Performing radio frequency alignment control using a model-based digital twin

Abstract

Radio frequency matching control is implemented using a model-based digital twin.SOLUTION: The method includes causing manufacturing equipment to generate an RF signal to energize a processing chamber associated with the manufacturing equipment. The method further includes receiving, from one or more sensors associated with the manufacturing equipment, current trace data associated with the RF signals. The method further includes updating impedance values of the digital replica associated with the manufacturing equipment based on the current trace data. The method further includes obtaining, from the digital replica, one or more outputs indicative of the prediction data. The method further includes causing performance of one or more corrective actions associated with the manufacturing equipment based on the predictive data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001]

[0001] The present disclosure relates to electrical components, and more particularly to implementing radio frequency matching control using a model-based digital twin. [Background technology]

[0002]

[0002] Products may be produced by performing one or more manufacturing processes using manufacturing equipment. For example, semiconductor manufacturing equipment may be used to produce semiconductor devices (e.g., substrates, wafers, etc.) through a semiconductor manufacturing process. In a conventional radio frequency (RF) plasma etching process used in the manufacturing of many semiconductor devices, an RF signal may be provided to a substrate processing chamber via an RF energy source. The RF signal may be generated and provided in a continuous wave mode or a pulsed wave mode. Due to a mismatch between the impedance of the RF energy source and the plasma formed in the processing chamber, some of the RF signal is reflected back to the RF energy source. As a result, the RF signal is used inefficiently, wasting energy, potentially damaging the RF energy source, and potentially causing inconsistency / non-repeatable issues with substrate processing. Therefore, there is a need for an RF matching control system that can achieve high repeatability in the delivery of the RF signal. Summary of the Invention

[0003]

[0003] The following presents a simplified summary of the present disclosure in order to provide a basic understanding of some aspects of the disclosure. This summary is not an exhaustive overview of the disclosure. It is not intended to identify key or critical elements of the disclosure, nor is it intended to delineate the scope of any particular implementation of the disclosure or the scope of the claims. The sole purpose of this summary is to present some concepts of the disclosure in a simplified form as a prelude to the more detailed description that is presented below.

[0004] In one aspect of the present disclosure, a method includes causing a manufacturing device to generate an RF signal to energize a process chamber associated with the manufacturing device. The method further includes receiving current trace data associated with the RF signal from one or more sensors associated with the manufacturing device. The method further includes updating an impedance value of a digital replica associated with the manufacturing device based on the current trace data. The method further includes obtaining one or more outputs from the digital replica indicative of predicted data. The method further includes causing implementation of one or more corrective actions associated with the manufacturing device based on the predicted data.

[0005] In another aspect of the disclosure, a system includes a memory and a processing device coupled to the memory, the processing device causing a manufacturing equipment to generate an RF signal to energize a process chamber associated with the manufacturing equipment. The processing device further receives current trace data associated with the RF signal from one or more sensors associated with the manufacturing equipment. The processing device further updates an impedance value of a digital replica associated with the manufacturing equipment based on the current trace data. The processing device further obtains one or more outputs from the digital replica indicative of predicted data. The processing device further causes implementation of one or more corrective actions associated with the manufacturing equipment based on the predicted data.

[0006] In another aspect of the present disclosure, a non-transitory machine-readable storage medium stores instructions that, when executed, cause a processing device to perform steps including causing a manufacturing equipment to generate an RF signal to energize a processing chamber associated with the manufacturing equipment. The steps further include receiving current trace data associated with the RF signal from one or more sensors associated with the manufacturing equipment. The steps further include updating an impedance value of a digital replica associated with the manufacturing equipment based on the current trace data. The steps further include obtaining one or more outputs from the digital replica indicative of predicted data. The steps further include causing implementation of one or more corrective actions associated with the manufacturing equipment based on the predicted data.

[0007]

[0007] The present disclosure is illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram illustrating an exemplary system architecture, according to certain embodiments. [Figure 2] FIG. 2 is a block diagram illustrating in greater detail an exemplary embodiment of a manufacturing device, in accordance with certain embodiments. [Figure 3] FIG. 1 illustrates a flow diagram for generating predictive data to enable corrective action to be taken, according to certain embodiments. [Figure 4A] 1 is a graph illustrating an exemplary optimization profile, according to certain embodiments. [Figure 4B] 1 is a graph illustrating an exemplary optimization profile, according to certain embodiments. [Figure 4C] 1 is a graph illustrating an exemplary optimization profile, according to certain embodiments. [Figure 5] 1 is a graph illustrating multiple tuning paths associated with two variable capacitors, according to certain embodiments. [Figure 6]FIG. 1 is a block diagram illustrating a computer system, according to certain embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0009]

[0014] Described herein are techniques directed to implementing radio frequency matching control using a model-based digital twin. Manufacturing equipment may be used to fabricate products (e.g., wafers). For example, the manufacturing equipment may execute a recipe for fabricating wafers by supplying RF energy, in the form of an RF signal, to a processing chamber to generate a plasma for etching. Often, the RF signal is coupled to the plasma in the processing chamber through a fixed or tunable matching network. These matching networks operate to minimize reflected RF signals by more closely matching the impedance of the plasma to the impedance of the RF energy source. The matching network allows the output of the RF source to be efficiently coupled to the plasma to maximize the amount of energy coupled to the plasma. Thus, the matching network allows the total impedance (e.g., plasma impedance, chamber impedance, and matching network impedance) to be the same or similar to the output impedance of the RF power supply.

[0010]

[0015] In some conventional systems, the RF match control profile of the matching network is designed based on a lookup table. This lookup table may be associated with process parameters such as the type of input frequency used to energize the process chamber (e.g., 2.2 MHz, 13.56 MHz, etc.), the type of process gas used (e.g., H2, He, Ar, O2, NF3, etc.), the type of process chamber used, etc. The control profile may be used to adjust one or more variable tuning elements (e.g., variable capacitors, variable inductors, or variable resistors) of the matching network. If the parameters of the matching network or the process chamber change (e.g., due to temperature, aging effects, corrosion, component failure, degradation, etc.), the control profile selected using the lookup table may not produce optimal results.

[0011]

[0016] The devices, systems, and methods disclosed herein provide real-time optimization of control profiles to achieve maximum power delivery with minimal tuning time using current sensor data. In one embodiment, a processing device causes a manufacturing equipment to generate an RF signal to energize a process chamber associated with the manufacturing equipment. The processing device may receive current trace data (e.g., voltage, current, etc.) associated with the RF signal from one or more sensors associated with the manufacturing equipment. The processing device may then update an impedance value of a digital replica associated with the manufacturing equipment based on the current trace data. An output indicative of predicted data may be obtained from the digital replica. The output indicative of predicted data may be generated using a trained machine learning model, a heuristic model, or a rule-based model. The processing device may implement one or more corrective actions associated with the manufacturing equipment based on the predicted data.

[0012]

[0017] In some embodiments, the matching network may include two adjustable (tunable) capacitors. The predicted data may include one or more tuning settings for the capacitors. The tuning settings may be a function of a reflection coefficient associated with the reflected RF signal and a time parameter. The corrective action may include adjusting one or both of the adjustable capacitors based on the predicted data.

