An etching prediction method, apparatus, device and medium

Abstract

The application provides an etching prediction method, device, equipment and medium, can obtain etching process parameters and a plurality of first etching profile data obtained by etching under the etching process parameters; the etching process parameters and the plurality of first etching profile data are input into an etching profile prediction model, and second etching profile data at a second time is output; the second time is after the first time; the etching profile prediction model is determined based on a long short-term memory network and a cascade combination layer, is a cascade recurrent neural network, can reduce a large amount of calculation based on a complex physical mechanism modeling in a traditional etching model, reduce repetitive manual trial and error work in calibration of the traditional etching model, reduce modeling time, improve simulation efficiency, in addition, improve the prediction accuracy of the etching profile data.

Description

Technical Field

[0001] This application relates to the semiconductor field, and in particular to an etching prediction method, apparatus, device, and medium. Background Technology

[0002] Modeling and predicting the etching process is a complex and difficult problem. Traditional methods for predicting the etching process mainly involve modeling the original physical mechanisms and chemical reactions. Monte Carlo models are generally used to simulate the particle trajectories and chemical reactions in the etching process. Although the models are quite accurate, they still require a lot of time to simulate a single etching process. Complex physical mechanisms require a lot of computation. In addition, model calibration also requires repeated manual trial and error, resulting in high time costs.

[0003] As process dimensions shrink and circuit size increases, computational complexity grows exponentially, leading to lower efficiency. Modeling a single-time etching profile also fails to reflect the temporal changes throughout the entire etching process. Therefore, providing a suitable etching prediction method has become an urgent technical problem to be solved. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide an etching prediction method, apparatus, device, and medium that reduces modeling time, improves simulation efficiency, and enhances the accuracy of etching profile prediction. The specific solution is as follows:

[0005] On the one hand, this application provides an etching prediction method, including:

[0006] Acquire etching process parameters and multiple first etching profile data obtained by etching under the etching process parameters, wherein each first etching profile data corresponds to each first time step;

[0007] The etching process parameters and multiple first etching profile data are input into the etching profile prediction model, and the second etching profile data at the second time point is output; the second time point is after the first time point; the etching profile prediction model is determined based on a long short-term memory network and a cascaded combination layer.

[0008] Specifically, the training process of the etching contour prediction model includes:

[0009] Obtain a first etching profile dataset; the first etching profile dataset includes multiple historical etching profile data, and the multiple historical etching profile data correspond one-to-one with multiple historical moments;

[0010] Based on the first etching profile dataset, the long short-term memory network is used to model and obtain the initial model for etching profile prediction.

[0011] Based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset, the initial etching profile prediction model is trained using the cascaded combination layer to obtain the trained etching profile prediction model.

[0012] Specifically, the step of training the initial etching profile prediction model using the cascaded combination layer based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset to obtain the trained etching profile prediction model includes:

[0013] Based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset, the initial etching profile prediction model is trained using the cascaded combination layer to obtain the target etching profile prediction model.

[0014] Obtain the actual etching profile obtained through process means and the actual etching process parameters corresponding to the actual etching profile;

[0015] The target etching profile prediction model is adjusted using the actual etching profile and the actual etching process parameters to obtain the etching profile prediction model.

[0016] Specifically, the step of training the initial etching profile prediction model using the cascaded combination layer based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset to obtain the trained etching profile prediction model includes:

[0017] The historical etching profile data and the historical etching process parameters are spliced ​​together and fully connected to obtain the trained etching profile prediction model.

[0018] Specifically, obtaining the first etched contour dataset includes:

[0019] The first etched profile dataset is generated using simulation software.

[0020] Specifically, generating the first etching contour dataset using simulation software includes:

[0021] A second etching profile dataset is generated using simulation software;

[0022] The second etching profile dataset is normalized to obtain the first etching profile dataset.

[0023] Specifically, the cascaded combination layer includes multiple splicing layers and multiple fully connected layers.

