Wafer overlay error measurement method based on Halcon

Abstract

The invention discloses a wafer overlay error measurement method based on Halcon, and belongs to the technical field of computer graphics. According to the method, based on an operator function provided by Halcon software, firstly, images of different areas of an overlay error mark are extracted from a wafer image, and a corresponding template is generated based on the images; and for the wafer image of which the overlay error is to be measured, searching a corresponding overlay error mark area by using the overlay error mark template, and realizing the overlay error measurement of a sub-pixel level, thereby calculating the offset between different layers. According to the method, the measurement precision and efficiency are remarkably improved, and the method can adapt to various complex chip scenes. Through step-by-step template search, error interference and calculation cost in a measurement process are effectively reduced. The method is excellent in performance in various chip environments, the measurement time of a single wafer image is shorter than one second, an efficient and reliable overlay error measurement solution is provided, and the method has wide industrial application prospects.

Description

Technical Field

[0001] The invention belongs to the technical field of computer graphics, relates to image-based error measurement, and specifically relates to a wafer overlay error measurement method based on Halcon. Background Art

[0002] Overlay error is a key indicator of the accuracy of alignment between photolithography layers during wafer manufacturing. Its importance lies in its direct impact on chip function and performance: if the error is too large, it may cause circuit short circuit or open circuit, seriously affecting the reliability of the device; at the same time, as the process advances to advanced nodes, the demand for precise control of overlay error is further increased. Any tiny error will amplify the negative impact on yield, thereby increasing production costs and extending the development cycle. Therefore, effectively controlling overlay error is crucial to improving semiconductor manufacturing yield, ensuring chip high performance and reducing production costs.

[0003] The measurement of overlay error is usually achieved through overlay marks, which are specific patterns designed on the wafer to assist in detecting the alignment deviation between different photolithography layers. Overlay marks are usually placed in the inactive area of ​​the wafer to avoid interfering with the device function. During the measurement process, the detection equipment captures the mark pattern through high-precision optical imaging technology, analyzes its displacement and rotation deviation in the X and Y directions, and thus calculates the alignment error between the layers.

[0004] Although neural networks have shown strong learning capabilities in some areas, they have many instability problems in industrial production. First, neural networks rely on large-scale labeled data for training, but it is costly to fully obtain labeled data in industrial production. Secondly, its measurement results are greatly affected by model parameters, data distribution and external environment, and stability is difficult to guarantee. In addition, the "black box" characteristics of neural networks make the results difficult to interpret, and their adaptability to new scenarios is limited, which may lead to uncontrollable measurement errors in production. The algorithm based on computer graphics uses a deterministic model, and the measurement results are highly repeatable and the error range is controllable. At the same time, this method can achieve sub-pixel measurement accuracy, meeting the requirements of industrial scenarios for real-time and efficiency. In addition, the logic of computer graphics technology is clear, the results are interpretable, and it is easy to analyze and trace problems, making it a reliable choice for large-scale industrial applications. In order to ensure the reliability and consistency of wafer overlay error measurement results, it is necessary to design a measurement method based on computer graphics. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes a wafer overlay error measurement method based on Halcon. By utilizing the operator functions provided by the Halcon software, different areas of the overlay error mark are framed and a model is established, thereby realizing wafer overlay error measurement based on computer graphics, which can improve the efficiency and stability of the measurement.

[0006] A wafer overlay error measurement method based on Halcon specifically comprises the following steps:

[0007] Step 1. Select a wafer image, establish a two-dimensional rectangular coordinate system with the top left corner of the image as the origin O, the horizontal direction as the X axis, and the vertical direction as the Y axis. Select an overlay error mark from the image and mark rectangular boxes 1, 2, and 3. Among them, rectangular box 1 contains only a complete overlay error mark, rectangular box 2 contains only the inner rectangular mark of the overlay error mark, and rectangular box 3 is located within the outer contour of the inner rectangular mark of the overlay error mark. Record the position information of the three rectangular boxes in the two-dimensional rectangular coordinate system XOY respectively.

[0008] Step 2: Calculate the center point coordinates of rectangular frames 1, 2, and 3, and then calculate the area between rectangular frames 1, 2 and rectangular frames 2, 3 to obtain an outer rectangular marking frame area and an inner rectangular marking frame area. The outer rectangular marking frame area and the inner rectangular marking frame area are used to crop the enhanced wafer image to obtain an enhanced outer rectangular marking frame image and an enhanced inner rectangular marking frame image.

