JP2004219597A - Mask defect inspection apparatus and mask defect inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a mask defect inspection apparatus, capable of detecting any defect of various kinds of patterns formed in a mask, at high sensitivity. <P>SOLUTION: The mask defect inspection device includes a photoelectric conversion element 6 for obtaining a mask image transmitted through a mask 2 to be inspected, an A/D conversion circuit 7 for generating multi-gradation mask image data corresponding to the obtained mask image, a multi-gradation generation part 8 for generating mask reference data by the multi-gradation of 2-gradation mask design data corresponding to the mask to be inspected, and a comparison circuit 9, which extracts a line connecting a predetermined specified gradation value part, as the pattern contour line, from the mask image data generated by the A/D conversion circuit 7, which extracts a line connecting a prescribed gradation value part, as the pattern contour line, from the mask reference data generated by the multi-gradation generation part 8, and which detects the defect of the mask 2 to be inspected, on the basis of the comparison result on the lengths of these pattern contour lines. <P>COPYRIGHT: (C)2004,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a mask defect inspection apparatus and a mask defect inspection method, and more particularly to an apparatus and method suitable for detecting a defect (size defect) such as a local line width error in a mask.
[0002]
[Prior art]
With the miniaturization of semiconductors, the size of patterns formed on photomasks for exposure has been steadily reduced. In recent years, the miniaturization of design rules for mask patterns has been further accelerated, and the development speed of various semiconductor devices has been accelerating. On the other hand, development of a shorter wavelength of an exposure apparatus has not been able to keep up with the technical difficulty. Therefore, a super-resolution technique has been widely used as a life extension measure of an exposure apparatus.
[0003]
While super-resolution technology has the advantage of making patterns finer, it has the advantage that line width errors and defects that occur on the mask are transferred to the wafer compared to when the exposure wavelength is shortened. There is a disadvantage that the influence on the width is large. Therefore, a highly accurate and defect-free mask is required correspondingly, and in the mask defect inspection, further improvement in defect detection sensitivity is required. In particular, when forming a pattern whose pattern size is equal to or less than the exposure wavelength, the influence on the transferability due to the mask line width variation is large, and it is necessary to detect a local line width error defect in the mask defect inspection. Is growing.
[0004]
As a conventional technique related to mask defect inspection, the amount of light transmitted through a certain region (pattern) in a mask is integrated, and the integrated amount of transmitted light is compared with reference data based on mask design data to determine a pattern defect. A detection method is known (see Patent Document 1).
[0005]
[Patent Document 1]
JP-A-2001-82947 (Claims 1, 3, and FIG. 1)
[0006]
[Problems to be solved by the invention]
However, in the above-described conventional technology, when a plurality of chips are included on one mask, it is only necessary to simply add and compare the light transmission amounts of certain areas of the chips. When comparing the amount with the reference data, it is necessary to correct the light transmission amount according to the size of the pattern. In addition, the above-described prior art targets a so-called closed pattern, such as a white pattern having translucency or a translucent pattern, among various patterns formed in a mask, and has a light-shielding property. It cannot be applied to a linear pattern or the like, and therefore has a drawback that it cannot cope with defect detection such as a line width error. Further, as another method, a pattern edge position shift amount is obtained by comparing image data of a mask to be inspected with mask design data, and when the shift amount exceeds a preset threshold, it is detected as a defect. However, with this method, it is difficult to accurately extract the edge position of the pattern, and as a result, there is a problem that the defect detection sensitivity is low.
[0007]
SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a mask defect inspection apparatus and a mask defect inspection apparatus capable of detecting defects of various patterns formed in a mask with high sensitivity. It is to provide an inspection method.
