A Coating Control Method, System and Equipment for Photoresist

Abstract

The present invention relates to a coating control method, system and device for photoresist, belonging to the field of semiconductor technology, and includes: collecting an image of a dispensing head, performing dispensing deviation analysis to obtain a first position deviation parameter; collecting an image of substrate clamping, performing clamping deviation analysis to obtain a second position deviation parameter and an angle deviation parameter, and calculating a position deviation parameter by combining the first position deviation parameter; performing spin coating prediction according to the position deviation parameter and the angle deviation parameter to obtain the edge glue amount distribution after spin coating of the substrate; generating edge glue removal parameters according to the edge glue amount distribution to obtain an edge glue removal parameter distribution, compensating according to the position deviation parameter and the angle deviation parameter to obtain a compensated edge glue removal parameter distribution, and performing edge glue removal control on multiple edge regions of the substrate to complete coating. The present invention solves the technical problems of uneven spin coating and low coating quality of photoresist in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor technology, and in particular, to a coating control method, system, and device for photoresist. Background Art

[0002] Spin coating is a photoresist coating method widely used in semiconductor manufacturing. Its basic principle is to rotate the substrate at a high speed so that the photoresist spreads evenly under the action of centrifugal force and forms a thin film with a specific thickness. However, in the actual production process, the uniformity of photoresist coating is affected by various factors, resulting in the formation of photoresist accumulation at the edge of the substrate, and different amounts of accumulated glue, resulting in uneven photoresist in different areas of the substrate, affecting subsequent development and etching production.

[0003] Therefore, there are technical problems in the prior art such as uneven spin coating of photoresist and low coating quality. Summary of the Invention

[0004] In view of the technical problems of uneven spin coating of photoresist and low coating quality in the prior art, the present invention provides a coating control method, system, and device for photoresist.

[0005] The technical solutions of the present invention to solve the above technical problems are as follows:

[0006] In a first aspect, the present invention provides a coating control method for photoresist, including: collecting an image of a dispensing head, performing dispensing deviation analysis, and obtaining a first position deviation parameter;

[0007] Collecting an image of substrate clamping, performing clamping deviation analysis, obtaining a second position deviation parameter and an angle deviation parameter, and calculating a position deviation parameter by combining the first position deviation parameter;

[0008] Performing spin coating prediction according to the position deviation parameter and the angle deviation parameter to obtain the edge glue amount distribution after spin coating of the substrate;

[0009] Generating edge glue removal parameters according to the edge glue amount distribution to obtain an edge glue removal parameter distribution, compensating according to the position deviation parameter and the angle deviation parameter to obtain a compensated edge glue removal parameter distribution, and performing edge glue removal control on multiple edge regions of the substrate to complete coating.

[0010] In a second aspect, the present invention provides a coating control system for photoresist, including:

[0011] A dispensing deviation analysis module for collecting an image of a dispensing head, performing dispensing deviation analysis, and obtaining a first position deviation parameter;

[0012] The clamping deviation analysis module is used to collect images of the substrate clamping, perform clamping deviation analysis, obtain the second position deviation parameter and the angle deviation parameter, and calculate the position deviation parameter by combining the first position deviation parameter;

[0013] The edge glue amount prediction module is used to perform spin coating prediction according to the position deviation parameter and the angle deviation parameter, and obtain the edge glue amount distribution after spin coating of the substrate;

[0014] The edge glue removal control module is used to generate edge glue removal parameters according to the edge glue amount distribution, obtain the edge glue removal parameter distribution, perform compensation according to the position deviation parameter and the angle deviation parameter, obtain the compensated edge glue removal parameter distribution, and perform edge glue removal control on multiple edge regions of the substrate to complete coating.

[0015] In a third aspect, the present application provides an electronic device, which includes: a processor; a memory for storing executable instructions of the processor; wherein, the processor is used to execute a coating control method for photoresist provided by the present application.

[0016] The beneficial effects of the present invention are as follows: The present invention performs dispensing deviation analysis by collecting images of the dispensing head and clamping deviation analysis by collecting images of the substrate clamping, so as to obtain the first position deviation parameter, the second position deviation parameter and the angle deviation parameter. Compared with the traditional method that relies on the mechanical accuracy of the equipment and regular maintenance, this method can monitor and quantify the deviation of the dispensing head and substrate clamping in real time, improving the consistency and controllability of photoresist coating. Through the collection and analysis of the dispensing head images, the present invention can accurately identify the offset of the dispensing position, avoid the offset of the spin coating center caused by the blockage or position drift of the dispensing head, reduce the problems of uneven photoresist thickness and edge accumulation, and thus improve the uniformity of spin coating and the stability of subsequent processing. By detecting the offset amount and angle deviation of the substrate clamping, compensating for the uneven spin coating caused by improper clamping, and optimizing the overall distribution of the photoresist. Based on the spin coating prediction model of the position deviation parameter and the angle deviation parameter, the edge glue amount distribution after spin coating of the substrate can be predicted in advance. Compared with the traditional method of setting by experience, this method can adjust the edge glue removal strategy more accurately in a data-driven manner, reducing etching defects caused by excessive or insufficient edge photoresist accumulation. At the same time, according to the predicted edge glue amount distribution, edge glue removal parameters are adaptively generated, and compensation calculations are performed in combination with the position and angle deviation parameters to achieve differential glue removal control of multiple edge regions, so as to ensure the edge glue removal effect while avoiding insufficient photoresist caused by excessive glue removal, improving the integrity of the edge pattern and the etching accuracy. In summary, through multi-source image data analysis, intelligent prediction and adaptive compensation, this method can effectively improve the quality of photoresist coating, reduce edge defects, and improve the stability and accuracy of photoresist coating and semiconductor manufacturing processes. Description of the Drawings

[0017] Figure 1 A flowchart showing a method for controlling the coating of photoresist provided by the present invention.

[0018] Figure 2 A schematic structural diagram of a coating control system for photoresist provided by the present invention.

[0019] Figure 3 A schematic structural diagram of an electronic device provided by the present invention.

[0020] Reference numerals: glue dropping deviation analysis module 11, clamping deviation analysis module 12, edge glue amount prediction module 13, edge glue removal control module 14, electronic device 200, memory 210, processor 220, first computer program 211. Detailed implementation manners

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0022] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality of" means two or more, unless otherwise specifically defined.

[0023] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or having more advantages than other embodiments. In order for any person skilled in the art to implement and use the present invention, the following description is given. In the following description, details are set forth for purposes of explanation. It should be understood that those skilled in the art can recognize that the present invention can be implemented without using these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope that conforms to the principles and features disclosed in the present invention.

