Modeling method of optical proximity correction model and optical proximity correction method

Abstract

The invention discloses a modeling method of an optical proximity correction model and an optical proximity correction method. The modeling method comprises the following steps: providing a plurality of bright field test patterns and a plurality of dark field test patterns; obtaining the number ratio of the bright field areas to the dark field areas in the actual layout; extracting a bright field test pattern and a dark field test pattern with the same proportion; and establishing an optical proximity correction model according to the extracted bright field test pattern and the corresponding bright field photoresist pattern as well as the feature size difference between the extracted dark field test pattern and the corresponding dark field photoresist pattern. The bright field test patterns and the dark field test patterns with the same proportion are sampled on the basis of the number proportion of the bright field areas and the dark field areas in the actual layout, and the photoacid diffusion length in the optical proximity correction model established on the theoretical basis has high pertinence, so that the accuracy of the photoacid diffusion length corresponding to the actual layout is effectively improved; the actual layout is optically corrected through the optical proximity correction model, and the optical proximity correction precision can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of semiconductor manufacturing, and in particular, to a method for modeling an optical proximity correction model and an optical proximity correction method. Background Art

[0002] In semiconductor manufacturing, as the design size continues to shrink, the diffraction effect of light becomes more and more obvious. The result is that the optical image of the design pattern degrades, and finally the actual pattern formed on the silicon wafer after lithography becomes different from the design pattern. This phenomenon is called OPE (Optical Proximity Effect).

[0003] One of the most significant effects in the OPE phenomenon is called the Critical Dimension Proximity Effect (CDProximity). Specifically, as the spatial period changes, the critical dimension of the lithography pattern changes with the optical diffraction effect. Eventually, the critical dimensions of the photomask patterns with the same critical dimension under different spatial periods are different in the lithography patterns after lithography. This phenomenon is very fatal to circuit designers, as it can cause changes in the performance of various devices that should be the same under different circuit layouts, ultimately leading to circuit failure.

[0004] In order to correct the OPE phenomenon, OPC (Optical Proximity Correction) has emerged. The core idea of OPC is to establish an optical proximity correction model based on the consideration of canceling the OPE phenomenon, and design the photomask pattern according to the optical proximity correction model. In this way, the lithography pattern after lithography exhibits the OPC phenomenon relative to the photomask pattern, but the OPC phenomenon is canceled when the optical proximity correction model is established. Therefore, the lithography pattern after lithography is close to the pattern that the designer hopes to form through lithography.

[0005] However, there are still many problems in the optical proximity correction process in the prior art. Summary of the Invention

[0006] The technical problem solved by the present invention is to provide a method for modeling an optical proximity correction model and an optical proximity correction method to improve the accuracy of optical proximity correction.

[0007] To solve the above problems, the technical solution of the present invention provides a method for modeling an optical proximity correction model, including: providing a plurality of bright-field test patterns and a plurality of dark-field test patterns; obtaining the quantity proportion of the bright-field area and the dark-field area in the actual layout; based on the quantity proportion of the bright-field area and the dark-field area, extracting the same proportion of the bright-field test patterns and the dark-field test patterns; establishing an optical proximity correction model according to the feature size differences between the extracted bright-field test patterns and the corresponding bright-field photoresist patterns, and between the extracted dark-field test patterns and the corresponding dark-field photoresist patterns.

[0008] Optionally, dividing the bright-field test patterns and the dark-field test patterns based on a first preset value; wherein, the duty cycle of the bright-field test patterns is greater than or equal to the first preset value; the duty cycle of the dark-field test patterns is less than the first preset value.

[0009] Optionally, the first preset value is 50%.

[0010] Optionally, a plurality of bright-field test patterns and a plurality of dark-field test patterns include: a plurality of different types of the bright-field test patterns and the dark-field test patterns.

[0011] Optionally, a plurality of different types of the bright-field test patterns and the dark-field test patterns include: the bright-field test patterns and the dark-field test patterns with periodically changing dense line spacings / gap spacings, the bright-field test patterns and the dark-field test patterns with linearly changing isolated line sizes / gap sizes, the bright-field test patterns and the dark-field test patterns of storage patterns, the bright-field test patterns and the dark-field test patterns of serpentine patterns, and the bright-field test patterns and the dark-field test patterns of T-shaped patterns, one or more of which.

