JP2005300581A - Mask inspecting device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a mask inspecting device for accurately quantatively detecting strips generated on a photomask. <P>SOLUTION: The mask inspection device comprises an inspection stage 2 for supporting a mask 1 to be inspected with regularly arranged plurality of patterns; a line sensor 4 for reading a pattern image of the mask 1 to be inspected, with a plurality of sensor pixels in Y direction; a stage drive part 3 for moving the inspection stage 2 in X direction; amplifying means (11 and 12) for amplifying to change an amplification rate, in response to the reading position of the sensor pixel every pattern arranging a sensor output signal obtained by the plurality of the sensor pixels in the X direction, when the pattern image with the line sensor 4 is read during moving a stage; means (131 and 132) for taking in the amplified sensor output signal in the X direction, by sampling of a plurality of numbers of times every pattern; means (135 and 136) for detecting stripe-like uneveness, on the basis of an integration value by integrating the fetched sensor output signal for each pattern row along the Y direction. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a mask inspection apparatus for inspecting defects such as a photomask.

  In video devices such as CCD (Charge Coupled Device), LCD (Liquid Crystal Display), etc., one of the items for judging non-defective devices is the result of imaging using these video devices or the screen on which video is output. There are streaks. This streak may occur due to a problem in the circuit of the device, but in many cases, streak-like unevenness is generated on the device during the device manufacturing stage, and this unevenness is streaked in the imaging result or the video output result. It will appear.

  Currently, video devices such as CCDs and LCDs are manufactured by transferring a mask pattern onto a wafer or glass substrate by a lithography technique using an exposure photomask. At that time, a cause of the occurrence of streak-like unevenness on the device is that the streak-like unevenness generated on the photomask (hereinafter referred to as “straight spot”) is directly transferred to the wafer or the glass substrate. Originally, masks with streaks should be excluded as defective products, but there is currently no device for mechanically judging streaks on the mask. For this reason, inspectors in charge of mask inspection rely on sensory inspection to visually determine the presence or absence of streaks while shining light on the mask with a projector.

  However, in the visual sensory inspection, the physical condition of the inspector and other factors, or variation (individual differences) in the detection sensitivity of the stripe unevenness among a plurality of inspectors occurs. For this reason, the pass / fail judgment of the mask may differ depending on the inspector. As a result, a mask with uneven stripes is shipped as a non-defective product, which causes a problem that unevenness occurs on the device.

  On the other hand, for example, the following Patent Document 1 describes an invention relating to a streak inspection method in which a shadow mask or the like used in a color television cathode ray tube is used as a sample to be inspected and streaks are detected using image data of the sample to be inspected. Has been.

Japanese Patent No. 3254288

  However, in the above prior art, the image of the entire shadow mask as a sample to be inspected is captured by a CCD camera, and the captured image data is processed by an image processing apparatus (secondary differential processing or the like) to inspect streaks. Therefore, the detection sensitivity (accuracy) is too low to detect stripes with a very fine pattern such as a pattern formed on a photomask.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a mask inspection apparatus capable of accurately and quantitatively detecting uneven stripes generated on a photomask.

  A mask inspection apparatus according to the present invention includes an inspection stage that supports an inspection mask in which a plurality of patterns are regularly arranged in a two-dimensional direction, and a pattern image of the inspection mask that is supported on the inspection stage. A reading sensor that reads pixels in the main scanning direction; a moving unit that relatively moves the inspection stage and the reading sensor in the sub-scanning direction; and a plurality of patterns when reading a pattern image by the reading sensor during movement by the moving unit. Amplifying means for amplifying the sensor output signal obtained by the sensor pixel by switching the amplification factor according to the reading position of the sensor pixel for each pattern arranged in the sub-scanning direction, and sub-scanning the sensor output signal amplified by this amplifying means Sampling means to capture by sampling multiple times for each pattern in the direction, and sensor output captured by the sampling means No. was integrated for each pattern row along the main scanning direction, in which and a detection means for detecting streaky irregularities on the basis of the integrated value.

  In the mask inspection apparatus according to the present invention, the sensor output signal amplified by the amplifying unit is captured by the sampling unit, and the captured sensor output signal is integrated for each pattern row along the main scanning direction, thereby obtaining the sub scanning direction. When the difference between integrated values between adjacent pattern rows is taken, the portion where pattern rows containing normal patterns are adjacent to each other, and the portion where pattern rows containing normal patterns and pattern rows containing abnormal patterns are adjacent Thus, a difference occurs in the difference between the integrated values. For this reason, it is possible to accurately determine the presence or absence of unevenness by obtaining the difference between the integrated values of the pattern rows adjacent in the sub-scanning direction.

