JP2009250777A - Surface inspection device and surface inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a surface inspection device constituted of a usual surface inspection device composed of one two-dimensional camera, one illuminator, an operation stage and an image processor all of which are generally easily obtained, for determining whether a flaw is projected and realizing inspection processing in a short time. <P>SOLUTION: The surface inspection device includes: a stage on which an inspection target is placed; an imaging device composed of one illumination optical system and one image acquiring part for acquiring a plane image; a moving control part for relatively moving the stage and the imaging device; an image memory part; a flaw detection part; an image aligning part for setting an image, of which the flaw is detected by the flaw detection part, as a first image and acquiring a second image containing the same flaw as the first image in positional relation different from the positional relation of the stage with the imaging device in acquirement of the first image to align the first image with the second image; and a projection determining part for comparing the aligned first and second images with each other and determining whether the flaw remains projected. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an apparatus for inspecting a surface defect for a planar object such as a semiconductor chip or a circuit pattern and an inspection method thereof.

A planar object having a pattern such as a semiconductor chip or a circuit pattern is detected for defects by a surface inspection apparatus composed of one two-dimensional camera, one illumination, a movable stage, an image processing apparatus, and the like. ing. As a method for detecting surface defects by image processing using a two-dimensional image captured by a two-dimensional camera such as a CCD camera, a method for comparing with a reference pattern prepared in advance, template matching based on a normalized correlation coefficient, etc. There are a number of known approaches.
In the surface inspection apparatus using such a two-dimensional camera, the detected defect is usually re-inspected after the foreign substance removal processing. This is because if the detected defect is the attachment of a removable foreign matter such as dust, it can be handled as a normal product by removing the foreign matter.
On the other hand, a three-dimensional inspection apparatus is also realized that acquires three-dimensional information of an inspection object, detects a defect, and can determine whether or not the detected defect is due to adhesion of foreign matter or the like. As a known technique for acquiring three-dimensional information of an object, a stereo method using two cameras and a three-dimensional measurement method using a laser are known. However, these techniques require a dedicated device such as a plurality of cameras and laser scanners, and thus have problems such as an increase in size of the device and an increase in cost. As means for solving these problems, the invention described in Patent Document 1 is known.

A schematic configuration relating to the invention described in Patent Document 1 is shown in FIG. In the present invention, the stereo method is applied by the image processing apparatus 110 using the special optical system 114 that bisects the reflected light from the inspection object 118 using one camera 112 and one illumination 116. Get 3D information.
JP 2002-213929 A

In surface inspection of planar inspection objects, a surface inspection device using a two-dimensional camera cannot determine whether the detected defect is adhesion of foreign matter, etc., so all detected defects are re-inspected after foreign matter removal processing There is a problem that the inspection processing time becomes long.
In the invention described in Patent Document 1, the problem of the inspection apparatus that acquires the three-dimensional information is solved, but the special optical system 114 is accurately adjusted to obtain the three-dimensional information of the inspection object. It is very difficult to adjust. Furthermore, since the special optical system 114 is generally difficult to obtain, a great amount of cost and time are required for production. Further, in order to acquire the three-dimensional information of the inspection object, many processes and calculations are required, so that the inspection processing time is longer than that of the surface inspection apparatus using a two-dimensional camera.
On the other hand, whether or not reinspection after the foreign substance removal process is necessary for the detected defect is sufficient if it is possible to determine whether or not the foreign substance is attached, and detailed three-dimensional information of the detected defect is not required.
In view of this, the present invention determines whether or not a defect is convex in a configuration of a normal surface inspection apparatus, which is generally a single two-dimensional camera, a single illumination, an operation stage, and an image processing apparatus, which are easily available. An object of the present invention is to provide a surface inspection apparatus that can perform inspection processing in a short time.

In order to solve the above-described problems, the invention described in claim 1 includes a stage on which an inspection object is placed, one illumination that irradiates light to the inspection target area of the inspection object, and an optical that expands the inspection target area. An imaging apparatus comprising: a system, an image acquisition unit that acquires a planar image of the inspection target area enlarged by the optical system; and the stage so that the planar image acquired by the imaging apparatus becomes the inspection target area. A movement control unit that relatively moves the imaging device; an image storage unit that stores an image sent from the imaging device; a defect detection unit that reads an image from the image storage unit and detects a defect; An image in which a defect is detected by the defect detection unit is set as a first image, and is the same as the first image in a positional relationship different from the positional relationship between the stage and the imaging device when the first image is acquired. Defects included An image registration unit that performs registration of the first image and the second image, and compares the first image and the second image that have been registered by the image registration unit. And a convex determining unit that determines whether or not the defect is in a convex state.
The invention described in claim 2 is the surface inspection apparatus according to claim 1, wherein the image alignment unit extracts a predetermined region not including the same defect from the first and second images, The first image and the second image are aligned using the extracted predetermined area.