[0009] The create_shape_model operator of the Halcon software is used to create an outer rectangular marker frame template and an inner rectangular marker frame template from the enhanced outer rectangular marker frame image and the inner rectangular marker frame image, respectively.

[0010] Step 3: For the wafer image that needs to be measured for overlay error, first enhance the image and establish a two-dimensional rectangular coordinate system X'O'Y' according to the method in step 1. Then, using the outer rectangular mark frame template obtained in step 2, find all outer rectangular mark frames that meet the template matching from the enhanced wafer image, and obtain the horizontal coordinate, vertical coordinate and rotation angle of the outer rectangular mark frame matching result relative to the outer rectangular mark frame template.

[0011] Step 4: Generate a position transformation matrix 2D1 from rectangular frame 1 to the outer rectangular frame based on the position information of the matching result of rectangular frame 1 and the outer rectangular frame, and use the position transformation matrix 2D1 to generate a rectangular frame on the enhanced wafer image, which has the same shape as rectangular frame 1 and contains the corresponding outer rectangular frame matching result. .

[0012] Step 5: Use a rectangular frame The enhanced wafer image is cropped, and the inner rectangular marking frame template obtained in step 2 is used on the cropped image to find the inner rectangular marking frame that matches the template, and the horizontal coordinate, vertical coordinate and rotation angle of the inner rectangular marking frame matching result relative to the inner rectangular marking frame template are obtained, and step 6 is entered. If there is no inner rectangular marking frame that matches the template, step 9 is entered.

[0013] Step 6: Generate a position transformation matrix 2D2 from rectangular frame 2 to the inner rectangular frame based on the position information of the matching result of rectangular frame 2 and the inner rectangular frame, and use the position transformation matrix 2D2 to generate a rectangular frame on the enhanced wafer image that has the same shape as rectangular frame 2 and contains the corresponding inner rectangular frame matching result. . Calculate the rectangular box , The area between them is detected by Gaussian line segment to obtain the outline of the outer rectangular marking box.

[0014] Step 7: Based on the position information of the rectangular frame 3 and the position transformation matrix 2D2, a rectangular frame having the same shape as the rectangular frame 3 and located within the matching result of the inner rectangular marking frame is generated on the enhanced wafer image. . Calculate the rectangular box , The area between them is detected by Gaussian line segment to obtain the contour of the inner rectangular marking box.

[0015] Step 8: Perform minimum rectangle fitting on the outer rectangular mark frame contour and the inner rectangular mark frame contour to obtain the outer rectangular mark prediction result and the inner rectangular mark prediction result. Use the area_center operator of the Halcon software to obtain the center coordinates of the inner and outer rectangular mark prediction results, and then calculate the deviation results of the center coordinates of the inner and outer rectangular mark prediction results in the X' axis direction and the Y' axis direction.

[0016] Step 9: According to the methods of steps 4 to 8, traverse each outer rectangular mark frame that meets the template matching in turn, and output the error measurement results of different overlay error marks on the image.

[0017] The present invention has the following beneficial effects:

[0018] 1. Optimize the calculation process and resource scheduling so that the overlay error measurement of a single wafer image takes less than one second, ensuring that the motion stage can complete target switching in the shortest time, thereby significantly improving the efficiency of industrial production and meeting the high-beat and high-precision production requirements.

[0019] 2. After obtaining the matching results of the outer rectangular mark frame, the measured image is cropped, and then the matching of the inner rectangular mark is performed. This can effectively filter out non-target features and focus on the analysis of key areas, ensuring the continuity and reliability of the entire production process and meeting the stringent requirements of the industrial environment.

[0020] 3. Enhance the image to accurately identify the target area. Using the operators provided by Halcon software, sub-pixel deviation calculation can be achieved, which improves the measurement resolution. Especially in complex image backgrounds or low-contrast conditions, it can accurately capture target features and achieve high-precision measurement.

[0021] 4. The traditional method is only optimized for a single target. For a wafer image with multiple overlay error marks, only a single target can be detected at a time. However, this method can directly detect multiple targets, improving measurement efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 is a schematic diagram of an overlay error mark;

[0023] Figure 2 Schematic diagram of a two-dimensional rectangular coordinate system XOY in the embodiment;

[0024] Figure 3 Rectangular frames 1, 2, and 3 are marked in the embodiment;

[0025] Figure 4 Mark the frame image for the enhanced inner and outer rectangles;

[0026] Figure 5 Schematic diagram of a wafer to be measured in the embodiment. DETAILED DESCRIPTION

[0027] The specific implementation of this method is described below with reference to the accompanying drawings.