  According to the mask inspection apparatus of the present invention, the sensor output signal amplified by the amplification unit is captured by the sampling unit, and the captured sensor output signal is integrated by the detection unit for each pattern row along the main scanning direction. By detecting the stripe unevenness based on the value, it is possible to accurately and quantitatively detect the stripe unevenness generated in the mask to be inspected.

  Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings.

  First, before describing the embodiment of the present invention, the cause of the occurrence of uneven stripes in a photomask (hereinafter simply referred to as a mask) will be described. In general, when a pattern is formed on a mask (mask blank) in a mask manufacturing process, an exposure drawing apparatus such as an electron beam drawing apparatus is used. The drawing method for drawing a pattern by the exposure drawing apparatus is roughly divided into two methods depending on the moving method of the stage supporting the mask.

  In the first method, as shown in FIG. 1A, the stage is continuously moved in the arrow direction from the start end to the end in the X direction, and when the stage movement position reaches the end in the X direction, the stage is moved in the Y direction. One step is moved, and this is repeated until the end in the Y direction. In this method, unevenness may occur in the boundary portion P1 when the stage is moved stepwise in the Y direction due to position variation or dimensional variation when the pattern is drawn.

  In the second method, as shown in FIG. 1B, the stage is moved step by step from the start end to the end in the X direction, and when the moving position of the stage reaches the end in the X direction, the stage is moved 1 in the Y direction. Step movement is performed and this is repeated until the end in the Y direction. In this method, due to position fluctuations and dimensional fluctuations when drawing a pattern, stripes occur at the boundary part P2 when the stage is moved stepwise in the X direction and at the boundary part P3 when stepped in the Y direction. There is.

  2 (A) to 2 (C) show examples of occurrence of uneven stripes on the mask. First, when it is assumed that no stripes are generated on a photomask to be inspected, a plurality of patterns having the same pattern size, position, and shape are formed on the mask in a two-dimensional direction (orthogonal biaxial direction; XY). (Direction) is regularly and repeatedly arranged. Further, each pattern is an opening pattern (etching pattern) by etching or the like, and the pattern shape is a substantially quadrangle (rectangle) in which corner portions (four corners) are rounded into an arc shape.

  In such a case, in the pattern arrangement of FIG. 2A, a step is generated in the pattern of the column indicated by the arrow, and the position of the pattern on the left side and the position of the right pattern in the figure are in the Y direction with the pattern having the step formed as a boundary. It's off. In the pattern arrangement of FIG. 2B, the dimension (horizontal width) of the pattern indicated by the arrow is smaller than the dimensions of the other patterns. In the pattern arrangement of FIG. 2C, the position of the pattern in the column indicated by the arrow is shifted in the X direction (right side in the figure) as compared with the other patterns. Such pattern step, pattern dimension variation, and pattern position variation all appear as Y-direction unevenness on the mask. In addition, when the pattern step, pattern dimension variation, and pattern position variation directionality is reversed by 90 °, stripes occur in the X direction on the mask.

  FIG. 3 is a schematic view showing the configuration of the mask inspection apparatus according to the embodiment of the present invention. In the figure, an inspection mask 1 is a photomask used for exposure of a photoresist in a lithography process when manufacturing an optical device such as a CCD or LCD. As described above, a plurality of patterns are repeatedly arranged regularly (equally spaced) in the two-dimensional direction on the inspection mask 1. Each pattern is for forming pixels for imaging in the CCD device and pixels for display in the LCD device.

  The inspection stage 2 supports the inspection mask 1 in a mounted state. For example, a vacuum suction hole (not shown) is provided on the mask mounting surface of the inspection stage 2 in order to support the inspection mask 1 in a fixed state. The inspection stage 2 is supported by a stage driving unit 3 so as to be movable in two orthogonal axes, that is, in the X direction and the Y direction. The stage drive unit 3 includes, for example, a first movement table that is movable in the X direction, a second movement table that is movable in the Y direction, and a first drive unit that moves the first movement table in the X direction. And a second driving unit that moves the second moving table in the Y direction. The moving table of the inspection stage 2 and the stage driving unit 3 may be directly connected using screws or the like, or may be connected with an intermediate member interposed therebetween.

  A one-dimensional line sensor (reading sensor) 4 is provided above the inspection stage 2. The line sensor 4 is configured using, for example, a CCD line sensor. In the longitudinal direction of the line sensor 4, a plurality of sensor pixels are arranged in a line (straight line) at a constant pitch. Each sensor pixel photoelectrically converts light incident on the sensor pixel, and thereby, signal charges accumulated in the pixel portion are sequentially transferred in the line direction (pixel column arrangement direction) and output. The line sensor 4 reads the pattern image of the mask 1 to be inspected supported by the inspection stage 2 with a plurality of sensor pixels in the main scanning direction. The line sensor 4 is arranged in parallel with the Y direction in a state of facing the inspection mask 1 supported on the inspection stage 2. Therefore, the Y direction corresponds to the “main scanning direction” when the line sensor 4 reads the pattern image, and the X direction perpendicular to the Y direction corresponds to the “sub scanning direction”.

