CN114488680A - OPC model data collection method, data collection system and OPC model optimization method - Google Patents

Abstract

The OPC model data collection method comprises the following steps: simulating a Manhattan layout of the test pattern by adopting an initial OPC model to obtain a simulation result; inputting the simulation result into a measuring tool, and establishing a characteristic dimension measuring standard; and acquiring the characteristic dimension of the actual wafer pattern of the test pattern by adopting the characteristic dimension measurement standard to obtain collected data. According to the OPC model data collection method and the data collection system, the initial OPC model is adopted to simulate the Manhattan layout of a test pattern, then the simulation result is adopted to establish the characteristic dimension measurement standard, the simulated layout has a small dimension difference with an actual wafer, the characteristic dimension measurement standard established by the method can be better matched with the actual wafer pattern, a good data collection result is ensured, the data collection is accurate, and a more accurate model calibration result can be obtained.

Description

OPC model data collection method, data collection system and OPC model optimization method

Technical Field

The invention relates to the technical field of semiconductor manufacturing, in particular to an OPC model data collection method, a data collection system and an OPC model optimization method.

Background

Photolithography is an important industrial step in the production of semiconductor devices that transfers a pattern structure printed on a photomask to the surface of a wafer. As integrated circuits evolve, semiconductor fabrication techniques continue to evolve toward smaller dimensions, and the feature sizes of semiconductor devices are even smaller than the wavelength of light from the light source used in the photolithography process. In this case, the pattern on the mask is deformed when transferred, that is, an Optical Proximity Effect (Optical Proximity Effect) occurs, due to a diffraction Effect of light. The optical proximity effect causes a large difference between the actual pattern projected onto the wafer and the designed target pattern, thereby affecting the lithography quality of the adjacent pattern regions on the mask pattern, and further affecting the circuit performance and the production yield.

In order to eliminate the influence of the Optical Proximity effect, an Optical Proximity Correction (OPC) method is generally used. The method corrects an original pattern to be exposed on a semiconductor substrate of a silicon wafer by using computer software to obtain a target pattern different from the original pattern, then a photomask is manufactured according to the target pattern, and when photoetching is carried out, the pattern obtained by projecting the photomask on the semiconductor substrate can be almost the same as the original pattern, thereby making up the problem caused by the optical proximity effect.

The current OPC model is corrected by using CD-SEM (Critical Dimension-Scanning Electron Microscope) to test and collect the Critical Dimension in the mask or the chip on the wafer, screening the relevant collected data and feeding back the data to the OPC model, and performing comprehensive operation by combining modeling information. However, due to the optical proximity effect of the wafer during data collection, the dimension of the measured pattern adopting the CD-SEM standard has a larger deviation from the dimension of the actual pattern, the characteristic dimension of the wafer cannot be accurately captured, the data collection and screening are complex, the time is consumed, the accuracy of feedback data is invisibly increased, the correction time is prolonged, and the production cost is increased.

Disclosure of Invention

In view of the above problems, an object of the present invention is to provide an OPC model data collection method, a data collection system, and an OPC model optimization method, which solve the problems in the prior art.

According to a first aspect of the present invention, there is provided an OPC model data collection method, including:

simulating a Manhattan layout of the test pattern by adopting an initial OPC model to obtain a simulation result;

inputting the simulation result into a measuring tool, and establishing a characteristic dimension measuring standard; and

and acquiring the characteristic dimension of the actual wafer pattern of the test pattern by adopting the characteristic dimension measurement standard to obtain collected data.

Optionally, after the step of acquiring the feature size of the actual wafer pattern of the test pattern by using the feature size measurement standard to obtain the collected data, the method further includes:

inputting the collected data into the initial OPC model for model optimization to obtain an iterative OPC model;

and adopting the iterative OPC model to collect data again.

Optionally, before the step of simulating the manhattan layout of the test pattern by using the initial OPC model to obtain a simulation result, the method further includes:

and establishing an initial OPC model according to the basic data and the historical collection data of the test pattern.

Optionally, after the step of collecting data again by using the iterative OPC model, the method further includes:

and inputting the collected data into the iterative OPC model, and calibrating the iterative OPC model.

