Prediction method and device for overlay mark evaluation index
Through machine learning, the prediction model of the engraving mark evaluation index is constructed, which solves the problem of low prediction accuracy of the grating diffraction model, and realizes high-precision engraving mark evaluation index prediction, which improves the accuracy of design quality evaluation and optimization.
Patent Information
- Application Number
- CN202410968182.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-18
- Publication Date
- 2024-11-12
AI Technical Summary
The method of predicting the evaluation index of the engraving mark by constructing grating diffraction models in the prior art has low accuracy, which is affected by unmodeled factors and complex physical quantities, resulting in a decrease in the prediction accuracy of the key evaluation index.
The machine learning method is used to construct a prediction model for the evaluation index of the engraving mark. By constructing the training data set, including the parameter information and evaluation index of the engraving mark, the prediction model is trained to replace the complex optical diffraction model, and high-precision prediction is used for neural networks and other models.
High-precision cross-mark evaluation index prediction is achieved, and the output index is closer to the actual physical model, improving the accuracy of design quality evaluation and optimization.
Smart Images

Figure CN118940620A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of semiconductor manufacturing technology, and in particular to a method, device, storage medium and electronic device for predicting an overlay mark evaluation index. Background Art
[0002] Overlay marks used in the field of semiconductor manufacturing technology can be used to measure overlay errors. The overlay mark prediction model is used to calculate various key evaluation indicators to quantify and evaluate the design quality of the overlay mark, thereby optimizing and improving the design in the next step.
[0003] The conventional method for establishing the overlay mark prediction model is based on the multi-layer periodic structure grating diffraction model of numerical calculation, including: Rigorous Coupled Wave Analysis (RCWA), Slice Absorption Method (SAM), Finite Difference Time Domain (FDTD), Finite Difference Frequency Domain (FDFD), etc. The establishment process of these analytical models will be affected by unmodeled factors, complex physical quantities or physical processes that are difficult to model, etc., resulting in a decrease in model accuracy and restricting the prediction accuracy of key evaluation indicators. Summary of the invention
[0004] The purpose of the embodiments of the present disclosure is to provide a method, device, storage medium and electronic device for predicting an overlay mark evaluation index, so as to solve the problem of low prediction accuracy of the method of predicting the overlay mark evaluation index by constructing physical numerical values such as a grating diffraction model in the prior art.
[0005] The embodiments of the present disclosure adopt the following technical scheme: a method for predicting an overlay mark evaluation index, comprising: constructing a training data set, wherein the training data set at least includes: parameter information of the overlay mark and the evaluation index of the overlay mark; using the parameter information in the training data set as input and the evaluation index in the training data set as output, training an evaluation index prediction model to obtain a trained evaluation index prediction model; and inputting the parameter information of the overlay mark to be evaluated into the evaluation index prediction model to obtain the evaluation index of the overlay mark to be evaluated.
[0006] The disclosed embodiment also provides a prediction device for overlay mark evaluation indicators, comprising: a data set construction module, used to construct a training data set, wherein the training data set at least includes: parameter information of the overlay mark and the evaluation indicator of the overlay mark; a training module, used to use the parameter information in the training data set as input and the evaluation indicator in the training data set as output to train an evaluation indicator prediction model to obtain a trained evaluation indicator prediction model; and a prediction module, used to input the parameter information of the overlay mark to be evaluated into the evaluation indicator prediction model to obtain the evaluation indicator of the overlay mark to be evaluated.
[0007] The embodiment of the present disclosure further provides a storage medium storing a computer program, which implements the steps of the above-mentioned method for predicting the overlay mark evaluation index when executed by a processor.
[0008] An embodiment of the present disclosure also provides an electronic device, comprising at least a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the above-mentioned method for predicting the overlay mark evaluation index when executing the computer program on the memory.
