Mask image generation method and device, electronic equipment and storage medium

Abstract

The invention discloses a mask image generation method and device, electronic equipment and a storage medium, and belongs to the field of photoetching. The mask image generation method for the photoetching machine comprises the following steps: acquiring a plurality of target process parameters; inputting the plurality of target process parameters into a multi-layer sensing layer of the target network model, and obtaining a target reconstructed image output by the multi-layer sensing layer; and inputting the target reconstructed image, the target expected mask image and the target illumination image into a generative adversarial network layer of the target network model, and obtaining a target mask image output by the generative adversarial network layer. The mask image generation method for the photoetching machine can be suitable for different photoetching machines and photoetching environments, adaptive adjustment is carried out, so that a more accurate mask image is obtained, the resolution of a photoetching system is improved, and the method has relatively high flexibility and universality.

Description

Technical Field

[0001] This application belongs to the field of lithography, and particularly relates to a method, device, electronic device, and storage medium for generating a mask image. Background Art

[0002] Lithography is a core link in the chip manufacturing process. As the resolution of the lithography system continues to increase, the diffraction and interference effects of light become more and more obvious, making the lithography patterns on the wafer become distorted and blurred. In related technologies, mainly the method of lithography simulation is used to improve the resolution. For example, the model-based optical proximity correction method (MB-OPC) is adopted. A set of algorithms is designed according to physical, chemical, and optical laws, and through multiple iterative calculations to improve the resolution of the lithography system. This method has low efficiency, great implementation difficulty, and is difficult to be applicable to complex and changeable lithography scenarios. Summary of the Invention

[0003] This application aims to solve at least one of the technical problems existing in the prior art. Therefore, this application provides a method, device, electronic device, and storage medium for generating a mask image, which can obtain a more accurate mask image, improve the resolution of the lithography system, and has high flexibility and universality.

[0004] In a first aspect, this application provides a method for generating a mask image for a lithography machine, the method including: Obtain a plurality of target process parameters; Input the plurality of target process parameters into the multi-layer perceptron layer of the target network model, and obtain a target reconstructed image output by the multi-layer perceptron layer; Input the target reconstructed image, the target desired mask image, and the target illumination image into the generative adversarial network layer of the target network model, and obtain a target mask image output by the generative adversarial network layer; wherein, The target mask image is used for exposure by the lithography machine, and the output end of the multi-layer perceptron layer is connected to the input end of the generative adversarial network layer; the target network model is trained based on a plurality of training samples.

[0005] According to the method for generating a mask image for a lithography machine of this application, by predicting the target mask image through the target process parameters, the target desired mask image, and the target illumination image, it can enable the target mask image to be adaptively adjusted based on different process parameters and different illumination images, so that in any lithography environment, a target mask image that can obtain the target desired mask image is generated, without multiple iterations, and has high computational efficiency; and it can be applicable to different lithography machines and lithography environments for adaptive adjustment, so as to obtain a more accurate mask image, improve the resolution of the lithography system, and has high flexibility and universality.

[0006] According to an embodiment of the present application, inputting the multiple target process parameters into the multi-layer perceptron layer of the target network model and obtaining the target reconstructed image output by the multi-layer perceptron layer includes: Input the multiple target process parameters into the multi-layer perceptron layer of the target network model. The multi-layer perceptron layer learns the influence weights of the target process parameters, and maps each target process parameter to a target feature vector, where the target feature vector is used to characterize the correlation relationship between the target process parameters; Reconstruct the target feature vector into a single-channel image to obtain the target reconstructed image.

[0007] According to an embodiment of the present application, obtaining the multiple target process parameters includes: Obtain the multiple target process parameters according to the type of the lithography machine and the target lithography environment.

[0008] According to an embodiment of the present application, the multiple target process parameters include: exposure dose and photoresist parameters.

[0009] According to an embodiment of the present application, the training samples are obtained based on the following steps: Obtain multiple sample process parameters corresponding to the sample lithography machine; Obtain the lithography result image obtained by the sample lithography machine processing the sample mask image under the multiple sample process parameters and the sample illumination image; Using the multiple sample process parameters, the sample illumination image, and the lithography result image as samples, and the sample mask image as a sample label, construct the training samples.

[0010] According to an embodiment of the present application, the target network model is trained based on multiple training samples, including: Input the multiple sample process parameters in the training samples into the multi-layer perceptron layer; input the lithography result image, the sample illumination image, and the sample mask image in the training samples into the generative adversarial network layer, and use the sample reconstructed image output by the multi-layer perceptron layer and the predicted mask image output by the generative adversarial network layer as targets to train the target network model.

