Optical proximity correction modeling method and system

Abstract

The present invention discloses an optical proximity correction modeling method and system. The present invention obtains multiple film layer data of the same layout pattern for different photoresist thicknesses, and pre-establishes different optical models using multiple film layer data and light source information; selects the corresponding optical model according to the photoresist thickness in the measurement data; constructs a photolithography model by the selected optical model, photoresist model, and mask model; randomly selects values ​​within the value range of optical model parameters, photoresist model parameters, and mask model parameters to generate multiple parameter combinations; uses multiple parameter combinations to perform photolithography simulation on the photolithography model through simulation software, evaluates the photolithography effects under different parameter combinations, and selects the optimal parameter combination. The present invention introduces different optical models in different areas of the film layer structure to reduce the probability of subsequent OPC repairs and improve the accuracy and efficiency of semiconductor manufacturing.

Description

Technical Field

[0001] The present invention belongs to the field of semiconductor technology, and in particular relates to an optical proximity correction (OPC) modeling method and a system thereof. Background Art

[0002] As a core link in the semiconductor manufacturing process, the essence of photolithography technology is to accurately transfer the pre-designed graphics to the wafer surface with the help of a mask, and then construct the complex and diverse microstructures of semiconductor devices. In the actual implementation of the photolithography process, the optical proximity effect is like an elusive "interference source", which always follows and poses a serious threat to the accuracy of the pattern transfer. The optical proximity effect originates from the basic physical properties of light. When light passes through the fine pattern on the mask, due to the diffraction and interference of light, the light intensity distribution formed on the wafer surface is not uniform, but presents a complex change pattern. This uneven light intensity distribution will directly cause the lines at the edge of the pattern to deform during the exposure process. The edges of the originally designed regular pattern may be bent, blurred or have dimensional deviations, which seriously affects the integrity and accuracy of the pattern.

[0003] In order to effectively deal with the challenges brought by the optical proximity effect and ensure that the graphics transferred to the wafer can be as close to the ideal graphics as possible, the optical proximity correction (OPC) technology came into being. OPC technology improves the accuracy and quality of lithography graphics by finely adjusting the mask graphics to compensate for the graphic deformation caused by the optical proximity effect. However, the effective implementation of OPC technology depends on an accurate OPC model, which needs to be able to accurately predict and simulate various physical phenomena in the lithography process to provide reliable guidance for OPC correction.

[0004] At present, the existing OPC models are generally constructed based on a single film stack. This construction method has exposed obvious limitations when facing the complex and changeable situations in actual production. In the manufacturing process of semiconductor devices, due to the existence of the previous layer structure, the thickness of the photoresist at different locations often varies significantly. The OPC model based on a single film layer structure cannot fully consider the changes in the thickness of these underlying photoresists, resulting in the model being unable to handle the same graphics with different underlying layers, and unable to accurately cover and correct these graphics. This means that after the OPC correction, there is still a large deviation between the actual graphics and the expected ideal graphics, which cannot meet the requirements of high-precision semiconductor manufacturing.

[0005] In actual production, when problems are found in the graphics after OPC correction, in order to compensate for this defect, it is usually necessary to take remedial measures such as grabbing specific structures to compensate for the bias. This process is not only cumbersome and complicated, but also requires remaking the mask, which will undoubtedly consume a lot of time and cost, seriously affecting production efficiency and increasing production costs. Summary of the invention

[0006] The object of the present invention is to overcome the above-mentioned deficiencies of the prior art and to provide an optical proximity correction modeling method and system thereof.

[0007] The present invention is implemented in this way. In a first aspect, the present invention provides an OPC modeling method, comprising the following steps:

[0008] Step S1, obtaining multiple film layer data of the same layout pattern with different photoresist thicknesses, and pre-establishing different optical models using the multiple film layer data and light source information; the film layer data includes the film layer structure composition and the thickness, refractive index n, and extinction coefficient k of each film layer;

[0009] Step S2, obtaining measurement data; selecting a corresponding optical model according to the photoresist thickness in the measurement data; constructing a lithography model (Litho Model) by using the selected optical model, the photoresist model, and the mask model;

[0010] Step S3, determining the value ranges of the optical model parameters, the photoresist model parameters, and the mask model parameters, and generating multiple parameter combinations by randomly selecting values ​​for each parameter within the value range; different parameter combinations contain different optical model parameter values, photoresist model parameters, and mask model parameters;

[0011] Step S4: Use multiple parameter combinations to perform lithography simulation on the lithography model constructed in step S2 through simulation software, evaluate the lithography effects under different parameter combinations, screen out the optimal parameter combination, and obtain the required model parameters.

