Wafer alignment method, wafer alignment device and readable storage medium

Abstract

The invention discloses a wafer alignment method, a wafer alignment device and a readable storage medium, which can be used in the technical field of semiconductors, and the method comprises the following steps: obtaining a line scanning image of a to-be-aligned wafer; identifying the crystal grain position in the line scanning image, and determining the deflection angle of the crystal grain; rotating the wafer to be aligned based on the deflection angle; obtaining a rotation image corresponding to the rotated to-be-aligned wafer; and aligning the to-be-aligned wafer based on the rotation image. Therefore, the deviation angle is determined by automatically identifying the positions of the crystal grains on the wafer, the coordinate deviation amount is obtained after rotation correction for wafer alignment, the whole process does not depend on a wafer alignment pattern, and the accurate wafer alignment efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of semiconductor technology, and in particular to a wafer alignment method, a wafer alignment device and a readable storage medium. Background Art

[0002] Wafer alignment is a key technology in the semiconductor chip manufacturing process. Its alignment efficiency, accuracy and stability have a significant impact on mass production as well as chip quality and performance.

[0003] Existing wafer alignment methods usually rely on pre-set wafer alignment patterns, which are used for positioning and calibration before processing or testing. Different wafers generally require different alignment patterns, so engineers need to spend more time creating template information during the process, resulting in low efficiency in wafer precision alignment.

[0004] Therefore, how to improve the efficiency of wafer precision alignment is a problem that technical personnel in this field need to solve urgently. Summary of the invention

[0005] Based on the above problems, the present application provides a wafer alignment method, a wafer alignment device and a readable storage medium, which automatically identify the position of the grains on the wafer and then determine the offset angle, and then obtain the coordinate offset after rotation to perform wafer alignment, thereby improving the efficiency of precise wafer alignment.

[0006] In a first aspect, an embodiment of the present application provides a wafer alignment method, comprising: obtaining a line-scanning image of a wafer to be aligned; identifying a grain position in the line-scanning image, and determining a deflection angle of the grain; rotating the wafer to be aligned based on the deflection angle; obtaining a rotated image corresponding to the rotated wafer to be aligned; and aligning the wafer to be aligned based on the rotated image.

[0007] Optionally, aligning the wafer to be aligned based on the rotated image includes: acquiring the measured coordinates of the grain in the rotated image, and comparing them with the standard coordinates of the grain to obtain a coordinate offset; and aligning the rotated wafer to be aligned based on the coordinate offset.

[0008] Optionally, identifying the position of grains in the line scan image includes: acquiring a template image of the grains; performing feature comparison between the line scan image and the template image to determine the grain area corresponding to each grain in the line scan image; and determining the grain position of each grain based on the grain area.

[0009] Optionally, identifying the position of grains in the line scan image includes: obtaining a target grayscale histogram corresponding to the line scan image; defining a cutting threshold based on the target grayscale histogram; the cutting threshold includes a lower threshold vector and an upper threshold vector; in combination with the cutting threshold, segmenting the line scan image according to a segmentation condition and generating a segmented image; determining all connected domains corresponding to the line scan image based on the segmented image; and determining the position of the grains based on all connected domains.

[0010] Optionally, obtaining the target grayscale histogram corresponding to the line scan image includes: based on the line scan image, obtaining the corresponding initial grayscale histogram with the frequency corresponding to each gray level as the statistical target; performing Gaussian smoothing on the initial grayscale histogram to obtain the target grayscale histogram corresponding to the line scan image.

[0011] Optionally, defining the cutting threshold based on the target grayscale histogram includes: constructing an extreme point set based on the target grayscale histogram; and defining the cutting threshold based on the extreme point set.

[0012] Optionally, constructing an extreme point set based on the target grayscale histogram includes: analyzing the target grayscale histogram to construct a maximum value set and a minimum value set; merging and sorting the maximum value set and the minimum value set to obtain an extreme point set; and arranging each extreme value in the extreme point set in ascending order.

[0013] Optionally, the method further includes: performing an opening operation and a closing operation on the segmented image respectively to obtain a first operation result and a second operation result; determining all connected domains corresponding to the line scan image based on the segmented image includes: determining all connected domains corresponding to the line scan image based on the segmented image, the first operation result and the second operation result; the expression corresponding to all connected domains is:

[0014] CX=ConnectedComponents(X), X∈{S, K, C};

[0015] Wherein, CX is the entire connected domain, S is the segmented image, K is the first operation result, and C is the second operation result.

[0016] Optionally, determining the position of the grain based on all connected domains includes: screening all connected domains based on a standard size of the grain to determine a target connected domain corresponding to the grain; and determining the position of the grain based on the target connected domain.

[0017] Optionally, determining the position of the grain based on the target connected domain includes: determining the center position coordinates of the target connected domain, and using the center position coordinates as the position of the grain.