A third aspect of the present invention is the surface inspection apparatus according to the first or second aspect, wherein the convexity determination unit is configured to detect the first and second images aligned by the image alignment unit. The same area including the same defect is compared, and it is determined whether or not the defect is in a convex state.
The invention described in claim 4 is the surface inspection apparatus according to claim 3, wherein the convexity determination unit has the same defect in the first and second images aligned by the image alignment unit. In the case where the sum of squares of pixel value differences in the same region including is larger than a preset threshold value, the defect is determined to be a convex state.
A fifth aspect of the present invention is the surface inspection apparatus according to any one of the first to fourth aspects, in which a single inspection object includes a plurality of inspection target areas, and the single inspection target area. When the defect detection unit detects a defect in the image and acquires the first image, the second image is acquired before the next inspection target area is imaged by the imaging device. Features.
The invention described in claim 6 is the surface inspection apparatus according to any one of claims 1 to 4, wherein when a plurality of inspection object areas are included in one inspection object, all of the inspection object areas are included. After the defect is detected by the defect detection unit and the first image is acquired, the second image is acquired.

The invention according to claim 7 is the surface inspection apparatus according to claim 1, wherein the surface inspection apparatus has a foreign matter removing device that removes foreign matter, and when the convexity determining unit determines that the defect is convex, A defect is detected again by the defect detection unit for the defect after performing the foreign substance removal process by the foreign substance removing apparatus.
The invention according to claim 8 performs image alignment between the first image and the second image obtained by imaging the inspection object having one defect from different angles, and The first image and the second image that are aligned are compared to determine whether or not the defect is in a convex state.

According to the inventions according to claims 1 to 6 and 8, it is possible to determine in a short time whether or not the defect is convex, and it is possible to reduce the inspection processing time.
According to the seventh aspect of the invention, it is possible to determine whether or not the defect is convex in a short time, and to reinspect only the convex defect after the foreign substance removal processing, thereby realizing a reduction in the inspection processing time and the steps. It is also possible to improve the yield.

An embodiment of the present invention will be described with reference to FIGS.
FIG. 1 is a diagram showing a configuration of a surface inspection apparatus according to the present invention. The surface inspection apparatus includes an inspection object 1, a stage 2 on which the inspection object 1 is placed and moved, an imaging apparatus 10 that images an inspection target area of the inspection object 1, an image processing apparatus 20, and a foreign substance removal apparatus (not shown). Composed. The inspection object 1 is a substantially planar object having a pattern such as a semiconductor chip or a circuit pattern. For example, the inspection object 1 is an object having a pattern on a plane as shown in FIG.
The imaging device 10 obtains a two-dimensional image of an inspection target area enlarged by the illumination 11 that irradiates light to the inspection object 1 for imaging, an optical system 12 that expands an observation image of the inspection object 1, and the optical system 12. The image acquisition unit 13 is configured. The illumination 11 can use LED illumination or the like, and the image acquisition unit 13 can use a CCD camera or the like.
The image processing device 20 acquires an image data sent from the imaging device 10 and stores the image data, a defect detection unit 22 that reads an image from the image storage unit 21 and detects a defect, and the stage 2 captures the image at different positions. The image alignment unit 23 that performs image alignment for the two images that include the same defect, and the two images that are aligned by the image alignment unit 23 are compared to determine whether the defect is convex or not. The convex determination unit 24 and a stage control unit 25 that controls the movement of the stage 2 are configured. The image processing apparatus 20 includes a CPU that performs various arithmetic processes, a hard disk that stores various data, a memory that stores data used for arithmetic processes, an external input / output unit that exchanges data with an external system, image processing results, It is a PC or the like configured by a display or the like that displays a captured image, and each of the functions is realized by a CPU executing a program stored in a hard disk or a memory.

