A method and apparatus for generating a curved mask pattern

Abstract

The application discloses a method and device for generating a curved mask pattern. Each time the processed Manhattan layout is moved, it corresponds to a unique equivalent curve layout. In the optimization iteration process, the equivalent curve layout is used to replace the Manhattan layout for lithography imaging simulation. When the optimization converges, the current segmented edge movement processed Manhattan layout is executed for curve processing to generate the final curved mask pattern as the optimized OPC result. In the embodiment of the application, the Manhattan pattern is always used as the optimization object in the optimization iteration process, but the equivalent curve layout is used as the input in the simulation stage. In this way, the control difficulty caused by directly modifying the curve layout is avoided, the curved mask pattern is stably and efficiently generated while the stability of the traditional OPC optimization is maintained, the curved mask result meeting the EPE constraint is obtained, and the wafer pattern transfer precision and the chip manufacturing yield are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to, but is not limited to, the semiconductor integrated circuit technology, and in particular to a method and device for generating a curved mask pattern. BACKGROUND

[0002] With the continuous advancement of advanced processes, the imaging error from layout to wafer is significantly increased, and the edge position error (EPE) and critical dimension (CD) drift caused by optical proximity effect (OPE) become the core bottleneck of yield.

[0003] The existing optical proximity correction (OPC) technology mostly takes Manhattan pattern as the optimization object, and it is difficult to generate a curved layout that meets the high-precision process requirements, resulting in limited correction accuracy at corner points and complex structures, and easy to produce mask rule check (MRC) violations. Although inverse lithography technology (ILT) can obtain a curved result, it has large calculation amount, is difficult to control and has high cost. How to stably and efficiently generate a curved mask pattern is a technical problem to be solved. SUMMARY

[0004] The present application provides a method and device for generating a curved mask pattern, which can stably and efficiently generate a curved mask pattern, and improve wafer pattern transfer accuracy and chip manufacturing yield.

[0005] The present application provides a method for generating a curved mask pattern, comprising:

[0006] Performing boundary analysis on the Manhattan layout to obtain the geometric properties of the corner points, and segmenting each geometric edge of the Manhattan layout to obtain a segmented edge;

[0007] Setting a sampling point for edge position error (EPE) evaluation on each segmented edge to generate a set of evaluation points;

[0008] Performing movement processing on the segmented edges of the Manhattan layout;

[0009] Performing curve processing according to the segmented edge information after movement processing and the geometric properties of the corner points to generate an equivalent curved layout corresponding to the Manhattan layout after movement processing of the segmented edges;

[0010] Performing lithography imaging simulation on the equivalent curved layout to calculate the EPE between the imaging profile and the target layout at each evaluation point in the set of evaluation points;

[0011] Determining whether the optimization end condition is met according to the obtained EPE value, and when the optimization end condition is met, proceeding to the next step; when the optimization end condition is not met, returning to the step of performing movement processing on the segmented edges of the Manhattan layout;

[0012] The Manhattan layout is curved to generate a curved mask layout.

[0013] In an exemplary embodiment, the Manhattan layout is a mask layout mainly in Manhattan geometry, which is obtained after a circuit layout design and an optical proximity correction (OPC) preprocessing.

[0014] In an exemplary embodiment, for each geometric edge with a length greater than a preset threshold, the geometric edge is segmented according to a segment interval determined according to a lithography resolution and a target CD.

[0015] For a geometric edge with a complex shape or an edge close to a lithography hotspot structure, the geometric edge is segmented according to a segment interval adaptively reduced according to a local shape complexity.

[0016] In an exemplary embodiment, the boundary analysis of the Manhattan layout to obtain geometric properties of the corner points includes:

[0017] The boundary analysis of the Manhattan layout is performed to identify a connection relationship of each geometric edge and obtain geometric properties of the corner points formed by adjacent geometric edges.

[0018] The geometric properties of each corner point include a concave-convex type, an internal angle, and adjacent edge length information.

[0019] In an exemplary embodiment, the generation of the evaluation point set includes:

[0020] On each segment edge, a set of sampling points is arranged to form a sampling point set of the segment edge; the sampling points are arranged along a normal direction or a tangent direction of the segment edge, and are used to calculate an EPE value on the segment edge.

[0021] A union set of the sampling point sets of all the segment edges constitutes the evaluation point set.

