JP2008064553A - Pattern inspection device and pattern inspecting method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a pattern inspection device enhanced in S/N ratio, and a pattern inspection method. <P>SOLUTION: When the contact pattern 102 of a mask 101 is inspected, the mask 101 is photographed to acquire a censor image and a reference image is formed on the basis of the plan data of the mask 101. Next, the difference image of the sensor image and the reference image is formed and a light quantity measuring area 104 is set in the region of a part of the contact pattern 102 in the difference image to evaluate the brightness of the light quantity measuring area 104. If a flaw 103 is present in the light quantity measuring area 104, the brightness of the light quantity measuring area 104 lowers as compared with a case where no flaw 103 is not present. By this constitution, the presence of a flaw 103 is judged. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a pattern inspection apparatus and a pattern inspection method, and in particular, whether or not there is a defect by comparing a sensor image obtained by imaging a pattern to be inspected with a reference image generated based on design data. The present invention relates to a pattern inspection apparatus and a pattern inspection method.

  Conventionally, photolithography has been frequently used in the manufacture of semiconductor devices, and in these photolithography, a mask for transferring a circuit pattern onto a semiconductor substrate is used. As such a mask, for example, a mask in which a light shielding film processed into a predetermined pattern is formed on a glass substrate is used. As a method for inspecting whether or not such a mask is manufactured as designed, an optical image (hereinafter referred to as “sensor image”) obtained by imaging the mask with an image sensor is used as mask design data. There is a method for inspecting for the presence or absence of defects as compared with a pseudo optical image (hereinafter referred to as “reference image”) created based on the above.

  However, in recent years, with the miniaturization of semiconductor devices, the mask has also been miniaturized, and the size thereof has become smaller than the wavelength of the laser beam used for inspection, so that it has become difficult to detect defects. In particular, a rectangular opening pattern (hereinafter referred to as “contact pattern”) formed in a mask for forming a contact hole of a semiconductor device is one of the finest patterns among patterns formed in a photomask. Detecting defects is particularly difficult.

  Therefore, a method for inspecting whether or not a defect exists in the contact pattern by comparing the total amount of light passing through the contact pattern between the sensor image and the reference image has been developed (for example, Patent Documents). 1). However, since this method uses a continuous light source as the light source, the S / N ratio (Signal per Noise ratio) is low, and the defect detection sensitivity is insufficient.

JP 2002-162729 A

  An object of the present invention is to provide a pattern inspection apparatus and a pattern inspection method having a high S / N ratio.

  According to an aspect of the present invention, a sensor unit that captures an object to be inspected to obtain a sensor image, a reference image generation unit that creates a difference image between the sensor image and a reference image, and the object to be detected in the difference image. There is provided a pattern inspection apparatus comprising: a determination unit that determines the presence or absence of a defect based on the brightness of a part of an area corresponding to an inspection target pattern of an inspection object.

  According to another aspect of the present invention, the step of capturing a test object to obtain a sensor image, the step of creating a difference image between the sensor image and a reference image, and the test object in the difference image And a step of determining the presence / absence of a defect based on the brightness of a part of the region corresponding to the inspection target pattern.

  According to the present invention, a pattern inspection apparatus and a pattern inspection method having a high S / N ratio can be obtained.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1A is a diagram illustrating a pattern inspection method according to this embodiment, and FIG. 1B is a diagram illustrating a pattern inspection method according to a comparative example.
As shown in FIGS. 1A and 1B, in the present embodiment and the comparative example, a mask 101 for manufacturing a semiconductor device is illustrated as an object to be inspected, and a pattern to be inspected includes a semiconductor device. The contact pattern 102 for forming a contact hole is illustrated. The contact pattern 102 is a square opening. A plurality of contact patterns 102 are formed on the mask 101. It is assumed that a light-shielding defect 103 exists at the edge portion of a certain contact pattern 102.