  Further, when the inspection stage 2 is moved in the X direction by the stage driving unit 3, the inspection stage 2 and the line sensor 4 are relatively moved in the sub-scanning direction. The inspection stage 2 corresponds to a “moving body”. However, instead of the inspection stage 2, the line sensor 4 may be a “moving body”, and the “moving means” may be configured by a linear drive system that moves the line sensor 4 in the X direction (sub-scanning direction). .

  Further, as shown in FIG. 4, a light source lamp 5, a half mirror 6, and a lens 7 are disposed above the inspection stage 2. The light source lamp 5 emits line-shaped light toward the half mirror 6. The light source lamp 6 is constituted by, for example, an elongated rod-shaped halogen lamp, and is arranged in a direction parallel to the line sensor 4 (direction along the Y direction). The half mirror 6 reflects the line-shaped light emitted from the light source lamp 5 toward the mask surface of the mask 1 to be inspected on the inspection stage 2 and transmits the light reflected by the mask surface toward the lens 7. It is what The lens 7 enlarges the image of the light transmitted through the half mirror 6 in the longitudinal direction (Y direction) of the line sensor 4.

  In FIG. 4, the light reflected by the mask surface of the mask 1 to be inspected is incident on the line sensor 4. Although not shown, the mask 1 to be inspected is not shown. For example, the line sensor 4 is disposed below the inspection stage 2 so that the opening is formed in the supporting inspection stage 2 and the line sensor 4 and the light source lamp 5 are opposed to each other in the vertical direction through the opening. By arranging the light source lamp 5 above the inspection stage 2, it is also possible to adopt a configuration in which light irradiated to the mask 1 to be inspected by the light source lamp 5 passes through the mask pattern and this transmitted light enters the line sensor 4. It is.

  Returning to FIG. 3, two detectors 8 and 9 for detecting the position of the inspection stage 2 in the X direction and the Y direction are provided outside the inspection stage 2. Each detector 8, 9 is constituted by a laser interferometer, for example. One detector 8 is disposed in a state facing the end surface of the inspection stage 2 in the X direction, and the other detector 9 is disposed in a state facing the end surface of the inspection stage 2 in the Y direction. The detectors 8 and 9 are electrically connected to the stage position monitoring unit 10. The stage position monitoring unit 10 monitors the position of the inspection stage 2 in the X direction based on the detection signal output from the detector 8, and the Y direction of the inspection stage 2 based on the detection signal output from the detector 9. The position of is monitored. The stage position monitoring unit 10 outputs position information (hereinafter, stage position information) of the inspection stage 2 detected according to the detection signals output from the detectors 8 and 9.

  The amplifier 11 amplifies the sensor output signal output from the line sensor 4. The amplification factor control unit 12 switches the amplification factor when the sensor output signal output from the line sensor 4 is amplified by the amplifier 11 based on the position information of the inspection stage 2 output from the stage position monitoring unit 10. is there. Switching of the amplification factor by the amplification factor controller 12 is performed while the inspection stage 2 is being moved by the stage driving unit 3. That is, when reading the pattern image of the mask 1 to be inspected by the line sensor 4 while moving the inspection stage 2 in the sub-scanning direction (X direction), the amplifier 11 outputs each sensor output signal obtained by the plurality of sensor pixels. The amplification factor controller 12 switches the amplification factor of the sensor output signal by the amplifier 11 in accordance with the reading position of the sensor pixel in the two-dimensional direction (X, Y direction) for each pattern arranged in the sub-scanning direction.

  The image processing unit 13 captures the sensor output signal amplified by the amplifier 11 by sampling, and detects the unevenness of the mask 1 to be inspected using the captured sensor output signal. The image processing unit 13 includes a sampling execution unit 131 that captures the sensor output signal, a sampling control unit 132 that controls the sampling timing of the sensor output signal by the sampling execution unit 131, and the sensor output signal captured by the sampling execution unit 131. A storage unit 133 to store, a calculation unit 134 that performs a predetermined calculation process (described later) using the sensor output signal stored in the storage unit 133, and a determination to determine the presence or absence of streaks based on the calculation result of the calculation unit 134 Unit 135 and an output unit 136 that outputs a determination result by the determination unit 135. Among these, the sampling execution part 131 and the sampling control part 132 comprise the "sampling means" in this invention. Further, the calculation unit 134 and the determination unit 135 are configured to detect streak using the sensor output signal sampled by the sampling execution unit 131 and stored in the storage unit 133, and constitute the “detection means” in the present invention. Is. The display unit 14 displays the detection result of the stripe unevenness by the image processing unit 13 as visible information on a monitor screen or the like.

  Subsequently, a mask inspection method using the mask inspection apparatus according to the embodiment of the present invention will be described.

