Photolithography mask optimization method, system, electronic device and storage medium

Abstract

The application provides a photolithography mask optimization method, system, electronic equipment and storage medium, comprising: cutting a mask plate graph to be optimized or a corresponding target exposure graph into a plurality of polygonal subblocks, and marking each subblock, so that each plurality of subblocks with a common vertex have an internal angle sum of 360 degrees at the vertex, and each subblock corresponds to each mark in a marked combination; and cutting the target exposure graph or the mask plate graph into corresponding subblocks; for current optimization, based on the difference between the corresponding subblocks in the exposure graph corresponding to the current mask plate graph and the corresponding subblocks in the target exposure graph, determining the moving direction and moving step length of each edge segment of the graph in the subblocks of the current mark in the current mask plate graph, performing edge movement, and obtaining an optimized graph; and if the optimized graph meets an iterative convergence condition, outputting the optimized graph. The application greatly improves the optimization efficiency of photolithography masks and has strong expansibility.

Description

Technical Field

[0001] This invention belongs to the fields of image processing and semiconductor technology, and more specifically, relates to a photolithography mask optimization method, system, electronic device and storage medium. Background Technology

[0002] Spatial graphics optimization design is a problem faced in many fields, including semiconductor layout design, image processing, computer graphics, and computer-aided design. With the advancement of science and technology, the area of ​​graphics requiring optimization is gradually increasing, the graphics are becoming more complex, and their numbers are surging. This necessitates the use of various optimization algorithms to improve the scope and efficiency of graphics optimization design.

[0003] In the process of pattern optimization design, optimization algorithms are typically used to adjust the shape of the pattern edges. An objective function is set, and the edges are adjusted to meet the optimization objective, thus completing the optimization design. In processes such as layout pattern optimization design, the mask layout is the pattern to be optimized, and the optimization objective is to ensure that the spatial image light intensity distribution or the developed photoresist pattern on the silicon wafer surface conforms to the expected target. In mask optimization, using gradients to guide the step size of pattern edge movement is a typical method. However, in practice, the degree of influence of different parts of the mask layout on the optimization objective function is often inconsistent. Directly using gradient calculation results for mask edge movement can lead to the entire mask optimization process failing to converge, or even causing irreversible and significant distortion of the pattern, resulting in low optimization efficiency. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide a photolithography mask optimization method, system, electronic device and storage medium, which aims to solve the problems of poor convergence and low efficiency in the optimization process of existing photolithography mask optimization methods.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a photolithography mask optimization method, comprising the following steps:

[0006] Step S101: The mask pattern to be optimized or the corresponding target exposure pattern is cropped into multiple regular polygonal partitions, and each partition is marked so that the sum of the interior angles of each of the multiple partitions sharing a vertex is 360 degrees, and each partition corresponds to each mark in the mark combination; the target exposure pattern or the mask pattern to be optimized is cropped into the corresponding partitions; the mask pattern to be optimized is used as the current mask pattern.

[0007] Step S102: Based on the difference between each partition block in the exposure pattern corresponding to the current mask pattern and each partition block in the target exposure pattern, determine the moving direction and moving step size corresponding to each edge segment of the pattern in each currently marked partition block in the current mask pattern.

[0008] Step S103: Based on the movement direction and movement step size corresponding to each edge segment of the graphic in each currently marked partition block, perform movement optimization on each currently marked partition block in the current mask layout to obtain the optimized graphic;

[0009] Step S104: If the difference between the exposure pattern corresponding to the optimized pattern and the target exposure pattern satisfies the iterative convergence condition, then output the optimized pattern; otherwise, update the optimized pattern to the current mask pattern, update the next mark of the current mark to the current mark, and proceed to step S102.

[0010] In an optional example, step S103 is preceded by:

[0011] The movement step size corresponding to each currently marked partition block is adjusted so that the dispersion between the adjusted movement step sizes is less than a preset threshold.

[0012] In an optional example, if the blocks of the mask pattern or the target exposure pattern are obtained based on direct cropping, then the blocks of the target exposure pattern or the mask pattern are obtained based on smooth cropping.

[0013] For the direct cropping method, there are no gaps between adjacent partition blocks and no region overlap; for the smooth cropping method, there is region overlap between adjacent partition blocks, and after summing the window functions of the smooth cropping of each partition block point by point, the value of each point must be 1.

[0014] In an optional example, if the blocks of the target exposure pattern are obtained based on a smooth clipping window function, step S102 specifically includes:

[0015] The edge position error between each block in the exposure pattern and each edge segment in the corresponding block in the target exposure pattern is multiplied by the window function of the smooth clipping of the corresponding block to obtain the edge position error of each block of the current mark after modulation. Then, based on this, the moving direction and moving step size of each edge segment in the current mask pattern are calculated.

[0016] If each partition of the mask layout is obtained based on a window function with smooth clipping, step S102 specifically includes:

[0017] Based on the edge position error between edge segments in the corresponding exposure pattern of each currently marked partition block, the moving direction and moving step size of each edge segment in the currently marked partition block in the current mask pattern are solved; the moving step size of each edge segment in the currently marked partition block is multiplied by the window function of the smooth clipping of the corresponding partition block to obtain the moving step size corresponding to each edge segment in the modulated current marked partition block.

[0018] In an optional example, for the smooth cropping method, the window function for smooth cropping of any partition block is specifically obtained by convolving the window function for direct cropping of the corresponding partition block with a Gaussian function.

[0019] In an optional example, the regular polygon is preferably a square or a regular hexagon.