  As shown in FIG. 1A, in this embodiment, when inspecting the contact pattern 102, a light amount measurement area 104 is set in a partial region of the contact pattern 102, and the light amount is determined for each light amount measurement area 104. Measure. Specifically, the mask 101 is imaged to acquire a sensor image, a difference image between the sensor image and the reference image is created, and the brightness of each light quantity measurement area 104 is evaluated in the difference image. At this time, if the defect 103 is present in the light quantity measurement area 104, the brightness of the light quantity measurement area 104 is lower than when the defect 103 is not present. The reference image is an image of a non-defective mask 101, for example, an image generated based on the design data of the mask 101.

  The light quantity measurement area 104 is set at a plurality of positions in each contact pattern 102 as necessary. For example, the light quantity measurement areas 104 are set for some or all of the four edge portions, the four corner portions, and the central portion of the contact pattern 102. Based on the evaluation result of each light quantity measurement area 104, the presence or absence of the defect 103 in the contact pattern 102 is determined. For example, when the presence of a defect is recognized in one or more light quantity measurement areas 104 in a certain contact pattern 102, it is determined that the defect 103 exists in this contact pattern 102.

  In contrast, as shown in FIG. 1B, in the comparative example, the brightness is evaluated for the entire region of the contact pattern 102 at a time. That is, the light quantity measurement area 105 is set so as to include the entire one contact pattern 102. Based on the brightness of the light quantity measurement area 105, the presence / absence of the defect 103 is determined.

  In the present embodiment, the S / N ratio of the output signal is high because the light quantity measurement area to be evaluated is locally provided in a part of the contact pattern 102 as compared with the comparative example. For this reason, the minute defect 103 can be detected with high accuracy. In addition, since the difference image between the sensor image and the reference image is created and the brightness is evaluated for the difference image, the sensor image and the reference image are each evaluated, and then compared with the method of comparing the evaluation results. Thus, the number of evaluations can be reduced, and the inspection efficiency can be improved.

  In the present embodiment, the light quantity measurement area 104 is set to include a plurality of pixels of the difference image, and a weight kernel is applied to the luminance levels of the plurality of pixels included in the light quantity measurement area 104 to calculate the evaluation value. And based on this evaluation value, you may determine the presence or absence of a defect. Thereby, the influence of the defect 103 on the brightness of the light quantity measurement area 104 can be emphasized, and an inspection with a higher S / N ratio can be performed.

  In the present embodiment, the position of the contact pattern 102 on the mask 101 may be specified by measuring the brightness of a plurality of regions in the reference image or sensor image. Thereby, the position of the light quantity measurement area 104 can be set automatically and accurately.

  Next, a specific example for realizing the present embodiment will be described. The pattern inspection apparatus according to this example is, for example, a photomask inspection apparatus for manufacturing a semiconductor device, and in particular, a mask inspection apparatus that inspects a contact pattern.

FIG. 2 is a block diagram illustrating a mask inspection apparatus according to this example.
FIG. 3 is a flowchart illustrating the mask inspection method according to this example.
FIG. 4A is a diagram illustrating a contact pattern which is an inspection target pattern of the mask inspection apparatus according to this specific example, and FIG. 4B illustrates a pattern represented by the contact pattern design data shown in FIG. (C) is a figure which illustrates the sensor image obtained by imaging the contact pattern shown to (a), (d) is the reference created from the design data shown to (b) It is a figure which illustrates an image, (e) is a figure which illustrates the difference image of the sensor image shown in (c), and the reference image shown in (d).

FIG. 5A is a diagram illustrating a position specifying template for positioning a contact pattern, and FIG. 5B is a view illustrating the position specifying template shown in FIG. 4A to the reference image shown in FIG. It is a figure which illustrates the case where it is made to match.
FIG. 6A is a diagram showing a measurement template that defines a light quantity measurement area, and FIG. 6B is a diagram in the case where the measurement template shown in FIG. 4A is matched with the difference image shown in FIG. (C) is a diagram illustrating a weight kernel.
4A to 4E, FIG. 5B, and FIG. 6B, the luminance level of each pixel of the image is shown in monochrome gradation, and the brighter pixels are displayed in white and darker. The pixels are displayed in black.

