Modeling method of OPC model
By adopting multi-dimensional numerical search and contribution guidance step size setting in OPC modeling, the problems of low efficiency and insufficient accuracy of parameter search in the existing technology are solved, and fast and accurate parameter screening and model error reduction are achieved.
Patent Information
- Application Number
- CN202510517155.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-23
- Publication Date
- 2025-08-01
AI Technical Summary
The modeling methods of existing OPC models are inefficient in parameter search and cannot effectively filter invalid step sizes, resulting in long search time and inaccurate parameters, resulting in large errors in the OPC model.
By establishing an initial model, use the corresponding first step size and second step size to perform multi-dimensional numerical search, and set the step size and numerical range based on the contribution degree of model parameters. First, rough search and then fine search to filter out the best parameter combination.
It realizes the fast and accurate screening of the best parameter values of the OPC model, shortens the search time, improves modeling accuracy and efficiency, and reduces model errors.
Smart Images
Figure CN120406040A_ABST
Abstract
Description
[0001] The present invention relates to the field of semiconductor technology, and particularly relates to a method for modeling an OPC model. Background Art
[0002] As the feature size of semiconductor devices continues to shrink, the problem of optical image distortion becomes increasingly serious, resulting in a decrease in product yield. To improve the problem of optical image distortion, the semiconductor integrated circuit manufacturing industry has introduced Optical Proximity Correction (OPC) technology, which compensates for optical distortion effects by changing the layout shape. In OPC technology, the most critical step is the establishment of the OPC model.
[0003] Currently, the common practice is to establish an initial OPC model, add the required model parameters, determine the type of parameters to be searched and the search range, set a fixed step size within the search range, and search for the optimal values of the parameters respectively. However, the existing methods cannot effectively filter out invalid step sizes, resulting in a long search time, waste of resources, and the searched parameters are not accurate enough, and the Root Mean Square (RMS) value of the error corresponding to the determined OPC model is relatively large. Summary of the Invention
[0004] One of the objectives of the present invention is to provide a method for modeling an OPC model, which can quickly, accurately, and effectively screen out the optimal parameter values to enable the obtained OPC model to achieve a smaller error and improve the accuracy of modeling.
[0005] To achieve the above objective, the present invention provides a method for modeling an OPC model, including: Establish an initial model, where the initial model includes a plurality of model parameters to be searched, the plurality of model parameters to be searched include optical parameters and photoresist parameters, and each model parameter to be searched has a corresponding initial numerical range; Within the corresponding initial numerical range, perform numerical searches on the plurality of model parameters to be searched simultaneously with a corresponding first step size, and use the numerical group corresponding to the plurality of model parameters when the model error is the minimum value as the initial values of the plurality of model parameters to be searched; Based on the initial values of the model parameters to be searched, set corresponding precise numerical ranges, and within the corresponding precise numerical ranges, perform numerical searches on the plurality of model parameters to be searched simultaneously with a corresponding second step size, and use the numerical group corresponding to the plurality of model parameters when the model error is the minimum value as the searched numerical values of the plurality of model parameters to be searched, and the second step size corresponding to the same model parameter is smaller than the corresponding first step size; and Set the finally found value of the model parameter to be searched into the initial model to obtain the OPC model.
[0006] Optionally, there is a correlation between multiple model parameters to be searched.
[0007] Optionally, based on the contribution degree of the model parameter to the model, set the first step length and the second step length corresponding to each model parameter to be searched; among them, the first step length corresponding to the model parameter with a large contribution degree is less than the first step length corresponding to the model parameter with a small contribution degree, and the second step length corresponding to the model parameter with a large contribution degree is less than the second step length corresponding to the model parameter with a small contribution degree.
[0008] Optionally, based on the contribution degree of the model parameter to the model, set the accurate numerical range corresponding to each model parameter to be searched; among them, the accurate numerical range corresponding to the model parameter with a large contribution degree is greater than the accurate numerical range corresponding to the model parameter with a small contribution degree.
[0009] Optionally, before setting the finally found value of the model parameter to be searched into the initial model to obtain the OPC model, repeat the step of setting a new accurate numerical range based on the value found for the model parameter to be searched, and within the corresponding new accurate numerical range, perform a numerical search for multiple model parameters to be searched with a step length smaller than the previous one, and use the numerical group corresponding to the minimum model error as the value found for multiple model parameters to be searched until the model error is less than the set specification.