Next, the surface inspection process will be described with reference to the flowchart shown in FIG.
In S <b> 1, the stage controller 25 instructs the stage 2 to move, and moves the inspection target area of the inspection object 1 to a position where the imaging apparatus 10 captures an image. After the movement of the stage 2 is completed, the imaging device 10 captures a two-dimensional image of the inspection target area and stores it in the image storage unit 21 as a first image. The defect detection unit 22 reads out the first image from the image storage unit 21, and uses a known defect detection technique such as a method of comparing the reference pattern image stored in advance with the first image and template matching based on a normalized correlation coefficient. Use and detect defects. Since the defect detection technique is not the gist of the present invention, detailed description is omitted. Here, a specific example of the first image when a defect is detected in S1 is shown in FIG. FIG. 4 is a first image obtained by enlarging the inspection target region of the inspection object shown in FIG.
In S2, it is confirmed whether or not a defect is detected in the first image. If no defect is detected, it is determined that the defect is normal (S3). On the other hand, if a defect is detected, in S4, the movement to a position that is different from the time of capturing the first image and that includes the detected defect in the image captured by the image capturing apparatus 10, The stage controller 25 instructs the stage 2. After the movement of the stage 2 is completed, an image is acquired by the imaging device 10 and stored in the image storage unit 21 as a second image. For the second image as well, the defect is detected by the defect detector 22 in the same manner as the first image.

First, specific examples of the second image in which the defect is detected in S2 are shown in FIGS. FIG. 5 is a second image corresponding to the first image shown in FIG. 4 when the detected defect is in a convex state. FIG. 6 is a second image corresponding to the first image shown in FIG. 4 when the detected defect is in a planar state.
Next, the cause of the difference between FIG. 5 and FIG. 6, which is a specific example of the second image, will be described.
FIG. 7 is a diagram showing a difference in optical path between the first image and the second image when the detected defect is in a convex state. As shown in FIG. 7, when the detected defect is in a convex state due to the attachment of a foreign substance or the like, the portion where the optical path is blocked by the foreign substance in the convex state is changed by moving the inspection object 1. In this example, the image of the inspection object and the foreign object do not overlap in the first image, but it can be understood that the image of the inspection object and the foreign object overlap in the second image. That is, the positional relationship between the pattern and the defect changes in the first image and the second image.
On the other hand, FIG. 8 is a diagram showing a difference in optical path between the first image and the second image when the detected defect is in a planar state. When the detected defect is a flat state such as a spot, even if the inspection object 1 is moved as shown in FIG. 8, the optical path is not blocked by the defect. Therefore, the positional relationship between the pattern and the defect does not change in the first image and the second image.

From the above, as shown in FIGS. 5 and 6, a difference occurs in the second image depending on whether or not the detected defect is convex.
In S5, the image alignment unit 23 extracts a predetermined area excluding an area including the same defect in the first image and the second image as an image alignment area, and the extracted first image and second image A difference in position between the first image and the second image is calculated using the image alignment region, and the first image and the second image are aligned.
The image alignment area is a pixel area outside the detected defect, and is a pixel area determined from the operation error amount of the stage 2 and the number of pixels necessary for image alignment. Specifically, as shown in FIG. 9, the image alignment region is a region surrounded by two thick lines in a substantially donut shape formed so as to surround the detected defect. The operation error amount of the stage 2 and the number of pixels necessary for image alignment need to be set according to the positioning accuracy of the stage and the pattern of the inspection object, and can be arbitrarily set. In addition, the pixel area located at the center of the substantially donut shape that is not used as the image alignment area can be an area extended outward by several pixels on the outer periphery of the detected defect.
As described above, the substantially donut-shaped pixel region excluding the region including the defect is extracted as the image alignment region because, as described above, when the defect is in the convex state, the first image and the second image are extracted. This is because, since the positional relationship between the defect and the pattern changes, if the first image and the second image are aligned using an image including the defective portion, the alignment accuracy is lowered.