[0012] Optionally, for each type, the bright-field test pattern and the corresponding dark-field test pattern are complementary patterns, and the sum of the duty cycle of the bright-field test pattern and the duty cycle of the corresponding dark-field test pattern is 100%.

[0013] Optionally, the method for obtaining the quantity proportion of the bright-field area and the dark-field area in the actual layout includes: dividing the actual layout into a plurality of graphic areas; obtaining the duty cycle in each graphic area; providing a second preset value, dividing the bright-field area and the dark-field area based on the second preset value, and defining the graphic area with a duty cycle greater than or equal to the second preset value as the bright-field area, and defining the graphic area with a duty cycle less than the second preset value as the dark-field area; counting the quantity of the bright-field area and the dark-field area, and taking the ratios of the quantity of the bright-field area and the dark-field area to the quantity of the graphic areas respectively as the quantity proportions of the bright-field area and the dark-field area in the actual layout.

[0014] Optionally, remove the graphic area with a duty cycle of 100%.

[0015] Optionally, the second preset value is 50%.

[0016] Optionally, the method for extracting the same proportion of the bright-field test patterns and the dark-field test patterns based on the quantity proportion of the bright-field area and the dark-field area includes: respectively extracting the same proportion of the bright-field test patterns and the dark-field test patterns from each type of the bright-field test patterns and the dark-field test patterns based on the quantity proportion of the bright-field area and the dark-field area.

[0017] Optionally, the method for forming the bright-field photoresist pattern and the dark-field photoresist pattern includes: manufacturing a photomask based on the bright-field test pattern, and performing exposure and development on the photoresist layer to obtain the bright-field photoresist pattern on the photoresist layer; manufacturing a photomask based on the dark-field test pattern, and performing exposure and development on the photoresist layer to obtain the dark-field photoresist pattern on the photoresist layer.

[0018] Optionally, the method for establishing the optical proximity correction model includes: mixing a plurality of extracted bright-field photoresist patterns and a plurality of dark-field test patterns as a whole, and the photoacid diffusion length in the optical proximity correction model is a single value.

[0019] Optionally, the method for establishing the optical proximity correction model includes: respectively obtaining the corresponding photoacid diffusion lengths based on the extracted bright-field test patterns and dark-field test patterns, and the photoacid diffusion length in the optical proximity correction model is a weighted average calculated based on the proportion of the bright-field test patterns and the dark-field test patterns, or is a gradient sequence.

[0020] Correspondingly, the technical solution of the present invention further provides an optical proximity correction method, including: providing an optical proximity correction model established by using the modeling method in any one of the above technical solutions; performing optical proximity correction on the actual layout based on the optical proximity correction model.

[0021] Further, when the photoacid diffusion length in the optical proximity correction model is a gradient sequence calculated based on the proportion of the bright-field test patterns and the dark-field test patterns, before performing optical proximity correction on the layout to be corrected, it further includes: obtaining the bright and dark field proportion in the layout to be corrected; calling the photoacid diffusion lengths corresponding to the bright-field test patterns and the dark-field test patterns in the gradient sequence with the same proportion as the bright and dark field proportion.

[0022] Compared with the prior art, the technical solution of the present invention has the following advantages:

[0023] In the modeling method of the optical proximity correction model of the technical solution of the present invention, by accurately analyzing the quantity ratio of the bright field region and the dark field region in the actual layout, the bright field test pattern and the dark field test pattern are extracted in the same ratio. Based on this, the established optical proximity correction model can significantly improve the pertinence of the photoacid diffusion length. This pertinence enables the optical proximity correction model to more accurately simulate the complex characteristics of photoacid diffusion in the actual process, thereby greatly improving the accuracy and reliability of the optical proximity correction model. Photoacid diffusion is one of the key factors affecting pattern accuracy, and the optical proximity correction model effectively reduces the pattern deviation caused by photoacid diffusion through accurate sampling and simulation. Applying the optical proximity correction model to the optical correction of the actual layout can significantly improve the correction accuracy.

[0024] Furthermore, the graphic region with a duty cycle of 100% is removed. The graphic region with a duty cycle of 100% means that there is no graphic filling in this graphic region. Therefore, the graphic region with a duty cycle of 100% is a graphic region without actual function and will not cause any impact on the lithography process. Moreover, the graphic region with a duty cycle of 100% does not need to be subjected to optical proximity correction. Therefore, only the graphic regions with actual functions need to be identified, and then correct sampling is performed to make it conform to the graphic distribution in the real circuit layout, thereby further improving the accuracy of the equivalent diffusion length in the optical proximity correction model.

