A comprehensive modeling platform for manufacturing equipment.

Abstract

The disclosed method includes interacting with a manufacturing system and processing chambers via a graphical user interface (GUI). It includes receiving user input and displaying data from the processing chambers in multiple modes, such as a first mode with actual processing operation data and a second mode with virtual processing operation data, and displaying the corresponding data in the GUI. The method further includes outputting a graphical view of the manufacturing system, comprising the processing chambers, transfer chambers, and load locks, in selectable view options, and loading virtual models associated with the selected processing chambers and view options. The method includes providing simulation inputs of virtual substrate processing operations to the second model to obtain predicted output data, both datasets being displayed in the GUI for analysis.

Description

Technical Field

[0001]

[0001] This disclosure relates to methods related to models used for evaluating manufacturing devices such as semiconductor devices. More specifically, this disclosure relates to a comprehensive modeling platform for manufacturing equipment.

Background Art

[0002]

[0002] Products can be manufactured by performing one or more manufacturing processes using manufacturing equipment. For example, a semiconductor manufacturing apparatus can be used to manufacture a substrate through a semiconductor manufacturing process. Products are manufactured with specific characteristics suitable for the target application. Various models such as digital twin models are used in various process control and prediction functions related to manufacturing equipment. Models associated with manufacturing equipment are trained and / or calibrated using data associated with the manufacturing equipment.

Summary of the Invention

[0003]

[0003] The following is a simplified summary of this disclosure and is provided to offer a basic understanding of some aspects of this disclosure. This summary is not an exhaustive overview of this disclosure. It is neither for identifying the main or important elements of this disclosure nor for defining any scope of specific embodiments of this disclosure or any scope of the claims. The sole purpose of this summary is to present, in a simplified form, some concepts of this disclosure as a prelude to the more detailed description that follows.

[0004]

[0004] In one aspect of the present disclosure, the method includes receiving a first user input via a graphical user interface (GUI) for a processing unit to view data related to a first processing chamber in a first chamber data mode. The data in the first chamber data mode includes data of processing operations performed within the first processing chamber. The method further includes providing first display data of the first chamber data mode in response to the receipt of the first user input for display on the GUI. The method further includes receiving a second user input to view data related to the first processing chamber in a second chamber data mode. The data in the second chamber data mode includes data of virtual processing operations performed by a virtual representation of the first processing chamber. The method further includes providing second display data of the second chamber data mode in response to the receipt of the second user input for display on the GUI.

[0005]

[0005] In another aspect of the present disclosure, the method includes outputting a view of a manufacturing system by an apparatus. The manufacturing system includes several processing chambers. The manufacturing system further includes one or more transport chambers. The manufacturing system further includes one or more load locks. The view includes a graphical representation of each processing chamber in the manufacturing system. The view further includes one or more view options for data associated with the processing chambers of the manufacturing system. The method further includes receiving input for a view option of a processing chamber from one or more view options of the processing chambers. The method further includes loading one or more virtual models associated with the selected processing chamber and the selected view option. Each processing chamber is associated with one or more virtual models configured for that processing chamber.

[0006]

[0006] In another aspect of the present disclosure, the method includes receiving a first user input to a first processing chamber via a GUI of a modeling platform. The method further includes providing first input data to a first model in response to receiving the first user input to the first processing chamber. The first input data includes sensor data from the first processing chamber. The method further includes obtaining first output data from the first model. The first output data includes prediction data related to the sensor data from the first processing chamber. The method further includes providing second input data to a second model in response to receiving the first user input to the first processing chamber. The second input data includes simulation inputs for virtual board processing operations. The method further includes obtaining second output data from the second model. The second output data includes prediction data for virtual board processing operations. The method further includes providing first and second output data for display on a GUI of a modeling platform.

[0007]

[0007] The present disclosure is illustrated, not limited, in the drawings of the attached drawings. [Brief explanation of the drawing]

[0008] [Figure 1] This is a block diagram illustrating an exemplary system architecture according to several embodiments. [Figure 2A] A block diagram of a system including an exemplary dataset generator for creating datasets for one or more supervised models, according to several embodiments, is shown. [Figure 2B] A block diagram of an exemplary dataset generator for creating a dataset to calibrate a physically based model, according to several embodiments, is shown. [Figure 3] This is a block diagram showing a system for generating output data according to several embodiments. [Figure 4A] This is a flowchart of a method for generating a model dataset according to several embodiments. [Figure 4B] This is a flowchart illustrating a method for utilizing a comprehensive modeling platform for data analysis related to processing chambers, according to several embodiments. [Figure 4C] This is a flowchart illustrating a method for utilizing a model using a comprehensive modeling platform, according to several embodiments. [Figure 4D] This is a flowchart illustrating a method for operating a comprehensive modeling platform according to several embodiments. [Figure 5A] This is an overview of a GUI for selecting a processing chamber for a comprehensive modeling platform, according to several embodiments. [Figure 5B] An overview of the GUI for the first mode of the comprehensive modeling platform, based on several embodiments, is shown. [Figure 5C] This section provides an overview of the GUI for a second mode of the comprehensive modeling platform, based on several embodiments. [Figure 6] This is a block diagram showing a computer system according to several embodiments. [Modes for carrying out the invention]

[0009]

[0020] This specification describes a technology relating to a platform for presenting information demonstrating the performance of processing chambers for improving substrate manufacturing. Manufacturing equipment is used to manufacture products such as substrates (e.g., wafers, semiconductors). Manufacturing equipment may include a manufacturing chamber or processing chamber for isolating the substrate from the environment. The characteristics of the manufactured substrate are to meet target values ​​in order to promote specific functions. Manufacturing parameters are selected to manufacture a substrate that meets the target characteristic values. Many manufacturing parameters (e.g., hardware parameters, process parameters, etc.) contribute to the characteristics of the processed substrate. The manufacturing system can control parameters by specifying setpoints for characteristic values, receiving data from sensors placed in the manufacturing chamber, and adjusting the manufacturing equipment until the sensor readings match the setpoints. Data generated by the physical processing chamber (e.g., sensor data, virtual sensor data, virtual measurement data, etc.) and data generated by the corresponding virtual processing chamber can be used for generating input / output mappings, designing new processing recipes, updating current processing recipes, etc. Physical-based models, statistical models, and trained machine learning models can be used to improve the performance of the manufacturing equipment.

[0010]

[0021] Models are used to provide an understanding of processing behavior. The quality of process performance can be understood based on the output of one or more models. Quality of performance may refer to the extent to which the processing behavior meets design objectives, such as threshold characteristic conditions of the processed substrate or threshold processing conditions within the chamber. Models may include physical-based models, such as physical-based digital twin models. Models may include database models, such as machine learning models. Models may include statistical models, holistic models, rule-based models, models, or combinations of model types.

[0011]

[0022] A model can be used to predict one or more characteristics of a substrate processing operation. These characteristics may include chamber conditions, substrate conditions, hardware conditions, etc. A model representing chamber performance quality may be based on physical or virtual chamber conditions. A model based on physical chamber conditions may be provided as input, such as data related to the performance of one or more processing operations. For example, manufacturing parameters, substrate design parameters, and / or sensor data may be provided to one or more models. One or more models can generate virtual sensor data, predicted substrate measurement data, or other types of data representing the performance quality of process operations. A model based on physical chamber conditions may be a physical-based model. A model based on physical chamber conditions may be calibrated with a physical-based model. A model based on physical chamber conditions may be a digital twin model. A model based on physical chamber conditions may be a machine learning model, database model, physical-based model, rule-based model, heuristic model, statistical model, etc. A model based on physical chamber conditions may receive data based on the physical performance of the substrate processing operation, which may be costly to generate in terms of processing materials, waste of processed substrates, energy, environmental impact, etc.

[0012]

[0023] A model based on virtual chamber conditions may be provided as input, in the form of one or more virtual processing input conditions. These virtual processing input conditions may or may not correspond to physical processing input conditions (e.g., manufacturing or hardware parameters). A model based on virtual chamber conditions may be a database model. A model based on virtual chamber conditions may be a machine learning model. A model based on virtual chamber conditions may be an artificial intelligence-based model. A model based on virtual chamber conditions may be a physical-based model, a heuristic model, a statistical model, a rule-based model, etc. The input conditions may be provided to the model and virtual substrate processing being executed. Outputs from one or more models may be provided that demonstrate the expected operational performance quality related to the given input conditions. While models based on virtual chamber conditions may be less expensive to execute, establishing the model, particularly a database model, may involve providing large amounts of training and / or calibration data, labeling the data, and ensuring consistency between virtual and physical chamber performance.

[0013]

[0024] In some systems, different types of models are accessed, used, and / or understood separately. For example, models for evaluating processing chamber quality and / or processing operation performance quality can be divided into models based on sensor data of processing operations performed in a physical processing chamber and models based on processing operations performed by a virtual processing chamber, such as a digital twin corresponding to the physical processing chamber. Modeling based on physical processing operations can be used for evaluating chamber performance, predicting substrate measurements, etc. Modeling based on virtual processing operations can be used for predicting the results of process recipes under development, predicting process conditions and / or substrate characteristics, etc. Such categories of models may have different structures (e.g., physical-based models and database models). Such categories may have different usefulness for different users. Models may be accessed in some systems through various platforms or applications. For users interested in chamber quality, multiple models may be used, which may be difficult or inconvenient to access through various model platforms.

[0014]

[0025] In some systems, various models may exist for a particular processing chamber, processing chamber type, processing operation, etc. For example, one or more models may be used to predict or evaluate substrate characteristics, one or more models may be used to predict the temperature of the processing chamber or substrate, a model may be used to predict magnetic field characteristics, and a model may be used to predict plasma characteristics. Such models may be accessible through different computing devices, different platforms, different applications, etc. Such model separation can cause inconvenience or difficulty in building an understanding of the quality of the processing chamber.

[0015]

[0026] In some systems, models associated with various processing chambers can be further separated. For example, a tool can include several processing chambers configured to perform different types of processing operations such as deposition operations, etching operations, annealing operations, lithography operations, and the like. The tool can include chambers for performing physical vapor deposition (PVD) operations, chemical vapor deposition (CVD) operations, atomic layer deposition (ALD) operations, etching operations, lithography operations, and the like. To view performance data associated with a processing tool that includes multiple chambers, access can be made to models associated with individual chambers and / or individual chamber types via various platforms, applications, and the like. Such separation of models can cause inconvenience or difficulty in presenting data indicative of the performance quality of the processing tool.

[0016]

[0027] In some systems, models associated with various tools or manufacturing facilities can be further separated. Similar to the separation of models associated with various chambers, accessing models or model data associated with various tools or various facilities by a user can be inconvenient, difficult, and / or expensive in terms of time, training, expertise, and the like. Such separation of models can cause inconvenience or difficulty in presenting data indicative of the performance quality of a manufacturing facility or a set of facilities.

[0017]

[0028] The methods and systems of the present disclosure can address one or more drawbacks of conventional solutions. Comprehensive manufacturing equipment performance quality applications, programs, platforms, etc. can be constructed. The comprehensive platform can provide users with access to many models related to the performance of manufacturing equipment. The comprehensive platform can provide access to models that predict various characteristics related to manufacturing equipment. The comprehensive platform can provide access to models related to various tools, processing operations, chambers, facilities, etc. The comprehensive platform can provide access to models that generate predictions based on the physical processing operations of manufacturing equipment. The comprehensive platform can provide access to models that generate predictions based on the virtual processing operations of manufacturing equipment. The comprehensive platform can provide various data displays for displaying process operation data, process operation performance quality data, etc. The comprehensive platform can be used to improve processing recipes. The comprehensive platform can be used in the development of process recipes. The comprehensive platform can be used to identify manufacturing defects. The comprehensive platform can be used to determine the root cause of manufacturing failures. The comprehensive platform can be used to identify the drift and / or aging of manufacturing equipment. The comprehensive platform can be used to identify the failure of one or more components of manufacturing equipment. The comprehensive platform can be used when recommending maintenance of manufacturing equipment. The comprehensive platform can be used when recommending replacement of one or more components of manufacturing equipment.

[0018]

[0029] The comprehensive platform can enable the selection of a target processing chamber. The selection of the target processing chamber can include the selection of a manufacturing facility, the selection of a tool including the target processing chamber, the selection of the type or category of the chamber, the selection of the type of process operation associated with the chamber, the selection of the type of substrate associated with the chamber, etc.

[0019]

[0030] A comprehensive platform may provide a list-type selection of processing chambers. A comprehensive platform may provide a graphical selection of processing chambers. A comprehensive platform may provide a combination or selection of graphical, list-type, or other types of selections. In some embodiments, a graphical representation of the selected tools may be presented, and the selection of target chambers may be performed by selecting graphical user interface (GUI) elements associated with the chambers from the graphical representation.

[0020]

[0031] Modeling capabilities associated with processing chambers may be provided to the user in response to chamber selection. The model may be configured to reflect the associated processing chamber. For example, a PVD chamber may be associated with a PVD virtual model, an ALD chamber with an ALD virtual model, and an etching chamber with an etching model. In some embodiments, a selection of modeling capability categories and data access may be presented. For example, a first selection may be associated with data collected from physical processing operations previously performed within the chamber (e.g., a "monitor" data mode). A second selection may be associated with data collected from virtual processing operations performed within the virtual chamber (e.g., a "exploration" data mode). Modeling operations may be performed in response to user selections, requests, commands, etc. In some cases, modeling operations may have been performed previously, and the output data of the modeling may be displayed in response to user actions.

[0021]

[0032] In some embodiments, one or more datasets are displayed via a GUI based on user selections such as chamber, model category, and target data. Further subdivision into categories may be performed by an inclusive platform. For example, the monitor data mode may be divided into virtual sensor data (e.g., sensor data provided from physically based and / or digital twin models based on chamber sensor data) and measured sensor data (e.g., data from chamber sensors, measurement tools, etc.). The exploration data mode may be divided into a prediction mode (e.g., inputs to one or more models include process / simulation knobs, and outputs include a display of performance quality) and an exploration mode (e.g., inputs to one or more models include target outputs for a process, and outputs include process knob inputs that are predicted to yield the target outputs). Various data presentations, including histograms, time traces, substrate characteristic maps, or other data presentation methods, may be provided via the GUI.

