Coating process parameter self-adaptive optimization method, vacuum coating machine and storage medium

Abstract

The application relates to a coating process parameter self-adaptive optimization method, a vacuum coating machine and a storage medium; the method comprises the following steps: predicting key indexes of a target process formula by using a digital twin model of an ion beam sputtering coating equipment, and optimizing coating process parameters according to the deviation between the key index prediction value and a design target to obtain optimal process parameters; then the method is executed on the physical equipment of the ion beam sputtering coating equipment; in the thin film deposition process, the digital twin model is updated by using real-time sensor parameters, the deposition rate of the thin film is predicted online by using the updated digital twin model, and the actual deposited film thickness is adjusted in real time according to the prediction result to be consistent with the design value; according to the technical scheme, the virtual first process prediction and verification mechanism and the entity execution feedback control scheme are adopted, the process controllability and repeatability are improved, the number of physical test times is greatly reduced, and the research and development cost is reduced and the process development cycle is shortened.

Description

Technical Field

[0001] This application relates to the field of vacuum coating technology, and in particular to an adaptive optimization method for coating process parameters, a vacuum coating machine, and a computer-readable storage medium. Background Technology

[0002] In vacuum coating, when performing ion beam sputtering coating process, the control method is mostly based on preset programs and offline optimization. Operators pre-set process parameters such as ion source power, vacuum degree, gas flow rate and deposition time according to theoretical models or historical experience, and then control the vacuum coating equipment to execute the process parameters in sequence.

[0003] However, because ion beam sputtering is a complex process involving multiple variables, nonlinearity, and strong coupling, key intermediate parameters during sputtering (such as sputtering yield, particle energy distribution, and particle spatial angular distribution) are difficult to measure in real time, leading to incomplete perception of the process status and a tendency for the process to become a "black box." Simultaneously, time-varying disturbances such as aging of vacuum coating equipment (e.g., ion source decay) and environmental fluctuations (e.g., temperature drift) can cause the actual deposition environment (especially the deposition rate) to gradually deviate from the preset values, resulting in poor process repeatability. When developing new coating processes, it is usually necessary to repeatedly conduct open-loop experiments involving "deposition-offline testing-readjustment," which is time-consuming, costly, and makes it difficult to find globally optimal parameters. Furthermore, the evaluation of coating performance is typically conducted after the process is completed, making it impossible to dynamically adjust and optimize process parameters during the deposition process.

[0004] It is evident that existing ion beam sputtering coating technology is prone to defects such as process "black box", difficulty in dealing with dynamic disturbances, reliance on experience for debugging, and lag in coating optimization, which affect the efficiency and effect of vacuum coating. Summary of the Invention

[0005] To address one of the aforementioned technical deficiencies, this application provides an adaptive optimization method for coating process parameters, a vacuum coating machine, and a computer-readable storage medium.

[0006] An adaptive optimization method for coating process parameters includes: The target process formula for the coating process is obtained, and the target process formula is input into the digital twin model of the ion beam sputtering coating equipment to predict and obtain the predicted values ​​of key indicators of the coating process. Obtain the deviation between the predicted value of the key indicator and the design target, and optimize the coating process parameters based on the deviation value to obtain the optimal process parameters; The optimal process parameters are sent to the ion beam sputtering coating equipment for execution, and the sensor parameters of key components of the ion beam sputtering coating equipment are collected in real time during the thin film deposition process and synchronized to the digital twin model. The model state of the digital twin is dynamically updated using the sensor parameters. The updated digital twin model is used to predict the deposition rate of the thin film online. Based on the prediction results, the ion beam sputtering coating equipment is adjusted in real time to ensure that the actual deposited film thickness is consistent with the design value.

[0007] In some embodiments, the adaptive optimization method for coating process parameters further includes: A simplified physical mechanism model is established based on the physical relationship of the coating of the ion beam sputtering coating equipment to describe the basic physical relationship between the input parameters and the coating parameters of the ion beam sputtering coating equipment. A data-driven model is obtained by training a time-series neural network using historical coating process data and real-time sensing data from an ion beam sputtering coating equipment. The input of the data-driven model is the current coating process parameters, and the output is the predicted values ​​of intermediate coating parameters and key indicators for a future period. The physical mechanism model and the data-driven model are fused to obtain a hybrid model, and a digital twin model is generated based on the hybrid model using digital twin technology.