[0013]

[0018] In some embodiments, the processing device may use the current trace data to update the digital replica in real time, e.g., in response to determining that the digital replica does not meet an accuracy threshold criterion based on the current trace data, the processing logic may perform optimizations on the digital replica.

[0014]

[0019] In some embodiments, the matching network may include an input impedance sensor and an output impedance sensor. Updating the impedance values ​​of the digital replica may include updating a process chamber model associated with the digital replica using current trace data from the output impedance sensor. Updating the impedance values ​​of the digital replica may further include updating a matching network model associated with the digital replica using current trace data from the input impedance sensor and the output impedance sensor.

[0015]

[0020] Aspects of the present disclosure provide technical advantages of significantly reducing the time required to achieve optimal settings, energy consumption, bandwidth used, processor overhead, etc. For example, conventional systems perform trial-and-error iterations to attempt to improve the repeatability of optimal RF power delivery. In this trial-and-error process, each iteration (e.g., generating instructions via trial-and-error, sending the instructions, receiving feedback, generating updated instructions via trial-and-error, etc.) requires increasing time, energy consumption, bandwidth, and processor overhead. The present disclosure provides reductions in time, energy consumption, bandwidth, and processor overhead by using signal processing, digital replicas, and machine learning models to obtain predictive data and allow corrective actions to be taken based on the predictive data, and by avoiding trial-and-error iterations. The present disclosure provides predictions of optimal parameter settings associated with a matching network to avoid mismatched and abnormal products, unplanned user time, and damage to manufacturing equipment.

[0016]

[0021] 1 is a block diagram illustrating an example system 100 (example system architecture) according to certain embodiments. System 100 includes a simulation system 110, a client device 120, a sensor 126, manufacturing equipment 130, measurement equipment 128, and a data store 140. Simulation system 110 may include a digital representation server 170, a server machine 180, and a prediction server 112.

[0017]

[0022] The manufacturing equipment 130 may include an RF generator 132, a matching network 134, a process chamber 136, and a controller 138. An RF signal may be generated by the RF generator 132 and transmitted to the matching network 134. The RF signal may then be applied to the process chamber 136 to ignite and maintain a plasma used in the etching process. In some embodiments, the RF generator 132 may generate one or more low RF signals (e.g., 2.2 MHz, 13.56 MHz, etc.) to energize the process chamber 136, one or more high RF signals (e.g., 24 MHz, 60 MHz, 100 MHz, etc.) to energize the process chamber 136, or any combination thereof. The RF generator 132 may be capable of pulsing the RF signal at a desired pulse rate, duty cycle, and phase angle. A controller 138 may be connected to the RF power generator 132 and the RF matching network 134 and may control (e.g., start, switch, shut off, etc.) the RF signal of the RF generator 132. Additionally, the controller 138 may be used to adjust the pulse rate, duty cycle, and phase angle of the RF signal.

[0018]

[0023] The matching network 134 may operate to minimize reflected RF energy by matching the impedance of the plasma used in the etching process to the impedance (e.g., 50 ohms) of the RF signal provided by the RF generator 132. The matching network 134 may include capacitive, inductive, and resistive elements. The matching network 134 may include one or more controllable, adjustable, or variable tuning elements (such as a variable capacitor, inductor, or resistor, or a combination thereof) to vary its parameters. For example, the matching network 134 may include a variable shunt capacitor and a variable series capacitor, both of which function as variable tuning elements. The variable capacitor (e.g., tuning capacitor or tunable capacitor) may be an electromotive vacuum capacitor operated by the controller 138 or any other controllable variable capacitor. In one embodiment, the matching network 134 may tune one or more of the controllable variable tuning elements so that the impedance associated with the matching network 134 and / or the processing chamber 136 increases or decreases toward the impedance (e.g., 50 ohms) associated with the RF generator 132.

[0019]

[0024] The RF matching network 134 may also include one or more electrical sensors 135. The electrical sensors 135 may be any type of RF voltage / current measurement device (e.g., a sensor, a probe, etc.) capable of providing sensor data 142 (e.g., a sensor value, a signature, trace data, etc.). In some embodiments, the electrical sensors 135 perform electrical measurements on a power supply conductor (e.g., a power supply line) coupled to the RF generator 132, the matching network 134, and / or the process chamber 136. The electrical sensors 135 sense characteristics of the power supply conductor (e.g., impedance, magnetic fluctuations of the power supply conductor, current, voltage, resistance, etc.) and convert such characteristics into sensor data 142 (e.g., trace data, historical trace data 146, current trace data 154, etc.). The electrical sensors 135 may also measure one or more of the following: current, alternating current (AC) magnitude, phase, waveform (e.g., AC waveform or pulse waveform), direct current (DC), non-sinusoidal AC waveform, voltage, etc.

[0020]

[0025] In one embodiment, the electrical sensors 135 include an input (source) impedance sensor and an output (load) impedance sensor. The input impedance sensor may determine the impedance associated with the RF signal at the matching network 134. The output impedance sensor may determine the impedance of the RF signal at its destination (e.g., the processing chamber 136). The manufacturing equipment 130 is described in more detail below with reference to FIG. 2.

[0021]

[0026] The sensors 126 may provide sensor data 142 (e.g., sensor values, characteristics, trace data, etc.) associated with the manufacturing equipment 130 (e.g., associated with the manufacturing equipment 130 producing a corresponding product, such as a wafer). The sensor data 142 may be used in relation to equipment health and / or product health (e.g., product quality). The manufacturing equipment 130 may produce a product according to a recipe or perform an operation over a period of time. The sensor data 142 received over a period of time (e.g., corresponding to at least a portion of a recipe or operation) may be referred to as trace data received over time from the various sensors 126 (e.g., historical trace data 146 or current trace data 154).

[0022]

[0027] The sensors 126 may include additional sensors that provide other types of sensor data 142. In some embodiments, the sensor data 142 may include one or more values ​​of temperature (e.g., heater temperature), spacing (SP), pressure, high frequency radio frequency (HFRF), electrostatic chuck (ESC) voltage, current, flow, power, voltage, etc. The sensor data 142 may be associated with or indicative of manufacturing parameters, such as hardware parameters (e.g., settings or components (e.g., size, type, etc.) of the manufacturing equipment 130) or process parameters of the manufacturing equipment. The sensor data 142 may be provided while the manufacturing equipment 130 is performing a manufacturing process (e.g., equipment readings as the product is processed). The sensor data 142 may be different for each product (e.g., for each wafer).

[0023]

[0028] In some embodiments, sensor data 142 (e.g., historical trace data 146, historical component data set 148, current trace data 154, current component data set 156, etc.) may be processed (e.g., by client device 120 and / or by components of simulation system 110). Processing sensor data 142 may include generating features. In some embodiments, features are patterns in sensor data 142 (e.g., slope, width, height, peaks, etc.) or combinations of sensor values ​​from sensor data 142 (e.g., impedance derived from voltage and current measurements, etc.). Sensor data 142 may include features, which may be used by components of simulation system 110 to perform simulation processing and / or to obtain simulation data 167 and / or prediction data 168 for implementing corrective actions.