[0009] The beneficial effects of the embodiments of the present disclosure are as follows: a machine learning method is adopted, parameter information of the overlay mark is used as the input parameter of the prediction model, and the evaluation index of the overlay mark is used as the output parameter of the prediction model. The complex optical diffraction model is replaced by training the evaluation index prediction model of the overlay mark, so that the trained prediction model can approximate the actual physical model of the overlay mark with high precision, so as to output the overlay mark evaluation index with higher precision. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] In order to more clearly illustrate one or more embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0011] Figure 1 It is a flow chart of the method for predicting the overlay mark evaluation index in the first embodiment of the present disclosure;
[0012] Figure 2 A schematic diagram of a design pattern of an overlay mark in the first embodiment of the present disclosure;
[0013] Figure 3 for Figure 2 A schematic cross-sectional view of the position of the middle dotted line AA';
[0014] Figure 4 A schematic diagram of the disturbance of the overlay marking process in the first embodiment of the present disclosure;
[0015] Figure 5 The shape of the aperture of the overlay measurement device in the first embodiment of the present disclosure is schematically shown;
[0016] Figure 6 It is a schematic diagram of the prediction result of the diffraction efficiency of the line-on-line structure in the X direction in the first embodiment of the present disclosure;
[0017] Figure 7 It is a schematic diagram of the prediction result of the diffraction efficiency of the line-on-line structure in the Y direction in the first embodiment of the present disclosure;
[0018] Figure 8 It is a schematic diagram of the stacking sensitivity prediction result of the line-on-line structure in the X direction in the first embodiment of the present disclosure;
[0019] Fig. 9 It is a schematic diagram of the stacking sensitivity prediction result of the line-on-line structure in the Y direction in the first embodiment of the present disclosure;
[0020] Fig.10 It is a schematic diagram of the prediction result of the diffraction efficiency of the line-on-trench structure diagram in the X direction in the first embodiment of the present disclosure;
[0021] Fig.11 It is a schematic diagram of the prediction result of the diffraction efficiency of the line-on-trench structure in the Y direction in the first embodiment of the present disclosure;
[0022] Fig.12 It is a schematic diagram of the stacking sensitivity prediction result of the line-on-trench structure in the X direction in the first embodiment of the present disclosure;
[0023] Fig.13 It is a schematic diagram of the prediction result of stacking sensitivity of the line-on-trench structure in the Y direction in the first embodiment of the present disclosure;
[0024] Fig.14 It is a schematic diagram of the structure of the prediction device of the overlay mark evaluation index in the second embodiment of the present disclosure. DETAILED DESCRIPTION
[0025] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the following will be combined with the drawings in one or more embodiments of this specification to clearly and completely describe the technical solutions in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.
[0026] Overlay marks used in the field of semiconductor manufacturing technology can be used for the measurement of overlay errors. The overlay mark prediction model is used to calculate various key evaluation indicators to quantify and evaluate the design quality of the overlay mark, so as to carry out the next step of design optimization and improvement. Conventional calculation methods for key evaluation indicators of overlay marks based on numerical calculations rely on accurate analytical models of overlay marks. Due to the existence of a large number of unmodeled factors, the accuracy of the prediction results based on the analytical model is restricted. Conventional prediction methods based on analytical models require modeling through approximate methods for complex physical quantities and physical processes in known overlay mark graphic structures, process parameters, and measurement equipment parameters, resulting in reduced accuracy of analytical models.
[0027] In order to solve the problems existing in the prior art, the first embodiment of the present disclosure provides a method for predicting the evaluation index of the overlay mark. By constructing a prediction model based on machine learning, it is possible in principle to approximate the actual physical model of the overlay mark with high precision, and through the training optimization of the data set, it is possible to achieve high-precision output of the evaluation index results of the overlay mark. Figure 1 The flowchart of the method for predicting the overlay mark evaluation index in this embodiment is shown, which mainly includes steps S10 to S30:
[0028] S10, construct training data set;
[0029] S20, taking the parameter information in the training data set as input and the evaluation index in the training data set as output, training the evaluation index prediction model to obtain a trained evaluation index prediction model;
[0030] S30, inputting parameter information of the overlay mark to be evaluated into an evaluation index prediction model to obtain an evaluation index of the overlay mark to be evaluated.
[0031] The evaluation index predicted in this embodiment is a series of indicators used to evaluate the design quality of the overlay mark itself after the design is completed, which at least includes any one or more of the following indicators: diffraction efficiency, stacking sensitivity and overlay error. The design of the overlay mark and the evaluation of the process presentation effect are carried out through the specific values or changing trends of the above evaluation indicators. In the process of design and implementation of an overlay mark, its shape design, process implementation and measurement equipment will affect the presentation quality of the overlay mark. Therefore, this embodiment constructs a training data set with the above-mentioned overlay mark related parameters as the training basis of the evaluation index prediction model. Specifically, the training data set constructed by this embodiment includes at least the parameter information of the overlay mark and the evaluation index of the overlay mark, wherein the parameter information of the overlay mark is mainly used as the input data when the evaluation index prediction model is trained, and the evaluation index of the overlay mark is mainly used as the output data when the evaluation index prediction model is trained.