[0008] In some embodiments, the physical mechanism model and the data-driven model are fused to obtain a hybrid model, including: The output of the physical mechanism model is used as prior knowledge to constrain and correct the prediction results of the data-driven model, or the data-driven model is used as an online identifier for unknown or difficult-to-measure parameters in the physical mechanism model to form a hybrid model.

[0009] In some embodiments, the target process formulation is input into a digital twin model of an ion beam sputtering coating equipment to predict and obtain predicted values ​​of key indicators of the coating process, including: Based on the digital twin model, the target process formulation is rapidly simulated in virtual space, and the thin film deposition process under the target process formulation is predicted to obtain the predicted values ​​of various key indicators of the coating process.

[0010] In some embodiments, obtaining the deviation between the predicted value of the key indicator and the design target, and optimizing the coating process parameters based on the deviation to obtain the optimal process parameters, includes: Compare the predicted values ​​of each key indicator with the design target and calculate the deviation value; Based on the deviation value, the coating process parameters are optimized in the virtual space until the predicted values ​​of the key indicators meet the design target requirements to obtain the optimal process parameters.

[0011] In some embodiments, an updated digital twin model is used to predict the deposition rate of the thin film online, and the ion beam sputtering equipment is adjusted in real time based on the prediction results to ensure that the actual deposited film thickness matches the design value, including: The deposition rate of the thin film is predicted online using an updated digital twin model, and the required compensation is calculated based on the prediction results and the theoretical value of the deposition rate. The ion beam sputtering coating equipment is adjusted in real time according to the compensation amount so that the actual deposited film thickness is consistent with the design value.

[0012] In some embodiments, the digital twin model is used to predict the deposition rate of the thin film online, and the required compensation amount is calculated based on the prediction results and the theoretical value of the deposition rate, including: The deposition rate of the thin film is predicted online using the digital twin model to obtain a predicted deposition rate value. When the predicted deposition rate value deviates from the theoretical deposition rate value, the deviation of the predicted deposition rate value relative to the theoretical deposition rate value is calculated, and the required deposition rate compensation amount is calculated based on the deviation amount.

[0013] In some embodiments, the step of adjusting the ion beam sputtering coating equipment in real time according to the compensation amount so that the actual deposited film thickness is consistent with the design value includes: The ion source power or deposition time command of the ion beam sputtering coating equipment is adjusted in real time according to the deposition rate compensation amount, so that the actual deposited film thickness is consistent with the film thickness design value corresponding to the target process formula.

[0014] In some embodiments, the adaptive optimization method for coating process parameters further includes: The spectral characteristics of the coated film are acquired in real time using a light control system; The actual values ​​of film thickness and refractive index are calculated using the digital twin model, and then compared with the design values. If there is a deviation between the actual value and the design value, the remaining deposition time of the current layer is adjusted according to the deviation, and the film system design of the subsequent uncoated layers is optimized and reconstructed online based on the latest material optical constants.

[0015] A vacuum coating machine includes: a vacuum chamber, an ion source, a vacuum pumping device, and a control system; The vacuum chamber, ion source, and vacuum pumping equipment are equipped with various sensors; the sensors collect sensor data and synchronize it to the control system through an edge computing gateway. The control system is used to execute the steps of the adaptive optimization method for coating process parameters.

[0016] A computer-readable storage medium storing at least one instruction, at least one program, a code set, or an instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded by a processor and the steps of the adaptive optimization method for coating process parameters are executed.

[0017] The technical solution of the above embodiment first uses a digital twin model of the ion beam sputtering coating equipment to predict key indicators of the target process formula. Based on the deviation between the predicted key indicator values ​​and the design target, the coating process parameters are optimized to obtain the optimal process parameters. Then, the process is executed on the physical equipment of the ion beam sputtering coating equipment. During the thin film deposition process, the digital twin model is updated using real-time sensor parameters. The updated digital twin model is used to predict the deposition rate of the thin film online, and the coating process parameters of the ion beam sputtering coating equipment are adjusted in real time based on the prediction results, so that the actual deposited film thickness is consistent with the design value. This technical solution adopts a virtual-first process prediction and verification mechanism and a physical execution feedback control scheme. By predicting unmeasurable intermediate parameters and dynamically compensating for the deposition rate, it significantly suppresses the impact of disturbances such as equipment aging and environmental fluctuations, ensuring a high degree of consistency in product performance between batches, improving process controllability and repeatability. At the same time, it transforms the traditional trial-and-error process debugging into virtual verification and optimization, greatly reducing the number of physical experiments, lowering R&D costs, and shortening the process development cycle.