[0022]

[0033] The system and method of this disclosure offer advantages over conventional methods. By providing a comprehensive platform for models related to the quality of manufacturing equipment, the cost of actions to improve processing quality can be reduced in terms of user time, training for user expertise, and the reliability and accuracy of adjustments made to the manufacturing equipment. Furthermore, by including multiple modeling frameworks in a single comprehensive platform, cross-checking becomes possible to monitor chamber degradation, model degradation, component failures, required maintenance, and chamber aging. By including access to many models such as different structures, designs, intentions, types, and categories in a single comprehensive platform, a more complete understanding of chamber performance becomes possible, potentially improving results related to chamber maintenance, chamber and recipe design, PCB design, and PCB manufacturing.

[0023]

[0034] In one aspect of the present disclosure, the method includes receiving a first user selection via a graphical user interface (GUI) for a processing unit to display data related to a first processing chamber in a first chamber data mode. The data in the first chamber data mode includes data of processing operations performed within the first processing chamber. The method further includes providing first display data of the first chamber data mode in response to the receipt of the first user selection for display on the GUI. The method further includes receiving a second user selection for displaying data related to the first processing chamber in a second chamber data mode. The data in the second chamber data mode includes data of virtual processing operations performed by a virtual representation of the first processing chamber. The method further includes providing second display data of the second chamber data mode in response to the receipt of the second user selection for display on the GUI.

[0024]

[0035] In another aspect of the present disclosure, the method includes outputting a view of a manufacturing system by an apparatus. The manufacturing system includes several processing chambers. The manufacturing system further includes one or more transfer chambers. The manufacturing system further includes one or more load locks. The view includes a graphical representation of each processing chamber in the manufacturing system. The view further includes one or more view options for data associated with the processing chambers of the manufacturing system. The method further includes receiving a selection of view options for a processing chamber from one or more view options among the processing chambers. The method further includes loading one or more virtual models associated with the selected processing chamber and the selected view options. Each processing chamber is associated with one or more virtual models configured for that processing chamber.

[0025]

[0036] In another aspect of this disclosure, the method includes receiving a first user selection of a first processing chamber via a GUI of a modeling platform. The method further includes providing first input data to a first model in response to receiving a first user selection of a first processing chamber. The first input data includes sensor data from the first processing chamber. The method further includes obtaining first output data from the first model. The first output data includes predictive data related to the sensor data from the first processing chamber. The method further includes providing second input data to a second model in response to receiving a first user selection of a first processing chamber. The second input data includes simulation inputs for virtual board processing operations. The method further includes obtaining second output data from the second model. The second output data includes predictive data for virtual board processing operations. The method further includes providing first and second output data for display on a GUI of a modeling platform.

[0026]

[0037] The embodiments are described with reference to a GUI for a modeling platform and discuss user selection of options displayed within and / or associated with the GUI (e.g., process parameters, modeling parameters, recipes, etc.) by clicking drop-down menus and / or other GUI elements. However, it should be understood that user input may be received in other ways as well. For example, user input may be received as a prompt to a large language model (LLM) or other artificial intelligence (AI) model (e.g., a generative model). The prompt may be provided by voice and / or text and may provide a description of the GUI element, process parameter, model parameter, etc., that the user wishes to select. The LLM or other AI model may process the prompt to determine the selection to be made within the application associated with the GUI and send instructions to the application associated with the GUI to implement the selection. For example, the prompt may indicate a set of model parameters via voice or text. The LLM or other AI model may process the prompt to determine the set of model parameters and then send instructions to the application associated with the GUI to select the set of model parameters.

[0027]

[0038] Language-Language-Based Models (LLMs) are a type of artificial intelligence model designed to understand, generate, and operate on human language. These models are trained on vast amounts of text data to learn linguistic patterns such as grammar, facts, and context. In LLMs, "large" refers to the size of the model, typically characterized by the number of parameters (e.g., variables or "weights") the model has, often in the billions or even trillions. An example of an LLM is GPT-4.

[0028]

[0039] In some embodiments, instead of a traditional user interface where the user enters input values ​​or selects input entries from a menu, the user provides input through a "conversational" interaction with the system (for example, using LLM).

[0029]

[0040] Figure 1 is a block diagram showing an exemplary system 100 (exemplary system architecture) in several embodiments. System 100 includes a client device 120, manufacturing equipment 124, sensors 126, measuring equipment 128, a prediction server 112, and a data store 140. The prediction server 112 may be part of the prediction system 110. The prediction system 110 may further include server machines 170 and 180.

[0030]

[0041] Sensor 126 can provide sensor data 142 related to manufacturing equipment 124 (for example, related to manufacturing corresponding products such as substrates by manufacturing equipment 124). Sensor data 142 can be used to verify the health of the equipment and / or the health of the product (e.g., product quality). Manufacturing equipment 124 can manufacture products according to a recipe or perform an operation over a period of time. In some embodiments, sensor data 142 may include one or more values ​​from optical sensor data, spectral data, temperature (e.g., heater temperature), spacing (SP), pressure, high-frequency radio frequency (HFRF), high-frequency (RF) matching voltage, RF matching current, RF matching capacitor position, electrostatic chuck (ESC) voltage, actuator position, current, flow, power, voltage, etc. Sensor data 142 may include historical sensor data 144 and current sensor data 146. Current sensor data 146 may be related to the product currently being processed, recently processed product, number of recently processed products, etc. Current sensor data 146 can be used as input to a model to generate, for example, predictive data 168, virtual sensor data 167, etc. Historical sensor data 144 may include data stored in relation to previously manufactured products. Historical sensor data 144 can be used to train and / or calibrate a model (e.g., model 190). Historical sensor data 144 and / or current sensor data 146 may include attribute data, such as manufacturing equipment ID or design label, sensor ID, type, and / or location, manufacturing equipment status label, such as current failure, service life, etc.

[0031]

[0042] The sensor data 142 may be associated with or indicate manufacturing parameters such as hardware parameters of the manufacturing equipment 124 (e.g., hardware settings, or indication of installed components, e.g., size, make, model, type, etc.) or process parameters of the manufacturing equipment 124 (e.g., heater settings, gas flow, etc.). Some data associated with hardware parameters and / or process parameters may be stored instead or additionally as manufacturing parameters 150, which may include past and current manufacturing parameters (e.g., associated with past processing runs). Manufacturing parameters 150 may indicate input settings to the manufacturing device (e.g., heater power, gas flow, etc.). The sensor data 142 and / or manufacturing parameters 150 may be provided while the manufacturing equipment 124 is performing a manufacturing process (e.g., readings from the equipment while processing a product). The sensor data 142 may differ for each product (e.g., for each substrate). The substrate may have characteristic values ​​(e.g., film thickness, film strain) measured by the measuring instrument 128. The measuring instrument 128 may include instruments included in a processing tool or chamber, measuring instruments integrated with a tool, measuring instruments in a standalone facility, etc. The measurement data 160 may be a component of the data store 140. The measurement data 160 may include historical measurement data (for example, measurement data related to previously processed products).

[0032]

[0043] In some embodiments, the measurement data 160 may be provided without using a standalone measurement facility, e.g., in-site measurement data (e.g., measurements or proxies for measurements collected during processing), integrated measurement data (e.g., measurements or proxies for measurements collected when the product is in the chamber or under vacuum but not during the processing operation), or in-line measurement data (e.g., data collected after the substrate has been removed from the vacuum). The measurement data 160 may include current measurement data (e.g., measurement data related to the currently or recently processed product).

[0033]

[0044] In some embodiments, sensor data 142, measurement data 160, or manufacturing parameters 150 may be processed (e.g., by a client device 120 and / or a prediction server 112). Processing the sensor data 142 may include generating features. In some embodiments, features are patterns in the sensor data 142, measurement data 160, and / or manufacturing parameters 150 (e.g., slope, width, height, peak, etc.). In some embodiments, features are combinations of values ​​from the sensor data 142, measurement data 160, and / or manufacturing parameters 150 (e.g., power derived from voltage and current). The sensor data 142 may include features, which may be used by the prediction component 114 to perform signal processing and / or to obtain prediction data 168. The prediction data 168 (e.g., including virtual sensor data 167, virtual board data 169, or other types of prediction data) may be used to perform corrective actions. The predictive data 168 may be presented to the user, for example, via a graphical user interface or via a client device 120.

[0034]

[0045] Each instance (e.g., set) of sensor data 142 may correspond to a product (e.g., a circuit board), a set of manufacturing equipment, or the type of circuit board manufactured by the manufacturing equipment. Similarly, each instance of measurement data 160 and manufacturing parameter 150 may correspond to a product, a set of manufacturing equipment, or the type of circuit board manufactured by the manufacturing equipment. The data store can further store information that associates sets of different data types, such as information indicating that sets of sensor data, sets of measurement data, and sets of manufacturing parameters are all associated with the same product, manufacturing equipment, circuit board type, etc.

[0035]

[0046] In some embodiments, the prediction system 110 can generate prediction data 168 using supervised machine learning. Supervised machine learning may include utilizing a machine learning model trained on labeled data. Labeled data may include sensor data labeled with measurement data, sensor data labeled with probe data to generate virtual sensor data, manufacturing parameters labeled with sensor data or measurement data, etc. Labeled data may include any training input data of interest in relation to the resulting data. A physically based model may be calibrated in a similar manner by modifying or adjusting one or more parameters of the physically based model using some of the labeled data. In some embodiments, the prediction system 110 can generate prediction data 168 using unsupervised machine learning. For example, the prediction data 168 may include the output from a machine learning model trained on unlabeled data, the output may include clustering results, principal component analysis, anomaly detection, etc. In some embodiments, the prediction system 110 can generate prediction data 168 using semi-supervised learning. For example, the training data may include a mixture of labeled and unlabeled data, etc.

[0036]

[0047] The client device 120, manufacturing equipment 124, sensor 126, measuring instrument 128, prediction server 112, data store 140, server machine 170, and server machine 180 may be connected to each other via a network 130 to generate prediction data 168 and perform corrective actions. In some embodiments, the network 130 may provide access to cloud-based services. The operations performed by the client device 120, prediction system 110, data store 140, etc., may be performed by a virtual cloud-based device.

[0037]

[0048] In some embodiments, network 130 is a public network providing client device 120 with access to prediction server 112, data store 140, and other publicly available computing devices. In some embodiments, network 130 is a private network providing client device 120 with access to manufacturing equipment 124, sensors 126, measuring instruments 128, data store 140, and other privately available computing devices. Network 130 may include one or more wide area networks (WANs), local area networks (LANs), wired networks (e.g., Ethernet networks), wireless networks (e.g., 802.11 networks or Wi-Fi networks), cellular networks (e.g., Long Term Evolution (LTE) networks), routers, hubs, switches, server computers, cloud computing networks, and / or combinations thereof.

[0038]

[0049] The client device 120 may include computing devices such as personal computers (PCs), laptops, mobile phones, smartphones, tablet computers, netbooks, network-connected televisions ("smart TVs"), network-connected media players (e.g., Blu-ray players), set-top boxes, over-the-top (OTT) streaming devices, and operator boxes. The client device 120 may include a corrective action component 122. The corrective action component 122 can receive user input of instructions related to the manufacturing equipment 124 (e.g., via a graphical user interface (GUI) displayed through the client device 120). In some embodiments, the corrective action component 122 transmits instructions to a prediction system 110, receives output from the prediction system 110 (e.g., prediction data 168), determines corrective action based on the output, and has the corrective action executed. In some embodiments, the corrective action component 122 retrieves sensor data 142 associated with the manufacturing equipment 124 (e.g., current sensor data) (e.g., from a data store 140, etc.) and provides sensor data 142 associated with the manufacturing equipment 124 (e.g., current sensor data 146) to the prediction system 110.

[0039]

[0050] In some embodiments, the corrective action component 122 receives instructions for corrective action from the prediction system 110 and causes the corrective action to be executed. Each client device 120 may include an operating system that enables the user to generate, display, or edit one or more data (e.g., displays related to manufacturing equipment 124, corrective actions related to manufacturing equipment 124, etc.). The client device 120 may run an application including a graphical user interface (e.g., a comprehensive modeling platform) to display data to the user and / or receive input from the user.

[0040]

[0051] The client device 120 can run one or more models (e.g., model 190). The client device 120 may run one or more machine learning models. The client device 120 may run one or more physics-based models. The client device 120 may run one or more statistical models, rule-based models, heuristic models, or other types of models. The client device 120 may run a comprehensive modeling platform to provide the user with access to one or more models.

[0041]

[0052] In some embodiments, measurement data 160 (e.g., historical measurement data) corresponds to historical characteristic data of a product (e.g., a product processed using historical sensor data 144 and manufacturing parameters related to historical manufacturing parameters 150), and prediction data 168 is associated with prediction characteristic data (e.g., prediction characteristic data of a product manufactured under conditions recorded by current sensor data 146 and / or current manufacturing parameters, or a product that should be manufactured under such conditions). In some embodiments, prediction data 168 is or includes prediction measurement data (e.g., virtual measurement data) of a product that is manufactured or manufactured according to conditions recorded as current sensor data 146, current measurement data, current measurement data, and / or current manufacturing parameters. In some embodiments, prediction data 168 is or includes any anomalies (e.g., an abnormal product, an abnormal component, an abnormal manufacturing equipment 124, abnormal energy usage, etc.) and optionally an indication of one or more causes of the anomaly. In some embodiments, the predictive data 168 is an indicator of change or drift over time in several components, such as manufacturing equipment 124, sensors 126, and measuring instruments 128. In some embodiments, the predictive data 168 is an indication of the end of life of components, such as manufacturing equipment 124, sensors 126, and measuring instruments 128. In some embodiments, the predictive data 168 is an indication of the progress of a processing operation being performed (e.g., used for process control).

