JPH07128841A - Method for inspecting defect of photomask - Google Patents

Abstract

PURPOSE:To fully automatically execute defect inspection of a photomask by forming alignment marks indicating the conditions for forming mask patterns along the mask patterns simultaneously with the formation of the mask patterns. CONSTITUTION:The alignment marks indicating the conditions for forming the mask patterns are formed on a glass substrate simultaneously with the formation of the mask patterns along the (x) direction and (y) direction which are the arranging directions of the mask patterns. The regions of the transparent substrate where the alignment marks are formed are optically scanned and the kinds and forming positions of the respective alignment marks are successively detected at the time of placing the photomask 1 formed with the mask patterns and the alignment marks on an x-y stage 11 and inspecting the defects of the respective mask patterns. The conditions for the defect inspection are then determined in accordance with the kinds and forming conditions of the inspected alignment marks and the defects of the photomask 1 are inspected in accordance with the determined conditions for the defect inspection.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a photomask defect inspection method for inspecting a mask pattern formed on a transparent substrate for defects.

[0002]

2. Description of the Related Art Photolithography is used in the manufacturing process of semiconductor devices and lead frames, and various photomasks are used in this photolithography. Usually, a photomask has a plurality of identical mask patterns formed on a transparent substrate. Since the presence or absence of defects in these mask patterns has a significant influence on the manufactured product, strict defect inspection is performed for each manufactured photomask.

According to the currently used photomask defect inspection method, when a photomask is manufactured, that is, when a mask pattern is formed on a glass substrate, the mask pattern formation conditions are stored in a magnetic disk or the like to perform defect inspection. At this time, the manufacturing condition stored in the magnetic disk is read out, and the operator manually inputs the read-out data into the inspection device, and then the defect inspection is performed.

[0004]

In order to perform a highly accurate defect inspection in the defect inspection of the photomask, it is necessary to accurately position the substrate on which the mask pattern is formed in the reference coordinate system on the inspection device side. . However, in the method of performing the defect inspection by reading the mask pattern formation conditions stored in the magnetic disk or the like, the reference coordinate system on the inspection device side and the substrate do not correspond accurately, and an error may occur in the defect inspection. is there. For example, when a glass substrate is mounted on the inspection device, the substrate moves in parallel or rotationally moves with respect to the coordinate axis of the inspection device and does not exactly match, resulting in a positioning error and inaccurate defect inspection. Will end up.

Further, if the manufacturing conditions of the mask pattern are stored in the magnetic disk and manually input into the inspection device at the time of defect inspection, there is a drawback that the data input work is complicated and takes a long time.

Therefore, the object of the present invention is to eliminate the above-mentioned defects, to automatically input the mask pattern manufacturing conditions to the inspection device, to perform the defect inspection work fully automatically, and to detect the glass substrate and the inspection device. It is an object of the present invention to provide a method of inspecting a defect of a photomask, which can accurately correspond to the coordinate system of.

[0007]

A method for inspecting a photomask defect according to the present invention is used for inspecting a photomask having a mask pattern formed on a transparent substrate for defects and forming a mask pattern on the transparent substrate. Alignment marks representing the mask pattern forming conditions are formed in a region of the transparent substrate where the mask pattern is not formed along the x direction and the y direction orthogonal thereto, and when inspecting defects of each mask pattern, Optically scans the area where the alignment mark is formed on the transparent substrate to sequentially detect the type and formation position of each alignment mark, and find and determine the defect inspection condition based on the detected type and position of the alignment mark. The defect of the photomask is inspected according to the defect inspection condition.

[0008]

According to the present invention, when the mask pattern is formed on the glass substrate, the alignment mark representing the condition for forming the mask pattern is simultaneously formed. As the alignment mark, for example, a chip center mark indicating the center of the mask pattern of one unit, a stay mark indicating the stay width,
It is possible to use a deletion area mark or the like that indicates an area that is not inspected. Then, prior to the defect inspection, the alignment mark is optically scanned, and the type and formation position of each alignment mark is read to obtain the mask pattern formation condition. Next, a defect inspection is performed based on the obtained forming conditions.