[0018] Optionally, determining the deflection angle of the grain includes: performing straight line fitting processing on the center points of all grains on the same row based on the least squares method to form an offset straight line; and comparing the offset straight line with a preset standard straight line to obtain the deflection angle of the grain.

[0019] Optionally, the obtaining of the rotated image corresponding to the wafer to be aligned after rotation includes: performing affine transformation processing on the line scan image in combination with the deflection angle to obtain the rotated image corresponding to the wafer to be aligned after rotation.

[0020] Optionally, the line scan image is an image taken by a TDI line array camera.

[0021] In the second aspect, an embodiment of the present application provides a wafer alignment device, comprising: a first acquisition module, used to acquire a line-scan image of the wafer to be aligned; an identification module, used to identify the position of the grain in the line-scan image and determine the deflection angle of the grain; a rotation module, used to rotate the wafer to be aligned based on the deflection angle; a second acquisition module, used to acquire the rotated image corresponding to the wafer to be aligned after rotation; and an alignment module, used to align the wafer to be aligned based on the rotated image.

[0022] Optionally, the alignment module includes: a comparison module and an alignment sub-module; the comparison module is used to obtain the measured coordinates of the grain in the rotated image, and compare them with the standard coordinates of the grain to obtain a coordinate offset; the alignment sub-module is used to align the rotated wafer based on the coordinate offset.

[0023] Optionally, the identification module includes: a third acquisition module, a definition module, a segmentation module, a first determination module and a second determination module; the third acquisition module is used to acquire a target grayscale histogram corresponding to the line scan image; the definition module is used to define a cutting threshold based on the target grayscale histogram; the cutting threshold includes a lower threshold vector and an upper threshold vector; the segmentation module is used to segment the line scan image according to the segmentation conditions in combination with the cutting threshold, and generate a segmented image; the first determination module is used to determine all connected domains corresponding to the line scan image based on the segmented image; the second determination module is used to determine the position of the grain based on the all connected domains.

[0024] In a third aspect, an embodiment of the present application provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the wafer alignment method as described above are implemented.

[0025] It can be seen from the above technical solutions that compared with the prior art, the present application has the following advantages:

[0026] The present application first obtains a line scan image of the wafer to be aligned, identifies the position of the grain in the line scan image, and determines the deflection angle of the grain. Then, the wafer to be aligned is rotated based on the deflection angle, and the rotated image corresponding to the rotated wafer to be aligned is obtained. Finally, the wafer to be aligned is aligned based on the rotated image. In this way, the position of the grain on the wafer is automatically identified and the offset angle is determined. After rotation, the coordinate offset is obtained to align the wafer. The whole process does not rely on the wafer alignment pattern, which improves the efficiency of precise wafer alignment. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A flowchart of a wafer alignment method provided in an embodiment of the present application;

[0028] Figure 2 A schematic diagram of a TDI line scan image provided in an embodiment of the present application;

[0029] Figure 3 A schematic diagram of determining a grain deflection angle provided in an embodiment of the present application;

[0030] Figure 4 A schematic diagram of a rotation image provided in an embodiment of the present application;

[0031] Figure 5 A schematic diagram of a coordinate offset structure provided in an embodiment of the present application;

[0032] Figure 6 A flow chart of a method for identifying grain positions in a TDI line scan image provided in an embodiment of the present application;

[0033] Figure 7 A schematic diagram of an input image provided in an embodiment of the present application;

[0034] Figure 8 A schematic diagram of an initial grayscale histogram provided in an embodiment of the present application;

[0035] Fig. 9 A schematic diagram of a target grayscale histogram provided in an embodiment of the present application;

[0036] Fig.10 A schematic diagram of a cut image provided in an embodiment of the present application;

[0037] Fig.11 A schematic structural diagram of a wafer alignment device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0038] As mentioned above, the existing wafer alignment method has the problem of low alignment efficiency. Specifically, the existing wafer alignment method relies on the wafer alignment pattern, so engineers need to create the template information of the alignment pattern in advance before the wafer is processed. However, different wafers generally require different alignment patterns. This operation method increases the time for engineers to establish template information, and there is a delay in information exchange, resulting in low efficiency in wafer precision alignment.

[0039] To solve the above problems, an embodiment of the present application provides a wafer alignment method, which first obtains a line scan image of the wafer to be aligned, identifies the position of the grain in the line scan image, and determines the deflection angle of the grain. Then, the wafer to be aligned is rotated based on the deflection angle, and the rotated image corresponding to the rotated wafer to be aligned is obtained. Finally, the wafer to be aligned is aligned based on the rotated image.

[0040] In this way, the offset angle is determined by automatically identifying the position of the grain on the wafer, and the coordinate offset is obtained after rotation to perform wafer alignment. The entire process does not rely on the wafer alignment pattern, thereby improving the efficiency of precise wafer alignment.

