Verification to improve the quality of maintenance of manufacturing equipment
A system using sensors and models to validate and correct maintenance processes in manufacturing equipment addresses unscheduled downtime by ensuring each step is satisfactorily performed, enhancing maintenance efficiency and reducing costs.
Patent Information
- Application Number
- JP2024522246
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-10-13
- Filing Date
- 2022-10-12
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2042-10-12
Smart Images

Figure 2024542937000001_ABST
Abstract
Description
[Technical field]
[0001] This disclosure relates to a maintenance process for manufacturing equipment. More particularly, this disclosure relates to a system and method for verifying one or more maintenance operations to improve the quality of maintenance of manufacturing equipment. [Background technology]
[0002] Processing equipment is used in many types of processing systems. Examples of processing equipment include etch chambers, deposition chambers, anneal chambers, implantation chambers, and the like. Typically, a substrate, such as a semiconductor wafer, is placed in the process chamber, and conditions within the process chamber are set and maintained to process the substrate. Some processing operations, such as polishing processes, may utilize equipment that is not limited to chambers. Processing of substrates (e.g., many wafers, many processing times, etc.) wears out components of processing equipment (also called manufacturing equipment). Manufacturing equipment, such as chambers, are often taken offline at predetermined intervals to perform preventive maintenance. The maintenance process may include many maintenance steps and / or operations. When preventive maintenance is performed satisfactorily, the manufacturing equipment may be returned to production (e.g., used to process substrates) as scheduled. When preventive maintenance is not performed satisfactorily, unscheduled downtime that does not produce profits may occur. Summary of the Invention
[0003] The following is a simplified summary of the disclosure to provide a basic understanding of some aspects of the disclosure. This summary is not an extensive overview of the disclosure. It is not intended to identify key or critical elements of the disclosure, nor is it intended to limit the scope of particular embodiments of the disclosure or the claims. Its sole purpose is to present some concepts of the disclosure in a simplified form as a prelude to the more detailed description that is presented later.
[0004] One aspect of the disclosure describes a system including a server and manufacturing equipment, the system configured to initiate a maintenance process for the manufacturing equipment. The maintenance process includes several maintenance phases. The system is further configured to provide instructions to a user device directing the user device to perform a first maintenance phase. The system is further configured to receive data derived from the performance of the first maintenance phase. The system is further configured to determine that the performance of the first maintenance phase was unsatisfactory. The system is further configured to cause the user device to display instructions for a corrective action to perform.
[0005] In another aspect of the disclosure, a method includes initiating, by a server, a maintenance process for manufacturing equipment. The maintenance process includes several maintenance phases associated with several components of the manufacturing equipment. The method further includes providing instructions to a user device directing the user device to perform a first phase of the maintenance process. The method further includes receiving data resulting from performance of the first maintenance phase. The method further includes determining that performance of the first maintenance phase was unsatisfactory. The method further includes causing corrective action to be performed.
[0006] In another aspect of the disclosure, a method includes receiving, by a user device, instructions to perform a first maintenance step of a maintenance process. The maintenance process includes a number of maintenance steps. The maintenance process is associated with manufacturing equipment. The method further includes providing data derived from the performance of the first maintenance step to a server. The method further includes receiving validation data from the server indicating that the performance of the first maintenance step was unsatisfactory. The method further includes causing a corrective action to be performed in consideration of the validation data.
[0007] In the figures of the accompanying drawings, the present disclosure is illustrated by way of example and not by way of limitation. [Brief description of the drawings]
[0008] [Figure 1] FIG. 1 is a block diagram illustrating an example system (example system architecture) according to some embodiments. [Diagram 2] FIG. 2 is a block diagram of an example dataset generator used to generate a dataset for a model, according to some embodiments. [Diagram 3] FIG. 1 is a block diagram illustrating a system for generating output data (e.g., prediction data) according to some embodiments. [Figure 4A-4F] 1 is a flow diagram of a method for causing corrective actions to be taken in connection with generating predictive data, according to some embodiments. [Diagram 5] FIG. 1 illustrates the operation of a system for improving maintenance processes, according to some embodiments. [Figure 6] FIG. 1 is a block diagram illustrating a computer system according to some embodiments. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0009] Described herein is technology directed to a system for initiating and verifying a stage of a maintenance process associated with manufacturing equipment. Manufacturing equipment is used to process substrates, such as semiconductor wafers. The characteristics of these substrates are controlled by the conditions under which the substrates are processed. Various components of the manufacturing equipment contribute to consistent processing of substrates. Substrate processing (e.g., processing several substrates, the time it takes to process substrates, etc.) results in wear on the components of the manufacturing equipment. To maintain consistent production of substrates, the worn components may be periodically maintained. In some embodiments, maintenance may be initiated based on the amount of time the manufacturing equipment has been used to process wafers, the amount of time the manufacturing equipment has been in production, or the like, since the last maintenance event. In some embodiments, maintenance may be initiated based on the number of wafers processed, sensor data indicative of equipment performance, metrology data of completed substrates, etc.
[0010] In some cases, the maintenance process may be performed quickly, for example, when a small number of components are to be maintained, when the components to be maintained are easily accessible, etc. In some cases, the maintenance process may be extensive, including many maintenance steps associated with many components of the manufacturing equipment. In some cases, verification that the maintenance process was performed satisfactorily (or that a particular step or steps or processes were performed satisfactorily) may be limited. In one example, the maintenance process may include venting a manufacturing chamber used to process substrates under vacuum (which may include performing cleaning procedures, purging procedures, turning off components such as heating and cooling devices, etc.), removing and / or performing maintenance on various components (e.g., process kits, chamber liners, lift pins, chucking assemblies, valves, actuators, windows, sensors, etc.), and reassembling the chamber. Verification that the maintenance process was performed satisfactorily (e.g., performing the maintenance resulted in equipment that meets specifications) may not be performed until the chamber is reassembled. Sensors may then verify that the chamber is properly pumped back down to vacuum, that seasoning of the chamber is proceeding properly, that acceptable substrates are being produced, etc.
