NCVM non-conductive vacuum coating process execution monitoring method and system

Abstract

The invention discloses an NCVM non-conductive vacuum coating process execution monitoring method and system, and relates to the related field of vacuum coating, and the method comprises the steps: carrying out the dynamic division of detection unit grids on the surface of a substrate; the surface of the self-adaptive grid base material is scanned in real time through an infrared thermal imager, and a dynamic temperature distribution matrix of the base material is obtained; according to the mapping relation between the coating thickness and the temperature gradient, predicting the dynamic temperature distribution matrix of the base material to obtain a coating uniformity prediction map; carrying out fusion processing on the dynamic temperature distribution matrix of the base material and the coating uniformity prediction map to obtain a cross validation grid; when the verification of any grid is not passed, generating a three-dimensional compensation vector; and the process parameters of the coating equipment are regulated and controlled in real time. The technical problems that the prediction of the coating uniformity is not accurate enough and the process parameters are difficult to regulate and control accurately in real time in the existing coating process execution monitoring are solved, and the technical effects of accurately predicting the coating uniformity and regulating and controlling the process parameters accurately in real time are achieved.

Description

Technical Field

[0001] This application relates to the field of vacuum coating, and in particular to a method and system for monitoring the execution of NCVM non-conductive vacuum coating process. Background Technology

[0002] In vacuum coating processes, coating uniformity is a key factor affecting product quality, especially for coatings on non-conductive materials (NCVM). Poor uniformity can lead to inconsistent film properties, impacting the product's appearance and functionality. Currently, the industry primarily assesses coating uniformity indirectly by monitoring the temperature distribution on the substrate surface and adjusts process parameters accordingly. However, existing methods rely solely on static analysis of temperature distribution, neglecting the dynamic correlation between temperature change rate and film thickness deviation. This results in inaccurate predictions of coating uniformity and makes it difficult to control process parameters accurately in real time, thus affecting coating quality.

[0003] Currently, the NCVM non-conductive vacuum coating process monitoring suffers from inaccurate prediction of coating uniformity, making it difficult to control process parameters in real time and with precision. Summary of the Invention

[0004] This application provides a method and system for monitoring the execution of an NCVM (Non-Conductive Vacuum Coating) process. The method involves dynamically dividing the substrate surface within the vacuum coating chamber into a grid of detection units to obtain an adaptive grid substrate surface. An infrared thermal imager is used to scan the adaptive grid substrate surface in real time, generating a dynamic temperature distribution matrix of the substrate. Based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix is ​​analyzed and predicted to generate a coating uniformity prediction map. The dynamic temperature distribution matrix and the coating uniformity prediction map are aligned and fused to generate a cross-validation grid. Each grid in the cross-validation grid is validated. If the validation fails, a three-dimensional compensation vector (including plasma gun angle, sputtering power, and residence time) is generated. The process parameters of the vacuum coating equipment are adjusted in real time based on the three-dimensional compensation vector to optimize coating quality. This method solves the technical problem of inaccurate prediction of coating uniformity in existing NCVM non-conductive vacuum coating process monitoring, which makes it difficult to adjust process parameters in real time and accurately. It achieves the technical effect of accurately predicting coating uniformity, thereby enabling real-time and precise control of process parameters.

[0005] This application provides a method for monitoring the execution of a non-conductive vacuum coating process using NCVM (Non-Conductive Vacuum Coating), comprising: dynamically dividing the surface of a substrate within a vacuum coating cavity into a detection unit grid to obtain an adaptive grid substrate surface; scanning the adaptive grid substrate surface in real time using an infrared thermal imager to obtain a dynamic temperature distribution matrix of the substrate; analyzing and predicting the dynamic temperature distribution matrix of the substrate based on the mapping relationship between coating thickness and temperature gradient to obtain a coating uniformity prediction map; aligning and fusing the dynamic temperature distribution matrix of the substrate and the coating uniformity prediction map to obtain a cross-validation grid; generating a three-dimensional compensation vector when any grid in the cross-validation grid fails verification; and adjusting the process parameters of the vacuum coating equipment in real time based on the three-dimensional compensation vector.

[0006] In one possible implementation, the substrate surface inside the vacuum coating cavity is dynamically divided into detection unit grids to obtain an adaptive grid substrate surface. The following processes are then performed: a conversion relationship between the surface curvature of the substrate and the grid density is preset; multiple surface curvatures of the current substrate are acquired in real time, and the adaptive grid density is calculated based on the conversion relationship; the substrate surface inside the vacuum coating cavity is dynamically divided into detection unit grids based on the adaptive grid density, and the geometric center point of each grid is set as a temperature detection reference point to obtain the adaptive grid substrate surface.

[0007] In a possible implementation, the substrate surface inside the vacuum coating cavity is dynamically divided into detection unit grids to obtain an adaptive grid substrate surface. The following processing is also performed: the temperature change rate is extracted based on the substrate dynamic temperature distribution matrix to obtain a substrate dynamic temperature change rate matrix; the substrate dynamic temperature change rate matrix is ​​traversed to extract unit grids with temperature change rates greater than a preset temperature change rate threshold to obtain temperature anomaly grids; the temperature anomaly grids are encrypted and reconstructed to update the adaptive grid substrate surface.

[0008] In a possible implementation, based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map. The following processing is performed: historical coating data is acquired, including historical substrate dynamic temperature distribution matrix and historical film thickness deviation value matrix, wherein the film thickness deviation value refers to the difference between the actual film thickness and the standard film thickness; the temperature gradient of adjacent grids is calculated based on the historical substrate dynamic temperature distribution matrix to obtain a historical substrate temperature gradient distribution matrix; the historical substrate temperature gradient distribution matrix and the historical film thickness deviation value matrix are input into a convolutional neural network model for training and learning to obtain the mapping relationship between coating thickness and temperature gradient; based on the mapping relationship, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map.

[0009] In a possible implementation, the historical substrate temperature gradient distribution matrix and the historical film thickness deviation matrix are input into a convolutional neural network model for training to obtain the mapping relationship between coating thickness and temperature gradient. The following processing is performed: a dual-branch convolutional neural network model is constructed, wherein the dual-branch convolutional neural network model includes a first branch and a second branch; the first branch receives the historical substrate temperature gradient distribution matrix and extracts spatial features through a convolution kernel; the second branch receives the historical substrate temperature gradient distribution matrix at the same position in the previous N production batches and extracts temporal features through an LSTM layer, wherein N≥3; the spatial features and the temporal features are weighted and concatenated to obtain a film thickness deviation prediction matrix; when the mean square error of the film thickness deviation prediction matrix and the historical film thickness deviation matrix is ​​lower than a preset mean square error threshold and the prediction accuracy reaches a preset accuracy threshold, training is stopped and the mapping relationship between coating thickness and temperature gradient is derived.

[0010] In a possible implementation, when any grid in the cross-validation grid fails validation, a three-dimensional compensation vector is generated, and the following processing is performed: absolute temperature and temperature rise rate are extracted from the cross-validation grid; if the absolute temperature does not meet a preset temperature range and / or the temperature rise rate does not meet a preset temperature rise rate range, the corresponding grid is marked as a dangerous grid; the film thickness deviation prediction value is extracted from the dangerous grid and cross-validated; if the film thickness deviation prediction value does not meet a preset film thickness deviation range, the validation fails, and a three-dimensional compensation vector is generated.