[0032] In the process of constructing the training data set, the parameter information of the overlay mark can be specifically formed by at least one of the following data information: pattern information of the overlay mark, process parameter information of the overlay mark, and parameter information of the overlay measurement equipment; and the evaluation index of the overlay mark is based on the evaluation requirements of the actual overlay mark, with one or more of diffraction efficiency, stacking sensitivity and overlay error as the evaluation index. All data in the above training data set can be obtained by conventional statistical or acquisition methods, and each (or each group) of the parameter information of the overlay mark should have the corresponding overlay mark evaluation index, so as to realize the training of the prediction model.
[0033] Specifically, the pattern information of the overlay mark is mainly used to describe the structure and parameters that the overlay mark is expected to present when designing, which at least includes: the geometric structure of the overlay mark and the geometric parameters of the geometric structure. For the conventional overlay mark design, it has two groups of marks located at different levels in the vertical direction, namely the reference layer mark located at the lower layer and the current layer mark located at the upper layer. The above two layers of marks have the same geometric pattern, but can have different subdivisions. Usually, the geometric structure of the overlay mark is formed by at least one grating structure. In this embodiment, the form, number and positional relationship between multiple grating structures of the grating structure are used as the geometric structure data of the overlay mark; in addition, the geometric structure data should also include the positional relationship between the two layers of marks, such as two use cases formed according to the different corresponding relationship between the grating line positions between the reference layer mark and the current layer mark, at least including: a structure in which the grating lines of the reference layer mark are aligned with the grating lines of the current layer mark (line-on-line), and a structure in which the grating lines of the reference layer are aligned with the grating trenches of the current layer (line-on-trench). Figure 2 A schematic diagram of a design pattern of an overlay mark is shown, wherein Figure 2The left side shows the pattern marked by the current layer. Figure 2 The right part is the pattern marked by the reference layer. Figure 2 Each black line in the figure represents a grating line in the grating structure, and the blank space between the grating lines corresponds to the grating groove. Figure 3 It shows Figure 2 The cross-sectional diagram at the location of the dotted line AA' is used to indicate the vertical positional relationship between the reference layer mark and the current layer mark. Figure 2 The overlay marks shown, no matter they are reference layer marks or current layer marks, each layer mark includes at least one grating structure independently arranged in the X direction and the Y direction in the plane to which it belongs, so as to jointly constitute the overlay marks.
[0034] The geometric parameters of the geometric structure used as the training data set in this embodiment mainly correspond to the geometric parameters of the grating structure. For each grating structure, the corresponding geometric parameters include at least one of the following: structure type, mark size, period, duty cycle, etc., and may also include other parameters used to describe the grating structure, such as: the subdivision of each grating line in the grating structure, which at least includes: horizontal subdivision, vertical subdivision, and two-dimensional subdivision in both horizontal and vertical directions. The above subdivisions can be implemented independently in the reference layer mark and the current layer mark, or can be implemented simultaneously in both layers of marks.
[0035] The process parameter information of the overlay mark is mainly used to describe the process treatment methods used in the actual preparation process of the overlay mark and the parameters designed in the process of process implementation. In this embodiment, the process parameter information of the overlay mark includes at least: process stacking parameters and process disturbance parameters. Different process treatments are required during the silicon wafer processing process. The typical process in a photolithography exposure cycle includes: deposition, gluing, exposure, development, etching, etc. The overlay mark is in different process stages and will go through multiple photolithography exposure cycles. Each photolithography exposure cycle corresponds to a mark layer. The type, quantity and parameters of the process layers contained in the mark layer can be different.