[0011] In some cases, after a maintenance event (e.g., after a step in a maintenance process is completed), the manufacturing equipment may fail a qualification test (e.g., a manufacturing or processing chamber is found not to meet a performance metric). In some cases, after failing a qualification test, the manufacturing equipment goes into an unscheduled downtime period, possibly including troubleshooting, additional maintenance, etc., to return the equipment to an operable state (e.g., meeting a performance metric). Between the initiation of the maintenance process and receiving validation that the maintenance process was not performed satisfactorily, many components may be adjusted, cleaned, replaced, maintained, etc. It may be difficult to predict which component or components are responsible for the manufacturing equipment not meeting a performance metric, which maintenance steps should be repeated, etc. A maintenance step may be chosen to be repeated, or the entire maintenance process may be repeated, to correct a problem introduced during the manufacturing process. Such an approach involves experimentation to return the equipment (e.g., a manufacturing chamber) to an operable state. Experimentation involves costs in terms of time spent, materials used, labor, etc. Furthermore, after an extended maintenance event, the manufacturing equipment may fail qualification tests again, which may result in, for example, further increased downtime, reduced productivity, and wasted materials. To further exacerbate these issues, some metrics for qualifying manufacturing equipment (e.g., measurements of substrates produced in a substrate manufacturing chamber) may not directly indicate the success of the maintenance process (e.g., a particular substrate may not meet performance metrics even though the maintenance was performed correctly), and qualifying manufacturing equipment using such metrics may lead to unnecessary downtime due to false negative qualification tests.
[0012] The methods and devices of the present disclosure address at least some of these deficiencies of conventional approaches. The present disclosure enables the use of a system that provides a stage-by-stage verification of the success or failure of a stage or operation of a maintenance process. In some embodiments, the system of the present disclosure includes a server. In some embodiments, the system further includes a user device. In some embodiments, the system further includes a sensor associated with the manufacturing equipment (e.g., a substrate processing chamber) and a connected smart fixture. The system may be configured to provide instructions to the user device directing the execution of a first maintenance stage of the maintenance process. The system may further be configured to provide information to the server directing the execution of the maintenance stage. The system may further be configured to determine whether the maintenance stage has been performed satisfactorily. The system may further be configured to provide a verification notification to the user device indicating the quality of the execution of the maintenance stage.
[0013] In some embodiments, verifying that the maintenance step was performed satisfactorily may include comparing data values received by the processing device to a threshold (e.g., comparing a pressure sensor value to a threshold to ensure that the chamber can achieve sufficient vacuum). In some embodiments, verifying that the maintenance step was performed satisfactorily may include providing data from the performance of the maintenance step to a physics-based model (e.g., comparing the operation of a mass flow controller to the operation of a digital twin of the mass flow controller). In some embodiments, verifying that the maintenance step was performed correctly may include providing data from the performance of the maintenance step to a trained machine learning model (e.g., pictures of components cleaned during the maintenance step, such as an electrostatic chuck, may be provided to the trained machine learning model for processing, and the images may be processed to determine whether a part was placed correctly or whether a fastener was used properly). In some embodiments, the user device may be configured to display information related to the verification of the satisfactory performance of the maintenance step to the user. In some embodiments, the user device may be further configured to display a suggested action to the user in view of the verification data. In some embodiments, the user device may be further configured to gather information (e.g., generate images of the manufacturing equipment) that can be used to determine whether one or more maintenance steps or operations were performed correctly.
[0014] In some embodiments, the method of the present disclosure further includes initiating, by the server, a maintenance process for the manufacturing equipment. The method further includes providing instructions to a user device directing the user device to perform a first maintenance phase of the maintenance process. The method further includes receiving (e.g., at the server computing device or controller) data derived from a first execution of the first maintenance phase. The method further includes determining that the first execution of the first maintenance phase was unsatisfactory. The method further includes causing the user device to display instructions of a corrective action to perform.
[0015] In some embodiments, the method of the present disclosure further includes receiving, by the user device, instructions directing the execution of a first maintenance phase of the maintenance process. The method further includes providing data from the first execution of the first maintenance phase to a server. The method further includes receiving validation data from the server indicating that the execution of the first maintenance phase was unsatisfactory. The method further includes causing corrective action to be performed in consideration of the validation data.
[0016] Aspects of the present disclosure result in technical advantages over conventional methods. Verifying the quality of maintenance after one stage of the maintenance process limits the number of components to double-check, the number of maintenance operations, etc. in case of failure of a certification text or verification failure. Furthermore, accumulated knowledge may be stored on a server and appropriate corrective actions may be provided to a user via a user device. This increases the accuracy and efficiency of troubleshooting a failed maintenance stage.
[0017] 1 is a block diagram illustrating an example system 100 (example system architecture) according to certain embodiments. System 100 includes user devices 120, manufacturing equipment 124, sensors 126, measurement equipment 128, connected tools 129, a prediction server 112, and a data store 140. Prediction server 112 may be part of a prediction system 110. Prediction system 110 may further include server machines 170 and 180.
[0018] The sensors 126 (e.g., sensors associated with the manufacturing equipment 124) may provide sensor data 142 associated with the manufacturing equipment 124 (e.g., indicating conditions associated with the manufacturing equipment 124, e.g., conditions associated with the manufacturing equipment 124 during a stage of a maintenance event). The sensor data 142 may further include data recorded by sensors during times other than a maintenance event, such as during production, qualification testing, etc. The sensor data 142 may further include data from sensors not associated with the manufacturing equipment 124. A connected instrument 129 (also referred to as a connected tool) may generate the data included in the sensor data 142. The connected instrument 129 may be a tool connected to the user device 120 and / or the server machine 170 via a wired or wireless connection. In one embodiment, the connected instrument 129 may include a power source and may also include a wireless module that connects the connected instrument 129 to the user device 120 and / or the server machine 170 via a wireless connection, such as a Bluetooth connection, a Wi-Fi connection, a Zigbee connection, etc. For example, a connected torque wrench may provide information from a sensor in the connected torque wrench indicating the torque with which the connected torque wrench tightened a component, a connected screwdriver may provide information indicating how many turns the connected screwdriver turned a screw, and the like. Additionally, the user device 120 may generate data included in the sensor data 142. For example, the user device 120 may include a camera that can be used to generate image or video data, which may be included in the sensor data 142. The sensor data 142 may be used, for example, to determine the health of the equipment and / or the health of a product (e.g., the quality of the product). The manufacturing equipment 124 may produce a product according to a recipe or by executing a run over a period of time. In some embodiments, the sensor data 142 may include one or more values of temperature (e.g., the temperature of a heating device), spacing (SP), pressure, high frequency radio frequency (HFRF), electrostatic chuck (ESC) voltage, current, flow rate (e.g., the flow rate of one or more gases), power, voltage, etc.The sensor data 142 may include historical sensor data 144 and current sensor data 146 .