[0011] In a possible implementation, the following processing is performed: if the verification fails, the temperature-film thickness co-verification condition must also be met. The co-verification condition is as follows: calculate the temperature anomaly degree based on the substrate material, the absolute temperature, and the temperature rise rate; calculate the film thickness anomaly degree based on the substrate material and the predicted film thickness deviation; multiply the temperature anomaly degree and the film thickness anomaly degree to obtain a co-verification factor; if the temperature anomaly degree is greater than a preset temperature anomaly degree threshold, the film thickness anomaly degree is greater than a preset film thickness anomaly threshold, and the co-verification factor is greater than a preset co-verification factor threshold, then the verification is determined to fail.

[0012] In a possible implementation, when any grid in the cross-validation grid fails validation, a three-dimensional compensation vector is generated, and the following processing is performed: extracting temperature anomaly information and film thickness anomaly information from the grid that failed validation; calculating the angle compensation amount of the plasma gun based on the temperature anomaly information; calculating the sputtering power compensation amount and the residence time compensation amount based on the film thickness anomaly information; and integrating the angle compensation amount, the sputtering power compensation amount, and the residence time compensation amount to generate the three-dimensional compensation vector.

[0013] In a possible implementation, the following processing is also performed: calculating an edge compensation coefficient based on the radius of curvature of the mesh in the edge region of the substrate; adjusting the three-dimensional compensation vector according to the edge compensation coefficient to obtain a three-dimensional strengthening compensation vector; after compensating the mesh in the edge region according to the three-dimensional strengthening compensation vector, initiating a secondary edge verification, re-measuring the mesh in the edge region using a laser thickness gauge, and obtaining the re-measuring result; and optimizing the coating of the mesh in the edge region based on the re-measuring result.

[0014] This application also provides an NCVM non-conductive vacuum coating process execution monitoring system, comprising: a detection unit grid dynamic division module for dynamically dividing the substrate surface in the vacuum coating cavity into detection unit grids to obtain an adaptive grid substrate surface; an infrared thermal imaging scanning module for real-time scanning of the adaptive grid substrate surface using an infrared thermal imager to obtain a substrate dynamic temperature distribution matrix; a coating uniformity prediction module for analyzing and predicting the substrate dynamic temperature distribution matrix based on the mapping relationship between coating thickness and temperature gradient to obtain a coating uniformity prediction map; an alignment and fusion processing module for aligning and fusion processing the substrate dynamic temperature distribution matrix and the coating uniformity prediction map to obtain a cross-validation grid; a three-dimensional compensation vector generation module for generating a three-dimensional compensation vector when any grid in the cross-validation grid fails verification; and a process parameter control module for real-time control of the process parameters of the vacuum coating equipment based on the three-dimensional compensation vector.

[0015] This application proposes a method and system for monitoring the execution of an NCVM (Non-Conductive Vacuum Coating) process. First, the surface of the substrate within the vacuum coating chamber is dynamically divided into a detection unit grid to obtain an adaptive grid substrate surface. Then, an infrared thermal imager scans the adaptive grid substrate surface in real time to obtain a dynamic temperature distribution matrix. Next, based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix is ​​analyzed and predicted to obtain a coating uniformity prediction map. Then, the dynamic temperature distribution matrix and the coating uniformity prediction map are aligned and fused to obtain a cross-validation grid. When any grid in the cross-validation grid fails verification, a three-dimensional compensation vector is generated. Finally, the process parameters of the vacuum coating equipment are adjusted in real time based on the three-dimensional compensation vector. This achieves the technical effect of accurately predicting coating uniformity, thereby enabling real-time and precise control of process parameters. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a flowchart illustrating a method for monitoring the execution of an NCVM non-conductive vacuum coating process, provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of a monitoring system for the execution of a non-conductive vacuum coating process (NCVM) provided in an embodiment of this application.

[0019] Figure labeling: Detection unit grid dynamic division module 10, infrared thermal imaging scanning module 20, coating uniformity prediction module 30, alignment and fusion processing module 40, three-dimensional compensation vector generation module 50, process parameter control module 60. Detailed Implementation

[0020] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below.

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0023] This application provides a method for monitoring the execution of an NCVM non-conductive vacuum coating process, such as... Figure 1 As shown, the method includes:

[0024] Step S100: Dynamically divide the substrate surface in the vacuum coating cavity into a detection unit grid to obtain an adaptive grid substrate surface.

[0025] Specifically, multiple high-precision position sensors and optical sensors are installed within the vacuum coating chamber to detect the shape, size, and position information of the substrate surface in real time. The sensor array is activated to acquire image data of the substrate surface. The acquired image data is transmitted to an image processing system, which uses computer vision technology to process the images acquired by the sensors (such as noise reduction and contrast enhancement). Based on the shape and size of the substrate, the system adaptively adjusts the size and distribution of the grid to ensure that each grid accurately reflects the local features of the substrate surface.

[0026] For example, assuming the substrate is an irregularly shaped plastic part, after the sensor array acquires an image of its surface, the image processing system divides the substrate surface into multiple grid cells of varying sizes based on the part's contour and surface features. If the substrate surface has complex curves, the grid will be distributed more densely in the curved areas to ensure accurate capture of local variations.

[0027] In one possible implementation, the surface of the substrate within the vacuum coating cavity is dynamically divided into detection unit grids to obtain an adaptive grid substrate surface. Step S100 further includes step S110, which presets the conversion relationship between the surface curvature of the substrate and the grid density. Specifically, the influence of grid density on coating uniformity under different surface curvatures is studied experimentally. Based on the experimental data, a mathematical model is established to describe the quantitative relationship between surface curvature and grid density. For example, when the surface curvature increases within a certain range, the grid density increases linearly; when the surface curvature exceeds a certain threshold, the rate of increase in grid density accelerates. This relationship can be described by a piecewise function or a nonlinear function, the specific form of which is determined based on the experimental data.

[0028] Step S120: The curvature of multiple surfaces of the current substrate is acquired in real time, and the adaptive mesh density is calculated based on the transformation relationship. Specifically, multiple high-precision sensors (such as optical sensors or laser scanners) are installed in the vacuum coating chamber to measure the surface curvature of the substrate in real time. The sensor array is activated to measure the surface curvature of multiple points on the substrate surface in real time. The surface curvature data collected by the sensors is transmitted to the data processing system, which calculates the adaptive mesh density corresponding to each measurement point according to the preset transformation relationship.

[0029] For example, the substrate surface has multiple regions, each with a different surface curvature. After the sensor array measures the surface curvature of these regions, the data processing system calculates the mesh density for each region according to a preset transformation relationship. For instance, if a region has a higher surface curvature, the mesh density calculated according to the transformation relationship will also be higher; if another region has a lower surface curvature, the mesh density will be lower.

[0030] Step S130: The substrate surface within the vacuum coating cavity is dynamically divided into detection unit grids according to the adaptive grid density. The geometric center point of each grid is set as a temperature detection reference point, resulting in an adaptive grid substrate surface. Specifically, the substrate surface is dynamically divided using image processing and grid division algorithms based on the calculated adaptive grid density. For each divided grid, its geometric center point is calculated and set as a temperature detection reference point. For example, the substrate surface is divided into multiple grids, and the size and shape of each grid are determined according to the adaptive grid density. For each grid, the method for calculating its geometric center point is: find the grid boundary and calculate the average position of all points within the boundary. This average position is the geometric center point of the grid, which is used as the temperature detection reference point.

