JP2001266126A - Method and device for detecting defect and method of manufacturing for mask - Google Patents

Abstract

PROBLEM TO BE SOLVED: To remove a pseudo defect due to the error of reference picture data and to highly precisely detect a micro defect. SOLUTION: At the time of comparing sensor picture data obtained by image- picking up a mask 6 by a sensor 4 with reference picture data and detecting the defect of a mask pattern formed on the mask 6, a data identification part 11 identifies an error occurred in reference picture data by using highly precise reference picture data generated by a highly precise reference picture generation part 9. A correction registration part 12 corrects/registers the result of the defect detection of the mask pattern based on the identification result.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a defect detecting method and apparatus for detecting a pattern defect such as a semiconductor chip or its mask pattern, a liquid crystal substrate or its mask pattern as an object to be inspected, as well as these defect detecting methods and their The present invention relates to a method for manufacturing a mask to which the apparatus is applied.

[0002]

2. Description of the Related Art Such a method of inspecting a defect of a pattern is described as follows.
For example, image data obtained by imaging a mask on which a pattern of a semiconductor chip is formed as an object to be inspected is compared with reference image data created from design data, and a defect in the mask pattern is determined based on the comparison result. Detected.

[0003]

SUMMARY OF THE INVENTION Recent OPC (Optica)
In a small pattern such as an l Proximity Correction pattern, a rounding error occurs when converting from design data to reference image data when converting the coordinates into pixel units. For example, a step pattern does not appear where a step pattern should appear as in the reference image data shown in FIG. 4A, and a rounding error a occurs in this part. Therefore, when this reference image data is compared with image data (sensor image) obtained by imaging the mask shown in FIG. 2B, a pseudo-defect is obtained regardless of whether the rounding error portion has no pattern defect on the mask. Is erroneously detected.

As described above, the coordinate values in the conversion from the design data to the reference image data into the coordinates in pixel units are originally integer values, but in order to reduce the rounding error a,
One pixel of image data is complemented to 1 / N. However, when one pixel is divided into 1 / N, the data amount becomes a square of N times, and the circuit for processing this data amount becomes large. For this reason, one pixel is actually 1/4.
Or, it is restricted by dividing into about 1/8 (N = 4, 8).

For example, FIG. 5A shows an example of an edge pattern in 4 × 4 pixels. One pixel of this image data is 1
/ N (for example, N = 4), as shown in the partial enlarged view of one pixel s in FIG. The value of the pixel at the edge of the portion is 3/4 since 12 pixels out of 16 pixels have the white level as shown in FIG. Similarly, the value of the pixel at the edge of each one pixel portion where the pattern edge takes half becomes 1/2. The value of the pixel at the edge of each one-pixel portion of the black level is 0.

However, as shown in FIG. 5 (b), if the pattern edge position is shifted by 1 / N pixel or less, that is, over the Z portion in one pixel, it is impossible to obtain the value of the edge pixel accurately. Can not.

Further, when the coordinate value is restricted and a rounding error occurs, the reference image data contains an error in the change in luminance at the edge portion. Furthermore, when converting to luminance, the luminance is
Since it is usually represented by 256 stages (8 bits), a quantization error occurs in calculation.

Therefore, the above-described error in the reference image data is erroneously detected as a pseudo defect when detecting a minute defect.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a defect detection method and apparatus capable of detecting minute defects with high accuracy by eliminating pseudo defects due to errors in reference image data.

Further, the present invention provides a mask manufacturing method capable of improving the reliability of mask manufacturing by detecting a small defect in a mask pattern with high accuracy by eliminating a pseudo defect due to an error in reference image data of the mask pattern. The purpose is to provide.

[0011]

According to a first aspect of the present invention, there is provided a defect detecting apparatus for detecting a defect in an object to be inspected by comparing image data obtained by imaging the object to be inspected with reference image data. In the method, an error generated in the reference image data is identified, and a result of the defect detection is corrected based on the identification result.

According to a second aspect of the present invention, in the defect detecting method according to the first aspect, a high-precision reference image data created from the image data and the design data is used for a peripheral image area including the error part. Are compared, and an error generated in the reference image data is identified based on the comparison result.

According to a third aspect of the present invention, in the defect detection method according to the first aspect, for a peripheral image area including the error portion, the reference image data and the reference image data including an error registered in advance are used. Are compared, and an error generated in the reference image data is identified based on the comparison result.

According to a fourth aspect of the present invention, in the defect detection method according to the first aspect, the object to be inspected is re-detected by changing a pixel resolution at the time of developing the design data into the reference image data. It is characterized by.