[0011] According to an embodiment of the present application, before obtaining the multiple target process parameters, the method further includes: Receive a first input from the user for at least some of the multiple candidate configuration modules in the OpenMMLab configuration interface; In response to the first input, construct the target network model based on the at least some modules.

[0012] Second aspect, the present application provides a mask image generation device for a lithography machine, the device comprising: A first processing module, configured to obtain a plurality of target process parameters; A second processing module, configured to input the plurality of target process parameters into a multi-layer perceptron layer of a target network model, and obtain a target reconstructed image output by the multi-layer perceptron layer; A third processing module, configured to input the target reconstructed image, a target desired mask image, and a target illumination image into a generative adversarial network layer of the target network model, and obtain a target mask image output by the generative adversarial network layer; wherein, The target mask image is used for the lithography machine to perform image correction, and an output end of the multi-layer perceptron layer is connected to an input end of the generative adversarial network layer; the target network model is trained based on a plurality of training samples.

[0013] According to the mask image generation device for a lithography machine of the present application, by predicting a target mask image through target process parameters, a target desired mask image, and a target illumination image, it is possible to adaptively adjust the target mask image based on different process parameters and different illumination images, so that in any lithography environment, a target mask image capable of obtaining the target desired mask image can be generated, without multiple iterations, and has a high calculation efficiency; and it can be applied to different lithography machines and lithography environments for adaptive adjustment, so as to obtain a more accurate mask image, improve the resolution of the lithography system, and has high flexibility and universality.

[0014] Third aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the processor executes the computer program, it implements the mask image generation method for a lithography machine as described in the first aspect above.

[0015] Fourth aspect, the present application provides a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the mask image generation method for a lithography machine as described in the first aspect above.

[0016] Fifth aspect, the present application provides a computer program product, comprising a computer program, and when the computer program is executed by a processor, it implements the mask image generation method for a lithography machine as described in the first aspect above.

[0017] One or more of the above technical solutions in the embodiments of the present application have at least one of the following technical effects: Predicting the target mask image through the target process parameters, the target desired mask image, and the target illumination image enables the target mask image to be adaptively adjusted based on different process parameters and different illumination images, so that in any lithography environment, a target mask image that can obtain the target desired mask image can be generated without multiple iterations, with high computational efficiency; and it can be applied to different lithography machines and lithography environments for adaptive adjustment, so as to obtain a more accurate mask image, improve the resolution of the lithography system, and have high flexibility and universality.

[0018] Furthermore, by learning the correlation relationship between the target process parameters and the importance score of the lithography result through a multi-layer perceptron layer, generating a target reconstruction image as one of the features for predicting the target mask image, the accuracy and effect of the predicted target mask image can be improved, and a matching mask image can be predicted based on different process parameters, with high flexibility and a wide range of applicable scenarios.

[0019] Still further, by using the OpenMMLab framework for training, only the required modules need to be selected from it, combined and adjusted to meet specific requirements, and the model training can be completed only by writing a configuration file and running it, reducing repetitive actions, improving the model construction efficiency, and lowering the usage threshold; in addition, the deep learning network module can be customized to receive various process parameters and illumination patterns, and jointly optimize with the target desired mask image, further improving the flexibility and construction efficiency of model construction.

[0020] The additional aspects and advantages of this application will be partly given in the following description, partly become obvious from the following description, or be understood through the practice of this application. Description of the Drawings

[0021] The above and / or additional aspects and advantages of this application will become obvious and easy to understand from the description of the embodiments in conjunction with the following drawings, where: Figure 1 is one of the flow diagrams of the mask image generation method for a lithography machine provided by an embodiment of this application; Figure 2 is another flow diagram of the mask image generation method for a lithography machine provided by an embodiment of this application; Figure 3 is the principle diagram of the mask image generation method for a lithography machine provided by an embodiment of this application; Figure 4 is the third flow diagram of the mask image generation method for a lithography machine provided by an embodiment of this application; Figure 5 is the structural diagram of the mask image generation device for a lithography machine provided by an embodiment of this application; Figure 6 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. Specific embodiments

[0022] Next, the technical solutions in the embodiments of the present application will be clearly described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are 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 those of ordinary skill in the art belong to the scope of protection of the present application.

[0023] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, rather than to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims means at least one of the connected objects, and the character " / " generally represents an "or" relationship between the associated objects before and after.

[0024] Next, in conjunction with the accompanying drawings, a mask image generation method for a lithography machine, a mask image generation device for a lithography machine, an electronic device, and a readable storage medium provided by an embodiment of the present application will be described in detail through specific embodiments and their application scenarios.

[0025] Among them, the mask image generation method for a lithography machine can be applied to a terminal, and can be specifically executed by hardware or software in the terminal.