  First, the inspection mask 1 is set on the inspection stage 2. At this time, on the inspection stage 2, the boundaries P1, P2, and P3 (see FIG. 1) of the mask 1 to be inspected, which may be a spot where unevenness occurs, are parallel or perpendicular to the stage moving direction during mask inspection. The mask 1 to be inspected is aligned with high accuracy so as to be arranged in the above.

  Next, the pattern image of the mask 1 to be inspected is continuously read by the line sensor 4 while moving the inspection stage 2 in one direction at a constant speed with the light source lamp 5 turned on. In this case, since the line sensor 4 is disposed along the Y direction, the stage driving unit 3 moves the inspection stage 2 in the X direction orthogonal to the Y direction. Thereby, the pattern image of the mask 1 to be inspected is read and scanned in the Y direction (main scanning direction) by the line sensor 4, and in parallel with this, the pattern image is moved in the X direction (sub scanning direction) by the movement of the inspection stage 2. Read and scan.

  Here, the resolution when the pattern image of the mask 1 to be inspected is read by the line sensor 4 will be described. Now, it is assumed that a pattern as shown in FIG. 5 is repeatedly arranged in the X direction at a constant pitch (P1 = P2 = P3 = P4) on the mask 1 to be inspected, and this pattern arrangement is also repeatedly developed in the Y direction. . In such a case, the resolution in the X direction is set according to the moving speed of the inspection stage 2 in the X direction and the sampling frequency when the sensor output signal output from the amplifier 11 is sampled by the image processing unit 13. Therefore, by appropriately setting these parameters, the resolution in the X direction can be obtained by synchronizing the sampling frequency of the image processing unit 13 with the pattern arrangement period (pitch) in the X direction. , P4), sampling is performed at least N times (N is an integer of 2 or more). Incidentally, in the example shown in the figure, the sampling is set to be performed five times within each pitch. Therefore, in the X direction, the sensor output signal is captured by sampling five times for each pattern.

  On the other hand, the resolution in the Y direction is set according to the arrangement pitch of the sensor pixels provided in the line sensor 4 and the magnification of the lens 7. Therefore, by appropriately setting these parameters, the resolution in the Y direction is set so that each pattern is read by at least two (a plurality) sensor pixels. Incidentally, in the case of the illustrated example, the pattern image of the mask to be inspected is set so as to read four sensor pixels S1, S2, S3, S4 arranged in the Y direction as one set (unit). The reading cycle Ly by the four sensor pixels S1, S2, S3, and S4 is set to be substantially the same as the pattern arrangement pitch in the Y direction. However, regarding the Y direction, it is not necessary to accurately synchronize the period Ly with the pattern arrangement pitch. Therefore, depending on the arrangement relationship between the sensor pixel and the pattern in the Y direction, two patterns may partially intervene at the part read by the four sensor pixels belonging to the common set.

  By setting the resolution of the line sensor 4 as described above, in FIG. 5, each pattern included in the pattern image of the mask 1 to be inspected has 5 pixels in the X direction, which is the sub-scanning direction, and the main scanning direction. Are sequentially read in the size (window) of 4 pixels in the Y direction, and a sensor output signal corresponding to the amount of light received by the sensor pixel is obtained for each pixel.

  Further, the sensor output signal of the line sensor 4 is taken into the amplifier 11 in units of one pixel and is amplified and output by the amplifier 11. At that time, the amplifier 11 amplifies the sensor output signal corresponding to each pixel position in accordance with the gain variable signal given from the gain control unit 12 with respect to the sensor output signal for the (5 × 4) pixels. Change the rate step by step. Specifically, the amplification factor is changed under the conditions shown in FIG. 6 according to the reading position by the sensor pixel, for example. That is, in the X direction, the amplification factors are set to increase in order of 1 ×, 2 ×, 3 ×, 4 ×, and 5 × according to the reading order of the sensor pixels, and in the Y direction, the sensor pixels S1, S2, According to the arrangement order of S3 and S4, the amplification factor is set to increase in order of 1, 2, 3, 4, and 4 in order. As a result, at each pixel position (reading position) of (5 × 4) pixels, a value obtained by multiplying the set value of the amplification factor in the X direction and the set value of the amplification factor in the Y direction is an effective amplification factor. Multiplied by

  For example, if the pixel located in the upper left corner of the (5 × 4) pixel is the first pixel in the X direction and the Y direction, both the magnification values in the X direction and the Y direction are set to “1” at this pixel position. Therefore, the value of the effective amplification factor multiplied by the sensor output signal is also “1”. In contrast, at the pixel position of the second pixel in the X direction and the third pixel in the Y direction, the magnification value in the X direction is set to “2” and the magnification value in the Y direction is set to “3”. The value of the effective amplification factor multiplied by the sensor output signal is “6”. Further, at the pixel position of the fifth pixel in the X direction and the second pixel in the Y direction, the magnification value in the X direction is set to “5” and the magnification value in the Y direction is set to “2”. The value of the effective amplification factor multiplied by is “10”.