[0024] In another aspect, embodiments of this application also provide an etching prediction apparatus, comprising:

[0025] The acquisition unit is used to acquire etching process parameters and multiple first etching profile data obtained by etching under the etching process parameters, wherein each first etching profile data corresponds to each first time moment;

[0026] The prediction unit is used to input the etching process parameters and multiple first etching profile data into the etching profile prediction model, and output the second etching profile data at a second time point; the second time point is located after the first time point; the etching profile prediction model is determined based on a long short-term memory network and a cascaded combination layer.

[0027] In another aspect, embodiments of this application provide a computer device, the computer device including a processor and a memory:

[0028] The memory is used to store program code and transmit the program code to the processor;

[0029] The processor is used to execute the methods described above according to the instructions in the program code.

[0030] In another aspect, embodiments of this application provide a computer-readable storage medium for storing a computer program for performing the methods described above.

[0031] This application provides an etching prediction method, apparatus, device, and medium. It can acquire etching process parameters and multiple first etching profile data obtained under the etching process parameters, with each first etching profile data corresponding to a first time step. The multiple first etching profile data can reflect the profile change trend during the etching process. The etching process parameters and the multiple first etching profile data are input into an etching profile prediction model, which outputs second etching profile data at a second time step. The second time step is located after the first time step. The etching profile prediction model is based on a long short-term memory network and a cascaded combined layer, and is a cascaded recurrent neural network. This model can reduce the large amount of computation required for modeling based on complex physical mechanisms in traditional etching models, reduce repetitive manual trial and error work in traditional etching model calibration, reduce modeling time, improve simulation efficiency, and further improve the prediction accuracy of etching profile data. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 A flowchart illustrating an etching prediction method provided in an embodiment of this application is shown.

[0034] Figure 2 A cross-sectional schematic diagram of an etched trench provided in an embodiment of this application is shown;

[0035] Figure 3 This illustration shows a structural schematic diagram of an etching profile prediction model provided in an embodiment of this application;

[0036] Figure 4 This illustration shows a structural schematic diagram of another etching profile prediction model provided in an embodiment of this application;

[0037] Figure 5 A structural block diagram of an etching prediction device provided in an embodiment of this application;

[0038] Figure 6 This is a structural diagram of a computer device provided in an embodiment of this application. Detailed Implementation

[0039] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the specific embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0040] Many specific details are set forth in the following description in order to provide a full understanding of this application. However, this application may also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0041] Secondly, this application provides a detailed description in conjunction with schematic diagrams. When detailing the embodiments of this application, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not limit the scope of protection of this application. In addition, actual fabrication should include three-dimensional spatial dimensions of length, width, and depth.

[0042] For ease of understanding, the following detailed description, in conjunction with the accompanying drawings, provides an etching prediction method, apparatus, device, and medium according to embodiments of this application.

[0043] refer to Figure 1 The diagram shown is a flowchart of an etching prediction method provided in an embodiment of this application. The method may include the following steps.

[0044] S101, acquire etching process parameters and multiple first etching profile data obtained by etching under the etching process parameters, each first etching profile data corresponding to each first time step.

[0045] In this embodiment, the etching process parameters can be various environmental parameters during the etching process, parameters that can have a certain impact on the etching result, such as the power applied during etching, voltage, pressure or temperature within the chamber, etc. The etching process parameters can be obtained and generally do not change during the etching process. The etching process parameters can include at least one of etching power, etching voltage, etching pressure, or etching temperature; for example, they can include etching power, etching voltage, etching pressure, and etching temperature.

[0046] Etching profile data can be the profile data of the cross-section of the etching trench, and a polar coordinate system can be established for the trench, with reference to... Figure 2 The diagram shown is a cross-sectional view of an etching trench provided in an embodiment of this application. It includes a structure to be etched 101 and a mask layer 102 located above the structure to be etched. The structure to be etched 101 has etching trenches. A polar coordinate system can be established with the center point of the etching trench surface as the origin O. At each etching moment, a set of etching contour data is provided. The distance between any point P on the contour of the etched trench and the origin O can be denoted as d, the angle between the PO line and the trench surface can be denoted as θ, and the point P can be denoted as (d, θ).

