US-12618788-B2 - System and method for denoising a region of interest of a pattern

Abstract

A system and method for denoising a grey scale image of a pattern on a substrate, including: scanning a gaussian beam consecutively across scanlines extending along an x direction, wherein the scan lines are disposed adjacent to each other in a y direction within a predetermined region of the pattern, obtaining waveforms from secondary electrons and backscattered electrons reflected from the scanned pattern; converting the waveforms into a grey scale image, and applying a gaussian weighted distribution for each of the scan lines of the grey scale image and the neighboring scan lines of the grey scale image to account for the impact of the neighboring scan lines to each scan line.

Inventors

• Vladislav Kaplan

Assignees

• ETROLOGY, LLC

Dates

Publication Date

20260505

Application Date

20230427

Claims (10)

1. 1 . A method of denoising secondary and backscattered electron images obtained from a predetermined area of interest of a pattern, the method comprising: scanning an electron beam over scan lines across the pattern a plural of times in an x direction, wherein the scan lines are disposed adjacent to each other in a y direction within a predetermined area of interest of the pattern; obtaining a secondary electron image of the predetermined area of interest and a backscattered electron image of the predetermined area of interest; and applying a weight of one to a normalized scan line and a gaussian weighed distribution to neighboring scan lines, wherein the gaussian weighted distribution is determined according to each neighboring scan line position and wherein the gaussian weighed distribution equation is: Weight ( x ) = e ( - 1 / 2 ) ⁢ ( x - μ σ ) 2 = 1 , σ represents the width of the gaussian, x represents the scan line being processed and μ represents the mean of the scan line.
2. 2 . The method according to claim 1 , wherein the gaussian weighted distribution for each scan line is applied using the equation: Weight ( x ) = e ( - 1 / 2 ) ⁢ ( x - μ σ ) 2 , where σ represents the width of the gaussian, μ represents the location on a maximum point, and “x” represents the line currently being processed.
3. 3 . The method according to claim 1 , wherein the pattern is formed in a substrate.
4. 4 . The method according to claim 1 , wherein each scan line begins at a first edge of the pattern and ends at a second edge of the pattern.
5. 5 . A non-transitory computer readable medium that stores instructions for: scanning an electron beam over scan lines across the pattern a plural of times in an x direction, wherein the scan lines are disposed adjacent to each other in a y direction within a predetermined area of interest of the pattern; obtaining a secondary electron image of the predetermined area of interest and a backscattered electron image of the predetermined area of interest; and applying a weight of one to a normalized scan line and a gaussian weighed distribution to neighboring scan lines, wherein the gaussian weighted distribution is determined according to each neighboring scan line position and wherein the gaussian weighed distribution equation is: Weight ( x ) = e ( - 1 / 2 ) ⁢ ( x - μ σ ) 2 = 1 , σ represents the width of the gaussian, x represents the scan line being processed and μ represents the mean of the scan line.
6. 6 . The non-transitory computer readable medium according to claim 5 , wherein the gaussian weighted distribution for each scan line is applied using the equation: Weight ( x ) = e ( - 1 / 2 ) ⁢ ( x - μ σ ) 2 , where σ represents the width of the gaussian, μ represents the location on a maximum point, and “x” represents the line currently being processed.
7. 7 . A system for denoising secondary and backscattered electron images obtained from a predetermined area of interest of a pattern, the system comprising: a secondary electron detector to detect secondary scattered electrons reflected from a pattern in response to scanning a gaussian beam across the pattern consecutive times within a region of interest of the pattern, and to store the detected secondary scattered electrons as waveforms; a backscatter electron detector to detect backscattered electrons reflected from a pattern in response to scanning a gaussian beam across the pattern consecutive times within a region of interest of the pattern, and store the backscattered electrons as waveforms; and a processor to convert the stored secondary electron waveforms and the backscattered electron waveforms together as a grey scale image of the region of interest of the pattern, to apply a weight of one to a normalized scan line of the grey scale image and a gaussian weighed distribution to neighboring scan lines of the grey scale image, wherein the gaussian weighted distribution is determined according to each neighboring scan line position and the gaussian weighed distribution equation is: Weight ( x ) = e ( - 1 / 2 ) ⁢ ( x - μ σ ) 2 = 1 , σ represents the width of the gaussian, x represents to the scan line being processed and μ represents the mean of the scan line.
8. 8 . The system according to claim 7 , wherein the processor applies the gaussian weighted distribution for each scan line using the equation: Weight ( x ) = e ( - 1 / 2 ) ⁢ ( x - μ σ ) 2 , where σ represents the width of the gaussian, μ represents the location on a maximum point, and “x” represents the line currently being processed.
9. 9 . The system according to claim 7 , wherein the pattern is formed in a substrate.
10. 10 . The system according to claim 7 , wherein the processor converts the stored secondary electron waveforms and the backscattered electron waveforms together as a grey scale image using a photomultiplier detector (PMT) and a video multiplier.