  The correspondence relationship between the sensor pixel reading position and the amplification factor (correspondence relationship between the pixel position and the magnification value) is not limited to the condition shown in FIG. 6, and the sensor pixel reading position (pixel position) is in the X direction and It only needs to be set under the condition that the amplification factor gradually changes (increases or decreases) as it is displaced in the Y direction. Accordingly, when the sensor output signal is processed using (5 × 4) pixels as one unit in the same manner as described above, the pixel position (reading position by the sensor pixel) under any of the conditions shown in FIGS. The amplification factor may be changed for each time, or the amplification factor may be changed by appropriately combining these conditions. Moreover, you may make it change an amplification factor using arbitrary functions.

  By the way, in the condition shown in FIG. 7A, by changing (increasing) the magnification value in the X direction and the magnification value in the Y direction equally by “1” with respect to the pixel position of the upper right corner, The amplification factor at each pixel position is set so as to change stepwise as it moves away from the reference pixel position in the X and Y directions. In the condition shown in FIG. 7B, the magnification value in the X direction and the magnification value in the Y direction are each changed (increased) equally by “1” based on the pixel position of the lower right corner. The gain at each pixel position is set to change stepwise as it moves away from the reference pixel position in the X direction and Y direction. Under the conditions shown in FIG. Further, by changing (increasing) the magnification value in the X direction and the magnification value in the Y direction equally by “1”, the amplification factor at each pixel position is changed from the reference pixel position to the X direction and the Y direction. It is set to change step by step as it leaves.

  On the other hand, under the condition shown in FIG. 7D, the magnification value in the X direction is uniformly changed (increased) by “1” with reference to the pixel position of the upper left corner, and the magnification value in the Y direction is set to “5”. The gain at each pixel position is set to change stepwise as it moves away from the reference pixel position in the X and Y directions. Further, under the condition shown in FIG. 7E, the magnification value in the X direction is uniformly changed (increased) by “power of 2” with reference to the pixel position of the upper left corner, and the magnification value in the Y direction is changed. By changing (increasing) evenly by “1”, the amplification factor at each pixel position is set to change stepwise as it moves away from the reference pixel position in the X and Y directions. ), The magnification value in the X direction is uniformly changed (increased) by “1” with reference to the pixel position of the upper left corner, and the magnification value in the Y direction is uniformly “power of 2”. By changing (increasing), the amplification factor at each pixel position is set to change stepwise as it moves away from the reference pixel position in the X and Y directions.

  Switching of the amplification factor by the amplification factor controller 12 is performed according to the setting conditions shown in FIG. 6, for example. That is, in the Y direction which is the main scanning direction, the amplification factor set according to the reading position by each sensor pixel is applied to each of the four sensor pixels grouped in advance. Further, in the X direction, which is the sub-scanning direction, the reading position in the X direction is grasped according to the stage position information output from the stage position monitoring unit 10, and the amplification factor set according to this reading position is applied. As a result, the sensor output signals (20 sensor output signals in total) corresponding to (5 × 4) pixels read by the four sensor pixels in each set are set to the amplification factors (effective) set according to the respective reading positions. Is amplified by the amplifier 11 according to the amplification factor.

  As a result, when the pattern image of the mask 1 to be inspected is read by the line sensor 4, the sensor output signal for (5 × 4) pixels including the normal pattern shown in FIG. 8A is shown in FIG. If the sensor output signal is obtained with such a value, the value of the sensor output signal after the sensor output signal is taken into the amplifier 11 and amplified under the above set condition is as shown in FIG.

  The sensor output signal output from the amplifier 11 is captured by the sampling execution unit 13 of the image processing unit 13. At that time, in the image processing unit 13, the sampling control unit 132 controls the sampling timing of the sensor output signal by the sampling execution unit 131 based on the stage position information output from the stage position monitoring unit 10. At that time, the sampling control unit 132 grasps the position of the inspection stage 2 moving in the X direction in real time based on the stage position information output from the stage position monitoring unit 10 and determines the movement origin position in the X direction. Each time the inspection stage 2 moves by a distance of 1 / N (N = 5 in the example of FIG. 5) of the pattern arrangement pitch Px (= P1 = P2 = P3 = P4) as a reference, the sensor output to the sampling execution unit 131 Instructs signal capture. At this time, the sensor output signal sampled by the sampling execution unit 131 is sequentially stored (accumulated) in the storage unit 133 until the reading of the pattern image in the X direction is completed.