[0020] Secondly, the present invention provides a photolithography mask optimization system, comprising:

[0021] The image partitioning module is used to cut the mask pattern to be optimized or the corresponding target exposure pattern into multiple regular polygonal partitions, and to mark each partition so that the sum of the interior angles of each of the multiple partitions sharing a vertex is 360 degrees, and each partition corresponds to each mark in the mark combination; and to cut the target exposure pattern or the mask pattern to be optimized into the corresponding partitions; and to use the mask pattern to be optimized as the current mask pattern;

[0022] The step size determination module is used to determine the movement direction and movement step size of each edge segment of the pattern in each currently marked partition in the current mask pattern based on the difference between each partition in the current exposure pattern and each partition in the target exposure pattern.

[0023] The edge adjustment module is used to optimize the movement of each marked partition in the current mask layout based on the movement direction and movement step size corresponding to each edge segment of the graphic in each currently marked partition, so as to obtain the optimized graphic.

[0024] The convergence judgment module is used to output the optimized image if the difference between the exposure image corresponding to the optimized image and the target exposure image meets the iterative convergence condition; otherwise, it updates the optimized image to the current mask pattern, updates the next mark of the current mark to the current mark, and returns to the execution step size determination module.

[0025] Thirdly, the present invention provides an electronic device comprising: at least one memory for storing a program; and at least one processor for executing the program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to execute the method described in the first aspect or any possible implementation thereof.

[0026] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to perform the method described in the first aspect or any possible implementation thereof.

[0027] Fifthly, the present invention provides a computer program product that, when run on a processor, causes the processor to perform the method described in the first aspect or any possible implementation thereof.

[0028] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art:

[0029] This invention provides a photolithography mask optimization method, system, electronic device, and storage medium. By employing a partitioning strategy, the influence of the pattern edges within each region on the optimization objective function is made approximately the same. Pattern optimization is performed in batches according to the partitioning method, ensuring effective optimization and adjustment of all parts of the mask. Different regions within the same batch are not adjacent, avoiding coupling effects between adjacent regions, ultimately achieving an effective photolithography mask optimization process and improving optimization efficiency. Furthermore, the partitioning optimization method proposed in this invention can effectively reduce computational load, significantly improving optimization speed for large-scale pattern optimization designs, and can be flexibly extended to pattern optimization designs in more fields. Attached Figure Description

[0030] Figure 1 This is one of the flowcharts illustrating the photolithography mask optimization method provided in this embodiment of the invention;

[0031] Figure 2 This is a schematic diagram of the rectangular partitioning strategy and corresponding marking method provided in the embodiments of the present invention;

[0032] Figure 3 This is a schematic diagram of the hexagonal cellular partitioning strategy and corresponding marking method provided in an embodiment of the present invention;

[0033] Figure 4 This is a flowchart illustrating the graphical partitioning optimization design provided in an embodiment of the present invention;

[0034] Figure 5 This is a schematic diagram of the partition window function in the direct truncation and clipping scheme provided in the embodiments of the present invention;

[0035] Figure 6 This is a schematic diagram of the clipping window function in the continuous smooth clipping scheme provided in the embodiments of the present invention;

[0036] Figure 7 This is a schematic diagram illustrating the unit decomposition and spatial uniformity of the clipping window function provided in this embodiment of the invention;

[0037] Figure 8 This is a schematic diagram of the corresponding partitioning mode provided in the embodiments of the present invention;

[0038] Figure 9 This is a schematic diagram of the second corresponding partitioning mode provided in the embodiments of the present invention;

[0039] Figure 10 This is a second schematic flowchart of the photolithography mask optimization method provided in this embodiment of the invention;

[0040] Figure 11 This is an architecture diagram of the photolithography mask optimization system provided in an embodiment of the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0042] In embodiments of the present invention, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" or "for example" in embodiments of the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0043] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more, for example, multiple processing units means two or more processing units, multiple elements means two or more elements, etc.

[0044] This invention provides a method for optimizing photolithography masks. Figure 1 This is a flowchart illustrating the photolithography mask optimization method provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes the following steps:

[0045] Step S101: The mask pattern to be optimized or the corresponding target exposure pattern is cropped into multiple regular polygonal partitions, and each partition is marked so that the sum of the interior angles of each of the multiple partitions sharing a vertex is 360 degrees, and each partition corresponds to each mark in the mark combination; the target exposure pattern or the mask pattern to be optimized is cropped into the corresponding partitions; the mask pattern to be optimized is used as the current mask pattern.

[0046] Step S102: Based on the difference between each partition block in the exposure pattern corresponding to the current mask pattern and each partition block in the target exposure pattern, determine the moving direction and moving step size corresponding to each edge segment of the pattern in each currently marked partition block in the current mask pattern.

[0047] Step S103: Based on the movement direction and movement step size corresponding to each edge segment of the graphic in each currently marked partition block, perform movement optimization on each currently marked partition block in the current mask layout to obtain the optimized graphic;

[0048] Step S104: If the difference between the exposure pattern corresponding to the optimized pattern and the target exposure pattern satisfies the iterative convergence condition, then output the optimized pattern; otherwise, update the optimized pattern to the current mask pattern, update the next mark of the current mark to the current mark, and proceed to step S102.

