Measurement method and system of overlay error, equipment and storage medium

Abstract

An overlay error measurement method, system and device, and a storage medium, the method comprising: providing a front layer structure and a current layer structure on the front layer structure, the front layer structure comprising a plurality of second alignment patterns, the current layer structure comprising a plurality of measurement areas corresponding to the second alignment patterns; setting a measurement interval of the current-layer pattern; in each measurement area of the current layer, a plurality of measurement patterns which are arranged in parallel at measurement intervals are obtained, and the measurement patterns comprise a first alignment pattern and other measurement patterns located on the two sides of the first alignment pattern; taking each measurement pattern as a measurement position, and acquiring a first alignment signal of the measurement pattern and a second alignment pattern of a front layer; acquiring the position information of the measurement pattern at the central position in the measurement patterns according to the plurality of first alignment signals; and obtaining an overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information. According to the invention, the accuracy of overlay error measurement is improved, and the measurement cost is low.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of semiconductor manufacturing, and in particular to a method and system, device and storage medium for measuring overlay error. Background Art

[0002] Overlay refers to the offset between the current layer and the previous layer during the photolithography process. With the development of integrated circuit manufacturing technology, the control requirements for overlay error are becoming more and more stringent. Excessive overlay error will cause short circuits and open circuits in devices, affecting product quality. However, overlay error is related to many factors such as the deformation of the mask and wafer, the projection system of the photolithography machine itself, and the displacement inconsistency of the wafer worktable. At the same time, external environmental factors will also affect it, such as temperature, humidity, vibration, etc.

[0003] Since the manufacture of integrated circuit chips is through the stacking of multiple layers, the structure formed by each photolithography process needs to be aligned with the previously formed layers, so the overlay accuracy between the layers directly affects the effectiveness and yield of the integrated circuit chip. In the preparation process of semiconductor chips, the process parameters for preparing integrated circuit chips can be adjusted according to the overlay error value to improve the effectiveness and yield of the integrated circuit chip. The deviation of the overlay alignment between layers will have a great impact on the performance of the device, and may even cause device failure, reduce the yield, waste time and money, and increase costs. Therefore, OVL calculation becomes more and more important. Measuring and calculating the accurate overlay error value is a key step to improve the yield. Summary of the invention

[0004] The problem solved by the embodiments of the present invention is to provide a method and system, device and storage medium for measuring overlay error, so as to improve the accuracy of overlay error measurement and reduce the measurement cost.

[0005] To solve the above problems, an embodiment of the present invention provides an overlay error measurement method, comprising: providing a front layer structure and a current layer structure located on the front layer structure, the front layer structure including a plurality of second alignment patterns, and the current layer structure including a plurality of measurement areas corresponding to the second alignment patterns; setting a measurement interval of the current layer pattern; in each measurement area of ​​the current layer, obtaining a plurality of measurement patterns arranged in parallel with the measurement interval, including a first alignment pattern, and other measurement patterns located on both sides of the first alignment pattern; taking each measurement pattern as a measurement position, obtaining a first alignment signal between the measurement pattern and the second alignment pattern of the front layer; obtaining position information of the measurement pattern at the center position in the measurement pattern according to the plurality of first alignment signals; and obtaining the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information.

[0006] Optionally, each measurement pattern is used as a measurement position to obtain a first alignment signal between the measurement pattern and a second alignment pattern of the front layer, wherein the first alignment signal includes an optical signal.

[0007] Optionally, obtaining position information of a measurement figure at a center position in the measurement figure based on multiple first alignment signals includes: performing dimensionality reduction processing on the multiple first alignment signals to obtain corresponding second alignment signals; and obtaining position information of a measurement figure at a center position in the measurement figure based on multiple second alignment signals.

[0008] Optionally, before performing dimensionality reduction processing on the multiple first alignment signals, it also includes: in each measurement area, simplifying the first alignment signal, subtracting the first alignment signal corresponding to the first alignment figure from the first alignment signal of each measurement figure; and in performing dimensionality reduction processing on the multiple first alignment signals, performing dimensionality reduction processing on the simplified first alignment signal.

[0009] Optionally, a principal component analysis method is used to perform dimensionality reduction processing on the multiple first alignment signals.

[0010] Optionally, obtaining position information of a measurement figure at a center position in the measurement figure based on multiple second alignment signals includes: selecting data corresponding to a dimension that best reflects the nonlinear characteristics of the overlay error among the multiple second alignment signals as a third alignment signal; and obtaining position information of a measurement figure at a center position in the measurement figure based on multiple third alignment signals.

[0011] Optionally, according to the key performance indicator, data corresponding to the dimension that best reflects the nonlinear characteristics of the overlay error among the multiple second alignment signals is selected as the third alignment signal.

[0012] Optionally, the position information of a measurement figure at a center position in the measurement figure is obtained according to multiple third alignment signals, including: taking multiple third alignment signals as output signals of machine learning and taking multiple first alignment signals as input signals of machine learning to obtain an algorithm model of machine learning; taking multiple first alignment signals as input signals and obtaining corresponding multiple fourth alignment signals according to the algorithm model; and obtaining the position information of a measurement figure at a center position in the measurement figure according to the multiple fourth alignment signals.

[0013] Optionally, obtaining position information of a measurement figure at a center position in the measurement figure according to multiple fourth alignment signals includes: performing function fitting on multiple fourth alignment signals of multiple measurement areas to obtain an ideal function; obtaining ideal signal values ​​corresponding to multiple measurement positions in each measurement area according to the ideal function; obtaining an updated machine learning algorithm model according to the multiple ideal signal values; obtaining updated multiple fourth alignment signals according to the algorithm model; judging whether the function morphology of the ideal function meets the feature requirements; if so, obtaining the position information of the measurement figure at the center position in the measurement figure according to the updated multiple fourth alignment signals; otherwise, returning to execute function fitting on multiple fourth alignment signals of multiple measurement areas to obtain the ideal function.

[0014] Optionally, a supervised learning algorithm or a neural network algorithm is used to obtain an algorithm model for machine learning.

[0015] Optionally, multiple fourth alignment signals of multiple measurement areas are subjected to function fitting to obtain an ideal function, including: in each measurement area, the multiple fourth alignment signals are subjected to decentralized processing along the data direction of the measurement interval; after the decentralized processing, the function type formed by the multiple fourth alignment signals is obtained; and the multiple fourth alignment signals of the multiple measurement areas are fitted into an ideal function using a fitting function of the same type as the function type.

[0016] Optionally, a fitting function of the same type as the function type is used to fit a plurality of fourth alignment signals of a plurality of measurement areas into an ideal function, wherein the order of the fitting function used this time is higher than the order of the fitting function used last time.

[0017] Optionally, an updated machine learning algorithm model is obtained based on multiple ideal signal values, including: taking multiple ideal signal values ​​as output signals of machine learning and taking multiple first alignment signals as input signals of machine learning, and updating the machine learning algorithm model; and obtaining multiple updated fourth alignment signals based on the algorithm model, including: taking multiple ideal signals as input signals and obtaining multiple updated fourth alignment signals based on the updated algorithm model.