[0031] This approach, by dynamically dividing the detection unit grid and adjusting the grid density according to the surface curvature, can capture changes on the substrate surface more precisely. Especially in areas with large curvature, it helps to improve the uniformity of the coating and reduce the problem of uneven coating thickness caused by irregular surface shape.

[0032] In one possible implementation, the surface of the substrate within the vacuum coating cavity is dynamically divided into detection unit grids to obtain an adaptive grid substrate surface. Step S100 further includes step S140, whereby the temperature change rate is extracted based on the substrate dynamic temperature distribution matrix to obtain a substrate dynamic temperature change rate matrix. Specifically, the substrate dynamic temperature distribution matrix is ​​read from a database, and time-series analysis is performed on the temperature data of each grid to calculate its temperature change rate. The calculated temperature change rate is stored as a new matrix, namely the substrate dynamic temperature change rate matrix.

[0033] For example, the substrate dynamic temperature distribution matrix records the temperature value of each grid at different time points. By calculating the temperature difference between adjacent time points and dividing by the time interval, the temperature change rate of each grid can be obtained. If the temperature of a certain grid at time point t1 is T1 and the temperature at time point t2 is T2, then the temperature change rate of that grid is (T2-T1) / (t2-t1).

[0034] Step S150: Traverse the dynamic temperature change rate matrix of the substrate and extract the cell grids with temperature change rates greater than a preset temperature change rate threshold to obtain temperature anomaly grids. Specifically, read the dynamic temperature change rate matrix of the substrate from the database, traverse each grid in the matrix, and extract its temperature change rate. Compare the temperature change rate of each grid with the preset temperature change rate threshold. If the temperature change rate of a certain grid is greater than the preset temperature change rate threshold, then the grid is marked as a temperature anomaly grid, and all temperature anomaly grids are stored in the database. For example, assuming the preset temperature change rate threshold is 3℃ / min, if the temperature change rate of a certain grid is 5℃ / min, then the grid will be marked as a temperature anomaly grid.

[0035] Step S160 involves reconstructing and densifying the temperature anomaly mesh and then updating the adaptive mesh substrate surface. Specifically, each temperature anomaly mesh is densified to increase its density, enabling more precise capture of temperature changes. The densified mesh is then reintegrated into the adaptive mesh substrate surface, updating the surface and ensuring that the geometric center point of all meshes remains a temperature detection reference point. For example, assuming the original mesh size of a temperature anomaly mesh is 1cm × 1cm, it is divided into smaller meshes, such as 0.5cm × 0.5cm, after densification. This allows for more precise capture of changes in the temperature anomaly region. This implementation, after densifying and reconstructing the temperature anomaly mesh, enables more precise capture of temperature changes, thus providing more accurate data support for subsequent coating uniformity prediction and process parameter control.

[0036] Step S200: The surface of the adaptive grid substrate is scanned in real time using an infrared thermal imager to obtain the dynamic temperature distribution matrix of the substrate.

[0037] Specifically, during the vacuum coating process, an infrared thermal imager installed inside the coating chamber is activated to ensure it covers the entire substrate surface. The imager scans the substrate surface at a set frequency (e.g., multiple times per second), capturing the surface temperature in real time. The imager outputs the scanned temperature data as images containing temperature information at different locations on the substrate surface. The data acquisition system converts the temperature image data into digital signals and transmits the data to the data processing system via a high-speed data transmission system. Upon receiving the temperature data from the imager, the data processing system preprocesses the data (e.g., noise reduction, calibration). The preprocessed temperature data is then organized according to an adaptive grid division to generate a dynamic temperature distribution matrix for the substrate. The temperature value of each grid is recorded in the matrix, with rows and columns corresponding to grid positions on the substrate surface, and the values ​​representing the real-time temperature of that grid. For example, assuming the substrate surface is divided into 100 grids, the dynamic temperature distribution matrix will record the real-time temperature value of each grid, forming a 10×10 matrix.

[0038] Step S300: Based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map.

[0039] Specifically, a mapping model between coating thickness and temperature gradient is pre-established through historical data analysis. The dynamic temperature distribution matrix of the substrate is input into this mapping model to predict coating uniformity, generating a coating uniformity prediction map. For example, assuming the mapping model indicates that areas with larger temperature gradients will have thinner coatings, if a certain area on the substrate surface has a large temperature gradient, the mapping model will predict that the coating thickness in that area will be thinner and mark it on the coating uniformity prediction map.

[0040] In one possible implementation, based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map. Step S300 further includes step S310, acquiring historical coating data, including a historical substrate dynamic temperature distribution matrix and a historical film thickness deviation value matrix, wherein the film thickness deviation value refers to the difference between the actual film thickness and the standard film thickness. Specifically, historical coating data, including the historical substrate dynamic temperature distribution matrix and the historical film thickness deviation value matrix, is read from the database. The collected historical coating data is cleaned and preprocessed to remove noise and outliers, ensuring the integrity and accuracy of the data. For example, the historical coating data contains records of multiple coating processes, each record including the temperature distribution of different grids on the substrate surface and the corresponding film thickness deviation value. For example, in a certain historical record, the temperature of a certain grid is 30℃, the actual film thickness is 1.2μm, and the standard film thickness is 1.0μm, then the film thickness deviation value of that grid is 0.2μm.

[0041] Step S320: Calculate the temperature gradient between adjacent grids based on the historical substrate dynamic temperature distribution matrix to obtain the historical substrate temperature gradient distribution matrix. Specifically, for each grid in the historical substrate dynamic temperature distribution matrix, calculate the temperature difference between it and its adjacent grids, and divide it by the grid spacing to obtain the temperature gradient. Store the calculated temperature gradient values ​​in a new matrix to form the historical substrate temperature gradient distribution matrix. For example, assuming the temperature of a certain grid is 30℃, the temperature of its adjacent grid is 32℃, and the grid spacing is 1cm, then the temperature gradient of this grid is (32-30) / 1 = 2℃ / cm.

[0042] Step S330: The historical substrate temperature gradient distribution matrix and the historical film thickness deviation matrix are input into a convolutional neural network model for training to obtain the mapping relationship between coating thickness and temperature gradient. Specifically, the historical substrate temperature gradient distribution matrix and the historical film thickness deviation matrix are used as inputs to the convolutional neural network model for training. The backpropagation algorithm is used to train the model, adjust model parameters, and optimize model performance. The prediction accuracy of the model is evaluated through methods such as cross-validation to ensure the reliability and accuracy of the model. The trained convolutional neural network model is then stored in a database. The convolutional neural network model contains multiple convolutional layers and pooling layers. Convolutional operations extract features of the temperature gradient distribution, and pooling operations reduce the feature dimensionality. The output of the model is the predicted film thickness deviation value. The backpropagation algorithm is used to adjust model parameters based on the difference between the predicted and actual values ​​to optimize model performance.

[0043] Step S340: Based on the mapping relationship, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map. Specifically, the current substrate dynamic temperature distribution matrix is ​​read from the database, the temperature gradient between adjacent grids is calculated to obtain the current substrate temperature gradient distribution matrix, and the current substrate temperature gradient distribution matrix is ​​input into a trained convolutional neural network model. The model outputs a predicted film thickness deviation matrix. Based on the predicted film thickness deviation matrix, a coating uniformity prediction map is generated.

