Inter-batch control at manufacturing system using machine learning

Through machine learning technology, inter-batch control is carried out in the manufacturing system, and process formula is predicted and adjusted, the problems of substrate characteristics offset and change in the prior art are solved, and manufacturing efficiency and product quality are improved.

CN119096204APending Publication Date: 2024-12-06APPLIED MATERIALS INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202380036676.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-04-27
Filing Date
2023-04-26
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

Existing manufacturing systems are difficult to effectively tune or modify process formulas in inter-batch process control, resulting in substrate characteristics offset, variation and variability, affecting manufacturing efficiency and product quality.

Method used

Using machine learning technology, by identifying data associated with the substrate, training the model to predict deviations in the metrological measurement values, determine modifications of the process formulation to optimize the substrate process and correct characteristic offsets.

Benefits of technology

The processing efficiency and product quality of the manufacturing system are improved, and the substrate characteristics can be detected and corrected earlier, additional processing operations are reduced, and the total delay and processing volume are reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119096204A_ABST
    Figure CN119096204A_ABST
Patent Text Reader

Abstract

First data associated with a first process performed for a first layer of a substrate is identified. The first layer will be further processed according to a second process. The first data is provided as input to a machine learning model trained to predict metrological measurements for layers of the substrate at the manufacturing system. An offset of the first set of metrology measurements for the first layer deviating from the target value after completion of the first process and / or the second process is determined. A modification to the recipe of the second process is determined in view of the determined offset and second data associated with a second substrate layer previously processed at the manufacturing system. The second process is updated based on the determined one or more modifications.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] Embodiments of the present disclosure relate generally to manufacturing systems and, more particularly, to batch-to-batch control at manufacturing systems using machine learning. Background Art

[0002] As the size of electronic components continues to shrink, substrate processing complexity continues to increase. Some techniques for manufacturing substrates may involve multiple different processes, with some advanced techniques (e.g., plasma etching) involving twenty or even more different processes. Multiple process control variables of a substrate process may affect the characteristics of a corresponding substrate after the substrate process is completed. Run-to-run (R2R) process control refers to techniques for modifying a process recipe between batches in order to minimize drift, variation, and / or variability of a substrate. A process control system may modify a process recipe (e.g., to optimize a process, cause a characteristic of a corresponding substrate to correspond to a target characteristic, etc.) by modifying or tuning settings associated with one or more process control variables associated with the process recipe. It may be difficult for a process control system to determine which process control variable settings to modify and / or how many process control variable settings to modify in order to optimize a corresponding process and / or cause a substrate characteristic to correspond to a target characteristic. Summary of the invention

[0003] Some embodiments described include a method for batch-to-batch (R2R) control at a manufacturing system using machine learning. The method includes identifying first data associated with a first process performed at the manufacturing system for a first layer of a substrate. The first layer of the substrate is to be further processed at the manufacturing system according to a second process. The method further includes providing the first data as an input to a machine learning model. The machine learning model is trained to predict metrology measurements for one or more layers of the substrate processed at the manufacturing system, and an offset of the predicted metrology measurements from a target metrology measurement. The method further includes determining, based on one or more outputs of the machine learning model, an offset by which a first set of metrology measurements for the first layer of the substrate deviates from a target set of metrology measurements after completing at least one of the first process or the second process. The method further includes determining one or more modifications to a process recipe for the second process in view of the determined offset and second data associated with the second layer of the substrate. The second layer of the substrate was previously processed at the manufacturing system according to a third process. The method further includes updating the process recipe for the second process based on the determined one or more modifications.

[0004] In some embodiments, a system includes a memory and a processing device coupled to the memory. The processing device is used to identify first data associated with a first process performed on a first layer of a substrate at a manufacturing system. The first layer of the substrate is to be further processed at the manufacturing system according to a second process. The processing device is further used to provide the first data as an input to a machine learning model. The machine learning model is trained to predict metrology measurements for one or more layers of the substrate processed at the manufacturing system, and the offset of the predicted metrology measurements from target metrology measurements. The processing device is further used to determine, based on one or more outputs of the machine learning model, an offset by which a first set of metrology measurements for the first layer of the substrate deviates from a target set of metrology measurements after completing at least one of the first process or the second process. The processing device is further used to determine one or more modifications to a process recipe for the second process in view of the determined offset and second data associated with the second layer of the substrate. The second layer of the substrate was previously processed at the manufacturing system according to a third process. The processing device is further used to update the process recipe for the second process based on the determined one or more modifications.

[0005] In some embodiments, a non-transitory computer-readable storage medium includes instructions that, when executed by a processing device, cause the processing device to identify first data associated with a first process performed on a first layer of a substrate at a manufacturing system. The first layer of the substrate is to be further processed at the manufacturing system according to a second process. The processing device is further configured to provide the first data as an input to a machine learning model. The machine learning model is trained to predict metrology measurements for one or more layers of the substrate processed at the manufacturing system, and an offset of the predicted metrology measurements from a target metrology measurement. The processing device is further configured to determine, based on one or more outputs of the machine learning model, an offset by which a first set of metrology measurements for the first layer of the substrate deviates from a target set of metrology measurements after completing at least one of the first process or the second process. The processing device is further configured to determine one or more modifications to a process recipe for the second process in view of the determined offset and second data associated with a second layer of the substrate. The second layer of the substrate was previously processed at the manufacturing system according to a third process. The processing device is further configured to update the process recipe for the second process based on the determined one or more modifications. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] The present disclosure is illustrated by way of example and not by way of limitation in the figures of the accompanying drawings, in which like references indicate similar components. It should be noted that different references to "an" or "one" embodiment in the present disclosure are not necessarily to the same embodiment, and such reference means at least one.

[0007] Figure 1An illustrative system architecture in accordance with aspects of the present disclosure is depicted.

[0008] Figure 2 is a block diagram of an example batch-to-batch (R2R) control engine according to aspects of the present disclosure.

[0009] Figure 3 is a flow chart of a method for performing R2R control at a manufacturing system using machine learning according to aspects of the present disclosure.

[0010] Figure 4 An example of using machine learning for R2R control at a manufacturing system according to aspects of the present disclosure is depicted.

[0011] Figure 5 is a flowchart of a method for training a machine learning model according to aspects of the present disclosure.

[0012] Figure 6 A block diagram of an illustrative computer system that operates in accordance with one or more aspects of the present disclosure is depicted. DETAILED DESCRIPTION

[0013] Embodiments described herein provide systems and methods for batch-to-batch (R2R) control at a manufacturing system using machine learning. In some cases, a substrate may undergo a large number of processes before it can be included in a final product. For example, lithography involves depositing a multilayer material onto a substrate surface and using the deposited layers to form a highly complex pattern (e.g., by etching away portions of one or more deposited layers). Before a complex pattern is formed on a substrate and the substrate can be included in a final product, the substrate may undergo multiple substrate processes (e.g., a process for depositing each corresponding layer onto a substrate, a process for etching away portions of one or more layers from the substrate surface, etc.). As the size of electronic components continues to shrink, the complexity of the pattern formed on the substrate increases, which increases the overall substrate processing complexity. For example, for some integrated circuits, before a target pattern is formed on a substrate surface, the substrate may undergo up to 50 or more lithography cycles (e.g., one or more etching processes before and / or after one or more deposition processes, etc.).

[0014] The manufacturing system may include process control subsystems (also referred to herein as process control systems or simply process controllers) that are configured to tune and / or modify settings of processes performed on a substrate (or a batch or group of substrates) so as to optimize the corresponding process and / or cause the characteristics of the substrate (or a batch of substrates) to correspond to target characteristics. Batch-to-batch (R2R) process control refers to techniques for tuning or modifying settings of a process recipe for a substrate process to minimize offsets, variations, and / or variability of the substrate. In some cases, characteristics of a substrate after an initial substrate process is completed may affect characteristics of a substrate after a subsequent substrate process is completed. For example, multiple deposition processes may be performed to deposit material layers on a substrate surface (e.g., according to the lithography cycle described above). Characteristics of an initial layer deposited on a substrate surface may affect characteristics of subsequent layers deposited on the initial layer. In addition, before one or more subsequent layers are deposited on the initial layer, one or more processes may be performed on the initial layer to prepare for subsequent layer processes. For example, a chemical mechanical polishing process may be performed on the initial layer to prepare the initial layer for deposition of subsequent layers. Such processes are referred to herein as upstream processes.