[0018] Furthermore, based on the aforementioned virtual-first process prediction and verification mechanism and physical execution feedback control, an adaptive closed-loop control with feedback optimization is further constructed to achieve "result-oriented" adaptive deposition and realize intelligent optimization of the coating process.

[0019] Additional aspects and advantages of this application will be set forth in part in the description which follows, and will become apparent from the description or may be learned by practice of this application. Attached Figure Description

[0020] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a hardware block diagram of an example adaptive optimization method for coating process parameters; Figure 2 This is a schematic diagram of the key components of an example ion beam sputtering coating equipment; Figure 3 This is a flowchart of an embodiment of an adaptive optimization method for coating process parameters; Figure 4 This is a flowchart of an adaptive optimization method for coating process parameters according to another embodiment; Figure 5 This is an example of an adaptive process optimization control architecture diagram; Figure 6 This is a schematic diagram of a digital twin construction scheme for an example ion sputtering coating equipment; Figure 7 This is a breakdown diagram of the structure and function of an example ion sputtering coating equipment; Figure 8 This is an example diagram of a Transformer-based temporal neural network architecture; Figure 9 This is a schematic diagram comparing the spectral curves of a virtual preview with those of an actual execution. Detailed Implementation

[0021] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0022] Those skilled in the art will understand that, unless otherwise stated, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the word “comprising” as used in this application’s specification means the presence of the stated feature, integer, step, or operation, but does not preclude the presence or addition of one or more other features, integers, steps, or operations.

[0023] The technical solution of this application aims to solve the technical defects of ion beam sputtering coating technology, such as the "black box" phenomenon of the process, the lag in process control and poor repeatability caused by dynamic disturbances or experience dependence; to achieve real-time and high-precision prediction of key but unmeasurable intermediate state parameters during sputtering; to conduct virtual debugging and optimization by predicting the impact of process parameter changes on the film formation results; and further, to establish an adaptive process control closed-loop scheme of "virtual first, physical execution, and feedback optimization" to achieve dynamic compensation of deposition rate and online self-tuning of process parameters.

[0024] In some embodiments, the technical solution of this application refers to Figure 1 As shown, Figure 1This is a hardware block diagram of an example adaptive optimization method for coating process parameters. Real-time control and data acquisition programs can be deployed on an industrial control computer (lower-level machine), and a digital twin control center can be deployed on a performance server (upper-level machine or cloud). For example, the digital twin control center can deploy a highly mapped digital twin based on digital twin technology, synchronized in real-time with the physical equipment of the ion beam sputtering coating equipment. This digital twin simulates the real ion beam sputtering coating equipment in virtual space, interacts with the user through a human-machine interface, and executes the technical solution of the adaptive optimization method for coating process parameters of this application.

[0025] In some embodiments, for constructing a digital twin, the parameters of key components of the ion beam sputtering coating equipment can be synchronized. This is achieved by installing sensors on the key components of the ion beam sputtering coating equipment, collecting real-time sensor data output by these sensors, and synchronizing it to the digital twin. Key components of the ion beam sputtering coating equipment can refer to components that play a crucial role in the coating process. The selection and classification of key components in practical applications depend on the different types of ion beam sputtering coating equipment and the requirements of the coating process.

[0026] refer to Figure 2 As shown, Figure 2 This is a schematic diagram of the key components of an example ion beam sputtering coating equipment. In this example, the key components of the ion beam sputtering coating equipment may include: an ion source system, a vacuum system, a target system, a cooling circulation system, an electrical system, and a workpiece disk mechanism; in addition, other key components may also be included, which will not be described in detail here.

[0027] For example, the sensors used may include: vibration sensors, temperature sensors, voltage / current sensors, vacuum sensors, flow sensors, acoustic emission sensors, etc.; the specific solution may be determined according to the type of key components and monitoring requirements.

[0028] For example, the ion source system may include a sputtering ion source and an auxiliary ion source. The ion source system can monitor the stability of the ion source beam, the stability of the working gas, and provide early warnings of abnormal discharges, short circuits, and component aging by setting voltage / current sensors, temperature sensors, acoustic emission sensors, flow sensors, etc.

[0029] For example, a vacuum system includes a vacuum chamber and a vacuum pump; wherein, the vacuum system can monitor vacuum fluctuations and temperature fluctuations in the vacuum chamber, abnormal voltage and current conditions of the vacuum pump, bearing wear, etc., by setting up vacuum sensors, temperature sensors, voltage / current sensors, acoustic emission sensors, etc., to ensure the basic environment of the process.