[0012] Preferably, the light source information in step S1 includes wavelength, numerical aperture, and light source type;

[0013] Preferably, the measurement data in step S1 is SEM data of the film structure to be measured;

[0014] Preferably, the thickness of the photoresist in the film layer structure to be tested in step S1 is affected by the previous layer structure;

[0015] More preferably, the front layer structure in step S1 includes an active area and a polysilicon layer covering a portion of the active area;

[0016] Preferably, the optical model parameters in step S3 include an imaging plane (Beamfocus) not considering the refraction of the photoresist and an actual imaging plane (DEF_START);

[0017] Preferably, the photoresistance model parameters in step S3 include diffusion length;

[0018] Preferably, the mask model parameters in step S3 include mask deviation and mask cutting angle;

[0019] Preferably, the process of evaluating the lithography effects under different parameter combinations and selecting the optimal parameter combination in step S4 is as follows:

[0020] First, the difference between the critical dimension of the lithography simulation and the critical dimension of the wafer is calculated, and then the root mean square value of all the differences is calculated; finally, the parameter combination with the smallest difference is selected as the optimal parameter combination.

[0021] In a second aspect, the present invention provides an OPC modeling system, comprising:

[0022] The optical model building module is responsible for obtaining multiple film layer data of the same layout pattern with different photoresist thicknesses, and pre-establishing different optical models using multiple film layer data and light source information;

[0023] The photolithography model building module is responsible for selecting the corresponding optical model from the different optical models built by the optical model building module according to the photoresist thickness in the measurement data; and building the photolithography model by the selected optical model, the photoresist model and the mask model;

[0024] The model parameter selection module is responsible for randomly selecting values ​​from the value range of optical model parameters, photoresist model parameters, and mask model parameters to generate multiple parameter combinations; using multiple parameter combinations to perform lithography simulation on the constructed lithography model through simulation software, evaluate the lithography effects under different parameter combinations, screen out the optimal parameter combination, and obtain the required model parameters.

[0025] The beneficial effects of the present invention are:

[0026] The present invention introduces different optical models taking into account the differences in actual film layer structures, thereby avoiding the problem that the ideal pattern cannot be achieved by using a lithography model that references the same optical model due to the different photoresist thicknesses in different areas of the front layer, reducing the probability of subsequent OPC repairs and improving the accuracy and efficiency of semiconductor manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0028] Figure 1 It is a schematic diagram of the structure of the front layer of a semiconductor device.

[0029] Figure 2 It is an ideal graph for OPC correction.

[0030] Figure 3 It is the actual pattern after OPC correction using the same lithography model.

[0031] Markings in the figure: 1. Active area; 2. Polysilicon; 3. Photoresist; 4. Ion implantation layer. DETAILED DESCRIPTION

[0032] As can be seen from the background technology, the existing OPC model cannot adapt to the different photoresist thicknesses at different positions of the front layer, resulting in large deviations when using the same photolithography model. Figure 1 First, make an active area 1 on the wafer substrate (it can be considered that the entire wafer surface is flat after this), and then make a polysilicon layer 2 on part of the active layer (which makes the entire wafer surface uneven after this), and then SiO2 and photoresist 3 are coated. That is, these wafers contain the front layer structure consisting of the active area 1 and the polysilicon layer 2 covering part of the active area. Because some of the ion implantation layer 4 is on the polysilicon layer 2, the thickness of the photoresist underneath it will be thinner and easier to expose, while the photoresist underneath the ion implantation layer 4 that is not on the polysilicon layer, that is, only on the active area 1, is thicker, and the same energy will not be able to expose it, such as Figure 3 As shown, compared with the ideal graph ( Figure 2 ) has a large deviation.