Optionally, the step of simulating the manhattan layout of the test pattern by using the initial OPC model to obtain the simulation result includes:

an initial OPC model is adopted to simulate the Manhattan layout of the test pattern with a one-dimensional structure to obtain a first simulation result,

the characteristic dimension measurement standard established by the first simulation result is adopted as a first characteristic dimension measurement standard, and the collected data acquired by the first characteristic dimension measurement standard is adopted as first collected data.

Optionally, the performing data collection again by using the iterative OPC model includes:

simulating the Manhattan layout of the test pattern of the two-dimensional structure by adopting the iterative OPC model to obtain a second simulation result;

inputting the second simulation result into a measuring tool, and establishing a second characteristic dimension measuring standard;

and acquiring the characteristic dimension of the actual wafer pattern of the test pattern of the two-dimensional structure by adopting the second characteristic dimension measurement standard to obtain second collected data.

Optionally, after the step of simulating the manhattan layout of the test pattern of the one-dimensional structure by using the initial OPC model to obtain the first simulation result, the method further includes:

and screening out the data with the simulation data size smaller than a threshold value in the first simulation result.

According to a second aspect of the present invention, there is provided an OPC model optimization method, comprising:

data collection is carried out by adopting the OPC model data collection method;

and optimizing the OPC model according to the difference between the collected characteristic dimension of the actual wafer pattern and the characteristic dimension of the test pattern.

According to a third aspect of the present invention, there is provided an OPC model data collection system comprising:

the first simulation module is used for simulating a Manhattan layout of a test pattern of a one-dimensional structure by adopting an initial OPC model to obtain a first simulation result;

the first standard establishing module is used for inputting the first simulation result into a measuring tool and establishing a first characteristic dimension measuring standard; and

and the first collection module is used for collecting the characteristic dimension of the actual wafer pattern of the test pattern by adopting the first characteristic dimension measurement standard to obtain first collection data.

Optionally, the OPC model data collection system further includes:

the initial model building module is used for building an initial OPC model according to basic data and historical collection data of the test pattern;

the data screening module is used for screening out the data of which the simulation data size is smaller than a threshold value in the first simulation result;

the iteration module is used for inputting the first collected data into the initial OPC model for model optimization to obtain an iteration OPC model;

the second simulation module is used for simulating the Manhattan layout of the test pattern of the two-dimensional structure by adopting the iterative OPC model to obtain a second simulation result;

the second standard establishing module is used for inputting the second simulation result into a measuring tool and establishing a second characteristic dimension measuring standard;

the second collection module is used for collecting the characteristic dimension of the actual wafer pattern of the test pattern of the two-dimensional structure by adopting the second characteristic dimension measurement standard to obtain second collection data;

and the calibration module is used for inputting the collected data into the iterative OPC model and calibrating the iterative OPC model.

The OPC model data collection method, the data collection system and the OPC model optimization method provided by the invention have the advantages that the initial OPC model is firstly adopted to simulate the Manhattan layout of a test pattern, then the obtained simulation result is adopted to establish the characteristic dimension measurement standard, and the actual dimension of a wafer is measured according to the established characteristic dimension measurement standard, so that the data collection is realized.

Furthermore, the data collection method of the OPC model comprises the steps of firstly simulating a Manhattan layout of a test pattern with a one-dimensional structure, establishing a characteristic dimension measurement standard according to a simulation result, collecting data according to the characteristic dimension measurement standard, then returning the collected result to the OPC model for optimization iteration to obtain an iterative OPC model, then simulating the Manhattan layout of the test pattern with a two-dimensional structure by using the iterative OPC model, establishing the characteristic dimension measurement standard according to the simulation result, collecting data according to the characteristic dimension measurement standard, realizing gradual collection from the one-dimensional data to the two-dimensional data, enabling the model to be more accurate during two-dimensional data collection by collecting and screening the one-dimensional data, reducing the difficulty of directly collecting the two-dimensional data, improving the collection precision of the two-dimensional data, and enabling the data collection result to be more accurate.

