Photoetching defect detection method and device based on deep learning

Abstract

The invention belongs to the technical field of artificial intelligence and optical detection, and particularly discloses a photoetching defect detection method and device based on deep learning, and the method comprises the steps: scanning each defective photoetching sample through a coherent Fourier scattering method, and obtaining a far-field differential data set, the far-field differential data set comprises defect label data and far-field differential data corresponding to each scanning point in each defective photoetching sample; performing asymmetric analysis based on the far-field differential data set to obtain left-right asymmetric information, up-down asymmetric information and differential summation information corresponding to each scanning point in each defective photoetching sample; and training a deep learning model by using the three-channel data. According to the method, after the diffraction information of the far field is collected, the three-channel data is obtained by using the differential asymmetric analysis method and serves as the input of the deep learning model, the deep learning model can automatically learn the characteristics of defect signals, the robustness of the system is improved, and photoetching defect detection through the neural network is realized.

Description

Technical Field

[0001] The present application belongs to the field of artificial intelligence and optical detection technology, and more specifically, to a lithography defect detection method and device based on deep learning. Background Art

[0002] Lithography defect detection has always been a key link in the semiconductor chip manufacturing process. With the continuous advancement of integrated circuit related manufacturing, the feature size on wafers and masks is shrinking, and the requirements for detection capabilities are also increasing. Lithography defects usually refer to pattern distortion or loss caused by various reasons during the lithography process. These defects will affect the subsequent etching, deposition and electrical performance, and reduce the yield and reliability of the device. Coherent Fourier Scatterometry (CFS) is a scanning-based scattering measurement technology. It was originally used to measure the shape parameters of periodic gratings with high precision. In recent years, it has been introduced to the detection of nanostructured particles in various bright field illumination configurations. A monochromatic light source with high temporal and spatial coherence is focused on the surface of the lithography sample to form an illumination spot, covering the entire plane in a multi-line scanning manner (there are multiple scanning points on each scanning line, and there is a certain overlap between the spots corresponding to each scanning point). In this way, the reflected light from the sample surface and the scattered light of the defect are coherently superimposed in the far field, and the far-field image is captured on the Fourier plane using a quadrant detector or CCD camera, and is differentiated from the far-field image of the defect-free sample. When a sample defect passes through the focused spot during the scanning process, it will cause an uneven distribution of far-field intensity, that is, the asymmetry of the differential signal. At the same time, it will be observed that the intensity and positive and negative signs of the asymmetric signal will change relative to the change of the position of the defect in the spot. Based on this property, the specific position of the defect on the lithography sample can be determined.

[0003] The diffraction measurement data of defective lithography wafers and masks is a "large" and "high-dimensional" big data. In addition to the scattering signal of the defect, the data also contains interference signals such as instrument noise, background noise, and readout noise, which leads to the appearance of indistinguishable erroneous signals in the far-field image and causes false detection. When the defect size is reduced to a certain level (sub-wavelength or sub-deep wavelength scale), the scattering signal caused by the defect is easily submerged in the background noise and is also prone to missed detection. In order to overcome the above problems, a more accurate and robust defect detection method is needed. In recent years, deep learning has been a hot research field and has been widely used in target detection, image classification and other fields. By reasonably constructing a neural network model, features can be extracted from the data autonomously, which is more robust than manually extracted features and has good high-dimensional data processing capabilities. The trained neural network can quickly detect data in the inference stage and can meet the real-time detection needs in industrial production. How to realize lithography defect detection through neural networks is a technical problem that needs to be solved in this field. Summary of the invention

[0004] In view of the defects of the prior art, the purpose of this application is to realize lithography defect detection through neural network.