[0026] The terminal includes, but is not limited to, portable communication devices such as mobile phones or tablet computers having a touch-sensitive surface (for example, a touch screen display and / or a touchpad). It should also be understood that in some embodiments, the terminal may not be a portable communication device, but a desktop computer having a touch-sensitive surface (for example, a touch screen display and / or a touchpad).

[0027] In the following various embodiments, a terminal including a display and a touch-sensitive surface is described. However, it should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, a mouse, and a joystick.

[0028] The mask image generation method for a lithography machine provided by an embodiment of the present application. The execution subject of the mask image generation method for the lithography machine can be an electronic device or a functional module or functional entity in the electronic device that can implement the mask image generation method for the lithography machine. The electronic devices mentioned in the embodiments of the present application include, but are not limited to, mobile phones, tablet computers, computers, cameras, and wearable devices, etc. Hereinafter, taking the electronic device as the execution subject as an example, the mask image generation method for the lithography machine provided by the embodiments of the present application will be described.

[0029] As Figure 1 shown, the mask image generation method for the lithography machine includes: step 110, step 120, and step 130.

[0030] It should be noted that the mask image generation method for the lithography machine is applied to the field of lithography applications.

[0031] Step 110: Obtain multiple target process parameters; In this step, the process parameters are the process parameters involved in the lithography process of the lithography machine, including but not limited to: exposure system parameters, resolution-related parameters, photoresist (photo resist) parameters, process control parameters, mask (reticle) parameters, and other key parameters, etc.

[0032] Among them, the exposure system parameters may include: light source wavelength (Wavelength), numerical aperture (NA, Numerical Aperture), exposure dose (Exposure Dose), and illumination mode, etc.

[0033] The photoresist (photo resist) parameters may include: photoresist type, sensitivity, and contrast, etc.

[0034] The process control parameters may include: depth of focus, scanning speed, and environmental control, etc.

[0035] The mask (reticle) parameters may include: mask magnification, phase-shift mask, and optical proximity effect correction, etc.

[0036] It can be understood that for different types of lithography machines, the process parameters involved may be different; for different lithography environments, the corresponding process parameters may also vary.

[0037] The target process parameters can be user-defined. For example, the user selects the corresponding process parameters as the target process parameters according to actual needs.

[0038] In some embodiments, step 110 may include: Obtain multiple target process parameters according to the type of the lithography machine and the target lithography environment.

[0039] In this embodiment, the target process parameters can be the actual parameters in the current lithography scenario, which can be determined according to the type of the lithography machine and the lithography environment, such as environmental temperature and humidity, shock and vibration suppression ability, etc.

[0040] In some embodiments, the multiple target process parameters may include: exposure dose and photoresist parameters.

[0041] Of course, in other embodiments, the target process parameters may further include: exposure dose, light source wavelength, photoresist type, sensitivity, contrast, scanning speed, mask magnification, and environmental temperature and humidity, etc.

[0042] Step 120: Input the multiple target process parameters into the multi-layer perceptron layer of the target network model, and obtain the target reconstructed image output by the multi-layer perceptron layer; In this step, the target network model can be a pre-constructed and trained neural network model, including three external input channels and one external output channel.

[0043] The target network model includes: a multi-layer perceptron layer and a generative adversarial network layer. Among them, the multi-layer perceptron layer corresponds to one external input channel, the generative adversarial network layer corresponds to two external input channels and one external output channel, and the internal output end of the multi-layer perceptron layer is connected to the internal input end of the generative adversarial network layer.

[0044] Among them, the multi-layer perceptron (MLP) layer is a feedforward artificial neural network, including multiple neuron layers, as Figure 4 shown, each circle corresponds to a neuron, and can learn complex non-linear mapping relationships.

[0045] The generative adversarial network (GAN) layer is a deep learning model that generates data through adversarial training, and may include, but is not limited to, image-to-image deep learning networks such as Pix2Pix or CycleGAN.

[0046] The target network model is trained based on multiple training samples. The specific training method will be described in the following embodiments and will not be elaborated here for the time being.

[0047] The target reconstructed image is an image that can represent the non-linear mapping relationship between the target process parameters, such as an image of 256×256; of course, in other embodiments, images of other sizes can also be set based on actual needs, and the present application does not make any limitations here.

[0048] In some embodiments, step 120 may include: Input multiple target process parameters into the multi-layer perceptron layer of the target network model. The multi-layer perceptron layer learns the influence weights of each target process parameter and maps each target process parameter into a target feature vector, which is used to characterize the correlation relationship between the target process parameters. Reconstruct the target feature vector into a single-channel image to obtain a target reconstructed image.