Drawings

The above and other objects, features and advantages of the present invention will become more apparent from the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

FIGS. 1 a-1 e respectively illustrate electron microscope images acquired from a plurality of data acquired by a conventional data collection method;

FIG. 2 shows a flow chart of an OPC model data collection method according to a first embodiment of the present invention;

FIGS. 3 a-3 c are diagrams illustrating comparison of an OPC model data collection method before and after simulation of a plurality of Manhattan layouts, respectively, according to an embodiment of the invention;

FIGS. 4 a-4 b are diagrams illustrating the results of data collection by two exemplary OPC model data collection methods according to embodiments of the present invention;

FIG. 5 shows a flow chart of an OPC model data collection method according to a second embodiment of the present invention;

FIG. 6 illustrates a simplified flow diagram of a method for OPC model optimization according to an embodiment of the present invention;

FIG. 7 shows a schematic block diagram of an OPC model data collection system in accordance with an embodiment of the present invention.

Detailed Description

Various embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Like elements in the various figures are denoted by the same or similar reference numerals. For purposes of clarity, the various features in the drawings are not necessarily drawn to scale. In addition, certain well known components may not be shown.

The following detailed description of the present invention is provided in connection with the accompanying drawings and examples.

Fig. 1a to 1e respectively show electron microscope images obtained by acquiring a plurality of data in a conventional data collection method.

The OPC model needs to collect data in the process of establishing, and the model is adjusted and optimized according to the collected data, so that the mask plate can be repaired. The most time-consuming part of the modeling process is data collection and data consolidation. Data collection is currently mainly performed by Hitachi (Hitachi) measuring instruments) and measuring tools of applied materials companies by inputting manhattan layout (GDS) of test patterns (GDS: graphic Data Stream) to identify the measurement location and boundary, and then establish the CD-SEM measurement standard offline, and then use the measurement standard to realize the online measurement of the actual wafer size to obtain the collected Data. And the data processing is to process collected data of measurement invalidity, measurement failure and measurement error, remove the problem data, then perform offline manual measurement correction or re-measurement, and re-collect data.

In the process of establishing the CD-SEM measurement standard by inputting the manhattan layout using the measurement tool, problems such as invalid data measurement, measurement failure, measurement error, etc. may occur, which will be described below with reference to the examples of fig. 1a to 1 e.

As shown in fig. 1a, the feature size of the measured pattern is too small, which easily causes the connection between the patterns, or the size is too small to capture, and the white dotted line in the figure is an anchor point, which does not measure the required data, resulting in ineffective measurement.

As shown in fig. 1b and fig. 1c, the white dotted line in the figure is a measurement point, and the pattern enclosed by the white solid line is an image of an actual wafer under a scanning electron microscope, it can be obviously observed that a part of the measurement point of the white dotted line has exceeded the position of the actual wafer to be measured, resulting in measurement failure. The reason for the measurement failure is mainly due to the complex pattern on the wafer, and the CD-MES measurement box built off-line cannot completely and correctly capture the boundary of the pattern.

As shown in fig. 1d and fig. 1e, the white dotted line is the boundary of the manhattan layout, and the pattern enclosed by the white solid line is the image of the actual wafer under the scanning electron microscope, and the two are overlapped. When the actual wafer measurement position and boundary are identified through the manhattan layout, it is obvious that the manhattan layout pattern has a larger deviation from the actual wafer pattern, such as a square boundary pattern and an elliptical boundary pattern in fig. 1d, and a square and a circular pattern in fig. 1 e. Then, the existing manhattan graph establishing measurement standard cannot intuitively reflect the graph shape of the actual wafer, so that the measurement point is selected incorrectly, and the measurement is incorrect.

In summary, in the data collection stage, the conventional OPC model uses the CD-SEM measurement standard established offline through the manhattan layout, and because the difference between the measurement standard and the actual wafer pattern is large, the actual shape of the wafer cannot be correctly represented, which causes a series of measurement problems, which causes difficulty in data collection and screening, and consumes time and cost. Therefore, the improved OPC model data collection method is provided, and the accuracy of data collection can be improved. As described in detail below in conjunction with fig. 2-7.