[0024] Embodiment 1, as Figure 1 shown, the embodiment of the present invention provides a method for controlling the coating of photoresist, and the method specifically includes the following steps:

[0025] S10: Collect an image of the dispensing head, perform dispensing deviation analysis, and obtain a first position deviation parameter.

[0026] In the embodiment of the present application, during the process of spin-coating photoresist, the photoresist is first evenly dispensed onto the center of the substrate by the dispensing head, and then diffuses to the entire surface of the substrate through rotation. If the dispensing head is partially blocked or deposits dry photoresist residues, it will cause limited flow of the photoresist, resulting in a decrease in the amount of dispensed photoresist or a deviation in the dropping position. After the dispensing deviates from the center, the spreading of the photoresist during the spin-coating process is uneven, resulting in significant differences in the thickness of the photoresist in different regions, and further leading to differences in the accumulation of the photoresist in different edge regions of the substrate. For example, the amount of photoresist accumulation is large in some edges and small in some edges.

[0027] Therefore, first collect an image of the dispensing head to perform dispensing deviation analysis and obtain a first position deviation parameter to obtain the deviation of the photoresist dispensing caused by partial blockage of the dispensing head, which is used as the data basis for subsequent analysis of the accumulation of the edge photoresist.

[0028] The step S10 in the method provided by the embodiment of the present application includes:

[0029] Collect an image of the dispensing head to obtain a dispensing image;

[0030] According to the historical data of photoresist spin-coating, collect a set of sample dispensing images, label the dispensing position deviation after dispensing in each sample dispensing image, and obtain a set of sample first position deviation parameters. Among them, the position deviation parameter includes an X-axis deviation parameter and a Y-axis deviation parameter;

[0031] Based on a convolutional neural network, construct a dispensing deviation recognizer;

[0032] Use the set of sample dispensing images and the set of sample first position deviation parameters to perform supervised training and testing on the dispensing deviation recognizer until the accuracy rate meets the accuracy threshold to complete the construction;

[0033] Input the dispensing image into the dispensing deviation recognizer to identify and obtain a first position deviation parameter.

[0034] In the embodiment of the present application, in order to accurately identify the deviation of the photoresist dispensing position, first collect an image of the dispensing head. Specifically, the image of the dispensing head can be collected in the front and left directions of the dispensing head. For example, a high-resolution industrial camera is used for collection to obtain a dispensing image.

[0035] Furthermore, machine learning technology is used to identify the dispensing image, and then identify the possible deviation situation after the dispensing head performs dispensing. Among them, first collect sample data to train the dispensing deviation recognizer for dispensing deviation recognition analysis.

[0036] According to the historical data of photoresist spin coating, images of the dispensing head are collected during the previous dispensing, and a set of sample dispensing images is obtained. Further, the deviation between the position of the dispensing head after dispensing and the standard dispensing position in each sample dispensing image is obtained and marked as the sample first position deviation parameter, and a set of sample first position deviation parameters is obtained. Among them, each first position deviation parameter includes an X-axis deviation parameter and a Y-axis deviation parameter.

[0037] Exemplarily, a coordinate system can be established at the standard dispensing position. For example, taking the standard dispensing position as the origin, an X-axis and a Y-axis coordinate system are constructed, and the X-axis and the Y-axis are perpendicular. Then, the position coordinates of the dispensing head after dispensing in each sample dispensing image are obtained as the position deviation parameter. For example, the standard dispensing position is the origin with coordinate values (0, 0), and the actual dispensing position is (0.12 mm, 0.15 mm), which is used as the first position deviation parameter.

[0038] Further, a dispensing deviation recognizer is constructed using a convolutional neural network (CNN, Convolutional Neural Network), which includes an input layer, multiple convolutional layers, multiple pooling layers, a fully connected layer, and an output layer. The convolutional layer uses a 3*3 convolutional kernel for image feature extraction, and the fully connected layer uses a mean squared error loss function to calculate the loss for training optimization.

[0039] The set of sample dispensing images and the set of sample first position deviation parameters are divided, for example, according to a ratio of 8:2, to obtain training data and test data. The training data is used to supervise the training of the dispensing deviation recognizer. The sample dispensing image is input, and the output first position deviation parameter is obtained. The error from the sample first position deviation parameter is calculated, and the loss is calculated through the mean squared error loss function. For example, the sum of the squared errors of the X-axis deviation parameter and the Y-axis deviation parameter is calculated as the loss. Then, the network parameters such as the weights of the fully connected layer are tuned according to the loss to reduce the loss until the accuracy meets the accuracy threshold. For example, if the accuracy meets or is greater than 95%, the test is qualified and the construction training is completed.

[0040] The currently collected dispensing image is input into the trained dispensing deviation recognizer, and the corresponding first position deviation parameter is recognized and output, such as the dispensing position deviation parameter of (0.12 mm, 0.15 mm).

[0041] Through the above steps, the embodiments of the present application can monitor in real time the deviation of the dispensing position caused by reasons such as dispensing head blockage during the dispensing process, use a convolutional neural network for recognition, achieve automatic and high-precision dispensing deviation recognition, thereby avoiding uneven spin coating caused by problems such as dispensing head blockage, improving the consistency of photoresist coating, and ensuring the accuracy and stability of subsequent processing.

[0042] S20: Collect an image of the substrate clamping, perform clamping deviation analysis, obtain a second position deviation parameter and an angle deviation parameter, and calculate a position deviation parameter by combining the first position deviation parameter.

[0043] In the embodiment of the present application, in the photoresist coating process, before the substrate enters the spin coating device, it needs to be fixed by a vacuum chuck or a mechanical fixture. The clamping deviation of the substrate will directly affect the uniformity of the photoresist after spin coating, and further affect the subsequent exposure and etching accuracy. At present, although regular maintenance and adjustment are carried out to ensure the accuracy of substrate clamping, due to inevitable errors, equipment wear and other problems, clamping offset may still occur, resulting in deviation of the dispensing position.

[0044] Therefore, the embodiment of the present application collects an image of the substrate clamping and performs deviation analysis caused by clamping deviation. Among them, clamping may also cause an angle between the substrate plane and the horizontal plane, resulting in deviation of the substrate angle, and further causing different angles in different regions of the substrate during spin coating, resulting in less photoresist accumulation in the region with a downward angle and more photoresist accumulation in the region with an upward angle, affecting the coating uniformity.

[0045] Perform clamping deviation analysis through the image of the substrate clamping, and obtain a second position deviation parameter and an angle deviation parameter caused by clamping.

