JP2010286500A - Device and method for inspecting sample - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an efficient inspection by preventing the occurrence of quasi-defects due to a position alignment error. <P>SOLUTION: A sample inspection device includes: a measurement image generating section for generating a measurement image by measuring a sample pattern; a comparator for comparing the measuring image with a reference image and determining a sample which is defective, when the difference between them exceeds a threshold; a position displacement measuring section for measuring a displacement amount between area images with a fixed area that are segmented from the measuring image and the reference image and determining reliability information, indicating the reliability of the position displacement amount between them; and a position aligning section for aligning the positions of the area images, by utilizing the reliability information and the position displacement amount. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an inspection for inspecting a pattern of a sample, and in particular, an inspection apparatus and an inspection for inspecting a pattern of a sample such as a photomask, a wafer, or a liquid crystal substrate used for manufacturing a semiconductor element or a liquid crystal display (LCD). It is about the method.

  As represented by a 1 gigabit class DRAM, a pattern constituting a large scale integrated circuit (LSI) is going to be on the order of submicron to nanometer. One of the major causes of a decrease in yield in the manufacture of LSI is a defect of a photomask used when an ultrafine pattern is exposed and transferred onto a semiconductor wafer by a photolithography technique. In particular, with the miniaturization of the pattern dimensions of LSIs formed on semiconductor wafers, the dimensions that must be detected as pattern defects have become extremely small. For this reason, an apparatus for inspecting such a defect has been developed.

  On the other hand, with the development of multimedia, liquid crystal display devices (LCDs) are increasing in size of the liquid crystal substrate to 500 mm × 600 mm or more and miniaturization of patterns such as TFTs formed on the liquid crystal substrate. Progressing. Therefore, it is required to inspect a very small pattern defect over a wide range. For this reason, there is an urgent need to develop a sample inspection apparatus for efficiently inspecting defects of a photomask used in manufacturing such a large area LCD pattern and a large area LCD in a short time.

In the background art, the inspection area is virtually divided into strip-shaped inspection stripes, and the inspection is executed by controlling the operation of the XYθ table so that the divided inspection stripes are continuously scanned. It is assumed that. In the actual comparison process, the comparison determination process is performed on the area where the inspection stripes are further subdivided, and after the images are aligned for each area, the comparison for determining the presence or absence of defects in the area is performed. An act has been made.

  However, when there is no pattern in the area, it is impossible to measure an accurate displacement amount, or even if there is only a pattern in the X axis direction, the displacement amount in the X axis direction cannot be measured. There are some cases. In addition, since the positional deviation of the area does not fluctuate extremely, even in such a case, more accurate alignment can be performed by using the positional deviation amount of the neighboring area (see Patent Document 1).

JP2000-147749

(1) The present invention is to perform an efficient sample inspection.
(2) Further, the present invention is to efficiently perform alignment of a sample image.

(1) In the embodiment of the present invention, a measurement image generation unit that measures a pattern of a sample to generate a measurement image, and a comparison that compares the measurement image with a reference image and determines a defect when the difference exceeds a threshold value In a sample inspection apparatus including a scanning unit, an area image of a certain area of the measurement image and the reference image is cut out, the amount of positional deviation between the area images is measured, and the reliability indicating the reliability of the positional deviation amount between the area images A misregistration measurement unit for obtaining information, and a registration unit for aligning area images using reliability information and misregistration amount ,
The reliability information is obtained by shifting the area images. When the gradient of the square sum average of the differences between the area images is large, the reliability of the deviation amount is high. When the gradient is small, the deviation amount is In the sample inspection device , the reliability is low .
(2) In the embodiment of the present invention, a measurement image is generated by measuring a pattern of a sample, an area image of a certain area of the measurement image and the reference image is cut out, and a positional deviation amount between the area images is measured. Then, reliability information indicating the reliability of the positional deviation amount between the area images is obtained, the area images are aligned using the reliability information and the positional deviation amount, and the area image between the measurement image and the reference image is obtained. In comparison, if the difference exceeds the threshold, it is determined as a defect ,
The reliability information is obtained by shifting the area images. When the gradient of the square sum average of the differences between the area images is large, the reliability of the deviation amount is high. When the gradient is small, the deviation amount is The sample inspection method has low reliability .