[0047] Specifically, the entire etching process can be divided into multiple stages according to time, with each stage corresponding to a specific moment. For example, it can be divided into 6 stages, corresponding to 0, 1, 2, 3, 4, and 5 in the diagram. For each stage, multiple polar coordinate points can be used for representation. For instance, a point can be taken every 10°, resulting in 19 polar coordinate points for the etching profile data at each moment. To make the etching profile data more accurately reflect the shape of the etched cross-section, a point can also be taken every 1°, resulting in 181 polar coordinate points for the etching profile data at each stage. The distance 'd' of each angle of the etching profile is uniformly sampled. In this way, the etching profile at each moment can be discretized into a multidimensional vector, and the multiple profiles recorded throughout the etching process can be considered a multivariate time series. For ease of description, the etching profile data at the first moment can be denoted as the first etching profile data, which may include multiple polar coordinate points.

[0048] Specifically, under the influence of etching process parameters, multiple first etching profile data can be obtained. These multiple first etching profile data can reflect the profile change trend during the etching process. Each first etching profile data corresponds to a first moment. In other words, multiple first etching profile data at multiple moments can be acquired as a set of etching profiles to predict the etching profile data at the next moment. For example, a set of etching profiles can include etching profile data corresponding to each second (i.e., the first moment) within the time period from the 1st second to the 10th second, for a total of 10 etching profile data. Subsequently, the etching profile data at the 11th second (i.e., the second moment) can be predicted based on these 10 etching profile data.

[0049] S102, the etching process parameters and multiple first etching profile data are input into the etching profile prediction model, and the second etching profile data at the second time point is output; the second time point is after the first time point; the etching profile prediction model is determined based on the long short-term memory network and cascaded combination layer.

[0050] In this embodiment, an etching contour prediction model can be determined based on a Long Short-Term Memory (LSTM) network and cascaded combined layers. LSTM is a variant of Recurrent Neural Networks (RNNs), which are recurrent neural networks that take sequence data as input, recursively move along the sequence's direction, and have all nodes (recurrent units) connected in a chain-like manner. LSTM networks are commonly used in natural language processing tasks and can also be used in time series prediction. LSTM introduces gate structures on top of RNNs, including input gates, output gates, and forget gates, mitigating the gradient vanishing and gradient exploding problems inherent in RNNs.

[0051] Specifically, the etching profile prediction model can be expressed as f(x) p ,x e ) = y, f is the mapping relationship represented by the network, containing two inputs x. p and x e x p It is the etched outline data of history, x e y represents the etching process parameters, and y represents the predicted etching profile data.

[0052] The etching process parameters and multiple first etching profile data are used as inputs to the etching profile prediction model, and the output is the second etching profile data at a second time point, which is a time point after the first time point. The etching environment remains unchanged between the second and first time points; that is, the etching process parameters at the second time point are the same as those at the first time point. Therefore, based on the known etching profiles and etching process parameters of the trenches at multiple first time points, the etching profile of the trenches at the second time point can be predicted, thus obtaining the second etching profile data. Of course, there can be one or multiple second time points, resulting in the prediction of multiple second etching profile data points. For example, the etching profile data from the 1st to the 10th second can be used to predict the etching profile data from the 11th to the 15th second, resulting in a set of etching profile data. This allows for accurate prediction of the etching profile at future time points, providing advance knowledge of the possible shape and size of the trenches. Furthermore, by calculating the difference between the second and first etching profile data, the amount of etching during the time period from the first to the second time point can be obtained. The etching profile prediction model is a cascaded recurrent neural network that can reduce the large amount of computation required for modeling based on complex physical mechanisms in traditional etching models, reduce repetitive manual trial and error in the calibration of traditional etching models, reduce modeling time, improve simulation efficiency, and improve the prediction accuracy of etching profile data.