[0049] In this embodiment, continue to refer to Figure 4 , taking multiple target process parameters including exposure dose, photoresist sensitivity, and other parameters as an example. Input the exposure dose, photoresist sensitivity, and other parameters into the multi-layer perceptron layer. By learning the influence weights of each parameter, map each parameter into a target feature vector that can characterize the non-linear relationship between various types of parameters, such as a tensor of 1×65536, whose total length is the same as the size of the target expected mask image and the target illumination image. Then reconstruct it into a target reconstructed image, such as a tensor arranged as 1×256×256.

[0050] The target reconstructed image obtained after reconstruction, the target expected mask image, and the target illumination image are connected into a three-channel image, such as a tensor of 1×256×256, and then input into the generative adversarial network layer of the target network model for processing to obtain a target mask image.

[0051] According to the mask image generation method for a lithography machine provided by the embodiments of the present application, by learning the correlation relationship between target process parameters and the importance score for the lithography result through the multi-layer perceptron layer, generating a target reconstructed image as one of the features for predicting the target mask image can improve the accuracy and effect of the predicted target mask image, and can predict a matching mask image based on different process parameters, with high flexibility and a wide range of applicable scenarios.

[0052] Step 130: Input the target reconstructed image, the target expected mask image, and the target illumination image into the generative adversarial network layer of the target network model, and obtain the target mask image output by the generative adversarial network layer.

[0053] In this step, the target illumination image can be determined based on the illumination mode, including but not limited to: traditional illumination image, off-axis illumination (OAI) image, and annular illumination image, etc., which are used to optimize the pattern contrast and resolution.

[0054] Figure 2Illustrates a lithography process flow chart. In the lithography process, it involves a light source, an optical system, an illumination image, a mask image, a lens, a photoresist, a wafer for placing the photoresist, and related process parameters (including but not limited to photoresist parameters, etc.). Among them, the mask image is placed on the mask plate. Under the process parameters and the illumination image, the lithography system processes the mask image to obtain a lithography result image. In the physical process of simulating lithography, it involves an Inverse Lithography Technology (ILT), as Figure 3 shown, that is, given a desired target mask pattern target, it is required that the model outputs what kind of mask pattern M will become the target mask pattern target after passing through the lithography system. This target mask pattern target corresponds to the target desired mask image of the present application, and the mask pattern M corresponds to the target mask image output by the desired target network model of the present application.

[0055] That is, the target desired mask image is the final mask result obtained by the desired lithography machine processing the mask image under the target process parameters and the target illumination image.

[0056] The target mask image is the desired wafer pattern, which is the image corresponding to the standard that the mask image needs to reach when the lithography machine processes the mask image under the target process parameters and the target illumination image and can meet the expected mask result.

[0057] The target mask image is used for the lithography machine to further expose.

[0058] In the actual application process, continuing to refer to Figure 4 , only by inputting the target process parameters, the target desired mask image, and the target illumination image into the target network model, the target network model can output the inverse lithography result that meets the requirements, that is, the target mask image.

[0059] This target mask image can be used in subsequent lithography scenarios. Under the target process parameters and the target illumination image, the lithography machine processes this target mask image to obtain a lithography result that meets the expectations.

[0060] The inventors found during the R & D process that in the related art, there is also a method of using machine learning, deep learning, neural network and other methods to train a model to simulate the physical process of lithography, that is, inputting the target mask and outputting the inverse lithography result by the model. However, in the case of a certain target mask, the output inverse lithography results are roughly the same, and in different lithography machines or lithography environments, there may be significant differences in the lithography results obtained by processing the same mask image, thus affecting the resolution of the lithography system and being unable to be applied to different lithography machines and lithography environments, having certain limitations.

[0061] According to the mask image generation method for a lithography machine provided by an embodiment of the present application, predicting a target mask image through target process parameters, a target desired mask image, and a target illumination image can enable the target mask image to be adaptively adjusted based on different process parameters and different illumination images. Thus, in any lithography environment, a target mask image that can obtain the target desired mask image can be generated without multiple iterations, having a high computational efficiency; and it can be applicable to different lithography machines and lithography environments for adaptive adjustment, thereby obtaining a more accurate mask image, improving the resolution of the lithography system, and having high flexibility and universality.

[0062] The acquisition of training samples, the construction of the target network model, and the training method will be described below.

[0063] In some embodiments, the training samples can be obtained based on the following steps: Obtain multiple sample process parameters corresponding to a sample lithography machine; Obtain the lithography result image obtained by the sample lithography machine processing the sample mask image under multiple sample process parameters and a sample illumination image; Using multiple sample process parameters, the sample illumination image, and the lithography result image as samples, and the sample mask image as a sample label, construct training samples.