[0015] As indicated above, the process control subsystem may tune and / or modify settings for a substrate process. Because each process performed on a substrate may affect the characteristics of the substrate, it is difficult for the process control subsystem to identify which settings should be tuned or modified and / or how each setting should be tuned or modified. In some systems, metrology data for the substrate may be collected after the process is completed, and the process control subsystem may decide which settings to modify and / or how to tune or modify settings for subsequent processes performed on the substrate based on the collected metrology data. However, collecting metrology data for each substrate or even a portion of the substrates in a batch may take a significant amount of time, which may reduce the overall throughput and increase the overall latency of the manufacturing system. Furthermore, conventional process control subsystems do not consider how the characteristics of the substrate are affected by upstream processes performed on the substrate that prepare the substrate for subsequent substrate processes (e.g., subsequent deposition processes, subsequent etching processes, etc.). Conventional process control systems do not detect variability and / or defects in substrate characteristics caused by upstream processes until after the substrate process is performed and metrology data is collected for the substrate. Thus, conventional process control systems do not tune or modify the settings of a substrate process in order to counteract changes and / or defects in substrate characteristics caused by upstream processes before initiating the substrate process, and the changes or defects may be present on the substrate after the substrate process is completed. In some cases, such systems may perform additional processing operations to correct such changes or defects. In other cases, such changes or defects cannot be corrected, and the substrate is thus unusable. Correcting the changes or defects by additional processing operations and / or removing unusable substrates from the manufacturing system reduces the overall throughput and efficiency and increases the overall latency of the manufacturing system.

[0016] Aspects of the present disclosure address the deficiencies of conventional techniques by providing systems and methods for batch-to-batch (R2R) control at a manufacturing system using machine learning. In some embodiments, a substrate (or a batch or group of substrates) may undergo multiple substrate processes (e.g., according to lithography techniques) at a manufacturing system. In some embodiments, each process may correspond to a corresponding layer of a substrate. For purposes of explanation and illustration only, a substrate process (e.g., for depositing one or more layers, etching away portions of one or more layers, etc.) performed on a layer of a substrate is referred to herein as a current layer process. Such a layer is referred to herein as a current layer. A substrate process previously performed on a layer of a substrate (e.g., before performing the current layer process) is referred to herein as a previous layer process. Such a layer is referred to herein as a previous layer or initial layer. A process performed to prepare a substrate for a current layer process is referred to herein as an upstream process. In some embodiments, examples of upstream processes may include polishing processes (e.g., chemical mechanical polishing processes, etc.), roughening processes, etching processes, deposition processes, and the like.

[0017] In some embodiments, the process control system may use machine learning techniques to determine one or more modifications to the current layer process to be performed on the substrate. For example, after completing the upstream process (e.g., and before starting the current layer process), the process control system may identify data associated with the upstream process. In some embodiments, the upstream process data may include data associated with one or more operations and / or settings associated with a process recipe of the upstream process, data collected by one or more sensors of the manufacturing system before, during, or after performing the upstream process, and the like. The process control system may provide the upstream process data as input to a machine learning model that is trained to predict metrology measurements for one or more layers of a substrate processed according to the upstream process and an offset of the predicted metrology measurements from a target metrology measurement for the substrate after the upstream process. In some embodiments, the machine learning model may be, or may otherwise correspond to, a variable recursive autoencoder model. In response to providing the upstream process data as input to the machine learning model, the process control system may obtain one or more outputs of the machine learning model and may determine an offset of the metrology measurement of the substrate from the target metrology measurement based on the obtained outputs. In some embodiments, the determined offset may correspond to a first correction factor to be applied by the process control system to tune or modify settings associated with a process recipe for the current layer process.

[0018] In some embodiments, metrology data may be collected (e.g., using a metrology device) after performing a previous layer process for a substrate. The process control system may determine an offset of the collected metrology data from a target metrology measurement value for a substrate after the previous layer process by comparing the collected metrology data with a target metrology measurement value. In some embodiments, the determined offset may correspond to a second correction factor that will be applied by the process control system to tune or modify settings associated with a process recipe for a current layer process. In additional or alternative embodiments, additional metrology data may be collected (e.g., using a metrology device) after a process recipe for a current layer process is performed for other substrates at a manufacturing system. As described above, the process control system may determine an offset of the additional metrology data from a target metrology measurement value for a substrate after the current layer process. As described above, the determined offset may correspond to a third correction factor that will be applied by the process control system to tune or modify settings associated with a process recipe for a current layer process.

[0019] In some embodiments, the process control system may determine one or more modifications to a process recipe for a current layer process in view of the first correction factor, the second correction factor, and / or the third correction factor, and may apply the one or more modifications to the process recipe. After processing the substrate according to the current layer process recipe, the process control system may identify metrology data collected for the substrate and may use the identified metrology data to further tune or modify settings associated with the current layer process recipe for future substrates processed at the manufacturing system. Further details on determining one or more modifications to the current layer process recipe and performing the current layer process are provided herein.

[0020] Embodiments of the present disclosure provide techniques for optimizing substrate processes and / or controlling the characteristics of a substrate after a substrate process using machine learning. Embodiments of the present disclosure utilize machine learning techniques to determine the offset of a metrology measurement value from a target metrology measurement value after an upstream process without removing the substrate from the manufacturing system to obtain such a measurement value. Thus, the total throughput of the manufacturing system is increased and the total latency of the manufacturing system is reduced. In addition, embodiments of the present disclosure enable the process control system to consider changes in the substrate caused by the upstream process when deciding whether to tune or modify the settings of the current layer process and / or how to tune or modify the settings of the current layer process. Thus, changes and / or defects in substrate characteristics can be detected and corrected earlier and after a smaller number of substrate processes, which further increases the total throughput and efficiency of the manufacturing system and further reduces the total latency of the manufacturing system.

[0021] Figure 1An illustrative system architecture 100 according to aspects of the present disclosure is depicted. In some embodiments, the system architecture 100 may be included as part of a manufacturing system for processing substrates. The system architecture 100 may include one or more client devices 120, manufacturing equipment 124, metrology equipment 128, a prediction server 112 (e.g., for generating prediction data, providing model adaptation, using a knowledge base, etc.), a computing system 150, and a data storage 140. The prediction server 112 may be part of the prediction system 110. The prediction system 110 may further include server machines 170 and 180. The manufacturing equipment 124 may include a sensor configured to collect data of a substrate being processed at the manufacturing system. In some embodiments, the manufacturing equipment 124 and the sensor may be part of a sensor system that includes a sensor server (e.g., a field service server (FSS) at a manufacturing facility) and a sensor identifier reader (e.g., a front opening unified pod (FOUP) radio frequency identification (RFID) reader of the sensor system). In some embodiments, the metrology device 128 may be part of a metrology system that includes a metrology server (e.g., a metrology database, a metrology folder, etc.) and a metrology identifier reader (e.g., a FOUP RFID reader of the metrology system). Figure 1 Although depicted as a separate component in FIG. 1 , the metrology apparatus 128 can be included as part of the manufacturing apparatus 124. For example, the manufacturing apparatus 124 can include a processing tool. One or more components of the metrology apparatus 128 can be integrated into one or more components or stations of the processing tool. For example, one or more components of the metrology apparatus 128 can be integrated into a factory interface, a load lock, a transfer chamber, a processing chamber, and / or one or more additional stations of a processing tool of the manufacturing apparatus 124.

[0022] The manufacturing equipment 124 may produce a product according to a recipe and / or perform a run over a period of time. The manufacturing equipment 124 may include one or more sensors configured to generate data (referred to as sensor data) of a substrate during a substrate process. The sensor data may include values ​​of one or more of the following: temperature (e.g., heater temperature), spacing (SP), pressure, high frequency radio frequency (HFRF), voltage of an electrostatic chuck (ESC), current, flow, power, voltage, etc. The sensor data may be associated with or indicate a manufacturing parameter, such as a hardware parameter, such as a setting or component (e.g., size, type, etc.) of the manufacturing equipment 124, or a process parameter of the manufacturing equipment 124. The sensor data (e.g., a device reading when processing a product) may be provided when the manufacturing equipment 124 performs a manufacturing process. The sensor data may be different for each substrate.

[0023] The metrology equipment 128 provides metrology data associated with substrates (e.g., wafers, etc.) processed by the manufacturing equipment 124. The metrology data may include one or more values ​​of the following: film property data (e.g., wafer space film properties), dimensions (e.g., thickness, height, etc.), dielectric constants, dopant concentrations, densities, defects, etc. In some embodiments, the metrology data may further include the following values: one or more surface profile property data (e.g., etching rate, etching rate uniformity, critical dimensions of one or more features included on the substrate surface, critical dimension uniformity across the substrate surface, edge placement error, etc.). The metrology data may be metrology data for a finished or semi-finished product. The metrology data may be different for each substrate. In some embodiments, the metrology equipment 128 may collect metrology data for each substrate processed at the manufacturing equipment 124. In other or similar embodiments, the metrology equipment 128 may collect metrology data for a portion of the substrates processed at the manufacturing equipment 124. For example, a batch of substrates may be processed at the manufacturing equipment 124. The metrology device 128 may collect metrology data for a portion of the substrates in the batch (e.g., 15% of the substrates in the batch, 20% of the substrates in the batch, etc.). In some embodiments, a system of the system architecture 100 (e.g., computing system 150, prediction system 110) may correlate the metrology data collected for the portion of the substrates in the batch as representative of the metrology data for each substrate in the batch.