[0022] In an exemplary embodiment, the method further includes: performing encrypted sampling at the corner point or the hotspot structure.

[0023] In an exemplary embodiment, the moving processing of the segment edges of the Manhattan layout includes:

[0024] According to an EPE result obtained by a previous lithography imaging simulation, each segment edge of the Manhattan layout is subjected to geometric position updating.

[0025] For an initial entry into optimization, a lithography simulation is performed with the Manhattan layout as an input to obtain an initial imaging profile; an EPE between the initial imaging profile and a target layout is calculated at each evaluation point; and a position of the segment edge is adjusted along a normal direction of the segment edge according to a sign and a size of the obtained EPE.

[0026] In an example, the generating the equivalent curve layout corresponding to the Manhattan layout after the moving processing of the segmented edges comprises:

[0027] The curve parameter is determined using the geometric information of the segmented edges and the geometric attribute of the corner point, and a smooth and continuous curve boundary is generated to constitute the equivalent curve layout corresponding to the Manhattan layout after the moving processing of the segmented edges.

[0028] In an example, the optimization end condition comprises any one of the following:

[0029] The absolute value of all EPEs is less than a preset threshold; or,

[0030] The change of the optimization objective function is less than a convergence threshold; or,

[0031] The maximum number of iterations is reached.

[0032] Embodiments of the present application also provide a computer readable storage medium storing computer executable instructions for executing the method for generating the curve mask pattern.

[0033] Embodiments of the present application further provide a computer device comprising a memory and a processor, wherein the memory stores instructions executable by the processor, for executing the steps of the method for generating the curve mask pattern.

[0034] Embodiments of the present application also provide a device for generating a curve mask pattern, comprising a first processing module, a generating module, a second processing module, a first curve module, a simulation module, and a second curve module, wherein:

[0035] The first processing module is configured to analyze the geometric attribute of the corner point by boundary analysis on the Manhattan layout, and to obtain the segmented edges by segmenting each geometric edge of the Manhattan layout;

[0036] The generating module is configured to set sampling points for EPE evaluation on each segmented edge, and to generate an evaluation point set;

[0037] The second processing module is configured to perform moving processing on the segmented edges of the Manhattan layout;

[0038] The first curve module is configured to perform curve processing according to the information of the moving processed segmented edges and the geometric attribute of the corner point, and to generate the equivalent curve layout corresponding to the Manhattan layout after the moving processing of the segmented edges;

[0039] The simulation module is configured to perform lithography imaging simulation on the equivalent curve layout, and to calculate the EPE between the imaging profile and the target layout at each evaluation point in the evaluation point set.

[0040] a judging module, configured to judge whether an optimization end condition is met according to the obtained EPE value, and execute the generating module when the optimization end condition is met; and return to the second processing module when the optimization end condition is not met;

[0041] a second curving module, configured to perform curving on the Manhattan layout after the current segment edge movement processing, to generate a curving mask pattern.

[0042] The method for generating a curving mask pattern provided in the embodiments of the present application has a unique equivalent curving layout corresponding to the Manhattan layout after each edge movement processing, and the shape of the equivalent curving layout can be flexibly controlled by parameters. In the optimization iteration process, the equivalent curving layout is used to replace the Manhattan layout to perform photolithography imaging simulation. When the optimization converges, curving processing is performed on the Manhattan layout after the current segment edge movement processing, to generate a final curving mask pattern as the optimized OPC result. In the embodiments of the present application, the Manhattan layout is always taken as the optimization object in the optimization iteration process, but the equivalent curving layout is taken as the input in the simulation stage. In this way, the control difficulty caused by directly modifying the curving layout is avoided, the curving mask pattern is stably and efficiently generated while the stability of the traditional OPC optimization is maintained, the curving mask result meeting the EPE constraint is obtained, and the wafer pattern transfer precision and the chip manufacturing yield are improved.

[0043] The method for generating a curving mask pattern provided in the embodiments of the present application, on one hand, extends the photolithography imaging simulation and output to a curving-friendly form without changing the traditional OPC mainstream process and edge movement mechanism, so that the optical effect of the curvature radius and the corner smoothness can directly participate in the optimization decision; on the other hand, a parameterized and adjustable curving channel is provided, so that the method can be quickly reconfigured according to the changes of the process, equipment and mask process, without introducing a new ILT infrastructure; and on the other hand, the optimization stability is maintained while the MRC violation and post-processing cost are reduced, and the repeated process of repairing and re-violation after optimization is avoided.