[0033] Therefore, this embodiment provides an OPC modeling method, comprising the following steps:

[0034] Step S1, obtaining multiple film layer data of the same layout pattern with different photoresist thicknesses, and pre-establishing different optical models using the multiple film layer data and light source information; the film layer data includes the film layer structure composition and the thickness, refractive index n, and extinction coefficient k of each film layer;

[0035] Specifically, the light source information in step S1 includes wavelength, numerical aperture, and light source type; the measurement data is SEM data of the film structure to be measured;

[0036] Specifically, the thickness of the photoresist in the film structure to be tested is affected by the front layer structure, wherein the front layer structure includes an active area and a polysilicon layer covering part of the active area. Therefore, in the film structure of the same layout pattern, the front layer structure has different photoresist thicknesses in different regions because the photoresist layer covers the active area and the polysilicon layer. Therefore, a corresponding optical model can be established for the photoresist thickness in different regions of the front layer, that is, different optical models are introduced in a film structure according to the photoresist thickness.

[0037] Step S2, obtaining measurement data; selecting a corresponding optical model according to the photoresist thickness in the measurement data; constructing a lithography model (Litho Model) by using the selected optical model, the photoresist model, and the mask model;

[0038] Step S3, determining the value ranges of the optical model parameters, the photoresist model parameters, and the mask model parameters, and generating multiple parameter combinations by randomly selecting values ​​for each parameter within the value range; the values ​​of the optical model parameters, the photoresist model parameters, and the mask model parameters included in different parameter combinations are different from each other;

[0039] Step S4: Use multiple parameter combinations to perform lithography simulation on the lithography model constructed in step S2 through simulation software, evaluate the lithography effects under different parameter combinations, screen out the optimal parameter combination, and obtain the required model parameters.

[0040] Specifically, the process of evaluating the lithography effects under different parameter combinations and selecting the optimal parameter combination is as follows:

[0041] First, the difference between the critical dimension of the lithography simulation and the critical dimension of the wafer is calculated, and then the root mean square value of all the differences is calculated; finally, the parameter combination with the smallest difference is selected as the optimal parameter combination.

[0042] The method of the present invention is applicable to the front layer structure, and has important significance in the photolithography process related to the front layer OPC modeling. The front layer can be one or more structural layers that already exist before the layer currently being modeled by OPC. The existence of these layers will affect the photolithography process of the current layer, such as the optical properties such as the propagation, reflection, and refraction of light between different layers, as well as the material properties of different layers (such as the refractive index, absorption coefficient, etc. of different layers), geometric shapes, and other factors that may be related to OPC modeling.

[0043] This embodiment also provides an OPC modeling system, including:

[0044] The optical model building module is responsible for obtaining multiple film layer data of the same layout pattern with different photoresist thicknesses, and pre-establishing different optical models using multiple film layer data and light source information;

[0045] The photolithography model building module is responsible for selecting corresponding optical models in different areas from different optical models built by the optical model building module according to the photoresist thickness in different areas of the front layer in the measurement data; and building a photolithography model from the selected optical model, photoresist model and mask model;

[0046] The model parameter selection module is responsible for randomly selecting values ​​from the value range of optical model parameters, photoresist model parameters, and mask model parameters to generate multiple parameter combinations; using multiple parameter combinations to perform lithography simulation on the constructed lithography model through simulation software, evaluate the lithography effects under different parameter combinations, screen out the optimal parameter combination, and obtain the required model parameters.