  Now, using the four sensor pixels belonging to the same group among the plural (many) sensor pixels provided in the line sensor 4, the normal pattern shown in FIG. 8A and the step shown in FIG. When a certain pattern is read, the sensor output signal for (5 × 4) pixels output from the line sensor 4 as a normal pattern reading result has a value as shown in FIG. Assume that the sensor output signal for (5 × 4) pixels output from the line sensor 4 as a pattern reading result has a value as shown in FIG. In this case, the value obtained by integrating the sensor output signals for (5 × 4) pixels is the same value (= 25.5) when a normal pattern is read and when a stepped pattern is read. On the other hand, when the sensor output signal is amplified under the setting conditions shown in FIG. 6, the sensor output signal (after amplification) when a normal pattern is read has a value as shown in FIG. The sensor output signal (after amplification) when a stepped pattern is read has a value as shown in FIG. In this case, the value obtained by integrating the sensor output signals (after amplification) for (5 × 4) pixels is different depending on whether a normal pattern is read or a stepped pattern is read. Specifically, the integrated value in the former case is 185, and the integrated value in the latter case is 201.

  Such a difference in the integrated value similarly occurs in the reading result by another set of sensor pixels in which four sensor pixels are set in the main scanning direction (Y direction). For example, if there is a step in the pattern of the mask 1 to be inspected in the X direction, sensor output signals (after amplification) for (5 × 4) pixels are integrated for each pattern as shown in FIG. The time value is different at the boundary between the stepped pattern portion and the normal pattern portion adjacent thereto. Therefore, the calculation unit 134 of the image processing unit 13 uses the sensor output signal (after amplification) captured by the sampling execution unit 131 and stored in the storage unit 133, and the sensor output signal is aligned along the main scanning direction (Y direction). Accumulate for each pattern row. Each pattern row is divided into groups of 5 pixels according to the number of times of sampling in the sub-scanning direction (X direction).

  At that time, the calculation unit 134 may integrate the sensor output signals (after amplification) for (5 × 4) pixels for each pattern, and then integrate this integrated value for each pattern row along the Y direction. Sensor output signals (after amplification) for (5 × 4) pixels may be integrated together for each pattern row along the Y direction. Regardless of which calculation method is employed, the integrated value obtained as the calculation result is the same value. However, in order to reduce variations in sensor output signals due to sensitivity differences between sensor pixels and random variations in pattern dimensions on the mask, it is preferable to integrate sensor output values for each pattern. In the example of FIG. 10, the integrated value of the sensor output signal corresponding to the leftmost pattern row is 518, the integrated value of the sensor output signal corresponding to the second pattern row from the left is 518, and the third value from the left. The integrated value of the sensor output signal corresponding to the pattern row of 559 is, the integrated value of the sensor output signal corresponding to the fourth pattern row from the left is 585, and the integrated value of the sensor output signal corresponding to the rightmost pattern row is 585.

  The calculation result of the integrated value calculated in this way by the calculation unit 134 is given to the determination unit 135. Then, the determination unit 135 determines the presence or absence of streaks based on the calculation result of the integrated value sent from the calculation unit 134. More specifically, the determination unit 135 first obtains a difference between integrated values between pattern rows adjacent in the sub-scanning direction (X direction). In the example of FIG. 10, the integrated value of the sensor output signal corresponding to the leftmost pattern row is 518, and the integrated value of the sensor output signal corresponding to the second pattern row from the left adjacent thereto is also 518. Therefore, these differences are “0”. Also, since the integrated value of the sensor output signal corresponding to the second pattern column from the left is 518 and the integrated value of the sensor output signal corresponding to the third pattern column from the left adjacent thereto is 559, these values are The difference is “41”. Further, since the integrated value of the sensor output signal corresponding to the third pattern row from the left is 559, and the integrated value of the sensor output signal corresponding to the fourth pattern row from the left adjacent thereto is 585, these values are The difference is “26”. Further, the integrated value of the sensor output signal corresponding to the fourth pattern row from the left is 585, and the integrated value of the sensor output signal corresponding to the rightmost pattern row adjacent thereto is also 585. Becomes “0”.

  Thus, by obtaining the difference of the integrated values between the adjacent pattern rows in the sub-scanning direction, an abnormal pattern having a step exists in the pattern image of the mask 1 to be inspected, and this abnormal pattern is in the Y direction. In the situation arranged side by side, the difference between the integrated value in the pattern sequence including the pattern in which the step is generated and the integrated value of the pattern sequence adjacent to the pattern sequence and including the normal pattern is adjacent to each other and It becomes larger than the difference between the integrated values of the pattern strings including the normal pattern.