[0024] The client device 120 includes a computing device such as a personal computer (PC), a laptop, a mobile phone, a smart phone, a tablet computer, a notebook computer, a network-connected television ("Smart TV"), a network-connected media player (e.g., a Blu-ray player), a set-top box, an over-the-top (OTT) streaming device, an operation box, etc. In some embodiments, metrology data may be received from the client device 120. In some embodiments, the client device 120 displays a graphical user interface (GUI), wherein the GUI enables a user to provide as input metrology measurements for a substrate processed at the manufacturing system. In other or similar embodiments, the client device 120 may display another GUI that enables a user to provide as input an indication of the type of substrate to be processed at the manufacturing system, the type of process to be performed on the substrate, and / or the type of equipment at the manufacturing system.

[0025] The data storage 140 may be a memory (e.g., random access memory), a drive (e.g., a hard disk, a flash drive), a database system, or another type of component or device capable of storing data. The data storage 140 may include multiple storage components (e.g., multiple drives or multiple databases) that may span multiple computing devices (e.g., multiple server computers). In some embodiments, the data storage 140 may store data associated with a process performed or to be performed on one or more substrates at the manufacturing equipment 124 (referred to herein as process data). The process data may include an indication of a process recipe associated with the process and / or settings for one or more operations of the process recipe. A process recipe refers to a set or series of operations and / or instructions associated with a process performed at the manufacturing equipment 124. For example, a deposition process recipe may include a set or series of operations and / or instructions associated with performing a deposition process on a substrate at a processing chamber of the manufacturing equipment 124. In some embodiments, a process recipe may include a set of operations and / or instructions associated with transporting a substrate to and / or out of a particular station of the manufacturing equipment 124. For example, a deposition process recipe may also include a set and / or series of operations associated with transporting a substrate into a processing chamber prior to initiating a deposition process (e.g., from a transfer chamber via a transfer chamber robot, etc.) and / or transporting a substrate out of a processing chamber after completing a deposition process (e.g., via a transfer chamber robot to a transfer chamber, etc.). In some embodiments, process data may refer to historical process data (e.g., process data associated with a previous process performed using the fabrication equipment 124) and / or current process data (e.g., process data associated with a current process being performed or to be performed using the fabrication equipment 124).

[0026] In additional or alternative embodiments, the data storage 140 may store data collected for a substrate by a conductive sensor at or coupled to the manufacturing equipment 124 before, during, and / or after performing a substrate process (referred to herein as sensor data). For example, a processing chamber may include one or more sensors (e.g., a temperature sensor, a spectral sensor, etc.) configured to collect data for a substrate and / or an environment within the processing chamber before, during, or after a substrate process. A computing system associated with the system architecture 100 (e.g., the prediction system 110, the computing system 150, a system controller for the manufacturing equipment 124, etc.) may receive the sensor data collected before, during, or after a substrate process and may store the sensor data at the data storage 140. In some embodiments, the sensor data may refer to historical sensor data (e.g., sensor data collected for a previous substrate processed according to a previous substrate process) and / or current sensor data (e.g., sensor data collected for a current substrate being processed or to be processed according to a current substrate process).

[0027] In some embodiments, the data storage 140 can store additional types of data. For example, the data storage can store metrology data associated with substrates processed using the manufacturing equipment 124. The metrology data can include historical metrology data (e.g., metrology measurements made for previous substrates processed using the manufacturing equipment 124) and / or current metrology data (e.g., metrology measurements made for current substrates processed using the manufacturing equipment 124). The data storage 140 can also store contextual data associated with one or more substrates at the manufacturing system (e.g., previous substrates, current substrates, etc.). The contextual data can include an identifier of a process recipe, an identifier of a substrate (and / or a batch of substrates), a preventive maintenance indicator, an identifier of an operator, and the like.

[0028] In some embodiments, the data storage 140 may be configured to store data that is not accessible to users of the manufacturing system (e.g., operators, engineers, etc.). For example, process data, sensor data, metrology data, and / or context data obtained for a substrate may not be accessible to users of the manufacturing system. In some embodiments, all data stored at the data storage 140 is not accessible to users of the manufacturing system (e.g., operators). In other or similar embodiments, a portion of the data stored at the data storage 140 is not accessible to the user, while another portion of the data stored at the data storage 140 is accessible to the user. In some embodiments, one or more portions of the data stored at the data storage 140 are encrypted using an encryption mechanism unknown to the user (e.g., the data is encrypted using a private encryption key). In other or similar embodiments, the data storage 140 includes a plurality of data storages, wherein data that is not accessible to the user is stored in one or more first data storages and data that is accessible to the user is stored in one or more second data storages.

[0029] The computing system 150 may include a batch-to-batch (R2R) control engine 152 and / or a prediction component 114. The R2R control engine 152 may be configured to tune and / or modify process recipe settings for a process performed (or to be performed) using the manufacturing equipment 124 so as to optimize the process and / or cause the characteristics of the substrate (e.g., after the process is completed) to correspond to target characteristics, etc. In some embodiments, the R2R control engine 152 may tune and / or modify the settings of the process recipe associated with the current process to be performed on the substrate at the manufacturing equipment 124 in view of data associated with previous processes performed on the substrate and / or data associated with upstream processes performed on the substrate. In some embodiments, one or more processes performed using the manufacturing equipment 124 may correspond to lithography techniques to form one or more complex patterns on the surface of the substrate. Each process may correspond to a corresponding layer of the substrate. The substrate process to be performed on the layer of the substrate (e.g., for depositing one or more layers, etching away portions of one or more layers, etc.) is referred to herein as the current layer process. As indicated above, such a layer is referred to herein as the current layer. A substrate process previously performed on a layer of the substrate (e.g., before performing a current layer process) is referred to herein as a previous layer process. As indicated above, such a layer is referred to herein as a previous layer or initial layer. A process performed to prepare the substrate for the current layer process is referred to herein as an upstream process (e.g., a polishing process, a roughening process, an etching process, etc.). In some embodiments, an upstream process can be performed on a previous layer (e.g., to prepare for deposition of the current layer on the previous layer, etc.).

[0030] In some embodiments, the R2R control engine 152 may determine one or more settings of a process recipe associated with a current layer process to be tuned and / or how much to tune the corresponding process recipe based on data associated with a previous layer of the substrate and / or data associated with an upstream process performed for the current layer of the substrate. In some embodiments, the R2R control engine 152 may obtain metrology data associated with a previous layer of a substrate (or a batch of substrates) (e.g., from a data storage 140) and may determine an offset by which the obtained metrology data deviates from target metrology data associated with the substrate. In some embodiments, the R2R control engine 152 may determine a first correction factor to be used to tune the current layer process recipe settings based on the determined offset. Further details on determining an offset associated with a previous layer metrology and determining a correction factor associated with a previous layer are provided herein.

[0031] In additional or alternative embodiments, the R2R control engine 152 may determine a second correction factor to be used to tune the current layer process recipe settings based on data associated with an upstream process performed for the current layer of the substrate. For example, the R2R control engine 152 may obtain process data and / or sensor data associated with the upstream process performed for the current layer of the substrate and may provide the obtained process data and / or sensor data to the prediction component 114. As described below, the prediction component 114 may provide the process data and / or sensor data as input to the trained machine learning model 190. In some embodiments, the machine learning model 190 may be trained to predict metrology measurements associated with the substrate and the offset of the metrology measurements from the target metrology measurements based on given process data and sensor data. Further details about the trained machine learning model 190 are provided herein. The prediction component 114 may provide one or more outputs of the trained machine learning model 190 to the R2R control engine 152. According to embodiments described herein, the R2R control engine 152 may determine a second correction factor based on one or more outputs of the trained machine learning model 190.

[0032] In some embodiments, the R2R control engine 152 can determine a third correction factor to be used to tune the current layer process recipe settings based on data associated with a previous substrate processed according to the current layer recipe. For example, the R2R control engine 152 can obtain metrology data collected for a previous substrate processed according to the current layer recipe and can determine an offset by which the obtained metrology data deviates from target metrology data associated with the substrate after completing the current layer process. In some embodiments, the third correction factor can correspond to the determined offset. Further details on determining the third correction factor are provided in further detail below.

[0033] In some embodiments, the R2R control engine 152 may determine which settings of the current layer process will be tuned in view of the first correction factor, the second correction factor, and / or the third correction factor. According to embodiments described herein, the R2R control engine 152 may additionally or alternatively determine how much to tune the settings of the current layer process in view of the first correction factor, the second correction factor, and / or the third correction factor. In some embodiments, the R2R control engine 152 may modify the process recipe associated with the current layer process based on the determined modification or tuning of the process recipe settings. In some embodiments, the R2R control engine 152 may provide an indication of the modified process recipe to a system controller associated with the manufacturing equipment 124. According to embodiments of the present disclosure, the system controller may perform the current layer process based on the modified process recipe. Further details about the R2R control engine 152 and modifying the current layer process recipe are provided herein.