According to a fifth aspect of the present invention, there is provided a defect detecting apparatus for detecting a defect of the inspection object by comparing image data obtained by imaging the inspection object with reference image data. A defect detection apparatus comprising: identification means for identifying an error occurring in data; and correction means for correcting the result of the defect detection based on the identification result by the identification means.

In the defect detecting apparatus according to the present invention, the identification means compares the image data with high-precision reference image data created from design data for a peripheral image area including the error portion. It has a function of identifying an error generated in the reference image data.

Further, in the defect detection device of the present invention, the identification means compares the reference image data with a previously registered reference image data including an error in a peripheral image region including the error portion. It has a function of identifying an error generated in the reference image data.

The defect detecting apparatus according to the present invention is characterized in that the defect detecting apparatus further comprises a re-inspection means for re-detecting the object to be inspected by changing a pixel resolution at the time of developing the design data into the reference image data. I do.

According to a sixth aspect of the present invention, there is provided a mask manufacturing method for forming a mask on a substrate, drawing a mask pattern on the film, and thereafter performing a defect inspection of the mask pattern to manufacture a mask. A step of identifying an error occurring in the reference image data, and a step of correcting the result of the defect detection based on the identification result of the step.

In the method of manufacturing a mask according to the present invention,
For a peripheral image area including the error portion, comparing the image data and high-precision reference image data created from the mask design data to identify an error generated in the reference image data. I do.

In the method of manufacturing a mask according to the present invention,
An error occurring in the reference image data is identified by comparing the reference image data with reference image data including a previously registered error in a peripheral image region including the error portion.

In the method of manufacturing a mask according to the present invention,
It is characterized in that the object to be inspected is re-detected by changing the pixel resolution at the time of development from the design data to the reference image data.

[0023]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (1) Hereinafter, a first embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a configuration diagram of a mask pattern defect detection apparatus. The design database 1 stores, for example, mask pattern design data 2 for manufacturing a semiconductor designed by a CAD system. The reference image data 3 is developed from the design data 2 stored in the design database 1 to create a black-and-white binary image, and then multiplied by a blur function equivalent to the characteristics of the optical system to obtain an image equivalent to the sensor 4. Is what you get. This reference image data 3 has been sent to the comparison unit 5.

On the other hand, above the mask 6, the sensor 4 is disposed. The sensor 4 is, for example, a CCD camera or a television camera, and captures an image of the mask 6 and outputs an image signal thereof. This image signal is
The data is A / D converted and sent to the comparison unit 5 as image data (hereinafter, referred to as sensor image data).

The comparator 5 compares the sensor image data with the reference image data to detect a minute defect of the mask 6, and determines the center coordinates (X, Y) and type of the minute defect as an inspection result 7. It has a function to obtain. The defects detected by the comparing unit 5 include a pseudo defect due to a rounding error of the reference image data.

The storage unit 8 stores sensor image data, reference image data, and sensor image data of a peripheral image area including coordinates (X, Y) at which a minute defect is detected as defect image data. Inspection results 7 such as the center coordinates (X, Y) and types of various defects are stored.

The high-precision reference image generator 9 retrieves the corresponding design data 2 from the design database 1 based on, for example, the name of the mask 6 and the like, and finds the center coordinates (X, Y) of the minute defect in the design data 2. ) Is read out only, and a black-and-white binary image is created by developing the partial design data with high accuracy from the partial design data, and then blurring equivalent to the characteristics of the optical system is performed. It has a function of generating high-precision reference image data of a predetermined small area equivalent to the sensor 4 by multiplying the functions.

The comparison unit 10 receives the high-precision reference image data generated by the high-precision reference image generation unit 9 again, reads out the sensor image data stored in the storage unit 8, and It has a function of comparing with sensor image data. Note that the high-precision reference image data is image data of a predetermined small area only around the periphery including the center coordinates (X, Y) of the defect. Therefore, when comparing with the sensor image data, the center coordinates (X, Y ) And compare.

The comparing section 10 also has a function of comparing the high-precision reference image data with the reference image data stored in the storage section 8 again.

The discrimination unit 11 does not detect any minute defect from the comparison result between the high-precision reference image data and the sensor image data by the comparison unit 10 again, and the high-precision reference image data is stored in the storage unit 8. If there is a difference between the reference image data and the reference image data, the reference image data has a rounding error, and has a function of recognizing that a defect at the center coordinates (X, Y) is a pseudo defect.