[0010] Optionally, for the same model parameter to be searched, when performing a numerical search within the new accurate numerical range, the step length is less than one-tenth of the previous search step length.
[0011] Optionally, the initial value is the central value of the accurate numerical range.
[0012] Optionally, the accurate numerical range is located within the corresponding initial numerical range and is smaller than the corresponding initial numerical range.
[0013] Optionally, the several optical parameters include the position of the light focusing plane and the optical signal gradient; the several photoresist parameters include the acid-base concentration, acid-base diffusion length, and acid-base reaction rate in the photoresist.
[0014] Optionally, the first step length corresponding to the same model parameter to be searched is more than ten times the corresponding second step length.
[0015] In the modeling method of the OPC model provided by the present invention, first, an initial model is established. The initial model includes multiple model parameters to be searched. The multiple model parameters to be searched include optical parameters and photoresist parameters. Each model parameter to be searched has a corresponding initial numerical range. Then, within the corresponding initial numerical range, with a corresponding first step size, numerical searches are simultaneously performed on the multiple model parameters to be searched. The numerical group corresponding to the multiple model parameters when the model error is the minimum value is used as the initial value of the multiple model parameters to be searched. After that, based on the initial value of the model parameters to be searched, a corresponding precise numerical range is set. Within the corresponding precise numerical range, with a corresponding second step size, numerical searches are simultaneously performed on the multiple model parameters to be searched. The numerical group corresponding to the multiple model parameters when the model error is the minimum value is used as the numerical values found for the multiple model parameters to be searched. The second step size corresponding to the same model parameter is smaller than the corresponding first step size. In this way, multi-dimensional numerical searches are simultaneously performed on the related optical parameters and photoresist parameters, and their numerical values are determined simultaneously, which can quickly and accurately obtain an optimal combination of multiple model parameters, helping to improve the modeling accuracy. Moreover, when modeling, first perform a rough search with a larger first step size, and then perform a fine search with a smaller second step size, which can effectively screen out invalid step sizes and shorten the search time, that is, it can quickly, accurately and effectively screen out the optimal parameter values to make the obtained OPC model reach a smaller error and improve the modeling accuracy.
[0016] Further, based on the contribution degree of the model parameters to the model, the first step size and the second step size corresponding to each model parameter to be searched are set, which can reduce the search time for the parameters with a small contribution degree to the model, help to shorten the search time, and improve the search efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 It is a flowchart of the modeling method of the OPC model provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0018] The following further details the modeling method of the OPC model proposed by the present invention in conjunction with the accompanying drawings and specific embodiments. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the drawings are all in a very simplified form and use non-precise scales, only for conveniently and clearly assisting in explaining the purpose of the embodiments of the present invention.
[0019] Figure 1 It is a flowchart of the modeling method of the OPC model provided by an embodiment of the present invention. Refer to Figure 1 As shown, the modeling method of the OPC model provided by the present application includes: Step S1, establish an initial model, where the initial model includes multiple model parameters to be searched. The multiple model parameters to be searched include optical parameters and resist parameters, and each model parameter to be searched has a corresponding initial numerical range; Step S2, within the corresponding initial numerical ranges, perform numerical searches on the multiple model parameters to be searched simultaneously with a corresponding first step size, and use the numerical group corresponding to the multiple model parameters when the model error is at the minimum value as the initial values of the multiple model parameters to be searched; Step S3, set the corresponding precise numerical ranges based on the initial values of the model parameters to be searched, perform numerical searches on the multiple model parameters to be searched simultaneously within the corresponding precise numerical ranges with a corresponding second step size, and use the numerical group corresponding to the multiple model parameters when the model error is at the minimum value as the searched numerical values of the multiple model parameters to be searched. The second step size corresponding to the same model parameter is smaller than the corresponding first step size; and Step S4, set the finally searched numerical values of the model parameters to be searched into the initial model to obtain an OPC model.
[0020] Specifically, in Step S1, first establish an initial model and add multiple model parameters to the initial model. Among them, at least some model parameters are uncertain and need to be searched, that is, the initial model includes multiple model parameters to be searched, and each model parameter to be searched has a corresponding initial numerical range.