[0008]
[Means for Solving the Problems]
A mask defect inspection apparatus according to the present invention includes a first generation unit that obtains a mask image transmitted through a mask to be inspected and generates multi-tone mask image data, and a two-tone mask corresponding to the mask to be inspected. A second connecting means for generating mask reference data by multiplying the design data into multiple gradations, and a line connecting predetermined predetermined gradation values from the mask image data generated by the first generating means are used as pattern outlines. First extracting means for extracting as a line, second extracting means for extracting a line connecting predetermined tone value portions as a pattern contour line from the mask reference data generated by the second generating means, Defect detecting means for comparing the length of the pattern contour extracted by the extracting means with the length of the pattern contour extracted by the second extracting means, and detecting a defect of the inspected mask based on the comparison result; And it has a configuration.
[0009]
In the mask defect inspection apparatus having the above configuration, the mask image transmitted through the inspection target mask is acquired by the first generation unit, and the multi-tone mask image data corresponding to the mask image is acquired by the first generation unit. Generated. The two-gradation mask design data corresponding to the inspection target mask is multi-graded by the second generation means, and the multi-gradation mask reference data corresponding to the mask design data is multiplied by the second generation means. Is generated. Further, a pattern contour is extracted by the first extracting means from the mask image data generated by the first generating means, and a second extracting means is extracted from the mask reference data generated by the second generating means. Is used to extract a pattern outline. Then, the lengths of the respective pattern contours are compared by the defect detection means, and based on the comparison result, the defect detection of the inspection target mask is performed.
[0010]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[0011]
FIG. 1 is a schematic diagram illustrating a configuration example of a mask defect inspection apparatus according to an embodiment of the present invention. In FIG. 1, a light source 1 emits inspection light for defect inspection to a mask 2 to be inspected. As the inspection light emitted from the light source 1, for example, light having a wavelength of 266 nm can be used. A condenser lens 3 is provided between the light source 1 and the inspection target mask 2.
[0012]
As the inspection target mask 2, for example, a light transmission type photomask in which a mask pattern including a light-shielding pattern portion is formed on a transparent substrate such as glass is used. The inspection target mask 2 is set horizontally on an inspection stage 4 movable in two horizontal axis directions (XY directions). An imaging lens (objective lens) 5 is disposed on the opposite side of the condenser lens 3 (below the inspection stage 4) with the inspection stage 4 interposed therebetween. The imaging lens 5 forms an image of the light transmitted through the inspection target mask 2 on the light receiving surface of the photoelectric conversion element 6 for capturing an image. The photoelectric conversion element 6 acquires a mask image corresponding to the pattern shape on the mask by receiving light transmitted through the inspection target mask 2.
[0013]
As the photoelectric conversion element 6, a CCD (Charge Coupled Device) image sensor can be used. The focus position of the transmitted light image by the imaging lens 5 can be automatically adjusted by moving the imaging lens 5 in the optical axis direction by a focus control circuit (not shown). The inspection mask 2 may be, for example, an electron beam exposure mask other than the light transmission type photomask. In this case, the mask image of the electron beam transmitted through the electron beam exposure mask is converted into a photoelectric conversion element. It will be taken in at 6.
[0014]
The photoelectric conversion element 6 generates analog image data (image signal) according to the amount of light received for each of a plurality of minute areas (read pixels) partitioned on the light receiving surface. On the other hand, an A / D conversion circuit (analog / digital conversion circuit) 7 digitizes the analog image data generated from the photoelectric conversion element 6 and converts the image data into two-dimensional multi-tone image data (hereinafter referred to as a mask). Image data). The A / D conversion circuit 7 constitutes a first generation unit together with the inspection optical system including the photoelectric conversion element 6. The mask image data generated by the A / D conversion circuit 7 is supplied to the multi-tone generation unit 8 and the comparison circuit 9, respectively.