Next, a method for calculating the difference in position between the first image and the second image will be described. As described above, the first image and the second image are images taken at different positions on the stage 2. If the position difference in the x-axis direction of the stage 2 at this time is a and the position difference in the y-axis direction is b, the coordinate value of the second image corresponding to the coordinate value (x, y) of the first image is (X + a, y + b). Therefore, when the pixel value of the first image is expressed by Expression 1, the corresponding pixel value of the second image can be expressed by Expression 2.
[Equation 1]

[Equation 2]

By using Equations 1 and 2 above, a and b that minimize the sum of squares of the difference between the pixel values of the first image and the second image in the image alignment region are obtained, so that the positions of the first image and the second image are determined. The difference (a, b) can be determined. Specifically, the difference (a, b) between the positions of the first image and the second image is obtained by obtaining a and b at which Equation 3 is minimized by the least square method.
[Equation 3]

By using (a, b) obtained as described above, the first image and the second image can be aligned by moving the first image by a in the x-axis direction and b in the y-axis direction.
When the operation error of the stage 2 is very small, the a and b are the difference in the movement position of the stage 2 instructed from the stage control unit 25 at the time of acquiring the first image and the second image. You can also.
In S6, the convex determination unit extracts, from the first image and the second image that have been subjected to image alignment, pixel regions that include the same defect and that have the region necessary for comparison as a defect neighboring region. Then, the difference amount of the pixel values in the defect vicinity region is obtained, and when the difference amount is equal to or larger than a predetermined threshold value, it is determined that the defect is in a convex state.
A specific example of the defect vicinity area extracted from the first image and the second image is shown in FIG. In this specific example, as a result of the alignment of the first image and the second image by the image alignment unit 23, the difference in position between the first image and the second image is (a, 0). In both the first image and the second image, a defect is included, and the defect vicinity area necessary for the determination of the convex state is m pixels × n pixels. In this case, if the x-axis position of the defect vicinity region in the first image is x to x + m and the y-axis position is y to y + n, the x-axis position of the defect vicinity region in the second image is x + a to x + a + m, y-axis. The position is from y to y + n.

By extracting the defect vicinity area as described above, the same defect is included in both the defect vicinity areas of the first image and the second image, and the defect vicinity areas of the first image and the second image correspond to each other. It can be easily grasped by obtaining the sum of squares of the difference in pixel values that the pixel area of the same portion is obtained and the positional relationship of the defect has changed with respect to the pattern of the inspection object.
Equation 4 is an equation for obtaining the sum of squares D of the pixel value difference using the defect vicinity region of the first image and the second image.
[Equation 4]

The sum of squares D of pixel value differences obtained by Equation 4 is compared with a preset threshold value L. If the square sum of pixel value differences D is greater than or equal to the threshold value L, the defect is determined to be convex. This is because, as described above, when the defect is in the convex state in the first image and the second image, the position of the pattern and the defect changes, and the square sum D of the difference between the pixel values in the common area becomes large.
The threshold value L is obtained by using a plurality of normal samples having no defect, obtaining the first image and the second image for a normal part without a defect, and taking a square sum D of pixel value differences as a region near the defect. Calculate and memorize many. The set of calculated sums of squares D of pixel value differences includes a plurality of variation factors such as the influence of imaging conditions and stage movements, and therefore is considered to follow a normal distribution. This distribution is a distribution of the sum of squares D of pixel value differences in a state where there is no convex state defect, and the upper deviation value may be a convex state defect. Therefore, the average and standard deviation are calculated, and the value of the upper percent point (for example, 99.9%) can be determined as the threshold value L according to the required detection accuracy.
Further, as a variation factor of the square sum D of the pixel value difference, when the pixel value inclination state (contrast of light and shade) of the image is large, that is, when a pixel group having a low value and a pixel group having a high value are adjacent to each other. Since a slight change in position results in a large pixel value difference, it is also possible to divide an image into regions and determine a threshold value for each region.

In S7, the defect determined not to be convex in S6 is determined to be a defect (S8). On the other hand, if it is determined to be convex, a foreign matter removal process is performed in S9. The content of the foreign matter removal processing is not a characteristic part of the present invention but is a known technique, and thus detailed description thereof is omitted. However, the foreign matter removal processing can be realized by sucking and removing the foreign matter with a powerful suction device.
In S10, the defect subjected to the foreign matter removal process is picked up again by the image pickup device 10 and stored in the image storage unit 21, and the defect detection unit 22 detects the defect using the image. If no defect is detected, it is determined in S11 that the foreign matter has been removed and can be handled as normal, and processing is performed as normal (S3). On the other hand, if a defect is detected again, it is determined in S11 that the defect cannot be removed, and the defect is processed (S8).
When the above processing is completed, the inspection processing is completed for one inspection target region. Next, the new inspection target area starts again from S1.
Further, as shown in the flowchart of FIG. 11, for a plurality of inspection target areas existing in the inspection object in S21, the first images of all the inspection target areas are acquired to detect defects, and then detected in S24. A second image including a defective defect may be acquired. In addition, since the following processes perform the same process as the above, description is abbreviate | omitted.