[0025] Furthermore, the method of extracting the bright field test pattern and the dark field test pattern in the same ratio based on the quantity ratio of the bright field region and the dark field region includes: based on the quantity ratio of the bright field region and the dark field region, the bright field test pattern and the dark field test pattern in the same ratio are respectively extracted from each type of the bright field test pattern and the dark field test pattern. By performing extraction sampling from each type of the bright field test pattern and the dark field test pattern, the types of sample patterns for modeling are made more comprehensive, thereby ensuring that the influence of different types of test patterns on the model can be fully considered during the modeling process, improving the accuracy and generalization ability of the optical proximity correction model, enabling it to better adapt to various complex situations, enhancing the reliability of the modeling process, avoiding the deviation of the optical proximity correction model caused by a single sample type, and thus improving the stability and applicability of the optical proximity correction model in different scenarios. Description of the Drawings

[0026] Figure 1 is the flowchart of the modeling method of the optical proximity correction model of the embodiment of the present invention;

[0027] Figures 2 to 3It is a schematic structural diagram of each step of the optical proximity correction model modeling method in the embodiments of the present invention. Detailed implementation manners

[0028] As described in the background art, there are still many problems in optical proximity correction in the prior art. The following will be specifically described.

[0029] The existing optical proximity correction model is based on a fixed photoacid diffusion length, and its establishment generally includes the following steps: Step 1, measure the photoacid diffusion length of the photoresist in a certain lithography process; Step 2, use the photoacid diffusion length obtained in Step 1 to establish an optical proximity correction model (the detailed steps of how to establish an optical proximity correction model are numerous and are all publicly available prior arts, so they will not be elaborated here); Step 3, use the optical proximity correction model obtained in Step 2 to design multiple mask patterns with different spatial periods and different critical dimensions; Step 4, lithograph the multiple mask patterns obtained in Step 3, measure the critical dimensions of each lithographed pattern, and for the case where the error between the critical dimensions of the lithographed pattern and the critical dimensions of the mask pattern in Step 3 is large, add many special rules for correction (the content of what special rules are used to correct the optical proximity correction model is numerous and are all publicly available prior arts, so they will not be elaborated here), and finally obtain a relatively accurate optical proximity correction model.

[0030] In Step 1 of the above method, there are two methods for measuring the photoacid diffusion length of the photoresist. The first method is to expose a standard test pattern, and then determine the diffusion length of the photoacid after exposure by measuring the photoacid concentration distribution after exposure. The diffusion length obtained by this method is the actual diffusion length. The second method is to expose a series of standard test patterns, then measure the critical dimensions of all the patterns, and then determine a series of diffusion length guess values through an equivalent Gaussian diffusion model (this model assumes that the photoacid concentration distribution is a Gaussian distribution, and defines the width of the corresponding Gaussian distribution when the diffusion concentration reaches a certain value as the corresponding diffusion length). Calculate the critical dimensions of all the test patterns for different diffusion length guess values respectively, and finally compare with the actual critical dimensions. The diffusion length guess value with the smallest error is taken as the photoacid diffusion length of the photoresist. The diffusion length obtained by this method is the equivalent diffusion length.

[0031] Generally, the Gaussian equivalent diffusion model is used for the chemical diffusion part, that is, after the photoacid is generated, it diffuses around, depending on the concentration gradient of the photoacid and the surrounding photoacid quencher concentration, and the distribution probability follows a Gaussian distribution, and the average diffusion length becomes the equivalent diffusion length. In the current optical proximity correction model, it is considered that the equivalent diffusion length is fixed. Therefore, when establishing and correcting the optical proximity correction model, the equivalent diffusion length is taken as a constant. Once the optical proximity correction modeling is completed, its value remains unchanged, and the same value is taken for all patterns in the TCC transfer matrix during correction.