[0042]

[0053] In some embodiments, a comprehensive modeling platform can be used to generate, display, and manipulate data related to one or more processing chambers. A comprehensive modeling platform can be used to run numerous models, including models of different types, structures, functions, and categories. The comprehensive modeling platform can run one or more models to generate virtual sensor data 167. The virtual sensor data 167 may be data modeled based on sensor data received from sensors 126 of the manufacturing equipment 124. The comprehensive modeling platform can run one or more models to generate virtual board data 169. The virtual board data 169 may include data generated by one or more models performing virtual board processing operations. For example, the virtual board data 169 may be data generated in relation to a digital twin model of a processing chamber, virtual processing operations of a virtual processing chamber, etc. As used herein, a digital twin model is a virtual model of a physical object (e.g., a processing chamber). It extends to the object's lifecycle and can simulate and monitor the object's behavior using real-time data transmitted from sensors on the object. A digital twin of a processing chamber can be configured to run in a manner that mimics the physical processing chamber when inputs corresponding to the physical and digital processing chambers (e.g., processing inputs corresponding to simulation inputs) are provided. The digital twin model may be, or may include, a physically based model, a database (e.g., machine learning) model, or a hybrid model (containing features of both a database model and a physically based model).

[0043]

[0054] A comprehensive modeling platform can enable the execution of numerous models related to substrate processing chambers. Chamber performance analysis may include monitoring the performance of physical processing operations performed within a physical chamber. Chamber performance analysis may include executing virtual processing operations within a virtual processing chamber (e.g., a digital twin). Chamber performance analysis may include monitoring the performance of virtual processing operations. Analysis of physical processing operations can enable the implementation of corrective actions by indicating, for example, recommended recipe updates, recommended maintenance, failures, and / or aging chamber components. Performance and analysis of virtual processing operations can offer similar advantages, in addition to providing a means to test substrate processing operations (e.g., during processing recipe development) without the costs associated with physical substrate processing, such as materials, energy, chamber time, resulting substrate disposal, and environmental impact. Furthermore, comparisons of physical and virtual chamber performance can provide indicators of aging and failure of chamber components, recommended maintenance, and recommended corrective actions.

[0044]

[0055] Implementing a manufacturing process that results in defective products can incur costs such as time, energy, products, parts, manufacturing equipment 124, and the costs of identifying defects and discarding defective products. By inputting sensor data 142 (e.g., manufacturing parameters used or planned to be used to manufacture a product) into a predictive system 110, receiving the output of predictive data 168, and taking corrective action based on the predictive data 168, system 100 can have the technical advantage of avoiding the costs of manufacturing, identifying, and discarding defective products.

[0045]

[0056] Executing a manufacturing process that results in the failure of components of the manufacturing equipment 124 can incur costs such as downtime, product damage, equipment damage, and expedited orders for replacement components. By inputting sensor data 142 (e.g., indicating manufacturing parameters used or planned to be used to manufacture a product), manufacturing parameters 150, measurement data, etc., into model 190, receiving the output of predictive data 168, and performing corrective actions (e.g., predictive operational maintenance such as component replacement, processing, and cleaning) based on the predictive data 168, system 100 may have the technical advantage of avoiding one or more costs such as unexpected component failure, unplanned downtime, productivity loss, unexpected equipment failure, unplanned failure, and product scrap. For example, monitoring the performance of components such as the manufacturing equipment 124, sensor 126, measuring instrument 128, and virtual and physical processing chambers over time can provide indicators of component degradation.

[0046]

[0057] Manufacturing parameters may not be optimal for manufacturing products, which can result in costly consequences such as increased resource consumption (e.g., energy, coolant, gas, etc.), increased time spent manufacturing products, increased component failures, and increased quantities of defective products. By inputting input data representing the manufacturing process into Model 190, receiving output from the model, and performing corrective actions (e.g., based on model-generated predictive data 168) to update the manufacturing parameters (e.g., setting optimal manufacturing parameters), System 100 can have the technical advantage of using optimal manufacturing parameters (e.g., hardware parameters, process parameters, optimal design) to avoid the costs resulting from suboptimal manufacturing parameters.

[0047]

[0058] Corrective actions may be associated with one or more of the following: computational processing control (CPC), statistical processing control (SPC) (e.g., SPC on electronic components to determine the process under control, SPC to predict the useful life of components, etc.), advanced processing control (APC), model-based processing control, preventive operational maintenance, design optimization, manufacturing parameter updates, manufacturing recipe updates, feedback control, machine learning modifications, etc.).

[0048]

[0059] In some embodiments, the corrective action includes providing an alarm (e.g., an alarm to stop or not run the manufacturing process if the predictive data 168 indicates a predicted anomaly, such as a product, part, or manufacturing equipment 124). In some embodiments, a machine learning model is trained to monitor the progress of the process execution (e.g., monitor in-sensor data to predict whether the manufacturing process is complete). In some embodiments, the machine learning model may send an instruction to terminate the process execution when the model determines that the process is complete. In some embodiments, the corrective action includes providing feedback control (e.g., modifying manufacturing parameters in response to predictive data 168 indicating a predicted anomaly). In some embodiments, the implementation of the corrective action includes ensuring that one or more manufacturing parameters are updated. In some embodiments, the implementation of the corrective action may include retraining the machine learning model associated with the manufacturing equipment 124. In some embodiments, the implementation of the corrective action may include training a new machine learning model associated with the manufacturing equipment 124. In some embodiments, the implementation of the corrective action may include recalibrating the physical-based model. In some embodiments, the implementation of corrective action may include, for example, providing the user with a display of abnormal data via a graphical user interface (GUI).

[0049]

[0060] The manufacturing parameters 150 may include hardware parameters (e.g., information indicating which components are installed in the manufacturing equipment 124, information indicating component replacement, information indicating the elapsed time of components, information indicating software version or update, etc.) and / or processing parameters (e.g., temperature, pressure, flow rate, current, voltage, gas flow, lift speed, etc.). In some embodiments, corrective action includes causing preventive operational maintenance (e.g., replacement, processing, cleaning, etc. of components of the manufacturing equipment 124). In some embodiments, corrective action includes causing design optimization (e.g., updating manufacturing parameters, manufacturing processes, manufacturing equipment 124, etc., for the optimized product). In some embodiments, corrective action includes updating the recipe (e.g., changing the timing of manufacturing subsystems that enter idle mode or active mode, changing the setpoints for various characteristic values, etc.).

[0050]

[0061] The prediction server 112, server machine 170, and server machine 180 may each include one or more computing devices, such as rack-mount servers, router computers, server computers, personal computers, mainframe computers, laptop computers, tablet computers, desktop computers, graphics processing units (GPUs), and application-specific integrated circuits (ASICs) (e.g., Tensor Processing Units (TPUs)). The computing techniques may include edge computing. The operation of the prediction server 112, server machine 170, server machine 180, data store 140, etc., may be performed by cloud computing services, cloud data storage services, etc.

[0051]

[0062] The prediction server 112 may include a prediction component 114. In some embodiments, the prediction component 114 may receive current sensor data 146 and / or current manufacturing parameters (e.g., received from a client device 120 and retrieved from a data store 140) and generate an output (e.g., prediction data 168) for performing corrective actions related to the manufacturing equipment 124 based on the current data. In some embodiments, the prediction data 168 may include one or more predicted dimensional measurements of the processed product. In some embodiments, the prediction component 114 may use one or more trained machine learning models 190 to determine the output for performing corrective actions based on the current data. In some embodiments, the prediction component 114 may utilize one or more physical-based models. In some embodiments, the prediction component 114 may utilize one or more statistical, heuristic, rule-based, or other types of models to generate the prediction data 168.

[0052]

[0063] The manufacturing equipment 124 may be associated with one or more machine cleaning models, e.g., model 190. The machine learning models associated with the manufacturing equipment 124 can perform a number of tasks, including process control, classification, and performance prediction. Model 190 may be trained using data related to the manufacturing equipment 124 or the products processed by the manufacturing equipment 124, e.g., sensor data 142 (e.g., collected by sensor 126), manufacturing parameters 150 (e.g., related to process control of the manufacturing equipment 124), measurement data 160 (e.g., generated by measurement equipment 128), etc. Other types of models may operate similarly. For example, a physical-based model associated with the manufacturing equipment 124 can be calibrated using data related to the manufacturing equipment 124 in a manner similar to training a machine learning model (e.g., updating one or more parameters of the model based on an error term between the model's output and a target output). In some embodiments, calibrating a physical-based model may consume far less data than training a machine learning model.

[0053]

[0064] One type of machine learning model that can be used to perform some or all of the above tasks is an artificial neural network, such as a deep neural network. An artificial neural network generally includes a feature representation component, which includes a classifier or a regression layer that maps features to a desired output space. For example, a convolutional neural network (CNN) has multiple layers of convolutional filters. Pooling is performed in the lower layers to address nonlinearity, and a multilayer perceptron is commonly added above the lower layers, with the top layer mapping the features extracted by the convolutional layers to produce a decision (e.g., a classification output).

[0054]

[0065] A recurrent neural network (RNN) is another type of machine learning model. Recurrent neural network models are designed to interpret a set of inputs that are inherently related to each other, such as time trace data or sequential data. The output of the perceptron in an RNN is fed back into the perceptron as input, generating the next output.

[0055]

[0066] Deep learning is a class of machine learning algorithms that use a cascade of multiple layers of nonlinear processing units for feature extraction and transformation. Each successive layer uses the output from the previous layer as input. Deep neural networks can be trained in supervised (e.g., classification) and / or unsupervised (e.g., pattern analysis) forms. Deep neural networks consist of layers, with different layers learning different levels of representation corresponding to different levels of abstraction. In deep learning, each level learns to transform its input data into a more abstract and complex representation. For example, in an image recognition application, the raw input is a matrix of pixels, the first representation layer can abstract the pixels and encode the edges, the second layer can synthesize and encode the arrangement of the edges, the third layer can encode higher-order levels of shape (e.g., teeth, lips, gums, etc.), and the fourth layer can recognize the role of scanning. Notably, the deep learning process can learn on its own which features are best placed at which levels. The "deep" in "deep learning" refers to the number of layers that transform the data. More precisely, deep learning systems have a very deep CAP (credit assignment path). The CAP is a chain of transformations from input to output. The CAP describes the latent causal relationship between the input and output. In the case of a feedforward neural network, the depth of the CAP can be the depth of the network, which can be the number of hidden layers plus one. For recurrent neural networks where a signal can propagate through a layer multiple times, the depth of the CAP is potentially unlimited.

[0056]

[0067] In some embodiments, the user utilizes a comprehensive modeling platform to provide a predictive component 114 having current sensor data 146, current measurement data, and / or current manufacturing parameters. The predictive component 114 can decompose the current data into a set of current data, provide the set of current data as input to a trained model 190, and perform signal processing to obtain an output from the trained model 190 showing predictive data 168. In some embodiments, the predictive component 114 receives measurement data of the substrate (e.g., predictive measurement data based on sensor data) and provides the measurement data to the trained model 190. For example, the current sensor data 146 may include sensor data showing measurements of the substrate (e.g., shape dimensions) and / or other patterns of features or structure of the substrate. The model 190 may accept data showing the substrate measurements and, as an output, generate a set of process inputs that are predicted to result in a substrate having a target shape.

[0057]

[0068] In some embodiments, the various models described in relation to Model 190 (e.g., supervised machine learning models, unsupervised machine learning models, physically based digital twin models, etc.) may be combined into a single model (e.g., an ensemble model) or they may be separate models.

[0058]

[0069] Data may be passed back and forth between several different models included in Model 190 and the prediction component 114. In some embodiments, some or all of these operations may instead be performed by different devices (e.g., client device 120, server machine 170, server machine 180, etc.). Those skilled in the art will understand that variations in data flow, such as which component performs which process and which model is provided with which data, are within the scope of this disclosure.

[0059]

[0070] The data store 140 may be memory (e.g., random access memory), drives (e.g., hard drives, flash drives), a database system, a cloud-accessible memory system, or another type of component or device capable of storing data. The data store 140 may include multiple storage components (e.g., multiple drives or multiple databases), and such storage components may be located across multiple computing devices (e.g., multiple server computers). The data store 140 may store sensor data 142, manufacturing parameters 150, measurement data 160, and prediction data 168.

[0060]

[0071] Sensor data 142 may include historical sensor data 144 and current sensor data 146. Sensor data may include sensor data time traces over the duration of the manufacturing process, correlations between data and physical sensors, pre-processed data such as averages and composite data, and data showing sensor performance over time (i.e., many manufacturing processes). Manufacturing parameters 150 and measurement data 160 may include similar features, e.g., historical measurement data and current measurement data. Historical sensor data 144, historical measurement data, and historical manufacturing parameters may be historical data (e.g., at least some of this data may be used to train or calibrate model 190). Current sensor data 146, current measurement data may be current data (e.g., at least some of which are input into model 190 following historical data) from which predictive data 168 is generated (e.g., to perform corrective actions).

[0061]

[0072] In some embodiments, the prediction system 110 further includes server machines 170 and 180. Server machine 170 includes a dataset generator 172 that can generate datasets (e.g., sets of data inputs and sets of target outputs) for training, validating, and / or testing one or more models 190, which include one or more machine learning models. Some operations of the dataset generator 172 are described below in detail with respect to Figures 2A-2B and 4A. In some embodiments, the dataset generator 172 can divide historical data (e.g., historical sensor data 144, historical manufacturing parameters, historical measurement data) into a training set (e.g., 60% of the historical data), a validation set (e.g., 20% of the historical data), and a test set (e.g., 20% of the historical data).