[0021] In this embodiment, the multiple model parameters to be searched include optical parameters and resist parameters. There is a correlation between the multiple model parameters to be searched. For example, there is a correlation between the optical parameters and the resist parameters. Exemplarily, the optical parameters include the position of the light focusing plane and the optical signal gradient, etc. The resist parameters include the acid-base concentration, acid-base diffusion length, and acid-base reaction rate in the photoresist, etc.
[0022] It should be noted that hereinafter, mainly the data search of the model parameters to be searched will be described. For the sake of concise description, in some places, the "model parameters to be searched" will be abbreviated as "model parameters".
[0023] Execute Step S2, within the corresponding initial numerical ranges, perform numerical searches on the multiple model parameters to be searched simultaneously with a corresponding first step size, and use the numerical group corresponding to the multiple model parameters when the model error is at the minimum value as the initial values of the multiple model parameters to be searched.
[0024] Specifically, each model parameter to be searched has a corresponding initial value range, and each model parameter performs data search within its initial value range with a corresponding first step size. The first step sizes corresponding to different model parameters can be different.
[0025] In the OPC model, the contribution degrees of different model parameters to the model are different, and the contribution degrees of the model parameters can be obtained by the method of controlling variables. Exemplarily, when confirming the contribution degrees of some model parameters, the remaining model parameters are kept unchanged, the values of the model parameters whose contribution degrees need to be confirmed are adjusted, and the change of the model error of the model is analyzed; when the change of the model error is small, it is confirmed that the contribution degrees of this part of model parameters are small; when the model error is large, it is confirmed that the contribution degrees of this part of model parameters are small.
[0026] In this embodiment, based on the contribution degrees of the model parameters to the model, the first step sizes corresponding to each model parameter to be searched are set, wherein the first step size corresponding to the model parameter with a large contribution degree is smaller than the first step size corresponding to the model parameter with a small contribution degree. In this way, the search time of the model parameters with small contribution degrees can be reduced, and the limited time can be used more for the search of the model parameters with large contribution degrees. Furthermore, the search time can be shortened and the search efficiency can be improved while ensuring the query accuracy.
[0027] In this embodiment, multi-dimensional search is performed on multiple model parameters to be searched simultaneously, and the numerical group corresponding to the multiple model parameters to be searched when the model error is the minimum value is used as the initial values of the multiple model parameters to be searched, rather than searching one model parameter one by one and then combining them. In this way, a better combination of multiple model parameters can be obtained quickly and accurately, which helps to improve the modeling accuracy. It should be noted that the numerical group (i.e., the model parameter numerical group) corresponding to the minimum value of the model error includes the values corresponding to each model parameter, and the value corresponding to the model parameter in the numerical group is used as the initial value of the model parameter.
[0028] In this embodiment, the model error is the root mean square (RMS) of the difference between the actual measured critical dimension of the wafer and the model simulation. This model error can be given by the error evaluation function in the OPC system.
[0029] Next, step S3 is executed. Based on the initial values of the model parameters to be searched, the corresponding precise value ranges are set, and within the corresponding precise value ranges, numerical search is performed on multiple model parameters to be searched simultaneously with a corresponding second step size. The numerical group corresponding to the multiple model parameters to be searched when the model error is the minimum value is used as the searched values of the multiple model parameters to be searched, and the second step size corresponding to the same model parameter is smaller than the corresponding first step size.
[0030] Specifically, an exact numerical range can be set corresponding to the initial value of each model parameter to be searched. Exemplarily, the initial value can be the central value of the exact numerical range, but is not limited thereto. For example, if the initial value is A0, the exact numerical range can be A0±Δx, where Δx is the tolerance and can be set according to the actual situation, and the tolerances corresponding to different model parameters can be different.
[0031] In this embodiment, the exact numerical range is within the corresponding initial numerical range and smaller than the corresponding initial numerical range.
[0032] In this embodiment, the exact numerical range corresponding to each model parameter to be searched can be set based on the contribution degree of the model parameter to the model; among them, the exact numerical range corresponding to the model parameter with a greater contribution degree can be larger than the exact numerical range corresponding to the model parameter with a smaller contribution degree. In this way, the search range for the model parameter with a greater contribution degree is larger, and the search range for the model parameter with a smaller contribution degree is smaller, which can effectively reduce the search time on the premise of ensuring the accuracy of the model.