[0018] Optionally, obtaining position information of a measurement figure at a center position in the measurement figure according to multiple fourth alignment signals includes: in each measurement area, obtaining the center point of the dot diagram according to the dot diagram formed by multiple fourth alignment signals; and obtaining the measurement interval corresponding to the center point.

[0019] Optionally, acquiring the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information includes: acquiring the negative value of the measurement interval corresponding to the center point as the overlay error.

[0020] Correspondingly, an embodiment of the present invention further provides an overlay error measurement system, comprising: a structure providing module, used to provide a front layer structure, and a current layer structure located on the front layer structure, the front layer structure including multiple second alignment patterns, and the current layer structure including multiple measurement areas corresponding to the second alignment patterns; a measurement interval setting module, used to set the measurement interval of the current layer pattern; a measurement pattern acquisition module, used to acquire, in each measurement area of ​​the current layer, multiple measurement patterns arranged in parallel at the measurement interval, including a first alignment pattern, and other measurement patterns located on both sides of the first alignment pattern; a first alignment signal acquisition module, used to acquire a first alignment signal between the measurement pattern and the second alignment pattern of the front layer with each measurement pattern as the measurement position; a position information acquisition module, used to acquire the position information of the measurement pattern at the center position in the measurement pattern according to the multiple first alignment signals; an overlay error acquisition module, used to acquire the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information.

[0021] Correspondingly, an embodiment of the present invention also provides a device, comprising at least one memory and at least one processor, wherein the memory stores one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the overlay error measurement method provided by an embodiment of the present invention.

[0022] Correspondingly, an embodiment of the present invention further provides a storage medium, wherein the storage medium stores one or more computer instructions, and the one or more computer instructions are used to implement the overlay error measurement method provided by the embodiment of the present invention.

[0023] Compared with the prior art, the technical solution of the embodiment of the present invention has the following advantages:

[0024] In the method for measuring overlay error provided by an embodiment of the present invention, a measurement interval of a current layer pattern is set, and in each measurement area of ​​the current layer, a plurality of measurement patterns arranged in parallel with the measurement interval are obtained, including a first alignment pattern and other measurement patterns located on both sides of the first alignment pattern, each measurement pattern is used as a measurement position, a first alignment signal between the measurement pattern and the second alignment pattern of the previous layer is obtained, and the position information of the measurement pattern at the center position in the measurement pattern is obtained according to the plurality of first alignment signals, and the overlay error between the first alignment pattern and the second alignment pattern of the previous layer is obtained according to the position information; by setting the measurement interval, the first alignment signal of each measurement position is detected, the change of the first alignment signal is detected, and the overlay error is obtained according to the mutation of the first alignment signal, that is, the overlay error is obtained by a nonlinear method, which is beneficial to improving the accuracy of overlay error measurement and has a low measurement cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of an embodiment of a method for measuring overlay error of the present invention;

[0026] Figure 2 is a schematic diagram corresponding to an embodiment of a method for measuring overlay error of the present invention;

[0027] Figure 3 is a functional block diagram of an embodiment of a system for measuring overlay error of the present invention;

[0028] Figure 4 It is a hardware structure diagram of an embodiment of the device provided by the present invention. DETAILED DESCRIPTION

[0029] It is known from the prior art that in the semiconductor structure manufacturing process, some process flows may cause the alignment marks on the wafer to be asymmetric, which results in the inability to accurately obtain the absolute overlay error between two adjacent layers of the structure during the optical overlay error measurement process.

[0030] In order to solve the above technical problem, an embodiment of the present invention provides a method for measuring overlay error. Figure 1 , shows a flow chart of an embodiment of the overlay error measurement method of the present invention.

[0031] In this embodiment, the overlay error measurement method includes the following basic steps:

[0032] Step S1: providing a front layer structure and a current layer structure located on the front layer structure, wherein the front layer structure includes a plurality of second alignment patterns, and the current layer structure includes a plurality of measurement areas corresponding to the second alignment patterns;

[0033] Step S2: setting the measurement interval of the current layer of graphics;

[0034] Step S3: in each measurement area of ​​the layer, a plurality of measurement patterns arranged in parallel at a measurement interval are obtained, including a first alignment pattern and other measurement patterns located on both sides of the first alignment pattern;

[0035] Step S4: taking each measurement pattern as a measurement position, obtaining a first alignment signal between the measurement pattern and the second alignment pattern of the front layer;

[0036] Step S5: acquiring position information of a measurement pattern at a center position in the measurement pattern according to a plurality of first alignment signals;

[0037] Step S6: acquiring an overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information.

[0038] In an embodiment of the present invention, a measurement interval is set to detect the first alignment signal at each measurement position, detect changes in the first alignment signal, and obtain the overlay error based on the mutation of the first alignment signal. That is, a nonlinear method is used to obtain the overlay error, which is beneficial to improving the accuracy of overlay error measurement and has a low measurement cost.

[0039] In order to make the above-mentioned purposes, features and advantages of the embodiments of the present invention more obvious and understandable, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0040] Figure 2 It is a corresponding schematic diagram of an embodiment of the overlay error measurement method of the present invention.

[0041] Execute step S1: provide a front layer structure and a current layer structure located on the front layer structure, wherein the front layer structure includes a plurality of second alignment patterns, and the current layer structure includes a plurality of measurement areas corresponding to the second alignment patterns.

[0042] The second alignment pattern of the front layer structure and the measurement area corresponding to the current layer structure and the second alignment pattern are used to align the current layer with the front layer to obtain the overlay error between the current layer and the front layer.

[0043] It should be noted that, in this embodiment, when the measurement area of ​​the current layer structure corresponds to the second alignment pattern of the previous layer structure, it means that the previous layer structure has multiple second alignment patterns at different positions, and each second alignment pattern has a corresponding measurement area in the current layer structure.

[0044] Execute step S2: set the measurement interval of the current layer of graphics.

[0045] In the current layer structure, the measurement interval of the current layer pattern is set to arrange the measurement pattern in the measurement area.

[0046] In this embodiment, in actual applications, the measurement interval is set according to specific needs. If the measurement interval is set smaller, the measurement accuracy of the overlay error can be higher. If the measurement accuracy requirement is not high, the size setting of the measurement interval can be appropriately relaxed.

[0047] Execute step S3: in each measurement area of ​​the layer, obtain a plurality of measurement patterns arranged in parallel at measurement intervals, including a first alignment pattern and other measurement patterns located on both sides of the first alignment pattern.

[0048] The measurement pattern is used to measure with the second alignment pattern of the front layer structure to obtain a first alignment signal, wherein the first alignment pattern in the measurement pattern is used to measure with the second alignment pattern to obtain an overlay error.

