Marker alignment method
By performing grayscale summation and contour interpolation on the marked image, combined with a two-dimensional Gaussian distribution, the problem of insufficient accuracy and stability in existing alignment techniques is solved, achieving more accurate calculation of alignment deviation and improving alignment accuracy and stability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- INST OF OPTICS & ELECTRONICS CHINESE ACAD OF SCI
- Filing Date
- 2022-12-29
- Publication Date
- 2026-06-12
AI Technical Summary
Existing alignment techniques are limited by the resolution of image sensors and the magnification of imaging lens groups. High-precision gratings are difficult to manufacture, resulting in poor alignment stability and a limited measurement range.
By summing the gray levels of the marked images by row or column, the average and standard deviation of the center coordinates of the template and substrate markings are calculated. Combined with gray level thresholding and contour gray level curve interpolation, statistical fusion is performed using a two-dimensional Gaussian distribution to calculate the alignment deviation.
It improves alignment accuracy and stability, expands the measurement range, and maintains the advantages of photometric alignment technology, such as wide measurement range and simple structure.
Smart Images
Figure CN116107177B_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of alignment detection technology, and in particular to a marker alignment method combining multiple algorithms. Background Technology
[0002] In the manufacturing process of integrated circuits, the pattern on the stencil is transferred to the substrate surface through a photochemical reaction during exposure. Therefore, exposure is often one of the most important steps in the entire integrated circuit manufacturing process. Modern integrated circuits generally adopt a three-dimensional multilayer structure, which means that multiple pattern transfers are required; this process is called overlay. During overlay, in order to ensure the functionality of the three-dimensional circuit structure, the layers must be precisely aligned; otherwise, the entire integrated circuit will fail.
[0003] After years of research and development, alignment technology has evolved in multiple directions. Currently, the mainstream alignment methods include photometric alignment and diffraction alignment. Photometric alignment uses an imaging lens group to collect alignment marks pre-processed on the template and substrate, and then calculates the alignment deviation through digital image processing. This method requires simple alignment mark structures and is less affected by process variations, substrate deformation, and particle size, thus offering advantages such as a large measurement range and high stability. However, its accuracy is limited by the image sensor resolution and the magnification of the imaging lens group. Furthermore, the image processing algorithm typically only provides a fixed result, making it highly susceptible to changes in algorithm parameters. In diffraction alignment, when the alignment light source illuminates the gratings processed on the template and substrate, the diffracted light from the two gratings interferes with each other, forming interference light correlated with the relative displacement of the template and silicon wafer. By receiving this interference light through a sensor, the alignment deviation can be calculated. The accuracy of diffraction alignment is limited only by the light source wavelength and the resolution of the measurement system, thus achieving very high precision. However, this method requires high-precision gratings as alignment marks, making mark fabrication difficult and highly susceptible to variations in process technology and substrate deformation, resulting in relatively poor stability. Furthermore, due to the periodicity of the gratings, the measurement range of this method is often relatively small. Summary of the Invention
[0004] To address the aforementioned technical problems, this disclosure provides a template-substrate alignment method, which at least partially solves the technical problems of poor alignment stability caused by limitations in image sensor resolution and magnification of imaging lens groups, as well as the difficulty in high-precision grating processing in existing alignment technologies.
[0005] Based on this, the first aspect of this disclosure provides a template and substrate alignment method, comprising: summing grayscale values of a marked image by row or column to obtain a first projection summation curve for each row and a second projection summation curve for each column; calculating a first average value and a first standard deviation of the center coordinates of the template mark and a second average value and a second standard deviation of the center coordinates of the substrate mark based on the first projection summation curve for each row and the second projection summation curve for each column; extracting the image contour of the marked image and adding a horizontal processing window or a vertical processing window to the image contour; extracting the first contour grayscale curve for each row in the horizontal processing window or the second contour grayscale curve for each column in the vertical processing window; calculating a third average value and a third standard deviation of the center coordinates of the template mark and a fourth average value and a fourth standard deviation of the center coordinates of the substrate mark based on the first contour grayscale curve for each row or the second contour grayscale curve for each column; statistically fusing the first average value, first standard deviation, second average value, second standard deviation, third average value, third standard deviation, fourth average value, and fourth standard deviation to obtain an alignment deviation; and aligning the template and substrate based on the alignment deviation.
[0006] According to embodiments of this disclosure, calculating the first average value and first standard deviation of the template mark center coordinates and the second average value and second standard deviation of the substrate mark center coordinates based on the first projection summation curve corresponding to each row and the second projection summation curve corresponding to each column includes: cutting the first peak value of the first projection summation curve by a first grayscale threshold, and cutting the second peak value of the second projection summation curve by a second grayscale threshold; calculating the first centroid corresponding to each peak value greater than the first grayscale threshold on the first projection summation curve, and calculating the second centroid corresponding to each peak value greater than the second grayscale threshold on the second projection summation curve; calculating the template mark center coordinates and the substrate mark center coordinates based on the first centroid and the second centroid; changing the size of the first grayscale threshold and the second grayscale threshold, and repeating the above operations to obtain multiple template mark center coordinates and substrate mark center coordinates; calculating the first average value and the first standard deviation based on the multiple template mark center coordinates, and calculating the second average value and the second standard deviation based on the multiple substrate mark center coordinates.
[0007] According to embodiments of this disclosure, calculating the center coordinates of the template mark and the center coordinates of the substrate mark based on the first centroid and the second centroid includes: determining the ordinates corresponding to each peak on the first projection summation curve whose peak value is greater than a first grayscale threshold based on the first centroid; determining the abscissas corresponding to each peak on the second projection summation curve whose peak value is greater than a second grayscale threshold based on the second centroid; and calculating the center coordinates of the template mark and the center coordinates of the substrate mark based on the ordinates and abscissas corresponding to each peak. According to embodiments of this disclosure, based on...