[0034] In some embodiments, the prediction system 110 includes a server machine 170 and a server machine 180. The server machine 170 includes a training set generator 172 that can generate a training data set (e.g., a data input set and a target output set) to train, validate, and / or test a machine learning model 190. As described above, the machine learning model 190 can be trained to predict metrology data associated with a substrate based on given process data and / or sensor data associated with an upstream process performed on a current layer of the substrate, and in some embodiments, predict an offset of the metrology data from target metrology data associated with the substrate. Some operations of the training set generator 172 are described below with respect to Figure 5 Detailed description In some embodiments, the training set generator 172 can divide the training data into a training set, a validation set, and a test set. In some embodiments, the prediction system 110 generates multiple sets of training data.

[0035] The server machine 180 includes a training engine 182, a validation engine 184, a selection engine 186, and / or a test engine 188. An engine may refer to hardware (e.g., circuitry, dedicated logic, programmable logic, microcode, a processing device, etc.), software (such as instructions running on a processing device, a general purpose computer system, or a dedicated machine), firmware, microcode, or a combination thereof. The training engine 182 may be capable of training a machine learning model 190. The machine learning model 190 may refer to a model artifact created by the training engine 182 using training set data including training inputs and corresponding target outputs (correct answers for the corresponding training inputs). The training engine 182 may discover patterns in the training data that map the training inputs to target outputs (predicted answers) and provide a machine learning model 190 that captures these patterns. In some embodiments, the machine learning model 190 uses one or more of the following: support vector machine (SVM), radial basis function (RBF), clustering, supervised machine learning, semi-supervised machine learning, unsupervised machine learning, K-nearest neighbor algorithm (k-NN), linear regression, random forest, neural network (e.g., artificial neural network), clustering technology (e.g., hierarchical clustering technology), association technology (e.g., prior technology), classification technology (e.g., decision tree, random forest technology, etc.), variable recursive autoencoder, etc.

[0036] The validation engine 184 may be capable of validating the trained machine learning models 190 using a corresponding set of features of the validation set from the training set generator 172. The validation engine 184 may determine the accuracy of each of the trained machine learning models 190 based on the corresponding set of features of the validation set. The validation engine 184 may discard trained machine learning models 190 having an accuracy that does not meet a threshold accuracy. In some embodiments, the selection engine 185 may be capable of selecting a trained machine learning model 190 having an accuracy that meets a threshold accuracy. In some embodiments, the selection engine 185 may be capable of selecting a trained machine learning model 190 having a highest accuracy of the trained machine learning models 190.

[0037] The testing engine 188 may be able to test the trained machine learning model 190 using a corresponding set of features of the test set from the training set generator 172. For example, a first trained machine learning model 190 trained using a first set of features of the training set may be tested using a first set of features of the test set. The testing engine 188 may determine, based on the test set, that the trained machine learning model 190 has the highest accuracy of all the trained machine learning models.

[0038] The prediction server 112 includes a prediction component 114 that can provide process data and / or sensor data associated with an upstream process performed on a current layer of a substrate as input to a trained machine learning model 190 and run the trained machine learning model 190 on the input to obtain one or more outputs. Figure 3 In detail, in some embodiments, the prediction component 114 can also extract metrology data from one or more outputs of the trained machine learning model 190, and use the extracted metrology data to determine one or more metrology measurements associated with the current layer of the substrate after the upstream process. In some embodiments, the prediction component 114 can also determine an offset of one or more metrology measurements from a target metrology measurement. In some embodiments, the prediction component 114 can provide the determined metrology measurement and / or the determined offset to the R2R control engine 152. In additional or alternative embodiments, the prediction component 114 can provide one or more outputs of the trained machine learning model 190 to the R2R control engine 152. In some embodiments, the R2R control engine 152 can determine one or more metrology measurements associated with the current layer of the substrate after the upstream process and / or an offset of one or more metrology measurements from the target metrology measurement. According to embodiments described herein, the R2R control engine 152 can determine a second correction factor based on the offset.

[0039] The client device 120, the manufacturing equipment 124, the metrology equipment 128, the prediction server 112, the data storage 140, the computing system 150, the server machine 170, and the server machine 180 can be coupled to each other via the network 130. In some embodiments, the network 130 is a public network that provides the client device 120 with access to the computing system 150, the prediction server 112, the data storage 140, and other publicly available computing devices. In some embodiments, the network 130 is a private network that provides the client device 120 with access to the manufacturing equipment 124, the metrology equipment 128, the data storage 140, and other privately available computing devices. The network 130 may include one or more wide area networks (WANs), local area networks (LANs), wired networks (e.g., Ethernet networks), wireless networks (e.g., 802.11 networks or Wi-Fi networks), cellular networks (e.g., Long Term Evolution (LTE) networks), routers, hubs, switches, server computers, cloud computing networks, and / or combinations thereof.

[0040] It should be noted that in some other embodiments, the functionality of server machines 170 and 180, and prediction server 112 may be provided by a smaller number of machines. For example, in some embodiments, server machines 170 and 180 may be integrated into a single machine, while in some other or similar embodiments, server machines 170 and 180, and prediction server 112 may be integrated into a single machine. In other or similar embodiments, server machines 170 and 180, prediction server 112, and / or computing system 150 may be integrated into a single machine or into one or more machines.

[0041] In general, in one embodiment the functions described as being performed by computing system 150, server machine 170, server machine 180, and / or prediction server 112 may also be performed on client device 120. Furthermore, the functionality attributed to a particular component may be performed by a different or multiple components operating together.

[0042] In embodiments, a "user" may be represented as a single individual. However, other embodiments of the present disclosure encompass an entity where a "user" is controlled by multiple users and / or automation sources. For example, a collection of independent users united as a group of administrators may be considered a "user."

[0043] Figure 2 is a block diagram of an example batch-to-batch (R2R) control engine 152 according to aspects of the present disclosure. Figure 2 , the R2R control engine 152 may include a previous layer control component 210, an upstream process component 212, a current layer feedforward control component 214, and / or a current layer feedback control component 216. In some embodiments, the R2R control engine 152 may be connected to the memory 250 (e.g., via a memory device associated with the memory device). Figure 1 In some implementations, memory 250 may correspond to one or more portions of data storage 140 .

[0044] As described above, the previous layer control component 210 can be configured to determine a first correction factor (e.g., the previous layer correction factor 260) to be used to tune and / or modify settings associated with the current layer process recipe. As described above, in some embodiments, the metrology device 128 can generate metrology data associated with a previous layer of a substrate processed using the manufacturing device 124. For example, after the previous layer substrate process is completed at the manufacturing device 124, the substrate can be transferred to the metrology device 128 and metrology data can be generated and stored as the previous layer metrology data 252 at the memory 250. It should be noted that in some embodiments, the metrology data may not be generated for each substrate processed using the manufacturing device 124. For example, a batch of two or more substrates can be processed using the manufacturing device 124. As described above, the metrology data 252 can be generated for a portion of the substrates included in the batch. In some embodiments, the metrology data 252 generated for the portion of the substrates included in the batch can be associated with each substrate in the batch. In additional or alternative embodiments, the metrology data 252 can be generated for each substrate processed using the manufacturing device 124. According to implementations described herein, previous layer control component 210 may obtain previous layer metrology data 252 and may determine previous layer correction factor 260 .

[0045] In some embodiments, the upstream process component 212 may be configured to determine a second correction factor (e.g., the upstream process correction factor 262) to be used to tune and / or modify the current layer process recipe settings. In some embodiments, as described above, the upstream process component 212 may obtain process data associated with an upstream process performed for a current layer of the substrate (e.g., the upstream process data 254) and / or sensor data associated with the upstream process (e.g., the upstream process sensor data 256) from the memory 250 and may provide the upstream process data 254 and / or the upstream process sensor data 256 as input to the machine learning model 190. In other or similar embodiments, as described above, the upstream process component 212 may provide the upstream process data 254 and / or the upstream process sensor data 256 to the prediction component 114 and the prediction component 114 may provide the data 254 and / or the data 356 as input to the machine learning model 190. As described above, the upstream process component 212 (and / or the prediction component 114) can determine metrology data associated with a current layer of the substrate based on one or more outputs of the machine learning model 190, and according to the embodiments described herein, the upstream process correction factor 262 can be determined based on the determined metrology data and / or the deviation of the determined metrology data from the target metrology data.