[0047] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0048] The front layer structure of this embodiment takes the active area and the polysilicon layer covering the active area as an example. The active area is the key area for realizing electrical functions in semiconductor devices, and the polysilicon layer is used as a gate electrode or interconnection structure. The above-mentioned front layer structure can define the active area on the substrate through process steps such as photolithography and etching. After the active area is formed, the polysilicon layer is deposited on the surface of the active area by chemical vapor deposition (CVD) and other process methods. During the CVD process, a suitable reaction gas (such as silane, etc.) is selected to uniformly deposit and crystallize silicon atoms on the surface of the active area under high temperature and specific reaction conditions to form a polysilicon layer. The deposited polysilicon layer may need to be further annealed to improve its crystallization quality and reduce crystal defects, thereby improving its electrical stability and reliability. For example, by using a rapid thermal annealing (RTA) process, the polysilicon layer is heated to a high temperature (such as about 1000°C) in a short time and cooled rapidly, which can effectively eliminate the stress and defects generated during the deposition process. This front layer structure composed of active area and polysilicon is of great significance in the subsequent photolithography process such as OPC modeling. Due to the existence of the front layer, light will be affected by its material properties (such as refractive index, absorption coefficient, etc.) and geometric shape (such as thickness of active area and polysilicon layer, surface flatness, etc.) during propagation, thus producing optical proximity effect. When modeling OPC, these factors must be fully considered to establish a model that can accurately reflect the influence of the front layer to ensure the accuracy of pattern transfer during the photolithography process. For example, by accurately measuring the optical parameters of the front layer and incorporating them into the calculation of the OPC model, the mask pattern can be reasonably corrected to compensate for the pattern deformation caused by the optical proximity effect, improve the accuracy and quality of the photolithography pattern, and ultimately achieve reliable manufacturing of high-performance semiconductor devices.

[0049] Before the above-mentioned front layer structure is subjected to ion implantation, photoresist treatment is usually required, which is a key step, mainly including photoresist coating, photoresist exposure, photoresist development and other operations. Photoresist coating is to evenly coat the liquid photoresist on the surface of the wafer (including the active area). Photoresist is a photosensitive material, and its function is to protect specific areas from ion implantation during the subsequent ion implantation process, thereby achieving selective doping. For example, when manufacturing transistors, through the masking effect of photoresist, ion implantation can be performed only in areas where source and drain electrodes need to be formed, while other areas are covered and protected by photoresist. Exposure is to use ultraviolet rays or other specific wavelengths of light to irradiate the photoresist through the mask, so that the photoresist undergoes a chemical reaction and changes its solubility in the developer. The mask contains a pre-designed circuit pattern, and the light is irradiated onto the photoresist through the transparent area on the mask, causing the photoresist in the corresponding area to undergo an exposure reaction, while the opaque area on the mask blocks the light, and the photoresist below is not exposed. In this way, the pattern on the mask is transferred to the photoresist, providing an accurate pattern template for subsequent development and ion implantation. Development is the process of removing the unexposed or exposed (depending on the type of photoresist) parts of the photoresist using a developer, thereby forming a three-dimensional pattern on the photoresist that corresponds to the mask pattern, that is, exposing the active area that needs to be implanted, while other areas are still protected by the photoresist.

[0050] Introducing OPC modeling into the photoresist process before ion implantation of the front-layer structure is a complex but critical step, which is essential for improving lithography accuracy and semiconductor device performance. The core goal of OPC modeling is to predict and correct the pattern deformation caused by the optical proximity effect (OPE) through accurate simulation and analysis of the lithography process, ensuring that the pattern formed on the wafer is as close to the design intent as possible. It is based on a deep understanding of light in the photoresist and the front-layer structure (including the active area and polysilicon on the active area). By establishing a mathematical model to describe these processes, OPC modeling can predict the lithography results before actual manufacturing, and adjust the mask pattern according to the predicted results, thereby improving the accuracy of the lithography process.

[0051] Construct an OPC model that includes the influence of the front layer structure. The model includes multiple sub-models such as the optical model, the photoresist model, and the mask model, which interact with each other to simulate the entire photolithography process. In the optical model, the propagation path of light in the front layer structure and the change in light intensity distribution are considered to accurately calculate the optical proximity effect caused by the front layer structure. In the photoresist model, the chemical changes of the photoresist during exposure, post-exposure baking (PEB) and development are simulated by combining the relevant data of the front layer structure and the performance data of the photoresist itself. According to the simulation results of the OPC model, the key parameters in the photoresist process are optimized. For example, adjust the parameters related to the photoresist reaction, such as the diffusion range and the acid-base concentration, to compensate for the optical proximity effect caused by the front layer structure and ensure that the photoresist forms an accurate pattern before ion implantation. If the OPC model predicts that the edge of the photoresist pattern will shrink due to the optical proximity effect under a certain front layer structure, then the exposure dose can be appropriately increased or the development time can be adjusted to expand the pattern size to make it closer to the design requirements.