When the identification unit 11 recognizes that the defect of the center coordinates (X, Y) is a pseudo defect, the correction registration unit 12 checks the center coordinates (X, Y) of the inspection results 7 stored in the storage unit 8. , Y) are removed and re-registered.

Next, the operation of the device configured as described above will be described.

In the mask manufacturing process, first, a Cr light-shielding film is formed on a substrate such as quartz glass by vapor deposition. Next, a mask pattern is drawn by irradiating the Cr light-shielding film on the substrate with an electron beam. Next, a defect inspection of the drawn mask pattern is performed. If there is no defect, manufacturing of the mask is completed, and if there is a defect, its correction is performed.

In the manufacturing process of such a mask, the defect inspection of the mask pattern is performed as follows.

The reference image data 3 is stored in the design database 1
Is developed from the design data 2 stored in the image data, a black-and-white binary image is created, and then multiplied by a blur function equivalent to the characteristics of the optical system to obtain an image equivalent to the sensor 4.

On the other hand, the sensor 4 captures an image of the mask 6 and outputs an image signal of the image. This image signal is A / D converted and sent to the comparison unit 5 as sensor image data.

The comparator 5 compares the sensor image data with the reference image data to detect a minute defect of the mask 6, and determines the center coordinates (X, Y) and type of the minute defect as an inspection result 7. obtain. The detected defect includes a pseudo defect due to a rounding error of the reference image data.

The sensor image data, the reference image data, the defect image data of the peripheral image area including the coordinates (X, Y) at which the minute defect is detected, the center coordinates (X, Y) of the minute defect, and the like. The inspection result 7 such as the type is stored in the storage unit 8
Is stored in

The high-precision reference image generator 9 retrieves the corresponding design data 2 from the design database 1 based on, for example, the name of the mask 6 and the like, and the center coordinates (X, Y) of the minute defect in the design data 2. ) Is read out only, and a monochrome binary image is created by developing the partial design data with high precision from the partial design data, and then multiplied by a blur function equivalent to the characteristic of the optical system to obtain the sensor 4. The same high precision reference image data of a predetermined small area is generated.

The comparison unit 10 receives the high-precision reference image data generated by the high-precision reference image generation unit 9 again, reads out the sensor image data stored in the storage unit 8, and The sensor image data is aligned with the center coordinates (X, Y) of the defect and compared.

The comparing section 10 again compares the high-precision reference image data with the reference image data stored in the storage section 8.

The discrimination section 11 does not detect a minute defect from the comparison result between the high-precision reference image data and the sensor image data by the comparison section 10 again, and the high-precision reference image data is stored in the storage section 8. If there is a difference between the reference image data and the reference image data, there is a rounding error in the reference image data, and the defect at the center coordinates (X, Y) is recognized as a pseudo defect.

When the identification unit 11 recognizes that the defect of the center coordinates (X, Y) is a pseudo defect, the correction registration unit 12 checks the center coordinates (X) of the inspection results 7 stored in the storage unit 8. , Y) are removed and re-registered.

As described above, in the first embodiment, a rounding error generated in the reference image data is detected, and a high accuracy generated from the sensor image data and the design data for the peripheral image area including the rounding error is detected. Since the defect is re-inspected by comparing with the reference image data, a rounding error of the reference image data, that is, a pixel pattern when developing from the design data to the reference image data in a minute pattern such as a recent OPC pattern is used. Pseudo defects due to rounding errors that occur when converting to unit coordinates can be eliminated,
Only minute defects in the mask 6 can be detected with high accuracy.
Thereby, the reliability of mask production can be improved.

Since the defect inspection result of the mask 6 is corrected based on the result of the defect re-inspection, a pseudo defect due to a rounding error of the reference image data is eliminated, and the inspection result of only a minute defect in the mask 6 is obtained. Obtainable.

(2) Next, a second embodiment of the present invention will be described with reference to the drawings. The same parts as those in FIG. 1 are denoted by the same reference numerals, and detailed description thereof will be omitted.

FIG. 2 is a block diagram of a mask pattern defect detecting apparatus. The reference pattern storage unit 13 stores a rounding error a for generating a pseudo defect as shown in FIG.
Is registered in advance. Since the shape of the mask pattern and the manufacturing method thereof are predetermined in the portion where the pseudo defect occurs, the coordinate position at which the defect occurs is almost fixed, and the shape of the reference image data that causes the pseudo defect is almost fixed. ing. The reference image data registered in advance is an image area of a predetermined small area around the area including the part where the rounding error occurs.

In the reference pattern storage section 13, reference pattern data (hereinafter referred to as high-precision reference image data) of a predetermined small area around the reference image data including a portion where a rounding error occurs is registered and stored in advance. Have been.

