Defect detection method

Abstract

The present invention provides a method for detecting defects, which includes obtaining a first defect scan result of a wafer, obtaining a first background contour map of the defective part according to the first defect scan result, and performing grayscale processing and digitization processing on the first background contour map; obtaining a second defect scan result of the wafer after subsequent processes, obtaining a second background contour map according to the second defect scan result, and performing grayscale processing and digitization processing on the second background contour map, wherein the standards for grayscale processing and digitization processing of the first and second background contour maps are the same; by comparing the first and second background contour maps, locating the part of the second background contour map that matches the first background contour map and obtaining an enlarged image of that part. The present invention enhances the accuracy of defect location detection, ensures the accuracy of qualitative analysis of defects, and helps improve the product yield.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and particularly to a method for detecting defects. Background Art

[0002] As the size of integrated circuit devices continues to shrink, the yield loss caused by defects is increasing. During the defect detection process, the method of defect location detection can observe the true morphology of certain defects at subsequent stations, and can also observe the change process of the defect morphology during the process. Thus, it can be further analyzed whether the defect will affect the yield, so it plays an important role in improving the yield. However, in practice, due to the existence of interference signals, some defects cannot find the accurate position through comparison at subsequent stations, so qualitative analysis of the defects cannot be carried out.

[0003] To solve the above problems, a new method for detecting defects is needed. Summary of the Invention

[0004] In view of the above-mentioned disadvantages of the prior art, the purpose of the present invention is to provide a method for detecting defects, which is used to solve the problem that due to the existence of interference signals in the prior art, some defects cannot find the accurate position through comparison at subsequent stations, so qualitative analysis of the defects cannot be carried out.

[0005] To achieve the above purpose and other related purposes, the present invention provides a method for detecting defects, including:

[0006] Step 1: Obtain the first defect scan result of the wafer, obtain the first background contour map at the defect according to the first defect scan result, and perform grayscale processing and digital processing on the first background contour map;

[0007] Step 2: Obtain the second defect scan result of the wafer after subsequent processes, obtain the second background contour map according to the second defect scan result, and perform grayscale processing and digital processing on the second background contour map. The standards for grayscale processing and digital processing of the first and second background contour maps are the same;

[0008] Step 3: By comparing the first and second background contour maps, locate the part of the second background contour map that matches the first background contour map and obtain an enlarged image of that part.

[0009] Preferably, in Step 1, the first background contour map is obtained by corresponding the picture of the first defect scan result to the layout design file of the wafer.

[0010] Preferably, in Step 2, the second background contour map is obtained by corresponding the picture of the second defect scan result to the layout design file of the wafer.

[0011] Preferably, the size of the second background contour map in step two is larger than that of the first background contour map.

[0012] Preferably, in step one, gray-scale values are assigned to the light intensities received by different patterns on the wafer according to the scanning machine, and then digital processing is performed according to the gray-scale values to obtain a digital image.

[0013] Preferably, in step two, gray-scale values are assigned to the light intensities received by different patterns on the wafer according to the scanning machine, and then digital processing is performed according to the gray-scale values to obtain a digital image.

[0014] Preferably, in step three, the position of the defect in the second background contour map is initially located by comparing the gray-scale values of the first and second background contour maps, and then the position of the defect in the second background contour map is further located by comparing the pixel points of the first and second background contour maps.

[0015] As described above, the method for detecting defects of the present invention has the following beneficial effects:

[0016] The present invention uses graphic processing and digital analysis methods to restore the pattern under the defect, digitize it and make a comparison, accurately find the position where the defect should be located, enhance the accuracy of defect fixed-point detection, ensure the accuracy of qualitative analysis of the defect, and help improve the product yield. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It shows a schematic diagram of the front-layer defect image of the present invention;

[0018] Figure 2 It shows a schematic diagram of the front-layer defect background contour of the present invention;

[0019] Figure 3 It shows a schematic diagram of the digitized front-layer defect background contour of the present invention;

[0020] Figure 4 It shows a schematic diagram of the in-layer defect image of the present invention;

[0021] Figure 5 It shows a schematic diagram of the in-layer defect background contour of the present invention;

[0022] Figure 6 It shows a schematic diagram of the digitized in-layer defect background contour of the present invention;

[0023] Figure 7 It shows a schematic diagram of the defect fixed-point detection position of the present invention;

[0024] Figure 8 It shows a schematic diagram of the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0025] The embodiments of the present invention are described below through specific examples. Those skilled in the art can easily understand the other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention.