[0049] As an example, the second alignment pattern of the front layer structure is a fin structure, and when the measurement pattern of the layer structure is a plurality of gate structures arranged in parallel, the gate structures are arranged in parallel along the extension direction of the fin structure and cross the fin structure.

[0050] In this embodiment, in each measurement area of ​​the layer, multiple measurement patterns are set, and the spacing between adjacent measurement patterns is the measurement interval, that is, each measurement pattern corresponds to different position information. Subsequently, by obtaining the first alignment signal between each measurement pattern and the second alignment pattern of the previous layer, the first alignment signal corresponding to each position information is obtained, thereby obtaining the position information of the measurement pattern at the center position of the measurement pattern.

[0051] As an example, the measurement interval is set to 1nm, and the position information of the first alignment pattern is taken as the origin, marked as OVL. Then in one measurement area, the position information of multiple measurement patterns are arranged in sequence: OVL+n, OVL+(n-1), OVL+(n-2)...OVL+0...OVL-n, OVL-(n+1), OVL-(n+2), and a total of 2n+3 measurement patterns are set.

[0052] It should be noted that, in the present embodiment, the first alignment figure is a graphic structure for measuring the overlay error. In the measurement area, multiple measurement figures are arranged on both sides of the first alignment figure among multiple measurement figures, but the first alignment figure may be located at the center position of the multiple measurement figures, or may not be located at the center position of the multiple measurement figures.

[0053] Execute step S4: take each measurement pattern as a measurement position, and obtain a first alignment signal between the measurement pattern and the second alignment pattern of the front layer.

[0054] The first alignment signal is used to determine the positional relationship between the measurement pattern and the second alignment pattern of the front layer, so as to obtain the position information of the measurement pattern located at the center position in the measurement pattern.

[0055] In this embodiment, each measurement pattern is used as a measurement position to obtain a first alignment signal between the measurement pattern and the second alignment pattern of the front layer, wherein the first alignment signal includes an optical signal.

[0056] The measurement is performed by acquiring optical signals, which is simple and easy to operate, and can be measured in real time when the layer structure is formed, so that the acquisition of overlay error is faster and more convenient.

[0057] In other embodiments, the first alignment signal may also be an electrical signal.

[0058] Execute step S5: obtain position information of the measurement pattern at the center position in the measurement pattern according to the plurality of first alignment signals.

[0059] The position information of the measurement pattern at the center position in the measurement pattern is obtained according to the plurality of first alignment signals, so that the overlay error between the first alignment pattern and the second alignment pattern is subsequently obtained according to the position information of the measurement pattern at the center position.

[0060] Specifically, in the present embodiment, the measurement figure at the center position in the measurement figure is the optimal position relative to the second alignment figure, that is, the measurement figure at the center position in the measurement figure is the reference position required for the process in the semiconductor manufacturing process, and the first alignment figure is the formation position of the first alignment figure in the layer structure. Moreover, according to the foregoing, the position information includes information on the relative position deviation of the measurement figure relative to the first alignment figure. Therefore, the overlay error can be subsequently obtained based on the position information of the measurement figure at the center position in the measurement figure.

[0061] In this embodiment, before the multiple first alignment signals are subsequently subjected to dimensionality reduction processing, the method further includes: in each measurement area, simplifying the first alignment signal, and subtracting the first alignment signal corresponding to the first alignment pattern from the first alignment signal of each measurement pattern.

[0062] Taking the first alignment signal corresponding to the first alignment pattern as a reference, simplifying the first alignment signal of each measurement pattern is beneficial to simplifying the subsequent dimensionality reduction process and improving the efficiency of the dimensionality reduction process.

[0063] In this embodiment, obtaining the position information of the measurement pattern at the center position in the measurement pattern according to the plurality of first alignment signals includes: performing dimensionality reduction processing on the plurality of first alignment signals to obtain corresponding second alignment signals.

[0064] The multiple first alignment signals are processed by dimensionality reduction to obtain the corresponding second alignment signals, and the position information of the measurement figure at the center position is subsequently obtained through the second alignment signals obtained after dimensionality reduction, which is conducive to simplifying the processing, reducing the amount of calculation data, shortening the calculation time, and improving the processing efficiency.

[0065] Correspondingly, in this embodiment, when performing dimensionality reduction processing on a plurality of first alignment signals, performing dimensionality reduction processing on the first alignment signal after the simplified processing is beneficial to improving processing efficiency.

[0066] In this embodiment, a principal component analysis method is used to perform dimensionality reduction processing on the plurality of first alignment signals.

[0067] The principal component analysis (PCA) method is used for dimensionality reduction processing, which can replace the original data with more dimensions with data with fewer dimensions, and at the same time can reflect the original data information with more dimensions, which is beneficial to retain the original information of multiple first alignment signals while achieving the dimensionality reduction effect.

[0068] Specifically, as an example, in this embodiment, there are m first alignment signals, each of which is n-dimensional. In this embodiment, the principal component analysis method is used to reduce the dimension of the multiple first alignment signals, including: subtracting the mean of each dimension from the dimension, and forming an n×m matrix X; calculating the covariance matrix c=1 / m XX T ; Calculate the eigenvalues ​​and eigenvectors of the covariance matrix, and arrange the n eigenvectors from large to small and from top to bottom into a matrix A according to the corresponding eigenvalues; Take the top k eigenvectors of matrix A to form a matrix P; Y = PX, which is the data after dimensionality reduction to K dimensions.

[0069] In this embodiment, the top k eigenvectors of the matrix A are taken according to the principal component contribution rate and the cumulative contribution rate. Generally, the greater the contribution rate of the principal component, the more information representing the retained original data.

[0070] Correspondingly, in this embodiment, the position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of second alignment signals.

[0071] In this embodiment, the position information of the measurement pattern at the center position in the measurement pattern is obtained according to multiple second alignment signals, including: selecting data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error in the multiple second alignment signals as the third alignment signal.

[0072] Selecting data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error among multiple second alignment signals as the third alignment signal is beneficial to obtaining a regular third alignment signal and filtering out messy information, thereby facilitating improving the accuracy of obtaining the position information of the measurement figure at the center of the measurement figure.

[0073] As an example, each measurement area has 12 measurement figures, and the third alignment signals of multiple measurement areas are respectively plotted as dot diagrams, each dot diagram is the third alignment signal corresponding to the 12 measurement figures in a measurement area, the horizontal axis is the position information of the measurement figure, and the vertical axis is the third alignment signal corresponding to the measurement figure.

[0074] Specifically, in this embodiment, data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error is used as the third alignment signal, and the regularity of the structure includes an odd function rule or an even function rule.

[0075] In this embodiment, the multiple third alignment signals obtained are one-dimensional signals.

[0076] In this embodiment, according to the key performance indicator, data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error among the multiple second alignment signals is selected as the third alignment signal.

[0077] According to the key performance indicators, the third alignment signal is obtained, which is simple and easy to operate, and can more accurately obtain the data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error.

[0078] Correspondingly, in this embodiment, the position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of third alignment signals.