[0053] In the embodiments of this application, before using the etching contour prediction model to predict the contour, an initial etching contour prediction model can be constructed, and the initial etching contour prediction model can be trained using a training set. After the training is completed, the etching contour prediction model is obtained. The model building and training process is described below, including steps S201-S203.

[0054] S201, Obtain the first etched contour dataset.

[0055] Each historical moment has corresponding historical etching profile data, and multiple historical etching profile data at multiple historical moments can be obtained. That is, multiple historical etching profile data correspond one-to-one with multiple historical moments, and multiple historical etching profile data can be used as the first etching profile dataset.

[0056] In one possible implementation, the first etching profile dataset can be obtained from the etching profile of the actual etching process, which can improve the accuracy of the first etching profile dataset.

[0057] In another possible implementation, simulation software can be used to build a simulation model, and historical etching contour data can be obtained through simulation calculations as the first etching contour dataset. This provides a large amount of historical etching contour data, which allows for extensive model training and improves the accuracy of the etching contour prediction model. In addition, actual etching contours are often obtained by cutting and scanning, which is highly destructive to the etched structure and yields relatively little data. Obtaining the first etching contour dataset through simulation can reduce the damage to the etched structure.

[0058] refer to Figure 3 The diagram shown is a structural schematic of an etching profile prediction model provided in an embodiment of this application. It includes a preprocessing module, an LSTM module, a cascaded combination layer, and an output module. The preprocessing module can be used for normalization processing. The LSTM module is used to model the etching profile. The cascaded combination layer can integrate etching process parameters into the network structure in a cascaded manner. The model is trained using etching profile data with different environmental parameters. The output module is used to output the prediction results.

[0059] In practical applications, the etching profile dataset can be normalized to simplify calculations. Specifically, the first etching profile dataset can be generated by simulation software. The etching profile dataset generated by the simulation software can be designated as the second etching profile dataset. The second etching profile dataset can then be normalized to obtain the first etching profile dataset.

[0060] Specifically, data normalization can be achieved using the mean-standard deviation method, and the calculation method can be expressed as follows:

[0061]

[0062] Where, x i For the historical etch contour data of the i-th historical moment before normalization, x i ′ represents the historical etching profile data at the i-th historical moment after normalization, μ represents the mean of multiple historical etching profile data, and s represents the standard deviation of multiple historical etching profile data.

[0063] Specifically, the first etch profile dataset obtained after normalization can be denoted as X={x0′,x1′,…,x i ′,…x t ′},x t ′ represents the historical etching profile data at the t-th historical moment after normalization.

[0064] S202, based on the first etching profile dataset, a long short-term memory network is used for modeling to obtain the initial model for etching profile prediction.

[0065] Long Short-Term Memory (LSTM) networks can be used to model the first etching profile dataset to obtain an initial model for etching profile prediction. The first etching profile dataset can be represented as M×N, where M represents that the historical profile data at each historical moment has M points, and N represents that there are N sets of data, i.e., N historical moments.

[0066] Specifically, refer to Figure 4 The diagram shown is a structural schematic of another etching profile prediction model provided in this application embodiment. The input data is a first etching profile dataset, which can be 181×48 sets of data. 181 indicates that the etching profile data at each historical moment is represented by 181 points, and 48 sets indicates that the first etching profile dataset contains a total of 48 sets of data, that is, it has 48 historical moments. LSTM can be used to model these 48 sets of data to obtain 128×48 sets of data after modeling, that is, 48 ​​sets of data with a data bit width of 128.

[0067] Specifically, the first etched contour dataset is connected through two layers of LSTM network to obtain the corresponding hidden state matrix H, which is calculated as H = F(F(x) + x).