[0052] Before actually applying the OPC model to the photoresist process before ion implantation in the front layer structure, strict model verification is required. By manufacturing test structures (such as lines with different line widths, spacing, etc.), using the same front layer structure and photoresist process as in actual production, the actual measurement results are compared with the OPC model prediction results. For example, the actual line width, spacing and other dimensions of the photoresist pattern in the test structure are measured and compared with the model prediction value to calculate the error range.

[0053] Based on the verification results, if a large deviation is found between the model prediction and the actual measurement, the OPC model is adjusted for feedback. This may involve re-examining the accuracy of data collection, optimizing the model algorithm, or adjusting the parameters in the model. Through continuous verification and adjustment cycles, the accuracy and reliability of the OPC model in the photoresist process before ion implantation of the front layer structure are gradually improved, ensuring that high-precision photoresist pattern formation can be stably achieved in large-scale production, providing a good foundation for subsequent ion implantation and other processes.

[0054] Specifically, this embodiment is directed to Figure 1 The front-layer structure is OPC modeled, and the implementation process is as follows:

[0055] Step S1: In the front layer structure of the same layout pattern, a polysilicon layer is grown on the upper part of the active area. Therefore, the current front layer structure has two different photoresist (PR) thicknesses. The SEM (scanning electron microscope) data required by the model is collected to construct two film stacks, which are recorded as film stack A and film stack B. Film stack A is Figure 1Film layer data set of active area, photoresist thickness is 3000A; film stack B is in Figure 1 Film dataset of polysilicon layer, photoresist thickness is 2000A;

[0056] Select representative wafers from the semiconductor manufacturing process, such as Figure 1 First, make an active area 1 on the wafer substrate (after completion, the entire wafer surface is flat), and then make a polysilicon layer 2 on part of the active layer (after completion, the entire wafer surface is uneven), and then SiO2 and photoresist 3 are coated. That is, these wafers contain the front layer structure consisting of the active area and the polysilicon layer covering it, and have completed some of the preliminary process steps (such as oxidation, deposition, etc.), which is consistent with the actual production situation. For example, in a logic chip manufacturing process, a wafer that has undergone the front-end process (such as the preliminary process of transistor manufacturing) is selected, and the structure and characteristics of its active area and polysilicon layer can reflect the actual production situation, so as to obtain accurate SEM data for OPC modeling.

[0057] If the wafer is too large to be directly observed by SEM, it may be necessary to cut it into smaller samples to facilitate subsequent operations. The cutting process must ensure the integrity of the sample and avoid damage to the active area and polysilicon layer structure.

[0058] Before SEM observation, the sample surface must be strictly cleaned to remove surface contaminants (such as particles, organic matter, etc.), because these contaminants will interfere with the quality of the SEM image and affect the accuracy of the data. A combination of multiple cleaning methods is used, such as the use of chemical reagents (such as diluted hydrofluoric acid to remove surface oxides, acetone and ethanol to remove organic matter) for cleaning, and then combined with ultrasonic cleaning technology to enhance the cleaning effect. The frequency and time of ultrasonic cleaning should be optimized according to the characteristics of the sample to avoid damage to the sample structure. For example, for samples with relatively fragile active areas and polysilicon layer structures, the ultrasonic frequency is selected between 40kHz and 80kHz, and the cleaning time is controlled between 5 and 10 minutes to ensure that the sample structure is protected while effectively removing contaminants.

[0059] Select the appropriate SEM working mode according to the sample characteristics that need to be observed (such as the microstructure of the active area and polysilicon layer, interface characteristics, etc.), and optimize the SEM parameter settings.

[0060] Determine the area for collecting SEM data, covering the key parts of the active area and polysilicon layer, including active areas at different locations (such as active areas near the edge and center of the chip, because there may be differences in process uniformity) and different parts of the polysilicon layer (such as the polysilicon gate area, the interface area with the active area, etc.). Use a systematic scanning method, such as line-by-line scanning or spiral scanning, to ensure that the collected data is representative and complete.

[0061] During the data collection process, the location information (such as coordinates) corresponding to each data point should be recorded so that the characteristics of the active area and polysilicon layer at different locations can be accurately associated when the data is analyzed later. At the same time, for each observation area, multiple sets of data (such as image data at different magnifications) are collected to fully understand the sample structure from macro to micro.

