X-ray small-angle scattering measurement method and device based on machine learning acceleration

Abstract

An X-ray small-angle scattering measurement method and device based on machine learning acceleration is used to speed up the simulation calculation of scattering signals when determining the critical dimensions of a sample, thereby reducing the time spent on solving the critical dimensions. The method includes: obtaining a measurement scattering diagram of the sample to be tested through a measurement system; the measurement scattering diagram is a scattering diagram formed after the X-rays of the measurement system are scattered by the sample to be tested under different positional relationships; the positional relationship is the azimuth angle and incident angle of the X-rays and the sample to be tested; analyzing the distribution characteristics of the scattering signals in the measurement scattering diagram of the sample to be tested to obtain a measurement characteristic signal that characterizes the critical dimensions of the sample to be tested; comparing the measurement characteristic signal with the simulation characteristic signal to determine the simulation characteristic signal with the smallest difference from the measurement characteristic signal; and taking a set of simulated critical dimensions corresponding to the simulation characteristic signal with the smallest difference from the measurement characteristic signal as the critical dimensions of the sample to be tested.

Description

Technical Field

[0001] The present application relates to the field of measurement technology, and in particular to a method and device for measuring small-angle X-ray scattering based on machine learning acceleration. Background Art

[0002] Critical Dimension (CD) refers to a special line pattern that reflects the width of the characteristic lines of integrated circuits, designed to evaluate and control the pattern processing accuracy of the process in the manufacture of integrated circuit photomasks and lithography processes. The Small Angle X-ray Scattering (SAXS) critical dimension measurement method uses an optical measurement method that measures the critical dimensions of a sample by analyzing the distribution characteristics of the scattered signal after the X-rays are scattered through the sample to form a scattered field containing the sample structure information.

[0003] As the structure of semiconductor devices becomes more and more complex, the dimensions that characterize the key dimensions of the physical model are also getting larger. This not only makes the calculation of simulated scattering signals time-consuming, but also increases the number of iterations of the optimization algorithm to find the optimal solution. Ultimately, the solution process is very slow and cannot meet industrial needs.

[0004] Therefore, there is an urgent need for a solution to speed up the simulation calculation of scattering signals when determining the critical dimensions of the sample, thereby reducing the time spent on solving the critical dimensions. Summary of the invention

[0005] The present application provides an X-ray small-angle scattering measurement method and device based on machine learning acceleration, which is used to speed up the simulation calculation of scattering signals when determining the critical dimensions of a sample, thereby reducing the time spent on solving the critical dimensions.

[0006] In a first aspect, the present application provides a method for determining a critical dimension of a sample, the method comprising: obtaining a measurement scattering diagram of a sample to be tested through a measurement system; the measurement scattering diagram is a scattering diagram formed after X-rays of the measurement system are scattered after passing through the sample to be tested at different positional relationships; the positional relationship is the azimuth angle and the incident angle of the X-rays and the sample to be tested; analyzing the distribution characteristics of the scattering signal in the measurement scattering diagram of the sample to be tested to obtain a measurement characteristic signal characterizing the critical dimension of the sample to be tested; comparing the measurement characteristic signal with a simulation characteristic signal to determine the simulation characteristic signal with the smallest difference from the measurement characteristic signal; and using a set of simulated critical dimensions corresponding to the simulation characteristic signal with the smallest difference from the measurement characteristic signal as the critical dimension of the sample to be tested.

[0007] In the above technical solution, a measurement characteristic signal characterizing the critical dimension of the sample to be tested is extracted from the measurement scatter diagram of the sample to be tested, and then the feature-extracted measurement characteristic signal is compared with the simulation characteristic signal, which can reduce the comparison dimension and thus reduce the time spent on solving the critical dimension.

[0008] In one possible design, the measured characteristic signal is compared with the simulated characteristic signal to determine the simulated characteristic signal having the smallest difference from the measured characteristic signal, including: generating a set of first simulated critical dimensions within the variation range of the critical dimensions of the sample to be measured; inputting the first simulated critical dimensions into a forward analysis model to obtain a first simulated characteristic signal corresponding to the first simulated critical dimensions; calculating a loss value between the first simulated characteristic signal and the measured characteristic signal; if the loss value is less than a set threshold, using the first simulated characteristic signal as the simulated characteristic signal having the smallest difference from the measured characteristic signal.