[0041] It should be noted that the wafer alignment method, wafer alignment device and readable storage medium provided in the present application can be applied to the field of semiconductor technology. The above is only an example and does not limit the application field of the wafer alignment method, wafer alignment device and readable storage medium provided in the present application.

[0042] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0043] Figure 1 A flowchart of a wafer alignment method provided in an embodiment of the present application. Figure 1 As shown, a wafer alignment method provided in an embodiment of the present application may include:

[0044] S101: Acquire a scanning image of the wafer to be aligned.

[0045] The line scan image may be an image taken by a TDI line array camera.

[0046] In practical applications, the pre-aligned wafer needs to be accurately positioned on the wafer stage before it can be processed to ensure that subsequent processing and testing can be carried out accurately. Figure 2 A schematic diagram of a TDI line scan image provided in an embodiment of the present application. Figure 2 As shown, firstly, a time delay integration (TDI) linear array camera is used to quickly scan the wafer to be aligned and obtain a scanned image. It can be understood that since the wafer stage carrying the wafer can perform linear motion in a two-dimensional direction relative to the camera, the TDI linear array camera has a linear scanning field of view, so the obtained scanned image will show the tilt of the grain. The TDI linear array camera can quickly execute the row and column scanning requirements of the wafer, thereby obtaining the corresponding row scan image, that is, the TDI row scan image of the wafer to be aligned.

[0047] S102: Identify the position of the crystal grain in the line scan image, and determine the deflection angle of the crystal grain.

[0048] In practical applications, if the wafer alignment is accurate, the sorting direction of the grains should be consistent with the horizontal field of view direction of the TDI linear array camera. Therefore, the embodiment of the present application can use the horizontal field of view direction of the TDI linear array camera as the standard direction for grain sorting, and of course, the cutting path direction of the wafer can also be used as the standard direction for grain sorting. Since the offset morphology of each grain is displayed in the TDI line scan image, the deflection angle of the grain can be determined based on the position of each grain by identifying the grain position in the TDI line scan image.

[0049] In addition, since there are different ways to determine the grain deflection angle, the embodiments of the present application can illustrate a possible determination method.

[0050] In one case, determining the deflection angle of the grain includes: fitting the center points of all the grains on the same row with straight lines based on the least squares method to form an offset straight line; comparing the offset straight line with a preset standard straight line (a straight line corresponding to the standard direction) to obtain the deflection angle of the grain.

[0051] Figure 3 A schematic diagram of determining a grain deflection angle provided in an embodiment of the present application. Figure 3 As shown in FIG. 1 , after determining the position of the grain in the TDI line scan image, multiple points are formed with the center of the grain, and the least squares method is used to fit the straight line to connect the center points of all the grains on the same row, thereby forming an offset straight line. At this time, a deviation angle is formed between the offset straight line and the preset standard straight line (horizontal straight line), that is, the deflection angle θ of the grain.

[0052] S103: Rotate the wafer to be aligned based on the deflection angle.

[0053] In practical applications, after the deflection angle θ is determined, the wafer stage can be controlled to drive the wafer to be aligned to rotate synchronously by the angle θ, thereby achieving the rotation of the wafer to be aligned.

[0054] S104: Obtaining a rotated image corresponding to the rotated wafer to be aligned.

[0055] In practical applications, after the wafer to be aligned is rotated, a rotated image corresponding to the rotated wafer to be aligned can be obtained by reacquiring a line scan image, and then the offset of the wafer to be aligned on the x-axis and the y-axis can be determined.

[0056] In addition, since there are different ways to obtain the rotated image, the embodiment of the present application can illustrate a possible acquisition method.

[0057] In one case, S104: obtaining a rotation image corresponding to the rotated wafer to be aligned may specifically include:

[0058] The line scan image is subjected to affine transformation processing in combination with the deflection angle to obtain a rotated image corresponding to the wafer to be aligned after rotation.

[0059] In practical applications, in order to improve efficiency, the embodiment of the present application can substitute the above-determined deflection angle θ into the TDI line scan image corresponding to the wafer to be aligned that has not been rotated and obtained initially, and then use the affine matrix to process the TDI line scan image through affine transformation and obtain the rotated image corresponding to the wafer to be aligned after rotation. Figure 4 A schematic diagram of a rotation image provided in an embodiment of the present application. Figure 4 As shown, all grains in the normalized image are not tilted.

[0060] S105: Align the wafer to be aligned based on the rotation image.

[0061] In actual applications, the straight line formed by connecting the grains in the same row in the rotation image is parallel to the X-axis, and the straight line formed by connecting the grains in the same column is parallel to the Y-axis. However, the position of each grain is generally not the standard position, so it is necessary to compare the measured position of the grain with the standard position based on the rotation image, and then move the wafer to be aligned to achieve wafer alignment.

[0062] In addition, since the methods for aligning the wafers to be aligned are not the same, the embodiment of the present application can illustrate a possible alignment method.