[0024]

[0029] Data store 140 may be memory (e.g., random access memory), a drive (e.g., a hard drive or flash drive), a database system, or another type of component or device capable of storing data. Data store 140 may include multiple storage components (e.g., multiple drives or multiple databases), which may be distributed across multiple computing devices (e.g., multiple server computers). Data store 140 may store sensor data 142, performance data 160, library data 166, simulation data 168, and prediction data 169. Sensor data 142 may include historical sensor data 144 and current sensor data 152. Historical sensor data may include historical trace data 146, a set of historical component data 148, and a historical component identifier 150. Current sensor data 152 may include current trace data 154, a set of current component data 156, and a current component identifier 158. Performance data 160 may include historical performance data 162 and current performance data 164. Historical sensor data 144 and historical performance data 162 may be historical data. Current sensor data 144 may be current data for which simulation data 167 and prediction data 168 are generated (e.g., to implement corrective actions). Current performance data 164 may be current data (e.g., to retrain a trained machine learning model).

[0025]

[0030] Performance data 160 may include data associated with manufacturing equipment 130 and / or products produced by manufacturing equipment 130. In some embodiments, performance data 160 may include indications of the lifespan (e.g., time to failure) of components of manufacturing equipment 130, manufacturing parameters of manufacturing equipment 130, maintenance of manufacturing equipment 130, energy use of components of manufacturing equipment 130, variations among components (e.g., of the same model number) of manufacturing equipment 130, etc. Performance data 160 may include indications of variations among components (e.g., of the same type or part number) of manufacturing equipment. Performance data 160 may indicate whether such variations (e.g., jitter, slope, peaks, etc.) contribute to product-to-product variations. Performance data 160 may indicate whether variations lead to improved wafers (e.g., better matched RF generators, better tuned feedback loops, newer firmware, better chips, etc.). Performance data 160 may be associated with the quality of products produced by manufacturing equipment 130. Metrology equipment 128 may provide performance data 160 (e.g., wafer characteristic data, yield, metrology data, etc.) associated with products (e.g., wafers) produced by manufacturing equipment 130. Performance data 160 may include one or more values ​​of film characteristic data (e.g., wafer spatial film characteristics), dimensions (e.g., thickness, height, etc.), dielectric constant, dopant concentration, density, defects, etc. Performance data 160 may be for finished products or semi-finished products. Performance data 160 may be different for each product (e.g., for each wafer). Performance data 160 may indicate whether a product meets a threshold quality (e.g., defective, not defective, etc.). Performance data 160 may also indicate the cause of failure to meet the threshold quality. In some embodiments, performance data 160 includes historical performance data 162, which corresponds to historical characteristic data of products (e.g., produced using manufacturing parameters associated with historical trace data 146).The sensor data 142, performance data 160, and library data 166 may be used for supervised and / or unsupervised machine learning.

[0026]

[0031] Simulation system 110 may include digital representation server 170, server machine 180, and prediction server 112. Prediction server 112, digital representation server 170, and server machine 180 may each include one or more computing devices, such as a rack-mounted server, a router computer, a server computer, a personal computer, a mainframe computer, a laptop computer, a tablet computer, a desktop computer, a graphics processing unit (GPU), an accelerator, an application-specific integrated circuit (ASIC) (e.g., a tensor processing unit (TPU)), or the like.

[0027]

[0032] Digital representation server 170 may be an algorithmic model that simulates manufacturing equipment 130. As an example, digital representation server 170 may be a digital replica model (e.g., a digital twin) of manufacturing equipment 130. Digital representation server 170 may use supervised machine learning, semi-supervised machine learning, unsupervised machine learning, or any combination thereof, to generate a virtual representation of the physical elements and dynamics of how manufacturing equipment 130 operates. Digital representation server 170 may be updated through reinforcement learning using periodic updates from sensors 126, electrical sensors 135, sensor data 142, performance data 160, library data 166, and / or any other data associated with generating and maintaining the digital replica model of manufacturing equipment 130.

[0028]

[0033] The digital representation server 170 may include a matching network model 172 and a processing chamber model 174. The matching network model 172 may be associated with the physics and dynamics of the matching network 134. The processing chamber model 174 may be associated with the physics and dynamics of the processing chamber 136.

[0029]

[0034] In some embodiments, the digital representation server 170 may generate simulation data 167. The simulation data 167 may include data used to determine how the manufacturing equipment 130 (e.g., the matching network 134, the processing chamber 136, etc.) will operate based on current or simulated parameters. The simulation data 167 may include electrical parameter data (e.g., impedance, voltage, current, resistance, reflection, signal reflection, etc.) associated with the matching network model 172 and the processing chamber model 174. The simulation data 167 may also include predicted characteristic data of a digital replica model of the manufacturing equipment 130 (e.g., of a product that is or was created using the current trace data 154). The simulation data 167 may also include predicted metrology data (e.g., virtual metrology data) of a product that is or was created using the current trace data 154. The simulation data 167 may also include an indication of anomalies (e.g., an abnormal product, an abnormal component, an abnormal manufacturing equipment 130, an abnormal energy usage, etc.) and one or more causes of such anomalies. The simulation data 167 may further include an indication of the end of life of components of the manufacturing equipment 130. The simulation data may be exhaustive and cover all mechanical and electrical aspects of the manufacturing equipment.

[0030]

[0035] The prediction server 112 may include a prediction component 114. In some embodiments, the prediction component 114 may receive simulation data 167 and current trace data 154 (e.g., process chamber flow, process chamber pressure, RF power, etc.) and generate output (e.g., predicted data 168) for implementing corrective actions associated with the manufacturing equipment 130. In some embodiments, the prediction component 114 may use one or more trained machine learning models 190 to determine the output for implementing corrective actions based on the simulation data 167 and the current trace data 154. In some embodiments, the prediction component 114 receives the simulation data 167 and the current trace data 154, performs signal processing to decompose the current trace data 154 into a set of current component data 156 mapped to a current component identifier 158, provides the set of current component data 156 and the current component identifier 158 as inputs to the trained machine learning model 190, and obtains an output from the trained machine learning model 190 indicative of the predicted data 168. Trained machine learning model 190 may include a single model or multiple models. In some embodiments, trained machine learning model 190 may use additional data from data store 140 (e.g., library data 166, performance data 160, sensor data 142, etc.).