[0008]
[0009] Calculate the first or second centroid, where x is the number of rows or columns of pixels in the marked image, S(x) is the sum of gray values of the pixels in the xth row or column, k is the number of the peak on the first projection summation curve whose peak value is greater than the first gray threshold or the number of the peak on the second projection summation curve whose peak value is greater than the second gray threshold, x1 and x2 are the starting and ending points of the kth peak after it is cut, and i is the number of times the first gray threshold is changed or the second gray threshold is changed.
[0010] According to embodiments of this disclosure, the number of peaks on the first projection summation curve with peak values greater than the first grayscale threshold and the number of peaks on the second projection summation curve with peak values greater than the second grayscale threshold are both 3; calculating the center coordinates of the template mark and the center coordinates of the substrate mark based on the ordinate and abscissa of each peak includes: based on
[0011] x p1 (i)=(x g1(i) +x g3(i) ) / 2
[0012] y p1 (i)=(y g1 (i)+y g3(i) ) / 2
[0013] x q1 (i)=x g2(i)
[0014] y q1 (i)=y g2(i)
[0015] Calculate the x-coordinate of the center coordinate of the template mark. p1 (i) and y-coordinate p1 (i) and the x-coordinate of the center coordinate of the substrate mark q1 (i) and y-coordinate q1 (i), p represents the template, q represents the substrate, x g1 (i), x g2 (i), x g3 (i) represents the abscissas of each peak on the second projection summation curve whose peak value is greater than the second grayscale threshold, y g1 (i), y g2 (i), y g3 (i) are the ordinates of each peak on the first projection summation curve whose peak value is greater than the first gray level threshold.
[0016] According to embodiments of this disclosure, calculating the third average value and third standard deviation of the template mark center coordinates and the fourth average value and fourth standard deviation of the substrate mark center coordinates based on the first contour grayscale curve corresponding to each row or the second contour grayscale curve corresponding to each column includes: interpolating the first contour grayscale curve according to a first interpolation threshold or interpolating the second contour grayscale curve according to a second interpolation threshold to obtain the coordinates of the contour boundary formed by the template and the substrate corresponding to the first contour grayscale curve or the coordinates of the second contour grayscale curve; calculating the template mark center coordinates and substrate mark center coordinates corresponding to each row or column based on the coordinates of the contour boundary formed by the template and the substrate corresponding to the first contour grayscale curve or the coordinates of the second contour grayscale curve; and calculating the third average value, third standard deviation, fourth average value, and fourth standard deviation based on the template mark center coordinates and substrate mark center coordinates corresponding to each row or column.
[0017] According to embodiments of this disclosure, the number of grayscale peaks in the first contour grayscale curve corresponding to each row or the second contour grayscale curve corresponding to each column is 6; calculating the center coordinates of the template mark and the center coordinates of the substrate mark corresponding to each row based on the coordinates of the contour boundary formed by the template and the substrate corresponding to the first contour grayscale curve includes: based on
[0018] x p2 (j)=(Cx j，1 +Cx j，2 +Cx j，5 +Cx j，6 ) / 4
[0019] x q2 (j)=(Cx j，3 +Cx j，4 ) / 2
[0020] Calculate the x-coordinate of the template marker center coordinates corresponding to the j-th pixel in the horizontal processing window. p2 (j) and the x-coordinate of the center coordinate of the substrate mark q2 (j), the y-coordinate of the center coordinate of the template mark p2 (j) and the ordinate y of the substrate mark center coordinates q2 (j) represents the ordinate of the j-th row corresponding to the first contour grayscale curve, which can be converted into coordinate values based on the row number of the pixels. Here, v represents the index of the grayscale peak, taking values from 1 to 6, and Cx... j，v Cx represents the x-coordinate of the v-th grayscale peak corresponding to the j-th row pixel after interpolation. j，v =(C jl，v +C jr，v ) / 2;
[0021] Based on the coordinates corresponding to the grayscale curve of the second contour formed by the outline boundary of the template and the substrate, the coordinates of the template mark center and the substrate mark center for each column are calculated, including:
[0022] according to
[0023] y p2 (w)=(Cy w，1 +Cy w，2 +Cy w，5 +Cy w，6 ) / 4
[0024] y q2 (w)=(Cy w，3 +Cy w，4 ) / 2
[0025] Calculate the y-coordinate of the template marker center corresponding to the w-th pixel in the vertical processing window. p2 (w) and the ordinate y of the center coordinates of the substrate mark q2 (w), the x-coordinate of the center coordinate of the template mark p2 (w) and the x-coordinate of the center coordinate of the substrate mark q2 (w) represents the x-coordinate of the w-th column corresponding to the second contour grayscale curve, which can be converted into coordinate values based on the column number of the pixels. Here, v represents the index of the grayscale peak, taking values from 1 to 6, and Cy... w，v Cy represents the interpolated ordinate of the v-th grayscale peak corresponding to the w-th pixel. w，v =(C wl，v +C wr，v ) / 2.
[0026] According to embodiments of this disclosure, statistical fusion of the first average, first standard deviation, second average, second standard deviation, third average, third standard deviation, fourth average, and fourth standard deviation to obtain the alignment deviation includes: calculating the precise coordinates of the template mark center based on a two-dimensional Gaussian distribution using the first average, first standard deviation, third average, and third standard deviation; calculating the precise coordinates of the substrate mark center based on a two-dimensional Gaussian distribution using the second average, second standard deviation, fourth average, and fourth standard deviation; and calculating the alignment deviation based on the precise coordinates of the template mark center and the substrate mark center.
[0027] According to embodiments of this disclosure, based on
[0028]
[0029] Calculate the precise coordinates of the center of the template mark or the precise coordinates of the center of the substrate mark, where [u1(x,y), σ1(x,y)] represents the two-dimensional Gaussian distribution corresponding to the first mean and the first standard deviation or the two-dimensional Gaussian distribution corresponding to the second mean and the second standard deviation, and [u2(x,y), σ2(x,y)] represents the two-dimensional Gaussian distribution corresponding to the third mean and the third standard deviation or the two-dimensional Gaussian distribution corresponding to the fourth mean and the fourth standard deviation.