[0019] The manufacturing equipment 124 may be configured for a maintenance event based on the maintenance data 150. The maintenance data 150 may relate to or indicate parameters for a maintenance step, such as hardware parameters (e.g., settings or components (e.g., size, type, etc.) of the manufacturing equipment 124) and / or process parameters of the manufacturing equipment (e.g., gas flow rates, pressure, heating and cooling controllers, etc.). The maintenance data 150 may further include instructions for various maintenance steps or operations of a maintenance process. The instructions may include descriptions in the form of written text, images, videos, sounds, means for controlling parameters of the manufacturing equipment 124, etc. The maintenance data 150 may include process recipes for performing the maintenance steps. The maintenance data may include metrics for assigning a score indicating success of the maintenance event, process, step, operation, etc. The maintenance data 150 may include instructions to execute in case of failure of verification, qualification, etc. of the maintenance event, process, step, operation, etc. The instructions may include instructions provided to a user (e.g., via user device 120) or directly to manufacturing equipment 124 to configure manufacturing equipment 124 according to certain parameters for maintenance. Maintenance data 150 may include historical maintenance data 152 and / or current maintenance data 154. Maintenance data 150 may indicate input settings for a manufacturing device for a portion of a maintenance process (e.g., heating power, gas flow rate, etc.). Sensor data 142 may be provided while maintenance associated with manufacturing equipment 124 is being performed (e.g., equipment readings as the maintenance process is being performed). Sensor data 142 may be different for each maintenance event.
[0020] In some embodiments, the sensor data 142 may be processed (e.g., by the user device 120, the prediction server 112, the server machine 170, etc.). Processing the sensor data 142 may include generating or determining features of the sensor data 142. In some embodiments, the features are patterns of the sensor data 142 (e.g., slope, width, height, peaks, etc.) or combinations of values from the sensor data 142 (e.g., power derived from voltage and current, etc.). The sensor data 142 may include features that may be used by the prediction component 116 to perform signal processing and / or to obtain the prediction data 168, and possibly to perform corrective actions. The corrective actions may include issuing an alert to a user, providing a user with maintenance instructions (e.g., instructions to perform one or more maintenance operations, instructions to repeat a maintenance operation, etc.), updating a process recipe (e.g., a maintenance process recipe), etc.
[0021] Each instance (e.g., set) of sensor data 142 may correspond to a particular maintenance event (e.g., execution of a maintenance process including several maintenance steps on manufacturing equipment 124), a type of maintenance (e.g., a particular set of manufacturing equipment 124, a type of chamber, a process including set maintenance steps, etc.), or the like. Maintenance data 150 may include sets having similar characteristics.
[0022] Data store 140 may include metrology data 169 collected using metrology equipment 128 and related to products processed using manufacturing equipment 124. Metrology data 169 may include sets having similar characteristics as sensor data 142 and maintenance data 150. Data store 140 may store information relating different data sets. For example, data may be stored that indicates that multiple data sets correspond to the same product (e.g., metrology data 169 representing measurements of the product, sensor data 142 relating to a maintenance event involving the manufacturing equipment used to process the product, etc.).
[0023] User devices 120, manufacturing equipment 124, sensors 126, measurement equipment 128, connected instruments 129, prediction server 112, data store 140, server machine 170 and server machine 180 may be coupled to each other via network 130 to generate predictive data 168 and, optionally, to perform corrective actions.
[0024] In some embodiments, network 130 is a public network that provides user device 120 with access to prediction server 112, data store 140, and / or other public computing devices. In some embodiments, network 130 is a private network that provides user device 120 with access to manufacturing equipment 124, sensors 126, measurement equipment 128, connected instruments 129, data store 140, and / or other private computing devices. Network 130 may include one or more wide area networks (WANs), local area networks (LANs), wired networks (e.g., Ethernet networks), wireless networks (e.g., 802.11 networks or Wi-Fi networks), cellular networks (e.g., Long Term Evolution (LTE) networks), routers, hubs, switches, server computers, cloud computing networks, and / or combinations thereof.
[0025] The user device 120 may include computing devices such as personal computers (PCs), laptops, mobile phones, smartphones, tablet computers, netbook computers, networked televisions ("smart TVs"), networked media players (e.g., Blu-ray players), set-top boxes, over-the-top (OTT) streaming devices, operator boxes, wearable computing devices, augmented / virtual reality devices, and the like. In some embodiments, the user device 120 is a mobile computing device such as a laptop, mobile phone, smartphone, tablet computer, or netbook computer. The user device 120 may collect one or more types of sensor data (e.g., generate images of portions of a process chamber after one or more maintenance operations) and send the sensor data to the prediction server 112 for processing. The user device may additionally or instead receive information from the server machine 170 and / or the prediction server 112 indicating whether a maintenance step or operation was successful (e.g., sensor data associated with a maintenance step met a maintenance criterion) or unsuccessful (e.g., sensor data associated with a maintenance step did not meet a maintenance criterion).
[0026] The user device 120 may include a corrective action component 122. The corrective action component 122 may receive user input (e.g., via a graphical user interface (GUI) displayed via the user device 120, voice control of the user device 120, etc.) regarding a maintenance operation of the manufacturing equipment 124. In some embodiments, a user utilizes a camera of the user device 120 to capture an image of a component associated with the maintenance operation. In some embodiments, a user may provide a description of the maintenance operation via text input, voice input, multiple choice selection, etc. In some embodiments, the corrective action component 122 transmits data regarding the maintenance operation to the predictive system 110 and receives output (e.g., predicted data 168) from the predictive system 110. In some embodiments, the corrective action component 122 determines a corrective action based on the output and causes the corrective action to be implemented. The corrective action may include issuing an alert to a user, providing instructions to a device (e.g., the user device 120, a device of the manufacturing equipment 124, etc.), updating a process recipe (e.g., a maintenance process recipe), etc.
[0027] In some embodiments, the predictive system 110 may further include a predictive component 116. The predictive component 116 may utilize data (e.g., sensor data 142) retrieved from the data store 140, the user device 120, etc. to generate predictive data 168. The predictive data 168 may include predicted instructions for the performance of equipment, components, etc., taking into account the sensor data 142 related to the maintenance of the equipment. The predictive data 168 may include instructions for the success of a maintenance step, operation, process, etc. In some embodiments, the predictive component 116 provides the predictive data 168 to the user device 120, which in turn causes a corrective action to be performed via the corrective action component 122 taking into account the predictive data 168. For example, the corrective action component 122 may provide maintenance instructions to a user via the user device 120 based on the sensor data 142 related to the maintenance operation. In some embodiments, the corrective action component 122 obtains sensor data 142 (e.g., current sensor data 146) related to the manufacturing equipment 124 (e.g., from a data store 140, etc.) and provides the sensor data 142 (e.g., current sensor data 146) related to the manufacturing equipment 124 to the predictive system 110.