[0032] However, actual experimental tests have found that the actual equivalent diffusion lengths for different patterns and different photoresists are actually variable. Therefore, different equivalent diffusion lengths are correspondingly used to simulate different patterns. However, it is only used as a variable parameter in the TCC transfer matrix to correct the optical proximity correction model error with large differences between bright and dark fields, and there is no correct sampling and modeling method. In the modeling process, the test patterns for optical proximity correction are mainly designed according to the pattern attributes of the current layer. For example, if the pattern is mainly composed of lines, it is a bright-field pattern; if the pattern is mainly composed of holes or gaps, it is a dark-field pattern. For example, if the bright-field pattern is dominant, the test pattern mainly consists of lines varying with the spatial period; if the dark-field pattern is dominant, the test pattern consists of gaps or holes varying with the spatial period. In order to balance the different equivalent diffusion lengths between bright and dark fields, the isolated / 1:1 dense bar pattern varying with the feature size and the isolated / 1:1 gap pattern varying with the feature size are used as standard patterns to assist. In this way, the actual change situation between bright and dark fields cannot be reflected. Therefore, the equivalent diffusion length in the established optical proximity correction model is also inaccurate, and the accuracy of the layout correction using this optical proximity correction model also needs to be further improved.

[0033] Based on this, the present invention provides a modeling method for an optical proximity correction model and an optical proximity correction method. By accurately analyzing the quantity ratio of the bright-field region and the dark-field region in the actual layout, the bright-field test pattern and the dark-field test pattern are extracted according to the same ratio. The optical proximity correction model established based on this can significantly improve the pertinence of the photoacid diffusion length. This pertinence enables the optical proximity correction model to more accurately simulate the complex characteristics of photoacid diffusion in the actual process, thereby greatly improving the accuracy and reliability of the optical proximity correction model. Photoacid diffusion is one of the key factors affecting pattern accuracy, and the optical proximity correction model effectively reduces the pattern deviation caused by photoacid diffusion through accurate sampling and simulation. Applying the optical proximity correction model to the optical correction of the actual layout can significantly improve the correction accuracy.

[0034] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description will be given to the specific embodiments of the present invention with reference to the accompanying drawings.

[0035] Figure 1 It is a flowchart of the modeling method for the optical proximity correction model according to an embodiment of the present invention, including:

[0036] Step S101, providing a plurality of bright-field test patterns and a plurality of dark-field test patterns;

[0037] Step S102, obtaining the quantity ratio of the bright-field region and the dark-field region in the actual layout;

[0038] Step S103: Extract the same proportion of the bright-field test patterns and the dark-field test patterns based on the proportion of the number of the bright-field regions and the dark-field regions.

[0039] Step S104: Establish an optical proximity correction model according to the feature size differences between the extracted bright-field test patterns and the corresponding bright-field photoresist patterns, and between the extracted dark-field test patterns and the corresponding dark-field photoresist patterns.

[0040] The following details the steps of the method for modeling the optical proximity correction model in conjunction with the accompanying drawings.

[0041] Figures 2 to 3 It is a schematic structural diagram of each step of the method for modeling the optical proximity correction model in an embodiment of the present invention.

[0042] Please refer to Figure 2 to provide a plurality of bright-field test patterns 100 and a plurality of dark-field test patterns 101.

[0043] In this embodiment, the bright-field test patterns 100 and the dark-field test patterns 101 are divided based on a first preset value; wherein, the duty cycle of the bright-field test pattern 100 is greater than or equal to the first preset value; the duty cycle of the dark-field test pattern 101 is less than the first preset value. Herein, the duty cycle is the ratio of the region without a pattern (i.e., the light-transmitting region) to the total area.

[0044] In this embodiment, the first preset value is 50%.

[0045] In this embodiment, the plurality of bright-field test patterns 100 and the plurality of dark-field test patterns 101 include: a plurality of different types of the bright-field test patterns 100 and the dark-field test patterns 101.

[0046] In this embodiment, the plurality of different types of the bright-field test patterns 100 and the dark-field test patterns 101 include: the bright-field test patterns 100 and the dark-field test patterns 101 with periodically changing dense line spacings / gap spacings (as shown in part a of Figure 2 ), the bright-field test patterns 100 and the dark-field test patterns 101 with linearly changing isolated line sizes / gap sizes (as shown in part b of Figure 2 ), the bright-field test patterns 100 and the dark-field test patterns 101 of storage patterns (not shown), the bright-field test patterns 100 and the dark-field test patterns 101 of serpentine patterns (not shown), and the bright-field test patterns 100 and the dark-field test patterns 101 of T-shaped patterns (not shown), or one or more of them.