[0030] For example, the sputtering system may include a target, a baffle, etc.; wherein, the sputtering system can monitor power changes during the sputtering process by setting voltage / current sensors, acoustic emission sensors, etc., and provide early warning of abnormal target consumption, cooling failure or sparking.

[0031] For example, the cooling circulation system can be equipped with temperature sensors, flow sensors, etc., to monitor the temperature, flow rate and pressure of the coolant, prevent other critical components from overheating, and ensure equipment safety.

[0032] For example, electrical systems can be equipped with temperature sensors, voltage / current sensors, acoustic emission sensors, etc., to monitor the current, voltage, and heating status of various electrical devices and prevent power supply failures.

[0033] For example, the workpiece disk mechanism can be equipped with voltage / current sensors, acoustic emission sensors, position sensors, etc., to monitor the stability and accuracy of the workpiece disk's revolution / rotation, ensuring coating uniformity; and to provide early warnings of mechanical jamming, bearing failure, etc.

[0034] By installing sensors on key components of the ion beam sputtering coating equipment and controlling them with a PLC controller, a multi-source sensor network is constructed. All sensor data can be synchronized to the industrial control computer in real time via the Industrial Internet of Things protocol. The industrial control computer is ultimately mapped to a digital twin control center deployed on a performance server, enabling comprehensive status perception of the ion beam sputtering coating equipment.

[0035] The embodiments of the adaptive optimization method for coating process parameters of this application are described in detail below.

[0036] refer to Figure 3 As shown, Figure 3 This is a flowchart of an embodiment of an adaptive optimization method for coating process parameters, including: Step S10: Obtain the target process formula for the coating process, and input the target process formula into the digital twin model of the ion beam sputtering coating equipment to predict and obtain the predicted values ​​of key indicators of the coating process.

[0037] In this step, before the actual coating, the operator can first input the target process formula into the pre-trained digital twin model of the ion beam sputtering coating equipment, and then use the digital twin model to predict the key indicators of the coating process.

[0038] In some embodiments, the process of predicting and obtaining the predicted values ​​of key indicators of the coating process in step S10 above may include the following: Operators input the target process formula into the digital twin in the virtual space through a human-computer interaction interface. After the digital twin processes the target process formula, it is input into the digital twin model of the pre-trained ion beam sputtering coating equipment to quickly simulate and predict the entire deposition process under the target process formula, and obtain the predicted values ​​of various key indicators of the coating process, such as deposition rate stability, film thickness uniformity and final spectral curve.

[0039] In some embodiments, a digital twin model can be a high-fidelity digital twin model constructed by hybridizing a physical mechanism model and a data-driven model; for example, its construction process can be as follows: (1) Constructing a physical mechanism model: A simplified physical mechanism model is established based on the physical relationship of the coating process in ion beam sputtering equipment to describe the fundamental physical relationship between the input parameters and coating parameters of the ion beam sputtering equipment.

[0040] For example, a simplified physical mechanism model is established based on physical relationships such as sputtering yield formula, plasma transport equation, and thin film growth kinetics. This physical mechanism model describes the fundamental physical relationship between process parameters such as ion source parameters (voltage, beam current), target properties, and vacuum degree and deposition rate and film uniformity.

[0041] (2) Constructing a data-driven model: A data-driven model is obtained by training a time-series neural network using historical coating process data and real-time sensing data from an ion beam sputtering coating equipment. The input of the data-driven model is the current coating process parameters, and the output is the predicted values ​​of intermediate coating parameters and key indicators for a future period.

[0042] For example, a temporal neural network based on the Transformer architecture can be trained using historical process data and real-time sensor data (vacuum level, temperature, and status of various components). The input of this temporal neural network is the process parameters at the current and historical moments, such as ion source parameters, target parameters, vacuum chamber parameters, and particle positions at the previous moment. The output is the predicted values ​​of key intermediate parameters and final process indicators for future moments. Among them, key intermediate parameters may include sputtering yield, particle flux distribution, etc., and final process indicators may include deposition rate, film thickness, etc. After training with a large amount of data, a data-driven model is formed.

[0043] In the process of building a data-driven model, sensor data can be processed and feature data can be extracted. After a period of time, historical feature data and current real-time feature data will be obtained. At this time, the historical feature data is used as the input of the temporal neural network, and the current real-time feature data is used as the output. The current real-time feature data is used for verification and training of the temporal neural network. The training technology of temporal neural networks is a conventional technique and will not be elaborated here.