[0030] According to embodiments of this disclosure, calculating the alignment deviation based on the precise coordinates of the template mark center and the precise coordinates of the substrate mark center includes:
[0031]
[0032]
[0033]
[0034] The alignment deviation was calculated, where the template and the substrate each have two alignment marks, and the precise coordinates of the two alignment marks on the template are respectively... D is the distance between the two alignment marks on the template, and the precise coordinates of the two alignment marks on the substrate are respectively... Δx is the lateral distance deviation between the template and the substrate, Δy is the longitudinal distance deviation between the template and the substrate, and Δθ is the rotation angle deviation between the template and the substrate.
[0035] The template-substrate alignment method provided according to the embodiments of this disclosure has at least the following beneficial effects:
[0036] The grayscale values of the marked image are summed by row or column using a projection summation method. A set of mark center coordinates is calculated based on the summation results. Then, the outline of the marked image is interpolated, and another set of mark center coordinates is calculated based on the interpolation results. Finally, the two sets of mark center coordinates are fused. This method fully considers the statistical characteristics of obtaining mark center coordinates under different parameters, making the calculated mark center coordinates of the template and substrate more accurate. This makes the calculated alignment deviation more accurate and improves the alignment accuracy of the template and substrate.
[0037] Furthermore, by setting different grayscale thresholds and cutting the grayscale summation curve multiple times, the average value of multiple sets of template and substrate mark center coordinates is calculated. By averaging multiple sets of template and substrate mark center coordinates calculated based on multiple processing windows, the accuracy of template and substrate mark center coordinate calculation is improved, thereby improving the alignment accuracy of the template and substrate.
[0038] Furthermore, the final coordinates of the template and substrate mark center are calculated based on a two-dimensional Gaussian distribution. This calculation considers not only the average value of the template and substrate mark center but also the variance, in order to ensure the accuracy of the final coordinates.
[0039] Furthermore, this method is based on photometric alignment measurement technology, and improves its digital image processing method to calculate alignment deviation. It not only maintains the advantages of existing photometric technology, such as wide measurement range and simple structure, but also improves measurement accuracy and stability. Attached Figure Description
[0040] The above and other objects, features and advantages of this disclosure will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:
[0041] Figure 1 A flowchart illustrating the template and substrate alignment method provided in an embodiment of this disclosure is shown schematically.
[0042] Figure 2A The diagram schematically illustrates the shape of alignment marks on the template and substrate provided in the embodiments of this disclosure.
[0043] Figure 2B The diagram schematically illustrates the shape of an alignment mark acquired by an image sensor according to an embodiment of the present disclosure.
[0044] Figure 3 The diagram illustrates a graph of grayscale projection summation provided in an embodiment of this disclosure.
[0045] Figure 4 A flowchart illustrating operation S102 provided in an embodiment of this disclosure is shown schematically.
[0046] Figure 5 A flowchart illustrating operation S403 provided in an embodiment of this disclosure is shown schematically.
[0047] Figure 6 The schematic diagram illustrates a windowed outline provided in an embodiment of this disclosure.
[0048] Figure 7 The illustration schematically shows a first contour grayscale curve provided in an embodiment of the present disclosure.
[0049] Figure 8 A flowchart illustrating operation S104 provided in an embodiment of this disclosure is shown schematically.
[0050] Figure 9 The diagram illustrates the interpolation principle of the contour grayscale curve provided in the embodiments of this disclosure.
[0051] Figure 10 A flowchart illustrating operation S105 provided in an embodiment of this disclosure is shown schematically. Detailed Implementation
[0052] To make the objectives, technical solutions, and advantages of this disclosure clearer, the following detailed description is provided in conjunction with specific embodiments and accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without inventive effort are within the scope of protection of this disclosure.
[0053] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.
[0054] In this disclosure, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection, an electrical connection, or a connection that allows communication between them; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this disclosure according to the specific circumstances.
[0055] In the description of this disclosure, it should be understood that the terms "longitudinal", "length", "circumferential", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this disclosure and simplifying the description, and do not indicate or imply that the subsystem or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this disclosure.
[0056] Throughout the accompanying drawings, identical elements are represented by the same or similar reference numerals. Conventional structures or constructions have been omitted where they may cause confusion in understanding this disclosure. Furthermore, the shapes, dimensions, and positional relationships of the components in the drawings do not reflect actual size, scale, or actual positional relationships. Additionally, any reference numerals placed between parentheses in the claims should not be construed as limiting the claims.
[0057] Similarly, to simplify this disclosure and aid in understanding one or more of the various aspects of the disclosure, in the above description of exemplary embodiments of the present disclosure, various features of the present disclosure are sometimes grouped together in a single embodiment, figure, or description thereof. The use of terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refers to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present disclosure. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.
[0058] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this disclosure, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.
[0059] The template and substrate alignment method disclosed herein involves using an image sensor (Charge Coupled Device, CCD camera) as the acquisition means, acquiring alignment marks on the template and substrate through an imaging lens group, then using digital image processing methods to obtain statistical characteristics of alignment deviation, fusing statistical characteristics under different parameters, calculating template and substrate alignment deviation, and controlling the workpiece stage to complete the alignment action.
[0060] Figure 1 A flowchart illustrating the template and substrate alignment method provided in an embodiment of this disclosure is shown schematically.
[0061] like Figure 1 As shown, the template and substrate alignment method may include, for example, operations S101 to S105.
[0062] In operation S101, the grayscale of the marked image is summed by row or column to obtain the first projection summation curve for each row and the second projection summation curve for each column.
[0063] In operation S102, based on the first projection summation curve corresponding to each row and the second projection summation curve corresponding to each column, the first average value and the first standard deviation of the center coordinates of the template mark and the second average value and the second standard deviation of the center coordinates of the substrate mark are calculated.