[0047] In this embodiment, the bright-field test pattern 100 and the corresponding dark-field test pattern 101 under each type are complementary patterns to each other, and the sum of the duty cycle of the bright-field test pattern 100 and the duty cycle of the corresponding dark-field test pattern 101 is 100%.

[0048] Please refer to Figure 3 , to obtain the quantity proportion of the bright-field area and the dark-field area in the actual layout 200.

[0049] In this embodiment, the method for obtaining the quantity proportion of the bright-field area and the dark-field area in the actual layout 200 includes: dividing the actual layout 200 into several graphic areas 201; obtaining the duty cycle of each graphic area 201; providing a second preset value, dividing the bright-field area and the dark-field area based on the second preset value, and defining the graphic area 201 with a duty cycle greater than or equal to the second preset value as the bright-field area, and defining the graphic area 201 with a duty cycle less than the second preset value as the dark-field area; counting the quantities of the bright-field area and the dark-field area, and taking the ratios of the quantities of the bright-field area and the dark-field area to the quantity of the graphic area 201 respectively as the quantity proportions of the bright-field area and the dark-field area in the actual layout 200.

[0050] In a specific embodiment, the size of the actual layout 200 is 26000um * 33000um, and it can be divided into several graphic areas 201 with sizes of 300 um * 300 um or 500 um * 500 um, and the duty cycle of each graphic area 201 is obtained respectively.

[0051] In this embodiment, it is necessary to remove the graphic area 201 with a duty cycle of 100%. The graphic area 201 with a duty cycle of 100% means that there is no any graphic filling in this graphic area. Therefore, the graphic area 201 with a duty cycle of 100% is a graphic area without actual function, which will not cause any impact on the lithography process, and the graphic area 201 with a duty cycle of 100% also does not need to be subjected to optical proximity correction. Therefore, only the graphic area 201 with actual function needs to be identified, and then correct sampling is performed to make it conform to the graphic distribution in the real circuit layout, so as to further improve the accuracy of the equivalent diffusion length in the optical proximity correction model.

[0052] In this embodiment, the second preset value is 50%.

[0053] It should be noted that in this embodiment, when estimating the layout range of the effective layout after removing the auxiliary graphics in the actual layout 200 for multiple products, the sampling ratio of the corresponding attributes during modeling is not lower than the minimum value of the layout range. This is because the influence of the auxiliary graphics is reflected in the pattern density rather than the bright and dark fields, so it is necessary to remove them to reduce interference terms. Selecting multiple products is to exclude design differences in different designs (such as the gate layer). Different designs and product uses will have a range of variations, so that one model can be applicable to multiple product designs, rather than each product having to be modeled independently.

[0054] Please continue to refer to Figure 2 and Figure 3 , and based on the proportion of the number of the bright field regions and the dark field regions, extract the same proportion of the bright field test patterns 100 and the dark field test patterns 101.

[0055] In this embodiment, the method of extracting the same proportion of the bright field test patterns 100 and the dark field test patterns 101 based on the proportion of the number of the bright field regions and the dark field regions includes: based on the proportion of the number of the bright field regions and the dark field regions, extract the same proportion of the bright field test patterns 100 and the dark field test patterns 101 from each type of the bright field test patterns 100 and the dark field test patterns 101 respectively.

[0056] In a specific embodiment, the bright field regions in the actual layout 200 account for 80%, and the corresponding dark field regions account for 20%. Therefore, when performing extraction sampling, it is also necessary to extract 80% of the bright field test patterns 100 and 20% of the dark field test patterns 101. When allocating to each type, it is also necessary to extract 80% of the bright field test patterns 100 and 20% of the dark field test patterns 101 in each type. Among them, 20% of the dark field test patterns 101 are the anti-patterns corresponding to part of the 80% of the bright field test patterns 100.

[0057] By performing extraction sampling from each type of the bright field test patterns 100 and the dark field test patterns 101, the types of sample patterns for modeling are made more comprehensive, so as to ensure that the influence of different types of test patterns on the model can be fully considered during the modeling process, thereby improving the accuracy and generalization ability of the optical proximity correction model, enabling it to better adapt to various complex situations, as well as enhancing the reliability of the modeling process, avoiding the deviation of the optical proximity correction model caused by a single sample type, and thus enhancing the stability and applicability of the optical proximity correction model in different scenarios.

[0058] Please continue to refer to Figure 2 and Figure 3, an optical proximity correction model is established based on the feature size differences between the extracted bright-field test patterns 100 and the corresponding bright-field photoresist patterns, and between the extracted dark-field test patterns 101 and the corresponding dark-field photoresist patterns.