[0049] It is understandable that the mask pattern to be optimized is the mask design that needs optimization, and the target exposure pattern is the desired exposure pattern. There is a positional correspondence between the mask pattern and the target exposure pattern. By constructing an objective function using the known target exposure pattern, the mask pattern can be iteratively corrected to ensure that the exposure pattern on the silicon wafer surface after exposure and development of the finally optimized mask pattern is close to the target exposure pattern. To improve the efficiency of iterative optimization, the mask pattern to be optimized can be set to have the same shape as the target exposure pattern. However, due to diffraction effects, the exposure pattern of the mask pattern after forward simulation will differ from the target exposure pattern, requiring further optimization.

[0050] The exposure pattern can be a spatial image or a photoresist pattern obtained by exposing and developing a mask. When performing step S101, the mask to be optimized can be cropped first, and then the target exposure pattern corresponding to the mask can be cropped according to the respective blocks of the mask. Alternatively, the target exposure pattern can be cropped first, and then the mask to be optimized can be cropped according to the respective blocks of the target exposure pattern. This embodiment of the invention does not specifically limit the specific cropping method.

[0051] Regular polygons can be, but are not limited to, equilateral triangles, squares, or regular hexagons. The number of markers in a combination can be set according to the number of multiple partitions sharing a vertex. For example, if the regular polygon is an equilateral triangle, the number of multiple partitions sharing a vertex (i.e., adjacent partitions) is 6, which can be marked as A, B, C, D, E, and F respectively; if the regular polygon is a square, the number of multiple partitions sharing a vertex is 4, which can be marked as A, B, C, and D respectively; if the regular polygon is a regular hexagon, the number of multiple partitions sharing a vertex is 3, which can be marked as A, B, and C respectively.

[0052] The difference can be specifically set according to the lithography mask evaluation index, which can include, but is not limited to, edge position error (EPE), image normalized logarithmic slope (NILS), process window, etc. The movement step size corresponding to each currently marked partition can be solved using, but is not limited to, gradient-based methods. Based on the difference, it is solved by backpropagation of each stage of the lithography model to obtain the adjoint gradient, thereby determining the movement direction and movement step size.

[0053] It should be noted that the graphics are optimized in batches according to the batch strategy corresponding to the partitioning scheme. In each iteration, only the edges of the graphics within the same batch region are adjusted, that is, only the edges of the currently labeled partition blocks are adjusted. Since each set of multiple partition blocks sharing a vertex corresponds to each label in the label combination, the partition blocks of the same label are not adjacent, which avoids mutual coupling influence. The area of ​​each partition block of the mask layout is the same, and its side length can be set according to the actual graphic size, but care must be taken not to set it too small, otherwise the partition blocks of the same label will be too close together. By subdividing the mask layout into partition blocks, it is ensured that the graphic edges in each region have a similar degree of influence on the optimization objective function.

[0054] Since the blocks of the same marker are not adjacent, the influence of the boundary is considered eliminated. Therefore, in the gradient calculation of the mask pattern edge for each region, only the corresponding target exposure pattern area is calculated. That is, only the influence of the mask pattern edge movement of each region on the corresponding area of ​​the target exposure pattern is considered, without considering the mutual influence between different regions. Compared with the existing method that needs to consider the influence of the entire mask pattern on a certain position later, this invention can significantly reduce the amount of computation and greatly improve the optimization speed for large-scale pattern optimization design.

[0055] The method provided in this invention uses a partitioning strategy to ensure that the influence of the graphic edges within each region on the optimization objective function is similar. Graphic optimization is performed in batches according to the partitioning method, allowing for effective optimization and adjustment of all parts of the mask. Different regions within the same batch are not adjacent, avoiding coupling effects between adjacent regions. This ultimately achieves an effective photolithography mask optimization process and improves optimization efficiency. Furthermore, the partitioning optimization method proposed in this invention can effectively reduce computational load, greatly improving optimization speed for large-scale graphic optimization designs, and can be flexibly extended to graphic optimization designs in more fields.

[0056] Based on the above embodiments, in practice, the degree of influence of different parts of the graph on the objective function is often inconsistent, and sometimes even differs by orders of magnitude. This inevitably leads to the following two situations: First, edges with a greater influence on the objective function can be effectively moved during optimization and eventually adjusted to the optimal position, while edges with a relatively smaller influence can almost never be moved, making it impossible to optimize and adjust them to the required position; Second, if edges with a smaller influence can be effectively moved, the step size of a single movement of edges with a larger influence will be too large, causing the entire optimization process to fail to converge, or even causing irreversible and significant distortion of the graph.

[0057] In this embodiment of the invention, the method prior to step S103 includes:

[0058] The movement step size corresponding to each currently marked partition block is adjusted so that the dispersion between the adjusted movement step sizes is less than a preset threshold.

[0059] It should be noted that for the partitions to be optimized in the current optimization batch, i.e., the currently marked partitions, if the calculated step size of each partition differs too much, even if the partitions of the same mark are not adjacent, too much movement may still cause mutual interference. Therefore, adjusting the step size of each currently marked partition to make the step size of each currently marked partition less discrete can further improve the convergence efficiency of mask optimization and obtain better optimization results.

[0060] The adjustment can be made by setting different step size modulation factors for each region based on the corresponding moving step size, thereby avoiding the problem of being unable to perform effective graphics optimization due to excessive differences in the degree of influence.

[0061] Furthermore, the movement step size of each edge segment of the graphic in each currently marked block can form a small matrix, which ultimately constitutes the overall movement matrix of the current mask layout. Specifically, the degree of dispersion between each movement step size can be measured by the condition number of the overall movement matrix.