  As shown in FIG. 2, the mask inspection apparatus 1 according to this example uses a photomask 11 for manufacturing a semiconductor device (hereinafter simply referred to as “mask”) as an object to be inspected. A large number of contact patterns for forming contacts on the semiconductor device are formed on the mask 11 as patterns to be inspected, and the mask inspection apparatus 1 inspects the presence or absence of defects in the contact patterns.

  As shown in FIG. 4A, the mask 11 has a square opening as the contact pattern 22. In this specific example, for example, it is assumed that a defect 23 that does not transmit light is generated inside an opening (contact pattern 22) through which light is transmitted.

  On the other hand, it is assumed that the design data of the mask 11 is stored in the database 12 (see FIG. 2). The portion corresponding to the contact pattern 22 in this design data represents a contact pattern 22a having no defect, as shown in FIG.

As shown in FIG. 2, the mask inspection apparatus 1 includes a sensor unit 2, a reference image generation unit 3, a difference image creation unit 4, a position specification unit 5, and a determination unit 6.
The sensor unit 2 is a device that captures a mask 11 that is an object to be inspected and acquires a sensor image. For example, the sensor unit 2 is provided with a moving table (not shown), a deep ultraviolet laser light source (not shown), an optical system (not shown), and a TDI (Time Delay and Integration) sensor (not shown). ing. The sensor unit 2 irradiates the mask 11 placed on the moving table with deep ultraviolet laser light and forms an image of the transmitted light on the TDI sensor via the optical system, while the moving table places the mask 11 on the TDI sensor. The sensor image is acquired by scanning the entire area of the mask 11 by moving the relative position of the mask 11. The TDI sensor is a sensor in which several one-dimensionally arranged rows of CCDs (Charge-Coupled Devices) are arranged in the moving direction of an object to be inspected. In the TDI sensor, the charge data of the elements in each column is transferred to the column in synchronism with the movement of the inspected object, so that an integrated value of the data can be obtained with respect to the moving direction, and the high resolution and low S / N are obtained. Ratio can be achieved.

The reference image generation unit 3 is a circuit that generates a reference image based on the design data of the mask 11 stored in the external database 12.
The difference image creation unit 4 is a circuit that receives a sensor image from the sensor unit 2 and a reference image from the reference image generation unit 3 and creates a difference image between the sensor image and the reference image.

  The position specifying unit 5 is a circuit that receives the reference image from the reference image generating unit 3 and specifies the position of the contact pattern 22 in the mask 11 by measuring the brightness of a plurality of regions in the reference image. For example, the position specifying unit 5 stores a position specifying template 24 as shown in FIG. The position specifying template 24 is provided with one cross-shaped opening 24a and a plurality of, for example, 16 openings 24b arranged in a square ring around the opening 24a. The positional relationship between the opening 24a and the opening 24b is such that when the position specifying template 24 is overlaid on the reference image and the opening 24a is positioned at the center of the contact pattern, the opening 24b is located in the peripheral area of the contact pattern. It is a relationship that is located. The opening 24a corresponds to five pixels arranged in a cross shape in the reference image, and each opening 24b corresponds to one pixel in the reference image.

  The determination unit 6 receives the difference image from the difference image creation unit 4, and receives information indicating the position of the contact pattern from the position specifying unit 5. Of the regions corresponding to the contact pattern in the difference image, some of the regions It is a circuit that measures brightness and determines the presence or absence of defects based on the measurement results. For example, the determination unit 6 includes a measurement template 25 (see FIG. 6A) that defines an area (light quantity measurement area) in which the brightness is measured, and a weight kernel 26 ( FIG. 6C is stored.

  As shown in FIG. 6A, in the measurement template 25, when this measurement template 25 is superimposed on a difference image or the like, it is aligned with each edge portion of the contact pattern 22 (see FIG. 4A). A total of four light quantity measurement areas 25a are provided at the positions. Each light quantity measurement area 25a is a cross-shaped opening corresponding to five pixels arranged in a cross shape in the difference image.