[0036] Based on the above description, the process stacking parameters are used to describe the various parameters of the above-mentioned different process stages, mainly including at least one of the following parameters: process type, material type, process layer thickness, material optical refractive index (n value), material optical extinction coefficient (k value). Usually, the n value and k value of the material are used at the same time; wherein, the process type corresponds to different process stages including at least one of the following: deposition, coating, exposure, development, and etching; parameters such as material type, material optical refractive index, and material optical extinction coefficient are determined according to the material preparation requirements of different levels; and the process layer thickness is designed according to the process requirements during the preparation of each level.
[0037] The process disturbance parameters are disturbance items that occur in the actual process preparation, mainly including at least one of the following: stack thickness, material optical refractive index, material optical extinction coefficient, side wall angle, top tilt, and bottom tilt. Figure 4 A schematic diagram of the disturbance of the overlay mark process is shown, where: Figure 4 The left side shows the ideal photolithography mark shape. Figure 4 The pattern in the middle is the side wall angle disturbance d1 produced after actual preparation. Figure 4 The pattern on the right is the top tilt disturbance d2 produced after actual preparation. Figure 3 The reference layer mark and the current layer mark shown have known, fixed offsets shift in the X and Y directions within the planes to which they belong, which are used to calculate the overlay error.
[0038] The overlay measurement equipment is a device that measures the overlay mark and calculates the key evaluation index results of the mark. It generates an incident light beam, receives a diffracted light beam formed by the overlay mark, and then calculates various key evaluation indicators based on the amplitude, frequency and phase information of the diffracted light beam. In this embodiment, the parameter information of the overlay measurement equipment includes at least: incident light wavelength, incident light bandwidth, incident light incident angle, incident light polarization state, and aperture shape; wherein, for any wavelength, a wavelength bandwidth centered on the wavelength can be set at the same time, and the overlay measurement equipment measures the overlay mark with incident light in this wavelength range; the incident light incident angle is the angle between the incident light and the surface normal of the overlay mark; the incident light polarization state can be mainly set to 0 degrees, 90 degrees, or both at the same time; the aperture shape can be set according to the measurement requirements, Figure 5 The schematic diagrams of the shapes of apertures of several overlay measurement equipment are shown. In the actual measurement process, any shape of aperture can be used alone on the measurement equipment, or several aperture shapes can be superimposed and used.
[0039] It should be noted that all data in the above training data set can be used as training information for the prediction model. At the same time, the training data used by the prediction model in the actual training process may include but is not limited to the above data.
[0040] Different from the conventional optical diffraction model, this embodiment adopts an evaluation index prediction model based on machine learning to predict the evaluation index. Specifically, the prediction model of this embodiment is a regression model established by a machine learning method, and includes at least one of the following models: a neural network model, a decision tree model, a random forest model, a support vector regression model, a Gaussian process regression model, and a k-nearest neighbor algorithm model. In actual implementation, any of the above models can be selected as the evaluation index prediction model. In some embodiments, the neural network model includes any of the following: a multilayer perceptron model, a residual neural network model, a convolutional neural network model, and a deep neural network model.
[0041] After constructing the training data set, the above-mentioned evaluation index prediction model can be trained, that is, the parameter information of the overlay mark in the training data set is used as input, and the evaluation index of the overlay mark in the training data set is used as output to train the prediction model, so that the prediction model optimizes the parameters in the model based on the difference between the actual output and the evaluation index in the training data set, until the number of iterations or the loss function value is minimized, the prediction model training is considered complete. After the model training is completed, when it is necessary to evaluate the overlay mark that has not been evaluated, the parameter information of the overlay mark to be evaluated can be input into the evaluation index prediction model, and the corresponding evaluation index of the overlay mark can be obtained, and the predicted evaluation index can be used as the basis for determining the quality of the current overlay mark. In the actual prediction process, the prediction model can simultaneously output the evaluation index under ideal conditions and the evaluation index under disturbance, so that the user can optimize the design, process or measurement stage of the overlay mark based on the comparison of the above two results.
[0042] In some embodiments, the prediction model can also gradually improve the prediction accuracy of the prediction model by expanding the capacity of the training data set, or the prediction model can be optimized in the actual prediction process based on the prediction results and the actual measurement results of the overlay marks to achieve a more accurate prediction result output.