[0064] In this embodiment, one sample lithography machine can correspond to multiple training samples. By setting multiple different sample lithography machines and using a similar method, multiple training samples can be obtained, thereby obtaining a dataset. Among them, the sample lithography machine is a lithography machine used to obtain sample data, the number of sample lithography machines can be multiple, and the categories or models of different sample lithography machines may also vary.

[0065] It can be understood that when obtaining the sample mask image, the sample mask image can be corrected by processing the obtained lithography result image in reverse until, under multiple sample process parameters and a sample illumination image, the sample lithography machine processes the finally corrected sample mask image to obtain a lithography result image with a high resolution. Use these multiple sample process parameters, the sample illumination image, the finally corrected sample mask image, and the lithography result image with a high resolution as a training sample.

[0066] Taking the construction of a dataset using the OpenMMLab module as an example, since the paired format image can only be divided into two parts on the left and right, where the left is domain_a and the right is domain_b, during the actual execution process, a Dataset class that meets the input requirements of this module can be written to adapt to the three-channel input features corresponding to the present application (i.e., the target desired mask image, the target illumination image, and multiple target process parameters) and the corresponding output result (i.e., the target mask image).

[0067] Specifically, for each data sample, three corresponding data files are respectively set, namely "sample name_target.png", "sample name_lightpattern.png", "sample name_otherparams.txt", and "sample name_ILT.png". The storage paths of these four types of files are the same.

[0068] In some other embodiments, four folders can also be created, namely target, lightpattern, otherparams, and ILT, and then files with the same name are stored in these four folders respectively.

[0069] In some embodiments, the target network model is trained based on multiple training samples, and may include: Inputting multiple sample process parameters in the training samples into a multi-layer perceptron layer; inputting the lithography result images, sample illumination images, and sample mask images in the training samples into a generative adversarial network layer, and training the target network model with the output of the multi-layer perceptron layer being the sample reconstruction image and the output of the generative adversarial network layer being the predicted mask image.

[0070] In this embodiment, the data set is input into the target network model, and the target network model is trained with the output of the multi-layer perceptron layer being the sample reconstruction image and the output of the generative adversarial network layer being the predicted mask image. During the training process, a cross-entropy loss function or a mean squared error loss function, etc., can be used to globally train the target network model based on the difference degree between the predicted mask image and the sample mask image, so as to obtain the weights of each neuron in the multi-layer perceptron layer and the model parameters of the adversarial network layer.

[0071] In some embodiments, the data set can also be divided into a training set and a test set, for example, divided according to a ratio of 7:3 or 8:2. The target network model is trained by the training set, and the trained target network model is tested and optimized by the test set.

[0072] In some embodiments, before step 110, the method may further include: Receiving a first input from the user for at least some of the multiple candidate configuration modules in the OpenMMLab configuration interface; In response to the first input, constructing a target network model based on at least some of the modules.

[0073] In this embodiment, the first input is used to select the required modules from the multiple candidate configuration modules provided by the OpenMMLab tool and configure the connection relationships between the modules.

[0074] Among them, the first input can be at least one of the following ways: First, the first input can be a touch operation, including but not limited to click operations, swipe operations, press operations, etc.

[0075] In this embodiment, receiving the user's first input can be receiving the user's touch operation on the display area of the terminal display screen.

[0076] To reduce the user's accidental operation rate, the action area of the first input can be limited to a specific area, such as the upper middle area of the OpenMMLab configuration interface; or when the OpenMMLab configuration interface is displayed, a target control is displayed on the current interface, and touching the target control can achieve the first input; or the first input is set as a continuous multiple tapping operation on the display area within a target time interval.

[0077] Second, the first input can be a physical button input.

[0078] In this embodiment, corresponding physical buttons are provided on the body of the terminal. Receiving the user's first input can be receiving the user's operation of pressing the corresponding physical button; the first input can also be a combined operation of pressing multiple physical buttons simultaneously.

[0079] Third, the first input can be a voice input.

[0080] Of course, in other embodiments, the first input can also be in other forms, including but not limited to character input, etc., which can be determined according to actual needs specifically, and the embodiments of this application do not limit this.

[0081] The module can include but not limited to: network layer, loss function, optimizer, etc.

[0082] The following is an explanation of the specific implementation methods.

[0083] 1. Set up the OpenMMLab environment.