[0015]
The multi-gradation generating unit 8 converts multi-gradation mask design data (such as CAD data) as a reference when forming a mask pattern of the inspection target mask 2 into multi-gradation to be compared with the mask image data. Is generated. When generating the mask reference data, the multi-gradation generation unit 8 converts the two-gradation mask design data into multi-levels so that the number of gradations of the mask reference data is equal to the number of gradations of the mask image data. Perform image processing to adjust. For example, assuming that the number of tones of the mask image data generated by the A / D conversion circuit 7 is eight, the multi-tone generation unit 8 sets two tones (white / black) in accordance with the number of tones. The mask design data of 8) is divided into eight stages to obtain multiple gradations, thereby generating mask reference data of 8 gradations. Further, the multi-tone generation unit 8 extracts a plurality of characteristic patterns from the mask image data, and for each of the extracted patterns, multiplies the mask design data corresponding to each pattern to multi-tone and sets the mask reference data. Generate.
[0016]
The comparison circuit 9 includes a first extraction unit, a second extraction unit, and a defect detection unit. The comparison circuit 9 includes multi-tone mask image data input from the A / D conversion circuit 7 and multi-level mask image data. Using the multi-tone mask reference data input from the tone generation unit 8, the defect detection of the inspection target mask 2 is performed, and the defect detection result is output as defect data. The defect data includes position data for specifying the position of a pattern detected as a defect in the inspection target mask 2 and defect identification data indicating the type of the detected defect.
[0017]
In the mask defect inspection apparatus having the above-described configuration, inspection light emitted from the light source 1 is condensed by the condenser lens 3 and is irradiated on the inspection target mask 2. At this time, the light transmitted through the inspection target mask 2 is captured as an image by the photoelectric conversion element 6 through the imaging lens 5. As a result, the photoelectric conversion element 6 acquires an image (mask image) of the transmitted light transmitted through the inspection target mask 2. The transmitted light image thus obtained is provided to the multi-gradation generation unit 8 and the comparison circuit 9 through the A / D conversion circuit 7. The comparison circuit 9 uses the mask image data input from the A / D conversion circuit 7 and the mask reference data input from the multi-tone generation unit 8 to generate the pattern formed on the inspection target mask 2. Defect detection is performed.
[0018]
Subsequently, a mask defect inspection method according to the embodiment of the present invention will be described. Here, a description will be given of an example of a mask defect inspection method using the mask defect inspection apparatus having the above configuration and a light transmission type photomask as the inspection target mask 2.
[0019]
First, the inspection light emitted from the light source 1 is applied to the inspection mask 2 with the light transmission type photomask to be the inspection mask 2 set on the inspection stage 4, whereby the transmitted light transmitted through the inspection mask 2 is transmitted. Is acquired by the photoelectric conversion element 6. At this time, the mask image of the transmitted light acquired by the photoelectric conversion element 6 is sent to the A / D conversion circuit 7, where multi-gradation mask image data is generated (first generation step). The mask image data generated by the A / D conversion circuit 7 is sent to the multi-tone generation unit 8 and the comparison circuit 9, respectively.
[0020]
Next, the multi-gradation generating unit 8 extracts, for example, some patterns arbitrarily from the image data supplied from the A / D conversion circuit 7 and performs multi-gradation processing corresponding to each pattern image. To generate mask reference data having the same number of tones as the mask image data by multiplying the two-gradation mask design data in accordance with the generated parameters. Generation process). The mask reference data generated by the multi-tone generation unit 8 is provided to the comparison circuit 9. Examples of the parameters for increasing the number of gradations include the number of gradations and gradation values representing the image of each pattern.
[0021]
Here, an example of a pattern shape represented by the mask image data is shown in FIG. FIG. 2B shows an example of a pattern shape represented by mask design data, and FIG. 2C shows an example of a pattern shape represented by mask reference data obtained by multiplying the mask design data by multiple gradations. Shown in
[0022]
Next, the comparison circuit 9 compares the mask image data supplied from the A / D conversion circuit 7 with the mask reference data supplied from the multi-tone generation unit 8, and based on the comparison result, the mask 2 to be inspected. Is detected. Hereinafter, specific processing contents performed by the comparison circuit 9 will be described.