[0062]

[0073] In some embodiments, the prediction system 110 generates multiple sets of features (for example, via prediction components 114). For example, a first set of features may correspond to a first set of sensor data (e.g., a first set of values ​​from the first set of sensors, a first pattern of values ​​from the first set of sensors) corresponding to each of the datasets (e.g., a training set, a validation set, and a test set), and a second set of features may correspond to a second set of sensor data (e.g., a second set of sensors different from the first set of sensors, a second set of values ​​different from the first set, a second pattern different from the first pattern) corresponding to each of the datasets.

[0063]

[0074] In some embodiments, Model 190 is provided with historical data as training data (for example, in the case of a machine learning model). In some embodiments, Model 190 is provided with historical data as calibration data (for example, in the case of a physical-based model). The type of data provided depends on the intended use of the model. For example, a machine learning model may be trained by providing historical sensor data 144 as training input and corresponding measurement data 160 as target output. In some embodiments, a large amount of data is used to train Model 190, for example, sensor and measurement data from hundreds of substrates may be used. In some embodiments, a relatively small amount of data may be used to calibrate a physical-based model 190.

[0064]

[0075] The server machine 180 includes a training engine 182, a validation engine 184, a selection engine 185, and / or a test engine 186. The engines (e.g., training engine 182, validation engine 184, selection engine 185, and test engine 186) may refer to hardware (e.g., a network, dedicated logic, programmable logic, microcode, a processor, etc.), software (e.g., a processor, a general-purpose computer system, or instructions executed on a dedicated machine), firmware, microcode, or a combination thereof. The training engine 182 may be capable of training or calibrating a model 190 using one or more sets of features associated with a training set from a dataset generator 172. The training engine 182 may generate multiple trained models 190, where each trained model 190 corresponds to a distinct set of features from the training set (e.g., sensor data from a distinct set of sensors). For example, the first trained model might be trained using all features (e.g., X1-X5), the second trained model might be trained using a first subset of features (e.g., X1, X2, X4), and the third trained model might be trained using a second subset of features that partially overlap with the first subset (e.g., X1, X3, X4, and 5).

[0065]

[0076] The validation engine 184 may be able to validate the trained models 190 using the features of the corresponding set of validation sets from the dataset generator 172. For example, a first trained machine learning model 190, trained using a first set of features from the training set, may be validated using a first set of features from the validation set. The validation engine 184 can determine the accuracy of each trained model 190 based on the corresponding set of features from the validation set. The validation engine 184 may discard trained models 190 that have an accuracy that does not meet a threshold accuracy. In some embodiments, the selection engine 185 may be able to select one or more trained models 190 that have an accuracy that meets a threshold accuracy. In some embodiments, the selection engine 185 may be able to select the trained model 190 with the highest accuracy.

[0066]

[0077] The test engine 186 can test the trained model 190 using the corresponding set of features from the test set generated by the dataset generator 172. For example, a first trained machine learning model 190, trained using a first set of features from the training set, can be tested using a first set of features from the test set. Based on the test set, the test engine 186 can determine which trained model 190 has the highest accuracy among all the trained models.

[0067]

[0078] In the case of a machine learning model, model 190 may refer to a model artifact created by the training engine 182 using a training set containing data inputs and corresponding target outputs (ground truth for each training input). Patterns in the dataset can be found that map data inputs to target outputs (ground truth), and machine learning model 190 is provided with mappings that capture these patterns. Machine learning model 190 may use one or more of the following: support vector machines (SVMs), radial basis functions (RBFs), clustering, supervised machine learning, semi-supervised machine learning, unsupervised machine learning, k-nearest neighbor algorithm (k-NN), linear regression, random forests, and neural networks (e.g., artificial neural networks, regressive neural networks).

[0068]

[0079] The prediction component 114 can provide current data to the model 190 and run the model 190 on the input to obtain one or more outputs. For example, the prediction component 114 can provide current sensor data 146 to the model 190 and run the model 190 on the input to obtain one or more outputs. The prediction component 114 may be able to determine (e.g., extract) prediction data 168 from the output of the model 190. The prediction component 114 may determine (e.g., extract) confidence data from the output that indicates the confidence level of an accurate predictor of the process related to the input data of the current sensor data 146 and / or the product manufactured or produced using the manufacturing equipment 124 with the current manufacturing parameters. The prediction component 114 or the corrective action component 122 may use the confidence data to determine whether to take corrective action related to the manufacturing equipment 124 based on the prediction data 168.

[0069]

[0080] Confidence data may include or indicate confidence that the prediction data 168 is an accurate prediction of a product or component related to at least a portion of the input data. In one embodiment, the confidence level is a real number from 0 to 1, where 0 indicates no certainty that the prediction data 168 is an accurate prediction of a product processed according to the input data or component health of the components of the manufacturing equipment 124, and 1 indicates absolute confidence that the prediction data 168 is an accurate prediction of a product characteristic processed according to the input data or component health of the components of the manufacturing equipment 124. In response to confidence data indicating confidence below a threshold level for a predetermined number of instances (e.g., percentage of instances, frequency of instances, total number of instances), the prediction component 114 may allow the trained model 190 to be retrained (e.g., based on current sensor data 146, current manufacturing parameters, etc.). In some embodiments, retraining may include generating one or more datasets (e.g., via a dataset generator 172) using historical and / or synthetic data.

[0070]

[0081] For illustrative purposes, aspects of this disclosure describe, without limitation, one or more trained machine learning models 190 that determine predictive data 168 using historical data (e.g., historical sensor data 144, historical manufacturing parameters), synthetic data 162, and input current data (e.g., current sensor data 146, current manufacturing parameters, and current measurement data) in one or more trained machine learning models. In other embodiments, heuristic models, physical-based models, or rule-based models are used to determine predictive data 168 (e.g., without using trained machine learning models). In some embodiments, such models may be trained using historical data and / or synthetic data. In some embodiments, these models may be retrained using a combination of true historical data and synthetic data. The predictive component 114 may monitor historical sensor data 144, historical manufacturing parameters, and measurement data 160. Any of the information described with respect to the data inputs 210A to 210B in Figures 2A to 2B may be monitored in or otherwise used in the heuristic models, physical-based models, or rule-based models.

[0071]

[0082] In some embodiments, the functions of client device 120, prediction server 112, server machine 170, and server machine 180 may be provided by fewer machines. For example, in some embodiments, server machines 170 and 180 may be integrated into a single machine, while in some other embodiments, server machine 170, server machine 180, and prediction server 112 may be integrated into a single machine. In some embodiments, client device 120 and prediction server 112 may be integrated into a single machine. In some embodiments, the functions of client device 120, prediction server 112, server machine 170, server machine 180, and data store 140 may be performed by a cloud-based service.

[0072]

[0083] In general, functions described in one embodiment as being performed by the client device 120, prediction server 112, server machine 170, and server machine 180 can, where appropriate, also be performed on the prediction server 112 in other embodiments. Furthermore, functions belonging to a particular component may be performed by different components or by multiple components working together. For example, in some embodiments, the prediction server 112 may determine corrective actions based on prediction data 168. In another example, the client device 120 may determine the prediction data 168 based on the output from a trained machine learning model.

[0073]

[0084] Furthermore, the functionality of a particular component can be performed by different or multiple components working together. One or more of the prediction server 112, server machine 170, or server machine 180 may be accessed as a service provided to other systems or devices via an appropriate application programming interface (API).

[0074]

[0085] In some embodiments, “User” may be represented as a single individual. However, other embodiments of the present disclosure may include “User” as an entity controlled by multiple users and / or automated sources. For example, a collection of individual users formed as a group of administrators may be considered “User”.

[0075]

[0086] Figures 2A and 2B show block diagrams of exemplary dataset generators 272A and 272B (e.g., dataset generator 172 in Figure 1) for creating datasets for training, testing, validating, calibrating, etc., models (e.g., model 190 in Figure 1) according to several embodiments. Each dataset generator 272 may be part of the server machine 170 in Figure 1. In some embodiments, several machine learning models associated with the manufacturing equipment 124 may be trained, used, and maintained (e.g., within the manufacturing facility). Several physical-based models associated with the manufacturing equipment 124 may be calibrated, used, and maintained. Many models associated with the manufacturing equipment 124 may be maintained in relation to and accessed through a comprehensive modeling platform associated with one or more processing chambers, processing tools, manufacturing facilities, etc. Each model may be associated with one dataset generator 272, and multiple models may share dataset generators 272, etc.

[0076]

[0087] Figure 2A shows a system 200A that includes a dataset generator 272A for creating datasets for one or more supervised models (e.g., a database digital twin model, such as model 190 in Figure 1). The dataset generator 272A can use historical data to create datasets (e.g., data input 210A, target output 220A). In some embodiments, an unsupervised machine learning model can be trained using a dataset generator similar to the dataset generator 272A. For example, the target output 220A may not be generated by the dataset generator 272A.

[0077]

[0088] The dataset generator 272A can be used to generate datasets for training machine learning models in association with a comprehensive modeling platform. The comprehensive modeling platform may include many machine learning models. Various machine learning models can predict various chamber conditions, substrate features, changes over time, substrate conditions, etc. For example, various models can predict changes in the shape and dimensions of the substrate over time, the properties of the plasma generated in the etching chamber, the magnetic field values ​​in the processing chamber, the temperature at various locations in the processing chamber and / or the substrate, substrate deformation, etc. Each such model can be trained in a similar manner. Datasets can be generated corresponding to the intended use of each model. As an example, a dataset generator is described for generating a dataset for a model configured to receive input data indicating the state inside the processing chamber and output predicted substrate feature data. Any other machine learning models associated with the processing chamber can be trained in a similar manner by providing corresponding data for training, validation, testing, etc. For example, a model configured to receive input setpoints for a plasma generator and generate output predictions for plasma generation quality may be provided as a training input set of plasma generation parameters and as a target output representation of the associated plasma generation quality, and a model configured to receive input simulation inputs (e.g., virtual processing parameters) and generate output predictions for substrate characteristics may provide simulation inputs as training inputs and obtain substrate characteristics as target outputs, etc. In some embodiments, an unsupervised machine learning model can perform one or more operations that do not depend on the labeled outputs used to construct the model, such as anomaly detection.

[0078]

[0089] The dataset generator 272A can generate datasets for training, testing, and validating a model (e.g., a machine learning model). In some embodiments, the dataset generator 272A can generate a dataset for a machine learning model. The machine learning model is provided with a set of historical sensor data 244A as a data input 210A. The machine learning model is provided with output board feature data 268 as a target output 220A. The machine learning model may be configured to accept sensor data as input data and generate board feature data as output, based on learning performed in response to the provided training data.

[0079]

[0090] In some embodiments, the dataset generator 272A generates a dataset (e.g., a training set, a validation set, a test set) containing one or more data inputs 210A (e.g., a training input, a validation input, a test input). The data inputs 210A may be provided to the training engine 182, the validation engine 184, or the test engine 186. The dataset may be used to train, validate, or test a model (e.g., model 190 in Figure 1).

[0080]

[0091] In some embodiments, data input 210A may include one or more datasets. For example, system 200A may generate a set of sensor data which may include one or more of the following: sensor data from one or more types of sensors, combinations of sensor data from one or more types of sensors, patterns from sensor data from one or more types of sensors, etc. The set of data input 210A may include data describing various aspects of manufacturing, e.g., combinations of measurement data and sensor data, combinations of measurement data and manufacturing parameters, combinations of several measurement data, some manufacturing parameter data and some sensor data, etc. Data input 210A may include virtual data or synthetic data, e.g., data generated by a model, data generated by a digital twin of a processing chamber, etc. For example, a first model may generate virtual sensor data associated with a substrate processing operation, and a second model may receive virtual sensor data as input to generate further outputs. The second model may be trained on a dataset containing virtual sensor data (e.g., data input 210A). Furthermore, target output 220A may include one or more datasets. For example, the first set may be data such as first substrate features, properties, materials, and geometry dimensions, and the second set of target output data may be data such as second substrate features. The target output 220A may include a set of measurement data derived from one or more types of sensors, a combination of measurement data derived from one or more types of sensors, and a set of substrate feature data such as patterns from the measurement data. In some embodiments, different models may be generated based on combinations of input and / or output data sets.

[0081]

[0092] In some embodiments, the dataset generator 272A may generate a first data input corresponding to a first set of historical sensor data 244A for training, validating, or testing a first machine learning model. The dataset generator 272A may generate a second data input corresponding to a second set of historical sensor data 244Z for training, validating, or testing a second machine learning model.

[0082]

[0093] In some embodiments, the dataset generator 272A generates a dataset (e.g., a training set, a validation set, a test set) which may include one or more data inputs 210A (e.g., a training input, a validation input, a test input) and one or more target outputs 220A corresponding to the data inputs 210A. The dataset may further include mapping data that maps the data inputs 210A to the target outputs 220A. The data inputs 210A may also be referred to as “features,” “attributes,” or “information.” In some embodiments, the dataset generator 272A can provide the dataset to the training engine 182, the validation engine 184, or the test engine 186, which will use the dataset to train, validate, or test a machine learning model (e.g., one of the machine learning models included in model 190, an ensemble model 190, etc.).

[0083]

[0094] Figure 2B shows a block diagram of an exemplary dataset generator 272B for creating a dataset for calibrating a physical-based model, according to several embodiments. A physical-based model can perform calculations based on the physics occurring within a processing chamber and predict results based on one or more inputs. A physical-based model may include one or more parameters used in model construction to tune the model's performance. For example, there may be additional physical effects not considered in the model, such as components that change the system's performance compared to the model, thermal or electrical contacts of components, or material defects (which may be within manufacturing thresholds). Calibration may be performed to tune one or more parameters of the physical-based model according to calibration data so that the virtual chamber represented by the physical-based model accurately represents the physical operation of the processing chamber in question.

