Post preventative maintenance chamber condition monitoring and simulation

Abstract

The subject matter of this specification can be implemented in, among other things, methods, systems, computer-readable storage medium. A method can include a processing device receiving sensor data indicating a state of an environment of a processing chamber processing a series of substrates according to a substrate processing procedure. The series of substrates may be processed subsequent to a preventive maintenance procedure. The processing device may determine a first set of values based on the sensor data. The set of values may indicate a likelihood an associated substrate comprises a first process result that meets a threshold condition of the substrate processing procedure. The processing device may further predict a first test result based on the first set of values. The first test result may indicate a likelihood a first substrate processed subsequent to processing the series of substrates comprises a second process result that meets the threshold condition.

Description

[Technical Field]

[0001] This specification generally relates to the monitoring and simulation of chamber conditions. More specifically, this specification relates to the monitoring and simulation of chamber conditions associated with chamber aging. [Previous Technology]

[0002] Substrate processing may include a series of processes for forming circuits in a semiconductor (e.g., a silicon wafer) according to a circuit design. These processes may be performed in a series of chambers. The successful operation of modern semiconductor manufacturing facilities is intended to facilitate a stable flow of wafers moving from one chamber to another during the process of forming circuits in the wafer. In processes that perform many substrate processes, the conditions of the processing chambers may be altered, which may result in the processed substrate failing to meet the desired conditions and results.

[0003] One such substrate processing may include plasma etching; plasma etching is a process that transfers a pattern in a photomask material layer to another layer (such as a conductive or dielectric material layer) beneath the photomask by removing a layering material from the wafer surface. This process inevitably produces various types of etching byproducts (such as silicon oxide and organic polymers), depending on the layering material and etching chemistry. Some byproducts deposit on the inner surface of the chamber where the plasma etching process is performed. The deposition of byproducts may affect etching performance, such as by depositing particles (e.g., flakes) onto the substrate or by reacting with the plasma and influencing the processing outcome.

[0004] To mitigate the impact of etching byproducts, preventative maintenance (such as chamber cleaning) can be employed to periodically remove deposits from the chamber walls. For this purpose, production in the chamber is halted, and a cleaning plasma (such as a CF₄+O₂ plasma used to clean silicon oxide deposits during silicon etching) is introduced into the chamber. This plasma reacts with the deposits, and the products of this reaction are pumped out of the chamber. However, following such chamber cleaning, it has been observed that the cleaned chamber walls make the chamber unsuitable for immediate wafer etching. Chamber aging is a process of etching a series of substrates (e.g., blank silicon wafers) to restore chamber conditions suitable for substrate processing. After chamber aging, a thin layer of silicon oxide covers the chamber walls. The chamber is then returned to production wafer etching until the next round of chamber cleaning and aging is required. [Summary of the Invention]

[0005] A method and system for monitoring and simulating after preventative maintenance are implemented. In some embodiments, a method performed by a processing device may include the steps of: receiving sensor data indicating the environmental state of a processing chamber, the processing chamber being used to process a series of substrates according to a chamber aging procedure; processing the series of substrates after performing a preventative maintenance procedure on the processing chamber; the processing device may determine a first set of values ​​based on the sensor data; each of the first set of values ​​may be associated with a specific substrate in the series of substrates; the first set of values ​​may indicate the likelihood that the associated substrates include a first processing result that satisfies a threshold condition of the substrate processing procedure; the processing device may predict a first test result based on the first set of values; the first test result may indicate the likelihood that the first substrate processed after processing the series of substrates includes a second processing result that satisfies a threshold condition; the processing device may prepare the first test result for presentation on a graphical user interface (GUI); and the processing device may modify the operation of the processing chamber based on the first test result.

[0006] In some embodiments, a method for training a machine learning model to predict the conditions of a processing chamber after a preventive maintenance procedure is provided. The method includes the step of generating training data for the machine learning model. The step of generating training data may include the step of identifying a first training input having first sensor data. The first sensor data indicates a first state of the first processing chamber after performing a preventive maintenance procedure. The step of generating training data may further include the step of identifying a first target output for the first training input. The first target output may indicate whether the substrate processed by the first processing chamber in the first state according to the first substrate processing procedure results in the first substrate having a first processing result that satisfies a first threshold condition associated with the first substrate processing procedure. The method may further include the step of providing training data to train the machine learning model on (i) a set of training inputs including the first training input or (ii) a set of target outputs including the first target output. The trained machine learning model may receive new inputs having new sensor data indicating a new state of the new processing chamber. The new output may indicate whether the new substrate processed by the new processing chamber in the new state according to the first substrate processing procedure results in the new substrate having a new processing result that satisfies the threshold condition.

[0007] In some embodiments, the above methods may be stored on a non-transitory machine-readable storage medium including instructions that, when executed by a processing device, cause the processing device to perform operations, such operations may include one or more steps described in the foregoing methods.

Implementation Method

[0017] Substrate processing may include a series of processes for forming circuits in a semiconductor (e.g., a silicon wafer) according to a circuit design. These processes may be performed in a series of chambers. The successful operation of modern semiconductor manufacturing facilities is intended to facilitate a stable flow of wafers moving from one chamber to another during the process of forming circuits in the wafer. In processes that perform many substrate procedures, the conditions in the processing chambers may deteriorate and cause the processed substrate to fail to meet the desired conditions or processing results (e.g., critical dimensions, processing uniformity, and thickness dimensions, etc.).

[0018] Preventive maintenance procedures (e.g., chamber cleaning) are typically used as part of a chamber recovery process to return the processing chamber to a state suitable for entering a substrate processing production mode (e.g., batch processing of substrates). A recovery process is typically used after the preventive maintenance procedure to prepare the chamber for production mode (e.g., "preheating" the chamber). A common recovery process is aging the processing chamber. Chamber aging is a procedure that involves processing a series of substrates (e.g., blank silicon wafers) to restore chamber conditions (e.g., coating the chamber walls) suitable for production substrate processing (e.g., processing substrates in a chamber with processing results that meet desired threshold criteria). After chamber aging, the chamber may operate in production mode for a period of time until another round of preventive maintenance and further chamber aging is required, or it is recommended to otherwise restore the processing chamber to its original state.

[0019] As previously described, the chamber aging process includes processing a series of substrates to prepare a chamber for performing a substrate processing procedure. The substrate processing procedure may include setting conditions for the chamber (e.g., temperature, pressure, vacuum conditions, and plasma conditions, etc.) to obtain a substrate with a target processing result (e.g., the processing result meets a threshold criterion associated with the substrate processing formulation).

[0020] Conventionally, after a preventative maintenance procedure, a series of substrates are processed in a processing chamber. After processing this series of substrates, a test sample is processed and evaluated. The test sample may include substrates processed in the processing chamber and evaluated (e.g., measuring the processing results) to determine whether the test sample contains substrates whose processing results meet threshold conditions. If the conditions(s) are not met, further aged substrates may be processed and another test sample may be processed to determine whether the chamber has been fully restored. This process may be repeated until the chamber conditions are met or a threshold amount of aged wafers has been used. The conventional method for determining chamber conditions as described above may require an unnecessary number of aged wafers (e.g., an "over-aged" chamber) when a first number of aged wafers are used but a smaller number is actually needed to restore the chamber to an operable state for substrate production. Furthermore, the conventional method may delay returning the chamber to production mode by waiting for test sample measurements and evaluations to determine whether the chamber has been sufficiently restored. The time spent on the recovery process is a loss of productivity because the tool or chamber is down (e.g., not operating in production mode). Furthermore, conventional methods do not provide in-situ chamber condition monitoring; conventional approaches rely on test samples and await evaluation of the test samples during aging.

[0021] The present application addresses these and other disadvantages of the prior art by providing methods and systems in various embodiments that can estimate chamber conditions without relying on the results of test sample examinations, provide quantitative information about chamber health and recovery after preventative maintenance procedures, and simulate chamber conditions with artificial parameter adjustments. In some embodiments, machine learning models may be used to determine chamber conditions and / or states during and / or after the aging process. In some embodiments, the present application enables health checks and preventative measures during the chamber aging sequence and early detection of abnormal chamber conditions before sample testing. One or more of these embodiments and / or implementations can reduce downtime of processing tools (e.g., processing chambers) and enable higher productivity of manufacturing equipment.

[0022] In exemplary embodiments, methods and systems for monitoring and simulation after preventative maintenance are performed. In some embodiments, a method performed by a processing device may include the steps of: receiving sensor data indicating the environmental state of a processing chamber, the processing chamber being used to process a series of substrates according to a chamber aging procedure; processing the series of substrates after performing a preventative maintenance procedure on the processing chamber; the processing device may determine a first set of values ​​based on the sensor data; each of the first set of values ​​may be associated with a specific substrate in the series of substrates; the first set of values ​​may indicate the likelihood that the associated substrate includes a first processing result that meets a threshold condition of the substrate processing procedure; the processing device may predict a first test result based on the first set of values; the first test result may indicate the likelihood that the first substrate processed after processing the series of substrates includes a second processing result that meets a threshold condition; the processing device may prepare the first test result for presentation on a graphical user interface (GUI); and the processing device may modify the operation of the processing chamber based on the first test result.

[0023] In an exemplary embodiment, a method for training a machine learning model to predict the conditions of a processing chamber after a preventive maintenance procedure is provided. The method includes the step of generating training data for the machine learning model. The step of generating training data may include the step of identifying a first training input having first sensor data. The first sensor data may indicate a first state of the first processing chamber after performing a preventive maintenance procedure. The step of generating training data may further include the step of identifying a first target output for the first training input. The first target output may indicate whether the substrate processed by the first processing chamber in the first state according to the first substrate processing procedure results in the first substrate having a first processing result that satisfies a first threshold condition associated with the first substrate processing procedure. The method may further include the step of providing training data to train the machine learning model on (i) a set of training inputs including the first training input or (ii) a set of target outputs including the first target output. The trained machine learning model may receive new inputs having new sensor data indicating a new state of the new processing chamber. The new output may indicate whether the new substrate processed by the new processing chamber in the new state according to the first substrate processing procedure results in the new substrate having a new processing result that satisfies the threshold condition.