[0044] Additional features and advantages of the application are set forth in the description that follows, and in part will become apparent from the description, or can be learned by practice of the application. The objectives and other advantages of the application will be realized and attained by the structure particularly pointed out in the description and claims. BRIEF DESCRIPTION OF DRAWINGS

[0045] The accompanying drawings are included to provide a further understanding of the technical scheme of the present application, and constitute a part of the specification, and are used to explain the technical scheme of the present application together with the embodiments of the present application, and do not constitute a limitation to the technical scheme of the present application.

[0046] Figure 1 A flowchart of a method for generating a curved mask pattern in an embodiment of the present application is shown in FIG. 1.

[0047] Figure 2 A block diagram of an apparatus for generating a curved mask pattern in an embodiment of the present application is shown in FIG. 2.

[0048] Figure 3 A diagram of a simulation result of a mask pattern in an embodiment of the present application is shown in FIG. 3. DETAILED DESCRIPTION

[0049] To make the objects, technical solutions and advantages of the present application clearer, the embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other in any manner without conflict.

[0050] In a typical configuration of the present application, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0051] The memory can include non-persistent memory in computer readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer readable media.

[0052] Computer readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology for storing information. The information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape disk storage or other magnetic storage device, or any other non-transmission medium that can be used to store information accessible by a computing device. According to the definition herein, computer readable media does not include non-transitory computer readable media (transitory media), such as modulated data signals and carriers.

[0053] The steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions. Also, although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0054] The boundary of the curve is continuously changing, and it is difficult to directly disassemble into simple edges. In the photolithography simulation process, calculating the accurate reflection and diffraction characteristics of the curve will greatly increase the calculation cost. The design of Edge-based OPC is based on linear geometric boundaries rather than continuous curves. The curve needs to be converted into a polygon composed of a finite number of line segments before it can be recognized by the standard OPC tool. In order to adapt to Edge-based OPC, the curve graph needs to be converted into a Manhattan polygon through approximation, so that Edge-based OPC can be used for Manhattan polygon, thereby realizing OPC on the curve layout.

[0055] Traditional OPC takes geometric edges, evaluation points and edge movement as the main line of optimization, and iteratively converges with the tight coupling of the compact model of imaging / photoresist. But its natural assumption is that the input and output are mainly Manhattan (orthogonal) polygons, and the shape generated after the edge moves is full of sawtooth / steps. Under high numerical aperture, strong non-linear photoresist and strict MRC constraints, violations are often triggered at corner points / narrow slits / sharp turns. MRC is a set of rules for mask manufacturing processes, used to limit the minimum line width, spacing and shape processing range of the mask. If these rules are violated, the mask may not be manufactured under current process conditions, or the yield is very low. ILT can naturally obtain curve / free shape from the image domain, and the graph quality is high, but the optimization space is extremely large, the control granularity is dispersed, and the calculation cost and engineering controllability are poor. Therefore, engineers often still need a lot of post-processing (smooth, denoise, MRC repair), and the coupling with the existing OPC tool chain is not smooth. Therefore, the industry is faced with a dilemma: either stick to stable Manhattan OPC and give up shape flexibility, or use high-cost ILT and sacrifice controllability and computing power / time. If a method can continue to use the mainstream process of OPC based on edge movement-simulation-evaluation, but obtain the benefits of curve layout at the simulation and output level, then both the engineering controllability and cost advantage are retained, and the imaging and MRC friendliness at high-end nodes are significantly improved.

[0056] Therefore, the embodiment of the present application provides a method for generating a curved mask graph, as shown in Figure 1 The method can include the following steps:

[0057] Step 100: Analyzing the boundary of the Manhattan layout to obtain the geometric properties of the corner points, and segmenting each geometric edge of the Manhattan layout to obtain a segmented edge.

[0058] The Manhattan layout is a mask layout mainly based on Manhattan geometry, and is generated in the circuit design or basic OPC stage. The Manhattan layout is an initial mask pattern obtained after conventional circuit layout design and OPC preprocessing, and is usually geometric data in GDSII / OASIS format.