[0009] In the above technical solution, the simulation characteristic signal corresponding to the simulation key dimension is obtained by the forward analysis model. Compared with using the traditional mathematical formula to solve the simulation characteristic signal corresponding to the simulation key dimension, the speed of calculating the simulation characteristic signal can be accelerated. Especially when the structure of the sample to be tested is complex, most of the time occupied by the simulation calculation in the solution process can be saved, and the iteration speed can be greatly improved.

[0010] In one possible design, the method also includes: if the loss value is greater than a set threshold, adjusting the value of the first simulation key dimension according to the loss value to obtain a set of second simulation key dimensions; inputting the second simulation key dimensions into the forward analysis model to obtain a second simulation characteristic signal corresponding to the second simulation key dimensions; calculating the loss value of the second simulation characteristic signal and the measured characteristic signal until the difference in the loss value is less than the set threshold.

[0011] In the above technical solution, by adjusting the value of the simulated critical dimension according to the loss value, the simulated critical dimension that meets the conditions can be quickly obtained, thereby accelerating the iteration speed in the process of solving the critical dimension of the sample to be tested.

[0012] In a possible design, before comparing the measured characteristic signal with the simulated characteristic signal, the method further includes: generating multiple groups of simulated key dimensions within the variation range of the key dimensions of the sample to be measured; generating simulated scattering diagrams of each group of simulated key dimensions in different positional relationships with reference to the measurement system; and extracting the simulated characteristic signals characterizing the key dimensions in the simulated scattering diagrams by analyzing the distribution characteristics of the scattering signals in the simulated scattering diagrams corresponding to each group of simulated key dimensions and the correlation between each group of simulated key dimensions and the corresponding simulated scattering diagrams.

[0013] In the above technical solution, the simulation characteristic signal representing the key dimension in the simulation scatter diagram corresponding to the simulated key dimension is extracted in the preprocessing stage, and the characteristic signal is measured and compared with the simulation characteristic signal in the actual measurement stage, which can reduce the comparison dimension and thus reduce the time spent on solving the key dimension.

[0014] In a possible design, based on the method of extracting simulated characteristic signals that characterize key dimensions in a simulated scattering map, a feature extraction model is obtained; the feature extraction model is a model that performs feature extraction functions under different positional relationships, or there are multiple feature extraction models, each of which is a model that performs feature extraction functions under a single positional relationship; and the distribution characteristics of the scattering signals in the measured scattering map of the sample to be tested are analyzed, including: analyzing the distribution characteristics of the scattering signals in the measured scattering map of the sample to be tested by using the feature extraction model.

[0015] In the above technical solution, the measurement feature signal characterizing the key size of the sample to be tested can be extracted from the measurement scatter diagram of the sample to be tested through the feature extraction model, and then the feature-extracted measurement feature signal is compared with the simulation feature signal, which can reduce the comparison dimension and thus reduce the time spent on solving the key size.

[0016] In one possible design, the method further includes: taking each group of simulated critical dimensions as input and taking characteristic signals characterizing the critical dimensions as output, and training the forward analysis model.

[0017] In the above technical solution, the simulation characteristic signal corresponding to the simulation critical dimension is determined by the forward analysis model obtained through training. Compared with using traditional mathematical formulas to solve the simulation characteristic signal corresponding to the simulation critical dimension, the speed of calculating the simulation characteristic signal can be accelerated. Especially when the structure of the sample to be tested is complex, most of the time occupied by the simulation calculation in the solution process can be saved, thereby greatly improving the iteration speed.

[0018] In one possible design, multiple groups of simulated key dimensions are generated within the variation range of the key dimensions of the sample to be tested, including: generating M groups of simulated key dimensions within the variation range of the key dimensions of the sample to be tested; wherein the number of M is positively correlated with the number of parameter dimensions and / or the size of the parameter range of the sample to be tested.

[0019] In the above technical solution, the larger the number of parameter dimensions and / or parameter range of the sample to be tested, the more simulation key dimensions are generated, which can make the sample data richer and enable the sample data set to include key dimensions that characterize the structure of the sample to be tested.