[0064] In operation S103, the image contour of the marked image is extracted, and a horizontal processing window or a vertical processing window is added to the image contour. The grayscale curve of the first contour corresponding to each row in the horizontal processing window or the grayscale curve of the second contour corresponding to each column in the vertical processing window is extracted.
[0065] In operation S104, based on the first contour grayscale curve corresponding to each row or the second contour grayscale curve corresponding to each column, the third average value and the third standard deviation of the center coordinates of the template mark and the fourth average value and the fourth standard deviation of the center coordinates of the substrate mark are calculated.
[0066] In operation S105, the first average value, the first standard deviation, the second average value, the second standard deviation, the third average value, the third standard deviation, the fourth average value, and the first standard deviation are statistically fused to obtain the alignment deviation, and the template and the substrate are aligned according to the alignment deviation.
[0067] It should be noted that operations S101 to S104 are not intended to strictly limit the execution order of the alignment method in this embodiment. Operations S101 to S102 can be executed first, followed by operations S103 to S104, or operations S103 to S104 can be executed first, followed by operations S101 to S102, or operations S103 to S104 can be executed simultaneously with operations S101 to S102.
[0068] For example, an image sensor can be used to acquire alignment marks on the template and substrate to obtain a mark image. The size of the mark image can be M×N, where M represents M pixels in the vertical direction and N represents N pixels in the horizontal direction.
[0069] Figure 2A The diagram schematically illustrates the shape of alignment marks on the template and substrate provided in the embodiments of this disclosure. Figure 2B The diagram schematically illustrates the shape of an alignment mark acquired by an image sensor according to an embodiment of the present disclosure.
[0070] like Figures 2A to 2B As shown, the alignment marks on the template can be square alignment marks, and the alignment marks on the substrate can be cross alignment marks. The alignment of the template and the substrate can be understood as the cross alignment mark being exactly in the center of the square alignment mark.
[0071] For example, the grayscale values of the marked image are summed row by row and column by column to obtain the projection summation curve S. The grayscale value corresponding to the pixel in the m-th row and n-th column of the marked image is I(m, n), and the formula for summing the grayscale values of the m-th row and n-th column is:
[0072]
[0073]
[0074] Among them, S row (m) represents the sum of the gray values corresponding to the pixels in the m-th row, S col (n) represents the sum of gray values corresponding to the pixels in the nth column, where the subscript row indicates the row and the subscript col indicates the column.
[0075] Figure 3 The diagram illustrates a graph of grayscale projection summation provided in an embodiment of this disclosure.
[0076] like Figure 3 As shown, if Figure 3 If the x-coordinates are m = 1, 2, ..., M, then the y-coordinate is S. row (m), if Figure 3 If the x-coordinates are n = 1, 2, ..., N, then the y-coordinates are S. col (n).
[0077] According to an embodiment of this disclosure, after obtaining the projection summation curve S corresponding to each row or column, a threshold cutting method is used to perform peak cutting on the first projection summation curve and the second projection summation curve, and then the first average value and the first standard deviation of the template mark center coordinates and the second average value and the second standard deviation of the substrate mark center coordinates are calculated based on the cutting results.
[0078] Figure 4 A flowchart illustrating operation S102 provided in an embodiment of this disclosure is shown schematically.
[0079] like Figure 4 As shown, the calculation of the first average value and first standard deviation of the center coordinates of the template mark and the second average value and second standard deviation of the center coordinates of the substrate mark in operation S102 may include operations S401 to S405.
[0080] In operation S401, the first projection summation curve is cut by the first grayscale threshold, and the second projection summation curve is cut by the second grayscale threshold.
[0081] In operation S402, the first centroid corresponding to each peak value greater than the first grayscale threshold on the first projection summation curve is calculated, and the second centroid corresponding to each peak value greater than the second grayscale threshold on the second projection summation curve is calculated.
[0082] In operation S403, the center coordinates of the template mark and the center coordinates of the substrate mark are calculated based on the first and second centroids.
[0083] In operation S404, the size of the first grayscale threshold and the second grayscale threshold are changed, and the above operation is repeated to obtain the center coordinates of multiple template marks and the center coordinates of substrate marks.
[0084] In operation S405, a first average value and a first standard deviation are calculated based on the center coordinates of multiple template marks, and a second average value and a second standard deviation are calculated based on the center coordinates of multiple substrate marks.
[0085] For example, if the initial grayscale threshold is selected as T(0), then the initial first grayscale threshold T corresponding to the first projection summation curve is... row (0)=(max(S row (m))+min(S row (m))) / 2, the initial second grayscale threshold T corresponding to the second projection summation curve. col (0)=(max(S col (n))+min(S col (n))) / 2.
[0086] To improve the accuracy of the marker center coordinate calculation, the first grayscale threshold can be changed multiple times to segment the first and second projection summation curves. The grayscale threshold for the i-th threshold segmentation can be denoted as T(i), and the first grayscale threshold corresponding to the first projection summation curve can be denoted as Ti. row (i) The second grayscale threshold corresponding to the second projection summation curve is T. col (i). For each cut, the center coordinates of the template mark and the center coordinates of the substrate mark can be calculated.
[0087] During the threshold cutting process, it is necessary to determine whether the number of cuts i has reached the set value a. Once the number of cuts i reaches the set value a, the first average, first standard deviation, second average, and second standard deviation corresponding to the center coordinates of the template mark and the substrate mark in group a can be calculated based on these coordinates. If the number of cuts i has not reached the set value a, the first and second grayscale thresholds are continuously adjusted to cut the first and second projection summation curves, and the center coordinates of the template mark and the substrate mark are calculated.