[0059] In an example, the segmenting each geometric edge to obtain a segment edge can include:

[0060] For a longer edge (with a length greater than a preset threshold), the segment interval can be determined according to the lithography resolution and the target CD. In an embodiment, the size of the segment interval can be determined according to the lithography resolution λ / NA or the target EPE sampling requirement, where λ is the wavelength of the light source, and NA is the numerical aperture reflecting the resolution capability of the optical system. Preferably, the segment interval is smaller than the resolution limit corresponding to λ / NA or the target EPE sampling interval.

[0061] For an edge with complex geometric shape or an edge close to a lithography hotspot structure, the segment interval can be adaptively reduced according to the local shape complexity, so as to improve the local optimization accuracy. The edge with complex geometric shape can include an edge with a sharp corner, a concave-convex transition, or a significant line width or interval change. The edge close to the lithography hotspot structure refers to an edge adjacent to a high EPE sensitive structure identified in the historical library or simulation analysis. In the embodiment of the present application, the segment interval is encrypted in the local neighborhood around the edge, which helps to improve the simulation evaluation accuracy and improve the overall optimization stability.

[0062] In an example, the boundary analysis of the Manhattan layout to obtain the geometric attributes of the corner point includes: performing boundary analysis on the Manhattan layout, identifying the connection relationship of each geometric edge, and obtaining the geometric attributes of the corner point formed by the adjacent geometric edges. In an embodiment, the geometric attributes of each corner point can include the concave-convex type, the interior angle, and the length information of the adjacent edges. This information is used to determine the rounding direction, the curvature radius, and the spline control range in the subsequent curve processing, so as to realize the parameterized and controllable curve generation.

[0063] Step 101: Set a sampling point for EPE evaluation on each segment edge to generate an evaluation point set.

[0064] On each segment edge, a set of sampling points is set, which can be arranged along the normal or tangential direction of the segment edge, for calculating the EPE value on the segment edge. A set of sampling points is referred to as a sampling point set of the segment edge.

[0065] The union set of the sampling point sets of all segment edges constitutes a global evaluation point set, which is used for subsequent EPE calculation and simulation evaluation.

[0066] In an example, step 101 can further include:

[0067] At the corner or hot spot structure (such as line end, island, narrow pitch area), the encryption sampling is performed, that is, the interval between adjacent sampling points is reduced, so that more sampling points are arranged to capture the imaging error change more finely.

[0068] Step 102: Moving processing is performed on the segmented edges of the Manhattan layout.

[0069] In an exemplary example, the moving processing performed on the segmented edges of the Manhattan layout can include:

[0070] According to the EPE result obtained from the last round of lithography imaging simulation, the geometric position update is performed on each segmented edge of the initial Manhattan layout in step 100. In an embodiment, the moving direction and step along the normal of the segmented edge are determined by the sign and size of the EPE, and the optimization methods such as gradient descent, quasi-Newton or linear programming can be used to calculate the optimal moving amount. After moving, a new geometric edge structure is generated, and the topology is kept unchanged. In this step, the EPE is reduced by adjusting the position of the segmented edge, and the matching of the current imaging profile and the target layout is realized.

[0071] In the embodiment of the present application, when initially entering the cycle optimization (i.e. the first moving processing after the first lithography imaging simulation), the initial Manhattan layout is input to perform lithography simulation to obtain the initial imaging profile; the EPE between the initial imaging profile and the target layout is calculated at each evaluation point; and the position of the segmented edge is adjusted along the normal of the segmented edge according to the sign and size of the EPE.

[0072] Step 103: Curve processing is performed according to the moving processed segmented edge information and the geometric properties of the corner points to generate an equivalent curve layout corresponding to the Manhattan layout after the moving processing of the segmented edges.

[0073] In an exemplary example, the curve processing parameter is determined using the geometric information of the segmented edge and the geometric properties (concave-convex type, interior angle and adjacent edge length) of the corner points, and a smooth and continuous curve boundary is generated to constitute an equivalent curve layout corresponding to the Manhattan layout after the moving processing of the segmented edges.

[0074] It should be noted that the curve processing can be implemented by using any one of the curve generation methods, and the specific implementation does not limit the protection scope of the present application. The present application emphasizes that the generated equivalent curve layout will be used in the subsequent simulation steps together with the generated evaluation point set to calculate the EPE and guide the subsequent moving optimization of the segmented edges.

[0075] Step 104: Lithography imaging simulation is performed on the equivalent curve layout, and the EPE between the imaging profile and the target layout at each evaluation point in the evaluation point set is calculated.