[0044] For example, assuming a grid in the current substrate dynamic temperature distribution matrix has a temperature of 31℃ and its adjacent grid has a temperature of 33℃, the calculated temperature gradient is 2℃ / cm. Inputting this data into a convolutional neural network model, the model outputs a predicted film thickness deviation of 0.15μm for that grid. Based on the prediction result, the film thickness deviation of that grid is marked in the coating uniformity prediction map. This implementation, by training the convolutional neural network model on historical data, can establish an accurate mapping relationship between coating thickness and temperature gradient. Compared to traditional empirical formulas or simple models, convolutional neural networks can capture more complex nonlinear relationships, thereby improving prediction accuracy.

[0045] In one possible implementation, the historical substrate temperature gradient distribution matrix and the historical film thickness deviation matrix are input into a convolutional neural network model for training to obtain the mapping relationship between coating thickness and temperature gradient. Step S330 further includes step S331, constructing a two-branch convolutional neural network model, wherein the two-branch convolutional neural network model includes a first branch and a second branch. Specifically, a convolutional neural network model containing two branches is constructed, which are used to extract spatial features and temporal features, respectively. The first branch is used to process static spatial features, and the second branch is used to process dynamic temporal features. The outputs of the two branches are fused by weighted concatenation. Specifically, the first branch uses convolutional layers and pooling layers to extract the spatial features of the historical substrate temperature gradient distribution matrix. The second branch uses a Long Short-Term Memory (LSTM) network layer to extract the temporal features of the historical substrate temperature gradient distribution matrix at the same location in the first N production batches. The outputs of the two branches are fused by weighted concatenation to form a film thickness deviation prediction matrix. A loss function and prediction accuracy are defined as training objectives, and preset mean squared error thresholds and accuracy thresholds are set.

[0046] In step S332, the first branch receives the historical substrate temperature gradient distribution matrix and extracts spatial features using a convolution kernel. Specifically, the historical substrate temperature gradient distribution matrix is ​​input into the convolutional layer of the first branch. The convolutional layer performs a convolution operation on the input matrix using a convolution kernel to extract local spatial features. A pooling layer performs a pooling operation on the output of the convolutional layer to reduce the feature dimensionality. The convolution and pooling operations are repeated to extract higher-level spatial features. The extracted spatial features are then output to the subsequent weighted concatenation module.

[0047] For example, with a 3×3 kernel and a stride of 1, the convolution operation slides across the input matrix, calculating the convolution value at each position. For instance, for a 5×5 temperature gradient distribution matrix, the convolution operation would generate a new feature matrix of size 3×3. Pooling layers further reduce the dimensionality of the feature matrix; for example, max pooling can reduce a 3×3 feature matrix to 2×2.

[0048] Step S333: The second branch receives the historical substrate temperature gradient distribution matrix at the same location from the first N production batches, and extracts the time features through an LSTM layer, where N≥3. Specifically, the historical substrate temperature gradient distribution matrix at the same location from the first N production batches is read from the database, and these matrices are input into the LSTM layer of the second branch. The LSTM layer processes the input time series data, extracts the time features, and outputs the extracted time features to the subsequent weighted splicing module.

[0049] For example, each of the first N production batches has a temperature gradient distribution matrix at the same location, forming a time series. LSTM layers process these matrices one by one, capturing long-term dependencies in the time series through gating mechanisms. For instance, if the temperature gradient at a certain location gradually increases in the first few batches, the LSTM layer can capture this trend and output it as a time feature.

[0050] Step S334: The spatial features and the temporal features are weighted and concatenated to obtain a film thickness deviation prediction matrix. Specifically, according to preset weights or weights learned by the model, the spatial features extracted by the first branch and the temporal features extracted by the second branch are weighted and concatenated to form a film thickness deviation prediction matrix. The film thickness deviation prediction matrix is ​​compared with the historical film thickness deviation matrix to calculate the mean squared error and prediction accuracy.

[0051] Step S335: When the mean squared error (MSE) of the predicted film thickness deviation matrix and the historical film thickness deviation matrix is ​​lower than a preset MSE threshold and the prediction accuracy reaches a preset accuracy threshold, training is stopped and the mapping relationship between coating thickness and temperature gradient is derived. Specifically, the mean squared error (MSE) is used as the loss function to measure the difference between the predicted film thickness deviation matrix and the historical film thickness deviation matrix. During training, the MSE between the predicted film thickness deviation matrix and the historical film thickness deviation matrix is ​​calculated, and the accuracy of the model in predicting film thickness deviation is also calculated. If the MSE is lower than the preset MSE threshold and the prediction accuracy reaches the preset accuracy threshold, training is stopped. The trained model is derived, and the mapping relationship between coating thickness and temperature gradient is obtained. For example, the preset MSE threshold is 0.01, and the accuracy threshold is 90%. During training, the model continuously adjusts parameters to optimize the MSE and prediction accuracy. When the MSE is less than 0.01 and the classification accuracy is greater than 90%, training stops, and the model is derived. This implementation uses a dual-branch convolutional neural network model to extract both spatial and temporal features, enabling a more comprehensive capture of factors influencing film thickness deviation. Compared to single feature extraction methods, this dual-branch model significantly improves prediction accuracy.

[0052] Step S400: Align and fuse the substrate dynamic temperature distribution matrix and the coating uniformity prediction map to obtain a cross-validation grid.

[0053] Specifically, the substrate dynamic temperature distribution matrix records the real-time temperature value of each grid on the substrate surface, and the coating uniformity prediction map annotates the predicted film thickness deviation value for each grid. Using an image alignment algorithm, the substrate dynamic temperature distribution matrix and the coating uniformity prediction map are spatially aligned to ensure a one-to-one correspondence between the grids in the two matrices. For example, if the substrate surface is divided into a 10×10 grid, the aligned matrix will ensure that the temperature value and the predicted film thickness deviation value for each grid accurately match. The data from the aligned substrate dynamic temperature distribution matrix and the coating uniformity prediction map are then fused, that is, the temperature information and the predicted film thickness deviation information are integrated into a unified grid, forming a cross-validation grid. For example, for each grid, the temperature value and the predicted film thickness deviation value can be combined into a feature vector.

[0054] Step S500: When any one of the cross-validation grids fails validation, a three-dimensional compensation vector is generated.

[0055] Specifically, each grid in the cross-validation grid is validated to determine if it meets the preset coating quality standards. If a grid fails validation, its location and validation result are recorded. A three-dimensional compensation vector is generated based on the validation results, including plasma gun angle compensation, sputtering power compensation, and residence time compensation. The plasma gun angle compensation adjusts the plasma gun's jet angle to change the coating direction and coverage. The sputtering power compensation adjusts the sputtering source power to change the coating deposition rate. The residence time compensation adjusts the plasma gun's dwell time on each grid to change the coating thickness.

[0056] For example, assuming that the coating thickness of a certain grid in the cross-validation grid is lower than the standard value, a three-dimensional compensation vector is generated based on the position and thickness deviation of the grid: the angle of the plasma gun is adjusted to +5° so that the coating can cover the area more uniformly; the sputtering power is increased by 10% to improve the deposition rate of the coating; and the dwell time is increased by 2 seconds to increase the coating thickness in the area.