[0031]

[0036] In some embodiments, the simulation system 110 further includes a server machine 180. The server machine 180 may use a dataset generator to generate one or more datasets (e.g., a set of data inputs and a set of target outputs) for training, validating, and / or testing the machine learning model(s) 190. In particular, the server machine 180 may include a training engine 182, a validation engine 184, a selection engine 185, and / or a test engine 186. The engines (e.g., the training engine 182, the validation engine 184, the selection engine 185, and the test engine 186) may refer to hardware (e.g., circuitry, dedicated logic, programmable logic, microcode, a processing device, etc.), software (e.g., instructions executed 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 one or more machine learning models 190 using one or more sets of features associated with the training sets from the dataset generator 172. The training engine 182 may generate multiple trained machine learning models 190, where each trained machine learning model 190 corresponds to a distinct set of features (e.g., sensor data from a distinct set of sensors) of the training set. For example, a first trained machine learning model may be trained using all features (e.g., X1-X5), a second trained machine learning model may be trained using a first subset of features (e.g., X1, X2, and X4), and a third trained machine learning model may be trained using a second subset of features (e.g., X1, X3, X4, and X5) that may partially overlap with the first subset of features.

[0032]

[0037] The validation engine 184 may be capable of validating the trained machine learning models 190 using corresponding feature sets of the validation set from the dataset generator. For example, a first trained machine learning model 190 trained using a first set of features of the training set may be validated using the first set of features of the validation set. The validation engine 184 may determine the accuracy of each of the trained machine learning models 190 based on the corresponding feature sets of the validation set. The validation engine 184 may discard trained machine learning models 190 with accuracies below a threshold accuracy. In some embodiments, the selection engine 185 may be capable of selecting one or more trained machine learning models 190 with accuracies that meet the threshold accuracy. In some embodiments, the selection engine 185 may be capable of selecting the trained machine learning model 190 with the highest accuracy among the trained machine learning models 190.

[0033]

[0038] The testing engine 186 may be capable of testing the trained machine learning models 190 using corresponding feature sets of the test set from the dataset generator. 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 186 may identify the trained machine learning model 190 with the highest accuracy among all of the trained machine learning models based on the test set.

[0034]

[0039] The machine learning model 190 may refer to a model artifact. This model artifact is created by the training engine 182 using a training set that includes data inputs and corresponding target outputs (correct answers for each training input). Patterns may be found in the dataset that map data inputs to target outputs (correct answers), and the machine learning model 190 is provided with a mapping that captures such patterns. The machine learning model 190 may use one or more of a support vector machine (SVM), a radial basis function (RBF), clustering, supervised machine learning, semi-supervised machine learning, unsupervised machine learning, a k-nearest neighbor algorithm (k-NN), linear regression, random forests, a neural network (e.g., an artificial neural network), etc.

[0035]

[0045] The prediction component 114 may provide the simulation data 167 and the current sensor data 152 to the trained machine learning model 190 and may run the trained machine learning model 190 on the inputs to obtain one or more outputs. The prediction component 114 may be capable of determining (e.g., extracting) prediction data 168 from the outputs of the trained machine learning model 190.

[0036]

[0041] Aspects of the present disclosure describe, by way of example and not limitation, training one or more machine learning models 190 and inputting simulation data 167 and sensor data 142 into the one or more trained machine learning models 190 to determine predicted data 168. In other implementations, a heuristic or rule-based model is used to determine predicted data 168 (e.g., without using a trained machine learning model). The prediction component 114 may monitor historical sensor data 144 and historical performance data 162. Any of the information described in connection with data from data store 140 may also be monitored or otherwise used in a heuristic or rule-based model.

[0037]

[0042] In some embodiments, the functionality of client device 120, prediction server 112, digital representation server 170, and server machine 180 may be provided by fewer machines. For example, in some embodiments, digital representation server 170 and server machine 180 may be combined into a single machine, while in some other embodiments, digital representation server 170, server machine 180, and prediction server 112 may be combined into a single machine. In some embodiments, client device 120 and prediction server 112 may be combined into a single machine.

[0038]

[0043] In general, functionality described in one embodiment as being performed by client device 120, prediction server 112, digital representation server 170, and server machine 180 may also be performed on prediction server 112 in other embodiments, as appropriate. Additionally, functionality performed by a particular component may be performed by different or multiple components operating together. For example, in some embodiments, prediction server 112 may determine corrective actions based on prediction data 168. In another example, client device 120 may determine prediction data 168 based on output from a trained machine learning model.

[0039]

[0044] Additionally, the functionality of a particular component may be performed by different or multiple components working together. One or more of prediction server 112, digital representation server 170, and server machine 180 may be accessed through a suitable application programming interface (API) as a service offered to other systems or devices.

[0040] In embodiments, a "user" may be represented as a single individual. However, other embodiments of the present disclosure encompass a "user" that is an entity controlled by multiple users and / or automated sources. For example, a set of individual users who are federated as a group of administrators may also be considered a "user."

[0041]

[0046] Embodiments of the present disclosure may be applied to data quality assessment, feature enhancement, model evaluation, virtual metrology (VM), predictive maintenance (PdM), marginal optimization, and the like.

[0042]

[0047] Although embodiments of the present disclosure are described with respect to generating predictive data 168 for implementing corrective actions at a manufacturing facility (e.g., a semiconductor manufacturing facility), the embodiments may also be applied generally to component characterization and monitoring. The embodiments may also be applied generally to various types of data-based characterization and monitoring.

[0043]

[0048] The client devices 120, manufacturing equipment 130, sensors 126, measurement equipment 128, prediction server 112, data store 140, digital representation server 170, and server machine 180 may be coupled to each other via network 105 to generate predictive data 168 and implement corrective actions.

[0044]

[0049] In some embodiments, network 105 is a public network that provides client device 120 with access to prediction server 112, data store 140, and other publicly available computing devices. In some embodiments, network 105 is a private network that provides client device 120 with access to manufacturing equipment 130, sensors 126, measurement equipment 128, data store 140, and other privately available computing devices. Network 105 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.

[0045]

[0050] The client device 120 may include a computing device such as a personal computer (PC), a laptop, a mobile phone, a smartphone, a tablet computer, a netbook 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 operator box, etc. The client device 120 may include a corrective action component 122. The corrective action component 122 may receive user input of an indication associated with the manufacturing equipment 130 (e.g., through a graphical user interface (GUI) displayed via the client device 120). In some embodiments, the corrective action component 122 sends an indication to the simulation system 110, receives output (e.g., prediction data 168) from the simulation system 110, determines a corrective action based on the output, and causes the corrective action to be implemented. In some embodiments, the corrective action component 122 obtains sensor data 142 (e.g., current trace data 154) associated with the manufacturing equipment 130 (e.g., from the data store 140, etc.) and provides the sensor data 142 (e.g., current trace data 154) associated with the manufacturing equipment 130 to the simulation system 110. In some embodiments, the corrective action component 122 stores the sensor data 142 in the data store 140, and the prediction server 112 retrieves the sensor data 142 from the data store 140. In some embodiments, the prediction server 112 may store the output (e.g., predicted data 168) of the trained machine learning model(s) 190 in the data store 140, and the client device 120 may retrieve the output from the data store 140. In some embodiments, the corrective action component 122 receives an indication of the corrective action from the simulation system 110 and causes the corrective action to be implemented.Each client device 120 may include an operating system that enables a user to one or more of create, view, or edit data (e.g., views associated with manufacturing equipment 130, corrective actions associated with manufacturing equipment 130, etc.).