[0028] A wafer overlay error measurement method based on Halcon, for Figure 1 The overlay error mark composed of the inner and outer rectangular marks shown in the figure is used for feature extraction and overlay error measurement. The specific steps are as follows:

[0029] Step 1: For wafer images of the same batch, the shape of the overlay error mark is consistent. Therefore, it is necessary to first make an overlay error mark template for the wafers of the same batch, and then find the overlay error mark in other images based on the template and measure the error.

[0030] Select a wafer image, first obtain its height and width, then establish a two-dimensional rectangular coordinate system XOY with the upper left corner of the image as the origin O, the horizontal direction as the X axis, and the vertical direction as the Y axis, such as Figure 2 As shown. Select one of the overlay error marks and mark the rectangular boxes 1, 2, and 3. Among them, rectangular box 1 only contains a complete overlay error mark, rectangular box 2 only contains the inner rectangular mark of the overlay error mark, and rectangular box 3 is located within the outer contour of the inner rectangular mark of the overlay error mark, as shown in Figure 3 As shown. Record the coordinates of the upper left vertex and the lower right vertex of the three rectangular boxes in the two-dimensional rectangular coordinate system XOY respectively.

[0031] Step 2: Based on the coordinate information obtained in step 1, the area_center operator of the Halcon software is used to calculate the center point coordinates of rectangular frames 1, 2, and 3, and then the difference operator is used to calculate the area between rectangular frames 1, 2 and rectangular frames 2, 3, to obtain the outer rectangular marking frame area and the inner rectangular marking frame area. The reduce_domain operator cuts the enhanced wafer image through the outer rectangular marking frame area and the inner rectangular marking frame area to obtain the enhanced outer rectangular marking frame image and the inner rectangular marking frame image, as shown in FIG. Figure 4 shown.

[0032] The create_shape_model operator of the Halcon software is used to create shape-based templates, outer rectangular marker frame templates and inner rectangular marker frame templates from the enhanced outer rectangular marker frame image and inner rectangular marker frame image, respectively.

[0033] In this embodiment, the wafer image is first converted into a grayscale image. :

[0034]

[0035] Among them, R, G, and B represent the R, G, and B channel values ​​of the wafer image respectively. Then a 3×3 filter is used to filter the grayscale image. Perform image enhancement with an enhancement factor of 1 to obtain an enhanced wafer image.

[0036] Step 3: For the wafer image that needs to be measured for overlay error, first enhance the image and establish a two-dimensional rectangular coordinate system X'O'Y' according to the method in step 1. Then, based on the outer rectangular marker frame template obtained in step 2, use the find_shape_model operator of the Halcon software to find all the outer rectangular marker frames that meet the template matching from the enhanced wafer image, and obtain the horizontal coordinate, vertical coordinate and rotation angle of the outer rectangular marker frame matching result relative to the outer rectangular marker frame template.

[0037] In this embodiment, only matching results with a confidence level higher than 0.9 are retained.

[0038] Step 4: Generate a position transformation matrix 2D1 from rectangular frame 1 to the outer rectangular frame based on the coordinate information of rectangular frame 1 saved in step 1 and the position information of the matching result of the outer rectangular frame. Use the position transformation matrix 2D1 to generate a rectangular frame with the same shape as rectangular frame 1 on the enhanced wafer image. , the rectangular frame The interior contains the corresponding outer rectangular marking box matching results.

[0039] Step 5: Use a rectangular frame The enhanced wafer image is cropped, and the find_shape_model operator is used again on the cropped image to find the inner rectangular marking frame that matches the template according to the inner rectangular marking frame template obtained in step 2, and the horizontal coordinate, vertical coordinate and rotation angle of the inner rectangular marking frame matching result relative to the inner rectangular marking frame template are obtained, and step 6 is entered. If there is no inner rectangular marking frame that matches the template, step 9 is entered.

[0040] Searching for the inner rectangular marking frame in the cropped image can avoid obtaining an erroneous inner rectangular marking frame matching result when performing template matching directly in the wafer image.