[0046] The current layer feed-forward control component 214 may be configured to determine one or more modifications to a process recipe associated with a current layer process to be performed on a substrate. In some embodiments, the current layer feed-forward control component 214 may determine one or more modifications based on a previous layer correction factor 260, an upstream process correction factor 262, and / or a current layer correction factor 264, and may store the determined one or more modifications as process recipe modification data 266 at the memory 250. Further details on determining one or more modifications to a process recipe are provided herein.

[0047] In some embodiments, the current layer feedback control component 216 can be configured to determine a third correction factor (e.g., current layer correction factor 264) to be used to tune and / or modify the current process recipe settings. In some embodiments, as described above, a previous substrate can be processed according to the current layer process recipe. The metrology device 128 can generate metrology data associated with the previous substrate after completing the current layer process, which can be stored at the memory 250 as current layer metrology data 258. According to embodiments described herein, the current layer feedback control component 216 can determine the current layer correction factor 264 in view of the current layer metrology data 258.

[0048] Figure 3 is a flow chart of a method 300 for performing R2R control at a manufacturing system using machine learning according to aspects of the present disclosure. The method 300 is performed by processing logic, which may include hardware (circuitry, dedicated logic, etc.), software (such as running on a general-purpose computer system or a dedicated machine), firmware, or some combination thereof. In one embodiment, the method 300 may be performed by a system architecture (such as Figure 1 In other or similar embodiments, one or more operations of method 300 may be performed by one or more other machines not depicted in the drawings. In some aspects, one or more operations of method 300 may be performed by R2R control engine 152 of computing system 150. In other or similar aspects, one or more operations of method 300 may be performed by system controller 228. In still other or similar aspects, one or more operations of method 300 may be performed by prediction component 114.

[0049] For simplicity of explanation, the method is depicted and described as a series of actions. However, the actions according to the present disclosure may occur in any order and / or simultaneously, and other actions are not presented and described herein. In addition, the method according to the disclosed target may not perform all the actions shown. In addition, it will be understood and appreciated by those skilled in the art that the method may alternatively be represented as a series of interrelated states via a state diagram or event. In addition, it should be understood that the method disclosed in this specification can be stored on an article of manufacture, so that such a method can be transmitted and transferred to a computing device. As used herein, the term article of manufacture is intended to cover a computer program accessible from any computer-readable device or storage medium.

[0050] At block 310, processing logic identifies first data associated with a first process performed at a manufacturing system for a first layer of a substrate. In some embodiments, the first process may correspond to an upstream process performed on a current layer of a substrate using manufacturing equipment 124. As described above, the upstream process may be performed to prepare the substrate for the process of the current layer process. In some embodiments, the upstream process may include a polishing process, a roughening process, an etching process, and the like. In one example, the upstream process may include a chemical mechanical polishing process to prepare a previous layer of a substrate for deposition of the current layer onto the previous layer.

[0051] Figure 4 Depicted is an example of using machine learning to perform R2R control at a manufacturing system according to aspects of the present disclosure. Figure 4 As shown in FIG. 4 , one or more previous layer processes 410 may be performed on a substrate (or a batch of substrates) at a manufacturing system using manufacturing equipment 124. Figure 4 1, one or more upstream processes 412 may be performed on a substrate (or a batch of substrates) using a manufacturing apparatus 124 (or other apparatus of a manufacturing system) (e.g., after completing one or more previous layer processes 410). Processing logic (e.g., an upstream process component 212 of an R2R control engine 152) may obtain data associated with the upstream process 412. In some embodiments, as described above, the upstream process component 212 may obtain the upstream process data from the memory 250 and / or the data storage 140. In other or similar embodiments, a processing component (e.g., a system controller 450) associated with the manufacturing apparatus 124 may obtain first data (e.g., from the memory 250 and / or the data storage 140) and may send the first data to the upstream process component 212 (e.g., via the network 130, via a bus, etc.).

[0052] The upstream process data may include recipe data 254 associated with the upstream process 412 performed on the substrate and / or sensor data 256 collected by sensors at or coupled to the manufacturing system before, during, and / or after performing the upstream process 412. According to the previous example, the recipe data 254 for the chemical mechanical polishing process may include an indication of one or more operations performed during the process, an indication of characteristics of a pad used for the chemical mechanical polishing process (e.g., pad material, pad diameter, pad hardness, pad compressibility, etc.), an indication of a rotation speed and / or rotation angle of the pad during the chemical mechanical polishing process, an indication of one or more materials used to perform the chemical mechanical polishing process and / or concentrations of such materials, a downward force associated with the chemical mechanical polishing process, speeds, relative speeds, and / or other kinematic characteristics of one or more moving parts involved in the chemical mechanical polishing process, substrate-pad interface characteristics, etc. The sensor data 256 may include data collected by sensors before, during, and / or after performing the chemical mechanical polishing process.

[0053] In some embodiments, the amount of upstream process data identified or otherwise obtained by the upstream process component 212 may be very large and / or may include data that does not correspond to a specific feature of the current layer of the substrate. Thus, in some embodiments, the upstream process component 212 may analyze the upstream process data to identify a portion of the upstream process data corresponding to the specific feature of the current layer of the substrate and extract the identified data as the first data. In some embodiments, the upstream process component 212 may include a feature analysis module 414, which is configured to analyze the upstream process data to identify data about the specific feature of the current layer of the substrate. The feature analysis module 414 may provide one or more portions of the upstream process data as input to one or more dimensionality reduction functions. The dimensionality reduction function may depict the upstream process data as a point in a real coordinate space and may identify unrelated variables of the data point. Data points corresponding to related variables may be associated with upstream process data about the specific feature of the current layer of the substrate. In some embodiments, the dimensionality reduction function may correspond to at least one of a principal component analysis function, a partial least squares analysis function, or an autoencoder function. The feature analysis module 414 may obtain one or more outputs of the dimensionality reduction function. In some embodiments, one or more outputs of the dimensionality reduction function may indicate data points corresponding to the relevant variables. In some embodiments, the one or more outputs may include an indication of one or more portions of the upstream process data regarding a particular feature of the current layer of the substrate. In some embodiments, the feature analysis module 414 may extract the first data from the one or more outputs.

[0054] Return to see Figure 3At block 312, processing logic (e.g., upstream process component 212) provides the first data as input to a machine learning model. In some embodiments, the machine learning model may correspond to machine learning model 190. As described above, the machine learning model 190 may be trained to predict metrology measurements of one or more layers of a substrate according to an upstream process based on given process data and / or sensor data, and in some embodiments, predict an offset of the predicted metrology measurements from a target metrology measurement of the substrate after the upstream process. In some embodiments, the machine learning model 190 may be trained to predict a metrology measurement of one or more layers of a substrate according to an upstream process based on given process data and / or sensor data. Figure 5 The described implementation trains the machine learning model 190.

[0055] In some embodiments, the upstream process component 212 may directly provide the first data as an input to the machine learning model 190 and may obtain one or more outputs of the model 190. In other or similar embodiments, the upstream process component 212 may provide the first data to the prediction component 114, and the prediction component 114 may provide the first data as an input to the machine learning model 190. In such embodiments, the prediction component 114 may obtain one or more outputs of the model 190, and in some embodiments, may provide the one or more outputs to the upstream process component 212. The output of the machine learning model 190 may include metrology data indicating one or more sets of metrology measurement values, and for each set of metrology measurement values, including an indication of a confidence level that the corresponding set of metrology measurement values ​​corresponds to a current layer of the substrate (e.g., after completing an upstream process and / or after completing a current layer process). In some embodiments, the metrology data may also indicate an offset of each set of metrology measurement values ​​from a target set of metrology measurement values. In some embodiments, the target set of metrology measurement values ​​may be provided to the system 100 by a user of the manufacturing system (e.g., a developer, engineer, operator, etc.). In other or similar embodiments, the target set of metrology measurements may be determined (eg, by the R2R control engine 152, by the system controller 450, etc.) in view of experimental data associated with previous substrates processed according to the upstream process recipe and / or the current layer process recipe.

[0056] In some embodiments, the upstream process component 212 and / or the prediction component 114 may identify a corresponding set of metrology measurement values ​​having a confidence level that satisfies a confidence criterion. In some embodiments, the confidence level may satisfy the confidence criterion if the confidence level meets and / or exceeds a threshold confidence level. In other or similar embodiments, the confidence level may satisfy the confidence criterion if the confidence level meets and / or exceeds the threshold confidence level and is higher than other confidence levels associated with other sets of metrology measurement values. In response to identifying the corresponding set of metrology measurement values ​​having a confidence level that satisfies the confidence criterion, the upstream process component 212 and / or the prediction component 114 may extract the extracted set of identified metrology measurement values ​​and / or the offset from one or more outputs of the machine learning model 190. In some embodiments, the output of the machine learning model 190 may include one or more sets of metrology measurement values ​​and may not include the offset of each set of metrology measurement values ​​from a target metrology measurement value. In such an embodiment, as described above, the upstream process component 212 and / or the prediction component 114 can extract a set of metrology measurement values ​​having a confidence level that satisfies the confidence criterion level, and can compare the extracted set of metrology measurement values ​​with the target metrology measurement value to determine the offset.