[0046]

[0051] Performing a manufacturing process that results in a defective product can be costly in terms of time, energy, product, components, manufacturing equipment 130, costs of identifying defects and discarding the defective product, etc. By inputting sensor data 142 (e.g., manufacturing parameters that are being used or will be used in manufacturing the product), receiving output of predictive data 168, and implementing corrective actions based on the predictive data 168, the system 100 can have the technical advantage of avoiding the costs of creating, identifying, and discarding defective products.

[0047]

[0052] Performing a manufacturing process that results in a component failure of manufacturing equipment 124 can be costly in terms of downtime, product damage, equipment damage, rush ordering of replacement components, etc. By inputting sensor data 142 (e.g., manufacturing parameters that are being or will be used in manufacturing a product), receiving output of predictive data 168, and taking corrective action (e.g., predicted operational maintenance such as replacing, treating, cleaning, etc. components) based on the predictive data 168, system 100 may have the technical advantage of avoiding one or more of the costs of unexpected component failure, unplanned downtime, lost productivity, unexpected equipment failure, product waste, etc.

[0048]

[0053] The corrective action may be associated with one or more of computational process control (CPC), statistical process control (SPC) (e.g., SPC on electronic components to determine process in control, SPC to predict component life, SPC for comparison to 3 sigma graphs, etc.), advanced process control (APC), model-based process control, preventative operational maintenance, design optimization, manufacturing parameter updates, feedback control, machine learning corrections, etc. In some embodiments, the corrective action may include adjusting one or more of the controllable variable tuning elements of the matching network 134, as described in more detail below.

[0049]

[0054] In some embodiments, the corrective action includes providing an alert (e.g., an alert to stop or not perform a manufacturing process when the predictive data 168 indicates a predicted anomaly, such as an anomaly in a product, component, or manufacturing equipment 130). In some embodiments, the corrective action includes providing feedback control (e.g., modifying a manufacturing parameter in response to the predictive data 168 indicating a predicted anomaly). In some embodiments, the corrective action includes providing machine learning (e.g., modifying one or more manufacturing parameters based on the predictive data 168). In some embodiments, implementing the corrective action includes causing an update to one or more manufacturing parameters.

[0050]

[0055] The manufacturing parameters may include hardware parameters (e.g., replacing a component, using a specific component, replacing a processing chip, updating firmware, etc.) and / or process parameters (e.g., temperature, pressure, flow, speed, current, voltage, gas flow, ramp-up speed, etc.). In some embodiments, the corrective action includes causing preventative operational maintenance (e.g., replacing, treating, cleaning, etc., components of the manufacturing equipment 130). In some embodiments, the corrective action includes causing design optimization (e.g., updating manufacturing parameters, manufacturing process, manufacturing equipment 130, etc., for an optimized product). In some embodiments, the corrective action includes updating a recipe (e.g., causing the manufacturing equipment 130 to go into idle mode, sleep mode, warm-up mode, etc.).

[0051]

[0056] 2 is a block diagram illustrating an exemplary embodiment of fabrication equipment 130 in more detail. The components of FIG. 2 are used as examples for illustrative purposes. Thus, various combinations of components (insulators, capacitors, resistors, etc.) may be used in embodiments of the present disclosure. As described above, fabrication equipment 130 includes RF generator 132, matching network 134, process chamber 136, and controller 138.

[0052]

[0057] The matching network 126 may include a first capacitor 226 (also referred to as “C1”) and a second capacitor 228 (also referred to as “C2”). Each of the first capacitor 226 and the second capacitor 228 may be a variable capacitor (tunable or adjustable) capable of tuning the overall impedance of the matching network 134. For example, in embodiments of the present disclosure, matching tuning may be performed by adjusting one or both of the first capacitor 226 and the second capacitor 228 so that the output impedance associated with the processing chamber 136 is adjusted toward the input impedance (e.g., 50 ohms) associated with the RF generator 132. This may maximize the power delivery from the RF generator 132 to the processing chamber 136. In some embodiments, the first capacitor 226 may further be a shunt capacitor, while the second capacitor 228 may be a series capacitor. Series capacitors may be used in transmission lines to provide series compensation for improved power handling capabilities. Shunt capacitors may be applied to electrical systems for multiple tasks in a single application. Additionally, in a series capacitor, reactive power generation is proportional to the square of the load current, while in a shunt capacitor, reactive power generation is proportional to the square of the voltage.

[0053]

[0058] The matching network 134 may include an input sensor 222 and an output sensor 224. The input sensor 222 may determine an impedance associated with the matching network 134. The output sensor 224 may determine an impedance of the RF signal in the process chamber 136.

[0054]

[0059] The matching network 126 may be associated with a first impedance 212 (also referred to as “Z1”) and a second impedance 214 (also referred to as “Z2”). The first impedance 212 may be associated with a parallel path of a first capacitor 226, a resistive element 234 (also referred to as “R1”), and an inductive element 232 (also referred to as “L1”) in the matching network 134. The first impedance 212 may be represented by Equation 1 below: where the term j represents the imaginary unit, the term w represents the angular frequency and is expressed as 2×π×f, where f is the frequency of the RF signal (e.g., 2.2 MHz, 13.56 MHz, 100 MHz, etc.). TIFF2026000982000002.tif17170

[0055]

[0060] The second impedance 214 may be associated with the second capacitor 228. In some embodiments, the second impedance 214 may also be associated with one or more passive devices (e.g., inductors, resistors, etc.). The second impedance 214 may be represented by Equation 2 below: TIFF2026000982000003.tif17170

[0056]

[0061] The processing chamber 136 may be associated with a third impedance 216 (also referred to as "Z3") associated with impedances from processing chamber components, transition line impedances, feed rod impedances (from the matching network 134 to the processing chamber 136), etc. As an example, Z3 may be represented using a real part Rp 238 and an imaginary part Lp 236 of the impedance associated with the processing chamber 136. Rp and Lp may be measured in real time using the input impedance sensor 224. The third impedance 216 may be represented by Equation 3 below: TIFF2026000982000004.tif14170

[0057]

[0062] Therefore, the total load impedance of the manufacturing equipment 130 may be expressed by Equation 4 below: TIFF2026000982000005.tif22170

[0058]

[0063] 3 is a flow diagram of a method 300 for generating predictive data to enable corrective action to be taken, according to certain embodiments. Method 300 may be performed by processing logic, which may include hardware (e.g., circuitry, dedicated logic, programmable logic, microcode, a processing device, etc.), software (e.g., instructions running on a processing device, a general-purpose computer system, or a dedicated machine), firmware, microcode, or a combination thereof. In some embodiments, method 300 may be performed in part by simulation system 110. In some embodiments, a non-transitory storage medium stores instructions that, when executed by a processing device (e.g., simulation system 110, server machine 180, prediction server 112, etc.), cause the processing device to perform method 300.

[0059]

[0064] For simplicity of explanation, method 300 is shown and described as a series of steps. However, steps in accordance with this disclosure may occur in various orders and / or simultaneously with and in conjunction with other steps not shown and described herein. Moreover, not all of the illustrated steps may be performed to implement method 300 in accordance with the disclosed subject matter. Additionally, those skilled in the art will understand and appreciate that method 300 may alternatively be represented as a series of interrelated states via a state diagram or state events.