[0084] (1) Create an environment openmmlab_generate using Anaconda and switch to this environment conda create -name openmmlab python=3.8 -y conda activate openmmlab (2) Install CUDA version 11.3 (3) Install pytorch 1.11.0 conda install pytorch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0 cudatoolkit=11.3 -c pytorch (4) Install openmim, and then install mmcv-full pip install -U openmim mim install mmcv-full (5) Use the git tool to clone the mmgeneration project to the local, enter the root directory of the project, and then install mmgeneration in an editable manner through pip install -v -e. (6) Run the example code to test the installation success or failure for installation verification.

[0085] 2. Writing and training the configuration script of OpenMMLab

[0086] (1) Dataset production 1) Write a program to generate a dataset in the form of paired images. The first two input parameters are the paths of the input and label, and the last is the location where the processed images are saved.

[0087] 2) Write a program to randomly sample files. The first parameter is the path of the file to be sampled, and the second and third are the paths of the training set and test set.

[0088] 3) Download the dataset provided by lithobench, then first run the first program to convert the image format, and then execute the second program to generate the training set and test set.

[0089] (2) Write a configuration script (take Pix2Pix as an example, CycleGAN is the same) 1) Run the provided by the official The configuration script of the size. Finally, all configurations of this configuration script will be generated in workdir. Rename this all configuration to a suitable file name and cut and paste it to a suitable path. Next, make modifications based on this configuration file.

[0090] 2) Model configuration modification: First, change the in_channel of Generator to 3, out_channel to 1, the in_channel of discriminator to 3, and change the configurations related to the input source and output source to target and ILT respectively.

[0091] 3) Dataset pipeline configuration: Both the training set and the test set are changed to the pair format, the flag is changed to grayscale, domain_a and domain_b are changed to target and ILT respectively, and the random cropping part is removed. Finally, the configurations of the training set and the test set are basically the same, except that the test set does not have the flip step.

[0092] 4) Set dataroot to the paths of the training set and the test set generated in (1). 5) Configure variables such as checkpoint_config, log_config, custom_hooks, exp_name, work_dir, total_iters, etc. according to your own needs.

[0093] (3) Model training: Configure the run and debug configuration options of the IDE.

[0094] (4) Model code writing: The input parameters of the constructor of the model should include the number of process parameters, the shape of the fully connected layers (an array, each element representing the size of the layer), the size of the input image and the output image, the parameters of the GAN architecture, etc.

[0095] The forward function of the model should be forward(self, x, y, z). The three parameters (x, y, z) represent the target expected mask image, the target illumination image, and the tensor corresponding to multiple target process parameters respectively. The vector of multiple target process parameters passes through a multi-layer perceptron to obtain the target reconstructed image, which is concatenated with the target expected mask image and the target illumination image to form a three-channel image and is input into the GAN model. Among them, the input channel number of this GAN model is determined to be 3 and the output channel number is 1 during the constructor stage.

[0096] The inventor also found during the R & D process that in related technologies, the code of the target network model is mostly written in Pytorch. Writing the code requires writing codes such as Dataset, Dataloader, training, and testing from scratch, which is relatively complex to operate and difficult to implement, and has a relatively high usage threshold.

[0097] According to the mask image generation method for a lithography machine provided by an embodiment of the present application, by using the OpenMMLab framework for training, only the required modules need to be selected from it, and different modules are combined and adjusted to meet specific requirements. Only by writing a configuration file and running it can the training of the model be completed, reducing repetitive actions, improving the model construction efficiency, and lowering the usage threshold. In addition, deep learning network modules can be customized to receive various process parameters and illumination images, and jointly optimize them with the target expected mask image, further improving the flexibility and construction efficiency of model construction.

[0098] For the mask image generation method for a lithography machine provided by an embodiment of the present application, the execution subject can be a mask image generation device for a lithography machine. In the embodiments of the present application, taking the mask image generation device for a lithography machine executing the mask image generation method for a lithography machine as an example, the mask image generation device for a lithography machine provided by the embodiments of the present application is described.

[0099] An embodiment of the present application also provides a mask image generation device for a lithography machine.

[0100] As Figure 5 shown, the mask image generation device for a lithography machine includes: a first processing module 510, a second processing module 520, and a third processing module 530.

[0101] The first processing module 510 is configured to obtain a plurality of target process parameters; The second processing module 520 is configured to input the plurality of target process parameters into the multi-layer perceptron layer of the target network model to obtain a target reconstructed image output by the multi-layer perceptron layer; The third processing module 530 is configured to input the target reconstructed image, the target expected mask image, and the target illumination image into the generative adversarial network layer of the target network model to obtain a target mask image output by the generative adversarial network layer; wherein, The target mask image is used for further exposure by the lithography machine, and the output end of the multi-layer perceptron layer is connected to the input end of the generative adversarial network layer; the target network model is trained based on a plurality of training samples.