[0020] In a second aspect, an embodiment of the present application provides a device for determining a critical dimension of a sample, the device comprising:

[0021] An acquisition module is used to acquire a measurement scattering diagram of the sample to be tested through a measurement system; the measurement scattering diagram is a scattering diagram formed after the X-rays of the measurement system are scattered by the sample to be tested under different positional relationships; the positional relationship is the azimuth angle and the incident angle of the X-rays and the sample to be tested;

[0022] A processing module, used for analyzing the distribution characteristics of the scattering signal in the measurement scattering diagram of the sample to be tested, and obtaining a measurement characteristic signal characterizing the critical dimension of the sample to be tested;

[0023] A comparison module is used to compare the measured characteristic signal with the simulated characteristic signal to determine the simulated characteristic signal with the smallest difference from the measured characteristic signal; and to use a set of simulated critical dimensions corresponding to the simulated characteristic signal with the smallest difference from the measured characteristic signal as the critical dimensions of the sample to be tested.

[0024] In one possible design, the comparison module, when comparing the measured characteristic signal with the simulated characteristic signal to determine the simulated characteristic signal with the smallest difference from the measured characteristic signal, is specifically used to generate a set of first simulated key dimensions within the variation range of the key dimensions of the sample to be tested; input the first simulated key dimensions into the forward analysis model to obtain the first simulated characteristic signal corresponding to the first simulated key dimensions; calculate the loss value between the first simulated characteristic signal and the measured characteristic signal; if the loss value is less than a set threshold, use the first simulated characteristic signal as the simulated characteristic signal with the smallest difference from the measured characteristic signal.

[0025] In one possible design, the comparison module is also used to adjust the value of the first simulation key dimension according to the loss value to obtain a set of second simulation key dimensions if the loss value is greater than a set threshold; input the second simulation key dimensions into the forward analysis model to obtain a second simulation characteristic signal corresponding to the second simulation key dimensions; calculate the loss value of the second simulation characteristic signal and the measured characteristic signal until the difference between the loss values ​​is less than the set threshold.

[0026] In a possible design, before comparing the measured characteristic signal with the simulated characteristic signal, a preprocessing module is used to generate multiple groups of simulated key dimensions within the variation range of the key dimensions of the sample to be measured; using the measurement system as a reference, a simulated scattering diagram of each group of simulated key dimensions under different positional relationships is generated; and by analyzing the distribution characteristics of the scattering signals in the simulated scattering diagram corresponding to each group of simulated key dimensions, as well as the correlation between each group of simulated key dimensions and the corresponding simulated scattering diagram, the simulated characteristic signal representing the key dimensions in the simulated scattering diagram is extracted.

[0027] In one possible design, the preprocessing module is also used to obtain a feature extraction model based on the method of extracting simulated characteristic signals that characterize key dimensions in the simulated scattering map; the feature extraction model is a model that performs feature extraction functions under different positional relationships, or, there are multiple feature extraction models, each of which is a model that performs feature extraction functions under a single positional relationship; the processing module, when parsing the distribution characteristics of the scattering signal in the measured scattering map of the sample to be tested, is specifically used to parse the distribution characteristics of the scattering signal in the measured scattering map of the sample to be tested through the feature extraction model.

[0028] In one possible design, a training module is used to take each group of simulated critical dimensions as input and take characteristic signals characterizing the critical dimensions as output to train the forward analysis model.

[0029] In one possible design, when the preprocessing module generates multiple groups of simulated key dimensions within the variation range of the key dimensions of the sample to be tested, it is specifically used to generate M groups of simulated key dimensions within the variation range of the key dimensions of the sample to be tested; wherein the number of M is positively correlated with the number of parameter dimensions of the sample to be tested and / or the size of the parameter range.

[0030] In a third aspect, an embodiment of the present application further provides a computing device, including:

[0031] A memory for storing program instructions;

[0032] A processor is used to call the program instructions stored in the memory and execute the method described in any possible design of the first aspect according to the obtained program instructions.

[0033] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, in which computer-readable instructions are stored. When a computer reads and executes the computer-readable instructions, the method described in any possible design of the above-mentioned first aspect is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 A schematic diagram of a process for generating a sample data set provided in an embodiment of the present application;

[0036] Figure 2A flowchart of a training process for a forward model provided in an embodiment of the present application;

[0037] Figure 3 A schematic diagram of a flow chart of a method for determining a critical dimension of a sample provided in an embodiment of the present application;

[0038] Figure 4 A schematic diagram of a process for generating a set of first simulation critical dimensions provided in an embodiment of the present application;

[0039] Figure 5 A schematic diagram of the structure of a sample to be tested provided in an embodiment of the present application;

[0040] Figure 6 A measured scatter diagram of a sample to be tested provided in an embodiment of the present application;

[0041] Figure 7 A schematic diagram of signal distribution after feature extraction provided in an embodiment of the present application;

[0042] Figure 8 A schematic diagram of the measurement results provided in the embodiments of the present application;

[0043] Fig. 9 A schematic diagram of the relationship between the loss function provided in the embodiment of the present application and the iteration time;

[0044] Fig.10 A schematic diagram of the structure of a device for determining a critical dimension of a sample provided in an embodiment of the present application;

[0045] Fig.11 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0047] In the embodiments of the present application, a plurality refers to two or more than two. The words "first", "second" and the like are only used for the purpose of distinguishing descriptions and cannot be understood as indicating or implying relative importance or order.