  Also, even if the abnormal pattern is not due to a step, by performing the same process as described above, the difference between the integrated value in the pattern sequence including the abnormal pattern and the integrated value of the pattern sequence adjacent to the normal pattern and including the normal pattern However, the difference is larger than the difference between the integrated values of the pattern strings adjacent to each other and including the normal pattern. For example, as shown in FIG. 11 (A), sensor output signals for (5 × 4) pixels including a pattern whose pattern size is deviated from the normal size are obtained as shown in FIG. 11 (B). When the value of the sensor output signal after the sensor output signal is amplified by the amplifier 11 becomes a value as shown in FIG. 11C, as shown in FIG. 12, the sensor output signal for (5 × 4) pixels. The value when (after amplification) is integrated for each pattern is different at the boundary between the pattern portion where the dimensional variation has occurred and the normal pattern portion adjacent thereto. Therefore, the difference (= 34) between the integrated value (= 484) in the pattern sequence including the pattern in which the dimensional variation has occurred and the integrated value (= 518) in the pattern sequence adjacent to the pattern sequence and including the normal pattern. Is larger than the difference (= 0) of the integrated values of the pattern rows adjacent to each other and including the normal pattern.

  Further, as shown in FIG. 13A, sensor output signals for (5 × 4) pixels including a pattern in which the pattern position is shifted from the normal position are obtained as shown in FIG. 13B. When the value of the sensor output signal after the sensor output signal is amplified by the amplifier 11 becomes a value as shown in FIG. 13C, as shown in FIG. 14, the sensor output signal for (5 × 4) pixels. The value when (after amplification) is integrated for each pattern is different at the boundary between the pattern portion where the position variation has occurred and the normal pattern portion adjacent thereto. Therefore, the difference (= 17) between the integrated value (= 535) in the pattern sequence including the pattern in which the position variation has occurred and the integrated value (= 518) in the pattern sequence adjacent to the pattern sequence and including the normal pattern. Is larger than the difference (= 0) of the integrated values of the pattern rows adjacent to each other and including the normal pattern.

  Therefore, the determination unit 135 sequentially calculates the difference between the integrated values between the adjacent pattern rows in the sub-scanning direction, and compares each difference with a predetermined value set in advance. If the difference between the integrated values is greater than or equal to a predetermined value, it is determined that the inspection mask 1 has uneven stripes. If the difference between the integrated values is less than the predetermined value, it is determined that there is no uneven stripes. The detection result of the stripe unevenness by the determination unit 135 is output from the output unit 136 and displayed as visible information on the display unit 14. The contents of the inspection result displayed on the display unit 14 may be contents for notifying the presence or absence of unevenness, or contents for notifying the quality of the mask 1 to be inspected.

  As described above, by performing the mask inspection using the mask inspection apparatus of the present embodiment, it is possible to accurately and quantitatively detect the stripe unevenness generated in the mask 1 to be inspected. Further, in the pattern image of the mask 1 to be inspected, the part of the pattern row in which the difference between the integrated values is equal to or greater than a predetermined value can be specified as the occurrence of the uneven stripe. Moreover, it is good also as what displays the generation | occurrence | production location of the uneven stripe specified in this way on the display part 14 as a test result.

  In addition, regarding the predetermined value that is a criterion for determining the presence or absence of stripe unevenness, the stripe unevenness detection accuracy required when inspecting the mask 1 to be inspected and the level of stripe unevenness allowed in the manufacturing process of a video system device (CCD, LCD, etc.) In consideration of the above, it may be experimentally set in advance. Further, the difference between the integrated values of the pattern rows adjacent in the sub-scanning direction may be calculated by the calculation unit 134, and the determination unit 135 may determine the presence or absence of unevenness based on the calculation result.

  Further, in the apparatus configuration of the above-described embodiment, the inspection stage 2 on which the inspection mask 1 is set is moved in the X direction with respect to the line sensor 4 arranged along the Y direction as described above, and the inspection mask. 1 is read by the line sensor 4 to detect a stripe unevenness along the Y direction orthogonal to the moving direction (X direction) of the inspection stage 2. Due to the drawing method shown in FIG. 1B, stripes may occur both in the vertical direction (position P2) and in the horizontal direction (position P3) of the mask.

  In order to cope with such a case, after performing the mask inspection as described above, the inspection target mask 1 set on the inspection stage 2 is rotated by 90 °, and the same mask inspection is performed again. Can be detected. The inspection mask 1 may be rotated by rotating the mask itself or the inspection stage 2. In addition to the line sensor 4 arranged along the Y direction, a line sensor (not shown) arranged along the X direction is provided, and the mask inspection is performed using the Y direction line sensor 4 in the same manner as described above. Thereafter, the mask stage is detected by moving the inspection stage 2 on which the mask 1 to be inspected 1 is set in the Y direction with respect to the line sensor in the X direction, whereby both vertical and horizontal stripes can be detected.