[0057] In one possible implementation, when any grid in the cross-validation mesh fails validation, a three-dimensional compensation vector is generated. Step S500 further includes step S510, extracting the absolute temperature and temperature rise rate from the cross-validation mesh. Specifically, the cross-validation mesh is read from the database, and for each grid, its absolute temperature value is extracted. The temperature rise rate of each grid is calculated, i.e., the temperature difference between adjacent time points divided by the time interval.

[0058] Step S520: If the absolute temperature does not meet the preset temperature range and / or the temperature rise rate does not meet the preset temperature rise rate range, the corresponding grid is marked as a dangerous grid. Specifically, the absolute temperature of each grid is compared with the preset temperature range to determine if it is within the allowable range. The temperature rise rate of each grid is compared with the preset temperature rise rate range to determine if it is within the allowable range. If the absolute temperature or temperature rise rate of a certain grid does not meet the preset range, then the grid is marked as a dangerous grid. For example, assuming the preset temperature range is [20℃, 40℃] and the temperature rise rate range is [0.1℃ / min, 1.0℃ / min]. For a certain grid, if its absolute temperature is 45℃ and / or its temperature rise rate is 1.2℃ / min, then the grid is marked as a dangerous grid.

[0059] Step S530: Extract the predicted film thickness deviation value from the hazardous grid and perform cross-validation. If the predicted film thickness deviation value does not meet the preset film thickness deviation range, the validation fails, and a three-dimensional compensation vector is generated. Specifically, for each hazardous grid, its predicted film thickness deviation value is extracted. The predicted film thickness deviation value is compared with the preset film thickness deviation range to determine whether the condition is met. If the predicted film thickness deviation value of a hazardous grid does not meet the preset range, the validation fails. For grids that fail validation, a three-dimensional compensation vector is generated, including the plasma gun angle compensation, sputtering power compensation, and residence time compensation. For example, assuming the preset film thickness deviation range is [-0.1μm, 0.1μm]. For a hazardous grid, if its predicted film thickness deviation value is 0.2μm, then the validation of that grid fails.

[0060] This implementation, by simultaneously considering absolute temperature, temperature rise rate, and predicted film thickness deviation, enables a more comprehensive assessment of the coating quality for each grid, effectively avoiding potential misjudgments that might arise from relying solely on predicted values. For example, even if the predicted film thickness deviation indicates that the film thickness in a certain area may exceed the standard range, if the temperature data (absolute temperature and temperature rise rate) for that area are both within the normal range, then that area may not require adjustment. Conversely, if the predicted film thickness deviation is within the normal range, but the temperature data is abnormal (e.g., excessively high absolute temperature and / or excessively rapid temperature rise rate), then that area may still have potential problems and requires further attention. This combined verification mechanism improves the system's accuracy in identifying problem areas and avoids misjudgments caused by a single data anomaly.

[0061] In one possible implementation, step S500 further includes step S540, where the verification failure also requires meeting the temperature-film thickness collaborative verification conditions, which are as follows: calculating the temperature anomaly degree based on the substrate material, the absolute temperature, and the temperature rise rate; calculating the film thickness anomaly degree based on the substrate material and the predicted film thickness deviation; multiplying the temperature anomaly degree and the film thickness anomaly degree to obtain a collaborative factor; if the temperature anomaly degree is greater than a preset temperature anomaly degree threshold, the film thickness anomaly degree is greater than a preset film thickness anomaly threshold, and the collaborative factor is greater than a preset collaborative factor threshold, then the verification is determined to fail.

[0062] Specifically, the characteristic parameters of the substrate material, as well as the absolute temperature and temperature rise rate for each hazardous grid, are read from the database. Based on the characteristics of the substrate material, the temperature anomaly degree for each hazardous grid is calculated. The temperature anomaly degree reflects the degree of deviation of the current temperature and temperature rise rate from the allowable range of the substrate material. For example, if the allowable absolute temperature range for the substrate material under the current process conditions is [20℃, 40℃], and the allowable temperature rise rate range is [0.1℃ / min, 1.0℃ / min], then the extent to which these ranges are exceeded will be quantified as the temperature anomaly degree.

[0063] The characteristic parameters of the substrate material and the predicted film thickness deviation for each critical grid are retrieved from the database. Based on the characteristics of the substrate material, the film thickness anomaly degree for each critical grid is calculated. The film thickness anomaly degree reflects the degree of deviation between the predicted film thickness deviation and the allowable range of the substrate material. For example, if the allowable film thickness deviation range for the substrate material is [-0.1 μm, 0.1 μm], then the extent to which it exceeds this range will be quantified as the film thickness anomaly degree.

[0064] The synergy factor is obtained by multiplying the temperature anomaly and film thickness anomaly of each hazardous grid. The synergy factor comprehensively reflects the degree of anomaly in both the temperature and film thickness dimensions. For example, if the temperature anomaly of a certain grid is 1.2 and the film thickness anomaly is 1.5, then the synergy factor is 1.2 × 1.5 = 1.8.

[0065] The calculated temperature anomaly, film thickness anomaly, and synergy factor are compared with preset thresholds. If the temperature anomaly is greater than the preset temperature anomaly threshold, the film thickness anomaly is greater than the preset film thickness anomaly threshold, and the synergy factor is greater than the preset synergy factor threshold, then the mesh verification is deemed unsuccessful.

[0066] For example, assume the allowable absolute temperature range for the substrate material under current process conditions is [20℃, 40℃], and the temperature rise rate range is [0.1℃ / min, 1.0℃ / min]. Assuming the absolute temperature of a certain grid is 45℃, the absolute temperature anomaly can be equal to (45-40) / (40-20) = 0.25. Assuming the temperature rise rate of this grid is 1.2℃ / min, the temperature rise rate anomaly can be equal to (1.2-1) / (1-0.1) ≈ 0.222. The temperature anomaly can be the maximum of the absolute temperature anomaly and the temperature rise rate anomaly, or a weighted average of the two. Assuming the temperature anomaly is the maximum of the two, then the temperature anomaly = 0.25. Assuming the allowable film thickness deviation range for the substrate material is [-0.1μm, 0.1μm], and assuming the predicted film thickness deviation value for this grid is 0.2μm. The film thickness anomaly can be equal to (0.2-0.1) / [0.1-(-0.1)] = 0.5. Therefore, the synergy factor = 0.25 × 0.5 = 0.125. Assume the preset temperature anomaly threshold is 1.0, the film thickness anomaly threshold is 1.0, and the synergy factor threshold is 1.5. Since temperature anomaly 0.25 < 1.0, film thickness anomaly 0.5 < 1.0, and synergy factor 0.125 < 1.5, the mesh verification is deemed successful.

[0067] This approach, by simultaneously considering both temperature and film thickness anomalies and calculating a synergy factor, can more accurately identify areas where problems truly exist. The synergy factor comprehensively reflects the degree of anomaly in both temperature and film thickness dimensions, enabling a more holistic assessment of the problem's severity. For example, even if the temperature and film thickness anomalies are not high individually, their combination may still indicate a potential problem. This collaborative verification mechanism avoids misjudgments that can result from relying solely on a single anomaly, improving the accuracy and reliability of the verification.