[0063] In one case, S105: aligning the wafer to be aligned based on the rotated image, specifically including: obtaining the measured coordinates of the grain in the rotated image, and comparing them with the standard coordinates of the grain to obtain the coordinate offset; aligning the rotated wafer to be aligned based on the coordinate offset.

[0064] In practical applications, the measured coordinates of any grain can be obtained and compared with the fixed coordinate position (standard coordinate) formed during photolithography to obtain the coordinate offset. The measured coordinates of all grains in a row can also be averaged and then compared with the average value of the standard coordinates of all grains in this row to obtain the coordinate offset. The latter method is more accurate than the former, so the latter method is generally used to calculate the coordinate offset. Figure 5 A schematic diagram of a coordinate offset structure provided in an embodiment of the present application. Figure 5 As shown in the figure, the center point of the grain is used as the coordinate position of the grain, the lower pattern is the rotated image corresponding to the wafer to be aligned after rotation, and the upper pattern is the grain image formed during lithography. In this way, the black coordinate system can be used to represent the measured coordinates corresponding to the grain in the rotated image, and the red coordinate system can be used to represent the standard coordinates of the grain. The coordinate offset (XY Offset) can be obtained by comparing the two. Then, the wafer stage can be used to synchronously drive the rotated wafer to be aligned to move on the X-axis and Y-axis based on the coordinate offset, thereby completing the wafer precision alignment process.

[0065] In summary, the present application first obtains a line scan image of the wafer to be aligned, identifies the position of the grain in the line scan image, and determines the deflection angle of the grain. Then, the wafer to be aligned is rotated based on the deflection angle, and the rotated image corresponding to the rotated wafer to be aligned is obtained. Finally, the wafer to be aligned is aligned based on the rotated image. In this way, the position of the grain on the wafer is automatically identified and the offset angle is determined. After rotation, the coordinate offset is obtained to align the wafer. The whole process does not rely on the wafer alignment pattern, which improves the efficiency of precise wafer alignment.

[0066] In addition, since there are different ways to identify the location of a die, the application embodiment may describe a possible identification method.

[0067] In one case, identifying the position of a grain in a line scan image includes: acquiring a template image of the grain; performing feature comparison between the line scan image and the template image to determine the grain area corresponding to each grain in the line scan image; and determining the grain position of each grain based on the grain area.

[0068] In practical applications, the template image of the grain contains features such as the size and shape of the grain. The embodiment of the present application can pre-store the template image corresponding to the grain. When identifying the position of the grain, the features of the template image can be compared with the regional features on the line scan image based on feature recognition technology, so as to determine the region corresponding to each grain in the line scan image, and then use the center of the region corresponding to each grain as the grain position of the grain.

[0069] In addition, since the methods for identifying the position of a die are not the same, the application embodiment may describe another possible identification method.

[0070] Figure 6 A flow chart of a method for identifying grain positions in a TDI line scan image provided in an embodiment of the present application. Figure 6 As shown, the method may include:

[0071] S601: Obtain a target grayscale histogram corresponding to the line scan image.

[0072] In practical applications, it is first necessary to count the frequency of each gray level in the input image (TDI line scan image of the wafer to be aligned) to obtain the corresponding target gray level histogram.

[0073] In addition, since there are different ways to obtain the target grayscale histogram, the embodiment of the present application can illustrate a possible acquisition method.

[0074] In one case, a target grayscale histogram corresponding to a line scan image is obtained, including: based on the line scan image, obtaining a corresponding initial grayscale histogram with the frequency corresponding to each gray level as a statistical target; performing Gaussian smoothing on the initial grayscale histogram to obtain a target grayscale histogram corresponding to the line scan image.

[0075] Figure 7 A schematic diagram of an input image provided in an embodiment of the present application. Figure 7 As shown, assume that the input image is I(x, y) and the grayscale value range is [0, 255]. Figure 8 A schematic diagram of an initial grayscale histogram provided in an embodiment of the present application. Figure 8 As shown, the X axis is the gray level and the Y axis is the frequency. The initial gray level histogram H(k) is generated with the frequency of each gray level as the target, where k represents the gray level interval between 0 and 255. Fig. 9 A schematic diagram of a target grayscale histogram provided in an embodiment of the present application. Fig. 9 As shown, in order to make the distribution of the histogram smoother, the embodiment of the present application can perform Gaussian smoothing on the initial grayscale histogram obtained by statistics to remove its noise. The common Gaussian function G(x) can be used during Gaussian processing to represent the value of the Gaussian function at position x. The target grayscale histogram H after smoothing smooth The expression of (k) is as follows:

[0076]

[0077] Among them, G(kj) is the Gaussian kernel. Compared with the initial grayscale histogram, the target grayscale histogram is smoother.