[0024] In some embodiments, the above methods may be stored on a non-transitory machine-readable storage medium including instructions that, when executed by a processing device, cause the processing device to perform operations, such operations may include one or more steps described in the foregoing methods.

[0025] FIG1 is a block diagram illustrating an exemplary system architecture 100 in which embodiments of the present application may operate. As shown in FIG1, the system architecture 100 includes a manufacturing system 102, a metering system 110, a client device 150, a data storage 140, a server 120, and a machine learning system 170. The machine learning system 170 may be part of the server 120. In some embodiments, one or more components of the machine learning system 170 may be wholly or partially integrated into the client device 150. Each of the manufacturing system 102, the metering system 110, the client device 150, the data storage 140, the server 120, and the machine learning system 170 may be hosted by one or more computing devices; the one or more computing devices include server computers, desktop computers, laptop computers, tablet computers, notebook computers, personal digital assistants (PDAs), mobile communication devices, mobile phones, handheld computers, or similar computing devices.

[0026] Manufacturing system 102, metering system 110, client device 150, data storage 140, server 120, and machine learning system 170 may be coupled to each other via network 160 (e.g., for performing the methods described herein). In some embodiments, network 160 is a private network that provides each element of system architecture 100 with access to each other and other privately available computing devices. Network 160 may include one or more wide area networks (WANs), local area networks (LANs), wired networks (e.g., Ethernet), wireless networks (e.g., 802.11 networks or Wi-Fi networks), cellular networks (e.g., LTE networks), routers, hubs, switches, server computers, and / or any combination thereof. Alternatively or additionally, any element of system architecture 100 may be integrated together or otherwise coupled without using network 160.

[0027] Client device 150 may be or include any personal computer (PC), laptop, mobile phone, tablet computer, laptop, networked TV ("smart TV"), networked media player (e.g., Blu-ray player), set-top box, set-top box (OOT) streaming device, and operator box, etc. Client device 150 may include browser 152, application 154, and / or other tools as described and executed by other systems of system architecture 100. In some embodiments, the client device 150 may, as described herein, process at various stages of the system architecture 100 instructions to access the manufacturing system 102, metrology system 110, data storage 140, server 120 and / or machine learning system 170 and transmit (e.g., transmit and / or receive) sensor data, processed data, data classification, chamber recovery data (e.g., recovery curves), processed result data and / or inputs and outputs of various processing tools (e.g., metrology tool 114, data preparation tool 116, classifier tool 124, chamber recovery tool 126 and / or chamber recovery components).

[0028] As shown in FIG1, the manufacturing system 102 includes a processing tool 104, a processing procedure 106, and a processing controller 108. The processing controller 108 can coordinate the operation of the processing tool 104 to execute one or more processing procedures 106. For example, various processing tools may include dedicated chambers such as etching chambers, deposition chambers (including chambers for atomic layer deposition, chemical vapor deposition, or plasma-enhanced versions thereof), annealing chambers, and / or the like. In another example, the machine may incorporate a sample transport system (e.g., a Selective Compliant Assembly Robotic Arm (SCARA) robot, a transfer chamber, a front-opening wafer transfer box (FOUP), a side storage box (SSP), and / or the like) to transport samples between the machine and processing steps.

[0029] Processing procedure 106 (or sometimes referred to as processing formula or processing step) may include various specifications for performing the operation by means of processing tool 104. For example, processing procedure 106 may include processing specifications such as the start-up duration of the processing operation, the processing tool used for the operation, the temperature, flow rate and pressure of the machine (e.g., chamber), and the deposition sequence, etc. In another example, the processing procedure may include instructions for transporting the sample to a further processing step or for measuring the sample by the metering system 110. One or more processing procedures 106 include preventative maintenance procedures and aging procedures.

[0030] The processing controller 108 may include devices designed to manage and coordinate the actions of the processing tool 104. In some embodiments, the processing controller 108 is associated with a processing recipe or a series of processing procedure 106 instructions that, when applied by design, result in a desired processing outcome for the substrate. For example, the processing recipe may be associated with processing the substrate to produce a target processing outcome (e.g., critical dimensions, thickness, and uniformity standards). The processing controller 108 may coordinate the execution of preventative maintenance procedures (e.g., chamber cleaning) and chamber recovery procedures (e.g., chamber aging) to return the chamber to a production state.

[0031] As shown in FIG. 1, the metrology system 110 includes a metrology tool 114 and a data preparation tool 116. The metrology tool 114 may include various sensors to measure processing results (e.g., critical dimensions, thickness, and uniformity, etc.) within the manufacturing system 102. For example, wafers processed in one or more processing chambers can be used to measure critical dimensions. The metrology tool 114 may also include means for measuring the processing results of substrates processed using the manufacturing system. For example, the processing results of substrates processed according to the processing recipe and / or actions performed by the processing controller 108, such as critical dimensions, thickness measurements (e.g., films from etching, deposition, etc.), can be evaluated. These measurements can also be used to measure chamber conditions during recovery processing, such as during aging processes that include processing a series of wafers (e.g., blank wafers).

[0032] The data preparation tool 116 may include processing methods for extracting features and / or generating synthetic / engineering data associated with the data measured by the metrology tool 114. In some embodiments, the data preparation tool 116 may identify correlations, patterns, and / or anomalies in metrology or processing performance data. For example, the data preparation tool 116 may perform feature extraction, wherein the data preparation tool 116 uses a combination of measurement data to determine whether a criterion is met. For example, the data preparation tool 116 may analyze multiple data points of relevant parameters (e.g., thickness, critical dimensions, defect rate, and plasma conditions, etc.) to determine whether rapid changes occur during iterations of aging treatment within the processing chamber. In some embodiments, the data preparation tool 116 performs normalization on various sensor data associated with various processing chamber conditions. Normalization may include processing incoming sensor data to make it appear similar across various chambers and various sensors used to acquire the data.

[0033] In some embodiments, the data preparation tool 116 may perform one or more of the following on measurement data (e.g., obtained by the measurement tool 114): processing control analysis, univariate limit violation analysis, or multivariate limit violation analysis. For example, the data preparation tool 116 may perform statistical processing control (SPC) using a statistically based approach to monitor and control the processing controller 108. For example, SPC may improve the efficiency and accuracy of preventative maintenance and / or wafer aging processes (e.g., by identifying data points that fall into and / or exceed control limits).

[0034] In some embodiments, the processing chamber can be measured throughout the aging process. In some embodiments, additional measurement data is acquired during a predetermined aging process. For example, additional sensors can be activated during or immediately after wafer processing, and / or currently activated sensors can acquire additional data. In some embodiments, the processing controller 108 can trigger measurements of the metering tool 114 based on an operation to be performed by the processing tool 104. For example, in response to one or more test wafers aging the wafer being processed during a transition period (during which, if the processing result meets the criteria within the processing procedure 106, the processing chamber completes its aging procedure and awaits the next production wafer to be processed), the process controller 108 can trigger (e.g., the metering tool 114) the initiation of one or more processing results.

[0035] In some embodiments, the extracted features, the generated synthetic / engineered data, and statistical analyses may be used in conjunction with the machine learning system 170 (e.g., to train, validate, and / or test the machine learning model 190). Additionally and / or alternatively, the data preparation tool 116 may output data to the server 120 for use by either the classifier tool 124 or the chamber reconstruction tool 126.

[0036] Data storage 140 may be memory (e.g., random access memory), drive (e.g., hard disk and flash drive), database system, or another type of component or device capable of storing data. Data storage 140 may store one or more historical data 142, including old sensor data 144 and / or old processing result data 146 of previous chamber conditions and processing results of the substrate processed under associated chamber conditions. In some embodiments, historical data 142 may be used to train, validate, and / or test the machine learning model 190 of machine learning system 170 (see, for example, FIG. 6 for an exemplary method).

[0037] Server 120 may include one or more computing devices, such as rack-mount servers, router computers, server computers, personal computers, mainframe computers, laptop computers, tablet computers, and desktop computers, etc. Server 120 may include classifier tool 124, chamber recovery tool 126, and data simulation tool 128.

[0038] Classifier tool 124 receives measurement data (e.g., chamber sensor data) from processing tool 104 and determines the state of the chamber throughout the chamber aging process. In some embodiments, classifier tool 124 receives raw sensor data from processing tool 104; in other embodiments, the raw sensor data is combined with synthetic data processed from data preparation tool 116. Classifier tool 124 may process the sensor data to determine a set of values ​​indicating the likelihood that processing the substrate under relevant chamber conditions will result in the substrate having a processing outcome that meets the threshold criteria corresponding to processing procedure 106. Classifier tool 124 may classify the state or condition of the processing chamber based on the received sensor data and a machine learning model trained in machine learning system 170. For example, the state of the processing chamber during aging may be processed to determine whether the chamber meets the conditions for transitioning from chamber recovery mode to chamber production mode. In some embodiments, classifier tool 124 includes a machine learning model that uses processed data (e.g., sensor data (e.g., from processing tool 104), synthetic and / or engineered data (e.g., from processing tool 104) and general processing parameter values ​​corresponding to processing procedure 106) to determine a set of values ​​indicating the likelihood that a chamber will meet a threshold condition during the aging process. As discussed later, the machine learning model may include a bootstrap aggregation model, a random forest tree decision tree model, or a partial least squares regression (PLS) model, and other models. The machine learning model may include ensemble modeling; ensemble modeling includes multiple models and utilizes a higher confidence model to perform a final classification of the received data.

[0039] The chamber recovery tool 126 may receive data from the classifier tool 124 and generate a chamber recovery curve associated with the recovery of a processed chamber, which is associated with preventive maintenance and / or aging procedures. The chamber recovery tool 126 may determine preventive maintenance indicators that indicate the likelihood of a chamber meeting the conditions for transitioning from a recovery operation mode to a production operation mode. For example, the chamber recovery tool may identify when the conditions of the chamber meet a threshold condition such that a substrate treated under the chamber conditions results in a substrate having a processing outcome that meets the threshold condition associated with one or more processing procedures 106. In some embodiments, the chamber recovery tool 126 may use a machine learning model that receives input data from the classifier tool 124 and outputs preventive maintenance indicator values ​​associated with the likelihood of a chamber meeting the threshold condition to transition to a production operation mode.