It is a block diagram for demonstrating the sample inspection apparatus and sample inspection method in embodiment. It is a conceptual diagram which shows the internal structure of the sample inspection apparatus in embodiment. It is a figure for demonstrating the acquisition procedure of an optical image. It is a figure of some area images. It is a figure which shows an example which calculates | requires deviation | shift amount and reliability information. It is a figure which shows the other example which calculates | requires deviation | shift amount and reliability information. It is a figure for demonstrating the procedure of a sample test | inspection. It is a figure which shows the other example which calculates | requires reliability information. It is a figure for demonstrating die-die (DD) comparison.

  Hereinafter, the details of the present invention will be described with reference to embodiments.

(Sample inspection equipment)
FIG. 1 is a block diagram illustrating a sample inspection apparatus and a sample inspection method according to an embodiment of the present invention. The sample inspection apparatus 10 inspects a pattern 12 of a sample such as a photomask, a wafer, or a liquid crystal substrate. The sample inspection apparatus 10 measures a sample pattern 12 by a measurement image generation unit 14 and generates a measurement image of optical data. Further, the sample inspection apparatus 10 generates a reference image by processing the design data 22 of the sample pattern by the reference image generation unit 24.

  The misregistration measurement unit 30 includes an area image cutout unit 300 that extracts an area image of a certain area from the generated measurement image and reference image, and a misregistration amount calculation unit 302 that measures the misregistration amount between the area images. A reliability information generation unit 304 that generates reliability information of the positional deviation amount of the area is provided. The reliability information is information indicating the certainty of the degree of coincidence between the positional deviation amount of the area images obtained by the positional deviation amount calculation unit 302 and the actual positional deviation amount. The storage unit 32 stores the positional deviation amount and reliability information of each area obtained by the positional deviation measurement unit 30. The alignment unit 34 refers to the reliability information read from the storage unit 32 and aligns the area image 16 of the measurement image and the area image 26 of the reference image. The comparison unit 36 compares the area images 16 and 26, and determines that the defect is a defect if the difference exceeds a threshold value.

  As will be described later, the measurement image generated by the measurement image generation unit 14 can be used as the reference image to be compared with the measurement image, in addition to the reference image obtained from the design data 22. Therefore, an image to be compared with the measurement image is referred to as a reference image including the reference image and the measurement image, and is described as a reference image (reference image and measurement image) in block 26.

  FIG. 2 is a conceptual diagram showing the internal configuration of the sample inspection apparatus 10. The sample inspection apparatus 10 inspects a pattern defect of the sample 100 using a substrate such as a mask or a wafer as the sample 100. The sample inspection apparatus 10 includes an optical image acquisition unit 110 and a control system circuit 150. The optical image acquisition unit 110 includes an autoloader 112, an illumination device 114 that generates illumination light, an XYθ table 116, an XYθ motor 118, a laser length measurement system 120, a magnifying optical system 122, a piezo element 124, and a light receiving unit 126 such as a photodiode array. Sensor circuit 128 and the like. In the control system circuit 150, a CPU 152 serving as a control computer is connected to a mass storage device 156, a memory device 158, a display device 160, a printing device 162, an autoloader control circuit 170, a table control circuit via a bus 154 serving as a data transmission path. 172, an autofocus control circuit 174, a development circuit 176, a reference circuit 178, a comparison circuit 180, a position circuit 182 and a displacement measurement circuit 184. The development circuit 176, the reference circuit 178, the comparison circuit 180, the position circuit 182 and the position deviation measurement circuit 184 are connected to each other as shown in FIG. The XYθ table 116 is driven by an XYθ motor 118 such as an X-axis motor, a Y-axis motor, or a θ-axis motor.