[0062] Quantitatively analyze the collected SEM raw data and extract the key dimensions of the pattern after wafer exposure and development, which is the wafer measurement value;

[0063] Store the processed data in a suitable format, establish a database or data file system, and facilitate data retrieval at any time during the subsequent OPC modeling process. The data storage structure should be designed reasonably to clearly classify and identify SEM data collected from different wafers, different areas, and different process steps to ensure data traceability and ease of management. For example, store data according to wafer batches, process steps (such as before lithography, after lithography, etc.), observation areas (such as different quadrants of the chip), etc., and record the time of data collection, SEM equipment parameters and other related information at the same time, to provide a comprehensive reference for data analysis and model optimization.

[0064] Step S2, performing model calibration tuning according to the two film stacks and light source information, and then constructing two optical models, denoted as Optical Model A and Optical Model B; the light source information includes wavelength, numerical aperture, light source type and other related parameters;

[0065] Step S3, obtaining the measurement data in Table 1; selecting the corresponding optical model according to the photoresist thickness in the measurement data; if the PR value is 3000, calling Optical model A, and if it is 2000, calling Optical model B;

[0066] Construct a lithography model (Litho Model) based on the selected optical model, photoresist model and mask model;

[0067] Table 1 Gauge

[0068] Measurement 1 Measurement 2 Measurement 3 Measurement 4 Measurement 5 Measurement 6 Measurement 7 Measurement 8 Measurement graphic name (Struc) 1D_Anochor_CD90_P180 1D_Anochor_CD90_P180 1D_Anochor_CD90_P180 1D_Anochor_CD90_P180 1D_Anochor_CD90_P180 1D_Line_CD80_P160 1D_Line_CD85_P170 1D_Line_CD90_P180 Array ID (GID) 1 2 3 4 5 6 7 8 X1 95199000 95199000 95199000 95199000 95199000 95299000 95399000 95299000 Y1 8940000 9040000 9140000 9240000 9340000 8940000 8940000 9040000 X2 95201000 95201000 95201000 95201000 95201000 95301000 95401000 95301000 Y2 8940000 9040000 9140000 9240000 9340000 8940000 8940000 9040000 Measuring the size of the position mask (Drawn) 90 90 90 90 90 80 85 90 Wafer measurement value (Meas) 89.89 90.15 89.93 90.72 89.63 80.69 84.70 90.83 Photoresist thickness 3000 2000 3000 2000 3000 2000 3000 2000

[0069] Among them, (X1, Y1) and (X2, Y2) represent the starting and ending coordinates of the measurement point respectively.

[0070] Step S4, determining the value ranges of optical model parameters, photolithography model parameters, mask model parameters, and mask parameters, and generating multiple parameter combinations by randomly selecting values ​​for each parameter within the value range; the values ​​of optical model parameters, photoresist model parameters, and mask model parameters included in different parameter combinations are different from each other; the optical model parameters include Beamfocus (without considering the photoresist refraction imaging plane) and DEF_START (actual imaging plane), and there is a fixed relationship between the two, and the value range is 0~the entire film layer structure thickness; the photoresist model parameters include photoresist-related parameters resist term (such as diffusion length, whose value is 0~200); the mask model parameters include mask deviation (value range -0.001~0.001) and cut angle (value range 0~0.012um).

[0071] Step S5: Use multiple parameter combinations to perform lithography simulation on the lithography model constructed in step S3 through simulation software, evaluate the lithography effects under different parameter combinations, screen out the optimal parameter combination, and obtain the required model parameters.

[0072] The process of evaluating the lithography effects under different parameter combinations and selecting the optimal parameter combination is as follows:

[0073] First, calculate the difference between the critical dimensions of lithography simulation and the critical dimensions of wafers; the critical dimensions of lithography simulation refer to the specific size parameters of lithography patterns obtained during lithography simulation, such as line width, spacing, hole diameter, etc. The critical dimensions of wafer etching are the corresponding graphic dimensions measured after the actual wafer exposure and development process. Calculating the difference between the two is an important basis for evaluating the lithography effect. For example, for a lithography pattern with a designed line width of 100nm, the line width obtained by lithography simulation may be 102nm, while the line width measured after wafer etching is 103nm, then the difference is 103nm -102nm = 1nm. This difference reflects the deviation of the graphic size during the lithography process. The smaller the difference, the more accurate the lithography model is in predicting the size of the lithography pattern. When calculating the difference, it is necessary to measure and statistically analyze multiple graphics at different locations on the wafer to obtain representative difference data, because different areas of the wafer may have different etching effects due to factors such as process uniformity.