[0079] In this embodiment, the position information of the measurement figure at the center position in the measurement figure is obtained according to multiple third alignment signals, including: using multiple third alignment signals as output signals of machine learning and multiple first alignment signals as input signals of machine learning to obtain an algorithm model of machine learning.

[0080] The third alignment signal is the data that meets the feature requirements after dimensionality reduction filtering, and the first alignment signal is the original data. Therefore, a machine learning algorithm model is established with multiple third alignment signals as output signals for machine learning and multiple first alignment signals as input signals for machine learning. A model that can filter out data that meets the feature requirements from the original data is trained, so that by subsequently inputting the original data into the machine learning algorithm model, output data that meets the feature requirements can be directly obtained through the algorithm model.

[0081] In this embodiment, a supervised learning algorithm or a neural network algorithm is used to obtain a machine learning algorithm model.

[0082] The supervised learning algorithm or neural network algorithm is used to obtain the algorithm model of machine learning, which is simple and easy to operate, and the model construction is relatively accurate.

[0083] In this embodiment, a plurality of first alignment signals are used as input signals, and a corresponding plurality of fourth alignment signals are acquired according to an algorithm model.

[0084] After establishing the machine learning algorithm model, taking multiple first alignment signals as input signals, a more accurate fourth alignment signal can be directly output through the model. The fourth alignment signal is used to obtain the position information of the measurement figure at the center position.

[0085] Correspondingly, in this embodiment, the position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of fourth alignment signals.

[0086] In this embodiment, obtaining the position information of the measurement pattern at the center position in the measurement pattern according to the plurality of fourth alignment signals includes: performing function fitting on the plurality of fourth alignment signals in the plurality of measurement areas to obtain an ideal function.

[0087] Function fitting is performed on multiple fourth alignment signals of multiple measurement areas to obtain an ideal function with more obvious morphological features, which is conducive to obtaining ideal data according to the ideal function to update the fourth alignment signal later.

[0088] In this embodiment, function fitting is performed on multiple fourth alignment signals of multiple measurement areas to obtain an ideal function, including: in each measurement area, decentralizing the multiple fourth alignment signals along the data direction of the measurement interval.

[0089] The decentralization process refers to moving the origin of the function composed of a plurality of fourth alignment signals from the origin of the position information to the position of the measurement pattern at the center position.

[0090] Decentralized processing can unify the function morphology of multiple fourth alignment signals, so that function fitting can be performed on the multiple fourth alignment signals later.

[0091] In this embodiment, after decentralization processing, a function type composed of multiple fourth alignment signals is obtained.

[0092] Function fitting is performed on the fourth alignment signal according to the type of function formed by the plurality of fourth alignment signals.

[0093] In this embodiment, a fitting function of the same type as the function type is used to fit a plurality of fourth alignment signals of a plurality of measurement areas into an ideal function.

[0094] As an example, if the function type formed by the multiple fourth alignment signals is an odd function, the multiple fourth alignment signals are fitted using an odd function; if the function type formed by the multiple fourth alignment signals is an even function, the multiple fourth alignment signals are fitted using an even function.

[0095] In this embodiment, a fitting function of the same type as the function type is used to fit a plurality of fourth alignment signals of a plurality of measurement areas into an ideal function, and the order of the fitting function used this time is higher than the order of the fitting function used last time.

[0096] When the order of the fitting function used this time is higher than the order of the fitting function used last time, it is possible to continuously iterate the fitting process, such as Figure 4 The changing trend indicated by the arrow makes the morphological characteristics of the ideal function more and more obvious.

[0097] In this embodiment, ideal signal values ​​corresponding to a plurality of measurement positions in each measurement area are obtained according to the ideal function.

[0098] The ideal signal value is used to update the machine learning algorithm model, so that the machine learning algorithm model becomes more accurate after each function fitting.

[0099] In this embodiment, an updated machine learning algorithm model is obtained based on multiple ideal signal values.

[0100] An updated machine learning algorithm model is obtained based on multiple ideal signal values, so that the machine learning algorithm model is continuously updated and iterated, and becomes more accurate.

[0101] In this embodiment, an updated machine learning algorithm model is obtained based on multiple ideal signal values, including: taking multiple ideal signal values ​​as output signals of machine learning and multiple first alignment signals as input signals of machine learning, and updating the machine learning algorithm model.

[0102] Using multiple ideal signal values ​​as the output signals of machine learning and multiple first alignment signals as the input signals of machine learning, the algorithm model of machine learning is updated, so that the algorithm model of machine learning is continuously updated and iterated according to actual data conditions, and becomes more accurate.

[0103] In this embodiment, multiple updated fourth alignment signals are acquired according to the algorithm model.

[0104] The updated plurality of fourth alignment signals are used to obtain position information of the measurement pattern at the center position in the measurement pattern.

[0105] In this embodiment, obtaining updated multiple fourth alignment signals according to the algorithm model includes: taking multiple ideal signals as input signals, and obtaining updated multiple fourth alignment signals according to the updated algorithm model.

[0106] According to the updated algorithm model, multiple updated fourth alignment signals are obtained, so that the fourth alignment signal is also continuously iterated and updated, and becomes more accurate.

[0107] In this embodiment, it is determined whether the function shape of the ideal function meets the characteristic requirements;

[0108] Whether it is necessary to return to iterative fitting processing is determined based on whether the characteristic performance of the function morphology of the ideal function meets the accuracy requirements of the characteristics.

[0109] Specifically, in this embodiment, if yes, the position information of the measurement pattern at the center position in the measurement pattern is obtained according to the updated multiple fourth alignment signals; otherwise, return to execute function fitting of multiple fourth alignment signals of multiple measurement areas to obtain an ideal function.

[0110] In this embodiment, obtaining the position information of the measurement pattern at the center position in the measurement pattern according to the plurality of fourth alignment signals includes: in each measurement area, obtaining the center point of the dot diagram according to the dot diagram formed by the plurality of fourth alignment signals.

[0111] In each measurement area, the center point of the dot diagram is obtained according to the dot diagram formed by multiple fourth alignment signals, which is intuitive and efficient. The position information of the center point is the position information of the measurement pattern at the center position among the multiple measurement patterns.

[0112] In this embodiment, the measurement interval corresponding to the center point is obtained.

[0113] The measurement interval corresponding to the center point is obtained, thereby obtaining the position information corresponding to the center point.

[0114] As an example, the measurement interval corresponding to the center point is +2, and accordingly, the position information corresponding to the center point is OVL+2.

[0115] Execute step S6: obtain the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information.

[0116] By acquiring the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information, the absolute overlay error between the first alignment pattern and the second alignment pattern of the front layer can be acquired more accurately.

[0117] In this embodiment, the measurement interval is set to detect the first alignment signal at each measurement position, detect the change of the first alignment signal, and obtain the overlay error according to the sudden change of the first alignment signal. That is, the overlay error is obtained by a nonlinear method, which is beneficial to improving the accuracy of the overlay error measurement and has a low measurement cost.