[0068] Specifically, assume the hidden state matrix is ​​H = {h1, h2, ..., h...} t}, where h i h represents the hidden state at historical moment i. i The calculation method mainly consists of the following equations from the LSTM unit:

[0069] i t =σ(W i ·[x t ,h t-1 ]+b i )

[0070] f t =σ(W f ·[x t ,h t-1 ]+b f )

[0071]

[0072]

[0073] o t =σ(W o ·[x t ,h t-1 ]+b o )

[0074] h t =o t *tanh(Ct )

[0075] Among them, i t f t o t Let x represent the outputs of the input gate, forget gate, and output gate at time t, respectively. t h represents the input at historical time t. t-1 h t Let C represent the hidden state at historical time t-1 and historical time t, respectively. t denoted by t, where W and b are trainable parameters, tanh is the hyperbolic tangent function, and σ is the sigmoid function.

[0076] S203, based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset, the initial etching profile prediction model is trained using a cascaded combination layer to obtain a trained etching profile prediction model.

[0077] In this embodiment, historical etching process parameters corresponding to the first etching profile dataset can be obtained. These historical etching process parameters represent the etching environment at a historical moment and can be normalized. A cascaded combination layer can be used to train the initial etching profile prediction model. When the training stopping condition is met, a trained etching profile prediction model is obtained.

[0078] Specifically, historical etching profile data and historical etching process parameters can be stitched together and fully connected to obtain a trained etching profile prediction model. The cascaded combination layer can include multiple stitching layers and multiple fully connected layers. The stitching layers stitch together the historical etching profile data and historical etching process parameters, and the fully connected layers perform fully connected processing on the stitched data. Figure 4 In this process, the output data of the LSTM network can be processed into one-dimensional data with a width of 128N, or 6144 bits. Since the width of the output data is limited to 1024, 1024 bits are taken from the 128N data each time and concatenated with the corresponding four historical etching process parameters (such as etching power, etching voltage, etching pressure, and etching temperature). Then, a fully connected operation is performed to obtain 512 bits of data. Next, the 512 bits of data are concatenated with the four historical etching process parameters again and a fully connected operation is performed to obtain 256 bits of data. The 256 bits of data are concatenated with the four historical etching process parameters again and a fully connected operation is performed to obtain 256 bits of data.

[0079] For example, suppose the 1024-bit input data is denoted as X = {x1, x2, ..., x...} 1024The input data is then concatenated with the historical etching process parameters to obtain the concatenated input data, which can be denoted as X′={x1,x2,…,x}. 1024 The algorithm first calculates the power, voltage, pressure, and temperature values, then performs a full connection operation between layers, a total of 3 full connections, outputting 256 points, which are finally sorted into 181 points, corresponding to one etch profile.

[0080] In one possible implementation, to improve the accuracy of the etching profile prediction model, transfer learning can be performed on the model, using a small amount of real etching data to adjust the model and obtain the final etching profile prediction model.

[0081] Specifically, the initial etching profile prediction model can be trained using a cascaded combination layer based on the first etching profile dataset and the corresponding historical etching process parameters. The resulting model is denoted as the target etching profile prediction model. Then, etching is performed using process methods. The actual etching profile can be obtained by observing the cross-section of the structure to be etched using a scanning electron microscope. The actual etching process parameters corresponding to the actual etching profile are the same as the historical etching process parameters.

[0082] The target etching profile prediction model is adjusted using actual etching profiles and actual etching process parameters. When the expected result is achieved, the etching profile prediction model is obtained. In this way, under the same etching process parameters, the model can be fine-tuned using real process profile data, which can improve the accuracy of model prediction.

[0083] Specifically, the performance of the etching profile prediction model can be evaluated using the mean absolute error (MAE), which is defined as:

[0084]

[0085] Among them, y i This represents historical etching profile data. This represents the predicted etching profile data.

[0086] Specifically, the neural network model in this invention is based on RNN. Replacing it with Recurrent Neural Networks (DNN), Convolutional Neural Networks (CNN), combinations thereof, or changing the number of layers in the network are all within the scope of this invention.

[0087] This invention can be accelerated using CUDA (Compute Unified Device Architecture) tools, which can further improve the efficiency of etching modeling.

[0088] This invention uses mean absolute error to evaluate the model, and any change in the form of the error function is within the scope of this invention.

[0089] This invention uses polar coordinates to represent etching contour data, which simplifies the representation of etching data. Except for the two-dimensional contour in the example, contour representations of other dimensions are within the scope of this invention.