[0046] The step S20 in the method provided by the embodiment of the present application includes:

[0047] Collect an image of the substrate clamping to obtain a substrate clamping image;

[0048] According to the historical data of substrate clamping, collect a set of sample substrate clamping images, label the position deviation and angle deviation of the substrate in each sample substrate clamping image, and obtain a set of sample second position deviation parameters and a set of sample angle deviation parameters. Among them, the angle deviation parameter includes the angle formed by the deviation of the substrate from the horizontal plane and the deviation direction;

[0049] Based on the convolutional neural network, construct a clamping deviation recognizer;

[0050] Use the set of sample substrate clamping images, the set of sample second position deviation parameters and the set of sample angle deviation parameters to perform supervised training and testing on the clamping deviation recognizer until the accuracy rate meets the accuracy rate threshold to complete the construction;

[0051] Input the substrate clamping image into the clamping deviation recognizer, and recognize and output a second position deviation parameter and an angle deviation parameter.

[0052] In the embodiments of the present application, during the photoresist coating process, the clamping deviation of the substrate may cause uneven thickness of the spin-coated photoresist. Therefore, it is necessary to detect and compensate for the clamping error through substrate clamping image analysis. For this purpose, first, an image of the substrate clamping is collected, and a clamping deviation recognizer is trained through a convolutional neural network (CNN) to identify the second position deviation parameter (i.e., the translational offset of the substrate in the X and Y directions) and the angular deviation parameter (i.e., the tilt angle and deviation direction of the substrate relative to the horizontal plane).

[0053] In the embodiments of the present application, an image of the substrate clamping is collected. For example, an industrial camera is used to collect images directly above, directly in front of, and to the left of the substrate clamping to obtain the substrate clamping image, and then the parameters of the clamping deviation of the substrate in different directions are analyzed.

[0054] Furthermore, according to the data of the substrate clamping within the historical time, the images taken during the previous substrate clamping are collected as sample substrate clamping images to obtain a set of sample substrate clamping images. And the position deviation and angular deviation of the substrate in each sample substrate clamping image are marked. For example, a coordinate system is constructed with the standard position as the origin, the coordinate system includes the X-axis and the Y-axis, and the coordinate values of the center of the substrate after actual clamping within the coordinate system are obtained, such as (-0.08 mm, -0.10 mm), as the sample second position deviation parameter. Then, a set of sample second position deviation parameters is obtained through marking. And the angle formed by the deviation of the plane where the substrate is located from the horizontal plane and the deviation direction in each sample substrate clamping image are marked. The formed angle is, for example, 5°, and the deviation direction is, for example, the direction in which the plane where the substrate is located forms an angle with the horizontal plane due to the warping of the substrate, such as the positive direction of the X-axis, or 45° deviating from the negative direction of the Y-axis in the positive direction of the X-axis. In this way, the sample angle and the sample deviation direction are marked as the sample angular deviation parameter, and a set of sample angular deviation parameters is obtained.

[0055] Furthermore, based on the convolutional neural network, a clamping deviation recognizer is constructed, which includes an input layer, multiple convolutional layers, multiple pooling layers, a fully connected layer, and an output layer. The convolutional layer uses a 3*3 convolutional kernel for image feature extraction, and the fully connected layer uses the mean squared error loss function to calculate the loss for training optimization.

[0056] Further, a supervised training and testing are performed on the clamped deviation identifier by using a set of sample substrate clamping images, a set of sample second position deviation parameters, and a set of sample angle deviation parameters. Specifically, each type of data in the set of sample substrate clamping images, the set of sample second position deviation parameters, and the set of sample angle deviation parameters is divided, for example, divided into training data and testing data according to a ratio of 8:2. Then, supervised training is performed using the training data. The sample substrate clamping image is input, and the output second position deviation parameter and angle deviation parameter are obtained. Then, the difference amplitude from the true sample second position deviation parameter and sample angle deviation parameter is calculated. Based on the loss function, the loss is calculated. Then, according to the loss, an adam optimizer is used to optimize the network parameters such as weights of the fully connected layer to reduce the loss and improve the accuracy. After iterative training and testing, if the accuracy meets the accuracy threshold, for example, is greater than or equal to 96%, the training is completed, and a clamped deviation identifier is constructed and obtained.

[0057] Based on the trained clamped deviation identifier, the currently acquired substrate clamping image is input. Through image convolution feature extraction and recognition, the output second position deviation parameter and angle deviation parameter are obtained as the analysis results of the position deviation and angle deviation caused by substrate clamping.

[0058] In the embodiment of the present application, by using deep learning, automatic and high-accuracy substrate position deviation and angle deviation can be realized. Furthermore, combined with the dispensing position deviation analyzed by the dispensing head, the possible position deviation of the photoresist in actual spin coating is analyzed, thereby providing a data basis for photoresist spin coating and edge degumming control and improving the quality of photoresist coating.

[0059] In the embodiment of the present application, after the first position deviation parameter caused by dispensing head blockage and the second position deviation parameter caused by clamping deviation are identified, a fusion of the two types of position deviation parameters is performed to obtain the final dispensing position deviation influence under the influence of multi-dimensional deviations.

[0060] Step S20 in the method provided by the embodiment of the present application further includes:

[0061] Extract the first X-axis deviation parameter, the first Y-axis deviation parameter, the second X-axis deviation parameter, and the second Y-axis deviation parameter within the first position deviation parameter and the second position deviation parameter;

[0062] According to the first X-axis deviation parameter, the first Y-axis deviation parameter, the second X-axis deviation parameter, and the second Y-axis deviation parameter, a combined X-axis deviation parameter and a combined Y-axis deviation parameter are calculated, and a position deviation parameter is integrally obtained.

[0063] In the embodiments of the present application, during the photoresist spin coating process, the deviation between the dispensing position and the substrate clamping position will affect the final uniformity of the photoresist. Therefore, it is necessary to comprehensively analyze the dispensing deviation and the substrate clamping deviation and calculate the final position deviation parameter.

[0064] In the embodiments of the present application, the X-axis deviation parameters and Y-axis deviation parameters within the first position deviation parameter and the second position deviation parameter are extracted to obtain the first X-axis deviation parameter, the first Y-axis deviation parameter, the second X-axis deviation parameter, and the second Y-axis deviation parameter.

[0065] Exemplarily, the first X-axis deviation parameter: 0.12 mm, the first Y-axis deviation parameter: 0.15 mm, that is, the actual dispensing position deviates 0.12 mm and 0.15 mm in the positive X-axis direction and the positive Y-axis direction from the standard position, and the second X-axis deviation parameter: -0.08 mm, the second Y-axis deviation parameter: -0.10 mm. That is, the actual dispensing position deviates 0.08 mm and 0.10 mm in the negative X-axis direction and the negative Y-axis direction from the standard position.