[0078] S602: Define a cutting threshold based on a target grayscale histogram; the cutting threshold includes a lower threshold vector and an upper threshold vector.

[0079] In practical applications, the grayscale values ​​of the same feature area on the wafer generally follow the same pattern. For example, the grayscale values ​​of each pixel on the cutting path are generally the same, and the grayscale values ​​of the pixels at the same position on each die are generally the same. In this regard, the die position can be determined based on the grayscale value distribution pattern, and then the cutting threshold, i.e., the lower threshold vector and the upper threshold vector, can be defined.

[0080] In addition, since there are different ways to define the cutting threshold, the embodiment of the present application can illustrate a possible definition method.

[0081] In one case, defining a cutting threshold based on a target grayscale histogram specifically includes: constructing an extreme point set based on the target grayscale histogram; and defining a cutting threshold based on the extreme point set.

[0082] In practical applications, the extreme points in the target grayscale histogram are determined by taking derivatives and finding zero crossing points, and the extreme points are used to construct an extreme point set to define the cutting threshold.

[0083] In addition, since there are different ways to construct an extreme point set, the embodiment of the present application can illustrate a possible construction method.

[0084] In one case, an extreme point set is constructed based on a target grayscale histogram, including: analyzing the target grayscale histogram to construct a maximum value set and a minimum value set; merging and sorting the maximum value set and the minimum value set to obtain an extreme point set; and arranging each extreme value in the extreme point set in ascending order.

[0085] In practical applications, H smooth(k) Solve the first derivative, and use the obtained maximum and minimum values to construct a maximum value set M = {m1, m2, …, mp} and a minimum value set N = {n1, n2, …, nq} respectively. Here, p and q are the numbers of maximum and minimum values respectively. Then merge the two sets M and N into a set R, R = M ∪ N = {m1, m2, …, mp, n1, n2, …, nq}. Since the value of the maximum is not necessarily greater than the value of the minimum, for the convenience of subsequent processing, the set R needs to be sorted in ascending order, and the set R′ = {r1, r2, …, rn} is obtained. Here, n = p + q, and r1 < r2 < … < rn. For example, assume that the calculation result of the minimum value set is LocalMin = [0, 76, 158], and the calculation result of the maximum value set is LocalMax = [65, 92, 255]. Then when merging the two sets M and N into a set R, the calculation result R = [LocalMin, LocalMax] = [0, 76, 158, 65, 92, 255]. Finally, after sorting in ascending order, the sorted result should be Sorted = [0, 65, 76, 92, 158, 255]. In this way, according to the sorted set R′, define the cutting threshold, and generate the threshold lower limit vector and the threshold upper limit vector. Specifically, the threshold lower limit vector L is L = {0} ∪ {r1 + 1, r2 + 1, …, rn + 1}, and the threshold upper limit vector U is U = R′ ∪ {255}. Continuing with the above example, then the calculation result of the threshold lower limit vector threshMin = [0, 66, 77, 93, 159], and the calculation result of the threshold upper limit vector threshMax = [65, 76, 92, 158, 255].

[0086] S603: Combine the cutting threshold, and segment the line-scanned image according to the segmentation condition, and generate a segmented image.

[0087] Fig.10 A schematic diagram of an image cutting provided by an embodiment of the present application. Combine Fig.10 As shown, segment the input image by combining the above threshold lower limit vector and threshold upper limit vector to obtain a segmented image. Specifically, if not taking the above calculation results as an example, the threshold lower limit vector and the threshold upper limit vector can be set as L = {l0, l1, …, lm} and U = {u0, u1, …, un} respectively. Here, li and uj respectively represent the specific values in the threshold lower limit vector and the upper limit vector. Then for each pixel I(x, y), a segmentation condition can be defined, and the expression is as shown in the following formula:

[0088]

[0089] Then generate a segmented image according to this segmentation condition, and the expression is as shown in the following formula:

[0090]

[0091] In this way, a segmented image is obtained after processing each pixel.

[0092] S604: Determine all connected regions corresponding to the line scan image based on the segmented image.

[0093] In practical applications, the coordinate positions of the grains can be determined by the connected domains. Therefore, after the segmented image is determined, all the connected domains corresponding to the input image can be calculated according to the connected domain calculation function.

[0094] In addition, due to different processing methods for segmenting images, the corresponding methods for determining connected domains are also different, so the embodiment of the present application can illustrate a possible determination method.

[0095] In one case, the method of the present application also includes: performing opening and closing operations on the segmented image respectively to obtain a first operation result and a second operation result; determining all connected domains corresponding to the line scan image based on the segmented image includes: determining all connected domains corresponding to the line scan image based on the segmented image, the first operation result and the second operation result.

[0096] The expression corresponding to the entire connected domain is:

[0097] CX=ConnectedComponents(X), X∈{S, K, C};

[0098] Where CX is the total connected domain, S is the segmented image, K is the result of the first operation, and C is the result of the second operation.