[0043] In order to better illustrate the prediction model function of this embodiment, Figure 2 Taking the overlay mark shown in FIG. 1 as an example, a deep neural network prediction model is established. The pattern information of the overlay mark during training includes: the usage of the overlay mark (corresponding to Figure 2 The overlay mark is a line-on-line structure), the geometric structure and geometric parameters of the overlay mark; the process parameter information of the overlay mark includes: the material and thickness of each process layer, as well as the process disturbance of the mark side wall angle and the process disturbance of the process layer thickness; the overlay measurement equipment parameters include: the wavelength of the incident light, the bandwidth of the incident light and the polarization state of the incident light; the evaluation indicators of the overlay mark include the diffraction efficiency and stacking sensitivity of the overlay mark in the X and Y directions in the plane where the silicon wafer surface is located.
[0044] Figures 6 to 9 An example of the prediction results of the deep neural network prediction model is shown, where Figure 6 The figure shows the predicted results of diffraction efficiency in the X direction. Figure 7 The figure shows the predicted results of diffraction efficiency in the Y direction. Figure 8 The figure shows the prediction results of stacking sensitivity in the X direction. Fig. 9The figure shows the prediction results of stacking sensitivity in the Y direction. Based on the above prediction results, it can be seen that when the process disturbance and measurement equipment parameters are different, there are certain differences in the values of the evaluation indicators. This difference represents the impact of the above parameter changes on the evaluation indicators. In addition, Figures 10 to 13 The prediction results corresponding to the overlay marks of the line-on-trench structure are shown, further verifying the above conclusions.
[0045] This embodiment adopts a machine learning approach, with the parameter information of the overlay mark as the input parameter of the prediction model and the evaluation index of the overlay mark as the output parameter of the prediction model. The complex optical diffraction model is replaced by training the evaluation index prediction model of the overlay mark, so that the trained prediction model can approximate the actual physical model of the overlay mark with high precision, so as to output the overlay mark evaluation index with higher precision.
[0046] Based on the same inventive concept, the second embodiment of the present disclosure provides a prediction device for an overlay mark evaluation index, and its structural schematic diagram is shown in FIG. Fig.14 As shown, it mainly includes: a data set construction module 10, which is used to construct a training data set, and the training data set at least includes: parameter information of the overlay mark and evaluation indicators of the overlay mark; a training module 20, which is used to use the parameter information in the training data set as input and the evaluation indicators in the training data set as output to train the evaluation indicator prediction model to obtain a trained evaluation indicator prediction model; and a prediction module 30, which is used to input the parameter information of the overlay mark to be evaluated into the evaluation indicator prediction model to obtain the evaluation indicator of the overlay mark to be evaluated.
[0047] Specifically, the parameter information of the overlay mark includes at least one of the following: pattern information of the overlay mark, process parameter information of the overlay mark, and parameter information of the overlay measurement equipment. The evaluation index of the overlay mark includes at least one of the following: diffraction efficiency, stacking sensitivity, and overlay error.
[0048] Specifically, the pattern information of the overlay mark includes at least: the geometric structure of the overlay mark and the geometric parameters of the geometric structure; wherein the geometric parameters include at least one of the following: structure type, mark size, period, duty cycle. The process parameter information of the overlay mark includes at least: process stack parameters and process disturbance parameters; wherein the process stack parameters include at least one of the following: process type, material type, process layer thickness, material optical refractive index, material optical extinction coefficient; process type includes at least one of the following: deposition, coating, exposure, development, etching; process disturbance parameters include at least one of the following: stack thickness, material optical refractive index, material optical extinction coefficient, side wall angle, top tilt, bottom tilt. The overlay measurement equipment parameter information includes at least: incident light wavelength, incident light bandwidth, incident light incident angle, incident light polarization state, and aperture shape.
[0049] Specifically, the evaluation index prediction model includes any one of the following: neural network model, decision tree model, random forest model, support vector regression model, Gaussian process regression model, k-nearest neighbor algorithm model; the neural network model includes any one of the following: multilayer perceptron model, residual neural network model, convolutional neural network model, deep neural network model.
[0050] This embodiment adopts a machine learning approach, with the parameter information of the overlay mark as the input parameter of the prediction model and the evaluation index of the overlay mark as the output parameter of the prediction model. The complex optical diffraction model is replaced by training the evaluation index prediction model of the overlay mark, so that the trained prediction model can approximate the actual physical model of the overlay mark with high precision, so as to output the overlay mark evaluation index with higher precision.