[0088] In this embodiment of the disclosure, it can be based on
[0089]
[0090] Calculate the first or second centroid g k (i), where x is the number of rows or columns of pixels in the marked image, S(x) is the sum of gray values of the pixels in the xth row or column, k is the number of the peak on the first projection summation curve whose peak value is greater than the first gray threshold or the number of the peak on the second projection summation curve whose peak value is greater than the second gray threshold, x1 and x2 are the starting point and ending point of the kth peak after it is cut, and i is the number of times the first gray threshold is changed or the number of times the second gray threshold is changed, that is, the number of cuts.
[0091] Figure 5 A flowchart illustrating operation S403 provided in an embodiment of this disclosure is shown schematically.
[0092] like Figure 5 As shown, the calculation of the center coordinates of the template mark and the center coordinates of the substrate mark based on the first and second centroids in operation S403 may include operations S501 to S503.
[0093] In operation S501, the ordinates of each peak on the first projection summation curve whose peak value is greater than the first grayscale threshold are determined based on the first centroid.
[0094] In operation S502, the horizontal coordinates of each peak on the second projection summation curve whose peak value is greater than the second grayscale threshold are determined based on the second centroid.
[0095] In operation S503, the center coordinates of the template mark and the center coordinates of the substrate mark are calculated based on the vertical coordinates and horizontal coordinates of each peak.
[0096] For example, please refer to the following: Figure 3 Each time the first projection summation curve is cut by a first grayscale threshold and the second projection summation curve is cut by a second grayscale threshold, the three peaks above the first and second grayscale thresholds are retained. That is, the extreme value T of T(i) is... max It should be less than the third highest peak value in the projection summation curve S, that is, T(i) is between T(0) and the third highest peak value. After the i-th iteration, the cutting threshold T(i+1) value for the (i+1)-th iteration is:
[0097]
[0098] In the above g k In the calculation formula (i), when x is the row number of pixels in the marked image, that is, when x is m = 1, 2, ..., M, g k (i) represents the ordinate of the three peaks g1(i), g2(i), and g3(i) corresponding to the first projection summation curve, denoted as y. g1 (i), y g2 (i), y g3 (i); when x is the column number of pixels in the labeled image, that is, when x is n = 1, 2, ..., N, g k (i) represents the abscissa of the three peaks g1(i), g2(i), and g3(i) corresponding to the second projection summation curve, denoted as x. g1 (i), x g2 (i), x g3 (i).
[0099] Therefore, based on
[0100] x p1 (i)=(x g1(i) +x g3(i) ) / 2
[0101] y p1 (i)=(y g1(i) +y g3(i) ) / 2
[0102] x q1 (i)=x g2(i)
[0103] y q1 (i)=y g2(i)
[0104] Calculate the x-coordinate of the center coordinate of the template mark. p1 (i) and y-coordinate p1 (i) and the x-coordinate of the center coordinate of the substrate mark q1 (i) and y-coordinate q1 (i), p represents the template, q represents the substrate, x g1 (i), x g2 (i), x g3 (i) represents the abscissas of each peak on the second projection summation curve whose peak value is greater than the second grayscale threshold, y g1 (i), y g2 (i), y g3 (i) are the ordinates of each peak on the first projection summation curve whose peak value is greater than the first gray level threshold.
[0105] After a number of cuts using the first grayscale threshold T(i), a groups of x can be calculated. p1 (i), y p1 (i) and the x-coordinate of the center coordinate of the substrate mark q1 (i) and y-coordinate q1 (i), the center coordinates of the a template markers can be denoted as:
[0106] [x p1 (1), y p1 (1)]、[x p1 (2), y p1 (2)]、…[x p1 (a), y p1 (a)]
[0107] The center coordinates of a substrate mark can be denoted as:
[0108] [x q1 (1), y q1 (1)]、[x q1 (2), y q1 (2)]、…[xq1 (a), y q1 (a)]
[0109] Calculate the average and variance of the center coordinates of the above 'a' template marks respectively to obtain the first average. and the first standard deviation (σ) px1 σ py1 ), calculate the average and variance of the center coordinates of the above a substrate marks respectively, and obtain the second average value. Second standard deviation (σ) qx1 σ qy1 ).
[0110] Based on the above operations, the statistical features of the center coordinates of the marker under the gray-scale summation curve were obtained.
[0111] According to embodiments of this disclosure, the process of extracting the first contour grayscale curve and the second contour grayscale curve in operation S103 can be as follows:
[0112] High-pass filtering is applied to the marked image to obtain high-frequency information, that is, the image contour is preserved while the background and padding are removed to obtain a contour map. Then, based on the known marker size information and the result of the projection summation algorithm, a processing window is added to the contour map to remove redundant information such as the background.
[0113] Figure 6 The schematic diagram illustrates a windowed outline provided in an embodiment of this disclosure.
[0114] like Figure 6 As shown, horizontal or vertical processing windows are added to the image outline, as indicated by the dashed boxes in the figure. Labels 1 and 2 represent horizontal processing windows, and labels 3 and 4 represent vertical processing windows.
[0115] For each horizontal processing window, extract the contours row by row within the horizontal processing window to obtain the first contour grayscale curve. For each vertical processing window, extract the contours column by column within the vertical processing window to obtain the second contour grayscale curve.
[0116] Figure 7 The illustration schematically shows a first contour grayscale curve provided in an embodiment of the present disclosure.
[0117] like Figure 7 As shown, in each row, each contour forms a grayscale peak, and the grayscale curve of each row of contours has a total of G. j，1 -G j，6 Six grayscale peaks. Based on the first contour grayscale curve of each row, the center coordinates of a template mark and the center coordinates of a substrate mark can be calculated. The specific calculation method is explained below.
[0118] Figure 8A flowchart illustrating operation S104 provided in an embodiment of this disclosure is shown schematically.
[0119] like Figure 8 As shown, the calculation of the third average value and third standard deviation of the center coordinates of the template mark and the fourth average value and fourth standard deviation of the center coordinates of the substrate mark in operation S104 may include operations S801 to S803.