[0076] In an exemplary example, step 104 can include:

[0077] The lithography imaging simulation is performed to obtain an imaging profile using the equivalent curve layout as the input of the lithography imaging simulation. At each evaluation point in the evaluation point set established in step 100, an EPE value between the imaging profile and the target layout is calculated to form error distribution data for optimization feedback.

[0078] In the embodiments of the present application, each Manhattan layout after the edge movement processing has a unique corresponding equivalent curve layout, and the shape of the equivalent curve layout can be flexibly controlled by parameters. The curve layout is used to replace the Manhattan layout to perform the lithography imaging simulation.

[0079] Step 105: Determine whether the optimization end condition is met according to the obtained EPE value. When the optimization end condition is met, go to step 106; when the optimization end condition is not met, return to step 102 to continue the movement processing of the segmented edge.

[0080] In this step, whether the current optimization converges is determined according to all the obtained EPE values. When any of the following optimization end conditions is met, the loop is determined to end:

[0081] All absolute values of the EPEs are less than a preset threshold; or,

[0082] The change of the optimization objective function is less than a convergence threshold; or,

[0083] The maximum number of iterations is reached.

[0084] If the above conditions are not met, return to step 102 to continue the movement processing of the segmented edge.

[0085] In an exemplary instance, the optimization objective function can consider both the EPE error term and the curve deformation constraint term, that is, the optimization in the embodiments of the present application not only includes the geometric optimization of the Manhattan edge movement, but also includes the EPE error optimization based on the curve simulation. When the change of the optimization objective function is less than the convergence threshold, it indicates that the imaging error caused by the current geometric edge movement and the curve deformation both tend to be stable, and at this time, the optimization process can be determined to converge. By including the stability of the curve mapping in the convergence judgment in the embodiments of the present application, the oscillation phenomenon caused by the instability of the curve fitting in the traditional Manhattan optimization is avoided, thereby significantly improving the certainty and controllability of the overall optimization process.

[0086] Step 106: Perform curve mapping on the Manhattan layout after the current segmented edge movement processing to generate a curve mask pattern.

[0087] When the optimization converges, the Manhattan layout after the current segmented edge movement processing is subjected to curve mapping to generate a final curve mask pattern as the OPC result after optimization in the same curve mapping manner as in step 103.

[0088] Compared with the OPC and ILT technologies in the related art, the embodiment of the present application modifies the traditional Edge-based OPC process, and introduces a curved layout instead of an original Manhattan layout as an input in the simulation stage. The traditional Edge-based OPC only uses the original layout in the imaging simulation, and the optimization object is consistent with the simulation input, which is difficult to capture the response of the real optical system to the curvature radius and the corner smoothing. In the embodiment of the present application, by introducing the curved post-layout (i.e. the equivalent curve layout) in the simulation stage, the lithography simulation can reflect the curved optical characteristics without changing the original edge movement mechanism. The embodiment of the present application quickly upgrades the traditional OPC process to the curved OPC process, without introducing the high-cost pixel-level optimization or the complex curve control point iteration of ILT, and maintains the stability and controllability of the edge movement main line.

[0089] The method for generating a curved mask pattern provided by the embodiment of the present application, on the one hand, extends the lithography imaging simulation and output to a curve-friendly form without changing the traditional OPC mainstream process and edge movement mechanism, so that the optical effects of the curvature radius and the corner smoothing can directly participate in the optimization decision; on the other hand, a parameterized and adjustable curved channel is provided, so that the method can be quickly reconfigured according to the changes of the process, equipment and mask process, without introducing a new ILT infrastructure; and on the other hand, the MRC violation and post-processing cost are reduced while maintaining the optimization stability, and the repeated process of repairing and re-violation after optimization is avoided.

[0090] The method for generating a curved mask pattern provided by the embodiment of the present application, the Manhattan layout after each edge movement processing corresponds to a unique equivalent curve layout, and the shape thereof can be flexibly controlled through parameters. In the optimization iteration process, the equivalent curve layout is used to replace the Manhattan layout to perform lithography imaging simulation. When the optimization converges, the Manhattan layout after the current segmented edge movement processing is executed for curve processing to generate a final curved mask pattern as the optimized OPC result. In the embodiment of the present application, the Manhattan pattern is always taken as the optimization object in the optimization iteration process, but the equivalent curve layout is used as the input in the simulation stage. In this way, the control difficulty brought by directly modifying the curve layout is avoided, the curved mask pattern is stably and efficiently generated while maintaining the stability of the traditional OPC optimization, the curved mask result meeting the EPE constraint is obtained, and the wafer pattern transfer precision and the yield of chip manufacturing are improved.