[0048] Before measuring the critical dimensions of the sample, preprocessing is performed, which includes: generating a sample data set, determining a feature extraction model, and training a forward analysis model.

[0049] Figure 1 A schematic diagram of a process for generating a sample data set provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the following steps are included:

[0050] Step 101 : generating multiple groups of simulated critical dimensions within a variation range of the critical dimensions of the sample to be tested.

[0051] Specifically, within the variation range of the key dimensions of the sample to be tested, M groups of simulated key dimensions are generated by the simulation program. The number of M is positively correlated with the number of parameter dimensions and / or the size of the parameter range of the sample to be tested. That is, the more parameter dimensions of the sample to be tested and the larger the parameter range, the more simulated key dimensions need to be generated.

[0052] Step 102: Using the measurement system as a reference, generate simulation scatter diagrams of each group of simulated key dimensions under different position relationships.

[0053] The measurement system includes a sample, a detector and an X-ray light source. The positional relationship is the azimuth and incident angle of the X-ray and the sample to be measured. According to the hardware configuration of the measurement system (such as the distance between the detector and the sample, the size of the light source, the size of the detector and other parameters), as well as the azimuth and incident angle when the sample to be measured is actually measured, the simulation calculation obtains the simulation scattering diagram of each group of simulated key dimensions under different positional relationships. Among them, a group of simulated key dimensions can use a simulation program to obtain the simulation scattering diagram of the group of simulated key dimensions under a group of different azimuths (greater than 1) and incident angles (greater than 10) at the corresponding different azimuths and incident angles.

[0054] Step 103: extracting simulation characteristic signals representing the key dimensions in the simulation scattering diagram by analyzing the distribution characteristics of the scattering signals in the simulation scattering diagram corresponding to each group of simulation key dimensions and the correlation between each group of simulation key dimensions and the corresponding simulation scattering diagram.

[0055] The feature extraction methods for extracting the simulation characteristic signals representing the key dimensions in the simulation scattering diagram include: finding the maximum value, non maximum suppression (NMS), Fourier transform, principal component analysis (PCA), independent component analysis (ICA), correlation analysis, autoencoder network and other methods. Through feature extraction, the characteristic signals most sensitive to the structural transformation of the sample to be tested can be obtained as the simulation characteristic signals representing the key dimensions.

[0056] By extracting the simulation characteristic signals representing the key dimensions in the simulation scattering diagram corresponding to each group of simulation key dimensions, a simulation measurement library with one-to-one correspondence between each group of simulation key dimensions and the simulation characteristic signals is formed. Based on the above method of extracting the simulation characteristic signals representing the key dimensions in the simulation scattering diagram, a feature extraction model is obtained.

[0057] Among them, the feature extraction model is a model that performs feature extraction functions under different position relationships.

[0058] Alternatively, there are multiple feature extraction models, each of which is a model that performs feature extraction under a single position relationship. That is to say, in the above feature extraction process, all incident angle and azimuth angle data can be uniformly analyzed for features to form a unified feature extraction model, or different incident angles and azimuth angles can be analyzed for features separately, and each incident angle and azimuth angle corresponds to a feature extraction model.

[0059] Furthermore, after a simulation measurement library in which each set of simulated critical dimensions corresponds to a simulation characteristic signal one by one, each set of simulated critical dimensions is used as input and the characteristic signal representing the critical dimensions is used as output to train a forward analysis model.

[0060] Among them, the forward analysis model can be traditional machine learning (support vector machine, random forest, decision tree, etc.), neural network (fully connected, convolutional network, recurrent neural network, etc.) and deep neural network (residual network,) and other machine learning models.