[0040] The data simulation tool 128, in coordination with the classifier tool 124 and the chamber recovery tool 126, identifies changes in the chamber state (e.g., readiness to transition to production mode, need for further aging, and need for further preventative maintenance) based on components of sensor data acquired using the manufacturing system 102, adjusted manually. The data simulation tool 128 can identify one or more sensor data parameters or combinations thereof that have a greater influence on the chamber recovery process. The data simulation tool can utilize the classifier tool 124 and the chamber recovery tool 126 to identify the state of the chamber under simulated conditions. The data simulation tool 128 can identify one or more processing priorities associated with the chamber recovery procedure. For example, one or more simulation models can determine one or more weights associated with various process parameters (e.g., temperature, pressure, and vacuum conditions, etc.).

[0041] As previously described, some embodiments of the classifier tool 124, chamber restoration tool 126, and / or data simulation tool 128 may use machine learning models to perform the methods described therein. The associated machine learning model may be generated (e.g., trained, validated, and / or tested) using the machine learning system 170. An exemplary description of the machine learning system 170 is described below in the context of using the machine learning system 170 to generate a machine learning model 190 associated with the classifier tool 124. However, it should be noted that this description is purely exemplary. Similar processing hierarchies and methods will be discussed further in conjunction with other embodiments for generating and performing machine learning models associated with the classifier tool 124, chamber restoration tool 126, and / or data simulation tool 128 individually and / or in combination with each other.

[0042] The machine learning system 170 may include one or more computing devices, such as rack-mount servers, router computers, server computers, personal computers, mainframe computers, laptop computers, tablet computers, and desktop computers, etc. The machine learning system 170 may include a chamber state component 194. In some embodiments, the chamber state component 194 may use historical data 142 to determine whether the state of the processing chamber meets the conditions for transitioning to a production operation mode. In some embodiments, the chamber state component 194 may use a trained machine learning model 190 to determine the state of the chamber. The trained machine learning model 190 may use historical data to determine the chamber state.

[0043] In some embodiments, the trained machine learning model 190 is stored on server machine 192 and / or server 120. Server machine 180 and / or server machine 192 may store one or more recently trained machine learning models (e.g., to perform model performance management). For example, one or more models trained over a period of time (e.g., the past 6 months) may be stored on server machine 192, and one or more final models for predicting chamber conditions may be stored on server 120 (e.g., to operate in conjunction with chamber recovery tool 126).

[0044] In some embodiments, the machine learning system 170 further includes server machine 172 and server machine 180. Server machines 172 and 180 may be one or more computing devices (such as rack servers, router computers, server computers, personal computers, mainframe computers, laptop computers, tablet computers, desktop computers, etc.), data storage (e.g., hard disks, memory databases), networks, software components, or hardware components.

[0045] Server machine 172 may include dataset generator 174; dataset generator 174 is capable of generating datasets (e.g., a set of data inputs and a set of target outputs) to train, validate, or test machine learning models. Dataset generator 174 may divide historical data 142 into a training set (e.g., 60 percent of historical data, or any other portion of historical data), a validation set (e.g., 20 percent of historical data, or some other portion of historical data), and a test set (e.g., 20 percent of historical data). In some embodiments, chamber state component 194 generates multiple sets of training data. For example, one or more sets of training data may include each dataset (e.g., training set, validation set, and test set).

[0046] Server machine 180 includes a training engine 182, a validation engine 184, and a testing engine 186. The training engine 182 may be able to train a machine learning model 190 using one or more old sensor data 144 and old processing result data 146 from historical data 142 (of data storage 140). In some embodiments, the machine learning model 190 may be trained using one or more outputs from data preparation tool 116, classifier tool 124, chamber recovery tool 126, and / or data simulation tool 128. For example, the machine learning 190 may be a hybrid machine learning model using sensor data and / or mechanical features (such as feature extraction, mechanical modeling, and / or statistical modeling (e.g., using data preparation tool 116)). The training engine 182 may generate multiple trained machine learning models 190, each corresponding to a different set of features for each training set.

[0047] The validation engine 184 can determine the accuracy of each trained machine learning model 190 based on the corresponding feature set of each training set. The validation engine 184 can discard trained machine learning models 190 with accuracy that does not meet the threshold accuracy. The testing engine 186 can determine the trained machine learning model 190 with the highest accuracy among all trained machine learning models based on the test (and optionally, validation) set.

[0048] In some embodiments, training data is provided to train the machine learning model 190, such that the trained machine learning model can receive new inputs having new sensor data indicating a new state of the new processing chamber. The new output can indicate whether the new substrate processed by the new processing chamber in the new state according to the first substrate processing procedure results in a new processing outcome that satisfies a threshold condition.

[0049] Machine learning model 190 may refer to a model created by training engine 182 using a training set, which includes data inputs and corresponding target outputs (historical results of the processing chamber under parameters associated with the target inputs). Patterns in the dataset can be discovered that map data inputs to target outputs (e.g., identifying connections between portions of sensor data and resulting chamber states), and mappings capturing these patterns are provided to machine learning model 190. Machine learning model 190 may use one or more of logistic regression, syntactic parsing, decision trees, or support vector machines (SVMs). Machine learning may consist of single-level linear or non-linear operations (e.g., SVMs) and / or may be neural networks.

[0050] The chamber state component 194 can provide current data (e.g., current sensor data associated with the state of the processing chamber during the aging process) as input to a trained machine learning model 190, and the chamber state component 194 can run the trained machine learning model 190 on the input to obtain one or more outputs; this one or more outputs include a set of values ​​indicating the probability (e.g., without wasting unwanted aging wafers) of satisfying a threshold condition at various points in the aging process to determine the optimal number of substrates to be processed during the aging process and to prepare the chambers for production. The chamber state component 194 may be able to identify confidence data from the output; this confidence data indicates the confidence level at which the state of the processing chamber at various points in the aging process satisfies the threshold condition to move to the production mode. In a non-limiting example, the confidence level is a real number between 0 and 1 (inclusive), where 0 indicates no confidence in one or more chamber states, and 1 indicates absolute confidence in the chamber state.

[0051] For illustrative and not limiting purposes, this application describes the training of a machine learning model and the use of a trained learning model with information relating to historical data 142. In other embodiments, a heuristic or rule-based model is used to determine the chamber state.

[0052] In some embodiments, the functionality of client device 150, server 120, data storage 140, and machine learning system 170 may be provided by a fewer number of machines than shown in FIG. 1. For example, in some embodiments, server machines 172 and 180 may be integrated into a single machine, while in some other embodiments, server machines 172, 180, and 192 may be integrated into a single machine. In some embodiments, machine learning system 170 may be provided entirely or partially by server 120.

[0053] Generally, the functions described in one embodiment as being performed by the client device 150, data storage 140, metering system 110, manufacturing system 102, and machine learning system 170 may also be performed on the server 120 in other embodiments (if appropriate). Furthermore, functions belonging to a particular component may be performed by multiple components, either different or running together.

[0054] In this embodiment, a "user" may be referred to as a single individual. However, other embodiments of this application cover a "user" as an entity controlled by multiple users and / or automated sources. For example, a group of individual users united as a group of administrators may be considered a "user".

[0055] Figures 2A and 2B are block diagrams illustrating chamber recovery processes 200A-200B that may operate according to embodiments of the present application. Chamber recovery processes 200A-200B may include performing preventative maintenance 202 on one or more processing tools. Preventative maintenance 202 may include any number of processes performed on one or more processing tools (e.g., processing chambers) to maintain the condition of the processing chamber over time. In some embodiments, preventative maintenance 202 may employ a proactive preventative maintenance procedure having one or more of the following objectives: minimizing downtime of the processing equipment, maintaining low maintenance costs (e.g., maintaining minor repairs rather than preventing major, less frequent repairs), extending the lifespan of the processing equipment, improving product quality (e.g., the processing results of substrates), improving processing efficiency (e.g., keeping tools in production mode for longer periods and reducing downtime of the processing equipment), and / or maintaining the safety and integrity of the equipment, etc.

[0056] In some embodiments, preventive maintenance 202 includes cleaning (e.g., wet cleaning maintenance) one or more processing chambers (e.g., removing buildup on the chamber walls due to processing procedures performed within the processing chamber). Preventive maintenance 202 may restore the conditions of one or more processing tools to an unaged state (e.g., a clean chamber that has not yet undergone aging). Replacement and / or cleaning of one or more processing tools may result in the processing tools operating in a clean (e.g., original) state; this state may be further fine-tuned by using the processing tools to perform a substrate processing procedure. For example, after performing preventive maintenance 202, the processing chamber may return to an unaged state, such as a new state or a clean state.

[0057] As shown in FIG2, the chamber recovery processes 200A to 200B include chamber initialization. Chamber initialization 204 may include preheating the chamber or activating one or more systems associated with the processing chamber. For example, chamber initialization 204 may be performed to prepare the processing chamber for receiving and processing a substrate. Chamber initialization may be associated with one or more processing procedures (e.g., processing procedure 106 of FIG1) that may include one or more processing parameters (e.g., temperature, pressure, and vacuum conditions, etc.).

[0058] As shown in Figure 2, the chamber recovery processes 200A to 200B include aging 206 of the processing chamber. Chamber aging 206 is a process of processing a series of substrates (e.g., blank silicon wafers) to restore chamber conditions suitable for processing production substrates. For example, in some embodiments, after chamber aging, a thin layer of silicon oxide covers the chamber walls. At aging 206, an initial number of aged wafers (e.g., 20 or more aged wafers) are processed. Chamber recovery process 200A uses test sample 208 to determine whether the chamber conditions are met for transition to production mode, while chamber recovery process 200B uses preventative maintenance (PM) model 212 to determine whether the chamber conditions are met for transition to production mode.