[0005] To achieve the above objectives, in a first aspect, the present application provides a lithography defect detection method based on deep learning, comprising:

[0006] Scan each defective lithography sample by a coherent Fourier scattering method to obtain a far-field differential data set, wherein the far-field differential data set includes defect label data corresponding to each defective lithography sample and far-field differential data corresponding to each scanning point in each defective lithography sample, and the far-field differential data is obtained by differentiating the diffraction light intensity distribution of the far field corresponding to the defective lithography sample and the diffraction light intensity distribution of the far field corresponding to the defect-free lithography sample with respect to the scanning point;

[0007] Based on the far-field differential data set, an asymmetric analysis is performed on the far-field differential data corresponding to each scanning point to obtain left-right asymmetric information, top-bottom asymmetric information and differential summation information corresponding to each scanning point in each defective lithography sample;

[0008] Based on the defect label data corresponding to each defective lithography sample and the left-right asymmetric information, top-bottom asymmetric information and differential summation information corresponding to each scanning point in each defective lithography sample, a deep learning model is trained, and the deep learning model is used to detect lithography defects in lithography samples.

[0009] In a possible implementation, the far-field differential data corresponding to the scanning point is obtained by the following steps:

[0010] At the scanning point, calculate the near-field electric field distribution after the illumination probe interacts with the lithography sample;

[0011] Based on the near-field electric field distribution, the far-field diffraction light intensity distribution corresponding to the defective lithography sample and the far-field diffraction light intensity distribution corresponding to the defect-free lithography sample are obtained through the Fraunhofer far-field propagation model;

[0012] Based on the diffraction light intensity distribution of the far field corresponding to the defective lithography sample and the diffraction light intensity distribution of the far field corresponding to the defect-free lithography sample, a difference is made to obtain far-field differential data corresponding to the scanning point.

[0013] In a possible implementation, at the scanning point, the near-field electric field distribution after the illumination probe interacts with the lithography sample is calculated, including calculating the near-field electric field distribution by the following formula:

[0014] U(x,y)=P(λ,ξ,η)S(ξ,η);

[0015] Among them, U(x,y) represents the near-field electric field distribution, (x,y) represents the coordinates in the near-field plane coordinate system, P(λ,ξ,η) represents the input illumination matrix, (ξ,η) represents the coordinates in the sample coordinate system, λ represents the illumination wavelength, and S(ξ,η) represents the sample complex refractive index function.

[0016] For a defective lithographic sample, the complex refractive index function is expressed as:

[0017] S(ξ,η)=S free (ξ,η)+D(ξ k , η k );

[0018] Where S free (ξ,η) represents the complex refractive index function of the defect-free lithography sample, D(ξ k ,η k ) represents the complex refractive index function of the defect, and the coordinates of the defect center are (ξ k ,η k ).

[0019] For a defect-free lithographic sample, the complex refractive index function is expressed as:

[0020] S(ξ,η)=S free (ξ,η).

[0021] In a possible implementation, the deep learning model includes a cascaded feature extraction module, a position prediction module, and a category prediction module, and accordingly, the defect label data includes a defect position label and a defect size category label;

[0022] The feature extraction module is used to extract defect features corresponding to the lithography sample based on left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to the scanning point;

[0023] The position prediction module is used to predict the defect position based on the defect characteristics corresponding to the lithography sample;

[0024] The category prediction module is used to predict the defect category based on the prediction information of the defect location and the defect characteristics corresponding to the lithography sample.

[0025] In one possible implementation, the loss function used to train the deep learning model is:

[0026] L=αL P +(1-α)L c ;

[0027] Among them, L represents the total loss, α represents the weighting coefficient, and L P represents the position loss, L c represents the class loss.

[0028] In one possible implementation, the position loss is calculated by the following formula:

[0029]

[0030] Among them, (ξ i ,η i ) represents the defect location predicted by the deep learning model, (ξ t ,η t ) represents the defect location indicated by the defect location label, n batch Represents the number of samples in a training batch. A sample includes the left-right asymmetric information, top-bottom asymmetric information, and differential summation information corresponding to each scanning point in a defective lithography sample.

[0031] In one possible implementation, the class loss is calculated using the following formula:

[0032]

[0033] Among them, y k represents the category predicted by the deep learning model, t k Indicates the category indicated by the defect size category label.