[0084]

[0095] System 200B, including dataset generator 272B (e.g., dataset generator 172 in Figure 1), creates datasets for one or more physically based models (e.g., a digital twin model, such as model 190 in Figure 1). Dataset generator 272B may create datasets (e.g., data input 210B) using historical data. An exemplary dataset generator 272B is configured to generate datasets for a physically based model configured to take input manufacturing parameter data and output it as output plasma quality data. Similar dataset generators (or similar operations of dataset generator 272B) can be used for physically based models configured to perform various functions (e.g., a physically based model configured to receive input sensor data and generate predictive measurement data, a machine learning model configured to receive input target measurement data (e.g., target microscope images) and output estimation conditions or processing operation recipes that can generate a device matching the input target data). Dataset generator 272B may share features and / or functions with dataset generator 272A. In some embodiments, the physically based model may be generated in multiple sequences, and multiple versions may be assumed, for example, using a different number of terms, parameters, etc. The selection of the physically based model to be used may include considerations such as runtime and accuracy.

[0085]

[0096] The dataset generator 272B can generate datasets for calibrating a physical-based model. The physical-based model is provided with a set of historical sensor data 245A as a data input 210B. Depending on the intended use of the model, the physical-based model may be provided with a target output 220B, for example, output plasma quality data 269 of the model configured to predict plasma generation quality based on sensor data.

[0086]

[0097] A physical-based model may include two or more separate models (for example, a physical-based model can be an ensemble model). The physical-based model may be configured to generate output data indicating the performance of the processing chamber, such as signs of anomalies present in the processing equipment.

[0087]

[0098] In some embodiments, the dataset generator 272B generates a dataset (e.g., a calibration set) containing one or more data inputs 210B. The data inputs 210B may also be referred to as “features,” “attributes,” or “information.” In some embodiments, the dataset generator 272B may provide the dataset to a calibration engine configured to calibrate a physically based model (e.g., Model 190 in Figure 1). Several embodiments for generating a training set are described further with reference to Figure 4A.

[0088]

[0099] In some embodiments, the dataset generator 272B can generate a first data input corresponding to a first set 245A of historical sensor data to calibrate a first physical-based model, and the dataset generator 272A can generate a second data input corresponding to a second set 245Z of historical sensor data to calibrate a second physical-based model.

[0089]

[0100] Data input 210B for calibrating the physical-based model may include information for a specific manufacturing chamber (e.g., for a particular substrate manufacturing machine). In some embodiments, data input 210B may include information for a specific type of manufacturing machine, e.g., manufacturing machine that shares certain characteristics. Data input 210B may include data related to a specific type of device (e.g., intended function, design, manufactured according to a specific recipe). By calibrating the physical-based model based on types such as machine, device, and recipe, the trained model may be able to generate plausible predictive data in several settings (e.g., for several different facilities, products, etc.).

[0090]

[0101] In some embodiments, a dataset can be generated, and a machine learning model can be trained, validated, or tested using the dataset, followed by further training, validation, testing, or tuning of the model (e.g., tuning weights or parameters related to the model's input data, such as connection weights in a neural network).

[0091]

[0102] Figure 3 is a block diagram showing a system 300 for generating output data (e.g., virtual sensor data 167 and / or virtual substrate data 169 in Figure 1) according to several embodiments. In some embodiments, the system 300 may be used in conjunction with a comprehensive modeling platform. In some embodiments, the system 300 may be used in conjunction with one or more machine learning models, including a comprehensive modeling platform. In some embodiments, the system 300 may be used in conjunction with one or more physical-based models, including a comprehensive modeling platform. The system 300 may be used in conjunction with a model that generates predictive data based on sensor data of performed substrate processing operations. The system 300 may be used in conjunction with a model that generates predictive data based on virtual substrate processing operations. The system 300 may be used in conjunction with these models and other types of models to generate data related to the target processing chamber of the manufacturing equipment 124 in Figure 1. In some embodiments, the system 300 may be used in conjunction with a model for determining corrective actions related to the manufacturing equipment. In some embodiments, the system 300 may be used in conjunction with a model for determining failures of the manufacturing equipment. In some embodiments, system 300 may be used in conjunction with a model to cluster or classify substrates. System 300 may also be used in conjunction with a model having different functions than those listed, associated with a manufacturing system.

[0092]

[0103] In block 310, system 300 (for example, a component of the prediction system 110 in Figure 1) can perform data partitioning (for example, via the dataset generator 172 of the server machine 170 in Figure 1) of data used for training, validating, and / or testing machine learning models. In block 310, system 300 can perform data partitioning of data used for calibrating physically based models. In block 310, system 300 can perform data partitioning of data used when configuring other types of models. Configuring data 364 may include any data used when configuring (e.g., training, calibration, etc.) models used by a comprehensive modeling platform. In some embodiments, configuring data 364 includes historical data such as historical measurement data, historical design data, historical classification data (e.g., classification of whether a product meets a performance threshold), and historical microscopic image data. Configuring data 364 can be partitioned in block 310 to generate training set 302, validation set 304, and test set 306. For example, the training set may be 60% of the training data, the validation set may be 20% of the training data, and the test set may be 20% of the training data. In some embodiments, fewer types of sets may be generated. For example, model validation may not be performed for some models, and mode testing may not be performed for some models. The training set 302 may be used as a calibration set for a physically based model or another type of model configuration set.

[0093]

[0104] The generation of the training set 302, validation set 304, and test set 306 can be tailored for specific applications. For example, the training set may consist of 60% of the training data, the validation set may consist of 20% of the training data, and the test set may consist of 20% of the training data. The system 300 can generate multiple sets of features for each of the training set, validation set, and test set. For example, if the configuration data 364 includes sensor data containing features derived from sensor data from 20 sensors (e.g., sensor 126 in Figure 1), and sensor data containing 10 manufacturing parameters (e.g., manufacturing parameters corresponding to the same processing execution as the sensor data from the 20 sensors), the sensor data may be split into a second set of feature sets containing 0 for sensors 1-10 and 1-20. The manufacturing parameters may also be split into sets such as a first set of manufacturing parameters containing parameters 1-5, and a second set of manufacturing parameters containing parameters 6-10. Target inputs, target outputs, both, or both can be split into sets. Multiple models can be trained on different datasets.

[0094]

[0105] In block 312, system 300 performs model training (e.g., via training engine 182 in Figure 1) using training set 302. Training of machine learning models and / or physically based models (e.g., digital twins) can be achieved in a supervised learning manner. A supervised learning manner involves providing a training dataset containing inputs labeled through the model, observing its output, defining the error (by measuring the difference between the output and the label value), and using techniques such as deep gradient direction and backpropagation to adjust the model's weights so that the error is minimized. In many applications, repeating this process across many labeled inputs in the training dataset yields a model that can produce the correct output when presented with inputs different from those present in the training dataset. In some embodiments, training of a machine learning model can be achieved in an unsupervised manner. For example, no labels or classifications may be provided during training. An unsupervised model may be configured to perform anomaly detection, result clustering, etc.

[0095]

[0106] For each training data item in the training dataset, the training data item can be input into a model (e.g., a machine learning model). The model can then process the input training data item (e.g., data values ​​provided by sensors in a manufacturing system) to generate an output. The output may include, for example, predicted substrate features. The output can be compared to the labels of the training data items (e.g., measured features of the substrate corresponding to input data from a measuring instrument).

[0096]

[0107] The processing logic may then compare the generated output (e.g., predicted substrate features) with the labels included in the training data items (e.g., measured substrate features). The processing logic determines the error (i.e., classification error) based on the difference between the output and the labels. The processing logic adjusts one or more weights and / or values ​​of the model based on the error.

[0097]

[0108] When training a neural network, an error term or delta may be determined for each node in the artificial neural network. Based on the above error, the artificial neural network adjusts one or more of its own parameters (weights for one or more inputs of a node) for one or more of its own nodes. Parameters may be updated in a backpropagation manner, such that the top layer nodes are updated first, followed by the nodes of the next layer, and so on. An artificial neural network contains multiple layers of "neurons," each layer receiving values ​​as input from the neurons of the previous layer. The parameters for each neuron include weights associated with the values ​​received from each neuron of the previous layer. Therefore, adjusting parameters may involve adjusting the weights assigned to each input for one or more neurons in one or more layers of the artificial neural network.

[0098]

[0109] When training a physically based model, one or more floating parameters may be included in the model and adjusted to fit calibration data. Floating parameters may be additional terms (e.g., higher-order terms in one or more equations describing the physics of the processing chamber) or may involve adjustments to core terms related to known physics. A physically based model may include terms and / or parameters related to any physical process, including material properties, thermal properties, electrical properties, and fluid dynamic properties related to the processing chamber. Adjustments to floating parameters may be made so that the error term between the output generated by the physically based model and the expected output based on the configuration data 364 is reduced or minimized. Determining the number of terms to include in the physically based model may involve considerations such as accuracy, execution time, and overfitting (e.g., the number of floating parameters compared to the amount of available calibration data).

[0099]

[0110] System 300 can train multiple models using multiple sets of features from the training set 302 (e.g., a first set of features from the training set 302, a second set of features from the training set 302, etc.). For example, System 300 can generate a first trained model using a first set of features in the training set (e.g., sensor data from sensors 1-10, measured values ​​1-10, etc.) and generate a second trained model using a second set of features in the training set (e.g., sensor data from sensors 11-20, measured values ​​11-20, etc.). In some embodiments, the first and second trained models can be combined to generate a third trained model (e.g., a predictor or synthetic data generator that is better than the first or second trained model itself). In some embodiments, the sets of features used for comparing the models may overlap (e.g., a first set of features that are sensor data from sensors 1-15 and a second set of features that are sensors 5-20). In some embodiments, hundreds of models can be generated, including models with various feature substitutions and model combinations.

[0100]

[0111] In block 314, system 300 performs model validation (e.g., via validation engine 184 in Figure 1) using validation set 304. System 300 can validate each of the trained models using the corresponding set of features in validation set 304. For example, system 300 may validate a first trained model using a first set of features in the validation set (e.g., sensor data from sensors 1-10 or measured values ​​1-10) and a second trained model using a second set of features in the validation set (e.g., sensor data from sensors 11-20 or measured values ​​11-20). In some embodiments, system 300 can validate hundreds of models generated in block 312 (e.g., models with various permutations of features, combinations of models, etc.). In block 314, system 300 can determine the accuracy of each of one or more trained models (e.g., via model validation) and determine whether one or more of the trained models have an accuracy that meets a threshold accuracy. In response to the determination that none of the trained models have the precision to meet the threshold precision, the flow returns to block 312, where system 300 trains models using different sets of features from the training set. In response to the determination that one or more of the trained models have the precision to meet the threshold precision, the flow continues to block 316. System 300 may discard trained models that have precision below the threshold precision (for example, based on the validation set).

[0101]

[0112] In block 316, system 300 performs model selection (e.g., via selection engine 185 in Figure 1) to determine which of the one or more trained models that satisfy the threshold precision has the highest precision (e.g., the selected model 308 based on the validation in block 314). In response to the determination that two or more trained models that satisfy the threshold precision have the same precision, the flow can return to block 312, where system 300 performs model training using a further refined training set corresponding to a further refined set of features to determine the trained model with the highest precision.

[0102]

[0113] In block 318, system 300 performs a model test (e.g., via the test engine 186 in Figure 1) using test set 306 to test the selected model 308. System 300 can test a first trained model using a first set of features in the test set (e.g., sensor data from sensors 1-10) to determine if the first trained model meets the threshold accuracy. Determining whether the first trained model meets the threshold accuracy can be obtained based on the first set of features in test set 306. In response that the accuracy of the selected model 308 does not meet the threshold accuracy, the flow proceeds to block 312, where system 300 performs model training (e.g., retraining) using a different training set corresponding to a different set of features. If the selected model 308 is overfitted to training set 302 and / or validation set 304, the accuracy of the selected model 308 may not meet the threshold accuracy. The accuracy of the selected model 308 may not meet the threshold accuracy if the selected model 308 is not applicable to other datasets, including the test set 306. Training with different features may include training with data from different sensors, different manufacturing parameters, etc. In response to the determination that the selected model 308 has accuracy that meets the threshold accuracy based on the test set 306, the flow proceeds to block 320. In at least block 312, the model can learn patterns in the training data in order to make predictions. In block 318, the system 300 can apply the model to the remaining data (e.g., the test set 306) to test the predictions.

[0103]

[0114] In block 320, system 300 receives current data 322 and uses a trained model (e.g., selected model 308) to determine (e.g., extract) predictive data 324 from the output of the trained model. Current data 322 may be a process, operation, or manufacturing parameter related to an operation in question. Current data 322 may be a manufacturing parameter related to a process such as development, redevelopment, or investigation. Current data 322 may be a manufacturing parameter related to a gas transport system. Current data 322 may be a manufacturing parameter that may influence the delay of a change in condition values ​​compared to the initiation of an operation that changes the condition. Current data 322 may be a manufacturing parameter related to gas supply and / or gas removal related to a substrate processing chamber. Corrective actions related to the manufacturing equipment 124 in Figure 1 may be performed taking predictive data 324 into consideration. In some embodiments, current data 322 may correspond to the same type of feature in the historical data used to train the model. In some embodiments, current data 322 corresponds to a subset of the types of features in the historical data used to train the selected model 308. For example, a machine learning model may be trained using several manufacturing parameters and configured to produce an output based on a subset of those parameters.

[0104]

[0115] In some embodiments, the performance of a model trained, validated, and tested by System 300 may degrade. For example, the manufacturing system associated with the trained machine learning model may undergo gradual or abrupt changes. Changes in the manufacturing system may further degrade the performance of the trained machine learning model. A new model may be generated to replace the degraded machine learning model. The new model may be generated by modifying the old model through retraining. For example, the performance of the model may be degraded by changes in the coating on the inner wall of the processing chamber, degradation of one or more components of the chamber (e.g., quartz or silicon carbide components), aging or failure of one or more components of the chamber (e.g., aging of lamps, arc discharge of heaters, etc.), or other changes that affect the performance of the chamber.