[0059] In some embodiments, as shown in FIG2A, after aging the chamber with a first number of wafers, a sample test 208 is performed and evaluated. The test sample may include a substrate processed by a processing chamber under post-aging conditions of the first number of substrates. The sample is measured, and various processing result criteria are evaluated based on the sample test. For example, the sample test 208 may be evaluated to determine whether the substrate includes processing results that meet threshold criteria; such threshold criteria include critical size, processing uniformity threshold, and thickness distribution requirements, etc. Processing and evaluating the sample test may involve the duration of downtime required to process the substrate and evaluate the processed substrate (e.g., measuring the processing results such as thickness and uniformity conditions across the entire substrate surface).

[0060] At block 210A, the chamber recovery process 200A can determine whether one or more threshold conditions for the processed chamber are met based on sample test 208. Threshold conditions may include criteria associated with one or more processing results of sample test 208. As previously described, sample test 208 can be evaluated to determine whether the substrate includes processing results that meet threshold criteria; such threshold criteria include critical dimensions, processing uniformity thresholds, and thickness distribution requirements, etc., any of which can be evaluation criteria used to determine whether the chamber is ready to transition to an operational production mode. The processing logic can determine that the state of the chamber does not meet the threshold conditions(s), and the processing logic can further perform additional aging 206 and another sample test 208 to determine whether the chamber has been properly aged (e.g., meets the threshold conditions). The process of aging 206, performing sample test 208, and evaluating whether the threshold conditions (e.g., block 210) for the processed chamber are met based on previous sample tests can be repeated until the chamber meets the threshold conditions or until a threshold number of aged wafers or aging iterations have been performed. When the threshold condition is met, the chamber recovery process 200A can continue to initialize the production mode of the processing chamber.

[0061] The chamber recovery process 200B includes performing preventative maintenance (PM), chamber initialization, and initial aging 206 as described above. As shown in FIG2B, the chamber recovery process 200B continues to monitor the health status of the chamber and assess the state of the treated chamber using the PM model 212. The PM model 212 may receive sensor data indicating the state of the treated chamber during aging 206. In some embodiments, the PM model includes one or more samples and / or features of a classifier tool 124, a chamber recovery tool 126, and / or a data simulation tool 128. The PM model 212 may output a notification indicating the state of the treated chamber. In some embodiments, the notification may include instructions for chamber recovery. For example, the PM model 212 may instruct authorization to perform a sample test 208 (e.g., to verify the results of the PM model 212).

[0062] At block 210, the chamber recovery process 200B receives output from the PM model 212 and determines further action to be taken in response to the received output. In some embodiments, the notification may include an instruction for chamber recovery. For example, the PM model 212 may instruct authorization to perform sample testing 208 (e.g., to verify the results of the PM model 212). The PM model 212 may output a notification that the processed chamber has not been recovered and that further aging of the chamber is required before entering production mode. The PM model 212 may output a notification indicating that the quality of preventive maintenance does not meet a threshold condition and that further preventive maintenance is required. In some embodiments, the chamber recovery process 200B may determine that the conditions for processing the chamber are met and that production mode can be initialized without using sample testing. In some embodiments, the chamber recovery process 200B may determine that previously performed preventive maintenance is insufficient and that further preventive maintenance may be needed to meet the threshold condition for processing the chamber without performing multiple aging iterations 206. For example, when the threshold condition is not met, the processing logic does not redirect the processing logic to aging 206, but instead redirects the processing back to preventive maintenance 202 (instead of performing aging that would not allow the chamber conditions to meet the threshold condition (e.g., wasting aging wafers).

[0063] In some embodiments, the chamber recovery process 200B uses less time (e.g., less downtime of the processing chamber) to return the chamber to production mode. For example, the chamber recovery process 200B may not require the execution of sample tests (e.g., processing and evaluation of test wafers) to reach a decision that the processing chamber meets one or more threshold conditions. In another example, the chamber recovery process 200B may determine that further preventative maintenance (rather than further aging of the processing chamber) is preferable to further restore the state of the chamber to meet one or more threshold conditions.

[0064] As will be further discussed in other embodiments, the chamber recovery process 200A can be used to generate training data for the PM model 212. Sensor data of the processing chamber can be tracked during the aging process 206, and data associated with the sample test 208 can be used to validate the output from the PM model 212. For example, the PM model 212 can determine the state of the processing chamber. The evaluation of the sample test processed under the same chamber conditions can be compared with the predictions of the PM model 212 to validate and further improve the PM model 212.

[0065] FIG3 is a block diagram illustrating a chamber monitoring and maintenance system 300 in which an embodiment of the present application may operate. The chamber monitoring and maintenance system 300 includes a substrate processing system 302 (e.g., manufacturing system 102 of FIG1). The substrate processing system 302 may include a processing tool 304 and (e.g., chamber data 306 acquired using the processing tool 304). The processing tool 304 may include chamber sensors (e.g., for measuring radio frequency (RF), temperature, and gases, etc.) and additional sensors (e.g., optical emission spectroscopy (OES) sensors). The chamber data 306 may include data indicating the state of the processed chamber. For example, the processing tool 304 may measure and monitor parameters such as RF, temperature, gas, and spectral data, etc.

[0066] In some embodiments, chamber data 306 is processed by data preparation tool 308. Data preparation tool 308 may receive chamber data (e.g., raw sensor data) and generate synthetic data associated with various combinations, correlations, and / or artificial parameters. Data preparation tool 308 may reduce the dimensionality of raw sensor data into groups or features. For example, data preparation tool 308 may generate features including rates of change of parameter values ​​and / or sets of parameter values. In some embodiments, data preparation tool 308 performs any one of partial least squares analysis, principal component analysis, multi-factor dimensionality reduction, nonlinear dimensionality reduction, and / or any combination thereof. In some embodiments, data preparation tool 308 is designed for edge detection of chamber data. For example, data preparation tool 308 includes methods designed to identify data points that change abruptly and / or have discontinuities. For example, processing the slope of one or more measured chamber parameters.

[0067] In some embodiments, the data preparation tool 308 performs normalization on various sensor data associated with various processing chamber conditions. Normalization may include processing the incoming sensor data to make it appear similar across various chambers and various sensors used to acquire the data. In some embodiments, the data preparation tool 308 may include the patterns and / or features of the data preparation tool 116 of FIG1.

[0068] In some embodiments, data preparation tool 308 is used to generate input for classifier model 310. As further discussed in other embodiments, classifier model 310 receives input data (e.g., raw sensor data, synthetic data, and outputs of other models, etc.) to determine the state of the processing chamber (e.g., classifying the processing chamber into one of a set of categories). In some embodiments, the classifier model receives sensor data and measurement data 307 indicating the state of the processing chamber when processing a series of substrates (e.g., aged wafers). The classifier model may output a set of values ​​indicating the probability that processing a substrate after an associated substrate in the series of substrates will result in the substrate having a processing outcome that satisfies (e.g., a threshold condition associated with a substrate processing procedure). For example, classifier model 310 may output a series of probability values, each probability value associated with one or more substrates in the series of substrates (e.g., aged wafers). The probability values ​​may indicate the probability that processing an associated substrate or a substrate after an associated substrate will result in the substrate having a processing outcome that satisfies one or more threshold conditions associated with a substrate processing procedure (e.g., a substrate processing recipe). In some embodiments, as will be further discussed in other embodiments, the classifier model 310 may be incorporated into a machine learning model (e.g., trained using FIG6 and / or implemented using FIG5).

[0069] As shown in FIG3, the chamber monitoring and maintenance system 300 may include a chamber recovery tool 312. The chamber recovery tool 312 may receive output from a classifier model 310. The chamber recovery tool 312 may approximate a chamber recovery curve (e.g., the chamber recovery curve of FIG4) by determining a series of PM indices associated with the output of the classifier model 310. In some embodiments, the series of PM indices is based on the analysis of current and previous data (e.g., historical data 142 of FIG1). The chamber recovery tool 312 may determine a threshold PM index; this threshold PM index distinguishes whether the chamber conditions are good enough to pass the sample test, whether more aging is required, or whether further preventative maintenance is required. A series of PM wafer PM index results in the aging process can be obtained, which enables in-situ chamber monitoring by the chamber monitoring and maintenance system 300.

[0070] In some embodiments, the chamber recovery tool 312 determines one or more PM indices associated with the state of the processing chamber. The chamber recovery tool 312 may perform statistical optimization calculations to determine a mapping between threshold conditions and the number of aged wafers and the PM indices. Exemplary methods for calculating the PM indices may include the following: where Yet is a classifier model result associated with each of the series of aged wafers in the aging run.

[0071] In some embodiments, the chamber recovery tool 312 outputs instructions to the processing controller 314 to change one or more operations associated with the substrate processing system 302. For example, the chamber recovery tool 312 may indicate the end of the aging process, and the processing controller may change the operating parameters of the substrate processing system to enter a production operation mode. In another example, the chamber recovery tool 312 may indicate that the processing chamber is almost recovered and requires further aging to improve the condition of the processing chamber to a point sufficient to enter a production operation mode. In yet another example, the chamber recovery tool 312 may indicate to the processing controller that the preventive maintenance quality is insufficient (e.g., failure to meet threshold conditions) and the processing controller 314 may instruct the substrate processing system 302 to perform further preventive maintenance.

[0072] In some embodiments, the chamber recovery tool 312 approaches the desired chamber conditions during the aging process. The chamber recovery tool 312 interleaves the probabilities received from the classifier model 310 and estimates the conditions of a chamber at a given point in time. These conditions can be represented as PM indices. As will be discussed further in conjunction with FIG4, these PM indices can be defined by a logarithmic function. The chamber recovery tool 312 can determine the number of substrates processed in the aging process (e.g., to meet the threshold conditions of the processed chamber). For example, the PM indices can be used to determine the optimal aging time (e.g., the optimal number of wafers used in the aging process).

[0073] As shown in FIG3, the processing controller 314 receives input from the chamber recovery tool 312 and provides instructions to the substrate processing system 302 to perform further operations. For example, the processing controller 314 may instruct the substrate processing system 302 to continue further aging of one or more processing chambers. In another example, the processing controller 314 may instruct the substrate processing system to end the aging of one or more processing chambers. In another example, the processing controller 314 may instruct the substrate processing system to perform preventive maintenance on one or more processing chambers.