[0034] In a second aspect, the present application provides a lithography defect detection device based on deep learning, comprising:

[0035] A data set construction module is used to scan each defective lithography sample by a coherent Fourier scattering method to obtain a far-field differential data set, wherein the far-field differential data set includes defect label data corresponding to each defective lithography sample and far-field differential data corresponding to each scanning point in each defective lithography sample. The far-field differential data is obtained by differentiating the diffraction light intensity distribution of the far field corresponding to the defective lithography sample and the diffraction light intensity distribution of the far field corresponding to the defect-free lithography sample for the scanning point;

[0036] A data set preprocessing module is used to perform asymmetry analysis on the far-field differential data corresponding to each scanning point based on the far-field differential data set, and obtain left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to each scanning point in each defective lithography sample;

[0037] The training module trains a deep learning model based on the defect label data corresponding to each defective lithography sample and the left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to each scanning point in each defective lithography sample. The deep learning model is used to detect lithography defects in lithography samples.

[0038] In a third aspect, the present application provides an electronic device comprising: at least one memory for storing programs; and at least one processor for executing the programs stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method described in the first aspect or any possible implementation of the first aspect.

[0039] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method described in the first aspect or any possible implementation of the first aspect.

[0040] In general, the above technical solutions conceived by this application have the following beneficial effects compared with the prior art:

[0041] (1) Deep learning is introduced into lithography defect detection based on coherent Fourier scattering (CFS). Defects of different positions and sizes are set on the lithography sample for multiple scans to obtain a sufficient data set. After collecting the far-field diffraction information, the differential asymmetric analysis method is used to obtain three-channel data as the input of the neural network. The neural network model will automatically learn the characteristics of the defect signal without manual extraction, thereby improving the robustness of the detection system.

[0042] (2) The identification of lithography defects is mainly reflected in two aspects: the determination of defect location and the identification of defect type. By setting two types of loss functions L in the network P and L cThe position prediction loss and category prediction loss are penalized separately, and the two loss functions are summed at a certain weighted ratio. After sufficient data training, the neural network can process a large amount of scan data at the same time, and has a faster processing speed to achieve more efficient detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is one of the flow charts of the deep learning-based lithography defect detection method provided in the embodiment of the present application;

[0044] Figure 2 This is the second flow chart of the deep learning-based lithography defect detection method provided in the embodiment of the present application;

[0045] Figure 3 is a schematic diagram of a photolithography pattern and a defect provided in an embodiment of the present application;

[0046] Figure 4 is a schematic diagram of a serpentine scanning path provided in an embodiment of the present application;

[0047] Figure 5 is a schematic diagram of a transmissive and reflective lithography defect detection system provided in an embodiment of the present application;

[0048] Figure 6 is a schematic diagram of left-right asymmetric information of each scanning point obtained from a complete scan provided by an embodiment of the present application;

[0049] Figure 7 is a schematic diagram of the up-down asymmetric information of each scanning point obtained from a complete scan provided by an embodiment of the present application;

[0050] Figure 8 is a schematic diagram of differential sum information of each scanning point obtained from a complete scan provided by an embodiment of the present application;

[0051] Fig. 9 is a schematic diagram of the structure of a lithography defect recognition network provided in an embodiment of the present application;

[0052] Fig.10 is a structural schematic diagram of a lithography defect detection device based on deep learning provided in an embodiment of the present application;

[0053] Fig.11 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0054] Throughout the drawings, the same reference numerals are used to denote the same elements or structures, wherein:

[0055] 1: He-Ne laser; 2: Attenuator; 3: Adjustable aperture; 4: Beam expander; 5: Reflector; 6: Adjustable aperture; 7: Focusing lens; 8: Lithography sample; 9: Camera; 10: Computer. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0057] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0058] In the description of the embodiments of the present application, unless otherwise specified, "multiple" means two or more than two. For example, multiple processing units refer to two or more processing units, etc.; multiple elements refer to two or more elements, etc.

[0059] First, the technical terms involved in the embodiments of the present application are introduced.

[0060] (1) Coherent Fourier Scattering (CFS) is a technique that uses a coherent light source to perform scattering experiments and analyzes the scattered signals through Fourier transform. This method is widely used in physics, chemistry, biology, and materials science to study the microstructure of matter and its dynamic processes.

[0061] (2) The Fraunhofer diffraction model is an important theoretical model that describes the propagation and diffraction of light waves far away from the light source or optical element. This model is usually used to analyze the diffraction pattern of light waves in the far field after passing through slits, optical elements or other obstacles.