[0057] As described above, in some embodiments, the machine learning model 190 may be a variational recurrent auto-encoder (VRAE) model. The VRAE model may be trained to extract data corresponding to a specific feature of the current layer from the upstream process data and predict the deviation of the metrology measurement value from the target metrology measurement value in view of the extracted feature. Thus, the upstream process data obtained by the upstream process component 212 may correspond to the first data provided as input to the VRAE model. For example, as in Figure 4 4. In some embodiments, the upstream process component 212 may include a VRAE engine 416. In response to obtaining the upstream process data, the VRAE engine 416 of the upstream process component 212 may directly provide the upstream process data as an input to the VRAE model, rather than providing the upstream data to the feature analysis module 414, as described above with respect to additional embodiments. The VRAE engine 416 may extract, from one or more outputs of the VRAE model, an offset by which a set of predicted metrology measurement values ​​for a current layer of the substrate deviate from a target metrology measurement value. In some embodiments, the VRAE model may be trained based on historical data collected for previous layer processes, upstream processes, and / or current layer processes performed at the manufacturing system, or other manufacturing system associated with the system architecture 100.

[0058] At block 314, processing logic (e.g., upstream process component 212) determines an offset of a first set of metrology measurement values ​​for a first layer of a substrate from a target set of metrology measurement values ​​after completing at least one of a first process (e.g., upstream process 412) or a second process (e.g., current layer process) based on one or more outputs of the machine learning model 190. As indicated above, in some embodiments, for each set of metrology measurement values, one or more outputs of the machine learning model 190 may include an indication of an offset of the corresponding set of metrology measurement values ​​from the target metrology measurement value. Thus, the upstream process component 212 may determine an offset of a set of metrology measurement values ​​for the first layer of the substrate based on an indication of the offset extracted from the one or more outputs of the machine learning model 190. In other or similar embodiments, as indicated above, the upstream process component 212 may compare the set of metrology measurement values ​​extracted from the one or more outputs of the machine learning model 190 with the target metrology measurement value to determine an offset. In some embodiments, the upstream process component 212 may further determine an offset based on one or more additional metrology measurement values ​​obtained for a previous substrate processed according to the upstream process recipe and / or the current layer process recipe. For example, the upstream process component 212 may obtain additional metrology measurements obtained for a previous substrate (e.g., from the data storage 140 and / or the memory 250) to determine if there is a trend (e.g., greater than a threshold number of lots) in which the metrology measurements for the previous substrate and / or the current substrate are deviating from the target metrology measurements. In some embodiments, the upstream process component 212 may determine the amount of the offset based on the determined trend.

[0059] At box 316, processing logic (e.g., current layer feedforward control component 214) determines one or more modifications to a process recipe for a second process (e.g., the current layer process) and second data associated with a second layer (e.g., the previous layer) of the substrate. As indicated above, in some embodiments, the offset of a set of metrology measurements associated with the current layer of the substrate can correspond to a correction factor (e.g., correction factor 262) that will be used to tune the current layer process recipe settings. For example, the offset of the set of metrology measurements can indicate that the difference between the predicted smoothness of the current layer of the substrate before initiating the current layer process (e.g., depositing the current layer, etc.) deviates from the target smoothness of the current layer by a specific number of units. Thus, the correction factor 262 can correspond to a difference or change in the predicted smoothness of the current layer from the target smoothness that will be corrected for tuning and / or modifying the settings of the current layer process recipe. As Figure 4 As shown in FIG. 2 , the determined correction factor 262 may be provided to the current layer feed-forward control component 214. According to embodiments described below, the feed-forward control component 214 may determine one or more modifications to the process recipe in view of the correction factor 262.

[0060] like Figure 4 4. As shown in FIG. 4, according to the previously described embodiments, metrology data 254 for a substrate processed according to a previous layer process 410 can be collected. The previous layer control component 210 can identify or otherwise obtain the metrology data 254 (e.g., from the data storage 140, from the memory 250, etc.) and can compare the metrology data 254 to target metrology data (e.g., target metrology measurements) associated with the substrate after completing the previous layer process 410. In some embodiments, according to the previously described embodiments, the target metrology data associated with the substrate after completing the previous layer process 410 can be provided by a user of the manufacturing system and / or determined based on experimental data collected for previous substrates processed according to the previous layer process 410. According to the previously described embodiments, the previous layer control component 210 can determine an offset of the metrology data 254 from the target metrology data. As described above with respect to the correction factor 262, the correction factor 260 can correspond to the determined offset.

[0061] If also in Figure 4 , the current layer feedforward control component 214 can obtain another correction factor 264 from the current layer feedback control component 216. As described above, the current layer feedback control component 216 can identify or otherwise obtain metrology data collected for a previous substrate processed according to the current layer process recipe. The current layer feedback control component 216 can compare the metrology data obtained for the previous substrate with the target metrology data (e.g., target metrology measurement values) for the substrate processed according to the current layer process recipe, and can determine an offset of the collected metrology measurement values ​​based on the comparison. According to the previously described embodiments, the target metrology data for the substrate processed according to the current layer process recipe can be provided by a user of the manufacturing system and / or determined based on experimental data collected for the previous substrate processed according to the current layer process recipe. As described above with respect to correction factors 260 and 262, correction factor 264 can correspond to the determined offset of the previous substrate processed according to the current layer substrate process.

[0062] As indicated above, the current layer feedforward control component 214 may determine one or more modifications to the current layer process recipe based on the correction factors 260, 262, and / or 264. In some embodiments, the current layer feedforward control component 214 may determine one or more modifications to the current layer process recipe in view of one or more rules associated with the current layer process. For example, a user of the manufacturing system (e.g., a developer, an operator, an engineer, etc.) may provide an indication of one or more modifications to the process recipe in view of changes detected for a corresponding layer of the substrate. The current layer feedforward control component 214 may compare the correction factors 260, 262, and / or 264 with the changes and / or corrections included in the provided modifications to determine a modification or tuning of one or more settings to be applied to the current layer process recipe. In another example, the current layer feedforward control component 214 (or another component of the R2R control engine 152 and / or system 100) may identify or otherwise obtain historical and / or experimental data associated with tuning and / or modifications to the substrate process recipe in response to changes detected for a corresponding layer of the substrate. In one example, the history and / or historical data may include an indication of changes or differences that occurred after completing the previous layer process 410, the upstream process 412, and / or the current layer process 420 according to the unmodified current layer process recipe, and an indication of one or more adjustments or modifications to the current layer process recipe in view of the changes or differences. The current layer feedforward control component 214 may compare the correction factors 260, 262, and / or 264 with the identified historical and / or experimental data to determine the modification or tuning of one or more settings to be applied to the current layer process recipe.

[0063] In additional or alternative embodiments, the current layer feedforward control component 214 may provide the correction factors 260, 262, and / or 264 as inputs to a machine learning model that is trained to predict one or more modifications to the process recipe based on one or more given correction factors so as to cause the characteristics of the substrate to correspond to the target characteristics. In some embodiments, the machine learning model may be trained using historical and / or experimental data associated with previous substrates processed at the manufacturing system according to the previous layer process 410, the upstream process 412, and / or the current layer process 420. In other or similar embodiments, the machine learning model may be trained based on historical data and / or experimental data collected for previous substrates processed at other manufacturing systems. The current layer feedforward control component 214 may obtain one or more outputs of the machine learning model. In some embodiments, the one or more outputs may include process modification data, including one or more sets of process recipe modifications, and for each corresponding set of process recipe modifications, including an indication of a confidence level that the corresponding set of process recipe modifications corresponds to the target characteristics of the substrate. The current layer feedforward control component 214 may identify a set of process recipe modifications having a confidence level that satisfies a confidence criterion (eg, exceeds a threshold confidence level, etc.).

[0064] At block 318, processing logic (e.g., current layer feedforward control component 214) updates the process recipe for the second process based on the determined one or more modifications. The current layer feedforward control component 214 may identify a process recipe associated with the current layer process (e.g., at data storage 140, memory 250, etc.) and may update one or more settings of the process recipe in light of the determined one or more modifications. Figure 4 , the modified process recipe 418 may correspond to the current layer process recipe modified according to the determined one or more modifications. In some embodiments, the current layer feedforward control component 214 may send the modified process recipe 418 to the system controller 450 (e.g., via the network 130, via a bus). In other or similar embodiments, the current layer feedforward control component 214 may store the modified process recipe 418 at the data storage 140 and / or the memory 250. The system controller 450 may access the modified process recipe 418 via the data storage 140 and / or the memory 250 (e.g., via the network 130, via a bus, etc.).