[0068] In one possible implementation, when any grid in the cross-validation mesh fails validation, a three-dimensional compensation vector is generated. Step S500 further includes step S550, extracting temperature anomaly information and film thickness anomaly information from the grids that failed validation. Specifically, a list of grids that failed validation is read from a database. For each grid that failed validation, its temperature anomaly information (absolute temperature, temperature rise rate) and film thickness anomaly information (predicted film thickness deviation value) are extracted. For example, suppose a grid fails validation, its absolute temperature is 45℃, its temperature rise rate is 1.2℃ / min, and its predicted film thickness deviation value is 0.2μm. This information is extracted and stored in a temporary database.

[0069] Step S560: Calculate the angle compensation amount of the plasma gun based on the temperature anomaly information. Specifically, read the temperature anomaly information (absolute temperature and / or temperature rise rate) from the temporary database, and use a preset compensation calculation model to calculate the angle compensation amount of the plasma gun based on the temperature anomaly information. For example, assume that the preset model indicates that when the absolute temperature exceeds 40°C, the angle of the plasma gun needs to be adjusted to reduce heat input. The specific compensation amount can be calculated using the following formula: Angle compensation amount = k × (Actual absolute temperature - Maximum allowable absolute temperature), where k is a model parameter.

[0070] Step S570: Calculate the sputtering power compensation and residence time compensation based on the film thickness anomaly information. Specifically, similar to step S560, read the film thickness anomaly information (predicted film thickness deviation value) from the temporary database, and use a preset compensation calculation model to calculate the sputtering power compensation and residence time compensation based on the film thickness anomaly information.

[0071] Step S580: Integrate the angle compensation amount, the sputtering power compensation amount, and the dwell time compensation amount to generate the three-dimensional compensation vector. Specifically, the calculated angle compensation amount, sputtering power compensation amount, and dwell time compensation amount are integrated into a single three-dimensional compensation vector, in the format (angle compensation amount, sputtering power compensation amount, dwell time compensation amount). This implementation method, by separately calculating the plasma gun angle compensation amount, sputtering power compensation amount, and dwell time compensation amount, and integrating them into a single three-dimensional compensation vector, provides a precise compensation strategy, improves the uniformity and consistency of the coating, and effectively reduces the defect rate.

[0072] In one possible implementation, step S500 further includes step S590, calculating an edge compensation coefficient based on the radius of curvature of the mesh in the edge region of the substrate. Specifically, the radius of curvature of the mesh in the edge region of the substrate is measured using a high-precision sensor or optical measurement equipment, and the measured radius of curvature is substituted into a preset compensation coefficient calculation model to calculate the edge compensation coefficient. For example, the preset compensation coefficient calculation model indicates that the edge compensation coefficient is inversely proportional to the radius of curvature.

[0073] Step S5100: Adjust the three-dimensional compensation vector according to the edge compensation coefficient to obtain a three-dimensional enhancement compensation vector. Specifically, adjust each component of the three-dimensional compensation vector according to the edge compensation coefficient. Integrate the adjusted compensation amounts into a new three-dimensional compensation vector to obtain the three-dimensional enhancement compensation vector.

[0074] Step S5110: After compensating the edge region mesh according to the three-dimensional strengthening compensation vector, a secondary edge verification is initiated. A laser thickness gauge is used to re-measure the edge region mesh, and the re-measurement result is obtained. Specifically, process parameters are adjusted according to the three-dimensional strengthening compensation vector to perform compensation on the edge region mesh. A laser thickness gauge is then used to re-measure the compensated edge region mesh, and the re-measurement result is obtained.

[0075] Step S5120: Optimize the coating of the edge region mesh based on the retest results. Specifically, based on the retest results, determine whether the film thickness is within the allowable range. If the film thickness deviation is less than the minimum allowable deviation, calculate a new compensation amount, adjust process parameters, and further optimize the coating quality of the edge region mesh. If the film thickness deviation is greater than the maximum allowable deviation, remove excess coating and then redeposit to ensure the film thickness is within the allowable range. This implementation method, by calculating the edge compensation coefficient and adjusting the three-dimensional compensation vector, can provide more precise compensation for the special shape of the edge region. This targeted compensation improves the coating quality of the edge region. Retesting using a laser thickness gauge and further optimizing process parameters based on the retest results ensures the uniformity of the film thickness in the edge region.

[0076] Step S600: Adjust the process parameters of the vacuum coating equipment in real time according to the three-dimensional compensation vector.

[0077] Specifically, the three-dimensional compensation vector is input into the process parameter control system. Based on this vector, the system adjusts the process parameters of the vacuum coating equipment: adjusting the plasma gun angle, the sputtering source power, and the plasma gun residence time on each grid. The adjusted process parameters are then fed back to the equipment to ensure the coating process follows the new parameters. Real-time monitoring of the coating process ensures the coating quality meets requirements. For example, if the three-dimensional compensation vector indicates a need to increase the coating thickness in a certain area, the system adjusts the plasma gun angle to +5°, increases the sputtering power by 10%, and increases the residence time by 2 seconds. These adjustments ensure that the coating thickness in that area reaches the standard value.

[0078] This application employs a method of dynamically dividing the substrate surface within a vacuum coating chamber into a detection unit grid to obtain an adaptive grid substrate surface. An infrared thermal imager is used to scan the adaptive grid substrate surface in real time, generating a dynamic temperature distribution matrix for the substrate. Based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix is ​​analyzed and predicted to generate a coating uniformity prediction map. The dynamic temperature distribution matrix and the coating uniformity prediction map are aligned and fused to generate a cross-validation grid. Each grid in the cross-validation grid is validated. If the validation fails, a three-dimensional compensation vector (including plasma gun angle, sputtering power, and residence time) is generated. The process parameters of the vacuum coating equipment are adjusted in real time based on the three-dimensional compensation vector to optimize coating quality. This method solves the technical problem of inaccurate prediction of coating uniformity in existing NCVM non-conductive vacuum coating process execution monitoring, which makes it difficult to adjust process parameters in real time and accurately. It achieves the technical effect of accurately predicting coating uniformity, thereby enabling real-time and accurate adjustment of process parameters.

[0079] In the above text, refer to Figure 1 A method for monitoring the execution of an NCVM non-conductive vacuum coating process according to an embodiment of the present invention is described in detail. Next, reference will be made to... Figure 2 This invention describes a monitoring system for the execution of an NCVM non-conductive vacuum coating process according to an embodiment of the present invention.

[0080] An NCVM (Non-Conductive Vacuum Coating) process execution monitoring system according to an embodiment of the present invention addresses the technical problem of inaccurate prediction of coating uniformity in existing NCVM process execution monitoring systems, which makes it difficult to control process parameters in real time and with precision. The system achieves the technical effect of controlling process parameters in real time and with precision by accurately predicting coating uniformity. The NCVM non-conductive vacuum coating process execution monitoring system includes: a detection unit grid dynamic division module 10, an infrared thermal imaging scanning module 20, a coating uniformity prediction module 30, an alignment and fusion processing module 40, a three-dimensional compensation vector generation module 50, and a process parameter control module 60.