[0062] Based on any of the above embodiments, if each partition of the mask pattern or target exposure pattern is obtained by direct cropping, then each partition of the target exposure pattern or mask pattern is obtained by smooth cropping.

[0063] For the direct cropping method, there are no gaps between adjacent partition blocks and no region overlap; for the smooth cropping method, there is region overlap between adjacent partition blocks, and after summing the window functions of the smooth cropping of each partition block point by point, the value of each point must be 1.

[0064] It is understood that the embodiments of the present invention propose two clipping schemes: direct clipping and smooth clipping. The direct clipping scheme involves directly clipping at the boundaries of each region during graphic partitioning. This means that graphic information exists within the boundaries but not outside, resulting in seamless clipping between adjacent regions. This can be achieved through geometric clipping or by multiplying the pixelated graphic representation with a two-dimensional window function for direct clipping point by point. Specifically, the graphic can be pixelated, and a set of "piecewise defined and supported truncation window functions" (hereinafter referred to as "piecewise window functions") can be selected, which are the window functions for direct clipping mentioned above. Each piecewise window function corresponds to a specific partition; the function takes a value of 1 inside the partition and a value of 0 outside the partition.

[0065] The smooth clipping scheme works as follows: each smooth clipping window function corresponds to a partition, which is multiplied point-by-point with the pixelated graphic to achieve partition selection. The function value of this window function approaches 0 when it approaches negative infinity and positive infinity, and its integral should be finite. The smooth clipping window function can be a Gaussian distribution function directly, or it can be obtained by convolving the direct clipping window function with a Gaussian function. Using this continuous and smooth clipping window function can ensure the continuity of graphic optimization and adjustment at the clipping edge, so as to avoid discontinuity of graphic edges when adjacent areas are stitched together. There will be regional overlap between the partitions obtained by smooth clipping. Compared with the corresponding partitions using direct clipping, the area size will be larger, which is equivalent to expanding outwards from the original partition, but the center position remains the same. If a steeper window function is selected for smooth clipping, the expanded area will be smaller, and if a smoother window function is selected, the expanded area will be larger.

[0066] It should be noted that although this direct cropping method is highly efficient, it can easily cause discontinuities at the edges of the graphics when adjacent areas are stitched together. Moreover, if the shape of the mask pattern is different from that of the regular polygon, the cropping edges can easily be broken. The smooth cropping method can solve this problem. Since there is a positional correspondence, only one of the two graphics, the mask pattern and the target exposure pattern, needs to use the smooth cropping method to solve the problem.

[0067] This invention divides the graphic to be optimized and the target graphic into corresponding partitions. Two different region expansion modes can be selected to establish the graphic correspondence: the mask pattern is directly cropped and the target exposure graphic is smoothly cropped, or the target exposure graphic is directly cropped and the mask pattern is smoothly cropped. Region expansion helps to take into account the influence of graphics outside each region on the graphic optimization design within the region, avoiding the optimization limitations caused by only considering the influence within the corresponding region.

[0068] This invention is not limited by graphic features and can be applied to, but is not limited to, Manhattan graphics, polygonal graphics, free-form curve graphics, etc.

[0069] Based on any of the above embodiments, if each partition of the target exposure pattern is obtained based on a window function with smooth cropping, step S102 specifically includes:

[0070] The edge position error between each block in the exposure pattern and each edge segment in the corresponding block in the target exposure pattern is multiplied by the window function of the smooth clipping of the corresponding block to obtain the edge position error of each block of the current mark after modulation. Then, based on this, the moving direction and moving step size of each edge segment in the current mask pattern are calculated.

[0071] If each partition of the mask layout is obtained based on a window function with smooth clipping, step S102 specifically includes:

[0072] Based on the edge position error between edge segments in the corresponding exposure pattern of each currently marked partition block, the moving direction and moving step size of each edge segment in the currently marked partition block in the current mask pattern are solved; the moving step size of each edge segment in the currently marked partition block is multiplied by the window function of the smooth clipping of the corresponding partition block to obtain the moving step size corresponding to each edge segment in the modulated current marked partition block.

[0073] After that, the step size is adjusted by a step size modulation factor to adjust the magnitude of the movement step size corresponding to each currently marked partition block, thereby further improving the optimization effect of each iteration and improving optimization efficiency.

[0074] It should be noted that the difference can specifically be the edge position error (EPE). Here, EPE can be the light intensity error at the edge position in the spatial image, or it can be the difference between the edge contours in the photoresist pattern. This embodiment of the invention does not specifically limit this.

[0075] Whether calculating the EPE between the overall exposure pattern corresponding to the current mask layout and the target exposure pattern, or calculating the EPE between corresponding blocks, the specific calculation method of EPE is not limited. For example, it can be to compare the area difference between a certain edge and its adjacent edge, or it can be to replace a certain edge and its adjacent edge with a point and compare the distance between the two points.

[0076] Based on any of the above embodiments, for the smooth cropping method, the window function for smooth cropping of any partition block is specifically obtained by convolving the window function for direct cropping of the corresponding partition block with a Gaussian function.

[0077] Understandably, if the mask pattern uses direct cropping, the smooth cropping window function for any block in the target exposure pattern is obtained by convolving the direct cropping window function of the corresponding block in the mask pattern with a Gaussian function; if the target exposure pattern uses direct cropping, the smooth cropping window function for any block in the mask pattern is obtained by convolving the direct cropping window function of the corresponding block in the target exposure pattern with a Gaussian function.

[0078] Based on any of the above embodiments, the regular polygon is preferably a square or a regular hexagon.

