A wafer appearance detection method and system based on process area

Abstract

The application provides a wafer appearance detection method based on process regions, comprising the following steps: obtaining a standard wafer on a wafer and performing process-based region slicing on the standard wafer; aligning a to-be-detected wafer on the wafer by selecting feature points of the standard wafer; extracting corresponding process regions of the to-be-detected wafer based on different process regions of the standard wafer; and performing defect detection on different process regions of the to-be-detected wafer by using corresponding process detection methods. The application extracts the to-be-detected wafer based on different processes, and uses different detection methods to perform defect detection on different process regions, thereby improving the detection accuracy and efficiency. Meanwhile, the same standard wafer is used as a comparison standard for all to-be-detected wafers, thereby improving the detection robustness.

Description

Technical Field

[0001] This invention relates to the field of semiconductor optical defect detection technology, and in particular to a method and system for inspecting the appearance of grains based on process regions. Background Technology

[0002] In recent years, with the rapid development of semiconductor integrated circuits, the performance of integrated circuit devices has also been rapidly improving. As the performance of integrated circuit devices improves, the manufacturing process has become increasingly complex. Currently, advanced integrated circuit manufacturing processes typically involve hundreds of steps, and a problem in any one of these steps can cause issues with the entire semiconductor integrated circuit chip. This manifests as the semiconductor integrated circuit failing to meet design requirements, and in severe cases, may even lead to the failure of the entire chip. Therefore, timely detection of problems in the product manufacturing process is particularly important in semiconductor integrated circuit manufacturing. Based on these considerations, the industry generally uses highly sensitive optical inspection equipment to perform real-time defect detection on products.

[0003] The principle of defect detection is to scan the wafer under test with a high-sensitivity optical inspection device to obtain optical images of the chips on the wafer, then convert the obtained optical images into data images, and then detect the location of wafer defects by comparing the data images of chips at different locations.

[0004] A prior art Chinese patent (application publication number: CN 107910276 A) discloses a wafer inspection method, characterized by comprising: step S1, scanning the surface of the wafer to obtain a surface image of the wafer; step S2, uniformly dividing the surface image into several comparison units; step S3, subtracting each pixel in each comparison unit from the corresponding pixel in the adjacent comparison unit; if the difference is negative and less than a preset negative threshold, the wafer surface corresponding to that pixel is judged as a defect; if the difference is positive and greater than a preset positive threshold, the wafer surface corresponding to that pixel is judged as a defect. Although this patent discloses a wafer defect detection method, it does not consider the classification and extraction of different process regions of the wafer and further targeted defect detection. Summary of the Invention

[0005] To address at least one defect or improvement requirement in the prior art, the present invention provides a method and system for inspecting the appearance of grains based on process regions.

[0006] The present invention provides a method for inspecting the appearance of grains based on process regions, comprising the following steps:

[0007] Obtain standard dies on the wafer and perform process-based dicing on the standard dies;

[0008] The test grains on the wafer are aligned by selecting the feature points of the standard grains;

[0009] Based on the different process regions of the standard grain, the corresponding process regions of the test grain are extracted;

[0010] Defect detection is performed on different process regions of the grain to be tested using the corresponding process detection methods.

[0011] According to the die appearance inspection method based on process region provided by the present invention, the defect detection of different process regions of the die to be tested using the corresponding process inspection method includes:

[0012] The different process regions of the die to be tested are compared with the corresponding process regions of the standard die, thereby performing defect detection on the different process regions of the die to be tested.

[0013] According to the die appearance inspection method based on process region provided by the present invention, the die to be tested has the same size as the standard die, or the area coverage of each process region of the standard die corresponds to the area coverage of the corresponding process region of the die to be tested.

[0014] According to the die appearance inspection method based on process region provided by the present invention, after performing defect detection on different process regions of the die to be tested using the corresponding process inspection method, the method further includes:

[0015] Historical data on defect detection of the grain under test is obtained, and the historical data is input into the classifier of the corresponding process for training. The trained classifier of the corresponding process is used to judge and output defects in different process regions of the grain under test.

[0016] According to the method for inspecting the appearance of grains based on process regions provided by the present invention, the process regions include one or more of adhesive strip regions, tooth regions, and display regions.

[0017] According to the die appearance inspection method based on process region provided by the present invention, the feature points are selected by selecting the four corner points on the rectangular display area of ​​the standard die as feature points.

[0018] According to the grain appearance inspection method based on process region provided by the present invention, the extraction method of the adhesive strip region is as follows:

[0019] The image of the crystal to be tested is subjected to threshold segmentation, and the adhesive strip region is extracted using the result of the threshold segmentation.