[0044] (3) Hybrid model fusion: The physical mechanism model and the data-driven model are fused to obtain a hybrid model, and a digital twin model is generated based on the hybrid model using digital twin technology.

[0045] For example, the output of the physical mechanism model can be used as prior knowledge to constrain and correct the prediction results of the data-driven model, or the data-driven model can be used as an online identifier for unknown / difficult-to-measure parameters in the physical mechanism model to form a hybrid model. The hybrid model can be integrated into a digital twin constructed based on digital twin technology to obtain a high-fidelity, extrapolable digital twin model.

[0046] Step S20: Obtain the deviation value between the predicted value of the key indicator and the design target, and optimize the coating process parameters based on the deviation value to obtain the optimal process parameters.

[0047] In this step, the design target is the target value that needs to be input in advance during coating. When there is a deviation between the predicted value of the key indicator and the design target, the coating process parameters are optimized based on the deviation value to obtain the optimal process parameters.

[0048] In some embodiments, the method for obtaining the optimal process parameters in step S20 above may include the following: The predicted values ​​of each key indicator are compared with the design target and the deviation value is calculated. Based on the deviation value, the coating process parameters are optimized in virtual space until the predicted values ​​of each key indicator meet the requirements of the design target to obtain the optimal process parameters.

[0049] For example, the prediction results are compared with the design target. If there is a deviation, the optimization loop is entered. The particle swarm optimization algorithm (PSO) or Bayesian optimization algorithm is used to iteratively optimize the coating process parameters such as ion source power, gas flow rate, and deposition time in the virtual space until the prediction results meet the design target requirements. Then, the corresponding coating process parameters are taken as the optimal process parameters.

[0050] As described in the above embodiments, a virtual-first process prediction and verification mechanism is established. Before coating, the process is first executed virtually on a digital twin, and then the process parameters are iteratively optimized in the virtual space to obtain the optimal process parameters before coating.

[0051] Step S30: The optimal process parameters are sent to the ion beam sputtering coating equipment for execution. During the thin film deposition process, the sensor parameters of the key components of the ion beam sputtering coating equipment are collected in real time and synchronized to the digital twin model.

[0052] In this step, the optimal process parameters are obtained by optimizing the aforementioned steps, and the ion beam sputtering coating equipment is controlled to execute the coating process. During the thin film deposition process, the sensor parameters of key components of the ion beam sputtering coating equipment are collected in real time, such as the vacuum degree and temperature of the vacuum chamber, and the real-time transmittance curve of the thin film detected by the optical control system. Then, these sensor parameters are synchronized to the digital twin model.

[0053] Step S40: The model state of the digital twin model is dynamically updated using the sensor parameters. The updated digital twin model is used to predict the deposition rate of the thin film online. Based on the prediction results, the ion beam sputtering coating equipment is adjusted in real time to ensure that the actual deposited film thickness is consistent with the design value.

[0054] In some embodiments, the adjustment process in step S40 may include the following: S401, the model state of the digital twin model is dynamically updated using the sensor parameters, making the predictions of the digital twin model more accurate.

[0055] S402 uses an updated digital twin model to predict the deposition rate of the thin film online, and calculates the required compensation based on the prediction results and the theoretical value of the deposition rate.

[0056] Specifically, the deposition rate of the thin film is predicted online using the digital twin model to obtain a predicted deposition rate value, which mainly predicts the deposition rate in the current and short future time. When the predicted deposition rate value deviates from the theoretical deposition rate value, the deviation of the predicted deposition rate value relative to the theoretical deposition rate value is calculated, and the required deposition rate compensation amount is calculated based on the deviation amount.

[0057] For example, if the deposition rate decreases due to the aging of the target material, the predicted deposition rate will deviate from the theoretical deposition rate. In this case, it is necessary to calculate the deviation and the compensation amount.

[0058] S403, The ion beam sputtering coating equipment is adjusted in real time according to the compensation amount so that the actual deposited film thickness is consistent with the design value.

[0059] For example, the ion source power or deposition time command of the ion beam sputtering coating equipment is adjusted in real time according to the deposition rate compensation amount, so that the actual deposited film thickness is consistent with the film thickness design value corresponding to the target process formulation.