[0120] In operation S801, the first contour grayscale curve is interpolated according to the first interpolation threshold or the second contour grayscale curve is interpolated according to the second interpolation threshold to obtain the coordinates of the contour boundary formed by the template and the substrate corresponding to the first contour grayscale curve or the coordinates corresponding to the second contour grayscale curve.
[0121] In operation S802, based on the coordinates corresponding to the first contour grayscale curve or the second contour grayscale curve of the contour boundary formed by the template and the substrate, the center coordinates of the template mark and the center coordinates of the substrate mark corresponding to each row or column are calculated.
[0122] In operation S803, the third average, third standard deviation, fourth average, and fourth standard deviation are calculated based on the center coordinates of the template mark and the center coordinates of the substrate mark corresponding to each row or column.
[0123] Figure 9 The diagram illustrates the interpolation principle of the contour grayscale curve provided in the embodiments of this disclosure.
[0124] For example, such as Figure 9 As shown, the explanation uses a horizontal processing window. Let the first interpolation threshold be Q(j), and G... j，v For the first contour grayscale curve of the vth contour grayscale peak of the j-th phase cell near Q(j) in the horizontal processing window, select the interpolation point [b, G] near Q(j). j (b)], [b+1, Gj(b+1)]; [c, G j (c)]、[c+1,G j (c+1)]. By locally linearizing the left and right sides of the v-th contour grayscale peak of the j-th phase element, the left and right boundaries of the contour can be obtained, as follows:
[0125]
[0126]
[0127] Among them, C jl，v C represents the left boundary of the v-th contour grayscale peak in the horizontal processing window. jr，v This represents the right boundary of the v-th contour grayscale peak in the horizontal processing window. The subscript l represents the left boundary, and the subscript r represents the right boundary. v = 1, 2, 3, 4, 5, 6.
[0128] Based on the contour boundary, it can be determined according to
[0129] x p2 (j)=(Cx j，1 +Cx j，2 +Cx j，5 +Cx j，6 ) / 4
[0130] x q2 (j)=(Cx j，3 +Cx j，4 ) / 2
[0131] Calculate the x-coordinate of the template marker center coordinates corresponding to the j-th pixel in the horizontal processing window. p2 (j) and the x-coordinate of the center coordinate of the substrate mark q2 (j), the y-coordinate of the center coordinate of the template mark p2 (j) and the ordinate y of the substrate mark center coordinates q2 (j) represents the ordinate of the j-th row corresponding to the first contour grayscale curve, which can be converted into coordinate values based on the row number of the pixels. Here, v represents the index of the grayscale peak, taking values from 1 to 6, and Cx... j，v Cx represents the x-coordinate of the v-th grayscale peak corresponding to the j-th row pixel after interpolation. j，v =(C jl，v +C jr，v ) / 2.
[0132] Similarly, in the vertical processing window, based on the contour boundary, it can be determined according to...
[0133] y p2 (w)=(Cy w，1 +Cy w，2 +Cy w，5 +Cy w，6 ) / 4
[0134] y q2 (w)=(Cy w，3 +Cy w，4 ) / 2
[0135] Calculate the y-coordinate of the template marker center corresponding to the w-th pixel in the vertical processing window. p2 (w) and the ordinate y of the center coordinates of the substrate mark q2 (w), the x-coordinate of the center coordinate of the template mark p2 (w) and the x-coordinate of the center coordinate of the substrate mark q2 (w) represents the x-coordinate of the w-th column corresponding to the second contour grayscale curve, which can be converted into coordinate values based on the column number of the pixels. Here, v represents the index of the grayscale peak, taking values from 1 to 6, and Cy...m，v Cy represents the interpolated ordinate of the v-th grayscale peak corresponding to the w-th pixel. w，v =(C wl，v +C wr，v ) / 2.
[0136] It should be understood that the principle of vertical window processing is similar to that of horizontal processing.
[0137] After processing all horizontal and vertical processing windows, the d groups of x can be calculated. p2 (j), y p2 (j) and x q2 (j), y q2 (j), where d represents the total number of rows or columns within the window, and the center coordinates of the d template markers can be denoted as:
[0138] [x p2 (1), y p2 (1)]、[x p2 (2), y p2 (2)]、…[x p2 (d), y p2 (d)]
[0139] The center coordinates of the d substrate marks can be denoted as:
[0140] [x q2 (1), y q2 (1)]、[x q2 (2), y q2 (2)]、…[x q2 (d), y q2 (d)]
[0141] The average and variance of the center coordinates of the above d template marks are calculated respectively to obtain the third average. and the third standard deviation (σ) p x2, σ p y2), calculate the average and variance of the center coordinates of the above d substrate marks respectively, and obtain the fourth average value. and the fourth standard deviation (σ) qx2 σ qy2 ).
[0142] Based on the above operations, the statistical features of the marker center coordinates under the contour curve were obtained.
[0143] Next, by fusing the statistical features of the marker center coordinates under the grayscale summation curve and the statistical features of the marker center coordinates under the contour curve, the precise coordinates of the template marker center and the substrate marker center can be obtained, and then the rotation deviation can be calculated.
[0144] Figure 10A flowchart illustrating operation S105 provided in an embodiment of this disclosure is shown schematically.
[0145] like Figure 10 As shown, operation S105 statistically integrates the first mean, first standard deviation, second mean, second standard deviation, third mean, third standard deviation, fourth mean, and first standard deviation to calculate a more accurate label center, which may include operations S1001 to S1003.
[0146] In operation S1001, based on a two-dimensional Gaussian distribution, the precise coordinates of the template mark center are calculated according to the first mean, the first standard deviation, the third mean, and the third standard deviation.
[0147] In operation S1002, based on a two-dimensional Gaussian distribution, the precise coordinates of the substrate mark center are calculated according to the second mean, the second standard deviation, the fourth mean, and the fourth standard deviation.
[0148] In operation S1003, the alignment deviation is calculated based on the precise coordinates of the template mark center and the precise coordinates of the substrate mark center.
[0149] For example, the average and variance of the center coordinates of the template and substrate marks obtained by the above two methods can be represented as a two-dimensional Gaussian distribution, which can then be used to...