[0099] In practical applications, performing opening and closing operations on the segmented image can segment the connected grains on the segmented image to form segmentation paths, which is more conducive to the subsequent calculation of connected domains. The opening operation K(S) and closing operation C(S) commonly used in image processing can respectively obtain the opening operation result (first operation result) and the closing operation result (second operation result). Then, combined with the above results S, K and C, let X∈{S, K, C} and calculate the connected domain. The expression of the connected domain is shown as follows:

[0100] CX=ConnectedComponents(X);

[0101] In this way, all connected domains corresponding to the TDI line scan image can be obtained.

[0102] S605: Determine the location of the grain based on all connected domains.

[0103] In practical applications, the connected domain corresponding to the grain can be determined from all connected domains in combination with the standard size corresponding to the grain, and then the position of the grain can be determined.

[0104] In addition, since there are different ways to determine the location of a grain, the embodiments of the present application can illustrate a possible determination method.

[0105] In one case, determining the position of the grain based on all connected domains includes: screening all connected domains based on a standard size of the grain to determine a target connected domain corresponding to the grain; and determining the position of the grain based on the target connected domain.

[0106] In practical applications, the grain sizes on different wafers have fixed standards. Therefore, all the calculated connected domains can be screened by the standard size of the grains to determine the target connected domain corresponding to the grains. Specifically, let the width and height of each connected domain be W i and H i , then there is (W i ,H i ) = GetDimensions(CX). Assume that the width of the input grain standard size is W in Height H in , with a tolerance range of Δ. Then, using this as the screening condition, the screening expression for connected domains is as follows:

[0107]

[0108] In this way, a connected domain with a standard size, namely a target connected domain, can be screened out from the input image. Then, the position of the grain is determined by the target connected domain. Specifically, the position of the target connected domain (the connected domain with a standard size) is the position of the grain.

[0109] In addition, since there are different ways to determine the location of a grain, the embodiment of the present application may describe another possible determination method.

[0110] In one embodiment, the location of the grain is determined based on the target connected domain, including:

[0111] The center position coordinates of the target connected domain are determined, and the center position coordinates are used as the position of the grain.

[0112] In practical applications, the calculated position of the target connected domain is the position of the grain. To facilitate selection and subsequent calculations, the embodiment of the present application uses the center position coordinates of the target connected domain as the coordinate position of the grain.

[0113] After comparing the coordinates of the grain with the standard coordinates, the offset in the X-axis and Y-axis directions can be determined. After obtaining the coordinate offset, the wafer to be aligned after rotation is synchronously driven to move on the X-axis and Y-axis based on the coordinate offset through the wafer stage, thereby completing the wafer's precise alignment process.

[0114] In summary, in the process of identifying the position of the grain, the target grayscale histogram corresponding to the line scan image is first obtained, and the cutting threshold is defined based on the target grayscale histogram. Among them, the cutting threshold includes a lower threshold vector and an upper threshold vector. Then, combined with the cutting threshold, the line scan image is segmented according to the segmentation condition, and a segmented image is generated. Finally, all connected domains corresponding to the line scan image are determined based on the segmented image, and the position of the grain is determined based on all connected domains. In this way, the position of the grain in the input image is determined by calculating the connected domain, which improves the accuracy of identifying the position of the grain.

[0115] Based on the wafer alignment method provided in the above embodiment, the embodiment of the present application further provides a wafer alignment device. The wafer alignment device is described below in conjunction with the embodiments and drawings.

[0116] Fig.11 A schematic diagram of the structure of a wafer alignment device provided in an embodiment of the present application. Fig.11 As shown, the wafer alignment device 1100 provided in the embodiment of the present application includes:

[0117] The first acquisition module 1101 is used to acquire a scanning image of the wafer to be aligned.

[0118] The identification module 1102 is used to identify the position of the grain in the line scan image and determine the deflection angle of the grain.

[0119] The rotation module 1103 is used to rotate the wafer to be aligned based on the deflection angle.

[0120] The second acquisition module 1104 is used to acquire a rotated image corresponding to the wafer to be aligned after rotation.

[0121] The alignment module 1105 is used to align the wafer to be aligned based on the rotation image.

[0122] The line scan image is an image taken by a TDI line array camera.

[0123] As an implementation manner, with respect to how to align the wafer to be aligned based on the rotation image, the alignment module 1105 includes: a comparison module and an alignment sub-module.

[0124] The comparison module is used to obtain the measured coordinates of the grain in the rotation image, and compare them with the standard coordinates of the grain to obtain the coordinate offset.

[0125] The alignment submodule is used to align the wafer to be aligned after rotation based on the coordinate offset.