[0062] (3) The Softmax-with-loss layer is a commonly used layer in deep learning, usually for multi-classification problems. It combines the softmax activation function and the cross-entropy loss function, and can simultaneously calculate the output probability distribution and loss value of the model in one forward propagation.

[0063] The embodiments of the present application are described below in conjunction with the drawings in the embodiments of the present application.

[0064] Figure 1 is one of the flow charts of the deep learning-based lithography defect detection method provided in the embodiment of the present application, such as Figure 1 As shown, the method can be applied to a coherent Fourier scattering detection system, and the method includes the following steps S101 to S103.

[0065] Step S101, scanning each defective lithography sample by a coherent Fourier scattering method to obtain a far-field differential data set, wherein the far-field differential data set includes defect label data corresponding to each defective lithography sample and far-field differential data corresponding to each scanning point in each defective lithography sample, and the far-field differential data is obtained by differentiating the diffraction light intensity distribution of the far field corresponding to the defective lithography sample and the diffraction light intensity distribution of the far field corresponding to the defect-free lithography sample for the scanning point;

[0066] Step S102, based on the far-field differential data set, performing asymmetry analysis on the far-field differential data corresponding to each scanning point, and obtaining left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to each scanning point in each defective lithography sample;

[0067] Step S103, based on the defect label data corresponding to each defective lithography sample and the left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to each scanning point in each defective lithography sample, a deep learning model is trained, and the deep learning model is used to detect lithography defects in lithography samples.

[0068] It is understandable that the present application introduces deep learning into lithography defect detection based on coherent Fourier scattering (CFS). Defects of different positions and sizes are set on the lithography sample for multiple scans to obtain a sufficient data set. After collecting the far-field diffraction information, the differential asymmetric analysis method is used to obtain three-channel data (left-right asymmetric information, top-bottom asymmetric information, and differential summation information) as the input of the deep learning model. The deep learning model can automatically learn the characteristics of the defect signal without manual extraction, thereby improving the robustness of the detection system. The trained deep learning model can be used to accurately detect lithography defects in lithography samples and realize lithography defect detection through neural networks.

[0069] Therefore, the deep learning-based lithography defect detection method provided in this application has the advantages of no need for manual feature extraction, fast and accurate positioning and classification, and effectively solves the problems of complex feature extraction and insufficient robustness of existing lithography defect detection methods, and can be applied to mask and wafer defect detection and other fields.

[0070] Figure 2 This is the second flow chart of the deep learning-based lithography defect detection method provided in the embodiment of the present application, such as Figure 2 As shown, the process includes the following steps S201 to S205.

[0071] Step S201, obtaining a far-field differential data set that completely scans the entire lithography sample. This step specifically includes:

[0072] S201.1 designs a lithography sample model S(ξ,η) with any pattern type, adds a defect at any position on the sample pattern as a defective lithography sample, and records the coordinates of the defect center position on the sample pattern (ξ1,η1) and the defect characteristic length D, such as Figure 3 As shown, Figure 3 The value range of k in is 1≤k≤n;

[0073] S201.2 constructs the input illumination matrix P(λ,ξ,η) and calculates the near-field electric field distribution U(x,y) after the illumination probe interacts with the sample. The calculation method is as follows:

[0074] U(x,y)=P(λ,ξ,η)S(ξ,η);

[0075] Where λ represents the illumination wavelength, (x, y) is the near-field plane coordinate system at a small distance from the sample, and the sample surface is regarded as a function S(ξ,η). S(ξ,η) can be the complex refractive index function of the sample. The near-field electric field distribution is the simple product of the illumination probe and the complex refractive index function of the sample, P(λ,ξ,η)S(ξ,η).