[0026] Please refer to Figure 8 , the present invention provides a method for detecting defects, including:

[0027] Step 1: Obtain the first defect scan result of the wafer to obtain a front-layer image as shown in Figure 1 . According to the first defect scan result, obtain the first background contour map of the defect area to obtain a contour map as shown in Figure 2 . Then, perform grayscale processing and digital processing on the first background contour map to obtain a digital image as shown in Figure 3 , that is, obtain the image of the defect area in the front layer of the wafer after grayscale processing and digital processing;

[0028] Specifically, image digitization is the process of converting an analog image with continuous spatial distribution and brightness values into a digital image that can be processed by a computer through sampling and quantization. To process an image on a computer, the real image (photos, pictorials, books, drawings, etc.) must first be digitized into a display and storage format that can be accepted by the computer, and then analyzed and processed by the computer. The process of image digitization mainly includes three steps: sampling, quantization, and coding.

[0029] Sampling:

[0030] The essence of sampling is to determine how many points are used to describe an image, and the quality of the sampling result is measured by the image resolution mentioned above. Simply put, a continuous image in two-dimensional space is equally spaced and divided into a rectangular grid structure in the horizontal and vertical directions. The resulting small squares are called pixel points. An image is sampled into a set composed of a finite number of pixel points. For example: an image with a resolution of 640*480 means that this image is composed of 640*480 = 307200 pixel points, and each small grid is a pixel point.

[0031] The sampling frequency refers to the number of samplings per second, which reflects the interval size between sampling points. The higher the sampling frequency, the more realistic the obtained image sample, the higher the image quality, but the required storage capacity is also larger.

[0032] When sampling, the selection of the sampling point interval size is very important, which determines the degree to which the sampled image can truly reflect the original image. Generally speaking, the more complex the picture and the richer the colors in the original image, the smaller the sampling interval should be. Since the sampling of a two-dimensional image is a generalization of one-dimensional sampling, according to the sampling theorem of signals, to accurately restore an image from the sampled samples, the Nyquist theorem for image sampling can be obtained: the sampling frequency of the image must be greater than or equal to twice the highest frequency component of the source image.

[0033] The following are several commonly used resolutions:

[0034] (1) Screen resolution. Screen resolution, also known as raster resolution, is used to determine how much information is displayed on the computer screen, measured in horizontal and vertical pixels. When the screen resolution is low (e.g., 640×480), there are fewer items displayed on the screen, but they are larger in size; when the screen resolution is high (e.g., 1600×1200), there are more items displayed on the screen, but they are smaller in size.

[0035] (2) Display resolution. Display resolution refers to the display mode that the computer display controller can provide. Display resolution is the number of pixels displayed on the screen. A resolution of 160×128 means that there are 160 pixels in the horizontal direction and 128 pixels in the vertical direction. In the case of the same screen size, the higher the resolution, the finer and more delicate the display effect.

[0036] (3) Image resolution. Image resolution refers to the amount of information stored in an image, which is the number of pixels per inch in the image. The unit of resolution is PPI (Pixels Per Inch), usually read as pixels per inch. The expression of image resolution is also "horizontal pixel number × vertical pixel number". It should be noted that there are different names for image resolution. In addition to the name of image resolution, it can also be called image size, image dimension, pixel dimension, and recording resolution. Here, the meanings of the words "size" and "dimension" are dual. They can refer to both the number of pixels (quantity size) and the size of the picture (side length or area size). Since in the case of the same display resolution, the higher the resolution of the image, the more pixels there are in the image, and the larger the size and area of the image, the image resolution can also be represented by image size and image dimension.

[0037] (4) Bit resolution. Bit resolution, also called bit depth, is used to measure the number of information bits stored in each pixel. This resolution determines the color information stored in each pixel of the image. For example, a 24-bit RGB image means that each of the primary colors R, G, and B of the image uses 8b, and the three together use 24b. And in an RGB image, each pixel has to record the information of the three primary colors R, G, and B. Therefore, the number of bits stored in each pixel is 24.