[0118] Specifically, in this embodiment, obtaining the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information includes: obtaining the negative value of the measurement interval corresponding to the center point as the overlay error.

[0119] As an example, the measurement interval corresponding to the center point is +2, and accordingly, the position information corresponding to the center point is OVL+2, and the overlay error between the first alignment pattern and the second alignment pattern of the front layer is -2.

[0120] Correspondingly, the present invention also provides a system for measuring overlay error. Figure 3 It is a functional block diagram of an embodiment of an optical proximity correction system of the present invention.

[0121] In this embodiment, the optical proximity correction system 50 includes: a structure providing module 501, which is used to provide a front layer structure and a current layer structure located on the front layer structure, the front layer structure includes a plurality of second alignment patterns, and the current layer structure includes a plurality of measurement areas corresponding to the second alignment patterns; a measurement interval setting module 502, which is used to set the measurement interval of the current layer pattern; a measurement pattern acquisition module 503, which acquires a plurality of measurement patterns arranged in parallel at the measurement interval in each measurement area of ​​the current layer, including a first alignment pattern and other measurement patterns located on both sides of the first alignment pattern; a first alignment signal acquisition module 504, which is used to acquire a first alignment signal between the measurement pattern and the second alignment pattern of the front layer with each measurement pattern as the measurement position; a position information acquisition module 505, which is used to acquire the position information of the measurement pattern at the center position in the measurement pattern according to the plurality of first alignment signals; and an overlay error acquisition module 506, which is used to acquire the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information.

[0122] Execute step S1: provide a front layer structure and a current layer structure located on the front layer structure, wherein the front layer structure includes a plurality of second alignment patterns, and the current layer structure includes a plurality of measurement areas corresponding to the second alignment patterns.

[0123] The second alignment pattern of the front layer structure and the measurement area corresponding to the current layer structure and the second alignment pattern are used to align the current layer with the front layer to obtain the overlay error between the current layer and the front layer.

[0124] It should be noted that, in this embodiment, when the measurement area of ​​the current layer structure corresponds to the second alignment pattern of the previous layer structure, it means that the previous layer structure has multiple second alignment patterns at different positions, and each second alignment pattern has a corresponding measurement area in the current layer structure.

[0125] Execute step S2: set the measurement interval of the current layer of graphics.

[0126] In the current layer structure, the measurement interval of the current layer pattern is set to arrange the measurement pattern in the measurement area.

[0127] In this embodiment, in actual applications, the measurement interval is set according to specific needs. If the measurement interval is set smaller, the measurement accuracy of the overlay error can be higher. If the measurement accuracy requirement is not high, the size setting of the measurement interval can be appropriately relaxed.

[0128] Execute step S3: in each measurement area of ​​the layer, obtain a plurality of measurement patterns arranged in parallel at measurement intervals, including a first alignment pattern and other measurement patterns located on both sides of the first alignment pattern.

[0129] The measurement pattern is used to measure with the second alignment pattern of the front layer structure to obtain a first alignment signal, wherein the first alignment pattern in the measurement pattern is used to measure with the second alignment pattern to obtain an overlay error.

[0130] As an example, the second alignment pattern of the front layer structure is a fin structure, and when the measurement pattern of the layer structure is a plurality of gate structures arranged in parallel, the gate structures are arranged in parallel along the extension direction of the fin structure and cross the fin structure.

[0131] In this embodiment, in each measurement area of ​​the layer, multiple measurement patterns are set, and the spacing between adjacent measurement patterns is the measurement interval, that is, each measurement pattern corresponds to different position information. Subsequently, by obtaining the first alignment signal between each measurement pattern and the second alignment pattern of the previous layer, the first alignment signal corresponding to each position information is obtained, thereby obtaining the position information of the measurement pattern at the center position of the measurement pattern.

[0132] As an example, the measurement interval is set to 1nm, and the position information of the first alignment pattern is taken as the origin, marked as OVL. Then in one measurement area, the position information of multiple measurement patterns are arranged in sequence: OVL+n, OVL+(n-1), OVL+(n-2)...OVL+0...OVL-n, OVL-(n+1), OVL-(n+2), and a total of 2n+3 measurement patterns are set.

[0133] It should be noted that, in the present embodiment, the first alignment figure is a graphic structure for measuring the overlay error. In the measurement area, multiple measurement figures are arranged on both sides of the first alignment figure among multiple measurement figures, but the first alignment figure may be located at the center position of the multiple measurement figures, or may not be located at the center position of the multiple measurement figures.

[0134] Execute step S4: take each measurement pattern as a measurement position, and obtain a first alignment signal between the measurement pattern and the second alignment pattern of the front layer.

[0135] The first alignment signal is used to determine the positional relationship between the measurement pattern and the second alignment pattern of the front layer, so as to obtain the position information of the measurement pattern located at the center position in the measurement pattern.

[0136] In this embodiment, each measurement pattern is used as a measurement position to obtain a first alignment signal between the measurement pattern and the second alignment pattern of the front layer, wherein the first alignment signal includes an optical signal.

[0137] The measurement is performed by acquiring optical signals, which is simple and easy to operate, and can be measured in real time when the layer structure is formed, so that the acquisition of overlay error is faster and more convenient.

[0138] In other embodiments, the first alignment signal may also be an electrical signal.

[0139] Execute step S5: obtain position information of the measurement pattern at the center position in the measurement pattern according to the plurality of first alignment signals.

[0140] The position information of the measurement pattern at the center position in the measurement pattern is obtained according to the plurality of first alignment signals, so that the overlay error between the first alignment pattern and the second alignment pattern is subsequently obtained according to the position information of the measurement pattern at the center position.

[0141] Specifically, in the present embodiment, the measurement figure at the center position in the measurement figure is the optimal position relative to the second alignment figure, that is, the measurement figure at the center position in the measurement figure is the reference position required for the process in the semiconductor manufacturing process, and the first alignment figure is the formation position of the first alignment figure in the layer structure. Moreover, according to the foregoing, the position information includes information on the relative position deviation of the measurement figure relative to the first alignment figure. Therefore, the overlay error can be subsequently obtained based on the position information of the measurement figure at the center position in the measurement figure.

[0142] In this embodiment, before the multiple first alignment signals are subsequently subjected to dimensionality reduction processing, the method further includes: in each measurement area, simplifying the first alignment signal, and subtracting the first alignment signal corresponding to the first alignment pattern from the first alignment signal of each measurement pattern.

[0143] Taking the first alignment signal corresponding to the first alignment pattern as a reference, simplifying the first alignment signal of each measurement pattern is beneficial to simplifying the subsequent dimensionality reduction process and improving the efficiency of the dimensionality reduction process.

[0144] In this embodiment, obtaining the position information of the measurement pattern at the center position in the measurement pattern according to the plurality of first alignment signals includes: performing dimensionality reduction processing on the plurality of first alignment signals to obtain corresponding second alignment signals.