[0060]

[0065] In block 310, processing logic implementing the method 300 causes the RF generator 132 to generate an RF signal to energize the processing chamber.

[0061]

[0066] At block 312, processing logic updates the impedance values ​​in the digital replica. In particular, processing logic may update the impedance associated with the match network model 172 based on current trace data generated from the input sensor 222 and update the impedance associated with the process chamber model 174 based on current trace data 154 generated from the output sensor 224.

[0062]

[0067] At block 314, processing logic may determine whether the digital replica meets an accuracy threshold criterion based on the current impedance value. For example, processing logic may use a cost function formula (e.g., root mean square difference) to determine whether the match network model 172 and the process chamber model 174 are within an acceptable error range.

[0063]

[0068] As an example, processing logic may utilize the following equation 5: where the term J1 represents a tolerance and the term Z input represents the impedance determined by the input sensor 222, and the term angle represents the frequency and phase of the RF signal. TIFF2026000982000006.tif15170

[0064]

[0069] It is understood that the closer the term J1 is to zero, the more accurate the matching network model 172 is to the parameters of the physical elements and dynamics of the matching network 134. Constraints may be placed on the terms associated with Equation 5. In some embodiments, the values ​​of the first capacitor 226 and the second capacitor 228 are subject to an upper limit (C1 max and C2 max ) and lower bound (C1 min and C2 min These values ​​are expressed as follows: C1 min <C1<C1 max ;C2 min <C2<C2 maxIn response to the digital replica meeting the threshold criteria (e.g., the match network model 172 and / or the process chamber model 174 are each within the tolerance range), processing logic proceeds to block 318. In response to the digital replica not meeting the threshold criteria, processing logic proceeds to block 316.

[0065]

[0070] At block 316, processing logic optimizes the digital replica. In one embodiment, processing logic may use Equation 5 to determine optimal values ​​for the resistive element R1 and / or the inductive element L1 used in matching network model 172. For example, processing logic may update matching network model 172 by adjusting one or more parameters (e.g., L1, R1, C1, C2, etc.) until the term J1 meets an accuracy threshold criterion. In one embodiment, processing logic uses an update algorithm to adjust one or more parameters (e.g., L p or R p The process chamber model 174 may be updated by adjusting the parameters (e.g., the process flow, the temperature, the temperature, the temperature, the temperature) and the like. The adjusted parameters may be stored as simulation data 167.

[0066]

[0071] At block 318, processing logic obtains one or more outputs from simulation system 110 that are indicative of predicted data. For example, processing logic may determine optimal values ​​for first capacitor 226 and second capacitor 228 to minimize the reflected RF signal. In some embodiments, processing logic may use prediction component 114 to perform a closed-loop simulation using simulation data 167 and model 190. Prediction component 114 may generate one or more optimization profiles that may indicate values ​​to which first capacitor 226 and second capacitor 228 are adjusted to minimize the reflected RF signal. In some embodiments, prediction component 114 may further utilize a tuning time parameter to minimize the time required to reach the optimal value.

[0067]

[0072] As an example, processing logic may use the following equation 6 to generate the optimized profile: TIFF2026000982000007.tif17170

[0068]

[0073] where the term J2 represents a cost function and is the total impedance (e.g., Z total ), and the term w1 may be set to the reflection coefficient ( TIFF2026000982000008.tif9170 or γ (gamma)), and the term T tune is the tuning time, the term w2 is a weighting factor associated with the tuning time, and the term Z0 represents the target impedance (e.g., the impedance of RF generator 132, such as 50 ohms). In some embodiments, the following constraints are imposed on Equation 6: TIFF2026000982000009.tif60170

[0069]

[0074] The prediction server may use Equation 6 to generate one or more optimization profiles for the first capacitor 226 and the second capacitor 228. The optimization profiles may indicate reflection coefficient values ​​for each tuning of the first capacitor 226 and the second capacitor 228 combination. The optimization profiles may also indicate tuning paths to reach target values ​​for the capacitors. The capacitor target values ​​may set parameters for the first capacitor 226 and the second capacitor 228 that have the smallest achievable reflection coefficient.

[0070]

[0075] 4A-C are graphs illustrating exemplary optimization profiles, according to certain embodiments. Specifically, FIG. 4A illustrates reflection coefficient (γ) values ​​for each of the initial tuning values ​​(e.g., 0-100%) of the first capacitor 226 and the second capacitor 228. Point 410 illustrates the minimum achievable reflection coefficient of the optimization profile associated with the first capacitor 226 tuned to approximately 86% and the second capacitor tuned to approximately 30% (the capacitor target value).

[0071]

[0076] 4B shows the time required to reach the capacitor target values ​​for each set of initial values ​​for the first capacitor 226 and the second capacitor 228. For example, if the first capacitor 226 has a tuning value of 50% and the second capacitor 228 has a tuning value of 0% (e.g., point 430), the corrective action component 122 requires 1.25 seconds to tune the first capacitor 226 and the second capacitor 228 to the capacitor target values ​​(e.g., 86% and 30%, respectively).

[0072]

[0077] 4C shows a gradient-based match tuning optimization profile, which may be defined by the following cost function: TIFF2026000982000010.tif17170

[0073]

[0078] By incorporating the partial derivative of the cost function into Equation 7, the control parameters of the optimization model can be determined. For example, the following equation shows the partial derivative of the cost function: TIFF2026000982000011.tif34170

[0074]

[0079] The partial derivatives may lead to a gradient control value that defines the gradient tuning. For example, the gradient control law may lead to the following gradient control: TIFF2026000982000012.tif24170TIFF2026000982000013.tif24170

[0075]

[0080] The positive parameters k1, k2, g1, and g2 are controller gains. The values ​​of c1 and c2 are thresholds for switching from gradient control to linear control. Switching between gradient control and linear control can be used, for example, to eliminate or otherwise reduce chatter or noise in the steady-state portion of the optimization profile.

[0076]

[0081] In some embodiments, the optimization profile may be based on a Jacobian-based RF match tuning algorithm. In the context of controlling RF match, the Jacobian may predict sudden changes in capacitance (e.g., first capacitor C1 226 and second capacitor C2 228 in FIG. 2 ) caused by past and future iterations of the optimization profile. As previously mentioned, the matching network (e.g., matching network 134 in FIG. 1 ) may operate to minimize reflected RF energy by matching the impedance of a plasma used in an etch process to the impedance of an RF signal. For example, a desired real-valued impedance (e.g., 50 ohms, 75 ohms, etc.) may be matched by the plasma. Using Equation 4, the input impedance may be expressed as: TIFF2026000982000014.tif18170

[0077]

[0082] From equation 12, the Jacobian matrix can be calculated as follows: TIFF2026000982000015.tif32170

[0078]

[0083] In some embodiments, a Jacobian controller may be used as defined by the following equation: TIFF2026000982000016.tif20170

[0079]

[0084] where C(t) = [C1C2], e(t) = [50-abs(Z);-angle(Z)], and K is a positive definite gain matrix (e.g., Equation 13). In a further embodiment, to adjust the loop gain of the amplitude-phase control system, the Jacobian J c can be corrected.