[0038] (5) Output resolution. The output resolution, also known as the printer resolution, refers to the number of dots per inch (dpi) produced by output devices such as printers. The output resolution determines the quality of the output image. The higher the output resolution, the fewer jagged edges in the print, and the smoother the grayscale tonal representation. The resolution of printers can reach 300 dpi or even 720 dpi (special paper is required); while the resolution of older laser printers models is usually 300 - 360 dpi. Due to the maturity of ultra-fine toner technology, the resolution of new laser printers can reach 600 - 1200 dpi, which is more than sufficient for professional typesetting output.

[0039] (6) Scanner resolution. The representation method of scanner resolution is similar to that of printers, generally also expressed in dpi. However, the dots here are sample points, which are different from the output points of printers. Generally, the horizontal resolution provided by scanners is higher than the vertical resolution. The resolution of desktop scanners can be divided into optical resolution and output resolution. The optical resolution refers to the image resolution actually scanned by the scanner hardware. Currently, the optical resolution of products on the market can reach 800 - 1200 dpi. The output resolution is the resolution generated after software enhancement and interpolation, which is approximately 3 - 4 times the optical resolution. So when you see a scanner claiming a resolution as high as 4800 dpi or 6400 dpi, this must refer to the output resolution.

[0040] The sampling frequency refers to the number of samplings per second, which reflects the interval size between sampling points. The higher the sampling frequency, the more realistic the obtained image samples, the higher the image quality, but the greater the required storage capacity.

[0041] When sampling, the selection of the sampling point interval size is very important, which determines the degree to which the sampled image can truly reflect the original image. Generally speaking, the more complex the picture and the richer the colors in the original image, the smaller the sampling interval should be. Since the sampling of two-dimensional images is a generalization of one-dimensional sampling, according to the sampling principle of signals, to accurately restore the image from the sampled samples, the Nyquist theorem for image sampling can be obtained: The sampling frequency of the image must be greater than or equal to twice the highest frequency component of the source image.

[0042] Quantization:

[0043] Quantization refers to the range of values to be used to represent each point after image sampling. The result of quantization is the total number of colors that the image can accommodate, which reflects the quality of sampling.

[0044] For example, if a point is stored in 4 bits, it means the image can only have 16 colors; if 16 bits are used to store a point, there are 2^16 = 65536 colors. Therefore, the larger the quantization bit number, the more colors the image can have, and naturally, a more detailed image effect can be produced. However, it will also occupy a larger storage space. The basic problem of both is the trade-off between visual effect and storage space.

[0045] Suppose there is a black-and-white grayscale photo. Since the grayscale changes are continuous both horizontally and vertically, it can be considered that there are infinitely many pixels, and the grayscale value at any point can have infinitely many possible values from black to white. By sampling at equal intervals along the horizontal and vertical directions, this analog image can be decomposed into approximately a finite number of pixels, and the value of each pixel represents the grayscale (brightness) of that pixel. Quantize the grayscale so that its value becomes a finite number of possible values.

[0046] An image that, after such sampling and quantization, shows a finite number of pixels distributed discretely in space and a finite number of discrete possible values in grayscale is called a digital image. As long as the number of sampling points in the horizontal and vertical directions is large enough and the quantization bit number is large enough, the quality of the digital image is no less than that of the original analog image.

[0047] The number of discrete values determined during quantization is called the quantization level. The number of binary digits required to represent the quantized color value (or brightness value) is called the quantization word length. Generally, 8 bits, 16 bits, 24 bits, or a higher quantization word length can be used to represent the colors of an image; the larger the quantization word length, the more accurately it can reflect the original colors of the image, but the capacity of the resulting digital image is also larger.

[0048] Compression coding:

[0049] The amount of image data obtained after digitization is extremely large, and coding techniques must be used to compress its information content. In a certain sense, coding compression technology is the key to realizing image transmission and storage. There are already many mature coding algorithms applied to image compression. Common ones include predictive coding, transform coding, fractal coding, wavelet transform image compression coding, etc.

[0050] When high-rate compression of the image information to be transmitted or stored is required, complex image coding techniques must be adopted. However, without a common standard as a basis, different systems cannot be compatible. Unless all the details of each coding method are exactly the same, it is very difficult to connect different systems.

[0051] In the embodiment of the present invention, in step one, a first background contour map is obtained through the layout design file (GDS) of the wafer corresponding to the picture of the first defect scan result.