[0033] In step S3, each model parameter to be searched is searched within its corresponding exact numerical range with a corresponding second step size, and the second step size corresponding to the same model parameter is smaller than the corresponding first step size. In this way, after the rough search in step S2, a fine search is performed in step S3, which can effectively filter out invalid step sizes and shorten the search duration, that is, the optimal parameter value can be quickly, accurately and effectively screened out to make the obtained OPC model have a smaller error and improve the accuracy of modeling.
[0034] The second step sizes corresponding to different model parameters can be the same or different. Exemplarily, based on the contribution degree of the model parameter to the model, the second step size corresponding to each model parameter to be searched is set, where the second step size corresponding to the model parameter with a greater contribution degree can be smaller than the second step size corresponding to the model parameter with a smaller contribution degree, which can shorten the data search time while ensuring the model accuracy.
[0035] Exemplarily, the first step size corresponding to the same model parameter to be searched can be more than ten times the corresponding second step size, but is not limited thereto.
[0036] After executing step S3, the step of setting a new exact numerical range based on the values found for the model parameters to be searched can be repeated, and within the corresponding new exact numerical range, the values of multiple model parameters to be searched are simultaneously searched with a step size smaller than the previous one, and the numerical group corresponding to the minimum model error is used as the values found for the multiple model parameters to be searched, until the model error remains unchanged or the model error is less than the set specification, which can make the values found for the model parameters more accurate.
[0037] Exemplarily, when performing a numerical search within a new precise numerical range for the same model parameter to be searched, the step size is less than one-tenth of the previous search step size, but is not limited thereto.
[0038] Next, step S4 is executed to set the finally searched numerical value of the model parameter to be searched into the initial model to obtain an OPC model.
[0039] The following takes the numerical search of a model parameter with an initial numerical range of 0 to 100 as an example for illustration.
[0040] In the prior art, within the initial numerical range of 0 to 100, a fixed step size of 2 (Increment) is used for numerical search. 51 steps need to be searched. The numerical value corresponding to the model parameter when the model error is the smallest is 52, that is, the optimal result of the model parameter is 52, and the corresponding RMS is 3.13.
[0041] When performing a numerical search using the modeling method provided in this embodiment, within the initial numerical range of 0 to 100, a rough numerical search is first performed with a first step size of 10. 11 steps need to be searched, and the best result found is 50. That is, 50 is used as the initial value of the model parameter. Then, based on the initial value of 50, a precise numerical range of 45 to 55 is set, and a second step size of 1 is used to perform a numerical search within the precise numerical range of 45 to 55. 11 steps need to be searched, and the best result found is 51. After that, based on the found numerical value of 51, a new precise numerical range of 50 to 52 is set, and a step size of 0.1 is used to perform a numerical search within the precise numerical range of 50 to 52. 21 steps need to be searched, and the best result found is 51.2, and the corresponding RMS is 3.09.
[0042] When performing a numerical search using the existing method, a total of 51 steps need to be searched, and the found result is 52. When performing a numerical search using the modeling method provided in this embodiment, a total of 43 steps need to be searched, and the found result is 51.2. It can be seen that when performing a numerical search for the model parameter to be searched using the modeling method provided in this embodiment, many invalid step sizes can be reduced, and the search time is relatively short, the obtained RMS is smaller, and thus the model accuracy obtained is higher.
[0043] In the modeling method of the OPC model provided by the present invention, an initial model is first established. The initial model includes multiple model parameters to be searched. The multiple model parameters to be searched include optical parameters and photoresist parameters. Each model parameter to be searched has a corresponding initial value range. Then, within the corresponding initial value ranges, with a corresponding first step size, numerical searches are simultaneously performed on the multiple model parameters to be searched. The numerical group corresponding to the multiple model parameters when the model error is the minimum value is used as the initial values of the multiple model parameters to be searched. After that, based on the initial values of the model parameters to be searched, corresponding precise value ranges are set. Within the corresponding precise value ranges, with a corresponding second step size, numerical searches are simultaneously performed on the multiple model parameters to be searched. The numerical group corresponding to the multiple model parameters when the model error is the minimum value is used as the numerical values found for the multiple model parameters to be searched. The second step size corresponding to the same model parameter is smaller than the corresponding first step size. In this way, multi-dimensional numerical searches are simultaneously performed on the related optical parameters and photoresist parameters, and their numerical values are determined simultaneously, which can quickly and accurately obtain an optimal combination of multiple model parameters, helping to improve the modeling accuracy. Moreover, during modeling, a rough search is first performed with a larger first step size, and then a fine search is performed with a smaller second step size, which can effectively filter out invalid step sizes and shorten the search time, that is, the best parameter values can be quickly, accurately and effectively screened out to make the obtained OPC model have a smaller error and improve the modeling accuracy.