[0028] 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 of the trained machine learning model 190 in the data store 140, and the user device 120 may retrieve the output from the data store 140. In some embodiments, the corrective action component 122 receives corrective action instructions from the prediction system 110 and causes the corrective action to be implemented. Each user device 120 may include an operating system that enables a user to perform one or more of generating, reviewing, or editing data (e.g., instructions associated with the manufacturing equipment 124, corrective actions associated with the manufacturing equipment 124, etc.). The user device 120 may further include a component for generating image or video data, which in some embodiments provides to the prediction component 116.
[0029] In some embodiments, the prediction system 110 may receive data related to a maintenance operation of a maintenance event. The prediction system 110 may generate validation data 164 indicating the results of a validation process for the maintenance operation. The validation data 164 may include a pass / fail indication, a grade, or a score for the maintenance operation, etc.
[0030] In some embodiments, metrology data 169 corresponds to historical characteristic data of a product (e.g., a product produced using manufacturing equipment and maintained according to maintenance data 150 recorded by sensors as sensor data 142), and predictive data 168 relates to predicted characteristic data (e.g., predicted characteristic data of a product produced by manufacturing equipment 124 or a product produced by manufacturing equipment 124). In some embodiments, predictive data 168 is or includes predicted metrology data (e.g., virtual metrology data) of a product produced by manufacturing equipment 124 or a product produced using manufacturing equipment 124. In some embodiments, predictive data 168 is or includes indications of anomalies (e.g., an abnormal product, an abnormal component, an abnormal manufacturing equipment, an abnormal energy usage, etc.) and / or one or more causes of those anomalies. In some embodiments, predictive data 168 includes indications of time changes or drifts of some components of manufacturing equipment 124, sensors 126, metrology equipment 128, or the like. In some embodiments, the predictive data 168 includes an end of life indication for a component of manufacturing equipment 124, a sensor 126, a measurement device 128, or the like.
[0031] In some embodiments, the predictive data 168 includes an indication of whether execution of a maintenance step of a maintenance process was performed satisfactorily, e.g., an indication of whether the results of the maintenance process, manufacturing equipment 124 will pass qualification testing, an indication of whether products processed using manufacturing equipment 124 will meet performance standards after execution of the maintenance process, etc.
[0032] Maintenance events that result in equipment not meeting performance standards can be costly in time, energy, materials, components, costs of identifying defects, discarded defective product, etc. By using sensor data 142 to generate predictions of success of a step or operation of a maintenance process, an entire maintenance process, etc., costs associated with isolating and correcting errors in a maintenance event may be significantly reduced. System 100 may provide the technical advantage of avoiding and / or reducing costs associated with troubleshooting associated with failed maintenance events.
[0033] Performing a maintenance process that results in a failure of a component of the manufacturing equipment 124 can be costly in downtime, damage to product, damage to equipment, rush-ordering replacement components, and the like. The maintenance operations may be evaluated before the maintenance process is completed by inputting sensor data 142 (e.g., data associated with the execution of a maintenance event) into the predictive component 116, receiving an output of predictive data 168 output by the predictive component 116, and taking action based on the output predictive data 168. Such individual evaluation of individual maintenance operations can increase the granularity of information regarding the execution of the maintenance process and can reduce downtime associated with the maintenance process. The predictive data 168 can also be used to perform corrective actions (e.g., predicted operational action maintenance, such as replacing, treating, cleaning, etc. of a component) based on the predictive data 168. The system 100 may have the technical advantage of avoiding one or more costs of unexpected component failure, unscheduled downtime, lost productivity, unexpected equipment failure, product waste, or the like. Additionally, in embodiments, predictive component 116 may compare predictive data sets from the same chamber over time to diagnose drifting or failing components (recorded as predictive data 168). Performance of components, such as manufacturing equipment 124, sensors 126, metrology equipment 128, and the like, may be monitored over time to provide an indication of deteriorating components.
[0034] As a result of performing a suboptimal maintenance procedure, manufacturing equipment may produce products that do not meet performance targets. Such suboptimal maintenance procedures may have costly consequences, such as increased resource (e.g., energy, coolant, gas, etc.) consumption, increased time to produce products, increased component failures, increased amount of defective products, etc. The system 100 may perform one or more corrective actions by inputting the sensor data 142 into a predictive component 116 (which may include, e.g., a trained machine learning model 190), receiving an output of the predictive data 168, and performing a corrective action (e.g., based on the predictive data 168) based on the predictive data 168. The corrective actions performed by the system 100 may include alerting a user, providing maintenance instructions to a user, updating a process recipe, scheduling additional maintenance, etc.
[0035] In one example, the corrective action may include updating the manufacturing parameters (e.g., setting optimal manufacturing parameters) based on the predictive data 168. By updating the manufacturing parameters according to the state of the manufacturing equipment after maintenance, the system 100 may have the technical advantage of using optimal manufacturing parameters (e.g., hardware parameters, process parameters) to avoid the costly consequences of less than optimal manufacturing parameters.
[0036] In some embodiments, the corrective action includes issuing an alert (e.g., an alert to stop or not perform one or more additional steps or operations of a maintenance process if the predictive data 168 indicates a predicted anomaly, such as an anomaly in the product, component, or manufacturing equipment 124). In one embodiment, the corrective action includes repeating a previously performed maintenance operation. In one embodiment, the corrective action includes replacing a component of the process chamber. In some embodiments, the corrective action includes providing feedback control (e.g., feedback control that alters a manufacturing parameter in response to the predictive data 168 indicating an anomaly). In some embodiments, the corrective action may be performed taking into account the output of the trained machine learning model. In some embodiments, performing the corrective action includes causing updates to one or more maintenance parameters to be performed.
[0037] Corrective actions may include instructions to a user to perform a hardware fix (e.g., replacing a component, using certain components, replacing a processing chip, updating firmware, etc.), instructions to manufacturing equipment to update maintenance parameters (e.g., temperature, pressure, flow rate, rate, current, voltage, gas flow rate, lift speed, etc.), updating settings (e.g., torque settings, etc.) of one or more connected tools 129, or the like. In some embodiments, corrective actions include updating recipes for maintenance operations, product processing operations, etc.
[0038] Prediction server 112, server machine 170, and server machine 180 may each include one or more computing devices, such as a rack mount 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 application specific integrated circuit (ASIC) (e.g., a tensor processing unit (TPU)), etc.