[0105]

[0116] Generating a new model may include providing additional configuration data 346. Generating a new model may further include providing current data 322, for example, data used by the model to make predictions. In some embodiments, the current data 322 provided for generating a new model may be marked with an indication of the accuracy of the predictions generated by the model based on the current data 322. Additional training data 346 may be provided in block 312 for model training, such as generating one or more new models, updating, retraining, and / or improving a selected model 308.

[0106]

[0117] The generation of new models (e.g., updating the parameters of a trained model) can be performed on any model associated with a manufacturing chamber, such as any digital twin model that replicates the characteristics of the manufacturing chamber. New model generation can be performed on physically based digital twin models, database (e.g., machine learning) digital twin models, hybrid models (e.g., including database and physically based features), and so on. New model generation can also be performed on models that receive chamber sensor data, models that receive manufacturing parameters, and models that receive simulation inputs (e.g., for generating a virtual substrate).

[0107]

[0118] In some embodiments, one or more of operations 310-320 may occur in various orders and / or in conjunction with other operations not presented and described herein. In some embodiments, one or more of operations 310-320 may not be performed. For example, in some embodiments, one or more of the data partitioning of block 310, model verification of block 314, model selection of block 316, or model testing of block 318 may not be performed.

[0108]

[0119] Figure 3 shows a system that can be used to configure, train, validate, test, and use one or more machine learning models. The machine learning models are configured to accept data as input (e.g., setpoints provided for manufacturing equipment, sensor data, measurement data, etc.) and to provide data as output (e.g., predictive data, corrective action data, classification data, etc.). Block splitting, training, validation, selection, testing, and use of the system 300 can be performed using different types of data, similar to training a second model. Retraining can also be performed using the current data 322 and / or additional training data 346. The system 300 can further be used to configure and calibrate one or more physical-based models, statistical models, rule-based models, heuristic models, etc.

[0109]

[0120] Figures 4A to 4D are flowcharts of methods 400A to 400D relating to the training and use of a comprehensive modeling platform according to specific embodiments. Methods 400A to 400D may be executed by processing logic which may include hardware (e.g., circuits, dedicated logic, programmable logic, microcode, processing units, etc.), software (e.g., processing units, general-purpose computer systems, or instructions executed on dedicated machines), firmware, microcode, or a combination thereof. In some embodiments, methods 400A to 400D may be executed in part by a prediction system 110. Method 400A may be executed in part by a prediction system 110 (e.g., server machine 170 and dataset generator 172 in Figure 1, dataset generators 272A to B in Figures 2A to B). The prediction system 110 may use method 400A to generate datasets for at least one of training, validating, or testing a machine learning model according to embodiments of this disclosure. The prediction system 110 may use method 400A to generate datasets for calibrating a physically based model. The prediction system 110 can use method 400A to generate a dataset for constructing another type of model to be used by a comprehensive modeling platform associated with a processing chamber. Methods 400B to 400D may be executed by the prediction server 112 (e.g., prediction component 114), client device 120, and / or server machine 180 (e.g., training, validation, and test operations may be executed by server machine 180). In some embodiments, a non-temporary machine-readable storage medium stores instructions, which, when executed by a processing device (e.g., the prediction system 110, server machine 180, prediction server 112, etc.), cause the processing device to execute one or more of methods 400A to 400C.

[0110]

[0121] For the sake of simplicity, methods 400A–400D are illustrated and described as a series of operations. However, the operations according to this disclosure may be performed in various orders and / or simultaneously with other operations not presented and described herein. Furthermore, not all illustrated operations are performed in order to carry out methods 400A–400D according to the disclosed subject matter. In addition, those skilled in the art will understand and recognize that methods 400A–400D may alternatively be represented as a series of interrelated states via a state diagram or events.

[0111]

[0122] Figure 4A is a flowchart of Method 400A for generating a model dataset in several embodiments. The dataset may be used for training, validating, or testing the model. The dataset may be generated for a machine learning model. The dataset may be generated to calibrate a physically based model. The dataset may be for generating a digital twin model of a processing chamber. The dataset may be used to construct another type of model in relation to a processing chamber. Referring to Figure 4A, in some embodiments, in block 401, the processing logic implementing Method 400A initializes the training set T to an empty set.

[0112]

[0123] In block 402, the processing logic generates a first data input (e.g., a first training input, a first validation input) which may include one or more of the following: sensors, manufacturing parameters, measurement data, etc. In some embodiments, the first data input may include a first set of features for the data type, and the second data input may include a second set of features for the data type (e.g., as described with respect to Figure 3). The input data may include historical data.

[0113]

[0124] In some embodiments, in block 403, the processing logic optionally generates a first target output for one or more of the data inputs (e.g., a first data input). In some embodiments, the inputs include one or more sensor data points, and the target output includes substrate feature data. In some embodiments, the first target output is predictive data. In some embodiments, the input data may be in the form of sensor data, and the target output may be a list of components that are likely to be defective, as in the case of a machine learning model configured to identify a defective manufacturing system. In some embodiments, no target output is generated (e.g., an unsupervised machine learning model that can group or find correlations in the input data, and does not need to provide a target output).

[0114]

[0125] In block 404, the processing logic optionally generates mapping data indicating input / output mappings. The input / output mapping (or mapping data) may refer to data inputs (e.g., one or more of the data inputs described herein), target outputs for the data inputs, and associations between one or more data inputs and target outputs. In some embodiments, such as those relating to machine learning models for which no target outputs are provided, block 404 may not be executed.

[0115]

[0126] In block 405, the processing logic adds the mapping data generated in block 404 to the dataset T, in some embodiments.

[0116]

[0127] In block 406, the processing logic branches based on whether the dataset T is sufficient for at least one of the following: training, validating, and / or testing a machine learning model, calibrating a physically based model, setting up another type of model, etc. If it is sufficient, the process execution proceeds to block 407; otherwise, it returns to block 402 and the process execution continues. Note that in some embodiments, the sufficiency of dataset T may be determined simply based on the number of inputs, in some embodiments it may be mapped to the outputs in the dataset, while in some other embodiments the sufficiency of dataset T may be determined based on one or more other criteria (e.g., a measure of the diversity of data examples, precision, etc.) in addition to the number of inputs.

[0117]

[0128] In block 407, the processing logic provides a dataset T (for example, to the server machine 180) to train, validate, calibrate, and / or test the model 190. The model 190 may be a machine learning model. The model 190 may be a physically based model. In some embodiments, dataset T is a training set and is provided to the training engine 182 of the server machine 180 to perform training. In some embodiments, dataset T is a validation set and is provided to the validation engine 184 of the server machine 180 to perform validation. In some embodiments, dataset T is a test set and is provided to the test engine 186 of the server machine 180 to perform testing. For a neural network, for example, input values ​​of a given input / output mapping (e.g., numerical values ​​associated with data input 210A) are input to the neural network, and output values ​​of the input / output mapping (e.g., numerical values ​​associated with target output 220A) are stored in the output nodes of the neural network. Next, the connection weights within the neural network are adjusted according to a learning algorithm (e.g., backpropagation), and this procedure is repeated for other input / output mappings in the dataset T. After block 407, the model (e.g., model 190) may be trained using the training engine 182 of server machine 180, validated using the validation engine 184 of server machine 180, or tested using the test engine 186 of server machine 180. The trained model may be implemented by a prediction component 114 (of the prediction server 112) to generate prediction data 168 for performing signal processing, to generate synthetic data 162, or to perform corrective actions related to manufacturing equipment 124.

[0118]

[0129] Figure 4B is a flowchart of Method 400B for utilizing a comprehensive modeling platform for data analysis related to a processing chamber, according to several embodiments. In block 410, the processing logic receives a first user selection via a GUI to display data related to a first processing chamber in a first chamber data mode (e.g., modeling data mode). Alternatively, the processing logic may receive a prompt (e.g., a text or voice prompt) indicating the data to display for the first processing chamber in the first chamber data mode. The prompt may be processed using an LLM, which may output instructions for making a selection to display data related to the first processing chamber. The GUI and / or LLM may run as part of an application (e.g., a comprehensive modeling platform). The first chamber data mode contains data on processing operations performed within the first processing chamber. In some embodiments, inputs may be received in other ways, via different types of user interfaces, including, for example, a text interface, a virtual reality interface, an augmented reality interface, etc.

[0119]

[0130] The data that can be viewed may be the output data of one or more models associated with the first processing chamber. The displayed data may be the output data of one or more calibrated physical-based models, trained machine learning models, or other types of models. The output data may be from one or more digital twin models in the first processing chamber.

[0120]

[0131] Generating data may include providing input to a model and receiving output from a model. The first chamber data mode may include providing data to one or more models related to processing operations performed within a first processing chamber. Input to the model may include sensor data from the first processing chamber. Sensor data may include sensors that enable the precise operation of one or more components of the processing chamber, such as sensors that provide feedback on actuator position, power supplied to components, etc. Sensor data may include sensors that report conditions within the processing chamber, such as temperature sensors and pressure sensors.

[0121]

[0132] The provision of data to a model, the execution of a model, etc., can be performed in response to user requests provided through an application (e.g., a comprehensive modeling platform). The user can select a processing recipe or processing operation to be executed within a first processing chamber, and the selection of one or more models may be performed at the time of user selection. In some embodiments, the user can select data from a specific category, and the associated models may be executed at the time of user selection. In some embodiments, the user can select a specific model for execution. The user can select a group of models to be executed, for example, to generate a selection of data for use in a relevant analysis application.

[0122]

[0133] In block 412, the processing logic provides first display data of a first chamber data mode (e.g., modeling data mode) in response to receiving a first user selection for display in the GUI. Displaying the first data may include generating one or more tables, graphics, or other visualizations of the data. Displaying the first display data may include generating graphs, plots, charts, etc., of the first display data. The first display data may be virtual sensor data based on the output of one or more models, one or more digital twins, etc. The first display data may be virtual sensor data based on the output of one or more database models, one or more machine learning models, etc.

[0123]

[0134] In block 414, the processing logic receives a second user selection to display data related to the first processing chamber in a second chamber data mode. The user selection may be received via a GUI. Alternatively, the processing logic may receive a prompt (e.g., a text or voice prompt) indicating data to display regarding the first processing chamber in the second chamber data mode. The prompt may be processed using an LLM, which may output instructions for making a selection to display data related to the first processing chamber. The second chamber data mode includes data for virtual processing operations performed by a virtual representation of the first processing chamber. The second chamber data mode may include data generated by a digital twin model of the processing chamber. The second chamber data mode may include data associated with the virtual processing chamber, virtual processing operations, virtual board, etc.

[0124]

[0135] The second chamber data mode may include generating virtual board data (e.g., data related to the virtual board generation process). Virtual board data may be generated by providing input data to one or more models (e.g., a physically based model, a machine learning model, etc.). Virtual board data may also be generated by providing simulation inputs to one or more models. Inputs may be provided to one or more models when the user selects the second chamber data mode, selects a processing chamber for data analysis, or selects one or more data analysis or generation techniques. The second display data may indicate, for example, the quality of performance of a virtual processing operation, regardless of whether the virtual processing operation meets one or more threshold performance conditions.

[0125]

[0136] In block 416, the processing logic provides second display data in a second chamber data mode in response to receiving a second user selection for display on the GUI. Displaying the second display data may share one or more characteristics with the operation described in relation to the display of the first display data.

[0126]

[0137] Figure 4C is a flowchart of Method 400C for utilizing a model using a comprehensive modeling platform, in several embodiments. In block 420, the processing logic receives a user selection of a first processing chamber or other input via the GUI of the modeling platform. Alternatively, the processing logic may receive a prompt (e.g., a text or voice prompt) indicating the first processing chamber. The prompt may be processed using an LLM, which may output an instruction to make a selection of the first processing chamber.

[0127]

[0138] In block 422, the processing logic provides first input data to a first model in response to user selection or reception of other inputs to the first processing chamber. The input data includes sensor data from the first processing chamber. In some embodiments, the input data may include manufacturing parameters associated with the first processing chamber, such as recipe setpoints. In some embodiments, the input data may include design features, such as target characteristics of the substrate. In some embodiments, providing the first input data may be performed based on further user selections. For example, providing the first input data may be performed based on user selections such as manufacturing facility, manufacturing tool, data mode (e.g., monitor mode or exploration mode), target data type, target model, or model category. The first model may be a trained machine learning model, a calibrated physics-based model, or another type of model. In some embodiments, the data provided to the model may include data output by another model, such as providing virtual sensor data based on chamber sensor data as input to the first model. For example, the output from a plasma prediction model may generate an output including ion density, and the ion density prediction may be provided to a feature model to generate a prediction of a pattern generated on the substrate.

[0128]

[0139] In block 424, the processing logic obtains first output data from the first model. The first output data includes predictive data related to sensor data from the first processing chamber. The first output data may include virtual sensor data. The first output data may provide more detailed information than that available from the physical sensors of the first processing chamber. The first output data may include substrate feature data, processing conditions, substrate conditions, etc.

[0129]

[0140] In block 426, the processing logic provides second input data for the second model. The second input data is provided in response to receiving user selection or other inputs of the first processing chamber, as described in relation to block 422, and may be further provided based on additional user selections. The second input data includes simulation inputs of virtual board processing operations. The simulation inputs may correspond to manufacturing parameters. The simulation inputs may correspond to processing recipe parameters. The second model may be a digital twin of the first processing chamber.

[0130]

[0141] In block 428, the processing logic obtains second output data from the second model. The second output data includes predicted data for the virtual board processing operation. The second data may include virtual board feature data. The second data may include processing conditions for the virtual processing conditions. The second data may include condition data for the virtual board being processed.

[0131]

[0142] In block 430, the processing logic provides first and second output data for display via the GUI of the modeling platform.

