A mask pre-alignment method for projection lithography based on motion compensation model

Abstract

The present invention discloses a method for pre-alignment of a projection lithography machine mask based on a motion compensation model, and belongs to the field of lithography. The present invention first establishes a motion compensation model for a mask stage to compensate for the rotational deviation caused by the movement of the R axis and the displacement errors in the X and Y directions. Then, the precise coordinates of the alignment mark are obtained based on the pre-alignment image algorithm, and the servo motor is driven in combination with the motion compensation model to make corresponding adjustments to the mask stage along the X, Y and R directions, and finally enter the alignment position. The present invention ensures the pre-alignment accuracy of the mask by precisely controlling the movement of the mask stage, effectively improves the overall accuracy of the alignment between the mask and the silicon wafer, and thus improves the stability and reliability of the lithography process. While reducing the alignment error, the present invention improves the production efficiency of the lithography process, and has broad industrial application prospects.

Description

Technical Field

[0001] The present invention belongs to the field of lithography, comprises an image processing algorithm, and is specifically a projection lithography machine mask pre-alignment method based on a motion compensation model, which is used for accurately and quickly realizing the mask pre-alignment in the projection lithography machine. Background Art

[0002] Projection lithography is the core equipment in modern integrated circuit manufacturing. It projects circuit patterns onto silicon wafers through masks for exposure, thereby forming tiny structures and realizing the manufacture of integrated circuits. In the lithography process, the precise alignment of the mask and the silicon wafer is the key factor to ensure the accurate projection of the final circuit pattern. Therefore, the pre-alignment technology of the mask is crucial to the overall accuracy and production efficiency of the lithography process. The mask stage is used in the projection lithography machine to fix and move the mask so that it is accurately aligned with the exposure area. In actual lithography operations, the X, Y translation and R-axis rotation of the mask stage have a decisive influence on the precise alignment of the mask. However, due to assembly errors and unavoidable factors in the debugging process, the X, Y and R-axis movements of the mask stage cannot be completely concentric, and the linear motion characteristics of the R-axis motor are difficult to directly correspond to the changes in the mask rotation angle, resulting in deviations in the mask alignment process. If this deviation is not effectively compensated, it will directly affect the exposure accuracy of the silicon wafer, ultimately causing the circuit pattern to shift, and even cause product defects in severe cases.

[0003] In the prior art, the alignment of the mask stage mostly relies on direct mechanical adjustment and simple image recognition algorithms. This method has low initial alignment accuracy, and the adjustment process is cumbersome and time-consuming. In particular, when there are large errors between the X, Y and R axis movements, the alignment effect is not ideal. In addition, the prior art often relies on complex mechanical structures or manual intervention to compensate for the mask rotation angle, making it difficult to achieve fast and automated alignment adjustment. These deficiencies limit the overall efficiency of the lithography process and increase the complexity of process debugging. In particular, in production with high-precision alignment requirements, the deficiencies of the prior art are particularly prominent. Summary of the invention

[0004] The present invention is a method for pre-aligning a mask of a projection lithography machine based on a motion compensation model, the purpose of which is to ensure the consistency of the mask coordinate system and the exposure coordinate system of the lithography machine, thereby providing an important premise for the precise exposure of silicon wafers. In a projection lithography machine, the mask stage is composed of three servo motors, which respectively control the movement of the mask in the three directions of X, Y and R (rotation). In the mask pre-alignment process of the lithography machine, firstly due to assembly and debugging, the mask cannot be completely concentric with the mask stage in the X and Y directions, and the linear motion of the R-axis motor cannot directly correspond to the rotation angle of the mask. Therefore, the present invention first establishes a motion compensation model for the mask stage to compensate for the rotation deviation caused by the movement of the R axis and the displacement error in the X and Y directions. Then, the precise coordinates of the alignment mark are obtained based on the pre-alignment image algorithm, and the servo motor is driven in combination with the motion compensation model, so that the mask stage is adjusted accordingly along the X, Y and R directions, and finally enters the alignment position. This algorithm ensures the pre-alignment accuracy of the mask by accurately controlling the movement of the mask stage, effectively improves the overall accuracy of the alignment between the mask and the silicon wafer, and thus improves the stability and reliability of the lithography process. The present invention improves the production efficiency of the photolithography process while reducing the alignment error, and has broad industrial application prospects.

