JPS61294818A - Defect checking method for blank of photomask - Google Patents

Abstract

PURPOSE:To improve the exactness, the reliability and the efficiency of inspection, by taking an image of fine defects of the photomask blank with the wide visual field of photographing, and processing the obtained image data. CONSTITUTION:The blank 6 is illuminated by the transparent illumination part constituted by the incandescent lamp 8 and the lens 7, and the TV camera 1 takes the image of the region to be checked. The image processing equipment 2 converts the output signal of the TV camera 1 into the digital image data by A/D conversion, and performs with various high-speed image processing containing image adding and subtracting, whose results are displayed by the monitor 4. The controlling part 3 controls the image processing equipment 2 and the sample-traveling mechanism constituted by the X-Y stage 10 and the driving part 5. When the sample is traveled, the video signal of the part containing no defects scarcely changes in spite of traveling. Accordingly the subtraction of the two image data before and after traveling eliminates the shading of the imaging system and the video signal variation caused by foreign bodies of the optical system. Thus only the values of defect part change locally, so the defects can be detected.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a method for inspecting a photomask blank used for forming patterns in IC, LSI, etc.
Especially the scratches that exist on blanks like this.

The present invention relates to a defect inspection method that can automatically detect defects such as deposits and pinholes.

[Background of the invention]

Photomask blank (hereinafter simply referred to as blank)
A thin metal film such as chrome, which serves as a light-shielding film, is formed to a uniform thickness on the surface of a transparent substrate such as a glass plate by a method such as vapor deposition. Therefore, the light-shielding film formed there must be free of scratches, adhesion of foreign matter such as dust, pinholes, etc. as much as possible.

Conventionally, such blank defect inspection has mainly been carried out by an inspector directly or visually using a microscope.

However, this visual inspection method requires more manpower as the number of blanks to be inspected increases, and since it is a sensory test, there are problems in that the inspection accuracy and reliability are lacking.

Therefore, in order to solve this problem, a method has been proposed in which defects are detected by using microscopic imaging means to image a blank and examine the video signal obtained, but it is difficult to detect defects of several microns. Therefore, it is necessary to increase the resolution, which makes it impossible to take a wide field of view and has the disadvantage that the inspection speed is low.

Furthermore, a laser scanning defect inspection method that utilizes the light scattering effect of defects has been proposed, but although it has excellent detection sensitivity, it cannot distinguish between defect types, such as pinholes and foreign objects. In addition, since a precise optical system is required, the equipment becomes expensive.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems and to provide an inspection method that can efficiently inspect defects in blanks and can also discriminate the types of defects.

[Summary of the invention]

In order to achieve this objective, the present invention provides an image processing method that contains defect information even if the video signal has a large field of view in which defects cannot be directly identified. The random noise reduction effect of the video signal by screen addition processing, which is We found that changes in the video signal due to system shading and optical system dust are erased, and the image data in non-defect areas becomes almost 0, and the value changes locally only in the defect area, which makes it possible to detect defects. It is characterized by the fact that it performs the following.

[Embodiments of the invention]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The method for inspecting defects in photomask blanks according to the present invention will be described in detail below with reference to illustrated embodiments.

FIG. 1 shows an embodiment of the present invention, in which a blank 6 is illuminated by a transmissive illumination unit consisting of an incandescent lamp 8 and a lens 7, which are lit by a DC power source 9, and the inspection area is photographed by a TV left camera. The image processing device 2 converts the output signal of the TV left camera into an A/D
It converts the data into digital image data, and performs various types of image processing, including screen addition and subtraction, at high speed using a frame memory and arithmetic unit. Note that the processing status at this time is displayed on the monitor 4. Reference numeral 3 denotes a control unit that controls the image processing device 2, the XY stage 10, and a sample moving mechanism composed of the drive unit 5.

FIG. 2 is an enlarged view of the blank 6, showing that a light shielding film 61 is provided on the surface of a glass substrate 60, and that pinholes 62 and foreign matter 63 are present.