[0066] Further, based on the first X-axis deviation parameter, the first Y-axis deviation parameter, the second X-axis deviation parameter, and the second Y-axis deviation parameter, the combined X-axis deviation parameter and the combined Y-axis deviation parameter are calculated. For example, the sum of the first X-axis deviation parameter and the second X-axis deviation parameter is calculated as the combined X-axis deviation parameter, and the sum of the first Y-axis deviation parameter and the second Y-axis deviation parameter is calculated as the combined Y-axis deviation parameter. For example, the combined X-axis deviation parameter is 0.12 mm +-0.08 mm = 0.04 mm, and the combined Y-axis deviation parameter is 0.15 mm +-0.10 mm = 0.05 mm.

[0067] In the embodiments of the present application, by fusing the two deviation parameters, the comprehensive offset amount of the dispensing position relative to the standard position during the photoresist spin coating process can be accurately calculated, ensuring that subsequent edge degluing and spin coating compensation are more accurate.

[0068] S30: According to the position deviation parameter and the angle deviation parameter, perform spin coating prediction to obtain the edge glue amount distribution after the substrate is spin coated.

[0069] In the embodiments of the present application, the position offset of the epoxy resin on the substrate (i.e., the overall offset in the X-axis and Y-axis directions) and the angle offset of the substrate (i.e., the tilt angle and direction of the substrate) will directly affect the distribution of the photoresist after spin coating. For example, due to the centrifugal force during the spin coating process, the amount of photoresist accumulated at the substrate edge in the direction where the epoxy resin is deviated is larger during spin coating, and the amount of photoresist accumulated at the substrate edge of the upward-tilted part of the substrate is also larger during spin coating. Therefore, in order to perform edge desizing treatment for different amounts of photoresist accumulated at different edge regions and ensure the uniformity of photoresist coating, spin coating prediction is performed based on the obtained position deviation parameters and angle deviation parameters to predict the edge glue amounts of different edge regions of the substrate after spin coating, forming an edge glue amount distribution as a reference for edge desizing treatment.

[0070] Step S30 in the method provided by the embodiments of the present application includes:

[0071] According to the spin coating record data within the historical time, collect the sample position deviation parameter set and the sample angle deviation parameter set, and collect the edge glue amounts of multiple edge regions of the substrate after spin coating under different sample position deviation parameters and sample angle deviation parameters, and label to obtain the sample edge glue amount distribution set;

[0072] Adopt a multi-layer feedforward neural network to construct a spin coating predictor;

[0073] Use the sample position deviation parameter set, the sample angle deviation parameter set, and the sample edge glue amount distribution set to perform supervised training and testing on the spin coating predictor until the accuracy rate meets the accuracy threshold to complete the construction;

[0074] Input the position deviation parameter and the angle deviation parameter into the spin coating predictor, and predict and output the edge glue amount distribution.

[0075] In the embodiments of the present application, machine learning is used to perform spin coating prediction based on the position deviation parameter and the angle deviation parameter to predict the edge photoresist accumulation amounts of multiple edge regions of the substrate after spin coating, and obtain the edge glue amount distribution.

[0076] Among them, first, according to the data log of the substrate spin coating within the historical time, collect the sample position deviation parameters and sample angle deviation parameters monitored and recorded previously to form a sample position deviation parameter set and a sample angle deviation parameter set, which can be obtained through the steps in the above content.

[0077] Furthermore, collect the edge glue amounts of multiple edge regions of the substrate after spin coating the photoresist under different sample position deviation parameters and sample angle deviation parameters. When the spin coating speed is the same, label as the sample edge glue amount distribution according to the positions of multiple edge regions to obtain the sample edge glue amount distribution set.

[0078] Exemplarily, the edge region of the substrate can be divided to obtain multiple edge regions. For example, according to the four directions of the upper, lower, left, and right of the substrate, four edge regions are divided. For example, if the photoresist thicknesses at the edge positions of the four edge regions obtained by testing are 1.85 μm, 2.10 μm, 2.30 μm, and 1.70 μm respectively, then a sample edge glue amount distribution of (1.85 μm, 2.10 μm, 2.30 μm, 1.70 μm) is constructed. Optionally, more edge regions can also be divided, such as edge regions in 8 directions, and the photoresist thicknesses at the edges of the substrate in the eight directions are tested as the edge glue amount to form an edge glue amount distribution.

[0079] Furthermore, a spin coating predictor is constructed using a multi-layer feedforward neural network, which includes an input layer, a hidden layer, and an output layer. The dimension of the input layer is 4, which is used to input data of X-axis offset, Y-axis offset, tilt angle, and deviation direction. The dimension of the output layer is, for example, 4, which is used to output the photoresist stacking thicknesses of the four edge regions. The hidden layer includes two layers, which include 128 and 64 nodes respectively, and the ReLU activation function is used. The spin coating predictor calculates the loss using the mean squared error loss function.

[0080] Furthermore, a sample position deviation parameter set, a sample angle deviation parameter set, and a sample edge glue amount distribution set are used to supervise the training and testing of the spin coating predictor. First, the sample position deviation parameter set, the sample angle deviation parameter set, and the sample edge glue amount distribution set are divided. For example, they are divided into training data and test data according to a ratio of 8:2. The training data is used to supervise the training of the spin coating predictor, and the test data is used for testing. Among them, during the training process, the sample position deviation parameter and the sample angle deviation parameter are input to obtain the output edge glue amount distribution, and the error from the true sample edge glue amount distribution is calculated. The mean squared error loss function is used to calculate the loss. For example, the sum of the squares of the edge glue amount errors of the four edge regions is calculated as the loss, and the network parameters, such as the weights and biases of the nodes in the hidden layer, are adjusted and optimized according to the loss to reduce the loss and improve the accuracy until the test accuracy meets the accuracy threshold, for example, greater than or equal to 95%, then the construction training of the spin coating predictor is completed.

[0081] The currently recognized position deviation parameter and angle deviation parameter are input into the trained spin coating predictor to predict and output the edge glue amounts of multiple edge regions, such as the photoresist thicknesses of the four edge regions, as the edge glue amount distribution.

[0082] Based on machine learning, the embodiments of this application perform spin coating prediction according to the position deviation parameter and the angle deviation parameter, obtain the edge glue amount distribution affected by the drop glue deviation and the substrate angle deviation, and use it as the reference data for edge degluing processing control, thereby improving the uniformity and coating quality of the photoresist coating edge degluing process.

[0083] S40: Generate edge glue removal parameters based on the edge glue amount distribution to obtain an edge glue removal parameter distribution. Compensate according to the position deviation parameter and the angle deviation parameter to obtain a compensated edge glue removal parameter distribution, and perform edge glue removal control on multiple edge regions of the substrate to complete coating.