[0102] According to the mask image generation device for a lithography machine provided by an embodiment of the present application, by predicting the target mask image through the target process parameters, the target expected mask image, and the target illumination image, the target mask image can be adaptively adjusted based on different process parameters and different illumination images, so that in any lithography environment, a target mask image that can obtain the target expected mask image can be generated without multiple iterations, and has a high calculation efficiency; and it can be applied to different lithography machines and lithography environments for adaptive adjustment, so as to obtain a more accurate mask image, improve the resolution of the lithography system, and has high flexibility and universality.

[0103] In some embodiments, the second processing module 520 may further be configured to: Input multiple target process parameters into the multi-layer perceptron layer of the target network model. The multi-layer perceptron layer learns the influence weights of the target process parameters and maps the target process parameters to target feature vectors, where the target feature vectors are used to characterize the correlation relationships between the target process parameters; Reconstruct the target feature vectors into a single-channel image to obtain a target reconstructed image.

[0104] In some embodiments, the first processing module 510 may further be configured to: Obtain multiple target process parameters according to the type of the lithography machine and the target lithography environment.

[0105] In some embodiments, the multiple target process parameters include: exposure dose and photoresist parameters.

[0106] In some embodiments, the apparatus may further include a fourth processing module, configured to: Obtain multiple sample process parameters corresponding to a sample lithography machine; Obtain a lithography result image obtained by the sample lithography machine processing a sample mask image under multiple sample process parameters and a sample illumination image; Construct a training sample with multiple sample process parameters, the sample illumination image, and the lithography result image as samples and the sample mask image as a sample label.

[0107] In some embodiments, the fourth processing module may further be configured to: Input the multiple sample process parameters in the training sample into the multi-layer perceptron layer; input the lithography result image, the sample illumination image, and the sample mask image in the training sample into the generative adversarial network layer, and train the target network model with the output of the multi-layer perceptron layer being the sample reconstructed image and the output of the generative adversarial network layer being the predicted mask image as the goal.

[0108] In some embodiments, the apparatus may further include a fifth processing module, configured to: Before obtaining the multiple target process parameters, receive a first input from the user for at least some of the multiple candidate configuration modules in the OpenMMLab configuration interface; In response to the first input, construct a target network model based on at least some of the modules.

[0109] The mask image generation device for a lithography machine in the embodiments of the present application may be an electronic device or a component in an electronic device, such as an integrated circuit or a chip. The electronic device may be a terminal or other devices other than terminals. Exemplarily, the electronic device may be a mobile phone, a tablet computer, a laptop computer, a handheld computer, a vehicle-mounted electronic device, a Mobile Internet Device (MID), an augmented reality (AR) / virtual reality (VR) device, a robot, a wearable device, an ultra-mobile personal computer (UMPC), a netbook, or a personal digital assistant (PDA), etc. It may also be a server, a Network Attached Storage (NAS), a personal computer (PC), a television (TV), a teller machine, or a self-service machine, etc. The embodiments of the present application do not make specific limitations.

[0110] The mask image generation device for a lithography machine in the embodiments of the present application may be a device with an operating system. The operating system may be an Android operating system, an IOS operating system, or other possible operating systems. The embodiments of the present application do not make specific limitations.

[0111] The mask image generation device for a lithography machine provided in the embodiments of the present application can implement Figures 1 to 4 each process implemented by the method embodiments. To avoid repetition, it will not be elaborated here.

[0112] In some embodiments, as Figure 6 shown, the embodiments of the present application further provide an electronic device 600, including a processor 601, a memory 602, and a computer program stored on the memory 602 and executable on the processor 601. When the program is executed by the processor 601, it implements each process of the above method embodiment for generating a mask image for a lithography machine and can achieve the same technical effect. To avoid repetition, it will not be elaborated here.

[0113] It should be noted that the electronic devices in the embodiments of the present application include the above-mentioned mobile electronic devices and non-mobile electronic devices.

[0114] An embodiment of the present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements each process of the above-mentioned embodiment of the mask image generation method for a lithography machine and can achieve the same technical effects. To avoid repetition, it will not be elaborated here.

[0115] Wherein, the processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes computer-readable storage media, such as computer read-only memory ROM, random access memory RAM, magnetic disk or optical disc, etc.

[0116] An embodiment of the present application also provides a computer program product, including a computer program, which implements the above-mentioned mask image generation method for a lithography machine when executed by a processor.

[0117] Wherein, the processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes computer-readable storage media, such as computer read-only memory ROM, random access memory RAM, magnetic disk or optical disc, etc.