  The measurement image generation unit 14 in FIG. 1 can be configured by the optical image acquisition unit 110 in FIG. Similarly, the reference image generation unit 24 can be configured by a development circuit 176 and a reference circuit 178. The misregistration measurement unit 30 can be configured by a misregistration measurement circuit 184, and performs processing such as an area image cutout unit 300, a misregistration amount calculation unit 302, and a reliability information generation unit 304. The alignment unit 34 and the comparison unit 36 can be configured by the comparison circuit 180, and perform alignment processing, comparison processing, defect detection processing, and the like. The storage unit 32 can use the mass storage device 156 or the memory device 158. The storage unit 32 can be arranged at an appropriate location in the control system circuit 150. However, it is convenient to incorporate the storage unit 32 in the misalignment measuring unit 30 that performs misalignment processing. In FIG. 2, description of components other than those necessary for describing the present embodiment is omitted. The sample inspection apparatus 10 usually includes other necessary configurations.

(Operation of optical image acquisition unit)
The sample 100 is automatically transported from the autoloader 112 driven by the autoloader control circuit 170 and placed on the XYθ table 116. The sample 100 is irradiated with light by the illumination device 114. Below the sample 100, an magnifying optical system 122, a light receiving unit 126, and a sensor circuit 128 are arranged. Light that has passed through the sample 100 such as an exposure mask forms an optical image on the light receiving unit 126 via the magnifying optical system 122 and is incident thereon. The autofocus control circuit 174 controls the piezo element 124 to absorb the deflection of the sample 100 and fluctuations in the Z-axis direction (perpendicular to the X-axis and Y-axis) of the XYθ table 116, and focus on the sample 100. Align. The XYθ table 116 is driven by the table control circuit 172 under the control of the CPU 152. It can be moved by a drive system such as a three-axis (XY-θ) motor 118 driven in the X-axis direction, Y-axis direction, and θ-direction. For example, step motors can be used as these X motor, Y motor, and θ motor. The movement position of the XYθ table 116 is measured by the laser length measurement system 120 and supplied to the position circuit 182. The optical image received by the light receiving unit 126 becomes electronic data of a measurement image by the sensor circuit 128. The optical image of the sample 100 output from the sensor circuit 128 is sent to the misalignment measuring circuit 184 together with data indicating the position of the sample 100 on the XYθ table 116 output from the position circuit 182. A measurement image and a reference image are sent from the misalignment measurement circuit 184 to the comparison circuit 180. 2 uses the transmitted light of the sample 100, the reflected light of the sample 100 can also be used. The sample 100 on the XYθ table 116 is automatically discharged by the autoloader control circuit 170 after completion of the inspection. Note that the optical image is, for example, 8-bit unsigned data, and expresses the brightness gradation of each pixel.

(Reference image generation)
The design data 22 used when forming the pattern of the sample 100 is stored in the mass storage device 156. The design data 22 is input from the mass storage device 156 to the expansion circuit 176 through the CPU 152. As a design data development process, the development circuit 176 converts the design data of the sample 100 into binary or multivalued original image data, and the original image data is sent to the reference circuit 178. The reference circuit 178 performs an appropriate filter process on the original image data to generate a reference image. It can be said that the measurement image obtained from the sensor circuit 128 is in a state in which the filter is activated by the resolution characteristic of the magnifying optical system 122, the aperture effect of the light receiving unit 126, and the like. In this state, there is a difference between the characteristics of the measurement image and the reference image, so the original image data on the design side is also filtered by the reference circuit 178 to match the measurement image. The positional deviation measurement circuit 184 takes in the measurement image and the reference image, cuts out an area image, measures the positional deviation amount, and obtains reliability information. The comparison circuit 180 refers to the reliability information, aligns the area images by shifting the positional deviation amount, compares them according to a predetermined algorithm, and determines the presence / absence of a defect.

