JPH0417361B2 - - Google Patents

Description

[Detailed Description of the Invention] (a) Technical Field of the Invention The present invention relates to an automatic inspection method for photomasks,
In particular, large-sized chips, such as memory devices, in which multiple repeating patterns occupy a large area, are arranged in multiple chip areas on one substrate.
The present invention relates to a method for automatically inspecting a reticle (multi-reticle) pattern.

(b) Background of the technology A photomask, which is composed of a large number of chip area patterns of the same type arranged on one substrate, produces an image of the chip area pattern formed on the reticle at a size 5 to 10 times the actual size. It is formed by reducing and projecting onto a mask substrate to the actual size using a step-and-repeat method.

Conventionally, this reticle had a structure in which one chip area pattern was formed on one reticle substrate, but recently, a gigantic reticle has a structure in which multiple chip area patterns of the same type are formed in alignment on one reticle substrate. The production efficiency of photomasks has been improved by reducing the number of step exposures.

Similar to a normal reticle, when a defect occurs in such a gigantic reticle, defects occur regularly in the chip area pattern formed on the photomask, and its quality is extremely degraded. Therefore, pattern inspection of the Guyang reticle is extremely important.

(c) Prior Art and Problems When automatically inspecting the Gu Young reticle pattern, a comparative inspection method similar to that commonly used for photomasks is used. This comparative inspection method scans the mask substrate with a pair of imaging means that simultaneously capture the same corresponding locations of two adjacent chip area patterns in the field of view, and compares the video signals obtained from each imaging means. This is a method of detecting the presence or absence of a difference in chip area patterns.

FIG. 1 is a block diagram schematically showing an example of the main functional configuration of an automatic mask inspection device used for inspecting the above-mentioned Guyang reticle. In the figure, 1 indicates the Guyang reticle to be inspected, C ₁ , C ₂ is a chip area pattern arranged on the gigantic reticle; 2 is an X-Y stage (used for scanning the field of view) that is slidable in the X-Y directions; 3a and 3b are a pair of illumination devices; 4a , 4b is a pair of microscopes, 5
a and 5b are a pair of photoelectric conversion elements that convert the optical images of the respective microscopes into electrical signals, and 6a and 6b are a pair of signals 2 that binarize the respective video electrical signals.
digitization units, 7a and 7b are a pair of screen memory forming units that form screen memories based on their respective binary video signals; 8a and 8b are a pair of screen memory recognition units that confirm the respective screen memory information; 9 is a pair of screen memory recognition units, respectively; 10 is a memory pattern shape comparison section that compares the screen memory information sent from the screen memory recognition section of , a defect signal output section 11 outputs a signal when two pieces of screen memory information do not match in the comparison section; 12 is a defect coordinate memory into which the coordinate position is written when the two screen memory information do not match in the comparing section; 13 is a stage drive section; and 13 is a stage drive section. .

Conventionally, the Guyang reticle is inspected by the automatic mask inspection device as described above.
The upper part shows chip area patterns C ₁ , C ₂ , C ₃ , C ₄ , C ₅ , C ₆ and C ₇ , C ₈ , which are arranged on the reticle.
Since the optical axis spacing d is equal to the pitch spacing of C ₉ ,
The field of view A and B of the pair of microscopes are scanned over the entire surface with the X direction as the main scanning direction, and the corresponding two adjacent chip area patterns captured in the field of view A and B at each sliced time point are The optical images of the same location were converted into electrical signals, binarized, and compared, and if no difference was detected, the Guyang reticle was determined to be good. (The broken line in the figure represents a scanning line, and m ₁ is an arrow indicating the scanning direction.) However, in the above conventional method, there is a problem that commonly occurs when drawing a chip area pattern on the gigantic reticle ( It is impossible to detect abnormal defects of the same type that commonly appear in each chip area pattern due to defects in pattern information, defects in writing equipment, etc. Therefore, gigang reticles with these types of abnormal defects are considered non-defective. It was determined that this was the case and was used in the manufacture of photomasks, resulting in a problem of deteriorating the quality of the photomasks.

(d) Purpose of the Invention The present invention has been made in view of the above-mentioned problems, and is particularly applicable to the inspection of gigang reticles in which the chip area pattern is mainly occupied by repetitive patterns, such as gigang reticles used in the manufacture of semiconductor memory devices. The objective is to provide an automatic mask inspection method that improves accuracy.

(e) Structure of the Invention In other words, the present invention provides a method for automatically inspecting a photomask in which a plurality of chip area patterns of the same type including repeating patterns are arranged in a row, in the same corresponding location of two chip area patterns. A first step in which the entire chip area pattern is sequentially scanned by a pair of imaging means that simultaneously bring the chip area into the field of view, and the image signals obtained by each imaging means are compared to detect a difference in the shapes of the two chip area patterns.
For the photomask in which no difference in pattern shape was detected in the inspection process and the first inspection process, the same corresponding locations of two repeating patterns in one chip area pattern are simultaneously brought into view. A pair of imaging means sequentially scans the repeating pattern area within the chip area pattern, and the video signals obtained by each imaging means are compared to identify any abnormalities in the shape of the one chip area pattern. and a second inspection step for detecting.

(f) Examples of the invention The method of the present invention will be described below with reference to the drawings.

FIG. 3 is a schematic top view showing an example of a Guyang reticle to which the mask inspection method of the present invention is applied.
FIGS. 4A to 4B are schematic plan views showing the steps of an embodiment of the method of the present invention.