[0132]

[0143] In block 432, the processing logic optionally receives a second user input regarding the selection of a second processing chamber (e.g., user selection via interaction with a GUI, or voice or text prompt). Various actions may be performed in relation to the second user selection. A first trained machine learning model (e.g., configured to generate predictions of the same characteristics of a substrate processing operation) and a third trained machine learning model corresponding to the second processing chamber may be provided as input data. A second trained machine learning model and a fourth trained machine learning model corresponding to the second processing chamber may be provided as input data. Input data may be provided for additional models. Additional models may be part of a number of data modeling modes (e.g., exploration mode and monitoring mode). Additional models may be machine learning models, physical-based models, or other types of models. Additional models may be, or include, a digital twin model of the second processing chamber. Additional outputs associated with the second processing chamber may be provided for display by the GUI. Data displayed by the GUI may indicate the quality of performance of the associated processing operations (e.g., physical substrate processing operations, virtual substrate processing operations).

[0133]

[0144] Figure 4D is a flowchart of Method 400D for operating a comprehensive modeling platform according to several embodiments. In block 440, the processing logic outputs a view of the manufacturing system, for example, via the display of a computing device. The manufacturing system includes a plurality of processing chambers, one or more transfer chambers, and one or more load locks. The view includes a graphical representation of each of the plurality of processing chambers and one or more view options for data associated with the processing chambers. The view may further include labeling of the processing chambers, which may include a name, identification information, chamber type, etc. Chamber types may include physical vapor deposition (PVD) chambers, chemical vapor deposition (CVD) chambers, atomic layer deposition (ALD) chambers, etching chambers, lithography chambers, annealing chambers, etc.

[0134]

[0145] In block 442, the processing logic receives inputs related to the selection of a processing chamber view option from one or more view options among a plurality of processing chambers. The view options include, for example, monitors for each of the processing chambers of the manufacturing system, and can explore data view modes.

[0135]

[0146] In block 444, the processing logic loads one or more virtual models associated with the selected processing chamber and selected view options. Each processing chamber is associated with one or more virtual models configured for that processing chamber. The models may include database models (e.g., trained machine learning models). Database models may be trained based on training data from the associated processing chamber. The models may include physical-based models. Physical-based models may be calibrated based on calibration data from the associated processing chamber.

[0136]

[0147] In block 446, the processing logic optionally receives inputs related to the selection of processing operations to be performed within the selected processing chamber and provides sensor data related to the processing operations to one or more virtual models.

[0137]

[0148] In block 448, the processing logic optionally provides input data to one or more virtual models. The input data may be associated with a processed substrate (e.g., a substrate processed in a selected processing chamber). The processing logic unit may further obtain output data from one or more virtual models. The output data may include predictions of the properties of the processed substrate. The output data may include predictions of the thickness of one or more layers of the processed substrate, e.g., the thickness of the material deposited on the substrate processed in a target processing operation. The output data may include predictions of the thickness over a portion of the substrate. The output data may include predictions of one or more properties of one or more layers of the substrate in one or more cross-sectional slices of the substrate. The output data may include one or more cross-sectional profiles of the substrate. The output data may include predictions of properties (e.g., film thickness, film profile, etc.) of the bottom region of a substrate trench, the sidewalls of a substrate trench, or the field adjacent to the substrate trench. In some embodiments, multiple predictions may be made in response to various time periods during the processing operation (e.g., film deposition). For example, predictions of film growth through deposition processes can be generated.

[0138]

[0149] In block 450, the processing logic optionally receives a first set of simulation inputs. The processing logic further provides the first set of simulation inputs to one or more virtual models. The processing logic further obtains outputs from one or more virtual models that represent virtual board processing operations.

[0139]

[0150] In block 452, the processing logic optionally displays one or more output data, for example, via a GUI. Optionally, the displayed data may include characteristic data of a virtual board or characteristic values ​​of a virtual processing chamber. Optionally, the displayed data may include predicted characteristic data of a physical board or predicted characteristic values ​​of a virtual processing chamber.

[0140]

[0151] Figure 5A is an overview of GUI 500A for selecting a processing chamber for a comprehensive modeling platform, according to several embodiments. GUI 500A includes a graphical representation of tools 502. In some embodiments, the operation of GUI 500A may be performed in list format, icon format, or other format. Tools 502 may have been previously selected, for example, through a previous view of the GUI. Tools 502 may extend to a group of tools pre-selected by the user (e.g., manufacturing facilities).

[0141]

[0152] The tool 502 includes a processing chamber 504. The processing chamber 504 can be of various types, models, and configurations. The processing chambers 504 can be of the same type, model, or configuration. The tool 502 further includes a load lock chamber 506. In some embodiments, some chambers of the tool 502 may have associated models, while other chambers of the tool 502 may not have associated models. For example, each processing chamber 504 may have one or more associated models, while the load lock chamber 506 may not have associated models.

[0142]

[0153] The chamber may include an associated selection user interface (UI) element 508. The selection element 508 allows the user to further select data for the chamber to display, a model of the chamber to perform, model data for the chamber to display, and so on. In some embodiments, the selection element 508 may provide the user with one or more options. For example, a first option may include a model based on data provided by sensors on a physical processing tool corresponding to a digital representation tool 502 (e.g., a monitor option). A second option may include a model based on a digital representation of the processing chamber (e.g., a search option).

[0143]

[0154] One or more chambers may include options for accessing model categories. The various selection elements 508 may include the same or different options according to available models, user needs or preferences, chamber type, category, or quality, or other metrics.

[0144]

[0155] In some embodiments, the operation of a first user selection or other input for displaying chamber data includes data derived from a model that utilizes chamber sensor data as input. For example, chamber sensor data may be provided to a model, and one or more virtual sensors may report on conditions of interest to the user that are not monitored by physical sensors. The model may be a physical-based, digital twin model, database model, machine learning model, statistical model, or another type of model. The model may predict the state of the processing chamber. The model may predict substrate measurements. The model may predict changes in state over time. For example, the model may predict changes in substrate measurement indicators during processing. The model may predict conditions at various time intervals during processing. The model can be further divided into groups such as models based on various physical phenomena (e.g., magnetic field models, heat conduction models, gas conduction models, etc.), models targeting predictions of various attributes (e.g., substrate feature modeling, processing chamber condition modeling, etc.), or various other groups. Groups of models may be presented through a different view of the GUI when selected by the user in a chamber or data view mode (e.g., monitor or explore mode). In some embodiments, in addition to model data based on chamber sensor data, sensor data and / or measurement data may also be displayed.

[0145]

[0156] A second user-selected action for viewing chamber data may include data derived from a model that utilizes virtual processing operations. Such a model may digitally mimic the operation of the corresponding physical chamber. For example, process inputs may be provided to the model, and the model may predict the results of providing such process inputs to the chamber. In some embodiments, the model may be executed at the user's request. For example, a machine learning model may be configured to receive simulation inputs (which may or may not reflect process inputs) and output measurement predictions. The model may be executed at the user's request based on user-selected simulation inputs.

[0146]

[0157] A model based on virtual processing behavior may be a digital twin model of the associated chamber (for example, the model may mimic the operation of the physical chamber). Using a digital twin of the chamber can enable users to quickly adjust recipes, run numerous trial recipes, and experiment with the impact of recipe parameters on output. Performing the same operation in a physical chamber can be very costly in terms of time, materials, waste of defective products, and the time and cost of measuring the product. Furthermore, a virtual chamber can report on parameters that cannot be determined within the physical chamber, such as processing conditions in areas inaccessible to sensors, processing conditions and / or substrate characteristics where sensors are unavailable, time-trace data of conditions, substrate features, or other characteristics that may be inconvenient or impossible to measure. Models based on virtual substrate processing behavior may include physical-based models, database models, machine learning models, and statistical models.

[0147]

[0158] In some embodiments, a set of virtual processing operation models may operate in various modes (e.g., various modeling data modes). For example, a first mode may be a prediction mode. The prediction mode may receive inputs related to process inputs as input. The inputs may be the same as the inputs or related quantities provided to the physical chamber. The prediction mode can then use one or more models to generate outputs that indicate the performance quality of the processing chamber. The outputs may include predictions of processing conditions, predictions of substrate measurements, predictions of substrate evolution during processing, etc. A second mode may be an exploration mode or a recipe design mode. The second mode may receive a set of chamber performance quality metrics as input. For example, the second mode may receive one or more target process conditions, one or more target substrate measurement metrics, etc. as input. A model running in exploration mode may provide simulation and / or process inputs that are predicted to satisfy specific input chamber performance quality criteria.

[0148]

[0159] In some embodiments, the two chambers 504 of tool 502 may be of the same type, category, model, design, etc. The two chambers 504 may be nominally identical (e.g., similarly maintained, or have nominally identical components installed) or may differ by one or more metrics. In some embodiments, separate models may be maintained for the two chambers 504. Separate models may allow for the consideration of differences between nominally identical chambers (e.g., even within manufacturing tolerances). Separate models may allow for the updating of various model parameters based on maintenance performed, measured chamber performance, updates of one or more components of the chambers, and chamber locations (e.g., which port of tool 502 the chamber is connected to, or which facility the chamber is located in). In some embodiments, hardware components may be tracked and / or traced based on hardware configuration documentation, hardware configuration information, etc. In some embodiments, the model may be automatically updated and / or verified for applicability by a computing device based on hardware documentation.

[0149]

[0160] In some embodiments, a single chamber may utilize a first set of components for some operations and different components for other operations. For example, a chemical vapor deposition chamber may use a first type of substrate holder for some substrates when used by certain applications, certain operations, certain users, etc., and a second type of substrate holder for other applications. The model associated with the chamber may be selected to take into account the target configuration of the installed components. The user can select installed components and / or target components. The user can select one or more models or groups of models that target the installed components or target components. Processing logic may determine the appropriate one or more models based on the components installed in the chamber, based on hardware documentation. A virtual or physical chamber may contain any number of tunable components, and a comprehensive modeling platform may be available to provide models that correspond to various types of components, various combinations of components, etc.

[0150]

[0161] The data can be displayed to the user via a GUI in any convenient way. For example, model data can be displayed to the user using histograms, bar graphs, scatter plots, substrate maps, time traces, etc. Exemplary data displays are shown in Figures 5B to 5C. In some embodiments, the data may be displayed via an augmented reality or virtual reality device instead of or in addition to the GUI.

[0151]

[0162] Figure 5B shows a diagram of GUI 500B in a first mode (e.g., monitor mode) of the comprehensive modeling platform according to several embodiments. GUI 500B may be associated with a target processing chamber, selected by the user, for example, through a selection portion of the comprehensive platform, as shown in Figure 5A. Any GUI view may differ from those shown in this disclosure in terms of specific data, plots, component placement, component inclusion, etc., which are left to the designer's customization of the comprehensive platform, user preferences / selections, etc. Several exemplary elements that may be included in monitor mode are included in GUI 500B. Model data such as various types, categories, and groupings may be displayed to the user. In addition, useful data (e.g., depending on the user's preferences and selections) may also be displayed. An example of data that may be displayed but is not related to process modeling may include a chamber component view 510. The chamber component view 510 may show a view of one or more related chamber components, provide tags or labels for various components, and display the effect of chamber conditions on one or more components.

[0152]

[0163] The monitor mode view may include the selection of data view option 512. The illustrated data view options include model output data and model input data, but various data view options are possible. Model output data (e.g., virtual sensors) includes data derived from chamber sensors. Chamber sensors may include sensors positioned to record the state within the chamber, sensors configured to measure the actions taken by chamber components (e.g., actuator position sensors, power supply sensors to components, etc.), or other types of sensors. The model output data can be received as input sensor data (related to previously performed physical processing actions) and can generate one or more metrics of the quality of performance of the actions as output.

[0153]

[0164] The data view option 512 may include model input data. Model input data may include sensor data provided to various models associated with the processing chamber. For example, sensor data used by one or more models to derive substrate measurements can be provided to the user, enabling visual checks for anomalies, unexpected values, etc.

[0154]

[0165] GUI500B may include recipe and recipe view selection elements 514. Selection elements 514 may be used to filter one or more recipes, process executions, process operations, etc., from the associated processing chambers. Selection elements 514 may allow the user to compare data related to various process executions, show changes from one process execution to another, and show differences between one or more target process executions and a baseline or golden run.

[0155]

[0166] GUI500B may include a model selection 516. The model selection 516 can provide a means for determining which models the comprehensive platform will run and which model data the comprehensive platform will display via GUI500B, etc. The model selection 516 can divide models into groups or categories such as substrate feature models, plasma models, substrate temperature models, substrate stress or strain models, electromagnetic models, plasma generation models, etc. For example, each of the selection boxes shown as part of the model selection 516 may correspond to a model category, may include labels (e.g., text labels, illustrative images, etc.), and may be selectable to adjust the model's behavior and / or display data, etc.

[0156]

[0167] GUI500B may further include a model data display area 518. The model data display area 518 may include one or more elements for displaying different model data related to one or more processing executions performed in the corresponding processing chamber. The model data display area 518 may be or include customizable panels, panes, regions, etc. The model data display area 518 may be separated into groups corresponding to groups of model selections 516.

[0157]

[0168] The model data display area 518 may include various types of data displays. For example, bar graphs, histograms, scatter plots, time trace plots, and board dimension prediction images may be displayed. The model data display area 518 may include data of interest to the user and may be customizable by the user to highlight or display target model data, for example. In some embodiments, each display unit of the model data display area 518 may correspond to a selection in the model selection 516. For example, five selected model categories displayed in the model selection 516 may correspond to five data display areas shown in the model data display area 518. In some embodiments, displays may be provided to assist the user in associating the displayed data with the model selection. For example, corresponding boxes may be colored, patterned, labeled, or otherwise displayed to indicate the connection between the model selection 516 and the model data display area 518.

[0158]

[0169] Various other data, selection options, UI elements, etc., may be included in GUI500B, for example, options related to unit selection, data presentation, etc. For example, feature shape prediction 520 or other aspects or indicators of interest may be included in GUI500B. One or more elements of GUI500B may be expandable. For example, the component view 510 can be expanded when the user selects for a more detailed view of one or more chamber components, and the panel of the model data display area 518 can be expanded when the user selects for a more detailed or granular display of model data.