[0060] The solutions described in the above embodiments significantly suppress the impact of disturbances such as equipment aging and environmental fluctuations by predicting unmeasurable intermediate parameters and dynamically compensating for deposition rates. This ensures a high degree of consistency in product performance between batches, improves process controllability and repeatability, and transforms traditional trial-and-error process debugging into virtual verification and optimization. This greatly reduces the number of physical tests, lowers R&D costs, and shortens the process development cycle.

[0061] To further improve product yield and intelligent optimization of coating processes, the technical solution of this application may also include an intelligent optimization scheme based on feedback optimization.

[0062] In some embodiments, reference Figure 4 As shown, Figure 4 This is a flowchart of an adaptive optimization method for coating process parameters according to another embodiment. The adaptive optimization method for coating process parameters in this application further includes: Step S50: The spectral characteristics of the coated film are acquired in real time, and the actual values ​​of film thickness and refractive index are calculated using the digital twin model. The film system design of the uncoated layer is optimized and reconstructed online based on the deviation between the actual values ​​and the design values.

[0063] In some embodiments, the above-described online optimization and refactoring technical solutions may include the following: S501 utilizes a light control system to acquire the spectral characteristics of the coated film in real time.

[0064] S502, the actual values ​​of film thickness and refractive index are calculated by inversion using the digital twin model, and the actual values ​​are compared with the design values.

[0065] S503, if there is a deviation between the actual value and the design value, adjust the remaining deposition time of the current layer according to the deviation, and optimize and reconstruct the film system design of the subsequent uncoated layers online based on the latest material optical constants.

[0066] As described in the above embodiments, the spectral characteristics of the grown film are monitored in real time by a broadband monitoring system. Through online film system reconstruction and parameter self-tuning, deposition deviations can be corrected in real time, ensuring that the optical performance of the final film accurately meets the design requirements. This avoids scrapping of the entire batch of products due to parameter drift, improves product yield, and can automatically search for the global optimal solution in a multi-dimensional, strongly coupled parameter space. This overcomes the limitations of manual experience-based optimization and discovers a better process window. Based on the previous virtual-first process prediction and verification mechanism and physical execution feedback control, this technical solution can further construct a feedback-optimized adaptive closed-loop control to achieve "result-oriented" adaptive deposition and realize intelligent optimization of the coating process.

[0067] To better illustrate the technical effects of the adaptive optimization method for coating process parameters in this application, the following describes an adaptive process optimization control scheme based on this method that realizes closed-loop control of "perception → decision → execution → optimization".

[0068] refer to Figure 5 As shown, Figure 5 This is an example of an adaptive process optimization control architecture diagram. In terms of system architecture, real-time control and data acquisition programs are deployed on the industrial control computer (lower-level machine); digital twin models and optimization algorithm engines are deployed on the performance server (upper-level machine or cloud); data interaction between the industrial control computer and the performance server can be achieved through low-latency protocols such as MQTT.

[0069] refer to Figure 6 As shown, Figure 6 This is a schematic diagram of a digital twin construction scheme for an ion sputtering coating equipment. The human-computer interaction module integrates a digital twin model, a 3D visualization rendering module, a data communication and management module, a data processing, integration and monitoring module, and a deep learning module, thus constructing a high-fidelity digital twin of the ion beam sputtering coating equipment.

[0070] For the functional classification of ion beam sputtering coating equipment, refer to Figure 7 As shown, Figure 7 This is an example of an ion sputtering coating equipment structure and functional breakdown diagram. As shown in the figure, the overall process control technology is divided into two parts: equipment control system scheme design and optical monitoring system scheme design. Then, the hardware framework design and construction, software control framework design, detailed software control function scheme design, hardware optical platform construction and testing, optical monitoring software function list sorting and calculation, and control technology scheme design are carried out. After the above design is completed, the equipment control system software development and optical monitoring system development are carried out. Finally, the functional debugging machine system testing and process experiment verification are carried out.

[0071] In the process of data-driven model training, reference Figure 8 As shown, Figure 8 This is an example of a Transformer-based temporal neural network architecture diagram. It collects a large amount of multi-process data from an ion beam sputtering coating equipment during normal operation, such as sensor logs, light control records, and final product inspection results, to train the Transformer neural network model offline, resulting in a data-driven model. Simultaneously, the sputtering yield model and deposition rate model based on physical formulas are encoded as physical mechanism models. By fusing the physical mechanism model and the data-driven model, a hybrid model is obtained. Based on digital twin technology, a digital twin model is generated on the basis of the hybrid model.