[0090] This invention uses simulation data for model pre-training and real data for transfer learning of the model. The modeling and optimization methods for processes such as etching are within the scope of this invention.

[0091] This application provides an etching prediction method that can acquire etching process parameters and multiple first etching profile data obtained under the etching process parameters, each first etching profile data corresponding to a first time step; input the etching process parameters and multiple first etching profile data into an etching profile prediction model, and output second etching profile data at a second time step; the second time step is located after the first time step; the etching profile prediction model is determined based on a long short-term memory network and a cascaded combined layer, which is a cascaded recurrent neural network. It can reduce the large amount of computation in traditional etching models based on complex physical mechanisms, reduce the repetitive manual trial and error work in the calibration of traditional etching models, reduce modeling time, improve simulation efficiency, and improve the prediction accuracy of etching profile data.

[0092] Based on the above etching prediction method, this application also provides an etching prediction device, referencing... Figure 5 The diagram shown is a structural block diagram of an etching prediction device provided in an embodiment of this application. The device may include:

[0093] Acquisition unit 201 is used to acquire etching process parameters and multiple first etching profile data obtained by etching under the etching process parameters, wherein each first etching profile data corresponds to each first time moment;

[0094] The prediction unit 202 is used to input the etching process parameters and multiple first etching profile data into the etching profile prediction model, and output the second etching profile data at a second time point; the second time point is located after the first time point; the etching profile prediction model is determined based on a long short-term memory network and a cascaded combination layer.

[0095] Specifically, the training process of the etching contour prediction model includes:

[0096] Obtain a first etching profile dataset; the first etching profile dataset includes multiple historical etching profile data, and the multiple historical etching profile data correspond one-to-one with multiple historical moments;

[0097] Based on the first etching profile dataset, the long short-term memory network is used to model and obtain the initial model for etching profile prediction.

[0098] Based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset, the initial etching profile prediction model is trained using the cascaded combination layer to obtain the trained etching profile prediction model.

[0099] Specifically, the step of training the initial etching profile prediction model using the cascaded combination layer based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset to obtain the trained etching profile prediction model includes:

[0100] Based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset, the initial etching profile prediction model is trained using the cascaded combination layer to obtain the target etching profile prediction model.

[0101] Obtain the actual etching profile obtained through process means and the actual etching process parameters corresponding to the actual etching profile;

[0102] The target etching profile prediction model is adjusted using the actual etching profile and the actual etching process parameters to obtain the etching profile prediction model.

[0103] Specifically, the step of training the initial etching profile prediction model using the cascaded combination layer based on the first etching profile dataset and the historical etching process parameters corresponding to the first etching profile dataset to obtain the trained etching profile prediction model includes:

[0104] The historical etching profile data and the historical etching process parameters are spliced ​​together and fully connected to obtain the trained etching profile prediction model.

[0105] Specifically, obtaining the first etched contour dataset includes:

[0106] The first etched profile dataset is generated using simulation software.

[0107] Specifically, generating the first etching contour dataset using simulation software includes:

[0108] A second etching profile dataset is generated using simulation software;

[0109] The second etching profile dataset is normalized to obtain the first etching profile dataset.

[0110] Specifically, the cascaded combination layer includes multiple splicing layers and multiple fully connected layers.

[0111] This application provides an etching prediction device. An acquisition unit acquires etching process parameters and multiple first etching profile data obtained under the etching process parameters, each first etching profile data corresponding to a first time step. A prediction unit inputs the etching process parameters and the multiple first etching profile data into an etching profile prediction model and outputs second etching profile data at a second time step. The second time step is located after the first time step. The etching profile prediction model is determined based on a long short-term memory network and a cascaded combined layer, which is a cascaded recurrent neural network. This model reduces the large amount of computation required for modeling based on complex physical mechanisms in traditional etching models, reduces repetitive manual trial and error work in traditional etching model calibration, reduces modeling time, improves simulation efficiency, and further improves the prediction accuracy of the etching profile data.