[0150]
[0151] Calculate the precise coordinates of the center of the template mark or the precise coordinates of the center of the substrate mark, where [u1(x,y), σ1(x,y)] represents the two-dimensional Gaussian distribution corresponding to the first mean and the first standard deviation or the two-dimensional Gaussian distribution corresponding to the second mean and the second standard deviation, and [u2(x,y), σ2(x,y)] represents the two-dimensional Gaussian distribution corresponding to the third mean and the third standard deviation or the two-dimensional Gaussian distribution corresponding to the fourth mean and the fourth standard deviation.
[0152] Then according to
[0153]
[0154]
[0155]
[0156] The alignment deviation was calculated, where the template and the substrate each have two alignment marks, and the precise coordinates of the two alignment marks on the template are respectively... D is the distance between the two alignment marks on the template, and the precise coordinates of the two alignment marks on the substrate are respectively... Δx is the lateral distance deviation between the template and the substrate, Δy is the longitudinal distance deviation between the template and the substrate, and Δθ is the rotation angle deviation between the template and the substrate.
[0157] Based on the alignment deviation calculated above, the template and the substrate can be aligned by moving Δx and Δy left and right and rotating Δθ.
[0158] It should be noted that the template-substrate alignment method described above is based on the alignment marks on the template being square alignment marks and the alignment marks on the substrate being cross alignment marks. It should be understood that even if the alignment marks on the template are set to cross alignment marks and the alignment marks on the substrate are set to square alignment marks, the template-substrate alignment method described above can still be used for alignment. The only difference is that when calculating the average and standard deviation of the center coordinates of the alignment marks on the template, a second average and second standard deviation, and a fourth average and a fourth standard deviation calculation method are used; when calculating the average and standard deviation of the center coordinates of the alignment marks on the substrate, a first average and first standard deviation, and a third average and a third standard deviation calculation method are used. Therefore, using cross alignment marks on the template and square alignment marks on the substrate for alignment is also within the scope of this disclosure.
[0159] The specific embodiments described above further illustrate the purpose, technical solutions, and beneficial effects of this disclosure. It should be understood that the above descriptions are merely specific embodiments of this disclosure and are not intended to limit this disclosure. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the protection scope of this disclosure.
Claims
1. A method for aligning a template with a substrate, characterized in that, include: The grayscale values of the marked image are summed by row or column to obtain the first projection summation curve for each row and the second projection summation curve for each column. Based on the first projection summation curve corresponding to each row and the second projection summation curve corresponding to each column, calculate the first average value and the first standard deviation of the center coordinates of the template mark, and the second average value and the second standard deviation of the center coordinates of the substrate mark. Extract the image contour of the marked image, and add a horizontal processing window or a vertical processing window to the image contour; Extract the first contour grayscale curve corresponding to each row in the horizontal processing window or the second contour grayscale curve corresponding to each column in the vertical processing window; Based on the first contour grayscale curve corresponding to each row or the second contour grayscale curve corresponding to each column, calculate the third average value and the third standard deviation of the template mark center coordinates and the fourth average value and the fourth standard deviation of the substrate mark center coordinates. The alignment deviation is obtained by statistically fusing the first average, the first standard deviation, the second average, the second standard deviation, the third average, the third standard deviation, the fourth average, and the fourth standard deviation. The template and substrate are aligned according to the alignment deviation.
2. The template and substrate alignment method according to claim 1, characterized in that, The calculation of the first average and first standard deviation of the template mark center coordinates and the second average and second standard deviation of the substrate mark center coordinates based on the first projection summation curve corresponding to each row and the second projection summation curve corresponding to each column includes: The first projection summation curve is cut using a first grayscale threshold, and the second projection summation curve is cut using a second grayscale threshold; Calculate the first centroid corresponding to each peak value greater than the first grayscale threshold on the first projection summation curve, and calculate the second centroid corresponding to each peak value greater than the second grayscale threshold on the second projection summation curve; Calculate the center coordinates of the template mark and the center coordinates of the substrate mark based on the first centroid and the second centroid; By changing the values of the first grayscale threshold and the second grayscale threshold, and repeating the above operation, multiple template mark center coordinates and substrate mark center coordinates are obtained; The first average value and the first standard deviation are calculated based on the center coordinates of the multiple template marks, and the second average value and the second standard deviation are calculated based on the center coordinates of the multiple substrate marks.
3. The template and substrate alignment method according to claim 2, characterized in that, The step of calculating the center coordinates of the template mark and the center coordinates of the substrate mark based on the first centroid and the second centroid includes: The ordinates of each peak on the first projection summation curve whose peak value is greater than the first gray level threshold are determined based on the first centroid. The abscissas of each peak on the second projection summation curve whose peak value is greater than the second gray threshold are determined based on the second centroid. The center coordinates of the template mark and the center coordinates of the substrate mark are calculated based on the vertical coordinates and horizontal coordinates of each peak.
4. The template and substrate alignment method according to claim 3, characterized in that, according to calculating the first barycenter or the second barycenter g k (i), wherein x is the row number or column number of the pixel points in the mark image, S(x) is the sum of the gray values of the xth row or column pixel points, k is the number of the peak with a peak value greater than the first gray threshold on the first projection sum curve or the number of the peak with a peak value greater than the second gray threshold on the second projection sum curve, x1 and x2 are the starting point and the ending point after the kth peak is cut, and i is the number of changes of the first gray threshold or the number of changes of the second gray threshold.