The pattern matching unit 14 performs pattern matching between the reference image data stored in the storage unit 8 and reference image data registered in advance, and when the mask patterns of the two reference image data match, the rounding error is reduced. It is determined that the reference image data is included, and the high-precision reference image data of the small area corresponding to the coordinates having the rounding error is extracted from the reference pattern storage unit 13 and again compared with the comparison unit 15.
Has the function of passing to

The comparison section 15 has a function of comparing the high-precision reference image data extracted from the reference pattern storage section 13 with the sensor image stored in the storage section 8 again.

Next, the operation of the device configured as described above will be described.

In the mask manufacturing process, first, a Cr light-shielding film is formed on a substrate such as quartz glass by vapor deposition. Next, a mask pattern is drawn by irradiating the Cr light-shielding film on the substrate with an electron beam. Next, a defect inspection of the drawn mask pattern is performed. If there is no defect, manufacturing of the mask is completed, and if there is a defect, its correction is performed.

In the manufacturing process of such a mask, the defect inspection of the mask pattern is performed as follows.

The reference image data 3 is stored in the design database 1
Is developed from the design data 2 stored in the image data, a black-and-white binary image is created, and then multiplied by a blur function equivalent to the characteristics of the optical system to obtain an image equivalent to the sensor 4.

On the other hand, the sensor 4 images the mask 6 and outputs an image signal of the image. This image signal is A / D converted and sent to the comparison unit 5 as sensor image data.

The comparator 5 compares the sensor image data with the reference image data to detect a minute defect of the mask 6, and determines the center coordinates (X, Y) and the type of the minute defect as an inspection result 7. obtain. The detected defect includes a pseudo defect due to a rounding error of the reference image data.

The sensor image data, the reference image data, the defect image data of the peripheral image area including the coordinates (X, Y) at which the minute defect is detected, the center coordinates (X, Y) of the minute defect, and the like. The inspection result 7 such as the type is stored in the storage unit 8
Is stored in

The pattern matching unit 14 performs pattern matching between the reference image data stored in the storage unit 8 and reference image data registered in advance, and when the mask patterns of the two reference image data match, the storage unit It is determined that the reference image data stored in No. 8 includes a rounding error, and high-precision reference image data of a small area corresponding to the coordinates having the rounding error is extracted from the reference pattern storage unit 13 and is again sent to the comparison unit 15. hand over.

The comparison unit 15 compares the high-precision reference image data extracted from the reference pattern storage unit 13 and the sensor image stored in the storage unit 8 with the high-precision reference image data and the sensor image data. The defect is aligned at the center coordinates (X, Y) of the defect and compared.

If the identifying unit 11 does not detect a minute defect from the comparison result by the comparing unit 15 again, the reference image data stored in the storage unit 8 has a rounding error, and the defect of the center coordinate (X, Y). Is recognized as a pseudo defect.

When the identification unit 11 recognizes that the defect at the center coordinates (X, Y) is a pseudo defect, the correction registration unit 12 checks the center coordinates (X) of the inspection results 7 stored in the storage unit 8. , Y) are removed and re-registered.

As described above, in the second embodiment, the defect image is re-inspected by comparing the reference image data with the pre-registered reference image data including the rounding error in the peripheral image area including the rounding error portion. I decided to do it,
As in the first embodiment, the rounding error of the reference image data, that is, when converting from design data to reference image data in a minute pattern such as a recent OPC pattern, conversion into coordinates in pixel units is performed. Pseudo defects due to the generated rounding error can be eliminated, and only minute defects in the mask 6 can be detected with high accuracy. Thereby, the reliability of mask production can be improved.

Since the defect inspection result of the mask 6 is corrected based on the result of the defect re-inspection, a pseudo defect due to a rounding error of the reference image data is eliminated, and the inspection result of only a minute defect in the mask 6 is obtained. Obtainable.

(3) Next, a third embodiment of the present invention will be described with reference to the drawings. The same parts as those in FIG. 1 are denoted by the same reference numerals, and detailed description thereof will be omitted.

FIG. 3 is a block diagram of a mask pattern defect detecting apparatus. The identification unit 11 does not detect a minute defect again from the comparison result of the high-precision reference image data and the sensor image data by the comparison unit 10, and also outputs the high-precision reference image data and the reference image data stored in the storage unit 8. 3, there is a rounding error in the reference image data 3, and when it is recognized that the defect at the center coordinates (X, Y) is a pseudo defect, the development from the design data 2 to the reference image data 3 is performed. It has a function of changing the pixel resolution (magnification) at the time of re-detection of a mask pattern defect.