[0076] S201.3 The far-field diffraction intensity distribution I(x′, y′) and I(x′, y′) of the defective and non-defective samples are obtained by using the Fraunhofer far-field propagation model. And do the difference The diffraction difference light intensity distribution ΔI(x′,y′) is obtained, which is the far-field difference data. The calculation method of the Fraunhofer far-field propagation model is as follows:

[0077]

[0078] Where (x′, y′) is the diffraction plane coordinate (corresponding to the detector target position), U′(x, y) is the light field amplitude distribution on the diffraction plane, k is the wave number, and z represents the diffraction propagation distance. The two terms before the integral term affect the phase and overall amplitude of the light field. In actual detection, the detector obtains the intensity distribution of the light field, that is, the square of the amplitude. The phase factor does not affect the intensity distribution of the diffraction field, and the main focus is on the relative intensity distribution. Therefore, in order to reduce the amount of calculation, the first two terms of the integral can be discarded. In addition, the integral term is mathematically equivalent to the Fourier transform of the near-field electric field distribution U(x, y), so the final calculation formula for solving the far-field diffraction intensity distribution is as follows:

[0079]

[0080] in Represents a Fourier transform operation.

[0081] S201.4 Move the illumination probe to the next scanning position, and repeat steps S201.2-S201.3 along the serpentine scanning path until the entire sample is scanned (that is, a complete scan is completed), while ensuring that the overlapping area of ​​adjacent illumination spots is not less than 60%, that is, the spatial overlap constraint. The scanning path is shown in the figure below: Figure 4 As shown;

[0082] S201.5 Keep the sample pattern unchanged, adjust the coordinate position of the defect on the sample (ξ1, η1) and the defect characteristic length D, obtain a new defective lithography sample, and repeat steps S201.2-S201.4 for the new defective lithography sample until the defect coordinate position is (ξ n ,η n ) is scanned, and the far-field differential data obtained from each scan is saved;

[0083] Figure 5 Schematic diagram of a transmissive and reflective lithography defect detection system provided in an embodiment of the present application. Figure 5 As shown, it includes a transmission detection system and a reflection detection system, and one of them can be selected according to actual conditions during specific implementation. The photolithography defect detection system is sequentially configured as follows: a helium-neon laser 1, an attenuation plate 2, an adjustable aperture 3, a beam expander 4, a reflector 5, an adjustable aperture 6, a focusing lens 7, a photolithography sample 8, a camera 9, and a computer 10.

[0084] Step S202, preprocessing the far-field differential data: performing asymmetric analysis on the far-field differential data corresponding to each scanning point in the data set, and using the defect coordinates and size of each complete scan of the sample plane as label data, which are used together as input to the neural network.

[0085] The method for asymmetric analysis of far-field differential data is as follows:

[0086]

[0087] Among them, ΔI L-R Indicates left-right asymmetric information, ΔI T-B Indicates up-down asymmetric information, ΔI sum Represents the difference sum information.

[0088] The embodiment of the present application uses a pixel-based CMOS detector to receive light intensity signals, and obtains a matrix in which each pixel corresponds to a light intensity value. The method for solving the left-right asymmetric information is to subtract the matrix corresponding to the far-field differential data ΔI(x′, y′) from the matrix corresponding to its horizontal flip data ΔI(-x′, y′), and for the matrix obtained after the subtraction, all elements in the matrix are summed and normalized. The method for solving the upper and lower asymmetric information is to subtract the matrix corresponding to the far-field differential data ΔI(x′, y′) from the matrix corresponding to its vertical flip data ΔI(x′, -y′), and for the matrix obtained after the subtraction, all elements in the matrix are summed and normalized. The method for solving the differential sum information is to sum all elements in the matrix corresponding to the far-field differential data ΔI(x′, y′), and normalize. After all scanning points are solved, two asymmetric matrices I for a single complete scan can be obtained. L-R ,I T-B and a difference sum matrix I sum , respectively expressed as follows:

[0089]

[0090] The superscripts of the elements in the matrix represent the coordinate sequence of the center position of each scanning point in a complete scan, such as (i, j) represents the scanning point in the i-th row and j-th column, and 1≤i≤m, and 1≤j≤m. Finally, the three matrices are combined to obtain three-channel data. It is expressed as follows:

[0091]

[0092] Here, k represents the kth complete scan. To ensure that the network has enough data sets for training, the number of scans is no less than 800.