[0039] The predictive server 112 may include a predictive component 116. The predictive component 116 may be used to generate predictive data 168. In some embodiments, the predictive data 168 includes an indication of the success or failure of a maintenance step of a maintenance process. In some embodiments, the predictive data 168 includes an indication of aging, degradation, etc. of components of the manufacturing equipment 124. In some embodiments, the predictive component 116 may incorporate sensor data recorded during the processing of a product using the manufacturing equipment 124 to predict characteristics of the product.
[0040] In some embodiments, the predictive component 116 includes one or more models 190 (which may be trained machine learning models such as digital twins and / or models based on physical phenomena), to which the sensor data 142 is provided. Although the discussion that follows often refers to a single model, and a single model is shown in FIG. 1, it will be understood that the model 190 may be many separate models, a collection of models, or the like. In some embodiments, each maintenance step for which validation is determined has its own validation model. The model 190 may be trained to output data (e.g., as validation data 164 or prediction data 168) that indicates the validation of the success of the maintenance step of the maintenance process. The 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), or the like. In some embodiments, the model 190 is configured to accept as input sensor data that indicates the execution of a maintenance step. In some embodiments, this may include sensor data 142 from sensors 126 associated with the manufacturing equipment 124, e.g., tracking data from one or more pressure sensors when a manufacturing chamber is vented or pumped down to vacuum. In some embodiments, the input of the model 190 may include sensor data 142 from sensors associated with the user device 120, e.g., video or image data, e.g., images of a reassembled part or a cleaned part, which are processed by the image processing machine learning model 190. In some embodiments, the sensor data 142 provided as input to the model 190 may include sensor data 142 from a connected instrument 129, such as a smart tool. The model 190 may be trained to provide an indication of success or failure of a maintenance step given the input data. The indication of success or failure may take the form of a pass / fail, a grade, a score or a rating, or the like.The score may indicate the likelihood that performing that maintenance step will result in the maintenance equipment meeting one or more qualification criteria, the likelihood that the maintenance equipment will produce a product that meets one or more performance metrics, etc.
[0041] In some embodiments, the models 190 may include one or more models based on physical phenomena. For example, determining the success or failure of a mass flow controller maintenance, installation, etc. may include providing sensor data associated with the controller to a digital twin model based on physical phenomena. Many of the characteristics discussed above with respect to machine learning models may apply to models based on physical phenomena, e.g., the number of models, the nature of the input data, the source of the input data, the target output of the models based on physical phenomena, etc. In some embodiments, the validation data 164 may be generated in consideration of a comparison between a single sensor value (or multiple sensor values) and one or more thresholds (e.g., finding at least one sensor value meets at least one threshold criterion, finding at least one sensor value does not meet at least one threshold criterion, etc.).
[0042] In some embodiments, an action may be taken taking into account the validation data 164. The generation of an instruction to perform the action may depend on the validation data 164 (e.g., pass, fail, grade or score, etc.). In some embodiments, after determining that the maintenance step was performed satisfactorily (e.g., the validation data indicates a pass of the performance), an instruction may be provided to the manufacturing equipment 124, for example, to proceed to the next maintenance step or operation, or the user device 120 may be instructed to display to the user the instructions for the next maintenance step (which may have been stored in the data store 140 as the maintenance data 150). Additionally or alternatively, the prediction server 112 may provide an instruction to prescribe a corrective action (e.g., to perform one or more previous maintenance operations again).
[0043] In some embodiments, action may be taken in light of the validation data 164 indicating that performance of a maintenance step did not pass validation. The prediction server 112 may provide instructions to the manufacturing equipment 124 to repeat a maintenance step (or portion of a maintenance step, multiple maintenance steps, etc.), cause the user device 120 to present instructions to the user regarding troubleshooting, restarting the maintenance step, etc.
[0044] The historical sensor data 144 may be used in combination with the current sensor data 146 and the manufacturing data 150 to detect drift, changes, aging, etc. of components of the manufacturing equipment 124. The predictive component 116 may use a combination and comparison of these data types to generate predictive data 168. In some embodiments, the predictive data 168 includes data that predicts the lifespan of components such as the manufacturing equipment 124, the sensors 126, etc. In some embodiments, the predictive data 168 may include updates to maintenance processes to reflect changes in the performance of one or more components.
[0045] Providing validation of the maintenance phase provides technical advantages over other techniques. By ensuring that various phases of the maintenance process meet validation standards, the scope of actions to investigate (e.g., troubleshooting) in the event of a failure may be significantly narrowed, which may result in faster green-to-green times (e.g., shorter maintenance events), higher first-pass rates (e.g., a higher percentage of maintenance processes that result in manufacturing equipment 124 passing qualification tests), and the like. Multiple communication events between the user device 120 and a server (e.g., prediction server 112) allow accumulated maintenance knowledge to be provided directly to the user (e.g., technician) in a timely manner. This facilitates quality maintenance work by enabling the ability to provide detailed instructions to the user both when the maintenance process is proceeding smoothly and when it is not. The prediction server 112 may be configured to handle sensor data 142 originating from different maintenance phases (e.g., different parts of the maintenance process, maintenance of different equipment, etc.) separately. The use of data from each stage may be tailored to produce a higher quality validation (e.g., accurate prediction) for that stage (e.g., by feeding the data to a different model, using a particular subset of sensor data 142, adjusting thresholds for comparison with the sensor data, etc.). Additionally, prediction server 112 and / or models 190 may be updated over time to produce stronger predictive data 168, more impactful corrective action data 166, etc.
[0046] In some embodiments, the predictive component 114 may receive the current sensor data 146 and perform pre-processing such as extracting patterns in the data or combining the data into new composite data. The predictive component 116 may then provide the data as input to the model 190. The predictive component 116 may receive an indication of the success or failure of performing a maintenance step of the maintenance process from the model 190. The predictive component 116 may then generate corrective action data 166 or cause a corrective action to be performed in response to detecting a failed maintenance step. The corrective action may include sending an alert to the user device 120. The corrective action may further include updating maintenance parameters of the manufacturing equipment 124. The corrective action may further include generating predictive data 168 indicating drift, aging, or failure of the chamber or tool.