With this structure, the defect inspection of the mask pattern can be carried out fully automatically, and the accuracy of the defect inspection is further improved. Furthermore, it is possible to automatically and automatically detect deleted areas or stay marks on which individual patterns such as control numbers are formed, and exclude these areas from the inspection target. The inconvenience of inputting can be eliminated.

[0010]

1 is a diagrammatic plan view showing an example of a photomask which is inspected for defects by a defect inspection method according to the present invention. The photomask 1 has eight mask patterns (chips) 3 to 10 of the same shape formed on a glass substrate 2, and these mask patterns are arranged in two rows in the x direction and four rows in the y direction. A space portion is formed between the rows of the mask patterns 3 to 6 and the rows of 7 to 10, and this area is a stay width which is a non-inspection area. Each mask pattern 3-10
Are areas 3a to 10a in which the respective management numbers are formed
And these areas are to be deleted areas that are not subject to inspection. In the present invention, these non-inspected areas are automatically detected.

In the present invention, alignment marks representing the mask pattern forming conditions are formed on the glass substrate 2 at the same time when the mask pattern is formed along the x direction and the y direction which are the arrangement directions of the mask pattern. In this example, as the alignment marks, the chip center marks C _x1 and C _x2 , C _{y1 to} C _y4 that display the center position of each mask pattern in the x direction and the y direction, and S ₁ and S ₂ that display the stay marks, And a deletion area mark D indicating the position and width of the deletion area in the x and y directions.
Three types of alignment marks of _x1 and D _x2 , D _{y1 to} D _y4 are used.

In order to optically detect these alignment marks, mark lines l _x and l _y are respectively formed in the x and y directions, and the alignment marks are displayed along these mark lines. The tip center mark is displayed as a thick and short line which is orthogonal to the mark lines l _x and l _y . In addition, the stay marks S ₁ and S ₂ are marked lines l
_The line extends only toward the bottom of _x . further,
The deletion area mark is displayed thickly above each mark line and displayed by a length corresponding to the length. With this configuration, the type and formation position of each alignment mark can be accurately read by optically scanning with a certain width along the mark lines in the x direction and the y direction.

FIG. 2 is a diagram showing the construction of an example of the alignment mark reading apparatus according to the present invention. The photomask 10 on which the mask pattern and the alignment mark are formed is placed on the xy stage 11. It is assumed that the alignment mark is formed in the y direction (in-plane direction) and the x direction (direction orthogonal to the paper plane) orthogonal thereto. Illumination light is projected from the illumination light source 12 toward the photomask 10 via the condenser lens 13, and the image of the alignment mark is passed through the projection lens 14 and the half mirror 15 to the first and second linear image sensors 16 and 17. To project each. The first linear image sensor 16 has the respective light receiving elements arranged along the y direction and is used for scanning in the x direction. Further, in the second linear image sensor 17, the respective light receiving elements are arranged in the x direction and used for the y direction scanning. The imaging length of these image sensors is, for example, 12 m on a photomask.
The alignment mark having a width of 6 mm can be imaged by setting m. The first and second image sensors 16 and 17 are sequentially read by the clock signal supplied from the synchronization signal generating circuit 18 via the switch 19, and the video signal thereof is read by the switch 20 and the A / D converter 2.
1 to the central processing unit 22 and the memory 23. The central processing unit 22 determines the type of each detected alignment mark, and instructs the memory 23 to store the type of the detected alignment mark and its address. The switches 19 and 20 serve to switch between x-direction scanning and y-direction scanning.