[0145] The multiple first alignment signals are processed by dimensionality reduction to obtain the corresponding second alignment signals, and the position information of the measurement figure at the center position is subsequently obtained through the second alignment signals obtained after dimensionality reduction, which is conducive to simplifying the processing, reducing the amount of calculation data, shortening the calculation time, and improving the processing efficiency.

[0146] Accordingly, in this embodiment, when performing dimensionality reduction processing on a plurality of first alignment signals, performing dimensionality reduction processing on the first alignment signal after the simplified processing is beneficial to improving processing efficiency.

[0147] In this embodiment, a principal component analysis method is used to perform dimensionality reduction processing on the plurality of first alignment signals.

[0148] The principal component analysis (PCA) method is used for dimensionality reduction processing, which can replace the original data with more dimensions with data with fewer dimensions, and at the same time can reflect the original data information with more dimensions, which is beneficial to retain the original information of multiple first alignment signals while achieving the dimensionality reduction effect.

[0149] Specifically, as an example, in this embodiment, there are m first alignment signals, each of which is n-dimensional. In this embodiment, the principal component analysis method is used to reduce the dimension of the multiple first alignment signals, including: subtracting the mean of each dimension from the dimension, and forming an n×m matrix X; calculating the covariance matrix c=1 / m XX T ; Calculate the eigenvalues ​​and eigenvectors of the covariance matrix, and arrange the n eigenvectors from large to small and from top to bottom into a matrix A according to the corresponding eigenvalues; Take the top k eigenvectors of matrix A to form a matrix P; Y = PX, which is the data after dimensionality reduction to K dimensions.

[0150] In this embodiment, the top k eigenvectors of the matrix A are taken according to the principal component contribution rate and the cumulative contribution rate. Generally, the greater the contribution rate of the principal component, the more information representing the retained original data.

[0151] Correspondingly, in this embodiment, the position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of second alignment signals.

[0152] In this embodiment, the position information of the measurement pattern at the center position in the measurement pattern is obtained according to multiple second alignment signals, including: selecting data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error in the multiple second alignment signals as the third alignment signal.

[0153] Selecting data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error among multiple second alignment signals as the third alignment signal is beneficial to obtaining a regular third alignment signal and filtering out messy information, thereby facilitating improving the accuracy of obtaining the position information of the measurement figure at the center of the measurement figure.

[0154] As an example, each measurement area has 12 measurement figures, and the third alignment signals of multiple measurement areas are respectively plotted as dot diagrams, each dot diagram is the third alignment signal corresponding to the 12 measurement figures in a measurement area, the horizontal axis is the position information of the measurement figure, and the vertical axis is the third alignment signal corresponding to the measurement figure.

[0155] Specifically, in this embodiment, data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error is used as the third alignment signal, and the regularity of the structure includes an odd function rule or an even function rule.

[0156] In this embodiment, the multiple third alignment signals obtained are one-dimensional signals.

[0157] In this embodiment, according to the key performance indicator, data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error among the multiple second alignment signals is selected as the third alignment signal.

[0158] According to the key performance indicators, the third alignment signal is obtained, which is simple and easy to operate, and can more accurately obtain the data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error.

[0159] Correspondingly, in this embodiment, the position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of third alignment signals.

[0160] In this embodiment, the position information of the measurement figure at the center position in the measurement figure is obtained according to multiple third alignment signals, including: using multiple third alignment signals as output signals of machine learning and multiple first alignment signals as input signals of machine learning to obtain an algorithm model of machine learning.

[0161] The third alignment signal is the data that meets the feature requirements after dimensionality reduction filtering, and the first alignment signal is the original data. Therefore, a machine learning algorithm model is established with multiple third alignment signals as output signals for machine learning and multiple first alignment signals as input signals for machine learning. A model that can filter out data that meets the feature requirements from the original data is trained, so that by subsequently inputting the original data into the machine learning algorithm model, output data that meets the feature requirements can be directly obtained through the algorithm model.

[0162] In this embodiment, a supervised learning algorithm or a neural network algorithm is used to obtain a machine learning algorithm model.

[0163] The supervised learning algorithm or neural network algorithm is used to obtain the algorithm model of machine learning, which is simple and easy to operate, and the model construction is relatively accurate.

[0164] In this embodiment, a plurality of first alignment signals are used as input signals, and a corresponding plurality of fourth alignment signals are acquired according to an algorithm model.

[0165] After establishing the machine learning algorithm model, taking multiple first alignment signals as input signals, a more accurate fourth alignment signal can be directly output through the model. The fourth alignment signal is used to obtain the position information of the measurement figure at the center position.

[0166] Correspondingly, in this embodiment, the position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of fourth alignment signals.

[0167] In this embodiment, obtaining the position information of the measurement pattern at the center position in the measurement pattern according to the plurality of fourth alignment signals includes: performing function fitting on the plurality of fourth alignment signals in the plurality of measurement areas to obtain an ideal function.

[0168] Function fitting is performed on multiple fourth alignment signals of multiple measurement areas to obtain an ideal function with more obvious morphological features, which is conducive to obtaining ideal data according to the ideal function to update the fourth alignment signal later.

[0169] In this embodiment, function fitting is performed on multiple fourth alignment signals of multiple measurement areas to obtain an ideal function, including: in each measurement area, decentralizing the multiple fourth alignment signals along the data direction of the measurement interval.

[0170] The decentralization process refers to moving the origin of the function composed of a plurality of fourth alignment signals from the origin of the position information to the position of the measurement pattern at the center position.

[0171] Decentralized processing can unify the function morphology of multiple fourth alignment signals, so that function fitting can be performed on the multiple fourth alignment signals later.

[0172] In this embodiment, after decentralization processing, a function type composed of multiple fourth alignment signals is obtained.

[0173] Function fitting is performed on the fourth alignment signal according to the type of function formed by the plurality of fourth alignment signals.

[0174] In this embodiment, a fitting function of the same type as the function type is used to fit a plurality of fourth alignment signals of a plurality of measurement areas into an ideal function.

[0175] As an example, if the function type formed by the multiple fourth alignment signals is an odd function, the multiple fourth alignment signals are fitted using an odd function; if the function type formed by the multiple fourth alignment signals is an even function, the multiple fourth alignment signals are fitted using an even function.

[0176] In this embodiment, a fitting function of the same type as the function type is used to fit a plurality of fourth alignment signals of a plurality of measurement areas into an ideal function, and the order of the fitting function used this time is higher than the order of the fitting function used last time.

[0177] When the order of the fitting function used this time is higher than the order of the fitting function used last time, it is possible to continuously iterate the fitting process, such as Figure 4 The changing trend indicated by the arrow makes the morphological characteristics of the ideal function more and more obvious.

[0178] In this embodiment, ideal signal values ​​corresponding to a plurality of measurement positions in each measurement area are obtained according to the ideal function.

[0179] The ideal signal value is used to update the machine learning algorithm model, so that the machine learning algorithm model becomes more accurate after each function fitting.