[0047] The data store 140 may be a memory (e.g., random access memory), a drive (e.g., hard drive, flash drive), a database system, or another type of component or device capable of storing data. The data store 140 may include multiple storage components (e.g., multiple drives or multiple databases) that may be across multiple computing devices (e.g., multiple server computers). The data store 140 may store sensor data 142, maintenance data 150, validation data 164, corrective action data 166, forecast data 168, and metered data 169. The sensor data 142 may include historical sensor data 144 and current sensor data 146. The sensor data may include pre-processed data such as time tracking of sensor data throughout the duration of a maintenance process, association of data with physical sensors, averages and composite data, and data indicative of sensor performance over time (i.e., across many maintenance processes). The metered data 169 may include similar characteristics. The maintenance data 150 may include historical maintenance data 152 and current maintenance data 154. The maintenance data may include maintenance process recipes, instructions for equipment, tools, instruments, users, etc., for the maintenance process, metrics for rating the maintenance process, maintenance steps, etc., or the like. The historical sensor data 144 and the historical maintenance data 152 may be historical data (e.g., at least a portion for training the machine learning model 190 and / or the model 190 based on physical phenomena). The current sensor data 146 may be current data (e.g., at least a portion for input into the model 190 following the historical data) for which the validation data 164 is generated (e.g., to perform corrective actions). The metrology data 169 may be metrology data of produced substrates, as well as sensor data, manufacturing data, and model data corresponding to those products. The metrology data 169 may be utilized to design a process for producing additional substrates. The prediction data 168 may include predictions of success or failure of the maintenance steps, processes, etc.The predictive data 168 may include predictions of what measurement data would have been derived from a set of input parameters. The predictive data 168 may also include data indicative of components of the system 100 aging and failing over time.
[0048] In some embodiments, the prediction system 110 further includes a server machine 170 and a server machine 180. The server machine 170 includes a dataset generator 172 that can generate datasets (e.g., a set of data inputs and a set of target outputs) for training, validating, and / or testing the model 190. Some operations of the dataset generator 172 are described in detail below with respect to FIGS. 2 and 4A. In some embodiments, the dataset generator 172 may divide the historical data (e.g., historical sensor data 144, historical maintenance data 152) and performance data (e.g., measurement data 169, corrective action data 166 of corrective actions related to the historical sensor or maintenance data, etc.) into a training set (e.g., 60 percent of the data), a validation set (e.g., 20 percent of the data), and a test set (e.g., 20 percent of the data). In some embodiments, the prediction system 110 generates multiple sets of features (e.g., via the prediction component 116). For example, the first set of features may correspond to a first set of types of sensor data (e.g., from a first set of sensors, a first combination of values from the first set of sensors, a first pattern of values from the first set of sensors) corresponding to each of the datasets (e.g., a training set, a validation set, and a test set), and the second set of features may correspond to a second set of types of sensor data (e.g., from a second set of sensors different from the first set of sensors, a second combination of values different from the first combination, a second pattern different from the first pattern) corresponding to each of the datasets.
[0049] The server machine 180 includes a training engine 182, a validation engine 184, a selection engine 185, and / or a test engine 186. The engines (e.g., training engine 182, validation engine 184, selection engine 185, and test engine 186) may refer to hardware (e.g., circuitry, dedicated logic circuitry, programmable logic circuitry, microcode, processing devices, etc.), software (e.g., instructions running on a processing device, general-purpose computer system, or dedicated machine), firmware, microcode, or a combination thereof. The training engine 182 may be capable of training a model 190 using one or more sets of features associated with a training set from the dataset generator 172. The training engine 182 may generate multiple trained models 190, where each trained model 190 corresponds to a different set of features of the training set (e.g., sensor data from a different set of sensors). For example, a first trained model may be trained using all features (e.g., X1-X5), a second trained model may be trained using a first subset of features (e.g., X1, X2, X4), and a third trained model may be trained using a second subset of features (e.g., X1, X3, X4, and X5), where the second subset of features may overlap with the first subset of features. A dataset generator 172 may receive the output of a trained model (e.g., 190) and compile the data into training, validation, and testing datasets and use those datasets to train a second model.
[0050] The validation engine 184 may be capable of validating the trained models 190 using a corresponding set of validation set features from the dataset generator 172. For example, a first trained model 190 trained using a first set of training set features may be validated using a first set of validation set features. The validation engine 184 may determine the accuracy of each of the trained models 190 based on the corresponding set of validation set features. The validation engine 184 may discard trained models 190 that have an accuracy that does not meet a threshold accuracy. In some embodiments, the selection engine 185 may be capable of selecting one or more trained models 190 that have an accuracy that meets a threshold accuracy. In some embodiments, the selection engine 185 may be capable of selecting the trained model 190 with the highest accuracy among the trained models 190.
[0051] The testing engine 186 may be capable of testing the trained models 190 using a corresponding set of test set features from the dataset generator 172. For example, a first trained model 190 trained using a first set of training set features may be tested using a first set of test set features. The testing engine 186 may determine the trained model 190 that has the highest accuracy among all of the trained models based on the test set.
[0052] In the case of a machine learning model, the model 190 may refer to a model artifact generated by the training engine 182 using a training set including data inputs and corresponding target outputs (correct answers for each corresponding training input). Patterns in the data set that map the data inputs to the target outputs (correct answers) may be found, and the machine learning model 190 is provided with a mapping that captures these patterns. In the case of a model based on physical phenomena, further training, validation, selection, and / or testing may be performed (e.g., to account for model inaccuracies, manufacturing tolerances of components of the manufacturing equipment 124, variable performance of components, aging of components, etc.). In the case of a model based on physical phenomena, the model 190 may be user-generated, and the model 190 is improved by operation of the server machine 180.
[0053] The prediction component 116 may provide the current sensor data 146 to a model 190 (e.g., a trained machine learning model, a model based on physical phenomena, etc.) and may execute the model 190 on the input to obtain one or more outputs. The prediction component 116 may determine (e.g., extract) prediction data 168 from the output of the model 190, and may determine (e.g., extract) confidence data from the output indicating a confidence that the prediction data 168 is an accurate predictor of a process associated with the input data for maintenance of the manufacturing equipment 124, a product produced using the manufacturing equipment 124 maintained according to the current sensor data 146, an indication of component life or aging, etc. The prediction component 116 or the corrective action component 122 may use this confidence data to decide whether to cause a corrective action associated with the manufacturing equipment 124 to be performed based on the prediction data 168.
[0054] The confidence data may include or indicate a confidence that the prediction data 168 is an accurate prediction of future performance, e.g., future performance of the manufacturing equipment 124. As an example, the prediction data 168 may indicate that a maintenance step was not performed correctly. The confidence data may indicate that the prediction data 168 is an accurate prediction for rating the performance of a maintenance step associated with at least a portion of the input data. In one example, the confidence is a real number between 0 and 1, inclusive, where 0 indicates no confidence that the prediction data 168 is an accurate prediction and 1 indicates absolute confidence that the prediction data 168 is accurately predicting the outcome of the maintenance step. In response to the confidence data indicating a confidence below a threshold level for a predetermined number of instances (e.g., a percentage of instances, a frequency of instances, a total number of instances, etc.), the prediction component 116 may retrain the model 190 (e.g., based on the current sensor data 146, the current maintenance data 154, etc.).