  Further, as shown in FIG. 6B, the weight kernel 26 is a kernel corresponding to five pixels arranged in a cross shape, and a “weight” for each pixel is set. Then, the total amount of values obtained by multiplying the luminance level of each pixel by the weight is calculated as a combo. For example, among the five pixels arranged in a cross shape, the weight “w” is set for the pixel 26a located at the crossed portion of the cross shape, and the pixel 26b located in the cross-shaped branch portion. Is assigned a weight of “1”. The weight w is a value larger than 1, for example, and is 4, 16, or 64, for example.

Next, the operation of the pattern inspection apparatus (mask inspection apparatus) according to this example configured as described above, that is, the pattern inspection method (mask inspection method) according to this example will be described.
First, as shown in step S <b> 1 of FIG. 3, the mask 11 is fixed to the moving stage of the sensor unit 2, the deep ultraviolet laser light source irradiates the mask 11 with the deep ultraviolet laser light, and the laser light transmitted through the mask 11. Is imaged on the TDI sensor via the optical system. In this state, by driving the moving table, the mask 11 is moved with respect to the sensor unit 2 to scan the entire area of the mask 11. Thereby, the mask 11 is imaged and a sensor image is acquired. At this time, the sensor image of the contact pattern 22 (see FIG. 4A) is as shown in FIG. The sensor unit 2 outputs this sensor image to the difference image creation unit 4.

  On the other hand, as shown in step S2, the reference image generation unit 3 reads the design data of the mask 11 from the database 12, and processes the design data to generate a reference image as shown in FIG. For example, the reference image generation unit 3 creates a simulated sensor image by simulating the imaging operation by the sensor unit 2 with respect to the defect-free mask represented by the design data. The reference image generation unit 3 outputs this reference image to the difference image creation unit 4 and the position specification unit 5.

  Next, as shown in step S3, the difference image creation unit 4 creates a difference image between the sensor image and the reference image. The difference image is, for example, an image as shown in FIG. 4E, and a portion corresponding to the contact pattern defect 23 (see FIG. 4A) is emphasized to some extent.

  Next, as shown in step S4, the position specifying unit 5 applies the position specifying template 24 shown in FIG. 5A to the reference image shown in FIG. That is, as shown in FIG. 5B, the position specifying template 24 is superimposed on the reference image to search for a position where the openings 24a become bright and all the openings 24b become dark. For example, in the reference image, when the luminance of each pixel is normalized so that black is 20 levels and white is 220 levels, the luminance level of the opening 24a is 100 levels or more, and the luminance level of the opening 24b is Search for a position that is 40 levels or less. Thereby, the position of the opening 24a can be aligned with the center of the contact pattern, and the position of the contact pattern can be specified.

  In this specification, the “brightness” of each pixel in the image is a value proportional to the amount of light received by each sensor element of the sensor unit 2, and the “brightness level” is a standardized “brightness”. Value. The “brightness” of the light quantity measurement area in the difference image refers to an amount reflecting the quantity of light received by the light quantity measurement area when the sensor image is acquired. For example, the average value of luminance, luminance level, and luminance level A value (convolution) calculated by applying a weight kernel to the luminance level.

  Next, as shown in step S5, the determination unit 6 applies the measurement template 25 shown in FIG. 6A to the difference image shown in FIG. That is, as shown in FIG. 6B, the determination unit 6 superimposes the template 25 on the difference image so that the light quantity measurement area 25a is positioned at the position of the contact pattern specified in the difference image in step S4. Thereby, only the luminance level of the pixel corresponding to one light quantity measurement area 25a is extracted from the pixels constituting the difference image. For example, only the luminance levels of five pixels located at one edge portion of the contact pattern and arranged in a cross shape are extracted.