FIG. 2 is a flowchart illustrating an OPC model data collection method according to a first embodiment of the present invention.

As shown in fig. 2, an OPC model data collection method is provided, which includes the following specific steps:

in step S101, an initial OPC model is established based on the basic data of the test pattern and the history collection data.

In this step, an initial OPC model is established based on the basic data and the history collected data of the test pattern collected and evaluated in the early process window, the OPC model is based on the history collected data of the test pattern, and the initial OPC model is subsequently used to complete optimization of data collection, which is also called data collection.

In step S102, an initial OPC model is used to simulate a manhattan layout of a test pattern to obtain a simulation result.

In the step, the initial OPC model established in the last step is adopted to simulate the Manhattan layout of the test pattern so as to obtain a simulated graph of the Manhattan layout, and the simulated graph has high shape coincidence degree with the actual graph of the wafer. Generally, generating an OPC test mask requires generating a plurality of GDS (Graphic data stream) files of a manhattan layout of a test pattern, each GDS file includes one type of OPC test pattern, and then a plurality of GDS files are arranged in an effective exposure area of the mask as required to generate a JDV (Job drop View, photomask data detection) file and output the JDV file. The Manhattan layout is matched with an ideal test pattern in shape and is a graph with square corners, the corners of the graph formed on the silicon substrate have certain radian due to the influence of light diffraction, and therefore the problems of measurement errors, failure and the like caused by the fact that the measuring points of the actual graph of the wafer are not easy to grab by directly establishing a CD-MES measuring standard through the Manhattan layout are solved. The simulation graph obtained by simulating the Manhattan layout by adopting the OPC model is actually a pattern with radian corners and extremely matched with the actual etched pattern of the wafer, so that the shape and the size of the actual graph of the wafer can be accurately captured by the re-established CD-MES measurement standard.

In step S103, the simulation result is input into the metrology tool to establish a feature size measurement standard.

In this step, the simulation result of the manhattan layout obtained in the previous step is input to a Hitachi Design gauge or an AMAT OPCC (applied materials company test tool) to establish a CD-SEM feature size measurement standard, and the feature size of the actual wafer pattern can be accurately grasped according to the feature size measurement standard established by the simulated pattern.

In step S104, the feature size of the actual wafer pattern of the test pattern is collected by using the feature size measurement standard to obtain collected data.

In this step, the feature size of the actual wafer pattern of the test pattern is collected as collected data by using the feature size measurement standard established in the previous step.

In step S105, the collected data is input into the initial OPC model for model optimization, so as to obtain an iterative OPC model.

In the step, the coincidence degree of the collected data obtained by adopting the characteristic dimension measurement standard and the characteristic dimension data of the actual graph of the wafer is high, the collected data is input into an initial OPC model for model optimization to obtain an iterative OPC model, and then a new model is adopted to collect data again, so that the obtained result is more accurate.

In step S106, data collection is resumed using the iterative OPC model.

In the step, the obtained iterative OPC model is adopted to collect data again to obtain a more accurate collection result, so that the OPC model is subjected to subsequent adjustment and optimization, the difference between the mask plate and the actual wafer size can be accurately obtained, the model is better corrected, and the wafer is more accurately etched.

The OPC model data collection method provided by the invention firstly adopts an initial OPC model to simulate the Manhattan layout of a test pattern, then adopts the obtained simulation result to establish a characteristic dimension measurement standard, measures the actual dimension of a wafer according to the established characteristic dimension measurement standard, and realizes data collection.

FIGS. 3 a-3 c are diagrams illustrating comparison between an OPC model data collection method and a plurality of Manhattan layouts before and after simulation, respectively, according to an embodiment of the invention.

Fig. 3 a-3 c reflect the comparison graphs before and after the simulation in step S102 of fig. 2, and the simulation of manhattan layout by using the OPC model can obtain a layout completely approximating the morphology of the actual wafer pattern, and the simulated layout is input into a test tool, so that a correct and reliable CD-SEM measurement standard can be established, and the correct measurement of the subsequent pattern can be realized.