[0023]
First, the comparison circuit 9 extracts a line connecting predetermined gradation values as a pattern contour from the mask image data supplied from the A / D conversion circuit 7 (first extraction step). Further, the comparison circuit 9 extracts a line connecting the predetermined gradation value portions as a pattern contour from the mask reference data provided from the multi-gradation generation section 8 (second extraction step). At this time, for example, when the number of gradations of the mask image data is 8 gradations of “0 to 7” and the predetermined gradation value is “4”, the comparison circuit 9 outputs the mask image data and the mask reference data. Among them, the lines connecting the parts each having the gradation value of 4 are extracted as the pattern contour lines. FIG. 3A shows the result of extracting the pattern contour (OL1) in the mask image data, and FIG. 3B shows the result of extracting the pattern contour (OL2) in the mask reference data.
[0024]
Subsequently, the comparison circuit 9 differentiates the pattern contour OL1 extracted from the mask image data as described above, thereby enclosing the pattern outline OL1 with the line segments a, b, c, and d as shown in FIG. The extracted rectangular pattern part is extracted. Further, in the comparison circuit 9, the pattern outline OL2 extracted from the mask reference data as described above is differentiated, so that the pattern outline OL2 is surrounded by line segments e, f, g, and h as shown in FIG. Extract the rectangular pattern part. Thus, a pattern to be inspected (inspection pattern) in the inspection target mask 2 is extracted.
[0025]
Next, in the comparison circuit 9, the total length (a + b + c + d) of the line segments a, b, c, and d including the pattern outline OL1 of the mask image data, and the line segment e and the line e including the pattern outline OL2 of the mask reference data. The total length (e + f + g + h) of f, g, and h is obtained. Of these, the line segments a, b, c, and d correspond to the outlines of the rectangular pattern surrounded by them, and the line segments e, f, g, and h correspond to the outlines of the rectangular pattern surrounded by them. It will be equivalent.
[0026]
Here, as shown in FIG. 4C, the total length of the line segments a, b, c and d is “L1”, and the total length of the line segments e, f, g and h is “L2”. Then, the comparison circuit 9 calculates the difference ΔL between the lengths L1 and L2, and compares the magnitude of the calculation result with a preset threshold Ls by an absolute value. If the calculated difference ΔL exceeds the threshold value Ls, the pattern represented by the pattern contour line OL1 (the rectangular pattern portion surrounded by the line segments a, b, c, and d) has a line width error defect. (Defect detection step). Thereafter, defect detection is performed on other rectangular patterns connected to the rectangular pattern by the same method as described above. In addition, pattern outlines are extracted for isolated patterns in the same manner as described above, and defect detection is performed by comparing their lengths.
[0027]
As described above, the pattern outline is extracted from the multi-tone mask image data, the pattern outline is also extracted from the multi-tone mask reference data, and the defect is determined based on the comparison result of the pattern outline lengths. By performing the detection, it is possible to detect various pattern defects, particularly, size defects such as line width errors with high sensitivity.
[0028]
In the above-described embodiment, the difference ΔL between the lengths L1 and L2 of the pattern contour lines is obtained, and the difference ΔL is compared with a preset threshold value Ls. The ratio (L1 ÷ L2) of the line lengths L1 and L2 may be obtained, and the ratio (L1 ÷ L2) may be compared with a preset threshold. In that case, the ratio of the lengths L1 and L2 of the pattern contour lines (L1 ÷ L2) is obtained without depending on the size of the original pattern, which is more preferable.
[0029]
Further, in the above embodiment, the light transmission type photomask is used as the inspection target mask 2. However, even when an electron beam exposure mask is used as the inspection target mask 2, the same inspection result is obtained. Obtainable.
[0030]
In addition, the mask defect inspection apparatus and the mask defect inspection method according to the present invention are widely applied to detecting not only light-shielding patterns but also size defects of various patterns such as a light-transmitting white pattern and a semi-transparent pattern. It is possible to do.