  In the above description, what is described as “˜circuit” or “˜process” can be configured by a computer-operable program. Or you may make it implement by not only the program used as software but the combination of hardware and software. Alternatively, a combination with firmware may be used. Alternatively, a combination of these may be realized. When configured by a program, the program is recorded on a recording medium such as a magnetic disk device, a magnetic tape device, an FD, or a ROM (Read Only Memory).

(Acquisition of inspection stripe image)
FIG. 3A is a diagram for explaining a measurement image acquisition procedure. The inspection area of the sample 100 is virtually divided by the scan width W in the Y-axis direction as shown in FIG. That is, the inspection area is virtually divided into a plurality of strip-shaped inspection stripes 101 having a scan width W. Further, the XYθ table 116 is controlled so that the divided inspection stripes 101 are continuously scanned. The XYθ table 116 moves along the X axis, and the measurement image is acquired as the inspection stripe 101. In FIG. 3A, the inspection stripe 101 has a rectangular shape having a scan width W in the Y-axis direction and a longitudinal direction having a length in the X-axis direction. The light receiving unit 126 continuously receives images with a scan width W as shown in FIG. After acquiring the image in the first inspection stripe 101, the light receiving unit 126 continuously inputs the image having the scan width W while moving the image in the second inspection stripe 101 in the opposite direction. The scan width W is about 2048 pixels, for example. When acquiring an image in the third inspection stripe 101, the image is moved in the direction opposite to the direction in which the image in the second inspection stripe 101 is acquired, that is, in the direction in which the image in the first inspection stripe 101 is acquired. Get an image.

(Sample displacement processing for sample inspection)
In the position shift process for sample inspection, the measurement image and the reference image are first cut out into areas of appropriate pixel sizes as shown in FIG. Usually, prior to the inspection, alignment (overall alignment) between the measurement image and the reference image is performed. This is performed using an appropriate dedicated mark, but may be performed using an arbitrary pattern edge or the like designated by the operator. Therefore, although the approximated positions of the cut out area images are aligned, there is a slight positional shift due to the distortion of the sample, the accuracy of the XYθ table 116, and the like. Therefore, the amount of positional deviation between the cut out area images (usually about 1024 pixel angles) is measured in sub-pixel units and adjusted so that the amount of deviation between them becomes small (area alignment), and then compared according to an appropriate algorithm. Determine the presence or absence of defects.

  There are various methods used for area alignment. For example, while shifting the pattern in the X and Y axis directions with an accuracy of about 1/16 pixel, the square sum of the differences between the two area images is taken. There is a technique for determining the position where the minimum is as the optimum position. In the case of the pattern shapes as shown in FIGS. 4A and 4B, there is often no problem, but when there is no pattern in the area or as shown in FIG. However, when the amount of misalignment is measured, it is sometimes impossible to measure the exact amount of misalignment if it is buried in noise. Even if there is a pattern, if only the pattern in the Y-axis direction as shown in FIG. 4D exists, the amount of displacement in the Y-axis direction may not be measured. Or, if there is an actual defect, it may be dragged by the defect to prevent accurate area alignment.

  However, considering the measurement pattern acquisition method, the positional displacement of the area does not fluctuate drastically, and even in such a case, it is possible to perform more accurate alignment by using the positional displacement amount of the neighboring area. Become. However, at that time, it is important whether or not the amount of positional deviation measured in the vicinity area can be trusted. Therefore, in the embodiment of the present invention, the amount of positional deviation between the area images in a certain area of the measurement image and the reference image is measured, and reliability information indicating the reliability of the amount of positional deviation between the area images is obtained. In addition, such a method of measuring misalignment and obtaining misalignment measurement result reliability information based on the misalignment can efficiently obtain highly reliable reliability information. In addition to the above-described method of taking the mean-square average of differences, there are methods such as absolute value averaging, signed average, and differences from adjacent pixels. The positional deviation measurement result reliability information is information from which reliability information can be obtained at the same time when measuring the positional deviation. The positional deviation measurement result reliability information is information included in the positional deviation amount measured by the reliability information.