[0084] In the embodiment of the present application, in the photoresist spin coating process, due to the above-mentioned glue dropping deviation and substrate alignment deviation, the phenomenon of uneven photoresist thickness often occurs in different edge regions of the substrate, affecting the stability of subsequent processes. Therefore, based on the edge glue amount distribution, appropriate edge glue removal parameters are calculated, and combined with the position deviation parameter and the angle deviation parameter for compensation to generate a compensated edge glue removal parameter distribution, ultimately achieving precise edge glue removal control and ensuring the uniformity of photoresist coating.

[0085] Step S40 in the method provided by the embodiment of the present application includes:

[0086] Collect a sample edge glue amount set according to the edge glue removal historical data of spin coating, configure corresponding sample edge glue removal parameters, and label to obtain a sample edge glue removal parameter set;

[0087] Construct a mapping relationship between the sample edge glue amount set and the sample edge glue removal parameter set to obtain an edge glue removal decision table;

[0088] Input multiple edge glue amounts in the edge glue amount distribution into the edge glue removal decision table, and map to obtain edge glue removal parameters for multiple edge regions, generating an edge glue removal parameter distribution.

[0089] In the embodiment of the present application, from the edge glue removal historical data of spin coating in the historical spin coating process records, collect the photoresist thickness at the edge of the substrate during previous photoresist coating as the sample edge glue amount to obtain a sample edge glue amount set. Then, according to different sample edge glue amounts, configure corresponding sample edge glue removal parameters, such as the thickness of the removed photoresist or the solvent spraying amount. The solvent can melt the photoresist to achieve the purpose of glue removal. The larger the solvent spraying amount, the greater the thickness of the removed photoresist.

[0090] Exemplarily, the sample edge glue amount is 2.30 μm, and the configured sample edge glue removal parameter is -0.30 μm, that is, the glue removal amount required to reach the standard photoresist thickness (such as 2.0 μm). The corresponding solvent spraying amount is, for example, 3.5 μL. In this way, configure and label to obtain a sample edge glue removal parameter set. The configuration of the sample edge glue removal parameters can be configured by those skilled in the art of photoresist coating to ensure that the edge glue amount after edge glue removal meets the coating requirements. In this way, configure and label to obtain a sample edge glue removal parameter set.

[0091] Further, establish the mapping relationship between the sample edge glue amount set and the sample edge degumming parameter set to obtain the edge degumming decision table. In the edge degumming decision table, each sample edge glue amount corresponds to a sample edge degumming parameter.

[0092] Part of the data in the edge degumming decision table is shown in Table 1.

[0093] Edge glue amount (μm) Glue removal parameter (μm) 2.50 -0.50 2.30 -0.30 2.10 -0.10 1.90 0

[0094] Table 1

[0095] Further, input the multiple edge glue amounts of multiple edge regions in the currently predicted edge glue amount distribution into the edge degumming decision table for mapping to obtain the corresponding multiple edge degumming parameters, and combine the multiple edge regions to generate the edge degumming parameter distribution. For example, the edge degumming parameter distribution corresponding to the edge glue amount distribution of (1.85 μm, 2.10 μm, 2.30 μm, 1.70 μm) is (0μm, -0.10μm, -0.30μm, 0μm). After inputting the edge glue amount into the edge degumming decision table, if there is the same sample edge glue amount, map the corresponding sample edge degumming parameter; if there is no same sample edge glue amount, map the sample edge degumming parameter corresponding to the closest sample edge glue amount to obtain the edge degumming parameter distribution.

[0096] Through the above steps, the embodiments of the present application map to obtain the edge degumming parameter distribution, which is used as the basic data for subsequent edge degumming control to improve the uniformity of photoresist coating and the accuracy and stability of subsequent lithography processes.

[0097] In the embodiments of the present application, during the edge degumming process, if the degumming amount in the edge degumming parameters is too large, it may cause excessive removal of the photoresist and too small thickness of the photoresist, affecting subsequent lithography processing. In the foregoing content of the embodiments of the present application, machine learning and deep learning are used to identify the deviation of glue dropping and predict the spin coating of the photoresist, and there may be some errors resulting in the larger the edge degumming parameters. In order to avoid excessive edge degumming amount, it is necessary to compensate and adjust the edge degumming parameter distribution to avoid excessive degumming amount.

[0098] Step S40 in the method provided by the embodiments of the present application further includes:

[0099] Retrieve in the spin coating database according to the position deviation parameter and the angle deviation parameter to obtain the position deviation occurrence rate and the angle deviation occurrence rate, and calculate to obtain the deviation occurrence rate;

[0100] Obtain the average deviation occurrence rate of different position deviation parameters and angle deviation parameters in the spin coating database;

[0101] Judge whether the deviation occurrence rate is greater than or equal to the average deviation occurrence rate;

[0102] If so, no compensation is made for the edge desmearing parameter distribution; if not, calculate the difference amplitude between the deviation occurrence rate and the average deviation occurrence rate as the compensation adjustment parameter;

[0103] Obtain a preset compensation coefficient;

[0104] Use the compensation adjustment parameter to perform an adjustment calculation on the preset compensation coefficient to obtain a compensation coefficient;

[0105] Use the compensation coefficient to perform a compensation adjustment calculation on the edge desmearing parameter distribution to obtain a compensated edge desmearing parameter distribution;

[0106] According to multiple compensated edge desmearing parameters in the compensated edge desmearing parameter distribution, perform edge desmearing control on multiple edge regions of the substrate to complete coating.

[0107] In the embodiment of the present application, according to the historical data of photoresist spin coating within a historical time, use the method in the foregoing content to process and obtain a set of sample position deviation parameters and a set of sample angle deviation parameters, or actually measure to obtain a set of sample position deviation parameters and a set of sample angle deviation parameters to construct a spin coating database, which stores the sample position deviation parameters and sample angle deviation parameters generated during previous photoresist coating.

[0108] Further, according to the current position deviation parameter and angle deviation parameter, retrieve in the spin coating database the occurrence rate of the same position deviation parameter and angle deviation parameter. For example, the respective occurrence rates are 2% and 3%, which are used as the position deviation occurrence rate and the angle deviation occurrence rate. The larger the position deviation occurrence rate and the angle deviation occurrence rate, the greater the probability that the current position deviation parameter and angle deviation parameter appear in photoresist spin coating, the more representative the current position deviation parameter and angle deviation parameter are, and the higher the accuracy of the analysis and processing.

[0109] Further, according to the position deviation occurrence rate and the angle deviation occurrence rate, calculate the deviation occurrence rate. For example, calculate the mean value of the two as the deviation occurrence rate, such as 2.5%.