[0072] Before coating begins, the operator inputs the design goals and initial process parameters of the new film system into the Human Machine Interface (HMI) for virtual pre-deposition; the operator then invokes the digital twin service of the digital twin control center to simulate the deposition process, displaying the predicted spectral evolution curves, film thickness distribution, etc., on the HMI; such as Figure 9 As shown, Figure 9 This is a schematic diagram comparing the spectral curves of a virtual simulation and the actual execution; the solid line is the theoretical design curve, and the dashed line is the actual coating completion curve; the parameters are automatically adjusted through optimization algorithms until the simulation results meet the standards.

[0073] Following virtual simulation, online adaptive control is implemented. The physical equipment of the ion beam sputtering coating system operates according to the optimized formula, and real-time sensor data is collected to drive synchronous updates of the digital twin model. The updated digital twin model is then used to predict the current and short-term deposition rates online. For example, if the time required for the transmittance curve of the current layer to reach the stopping point of the current coating layer is 5% longer than theoretically predicted, indicating a lower actual deposition rate, the digital twin model quickly determines the deviation based on real-time vacuum level, ion source current, and other data. The deviation is attributed to a slight attenuation of the ion source. The ion beam sputtering coating system can then automatically increase the deposition time of subsequent layers by 5% or slightly increase the ion source power to compensate for the deposition rate deviation. Simultaneously, the spectral characteristics of the coating layer are acquired in real-time using the optical control system. Based on the latest refractive index obtained through inversion, the subsequent film system design is fine-tuned to ensure that the final optical performance remains unchanged.

[0074] In summary, the technical solutions of the above embodiments break the limitations of physical space by using digital twins, and achieve safe and efficient process optimization and verification through "virtual first". By forming a closed loop through "physical execution and real-time feedback", the model is continuously corrected and forward-looking compensation decisions are made using real-time data to drive the model, thereby upgrading the traditional "sequential + set value" control to adaptive intelligent control based on model prediction and real-time data.

[0075] The following describes an embodiment of a vacuum coating machine.

[0076] This application provides a vacuum coating machine, which mainly includes: a vacuum chamber, an ion source, a vacuum pumping device, and a control system; wherein, the vacuum chamber, the ion source, and the vacuum pumping device are equipped with a variety of sensors; each sensor synchronizes the collected sensing data to the control system through an edge computing gateway; the control system includes the aforementioned real-time digital twin control center, used to execute the steps of the coating process parameter adaptive optimization method of any embodiment of this application.

[0077] The vacuum coating machine of this application employs a virtual-first process prediction and verification mechanism and a physical execution feedback control scheme. By predicting unmeasurable intermediate parameters and dynamically compensating for deposition rates, it significantly suppresses the impact of disturbances such as equipment aging and environmental fluctuations, ensuring high consistency of product performance between batches, improving process controllability and repeatability. Simultaneously, it transforms traditional trial-and-error process debugging into virtual verification and optimization, greatly reducing the number of physical experiments, lowering R&D costs, and shortening the process development cycle. Furthermore, based on the aforementioned virtual-first process prediction and verification mechanism and physical execution feedback control, it further constructs an adaptive closed-loop control with feedback optimization to achieve "result-oriented" adaptive deposition, realizing intelligent optimization of the coating process.

[0078] The following describes an embodiment of a computer-readable storage medium.

[0079] The computer-readable storage medium provided in this application is used to implement the functions of the adaptive optimization method for coating process parameters. This computer-readable storage medium stores at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, at least one program, code set, or instruction set is loaded by a processor and executes the steps of the adaptive optimization method for coating process parameters of any embodiment.

[0080] In an exemplary embodiment, the computer-readable storage medium may be a non-transitory computer-readable storage medium that includes instructions, such as a memory that includes instructions. For example, a non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0081] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. An adaptive optimization method for coating process parameters, characterized in that, include: The target process formula for the coating process is obtained, and the target process formula is input into the digital twin model of the ion beam sputtering coating equipment to predict and obtain the predicted values ​​of key indicators of the coating process. Obtain the deviation between the predicted value of the key indicator and the design target, and optimize the coating process parameters based on the deviation value to obtain the optimal process parameters; The optimal process parameters are sent to the ion beam sputtering coating equipment for execution, and the sensor parameters of key components of the ion beam sputtering coating equipment are collected in real time during the thin film deposition process and synchronized to the digital twin model. The model state of the digital twin is dynamically updated using the sensor parameters. The updated digital twin model is used to predict the deposition rate of the thin film online. Based on the prediction results, the ion beam sputtering coating equipment is adjusted in real time to ensure that the actual deposited film thickness is consistent with the design value.