The clock signal from the synchronizing signal generating circuit 18 is also supplied to the drive control circuit 24, and the drive control circuit generates x-direction and y-direction stage feed pulses for moving the xy stage 11 in the x-direction and the y-direction. Occur.
The x-direction stage feed pulse and the y-direction stage feed pulse are supplied to the x-direction drive circuit 25 and the y-direction drive circuit 26, respectively, and the xy stage 11 is moved in the x-direction and the y-direction to perform x-direction scanning and y-direction scanning. . The xy stage 11 has an x-axis origin sensor 27 and a y-axis origin sensor 2
8 are provided, and the xy stage 11 moves in the x-axis direction and y
The arrival of the origin in the axial direction is detected, and these detection signals are supplied to the memory 23 via the switch 29, so that the memory 2
By resetting the address of 3, it is possible to match the address of the memory with the coordinates of the photomask. These origin sensors can be composed of photoelectric sensors, for example. The clock signal from the synchronizing signal generation circuit 18 is also supplied to the memory 23 and the central processing unit 22, and the clock signal is used to supply the first and second image sensors 16 and 1 to each other.
The video signal supplied from 7 is stored in the memory 23. The x-direction stage feed pulse and the y-direction stage feed pulse generated from the drive control circuit 24 are also supplied to the memory 23 and used to increment the memory address. With this configuration, the photomask 1
Since 0 is scanned along the x direction and the y direction over a width of 6 mm, the type of alignment mark formed on the mark line and the formation position thereof can be accurately detected.

Next, the x-direction and y-direction scanning will be described.
To do. Alignment marks and mark lines are in Photomass
The distance from the edge of the glass substrate
Ku. In addition, the scanning start position in the x direction and the y direction, the scanning length and
And the scanning speed are predetermined. Formed on photomask 1
The management number of the photomask (12 in FIG. 1)
(Indicated by 345) and then in x and y directions.
Read each alignment mark. First, xy
The stage 11 is moved in the x direction for sub-scanning, and the first scan
Main scanning is performed by reading out the near image sensor 16,
Perform two-dimensional scanning of the photomask 10 over a width of 6 mm.
U At this time, each switch 19, 20, 29 scans in the x direction.
Switch to. As shown in Figure 1,
Pcenter Mark C_X1Are detected and are deleted one after another
Ku D_X1, Stay mark S₁And S ₂, Chip Centama
C_x2, Deletion area D_x2Is read and the seed of each mark
The class and address are stored in the memory 23. The origin sensor
The address of the memory 23 is reset by the detection signal from the server 27.
set. Similarly, in the y-direction scanning, as shown in FIG.
So that the tip center mark C_y1, Delete area mark
D_y1, Chip center mark C_y2, Delete area mark D
_y2, Chip center mark C_y3, Delete area mark
D_y3, Chip center mark C_y4, And delete area mer
Ku D_y4Are sequentially read.

Next, the condition setting for the defect inspection by reading the alignment mark will be described. As shown in FIG. 3, when the photomask is placed on the xy stage, the upper left edge of the photomask 1 is determined as the origin, and the reference mark R serving as the scanning reference is determined by the photosensor. Are defined as R _x and R _y , and the absolute position is obtained using the coordinates of the reference mark R. Further, it is assumed that scanning in the x and y directions is performed along the direction indicated by the arrow. First, the reading of the alignment mark in the x direction will be described. (I) Stay width The stay width SW (x) is determined from the addresses of the stay marks S ₁ and S _{2 according} to the following equation.

[Equation 1] SW (x) = S _x2 −S _X1 (ii) Start point of inspection The start point that determines the start position of the inspection is determined by the coordinates of the upper left edge of the mask pattern 3 in FIG. The starting point of this inspection is the scanning start position of the defect inspection. Since this inspection start point is located on the cutting line of each chip and cannot be formed directly on the substrate, it is judged from the detected alignment mark. The x coordinate of this starting point is determined based on the following equation.

[Number 2] _{ST (x) = C X1 -} {(C x2 -C X1) - (S 2 -S 1)} Chiippu number CN _x Chiippu number at _{× 1/2 + R x (} iii) x direction was detected Determined from the number of chip center marks. (Iv) Chip pitch The chip pitch CS (x) in the x direction is determined based on the following equation.

## EQU00003 ## CS (x) = _{C.sub.x2- C.sub.X1} (v) Deletion Width As the deletion width in the x direction, the value of the deletion area mark _D.sub.xn can be used as it is.

Next, the setting of the inspection condition in the y direction will be described. (I) Stay width Since no stay mark exists in the y direction, the stay width becomes zero, and it is determined that each chip exists continuously in the y direction. (Ii) Starting point of inspection

## EQU4 ## ST (y) = R _y − {C _y4 + (C _y4 −C _y3 ) / 2} (iii) Chip pitch The chip pitch CS (y) is calculated based on the following equation.