[0061] It should be noted that if the feature extraction model is a model that performs feature extraction functions under different positional relationships (i.e., different incident angles and different azimuths), then the forward analysis model should also be a model that can obtain characteristic signals characterizing the key dimensions from the key dimensions under different positional relationships; if there are multiple feature extraction models, and each feature extraction model is a model that performs feature extraction functions under a single positional relationship (i.e., one incident angle and one azimuth), then there should also be multiple forward analysis models, and each forward analysis model can obtain characteristic signals characterizing the key dimensions from the key dimensions under a single positional relationship.

[0062] Specifically, the training process of the forward model refers to Figure 2, input the simulated key dimension into the forward analysis model, output its corresponding simulation feature signal, compare the loss value of the feature signal representing the key dimension after feature extraction with the simulation feature signal output by the forward analysis model, if the loss value reaches the standard, terminate the training and save the forward analysis model at this time; if the loss value does not meet the standard, adjust the relevant parameters of the forward analysis model, output the simulation feature signal corresponding to the simulated key dimension again through the adjusted forward analysis model, compare the loss value of the feature signal representing the key dimension after feature extraction with the simulation feature signal output by the forward analysis model, until the loss value reaches the standard.

[0063] Figure 3 A schematic diagram of a method for determining a critical dimension of a sample provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, the method comprises the following steps:

[0064] Step 301: Obtain a measurement scatter diagram of a sample to be measured through a measurement system.

[0065] The measurement system includes a sample, a detector and an X-ray light source. The measurement scatter diagram is a scatter diagram formed after the X-rays of the measurement system are scattered by the sample to be measured under different positional relationships, wherein the positional relationship is the azimuth angle and the incident angle of the X-rays and the sample to be measured.

[0066] In the above positional relationship, the number of azimuth angles is less than the number of incident angles. That is to say, when only one azimuth angle of the sample to be tested is collected, it is necessary to collect measurement data of multiple incident angles at the azimuth angle; when multiple azimuth angles of the sample to be tested are collected, it is necessary to collect measurement data of multiple incident angles at each azimuth angle.

[0067] Step 302: Analyze the distribution characteristics of the scattering signal in the measurement scattering diagram of the sample to be tested to obtain a measurement characteristic signal representing a critical dimension of the sample to be tested.

[0068] Specifically, the distribution characteristics of the scattering signal in the measured scattering map of the sample to be tested can be analyzed by the feature extraction model obtained by preprocessing.

[0069] Step 303: Compare the measured characteristic signal with the simulated characteristic signal to determine the simulated characteristic signal having the smallest difference with the measured characteristic signal.

[0070] Specifically, it can be done by Figure 4 The method shown compares the measured characteristic signal with the simulated characteristic signal and determines the simulated characteristic signal having the smallest difference from the measured characteristic signal.

[0071] Step 401: Generate a set of first simulated critical dimensions within a variation range of the critical dimensions of the sample to be tested.

[0072] Step 402: Input the first simulation critical dimension into the forward analysis model to obtain a first simulation characteristic signal corresponding to the first simulation critical dimension.

[0073] Step 403: Calculate the loss value of the first simulation characteristic signal and the measured characteristic signal.

[0074] Step 404: Determine whether the loss value is less than or equal to a set threshold.

[0075] If yes, execute step 405; otherwise, execute step 406.

[0076] Step 405: If the loss value is less than or equal to the set threshold, the first simulated characteristic signal is used as the simulated characteristic signal with the smallest difference from the measured characteristic signal.

[0077] Step 406: Adjust the value of the first simulation critical dimension.

[0078] If the loss value is greater than the set threshold, the value of the first simulated critical dimension is adjusted according to the loss value to obtain a set of second simulated critical dimensions. The second simulated critical dimensions are input into the forward analysis model to obtain a second simulated characteristic signal corresponding to the second simulated critical dimensions. The loss value of the second simulated characteristic signal and the measured characteristic signal is calculated until the difference in the loss value is less than the set threshold.

[0079] Step 304: Use a set of simulated critical dimensions corresponding to the simulated characteristic signals with the smallest difference from the measured characteristic signals as the critical dimensions of the sample to be tested.

[0080] Finally, a set of simulated critical dimensions corresponding to the simulated characteristic signals having the smallest difference from the measured characteristic signals can be used as the critical dimensions of the sample.

[0081] The accuracy and calculation speed of the method for determining the critical dimensions of a sample provided in the embodiment of the present application are verified below. The structure of the sample to be measured is as follows: Figure 5 As shown, top view Figure 5-1 is a six-fold symmetric periodic array of circular holes. Figure 5-2 is the x-direction cross-sectional view of a hole, Figure 5-3 This is a cross-sectional view of a hole in the y direction. The key dimensions that need to be measured are the cross-sectional width of the circular hole in the x and y directions at different heights.