[0081] The detection unit grid dynamic division module 10 is used to dynamically divide the substrate surface in the vacuum coating cavity into detection unit grids to obtain an adaptive grid substrate surface; the infrared thermal imaging scanning module 20 is used to scan the adaptive grid substrate surface in real time using an infrared thermal imager to obtain a substrate dynamic temperature distribution matrix; the coating uniformity prediction module 30 is used to analyze and predict the substrate dynamic temperature distribution matrix based on the mapping relationship between coating thickness and temperature gradient to obtain a coating uniformity prediction map; the alignment and fusion processing module 40 is used to align and fuse the substrate dynamic temperature distribution matrix and the coating uniformity prediction map to obtain a cross-validation grid; the three-dimensional compensation vector generation module 50 is used to generate a three-dimensional compensation vector when any grid in the cross-validation grid fails verification; and the process parameter control module 60 is used to control the process parameters of the vacuum coating equipment in real time according to the three-dimensional compensation vector.

[0082] The specific configuration of the detection unit grid dynamic division module 10 will be described in detail below. As mentioned above, the detection unit grid is dynamically divided on the surface of the substrate in the vacuum coating cavity to obtain an adaptive grid substrate surface. The detection unit grid dynamic division module 10 may further include: a conversion relationship preset unit for presetting the conversion relationship between the surface curvature of the substrate and the grid density; an adaptive grid density calculation unit for real-time acquisition of multiple surface curvatures of the current substrate and calculation of the adaptive grid density according to the conversion relationship; and a detection unit grid dynamic division unit for dynamically dividing the substrate surface in the vacuum coating cavity according to the adaptive grid density, setting the geometric center point of each grid as a temperature detection reference point to obtain an adaptive grid substrate surface.

[0083] The detection unit grid dynamic division module 10, which dynamically divides the substrate surface within the vacuum coating cavity into a detection unit grid to obtain an adaptive grid substrate surface, may further include: a temperature change rate extraction unit for extracting the temperature change rate based on the substrate dynamic temperature distribution matrix to obtain a substrate dynamic temperature change rate matrix; a temperature anomaly grid acquisition unit for traversing the substrate dynamic temperature change rate matrix and extracting unit grids with temperature change rates greater than a preset temperature change rate threshold to obtain temperature anomaly grids; and a grid encryption and reconstruction unit for updating the adaptive grid substrate surface after encryption and reconstruction of the temperature anomaly grids.

[0084] The specific configuration of the coating uniformity prediction module 30 will be described in detail below. As mentioned above, based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map. The coating uniformity prediction module 30 may further include: a historical coating data acquisition unit for acquiring historical coating data, including a historical substrate dynamic temperature distribution matrix and a historical film thickness deviation value matrix, wherein the film thickness deviation value refers to the difference between the actual film thickness and the standard film thickness; an adjacent grid temperature gradient calculation unit for calculating the adjacent grid temperature gradient based on the historical substrate dynamic temperature distribution matrix to obtain a historical substrate temperature gradient distribution matrix; a convolutional training and learning unit for inputting the historical substrate temperature gradient distribution matrix and the historical film thickness deviation value matrix into a convolutional neural network model for training and learning to obtain the mapping relationship between coating thickness and temperature gradient; and a mapping unit for analyzing and predicting the substrate dynamic temperature distribution matrix based on the mapping relationship to obtain a coating uniformity prediction map.

[0085] The process involves inputting the historical substrate temperature gradient distribution matrix and the historical film thickness deviation matrix into a convolutional neural network model for training to obtain the mapping relationship between coating thickness and temperature gradient. The convolutional training unit may further include: a dual-branch convolutional neural network model construction subunit for constructing a dual-branch convolutional neural network model, wherein the dual-branch convolutional neural network model includes a first branch and a second branch; a spatial feature extraction subunit for receiving the historical substrate temperature gradient distribution matrix in the first branch and extracting spatial features through a convolution kernel; a temporal feature extraction subunit for receiving the historical substrate temperature gradient distribution matrix at the same location in the previous N production batches in the second branch and extracting temporal features through an LSTM layer, where N≥3; a weighted splicing subunit for weighted splicing of the spatial features and the temporal features to obtain a film thickness deviation prediction matrix; and a mapping relationship derivation subunit for stopping training and deriving the mapping relationship between coating thickness and temperature gradient when the mean square error of the film thickness deviation prediction matrix and the historical film thickness deviation matrix is ​​lower than a preset mean square error threshold and the prediction accuracy reaches a preset accuracy threshold.

[0086] The specific configuration of the three-dimensional compensation vector generation module 50 will be described in detail below. As mentioned above, when any grid in the cross-validation grid fails verification, a three-dimensional compensation vector is generated. The three-dimensional compensation vector generation module 50 may further include: a temperature information extraction unit for extracting absolute temperature and temperature rise rate from the cross-validation grid; a dangerous grid marking unit for marking the corresponding grid as a dangerous grid if the absolute temperature does not meet a preset temperature range and / or the temperature rise rate does not meet a preset temperature rise rate range; and a cross-validation unit for extracting film thickness deviation prediction values ​​from the dangerous grids for cross-validation. If the film thickness deviation prediction value does not meet a preset film thickness deviation range, the verification fails, and a three-dimensional compensation vector is generated.

[0087] The three-dimensional compensation vector generation module 50 may further include: a temperature-film thickness co-verification unit for requiring that the temperature-film thickness co-verification conditions be met even if the verification fails. The co-verification conditions are as follows: calculate the temperature anomaly degree based on the substrate material, the absolute temperature, and the temperature rise rate; calculate the film thickness anomaly degree based on the substrate material and the predicted film thickness deviation; multiply the temperature anomaly degree and the film thickness anomaly degree to obtain a co-verification factor; if the temperature anomaly degree is greater than a preset temperature anomaly degree threshold, the film thickness anomaly degree is greater than a preset film thickness anomaly threshold, and the co-verification factor is greater than a preset co-verification factor threshold, then the verification is determined to fail.

[0088] When any grid in the cross-validation grid fails validation, a three-dimensional compensation vector is generated. The three-dimensional compensation vector generation module 50 may further include: an anomaly information extraction unit for extracting temperature anomaly information and film thickness anomaly information from the grids that failed validation; an angle compensation amount calculation unit for calculating the angle compensation amount of the plasma gun based on the temperature anomaly information; a sputtering power compensation amount and residence time compensation amount calculation unit for calculating the sputtering power compensation amount and residence time compensation amount based on the film thickness anomaly information; and a three-dimensional compensation vector generation unit for integrating the angle compensation amount, the sputtering power compensation amount, and the residence time compensation amount to generate the three-dimensional compensation vector.

[0089] The three-dimensional compensation vector generation module 50 may further include: an edge compensation coefficient calculation unit for calculating an edge compensation coefficient based on the radius of curvature of the mesh in the edge region of the substrate; a three-dimensional compensation vector adjustment unit for adjusting the three-dimensional compensation vector based on the edge compensation coefficient to obtain a three-dimensional strengthening compensation vector; an edge secondary verification unit for performing compensation on the mesh in the edge region based on the three-dimensional strengthening compensation vector, and then initiating edge secondary verification by re-measuring the mesh in the edge region using a laser thickness gauge to obtain the re-measuring result; and an edge coating optimization unit for optimizing the coating of the mesh in the edge region based on the re-measuring result.

