Image analysis – Applications – Manufacturing or product inspection
Reexamination Certificate
2001-01-26
2004-09-21
Bali, Vikkram (Department: 2623)
Image analysis
Applications
Manufacturing or product inspection
C382S146000, C382S147000
Reexamination Certificate
active
06795573
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an inspection method and apparatus and, more particularly, to an inspection method and apparatus for inspecting the formation state of a pattern on an object on which repetitive patterns are formed.
2. Description of the Related Art
In the manufacturing processes of semiconductor devices, liquid crystal display devices, and the like, circuit patterns and the like are formed sequentially on a substrate such as a wafer or a glass plate (to be referred to as a “substrate” or “wafer” hereinafter as needed). And an inspection apparatus for checking the formation state of the patterns is used in a predetermined step in the manufacturing process. As such inspection apparatus, an optical image inspection apparatus using light such as a laser beam, and an electron image inspection apparatus such as a scanning microscope using an electron beam have been put into practical use.
On the substrate of the semiconductor device or the like, identical patterns are periodically formed in each unit of so-called shot area. In a memory device or a liquid crystal display device, an identical pattern is periodically formed even in a single shot area.
As a technique for detecting any foreign matter or pattern defects (to be referred to as “pattern defects” hereinafter) on the substrate surface on which periodic repetitive patterns must be formed, a technique for comparing a raw image which is an optical or electron image obtained by the inspection apparatus and a shift image obtained by shifting the raw image by the repetition pitch (to be referred to as a “neighbor comparison method” hereinafter) has been proposed. And the neighbor comparison method is prevalently used as the inspection method of the formation state of periodic patterns. In such neighbor comparison method, a binary image having the number of gray level=2 is conventionally used, but a gray image with 3 or more gray level or continuous gray level (to be referred to as a “multi-gray level image” hereinafter) is often used today. In the neighbor comparison method, pattern defects or the like are estimated to be present at an image position where the difference value as a comparison result becomes equal to or larger than a predetermined value (threshold value).
As described above, in the conventional neighbor comparison method, actually formed patterns are compared. The actually formed patterns inevitably include errors from an expectation pattern which is to be originally formed upon pattern formation. For this reason, even when the difference between the signal levels (gray levels) of the raw and shift images at their identical positions is small, the differences between each signal level of the two images and the signal level (to be referred to as an “expectation level” hereinafter) of the expectation pattern are not always small. Even when the difference between the signal levels (gray levels) of the raw and shift images at their identical positions is large, the differences between each signal level of the two images and the expectation level are not always large.
That is, according to the conventional neighbor comparison method, even when the signal level at each image position has a large difference from the expectation level, pattern defects or the like are often not estimated to be present. In this case, even when pattern defects are present, they cannot be recognized. On the other hand, even when the signal level at each image position is not largely different from the expectation level, pattern defects are estimated to be present. In this case, even when no pattern defects are present, a false detection of the pattern defects occurs.
As described above, a multi-gray level image is prevalently used, and the difference between the signal levels at each image position is used, but binary information indicating whether or not the “difference” value is larger than a threshold value is merely obtained. That is, only basically the same information as that obtained using a binary image is obtained. For this reason, although a multi-gray level image is used, information included in the “difference” value is not always fully utilized. That is, a technique for accurately inspecting the substrate surface, on which periodic repetitive patterns are to be formed, for pattern defects by fully utilizing information obtained by a multi-gray level image is demanded.
SUMMARY OF THE INVENTION
The present invention has been made in consideration of the above situation, and has as its object to provide an inspection method and apparatus which can accurately inspect the formation state of periodic repetitive patterns on an object.
According to the first aspect of the present invention, there is provided an inspection method for inspecting an object on which a specific pattern is periodically and repetitively formed along a predetermined direction, comprising the steps of: picking-up an image of the object using not less than three gray levels; and obtaining formation information of the specific pattern by statistically analyzing a difference between a raw image obtained as an image pick-up result obtained in the image picking-up step, and a reference image.
According to this method, since the difference between the raw image which is an image pick-up result of an object picked-up as multi-gray level data, and the reference image is statistically analyzed to obtain the formation information of the specific pattern, the formation information of the specific pattern can be obtained by effectively using information contained in the multi-gray level image. Hence, the formation state of periodic repetitive patterns on the object can be accurately inspected.
In the inspection method of the present invention, the step of obtaining the formation information comprises: generating data points, which are defined at as data sets of gray levels at identical positions in said raw and reference images, in a coordinate space which has coordinate axes corresponding to values of the gray levels in said raw and reference images; and obtaining pattern formation information, based on a distribution of said data points in said coordinate space.
In the inspection method of the present invention, upon obtaining the formation information, (N−1) (N is an integer equal to or larger than 2) shift images are obtained by shifting the raw image obtained as the image pick-up result in the image pick-up step by integer multiples of a repetition period in a repetition direction of the specific pattern in the image pick-up result; sets of gray levels at identical positions in N images including the raw image and (N−1) shift images are defined as data points, and data points corresponding to positions in overlapping regions of the N images are plotted in an N-dimensional coordinate space; and pattern formation information of the object is obtained on the basis of a state of a distribution of the data points in the N-dimensional coordinate space.
In such case, based on the raw image which is obtained by picking-up the object and has three or more gray levels, (N−1) shift images are obtained by shifting the raw image in the repetition direction by integer multiples of the repetition period of the specific pattern. Sets of gray levels at identical positions of N images consisting of the raw image and (N−1) shift images are defined as data points in the N-dimensional coordinate space, and data points at respectively positions in overlapping regions of the N images are plotted in the N-dimensional coordinate space.
The plotted data points are distributed around a straight line or a curve (to be generally referred to as an “expectation line” hereinafter) formed by a set of data points of those similarly plotted in an expectation pattern (to be referred to as “expectation data points” hereinafter). When, for example, repetitive patterns formed are exactly the same, and are expected to be simultaneously picked up under identical conditions, the expectation line as a set of expec
Bali Vikkram
Nikon Corporation
LandOfFree
Inspection method and apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Inspection method and apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Inspection method and apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3220366