[0005] The specific technical solutions proposed by the present invention are as follows:

[0006] A method for pre-aligning a projection lithography machine mask based on a motion compensation model, the method comprising the following steps:

[0007] Step S1: The R-axis motion of the mask stage is measured by using the left and right mask pre-alignment cameras of the projection lithography machine and a special mask plate used for measurement, and an analysis and calculation is performed based on the measurement results to establish a mask stage R-axis motion compensation model.

[0008] Step S2: Based on the mask pre-alignment image algorithm and in combination with the mask stage R-axis motion compensation model, the pre-alignment positioning mark on the mask is identified and the adjustment amount is calculated so that it can quickly and accurately enter the exposure setting position.

[0009] The beneficial effects of the present invention are as follows:

[0010] 1. The present invention improves the accuracy of mask pre-alignment and ensures the consistency of the mask and exposure coordinate systems. By establishing a motion compensation model for the mask stage, the rotation deviation caused by the R-axis movement and the displacement errors in the X and Y directions are compensated, ensuring the consistency of the mask coordinate system and the exposure coordinate system of the lithography machine, thereby providing a stable foundation for the precise exposure of silicon wafers.

[0011] 2. The present invention accurately controls the movement of the mask stage and improves the stability and reliability of the photolithography process. By combining the mask pre-alignment image algorithm with the mask stage motion compensation model, the mask pre-alignment mark can be accurately identified, and the servo motor can be driven to quickly and accurately adjust the position of the mask stage, ensuring the accuracy of the mask pre-alignment process and greatly improving the stability and reliability of the photolithography process.

[0012] 3. The present invention reduces mask alignment errors and improves production efficiency. Through compensation model calculation and image algorithm, the alignment error caused by the non-concentricity between the mask and the mask stage and the R-axis motion characteristics is reduced, thereby reducing the need for multiple adjustments, shortening the alignment time, and significantly improving the production efficiency of the lithography process.

[0013] 4. The present invention is applicable to a variety of mask pre-alignment scenarios. By introducing compensation models and image processing technology, it not only improves the alignment accuracy, but also adapts to production processes with different size and accuracy requirements, and has broad industrial application potential.

[0014] 5. In the present invention, firstly, by accurately compensating for the error generated by the R-axis rotation, the rotation deviation problem caused by the non-concentricity of the mask and the mask stage in the X and Y directions is solved, so that the mask and the exposure coordinate system can be more accurately aligned, thereby significantly improving the alignment accuracy. Secondly, the present invention can automatically calculate and adjust the position of the mask stage through an automated error compensation model and image recognition algorithm, reduce manual intervention, improve the degree of automation of the operation, reduce dependence on the operator's experience, and ensure the rapidity and accuracy of alignment. In addition, the present invention effectively reduces the errors generated during the mask pre-alignment process, reduces the need for multiple adjustments, significantly shortens the alignment time, and thereby improves the overall production efficiency of the lithography process. Finally, the compensation model and image processing technology of the present invention have wide adaptability, can meet production processes with different size and precision requirements, have higher flexibility and versatility, and provide broad prospects and possibilities for industrial applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Schematic diagram of the measurement scale in the left and right camera interfaces for mask pre-alignment;

[0016] Figure 2 Schematic diagram of the method for calculating the mask rotation angle;

[0017] Figure 3 This is a schematic diagram of the R-axis motion compensation model;

[0018] Figure 4 Schematic diagram of the left and right side marking images of the mask. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in each embodiment of the present invention described below can be combined with each other as long as they do not conflict with each other. To achieve the above-mentioned purpose, the present invention adopts the following technical scheme.

[0020] The present invention is a projection lithography machine mask pre-alignment method based on a motion compensation model, comprising the following steps:

[0021] Step S1: The R-axis motion of the mask stage is measured by using the left and right mask pre-alignment cameras of the projection lithography machine and the mask plate used for measurement, and an analysis and calculation is performed based on the measurement results to establish a mask stage R-axis motion compensation model.