  Then, as shown in FIG. 6C, the weight kernel 26 is applied to the luminance level of the five pixels extracted by the opening 25a to calculate the convolution. That is, among the five pixels arranged in the cross shape, the luminance level of the pixel 26a located at the crossing portion of the cross shape is multiplied by the weight “w”, and the luminance level of the pixel 26b located in the cross-shaped branch portion. Is multiplied by the weight “1”, and the product of the luminance level and the weight of each pixel is added to calculate a combo. The weight w is set to a value larger than 1, for example. If this evaluation value is equal to or greater than a predetermined reference value, it is determined that there is no defect in the edge portion. If the evaluation value is less than the reference value, the edge portion Judge that a defect exists. This determination is performed for each of the four edge portions of the contact pattern. In addition to the edge portion, the above-described determination may be performed for the center portion of the contact pattern and the four corner portions.

  Next, as shown in step S6, an OR operation is performed on the determination result of the presence / absence of defects at each edge portion performed in step S5, and if a defect is recognized at any one or more edge portions, this contact is determined. If it is determined that the pattern has a defect and no defect is found in any edge portion, it is determined that the contact pattern has no defect. In addition, when judging also about the center part and the corner part in addition to the edge part, it is judged that this contact hole has no defect only when no defect is recognized in all parts. In the case, it is determined that a defect has occurred.

  Next, the process proceeds to step S7, where it is determined whether or not the determination has been completed for all contact patterns in the mask 11. If an uninspected contact pattern remains, the process returns to step S4, one is selected from the uninspected contact patterns, position specification (step S4), brightness evaluation (step S5), and determination (step) S6) is performed. On the other hand, when the inspection for all the contact patterns is finished, the inspection of the mask 11 is finished.

  When a plurality of masks having the same shape are to be inspected, the design data of these masks is common, so if the reference image obtained when the first mask is inspected is stored, the second and subsequent sheets are stored. When inspecting this mask, it is not necessary to newly generate a reference image. That is, in the second and subsequent mask inspections, the step shown in step S2 in FIG. 3 can be omitted.

Next, the effect of this example will be described.
In this example, the light quantity of the entire contact pattern is not evaluated, but the brightness of only a part of the contact pattern, for example, one edge portion is locally evaluated. It can be detected with high sensitivity.
Also, by applying a weight kernel to the luminance level of some extracted areas, the difference between the brightness of this area and the brightness of surrounding areas in the difference image can be emphasized, so there is no defect Can be detected with higher sensitivity.
Furthermore, by applying the position specifying template to the reference image, the position of the contact pattern on the mask can be specified accurately and efficiently. Thereby, the precision and efficiency of inspection can be further improved.
Thus, according to this example, a mask inspection apparatus and a mask inspection method having a high S / N ratio can be realized.

  In this example, an example in which a light-shielding defect exists inside the opening has been described. However, the mask inspection apparatus according to this example has a light-transmitting property that exists outside the opening. Such a defect, for example, a defect in which the light shielding film on the glass substrate is lost in the mask 11 can be detected.

Next, a modification of this example will be described.
First, a first modification will be described.
FIG. 7 is a diagram illustrating a measurement template in this modification.
As shown in FIG. 7, in this modification, a measurement template 35 is used instead of the measurement template 25 (see FIG. 6A) in the above-described specific example. In the measurement template 35, light amount measurement areas 35a corresponding to one pixel of the difference image are provided at four positions corresponding to the respective edge portions of the contact pattern. In this modification, based on the luminance level of one pixel extracted from the light quantity measurement area 35a, it is determined whether or not there is a defect at this edge portion.

  In this modification, it is not necessary to provide a weight kernel because the presence / absence of a defect is determined based on the luminance level of one pixel. Configurations, operations, and effects other than those described above in the present modification are the same as in the above-described specific example.

Next, a second modification of this example will be described.
FIG. 8 is a diagram illustrating a measurement template in this modification.
As shown in FIG. 8, in this modification, the measurement template 45 is used instead of the measurement template 25 (see FIG. 6A) in the above-described specific example. In the measurement template 45, rectangular light amount measurement areas 45a are provided at four positions corresponding to the respective edge portions of the contact pattern. The light quantity measurement area 45a corresponds to nine pixels arranged in 3 rows and 3 columns in the difference image. In FIG. 8, only one light amount measurement area 45a is shown for easy viewing of the drawing.