(First embodiment)
5 (A) and 5 (B) are different from each other in the X-axis direction and the Y-axis direction with respect to the pattern of FIG. 4 (B) while shifting the pattern with an accuracy of 1/16 pixel. This is obtained by taking the square sum of the differences between two images. In the X-axis direction, the square sum average value is the smallest at a “deviation” of −1/16 pixel, and the square sum average value has a slope in both the negative and positive directions with a gradient. It is getting bigger gradually. In this case, it is considered that the area images coincide with each other with a displacement amount of −1/16 pixel in the X-axis direction. Also, in the Y-axis direction, the value of the square sum average is the smallest at a “deviation” of +1/16 pixel, and the slope of the square sum average value has a gradient in both the minus direction and the plus direction centered on it. It is getting bigger gradually. In this case, it is considered that the area images coincide with each other with a positional deviation amount of +1/16 pixel in the Y-axis direction. With respect to the pattern of FIG. 4B, peaks appear in both the X-axis direction and the Y-axis direction as shown in FIGS. 5A and 5B.

  6 (A) and 6 (B) are similar to FIGS. 5 (A) and 5 (B) in the X-axis direction and the Y-axis direction, respectively, with respect to the pattern as shown in FIG. 4 (D). Thus, the area images are obtained by taking the square sum average of the difference between the two images while shifting the pattern with an accuracy of 1/16 pixel. In FIG. 4D, with respect to the X-axis direction, as shown in FIG. 6A, the area images coincide with a positional deviation amount of ± 0 pixels, and a peak appears, but in the Y-axis direction, On the other hand, no peak appears as shown in FIG.

  The reliability of the “deviation amount” can be known from the gradients of the graphs shown in FIGS. 5 (A) and 5 (B) and FIGS. 6 (A) and 6 (B). When the gradient of the graph is large, the reliability of the “deviation amount” is high, and when the gradient is small, the reliability of the “deviation amount” is low. For example, in the case of FIG. 5A, the gradient of the graph is large, and the reliability of the “deviation amount” of −1/16 pixel is high, and also in the case of FIG. 5B, the gradient is higher than that of FIG. Slightly smaller, the reliability of the “deviation amount” of +1/16 pixel is slightly lowered. In the case of FIG. 6A, the gradient of the graph is larger than that in FIG. 5A, and the reliability of the “deviation amount” of ± 0 pixels is further increased.

  In this way, the reliability in the X and Y directions is divided into a plurality of levels, for example, 10 levels based on the magnitude relationship between the peak appearance pattern and the gradient, and output as reliability information for each area. This reliability information is reflected in its own area, but can be easily reflected in nearby areas. For example, the reliability information is stored in a table as shown in Table 1. In Table 1, the reliability is divided into 10 levels, and the larger the numerical value, the higher the reliability. A plurality of types of reliability information can be provided in each area.

(Sample inspection flow)
FIG. 7 shows the flow of the pattern inspection of the sample. As shown in FIG. 3A, a measurement image of the inspection stripe is acquired (step S1). Next, as shown in FIG. 3B, a certain area is cut out from the measurement image and the reference image (step S2). A positional deviation amount between the area images of the measurement image and the reference image that have been cut out is calculated (step S3).

  As an example, the amount of positional deviation between the area images of the measurement image and the reference image can be obtained from FIGS. 5A and 5B and FIGS. 6A and 6B as described above. In FIG. 5A, a positional deviation amount of −1/16 pixel in the X-axis direction is obtained, and in FIG. 5B, a positional deviation amount of +1/16 pixel in the Y-axis direction is obtained, and FIG. Then, the positional deviation amount of ± 0 pixels in the X-axis direction is obtained. However, in FIG. 6B, since there is no peak in the Y-axis direction, the amount of positional deviation cannot be obtained.