[0110] Further, obtain the average deviation occurrence rate of different position deviation parameters and angle deviation parameters in the spin coating database. For example, retrieve the deviation occurrence rate of each different set of position deviation parameters and angle deviation parameters in the spin coating database, and then calculate the mean value as the average deviation occurrence rate, such as 1%.

[0111] Further, it is determined whether the deviation occurrence rate is greater than or equal to the average deviation occurrence rate. If so, it indicates that the deviation occurrence rates of the current position deviation parameter and the angle deviation parameter are relatively large, belonging to common cases of dispensing deviation and substrate angle deviation, with a relatively large amount of corresponding sample data and a high accuracy rate for spin coating prediction. There is no need to compensate for the distribution of the edge debonding parameters, and edge debonding treatment can be directly performed.

[0112] If not, it indicates that the deviation occurrence rates of the current position deviation parameter and the angle deviation parameter are relatively small, belonging to rare cases of dispensing deviation and substrate angle deviation, with a relatively small amount of corresponding sample data and a low accuracy rate for spin coating prediction. It is necessary to compensate for the distribution of the edge debonding parameters to avoid excessive debonding amount during edge debonding treatment, which may result in too little photoresist remaining after edge debonding and affect lithography processing.

[0113] Exemplarily, calculate the difference amplitude between the deviation occurrence rate and the average deviation occurrence rate. For example, calculate the absolute difference between the deviation occurrence rate and the average deviation occurrence rate, and calculate the ratio of the absolute difference to the average deviation occurrence rate as the difference amplitude. For example, if the deviation occurrence rate is 0.8% and the average deviation occurrence rate is 1.0%, then the difference amplitude is 0.2% / 1.0% = 0.2, which is used as the compensation adjustment parameter.

[0114] Further, obtain a preset compensation coefficient, specifically a coefficient preset for compensating and adjusting the reduction of the edge debonding parameters. For example, it is 20%, that is, reduce the debonding amount within 20% of the edge debonding parameters, or reduce the spraying amount of the debonding solvent by 20%. For example, if the edge debonding parameter is -0.030μm, the edge debonding parameter after compensation and adjustment according to the preset compensation coefficient is -0.030μm * (1 - 20%) = -0.024μm.

[0115] Using this compensation adjustment parameter, perform adjustment calculations on the preset compensation coefficient to obtain the compensation coefficient. For example, multiply the sum of the compensation adjustment parameter and 1 by the preset compensation coefficient to complete the adjustment calculation and obtain the compensation coefficient. In this way, the fewer the position deviation parameter and the angle deviation parameter, the larger the corresponding compensation adjustment parameter, the larger the adjusted compensation coefficient, and the greater the amplitude of reducing the debonding amount, so as to avoid too much debonding resulting in too small a photoresist thickness. For example, if the compensation adjustment parameter is 0.2, perform adjustment calculations on the preset compensation coefficient, and the adjusted compensation coefficient is (1 + 0.2) * 20% = 24%.

[0116] Further, the adjusted compensation coefficient is used to perform compensation adjustment calculations on multiple edge degluing parameters within the edge degluing parameter distribution. For example, 1 - 24% = 76% is used to multiply each edge degluing parameter to complete the compensation adjustment and obtain the compensated edge degluing parameter distribution. For example, if the edge degluing parameter is -0.030 μm, then the compensated edge degluing parameter after compensation is -0.030 μm * 76% = -0.0228 μm.

[0117] Further, according to the multiple compensated edge degluing parameters within the compensated edge degluing parameter distribution after compensation adjustment, edge degluing control is respectively performed on multiple edge regions of the substrate, such as removing photoresist with different thicknesses to complete coating. Exemplarily, after spin coating is completed, multiple compensated edge degluing parameters can be used to perform edge degluing treatment on different edge regions of the substrate to complete photoresist spin coating and edge degluing treatment, and complete photoresist coating.

[0118] A method for controlling the coating of photoresist provided by an embodiment of the present invention has at least the following technical effects: By collecting images of the dispensing head for dispensing deviation analysis and collecting images of the substrate clamping for clamping deviation analysis, the first position deviation parameter, the second position deviation parameter, and the angle deviation parameter are obtained. Compared with the traditional method that relies on the mechanical accuracy of the equipment and regular maintenance, this method can monitor and quantify the deviation of the dispensing head and the substrate clamping in real time, improving the consistency and controllability of photoresist coating. Through the image collection and analysis of the dispensing head, the offset of the dispensing position can be accurately identified, avoiding the offset of the spin coating center caused by the blockage or position drift of the dispensing head, reducing the problems of uneven photoresist thickness and edge accumulation, and thus improving the uniformity of spin coating and the stability of subsequent processing. Detecting the offset amount and angle deviation of the substrate clamping, compensating for the uneven spin coating caused by improper clamping, and optimizing the overall distribution of the photoresist. Based on the spin coating prediction model of the position deviation parameter and the angle deviation parameter, the edge glue amount distribution after the substrate is spin coated can be predicted in advance. Compared with the traditional method of setting by experience, this method can more accurately adjust the edge degluing strategy in a data-driven manner, reducing etching defects caused by excessive or insufficient edge photoresist accumulation. At the same time, according to the predicted edge glue amount distribution, the edge degluing parameters are adaptively generated, and compensation calculations are combined with the position and angle deviation parameters to achieve differential degluing control of multiple edge regions, thereby ensuring the edge degluing effect while avoiding insufficient photoresist caused by excessive degluing and improving the integrity of the edge pattern and the etching accuracy. In summary, this method can effectively improve the quality of photoresist coating, reduce edge defects, and improve the stability and accuracy of photoresist coating and semiconductor manufacturing processes through multi-source image data analysis, intelligent prediction, and adaptive compensation.

[0119] Example 2, as Figure 2As shown, with the same inventive concept as a coating control method for photoresist in Embodiment 1, the embodiment of the present invention also provides a coating control system for photoresist. The explanations in Embodiment 1 for a coating control method for photoresist are also applicable to a coating control system for photoresist. The system includes:

[0120] A glue dropping deviation analysis module 11, configured to collect an image of the glue dropping head, perform glue dropping deviation analysis, and obtain a first position deviation parameter;

[0121] A clamping deviation analysis module 12, configured to collect an image of the substrate clamping, perform clamping deviation analysis, obtain a second position deviation parameter and an angle deviation parameter, and calculate a position deviation parameter in combination with the first position deviation parameter;

[0122] An edge glue amount prediction module 13, configured to perform spin coating prediction according to the position deviation parameter and the angle deviation parameter, and obtain the edge glue amount distribution after spin coating of the substrate;

[0123] An edge glue removal control module 14, configured to generate edge glue removal parameters according to the edge glue amount distribution, obtain an edge glue removal parameter distribution, perform compensation according to the position deviation parameter and the angle deviation parameter, obtain a compensated edge glue removal parameter distribution, and perform edge glue removal control on multiple edge regions of the substrate to complete coating.