[0051] Based on the same inventive concept, a third embodiment of the present disclosure provides a storage medium, which can be installed in any semiconductor manufacturing and measurement equipment, and is specifically a computer-readable medium, storing a computer program, and when the computer program is executed by a processor, the method provided by any embodiment of the present disclosure is implemented, including the following steps S31 to S33:
[0052] S31, constructing a training data set, wherein the training data set at least includes: parameter information of the overlay mark and evaluation index of the overlay mark;
[0053] S32, using the parameter information in the training data set as input and the evaluation index in the training data set as output, training the evaluation index prediction model to obtain a trained evaluation index prediction model; and
[0054] S33, inputting parameter information of the overlay mark to be evaluated into an evaluation index prediction model to obtain an evaluation index of the overlay mark to be evaluated.
[0055] Specifically, the parameter information of the overlay mark includes at least one of the following: pattern information of the overlay mark, process parameter information of the overlay mark, and parameter information of the overlay measurement equipment. The evaluation index of the overlay mark includes at least one of the following: diffraction efficiency, stacking sensitivity, and overlay error.
[0056] Specifically, the pattern information of the overlay mark includes at least: the geometric structure of the overlay mark and the geometric parameters of the geometric structure; wherein the geometric parameters include at least one of the following: structure type, mark size, period, duty cycle. The process parameter information of the overlay mark includes at least: process stack parameters and process disturbance parameters; wherein the process stack parameters include at least one of the following: process type, material type, process layer thickness, material optical refractive index, material optical extinction coefficient; process type includes at least one of the following: deposition, coating, exposure, development, etching; process disturbance parameters include at least one of the following: stack thickness, material optical refractive index, material optical extinction coefficient, side wall angle, top tilt, bottom tilt. The overlay measurement equipment parameter information includes at least: incident light wavelength, incident light bandwidth, incident light incident angle, incident light polarization state, and aperture shape.
[0057] Specifically, the evaluation index prediction model includes any one of the following: neural network model, decision tree model, random forest model, support vector regression model, Gaussian process regression model, k-nearest neighbor algorithm model; the neural network model includes any one of the following: multilayer perceptron model, residual neural network model, convolutional neural network model, deep neural network model.
[0058] This embodiment adopts a machine learning approach, with the parameter information of the overlay mark as the input parameter of the prediction model and the evaluation index of the overlay mark as the output parameter of the prediction model. The complex optical diffraction model is replaced by training the evaluation index prediction model of the overlay mark, so that the trained prediction model can approximate the actual physical model of the overlay mark with high precision, so as to output the overlay mark evaluation index with higher precision.
[0059] The fourth embodiment of the present disclosure provides an electronic device, which can be any semiconductor manufacturing and measurement equipment, and at least includes a memory and a processor, and a computer program is stored in the memory. When the processor executes the computer program in the memory, the method provided in any embodiment of the present disclosure is implemented. Exemplarily, the steps of the electronic device computer program are as follows S41 to S43:
[0060] S41, constructing a training data set, wherein the training data set at least includes: parameter information of the overlay mark and evaluation index of the overlay mark;
[0061] S42, using the parameter information in the training data set as input and the evaluation index in the training data set as output, training the evaluation index prediction model to obtain a trained evaluation index prediction model; and
[0062] S43, inputting parameter information of the overlay mark to be evaluated into the evaluation index prediction model to obtain the evaluation index of the overlay mark to be evaluated. Specifically, the parameter information of the overlay mark includes at least one of the following: pattern information of the overlay mark, process parameter information of the overlay mark, and parameter information of the overlay measurement equipment. The evaluation index of the overlay mark includes at least one of the following: diffraction efficiency, stacking sensitivity, and overlay error.
[0063] Specifically, the pattern information of the overlay mark includes at least: the geometric structure of the overlay mark and the geometric parameters of the geometric structure; wherein the geometric parameters include at least one of the following: structure type, mark size, period, duty cycle. The process parameter information of the overlay mark includes at least: process stack parameters and process disturbance parameters; wherein the process stack parameters include at least one of the following: process type, material type, process layer thickness, material optical refractive index, material optical extinction coefficient; process type includes at least one of the following: deposition, coating, exposure, development, etching; process disturbance parameters include at least one of the following: stack thickness, material optical refractive index, material optical extinction coefficient, side wall angle, top tilt, bottom tilt. The overlay measurement equipment parameter information includes at least: incident light wavelength, incident light bandwidth, incident light incident angle, incident light polarization state, and aperture shape.