Description

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT Not applicable. COPYRIGHT NOTICE A portion of this disclosure contains material which is subject to copyright protection. The copyright owner has no objection to the photocopy reproduction by anyone of the patent document or the patent disclosure in exactly the form it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever. 37 C.F.R 1.71 (d). BACKGROUND OF THE INVENTIVE CONCEPT 1. Field of the Invention The present inventive concept relates to denoising of critical dimension scanning electron microscopy. More particularly, but not exclusively, this inventive concept relates to denoising of critical dimension scanning electron microscopy by gaussian weighted averaging of scanlines of a pattern. Description of the Related Art Measuring a pattern, generally formed within a substrate, in order to detect imperfections which can cause faulty circuits, is a well-known process. This process is referred to as critical dimension (CD) scanning electron microscopy. Generally, in CD scanning electron microscopy, an electron image is obtained from an area of a sample to be tested. The area to be sampled comprises a pattern and the electron image obtained comprises multiple lines scanned with an electron beam across the pattern, where each scanned line includes information obtained by scanning the electron beam over a scanline across the pattern and between the vertical edges of the pattern (a leading edge and a trailing edge of the pattern). For each scan line two peaks (representing the edges) must be located, and the distance between the peaks (the critical dimension (CD)) is measured. Waveforms are obtained and stored as result of the interaction of scanning an electron beam on a surface along the scan lines across the pattern, and these waveforms are converted into a grey scale image of the pattern. However, these resulting waveforms converted into a grey scale image, in most cases, are very noisy and difficult to read. Furthermore, measuring each scanline separately in order to determine the edges of the pattern creates a highly inaccurate measurement. Accordingly, “denoising” of the obtained grey scale image is a highly important part of image preprocessing techniques. A well-known technique for denoising is called averaging. Averaging is a process of taking all values sitting at the same location X between neighboring scan lines and replacing a particular scan line of interest with averaged values for each X. The resulting averaged scan line is generally much cleaner and easier for reading to determine critical dimensions (CD) of a pattern or other object. This process is generally performed for all scan lines in the resulting waveform to obtain a new waveform which is cleaner and easier to measure. FIG. 1 illustrates one process of denoising by averaging. Here the process is performed by averaging all the scan lines within a region of interest (ROI) (i.e., a designated window) with the same scale factor. In other words, any scan line before averaging will be provided with the same weight. The averaging of all scan lines is performed using the following equation: x_=1n⁢∑i=1nxi=1n⁢(x1+…+xn), where x1, x2 . . . xn are the scanlines. The results from averaging each scan line is poor sensitivity or a poor signal-to-noise ratio (SNR), since neighboring scan lines throughout the entire window are averaged with the same weight as the center scan line, while in fact neighboring scan lines have an effect, although less intense, on the center scan line being averaged. FIG. 2 illustrates another process of denoising by averaging. Here the process is performed by using a “sliding window” within the region of interest. More specifically, the sliding window moves one scan line down at a time, each time covering the same number of scan lines, but progressing downward until each scan line is averaged. The averaging of all scan lines is performed, as above, using the following equation: x_=1n⁢∑i=1nxi=1n⁢(x1+…+xn), where x1, x2 . . . xn are the scanlines. The results from averaging each scan line via the “sliding window” method is also poor sensitivity or a poor SNR, since neighboring scan lines throughout the entire window are averaged with the same weight as the center scan line, while in fact neighboring scan lines have an effect, although less intense, on the center scan line being averaged. Accordingly, there is a need to perform a process of denoising scan lines by taking into account the positioning of the scan lines, and therefore the impact of the scan lines depending on the proximity of the neighboring scan lines from the scan line being averaged. Accordingly, there is also a need to perform a process of denoising scan lines by applying a gaussian weighted scale to the scan lines within a region of interest of the scan lines detected. SUMMARY OF THE INVENTIVE CONCEPT The present general invent