[0055] For purposes of illustration and not limitation, aspects of the disclosure describe training one or more models 190 using historical data (e.g., historical sensor data 144, historical maintenance data 152) and inputting current data (e.g., current sensor data 146, current maintenance data 154) into the one or more trained models 190 to determine predicted data 168, validation data 164, corrective action data 166, etc. In other implementations, a heuristic or rule-based model is used (e.g., without using a trained machine learning model) to determine the predicted data. The predictive component 116 may monitor the historical sensor data 144, historical maintenance data 152, and metering data 169. Any of the information described with respect to the data input 210 of FIG. 2 may be monitored by the heuristic or rule-based model or used in other ways.
[0056] In some embodiments, the functionality of user device 120, prediction server 112, server machine 170, and server machine 180 may be provided by fewer machines. For example, in some embodiments, server machines 170 and 180 may be combined into a single machine, and in other embodiments, server machine 170, server machine 180, and prediction server 112 may be combined into a single machine. In some embodiments, user device 120 and prediction server 112 may be combined into a single machine.
[0057] In general, functions described as being performed by user device 120, prediction server 112, server machine 170, and server machine 180 in one embodiment may be performed on prediction server 112 in other embodiments, where appropriate. Additionally, functions attributed to particular components 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, user device 120 may determine prediction data 168 based on output from model 190.
[0058] Additionally, the functionality of a particular component may be performed by different or multiple components operating together. One or more of prediction server 112, server machine 170, or server machine 180 may be accessed as a service offered to other systems or devices through an appropriate application programming interface (API).
[0059] In embodiments, a "user" may be expressed as a single individual. However, other embodiments of the present disclosure encompass a "user" being an entity controlled by multiple users and / or automated sources. For example, a collection of individual users combined into a group of administrators may be considered a "user."
[0060] Embodiments of the present disclosure may be applied to data quality assessment, feature enhancement, model evaluation, virtual metrology (VM), predictive maintenance (PdM), marginal optimization, or the like.
[0061] Although embodiments of the present disclosure are discussed with respect to generating predictive data 168 for performing corrective actions at a manufacturing facility (e.g., a semiconductor manufacturing facility), the embodiments may also be generally applied to improved accuracy of staged processes where validation may be performed at various stages.
[0062] 2 is a block diagram of an example dataset generator 272 (e.g., dataset generator 172 of FIG. 1 ) used to generate a dataset for a model (e.g., machine learning model 190 of FIG. 1 , physically-based model 190, etc.), according to certain embodiments. The dataset generator 272 may be part of server machine 170 of FIG. 1 . In some embodiments, system 100 of FIG. 1 includes multiple models. In such cases, each model may have a separate dataset generator, or the models may share a dataset generator.
[0063] 2, a system 200 including a dataset generator 272 (e.g., dataset generator 172 of FIG. 1) generates a dataset for a model (e.g., model 190 of FIG. 1). The dataset generator 272 may generate the dataset using data retrieved as output from various sensing devices. In some embodiments, the dataset generator 272 generates training inputs by selecting a subset of historical data from sensing devices, possibly including sensors associated with manufacturing equipment, sensors associated with user devices, sensors associated with connected tools, etc. The dataset generator 272 may generate target outputs from indications of success of maintenance steps associated with the sensor data used to generate the input data (e.g., validation data 164 of FIG. 1). For example, the validation indications 230 may include a classification of whether the maintenance steps associated with the sensor data used as input resulted in a successful maintenance process, which troubleshooting or corrective actions taken in response to the input data led to the success, etc.
[0064] It is within the scope of this disclosure that the target output 220 may be expressed in a variety of different ways. The verification instructions 230 may include pass / fail, a grade, or a score, etc. The verification instructions 230 may include, for example, instructions for actions to be taken in response to the verification, such as instructions to move to the next maintenance stage, suggested troubleshooting actions, indications of product characteristics, an indication of the effect of the maintenance on the operation of the manufacturing equipment, etc.
[0065] 2, in some embodiments, a dataset generator 272 generates a dataset (e.g., a training set, a validation set, a test set) that includes one or more data inputs 210 (e.g., training inputs, validation inputs, test inputs) and may include one or more target outputs 220 corresponding to the data inputs 210. The dataset may further include mapping data that maps the data inputs 210 to the target outputs 220. The data inputs 210 may also be referred to as "features," "attributes," or "information." In some embodiments, the dataset generator 272 may provide the dataset to the training engine 182, validation engine 184, or testing engine 186 of FIG. 1, where the dataset is used to train, validate, or test the model 190 of FIG. 1. Some embodiments of generating a training set are further described with respect to FIG. 4A.
[0066] In some embodiments, the dataset generator 272 may generate a first data input corresponding to the first set of simulated sensor data 262A for training, validating or testing a first model, and the dataset generator 272 may generate a second data input corresponding to the second set of simulated sensor data 262B for training, validating or testing a second model.
[0067] In some embodiments, a dataset generator 272 may perform operations on one or more of the data inputs 210 and the target outputs 220. The dataset generator 272 may extract patterns from the data (slope, curvature, etc.), combine the data (average, feature generation, etc.), or group simulated sensors to train separate models.
[0068] The data input 210 and target output 220 for training, validating, or testing a model may include information for a particular manufacturing chamber (e.g., a particular semiconductor wafer manufacturing chamber). The data input 210 and target output 220 may include information for a particular manufacturing chamber design (e.g., to be used for all chambers of that design).
[0069] In some embodiments, the information used to train the model may be from a particular type of manufacturing equipment (e.g., manufacturing equipment 124 in FIG. 1 ) having particular characteristics of the manufacturing equipment, which may enable the trained model to determine outcomes for a particular group of manufacturing equipment 124 based on inputs for current sensor data (e.g., current sensor data 146) associated with one or more components sharing the characteristics of that particular group. In some embodiments, the information used to train the model may be information for components from more than one manufacturing equipment, which may enable the trained model to determine outcomes for a component based on inputs from one manufacturing equipment. In some embodiments, the information used to train the model may be specific information for a particular maintenance process, maintenance phase, etc.
[0070] In some embodiments, following generating a dataset and using the dataset to train, validate, or test a model, the model may be further trained, validated, or tested, or tuned (e.g., based on new sensor data, validation data, measurement data, maintenance data, etc.).