  Furthermore, although the stage 2 is moved in the above embodiment, the imaging device 10 may be moved.

It is a block diagram of the surface inspection apparatus which concerns on this invention. It is a figure showing the specific example of the test | inspection object of the surface inspection apparatus which concerns on this invention. It is a flowchart showing the processing content of the surface inspection apparatus which concerns on this invention. It is the figure which showed the specific example of the 1st image imaged with the surface inspection apparatus which concerns on this invention. It is the figure which showed the specific example of the 2nd image corresponding to the 1st image shown in FIG. 4 when the defect is a convex state imaged with the surface inspection apparatus which concerns on this invention. It is the figure which showed the specific example of the 2nd image corresponding to the 1st image shown in FIG. 4 when the defect is a plane state imaged with the surface inspection apparatus which concerns on this invention. It is the figure which showed the difference in the optical path of a 1st image and a 2nd image, when the detected defect is a convex state. It is the figure which showed the difference in the optical path of a 1st image and a 2nd image, when the detected defect is a plane state. It is the figure which showed the specific example of the image alignment area extracted with the surface inspection apparatus which concerns on this invention. It is the figure which showed the specific example of the defect vicinity area | region extracted from the 1st image and the 2nd image. It is another flowchart showing the processing content of the surface inspection apparatus which concerns on this invention. It is a block diagram of the three-dimensional inspection apparatus in the invention described in Patent Document 1.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Inspection object 2 Stage 10 Imaging device 11 Illumination 12 Optical system 13 Imaging device 20 Image processing device 21 Image memory | storage part 22 Defect detection part 23 Image alignment part 24 Convex determination part 25 Stage control part

Claims (8)

A stage for placing an inspection object;
From one illumination that irradiates light to the inspection target area of the inspection object, an optical system that expands the inspection target area, and one image acquisition unit that acquires a planar image of the inspection target area enlarged by the optical system, An imaging device,
A movement control unit that relatively moves the stage and the imaging device so that a planar image acquired by the imaging device becomes an inspection target region;
An image storage unit for storing an image sent from the imaging device;
A defect detection unit that reads an image from the image storage unit and detects a defect;
An image in which a defect is detected by the defect detection unit is defined as a first image, and is the same as the first image in a positional relationship different from the positional relationship between the stage and the imaging device when the first image is acquired. An image alignment unit that obtains a second image including the defect and aligns the first image and the second image;
A convexity determination unit that compares the first image and the second image that have been aligned by the image alignment unit and determines whether or not the defect is in a convex state;
A surface inspection apparatus characterized by comprising:   The image alignment unit extracts a predetermined area that does not include the same defect from the first and second images, and aligns the first image and the second image using the extracted predetermined area. The surface inspection apparatus according to claim 1, wherein the surface inspection apparatus is performed.   The convexity determination unit compares the same area including the same defect in the first and second images aligned by the image alignment unit, and determines whether or not the defect is in a convex state. The surface inspection apparatus according to claim 1, wherein the surface inspection apparatus is characterized.   When the convexity determination unit has a sum of squares of differences in pixel values of the same region including the same defect in the first and second images aligned by the image alignment unit greater than a preset threshold value The surface inspection apparatus according to claim 3, wherein the defect is determined as a convex state.   In the case where a plurality of inspection target areas are included in one inspection object, and when the defect detection unit detects a defect for the image of the one inspection target area and acquires the first image, 5. The surface inspection apparatus according to claim 1, wherein the second image is acquired before a next inspection target region is imaged by the imaging apparatus. 6.   In a case where a plurality of inspection target areas are included in one inspection object, the second image is obtained after the defect detection unit detects defects for all the inspection target areas and acquires the first image. 5. The surface inspection apparatus according to claim 1, wherein the surface inspection apparatus acquires one of the following:   A foreign matter removing device that removes foreign matter, and when the defect is determined to be convex by the convexity determining unit, after the foreign matter removing process is performed by the foreign matter removing device, the defect detecting unit again detects the defect The surface inspection apparatus according to claim 1, wherein detection is performed. Image alignment of the first image and the second image obtained by imaging the inspection object having one defect from different angles;
Comparing the first and second images, where the images are aligned, to determine whether the defect is in a convex state;
A surface inspection method characterized by the above.

Images