[0124] Furthermore, a coating control system for photoresist is also used for:

[0125] Collect an image of the glue dropping head to obtain a glue dropping image;

[0126] According to the photoresist spin coating historical data, collect a set of sample glue dropping images, label the glue dropping position deviation after glue dropping in each sample glue dropping image, and obtain a set of sample first position deviation parameters, where the position deviation parameter includes an X-axis deviation parameter and a Y-axis deviation parameter;

[0127] Build a glue dropping deviation recognizer based on a convolutional neural network;

[0128] Use the set of sample glue dropping images and the set of sample first position deviation parameters to perform supervised training and testing on the glue dropping deviation recognizer until the accuracy rate meets the accuracy rate threshold to complete the construction;

[0129] Input the glue dropping image into the glue dropping deviation recognizer to identify and obtain a first position deviation parameter.

[0130] Furthermore, a coating control system for photoresist is also used for:

[0131] Collect an image of the substrate clamping to obtain a substrate clamping image;

[0132] According to the substrate clamping historical data, collect a set of sample substrate clamping images, label the position deviation and angle deviation of the substrate in each sample substrate clamping image to obtain a set of sample second position deviation parameters and a set of sample angle deviation parameters, where the angle deviation parameters include the included angle formed by the deviation of the substrate from the horizontal plane and the deviation direction;

[0133] Based on a convolutional neural network, construct a clamping deviation recognizer;

[0134] Use the set of sample substrate clamping images, the set of sample second position deviation parameters and the set of sample angle deviation parameters to perform supervised training and testing on the clamping deviation recognizer until the accuracy rate meets the accuracy threshold to complete the construction;

[0135] Input the substrate clamping image into the clamping deviation recognizer, and recognize and output the second position deviation parameters and the angle deviation parameters.

[0136] Furthermore, a coating control system for photoresist is also used for:

[0137] Extract the first X-axis deviation parameter, the first Y-axis deviation parameter, the second X-axis deviation parameter and the second Y-axis deviation parameter in the first position deviation parameter and the second position deviation parameter;

[0138] According to the first X-axis deviation parameter, the first Y-axis deviation parameter, the second X-axis deviation parameter and the second Y-axis deviation parameter, calculate and obtain the combined X-axis deviation parameter and the combined Y-axis deviation parameter, and integrate to obtain the position deviation parameter.

[0139] Furthermore, a coating control system for photoresist is also used for:

[0140] According to the spin coating record data within the historical time, collect a set of sample position deviation parameters and a set of sample angle deviation parameters, and collect the edge glue amounts of multiple edge regions of the substrate after spin coating under different sample position deviation parameters and sample angle deviation parameters, and label to obtain a set of sample edge glue amount distributions;

[0141] Use a multi-layer feedforward neural network to construct a spin coating predictor;

[0142] Use the set of sample position deviation parameters, the set of sample angle deviation parameters, and the set of sample edge glue amount distributions to perform supervised training and testing on the spin coating predictor until the accuracy rate meets the accuracy threshold to complete the construction;

[0143] Input the position deviation parameter and the angle deviation parameter into the spin coating predictor, and predict and output the edge glue amount distribution.

[0144] Furthermore, a coating control system for photoresist is also used for:

[0145] Collect a set of sample edge glue amounts according to the historical data of edge glue removal during spin coating, configure corresponding sample edge glue removal parameters, and label to obtain a set of sample edge glue removal parameters;

[0146] Construct the mapping relationship between the set of sample edge glue amounts and the set of sample edge glue removal parameters to obtain an edge glue removal decision table;

[0147] Input multiple edge glue amounts within the edge glue amount distribution into the edge glue removal decision table, map to obtain edge glue removal parameters for multiple edge regions, and generate an edge glue removal parameter distribution.

[0148] Furthermore, a coating control system for photoresist is also used for:

[0149] Retrieve in the spin coating database according to the position deviation parameter and the angle deviation parameter to obtain the position deviation occurrence rate and the angle deviation occurrence rate, and calculate to obtain the deviation occurrence rate;

[0150] Obtain the average deviation occurrence rate of different position deviation parameters and angle deviation parameters in the spin coating database;

[0151] Judge whether the deviation occurrence rate is greater than or equal to the average deviation occurrence rate;

[0152] If so, do not compensate the edge glue removal parameter distribution. If not, calculate the difference amplitude between the deviation occurrence rate and the average deviation occurrence rate as the compensation adjustment parameter;

[0153] Obtain a preset compensation coefficient;

[0154] Use the compensation adjustment parameter to perform adjustment calculation on the preset compensation coefficient to obtain a compensation coefficient;

[0155] Use the compensation coefficient to perform compensation adjustment calculation on the edge glue removal parameter distribution to obtain a compensated edge glue removal parameter distribution;

[0156] Perform edge glue removal control on multiple edge regions of the substrate according to multiple compensated edge glue removal parameters in the compensated edge glue removal parameter distribution to complete coating.

[0157] Example three, please refer to Figure 3 , Figure 3 is a schematic diagram of an embodiment of an electronic device provided by an embodiment of the present invention. As Figure 3 shown, an electronic device 200 provided by an embodiment of the present invention includes a memory 210, a processor 220, and a first computer program 211 stored on the memory 210 and executable on the processor 220. When the processor 220 executes the first computer program 211, it implements a coating control method for photoresist.

[0158] It should be noted that in the above embodiments, the descriptions of the various embodiments have their own focuses. For the parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0159] Those skilled in the art should understand that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0160] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded computers, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices generate means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0161] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0162] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operating steps are performed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0163] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic creative concepts.

[0164] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalent technologies, the present invention also intends to include these changes and modifications.