[0093] Figure 6-8 This is an illustration of the three-channel data obtained from a complete scan: Figure 6 , Figure 7 , Figure 8 They are the left-right asymmetric information, the up-down asymmetric information, and the difference summation information of each scanning point in a single scan. Figure 6-8 The units of the coordinate axis are explained in the figure. The coordinates in the two directions on the horizontal plane represent the scanning point positions corresponding to a single complete scan, and the vertical coordinate represents the normalized signal intensity corresponding to each scanning point. Figure 6 , Figure 7 It can be seen that when the defect position is scanned, the asymmetric information will have positive and negative peaks along the corresponding direction. When the scanning line is offset from the defect position, the peak height will decrease. Figure 8From the differential summation information, we can see that the closer the scanning point is to the defect position, the higher the differential summation result will be; the larger the defect feature size, the larger the width of the generated signal; the above have very obvious characteristics and are easier to extract using neural networks.

[0094] Furthermore, at the end of each scan, the defect location and size need to be recorded as label data T k ,

[0095] It is expressed as follows:

[0096] T k =[P(ξ k ,η k ),D k ];

[0097] Where P represents the location of the defect, (ξ k ,η k ) represents the coordinate position of the defect center on the sample, D k Indicates the size of the defect.

[0098] Finally, the preprocessed training data and label data are represented as follows:

[0099]

[0100] T=[T1,T2 … T n ];

[0101] Step S203, establish a prediction network for defect location and size. Contains feature extraction, location prediction and category prediction parts, and the output is the defect location coordinates and defect category. The network structure includes but is not limited to LeNet, AlexNet, VGGNet and other networks, including convolutional layers, pooling layers, batch normalization layers, fully connected layers, etc. The feature extraction module needs to be determined according to the data scale. In the embodiment of the present application, the feature extraction module is composed of a convolutional layer, a pooling layer and a ReLU activation function layer (the number of layers can be further increased considering the input information dimension); the position prediction module is composed of a BatchNorm (batch normalization) layer and a fully connected layer, and outputs the predicted defect location coordinates on the sample; the category prediction module is composed of a fully connected layer and a Softmax-with-loss layer, and outputs the predicted defect category. The network structure is as follows Fig. 9 shown.

[0102] The defect classification method is based on the size range of the defect characteristic length D. The endpoint values ​​of each interval need to be set in advance. For example, if the interval endpoints are set to Ω = [40, 60, 80, 100, 120], the unit is nm, then the defects are divided into 4 categories, respectively, in the four intervals corresponding to every two endpoints.

[0103] Step S204, setting the network loss function and parameter update method: the loss functions used in the network include the position loss function L P And the category loss function L C , the two loss functions are expressed as follows:

[0104]

[0105] The loss function L for the defect location P The type is selected as the root mean square error, which represents the predicted defect location (ξ i ,η i ) and the actual position (ξ t ,η t ). The defect type is represented by one-hot, the correct solution label is 1, and the rest are 0. For example, at the above-set interval endpoint Ω=[40,60,80,100,120], the one-hot representation of the correct solution label of the defect size in the range of [60,80) is t k =[0,1,0,0]. The defect type loss function is the cross entropy error, y k is the output of the neural network, t k is the label data. This network adopts mini-batch learning, that is, each time learning will randomly select a batch of n batch (For example, batch =64) data set for learning, it will also output a number of n batch prediction results, so it is necessary to average the loss functions to obtain the “average loss function” of a single data.

[0106] Finally, the total loss function is the position loss function L P And the category loss function L C The weighted addition of is expressed as follows:

[0107] L=αL P +(1-α)L c ;

[0108] The Adam optimizer is used to update the parameters of the network, and the moment estimation decay exponents β1 and β2 are set to 0.9 and 0.999.

[0109] Step S205, training the network and adjusting system hyperparameters: The number of neurons in the neural network needs to be determined according to the dimension of the input data. In determining the initial weights of each neuron, in order to prevent the problem of gradient disappearance, "He Initialization" is selected. The number of neuron nodes in the current layer is n. neuron When the standard deviation is Gaussian distribution.

[0110] The best learning rate is found by performing multiple trainings. Considering that a complete learning may take a long time, the range of learning rates is specified on a logarithmic scale and random sampling is performed from this range. For example, the range of learning rates is set to 10 -3 ~10 3 , the learning rate for each complete training is set to That is, parameter n lr Each time, random sampling is performed in the range of [-3,3], and a relatively suitable result is finally found.