[0074] As shown in FIG3, the chamber restoration tool 312 can output instructions to be displayed on a graphical user interface (GUI). The GUI can output chamber state predictions on the GUI. In some embodiments, the chamber restoration tool 312 outputs instructions associated with one or more operations performed on the substrate processing system to be displayed on the GUI 316. For example, the GUI 316 can display a notification to continue further aging of one or more processes. In another example, the GUI 316 can display instructions to end the aging of one or more processing chambers. In another example, the GUI 316 can display a notification to perform preventive maintenance on one or more processing chambers.

[0075] Figure 4 depicts a graph 400 illustrating chamber recovery curves according to some embodiments of this application. Graph 400 depicts chamber recovery curves (e.g., generated by chamber recovery tool 312). The recovery curve may include a first data sequence 406 indicating each PM index associated with a separate aged wafer 404. The recovery curve may indicate a threshold 408 PM index, where the conditions of the processing chamber indicate the satisfaction of the chamber recovery process. For example, the intersection between the threshold 408 and the first data sequence 406 may indicate the optimal number of aged wafers used in an aging run, making the processing chamber ready for production mode. Specifically, the processing chamber may process subsequent substrates, and the substrates are likely to have processing results that satisfy the threshold conditions of the substrate processing procedure. In some embodiments, the first data sequence 402 may intersect with the threshold 408 and indicate that the chamber is ready for execution in production mode. In other embodiments, the data sequences in other recovery curves may not intersect with the threshold and may indicate that the chamber is not ready for production mode and may require further preventative maintenance and / or chamber aging.

[0076] Figures 5 to 7 depict flowcharts illustrating exemplary methods 500-700 related to training and / or using machine learning models associated with chamber data according to some embodiments of this application. For simplicity of explanation, methods 500-700 are depicted and described as a series of actions. However, actions according to this application may occur in various orders and / or simultaneously with other actions not presented and described herein. Furthermore, implementing methods 500-700 according to the disclosed subject matter does not require performing all the actions shown. In addition, those skilled in the art will understand and appreciate that methods 500-700 may alternatively be represented as a series of interrelated states via state diagrams or events.

[0077] Figure 5 depicts a flowchart of an example method 500 for monitoring and predicting conditions in a processing chamber according to some embodiments of this application. Method 500 is executed by processing logic; the processing logic may include hardware (e.g., circuitry and dedicated logic, etc.), software (such as running on a general-purpose computer system or a dedicated machine), or any combination thereof. In one embodiment, the method is executed using a server 120 and the trained machine learning model 190 of Figure 1, while in some other embodiments, one or more blocks of Figure 5 may be executed by one or more other machines not depicted in the figure.

[0078] Method 500 may include the steps of: receiving sensor data (e.g., associated with a processing chamber performing an aging process), and processing the sensor data using a trained machine learning model 190. The trained model may be configured to generate one or more outputs based on the sensor data; the one or more outputs indicate (i) the state of the processing chamber and (ii) a confidence level that the state accordingly represents the actual state of the processing chamber. The state of the processing chamber may indicate whether the processing chamber has recovered and is ready for production operation, whether the processing chamber requires further aging, and / or whether the processing chamber requires further preventative maintenance.

[0079] At block 502, sensor data indicating the environmental state of the processing chamber is received. The state of the processing chamber is measured during the aging process (e.g., when processing a series of substrates). The sensor data can be raw data, or it can be processed using one or more of feature extraction, mechanical models, and / or statistical models to prepare the sensor data for input into a machine learning model. The sensor data may indicate one or more parameters of the processing chamber (e.g., temperature, pressure, vacuum conditions, and spectral data, etc.).

[0080] In some embodiments, the sensor data further includes synthetic data or data constructed from raw sensor data. For example, as described in previous embodiments, various engineering tools can perform feature extraction and / or create combinations of artificial and / or virtual parameters. Feature extractors (e.g., data preparation tool 116 of FIG1) can create various features by performing variable analysis (such as processing control analysis, univariate limit violation analysis, and / or multivariate limit violation analysis) on the raw sensor data. In some embodiments, sensor data in multiple processing chambers and / or processing recipes are normalized to create a similar dataset with a common basis.

[0081] At box 504, sensor data is provided as input to a trained machine learning model; and at box 506, one or more outputs are obtained from the trained machine learning model. At box 508, a first set of values ​​is determined based on the outputs(s) from the machine learning model. The first set of values ​​indicates the likelihood that the associated substrate includes a first processing result that satisfies a threshold condition.

[0082] In some embodiments, the processing logic determines a first quantity based on a first set of values. The first value may indicate a threshold number of substrates processed after a preventative maintenance procedure causes the processing chamber to process subsequent substrates. These substrates include second processing results that meet threshold conditions associated with the substrate processing procedure (e.g., uniformity requirements, critical size criteria, and film thickness criteria, etc.). The threshold number of substrates and subsequent substrates are processed according to the substrate processing procedure. The first quantity indicates how many aged wafers are needed to restore the state of the processing chamber to transition to production operation mode.

[0083] At box 510, a first set of values ​​is used to predict a first test result. The first test result indicates the likelihood that the first substrate processed after processing a series of substrates includes a second processing result that meets a threshold condition. The first test result can be predicted based on a comparison between a first quantity and a second quantity (e.g., the difference between the first quantity and the second quantity). For example, a threshold number of substrates is compared with the current number of substrates to determine whether a sufficient number of processed wafers have been processed to restore the state of the processing chamber after preventative maintenance. In some embodiments, the difference between the identified quantities can be displayed on the GUI (e.g., to indicate over-aging or under-aging of one or more processing chambers).

[0084] In some embodiments, the processing logic determines a recovery curve corresponding to the processing chamber. The recovery curve may include an index distribution indicating the probability that a first substrate processed after processing a series of substrates will result in a second processing outcome that meets a threshold condition. The recovery curve may be used to predict a first test result. The recovery curve may include a cumulative distribution function associated with a set of values ​​received in one or more outputs of a machine learning model.

[0085] At block 512, a first test result may optionally be prepared for presentation on a graphical user interface (GUI). For example, the first test result may include notifications that further aging of the processing chamber is required, aging should be terminated, and / or preventative maintenance of the processing chamber is necessary. At block 514, the operation of the processing chamber may be modified based on the first test result. For example, the processing logic may transmit instructions to one or more processing controllers to modify one or more operations of the processing device (e.g., continue aging, end aging, and begin preventative maintenance procedures, etc.).

[0086] FIG6 is an exemplary illustration of the training phase of a machine learning system according to some embodiments of the present application. According to embodiments of the present application, the machine learning system 170 may use method 600 to perform at least one of training a machine learning model, validating a machine learning model, or testing a machine learning model. In some embodiments, one or more operations of method 600 may be performed by a dataset generator 174 of a server machine 172 as described with respect to FIG1. ​​It may be noted that the components described with respect to FIG1 can be used to illustrate the pattern of FIG6.

[0087] At block 602, the processing logic initializes the training set. At block 604, the processing logic identifies a first data input (e.g., a first training input and a first verification input) including (as described with respect to Figures 1 through 5) a first sensor. The first data input may include data indicating the environmental state of the processing chamber during aging. In some embodiments, the first data input may include any chamber environmental parameters (e.g., temperature, pressure, vacuum conditions, plasma conditions, and spectral conditions, etc.). The first target input may be identified from (e.g., via) old sensor data 144 of historical data 142 stored in data storage 140.

[0088] At block 606, the processing logic identifies a first target output of one or more data inputs (e.g., a first data input). The first target output indicates whether the first substrate processed by the first processing chamber in a first state (e.g., the first data input) according to the first substrate processing procedure results in the first substrate having a first processing result that satisfies a first threshold condition associated with the first substrate processing procedure. The first target input can be identified from sensor data 144 of the historical data 142 (of the data storage 140).

[0089] At block 608, the processing logic optionally generates mapping data indicating the input / output mapping. The input / output mapping (or mapping data) may refer to data inputs (e.g., one or more data inputs as described herein), target outputs of data inputs (e.g., one or more data inputs as described herein), target outputs of data (e.g., where the target output identifies an associated chamber state), and the association between (multiple) data inputs and target outputs.

[0090] At box 610, the processing logic will add the mapping data generated in box 504 to the data set T.

[0091] At block 612, the processing logic branches based on whether the dataset T is sufficient to train, validate, or test at least one of the machine learning models 190. If yes ("Yes" branch), execution proceeds to block 614; if no ("No" branch), execution continues back to block 604. It should be noted that in some embodiments, the sufficiency of the dataset T may be determined simply based on the number of input / output mappings in the dataset; while in some other embodiments, the sufficiency of the dataset T may be determined based on one or more other criteria besides or instead of the number of input / output mappings (e.g., measures of the diversity and accuracy of data examples, etc.).

[0092] At block 610, the processing logic provides a dataset T to train, validate, or test the machine learning model 190. In some embodiments, the dataset T is a training set, which is trained via a training engine 182 provided to the server machine 180. In some embodiments, the dataset T is a validation set, which is validated via a validation engine 184 provided to the server machine 180. In some embodiments, the dataset T is a test set, which is tested via a test engine 186 provided to the server machine 180. For example, in the case of a neural network, the input values ​​(e.g., numerical values ​​associated with the data input) of a given input / output mapping are input to the neural network, and the output values ​​(e.g., numerical values ​​associated with the target output) of the input / output mapping are stored in the output nodes of the neural network. The connection weights in the neural network are then adjusted according to a learning algorithm (e.g., backpropagation, etc.), and this process is repeated for other input / output mappings in the dataset T. Following box 614, the machine learning model (e.g., machine learning model 190) can be at least one of being trained using training engine 182 of server machine 180, verified using verification engine 184 of server machine 180, or tested using test engine 186 of server machine 180. The trained machine learning model can be implemented by the chamber state component 194 (of server machine 192) to identify the chamber state of the chamber (e.g., the probability that the substrate processed in the chamber has a processing result that meets a threshold condition) based on received sensor data.