[0091] The present application also provides a computer-readable storage medium storing computer executable instructions for executing the method for generating a curved mask pattern according to any one of the above.

[0092] The application further provides a computer device comprising a memory and a processor, wherein the memory stores instructions executable by the processor, and the instructions comprise the steps of the method for generating the curved mask pattern.

[0093] Figure 3 For the schematic diagram of the simulation result of the mask pattern in the embodiments of the application, the difference between the traditional Manhattan OPC and the curved OPC provided by the embodiments of the application is shown. Figure 3 In the figure, the pattern enclosed by the black solid line 31 is the design layout, that is, the target pattern, which represents the structure shape that is expected to be imaged on the wafer in an ideal case; the blue dotted line contour 32 is the actual contour after lithography imaging, which is affected by OPE and shows the imaging error (EPE), and the contour is usually generated by a lithography simulation tool (such as LithoSim, Sentaurus Lithography, etc.); the green dash-dot line boundary 33 area represents the optimization result of the traditional Manhattan OPC (Edge-based OPC), which is characterized by edges composed of horizontal and vertical line segments, limited shape and sharp corners; the red dash-dot line boundary 34 area represents the result obtained by using the curved OPC process of the application. Compared with the Manhattan result, the contour is smoother and the distance between adjacent patterns is larger, which reflects the high degree of freedom of the curved edge in the imaging space and the better optical fitting effect.

[0094] From Figure 3It can be seen that at multiple positions, the distance between the graphs of the curved OPC results is larger than that of the Manhattan OPC results, and the overall shape is more smooth and natural, indicating that it has higher geometric freedom and better optical matching ability. This curved feature brings significant technical advantages. First, the curved OPC result is easier to pass the mask rule check (MRC). In the case of obtaining the same EPE index, the distance between the edges and the corners of the curved shape is larger, and the adjustable space is wider, so it is less likely to have MRC violation problems, reducing the repeated process of post-processing repair and re-verification. Second, since the curved shape has a larger adjustable margin, the optimization algorithm can search for solutions in a wider space, which can obtain a better MEEF index and better focus / dose margin, which means that the mask has better process stability and a larger process window under different optical conditions, thereby effectively improving the yield of chip manufacturing. Third, the continuity of the curved boundary enables the layout to maintain higher critical dimension (CD) uniformity during lithographic imaging, and the uniformity of CD directly determines the electrical performance and consistency of the device, so the curved OPC scheme can effectively improve the overall performance of the device. At the same time, since the curved boundary provides more flexible adjustment capability, it also enables the final result to further reduce the EPE error and improve the consistency between the imaging profile and the target layout.

[0095] In addition, for silicon optoelectronic devices, the device structure usually requires a continuous curved form, such as waveguide, coupler and other structures, and the traditional Manhattan OPC cannot generate a layout that meets the optical continuity requirements of such devices, while the curved OPC result provided by the present application can naturally meet the process specification requirements of these advanced silicon photonic devices.

[0096] To further illustrate the advantages of the embodiments of this application, the following compares three existing typical methods for obtaining curve-based results. The first method is ILT. ILT uses pixels as the optimization object, and the result obtained through image domain inversion or combined with neural networks naturally has curve-based characteristics. However, this method needs to handle pixel-level high-dimensional optimization problems, which involves huge computational load, and the optimization result is often a pixelated image, requiring additional curve fitting steps to convert it into a layout graphic, resulting in poor local controllability and high implementation cost. The second method is curve control point optimization, such as using Bézier curves as the operation object, and achieving graphic correction by adjusting curve control points. This method can theoretically achieve continuous curve control, but the control point freedom is too high, and moving the control point often affects the shape of the entire curve, making it difficult to achieve precise local adjustment. At the same time, the optimization algorithm is complex and has poor convergence, making it difficult to apply in engineering. The third method is the scheme adopted in the embodiments of this application. The method for generating curve-based mask graphics provided by the embodiments of this application maintains the edge-moving mechanism of traditional Edge-based OPC, and only converts the Manhattan layout into a curve-based layout for lithography simulation during the simulation stage, thereby obtaining the curve OPC result without increasing the algorithm complexity. The operation object in the entire optimization process is still the edges of the original layout. It is only necessary to determine the movement direction and step size of each edge. Therefore, the algorithm complexity is consistent with that of traditional OPC, while the final result has high precision and high process matching.