[0111] The deep learning-based lithography defect detection device provided in the present application is described below. The deep learning-based lithography defect detection device described below and the deep learning-based lithography defect detection method described above can be referenced to each other.

[0112] Fig.10 is a schematic diagram of the structure of a lithography defect detection device based on deep learning provided in an embodiment of the present application, such as Fig.10 As shown, the device includes: a data set construction module 20, a data set preprocessing module 30 and a training module 40. Among them:

[0113] The data set construction module 20 is used to scan each defective lithography sample by a coherent Fourier scattering method to obtain a far-field differential data set, wherein the far-field differential data set includes defect label data corresponding to each defective lithography sample and far-field differential data corresponding to each scanning point in each defective lithography sample, wherein the far-field differential data is obtained by differentiating the diffraction light intensity distribution of the far field corresponding to the defective lithography sample and the diffraction light intensity distribution of the far field corresponding to the defect-free lithography sample for the scanning point;

[0114] The data set preprocessing module 30 is used to perform asymmetry analysis on the far-field differential data corresponding to each scanning point based on the far-field differential data set, and obtain left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to each scanning point in each defective lithography sample;

[0115] The training module 40 trains a deep learning model based on the defect label data corresponding to each defective lithography sample and the left-right asymmetry information, the top-bottom asymmetry information and the differential summation information corresponding to each scanning point in each defective lithography sample. The deep learning model is used to detect lithography defects in lithography samples.

[0116] It can be understood that the detailed functional implementation of each of the above-mentioned units / modules can be found in the introduction of the aforementioned method embodiment, and will not be repeated here.

[0117] It should be understood that the above-mentioned device is used to execute the method in the above-mentioned embodiment. The implementation principle and technical effect of the corresponding program module in the device are similar to those described in the above-mentioned method. The working process of the device can refer to the corresponding process in the above-mentioned method, which will not be repeated here.

[0118] Based on the method in the above embodiment, an embodiment of the present application provides an electronic device, Fig.11 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, such as Fig.11 As shown, the electronic device may include: a processor (Processor) 810, a communication interface (CommunicationsInterface) 820, a memory (Memory) 830 and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other through the communication bus 840. The processor 810 can call the logic instructions in the memory 830 to execute the method in the above embodiment.

[0119] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application.

[0120] Based on the method in the above embodiment, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program runs on a processor, the processor executes the method in the above embodiment.

[0121] Based on the method in the above embodiment, an embodiment of the present application provides a computer program product. When the computer program product runs on a processor, the processor executes the method in the above embodiment.

[0122] It is understandable that the processor in the embodiment of the present application may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0123] The method steps in the embodiments of the present application can be implemented by hardware or by a processor executing software instructions. The software instructions can be composed of corresponding software modules, and the software modules can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, mobile hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an ASIC.

[0124] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer instructions may be transmitted from a website site, a computer, a server or a data center to another website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or a data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

[0125] It should be understood that the various numerical numbers involved in the embodiments of the present application are only used for the convenience of description and are not used to limit the scope of the embodiments of the present application.

[0126] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A lithography defect detection method based on deep learning, characterized in that: include: Scan each defective lithography sample by a coherent Fourier scattering method to obtain a far-field differential data set, wherein the far-field differential data set includes defect label data corresponding to each defective lithography sample and far-field differential data corresponding to each scanning point in each defective lithography sample, and the far-field differential data is obtained by differentiating the diffraction light intensity distribution of the far field corresponding to the defective lithography sample and the diffraction light intensity distribution of the far field corresponding to the defect-free lithography sample with respect to the scanning point; Based on the far-field differential data set, an asymmetric analysis is performed on the far-field differential data corresponding to each scanning point to obtain left-right asymmetric information, top-bottom asymmetric information and differential summation information corresponding to each scanning point in each defective lithography sample; Based on the defect label data corresponding to each defective lithography sample and the left-right asymmetric information, top-bottom asymmetric information and differential summation information corresponding to each scanning point in each defective lithography sample, a deep learning model is trained, and the deep learning model is used to detect lithography defects in lithography samples.