[0079] Understandably, if the regular polygon is an equilateral triangle, the number of iterations will be too many. To further improve optimization efficiency, the regular polygon is preferably a square or a regular hexagon. Figure 2 This is a schematic diagram of the rectangular partitioning strategy and corresponding marking method provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the hexagonal cellular partitioning strategy and corresponding marking method provided in an embodiment of the present invention, as shown below. Figure 2 and Figure 3 As shown, this embodiment of the invention provides two partitioning strategies to choose from: a rectangular partitioning strategy and a hexagonal honeycomb partitioning strategy. In the rectangular partitioning strategy, each partition block is a square, while in the hexagonal honeycomb partitioning strategy, each partition block is a regular hexagon.

[0080] Based on any of the above embodiments, the present invention proposes a graphical partitioning optimization design method, as well as several typical partitioning schemes and corresponding batch optimization strategies. Figure 4 This is a flowchart illustrating the graphical partitioning optimization design provided in an embodiment of the present invention. The method includes the following steps:

[0081] Step 1: Select the partitioning strategy and clipping scheme, set the relevant parameters, partition and clip the graphic to be optimized, and mark the partition blocks according to the corresponding marking method;

[0082] Step 2: Using the pre-established graphic optimization solution model, calculate the graphic edge movement reference, i.e., the movement step size, for each region with the corresponding number in each iteration, set the movement step size modulation factor for each region, modulate to obtain the final edge movement step size, and move the graphic edge accordingly.

[0083] Step 3: Based on the optimization algorithm evaluation criteria or iterative convergence conditions, determine whether the optimization meets the requirements. If yes, stop the iteration and output the optimized graph as the optimization design result; if no, return to step 2 and enter the next iteration to calculate and move each partition block with the next number. The optimization algorithm evaluation criteria are determined by setting the iteratively optimized graph in a pre-established forward model for calculation and judging whether the model's output parameters meet the requirements.

[0084] Furthermore, the partitioning strategy proposed in this invention can employ rectangular partitioning or hexagonal honeycomb partitioning, such as... Figure 2 and Figure 3 As shown, the specific partitioning and marking schemes, and the corresponding batching strategies are as follows:

[0085] The rectangular partitioning scheme is as follows: Figure 2 As shown, set the side length of the rectangle. d The distance between the center points of adjacent regions is d The graphic to be optimized is divided into sections according to the arrangement of rectangles. According to the marking method shown in the figure, four adjacent rectangles are marked as A, B, C, and D respectively. During the iteration process, the order is A→B→C→D→A→..., and each iteration only optimizes the edges of the graphic in the same marked area. That is, the first time only the graphic in each area A is moved, the second time only the graphic in each area B is moved, and so on.

[0086] The hexagonal cellular partitioning scheme is as follows: Figure 3 As shown, set the side length of the regular hexagon. d The distance between the center points of adjacent areas is The pattern is divided according to the honeycomb structure; according to the marking method shown in the figure, three adjacent hexagons are marked as A, B, and C respectively; the iteration process is carried out in the order of A→B→C→A→..., and each iteration only optimizes the edge of the pattern in the same marked area.

[0087] Furthermore, the cutting scheme proposed in this invention can employ seamless cutting by directly severing at the boundary, or continuous smooth cutting. Specific cutting schemes are as follows:

[0088] The seamless clipping scheme that directly truncates at the boundary involves directly truncating at the boundary of each region when partitioning the graphic. This means that graphic information exists within the boundary but not outside. Optionally, this scheme can be implemented by directly clipping the original graphic using geometric clipping; alternatively, it can be implemented by multiplying a pixelated graphic with a two-dimensional rectangular window function, i.e., a partitioning truncation window function, point by point. Figure 5 This is a schematic diagram of the partition window function in the direct truncation and clipping scheme provided in this embodiment of the invention, as shown below. Figure 5 As shown, the partitioned truncated window function can be expressed as For each region, a corresponding partition truncation window function is set, and the function is multiplied point by point with the pixelated graphic to achieve partition clipping.

[0089] The continuous smooth clipping scheme uses a continuously smooth window function as the clipping function. For each partition, a corresponding window function is set with the geometric center coordinates of the region as the function center. This window function is then multiplied point-by-point with the pixelated representation of the graphic to achieve partition selection. The clipping window function's value approaches 0 as it approaches negative and positive infinity, and its integral should be finite. For example, but not limited to, a Gaussian-filtered partition truncation window function can be used as the selected clipping window function.

[0090] It should be noted that, Figure 6 This is a schematic diagram of the clipping window function in the continuous smooth clipping scheme provided in the embodiments of the present invention. Figure 7 This is a schematic diagram illustrating the unit decomposition and spatial uniformity of the clipping window function provided in an embodiment of the present invention, as shown below. Figure 6 and Figure 7 As shown, the selected set of smooth clipping window functions is a unit decomposition function based on spatial distribution. It needs to satisfy overall spatial uniformity, meaning that after summing the set of smooth clipping window functions point-by-point, its value is the same at all points in space. When using this type of function for clipping, the influence weight value of each point in each region of the graphic is obtained by applying the value of the clipping window function corresponding to that region at that point. In other words, by applying the clipping window function, the overlapping parts of adjacent regions are decomposed into units. Therefore, excessive or insufficient movement of the graphic edges in overlapping regions can be avoided during smooth clipping.