[0093] In some embodiments, the generated training dataset is used to train a machine learning model and / or a physical model. The trainable model receives chamber sensor data as input. The model can output the state of the chamber (e.g., the probability that a substrate processed within the chamber has a processing result that satisfies a threshold condition) based on the received sensor data. In embodiments, the model is agnostic to the individual processing chambers used in the production mode, the preventative maintenance performed, and / or the processing formulation. Therefore, the model can be generated based on training data (generated from aging processes run on the first processing chamber), and then used for the second processing chamber without performing any transformation learning on the model for adjusting the second processing chamber.

[0094] In one embodiment, the trained machine learning model is a regression model trained using regression. Examples of regression models are regression models trained using linear regression or Gaussian regression. Given known values ​​of variable X, the regression model predicts the value of Y. Regression analysis can be used to train the regression model; regression analysis may include interpolation and / or extrapolation. In one embodiment, least squares is used to estimate the parameters of the regression model. Alternatively, Bayesian linear regression, partial least squares regression, percentage regression, minimum absolute deviation, nonparametric regression, context optimization, and / or distance metric learning can be performed to train the regression model. In one embodiment, the trained machine learning model includes a support vector machine or other types of machine learning models.

[0095] In one embodiment, the trained machine learning model employs ensemble modeling. Ensemble modeling may include training multiple models (e.g., two or more related but different analytical models) and synthesizing the results of each model into (e.g., improving accuracy and reducing from any one machine learning model) a single score or distribution. For example, the trained machine learning model may include a bootstrap aggregation ("Bagging") model. Bagging may include an aggregation phase and a bootstrap phase for each individual model. Bootstrap includes sampling or creating a dataset (e.g., using method 600) in which each sample (e.g., sensor data associated with performing a aging procedure on a processing chamber) is selected and replaced within a broader dataset (e.g., randomly). As previously described, multiple models may be generated for each model based on different subsets of a broader total training dataset (e.g., historical data 142). The multiple models may output synthesized (e.g., aggregated) results to form a total score (e.g., total confidence level) associated with the recognition output (e.g., predicted chamber state) of the model ensemble. In another example, the machine learning model may include a random forest model. Machine learning models utilize multiple decision trees or multiple types of analytical models designed to predict outputs based on different variables and rules. Random forest models can combine multiple decision trees that analyze different sample data, evaluate different factors, or weight common variables by different amounts. The results of multiple decision trees can be transformed into a synthetic result (e.g., simple averaging or via further weighted aggregation) to predict the state of a chamber based on input sensor data.

[0096] In one embodiment, the trained machine learning model is an artificial neural network (also simply referred to as a neural network). The artificial neural network may be, for example, a convolutional neural network (CNN) or a deep neural network. In one embodiment, the processing logic performs supervised machine learning to train the neural network.

[0097] Artificial neural networks generally include feature representation components; these feature representation components have classifier or regression layers that map features to a target output space. For example, convolutional neural networks (CNNs) host multiple layers of convolutional filters. Pooling is performed at lower layers and can solve non-linear problems; typically, a multilayer perceptron is attached above the lower layers, which maps the top-level features extracted by the convolutional layers to decisions (e.g., classification outputs). Neural networks can be deep networks with multiple hidden layers or shallow networks with zero or a few (e.g., 1 to 2) hidden layers. Deep learning is a class of machine learning algorithms that use a series of multilayer non-linear processing units for feature extraction and transformation. Each successive layer uses the output of the previous layer as input. Neural networks can be learned in a supervised (e.g., classification) and / or unsupervised (e.g., pattern analysis) manner. Some neural networks (e.g., deep neural networks) include multilayer structures where different layers learn representations corresponding to different levels of abstraction. In deep learning, each level learns to transform its input data into a slightly more abstract and complex representation.

[0098] Supervised learning can be used to train neural networks. Supervised learning involves feeding a training dataset consisting of labeled inputs into the network, observing its output, defining the error (by measuring the difference between the output and the labeled value), and using techniques such as deep gradient descent and backpropagation to adjust the weights of all layers and nodes in the network to minimize the error. In many applications, repeating this process on many labeled inputs in the training dataset produces a network that can produce the correct output when the inputs differ from those present in the training dataset.

[0099] The trained machine learning model can be periodically or continuously retrained to enable continuous learning and improvement of the trained machine learning model. The model can generate outputs based on inputs, perform actions based on outputs, and measure the results of actions. In some cases, the results of actions are measured within seconds or minutes; in others, measuring the results of actions takes longer. For example, one or more additional processes can be performed before the results of actions can be measured. Actions and the results of actions can indicate whether the output is the correct output and / or the difference between what the output should be and what the output is. Therefore, actions and the results of actions can be used to determine the target output that can be used as a label for sensor measurement. Once the results of actions are determined, the inputs (e.g., sensor data), the output of the trained machine learning model (e.g., chamber state), and the target result (e.g., the target processing result of the substrate) and the actual measurement result (e.g., the measurement processing result of the substrate) can be used to generate new training data items. The new training data items can then be used to further train the trained machine learning model. In some embodiments, this retraining process can be performed on a tool on a manufacturing system (e.g., (multiple) processing chambers).

[0100] Figure 7 illustrates a model training workflow 705 and a model application workflow 717 for chamber monitoring and simulation according to the present application. In an embodiment, the model training workflow 705 may be executed on a server; the server may or may not include a chamber classifier and / or chamber recovery predictor application, and the trained model is provided to the chamber classifier and / or the chamber recovery predictor application that executes the model application workflow 717. The model training workflow 705 and the model application workflow 717 may be executed by processing logic executed by a processor of a computing device (e.g., server 120 of Figure 1). One or more of these workflows 705, 717 may be implemented, for example, by one or more machine learning modules implemented as a processing device and / or other software and / or firmware executed on the processing device.

[0101] Model training workflow 705 is used to train one or more machine learning models (e.g., deep learning models) to perform one or more decisions, predictions, and modifications (etc.) associated with chamber classification and chamber recovery prediction (e.g., determining whether a treated chamber has been properly aged and is ready for production operation). Model application workflow 717 is used to apply one or more trained machine learning models to perform decisions and / or adjustments (etc.) to chamber data (e.g., raw sensor data and synthetic data indicating the state of the treated chamber). One or more machine learning models may receive processing result data (e.g., chamber state and / or recovery curves).

[0102] This document describes various machine learning outputs. A specific number and arrangement of machine learning models are described and shown. However, it should be understood that the number and type of machine learning models used, as well as the arrangement of such models, can be modified to achieve the same or similar end results. Therefore, the arrangements of the machine learning models described and shown are merely examples and should not be construed as limiting.

[0103] In embodiments, one or more machine learning models are trained to perform one or more of the following tasks. Each task may be performed by a separate machine learning model. Alternatively, a single machine learning model may perform each task or a subset of tasks. Additionally or alternatively, different machine learning models may be trained to perform different combinations of tasks. In an example, one or more machine learning models may be trained, wherein the trained machine learning (ML) model is a single shared neural network having multiple shared layers and multiple higher-level different output layers, wherein each output layer outputs different predictions, classifications, and recognitions, etc. One or more trained machine learning models may be trained to perform tasks as follows: a. A chamber classifier – as previously described, capable of receiving and processing various input data; various input data such as sensor data, preprocessed data, synthetic data indicating the state of the processing chamber during aging processes (e.g., processing a series of aged wafers). The chamber classifier may output various values ​​corresponding to various aged wafers, indicating the probability that substrate processing under corresponding chamber conditions results in a processing outcome that satisfies threshold conditions associated with the substrate processing (e.g., substrate and / or manufacturing formulation). b. Chamber Recovery Predictor – The chamber recovery predictor receives data from a chamber classifier and generates a chamber recovery curve associated with the recovery of a processed chamber, which is associated with preventative maintenance and / or aging treatments. The chamber recovery predictor determines preventative maintenance indicators that indicate the likelihood of a chamber meeting conditions for transitioning from a recovery operation mode to a production operation mode. For example, the chamber recovery predictor can identify when chamber conditions meet threshold conditions such that substrates treated under those chamber conditions result in substrates having a processing outcome that meets threshold conditions associated with one or more processing procedures. In some embodiments, the chamber recovery predictor outputs a preventative maintenance indicator value associated with the likelihood of a chamber meeting threshold conditions to transition to a production operation mode.

[0104] 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. Artificial neural networks generally consist of feature representation components with classifier or regression layers that map features to a desired output space. For example, a convolutional neural network (CNN) hosts multiple layers of convolutional filters. Pooling is performed at lower layers and can solve non-linear problems; typically, multiple perceptrons are attached above the lower layers, which map the top-level features extracted by the convolutional layers to a decision (e.g., classification output). Deep learning is a class of machine learning algorithms that uses a series of multiple layers of non-linear processing units for feature extraction and transformation. Each successive layer uses the output of the previous layer as input. Deep neural networks can be learned in a supervised (e.g., classification) and / or unsupervised (e.g., pattern analysis) manner. Deep neural networks consist of a multi-layered structure where different layers learn different levels of abstraction corresponding to different levels of representation. In deep learning, each layer learns to transform its input data into a slightly more abstract and complex representation. For example, in plasma processing tuning, the initial input can be chamber data (e.g., raw sensor chamber measurement parameters, such as temperature, pressure, and vacuum conditions within the chamber); the second layer can comprise feature data associated with parameters related to the substrate processing procedure or processing recipe; and the third layer can include historical thresholds and chamber states under similar conditions. It is noteworthy that deep learning processes can learn on their own which features are best placed at which level. The "depth" in "deep learning" refers to the number of layers involved in data transformation. More precisely, deep learning systems have a considerable reward distribution path (CAP) depth. CAP is a transformation chain from input to output. CAP describes the underlying causal relationship between the input and output. For feedforward neural networks, the CAP depth can be the network depth or the number of hidden layers plus one. For recurrent neural networks, where a signal may propagate through multiple layers, the CAP depth can be infinite.