[0022] Step S2: Based on the mask pre-alignment image algorithm and in combination with the mask stage R-axis motion compensation model, the pre-alignment positioning mark on the mask is identified and the adjustment amount is calculated so that it can quickly and accurately enter the exposure setting position.

[0023] The present invention ensures the pre-alignment accuracy of the mask by precisely controlling the movement of the mask stage, effectively improves the overall alignment accuracy between the mask and the silicon wafer, and further improves the stability and reliability of the photolithography process. The present invention improves the production efficiency of the photolithography process while reducing the alignment error, and has broad industrial application prospects.

[0024] Specifically, step 1 includes the following steps:

[0025] Step S11: The measurement tool for establishing the R-axis compensation motion model of the mask stage is a specific mask with scale values. In the left and right mask pre-alignment cameras, Figure 1 As shown in (a) and (b), the left and right mask pre-alignment cameras each have a set of measurement scale values, where the longitudinal scale value is used to measure the Y-direction change after the R-axis moves, and the transverse scale value is used to measure the X-direction change after the R-axis moves. The marks of the left and right mask pre-alignment cameras are the left and right side marks.

[0026] First determine the initial position of the R axis as , and the marks in the left and right mask pre-alignment cameras, that is, the initial Y values ​​of the left and right marks and and the initial value of X direction and , according to the step distance Move the R axis i steps, and then measure the Y and X coordinates of the left mark and the right mark after moving in the Y and X directions, and then calculate the Y and X changes of the left mark according to the formula and And the Y and X changes of the camera on the right and .

[0027] The formula for calculating the change in the Y direction of the R-axis movement of the mask stage is: R-axis position (mm) is: ; R axis movement (mm): ; The coordinates (um) of the left mark after moving in the Y direction are: , the coordinates (um) of the mark on the right after moving in the Y direction are: ; The Y-axis change (um) marked on the left is: ; The Y-axis change (um) marked on the right is: ; The average cumulative change in the Y direction (um) is: .

[0028] The formula for calculating the change in the X direction of the R-axis movement of the mask stage is: R-axis position (mm) is: ; R axis movement (mm): ; The coordinates (um) of the left mark after moving in the X direction are: ; The change in the X direction (um) marked on the left is: ; The coordinates (um) of the mark on the right after moving in the X direction are: ; The X-axis change (um) marked on the right is: ; The average cumulative change in the X direction (um) is: .

[0029] Step S12: According to Figure 2 The R-axis movement can be calculated using formula (1): The angle at which the mask rotates after the distance Then according to The corresponding R-axis movement By fitting the reverse function, the R-axis motion can be obtained. The linear function relationship with the rotation angle is shown in formula (2).

[0030] (1)

[0031] (2)

[0032] in, is the mask rotation radius; is the absolute value of the Y-axis change on the left. The absolute value of the Y-axis change is marked on the right; , is the correlation coefficient of a linear function.

[0033] Step S13: Since the mask and the mask stage cannot be completely concentric, it is necessary to calculate the errors in the Y and X directions when the R axis moves. This error will be compensated by the X and Y axes of the mask stage in the mask pre-alignment part.

[0034] When calculating the Y-axis error, we first need to get the spatial position of the mask based on the measurement situation. For example, during the measurement process, when the R axis moves in the positive direction, the mask moves clockwise, and the Y-axis movement of the left mark is less than that of the right mark, which means that the left side is the short axis in the rotation, the right side is the long axis, and the center of motion is biased to the left. When the R axis moves in the positive direction, both the left and right marks move to the right, so it can be determined that the rotation center of the mask stage is below the center of the mask. Based on this situation, we can establish the following: Figure 3 The motion model of the mask shown, where and Respectively represent the left and right endpoints of the mask center horizontal line, represents the midpoint of the mask center horizontal line, represents the rotation center of the mask stage, represents the intersection of the horizontal lines of the mask center before and after rotation, for The midpoint of and Respectively represent the changes in the left and right marks after rotation, and is the Y-direction offset caused by the mask rotation. Due to the measurement error, in order to accurately calculate the Y-direction offset , cannot be used directly , but calculate The length of The offset is calculated, for Relative to The rotation angle of The length of is calculated by formula (3), where , are the mask rotation radius, combined with all And the corresponding R-axis movement By fitting the reverse function, we can get the relationship between the Y-axis offset and the R-axis motion, as shown in formula (4).