[0097] Therefore, the method for generating curved mask patterns provided in this application achieves an excellent balance between computational complexity, optimization controllability, and result quality. On the one hand, its algorithm complexity is basically the same as that of traditional Edge-based OPC, facilitating rapid integration into existing OPC toolchains; on the other hand, the resulting image closely resembles the free-form curved mask generated by ILT, possessing higher imaging accuracy and lithography process compatibility. The method for generating curved mask patterns provided in this application achieves high-quality curved mask generation without introducing high-cost ILT infrastructure, combining stability, efficiency, and engineering feasibility.

[0098] This application also provides an apparatus for generating curveled mask patterns, such as... Figure 2 As shown, it may include: a first processing module, a generation module, a second processing module, a first curve conversion module, a simulation module, and a second curve conversion module; wherein,

[0099] The first processing module is used to perform boundary analysis on the Manhattan map to obtain the geometric properties of the corner points, and to segment each geometric edge of the Manhattan map to obtain segmented edges.

[0100] The generation module is used to set sampling points for EPE evaluation on each segment edge and generate a set of evaluation points;

[0101] a second processing module, configured to perform a moving process on the segmented edges of the Manhattan layout;

[0102] a first curve processing module, configured to perform a curve processing on the segmented edge information after the moving process and the geometric attribute of the corner point, to generate an equivalent curve layout corresponding to the Manhattan layout after the moving process of the segmented edge;

[0103] a simulation module, configured to perform a lithography imaging simulation on the equivalent curve layout, to calculate the EPE between the imaging profile and the target layout at each evaluation point in the evaluation point set;

[0104] a judging module, configured to judge whether an optimization end condition is met according to the obtained EPE value, and execute the generating module when the optimization end condition is met, and return to the second processing module when the optimization end condition is not met;

[0105] a second curve processing module, configured to perform a curve processing on the Manhattan layout after the moving process of the current segmented edge, to generate a curve mask pattern.

[0106] In an exemplary instance, the optimization end condition includes any of the following:

[0107] all absolute values of the EPEs are less than a preset threshold; or,

[0108] a variation of the optimization target function is less than a convergence threshold; or,

[0109] a maximum iteration number is reached.

[0110] The device for generating a curve mask pattern provided by the embodiments of the present application has a unique equivalent curve layout corresponding to the Manhattan layout after each edge moving process, and the shape of the equivalent curve layout can be flexibly controlled by parameters. In the optimization iteration process, the equivalent curve layout is used to replace the Manhattan layout to perform a lithography imaging simulation. When the optimization converges, a curve processing is performed on the Manhattan layout after the moving process of the current segmented edge, to generate a final curve mask pattern as the optimized OPC result. In the embodiments of the present application, the Manhattan layout is always used as the optimization object in the optimization iteration process, but the equivalent curve layout is used as the input in the simulation stage. In this way, the control difficulty caused by directly modifying the curve layout is avoided, the curve mask pattern is stably and efficiently generated while the stability of the traditional OPC optimization is maintained, the curve mask result meeting the EPE constraint is obtained, and the wafer pattern transfer precision and the chip manufacturing yield are improved.

[0111] Although the present application has been described with reference to the above embodiments, the contents described are merely employed embodiments for facilitating the understanding of the present application, and are not intended to limit the present application. Any modification and change in the form and details can be made by any person skilled in the art without departing from the spirit and scope of the present application, and the patent protection scope of the present application shall be subject to the scope defined by the appended claims.