  From the gradients of the graphs shown in FIGS. 5A and 5B and FIGS. 6A and 6B, the reliability of the “deviation amount” can be known as described above. In the case of FIG. 6B, there is no change in the value of the root mean square, and the “deviation amount” cannot be obtained. Naturally, since the gradient is zero, the reliability cannot be obtained. In this way, the positional deviation amount and the reliability information are written in the storage unit 32 for each calculated area (step S4).

  As shown in FIG. 6B, when the “positional deviation amount” is unknown or the reliability is low, an area image with high reliability in the vicinity of this area is investigated, and an area image with high optimal reliability is checked. Is obtained, alignment processing is performed using the “position shift amount” (step S5). When the alignment is completed, the area image of the measurement image and the area image of the reference image are compared, and if the difference is larger than a predetermined threshold, it is determined as a defect (step S6). This procedure is performed in the same stripe area. When there is no comparison area (step S7), the process proceeds to the next stripe. When there is no stripe to be inspected (step S8), the inspection is terminated.

(Second Embodiment)
The calculation of the reliability information can be estimated from the pattern shape in addition to FIGS. 5 (A) and 5 (B) and FIGS. 6 (A) and 6 (B). As an example, a method for taking a difference from an adjacent pixel will be described. For example, when a difference (differentiation) is taken in the X-axis direction and the Y-axis direction for a pattern shape as shown in FIG. 8A (shown with 5 × 5 pixels for simplicity), FIG. ), As shown in FIG. Since the difference value increases when crossing the edge of the pattern, it can be estimated that the pattern of FIG. 8A has a vertical edge as shown in FIG. 4D. The reliability information for each direction of the target area can be created by comprehensively obtaining the certainty for each direction from these estimation results. If there is a large defect that affects the area alignment, the pattern shape information differs between the two images. The value of reliability information of such an area should be extremely lowered. By doing so, it is possible to perform inspection efficiently without being dragged by a defect and without being influenced by the pattern shape in the area.

(Third embodiment)
After obtaining the inspection stripe image, the area unit inspection is performed. As shown in FIG. 3B, there are a plurality of adjacent areas in the inspection stripe. Although it is fundamental to refer to reliability information of a closer area, there are cases where it is more appropriate to adopt reliability information of a farther area when the pattern shape is complicated. However, in the case of a distant area, it must be considered that the system is more affected by fluctuations in the system. For this reason, priority is not given to those having high reliability information, but a weight corresponding to the distance is given to the reliability information. For example, by assigning a weight inversely proportional to the distance to the reliability information, the reliability is lowered if the area image has a high reliability information but is far away.

(Fourth embodiment)
Up to now, reference has been made to the reference of area information within the inspection stripe, but the pattern shape may show a similar tendency in the direction across the inspection stripe. For this reason, reference to reliability information across inspection stripes should be taken into consideration. In this case, the weight is changed between the same inspection stripe and another inspection stripe, and the weight of the other inspection stripe is made smaller than that of the same inspection stripe to reduce the reliability.

(Fifth embodiment)
The operation of the XYθ table 116 is alternated between the forward direction (FWD direction) and the reverse direction (BWD direction) as shown in FIG. Depending on the traveling accuracy of the XYθ table 116, the track may be slightly different between forward traveling and reverse traveling. In such a case, if the previous inspection stripe, that is, the positional deviation information when the XYθ table 116 operates in the reverse direction is reflected, the positional deviation may be increased. Therefore, it is preferable to use the inspection stripe of the image acquired by operating in the same direction to reflect the positional deviation information.