[0059] By accurately analyzing the quantity ratio of the bright-field area and the dark-field area in the actual layout 200, the bright-field test patterns 100 and the dark-field test patterns 101 are extracted in the same proportion. Based on this, the established optical proximity correction model can significantly improve the pertinence of the photoacid diffusion length. This pertinence enables the optical proximity correction model to more accurately simulate the complex characteristics of photoacid diffusion in the actual process, thereby greatly improving the accuracy and reliability of the optical proximity correction model. Photoacid diffusion is one of the key factors affecting pattern accuracy, and the optical proximity correction model effectively reduces the pattern deviation caused by photoacid diffusion through accurate sampling and simulation. Applying the optical proximity correction model to the optical correction of the actual layout 200 can significantly improve the correction accuracy.

[0060] In this embodiment, the method for forming the bright-field photoresist pattern and the dark-field photoresist pattern includes: fabricating a photomask based on the bright-field test pattern 100, and performing exposure and development on the photoresist layer to obtain the bright-field photoresist pattern on the photoresist layer; fabricating a photomask based on the dark-field test pattern 101, and performing exposure and development on the photoresist layer to obtain the dark-field photoresist pattern on the photoresist layer.

[0061] In this embodiment, the method for establishing the optical proximity correction model includes: mixing a number of extracted bright-field photoresist patterns and a number of dark-field test patterns 101 into a whole, and the photoacid diffusion length in the optical proximity correction model is a single value, thereby simplifying the calculation process.

[0062] In other embodiments, the method for establishing the optical proximity correction model may further include: respectively obtaining the corresponding photoacid diffusion lengths based on the extracted bright-field test patterns and dark-field test patterns, and the photoacid diffusion length in the optical proximity correction model is a weighted average calculated based on the ratio of the bright-field test patterns and the dark-field test patterns, or is a gradient sequence.

[0063] Among them, when the photoacid diffusion length in the optical proximity correction model is the weighted average calculated based on the ratio of the bright-field test pattern and the dark-field test pattern, the method for establishing the optical proximity correction model includes: establishing a first optical proximity correction model based on a plurality of extracted bright-field photoresist patterns, where the first optical proximity correction model has a first photoacid diffusion length; establishing a second optical proximity correction model based on a plurality of extracted dark-field photoresist patterns, where the second optical proximity correction model has a second photoacid diffusion length; assigning corresponding weights to the first photoacid diffusion length and the second photoacid diffusion length respectively based on the ratio of the bright-field test pattern and the dark-field test pattern, and performing weighted summation on the first photoacid diffusion length and the second photoacid diffusion length to obtain the average value of the photoacid diffusion length, and the photoacid diffusion length in the established optical proximity correction model is the weighted average.

[0064] When the photoacid diffusion length in the optical proximity correction model is a gradient sequence calculated based on the ratio of the bright-field test pattern and the dark-field test pattern, the method for establishing the optical proximity correction model includes: establishing a first optical proximity correction model based on a plurality of extracted bright-field photoresist patterns, where the first optical proximity correction model has a first photoacid diffusion length; establishing a second optical proximity correction model based on a plurality of extracted dark-field photoresist patterns, where the second optical proximity correction model has a second photoacid diffusion length, and the photoacid diffusion length in the established optical proximity correction model is the gradient sequence.