5. The template and substrate alignment method according to claim 4, characterized in that, The number of peaks on the first projection summation curve with a peak value greater than the first grayscale threshold and the number of peaks on the second projection summation curve with a peak value greater than the second grayscale threshold are both 3. The step of calculating the center coordinates of the template mark and the center coordinates of the substrate mark based on the ordinate and abscissa of each peak includes: according to x p1 (i)=(x g1 (i)+x g3 (i)) / 2 y p1 (i)=(y g1 (i)+y g3 (i)) / 2 x q1 (i)=x g2 (i) y q1 (i)=y g2 (i) Calculate the x-coordinate of the center coordinate of the template mark. p1 (i) and y-coordinate p1 (i) and the x-coordinate of the center coordinate of the substrate mark q1 (i) and y-coordinate q1 (i), p represents the template, q represents the substrate, x g1 (i), x g2 (i), x g3 (i) are the x-coordinates of each peak on the second projection summation curve whose peak value is greater than the second grayscale threshold, respectively. g1 (i), y g2 (i), y g3 (i) are the ordinates of each peak on the first projection summation curve whose peak value is greater than the first gray level threshold.
6. The template and substrate alignment method according to claim 1, characterized in that, The step of calculating the third average value and third standard deviation of the template mark center coordinates and the fourth average value and fourth standard deviation of the substrate mark center coordinates based on the first contour grayscale curve corresponding to each row or the second contour grayscale curve corresponding to each column includes: Interpolate the first contour grayscale curve according to the first interpolation threshold or interpolate the second contour grayscale curve according to the second interpolation threshold to obtain the coordinates of the contour boundary formed by the template and the substrate corresponding to the first contour grayscale curve or the coordinates corresponding to the second contour grayscale curve. Based on the coordinates of the outline boundary formed by the template and the substrate corresponding to the first outline grayscale curve or the second outline grayscale curve, calculate the center coordinates of the template mark and the center coordinates of the substrate mark corresponding to each row or column. The third average, the third standard deviation, the fourth average, and the fourth standard deviation are calculated based on the center coordinates of the template mark and the center coordinates of the substrate mark corresponding to each row or column.
7. The template and substrate alignment method according to claim 6, characterized in that, The number of grayscale peaks in the first contour grayscale curve corresponding to each row or the second contour grayscale curve corresponding to each column is 6. Based on the coordinates corresponding to the grayscale curve of the first contour formed by the template and substrate, the coordinates of the template mark center and the substrate mark center for each row are calculated, including: according to x p2 (j)=(Cx j，1 +Cx j，2 +Cx j，5 +Cx j，6 ) / 4 x q2 (j)=(Cx j，3 +Cx j，4 ) / 2 Calculate the x-coordinate of the template marker center coordinates corresponding to the j-th pixel in the horizontal processing window. p2 (j) and the x-coordinate of the center coordinate of the substrate mark q2 (j), the y-coordinate of the center coordinate of the template mark p2 (j) and the ordinate y of the substrate mark center coordinates q2 (j) represents the ordinate of the j-th row corresponding to the first contour grayscale curve, which can be obtained by converting the row number of the pixel into coordinate values, where v represents the index of the grayscale peak, taking the value of an integer from 1 to 6, and C xj，v C represents the x-coordinate of the v-th grayscale peak corresponding to the j-th row pixel after interpolation. xj，v =(C jl，v +C jr，v ) / 2; Based on the coordinates corresponding to the grayscale curve of the second contour formed by the outline boundary of the template and the substrate, the coordinates of the template mark center and the substrate mark center for each column are calculated, including: according to y p2 (w)=(Cy w，1 +Cy w，2 +Cy w，5 +Cy w，6 ) / 4 y q2 (w)=(Cy w，3 +Cy w，4 ) / 2 Calculate the y-coordinate of the template marker center corresponding to the w-th pixel in the vertical processing window. p2 (w) and the ordinate y of the center coordinates of the substrate mark q2 (w), the x-coordinate of the center coordinate of the template mark p2 (w) and the x-coordinate of the center coordinate of the substrate mark q2 (w) represents the x-coordinate of the w-th column corresponding to the grayscale curve of the second contour. It can be obtained by converting the column number of the pixels into coordinates. Here, v represents the index of the grayscale peak, taking integer values from 1 to 6. Cy w，v Cy represents the interpolated ordinate of the v-th grayscale peak corresponding to the w-th pixel. w，v =(C wl，v +C wr，v ) / 2.
8. The template and substrate alignment method according to claim 1, characterized in that, The statistical fusion of the first average, the first standard deviation, the second average, the second standard deviation, the third average, the third standard deviation, the fourth average, and the fourth standard deviation yields the alignment deviation, which includes: Based on a two-dimensional Gaussian distribution, the precise coordinates of the template mark center are calculated according to the first mean, the first standard deviation, the third mean, and the third standard deviation; Based on a two-dimensional Gaussian distribution, the precise coordinates of the substrate mark center are calculated according to the second average value, the second standard deviation, the fourth average value, and the fourth standard deviation. The alignment deviation is calculated based on the precise coordinates of the center of the template mark and the precise coordinates of the center of the substrate mark.
9. The template and substrate alignment method according to claim 8, characterized in that, according to Calculate the precise coordinates of the center of the template mark or the precise coordinates of the center of the substrate mark, where [u1(x,y), σ1(x,y)] represents the two-dimensional Gaussian distribution corresponding to the first average and the first standard deviation or the two-dimensional Gaussian distribution corresponding to the second average and the second standard deviation, and [u2(x,y), σ2(x,y)] represents the two-dimensional Gaussian distribution corresponding to the third average and the third standard deviation or the two-dimensional Gaussian distribution corresponding to the fourth average and the fourth standard deviation.
10. The template and substrate alignment method according to claim 9, characterized in that, The step of calculating the alignment deviation based on the precise coordinates of the template mark center and the precise coordinates of the substrate mark center includes: The alignment deviation is calculated, wherein the template and the substrate each have two alignment marks, and the precise coordinates of the two alignment marks of the template are respectively... D is the distance between the two alignment marks on the template, and the precise coordinates of the two alignment marks on the substrate are respectively... Δx is the lateral distance deviation between the template and the substrate, Δy is the longitudinal distance deviation between the template and the substrate, and Δθ is the rotation angle deviation between the template and the substrate.