[0105] In one embodiment, one or more machine learning models are recurrent neural networks (RNNs). An RNN is a type of neural network that includes time-dependent memory. An RNN can learn an input-output mapping that depends on current and past inputs. The RNN will resolve past and future sensor and / or processing result measurements and make predictions based on this continuous sensor information (e.g., sensor data indicating constantly changing conditions in the processing chamber). The RNN can be trained using a training dataset to generate a fixed number of outputs (e.g., to determine a set of substrate processing rates, and to determine chamber conditions, production mode requirements, and condition thresholds). One type of RNN that can be used is a long short-term memory (LSTM) neural network.

[0106] Supervised learning can be used to train neural networks. Supervised learning involves feeding a training dataset consisting of labeled inputs into the network, observing its output, defining the error (by measuring the difference between the output and the labeled value), and using techniques such as deep gradient descent and backpropagation to adjust the weights of all layers and nodes in the network to minimize the error. In many applications, repeating this process on many labeled inputs in the training dataset produces a network that can produce the correct output when the inputs differ from those present in the training dataset.

[0107] For the model training workflow 705, a training dataset should be formed using a dataset containing hundreds, thousands, tens of thousands, hundreds of thousands, or more chamber data 710 (e.g., sensor data and synthetic data indicating the state of the relevant processing chambers). In embodiments, the training dataset may also include associated processing result data 712 (e.g., measurement parameters of the substrate (e.g., critical dimensions, uniformity requirements, and film thickness results, etc.)) used to form the training dataset, wherein each data point may include various labels or classifications of one or more types of useful information. For example, each case may include data indicating that one or more processing chambers are undergoing aging, and associated processing results of the substrate evaluated during and / or after aging. This data may be processed to generate one or more training datasets 736 for training one or more machine learning models. Machine learning models may be trained, for example, to automate a chamber classifier (e.g., chambers meet conditions for operation in production mode) and / or predict the threshold number of substrates that will operate in production mode when a chamber meets threshold conditions.

[0108] In one embodiment, generating one or more training datasets 736 includes performing a chamber recovery process 200A by performing preventative maintenance 202 and aging 206 and evaluating sample tests 208 to determine whether a chamber meets a threshold condition (e.g., block 210). One or more labels can be used for various iterations of the chamber recovery process 200A to identify sample tests 208 and measured chamber conditions as meeting the threshold condition to initialize a production mode, initialize further aging, and / or initialize further preventative maintenance corresponding to the processed chamber. The labels used may depend on what a particular machine learning model will be trained to do. In some embodiments, as described in other embodiments, chamber data, processing results, and / or sample test results may be represented as vectors, and processing rates may be represented as one or more matrices.

[0109] To perform training, the processing logic inputs (multiple) training datasets 736 into one or more untrained machine learning models. The machine learning models can be initialized before the first input is fed into them. The processing logic trains (multiple) untrained machine learning models based on the (multiple) training datasets to generate one or more trained machine learning models that perform the various operations described above.

[0110] Training can be performed by inputting one or more of the chamber data 710 and the processing result data 712 into the machine learning model one at a time.

[0111] Machine learning models process inputs to generate outputs. Artificial neural networks consist of an input layer composed of values ​​from data points. The next layer is called a hidden layer, and each node in the hidden layer receives one or more input values. Each node contains parameters (e.g., weights) applied to the input values. Thus, each node essentially feeds the input values ​​into a multivariable function (e.g., a nonlinear mathematical transformation) to produce an output value. The next layer may be another hidden layer or an output layer. In either case, nodes in the next layer receive output values ​​from nodes in the previous layer; each node applies weights to these values ​​and then generates its own output value. This can be done at each layer. The final layer is the output layer, where there is a node for each class, prediction, and / or output that the machine learning model can produce.

[0112] Therefore, the output may include one or more predictions or inferences. For example, the output prediction or inference may include one or more modifications to plasma processing data (e.g., modifications to one or more plasma exposure periods). The processing logic may process the substrate using an updated recipe (e.g., using identified modifications) and receive an updated thickness distribution. The processing logic may compare the updated thickness distribution with a target thickness distribution and determine whether a threshold criterion is met (e.g., the thickness value measured across the wafer surface falls within a target threshold window). The processing logic determines an error (i.e., classification error or prediction error) based on the difference between the identified chamber state or recovery prediction and the evaluated processing result (e.g., sample test result). The processing logic adjusts the weights of one or more nodes in the machine learning model based on the error. An error term or increment may be determined for each node in the artificial neural network. Based on this error, the artificial neural network adjusts one or more parameters (weights of one or more inputs to the node) of one or more of its nodes. The parameters may be updated using backpropagation, such that the nodes of the highest layer are updated first, followed by the nodes of the next layer, and so on. Artificial neural networks consist of multiple layers of "neurons," where each layer receives input values ​​from neurons in previous layers. The parameters of each neuron include weights associated with the values ​​received from each neuron in previous layers. Therefore, adjusting the parameters can include adjusting the weights assigned to each input of one or more neurons in one or more layers of the artificial neural network.

[0113] Once the model parameters have been optimized, model validation can be performed to determine whether the model has been improved and to determine the current accuracy of the deep learning model. After one or more rounds of training, the processing logic can determine whether a stopping criterion has been met. The stopping criterion can be a target accuracy level, the number of targets processed from the training dataset, the target change of parameters at one or more previous data points, a combination of the above, and / or other criteria. In one embodiment, the stopping criterion is met when at least a minimum number of data points have been processed and at least a threshold accuracy has been reached. The threshold accuracy can be, for example, 70%, 80%, or 90% accuracy. In one embodiment, the stopping criterion is met if the accuracy of the machine learning model has stopped improving. If the stopping criterion is not met, further training is performed. If the stopping criterion has been met, training can be completed. Once the machine learning model has been trained, a retained portion of the training dataset can be used to test the model.

[0114] Once one or more trained machine learning models 738 are generated, they can be stored in model storage 745 and added to the substrate processing rate determination and / or processing adjustment application. The substrate processing rate determination and / or processing adjustment application can then use one or more trained ML models 738 and additional processing logic to implement an automatic mode; in the automatic mode, manual user input is minimized or, in some cases, even eliminated.

[0115] For the model application workflow 717, according to one embodiment, input data 762 may be input into a processing rate determiner 767; the processing rate determiner 767 may include a trained machine learning model. Based on the input data 762, the processing rate determiner 767 outputs information indicating the likelihood that the associated substrate includes a first processing result that satisfies a threshold condition of the substrate processing procedure.

[0116] According to one embodiment, input data 762 may be input to a chamber recovery predictor 764; the chamber recovery predictor 764 may include a trained machine learning model. Based on the input data 762, the chamber recovery predictor 764 outputs a chamber recovery curve, a threshold number of aged wafers required for the chamber to meet threshold conditions to operate in a production operation mode, and / or a first test result indicating the likelihood that a first substrate processed after processing a series of substrates will include a processing result that meets the threshold conditions (e.g., which are associated with the substrate processing formulation).

[0117] FIG8 depicts a block diagram of an exemplary computing device 800 operating in one or more modes according to this application. In various illustrative examples, various components of the computing device 800 may represent various components of the client device 150, metering system 110, server 120, data storage 140, manufacturing system 102 and machine learning system 170 as shown in FIG1.

[0118] The exemplary computing device 800 may be connected to other computer devices in a LAN, intranet, extranet, and / or internet. The computing device 800 may operate as a server in a client-server network environment. The computing device 800 may be a personal computer (PC), a set-top box (STB), a server, a network router, a switch, or a bridge, or any device capable of executing (sequentially or otherwise) a set of instructions that specifically specify the actions to be taken by the device. Furthermore, although only a single exemplary computing device has been described, the term "computer" should also be considered as any collection of computers that individually or jointly execute a set (or more) of instructions to perform any one or more methods discussed herein.

[0119] An exemplary computing device 800 may include a processing device 802 (also referred to as a processor or CPU), main memory 804 (e.g., read-only memory (ROM), flash memory, dynamic random access memory (DRAM) such as synchronous DRAM (SDRAM), etc.), static memory 806 (e.g., flash memory and static random access memory (SRAM), etc.) and auxiliary memory (e.g., data storage device 818), which can communicate with each other through bus 830.

[0120] Processing device 802 represents one or more general-purpose processing devices, such as microprocessors and central processing units, etc. More specifically, processing device 802 may be a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor implementing other instruction sets, or a processor implementing a combination of instruction sets. Processing device 802 may also be one or more special-purpose processing devices, such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), and network processors, or the like. According to one or more embodiments of this application, processing device 802 may be configured to execute instructions implementing methods 500 and 600 shown in Figures 5 and 6.

[0121] The exemplary computing device 800 may further include a network interface device 708, which may be communicatively coupled to a network 820. The exemplary computing device 800 may further include a video display 810 (e.g., a liquid crystal display (LCD), a touch screen, or a cathode ray tube (CRT)), an alphanumeric input device 812 (e.g., a keyboard), a cursor control device 814 (e.g., a mouse), and a sound signal generation device 816 (e.g., a speaker).

[0122] The data storage device 818 may include a machine-readable storage medium (or, more specifically, a non-transitory machine-readable storage medium) 828 thereon storing one or more sets of executable instructions 822. According to one or more embodiments of this application, the executable instructions 822 may include executable instructions associated with the execution methods 500, 600 shown in Figures 5 and 6.

[0123] The executable instructions 822 may also reside wholly or at least partially in the main memory 804 and / or the processing device 802 during execution by the exemplary computing device 800; the main memory 804 and the processing device 802 also constitute a computer-readable storage medium. The executable instructions 822 may be further transmitted or received over a network via a network interface device 808.

[0124] Although Figure 8 shows a computer-readable storage medium 828 as a single medium, the term "computer-readable storage medium" should be understood to include a single medium or multiple media (e.g., a centralized or distributed database, and / or associated cache and server) that store one or more sets of operating instructions. The term "computer-readable storage medium" should also be understood to include any medium capable of storing or encoding a set of instructions for machine execution; such instructions cause the machine to perform any one or more of the methods described herein. Therefore, the term "computer-readable storage medium" should be understood to include, but is not limited to, solid-state memory, optical and magnetic media.