[0020] According to the grain appearance inspection method based on process region provided by the present invention, the extraction method of the tooth region is as follows:

[0021] The image of the crystal grain to be tested is subjected to threshold segmentation, and the tooth region is extracted using the result of the threshold segmentation.

[0022] According to the grain appearance inspection method based on process region provided by the present invention, the method for extracting the display area is as follows:

[0023] The image of the grain to be tested is segmented by thresholding, and the display area is extracted using the result of threshold segmentation combined with morphological processing methods.

[0024] This invention provides a die appearance inspection system based on process region, comprising:

[0025] Standardization module: used to acquire standard dies on the wafer and perform process-based dicing on the standard dies;

[0026] Alignment module: used to align the test grains on the wafer by selecting feature points of the standard grain;

[0027] Extraction module: used to extract the corresponding process region of the test grain based on different process regions of the standard grain;

[0028] Detection module: used to perform defect detection on different process regions of the die to be tested using the detection methods corresponding to the process.

[0029] The beneficial effects of this invention are as follows:

[0030] (1) By performing process-based region extraction on the grains to be tested and using different detection methods for different process regions to detect defects, the accuracy and efficiency of the detection are improved.

[0031] (2) Using the same standard grain as the comparison standard for all grains to be tested improves the robustness of the detection. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0033] Figure 1 This is the flowchart of the die appearance inspection method based on process region of the present invention. Figure 1 ;

[0034] Figure 2 This is the flowchart of the die appearance inspection method based on process region of the present invention. Figure 2 ;

[0035] Figure 3 This is a schematic diagram of the adhesive strip region on the grain of the present invention;

[0036] Figure 4 This is a schematic diagram of the tooth region on the grain of the present invention;

[0037] Figure 5 This is a schematic diagram of the display area on the die of the present invention;

[0038] Figure 6 This is a schematic diagram illustrating the selection of feature points on the standard grain of the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0040] To facilitate understanding of this invention, a brief introduction to basic wafer knowledge will be given first:

[0041] 1. Wafer: A silicon wafer used to fabricate silicon semiconductor circuits. Its raw material is silicon. High-purity polycrystalline silicon is dissolved and doped with silicon crystal seeds, then slowly pulled out to form a cylindrical silicon ingot. After grinding, polishing, and slicing, the silicon ingot forms a silicon wafer, also known as a crystal.

[0042] 2. Grain: A lattice-shaped wafer unit that can be divided into after processing, also called a bare die.

[0043] 3. Chip: A product made by processing a die through testing, packaging and other processes.

[0044] 4. Standard grain: The grain with the lowest defect rate selected from multiple grains that can be divided from the wafer.

[0045] 5. Grains to be tested: Other grains on the same wafer besides the standard grains that need to be tested for defects.

[0046] 6. Chip manufacturing process (or wafer manufacturing process, die manufacturing process, abbreviated as "process"): A chip contains a large number of transistors. A transistor includes a source, drain, and gate. The minimum width of the gate (gate length) is a value in the XX nm process technology, which is the chip's manufacturing process, or simply the process. The chip manufacturing process characterizes the linearity or integration density of transistors on the chip. The smaller the process, the smaller the linearity of the transistors and the greater the integration density.

[0047] The industry typically uses highly sensitive optical inspection equipment to perform real-time defect detection on wafers. The principle of defect detection is to scan the wafer under test with the highly sensitive optical inspection equipment to obtain optical images of the dies on the wafer, then convert the obtained optical images into data images, and then perform subsequent comparison operations to detect defects in the dies or wafer.

[0048] This invention also begins by acquiring images of the grains using existing optical inspection equipment, followed by relevant extraction and comparison operations, ultimately performing defect detection on the grains. The detailed process of image extraction and data conversion for defect detection will not be elaborated further below.

[0049] like Figure 1 As shown, the present invention provides a method for inspecting the appearance of grains based on process regions, comprising the following steps:

[0050] Obtain standard dies on the wafer and perform process-based dicing on the standard dies;

[0051] The test grains on the wafer are aligned by selecting the feature points of the standard grains;

[0052] Based on the different process regions of the standard grain, the corresponding process regions of the test grain are extracted;

[0053] Defect detection is performed on different process regions of the grain to be tested using the corresponding process detection methods.