[0044] The above description is only a description of the preferred embodiments of the present invention, and does not limit any scope of the rights of the present invention. Any person skilled in the art can make possible changes and modifications to the technical solution of the present invention by using the methods and technical contents disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and decorations made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention all belong to the protection scope of the technical solution of the present invention.
Claims
1. A modeling method for an OPC model, characterized in that, Including: An initial model is established. The initial model includes a plurality of model parameters to be searched. The plurality of model parameters to be searched include optical parameters and photoresist parameters. Each of the model parameters to be searched has a corresponding initial numerical range; Within the corresponding initial numerical range, with a corresponding first step size, numerical searches are simultaneously performed on the plurality of model parameters to be searched. The numerical group corresponding to the plurality of model parameters to be searched when the model error is at the minimum value is used as the initial values of the plurality of model parameters to be searched; Based on the initial values of the model parameters to be searched, corresponding precise numerical ranges are set. Within the corresponding precise numerical ranges, with a corresponding second step size, numerical searches are simultaneously performed on the plurality of model parameters to be searched. The numerical group corresponding to the plurality of model parameters to be searched when the model error is at the minimum value is used as the numerical values found for the plurality of model parameters to be searched. The second step size corresponding to the same model parameter is smaller than the corresponding first step size; And The finally found numerical values of the model parameters to be searched are set into the initial model to obtain an OPC model.
2. The modeling method of the OPC model according to claim 1, characterized in that, There is a correlation between the plurality of model parameters to be searched.
3. The modeling method of the OPC model according to claim 1, wherein Based on the contribution degree of the model parameters to the model, the corresponding first step size and second step size of each model parameter to be searched are set; wherein, the first step size corresponding to the model parameter with a larger contribution degree is smaller than the first step size corresponding to the model parameter with a smaller contribution degree, and the second step size corresponding to the model parameter with a larger contribution degree is smaller than the second step size corresponding to the model parameter with a smaller contribution degree.
4. The modeling method of the OPC model according to claim 1, characterized in that, Based on the contribution degree of the model parameters to the model, the corresponding precise numerical ranges of each model parameter to be searched are set; wherein, the precise numerical range corresponding to the model parameter with a larger contribution degree is larger than the precise numerical range corresponding to the model parameter with a smaller contribution degree.
5. The modeling method of the OPC model according to claim 1, characterized in that, Before setting the finally found numerical values of the model parameters to be searched into the initial model to obtain an OPC model, the step of repeatedly setting new precise numerical ranges based on the found numerical values of the model parameters to be searched, within the corresponding new precise numerical ranges, simultaneously performing numerical searches on the plurality of model parameters with a step size smaller than the previous one, and using the numerical group corresponding to the minimum model error as the found numerical values of the plurality of model parameters to be searched is repeated until the model error is less than the set specification.
6. The modeling method of the OPC model according to claim 5, characterized in that, For the same model parameter to be searched, when performing a numerical search within the new precise numerical range, the step size is less than one-tenth of the previous search step size.
7. The modeling method of the OPC model according to claim 1, characterized in that, The initial value is the central value of the precise numerical range.
8. The modeling method of the OPC model according to claim 1, characterized in that The precise numerical range is located within the corresponding initial numerical range and is smaller than the corresponding initial numerical range.
9. The modeling method of the OPC model according to claim 1, characterized in that, The several optical parameters include the light focusing plane position and the optical signal gradient; the several photoresist parameters include the acid-base concentration, acid-base diffusion length, and acid-base reaction rate in the photoresist.
10. The modeling method of the OPC model according to claim 1, characterized in that, The first step size corresponding to the same model parameter to be searched is more than ten times the corresponding second step size.
Citation Information
Patent Citations
Optical proximity correction method
CN104570585A