[0091] Furthermore, in processes such as layout pattern optimization design, the mask layout is the pattern to be optimized, and the optimization goal is to ensure that the exposed pattern on the silicon wafer surface after exposure and development meets the expected target. In this optimization design process, the objective function is based on the exposed pattern, which is also a spatial pattern, called the optimization target pattern, and this pattern has a positional correspondence with the layout pattern.

[0092] To address this type of problem, optionally, the graphic to be optimized and the target graphic can be partitioned into corresponding regions. During the calculation process, for each region's edge gradient calculation of the graphic to be optimized, only the corresponding target graphic region is calculated; that is, only the impact of the edge movement of the graphic to be optimized within each region on the corresponding region of the target graphic is considered. Furthermore, the partitioning of the graphic to be optimized and the target graphic into corresponding regions can employ two different region expansion modes to establish the graphic correspondence.

[0093] Figure 8 This is a schematic diagram of the corresponding partitioning mode provided in the embodiments of the present invention, such as... Figure 8 As shown, in the first corresponding partitioning mode, the image to be optimized 101 is seamlessly cropped, and the target image 102 is continuously and smoothly cropped accordingly. Therefore, adjacent regions in the target image 102 have some overlap. The cropping window function used for the continuous smooth cropping is set to a smooth cropping window function whose center coincides with the center of the seamless cropping region of the image to be optimized. Accordingly, each partition of the image to be optimized corresponds to an extended region on the target image, for example... Figure 8 As shown, the dashed extended region 104 in the target optimization region 102 corresponds to one of the A partitions 103 in the target optimization region 101. For each extended region, when calculating the corresponding moving reference of the target optimization region, the EPE value taken on the target optimization region should be the point-by-point multiplication of the original EPE0 of that region with the smooth clipping window function. W i That is, EPE = EPE0⊙ W i , where ⊙ represents point-by-point multiplication.

[0094] Figure 9 This is a schematic diagram of the second corresponding partitioning mode provided in the embodiments of the present invention, such as... Figure 9 As shown, in the second corresponding partitioning mode, the target graphic 102 is seamlessly cropped, while the graphic to be optimized 101 is continuously and smoothly cropped. Therefore, adjacent regions in the target graphic 102 have some overlap. The cropping window function used for the continuous smooth cropping is set to a smooth cropping window function whose center coincides with the center of the seamless cropping region of the target graphic. Accordingly, each partition of the target graphic corresponds to an extended region on the graphic to be optimized, for example... Figure 9 As shown, one of the A partitions 106 in the target graphic 102 corresponds to the dashed extended region 105 in the graphic to be optimized 101. For each extended region, after solving for the initial moving reference δM0 of the graphic to be optimized, it is compared with the smooth clipping window function of that region. W i By multiplying each point, we obtain the movement reference δM = δM0⊙ for this region. W i, where ⊙ represents point-by-point multiplication.

[0095] Further, as described in step 2, setting the step-size modulation factor for each region and moving the edge of the graphic according to the modulation result, wherein the step-size modulation factor... β The value is set based on the moving reference values ​​at various points along the edges of each region's graphic, and the corresponding step size modulation factor for each region is... β The values ​​can vary. The moving references at each edge of the graphic in each region are calculated based on a pre-established graphic optimization solution model, and can be, but is not limited to, a gradient-based solution method.

[0096] Based on any of the above embodiments, the partitioning optimization method proposed in this invention is applicable to the optimization design process of graphics. The following uses the optimization of mask layout in the field of semiconductor technology as an example to illustrate the optimization method proposed in this invention.

[0097] The graphic partitioning optimization method described in this invention is used to optimize the mask layout. Figure 10 This is the second flowchart illustrating the photolithography mask optimization method provided in this embodiment of the invention. The specific implementation steps are as follows:

[0098] Step S1: Select a rectangular or hexagonal partitioning strategy and set the side length parameters of the rectangle or hexagon. d .

[0099] Step S2: Select the corresponding partition mode, where the mask pattern M is the graphic to be optimized, and the exposure graphic W calculated by exposure and development simulation is the target optimized graphic. Select the first (or second) corresponding partition mode.

[0100] Step S3: According to the corresponding partitioning mode selected in step S2 and the partitioning strategy and partitioning geometry parameters in step S1, seamlessly crop the mask pattern M (or target exposure pattern W) and mark each area with batch number.

[0101] Step S4: Based on the cropping in step S3, set the corresponding smooth cropping window function to smoothly crop the target exposure pattern W (or mask pattern M), and mark each extended area with the corresponding batch number.

[0102] Step S5: According to the partition correspondence between the mask layout M and the target exposure pattern W, the mask movement of each region of the same batch number is solved with reference to δM.

[0103] Preferably, the solution method may employ, but is not limited to, a gradient-based solution method. It should be noted that when calculating the mask movement reference δM for each region using a gradient-based solution method, the corresponding calculated value needs to be modulated by a smoothing clipping window function.

[0104] Preferably, the gradient-based solution method can employ the adjoint gradient solution method. The adjoint gradient solution method can be calculated in, but is not limited to, the following two ways: First, after simulating the photolithography exposure and development process in a forward simulation, edge contour points or contour curves are extracted from the simulated exposure pattern. The EPE is calculated by comparing the simulated contour points or contour curves with the corresponding target points or target curves on the target exposure pattern. In this method, the adjoint gradient is obtained through the reverse propagation of the EPE through each stage of the photolithography model, serving as the mask movement reference δM. Second, after the forward photolithography simulation, simulated signal values ​​(i.e., light intensity values) are extracted at the position of a predetermined target point, and signal errors are calculated by comparing the simulated signal values ​​with a predetermined threshold. The signal errors are obtained through the reverse propagation of the photolithography model through each stage, serving as the mask movement reference δM.