[0054] First, a standard die with the lowest defect rate is selected on the wafer. Since the distribution of the process regions of the die on this wafer is fixed, the standard die is first diced based on the process region. Then, feature points on the standard die are selected to accurately align the die to be tested. After alignment, the corresponding process regions of the die to be tested are extracted based on the different process regions of the standard die. Finally, different detection methods for the corresponding process are used to detect defects in different process regions of the die to be tested.

[0055] By performing process-based region extraction on the test grains and employing different detection methods for different process regions, the accuracy and efficiency of the inspection were improved. Furthermore, using the same standard grain as a comparison standard for all test grains enhanced the robustness of the inspection.

[0056] In one embodiment, defect detection of different process regions of the die under test using a corresponding process detection method includes: comparing different process regions of the die under test with the corresponding process regions of the standard die, and then performing defect detection on the different process regions of the die under test.

[0057] In one embodiment, the size of the die to be tested is the same as that of the standard die, or the area coverage of each process region of the standard die is consistent with the area coverage of the corresponding process region of the die to be tested.

[0058] The size of the test die and the standard die diced from the same wafer is consistent, and the area covered by each process region of the standard die is consistent with the area covered by the corresponding process region of the test die.

[0059] In one embodiment, after performing defect detection on different process regions of the die under test using the corresponding process detection method, the method further includes: acquiring historical data of defect detection of the die under test, inputting the historical data into the corresponding process classifier for training, and using the trained corresponding process classifier to judge and output defects in different process regions of the die under test.

[0060] As mentioned earlier, a large amount of historical data on defect detection in the target die has been acquired. To improve detection efficiency and enhance the targeting of the detection, this historical data—that is, defect data detected in different process regions on a large number of target dies—can be input into the corresponding SVM (support vector machine) classifiers for training. The trained classifiers for the corresponding process regions are then used to determine and output the defect types in different process regions of the target die. The overall flowchart of die defect detection is as follows: Figure 2 As shown.

[0061] In one embodiment, the process area includes one or more of a strip area, a tooth area, and a display area, with different processes for each area.

[0062] like Figure 3 As shown, the area covered by the adhesive strip is darker than the surrounding area. Figure 3 The area covered by the irregular borders in the middle and outer edges.

[0063] like Figure 4 As shown, the tooth area is the region within the adhesive strip area that is not covered by the adhesive, and it is brighter than the adhesive strip area. Figure 4 Some narrow linear regions covered by the teeth-like structures.

[0064] like Figure 5 As shown, the display area is the central rectangular section, which has a relatively regular shape and is darker than the surrounding underlying circuitry. Figure 5 The rectangular area shown in the middle.

[0065] In one embodiment, the feature points are selected by choosing the four corner points of the rectangular display area on the standard die as feature points. At the center of the die, there is a rectangular display area. More preferably, four rectangular areas with a size of 80 pixels centered on the four corner points are used as feature areas to align and extract the corresponding process area of ​​the die under test. The alignment accuracy can reach the sub-pixel level.

[0066] In one embodiment, the method for extracting the adhesive strip region is as follows: threshold segmentation is performed on the image of the crystal to be tested, and the adhesive strip region is extracted using the result of the threshold segmentation.

[0067] In one embodiment, the method for extracting the tooth region is as follows: threshold segmentation is performed on the image of the crystal grain to be tested, and the tooth region is extracted using the result of the threshold segmentation.

[0068] In one embodiment, the method for extracting the display area is as follows: threshold segmentation is performed on the image of the crystal grain to be tested, and the display area is extracted using the result of the threshold segmentation combined with morphological processing methods.

[0069] Image thresholding is a widely used image segmentation technique that utilizes the difference in grayscale characteristics between the target region and the background region in an image. The image is viewed as a combination of two types of regions (target region and background region) with different grayscale levels. A reasonable threshold is selected to determine whether each pixel in the image should belong to the target region or the background region, thereby generating a corresponding binary image.

[0070] Morphological image processing (or mathematical morphology) is one of the most widely used techniques in image processing. It primarily extracts image components from an image that are meaningful in representing and depicting the shape of regions, enabling subsequent recognition work to capture the most essential (most discriminative) shape features of the target object, such as boundaries and connected regions. Techniques like thinning, pixelation, and deburring are also frequently used in image preprocessing and post-processing, serving as powerful supplements to image enhancement techniques. Morphological operations on grayscale images include grayscale erosion, grayscale dilation, grayscale opening, and grayscale closing.