[0082] The measured scattering diagrams at three different incident angles obtained by the measurement system are shown in Figure 2. Figure 6 As shown in the figure, the number and position of bright spots in the figure contain the periodic and structural distribution information of the sample to be tested, and the brightness contains the shape information of the sample to be tested. The size of the sample to be tested can be analyzed through the information represented by each signal in the measured scattering diagram at different incident angles. The signal distribution after feature extraction of the measured scattering diagram is shown in the figure. Figure 7 shown. Figure 7 The figure includes the first 2000 most obvious characteristic measurement signal points (points in the figure) found after feature extraction of the measurement scatter diagram and the measurement characteristic signal after fitting (curve in the figure).

[0083] The simulated signal that is found by the optimization algorithm to have the smallest difference from the measured signal has smaller differences in key features.

[0084] The measurement results obtained by using the method for determining the critical dimensions of a sample provided in the embodiment of the present application are as follows: Figure 8 As shown, Figure 8-1 is the width distribution of the x and y sections at different heights, Figure 8-2 The comparison results of the current method and the existing method are shown in Table 1. By comparing the two methods, it is found that the measurement method provided by the embodiment of the present application does not reduce the measurement accuracy, and the measurement results remain the same as those of the existing method.

[0085] like Fig. 9 The relationship between the loss function and the iteration time of the two methods in the iteration process is shown. Figure 9-1 The existing method takes about 1500 seconds to measure the sample. Figure 9-2 After using the method for determining the critical dimensions of a sample provided in an embodiment of the present application, it only takes about 100 seconds to obtain the measurement result. The measurement method provided in an embodiment of the present application can obtain the same accuracy of the measurement result as the existing method, but significantly reduces the time spent on solving the critical dimensions. At the same time, the measurement method provided in an embodiment of the present application is more than 10 times faster than the existing method after using 2 GPUs for acceleration when only one GPU hardware configuration is used.

[0086] Based on the same technical concept, Fig.10 The schematic diagram of the structure of a device for determining the critical dimension of a sample provided in an embodiment of the present application is shown as an example. Fig.10 As shown, the device 1000 includes:

[0087] The acquisition module 1001 is used to acquire a measurement scattering diagram of the sample to be tested through a measurement system; the measurement scattering diagram is a scattering diagram formed after the X-rays of the measurement system are scattered by the sample to be tested under different positional relationships; the positional relationship is the azimuth angle and the incident angle of the X-rays and the sample to be tested;

[0088] A processing module 1002 is used to analyze the distribution characteristics of the scattering signal in the measurement scattering diagram of the sample to be tested, and obtain a measurement characteristic signal representing the critical dimension of the sample to be tested;

[0089] The comparison module 1003 is used to compare the measured characteristic signal with the simulated characteristic signal to determine the simulated characteristic signal with the smallest difference from the measured characteristic signal; and use a set of simulated critical dimensions corresponding to the simulated characteristic signal with the smallest difference from the measured characteristic signal as the critical dimensions of the sample to be tested.

[0090] In one possible design, the comparison module 1003, when comparing the measured characteristic signal with the simulated characteristic signal to determine the simulated characteristic signal with the smallest difference from the measured characteristic signal, is specifically used to generate a set of first simulated critical dimensions within the variation range of the critical dimensions of the sample to be tested; input the first simulated critical dimensions into the forward analysis model to obtain the first simulated characteristic signal corresponding to the first simulated critical dimensions; calculate the loss value between the first simulated characteristic signal and the measured characteristic signal; if the loss value is less than a set threshold, use the first simulated characteristic signal as the simulated characteristic signal with the smallest difference from the measured characteristic signal.

[0091] In one possible design, the comparison module 1003 is also used to adjust the value of the first simulation key dimension according to the loss value to obtain a set of second simulation key dimensions if the loss value is greater than a set threshold; input the second simulation key dimensions into the forward analysis model to obtain a second simulation characteristic signal corresponding to the second simulation key dimensions; calculate the loss value of the second simulation characteristic signal and the measured characteristic signal until the difference between the loss values ​​is less than the set threshold.