In fig. 3 a-3 c, the manhattan layout of the test pattern is on the left and the simulated pattern is on the right. It can be seen by comparison that the simulation graph has a larger change than the shape of the manhattan layout, and can reflect the appearance of the actual wafer more truly, and the problems of ineffective measurement, measurement failure and measurement error in the measurement process in fig. 1a to fig. 1e can be basically solved through the CD-MES standard established by the simulation graph. Meanwhile, the accuracy of measurement is improved, a large amount of time and cost are saved, and a reliable basis is provided for the automatic off-line establishment of the CD-SEM characteristic dimension measurement standard.

FIGS. 4 a-4 b are diagrams illustrating the results of data collection by two exemplary OPC model data collection methods according to embodiments of the present invention.

Fig. 4 a-4 b show the display results under an electron microscope for capturing the actual wafer pattern by the data collection method according to the first embodiment of the present invention. In the figure, a black solid line indicates measurement points where data capture is performed by using a characteristic dimension measurement standard, and a black pattern indicates a display pattern of an actual wafer under an electron microscope. It can be seen that the coincidence degree between the captured boundary of the measuring point and the graph of the actual wafer is extremely high, and the wafer size can be accurately acquired, which further verifies that the characteristic size of the actual wafer can be accurately measured at one time by using the OPC model data collection method of the first embodiment of the present invention, reduces the time consumed in the data collection and data arrangement processes, reduces the difficulty of data arrangement, and well solves the problems in the prior art.

FIG. 5 shows a flowchart of an OPC model data collection method according to a second embodiment of the present invention.

The OPC model data collection method according to the second embodiment of the present invention includes steps S201 to S210, which are described step by step below. The embodiment is basically similar to the first embodiment, and is different from the first embodiment in that in the present embodiment, data of a test pattern with a one-dimensional structure is firstly used for simulation, then the obtained collected data is returned to an OPC model for optimization iteration, then the test pattern with a two-dimensional structure is used for simulation, and the collected data is obtained again, so that the accuracy of the collected data is ensured, the accuracy of the model is improved, and the accuracy of a final data collection result is improved.

In step S201, an initial OPC model is built based on the basic data of the test pattern and the history collection data.

This step is the same as step S101 in the embodiment of fig. 2, and is not described again here.

In step S202, an initial OPC model is used to simulate a manhattan layout of a test pattern of a one-dimensional structure, so as to obtain a first simulation result.

In the step, an initial OPC model is utilized, a test pattern of a one-dimensional structure is selected as a first part measuring point, and a Manhattan layout of the test pattern of the one-dimensional structure is simulated by the initial OPC model to obtain a first simulation result. Generally, a test pattern may be divided into a one-dimensional pattern and a two-dimensional pattern, wherein the one-dimensional pattern refers to a pattern having periodicity only in one direction, and the structure is relatively simple, such as lines; the two-dimensional pattern refers to a pattern having periodicity in both directions, and has a complicated structure, such as a contact hole, an L-pattern, a stitch pattern, and the like. The one-dimensional graph has simpler lines, so that the test is simpler than that of a two-dimensional graph, and a simulation result can be obtained more easily.

In step S203, the data in the first simulation result, the simulation data size of which is smaller than the threshold value, is screened out.

In the step, a graph with an excessively small data size is screened out according to a simulation result of a test pattern with a one-dimensional structure, namely when the data size of the graph obtained after simulation is smaller than a certain threshold value, the graph data is deleted, so that subsequent measurement is invalid, the problem that measurement of part of points is invalid can be solved by establishing a characteristic size measurement standard after the data are screened out, and repeated screening in a data sorting stage is avoided.

In step S204, the first simulation result is input into a metrology tool to establish a first feature size metrology standard.

Similar to step S103, the first simulation result of the one-dimensional structure test pattern is input to a test tool to establish a first characteristic dimension measurement standard (CD-SEM measurement standard).

In step S205, a first collected data is obtained by collecting a feature size of an actual wafer pattern of the test pattern using a first feature size measurement standard.

In this step, the first characteristic dimension measurement standard established in the previous step is used to collect the characteristic dimension of the actual wafer pattern of the test pattern, and the obtained data is used as the first collected data.