[0112] In another aspect, embodiments of this application provide a computer device, with reference to Figure 6 The diagram shown is a structural diagram of a computer device provided in an embodiment of this application. The computer device includes a processor 310 and a memory 320.

[0113] The memory 320 is used to store program code and transmit the program code to the processor 310;

[0114] The processor 310 is used to execute the method provided in the above embodiments according to the instructions in the program code.

[0115] The computer device may include a terminal device or a server, and the aforementioned apparatus may be configured in the computer device.

[0116] In another aspect, embodiments of this application also provide a storage medium for storing a computer program for executing the methods provided in the above embodiments.

[0117] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by program instructions in hardware. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium can be at least one of the following media: read-only memory (ROM), RAM, magnetic disk, or optical disk, etc., and other media capable of storing program code.

[0118] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. In particular, the apparatus embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0119] The above description is merely a preferred embodiment of this application. Although this application has disclosed preferred embodiments above, it is not intended to limit this application. Any person skilled in the art can make many possible variations and modifications to the technical solutions of this application using the methods and techniques disclosed above, or modify them into equivalent embodiments with equivalent changes, without departing from the scope of the technical solutions of this application. Therefore, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of this application without departing from the content of the technical solutions of this application shall still fall within the protection scope of the technical solutions of this application.

Claims

1. An etch prediction method, comprising: The method comprises: acquiring etching process parameters and a plurality of first etching profile data obtained by etching under the etching process parameters, each of the first etching profile data corresponding to each first time point; inputting the etching process parameters and the plurality of first etching profile data into an etching profile prediction model to output second etching profile data at a second time point; the second time point is located after the first time point; the etching profile prediction model is determined based on a long short-term memory network and a cascade combination layer; the etching process parameters at the second time point are the same as the etching process parameters at the first time point; the cascade combination layer comprises a plurality of splicing layers and a plurality of fully connected layers, the splicing layers being used for splicing etching profile data and etching process parameters, and the fully connected layers being used for fully connected processing of data formed after splicing.

2. The method of claim 1, wherein, The training process of the etching profile prediction model comprises: acquiring a first etching profile data set; the first etching profile data set comprises a plurality of historical etching profile data corresponding to a plurality of historical time points; modeling based on the first etching profile data set by using the long short-term memory network to obtain an etching profile prediction initial model; training the etching profile prediction initial model based on the first etching profile data set and historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model.

3. The method of claim 2, wherein, The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model.

4. The method of claim 2, wherein, The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model.

5. The method of claim 2, wherein, The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model.

6. The method of claim 5, wherein, The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises:

7. An etching prediction apparatus characterized by comprising: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model based on the first etching profile data set and the historical etching process parameters corresponding to the first etching profile data set by using the cascade combination layer to obtain the trained etching profile prediction model. The training process of the etching profile prediction model comprises: training the etching profile prediction initial model An acquisition unit is configured to acquire etching process parameters and a plurality of first etching profile data obtained by etching under the etching process parameters, each of the first etching profile data corresponding to each first time point; A prediction unit is configured to input the etching process parameters and the plurality of first etching profile data into an etching profile prediction model, and output second etching profile data at a second time point; The second time point is located after the first time point; the etching profile prediction model is determined based on a long short-term memory network and a cascade combination layer; the etching process parameters at the second time point are the same as the etching process parameters at the first time point; the cascade combination layer includes a plurality of splicing layers and a plurality of fully connected layers, the splicing layers are configured to splice the etching profile data and the etching process parameters, and the fully connected layers are configured to perform full connection processing on the data formed after splicing.

8. A computer device, comprising: The computer device includes a processor and a memory: The memory is configured to store program code and transmit the program code to the processor; The processor is configured to execute the method according to the instructions in the program code.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium is configured to store a computer program, and the computer program is configured to execute the method according to any one of claims 1-6. The computer readable storage medium is configured to store a computer program, and the computer program is configured to execute the method according to any one of claims 1-6.

Images