With such a configuration, as in the first embodiment, the identification unit 11 does not detect a minute defect from the comparison result between the high-precision reference image data and the sensor image data. In addition, there is a difference between the high-precision reference image data and the reference image data stored in the storage unit 8, the reference image data has a rounding error, and the center coordinates (X,
When the defect Y) is recognized as a pseudo defect, the identification unit 11 changes the pixel resolution at the time of development from the design data 2 to the reference image data 3 and causes the mask pattern defect to be re-detected.

Thus, the comparing section 5 again detects the minute defect of the mask 6 by comparing the sensor image data with the reference image data, and determines the center coordinates (X, Y) of the minute defect.
And the type are obtained as the inspection result 7.

As described above, in the third embodiment, when it is recognized that the reference image data 3 has a rounding error and the defect at the center coordinates (X, Y) is a pseudo defect,
Since the pixel resolution at the time of development from the design data 2 to the reference image data 3 is changed to re-detect the mask pattern defect, in addition to the effect of the first embodiment, the identification unit 11 performs continuous detection. When it is determined that a pseudo defect has occurred, it takes time to develop a number of pieces of reference image data around the defect.
By performing the re-detection while changing the pixel resolution at the time of the development, the rounding error of the reference image data is eliminated, and the inspection time can be reduced.

The present invention is not limited to the first to third embodiments, but may be modified as follows.

For example, the present invention can be applied to inspection of a semiconductor chip and detection of a pattern defect such as a liquid crystal substrate and its mask pattern.

[0072]

As described above in detail, according to the present invention, it is possible to provide a defect detecting method and apparatus capable of detecting a minute defect with high accuracy by eliminating a pseudo defect due to an error in reference image data.

Further, according to the present invention, a mask manufacturing method capable of improving the reliability of mask manufacturing by detecting small defects in the mask pattern with high accuracy by eliminating pseudo defects due to errors in reference image data of the mask pattern. We can provide a method.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing a first embodiment in which a defect detection device according to the present invention is applied to inspection of a mask pattern.

FIG. 2 is a configuration diagram showing a second embodiment in which the defect detection device according to the present invention is applied to inspection of a mask pattern.

FIG. 3 is a configuration diagram showing a third embodiment in which the defect detection device according to the present invention is applied to inspection of a mask pattern.

FIG. 4 is a schematic diagram of reference image data including a rounding error.

FIG. 5 is a schematic diagram for explaining complementation for one pixel.

[Explanation of symbols]

 1: Design database 4: Sensor 5: Comparison unit 6: Mask 8: Storage unit 9: High-accuracy reference image generation unit 10: Comparison unit 11: Identification unit 12: Correction registration unit 13: Reference pattern storage unit 14: Pattern matching Part 15: Comparison part again

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 2G051 AA56 AB02 AC21 CA04 EB09 5B057 AA03 BA11 CA16 CB16 CE02 CH07 CH11 DA03 DA06 DC16 DC32 5C054 AA05 CC02 EB05 EB07 FC01 HA03 HA05 5L096 BA03 BA20 CA03 EA27 JA03

Claims (6)

[Claims] 1. A defect detection method for detecting a defect of the object to be inspected by comparing image data obtained by imaging the object to be inspected with reference image data, wherein an error generated in the reference image data is identified. And correcting the result of the defect detection based on the identification result. 2. The image data in the surrounding image area including the error part is compared with high-precision reference image data created from design data, and generated in the reference image data based on the comparison result. 2. The method according to claim 1, wherein an error is identified. 3. The method according to claim 1, further comprising comparing the reference image data with reference image data including an error registered in advance in a peripheral image area including the error portion, and generating the reference image data based on the comparison result. 2. The method according to claim 1, wherein an error is identified. 4. The defect detection method according to claim 1, wherein the object to be inspected is re-detected by changing a pixel resolution at the time of developing the design data into the reference image data. 5. A defect detection apparatus for detecting a defect of the inspection object by comparing image data obtained by imaging the inspection object with reference image data, wherein an error generated in the reference image data is identified. A defect detecting device, comprising: an identifying unit that performs the defect detection; and a correcting unit that corrects the result of the defect detection based on a result of the identification performed by the identifying unit. 6. A method of manufacturing a mask, comprising forming a film on a substrate, drawing a mask pattern on the film, and then performing a defect inspection of the mask pattern to manufacture a mask. And a step of correcting a result of the defect detection based on a result of the identification in the step.