Next, the operation of this embodiment will be explained.

In this embodiment, the direction in which the blank 6 is moved via the drive unit 5 is the scanning line direction of the TV left camera.
This is a case where the moving distance is an integral multiple of the pixel pitch.

First, FIG. 3(a) is a diagram showing the light transmission weight distribution on a straight line passing through the defective area of the blank 6, and 14 is a diagram showing the defect (hereinafter referred to as white) caused by the pinhole 62 of the light-shielding film 61 shown in FIG. 15 indicates the change in light transmittance caused by the defect (hereinafter referred to as a black point) caused by the foreign substance 63 shown in FIG. 2, and FIG. This figure shows the video signal obtained when
It shows gradual signal changes (shading) due to uneven sensitivity on the imaging surface, random noise generated in the video signal processing circuit in the TV right camera, and local signal changes16 due to dust attached to the optical system. .

In addition, FIG. 3 (C1 is a diagram showing the result of screen addition processing by the image processing device 2, and it shows how the random noise component in FIG. 3(b) is reduced to lly when the number of additions is N. There is.

FIG. 3(d) shows the result of screen addition processing by moving the blank 6. The signal due to the defect moves as the blank 6 moves, but the shading of the non-life insurance system and the 16
A signal like the one shown in shows that there is no change.

Figure 3 (el shows the result of subtracting the image data of Figure 3 (dl) from Figure 3 (C), and shows the shading included in the data of Figures 3 (C) and (d) and the like shown in Figure 16. The figure shows how the components that do not change due to the movement of the blank 6 are erased, leaving only the signal due to the change in the light transmittance of the blank 6 and the reduced random noise component.The signal due to the defect is the amount of movement of the blank 6. At the position where the pixel is scattered according to the pixel count, the difference in value from the average value in the vicinity is almost the same and the sign appears reversed, and the order of the reversal is the type of defect (white spot).

The reverse appearance is clearly shown by the black dots.

By the way, in the above embodiment, an example is explained in which subtraction is performed between two screens of image data in which screen addition processing is performed before and after the blank 6 is moved, but as another embodiment, The moved image data may be sequentially subtracted from the added data by the same number of frames as the addition, and even in this case, the result is exactly the same. According to this embodiment, the frame memory can also be processed using only one side.

Therefore, if you observe the image data processed above on the monitor 4, you can easily recognize the defect because the brightness changes locally only in the defective area, and furthermore, the brightness changes in the order of brightness and darkness around the defect. Defect types (white dots, black dots) can be identified. In addition, from this image data, subtraction or differentiation processing of the neighborhood average value, especially when the movement of the blank is a parallel movement, image processing that calculates the difference between corresponding pixels before and after the movement of one point on the blank is performed. When this is done, gradual change components of the image data are also removed as shown in FIG. 3(f), and as a result, defects can be automatically detected by comparison with a predetermined d threshold.

Next, the moving distance and direction of the blank will be explained.

Figure 4 is a diagram explaining the positional relationship between the defect 17.18 and the pixel P, and the image data D after addition and subtraction. An example when it is not an integer multiple is shown. In these figures, 17 represents a defect before movement, and 18 represents a defect after movement.

As is clear from these figures, Fig. 4(a)
In this case, the image data due to the defect is vertically symmetrical with respect to the neighborhood average value, but when it is not an integral multiple, as shown in the example in Figure 4 (bl), the rate at which the image data due to the defect is distributed to adjacent pixels changes before and after movement. The vertical symmetry with respect to the neighborhood average value is lost.This causes errors in automatic detection processing.Therefore, the direction and distance of specimen movement must be determined based on the position of the defect and the pixel. It is desirable to set the relationship to be the same before and after the movement. For example, when moving in the pixel arrangement direction, good detection results can be obtained by setting it to an integral multiple of the pixel pitch in that direction.

Next, as an embodiment of the present invention, a method for reducing fluctuations in the defect signal level caused by the positional relationship between the defect and the pixel will be described.