[0041] Step 6: Generate a position transformation matrix 2D2 from rectangular frame 2 to the inner rectangular frame based on the coordinate information of rectangular frame 2 saved in step 1 and the position information of the inner rectangular frame matching result, and use the position transformation matrix 2D2 to generate a rectangular frame with the same shape as rectangular frame 2 on the enhanced wafer image. , the rectangular frame The interior contains the corresponding inner rectangular marking box matching results.

[0042] Calculate the rectangle , The area between the two is used to obtain the predicted area of ​​the outer rectangular marking frame, the enhanced wafer image is cropped, and then the operator lines_gauss is used to perform Gaussian line segment detection to obtain the contour line of the outer rectangular marking frame.

[0043] Step 7: Based on the coordinate information of the rectangular frame 3 saved in step 1 and the position transformation matrix 2D2 from the rectangular frame 2 to the inner rectangular marking frame, a rectangular frame with the same shape as the rectangular frame 3 is generated on the enhanced wafer image. , thus ensuring the rectangular frame , The relative position relationship between them is consistent with the relative position relationship between rectangular frames 2 and 3. Located within the inner rectangular marker box matching result.

[0044] Calculate the rectangle , The area between the two is obtained to obtain the predicted area of ​​the inner rectangular marking frame, the enhanced wafer image is cropped, and then the operator lines_gauss is used to perform Gaussian line segment detection to obtain the contour line of the inner rectangular marking frame.

[0045] Step 8. Perform minimum rectangle fitting on the outer rectangle mark frame contour and the inner rectangle mark frame contour to obtain the outer rectangle mark prediction result and the inner rectangle mark prediction result. Use the area_center operator of the Halcon software to obtain the center coordinates of the inner and outer rectangle mark prediction results. The result is accurate to three decimal places. Then calculate the deviation results of the center coordinates of the inner and outer rectangle mark prediction results in the X' axis direction and the Y' axis direction, and store them in the arrays Res_dx and Res_dy.

[0046] Step 9: According to the methods of steps 4 to 8, traverse each outer rectangular mark frame that meets the template matching in turn, obtain the prediction results of the inner and outer rectangular marks, and calculate the error measurement results of different overlay error marks on the image.

[0047] For Figure 5 For the wafer image to be measured shown in the figure, first use the outer rectangular marking frame template obtained in step 2 to find all the outer rectangular marking frames that meet the template matching, and then traverse the interior of the outer rectangular marking frame matching results a, b, and c in turn, and use the inner rectangular marking frame template obtained in step 2 to find the inner rectangular marking frame that meets the template matching. Since there is no inner rectangular marking frame that meets the template matching in the outer rectangular marking frame matching result b, it means that the outer rectangular marking frame matching result b is not a correct overlay error mark. Therefore, the error calculation results of the two overlay error marks in the figure are finally output.

[0048] This method can measure the overlay error of a single wafer image in less than 1 second, and can complete the measurement of multiple overlay error marks in the image at one time. Its stability and measurement efficiency can meet the application requirements of industrial grade. The neural network-based method not only requires a large amount of labeled data for model training, but also takes about 2 seconds to measure wafer images with multiple overlay error marks. At the same time, the interpretability of the neural network is relatively weak, which also limits its promotion in industrial scenarios.

Claims

1. A wafer overlay error measurement method based on Halcon, which extracts features of an overlay error mark composed of two inner and outer rectangular marks in a wafer image and measures the overlay error, characterized in that: The specific steps are as follows: Step 1, use rectangular frames 1, 2, and 3 to mark an overlay error mark in the wafer image, wherein the area between rectangular frames 2 and 3 only includes the contour line of the inner rectangular mark, which is recorded as the inner rectangular mark frame area; the area between rectangular frames 1 and 2 only includes the contour line of the outer rectangular mark, which is recorded as the outer rectangular mark frame area; record the position information of each rectangular frame; Step 2: For the images in the inner and outer rectangular marking frame areas, use the create_shape_model operator of the Halcon software to create inner and outer rectangular marking frame templates; Step 3: For wafer images that need to be measured for overlay error, first use the outer rectangular mark frame template to find all outer rectangular mark frames that match the template match and record the position information; then use the inner rectangular mark frame template to find the inner rectangular mark frame that matches the template match within the image range of the outer rectangular mark frame that matches the template match, record the position information, and discard the outer rectangular mark frame that does not have an inner rectangular mark frame inside; Step 4: For the inner and outer rectangular marking frame matching results after step 3, a rectangular frame is generated on the wafer image that needs to be overlay error measured according to the relative position relationship between the rectangular frame defined in step 1 and the matching result. , , , for the rectangular box , With rectangular frame , Gaussian line segment detection is performed on the area between to obtain the prediction results of the inner and outer rectangular marks, and the offset between the center coordinates of the inner and outer rectangular mark prediction results is calculated to complete the overlay error measurement.