[0105] Step S6: Based on the mask layout of each partition modulated by the smooth clipping window function, move the reference δM and set the step size modulation factor for that partition. β Update and adjust the mask layout graphic: M = M + β δM. The step size modulation factor for each region. β The settings can be configured based on the value of δM in that region, with the goal of aligning each region... β The overall movement matrix condition number of the mask layout composed of δM is better.

[0106] Step S7: Using the mask pattern updated in step S6, i.e., the optimized pattern, perform simulated exposure and development calculations to obtain the simulated exposure pattern W. ’ And calculate the edge position error EPE = W between each location and the target exposure pattern W. ’ - W.

[0107] Step S8: Determine whether the iterative convergence condition is met. If yes, proceed to step S9; otherwise, proceed to step S10. The iterative convergence condition can be one or more of the following: the weighted sum of the simulated exposure and development EPE of each edge segment of the entire mask pattern is less than a certain set value; the EPE of each edge segment is less than a certain set value; the number of iterations reaches a set maximum value, etc. This invention is not limited to the setting of the iterative convergence condition.

[0108] Step S9: Stop the iteration and output the current mask layout as the optimization result.

[0109] Step S10: Jump to step S5 and proceed to the next iteration to optimize and adjust each region of the next batch number.

[0110] Based on any of the above embodiments, the present invention provides a photolithography mask optimization system. Figure 11This is an architecture diagram of the photolithography mask optimization system provided in an embodiment of the present invention, as shown below. Figure 11 As shown, the system includes:

[0111] The image partitioning module 1110 is used to cut the mask pattern to be optimized or the corresponding target exposure pattern into multiple regular polygonal partitions, and to mark each partition so that the sum of the interior angles of each of the multiple partitions sharing a vertex is 360 degrees, and each partition corresponds to each mark in the mark combination; and to cut the target exposure pattern or the mask pattern to be optimized into the corresponding partitions; and to use the mask pattern to be optimized as the current mask pattern.

[0112] The step size determination module 1120 is used to determine the moving direction and moving step size corresponding to each edge segment of the pattern in each currently marked partition in the current mask pattern based on the difference between each partition in the current exposure pattern and each partition in the target exposure pattern.

[0113] The edge adjustment module 1130 is used to optimize the movement of each marked partition block in the current mask layout based on the movement direction and movement step size corresponding to each edge segment of the graphic in each currently marked partition block, so as to obtain the optimized graphic.

[0114] The convergence judgment module 1140 is used to output the optimized image if the difference between the exposure image corresponding to the optimized image and the target exposure image meets the iterative convergence condition; otherwise, it updates the optimized image to the current mask pattern, updates the next mark of the current mark to the current mark, and returns to the execution step size determination module 1120.

[0115] It is understood that the detailed functional implementation of each of the above modules can be found in the description of the aforementioned method embodiments, and will not be repeated here.

[0116] Based on the methods described in the above embodiments, this invention provides an electronic device. The device may include at least one memory for storing a program and at least one processor for executing the program stored in the memory. When the program stored in the memory is executed, the processor performs the methods described in the above embodiments.

[0117] Based on the methods in the above embodiments, this embodiment of the invention provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0118] Based on the methods in the above embodiments, this embodiment of the invention provides a computer program product that, when run on a processor, causes the processor to execute the methods in the above embodiments.

[0119] It is understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The general-purpose processor can be a microprocessor or any conventional processor.

[0120] The method steps in these embodiments of the invention can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can reside in an ASIC.

[0121] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0122] It is understood that the various numerical designations used in the embodiments of the present invention are merely for the convenience of description and are not intended to limit the scope of the embodiments of the present invention.

[0123] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A photolithography mask optimization method, characterized by, Includes the following steps: Step S101: Cut the mask pattern to be optimized or the corresponding target exposure pattern into multiple regular polygonal partitions and mark each partition so that the sum of the interior angles of each of the multiple partitions sharing a vertex is 360 degrees and each partition corresponds to each mark in the mark combination. The target exposure pattern or the mask pattern to be optimized is then cropped into corresponding partition blocks; the mask pattern to be optimized is then used as the current mask pattern. Step S102: Based on the difference between each partition block in the exposure pattern corresponding to the current mask pattern and each partition block in the target exposure pattern, determine the moving direction and moving step size corresponding to each edge segment of the pattern in each currently marked partition block in the current mask pattern. Step S103: Based on the movement direction and movement step size corresponding to each edge segment of the graphic in each currently marked partition block, perform movement optimization on each currently marked partition block in the current mask layout to obtain the optimized graphic; Step S104: If the difference between the exposure pattern corresponding to the optimized pattern and the target exposure pattern satisfies the iterative convergence condition, then output the optimized pattern; otherwise, update the optimized pattern to the current mask pattern, update the next mark of the current mark to the current mark, and proceed to step S102. The target exposure pattern or the mask layout to be optimized is cropped into corresponding partition blocks, including any of the following: The first corresponding partitioning mode: seamlessly crop the graphic to be optimized, and continuously and smoothly crop the target graphic to make adjacent regions in the target graphic partially overlap; the cropping window function used for continuous smooth cropping is set to a smooth cropping window function whose center coincides with the center of the seamless cropping region of the graphic to be optimized; each partition of the graphic to be optimized corresponds to an extended region on the target graphic; for each extended region, the edge placement error value taken on the target graphic is obtained by multiplying the original edge placement error value of the region by the smooth cropping window function point by point; The second corresponding partitioning mode: seamlessly cropping the target graphic and continuously and smoothly cropping the graphic to be optimized, so that there is partial overlap between adjacent regions in the graphic to be optimized; the cropping window function used for continuous smooth cropping is set to a smooth cropping window function whose center coincides with the center of the seamless cropping region of the target graphic; each partition of the target graphic corresponds to an extended region on the graphic to be optimized; for each extended region, an initial movement reference of the graphic to be optimized is determined, and the initial movement reference is multiplied point by point by the smooth cropping window function of the region to obtain the movement reference of the region; Wherein, the graphic to be optimized is a mask pattern, and the target graphic to be optimized is a target exposure graphic.