Figure 5(a) shows that the defect is in the center of the pixel, and Figure 5(b) shows that the defect is in the center of the pixel.
) indicates a state in which defects were present on each of the contact points of four pixels. Even if the defect is the same, in the case of Fig. 5(a), the signal due to the defect is almost concentrated in one pixel, and in the case of Fig. 5(mountain), it is dispersed among four pixels. In some cases, a difference of approximately four times occurs in the defect signal level, resulting in a problem that the reproducibility of defect detection is low. However, most of the dispersion of such a defect signal to surrounding pixels falls within a 3×3 pixel area centered around the defect, and its influence on the outside can be ignored. Therefore, if a spatial filter process used in the field of image processing technology is used to add neighboring pixels of up to 3 x 3 pixels to the image data after addition and subtraction, the sum of the defect signals dispersed in the surrounding pixels at the defective part is Therefore, it is possible to reduce changes in the defect signal level due to the positional relationship between the defect and the pixel. Next, a method for detecting defects and determining types from image data after addition and subtraction in an embodiment of the present invention will be described.

In Figure 6(a), the direction of movement of the blank 6 is the same as the pixel scanning direction, the distance is twice the pixel pitch, and after performing screen addition processing, the blank 6 is moved to the right in the figure, and the same frame as the addition is performed. 20 shows an example of image data obtained by subtracting the white spot defect 21, which shows a local change in image data due to a black spot defect, and the other parts show a gradual change in the transmittance of the blank 6. It represents the change (low frequency change) in the image data in response to the change, which is shown as a neighborhood change value 28.

Here, if we look at the image data of the defective part in detail, we can see that the pixel AC22, 22') corresponding to the defect during screen addition and the pixel B (23, 23') corresponding to the screen subtraction are blank as pixels with defect information. The pixels appear at a displacement distance of 6, that is, the number of pixels (2 pixels) corresponding to 2 pixel pitches, and each pixel data is IA.

Ill, and if Ic is the neighborhood average value of pixel C (24, 24') corresponding to its midpoint, TA-1, and 1s-Tc
have opposite signs and almost the same absolute value, and the order of the signs corresponds to the type of defect (white dot, black dot). It can be seen that the absolute value of il is proportional to the size of the defect. Also, I
It can also be seen that the difference between the average value of A and 7 and I is a sufficiently small ratio with respect to the value of IA-111. Therefore, image data processing that yields IA-I,+, for example, in FIG. 6 (I Tj for al) = I (j-1
1-I (j+11) is calculated as shown in Figure 6(b), and ■(j-11+■3, .1. are each 1.
4. This is data indicating the type and size of the defect for the pixel (25, 25') that becomes Ill, and since the low frequency change mentioned above is also reduced, if compared with a constant threshold value 27,
It becomes possible to detect defects and determine their types.

However, at this time, as shown in FIG. 6(b), the signs are opposite on both sides of the pixel to be detected as a defect, and the value is 1.
-1. Since data (26, 26') that is half of causing inconvenience.

Therefore, in the above IA, IB+ rC (■, +I,
l)/2 1c is 1A-1! For example, by taking advantage of the fact that the ratio is below a certain level with respect to +,
<J+ = (I, J-+, +T<;+n)
/2 1 ti.

(However, IU) is I<j, (D neighborhood average value)
6 (C1) is calculated, and compared with a predetermined threshold value 27, among the defect detection pixels including false defects, the above (rA+Ill
)/2-Ic is less than or equal to a predetermined value, the false defects can be excluded, and stable defect detection and type discrimination can be performed.

Next, to show an example of defect inspection of a blank according to the above embodiment, a photomask blank whose light-shielding film has a transmittance of 0.1% (density 3) is exposed to transmitted light illumination as shown in the first figure. Set the light intensity and the aperture of the photographic lens so that the video signal of the defect-free area is close to the maximum value of the A/D converter, and set the TV using an image pickup tube with a maximum shadow field of view of 10cm x 10cn+. When photographing with a camera, we were able to detect and judge pinholes in the light-shielding film with a diameter of 2 microns in approximately 5 seconds, including the total photographing time and image processing time.