[0092] In a possible design, before comparing the measured characteristic signal with the simulated characteristic signal, the preprocessing module 1004 is used to generate multiple groups of simulated key dimensions within the variation range of the key dimensions of the sample to be measured; using the measurement system as a reference, generate simulated scattering diagrams of each group of simulated key dimensions under different positional relationships; and extract the simulated characteristic signal representing the key dimensions in the simulated scattering diagram by analyzing the distribution characteristics of the scattering signals in the simulated scattering diagram corresponding to each group of simulated key dimensions and the correlation between each group of simulated key dimensions and the corresponding simulated scattering diagram.

[0093] In one possible design, the preprocessing module 1004 is also used to obtain a feature extraction model based on the method of extracting simulated characteristic signals that characterize key dimensions in the simulated scattering map; the feature extraction model is a model that performs feature extraction functions under different positional relationships, or, the feature extraction models are multiple, each feature extraction model is a model that performs feature extraction functions under a single positional relationship; the processing module 1002, when parsing the distribution characteristics of the scattering signal in the measured scattering map of the sample to be tested, is specifically used to parse the distribution characteristics of the scattering signal in the measured scattering map of the sample to be tested through the feature extraction model.

[0094] In one possible design, the training module 1005 is used to take each group of simulated critical dimensions as input and take characteristic signals representing the critical dimensions as output to train the forward analysis model.

[0095] In one possible design, when the preprocessing module 1004 generates multiple groups of simulated key dimensions within the variation range of the key dimensions of the sample to be tested, it is specifically used to generate M groups of simulated key dimensions within the variation range of the key dimensions of the sample to be tested; wherein the number of M is positively correlated with the number of parameter dimensions of the sample to be tested and / or the size of the parameter range.

[0096] Based on the same technical concept, the present application embodiment provides a computing device, such as Fig.11 As shown, it includes at least one processor 1101 and a memory 1102 connected to the at least one processor. The specific connection medium between the processor 1101 and the memory 1102 is not limited in the embodiment of the present application. Fig.11 For example, the processor 1101 and the memory 1102 are connected via a bus. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0097] In the embodiment of the present application, the memory 1102 stores instructions that can be executed by at least one processor 1101. The at least one processor 1101 can execute the above-mentioned method for determining the critical size of the sample by executing the instructions stored in the memory 1102.

[0098] Among them, the processor 1101 is the control center of the computing device, which can use various interfaces and lines to connect various parts of the computer device, and perform resource settings by running or executing instructions stored in the memory 1102 and calling data stored in the memory 1102.

[0099] Optionally, the processor 1101 may include one or more processing units, and the processor 1101 may integrate an application processor and a modem processor, wherein the application processor mainly processes an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the modem processor may not be integrated into the processor 1101. In some embodiments, the processor 1101 and the memory 1102 may be implemented on the same chip, and in some embodiments, they may also be implemented separately on separate chips.

[0100] Processor 1101 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware processor for execution, or can be executed by a combination of hardware and software modules in the processor.

[0101] Memory 1102, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules. Memory 1102 may include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory, random access memory (Random Access Memory, RAM), static random access memory (Static Random Access Memory, SRAM), programmable read-only memory (Programmable Read Only Memory, PROM), read-only memory (Read Only Memory, ROM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), magnetic memory, disk, optical disk, etc. Memory 1102 is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 1102 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0102] Based on the same technical concept, an embodiment of the present application also provides a computer-readable storage medium, which stores a computer-executable program. The computer-executable program is used to enable a computer to execute any of the methods listed above to determine the critical dimensions of a sample.

[0103] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0104] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0105] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0106] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0107] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0108] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for determining a critical dimension of a sample, characterized in that: The method comprises: Obtaining a measurement scattering diagram of the sample to be measured by the measurement system; the measurement scattering diagram is a scattering diagram formed after the X-ray of the measurement system is scattered by the sample to be measured at different positional relationships; the positional relationship is the azimuth angle and the incident angle of the X-ray and the sample to be measured; Analyzing the distribution characteristics of the scattering signal in the measurement scattering diagram of the sample to be tested to obtain a measurement characteristic signal characterizing the critical dimension of the sample to be tested; Generate a simulated critical dimension within the variation range of the critical dimension of the sample to be tested, and then determine a simulated characteristic signal corresponding to the simulated critical dimension, compare the measured characteristic signal with the simulated characteristic signal, and determine the simulated characteristic signal having the smallest difference from the measured characteristic signal; A set of simulated critical dimensions corresponding to the simulated characteristic signals having the smallest difference from the measured characteristic signals is used as the critical dimensions of the sample to be tested.