[0065] In some embodiments, the system controller 450 may perform a current layer process 420 on a substrate using the manufacturing equipment 124 according to the modified process recipe 418. The system controller 450 may perform the current layer process 420 by executing one or more instructions associated with the modified process recipe 418. After completing the current layer process 420, the substrate may be transferred to the metrology equipment 128 and the metrology equipment 128 may collect metrology data 422 associated with the substrate. In some embodiments, the current layer feedback control component 216 may determine an offset of the metrology data 422 from target metrology data associated with the substrate after performing the current layer process 420 in accordance with previously described embodiments. In some embodiments, the current layer feedback control component 216 may determine another correction factor 264 based on the determined offset of the metrology data 422. The determined correction factor 264 may be used by the current layer feedforward control component 214 to determine one or more modifications or adjustments to be applied to the modified process recipe 418 before performing the current layer process 420 on future substrates at the manufacturing system.

[0066] Figure 5 1 is a flow chart of a method 500 for training a machine learning model (e.g., machine learning model 190, etc.) according to aspects of the present disclosure. The method 500 is performed by processing logic, which may include hardware (circuitry, dedicated logic, etc.), software (such as running on a general-purpose computer system or a dedicated machine), firmware, or some combination thereof. In one embodiment, the method 500 may be performed by a system architecture (such as Figure 1 In other or similar embodiments, one or more operations of method 500 may be performed by one or more other machines not depicted in the drawings. In some aspects, one or more operations of method 500 may be performed by prediction server 112 of prediction system 110. In other or similar aspects, one or more operations of method 500 may be performed by R2R control engine 152 of computing system 150. In still other or similar aspects, one or more operations of method 500 may be performed by system controller 450.

[0067] For simplicity of explanation, the method is depicted and described as a series of actions. However, the actions according to the present disclosure may occur in any order and / or simultaneously, and other actions are not presented and described herein. In addition, the method according to the disclosed target may not perform all the actions shown. In addition, it will be understood and appreciated by those skilled in the art that the method may alternatively be represented as a series of interrelated states via a state diagram or event. In addition, it should be understood that the method disclosed in this specification can be stored on an article of manufacture, so that such a method can be transmitted and transferred to a computing device. As used herein, the term article of manufacture is intended to cover a computer program accessible from any computer-readable device or storage medium.

[0068] At box 510, processing logic initializes the training set T to an empty set (e.g., {}). At box 512, processing logic identifies historical data associated with a historical first process performed at the manufacturing system for a first layer of a previous substrate. In some embodiments, according to the previously described embodiments, the historical first process may correspond to a historical upstream process that was performed to prepare the substrate for a historical current layer process. The historical data may include historical process data associated with the historical process and / or historical sensor data collected before, during, or after performing the historical process. In some embodiments, the historical data may further include contextual data associated with the historical process. For example, the history may include an indication of a processing chamber in which the historical process was performed, an indication of a batch associated with the previous substrate, and the like. In some embodiments, as described above, processing logic may identify historical data from data storage 140 and / or memory 250.

[0069] At block 514, processing logic identifies a set of historical metrology measurements obtained for a previous substrate after completing at least one of a historical first process or a historical second process for the first layer at the manufacturing system. In some embodiments, the historical second process may correspond to a historical current layer process. After completing the historical first process and / or the historical second process, historical metrology measurements may be generated (e.g., using metrology equipment 128) for the previous substrate and stored at data storage 140 and / or memory 250. According to previously described embodiments, processing logic may identify a set of historical metrology measurements from data storage 140 and / or memory 250.

[0070] At block 516, processing logic generates first training data based on the identified historical data associated with the historical first process. As described above, in some embodiments, the machine learning model 190 may be a variable recursive autoencoder model. In such embodiments, the first training data may include the identified historical data associated with the historical first process. In other or similar embodiments, processing logic may provide the identified historical data as input to one or more dimensionality reduction functions. The dimensionality reduction functions may correspond to the dimensionality reduction functions associated with the historical first process. Figure 4 The dimensionality reduction function described by the feature analysis module 414 of FIG. The processing logic may obtain one or more outputs of the dimensionality reduction function. The one or more outputs of the dimensionality reduction function may correspond to a portion of the historical data of a specific feature of the current layer regarding the history of the previous substrate.

[0071] The processing logic may include the relevant portion of the historical data in the generated first training data. In additional or alternative embodiments, the processing logic may also include in the first training data an indication of a particular data cluster corresponding to a historical process performed for a previous substrate. In an illustrative example, the processing logic may provide the context data, the historical data, and / or the relevant portion of the historical data as input to one or more data clustering functions. The data clustering functions may be configured to determine which data clusters are assigned the relevant portion of the historical data in view of the context data and / or one or more features associated with the historical data. In some embodiments, the historical data may be clustered in view of the type of equipment used to perform the historical process (e.g., indicated by the context data), the type of defects and / or variations detected for the previous substrate (e.g., as indicated by the historical process data and / or historical sensor data), and / or one or more settings of a process recipe for the historical process (e.g., as indicated by the historical process data). The processing logic may extract an identifier associated with a particular cluster associated with the relevant portion of the historical data from one or more outputs of the data clustering function and may include the extracted identifier with the relevant portion of the historical data in the generated first training data. It should be noted that in additional or alternative embodiments, the processing logic may not provide the relevant portion of the historical data (and / or the historical context data and / or the historical data) as input to the clustering function. Instead, the relevant portion of the historical data may be clustered by a machine learning model (e.g., according to one or more machine learning techniques).

[0072] At box 518, processing logic may generate second training data based on the identified set of historical metrology measurements. In some embodiments, the generated second training data may include an indication of one or more identified sets of historical metrology measurements. At box 520, processing logic generates an image between the first training data and the second training data. At box 522, processing logic adds the image to training set T. At box 524, processing logic determines whether training set T includes a sufficient amount of training data to train the machine learning model. It should be noted that in some embodiments, the sufficiency of training set T may be determined simply based on the number of images in the training set, while in some other embodiments, in addition to or in lieu of the number of input / output images, the sufficiency of training set T may be determined based on one or more other criteria (e.g., a measure of the diversity of training instances, etc.). In response to determining that the training set does not include a sufficient amount of training data to train the machine learning model, method 500 returns to box 512. In response to determining that training set T includes a sufficient amount of training data to train the machine learning model, method 500 continues to box 528.

[0073] At block 528, processing logic provides the training set T to train the machine learning model. In one embodiment, the training set T is provided to the training engine 182 of the server machine 180 to perform the training. In the case of a neural network, for example, input values ​​for a given input / output image are input to the neural network, and output values ​​of the input / output image are stored in output nodes of the neural network. The connection weights in the neural network are then adjusted according to a learning algorithm (e.g., back propagation, etc.), and the process is repeated for other input / output images in the training set T. After block 528, according to the embodiments described above, the machine learning model 190 can be used to predict metrology measurements associated with the substrate and the offset of the metrology measurements from the target metrology measurements based on the given process data and / or sensor data.

[0074] Figure 6A block diagram of an illustrative computer system 600 operating in accordance with one or more aspects of the present disclosure is depicted. In alternative embodiments, the machine may be connected (e.g., networked) to other machines in a local area network (LAN), an intranet, an extranet, or the Internet. The machine may operate as a server or a client machine in a client-server network environment, or as a peer machine in a peer-to-peer (or distributed) network environment. The machine may be a personal computer (PC), a tablet computer, a set-top box (STB), a personal digital assistant (PDA), a cellular phone, a network appliance, a server, a network router, a switch or a bridge, or any machine capable of executing a set of instructions (continuously or otherwise) that specifies actions to be taken by that machine. In addition, although only a single machine is shown, the term "machine" shall also be construed to include any collection of machines (e.g., computers) that independently or jointly execute an instruction set (or multiple instruction sets) to perform any one or more of the methods discussed herein. In embodiments, the computing device 600 may correspond to Figure 1 The prediction server 112 and / or computing system 150, Figure 4 The system controller 450 and / or another processing device of the manufacturing system 100.

[0075] The example computing system 600 includes a processing device 602, a main memory 604 (e.g., read-only memory (ROM), flash memory, dynamic random access memory (DRAM) such as synchronous DRAM (SDRAM)), etc.), a static memory 606 (e.g., flash memory, static random access memory (SRAM), etc.), and an auxiliary memory (e.g., a data storage device 628), which communicate with each other via a bus 608.

[0076] Processing device 602 may represent one or more general purpose processors, such as a microprocessor, a central processing unit, or the like. More specifically, processing device 602 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 602 may also be one or more special purpose processing devices, such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), a network processor, or the like. Processing device 602 may also be or include a system on a chip (SoC), a programmable logic controller (PLC), or other types of processing devices. Processing device 602 is configured to execute processing logic for performing the operations and steps discussed herein.