  For example, if the first inspection stripe is a forward operation, the second inspection stripe is a backward operation, and the third inspection stripe is a forward operation. The positional deviation information of one inspection stripe is reflected.

  Since the comparison processing and the operation of the XYθ table 116 are not always synchronized, it is also effective to reflect the positional deviation information of the first inspection stripe in the positional deviation information reference of the fifth inspection stripe. . When using a forward inspection stripe by changing the weight in the direction of the inspection stripe, the reliability is increased by increasing the weight compared to the inspection stripe in the reverse direction. Furthermore, more accurate reliability can be obtained by combining with the weight of distance.

(Sixth embodiment)
When the figure shape in the area is a pattern as shown in FIG. 4D, it is difficult to find the optimum position in the Y-axis direction, but there are areas as shown in FIGS. 4A and 4B in the vicinity. If it exists, the corresponding area can be estimated from the Y position. Considering the influence of fluctuations in the system such as the sample inspection apparatus, the reflection amount in the neighborhood area is not limited to the sub-pixel unit but is limited to the pixel unit, and the alignment of the sub-pixel unit is performed within the processing area. Thereby, the noise reduction effect can be acquired. If only integer pixels are used, accumulation of errors can be prevented.

(Seventh embodiment)
In the above embodiments, the reliability information in the X and Y axis directions has been described. However, it is possible to measure both the X and Y axis directions for an oblique pattern as shown in FIG. 4B. In the above-described pattern shape determination, the directionality is not clear only in the X and Y axis directions, and it tends to be difficult to increase the reliability level. In preparation for such a case, in addition to the X and Y axis directions, it may be more appropriate to have the reliability information of the shape in the 45 degree direction with respect to the X and Y axes. Since there are diagonal lines other than 45 degrees as the pattern shape, reliability information may be obtained for several directions. However, if there are too many directions, it becomes complicated, so it is better to narrow down to an appropriate number of directions.

(Eighth embodiment)
In the embodiments described so far, the die-database (DB) comparison inspection apparatus that compares the measurement image with the reference image has been described. However, there is a die-die (DD) comparison inspection method that compares the measurement images. In the case of die-to-die comparison, one of the comparison measurement images is a reference image. In this case, it is assumed that there are two or more regions drawn from the same design data in the sample 100 as shown in FIG. A sample inspection apparatus for performing a die-to-die comparison inspection has two or more light sources and simultaneously captures an image to be compared and one light source, but has a memory for at least one inspection stripe, Some of the images to be compared are cut out from the captured images. Even in the case of such a die-to-die comparison inspection, as described in the embodiments so far, it is effective to reflect the positional deviation amount and reliability information of the neighboring area.

  By performing the alignment of the area with reference to the positional deviation amount and the reliability information of the neighboring area, it is possible to suppress the generation of pseudo defects due to the alignment error and perform an efficient inspection. A more reliable reliability can be obtained by combining a plurality of reliability information. Conventionally, when the amount of misalignment in the comparison target area cannot be measured accurately, the misalignment may not be successful by simply referring to the amount of misalignment in the neighboring area. Can be solved.

  The embodiments have been described above with reference to specific examples. However, the present invention is not limited to these specific examples. In each embodiment, the inspection position is scanned by moving the XYθ table 116. However, the XYθ table 116 may be fixed and other optical systems may be moved. That is, relative movement may be performed. In addition, although descriptions are omitted for parts and the like that are not directly required for the description of the present invention, such as a device configuration and a control method, a required device configuration and a control method can be appropriately selected and used. In addition, all sample inspection apparatuses that include the elements of the present invention and whose design can be appropriately changed by those skilled in the art are included in the scope of the present invention.