In step S206, the first collected data is input into the initial OPC model for model optimization, so as to obtain an iterative OPC model.

The step is the same as the step S105, and the first collected data is input into the initial OPC model again to perform the optimization iteration of the model, so as to perform the simulation of the test pattern of the two-dimensional structure and the establishment of the measurement standard, thereby reducing the difficulty and pressure for directly establishing the measurement standard of the test pattern of the two-dimensional structure, and improving the accuracy of the measurement standard.

Step S106 of fig. 2 may include steps S207 to S209 of the present embodiment, and the specific steps are as follows.

In step S207, an iterative OPC model is used to simulate the manhattan layout of the test pattern of the two-dimensional structure, so as to obtain a second simulation result.

In the step, the iterative OPC model is adopted to simulate the Manhattan layout of the test pattern of the two-dimensional structure again to obtain a second simulation result of the two-dimensional graph.

In step S208, the second simulation result is input into the metrology tool to establish a second feature size measurement standard.

In this step, the second simulation result of the two-dimensional structure test pattern is input to the test tool to establish the second characteristic dimension measurement standard (CD-SEM measurement standard), which mainly solves the problems of measurement failure and measurement error shown in fig. 1 b-1 e.

In step S209, the feature size of the actual wafer pattern of the test pattern of the two-dimensional structure is collected by using the second feature size measurement standard, so as to obtain second collected data.

In this step, the second characteristic dimension measurement standard established in the previous step is used to collect the characteristic dimension of the actual wafer pattern of the test pattern, and the obtained data is used as second collected data. Thus, data collection for both one-dimensional graphics and two-dimensional graphics is completed.

In step S210, the collected data is input into the iterative OPC model, and the iterative OPC model is calibrated.

Through the steps S201-S209, the data collection is basically completed, the data arrangement basically does not need to take time, the problem of inaccurate data measurement in the prior art is well solved, and the final calibration and calibration of the OPC model can be directly performed through the steps. Therefore, in this step, the two collected data are input into the iterative OPC model, and the iterative OPC model is calibrated.

Of course, the calibration of the iterative OPC model can also be performed directly at the last step of the first embodiment.

The data collection method of the OPC model comprises the steps of firstly simulating a Manhattan layout of a test pattern of a one-dimensional structure, establishing a characteristic dimension measurement standard according to a simulation result, collecting data according to the characteristic dimension measurement standard, then returning the collected result to the OPC model for optimization iteration to obtain an iteration OPC model, then simulating the Manhattan layout of the test pattern of the two-dimensional structure by using the iteration OPC model, establishing the characteristic dimension measurement standard according to the simulation result, collecting data according to the characteristic dimension measurement standard, realizing gradual collection from one-dimensional data to two-dimensional data, enabling the model to be more accurate during two-dimensional data collection by collecting and screening the one-dimensional data, reducing the difficulty of directly collecting the two-dimensional data, improving the collection precision of the two-dimensional data, and enabling the data collection result to be more accurate.

FIG. 6 is a simplified flowchart illustrating an OPC model optimization method according to an embodiment of the present invention. The OPC model optimization method of the present embodiment actually uses the above-described OPC model data collection method, and then optimizes the model using the collected data. The method comprises the following specific steps:

in step S301, an OPC model is built using the basic data, the historical data, and the finite element analysis data.

Similar to the embodiments of fig. 2 and 5, a preliminary OPC model is created using basic data, historical data, finite element analysis data, etc.

In step S302, test pattern data is simulated using an OPC model.

In this step, the OPC model established in the previous step is adopted to simulate the Manhattan layout of the test pattern, and here, the one-dimensional graph and the two-dimensional graph can also be simulated respectively.

In step S303, the simulation result is inputted into the CD-SEM to establish the measurement standard, and data is collected according to the measurement standard to complete the data simulation and test.

In this step, a CD-SEM measurement standard is established by adopting a simulation result, so that data acquisition is carried out, and the simulation and the test of the data are completed. After step S303, returning to step S301, implementing optimization iteration of the model, and re-executing steps S302 and S303 to complete re-collection of data after iteration.

In step S304, data collection is completed.