  In the present modification, the average value of the luminance levels of the nine pixels extracted from the light quantity measurement area 45a is used as an evaluation value, and whether or not there is a defect at this edge portion is determined based on this evaluation value. to decide. Configurations, operations, and effects other than those described above in the present modification are the same as in the above-described specific example. The size of the light quantity measurement area 45a is not limited to the size corresponding to the pixels of the 3 rows and 3 columns of the difference image, and other sizes such as 2 rows 2 columns, 2 rows 3 columns, 4 rows 4 columns, and the like. It may be.

Next, a third modification of this example will be described.
In this modification, the measurement template 45 shown in FIG. 8 is used as the measurement template, and the weight kernel is applied to the luminance levels of nine pixels extracted by the measurement template 45. In this weight kernel, the weight multiplied by the luminance level of the pixel located in the central portion of the rectangular light quantity measurement area 45a is larger than the weight multiplied by the luminance level of the pixel located in the peripheral portion of the rectangle.

For example, in this modification, a weight kernel along a Gaussian distribution is applied. That is, the coordinates of the center of a pixel when the center of the light quantity measurement area 45a is the origin is (i, j), the distance from the center of the light quantity measurement area 45a to the center of this pixel is D, and the standard of the luminance level. When the deviation is σ, the circumference is π, and the weights of the pixels are W ₁ (D) and W ₂ (D), the following Equation 1 and Equation 2 are defined, and the parameter P is set. The following formula 3 is defined by simplifying the following formula 2.

  Then, the weight of each pixel is determined according to Equation 3 above. At this time, the parameter P is set to 0, 1, 2, 5, or 10, for example. Configurations, operations, and effects other than those described above in the present modification are the same as in the above-described specific example. Also by this modification, the same effect as the above-mentioned specific example can be acquired.

Next, a comparative example of the present invention will be described.
FIG. 9 is a diagram showing a light amount measurement area in this comparative example.
As shown in FIG. 9, in this comparative example, a light quantity measurement area 61 that includes the entire contact pattern is set for the sensor image and the reference image, and the total light quantity that has passed through the contact pattern to be inspected is measured. Then, the total amount of light is compared between the sensor image and the reference image to determine the presence or absence of a defect. However, this method cannot detect the presence or absence of defects with high sensitivity.

The reason will be described below.
FIG. 10 is a graph showing the luminance profiles in the sensor image and the reference image, with the horizontal axis representing the position in the contact pattern and the vertical axis representing the luminance.
FIG. 11 is a graph showing luminance difference profiles in the sensor image and the reference image, with the horizontal axis representing the position in the contact pattern and the vertical axis representing the luminance difference.
Note that “+ side noise” shown in FIGS. 10 and 11 indicates noise in which the total light amount of the sensor image is larger than the total light amount of the reference image, and “− side noise” refers to the total light amount of the sensor image. The noise is smaller than the total amount of light in the image. Also, the vertical axis (luminance difference) in FIG. 11 indicates a value obtained by subtracting the luminance of the reference image from the luminance of the sensor image, the value of “+ side noise”, and the value of “−side noise”.

  As shown in FIG. 10, even when taking the luminance profile of the measurement result, the difference between the sensor image and the reference image does not appear clearly. Therefore, as shown in FIG. 11, when a luminance difference profile is taken, a negative peak 62 is observed at a position corresponding to a defect in the sensor image. The magnitude of the “+ side noise” and the “− side noise” is about 5%, but the height of the peak 62 is about 15%.

For this reason, as described in the above specific example, if the measured luminance level is locally extracted from a partial region of the contact pattern, the peak 62 caused by the defect can be detected separately from the noise. Is possible. Therefore, a defect can be detected.
On the other hand, as described in the comparative example, in the method for evaluating the total amount of light that has passed through the contact pattern, the luminance variation due to the defect is buried in noise, and it is difficult to detect the defect. .
As a result, according to the specific example of the present invention, the defect can be detected with higher sensitivity than the comparative example. In one example, according to the specific example of the present invention, the S / N ratio can be increased by 1.8 times compared to the comparative example.