[0125] Some parts of the above [implementation] are presented based on algorithms and symbolic representations of operations on data bits in computer memory. These algorithmic descriptions and representations are means used by those skilled in the art of data processing to most effectively communicate their work to others skilled in the art. Algorithms are generally considered herein as self-consistent sequences of steps that lead to desired results. Steps are those that require physical operations on physical quantities. Typically, but not necessarily, these quantities take the form of electrical or magnetic signals that can be stored, transmitted, combined, compared, and otherwise manipulated. Sometimes, primarily for common reasons, it has proven convenient to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, or the like.

[0126] However, it should be remembered that all these and similar terms will be associated with appropriate physical quantities and are merely convenient notations applied to those quantities. Unless otherwise specifically stated, as will be apparent from the following discussion, it can be understood that throughout this specification, the use of terms such as “identify,” “determine,” “store,” “adjust,” “cause,” “return,” “compare,” “create,” “stop,” “load,” “copy,” “throw,” “replace,” and “execute,” or similar terms, refers to the actions and processes of a computer system or similar electronic computing device (which processes and converts data represented as physical (electronic) quantities in computer system registers and memory into other data similarly represented as physical quantities in computer system memory or registers or other such information storage, transmission, or display devices).

[0127] Examples of this application also relate to apparatus for performing the methods described herein. This apparatus may be specifically constructed for the desired purpose, or it may be a general-purpose computer system selectively programmed by a computer program stored in a computer system. Such a computer program may be stored in a computer-readable storage medium; the computer-readable storage medium may be, for example, but not limited to, any type of disc including optical discs, CD-ROMs and ROMs, random access memory (RAM), erasable programmable read-only memory (EPROM), electronically erasable programmable read-only memory (EEPROM), magnetic disk storage media, optical storage media, flash memory devices, other types of machine-accessible storage media, or any type of media suitable for storing electronic instructions; each of which is coupled to a computer system bus.

[0128] The methods and displays presented herein are not inherently associated with any particular computer or other device. Various general-purpose systems may be used with the programs taught herein; or it may prove convenient to construct more specialized devices to perform the required method steps. The necessary structures for various such systems will appear in the description below. Furthermore, the scope of this application is not limited to any particular programming language. It should be understood that the teachings of this application may be implemented using a variety of programming languages.

[0129] It should be understood that the above description is intended to be illustrative and not restrictive. Many other implementation examples will be apparent to those skilled in the art after reading and understanding the above description. Although specific examples are described in this application, it will be appreciated that the systems and methods of this application are not limited to the examples described herein, but can be implemented with modifications within the scope of the claims. Therefore, the specification and drawings should be considered illustrative and not restrictive. Thus, the scope of this application should be determined with reference to the appended claims and the full scope of equivalents enjoyed by such claims. [Simplified Explanation of the Diagram]

[0008] The nature and implementation of this application will be more fully understood from the detailed description and accompanying drawings given below (the drawings are intended to illustrate the nature and implementation by way of example rather than limitation).

[0009] Figure 1 is a block diagram illustrating an exemplary system architecture in which embodiments of the present application may operate.

[0010] Figures 2A and 2B are block diagrams illustrating the chamber recovery processes 200A to 200B in which embodiments of the present application may operate.

[0011] Figure 3 is a block diagram showing a chamber monitoring and maintenance system 200 in which an embodiment of the present application may operate.

[0012] Figure 4 depicts a chamber recovery curve according to some embodiments of this application.

[0013] Figure 5 depicts a flowchart of an exemplary method for monitoring and predicting the conditions of a processing chamber according to some embodiments of this application.

[0014] Figure 6 is an exemplary illustration of the training phase of a machine learning system according to some embodiments of this application.

[0015] Figure 7 illustrates the model training workflow and model application workflow for chamber monitoring and simulation according to the present application.

[0016] FIG8 is a block diagram of an exemplary computing device operating in one or more modes according to the present application. [Biomaterial Storage]

[0131] Domestic storage information (please note in order of storage institution, date, and number): None. International storage information (please note in order of storage country, institution, date, and number): None.

Claims

1. A method comprising the steps of: receiving, by means of a processing device, sensor data indicating a state of an environment of a processing chamber, the processing chamber processing a series of substrates according to a substrate processing procedure, the series of substrates being processed after a preventive maintenance procedure performed on the processing chamber; determining, by means of the processing device, a first set of values ​​based on the sensor data, each of the first set of values ​​being associated with one of the series of substrates, and the first set of values ​​indicating a probability that an associated substrate includes a first processing result satisfying a threshold condition of the substrate processing procedure; predicting, by means of the processing device, a first test result based on the first set of values, the first test result indicating a probability that a first substrate processed after processing the series of substrates includes a second processing result satisfying the threshold condition; and performing, by means of the processing device, at least one of a) preparing the first test result for presentation on a graphical user interface (GUI) or b) modifying the processing chamber based on the first test result.

2. The method of claim 1, further comprising the steps of: determining a first quantity based on the first set of values, the first quantity indicating a threshold number of substrates processed after the preventive maintenance procedure resulting in the processing chamber processing a subsequent substrate having a second processing result that satisfies the threshold condition, wherein the threshold number of substrates and the subsequent substrate are processed according to the substrate processing procedure.

3. The method of claim 2, wherein the series of substrates includes a second number of substrates, the method further comprising the steps of: comparing the first number with the second number, wherein the first test result is predicted based on a comparison between the first number and the second number.

4. The method as described in claim 3 further includes the steps of: determining a first difference between the first quantity and the second quantity; and preparing to present a notification indicating the first difference on the GUI.

5. The method as described in claim 1, further comprising the step of: determining, based on the first test result, that the processing chamber meets one or more conditions of a production operation mode; 6. The method of claim 1, further comprising the steps of: determining a recovery curve corresponding to the processing chamber, the recovery curve including an index distribution indicating a change in the probability that the first substrate processed after processing the series of substrates includes the second processing result satisfying the threshold condition, wherein the index distribution is used to further predict the first test result.

7. The method as described in claim 6, wherein the recovery curve further includes a cumulative distribution function associated with the first set of values.

8. The method of claim 1, further comprising the steps of: using the sensor data as input to a machine learning model; and obtaining one or more outputs of the machine learning model, the one or more outputs indicating the first set of values.

9. The method of claim 1, wherein the machine learning model comprises at least one of a bootstrap aggregation model, a random forest model, a decision tree model, or a partial least squares regression (PLS) model.

10. A method for training a machine learning model to predict a condition of a processing chamber, the method comprising the steps of: generating training data for the machine learning model, wherein the step of generating the training data includes the following steps: Identify a first training input having first sensor data indicating a first state of a first processing chamber after a preventive maintenance procedure; and identify a first target output for the first training input, wherein the first target output indicates whether a first substrate processed by the first processing chamber in the first state according to a first substrate processing procedure results in the first substrate having a first processing result that satisfies a first threshold condition associated with the first substrate processing procedure; and provide the training data to train the machine learning model on (i) a set of training inputs including the first training input and (ii) a set of target outputs including the first target output, wherein the trained machine learning model will receive a new input having new sensor data indicating a new state of a new processing chamber, the new output indicating whether a new substrate processed by the new processing chamber in the new state according to the first substrate processing procedure results in the new substrate having a new processing result that satisfies the threshold condition.

11. The method of claim 10, wherein the first sensor data includes one or more measurements acquired after processing a first series of substrates in the first processing chamber.

12. The method of claim 10, wherein the first sensor data includes one or more measurements acquired before the first processing chamber processes a first series of substrates.

13. The method of claim 10, wherein each training input in the set of training inputs is mapped to a target output in the set of target outputs.

14. A non-transitory machine-readable storage medium comprising instructions that, when executed by a processing device, cause the processing device to perform operations including: receiving sensor data indicating a state of an environment of a processing chamber, the processing chamber processing a series of substrates according to a substrate processing program, the series of substrates being processed after a preventive maintenance program performed on the processing chamber; determining a first set of values ​​based on the sensor data, each of the first set of values ​​being associated with one of the series of substrates, and the first set of values ​​indicating a probability that an associated substrate includes a first processing result satisfying a threshold condition of the substrate processing program; predicting a first test result based on the first set of values, the first test result indicating a probability that a first substrate processed after processing the series of substrates includes a second processing result satisfying the threshold condition; and performing at least one of a) preparing the first test result for presentation on a graphical user interface (GUI) or b) changing the processing chamber based on the first test result.

15. The non-transitory machine-readable storage medium as described in claim 14, the operation further includes the following operation: determining a first quantity based on the first set of values, the first quantity indicating a threshold number of substrates processed after the preventive maintenance procedure resulting in the processing chamber processing a subsequent substrate having a second processing result that satisfies the threshold condition, wherein the threshold number of substrates and the subsequent substrate are processed according to the substrate processing procedure.

16. The non-transitory machine-readable storage medium as claimed in claim 15, wherein the series of substrates includes a second number of substrates, wherein the operations further include the following operations: comparing the first number with the second number, wherein a first test result is predicted based on a comparison between the first number and the second number.

17. The non-transitory machine-readable storage medium as described in claim 16, the operations further include: determining a first difference between the first quantity and the second quantity; and preparing to present a notification indicating the first difference on the GUI.

18. The non-transitory machine-readable storage medium as described in claim 14, the operation further comprising: determining a recovery curve corresponding to the processing chamber, the recovery curve including an index distribution indicating a change in the probability that the first substrate processed after processing the series of substrates includes the second processing result satisfying the threshold condition, wherein the index distribution is used to further predict the first test result.

19. The non-transitory machine-readable storage medium as described in claim 14, the operations further comprising: using the sensor data as input to a machine learning model; and obtaining one or more outputs of the machine learning model, the one or more outputs indicating the first set of values.

20. The non-transitory machine-readable storage medium as described in claim 19, wherein the machine learning model includes at least one of a bootstrap aggregation model, a random forest model, a decision tree model, or a partial least squares regression (PLS) model.

Images