[0080]

[0085] In a further embodiment, the Jacobian matrix J c A damping constant λ may be used to prevent singularities in the inverse calculation of λ and further facilitate processing logic's identification of the inverse Jacobian matrix. In another embodiment, a non-direct matrix solver may be used to calculate and update the tuning values ​​without directly solving for the inverse Jacobian matrix. For example, methods such as matrix least squares solvers, iterative solvers, and matrix decomposition solvers (e.g., upper and lower triangular solvers, singular value decomposition, etc.) may be used, as detailed in other embodiments.

[0081]

[0086] In some embodiments, variance control of the optimization profile may be utilized through singular value decomposition (SVD). For example, the SVD of the Jacobian matrix may be expressed as: TIFF2026000982000017.tif15170

[0082]

[0087] U and V are orthogonal matrices, and D is a diagonal matrix. For example, D has non-zero values ​​along the diagonal elements of the matrix (e.g., σ along the diagonal). i =d i,_j ) in the absence of singularities, the diagonal elements decrease down the rows and across the columns of the matrix (e.g., σ1 ≥ σ2 ... > 0). Setting D as a diagonal matrix results in decentralized control of the optimization model. For example, the impedance amplitude and phase control loops can be tuned independently by adjusting one or more gains of the diagonal matrix D.

[0083]

[0088] In some embodiments, an RF matching circuit optimization model may not meet an accuracy threshold. For example, the optimization models described herein (e.g., FIGS. 4A-C) may rely on values ​​(e.g., theoretical or generalized values) that misrepresent the unique characteristics of a particular RF matching circuit. In some embodiments, calibration, such as the following, may be used to more accurately measure parameters and values ​​to determine and predict a more accurate optimization profile.

[0084]

[0089] The processing logic may follow a calibration procedure to update the optimization parameters (e.g., to improve the precision and accuracy of the optimization profile). The processing logic may use an RF match auto-tuning mode to find tuning positions for a first capacitor (e.g., shunt capacitor C1) and a second capacitor (e.g., series capacitor C2). The processing logic may measure the Jacobian using a perturbation calculation of the tuned impedance position. For example, the processing logic may move C1 / C2 to a matched position (e.g., by setting the RF match to manual control). The processing logic may calculate the impedance across the C1 delta without making any changes to C2 (e.g., TIFF2026000982000018.tif7170=0.5). In this example, the processing logic changes C1 to C1+ TIFF2026000982000019.tif7170 and reads the input impedance R1 for a short period of time (e.g., 10 seconds). The processing logic sets C1 to C1- TIFF2026000982000020.tif7170 and read the matched input impedance R2 over a short period of time (e.g., 10 seconds). Processing logic may calculate the impedance across the C2 delta without making any changes to C1 (e.g., TIFF2026000982000021.tif7170=0.5). In this example, the processing logic converts C2 to C2+ TIFF2026000982000022.tif7170 and reads the input impedance R3 for a short period of time (e.g., 10 seconds). The processing logic sets C2 to C2- TIFF2026000982000023.tif7170 and read the matched input impedance R4 over a short period of time (e.g., 10 seconds). Processing logic may then calculate the Jacobian using the following equation and the parameter values ​​set forth in this calibration example: TIFF2026000982000024.tif72170

[0085]

[0090] The Jacobian specified in Equation 20 can then be used in conjunction with other optimization profiles (eg, SVD, damping, etc.) as described herein.

[0086]

[0091] In some embodiments, processing logic may continuously generate predicted data (e.g., an optimized profile) until the predicted data meets a threshold criterion based on weighting factors. Specifically, the weighting factors may be set to 1 (e.g., w1 + w2 = 1 or 100%). A user may change the weighting factors based on a preference for minimum reflected power over tuning time. For example, a user may set w2 to a high value if minimizing match tuning time is a priority. A user may set w1 to a high value if minimizing reflected power is a priority.

[0087]

[0092] At block 320, processing logic causes corrective action to be taken based on the prediction data. For example, processing logic may select a tuning path from the optimization profile. Controller 138 may adjust first capacitor 226 and / or second capacitor 228 based on the tuning path. FIG. 5 is a graph illustrating multiple tuning paths associated with first capacitor 226 and / or second capacitor 228. Specifically, FIG. 5 illustrates tuning paths in which initial tuning values ​​(e.g., from 0 to 100%) of first capacitor 226 and second capacitor 228 each move toward a capacitor target value.

[0088]

[0093] 6 is a block diagram illustrating a computer system 600 according to certain embodiments. In some embodiments, computer system 600 may be connected to other computer systems (e.g., via a network such as a local area network (LAN), an intranet, an extranet, or the Internet). Computer system 600 may operate as a server or a client computer in a client-server environment, or as a peer computer in a peer-to-peer or distributed network environment. Computer system 600 may be provided by a personal computer (PC), a tablet PC, a set-top box (STB), a personal digital assistant (PDA), a mobile phone, a web appliance, a server, a network router, switch, or bridge, or any device capable of executing a set of instructions (sequential or otherwise) that specify actions to be taken by the device. Furthermore, the term "computer" is intended to include any collection of computers that individually or collectively execute a set of instructions (or multiple sets) to perform any one or more of the methods described herein.

[0089]

[0094] In a further aspect, computer system 600 may include a processing device 602, a volatile memory 604 (e.g., random access memory (RAM)), a non-volatile memory 606 (e.g., read-only memory (ROM) or electrically erasable programmable ROM (EEPROM)), and a data storage device 616, which may communicate with each other via a bus 608.

[0090]

[0095] The processing device 602 may be provided by one or more processors, such as a general-purpose processor (e.g., a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a microprocessor implementing another type of instruction set, or a microprocessor implementing a combination of multiple types of instruction sets), or a specialized processor (e.g., an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), or a network processor).

[0091]

[0096] Computer system 600 may further include a network interface device 622 (e.g., coupled to a network 674). Computer system 600 may also include a video display unit 610 (e.g., an LCD), an alphanumeric input device 612 (e.g., a keyboard), a cursor control device 614 (e.g., a mouse), and a signal generating device 620.

[0092]

[0097] In some implementations, the data storage device 616 may include a non-transitory computer-readable storage medium 624. The non-transitory computer-readable storage medium 624 may store instructions 626 for implementing the methods described herein, encoding any one or more of the methods or functions described herein, including instructions encoding the components of FIG. 1 (e.g., corrective action 122, prediction component 114, etc.).

[0093]

[0098] The instructions 626 may reside, completely or partially, in the volatile memory 604 and / or the processing device 602 during execution by the computer system 600, and thus the volatile memory 604 and the processing device 602 may constitute machine-readable storage media.