[0065] It should be noted that in the field of lithography technology, the optical proximity correction algorithm is one of the key technologies for achieving high-precision pattern transfer. The core of this algorithm lies in accurately modeling and correcting the complex conversion relationship between the mask pattern M(x, y) and the photoresist topography M'(x, y). Specifically, the mask pattern M(x, y) does not directly appear on the photoresist in an ideal state, but undergoes a series of complex physical and chemical processes. Specifically, these include: the optical diffraction effect causes the light to change in direction and intensity when passing through the mask; the concentration distribution of the photoacid (abbreviated as PA) formed by the diffraction light passing through the mask propagating, absorbing, and undergoing photochemical reactions with the photoresist inside the photoresist; the physical / chemical diffusion phenomenon caused by the photoacid reacting with the chemical substances in the photoresist leads to changes in the distribution of chemical reaction substances inside the photoresist; the development process further affects the topography by dissolving and removing the photoresist through the reaction of the developer with the photoacid. These factors act together to finally transform the mask pattern M(x, y) into the photoresist topography M'(x, y), and the TCC (Transmission Cross Coefficient) transmission matrix is a powerful mathematical tool that can establish a matrix containing physical and chemical parameters in complex processes such as optical propagation, chemical reactions, photoacid diffusion, and development dissolution, approximately characterizing the deformation result of the original pattern after the above complex processes. Therefore, in the modeling process, a large number of test mask patterns M(x, y) with different characteristics need to be designed and manufactured, and then the corresponding photoresist topography M'(x, y) is obtained through high-precision measurement means. Next, the mapping relationship between the two is obtained through data fitting, and the difference δ(x, y) between the theoretical fitting value of M(x, y) and the actual M'(x, y) is calculated based on the mapping relationship. This difference intuitively reflects the graphic deviation between the actual lithography process and the calculated fitting prediction value. By optimizing and adjusting the numerical values of many parameters included in the mapping relationship through the algorithm, a relatively small δ(x, y) can be obtained for the vast majority of test patterns, and this mapping relationship is the OPC basic model under this process condition. In addition, many optical system parameters and process-related parameters need to be comprehensively considered in this mapping relationship. Among them, the optical system parameters include the exposure dose D-dose, which determines the total amount of light energy received by the photoresist; the numerical aperture NA, which affects the resolution and imaging ability of the optical system; the coherence coefficient σ, which characterizes the influence of the coherence of the light source on the imaging quality; in addition, there is also the diffusion length L inside the photoresist, which is closely related to the chemical diffusion process; the reaction threshold energy Eth, which determines the minimum energy required for the photoresist to undergo a chemical reaction, etc. In fact, there are dozens of process-related parameters involved, and each parameter affects the final topography of the pattern to varying degrees.

[0066] After the deviation and related parameters are determined, the key step in modeling is to use these differences δ(x, y) and numerous parameters to inversely deduce the determined values of each parameter in the TCC transfer matrix through complex mathematical inversion methods. However, this calculation process is extremely complex and usually no analytical solution can be obtained. Therefore, it is necessary to rely on powerful computer algorithms and solve by numerical methods such as iterative calculation of each parameter. The goal is to determine all components of the TCC transfer matrix so that the root mean square error and the maximum deviation max(δ(x, y)) of all differences δ(x, y) reach the minimum

[0067] Correspondingly, an embodiment of the present invention also provides an optical proximity correction method, including: providing an optical proximity correction model established by using the modeling method in any one of the above embodiments; performing optical proximity correction on the layout to be corrected based on the optical proximity correction model

[0068] It should be noted that when the photoacid diffusion length in the optical proximity correction model is a gradient sequence calculated based on the ratio of the bright-field test pattern 100 and the dark-field test pattern 101, before performing optical proximity correction on the layout to be corrected, it further includes: obtaining the clear ratio of the bright and dark fields in the layout to be corrected; calling the photoacid diffusion length corresponding to the ratio of the bright-field test pattern 100 and the dark-field test pattern 101 in the gradient sequence that is the same as the clear ratio of the bright and dark fields, and performing optical proximity correction on the layout to be corrected through the optical proximity correction model corresponding to the photoacid diffusion length. Wherein, the layout to be corrected and the actual layout 200 used in the modeling can be the same product or different products

[0069] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention should be subject to the scope defined by the claims

Claims

1. A method for modeling an optical proximity correction model, characterized in that Including: Providing a plurality of bright-field test patterns and a plurality of dark-field test patterns; Obtaining the quantity proportion of the bright-field area and the dark-field area in the actual layout; Based on the quantity proportion of the bright-field area and the dark-field area, extracting the same proportion of the bright-field test patterns and the dark-field test patterns; Establishing an optical proximity correction model according to the feature size differences between the extracted bright-field test patterns and the corresponding bright-field photoresist patterns, and between the extracted dark-field test patterns and the corresponding dark-field photoresist patterns.

2. The modeling method of the optical proximity correction model according to claim 1, characterized in that Dividing the bright-field test patterns and the dark-field test patterns based on a first preset value; wherein, the duty cycle of the bright-field test pattern is greater than or equal to the first preset value; the duty cycle of the dark-field test pattern is less than the first preset value.

3. The modeling method of the optical proximity correction model according to claim 2, characterized in that The first preset value is 50%.