[0035] (3)

[0036] (4)

[0037] When calculating the X-direction error, since the changes in the left and right markers are almost the same each time, the average cumulative change in the X-direction can be used directly. The corresponding R-axis movement By performing reverse function fitting, the relationship between the X-axis offset and the R-axis motion is obtained, as shown in formula (5).

[0038] (5)

[0039] in, is the mask Y error, is the mask X-direction error, , , , is the correlation coefficient of a linear function; thus, the mask stage compensation motion model is constructed.

[0040] Further, step S2 includes the following steps:

[0041] Step S21: First, load the alignment parameters, which include the set alignment coordinates of the left and right marks of the mask. and , the distance between the left and right marks of the mask , Mask mark space size , mask marking template image.

[0042] Step S22: Perform image recognition on the marks in the left and right mask pre-alignment cameras, as shown in the control software interface. Figure 4 As shown in (a) and (b), the red cross frame coordinates are the set alignment coordinates of the left and right marks of the mask and ,The process of mask image recognition algorithm is shown as follows.

[0043] 1) First, calculate the distribution map of the grayscale value of the mask mark template image, select the threshold to binarize the image to obtain the cross mark part, use the opening operation to smooth and reduce the noise of the marked area, and then use the closing operation to enhance the image features, and save the marked part as the alignment template.

[0044] 2) Detect the outline of the cross mark in the mask mark template image, save the result in a boundary point set, and then perform the minimum circumscribed matrix operation on the boundary point set to obtain the pixel side length of the mask alignment mark and .

[0045] 3) When processing the left and right mask pre-alignment camera detection images, first perform Gaussian filtering on the detection images to remove background noise, and then use a closing operation to highlight the features of the area to be inspected.

[0046] 4) Match and identify the mask marks in the detection image. First, set arrive The search angle range is 0.98 to 1.02 scale ratio, and the alignment template is used to perform matching search in the detection image, and the matching score is calculated. Then, the matching results are filtered according to the minimum matching score and the maximum matching number, and overlap filtering is performed according to the maximum overlap, and then the matching results are sorted and filtered according to the greediness to obtain the final matching result. The highest score is used as the mask alignment mark matching result, and finally the pixel center coordinates of the mask alignment mark are obtained. and And the cross template scaling and .

[0047] Step S23: Calculate the mask mark space and drive the mask stage. First, calculate the size ratio of the left and right mask pre-alignment camera pixels to the space according to equation (6): and Then, according to formula (7), the difference between the left and right alignment marks of the mask and the set alignment coordinates in the X and Y directions is calculated respectively. and , when there is When , it means that the mask has an angle at this time, and the angle value is calculated according to formula (8) Then, according to equations (2), (4), and (5), the movement amount of the mask stage R axis is calculated: And the resulting Y and X offsets and , and then calculate the movement of the X-axis and Y-axis respectively according to equations (9) and (10) and , and finally the movement amount can be input and the mask stage can be driven to move into the alignment position.

[0048] (6)

[0049] (7)

[0050] (8)

[0051] (9)

[0052] (10)

[0053] In formula (6), (7), the subscripts Left and Right are omitted; , They are The X-axis and Y-axis offsets, , They are The X and Y axis offsets.