Claims

1. A coating control method for photoresist, characterized in that The method includes: Collecting an image of the dispensing head, performing dispensing deviation analysis, and obtaining a first position deviation parameter; Collecting an image of the substrate clamping, performing clamping deviation analysis, obtaining a second position deviation parameter and an angle deviation parameter, and calculating a position deviation parameter based on the first position deviation parameter and the second position deviation parameter; Performing spin coating prediction based on the position deviation parameter and the angle deviation parameter to obtain the edge glue amount distribution after spin coating of the substrate; Generating edge glue removal parameters based on the edge glue amount distribution to obtain an edge glue removal parameter distribution, compensating according to the position deviation parameter and the angle deviation parameter to obtain a compensated edge glue removal parameter distribution, and performing edge glue removal control on multiple edge regions of the substrate to complete coating, including: Retrieving in the spin coating database according to the position deviation parameter and the angle deviation parameter to obtain a position deviation occurrence rate and an angle deviation occurrence rate, and calculating a deviation occurrence rate; Obtaining the average deviation occurrence rate of different position deviation parameters and angle deviation parameters in the spin coating database; Determining whether the deviation occurrence rate is greater than or equal to the average deviation occurrence rate; If so, no compensation is performed on the edge glue removal parameter distribution; if not, calculating the difference amplitude between the deviation occurrence rate and the average deviation occurrence rate as a compensation adjustment parameter; Obtaining a preset compensation coefficient; Adjusting and calculating the preset compensation coefficient by using the compensation adjustment parameter to obtain a compensation coefficient; Performing compensation adjustment calculation on the edge glue removal parameter distribution by using the compensation coefficient to obtain a compensated edge glue removal parameter distribution; Performing edge glue removal control on multiple edge regions of the substrate according to multiple compensated edge glue removal parameters in the compensated edge glue removal parameter distribution to complete coating, wherein the spin coating database stores sample position deviation parameters and sample angle deviation parameters obtained by processing historical data of photoresist spin coating within a historical time, or sample position deviation parameters and sample angle deviation parameters obtained by actual measurement.

2. The coating control method for photoresist according to claim 1, wherein Collecting an image of the dispensing head, performing dispensing deviation analysis, and obtaining a first position deviation parameter, including: Collecting an image of the dispensing head to obtain a dispensing image; Collecting a set of sample dispensing images according to the historical data of photoresist spin coating, and marking the dispensing position deviation after dispensing in each sample dispensing image to obtain a set of sample first position deviation parameters, wherein the first position deviation parameter includes a first X-axis deviation parameter and a first Y-axis deviation parameter; Constructing a dispensing deviation identifier based on a convolutional neural network; Supervising, training, and testing the dispensing deviation identifier by using the set of sample dispensing images and the set of sample first position deviation parameters until the accuracy rate meets the accuracy threshold to complete the construction; Inputting the dispensing image into the dispensing deviation identifier to identify and obtain the first position deviation parameter.

3. The coating control method for photoresist according to claim 1, wherein, Collecting an image of the substrate clamping, performing clamping deviation analysis, and obtaining a second position deviation parameter and an angle deviation parameter, including: Collecting an image of the substrate clamping to obtain a substrate clamping image; According to the substrate clamping historical data, collect a set of sample substrate clamping images, label the substrate position deviation and angle deviation in each sample substrate clamping image, and obtain a set of sample second position deviation parameters and a set of sample angle deviation parameters. Among them, the angle deviation parameter includes the included angle formed by the deviation of the substrate from the horizontal plane and the deviation direction. Based on the convolutional neural network, construct a clamping deviation recognizer. Use the set of sample substrate clamping images, the set of sample second position deviation parameters, and the set of sample angle deviation parameters to perform supervised training and testing on the clamping deviation recognizer until the accuracy rate meets the accuracy threshold, and complete the construction. Input the substrate clamping image into the clamping deviation recognizer, and recognize and output the second position deviation parameter and the angle deviation parameter.

4. The coating control method for photoresist according to claim 1, wherein According to the first position deviation parameter and the second position deviation parameter, calculate and obtain the position deviation parameter, including: Extract the first X-axis deviation parameter, the first Y-axis deviation parameter, the second X-axis deviation parameter, and the second Y-axis deviation parameter in the first position deviation parameter and the second position deviation parameter. According to the first X-axis deviation parameter, the first Y-axis deviation parameter, the second X-axis deviation parameter, and the second Y-axis deviation parameter, calculate and obtain the combined X-axis deviation parameter and the combined Y-axis deviation parameter, and integrate to obtain the position deviation parameter.

5. The coating control method for photoresist according to claim 1, wherein According to the position deviation parameter and the angle deviation parameter, perform spin coating prediction to obtain the edge glue amount distribution after spin coating of the substrate, including: According to the spin coating record data within the historical time, collect a set of sample position deviation parameters and a set of sample angle deviation parameters, and collect the edge glue amounts of multiple edge regions of the substrate after spin coating under different sample position deviation parameters and sample angle deviation parameters, and label to obtain a set of sample edge glue amount distributions. Use the multi-layer feedforward neural network to construct a spin coating predictor. Use the set of sample position deviation parameters, the set of sample angle deviation parameters, and the set of sample edge glue amount distributions to perform supervised training and testing on the spin coating predictor until the accuracy rate meets the accuracy threshold, and complete the construction. Input the position deviation parameter and the angle deviation parameter into the spin coating predictor, and predict and output the edge glue amount distribution.

6. The coating control method for photoresist according to claim 1, wherein, According to the edge glue amount distribution, generate edge degluing parameters to obtain the edge degluing parameter distribution, including: According to the edge degluing historical data of spin coating, collect a set of sample edge glue amounts, configure corresponding sample edge degluing parameters, and label to obtain a set of sample edge degluing parameters. Construct the mapping relationship between the set of sample edge glue amounts and the set of sample edge degluing parameters to obtain an edge degluing decision table. Input the multiple edge glue amounts in the edge glue amount distribution into the edge degluing decision table, and map to obtain the edge degluing parameters of multiple edge regions, and generate the edge degluing parameter distribution.

7. A coating control system for photoresist, characterized in that, The system is used to execute the method according to any one of claims 1-6. The system includes: A glue dropping deviation analysis module, which is used to collect images of the glue dropping head, perform glue dropping deviation analysis, and obtain the first position deviation parameter. The clamping deviation analysis module is used to collect the images of the substrate clamping, conduct clamping deviation analysis, obtain the second position deviation parameter and the angle deviation parameter, and calculate the position deviation parameter according to the first position deviation parameter and the second position deviation parameter; The edge glue amount prediction module is used to conduct spin coating prediction according to the position deviation parameter and the angle deviation parameter, and obtain the edge glue amount distribution after the substrate is spin coated; The edge glue removal control module is used to generate edge glue removal parameters according to the edge glue amount distribution to obtain the edge glue removal parameter distribution, conduct compensation according to the position deviation parameter and the angle deviation parameter to obtain the compensated edge glue removal parameter distribution, and conduct edge glue removal control on multiple edge areas of the substrate to complete coating.

8. An electronic device, characterized in that, The electronic device includes: A memory for storing computer software programs; A processor for reading and executing the computer software program, thereby implementing a coating control method for photoresist according to any one of claims 1-6.

Images