Claims

1. A method of generating a curved mask pattern, characterized by, The method comprises the following steps: Boundary analysis is performed on the Manhattan layout to identify the connection relationship of each geometric edge and obtain the geometric attributes of the corner points formed by adjacent geometric edges, the geometric attributes of each corner point including concave-convex type, interior angle, and adjacent edge length information; Each geometric edge of the Manhattan layout is segmented to obtain a segmented edge; Sampling points for edge position error (EPE) evaluation are set on each segmented edge to generate an evaluation point set; The segmented edge of the Manhattan layout is subjected to a moving process; Curve parameters are determined according to the geometric information of the moving processed segmented edge and the geometric attributes of the corner points, and the segmented edge is subjected to a curve processing to generate a smooth and continuous curve boundary, thereby forming an equivalent curve layout corresponding to the Manhattan layout after the moving process of the segmented edge; Photolithography imaging simulation is performed on the equivalent curve layout, and the EPE between the imaging profile and the target layout is calculated at each evaluation point in the evaluation point set; Whether the optimization end condition is met is determined according to the obtained EPE value, and when the optimization end condition is met, the next step is entered; when the optimization end condition is not met, the step of moving the segmented edge of the Manhattan layout is returned to; The Manhattan layout after the moving process of the current segmented edge is subjected to curve processing to generate a curve mask pattern.

2. The method of claim 1, wherein, The Manhattan layout is a mask layout mainly in Manhattan geometry, which is an initial mask pattern obtained after circuit layout design and optical proximity correction (OPC) preprocessing.

3. The method of claim 1, wherein, For edges with a length greater than a preset threshold, each geometric edge is segmented according to a segment interval determined according to lithography resolution and target CD; For edges with complex geometric shapes or edges close to lithography hot spots, each geometric edge is segmented according to a segment interval adaptively reduced according to local shape complexity.

4. The method of claim 1, wherein, The evaluation point set is generated by: On each segmented edge, a group of sampling points is set to form a sampling point set of the segmented edge; the sampling points are arranged along the normal or tangential direction of the segmented edge to calculate the EPE value on the segmented edge; The union set of the sampling point sets of all the segmented edges constitutes the evaluation point set.

5. The method of claim 4, further comprising: Encryption sampling is performed at the corner points or hot spot structures.

6. The method of claim 1, wherein, The moving process of the segmented edge of the Manhattan layout comprises: According to the EPE result obtained by the last round of lithography imaging simulation, the geometric position of each segmented edge of the Manhattan layout is updated; For the initial optimization, a lithography simulation is performed with the Manhattan layout as the input to obtain an initial imaging profile; the EPE between the initial imaging profile and the target layout is calculated at each evaluation point; and the position of the segmented edge is adjusted along the normal direction of the segmented edge according to the sign and size of the obtained EPE.

7. The method of claim 1, wherein, The optimization end condition includes any one of the following: The absolute value of all EPEs is less than a preset threshold; or The change of the optimization objective function is less than a convergence threshold; or The maximum number of iterations is reached.

8. A computer readable storage medium storing computer executable instructions for executing the method for generating a curve mask pattern according to any one of claims 1-7.

9. A computer device comprising a memory and a processor, wherein, The memory stores instructions executable by the processor for performing the steps of the method of generating a curve-shaped mask pattern according to any one of claims 1-7.

10. An apparatus for generating a curved mask pattern, the apparatus comprising: The method comprises: a first processing module, a generating module, a second processing module, a first curve-shaping module, a simulation module, and a second curve-shaping module; wherein the first processing module is configured to perform boundary analysis on the Manhattan layout, identify the connection relationship of each geometric edge, and obtain the geometric properties of the corner points formed by adjacent geometric edges, the geometric properties of each corner point including concave-convex type, interior angle, and adjacent edge length information; and segment each geometric edge of the Manhattan layout to obtain a segmented edge; the generating module is configured to set a sampling point for EPE evaluation on each segmented edge to generate a set of evaluation points; the second processing module is configured to perform movement processing on the segmented edges of the Manhattan layout; the first curve-shaping module is configured to determine curve-shaping parameters according to the geometric information of the movement-processed segmented edges and the geometric properties of the corner points, and perform curve-shaping processing on the segmented edges to generate a smooth and continuous curve boundary, thereby constituting an equivalent curve layout corresponding to the Manhattan layout after movement processing of the segmented edges; the simulation module is configured to perform lithography imaging simulation on the equivalent curve layout, and calculate the EPE between the imaging profile and the target layout at each evaluation point in the set of evaluation points; the judging module is configured to determine whether the optimization end condition is met according to the obtained EPE value, execute the generating module when the optimization end condition is met, and return to the second processing module when the optimization end condition is not met; the second curve-shaping module is configured to perform curve-shaping on the Manhattan layout after movement processing of the current segmented edge to generate a curve-shaped mask pattern.

Images