[0180] In this embodiment, an updated machine learning algorithm model is obtained based on multiple ideal signal values.

[0181] An updated machine learning algorithm model is obtained based on multiple ideal signal values, so that the machine learning algorithm model is continuously updated and iterated, and becomes more accurate.

[0182] In this embodiment, an updated machine learning algorithm model is obtained based on multiple ideal signal values, including: taking multiple ideal signal values ​​as output signals of machine learning and multiple first alignment signals as input signals of machine learning, and updating the machine learning algorithm model.

[0183] Using multiple ideal signal values ​​as the output signals of machine learning and multiple first alignment signals as the input signals of machine learning, the algorithm model of machine learning is updated, so that the algorithm model of machine learning is continuously updated and iterated according to actual data conditions, and becomes more accurate.

[0184] In this embodiment, multiple updated fourth alignment signals are acquired according to the algorithm model.

[0185] The updated plurality of fourth alignment signals are used to obtain position information of the measurement pattern at the center position in the measurement pattern.

[0186] In this embodiment, obtaining updated multiple fourth alignment signals according to the algorithm model includes: taking multiple ideal signals as input signals, and obtaining updated multiple fourth alignment signals according to the updated algorithm model.

[0187] According to the updated algorithm model, multiple updated fourth alignment signals are obtained, so that the fourth alignment signal is also continuously iterated and updated, and becomes more accurate.

[0188] In this embodiment, it is determined whether the function shape of the ideal function meets the characteristic requirements;

[0189] Whether it is necessary to return to iterative fitting processing is determined based on whether the characteristic performance of the function morphology of the ideal function meets the accuracy requirements of the characteristics.

[0190] Specifically, in this embodiment, if yes, the position information of the measurement pattern at the center position in the measurement pattern is obtained according to the updated multiple fourth alignment signals; otherwise, return to execute function fitting of multiple fourth alignment signals of multiple measurement areas to obtain an ideal function.

[0191] In this embodiment, obtaining the position information of the measurement pattern at the center position in the measurement pattern according to the plurality of fourth alignment signals includes: in each measurement area, obtaining the center point of the dot diagram according to the dot diagram formed by the plurality of fourth alignment signals.

[0192] In each measurement area, the center point of the dot diagram is obtained according to the dot diagram formed by multiple fourth alignment signals, which is intuitive and efficient. The position information of the center point is the position information of the measurement pattern at the center position among the multiple measurement patterns.

[0193] In this embodiment, the measurement interval corresponding to the center point is obtained.

[0194] The measurement interval corresponding to the center point is obtained, thereby obtaining the position information corresponding to the center point.

[0195] As an example, the measurement interval corresponding to the center point is +2, and accordingly, the position information corresponding to the center point is OVL+2.

[0196] Execute step S6: obtain the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information.

[0197] By acquiring the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information, the absolute overlay error between the first alignment pattern and the second alignment pattern of the front layer can be acquired more accurately.

[0198] In this embodiment, the measurement interval is set to detect the first alignment signal at each measurement position, detect the change of the first alignment signal, and obtain the overlay error according to the sudden change of the first alignment signal. That is, the overlay error is obtained by a nonlinear method, which is beneficial to improving the accuracy of the overlay error measurement and has a low measurement cost.

[0199] Specifically, in this embodiment, obtaining the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information includes: obtaining the negative value of the measurement interval corresponding to the center point as the overlay error.

[0200] As an example, the measurement interval corresponding to the center point is +2, and accordingly, the position information corresponding to the center point is OVL+2, and the overlay error between the first alignment pattern and the second alignment pattern of the front layer is -2.

[0201] Accordingly, an embodiment of the present invention further provides a device, which can implement the overlay error measurement method provided by the embodiment of the present invention by loading the above-mentioned overlay error measurement method in the form of a program. An optional hardware structure of the terminal device provided by the embodiment of the present invention can be as follows Figure 4As shown, it includes: at least one processor 01, at least one communication interface 02, at least one memory 03 and at least one communication bus 04.

[0202] In this embodiment, the number of processor 01, communication interface 02, memory 03, and communication bus 04 is at least one, and the processor 01, communication interface 02, and memory 03 communicate with each other through the communication bus 04. The communication interface 02 can be an interface of a communication module for network communication, such as an interface of a GSM module. The processor 01 may be a central processing unit CPU, or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement an embodiment of the present invention. The memory 03 may include a high-speed RAM memory, and may also include a non-volatile memory (NVM), such as at least one disk storage. Among them, the memory 03 stores one or more computer instructions, and the one or more computer instructions are executed by the processor 01 to implement the overlay error measurement method provided in an embodiment of the present invention.

[0203] It should be noted that the above-mentioned terminal device may also include other devices (not shown) that may not be necessary for understanding the contents disclosed in the embodiments of the present invention; given that these other devices may not be necessary for understanding the contents disclosed in the embodiments of the present invention, the embodiments of the present invention will not introduce them one by one.

[0204] An embodiment of the present invention further provides a storage medium, wherein the storage medium stores one or more computer instructions, and the one or more computer instructions are used to implement the overlay error measurement method provided by the embodiment of the present invention.

[0205] In the overlay error measurement method provided in an embodiment of the present invention, a measurement interval is set to detect the first alignment signal at each measurement position, detect the change of the first alignment signal, and obtain the overlay error according to the sudden change of the first alignment signal. That is, a nonlinear method is used to obtain the overlay error, which is beneficial to improving the accuracy of overlay error measurement and has low measurement cost.

[0206] The above-mentioned embodiments of the present invention are combinations of elements and features of the present invention. Unless otherwise mentioned, the elements or features may be considered as optional. Each element or feature may be put into practice without being combined with other elements or features. In addition, the embodiments of the present invention may be constructed by combining some elements and / or features. The order of operations described in the embodiments of the present invention may be rearranged. Some constructions of any one embodiment may be included in another embodiment, and may be replaced by the corresponding construction of another embodiment. It is obvious to those skilled in the art that claims that do not have a clear reference relationship to each other in the attached claims may be combined into embodiments of the present invention, or may be included as new claims in the amendment after submitting this application.

[0207] Embodiments of the present invention may be implemented by various means such as hardware, firmware, software or a combination thereof. In a hardware configuration, the method according to an exemplary embodiment of the present invention may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, etc. In a firmware or software configuration, embodiments of the present invention may be implemented in the form of modules, processes, functions, etc. The software code may be stored in a memory unit and executed by a processor. The memory unit is located inside or outside the processor and may send data to the processor and receive data from the processor via various known means.

[0208] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0209] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.