[0064] Specifically, the evaluation index prediction model includes any one of the following: neural network model, decision tree model, random forest model, support vector regression model, Gaussian process regression model, k-nearest neighbor algorithm model; the neural network model includes any one of the following: multilayer perceptron model, residual neural network model, convolutional neural network model, deep neural network model.
[0065] This embodiment adopts a machine learning approach, with the parameter information of the overlay mark as the input parameter of the prediction model and the evaluation index of the overlay mark as the output parameter of the prediction model. The complex optical diffraction model is replaced by training the evaluation index prediction model of the overlay mark, so that the trained prediction model can approximate the actual physical model of the overlay mark with high precision, so as to output the overlay mark evaluation index with higher precision.
[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present disclosure.
Claims
1. A prediction method for overlay mark evaluation index, It is characterized in that include: Constructing a training data set, wherein the training data set at least includes: parameter information of the overlay mark and evaluation index of the overlay mark; Taking the parameter information in the training data set as input and the evaluation index in the training data set as output, training the evaluation index prediction model to obtain a trained evaluation index prediction model; and The parameter information of the overlay mark to be evaluated is input into the evaluation index prediction model to obtain the evaluation index of the overlay mark to be evaluated.
2. The method for predicting the overlay mark evaluation index according to claim 1, It is characterized in that The parameter information of the overlay mark includes at least one of the following: pattern information of the overlay mark, process parameter information of the overlay mark, and parameter information of an overlay measurement device.
3. The method for predicting the overlay mark evaluation index according to claim 2, It is characterized in that The pattern information of the overlay mark includes at least: the geometric structure of the overlay mark and the geometric parameters of the geometric structure; wherein, The geometric parameters include at least one of the following: structure type, mark size, period, and duty cycle.
4. The method for predicting the overlay mark evaluation index according to claim 2, It is characterized in that The process parameter information of the overlay mark includes at least: process stacking parameters and process disturbance parameters; wherein, The process stacking parameters include at least one of the following: process type, material type, process layer thickness, material optical refractive index, material optical extinction coefficient; The process type includes at least one of the following: deposition, coating, exposure, development, and etching; The process disturbance parameter includes at least one of the following: stack thickness, material optical refractive index, material optical extinction coefficient, side wall angle, top tilt, and bottom tilt.
5. The method for predicting the overlay mark evaluation index according to claim 2, It is characterized in that The overlay measurement equipment parameter information at least includes: incident light wavelength, incident light bandwidth, incident light incident angle, incident light polarization state, and aperture shape.
6. The method for predicting an overlay mark evaluation index according to claim 1, It is characterized in that The evaluation index of the overlay mark includes at least one of the following: diffraction efficiency, stacking sensitivity and overlay error.
7. The method for predicting an overlay mark evaluation index according to any one of claims 1 to 6, It is characterized in that The evaluation index prediction model includes any one of the following: neural network model, decision tree model, random forest model, support vector regression model, Gaussian process regression model, k-nearest neighbor algorithm model; The neural network model includes any one of the following: a multi-layer perceptron model, a residual neural network model, a convolutional neural network model, and a deep neural network model.
8. A prediction device for overlay mark evaluation index, It is characterized in that include: A data set construction module, used to construct a training data set, wherein the training data set at least includes: parameter information of the overlay mark and evaluation index of the overlay mark; A training module, used to train the evaluation index prediction model using the parameter information in the training data set as input and the evaluation index in the training data set as output to obtain a trained evaluation index prediction model; and The prediction module is used to input the parameter information of the overlay mark to be evaluated into the evaluation index prediction model to obtain the evaluation index of the overlay mark to be evaluated.
9. A storage medium storing a computer program, It is characterized in that When the computer program is executed by a processor, the steps of the method for predicting an overlay mark evaluation index according to any one of claims 1 to 7 are implemented.
10. An electronic device comprising at least a memory and a processor, wherein the memory stores a computer program. It is characterized in that The processor implements the steps of the method for predicting an overlay mark evaluation index according to any one of claims 1 to 7 when executing the computer program on the memory.