[0071] In this embodiment, an image of the crystal to be tested is first acquired. Then, based on the differences in grayscale characteristics of the adhesive strip region, tooth region, display region, and their surrounding regions, a threshold segmentation method is used to extract the adhesive strip region, tooth region, and display region. Preferably, the extraction grayscale thresholds for the adhesive strip region, tooth region, and display region can be set to 60, 150, and 100, respectively. Thus, when the grayscale value of a pixel in a certain region is greater than the corresponding extraction grayscale threshold, the grayscale value of that pixel is set to 255; otherwise, its grayscale value is set to 0. This binarization process, i.e., selecting a threshold, converts the image into a black and white binary image, and then performs image segmentation and edge extraction. For the display region, since it has obvious boundaries, its extraction requires mathematical morphological processing including operations such as erosion and dilation to enhance the target and separate it from the background. To extract the target boundary in the binary image including the display region, all points inside the target can be deleted (set to the background color). Preferably, the original display area and the surrounding binary image can be scanned line by line. If a black pixel is found to have eight black pixels in its eight neighborhoods, then that pixel is an internal point and is deleted from the target image. In effect, this is equivalent to using a 3x3 structuring element to erode the original image, so that only those internal points with eight black pixels in their eight neighborhoods are retained. Subtracting the eroded image from the original image then precisely deletes these internal points, leaving only the boundary pixels.

[0072] The present invention also provides a grain appearance inspection system based on process region, which can implement the above-mentioned corresponding inspection methods, and the inspection system includes the following modules:

[0073] Standardization module: used to acquire standard dies on the wafer and perform process-based dicing on the standard dies;

[0074] Alignment module: used to align the test grains on the wafer by selecting feature points of the standard grain;

[0075] Extraction module: used to extract the corresponding process region of the test grain based on different process regions of the standard grain;

[0076] Detection module: used to perform defect detection on different process regions of the die to be tested using the detection methods corresponding to the process.

[0077] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for inspecting the appearance of grains based on process regions, characterized in that, Includes the following steps: Obtain standard dies on the wafer and perform process-based dicing on the standard dies; The test grains on the wafer are aligned by selecting the feature points of the standard grains; Based on the different process regions of the standard grain, the corresponding process regions of the test grain are extracted; Defect detection is performed on different process regions of the grain to be tested using the corresponding process detection methods.

2. The method for inspecting the appearance of grains based on process regions according to claim 1, characterized in that, The defect detection of different process regions of the test grain using corresponding process detection methods includes: The different process regions of the die to be tested are compared with the corresponding process regions of the standard die, thereby performing defect detection on the different process regions of the die to be tested.

3. The method for inspecting the appearance of grains based on process regions according to claim 1, characterized in that, The size of the die to be tested is the same as that of the standard die, or the area coverage of each process region of the standard die corresponds to the area coverage of the corresponding process region of the die to be tested.

4. The method for inspecting the appearance of grains based on process regions according to claim 1, characterized in that, After performing defect detection on different process regions of the test grain using the corresponding process detection method, the method further includes: Historical data on defect detection of the grain under test is obtained, and the historical data is input into the classifier of the corresponding process for training. The trained classifier of the corresponding process is used to judge and output defects in different process regions of the grain under test.

5. The method for inspecting the appearance of grains based on process regions according to claim 1, characterized in that, The process area includes one or more of the adhesive strip area, the tooth area, and the display area.

6. The method for inspecting the appearance of grains based on process regions according to claim 5, characterized in that, The feature points are selected by choosing the four corner points of the rectangular display area on the standard grain as feature points.

7. The method for inspecting the appearance of grains based on process regions according to claim 5, characterized in that, The method for extracting the adhesive strip area is as follows: The image of the crystal to be tested is subjected to threshold segmentation, and the adhesive strip region is extracted using the result of the threshold segmentation.

8. The method for inspecting the appearance of grains based on process regions according to claim 5, characterized in that, The method for extracting the tooth region is as follows: The image of the crystal grain to be tested is subjected to threshold segmentation, and the tooth region is extracted using the result of the threshold segmentation.

9. The method for inspecting the appearance of grains based on process regions according to claim 5, characterized in that, The method for extracting the display area is as follows: The image of the grain to be tested is segmented by thresholding, and the display area is extracted using the result of threshold segmentation combined with morphological processing methods.

10. A grain appearance inspection system based on process region, characterized in that, include: Standardization module: used to acquire standard dies on the wafer and perform process-based dicing on the standard dies; Alignment module: used to align the test grains on the wafer by selecting feature points of the standard grain; Extraction module: used to extract the corresponding process region of the test grain based on different process regions of the standard grain; Detection module: used to perform defect detection on different process regions of the die to be tested using the detection methods corresponding to the process.

Images