  Further, a variable mechanism for changing the direction of the line sensor 4 is provided, and the line sensor 4 is arranged along the X direction by changing the direction of the line sensor 4 by 90 ° in one direction and the other direction by this variable mechanism. , And can be arranged along the Y direction. In the mask inspection apparatus provided with this variable mechanism, the inspection stage 2 is moved in the Y direction when the line sensor 4 is disposed along the X direction, and the inspection stage is disposed when the line sensor 4 is disposed along the Y direction. By switching the stage moving direction so as to move 2 in the X direction, it is possible to detect both vertical and horizontal stripes using the common line sensor 4.

  Further, when the line sensor is arranged along at least one of the X direction and the Y direction, for example, when the line sensor 4 is arranged along the Y direction as shown in FIG. 3, the line sensor 4 is arranged in the X direction ( By disposing a plurality at a predetermined interval in the sub-scanning direction), it is possible to shorten the time required for mask inspection. As a specific example, when two line sensors are arranged parallel to each other along the Y direction, and the interval between these two line sensors is set to ½ of the mask dimension of the mask 1 to be inspected in the stage moving direction, Compared with the case where the pattern image of the mask 1 to be inspected is read by one line sensor, the amount of movement of the inspection stage 2 in the X direction can be reduced to almost half. Further, by parallel processing the sensor output signals output from the respective line sensors, the signal processing time can be shortened to almost half.

It is a figure which shows an example of the pattern drawing system in a mask manufacturing process. It is a figure which shows the example of generation | occurrence | production of the stripe unevenness on a mask. It is the schematic which shows the structure of the mask inspection apparatus which concerns on embodiment of this invention. It is the schematic which shows the arrangement | positioning state of the optical system of a mask inspection apparatus. It is a figure which shows the example of a setting of the resolution when reading a pattern image by a mask test | inspection. It is a figure which shows an example of the switching conditions of an amplification factor. It is a figure which shows the other example of the switching conditions of an amplification factor. It is a figure which shows the relationship between the sensor output signal before and behind amplification obtained when a normal pattern and this are read. It is a figure which shows the relationship between the sensor output signal before and behind amplification obtained when the pattern in which the level | step difference produced and this was read. It is a figure which shows the relationship between the pattern image containing the pattern row | line | column in which the level | step difference produced, and the integrated value for every pattern row | line | column. It is a figure which shows the relationship between the sensor output signal before and after amplification obtained when the pattern in which the dimension fluctuation | variation produced was read. It is a figure which shows the relationship between the pattern image containing the pattern row | line | column which the dimension fluctuation | variation produced, and the integrated value for every pattern row | line | column. It is a figure which shows the relationship between the sensor output signal before and after amplification obtained when the pattern which the position fluctuation produced, and this is read. It is a figure which shows the relationship between the pattern image containing the pattern row | line | column which the position fluctuation | variation produced, and the integrated value for every pattern row | line | column.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Mask to be inspected, 2 ... Inspection stage, 3 ... Stage drive part, 4 ... Line sensor, 10 ... Stage position monitoring part, 11 ... Amplifier, 12 ... Amplification rate control part, 13 ... Image processing part, 131 ... Execution of sampling , 132 ... Sampling control unit, 133 ... Storage unit, 134 ... Calculation unit, 135 ... Determination unit

Claims (6)

An inspection stage for supporting an inspection mask in which a plurality of patterns are regularly arranged in a two-dimensional direction;
A reading sensor that reads a pattern image of the inspection target mask supported by the inspection stage in a main scanning direction with a plurality of sensor pixels;
Moving means for relatively moving the inspection stage and the reading sensor in a sub-scanning direction;
When the pattern image is read by the reading sensor during movement by the moving unit, sensor output signals obtained by the plurality of sensor pixels are set according to the reading position of the sensor pixel for each pattern arranged in the sub-scanning direction. Amplifying means for switching and amplifying the amplification factor;
Sampling means for capturing the sensor output signal amplified by the amplifying means by sampling a plurality of times for each pattern in the sub-scanning direction;
A mask inspection apparatus comprising: a detection unit that integrates the sensor output signal captured by the sampling unit for each pattern row along the main scanning direction, and detects streak-like unevenness based on the integration value. . The detection means detects the streaky unevenness by obtaining a difference of the integrated values between pattern rows adjacent in the sub-scanning direction and comparing the difference with a predetermined value set in advance. The mask inspection apparatus according to claim 1. The amplifying unit moves the inspection stage and the reading sensor relative to each other in the sub-scanning direction by the moving unit using either the inspection stage or the reading sensor as a moving body. And a gain control means for switching the gain when the sensor output signal is amplified based on the detection result of the position detection means during movement by the movement means. The mask inspection apparatus according to claim 1. The mask inspection apparatus according to claim 1, further comprising a variable unit that changes a direction of the reading sensor. The mask inspection apparatus according to claim 1, wherein a plurality of the reading sensors are arranged at regular intervals in the sub-scanning direction. The mask inspection apparatus according to claim 2, wherein the detection unit determines that streaky unevenness exists when the difference is equal to or greater than a predetermined value.

Images