[0159]

[0170] Figure 5C shows an overview of GUI500C for displaying model data in a second mode (e.g., exploration mode) of a comprehensive modeling platform, according to several embodiments. GUI500C may be displayed in response to a user selection of the second mode. GUI500C may be used by the user to view modeling data related to virtual board processing chambers, virtual board processing operations, virtual boards, etc. GUI500C is an exemplary arrangement and may include different UI elements, may not include the UI elements discussed, and UI elements may be spatially rearranged.

[0160]

[0171] GUI500C includes a recipe setting 520. Recipe setting 520 may include fields that the user can use to select modeling parameters. Recipe setting 520 may include one or more fields for selecting model inputs, adjusting simulation knobs, adjusting virtual processing knobs, etc. Fields in recipe setting 520 may include editable fields, dropdown selections, checkboxes, value sliders, or other types of UI elements to accept user specifications. Recipe setting 520 may further include options for model selection, for example, if multiple sets of models are available in the processing chamber. In embodiments, the user can perform recipe setting by interacting with GUI500C to input selections or adjustments of model parameters, simulation knobs, model inputs, virtual processing knobs, checkboxes, value sliders, etc. In some embodiments, the user can perform recipe setting by providing prompts as input. In some embodiments, prompts are processed by a Large Language Model (LLM) to interpret the prompts and determine the parameters, model inputs, virtual processing knobs, etc. to be used in recipe setting. LLM may be incorporated into an application that presents GUI500C, and / or interface with an application to provide the application with commands to execute recipe settings according to received prompts.

[0161]

[0172] GUI500C includes an operational UI element 526. The operational UI element 526 may include UI elements for applying the selection of the recipe setting 520, UI elements for executing one or more models according to the recipe setting 520, and so on. The operational UI element 526 can display data by GUI500C. The operational UI element 526 can execute one or more models. The operational UI element 526 can cause the virtual processing chamber to execute one or more virtual processing operations. The operational UI element 526 can cause the virtual processing chamber to process a virtual board according to the recipe setting 520.

[0162]

[0173] GUI500C includes a recipe preview 522. The recipe preview 522 can be used, for example, to record the recipe input of the last virtual processing operation performed. The recipe preview 522 can allow the user to modify the recipe settings 520 while maintaining the display of recipe data associated with the previous virtual board, recipe data associated with other model data displayed by GUI500C, etc.

[0163]

[0174] GUI500C may include a further mode selection 524. Mode selection 524 may allow access to any other selections that may be applicable to different models, different frameworks for models, different data types, or comprehensive modeling platforms. Figure 5C includes a first selection for running one or more models based on recipe input. Figure 5C also includes a second selection for the user to input one or more desired processing quality measurements (e.g., target substrate measurements, target processing conditions, etc.) and run one or more models to determine the expected simulation and / or process inputs to generate the desired quality measurements.

[0164]

[0175] GUI500C further includes a model data display area 528. The model data display area 528 may share one or more features with the model data display area 518 in Figure 5B, and various graphs, lists, charts, etc., may be used to display data related to various models running as part of a comprehensive modeling platform. The model data display area 528 may include detailed data (e.g., wafer characteristic maps, time trace data, etc.) as well as summary data (e.g., average wafer characteristics, significant condition markers, etc.).

[0165]

[0176] Figure 6 is a block diagram showing computer system 600 in several embodiments. In some embodiments, computer system 600 may be connected to other computer systems (for example, via a network such as a local area network (LAN), intranet, extranet, or the Internet). Computer system 600 may operate as a server or client computer in a client-server environment, or as a peer computer in a peer-to-peer or distributed network environment. Computer system 600 may be provided by a personal computer (PC), tablet PC, set-top box (STB), personal digital assistant (PDA), mobile phone, web appliance, server, network router, switch or bridge, or any device capable of executing a set of instructions (sequential or otherwise) that specify the actions that the device should perform. Furthermore, the term “computer” may include any collection of computers that individually or collectively execute a set of instructions (or sets of instructions) to perform any one or more of the methods described herein, and may include cloud-based devices.

[0166]

[0177] In a further embodiment, the computer system 600 may include a processing unit 602, a volatile memory 604 (e.g., random access memory (RAM)), a non-volatile memory 606 (e.g., read-only memory (ROM) or electrically erasable programmable ROM (EEPROM)), and a data storage device 618, all of which can communicate with each other via a bus 608.

[0167]

[0178] The processing unit 602 can be provided by one or more processors, which are, for example, general-purpose processors (e.g., composite instruction set arithmetic (CISC) microprocessors, reduced instruction set arithmetic (RISC) microprocessors, very long instruction word (VLIW) microprocessors, microprocessors implementing other types of instruction sets, microprocessors implementing other types of combinations), or dedicated processors (e.g., application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), or network processors).

[0168]

[0179] The computer system 600 may further include a network interface device 622 (for example, connected to a network 674). The computer system 600 may also include a video display unit 610 (e.g., an LCD), an alphanumeric input device 612 (e.g., a keyboard), a cursor control device 614 (e.g., a mouse), and a signal generation device 620. The video display unit 610 may display a GUI that enables user interaction with a comprehensive modeling platform.

[0169]

[0180] In some embodiments, the data storage device 618 may include a non-temporary computer-readable storage medium 624 capable of storing instructions 626 that encode any one or more of the methods or functions described herein, including instructions for encoding the components of Figure 1 (e.g., the prediction component 114, the corrective action component 122, the model 190, etc.) and for implementing the methods described herein. The non-temporary computer-readable storage medium 624 may store instructions that encode a comprehensive modeling platform for running one or more models in association with a processing chamber.

[0170]

[0181] Instruction 626 may reside entirely or partially in the volatile memory 604 and / or the processing unit 602 while being executed by the computer system 600, so that the volatile memory 604 and the processing unit 602 may also constitute a machine-readable storage medium.

[0171]

[0182] In the illustrative examples, computer-readable storage medium 624 is shown as a single medium, but the term “computer-readable storage medium” includes a single or multiple mediums that store one or more sets of executable instructions (e.g., centralized or distributed databases, and / or associated caches and servers). The term “computer-readable storage medium” also includes any tangible medium capable of storing or encoding a set of instructions that a computer executes, causing a computer to perform any one or more of the methods described herein. The term “computer-readable storage medium” includes, but is not limited to, solid-state memory, optical media, and magnetic media.

[0172]

[0183] The methods, components, and features described herein may be implemented by separate hardware components or incorporated into the functionality of other hardware components such as ASICS, FPGAs, DSPs, or similar devices. In addition, the methods, components, and features may be implemented by firmware modules or functional networks within hardware devices. Furthermore, the methods, components, and features may be implemented in any combination of hardware devices and computer program components, or in computer programs.

[0173]

[0184] Unless otherwise specified, terms such as “receive,” “execute,” “provide,” “acquire,” “cause,” “access,” “determine,” “add,” “use,” “train,” “reduce,” “generate,” and “correct” refer to computer system operations and processes that manipulate and convert data represented as physical (electronic) quantities within a computer system into other data similarly represented as physical quantities within computer system memory or registers. Furthermore, in this specification, terms such as “first,” “second,” “third,” and “fourth” are presented as labels to distinguish between various elements and do not necessarily have an orderly meaning according to their numerical designation.

[0174]

[0185] The embodiments described herein also relate to apparatus for carrying out the methods described herein. This apparatus may include a general-purpose computer system that is specifically constructed for carrying out the methods described herein or is selectively programmed by a computer program stored in the computer system. Such a computer program may be stored in a computer-readable tangible storage medium.

[0175]

[0186] The methods and exemplary embodiments described herein are not inherently related to any particular computer or other device. A wide variety of general-purpose systems may be used in accordance with the teachings described herein, or it may be convenient to construct more specialized devices to perform each of the methods and / or their individual functions, routines, subroutines, or operations. Examples of the construction of such a wide variety of systems are explicitly shown above.

[0176]

[0187] The above description is illustrative and not limiting. While this disclosure has been described with reference to certain exemplary examples and embodiments, it should be acknowledged that this disclosure is not limited to those examples and embodiments. The scope of this disclosure should be determined in reference to the entire scope of equivalents to which these claims are granted, together with the subsequent claims.

Claims

1. It is a method, The processing unit receives a first user input via a graphical user interface (GUI) to display data related to a first processing chamber in a first chamber data mode, wherein the data in the first chamber data mode includes data of processing operations performed within the first processing chamber. To display in the GUI, in response to receiving a first user selection, first display data of the first chamber data mode is provided. Receiving a second user input for displaying data associated with the first processing chamber in a second chamber data mode, wherein the data in the second chamber data mode includes data of virtual processing operations performed by a virtual representation of the first processing chamber. To display in the GUI, in response to receiving a second user selection, the second display data of the second chamber data mode is provided. Methods that include...

2. Receiving a third user input to display data related to the second processing chamber in either the first chamber data mode or the second chamber data mode, To display in the GUI, a third display data is provided in response to the receipt of a third user selection. The method according to claim 1, further comprising:

3. The first display data is generated by To provide the physical-based digital twin model of the first processing chamber with sensor data of the processing operations performed within the first processing chamber, To obtain the output from the physical-based digital twin model that shows the performance quality of the aforementioned processing operation, Includes, The first display data includes the output from the digital twin model, The method according to claim 1.

4. The second display data is generated by To provide the trained machine learning model associated with the first processing chamber with input data for the virtual processing operation, To obtain the output from the trained machine learning model that indicates the performance quality of the virtual processing operation. Includes, The second display data includes the output from the trained machine learning model, The method according to claim 1.

5. It is a method, The processing device outputs a view of a manufacturing system comprising a plurality of processing chambers, one or more transfer chambers, and one or more load locks, wherein the view outputs a view of the manufacturing system including, for each of the plurality of processing chambers, a graphical representation of the processing chamber and one or more view options for data associated with the processing chamber. The system receives input for the view option of the selected processing chamber from the one or more view options of the aforementioned processing chambers, Loading one or more virtual models associated with the selected processing chamber and the view options, wherein each processing chamber loads one or more virtual models associated with the one or more virtual models configured for that processing chamber. Methods that include...

6. Receiving input for processing operations to be performed within the selected processing chamber, To provide sensor data related to the processing operation to one or more virtual models, To display the data output by one or more virtual models based on the aforementioned sensor data. The method according to claim 5, further comprising:

7. Providing input data to one or more virtual models, wherein the input data is associated with the processed board, Obtaining output data from one or more virtual models, wherein the output data includes a prediction of the cross-sectional profile of the processed substrate. The method according to claim 5, further comprising:

8. The prediction of the cross-sectional profile is The bottom region of the trench of the processed substrate, The side wall of the trench of the processed substrate, or The field adjacent to the trench of the processed substrate, This includes predicting the thickness of the layers of the processed substrate in one or more of the following: The method according to claim 7.

9. The method according to claim 7, wherein the prediction of the cross-sectional profile includes a first prediction of the cross-sectional profile when a first portion of a processing operation related to the processed substrate is performed, and a second prediction of the cross-sectional profile when the first portion of the processing operation and the second portion of the processing operation are performed.

10. Receiving the first set of simulation inputs, To provide the first set of simulation inputs to the one or more virtual models, Obtaining first output data based on a first set of simulation inputs from one or more virtual models that represent virtual board processing operations, To display the first output data, The method according to claim 5, further comprising:

11. The method according to claim 10, wherein the first output data includes characteristic data of a virtual board associated with the virtual board processing operation.

12. The method according to claim 10, wherein the first output data includes characteristic values ​​of a virtual processing chamber associated with the virtual board processing operation.

13. The first processing chamber among the plurality of processing chambers includes a physical vapor deposition (PVD) chamber, and the one or more virtual models associated with the first processing chamber are one or more PVD virtual models. The method according to claim 5, wherein the second processing chamber among the plurality of processing chambers comprises an atomic layer deposition (ALD) chamber, and the one or more virtual models associated with the first processing chamber are one or more ALD virtual models.

14. It is a method, The modeling platform receives a first user input to the first processing chamber via a graphical user interface (GUI), Providing first input data to a first model in response to receiving a first user input from the first processing chamber, wherein the first input data includes sensor data from the first processing chamber. Obtaining first output data from the first model, wherein the first output data includes prediction data associated with the sensor data from the first processing chamber. Providing a second input data to a second model in response to the reception of the first user input in the first processing chamber, wherein the second input data includes a simulation input of a virtual board processing operation. Obtaining second output data from the second model, wherein the second output data includes prediction data of the virtual board processing operation, To provide the first output data and the second output data for display on the GUI of the modeling platform, Methods that include...

15. The method according to claim 14, wherein the first modeling data mode includes the first model, and the first modeling data mode provides input data from a substrate processing operation performed in the first processing chamber to a plurality of models.

16. The method according to claim 14, wherein the first modeling data mode includes the second model, and the second modeling data mode includes providing a plurality of models with simulation inputs for virtual board processing operations.

17. Receiving a second user input in the second processing chamber and Providing a third model in response to receiving the second user input of the second processing chamber, wherein the third input data includes sensor data from the second processing chamber, and the third model is configured to perform the same function in relation to the second processing chamber as the first model is configured to perform in relation to the first processing chamber. Obtaining a third output from the aforementioned third model, To provide the third output for display on the GUI of the modeling platform, The method according to claim 14, including the method described in claim 14.

18. In response to the first user input, a fourth input data is provided to a calibrated physical-based model associated with the first processing chamber, Obtaining a third output from the calibrated physical-based model, To provide the third output for display on the GUI of the modeling platform, The method according to claim 14, including the method described in claim 14.

19. The method according to claim 14, wherein at least one of the first model or the second model includes a trained machine learning model.

20. The method according to claim 14, wherein at least one of the first model or the second model includes a calibrated physically based model.

Images