[0126] As an implementation method, regarding how to identify the position of a die, the identification module 1102 includes: a third acquisition module, a definition module, a segmentation module, a first determination module, and a second determination module;

[0127] A third acquisition module is used to acquire a target grayscale histogram corresponding to the line scan image;

[0128] A definition module is used to define a cutting threshold based on a target grayscale histogram; the cutting threshold includes a lower threshold vector and an upper threshold vector.

[0129] The segmentation module is used to segment the line scan image according to the segmentation conditions in combination with the cutting threshold, and generate a segmented image.

[0130] The first determination module is used to determine all connected domains corresponding to the line scan image based on the segmented image.

[0131] The second determination module is used to determine the position of the grain based on the entire connected domain.

[0132] As an implementation method, regarding how to obtain a target grayscale histogram, the third acquisition module is specifically used to: based on the line scan image, obtain the corresponding initial grayscale histogram with the frequency corresponding to each grayscale level as the statistical target. Gaussian smoothing is performed on the initial grayscale histogram to obtain a target grayscale histogram corresponding to the line scan image.

[0133] As an implementation method, regarding how to define the cutting threshold, the above-mentioned definition module specifically includes: a construction module and a definition sub-module; the construction module is used to construct an extreme point set based on the target grayscale histogram; the definition sub-module is used to define the cutting threshold based on the extreme point set.

[0134] As an implementation method, regarding how to construct an extreme point set, the above-mentioned construction module is specifically used to: analyze the target grayscale histogram to construct a maximum value set and a minimum value set; merge and sort the maximum value set and the minimum value set to obtain an extreme point set; and arrange the extreme values ​​in the extreme point set in ascending order.

[0135] As a real-time method, regarding how to process the segmented image, the wafer alignment device 1100 further includes: a processing module; the processing module is used to perform an opening operation and a closing operation on the segmented image respectively to obtain a first operation result and a second operation result.

[0136] Correspondingly, the first determination module is specifically used to: determine all connected domains corresponding to the line scan image based on the segmented image, the first operation result and the second operation result;

[0137] The expression corresponding to all connected domains is:

[0138] CX=ConnectedComponents(X), X∈{S, K, C};

[0139] Wherein, CX is the entire connected domain, S is the segmented image, K is the first operation result, and C is the second operation result.

[0140] As an implementation method, regarding how to determine the position of the grain, the above-mentioned second determination module specifically includes: a screening module and a second determination submodule; the screening module is used to screen all the connected domains based on the standard size of the grain to determine the target connected domain corresponding to the grain; the second determination submodule is used to determine the position of the grain based on the target connected domain.

[0141] As an implementation manner, with respect to how to determine the position of a grain, the second determination submodule is specifically used to: determine the center position coordinates of the target connected domain, and use the center position coordinates as the position of the grain.

[0142] As an implementation method, regarding how to determine the deflection angle of the grain, the above-mentioned identification module 1102 also includes: a third determination module; the third determination module is used to fit the center points of all the grains on the same row with a straight line based on the least squares method to form an offset straight line; the offset straight line is compared with a preset standard straight line to obtain the deflection angle of the grain.

[0143] As an implementation method, regarding how to obtain a rotated image, the second acquisition module 1104 is specifically used to: perform affine transformation processing on the line scan image in combination with the deflection angle to obtain a rotated image corresponding to the wafer to be aligned after rotation.

[0144] As an implementation method, regarding how to identify the position of a die, the identification module 1102 further includes: a fourth acquisition module, a feature comparison module, and a die position determination module;

[0145] A fourth acquisition module, used for acquiring a template image of a grain;

[0146] A feature comparison module, used for performing feature comparison between the line scan image and the template image to determine the grain area corresponding to each grain in the line scan image;

[0147] The grain position determination module is used to determine the grain position of each grain based on the grain area.

[0148] In summary, the present application first obtains a line scan image of the wafer to be aligned, identifies the position of the grain in the line scan image, and determines the deflection angle of the grain. Then, the wafer to be aligned is rotated based on the deflection angle, and the rotated image corresponding to the rotated wafer to be aligned is obtained. Finally, the wafer to be aligned is aligned based on the rotated image. In this way, the position of the grain on the wafer is automatically identified and the offset angle is determined. After rotation, the coordinate offset is obtained to align the wafer. The whole process does not rely on the wafer alignment pattern, which improves the efficiency of precise wafer alignment.

[0149] In addition, the present application also provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the wafer alignment method described above are implemented.

[0150] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A wafer alignment method, characterized in that: The method comprises: Acquire a scanning image of the wafer to be aligned; Identifying the position of a crystal grain in the line scan image and determining the deflection angle of the crystal grain; Rotating the wafer to be aligned based on the deflection angle; Acquire a rotation image corresponding to the wafer to be aligned after rotation; The wafer to be aligned is aligned based on the rotation image.

2. The method according to claim 1, characterized in that The step of aligning the wafer to be aligned based on the rotation image comprises: Acquire the measured coordinates of the grain in the rotation image, and compare them with the standard coordinates of the grain to obtain the coordinate offset; The rotated wafer to be aligned is aligned based on the coordinate offset.