2. A wafer overlay error measurement method based on Halcon as claimed in claim 1, characterized in that: For a wafer image, a two-dimensional rectangular coordinate system XOY is established with its upper left corner vertex as the origin O, the horizontal direction as the X-axis, and the vertical direction as the Y-axis. The coordinates of the upper left vertex and the lower right vertex of the rectangular box in the two-dimensional rectangular coordinate system XOY are recorded as the position information of the rectangular box.

3. A wafer overlay error measurement method based on Halcon as claimed in claim 1, characterized in that: Perform image enhancement on the wafer image, first converting it from RGB image to grayscale image Then, a filter with a size of 3×3 and an enhancement factor of 1 is used to filter the grayscale image. Perform processing enhancement.

4. A wafer overlay error measurement method based on Halcon as claimed in claim 1 or 3, characterized in that: The grayscale image for: ； Among them, R, G, and B represent the R, G, and B channel values ​​of the original wafer image, respectively.

5. The wafer overlay error measurement method based on Halcon as claimed in claim 1, characterized in that: The rectangular frame 1 contains only one complete overlay error mark, the rectangular frame 2 contains only the inner rectangular mark of the overlay error mark, and the rectangular frame 3 is located within the outer contour of the inner rectangular mark of the overlay error mark.

6. A wafer overlay error measurement method based on Halcon as claimed in claim 1 or 5, characterized in that: The area_center operator of the Halcon software is used to calculate the coordinates of the center points of rectangular frames 1, 2, and 3 respectively, and then the difference operator is used to calculate the area between rectangular frames 1, 2 and rectangular frames 2, 3 respectively to obtain the outer rectangular marked frame area and the inner rectangular marked frame area; then the reduce_domain operator is used to crop the wafer image to obtain the image of the outer rectangular marked frame area and the image of the inner rectangular marked frame area.

7. The wafer overlay error measurement method based on Halcon as claimed in claim 1, characterized in that: Use the find_shape_model operator to find the inner and outer rectangular marking boxes that match the template from the enhanced wafer image.

8. A wafer overlay error measurement method based on Halcon as claimed in claim 1 or 7, characterized in that: Only matches with a confidence level higher than 0.9 are retained.

9. The wafer overlay error measurement method based on Halcon as claimed in claim 1, characterized in that: According to the position information of the matching result of rectangular frame 1 and the outer rectangular marking frame, as well as the position information of the matching result of the outer rectangular marking frame, a position transformation matrix 2D1 from rectangular frame 1 to the outer rectangular marking frame is generated, and a rectangular frame having the same shape as rectangular frame 1 and containing the corresponding matching result of the outer rectangular marking frame is generated using the position transformation matrix 2D1. ; According to the position information of the matching result of rectangular frame 2 and the inner rectangular marking frame, a position transformation matrix 2D2 from rectangular frame 2 to the inner rectangular marking frame is generated, and a rectangular frame with the same shape as rectangular frame 2 and containing the corresponding inner rectangular marking frame matching result is generated using the position transformation matrix 2D2. ; According to the position information of rectangular frame 3 and the position transformation matrix 2D2, a rectangular frame with the same shape as rectangular frame 3 and located within the matching result of the inner rectangular marking frame is generated. .

10. A wafer overlay error measurement method based on Halcon as claimed in claim 1 or 9, characterized in that: Calculate the rectangle , In the area between, use the operator lines_gauss to perform Gaussian line segment detection to obtain the outline of the outer rectangular marking frame; calculate the rectangular frame , In the area between, use the operator lines_gauss to perform Gaussian line segment detection to obtain the contour of the inner rectangular marking box; Perform minimum rectangle fitting on the outer rectangular marking frame contour and the inner rectangular marking frame contour to obtain the outer rectangular marking prediction result and the inner rectangular marking prediction result. Use the area_center operator to obtain the center coordinates of the inner and outer rectangular marking prediction results, and then calculate the offset results of the center coordinates of the inner and outer rectangular marking prediction results in the X-axis direction and the Y-axis direction.

Images