[0077] The computing device 600 may further include a network interface device 622 for communicating with a network 664. The computing device 600 may also include a video display unit 610 (e.g., a liquid crystal display (LCD) or a cathode ray tube (CRT)), an alphanumeric input device 612 (e.g., a keyboard), a cursor control device 614 (e.g., a mouse), and a signal generating device 620 (e.g., a speaker).

[0078] The data storage component 628 may include a machine-readable storage medium (or more specifically, a non-transitory computer-readable storage medium) 624 on which a set of one or more instructions 626 for any one or more methods or functions described herein is stored. A non-transitory storage medium refers to a storage medium other than a carrier wave. The instructions 626 may also reside completely or at least partially in the main memory 604 and / or in the processing device 602 during execution thereof by the computer device 600, the main memory 604, and the processing device 602, which also constitute a computer-readable storage medium.

[0079] The computer-readable storage medium 624 may also be used to store the model 190 and data used to train the model 190. The computer-readable storage medium 624 may also store a software library containing methods that call the model 190. Although the computer-readable storage medium 624 is illustrated as a single medium in the example embodiment, the term "computer-readable storage medium" should be considered to include a single medium or multiple media (e.g., a centralized or distributed database, and / or associated caches and servers) that store one or more sets of instructions. The term "computer-readable storage medium" should also be considered to include any medium that can store or encode a set of instructions for execution by a machine and cause the machine to perform any one or more methodologies of the present disclosure. The term "computer-readable storage medium" should thus be considered to include, but is not limited to, solid-state memories, and optical and magnetic media.

[0080] The foregoing description has been described several specific details, such as examples of specific systems, parts, methods, etc., in order to provide a good understanding of several embodiments of the present disclosure. However, it will be apparent to those skilled in the art that at least some embodiments of the present disclosure can be put into practice without these specific details. In other cases, known parts or methods are not described in detail and are provided in a simple block diagram format to avoid unnecessary confusion of the present disclosure. Therefore, the specific details of elaboration are only exemplary. Specific embodiments can be changed from these exemplary details and are still expected to be within the scope of the present disclosure.

[0081] Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Therefore, the appearance of the phrase "in one embodiment" or "in an embodiment" in various places throughout this specification does not necessarily refer to the same embodiment. In addition, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or". When the term "about" or "approximately" is used herein, this is intended to mean that the nominal value provided is accurate within ±10%.

[0082] Although the operations of the methods herein are illustrated and described in a particular order, the order of operations of each method may be changed so that some operations may be performed in reverse order, so that some operations may be performed at least partially simultaneously with other operations. In another embodiment, instructions or sub-operations of different operations may be performed in an intermittent and / or alternating manner.

[0083] It will be understood that the above description is intended to be illustrative rather than restrictive. After reading and understanding the above description, numerous other embodiments will be apparent to those skilled in the art. Thus, the scope of the present disclosure should be determined with reference to the appended claims together with the full scope of equivalents given by such claims.

Claims

1. A method comprising the following steps: identifying first data associated with a first process performed at a manufacturing system on a first layer of a substrate, wherein the first layer of the substrate is to be further processed according to a second process at the manufacturing system; providing the first data as input to a machine learning model, wherein the machine learning model is trained to predict metrology measurements for one or more layers of a substrate processed at the manufacturing system; determining, based on one or more outputs of the machine learning model, an offset of a first set of metrology measurement values ​​for the first layer of the substrate from a set of target metrology measurement values ​​after completing at least one of the first process or the second process; determining one or more modifications to a process recipe for the second process in view of the determined offset and second data associated with a second layer of the substrate, wherein the second layer of the substrate was previously processed according to a third process at the manufacturing system; as well as The process recipe of the second process is updated based on the determined one or more modifications.

2. The method of claim 1 , wherein the one or more modifications to the process recipe of the second process are determined further in view of third data comprising one or more metrology measurements associated with another substrate processed according to the second process at the manufacturing system.

3. The method of claim 1 , wherein the step of identifying the first data associated with the first process performed on the first layer of the substrate comprises the steps of: providing as input to one or more dimensionality reduction functions a data set associated with the first process performed on the first layer of the substrate; and The first data is extracted from outputs of the one or more dimensionality reduction functions.

4. The method of claim 3, wherein the one or more dimensionality reduction functions include at least one of a principal component analysis function, a partial least squares analysis function, or an autoencoder function.

5. The method of claim 1, wherein the machine learning model is a variational autoencoder model.

6. The method of claim 1, further comprising the steps of: identifying a second set of metrology measurements for the first layer of the substrate, wherein the second set of metrology measurements for the first layer of the substrate is generated after completing a substrate process for the first layer of the substrate according to the updated process recipe; and At least one of the updated process recipe or another process recipe for the third process is updated in view of the identified second set of metrology measurements.

7. The method of claim 1, wherein the process recipe corresponds to one or more operations associated with a lithographic process. 8 . The method of claim 7 , wherein the first process performed on the first layer of the substrate comprises at least one of a chemical mechanical polishing process, an etching process, or a deposition process.

9. The method of claim 1, wherein the first process comprises a chemical mechanical polishing process, the second process comprises one or more first operations of a lithography process, and the third process comprises one or more second operations of the lithography process performed prior to the one or more first operations.

10. A system comprising: Memory; and A processing device, coupled to the memory, the processing device being configured to: identifying first data associated with a first process performed at a manufacturing system on a first layer of a substrate, wherein the first layer of the substrate is to be further processed according to a second process at the manufacturing system; providing the first data as input to a machine learning model, wherein the machine learning model is trained to predict metrology measurements for one or more layers of a substrate processed at the manufacturing system; determining, based on one or more outputs of the machine learning model, an amount by which a first set of metrology measurement values ​​for the first layer of the substrate deviates from a target set of metrology measurement values ​​after completing at least one of the first process or the second process; determining one or more modifications to a process recipe for the second process in view of the determined offset and second data associated with a second layer of the substrate, wherein the second layer of the substrate was previously processed according to a third process at the manufacturing system; as well as The process recipe of the second process is updated based on the determined one or more modifications.

11. The system of claim 10, wherein the one or more modifications to the process recipe of the second process are determined further in view of third data comprising one or more metrology measurements associated with another substrate processed according to the second process at the manufacturing system.

12. The system of claim 10, wherein to identify the first data associated with the first process performed on the first layer of the substrate, the processing device is configured to: providing as input to one or more dimensionality reduction functions a data set associated with the first process performed on the first layer of the substrate; and The first data is extracted from outputs of the one or more dimensionality reduction functions.

13. The system of claim 12, wherein the one or more dimensionality reduction functions include at least one of a principal component analysis function, a partial least squares analysis function, or an autoencoder function.

14. The system of claim 10, wherein the machine learning model is a variational autoencoder model.

15. The system of claim 10, wherein the processing device is further configured to: identifying a second set of metrology measurements for the first layer of the substrate, wherein the second set of metrology measurements for the first layer of the substrate is generated after completing a substrate process for the first layer of the substrate according to the updated process recipe; and At least one of the updated process recipe or another process recipe for the third process is updated in view of the identified second set of metrology measurements.

16. The system of claim 10, wherein the first process comprises a chemical mechanical polishing process, the second process comprises one or more first operations of a lithography process, and the third process comprises one or more second operations of the lithography process performed prior to the one or more first operations.

17. A non-transitory computer readable medium comprising instructions which, when executed by a processing device, cause the processing device to: identifying first data associated with a first process performed at a manufacturing system on a first layer of a substrate, wherein the first layer of the substrate is to be further processed according to a second process at the manufacturing system; providing the first data as input to a machine learning model, wherein the machine learning model is trained to predict metrology measurements for one or more layers of a substrate processed at the manufacturing system; determining, based on one or more outputs of the machine learning model, an amount by which a first set of metrology measurement values ​​for the first layer of the substrate deviates from a target set of metrology measurement values ​​after completing at least one of the first process or the second process; determining one or more modifications to a process recipe for the second process in view of the determined offset and second data associated with a second layer of the substrate, wherein the second layer of the substrate was previously processed according to a third process at the manufacturing system; as well as The process recipe of the second process is updated based on the determined one or more modifications.

18. The non-transitory computer readable medium of claim 17, wherein the one or more modifications to the process recipe of the second process are determined further in view of third data comprising one or more metrology measurements associated with another substrate processed according to the second process at the manufacturing system.

19. The non-transitory computer readable medium of claim 17, wherein to identify the first data associated with the first process performed on the first layer of the substrate, the processing device is configured to: providing as input to one or more dimensionality reduction functions a data set associated with the first process performed on the first layer of the substrate; and The first data is extracted from outputs of the one or more dimensionality reduction functions.

20. The non-transitory computer-readable medium of claim 17, wherein the first process comprises a chemical mechanical polishing process, the second process comprises one or more first operations of a lithography process, and the third process comprises one or more second operations of the lithography process performed prior to the one or more first operations.