Claims

1. A method for measuring overlay error, characterized in that: include: Providing a front layer structure and a current layer structure located on the front layer structure, wherein the front layer structure includes a plurality of second alignment patterns, and the current layer structure includes a plurality of measurement areas corresponding to the second alignment patterns; Set the measurement interval of the current layer graphics; In each measurement area of ​​the layer, a plurality of measurement patterns arranged in parallel at the measurement interval are obtained, including a first alignment pattern and other measurement patterns located on both sides of the first alignment pattern; Taking each of the measurement patterns as a measurement position, obtaining a first alignment signal between the measurement pattern and a second alignment pattern of a front layer; Acquire position information of a measurement pattern at a center position in the measurement pattern according to the plurality of first alignment signals; An overlay error between the first alignment pattern and the second alignment pattern of the front layer is obtained according to the position information.

2. The method for measuring overlay error as claimed in claim 1, characterized in that: Taking each of the measurement patterns as a measurement position, a first alignment signal between the measurement pattern and a second alignment pattern of a front layer is obtained, wherein the first alignment signal includes an optical signal.

3. The method for measuring overlay error as claimed in claim 1, wherein: Acquiring position information of a measurement pattern at a center position in the measurement pattern according to the plurality of first alignment signals, comprising: Performing dimensionality reduction processing on the plurality of first alignment signals to obtain corresponding second alignment signals; The position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of the second alignment signals.

4. The method for measuring overlay error as claimed in claim 3, characterized in that: Before performing dimensionality reduction processing on the plurality of first alignment signals, the method further includes: in each of the measurement areas, performing simplification processing on the first alignment signal, and subtracting the first alignment signal corresponding to the first alignment pattern from the first alignment signal of each of the measurement patterns; In performing dimensionality reduction processing on the plurality of first alignment signals, dimensionality reduction processing is performed on the first alignment signal after the simplification processing.

5. The method for measuring overlay error as claimed in claim 3, characterized in that: The principal component analysis method is used to perform dimension reduction processing on the plurality of first alignment signals.

6. The method for measuring overlay error as claimed in claim 3, characterized in that: Acquiring position information of a measurement pattern at a center position in the measurement pattern according to the plurality of second alignment signals, comprising: Selecting data corresponding to a dimension that best reflects the nonlinear characteristics of the overlay error from among the plurality of second alignment signals as a third alignment signal; The position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of the third alignment signals.

7. The method for measuring overlay error as claimed in claim 6, wherein: According to the key performance indicator, data corresponding to the dimension that can best reflect the nonlinear characteristics of the overlay error among the plurality of the second alignment signals is selected as the third alignment signal.

8. The method for measuring overlay error as claimed in claim 6, wherein: Acquiring position information of a measurement pattern at a center position in the measurement pattern according to the plurality of third alignment signals, comprising: Using the plurality of the third alignment signals as output signals of machine learning and using the plurality of the first alignment signals as input signals of machine learning to obtain an algorithm model of machine learning; Taking the first alignment signals as input signals, and acquiring corresponding fourth alignment signals according to the algorithm model; The position information of the measurement pattern at the center position in the measurement pattern is acquired according to the plurality of fourth alignment signals.

9. The method for measuring overlay error as claimed in claim 8, characterized in that: Acquiring position information of a measurement pattern at a center position in the measurement pattern according to the plurality of fourth alignment signals, comprising: Performing function fitting on a plurality of fourth alignment signals of a plurality of the measurement areas to obtain an ideal function; According to the ideal function, obtaining ideal signal values ​​corresponding to a plurality of measurement positions in each of the measurement areas; Acquire an updated algorithm model of the machine learning according to the plurality of ideal signal values; Acquire updated multiple fourth alignment signals according to the algorithm model; Determining whether the function shape of the ideal function meets the characteristic requirements; If yes, acquiring the position information of the measurement pattern at the center position in the measurement pattern according to the updated plurality of fourth alignment signals; Otherwise, the process returns to executing function fitting for the plurality of fourth alignment signals of the plurality of measurement areas to obtain an ideal function.

10. The method for measuring overlay error according to claim 8 or 9, characterized in that: A supervised learning algorithm or a neural network algorithm is used to obtain the algorithm model of the machine learning.

11. The method for measuring overlay error as claimed in claim 9, wherein: Performing function fitting on the plurality of fourth alignment signals of the plurality of measurement areas to obtain an ideal function, comprising: performing decentralization processing on the plurality of fourth alignment signals along the data direction of the measurement interval in each of the measurement areas; After performing the decentralization process, obtaining a function type composed of a plurality of the fourth alignment signals; The fourth alignment signals of the measurement areas are fitted into an ideal function by using a fitting function of the same type as the function.

12. The method for measuring overlay error according to claim 11, wherein: A fitting function of the same type as the function type is used to fit a plurality of fourth alignment signals of a plurality of the measurement areas into an ideal function, wherein the order of the fitting function used this time is higher than the order of the fitting function used last time.

13. The method for measuring overlay error as claimed in claim 9, wherein: Acquiring an updated algorithm model of the machine learning according to the multiple ideal signal values, comprising: taking the multiple ideal signal values ​​as output signals of the machine learning and taking the multiple first alignment signals as input signals of the machine learning, and updating the algorithm model of the machine learning; Acquiring updated multiple fourth alignment signals according to the algorithm model includes: taking the multiple ideal signals as input signals, and acquiring updated multiple fourth alignment signals according to the updated algorithm model.

14. The method for measuring overlay error according to claim 8 or 9, characterized in that: Acquiring position information of a measurement pattern at a central position in the measurement pattern according to the plurality of fourth alignment signals, comprising: acquiring a central point of the dot diagram in each measurement area according to a dot diagram formed by the plurality of fourth alignment signals; Obtain the measurement interval corresponding to the center point.

15. The method for measuring overlay error according to claim 14, wherein: Acquiring an overlay error between the first alignment pattern and a second alignment pattern of a front layer according to the position information includes: acquiring a negative value of a measurement interval corresponding to the center point as the overlay error.

16. A system for measuring overlay error, characterized in that: include: A structure providing module, used for providing a front layer structure and a current layer structure located on the front layer structure, wherein the front layer structure includes a plurality of second alignment patterns, and the current layer structure includes a plurality of measurement areas corresponding to the second alignment patterns; The measurement interval setting module is used to set the measurement interval of the current layer graphics; A measurement pattern acquisition module, which acquires, in each measurement area of ​​the current layer, a plurality of measurement patterns arranged in parallel at the measurement interval, including a first alignment pattern and other measurement patterns located on both sides of the first alignment pattern; A first alignment signal acquisition module, used for taking each of the measurement patterns as a measurement position and acquiring a first alignment signal between the measurement pattern and a second alignment pattern of a front layer; a position information acquisition module, configured to acquire position information of a measurement pattern at a center position in the measurement pattern according to a plurality of the first alignment signals; The overlay error acquisition module is used to acquire the overlay error between the first alignment pattern and the second alignment pattern of the front layer according to the position information.

17. A device, characterized in that It comprises at least one memory and at least one processor, wherein the memory stores one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method for measuring the overlay error as described in any one of claims 1-15.

18. A storage medium, characterized in that: The storage medium stores one or more computer instructions, and the one or more computer instructions are used to implement the method for measuring overlay error as described in any one of claims 1-15.

Images