  The contents of the present invention have been described above with reference to the embodiment and specific examples and modifications thereof, but the present invention is not limited to these examples. For example, in the above-described specific example, an example in which a mask is inspected using transmitted light has been shown, but reflected light can also be used. In the above-described specific example, an example in which imaging is performed using a TDI sensor has been described. However, the present invention is not limited to this, and another one-dimensional sensor or a two-dimensional sensor may be used. Furthermore, the configurations of the position specifying template, the measurement template, and the weight kernel are not limited to the above-described specific examples and modifications, and the S / N is determined according to the type, shape, size, and the like of the pattern to be inspected. A configuration in which the ratio is as high as possible may be set as appropriate.

  Further, in the specific example described above, an example in which the position specifying template is applied to the reference image to specify the position of the contact pattern is shown, but the position specifying template may be applied to the sensor image. Furthermore, in the mask inspection apparatus of the present invention, the position of the contact pattern can be specified based only on the design data without providing the position specifying unit. Furthermore, in the above-described embodiments and specific examples, an example in which a reference image is generated based on design data has been shown. However, the present invention is not limited to this, and for example, a sensor image of a non-defective mask is used as a reference image. Also good. That is, the present invention can be applied not only to the pattern inspection apparatus of the die-to-database system but also to the pattern inspection apparatus of the die-to-die system and the cell-to-cell system.

  Furthermore, in the above-described embodiments and specific examples, an example in which a contact pattern of a mask for manufacturing a semiconductor device is inspected is shown. However, the inspection object is not limited to this, and for example, other patterns of the mask are inspected. Alternatively, an object other than the mask may be inspected. Furthermore, those in which the person skilled in the art appropriately changed the design and added and deleted the constituent elements of the above-described embodiment, specific examples, modified examples, and examples in which these were combined are also included in the scope of the present invention. And within the scope of the present invention.

(A) is a figure which illustrates the pattern inspection method which concerns on embodiment of this invention, (b) is a figure which illustrates the pattern inspection method which concerns on a comparative example. It is a block diagram which illustrates the mask inspection apparatus which concerns on the specific example of this invention. It is a flowchart figure which illustrates the mask inspection method which concerns on this example. (A) is a figure which illustrates the contact pattern which is a test object pattern of the mask inspection apparatus which concerns on this example, (b) is a figure which illustrates the pattern which the design data of the contact pattern shown to (a) represents (C) is a diagram illustrating a sensor image obtained by imaging the contact pattern shown in (a), and (d) is a reference image created from the design data shown in (b). It is a figure which illustrates, (e) is a figure which illustrates the difference image of the sensor image shown to (c), and the reference image shown to (d). (A) is a figure which illustrates the template for position specification for positioning a contact pattern, (b) matches the template for position specification shown in (a) with the reference image shown in FIG.4 (b). It is a figure which illustrates the case. (A) is a figure which shows the measurement template which prescribes | regulates a light quantity measurement area, (b) illustrates the case where the measurement template shown to (a) is matched with the difference image shown to FIG.4 (e). (C) is a figure which illustrates a weight kernel. It is a figure which illustrates the template for a measurement in the 1st modification of this example. It is a figure which illustrates the template for a measurement in the 2nd modification of this example. It is a figure which shows the light quantity measurement area in a comparative example. It is a graph which shows the brightness | luminance profile in a sensor image and a reference image, taking the position in a contact pattern on a horizontal axis, and taking a brightness | luminance on a vertical axis | shaft. It is a graph which shows the brightness | luminance difference profile in a sensor image and a reference image by taking the position in a contact pattern on a horizontal axis, and taking a brightness | luminance difference on a vertical axis | shaft.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Mask inspection apparatus, 2 sensor part, 3 reference image production | generation part, 4 difference image creation part, 5 position specification part, 6 determination part, 11 mask, 12 database, 22, 22a contact pattern, 23 defect, 24 position specification template 24a, 24b Opening, 25, 35, 45 Measurement template, 25a, 35a, 45a, 61 Light measurement area, 26 Weight kernel, 26a, 26b Pixel, 62 Peak, 101 Mask, 102 Contact pattern, 103 Defect, 104 , 105 Light intensity measurement area