2. The method according to claim 1, characterized in that Generate a simulated critical dimension within a variation range of the critical dimension of the sample to be tested, and then determine a simulated characteristic signal corresponding to the simulated critical dimension, compare the measured characteristic signal with the simulated characteristic signal, and determine a simulated characteristic signal having the smallest difference from the measured characteristic signal, including: generating a set of first simulated critical dimensions within a variation range of the critical dimensions of the sample to be tested; Inputting the first simulation critical dimension into a forward analysis model to obtain a first simulation characteristic signal corresponding to the first simulation critical dimension; Calculating a loss value between the first simulation characteristic signal and the measured characteristic signal; If the loss value is less than a set threshold, the first simulated characteristic signal is used as the simulated characteristic signal having the smallest difference with the measured characteristic signal.

3. The method according to claim 2, characterized in that The method further comprises: If the loss value is greater than a set threshold, adjusting the value of the first simulated critical dimension according to the loss value to obtain a set of second simulated critical dimensions; Inputting the second simulation critical dimension into the forward analysis model to obtain a second simulation characteristic signal corresponding to the second simulation critical dimension; The loss values ​​of the second simulation characteristic signal and the measured characteristic signal are calculated until the difference between the loss values ​​is less than a set threshold.

4. The method according to any one of claims 1 to 3, characterized in that: Before comparing the measured characteristic signal with the simulated characteristic signal, the method further includes: generating multiple sets of simulated critical dimensions within a variation range of the critical dimensions of the sample to be tested; Using the measurement system as a reference, generating simulation scatter diagrams of each group of simulated critical dimensions under different positional relationships; By analyzing the distribution characteristics of the scattering signals in the simulation scattering diagram corresponding to each group of simulated critical dimensions and the correlation between each group of simulated critical dimensions and the corresponding simulation scattering diagram, the simulation characteristic signals representing the critical dimensions in the simulation scattering diagram are extracted.

5. The method according to claim 4, characterized in that The method further comprises: Based on the method of extracting the simulation characteristic signal representing the key size in the simulation scattering map, a feature extraction model is obtained; the feature extraction model is a model that performs feature extraction functions under different positional relationships, or the feature extraction model is multiple, each feature extraction model is a model that performs feature extraction functions under a single positional relationship; Analyzing the distribution characteristics of the scattering signal in the measured scattering diagram of the sample to be tested, including: The distribution characteristics of the scattering signal in the measured scattering map of the sample to be tested are analyzed by the feature extraction model.

6. The method according to claim 4, characterized in that The method further comprises: Each set of simulated critical dimensions is taken as input, and the characteristic signals representing the critical dimensions are taken as output to train the forward analysis model.

7. The method according to claim 4, characterized in that Generate multiple sets of simulated critical dimensions within the range of critical dimensions of the sample to be tested, including: Within the variation range of the key dimensions of the sample to be tested, M groups of simulated key dimensions are generated; wherein the number of M is positively correlated with the number of parameter dimensions and / or the size of the parameter range of the sample to be tested.

8. A device for determining a critical dimension of a sample, characterized in that: include: An acquisition module is used to acquire a measurement scattering diagram of the sample to be tested through a measurement system; the measurement scattering diagram is a scattering diagram formed after the X-rays of the measurement system are scattered by the sample to be tested under different positional relationships; the positional relationship is the azimuth angle and the incident angle of the X-rays and the sample to be tested; A processing module, used for analyzing the distribution characteristics of the scattering signal in the measurement scattering diagram of the sample to be tested, and obtaining a measurement characteristic signal characterizing the critical dimension of the sample to be tested; A comparison module is used to generate simulated critical dimensions within the variation range of the critical dimensions of the sample to be tested, and then determine a simulation characteristic signal corresponding to the simulated critical dimensions, compare the measured characteristic signal with the simulation characteristic signal, and determine the simulation characteristic signal with the smallest difference from the measured characteristic signal; and use a group of simulated critical dimensions corresponding to the simulation characteristic signal with the smallest difference from the measured characteristic signal as the critical dimensions of the sample to be tested.

9. A computing device, characterized in that include: A memory for storing program instructions; A processor is used to call the program instructions stored in the memory, and execute the method according to any one of claims 1 to 7 according to the obtained program instructions.

10. A computer-readable storage medium, characterized in that: The method comprises computer-readable instructions, and when a computer reads and executes the computer-readable instructions, the method according to any one of claims 1 to 7 is implemented.

Images