[0074] Then, in order to comprehensively evaluate the consistency and accuracy of the lithography pattern size on the entire wafer, the root mean square value of all differences is calculated; for example, if there is a set of difference data of [0.5nm, 1.2nm, -0.8nm, 2.0nm, -1.0nm], its average value is 0.38nm, and the root mean square value is 1.14nm. The smaller the root mean square value, the more concentrated the size distribution of the lithography pattern on the wafer, and the better the stability and repeatability of the lithography process. By calculating the root mean square value, the lithography effects under different parameter combinations can be quantitatively compared, providing an objective basis for screening the optimal parameter combination.

[0075] Finally, the parameter combination with the smallest difference is selected as the optimal parameter combination.

[0076] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. An optical proximity correction modeling method, characterized in that The method comprises the following steps: Step S1, in the film layer structure of the same layout pattern, the front layer structure has different photoresist thicknesses in different regions because the photoresist layer is covered on the active area and the polysilicon layer, and a plurality of film layer data of different photoresist thicknesses in different regions of the front layer in a film layer structure are obtained, and different optical models are pre-established using the plurality of film layer data and light source information; wherein the film layer data includes the film layer structure composition and the thickness, refractive index, and extinction coefficient of each film layer; Step S2, obtaining measurement data; selecting a corresponding optical model according to the photoresist thickness in the measurement data; constructing a photolithography model based on the selected optical model, the photoresist model, and the mask model; Step S3, determining the value ranges of the optical model parameters, the photoresist model parameters, and the mask model parameters, and generating a plurality of parameter combinations by randomly selecting values ​​for each parameter within the value range; the optical model parameter values, photolithography model parameters, and mask model parameters included in different parameter combinations are different from each other; Step S4: Use multiple parameter combinations to perform lithography simulation on the lithography model constructed in step S2 through simulation software, evaluate the lithography effects under different parameter combinations, screen out the optimal parameter combination, and obtain the required model parameters.

2. The optical proximity correction modeling method according to claim 1, characterized in that: The light source information includes wavelength, numerical aperture, and light source type.

3. The optical proximity correction modeling method according to claim 1, characterized in that: The measurement data is SEM data of the film layer structure to be measured.

4. The optical proximity correction modeling method according to claim 1, characterized in that: The optical model parameters include an imaging plane that does not consider the refraction of the photoresist and an actual imaging plane.

5. The optical proximity correction modeling method according to claim 1, characterized in that: The photoresist model parameters include diffusion length.

6. The optical proximity correction modeling method according to claim 1, characterized in that: The mask model parameters include mask deviation and mask cutting angle.

7. The optical proximity correction modeling method according to claim 1, characterized in that: The process of evaluating the lithography effects under different parameter combinations and selecting the optimal parameter combination is as follows: First, the difference between the critical dimension of the lithography simulation and the critical dimension of the wafer is calculated, and then the root mean square value of all the differences is calculated; finally, the parameter combination with the smallest difference is selected as the optimal parameter combination.

8. An optical proximity correction modeling system based on the method according to any one of claims 1 to 7, characterized in that include: The optical model building module is responsible for obtaining multiple film layer data of the same layout pattern with different photoresist thicknesses, and pre-establishing different optical models using multiple film layer data and light source information; The photolithography model building module is responsible for selecting the corresponding optical model from the different optical models built by the optical model building module according to the photoresist thickness in the measurement data; and building the photolithography model by the selected optical model, the photoresist model and the mask model; The model parameter selection module is responsible for randomly selecting values ​​from the value range of optical model parameters, photoresist model parameters, and mask model parameters to generate multiple parameter combinations; using multiple parameter combinations to perform lithography simulation on the constructed lithography model through simulation software, evaluate the lithography effects under different parameter combinations, screen out the optimal parameter combination, and obtain the required model parameters.

Images