Through the steps, the data collection process is completed, and then model optimization is realized through the collected data.

In step S305, a difference analysis is performed using the collected data, and model optimization is realized.

In this step, difference analysis is performed by using the collected characteristic dimension of the actual wafer and the characteristic dimension of the test pattern, so as to realize correction and optimization of the model. Specifically, the data collection method is used to sample the feature sizes of the original pattern and the actual wafer pattern lithographically printed on the semiconductor substrate, the data are input into an OPC model, the model is trained, the parameters of the model are adjusted to minimize the error between the feature size of the simulated pattern and the feature size of the actual pattern, and the model of the parameters is used to correct the mask pattern of the actual pattern. After the mask plate is corrected by the OPC model, the difference between the actual wafer pattern and the ideal wafer pattern is very small, and the wafer etching result is accurate.

FIG. 7 shows a schematic block diagram of an OPC model data collection system in accordance with an embodiment of the present invention.

As shown in fig. 7, the OPC model data collection system 700 includes: an initial model building module 701, a first simulation module 702, a data sifting module 703, a first standard building module 704, a first collecting module 705, an iteration module 706, a second simulation module 707, a second standard building module 708, a second collecting module 709, and a calibration module 710.

The initial model establishing module 701 establishes an initial OPC model according to basic data and historical collection data of the test pattern; the first simulation module 702 simulates a Manhattan layout of a test pattern of a one-dimensional structure by adopting an initial OPC model to obtain a first simulation result; the data screening module 703 screens out data of which the simulation data size is smaller than a threshold value in the first simulation result; the first standard establishing module 704 inputs the first simulation result into the metrology tool to establish a first characteristic dimension metrology standard; the first collection module 705 collects the feature size of the actual wafer pattern of the test pattern by using the first feature size measurement standard to obtain first collection data; the iteration module 706 inputs the first collected data into the initial OPC model for model optimization to obtain an iteration OPC model; the second simulation module 707 simulates the manhattan layout of the test pattern of the two-dimensional structure by adopting an iterative OPC model to obtain a second simulation result; the second standard establishing module 708 inputs the second simulation result into the metrology tool to establish a second characteristic dimension metrology standard; the second collection module 709 collects the feature size of the actual wafer pattern of the test pattern of the two-dimensional structure by using a second feature size measurement standard to obtain second collected data; the calibration module 710 inputs the collected data into an iterative OPC model to calibrate the iterative OPC model.

In summary, according to the OPC model data collection method, the data collection system, and the OPC model optimization method provided by the present invention, the initial OPC model is used to simulate the manhattan layout of the test pattern, and then the characteristic dimension measurement standard is established using the obtained simulation result, and the actual dimension of the wafer is measured using the established characteristic dimension measurement standard, so as to achieve data collection. The OPC model data collection method is adopted to collect data, so that the OPC model is optimized, the collected result is closer to the size of an actual wafer, the real difference between a test pattern and the actual wafer can be effectively calculated, and the optimization of the OPC model is better completed.

Furthermore, firstly, a Manhattan layout of a test pattern with a one-dimensional structure is adopted for simulation, a characteristic dimension measurement standard is established according to a simulation result, data is collected according to the characteristic dimension measurement standard, then the collected result is returned to an OPC model for optimization iteration to obtain an iteration OPC model, then the Manhattan layout of the test pattern with the two-dimensional structure is simulated by using the iteration OPC model, the characteristic dimension measurement standard is established according to the simulation result, the data is collected according to the characteristic dimension measurement standard, the gradual collection from the one-dimensional data to the two-dimensional data is realized, the model is more accurate during the two-dimensional data collection through the collection and screening of the one-dimensional data, the difficulty of directly collecting the two-dimensional data is reduced, the collection precision of the two-dimensional data is improved, and the data collection result is more accurate. Meanwhile, in the process of one-dimensional data collection, graphic data with too small simulation result size are screened out, the influence of the data on the collection result is avoided, the difficulty of subsequent data removal is reduced, and the accuracy of two-dimensional data collection can be ensured.