[0094]

[0099] Although the computer-readable storage medium 624 is illustrated as a single medium in the illustrative embodiment, the term "computer-readable storage medium" is intended 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 executable instructions. The term "computer-readable storage medium" is also intended to include any tangible medium that is capable of storing or encoding a set of instructions that are executed by a computer to cause the computer to perform any one or more of the methods described herein. The term "computer-readable storage medium" includes, but is not limited to, solid-state memory, optical media, and magnetic media.

[0095]

[0100] The methods, components, and features described herein may be implemented by multiple separate hardware components or may be integrated into the functionality of other hardware components (such as ASICS, FPGAs, DSPs, or similar devices). In addition, the methods, components, and features may be implemented by firmware modules or functional circuitry within a hardware device. Furthermore, the methods, components, and features may be implemented in any combination of hardware devices and computer program components, or in a computer program.

[0096]

[0101] Unless otherwise indicated, words such as "receiving," "performing," "providing," "obtaining," "causing," "accessing," "determining," "adding," "using," "training," and the like refer to actions and processes performed or implemented by a computer system that manipulates and transforms data represented as physical (electronic) quantities in the computer system's registers and memory into other data similarly represented as physical quantities in the computer system's memory or registers or other devices that store, transmit, or display such information. Additionally, as used herein, the terms "first," "second," "third," "fourth," and the like are intended to distinguish between various elements and may not have a sequential meaning according to their numerical designations.

[0097]

[0102] The examples described herein also relate to apparatus for performing the methods described herein. This apparatus may be specially constructed to perform the methods described herein, or may comprise a general-purpose computer system selectively programmed by a computer program stored on the computer system. Such a computer program may be stored on a computer-readable tangible storage medium.

[0098]

[0103] The methods and embodiments described herein are not inherently related to any particular computer or other apparatus. Many different general-purpose systems may be used in accordance with the teachings described herein, or it may prove convenient to construct more specialized apparatus to perform the methods described herein and / or each of their individual functions, routines, subroutines, or steps. Examples of structures for a variety of these systems are set forth above.

[0099]

[0104] The above description is intended to be illustrative, not limiting. While the present disclosure has been described with reference to particular examples and implementations, it will be recognized that the present disclosure is not limited to the described examples and implementations. The scope of the present disclosure should be determined with reference to the following claims, along with the full scope of equivalents to which such claims are entitled.

Claims

1. causing a manufacturing device to generate an RF signal to energize a processing chamber associated with the manufacturing device; receiving current trace data associated with the RF signal from one or more sensors associated with the manufacturing equipment; updating impedance values ​​of digital replicas associated with the manufacturing equipment based on the current trace data; obtaining one or more outputs from the digital replica indicative of predicted data; causing implementation of one or more corrective actions associated with the manufacturing equipment based on the predictive data; A method comprising:

2. 10. The method of claim 1, further comprising: performing optimization of the digital replica in response to determining that the digital replica does not meet an accuracy threshold criterion based on the current trace data.

3. updating the impedance value of the digital replica; 10. The method of claim 1, further comprising updating a process chamber model associated with the digital replica using trace data from an output impedance sensor, the output impedance sensor being associated with a matching network of the manufacturing equipment.

4. updating the impedance value of the digital replica; 2. The method of claim 1, further comprising: updating a matching network model associated with the digital replica using trace data from an input impedance sensor and an output impedance sensor, the input impedance sensor and the output impedance sensor being associated with a matching network of the manufacturing equipment.

5. The method of claim 1 , wherein the one or more outputs indicative of the predicted data are generated using a trained machine learning model.

6. The method of claim 1 , wherein the predictive data includes one or more tuning settings for one or more variable capacitors associated with a matching network of the manufacturing equipment.

7. The method of claim 1 , wherein the corrective action comprises adjusting one or more variable capacitors associated with a matching network of the manufacturing equipment based on the predictive data.

8. Memory and a processing device coupled to the memory, the processing device: causing a manufacturing device to generate an RF signal to energize a processing chamber associated with the manufacturing device; receiving current trace data associated with the RF signal from one or more sensors associated with the manufacturing equipment; updating impedance values ​​of digital replicas associated with the manufacturing equipment based on the current trace data; obtaining one or more outputs from the digital replica indicative of predicted data; and causing implementation of one or more corrective actions associated with the manufacturing equipment based on the predictive data. system.

9. the processing device further comprising: The system of claim 8 , further comprising: performing optimization of the digital replica in response to determining that the digital replica does not meet an accuracy threshold criterion based on the current trace data.

10. To update the impedance value of the digital replica, the processing device further 10. The system of claim 8, wherein trace data from an output impedance sensor is used to update a process chamber model associated with the digital replica, the output impedance sensor being associated with a matching network of the manufacturing equipment.

11. To update the impedance value of the digital replica, the processing device further 9. The system of claim 8, wherein trace data from an input impedance sensor and an output impedance sensor are used to update a matching network model associated with the digital replica, the input impedance sensor and the output impedance sensor being associated with a matching network of the manufacturing equipment.

12. The system of claim 8 , wherein the one or more outputs indicative of the predicted data are generated using a trained machine learning model.

13. 10. The system of claim 8, wherein the predictive data includes one or more tuning settings for one or more variable capacitors associated with a matching network of the manufacturing equipment.

14. 10. The system of claim 8, wherein the corrective action comprises adjusting one or more variable capacitors associated with a matching network of the manufacturing equipment based on the predictive data.

15. A non-transitory machine-readable storage medium storing instructions that, when executed, cause a processing device to: causing a manufacturing device to generate an RF signal to energize a processing chamber associated with the manufacturing device; receiving current trace data associated with the RF signal from one or more sensors associated with the manufacturing equipment; updating impedance values ​​of digital replicas associated with the manufacturing equipment based on the current trace data; obtaining one or more outputs from the digital replica indicative of predicted data; and causing implementation of one or more corrective actions associated with the manufacturing equipment based on the predictive data. Non-transitory machine-readable storage medium.

16. The process further comprises:

16. The non-transitory machine-readable storage medium of claim 15, further comprising: performing optimization of the digital replica in response to determining that the digital replica does not meet an accuracy threshold criterion based on the current trace data.

17. updating the impedance value of the digital replica; 16. The non-transitory machine-readable storage medium of claim 15, further comprising updating a processing chamber model associated with the digital replica using trace data from an output impedance sensor, the output impedance sensor being associated with a matching network of the manufacturing equipment.

18. updating the impedance value of the digital replica; 16. The non-transitory machine-readable storage medium of claim 15, further comprising updating a matching network model associated with the digital replica using trace data from an input impedance sensor and an output impedance sensor, the input impedance sensor and the output impedance sensor being associated with a matching network of the manufacturing equipment.

19. 16. The non-transitory machine-readable storage medium of claim 15, wherein the one or more outputs indicative of the predictive data are generated using a trained machine learning model.

20. 16. The non-transitory machine-readable storage medium of claim 15, wherein the predictive data includes one or more tuning settings of one or more variable capacitors associated with a matching network of the manufacturing equipment, and the corrective action includes adjusting one or more variable capacitors associated with a matching network of the manufacturing equipment based on the predictive data.

Images