## _EQU00005 ## CS (y) = ( _{Cyn- Cy1} ) / ( _Cyn- 1) = ( _{Cy4- Cy1} ) / 3 Other items can be detected in the same manner as the alignment mark in the x direction. it can.

Next, the rotation measurement will be described. As shown in FIG. 4, when forming a mask pattern on a glass substrate, the arrangement direction of the mask pattern may be formed at an angle to the edge of the glass substrate. When the arrangement direction of the mask pattern is inclined with respect to the edge of the substrate,
The alignment mark formed at the same time as the mask pattern is inclined to the edge of the glass substrate. However, if the mask pattern and the alignment mark are tilted, the mask pattern and the alignment mark are also tilted with respect to the coordinate system of the example of the inspection apparatus, and accurate defect inspection cannot be performed. Therefore, the tilt angles of the mark lines l _x and l _y of the alignment mark with respect to the edge of the substrate, that is, the tilt angles with respect to the xy coordinate system of the inspection apparatus are obtained. In addition, the arrangement direction of the mask pattern is
Since it coincides with the mark line of the alignment mark formed at the same time as the mask pattern, the inclination angle of the mark line is obtained, and the xy stage is rotationally corrected based on the obtained angle, whereby the arrangement direction of the mask pattern is determined by the inspection apparatus. Can be exactly matched to the coordinate system.

The rotation θ is obtained as follows. If the xy coordinate values detected by the two chip center marks are (x ₁ , y ₁ ), (x ₂ , y ₂ ),

[Equation 6] In this example, the coordinates are calculated using the coordinates of the two chip center marks, but it is also preferable to calculate using the coordinates of a large number of alignment marks by the least squares method.

Next, a process of performing a defect inspection based on the read mask pattern formation data will be described. FIG. 5 is a diagram showing a configuration of an example of the defect inspection apparatus according to the present invention. In this example, two inspection heads are used to optically scan two adjacent mask patterns,
A comparative inspection method for detecting defects while comparing the results is used. FIG. 6 shows a photomask 10 with two inspection heads.
3 is a diagrammatic plan view showing a state of optical scanning of FIG. The defect inspection conditions are set prior to the defect inspection. First, if the mask pattern is rotating with respect to the coordinate system of the inspection apparatus, rotation correction is performed. This rotation correction is performed by rotating the xy stage 11 by an angle determined so as to match the coordinate system of the inspection device.

Next, the two inspection heads are respectively positioned at the inspection start points S ₁ and S ₂ , and then a defect is detected by optical scanning. These start points S ₁ and S ₂ are the upper left edge of the adjacent mask pattern. Figure 5
Let's go back and explain. The photomask 10 is placed on the xy stage 11 which moves intermittently at a predetermined pitch along the x direction. The illumination light from the illumination light source 31 is divided into two optical fiber bundles 31a and 31b and a condenser lens 32, respectively.
The light is projected toward the photomask 10 as first and second illumination light via a and 32b. These first and second
The distance in the y-direction between the exit ends of the optical fibers is set equal to the chip pitch, and they are located at the same coordinates in the x-direction.
Illumination light from the first and second optical fibers 31a and 31b is transmitted through the photomask 10 and is emitted into the first and second optical fibers, respectively.
The light enters the first and second linear image sensors 34a and 34b via the projection lenses 33a and 33b, respectively. First and second linear image sensors 34a and 3
In 4b, a plurality of light receiving elements are arranged along the y direction and are sequentially driven at a predetermined reading frequency. In this example, the driving scan of these linear image sensors is the main scan, and the sub-scan is performed by moving the xy stage 11 in the x direction. The video signals read from the first and second linear image sensors 34a and 34b are supplied to the video monitors 36a and 36b after being amplified by the amplifiers 35a and 35b, respectively, and are also supplied to the differential circuit 37. Therefore, the first
The image picked up by the inspection head is displayed on the first video monitor 36a, and the image picked up by the second inspection head is displayed on the second video monitor 36b. The differential circuit 37 includes the first linear image sensor 34.
video signal from a and the second linear image sensor 34b
The difference signal is generated by comparing with the video signal from. This difference signal is supplied to the first and second comparison circuits 38a and 38b, respectively. These comparison circuits perform a function of detecting whether or not the input difference signal exceeds a predetermined limit value, and the first comparison circuit 38a determines whether or not the difference signal exceeds a positive upper limit value and determines whether or not the second comparison circuit. 38b determines whether the negative lower limit value is exceeded. The comparison results from the comparison circuits 38a and 38b are supplied to the central processing unit 39.