DESCRIPTION OF SYMBOLS 10 ... Sample inspection apparatus 12 ... Sample pattern 14 ... Measurement image generation part 16 ... Measurement image area image 22 ... Sample pattern design data 24 ... Reference image generation part 26 ... Area image 30 of a standard image (reference image or measurement image) ... Misalignment measuring unit 300 ... Area image cutting unit 302 ... Misalignment amount calculating unit 302 ... Reliability information generating unit 32 ... A storage unit 34 of positional deviation amount and reliability information in each area ... a positioning unit 36 ... a comparison unit 100, a sample 101, an inspection stripe 110, an optical image acquisition unit 112, an autoloader 114, Illuminating device 116... XYθ table 118 • XYθ motor 120 • Laser measuring system 122 • Enlargement optical system 124 • Piezo element 126 • Light receiving unit 128 • Sensor circuit 150 • Control circuit Control circuit 152 .. CPU
154 ·· Bus 156 · · Mass storage device 158 · · Memory device 160 · · Display device 162 · · Printing device 170 · · Autoloader control circuit 172 · · Table control circuit 174 · · Autofocus control circuit 176 · · Expansion circuit 178 .. Reference circuit 180 .. Comparison circuit 182 .. Position circuit 184.

Claims (8)

In a sample inspection apparatus comprising a measurement image generation unit that measures a pattern of a sample and generates a measurement image, and a comparison unit that compares the measurement image with a reference image and determines a defect when the difference exceeds a threshold value,
A positional deviation measurement unit that cuts out an area image of a certain area of the measurement image and the reference image, measures the positional deviation amount between the area images, and obtains reliability information indicating the reliability of the positional deviation amount between the area images;
An alignment unit that aligns the area images using the reliability information and the positional deviation amount , and
The reliability information is obtained by shifting the area images. When the gradient of the square sum average of the differences between the area images is large, the reliability of the deviation amount is high. When the gradient is small, the deviation amount is Sample inspection equipment with low reliability .   When aligning the area images to be aligned, the alignment unit refers to the reliability information of the neighboring area image. If the reliability of the neighboring area image is high, the positional deviation amount of the neighboring area image The sample inspection apparatus according to claim 1, wherein the area images are aligned using each other. Positional deviation amount and includes a storage unit that holds the reliability information,
The sample inspection apparatus according to claim 1, wherein the storage unit has a plurality of positional deviation amounts and reliability information thereof.   When aligning the alignment target area images, the alignment unit obtains the reliability by weighting the reliability information of neighboring area images based on the distance from the alignment target area, The sample inspection apparatus according to claim 1, wherein, when the reliability of the neighboring area images is high, the area images are aligned using the amount of positional deviation of the neighboring area images. In the sample inspection apparatus in which the operation of the XYθ table is controlled so that the pattern of the sample is virtually divided into strip-shaped inspection stripes and the inspection stripes are continuously scanned,
The sample inspection apparatus according to claim 1, wherein the alignment unit refers to reliability information of an area image obtained with an inspection stripe other than the inspection stripe to which the area image to be compared belongs. In the sample inspection apparatus in which the operation of the XYθ table is controlled so that the pattern of the sample is virtually divided into strip-shaped inspection stripes and the inspection stripes are continuously scanned,
The alignment unit refers to reliability information of an area image obtained by an inspection stripe other than the inspection stripe to which the comparison area image having the same operation direction of the XYθ table belongs. The sample inspection apparatus described. Reliability information has separately in the X-axis direction and the Y-axis direction, sample inspection apparatus according to any one of claims 1 to 6. Measure the pattern of the sample to generate a measurement image,
Cut out the area image of a certain area of the measurement image and the reference image,
Measure the amount of misalignment between area images,
Obtain reliability information indicating the reliability of the amount of misalignment between area images,
Use the reliability information and misregistration amount to align the area images,
Compare the area image of the measurement image and the reference image, and if the difference exceeds the threshold, determine the defect ,
The reliability information is obtained by shifting the area images. When the gradient of the square sum average of the differences between the area images is large, the reliability of the deviation amount is high. When the gradient is small, the deviation amount is Sample inspection method with low reliability .

Images