[0031]
【The invention's effect】
As described above, according to the present invention, a size defect such as a local line width error can be detected with high sensitivity when performing a defect inspection of a mask to be inspected on which various patterns are formed.
[Brief description of the drawings]
FIG. 1 is a schematic diagram illustrating a configuration example of a mask defect inspection apparatus according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating an example of processing for generating mask image data and mask reference data.
FIG. 3 is a diagram illustrating an example of a pattern contour extraction process.
FIG. 4 is a diagram illustrating an example of comparison processing of the length of a pattern contour line.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Light source, 2 ... Inspection mask, 3 ... Condensing lens, 4 ... Inspection stage, 5 ... Imaging lens, 6 ... Photoelectric conversion element, 7 ... A / D conversion circuit, 8 ... Multi-tone generation part, 9 ... Comparison circuit

Claims (8)

First generating means for obtaining a mask image transmitted through the inspection target mask and generating multi-tone mask image data;
Second generating means for generating mask reference data by multiplying two-gradation mask design data corresponding to the inspection target mask into multiple gradations;
First extracting means for extracting, from the mask image data generated by the first generating means, a line connecting predetermined predetermined tone value portions as a pattern contour;
Second extraction means for extracting a line connecting the predetermined gradation value portion as a pattern contour from the mask reference data generated by the second generation means;
The length of the pattern outline extracted by the first extraction unit is compared with the length of the pattern outline extracted by the second extraction unit, and the defect detection of the inspection target mask is performed based on the comparison result. And a defect detecting unit for performing the inspection. 2. The method according to claim 1, wherein the second generation unit multiplies the mask design data so that the number of tones of the mask reference data is equal to the number of tones of the mask image data. 2. The mask defect inspection device according to 1. The defect detecting means obtains a difference between the length of the pattern contour extracted by the first extracting means and the length of the pattern contour extracted by the second extracting means, and the difference is set to a predetermined threshold value. 2. The mask defect inspection apparatus according to claim 1, wherein when the number exceeds the threshold, the pattern represented by the pattern contour is detected as a defect. The defect detecting means obtains a ratio between a length of the pattern contour extracted by the first extracting means and a length of the pattern contour extracted by the second extracting means, and this ratio is set to a predetermined threshold value. 2. The mask defect inspection apparatus according to claim 1, wherein when the number exceeds the threshold, the pattern represented by the pattern contour is detected as a defect. A first generation step of obtaining a mask image transmitted through the inspection target mask and generating multi-tone mask image data;
A second generation step of generating mask reference data by multiplying two-tone mask design data corresponding to the inspection target mask into multiple gradations;
A first extraction step of extracting a line connecting a predetermined gradation value portion as a pattern contour line from the mask image data generated in the first generation step;
A second extraction step of extracting a line connecting the predetermined gradation value portion as a pattern contour line from the mask reference data generated in the second generation step;
The length of the pattern contour extracted in the first extraction step is compared with the length of the pattern contour extracted in the second extraction step, and the defect detection of the inspection target mask is performed based on the comparison result. Performing a defect detection step. 2. The method according to claim 1, wherein in the second generating step, the mask design data is multi-grayed so that the number of tones of the mask reference data is equal to the number of tones of the mask image data. 5. The mask defect inspection method according to 5. In the defect detecting step, a difference between the length of the pattern contour extracted in the first extracting step and the length of the pattern contour extracted in the second extracting step is obtained, and the difference is determined by a preset threshold value. 6. The mask defect inspection method according to claim 5, wherein when the number exceeds the threshold value, the pattern represented by the pattern outline is detected as a defect. In the defect detecting step, a ratio between the length of the pattern contour extracted in the first extracting step and the length of the pattern contour extracted in the second extracting step is obtained, and the ratio is set to a predetermined threshold value. 6. The mask defect inspection method according to claim 5, wherein when the number exceeds the threshold value, the pattern represented by the pattern outline is detected as a defect.

Images