2. The adaptive optimization method for coating process parameters according to claim 1, characterized in that, Also includes: A simplified physical mechanism model is established based on the physical relationship of the coating of the ion beam sputtering coating equipment to describe the basic physical relationship between the input parameters and the coating parameters of the ion beam sputtering coating equipment. A data-driven model is obtained by training a time-series neural network using historical coating process data and real-time sensing data from an ion beam sputtering coating equipment. The input of the data-driven model is the current coating process parameters, and the output is the predicted values ​​of intermediate coating parameters and key indicators for a future period. The physical mechanism model and the data-driven model are fused to obtain a hybrid model, and a digital twin model is generated based on the hybrid model using digital twin technology.

3. The adaptive optimization method for coating process parameters according to claim 1, characterized in that, The hybrid model is obtained by fusing the physical mechanism model with the data-driven model, including: The output of the physical mechanism model is used as prior knowledge to constrain and correct the prediction results of the data-driven model, or the data-driven model is used as an online identifier for unknown or difficult-to-measure parameters in the physical mechanism model to form a hybrid model.

4. The adaptive optimization method for coating process parameters according to claim 1, characterized in that, The target process formulation is input into the digital twin model of the ion beam sputtering coating equipment to predict and obtain the predicted values ​​of key indicators of the coating process, including: Based on the digital twin model, the target process formulation is rapidly simulated in virtual space, and the thin film deposition process under the target process formulation is predicted to obtain the predicted values ​​of various key indicators of the coating process.

5. The adaptive optimization method for coating process parameters according to claim 4, characterized in that, Obtain the deviation between the predicted value of the key indicator and the design target, and optimize the coating process parameters based on the deviation to obtain the optimal process parameters, including: Compare the predicted values ​​of each key indicator with the design target and calculate the deviation value; Based on the deviation value, the coating process parameters are optimized in the virtual space until the predicted values ​​of the key indicators meet the design target requirements to obtain the optimal process parameters.

6. The adaptive optimization method for coating process parameters according to claim 5, characterized in that, An updated digital twin model is used to predict the deposition rate of thin films online, and the ion beam sputtering equipment is adjusted in real time based on the prediction results to ensure that the actual deposited film thickness matches the design value. This includes: The deposition rate of the thin film is predicted online using an updated digital twin model, and the required compensation is calculated based on the prediction results and the theoretical value of the deposition rate. The ion beam sputtering coating equipment is adjusted in real time according to the compensation amount so that the actual deposited film thickness is consistent with the design value.

7. The adaptive optimization method for coating process parameters according to claim 6, characterized in that, The digital twin model is used to predict the deposition rate of the thin film online, and the required compensation is calculated based on the prediction results and the theoretical value of the deposition rate, including: The deposition rate of the thin film is predicted online using the digital twin model to obtain a predicted deposition rate value. When the predicted deposition rate value deviates from the theoretical deposition rate value, the deviation of the predicted deposition rate value relative to the theoretical deposition rate value is calculated, and the required deposition rate compensation amount is calculated based on the deviation amount. The step of adjusting the ion beam sputtering coating equipment in real time according to the compensation amount to ensure that the actual deposited film thickness matches the design value includes: The ion source power or deposition time command of the ion beam sputtering coating equipment is adjusted in real time according to the deposition rate compensation amount, so that the actual deposited film thickness is consistent with the film thickness design value corresponding to the target process formula.

8. The adaptive optimization method for coating process parameters according to any one of claims 1 to 7, characterized in that, Also includes: The spectral characteristics of the coated film are acquired in real time using a light control system; The actual values ​​of film thickness and refractive index are calculated using the digital twin model, and then compared with the design values. If there is a deviation between the actual value and the design value, the remaining deposition time of the current layer is adjusted according to the deviation, and the film system design of the subsequent uncoated layers is optimized and reconstructed online based on the latest material optical constants.

9. A vacuum coating machine, characterized in that, include: Vacuum chamber, ion source, vacuum pumping equipment, and control system; The vacuum chamber, ion source, and vacuum pumping equipment are equipped with various sensors; the sensors collect sensor data and synchronize it to the control system through an edge computing gateway. The control system is used to execute the steps of the adaptive optimization method for coating process parameters according to any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The storage medium stores at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, the at least one program, the code set, or instruction set is loaded by a processor and executed by the steps of the adaptive optimization method for coating process parameters according to any one of claims 1-8.

Images