4. The modeling method of the optical proximity correction model according to claim 1, wherein, The plurality of bright-field test patterns and the plurality of dark-field test patterns include: a plurality of different types of the bright-field test patterns and the dark-field test patterns.

5. The modeling method of the optical proximity correction model according to claim 4, characterized in that, The plurality of different types of the bright-field test patterns and the dark-field test patterns include: the bright-field test patterns and the dark-field test patterns with periodically changing dense line spacings / gap spacings, the bright-field test patterns and the dark-field test patterns with linearly changing isolated line sizes / gap sizes, the bright-field test patterns and the dark-field test patterns of storage patterns, the bright-field test patterns and the dark-field test patterns of serpentine patterns, and the bright-field test patterns and the dark-field test patterns of T-shaped patterns, being one or more of them.

6. The modeling method of the optical proximity correction model according to claim 4, characterized in that, Under each type, the bright-field test pattern and the corresponding dark-field test pattern are complementary patterns to each other, and the sum of the duty cycle of the bright-field test pattern and the duty cycle of the corresponding dark-field test pattern is 100%.

7. The modeling method of the optical proximity correction model according to claim 1, wherein The method for obtaining the quantity proportion of the bright-field area and the dark-field area in the actual layout includes: dividing the actual layout into a plurality of graphic regions; obtaining the duty cycle in each graphic region; providing a second preset value, dividing the bright-field area and the dark-field area based on the second preset value, and defining the graphic regions with duty cycles greater than or equal to the second preset value as the bright-field area, and defining the graphic regions with duty cycles less than the second preset value as the dark-field area; counting the quantities of the bright-field area and the dark-field area, and taking the ratios of the quantities of the bright-field area and the dark-field area to the quantity of the graphic regions respectively as the quantity proportions of the bright-field area and the dark-field area in the actual layout.

8. The modeling method of the optical proximity correction model according to claim 7, characterized in that Removing the graphic regions with a duty cycle of 100%.

9. The modeling method of the optical proximity correction model according to claim 7, characterized in that The second preset value is 50%.

10. The modeling method of the optical proximity correction model according to claim 4, characterized in that, The method for extracting the same proportion of the bright-field test patterns and the dark-field test patterns based on the quantity proportion of the bright-field area and the dark-field area includes: based on the quantity proportion of the bright-field area and the dark-field area, respectively extracting the same proportion of the bright-field test patterns and the dark-field test patterns from each type of the bright-field test patterns and the dark-field test patterns.

11. The modeling method of the optical proximity correction model according to claim 1, characterized in that, The method for forming the bright-field photoresist pattern and the dark-field photoresist pattern includes: fabricating a photomask based on the bright-field test pattern, and performing exposure and development on the photoresist layer to obtain the bright-field photoresist pattern on the photoresist layer; fabricating a photomask based on the dark-field test pattern, and performing exposure and development on the photoresist layer to obtain the dark-field photoresist pattern on the photoresist layer.

12. The modeling method of the optical proximity correction model according to claim 1, characterized in that The method for establishing the optical proximity correction model includes: mixing a plurality of extracted bright-field photoresist patterns and a plurality of dark-field test patterns as a whole, and the photoacid diffusion length in the optical proximity correction model is a single value.

13. The modeling method of the optical proximity correction model according to claim 1, characterized in that, The method for establishing the optical proximity correction model includes: respectively obtaining corresponding photoacid diffusion lengths based on the extracted bright-field test pattern and the dark-field test pattern, and the photoacid diffusion length in the optical proximity correction model is a weighted average calculated based on the proportions of the bright-field test pattern and the dark-field test pattern, or is a gradient sequence.

14. An optical proximity correction method, characterized in that, Includes: Providing an optical proximity correction model established by using the modeling method according to any one of claims 1 to 13; Performing optical proximity correction on the layout to be corrected based on the optical proximity correction model.

15. The optical proximity correction method according to claim 14, wherein When the photoacid diffusion length in the optical proximity correction model is a gradient sequence calculated based on the proportions of the bright-field test pattern and the dark-field test pattern, before performing optical proximity correction on the layout to be corrected, it further includes: obtaining the bright and dark field proportions in the layout to be corrected; calling the photoacid diffusion lengths corresponding to the bright-field test pattern and the dark-field test pattern in the gradient sequence whose proportions are the same as the bright and dark field proportions.

Images