[0090] The NCVM non-conductive vacuum coating process execution monitoring system provided in this embodiment of the invention can execute the NCVM non-conductive vacuum coating process execution monitoring method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0091] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0092] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for monitoring the execution of an NCVM non-conductive vacuum coating process, characterized in that, The method includes: The substrate surface inside the vacuum coating cavity is dynamically divided into detection unit grids to obtain an adaptive grid substrate surface; The surface of the adaptive grid substrate is scanned in real time using an infrared thermal imager to obtain the dynamic temperature distribution matrix of the substrate; Based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map. The dynamic temperature distribution matrix of the substrate and the predicted uniformity of the coating are aligned and fused to obtain a cross-validation grid. When any one of the cross-validation grids fails validation, a three-dimensional compensation vector is generated. The process parameters of the vacuum coating equipment are adjusted in real time based on the three-dimensional compensation vector.

2. The method for monitoring the execution of an NCVM non-conductive vacuum coating process as described in claim 1, characterized in that, The substrate surface within the vacuum coating cavity is dynamically divided into detection unit grids to obtain an adaptive grid substrate surface, including: The conversion relationship between the surface curvature of the preset substrate and the mesh density; The curvature of multiple surfaces of the current substrate is acquired in real time, and the adaptive mesh density is calculated based on the transformation relationship. The substrate surface inside the vacuum coating cavity is dynamically divided into detection unit grids according to the adaptive grid density, and the geometric center point of each grid is set as the temperature detection reference point to obtain the adaptive grid substrate surface.

3. The method for monitoring the execution of an NCVM non-conductive vacuum coating process as described in claim 2, characterized in that, The substrate surface within the vacuum coating cavity is dynamically divided into detection unit grids to obtain an adaptive grid substrate surface, which also includes: Based on the dynamic temperature distribution matrix of the substrate, the temperature change rate is extracted to obtain the dynamic temperature change rate matrix of the substrate. Traverse the dynamic temperature change rate matrix of the substrate and extract the cell grids with temperature change rates greater than a preset temperature change rate threshold to obtain temperature anomaly grids. The adaptive mesh substrate surface is updated after the temperature anomaly mesh is encrypted and reconstructed.

4. The method for monitoring the execution of an NCVM non-conductive vacuum coating process as described in claim 1, characterized in that, Based on the mapping relationship between coating thickness and temperature gradient, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map, including: Acquire historical coating data, including historical substrate dynamic temperature distribution matrix and historical film thickness deviation matrix, where film thickness deviation refers to the difference between actual film thickness and standard film thickness; The temperature gradient of adjacent grids is calculated based on the historical substrate dynamic temperature distribution matrix to obtain the historical substrate temperature gradient distribution matrix. The historical substrate temperature gradient distribution matrix and the historical film thickness deviation matrix are input into a convolutional neural network model for training and learning to obtain the mapping relationship between film thickness and temperature gradient. Based on the mapping relationship, the dynamic temperature distribution matrix of the substrate is analyzed and predicted to obtain a coating uniformity prediction map.

5. The method for monitoring the execution of an NCVM non-conductive vacuum coating process as described in claim 4, characterized in that, The historical substrate temperature gradient distribution matrix and the historical film thickness deviation matrix are input into a convolutional neural network model for training to obtain the mapping relationship between coating thickness and temperature gradient, including: Construct a two-branch convolutional neural network model, wherein the two-branch convolutional neural network model includes a first branch and a second branch; The first branch receives the historical substrate temperature gradient distribution matrix and extracts spatial features through a convolution kernel; The second branch receives the historical substrate temperature gradient distribution matrix at the same location in the previous N production batches, and extracts the time features through the LSTM layer, where N≥3; The spatial features and the temporal features are weighted and concatenated to obtain the film thickness deviation prediction matrix; When the mean square error of the film thickness deviation prediction matrix and the historical film thickness deviation matrix is ​​lower than the preset mean square error threshold and the prediction accuracy reaches the preset accuracy threshold, training is stopped and the mapping relationship between the coating thickness and the temperature gradient is derived.

6. The method for monitoring the execution of an NCVM non-conductive vacuum coating process as described in claim 1, characterized in that, When any grid in the cross-validation grid fails validation, a three-dimensional compensation vector is generated, including: Extract the absolute temperature and temperature rise rate from the cross-validation grid; If the absolute temperature does not meet the preset temperature range and / or the temperature rise rate does not meet the preset temperature rise rate range, the corresponding grid will be marked as a dangerous grid. The predicted film thickness deviation is extracted from the dangerous grid and cross-validated. If the predicted film thickness deviation does not meet the preset film thickness deviation range, the validation fails and a three-dimensional compensation vector is generated.

7. The method for monitoring the execution of an NCVM non-conductive vacuum coating process as described in claim 6, characterized in that, If the verification fails, the temperature-film thickness collaborative verification condition must also be met. The collaborative verification condition is as follows: The temperature anomaly is calculated based on the substrate material, the absolute temperature, and the rate of temperature rise. The film thickness anomaly is calculated based on the substrate material and the predicted film thickness deviation. Multiply the temperature anomaly and the film thickness anomaly to obtain the synergy factor; If the temperature anomaly is greater than a preset temperature anomaly threshold, the film thickness anomaly is greater than a preset film thickness anomaly threshold, and the synergy factor is greater than a preset synergy factor threshold, then the verification is deemed unsuccessful.

8. The method for monitoring the execution of an NCVM non-conductive vacuum coating process as described in claim 1, characterized in that, When any grid in the cross-validation grid fails validation, a three-dimensional compensation vector is generated, including: Extract temperature anomaly and film thickness anomaly information from the grid that failed the verification; Calculate the angle compensation amount of the plasma gun based on the temperature anomaly information; Based on the film thickness anomaly information, calculate the sputtering power compensation and residence time compensation. The angle compensation, sputtering power compensation, and dwell time compensation are integrated to generate the three-dimensional compensation vector.

9. The method for monitoring the execution of an NCVM non-conductive vacuum coating process as described in claim 8, characterized in that, Also includes: Calculate the edge compensation coefficient based on the radius of curvature of the mesh in the edge region of the substrate; The three-dimensional compensation vector is adjusted according to the edge compensation coefficient to obtain the three-dimensional enhancement compensation vector; After performing compensation on the edge region mesh according to the three-dimensional strengthening compensation vector, the edge secondary verification is initiated, and the edge region mesh is re-measured using a laser thickness gauge to obtain the re-measurement result. The edge region mesh is optimized by coating based on the retest results.

10. A monitoring system for the execution of an NCVM non-conductive vacuum coating process, characterized in that, The system is used to implement the NCVM non-conductive vacuum coating process execution monitoring method according to any one of claims 1-9, the system comprising: The detection unit grid dynamic division module is used to dynamically divide the substrate surface inside the vacuum coating cavity into detection unit grids to obtain an adaptive grid substrate surface. An infrared thermal imaging scanning module is used to scan the surface of the adaptive grid substrate in real time using an infrared thermal imager to obtain a dynamic temperature distribution matrix of the substrate. The coating uniformity prediction module is used to analyze and predict the dynamic temperature distribution matrix of the substrate based on the mapping relationship between coating thickness and temperature gradient, and obtain a coating uniformity prediction map. The alignment and fusion processing module is used to align and fuse the dynamic temperature distribution matrix of the substrate and the coating uniformity prediction map to obtain a cross-validation grid. A three-dimensional compensation vector generation module is used to generate a three-dimensional compensation vector when any one of the grids in the cross-validation grid fails the validation. The process parameter control module is used to adjust the process parameters of the vacuum coating equipment in real time according to the three-dimensional compensation vector.

Images