in the Y direction; and a determination step of determining whether the calculated intervals in the X and Y directions respectively match predetermined scribe line widths, Two center coordinate values X_1 and X_ that are determined to match the scribe line width in the X and Y directions, respectively.
2. Coordinate position based on Y_1, Y_2 (X_1, Y_1
), (X_2, Y_1), (X_1, Y_2), (X_
2. A method for detecting the position of a scribe line intersection area, characterized in that Y_2) is the position of a corner of the scribe line intersection area. (4) Integrate the gradation values in order for each pixel column in the X and Y directions perpendicular to the scribe line for the multi-gradation image data of the scribe line intersection area on the substrate surface, which is magnified and imaged by the epi-illumination microscope optical system. an integral processing process for performing the processing; a minimum value detection process for detecting the two minimum integral values from among the group of integral values in the X and Y directions obtained in the above integration process; an interval calculation step of calculating an interval in the X direction and an interval in the Y direction from coordinate values corresponding to the two detected integral values; and a step of calculating the interval in the X direction and the interval in the Y direction from coordinate values corresponding to the two detected integral values; A first judgment process to determine whether or not they match; and in the first judgment process, if the two do not match, a differential process is performed on the integral values in the X and Y directions obtained in the integration process, respectively. an inflection point calculation process of calculating two inflection points in each of the X and Y directions from the results of the differential processing obtained in each of the X and Y directions, and the calculated X and Y directions. an interval calculation step of calculating an interval in the X direction and an interval in the Y direction from coordinate values corresponding to each of the two inflection points; and a step of calculating the interval in the X direction and the interval in the Y direction from the coordinate values corresponding to each of the two inflection points; a second judgment step of judging whether or not they match; and coordinate values X_1, X_2, Y_1 of the two integral values that were judged to match the scribe line width in the X and Y directions, respectively, in the first judgment step. , Y_2 coordinate positions (X_1, Y_1), (X_2, Y_1), (X_1
, Y_2), (X_2, Y_2) or the coordinate values X_1 of two inflection points determined to match the scribe line width in the X and Y directions in the second determination process, respectively.
′, X_2′, Y_1′, Y_2′ coordinate position (
X_1', Y_1'), (X_2', Y_1'), (X
_1', Y_2') and (X_2', Y_2') are positions of corners of the scribe line intersection area.