[0052] In an embodiment of the present invention, since different patterns reflect and refract light differently, in Step 1, a gray-scale value is assigned according to the light intensity received by different patterns on the wafer by the scanner stage, and then digital processing is performed based on the gray-scale value to obtain a digital image.

[0053] Step 2: Obtain the second defect scan result of the wafer after subsequent processes to obtain the in-layer image as shown in Figure 4 According to the second defect scan result, obtain the second background contour map, and obtain the contour map as shown in Figure 5 Perform gray-scale processing and digital processing on the second background contour map to obtain the digital image as shown in Figure 6 The standards for gray-scale processing and digital processing of the first and second background contour maps are the same, that is, the entire image of the previous layer of the wafer including the defective part and subjected to gray-scale processing and digital processing is obtained;

[0054] In an embodiment of the present invention, in Step 2, the second background contour map is obtained by corresponding the picture of the second defect scan result to the layout design file of the wafer.

[0055] In an embodiment of the present invention, the size of the second background contour map in Step 2 is larger than that of the first background contour map.

[0056] In an embodiment of the present invention, since different patterns reflect and refract light differently, in Step 2, a gray-scale value is assigned according to the light intensity received by different patterns on the wafer by the scanner stage, and then digital processing is performed based on the gray-scale value to obtain a digital image.

[0057] Step 3: Refer to Figure 7 By comparing the first and second background contour maps, locate the part of the second background contour map that matches the first background contour map and obtain the magnified image at that location.

[0058] In an embodiment of the present invention, in Step 3, the position of the defect in the second background contour map is initially located by comparing the gray-scale values of the first and second background contour maps, that is, the defect and part of the surrounding area, and then the specific position of the defect in the second background contour map is further accurately located by comparing the pixel points of the first and second background contour maps.

[0059] It should be noted that the diagrams provided in this embodiment only illustrate the basic concept of the present invention in a schematic manner. Therefore, only the components related to the present invention are shown in the diagrams, rather than being drawn according to the number, shape, and size of the components in actual implementation. The types, quantities, and proportions of the components in actual implementation can be arbitrarily changed, and the component layout type may also be more complex.

[0060] In summary, the present invention uses graphic processing and digital analysis methods to restore the pattern under the defect, digitize it and make a comparison, accurately find the position where the defect should be located, enhance the accuracy of defect location detection, ensure the accuracy of qualitative analysis of the defect, and help improve the product yield. Therefore, the present invention effectively overcomes various disadvantages in the prior art and has high industrial utilization value.

[0061] The above embodiments are only illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Any person familiar with this technology can modify or change the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or changes made by those with ordinary knowledge in the technical field without departing from the spirit and technical ideas disclosed by the present invention should still be covered by the claims of the present invention.

Claims

1. A method for detecting defects, characterized in that, At least including: Step 1: Obtain the first defect scan result of the wafer, obtain the first background contour map at the defect according to the first defect scan result, and perform grayscale processing and digital processing on the first background contour map; Step 2: Obtain the second defect scan result of the wafer after subsequent processes, obtain the second background contour map according to the second defect scan result, and perform grayscale processing and digital processing on the second background contour map. The standards for grayscale processing and digital processing of the first and second background contour maps are the same; Step 3: By comparing the first and second background contour maps, locate the part of the second background contour map that matches the first background contour map and obtain an enlarged image thereof. The comparison includes: initially locating the position of the defect in the second background contour map by comparing the gray scale values of the first and second background contour maps, and then further locating the position of the defect in the second background contour map by comparing the pixel points of the first and second background contour maps.

2. The defect detection method according to claim 1, wherein: In Step 1, the first background contour map is obtained by corresponding the picture of the first defect scan result to the layout design file of the wafer.

3. The defect detection method according to claim 1, characterized in that: In Step 2, the second background contour map is obtained by corresponding the picture of the second defect scan result to the layout design file of the wafer.

4. The defect detection method according to claim 1, wherein: In Step 2, the size of the second background contour map is larger than that of the first background contour map.

5. The defect detection method according to claim 1, characterized in that: In Step 1, gray scale values are assigned to the intensities of light received by different patterns on the wafer according to the scanning machine, and then digital processing is performed according to the gray scale values to obtain a digital image.

6. The defect detection method according to claim 1, wherein: In Step 2, gray scale values are assigned to the intensities of light received by different patterns on the wafer according to the scanning machine, and then digital processing is performed according to the gray scale values to obtain a digital image.

Images