2. The method of claim 1, wherein, The steps preceding step S103 also include: Adjust the movement step size corresponding to each currently marked partition block so that the dispersion between the adjusted movement step sizes is less than a preset threshold.

3. The method of claim 1, wherein, If each partition of the mask pattern or target exposure pattern is obtained by direct cropping, then each partition of the target exposure pattern or mask pattern is obtained by smooth cropping. For the direct cropping method, there are no gaps between adjacent partition blocks and no region overlap; for the smooth cropping method, there is region overlap between adjacent partition blocks, and after summing the window functions of the smooth cropping of each partition block point by point, the value of each point must be 1.

4. The method of claim 3, wherein, If each partition of the target exposure pattern is obtained based on a window function with smooth cropping, step S102 specifically includes: The edge position error between each block in the exposure pattern and each edge segment in the corresponding block in the target exposure pattern is multiplied by the window function of the smooth clipping of the corresponding block to obtain the edge position error of each block of the current mark after modulation. Then, based on this, the moving direction and moving step size of each edge segment in the current mask pattern are calculated. If each partition of the mask layout is obtained based on a window function with smooth clipping, step S102 specifically includes: Based on the edge position error between edge segments in the corresponding exposure pattern of each currently marked partition block, the moving direction and moving step size of each edge segment in the currently marked partition block in the current mask pattern are solved; the moving step size of each edge segment in the currently marked partition block is multiplied by the window function of the smooth clipping of the corresponding partition block to obtain the moving step size corresponding to each edge segment in the modulated current marked partition block.

5. The method of claim 3, wherein, For the smooth cropping method, the window function for smooth cropping of any partition block is specifically obtained by convolving the window function for direct cropping of the corresponding partition block with a Gaussian function.

6. The method according to any one of claims 1 to 5, characterized in that, The regular polygon is a square or a regular hexagon.

7. A photolithography mask optimization system, characterized in that, include: The graphic partitioning module is used to cut the mask pattern to be optimized or the corresponding target exposure pattern into multiple regular polygonal partitions and to mark each partition so that the sum of the interior angles of each of the multiple partitions sharing a vertex is 360 degrees and each partition corresponds to each mark in the mark combination. The target exposure pattern or the mask pattern to be optimized is then cropped into corresponding partition blocks; the mask pattern to be optimized is then used as the current mask pattern. The step size determination module is used to determine the movement direction and movement step size of each edge segment of the pattern in each currently marked partition in the current mask pattern based on the difference between each partition in the current exposure pattern and each partition in the target exposure pattern. The edge adjustment module is used to optimize the movement of each marked partition in the current mask layout based on the movement direction and movement step size corresponding to each edge segment of the graphic in each currently marked partition, so as to obtain the optimized graphic. The convergence judgment module is used to output the optimized image if the difference between the exposure image corresponding to the optimized image and the target exposure image meets the iterative convergence condition; otherwise, the optimized image is updated to the current mask pattern, the next mark of the current mark is updated to the current mark, and the execution step size determination module is returned. The target exposure pattern or the mask layout to be optimized is cropped into corresponding partition blocks, including any of the following: The first corresponding partitioning mode: seamlessly crop the graphic to be optimized, and continuously and smoothly crop the target graphic to make adjacent regions in the target graphic partially overlap; the cropping window function used for continuous smooth cropping is set to a smooth cropping window function whose center coincides with the center of the seamless cropping region of the graphic to be optimized; each partition of the graphic to be optimized corresponds to an extended region on the target graphic; for each extended region, the edge placement error value taken on the target graphic is obtained by multiplying the original edge placement error value of the region by the smooth cropping window function point by point; The second corresponding partitioning mode: seamlessly cropping the target graphic and continuously and smoothly cropping the graphic to be optimized, so that there is partial overlap between adjacent regions in the graphic to be optimized; the cropping window function used for continuous smooth cropping is set to a smooth cropping window function whose center coincides with the center of the seamless cropping region of the target graphic; each partition of the target graphic corresponds to an extended region on the graphic to be optimized; for each extended region, an initial movement reference of the graphic to be optimized is determined, and the initial movement reference is multiplied point by point by the smooth cropping window function of the region to obtain the movement reference of the region; Wherein, the graphic to be optimized is a mask pattern, and the target graphic to be optimized is a target exposure graphic.

8. An electronic device, characterized in that, include: At least one memory for storing programs; At least one processor is configured to execute a program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is run on the processor, it causes the processor to perform the method as described in any one of claims 1-6.

10. A computer program product, characterized in that, When the computer program product is run on a processor, the processor causes the processor to perform the method as described in any one of claims 1-6.

Images