[0118] Another embodiment of the present application provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement each process of the above-mentioned embodiment of the mask image generation method for a lithography machine and can achieve the same technical effects. To avoid repetition, it will not be elaborated here.

[0119] It should be understood that the chip mentioned in the embodiment of the present application may also be referred to as a system-on-chip, system chip, chip system or system-on-chip, etc.

[0120] It should be noted that in this article, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a..." does not exclude the existence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the methods and devices in the embodiments of the present application are not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in a reverse order according to the functions involved. For example, the described methods may be performed in an order different from that described, and various steps may be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0121] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described method of the embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases, the former is a better implementation. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art can be embodied in the form of a computer software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions for causing a terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present application.

[0122] The embodiments of the present application have been described above in conjunction with the accompanying drawings. However, the present application is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present application, those of ordinary skill in the art can also make many forms without departing from the purpose of the present application and the scope protected by the claims, and all of them belong to the protection scope of the present application.

[0123] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0124] Although the embodiments of the present application have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and purposes of the present application. The scope of the present application is defined by the claims and their equivalents.

Claims

1. A method for generating a mask image for a lithography machine, characterized in that Including: Obtain a plurality of target process parameters; Input the plurality of target process parameters into the multi-layer perceptron layer of the target network model, and obtain a target reconstructed image output by the multi-layer perceptron layer; Input the target reconstructed image, the target desired mask image, and the target illumination image into the generative adversarial network layer of the target network model, and obtain a target mask image output by the generative adversarial network layer; wherein, The target mask image is used for the exposure of the lithography machine, and the output end of the multi-layer perceptron layer is connected to the input end of the generative adversarial network layer; the target network model is trained based on a plurality of training samples.

2. The method for generating a mask image for a lithography machine according to claim 1, characterized in that, The step of inputting the plurality of target process parameters into the multi-layer perceptron layer of the target network model and obtaining the target reconstructed image output by the multi-layer perceptron layer includes: Input the plurality of target process parameters into the multi-layer perceptron layer of the target network model, and let the multi-layer perceptron layer learn the influence weights of the target process parameters, and map each target process parameter to a target feature vector, where the target feature vector is used to characterize the correlation relationship between the target process parameters; Reconstruct the target feature vector into a single-channel image to obtain the target reconstructed image.

3. The mask image generation method for a lithography machine according to claim 1, wherein, The step of obtaining a plurality of target process parameters includes: According to the type of the lithography machine and the target lithography environment, obtain the plurality of target process parameters.

4. The mask image generation method for a lithography machine according to any one of claims 1-3, characterized in that, The plurality of target process parameters include: exposure dose and photoresist parameters.

5. The method for generating a mask image for a lithography machine according to any one of claims 1-3, characterized in that, The training samples are obtained based on the following steps: Obtain a plurality of sample process parameters corresponding to the sample lithography machine; Obtain a lithography result image obtained by the sample lithography machine processing a sample mask image under the plurality of sample process parameters and the sample illumination image; Using the plurality of sample process parameters, the sample illumination image, and the lithography result image as samples, and the sample mask image as a sample label, construct the training samples.

6. The method for generating a mask image for a lithography machine according to any one of claims 1-3, characterized in that, The target network model is trained based on a plurality of training samples, including: Input the plurality of sample process parameters in the training samples into the multi-layer perceptron layer; input the lithography result image, the sample illumination image, and the sample mask image in the training samples into the generative adversarial network layer, and use the sample reconstructed image output by the multi-layer perceptron layer and the predicted mask image output by the generative adversarial network layer as targets to train the target network model.

7. The mask image generation method for a lithography machine according to any one of claims 1-3, characterized in that, Before obtaining the plurality of target process parameters, the method further includes: Receive a first input from a user for at least some of a plurality of candidate configuration modules in the OpenMMLab configuration interface; In response to the first input, construct the target network model based on the at least some modules.

8. A mask image generation device for a lithography machine, characterized in that, Including: A first processing module, configured to obtain a plurality of target process parameters; A second processing module, configured to input the plurality of target process parameters into the multi-layer perceptron layer of the target network model, and obtain a target reconstructed image output by the multi-layer perceptron layer; A third processing module, configured to input the target reconstructed image, the target desired mask image, and the target illumination image into the generative adversarial network layer of the target network model, and obtain a target mask image output by the generative adversarial network layer; wherein, The target mask image is used for the image correction by the lithography machine, and the output end of the multi-layer perceptron layer is connected to the input end of the generative adversarial network layer; the target network model is trained based on a plurality of training samples.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the mask image generation method for a lithography machine according to any one of claims 1-7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the mask image generation method for a lithography machine according to any one of claims 1-7.

Images