When there are no defects in the two mask patterns adjacent to each other, the linear image sensors 35a and 35b are used.
Since the video outputs from the same take the same value, the output signal from the differential circuit 37 becomes zero. On the other hand, when either one of the mask patterns has a defect, the output from the differential circuit 37 becomes a positive or negative signal, and the central processing unit 39 determines that it is a defect. The central processing unit 39 supplies the defect determination result together with its address to the defect monitor 40,
The output of the differential circuit 37 is also supplied to the defect monitor 40. According to this structure, the image of the defect is displayed on the defect monitor, so that the operator can accurately recognize the existence and the form of the defect.

The present invention is not limited to the above-described embodiments, and various modifications and changes can be made. For example, in the above-described embodiment, an example of inspecting a photomask for a defect in which a plurality of mask patterns are formed on one transparent substrate along the x and y directions has been described. The present invention can be applied to a case where a plurality of mask patterns are formed, and can also be applied to a defect inspection of a photomask in which one mask pattern is formed on a substrate.

[0024]

As described above, according to the present invention, the alignment mark indicating the condition for forming the mask pattern is formed along the mask pattern at the same time when the mask pattern is formed. It can be done automatically. In addition, since the coordinate system when forming the mask pattern and the coordinate system on the defect inspection device side exactly match, the accuracy of the defect inspection can be further improved.

[Brief description of drawings]

FIG. 1 is a plan view showing the configuration of an example of a photomask used in a defect inspection method according to the present invention.

FIG. 2 is a diagram showing a configuration of an example of an alignment mark reading device according to the present invention.

FIG. 3 is a diagram showing a reading scan of an alignment mark.

FIG. 4 is a diagram showing a state in which a mask pattern is displaced from a glass substrate.

FIG. 5 is a diagram showing a configuration of an example of a defect inspection apparatus used in the present invention.

FIG. 6 is a diagram showing a state of optical scanning in the defect inspection apparatus.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Photomask 3-10 Mask pattern 11 xy stage 12 Illumination light source 14 Projection lens 15 Half mirror 16,17 Linear image sensor 18 Synchronous signal generation circuit 19,20,29 Switch 22 Central processing unit 23 Memory 24 Drive control circuit 25 x direction Drive circuit 26 y-direction drive circuit

Claims (2)

[Claims] 1. When inspecting a photomask having a mask pattern formed on a transparent substrate for defects, an alignment mark representing the conditions for forming the mask pattern is formed on the transparent substrate when the mask pattern is formed on the transparent substrate. The mask pattern is formed in the area not formed along the x direction and the y direction orthogonal thereto, and when inspecting the defect of each mask pattern, the area where the alignment mark of the transparent substrate is formed is optically formed. The feature is that the type and formation position of each alignment mark is sequentially detected by scanning, the defect inspection condition is obtained based on the detected alignment mark type and formation position, and the defect of the photomask is inspected according to the obtained defect inspection condition. Photomask defect inspection method. 2. When inspecting a defect of a photomask in which a plurality of mask patterns are regularly formed on a transparent substrate, when forming a mask pattern, an alignment mark indicating a condition for forming the mask pattern, It is formed along the x direction which is the arrangement direction of the mask pattern and the y direction which is orthogonal thereto, and when inspecting the defect of each mask pattern, the region of the transparent substrate where the alignment mark is formed is optically scanned. The defect inspection condition is determined based on the detected alignment mark type and formation position, the defect inspection condition is determined based on the detected alignment mark type and formation position, and the photomask defect is inspected based on the determined defect inspection condition. Characteristic photomask defect inspection method.