While embodiments in accordance with the invention have been described above, these embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments described. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention and various embodiments with various modifications as are suited to the particular use contemplated. The invention is limited only by the claims and their full scope and equivalents.

Claims (10)

1. An OPC model data collection method comprising:

simulating a Manhattan layout of the test pattern by adopting an initial OPC model to obtain a simulation result;

inputting the simulation result into a measuring tool, and establishing a characteristic dimension measuring standard; and

and acquiring the characteristic dimension of the actual wafer pattern of the test pattern by adopting the characteristic dimension measurement standard to obtain collected data.

2. The OPC model data collection method of claim 1, wherein after the step of collecting data by collecting the feature sizes of the actual wafer patterns of the test patterns using the feature size measurement criteria, further comprising:

inputting the collected data into the initial OPC model for model optimization to obtain an iterative OPC model;

and adopting the iterative OPC model to collect data again.

3. The OPC model data collection method of claim 1, wherein before the step of simulating the manhattan layout of the test pattern using the initial OPC model to obtain the simulation result, further comprising:

and establishing an initial OPC model according to the basic data and the historical collection data of the test pattern.

4. The OPC model data collection method of claim 2, wherein after the step of re-performing data collection using the iterative OPC model, further comprising:

and inputting the collected data into the iterative OPC model, and calibrating the iterative OPC model.

5. The OPC model data collection method of claim 2, wherein the step of simulating the manhattan layout of the test pattern using the initial OPC model to obtain the simulation result comprises:

an initial OPC model is adopted to simulate the Manhattan layout of the test pattern with a one-dimensional structure to obtain a first simulation result,

the characteristic dimension measurement standard established by the first simulation result is adopted as a first characteristic dimension measurement standard, and the collected data acquired by the first characteristic dimension measurement standard is adopted as first collected data.

6. The OPC model data collection method of claim 5, wherein re-performing data collection using the iterative OPC model comprises:

simulating the Manhattan layout of the test pattern of the two-dimensional structure by adopting the iterative OPC model to obtain a second simulation result;

inputting the second simulation result into a measuring tool, and establishing a second characteristic dimension measuring standard;

and acquiring the characteristic dimension of the actual wafer pattern of the test pattern of the two-dimensional structure by adopting the second characteristic dimension measurement standard to obtain second collected data.

7. The OPC model data collection method of claim 5, wherein after the step of simulating the manhattan layout of the test pattern of the one-dimensional structure using the initial OPC model to obtain the first simulation result, further comprising:

and screening out the data with the simulation data size smaller than a threshold value in the first simulation result.

8. An OPC model optimization method, comprising:

carrying out data acquisition by adopting the OPC model data collection method of claims 1-7;

and optimizing the OPC model according to the difference between the collected characteristic dimension of the actual wafer pattern and the characteristic dimension of the test pattern.

9. An OPC model data collection system comprising:

the first simulation module is used for simulating a Manhattan layout of a test pattern of a one-dimensional structure by adopting an initial OPC model to obtain a first simulation result;

the first standard establishing module is used for inputting the first simulation result into a measuring tool and establishing a first characteristic dimension measuring standard; and

and the first collection module is used for collecting the characteristic dimension of the actual wafer pattern of the test pattern by adopting the first characteristic dimension measurement standard to obtain first collection data.

10. The OPC model data collection system of claim 9, further comprising:

the initial model building module is used for building an initial OPC model according to basic data and historical collection data of the test pattern;

the data screening module is used for screening out the data of which the simulation data size is smaller than a threshold value in the first simulation result;

the iteration module is used for inputting the first collected data into the initial OPC model for model optimization to obtain an iteration OPC model;

the second simulation module is used for simulating the Manhattan layout of the test pattern of the two-dimensional structure by adopting the iterative OPC model to obtain a second simulation result;

the second standard establishing module is used for inputting the second simulation result into a measuring tool and establishing a second characteristic dimension measuring standard;

the second collection module is used for collecting the characteristic dimension of the actual wafer pattern of the test pattern of the two-dimensional structure by adopting the second characteristic dimension measurement standard to obtain second collection data;

and the calibration module inputs the collected data into the iterative OPC model and calibrates the iterative OPC model.

Images