The mask inspection method of the present invention has a large repeating pattern area in which many cell patterns (unit patterns) of the same type are repeatedly arranged, such as a gigantic reticle used in manufacturing semiconductor memory devices. This method is applied to a gigantic reticle in which a plurality of chip area patterns occupying a large area are aligned and arranged on one mask substrate (glass substrate). Fig. 3 is a plan view schematically showing an example of this, and G in the figure is a gigantic reticle substrate;
C ₁ to _{C 9} are chip area patterns, R ₁ to _{R 9} are repeating pattern areas, and P ₁ to _{P 9} are areas in which non-repetitive element patterns such as peripheral circuits are arranged.

The automatic mask inspection apparatus used in the method of the present invention is a comparative inspection apparatus similar to the conventional one as shown in FIG.

In the method of the present invention, when inspecting a gigang reticle as shown in FIG. 3 above, as shown in FIG. A pair of microscopes is adjusted to have an optical axis interval d ₁ equal to the arrangement pitch interval in the X direction of the chip area patterns C ₁ to C ₉ arranged, and the same corresponding locations of the adjacent area patterns are included in the field of view. Using fields of view A and B, scan the entire surface of the Guyang reticle G with the X direction as the main scanning direction (m ₁ is an arrow in the reticle surface scanning direction), and scan each sliced time point. Converts images of the corresponding same locations of two adjacent chip area patterns captured in fields of view A and B into electrical signals, converts the image electrical signals into binarized values, writes them into screen memory, and simultaneously compares them. If a difference is detected, its coordinate position is recorded in the defect coordinate memory. A Guyang reticle whose entire surface has been scanned and no mismatched location is detected is determined to be a good product for the time being, and a reticle where a mismatched location is detected is determined to be defective. The above is the first test.

Next, in the method of the present invention, the Guyang reticle determined to be good in the first test is subjected to a second test as described below. That is, one chip area pattern of the Guyang reticle,
For example, regarding C ₁ , as _shown in FIG _.
The entire surface is placed in the fields of view D and E of a pair of microscopes, which have an optical axis distance d ₂ that is an integral multiple of the arrangement pitch of the repeating pattern _U , and which, for example, cover the same corresponding locations of different repeating patterns U D and U _E. Therefore, for example, by using the main scanning method in the direction perpendicular to the array axis of the microscope (the Y direction in the figure) as indicated by the arrow _m2 , the field of view D is scanned at each sliced time point. The video signals at the same locations corresponding to the two repeating patterns captured within E and E are binarized and sequentially compared, and
If no difference is detected between the unit patterns (repetitive patterns) over the entire area _R1 , the repeated patterns included in all chip area patterns _C1 to _C9 arranged on the gigantic reticle G are It is determined that there is no abnormal defect in the regions _R1 to _R9 . This determination is possible because the reason for abnormal pattern defects in gigantic reticles is mainly due to defects in information and defects in the drawing device during pattern drawing, and therefore, the same type of abnormal defects are caused by the respective chip area patterns. By having characteristics that occur in common.

In the method of the present invention, it is confirmed by the first inspection that there is no difference between the respective chip area patterns arranged on the gigantic reticle, and further, by the second inspection, the chip area patterns are After confirming that there are no abnormal defects in the repeating pattern area that occupies most of the area in the area pattern,
Determine the reticle as good.

Note that in the second inspection step, the main scanning direction of the microscope is not limited to the direction of the above embodiment, but is determined in a direction in which the scanning time can be shortened depending on the shape of the repeated pattern region.

(g) Effects of the Invention As explained above, the automatic mask inspection method of the present invention can be used in the manufacture of semiconductor memory devices, etc. in which the area where the repeated pattern is provided represents most of the chip area. The quality of the gigantic reticle is determined by taking into account the comparison inspection results between repeated patterns and the comparison inspection results between chip area patterns.

Therefore, according to the present invention, the inspection accuracy of the Guyang reticle is improved, and the manufacturing yield of semiconductor memory devices and the like is improved.

[Brief explanation of drawings]

Figure 1 is a block diagram schematically showing an example of the main functional configuration of an automatic mask inspection device, Figure 2 is a process plan diagram schematically showing an automatic mask inspection method, and Figure 3 is an example of a gigantic reticle. FIGS. 4A to 4B are schematic plan views showing one embodiment of the automatic inspection method for a gigantic reticle according to the present invention. In the figure, G is a gigantic reticle, C ₁ to _{C 9} are chip area patterns, R ₁ to R ₉ are repeating pattern areas, A, B, D, and E are fields of view of a pair of microscopes, respectively, and U, U _D , U _E is a unit repeating pattern,
d ₁ and d ₂ are the distances between the optical axes of the microscope, and m ₁ and m ₂ are the scanning direction arrows.

Claims (1)

[Claims] 1. When automatically inspecting a photomask in which a plurality of chip area patterns of the same type including repeating patterns are arranged in an array, the same corresponding locations of two chip area patterns are simultaneously viewed. A first step that sequentially scans all the chip area patterns using a pair of image pickup means, and compares the video signals obtained by each image pickup means to detect a difference in the shapes of the two chip area patterns.
and the photomask in which no difference in pattern shape was detected in the first inspection step,
A pair of imaging means that simultaneously view the same corresponding locations of two repeating patterns within one chip area pattern sequentially scan the repeating pattern arrangement area within the chip area pattern, and a second inspection step of comparing video signals obtained by the imaging means to detect an abnormality in the shape of the one chip region pattern.