Furthermore, as the TV photographing device and illumination method used in the present invention, any device can be used as long as it can obtain a video signal including defect information. TV cameras using channels and high-sensitivity cameras using channel plates, etc.
Illumination methods include transmitted light illumination, transmitted dark-field illumination, reflected dark-field illumination, regular reflection illumination, etc.As for the properties of illumination light, coherence, spectral characteristics, polarization, etc. can be used, and any of them can be used for detection. S/N of defect signal
Defect inspection can be performed by selecting a signal level that increases the signal level due to defects that do not need to be detected, or a signal level that is low due to defects that do not need to be detected.Furthermore, defect detection can be performed sequentially under multiple imaging conditions and the results can be examined. It is also possible to identify the type of defect in detail.

For example, when a black spot is detected with transmitted light illumination, it is impossible to determine whether it is a foreign object on the light-shielding film or a foreign substance attached to the surface. Since the foreign object on the back side is not detected, the position of the foreign object can be recognized by comparing it with the inspection results using reflected dark field illumination.

〔Effect of the invention〕

As explained above, according to the present invention, minute defects in photomask blanks are automatically detected by photographing them in a wide field of view and processing the image data. Therefore, the disadvantages of the conventional technology can be eliminated and effects such as improved inspection accuracy, reliability, and efficiency can be obtained.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the photomask blank defect inspection method according to the present invention, FIG. 2 is an explanatory diagram of the photomask blank and defects, and FIG. 3 (al to (f)
l, Figure 4 (at, (bl, Figure 5 (al, (b)
, and FIGS. 6(a) to (C1) are explanatory diagrams showing the operation of an embodiment of the present invention. 1...TV camera, 2...image processing device, 3...
Control unit, 4... Monitor, 5... Drive unit, 6... Photomask blank, 7... Lens, 8... Incandescent lamp, 9... DC power supply, 10... XY stage. Figure 1 Figure 2 Figure 3 Figure 4 Figure 5

Claims (5)

[Claims] (1) In a photomask blank defect inspection method using a video signal, providing means for moving the photomask blank to be inspected by a predetermined distance in a predetermined direction within the field of view of the imaging system for imaging, Subtract one frame of second image data based on a video signal captured after the photomask blank to be inspected is captured from one frame of first image data based on a video signal captured before the photomask blank to be inspected is moved. 1. A method for inspecting defects in a photomask blank, characterized in that a defect detection process is performed based on third image data obtained by performing a defect detection process. (2) Claim 1 is characterized in that both the first image data and the second image data are image data obtained by adding each pixel data of a video signal obtained by capturing images a plurality of times. A method for inspecting photomask blanks for defects. (3) In claim 1, it is provided that the predetermined direction is parallel to the horizontal scanning direction during imaging, and the predetermined distance is an integral multiple of the pixel pitch on the imaging surface. Features: Defect inspection method for photomask blanks. (4) A defect in a photomask blank according to claim 1, wherein the defect detection process is configured to include an addition process of neighboring pixel data of up to 3×3 pixels. Inspection method. (5) In claim 1, the defect detection process detects, with respect to the third image data, a point on the photomask blank to be inspected that corresponds to a pixel corresponding to the first image data. A, the pixel corresponding to the above second image data is B, and the pixel corresponding to the midpoint of these pixels A and B is C, then the neighborhood average value of this pixel C and the pixel A
If the difference between the average values of the data of pixels A and B is sufficiently small compared to the difference between the data of pixels A and B, and the difference between the data of pixels A and B is greater than a predetermined level, the point is determined to be defective, and the pixel is determined to be defective. A defect inspection method for a photomask blank, characterized in that the number and size of defects are recognized based on the sign and absolute value of the difference between data A and B.