Claims

1. A projection lithography machine mask pre-alignment method based on a motion compensation model, characterized in that: The method comprises the following steps: Step S1: measuring the R-axis motion of the mask stage through the left and right mask pre-alignment cameras of the projection lithography machine and the mask plate used for measurement, and analyzing and calculating according to the measurement results to establish a compensation model for the R-axis motion of the mask stage; Step S2: Based on the mask pre-alignment image algorithm and in combination with the mask stage R-axis motion compensation model, the pre-alignment positioning mark on the mask is identified and the adjustment amount is calculated to enter the exposure setting position; The step S1 comprises the following steps: Step S11: using a specific mask with scale values, respectively measuring the changes in the X and Y directions after the R-axis is moved in the left and right mask pre-alignment cameras, the longitudinal scale value of the specific mask is used to measure the Y-direction change after the R-axis is moved, and the transverse scale value is used to measure the X-direction change. After moving the R-axis according to a preset step distance, the X and Y coordinates of the left and right marks in the left and right mask pre-alignment cameras are recorded, and the corresponding changes are calculated; Step S12: Calculate the rotation angle of the mask according to the measured changes of the left and right side marks, and perform reverse fitting based on the corresponding relationship between the R-axis movement and the mask rotation angle to obtain a linear function relationship between the R-axis movement and the mask rotation angle; Step S13: Establishing a motion model of the mask, calculating the Y-direction and X-direction offsets according to the changes in the left and right side marks before and after the mask is rotated, and obtaining the relationship between the Y-direction and X-direction offsets and the R-axis motion through a fitting method, and constructing a mask stage R-axis motion compensation model; and Respectively represent the left and right endpoints of the mask center horizontal line, represents the midpoint of the mask center horizontal line, for The midpoint of and Respectively represent the changes in the left and right marks after rotation, and is the Y-direction offset caused by mask rotation; The offset is calculated, for Relative to The rotation angle of The length of is calculated by formula (3), where , are the mask rotation radius; （3）。 2. The method for pre-aligning a projection lithography mask based on a motion compensation model according to claim 1, characterized in that: Step S11 is further specifically as follows: a specific mask with scale values ​​has a set of measurement scale values ​​in the left and right mask pre-alignment cameras, wherein the longitudinal scale value is used to measure the Y-direction change after the R-axis moves, and the transverse scale value is used to measure the X-direction change after the R-axis moves; First determine the initial position of the R axis as , and the marks in the left and right mask pre-alignment cameras, that is, the initial Y values ​​of the left and right marks and and the initial value of X direction and , according to the step distance Move the R axis i steps, and then measure the Y and X coordinates of the left mark and the right mark after moving in the Y and X directions, and then calculate the Y and X changes of the left mark according to the formula and And the Y and X changes of the camera on the right and ; The formula for calculating the change in the Y direction of the R-axis movement of the mask stage is: The R-axis position is: ; The R-axis movement is: ; The coordinates of the left mark after moving in the Y direction are: , the coordinates of the right marker after moving in the Y direction are: ; The Y-axis change of the left mark is: ; The Y-axis change on the right is: ; The average cumulative change in the Y direction is: ; The formula for calculating the change in the X direction of the R-axis movement of the mask stage is: The R-axis position is: ; The R-axis movement is: ; The coordinates of the left mark after moving in the X direction are: ; The X-axis change on the left is: ; The coordinates of the right marker after moving in the X direction are: ; The X-axis change on the right is: ; The average cumulative change in the X direction is: .

3. The method for pre-aligning a projection lithography mask based on a motion compensation model according to claim 2, characterized in that: Step S12 is further specifically as follows: according to formula (1), the R-axis movement is calculated The angle at which the mask rotates after the distance , then according to The corresponding R-axis movement Perform reverse function fitting to obtain the R-axis motion The linear function relationship with the rotation angle is shown in formula (2); （1） （2） in, is the mask rotation radius; is the absolute value of the Y-axis change on the left. is the absolute value of the Y-axis change on the right side. , is the correlation coefficient of a linear function.

4. The method for pre-aligning a projection lithography mask based on a motion compensation model according to claim 3, characterized in that: Step S13 is further specifically as follows: compensating for the non-concentricity between the mask and the mask stage, including: calculating the corresponding errors in the Y direction and the X direction when the R axis moves, and the errors are compensated by the X axis and the Y axis of the mask stage in the mask pre-alignment part; When calculating the mask Y-direction error, use the average cumulative change in the Y direction Corresponding R-axis movement Perform reverse function fitting to obtain the relationship between the Y-axis offset and the R-axis motion, as shown in formula (4): （4） When calculating the mask X-direction error, use the average cumulative change in the X-direction The corresponding R-axis movement By performing reverse function fitting, the relationship between the X-axis offset and the R-axis motion is obtained, as shown in formula (5): （5） in, is the mask Y error, is the mask X-axis error, , , , is the correlation coefficient of a linear function; thus, the R-axis motion compensation model of the mask stage is constructed.