2. The method for lithography defect detection based on deep learning according to claim 1, characterized in that: The far-field differential data corresponding to the scanning points are obtained by the following steps: At the scanning point, calculate the near-field electric field distribution after the illumination probe interacts with the lithography sample; Based on the near-field electric field distribution, the far-field diffraction light intensity distribution corresponding to the defective lithography sample and the far-field diffraction light intensity distribution corresponding to the defect-free lithography sample are obtained through the Fraunhofer far-field propagation model; Based on the diffraction light intensity distribution of the far field corresponding to the defective lithography sample and the diffraction light intensity distribution of the far field corresponding to the defect-free lithography sample, a difference is made to obtain far-field differential data corresponding to the scanning point.

3. The method for lithography defect detection based on deep learning according to claim 2, characterized in that: The calculating of the near-field electric field distribution after the illumination probe interacts with the photolithography sample at the scanning point includes calculating the near-field electric field distribution by the following formula: U(x,y)=P(λ,ξ,η)S(ξ,η); Among them, U(x,y) represents the near-field electric field distribution, (x,y) represents the coordinates in the near-field plane coordinate system, P(λ,ξ,η) represents the input illumination matrix, (ξ,η) represents the coordinates in the sample coordinate system, λ represents the illumination wavelength, and S(ξ,η) represents the sample complex refractive index function.

4. The method for lithography defect detection based on deep learning according to any one of claims 1 to 3, characterized in that: The deep learning model includes a cascaded feature extraction module, a position prediction module and a category prediction module, and accordingly, the defect label data includes a defect position label and a defect size category label; The feature extraction module is used to extract defect features corresponding to the lithography sample based on left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to the scanning point; The position prediction module is used to predict the defect position based on the defect characteristics corresponding to the lithography sample; The category prediction module is used to predict the defect category based on the prediction information of the defect position and the defect characteristics corresponding to the lithography sample.

5. The method for lithography defect detection based on deep learning according to claim 4, characterized in that: The loss function used to train the deep learning model is: L=αL P +(1-α)L c ; Among them, L represents the total loss, α represents the weighting coefficient, and L P represents the position loss, L c represents the class loss.

6. The method for lithography defect detection based on deep learning according to claim 5, characterized in that: The position loss is calculated by the following formula: Among them, (ξ i ,η i ) represents the defect location predicted by the deep learning model, (ξ t ,η t ) represents the defect location indicated by the defect location label, n batch Represents the number of samples in a training batch. A sample includes the left-right asymmetric information, top-bottom asymmetric information, and differential summation information corresponding to each scanning point in a defective lithography sample.

7. The method for lithography defect detection based on deep learning according to claim 5, characterized in that: The class loss is calculated by the following formula: Among them, y k represents the category predicted by the deep learning model, t k Indicates the category indicated by the defect size category label.

8. A lithography defect detection device based on deep learning, characterized in that: include: A data set construction module is used to scan each defective lithography sample by a coherent Fourier scattering method to obtain a far-field differential data set, wherein the far-field differential data set includes defect label data corresponding to each defective lithography sample and far-field differential data corresponding to each scanning point in each defective lithography sample. The far-field differential data is obtained by differentiating the diffraction light intensity distribution of the far field corresponding to the defective lithography sample and the diffraction light intensity distribution of the far field corresponding to the defect-free lithography sample for the scanning point; A data set preprocessing module is used to perform asymmetry analysis on the far-field differential data corresponding to each scanning point based on the far-field differential data set, and obtain left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to each scanning point in each defective lithography sample; The training module trains a deep learning model based on the defect label data corresponding to each defective lithography sample and the left-right asymmetry information, top-bottom asymmetry information and differential summation information corresponding to each scanning point in each defective lithography sample. The deep learning model is used to detect lithography defects in lithography samples.

9. An electronic device, characterized in that: include: at least one memory for storing a computer program; At least one processor is used to execute the program stored in the memory. When the program stored in the memory is executed, the processor is used to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program runs on a processor, the processor is caused to execute the method according to any one of claims 1 to 7.

Images