Claims (15)

A sensor unit that captures an image of the inspected object and obtains a sensor image;
A reference image generation unit for creating a difference image between the sensor image and the reference image;
A determination unit that determines the presence or absence of a defect based on the brightness of a partial region of the region corresponding to the inspection target pattern of the inspection object in the difference image;
A pattern inspection apparatus comprising:   The pattern inspection apparatus according to claim 1, further comprising a reference image generation unit configured to generate the reference image based on design data of the object to be inspected. The partial area is an area including one or a plurality of pixels of the difference image,
The pattern inspection apparatus according to claim 1, wherein the determination unit determines the presence / absence of the defect based on an average value of luminance levels of the pixels. The partial area is an area including a plurality of pixels of the difference image,
The pattern inspection apparatus according to claim 1, wherein the determination unit determines the presence or absence of the defect based on an evaluation value calculated by applying a weight kernel to a luminance level of the plurality of pixels. . The partial area is a cross-shaped area;
5. The weight multiplied by the luminance level of the pixel located at the intersection of the cross in the weight kernel is greater than the weight multiplied by the luminance level of the pixel located at the branch of the cross. Pattern inspection equipment. The partial area is a rectangular area;
5. The pattern according to claim 4, wherein a weight multiplied by a luminance level of a pixel located in a central portion of the rectangle in the weight kernel is larger than a weight multiplied by a luminance level of a pixel located in a peripheral portion of the rectangle. Inspection device.   2. The apparatus according to claim 1, further comprising a position specifying unit that specifies a position of the inspection target pattern in the inspection object by measuring brightness of a plurality of regions in the reference image or the sensor image. The pattern inspection apparatus according to any one of 6. Capturing a sensor image by imaging the object to be inspected;
Creating a difference image between the sensor image and a reference image;
Determining the presence or absence of defects based on the brightness of a partial area of the area corresponding to the inspection target pattern of the inspection object in the difference image;
A pattern inspection method comprising:   9. The pattern inspection method according to claim 8, wherein the reference image is generated based on design data of the object to be inspected. The partial area is an area including one or a plurality of pixels of the difference image,
The pattern inspection method according to claim 8 or 9, wherein, in the determining step, the presence or absence of the defect is determined based on an average value of luminance levels of the pixels. The partial area is an area including a plurality of pixels of the difference image,
10. The pattern inspection according to claim 8, wherein in the determining step, the presence or absence of the defect is determined based on an evaluation value calculated by applying a weight kernel to a luminance level of the plurality of pixels. Method. The partial area is a cross-shaped area;
12. The weight multiplied by the luminance level of the pixel located at the crossing portion of the cross shape in the weight kernel is made larger than the weight multiplied by the luminance level of the pixel located at the branch portion of the cross shape. The pattern inspection method described. The partial area is a rectangular area;
The weight multiplied by the luminance level of the pixel located in the central portion of the rectangle in the weight kernel is set larger than the weight multiplied by the luminance level of the pixel located in the peripheral portion of the rectangle. Pattern inspection method.   14. The method according to claim 18, further comprising a step of specifying a position of the inspection target pattern on the object to be inspected by measuring brightness of a plurality of regions in the reference image or the sensor image. The pattern inspection method according to any one of the above. The object to be inspected is a mask for manufacturing a semiconductor device,
The inspection target pattern is a contact pattern for forming a contact hole in the semiconductor device,
The partial region is an edge portion of the contact pattern;
9. In the determining step, if it is determined that at least one edge portion of the edge portions of the contact pattern is defective, it is determined that the contact pattern is defective. The pattern inspection method according to any one of 14.