3. The method according to claim 1, characterized in that The step of identifying the position of a grain in the line scan image comprises: Acquire a template image of the grain; Performing feature comparison between the line scan image and the template image to determine the grain area corresponding to each grain in the line scan image; A die position of each die is determined based on the die area.

4. The method according to claim 1, characterized in that The step of identifying the position of a grain in the line scan image comprises: Acquire a target grayscale histogram corresponding to the line scan image; A cutting threshold is defined based on the target grayscale histogram; the cutting threshold includes a lower threshold vector and an upper threshold vector; In combination with the cutting threshold, the line scan image is segmented according to the segmentation condition, and a segmented image is generated; Determine all connected domains corresponding to the line scan image based on the segmented image; The positions of the grains are determined based on the entire connected domains.

5. The method according to claim 4, characterized in that The step of acquiring a target grayscale histogram corresponding to the line scan image comprises: Based on the line scan image, taking the frequency corresponding to each gray level as a statistical target, obtaining a corresponding initial gray histogram; The initial grayscale histogram is subjected to Gaussian smoothing to obtain a target grayscale histogram corresponding to the line scan image.

6. The method according to claim 4, characterized in that Defining a cutting threshold based on the target grayscale histogram includes: Constructing an extreme point set based on the target grayscale histogram; A cutting threshold is defined based on the extreme point set.

7. The method according to claim 6, characterized in that The step of constructing an extreme point set based on the target grayscale histogram includes: Analyze the target grayscale histogram to construct a maximum value set and a minimum value set; The maximum value set and the minimum value set are merged and sorted to obtain an extreme value point set; each extreme value in the extreme value point set is arranged in ascending order.

8. The method according to claim 4, characterized in that The method further comprises: Performing an opening operation and a closing operation on the segmented image respectively to obtain a first operation result and a second operation result; The determining, based on the segmented image, all connected domains corresponding to the line scan image comprises: Determine all connected domains corresponding to the line scan image based on the segmented image, the first operation result, and the second operation result; The expression corresponding to all connected domains is: CX=ConnectedComponents(X), X∈{S, K, C}; Wherein, CX is the entire connected domain, S is the segmented image, K is the first operation result, and C is the second operation result.

9. The method according to claim 4, characterized in that The determining the position of the grain based on the entire connected domains comprises: Screening all the connected domains based on the standard size of the grains to determine the target connected domain corresponding to the grains; The location of the grain is determined based on the target connected domain.

10. The method according to claim 9, characterized in that The determining the position of the grain based on the target connected domain comprises: The center position coordinates of the target connected domain are determined, and the center position coordinates are used as the position of the grain.

11. The method according to claim 1, characterized in that: The step of determining the deflection angle of the grain comprises: Fitting straight lines is performed on the center points of all grains on the same row based on the least square method to form offset straight lines; The offset straight line is compared with a preset standard straight line to obtain the deflection angle of the grain.

12. The method according to claim 1, characterized in that The step of obtaining a rotated image corresponding to the wafer to be aligned after rotation includes: Affine transformation is performed on the line scan image in combination with the deflection angle to obtain a rotated image corresponding to the wafer to be aligned.

13. The method according to claim 1, characterized in that The line scan image is an image taken by a TDI line array camera.

14. A wafer alignment device, characterized in that: include: A first acquisition module is used to acquire a scanning image of the wafer to be aligned; An identification module, used for identifying the position of a crystal grain in the line scan image and determining the deflection angle of the crystal grain; A rotation module, used for rotating the wafer to be aligned based on the deflection angle; A second acquisition module is used to acquire a rotation image corresponding to the wafer to be aligned after rotation; An alignment module is used to align the wafer to be aligned based on the rotation image.

15. The wafer alignment device according to claim 14, characterized in that: The alignment module includes: a comparison module and an alignment submodule; The comparison module is used to obtain the measured coordinates of the grain in the rotation image, and compare them with the standard coordinates of the grain to obtain the coordinate offset; The alignment submodule is used to align the rotated wafer to be aligned based on the coordinate offset.

16. The wafer alignment device according to claim 14, characterized in that: The identification module includes: a third acquisition module, a definition module, a segmentation module, a first determination module and a second determination module; The third acquisition module is used to acquire a target grayscale histogram corresponding to the line scan image; The definition module is used to define a cutting threshold based on the target grayscale histogram; the cutting threshold includes a lower threshold vector and an upper threshold vector; The segmentation module is used to segment the line scan image according to the segmentation condition in combination with the cutting threshold, and generate a segmented image; The first determination module is used to determine all connected domains corresponding to the line scan image based on the segmented image; The second determination module is used to determine the position of the grain based on the entire connected domain.

17. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the wafer alignment method according to any one of claims 1 to 13 are implemented.

Images