5. The method for pre-aligning a projection lithography mask based on a motion compensation model according to claim 4, characterized in that: The step S2 comprises the following steps: Step S21: loading mask alignment parameters, including set alignment coordinates of left and right mask marks, distance between left and right mask marks, mask mark space size, and left and right mask mark template images; Step S22: using an image algorithm to identify the left and right side marks of the mask in the left and right mask pre-alignment cameras, the image algorithm includes: firstly, binarizing and performing morphological operations on the left and right side mark template images of the mask to remove noise, and then performing a matching search on the contours of the left and right mask alignment camera detection images based on a shape matching algorithm, screening out the most suitable matching results, and obtaining the pixel side lengths and coordinates of the left and right side marks of the mask; Step S23: Calculate the X and Y offsets of the left and right side marks according to the pixel and space size ratio of the left and right mask alignment cameras. If there is an angle deviation, further calculate the movement of the mask stage R axis and the corresponding X and Y offsets.

6. The method for pre-aligning a projection lithography mask based on a motion compensation model according to claim 5, characterized in that: The step S21 is further specifically as follows: Mask alignment parameters include setting alignment coordinates for the left and right side marks of the mask and , Distance between the left and right marks of the mask , Left and right side marking space dimensions .

7. The method for pre-aligning a projection lithography mask based on a motion compensation model according to claim 6, characterized in that: The step S22 is further specifically as follows: Image recognition is performed on the left and right side marks of the mask in the left and right mask pre-alignment cameras. The mask image recognition algorithm process is as follows: Step S221: firstly, calculate the distribution map of the grayscale values ​​of the marked template images on the left and right sides of the mask, select a threshold to perform a binarization operation on the image to obtain the cross marked part, use an open operation to smooth and reduce noise on the marked area, then use a closed operation to enhance the image features, and save the marked part as the alignment template; Step S222: Detect the outline of the cross mark in the left and right side mark template images of the mask, save the result in a boundary point set, and then perform the minimum circumscribed matrix operation on the boundary point set to obtain the pixel side lengths of the left and right side marks of the mask. and ; Step 223: When processing the detection image of the left and right camera of the mask, firstly perform Gaussian filtering on the detection image to remove background noise, and then use a closing operation to highlight the features of the area to be inspected; Step 224: Match and identify the mask mark in the detection image, set the search angle range and zoom ratio, use the alignment template graphic to perform a matching search in the detection image, calculate the matching score, and then filter the matching results according to the minimum matching score and the maximum matching number, and perform overlapping filtering according to the maximum overlap, and then sort and filter the matching results according to the greediness to obtain the final matching result. The one with the highest score is used as the mask alignment mark matching result, and finally the pixel center coordinates of the left and right side marks of the mask are obtained. and and template scaling and .

8. The method for pre-aligning a projection lithography machine mask based on a motion compensation model according to claim 7, characterized in that: The step S23 is further specifically as follows: Calculate the mask mark space and drive the mask stage. First, calculate the size ratio of the left and right mask pre-alignment camera pixels to the space according to equation (6): and Then, according to formula (7), the difference between the left and right side marks of the mask and the set alignment coordinates in the X and Y directions is calculated respectively. and , when there is When , it means that the mask has an angle at this time, and the angle value is calculated according to formula (8) Then, according to equations (2), (4), and (5), the movement amount of the mask stage R axis is calculated: And the resulting Y and X offsets and , and then calculate the movement of the X-axis and Y-axis respectively according to equations (9) and (10) and , finally the movement amount can be input and the mask stage can be driven to move into the alignment position; （6） （7） （8） （9） （10） In formula (6), (7), the subscripts Left and Right are omitted; , They are The X-axis and Y-axis offsets, They are The X and Y axis offsets.

Images