[0071] 3 is a block diagram illustrating a system 300 for generating output data (e.g., prediction data 168 and / or validation data 164 of FIG. 1 ) according to some embodiments. System 300 may be used to determine a prediction of successful execution of a maintenance step of a maintenance process pre-configured in association with manufacturing equipment. System 300 may be used to determine instructions to execute in response to execution of the maintenance step.
[0072] 3, at block 310, a system 300 (e.g., a component of the predictive system 110 of FIG. 1) performs data partitioning (e.g., via the dataset generator 172 of the server machine 170 of FIG. 1 ) of the sensor data 360 (e.g., extracting features from historical sensor data) and validation data 362 (e.g., historical maintenance verification instructions, corrective action instructions related to historical maintenance) to generate a training set 302, a validation set 304, and a test set 306. For example, the training set may be 60% of the sensor and validation data, the validation set may be 20% of the sensor and validation data, and the test set may be 20% of the sensor and validation data.
[0073] At block 312, the system 300 performs model training (e.g., via the training engine 182 of FIG. 1 ) using the training set 302. The system 300 may train multiple models using multiple sets of features of the training set 302 (e.g., a first set of features including a group of sensors in the training set 302, a second set of features including a different group of sensors in the training set 302, etc.). For example, the system 300 may train a machine learning model to generate a first trained machine learning model using the first set of features in the training set and to generate a second trained machine learning model using the second set of features in the training set (e.g., different data than the data used to train the first machine learning model). In some embodiments, the first trained machine learning model and the second trained machine learning model may be combined to generate a third trained machine learning model (e.g., which may be a better predictor than the first or second trained machine learning models, alone). In some embodiments, the sets of features used in comparing the models may overlap (e.g., one model may be trained using simulated sensors 1-15, a second model may be trained using simulated sensors 10-20, etc.). In some embodiments, hundreds of models may be generated, including models with various permutations of features and combinations of models.
[0074] At block 314, the system 300 performs model validation (e.g., via validation engine 184 in FIG. 1 ) using the validation set 304. The system 300 may validate each of the trained models using a corresponding set of features from the validation set 304. For example, the validation set 304 may use the same subset of sensors used in the training set 302 but for different maintenance events. In some embodiments, the system 300 may validate hundreds of models (e.g., models with various permutations of features, combinations of models, etc.) generated at block 312. At block 314, the system 300 may determine the accuracy of each of the one or more trained models (e.g., via model validation) and may determine whether one or more of the trained models have an accuracy that meets a threshold accuracy. In response to determining that none of the trained models have an accuracy that meets the threshold accuracy, flow returns to block 312, where the system 300 performs model training using a different set of features from the training set. In response to determining that one or more of the trained models have an accuracy that meets the threshold accuracy, flow proceeds to block 316. The system 300 may discard trained models that have an accuracy below the threshold accuracy (e.g., based on a validation set).
[0075] At block 316, the system 300 performs model selection (e.g., via selection engine 185 of FIG. 1 ) to determine which model of the one or more trained models that meet the threshold accuracy has the highest accuracy (e.g., selected model 308 based on the check of block 314). In response to determining that two or more of the trained models that meet the threshold accuracy have the same accuracy, flow may return to block 312, where the system 300 performs model training to determine the trained model with the highest accuracy using a further refined training set corresponding to the further refined set of features.
[0076] At block 318, the system 300 performs model testing (e.g., via the test engine 186 of FIG. 1 ) using the test set 306 to test the selected model 308. The system 300 may test the first trained model using a first set of features in the test set (e.g., simulated sensors 1-15) to determine that the first trained model meets a threshold accuracy (e.g., based on the first set of features of the test set 306). In response to the accuracy of the selected model 308 not meeting the threshold accuracy (e.g., the selected model 308 is too well-fitted to the training set 302 and / or the validation set 304 and cannot be applied to other data sets, such as the test set 306), flow proceeds to block 312, where the system 30A performs model training (e.g., retraining) using a different training set, possibly a different training set corresponding to a different set of features, or a reorganization of the substrate divided into training, validation, and test sets. In response to determining that the selected model 308 has an accuracy that meets the threshold accuracy based on the test set 306, flow proceeds to block 320. At least in block 312, the model may learn patterns in the sensor data to make predictions, and in block 318, the system 300 may apply the model to the remaining data (e.g., the test set 306) to test the predictions.
[0077] In block 320, the system 300 receives current sensor data 346 (e.g., current sensor data 146 of FIG. 1 ) using the trained model (e.g., selected model 308) and determines (e.g., extracts) predicted data 366 (e.g., predicted data 168 of FIG. 1 ) from the output of the trained model in order to perform an action (e.g., perform corrective action associated with manufacturing equipment 124 of FIG. 1 , issue an alert to user device 120 of FIG. 1 , etc.).
[0078] In some embodiments, the machine learning model is retrained by providing additional data to further train the model. Current sensor data 352 may be provided in block 312. Additional validation data 361 may be provided as well. These data may differ from the data originally used to train the model by including input parameters that were not part of the original training, combinations of input parameters outside the parameter space spanned by the original training, or these data may be updated to reflect chamber specific knowledge (e.g., differences from an ideal chamber due to manufacturing tolerance ranges, aging components, etc.). Based on this data, the selected model 308 may be retrained.
[0079] In some embodiments, one or more of operations 310-320 may be performed in various orders and / or together with other operations not presented and described herein. In some embodiments, one or more of operations 310-320 may not be performed. For example, in some embodiments, one or more of data partitioning of block 310, model validation of block 314, model selection of block 316, or model testing of block 318 may not be performed.
[0080] 4A-F are flow diagrams of methods 400A-F associated with generating prediction and / or validation data for performing corrective actions and / or determining whether a maintenance step or operation of a maintenance process has been successfully performed, according to some embodiments. Methods 400A-F may be performed by processing logic, which may include hardware (e.g., circuits, dedicated logic circuits, programmable logic circuits, microcode, processing devices, etc.), software (e.g., instructions executed on a processing device, general-purpose computer system, or dedicated machine), firmware, microcode, or combinations thereof. In some embodiments, methods 400A-F may be performed in part by prediction system 110 of FIG. 1. Method 400A may be performed in part by prediction system 110 (e.g., server machine 170 and data set generator 172 of FIG. 1, data set generator 272 of FIG. 2). The prediction system 110 may use the method 400A to generate a data set for performing at least one of training, validating, or testing a model, e.g., a model based on physical phenomena and / or a machine learning model, according to an embodiment of the present disclosure. The method 400B may be performed by the server machine 180 (e.g., training engine 182, etc.). The method 400C may be performed by the prediction server 112 (e.g., prediction component 116). The method 400D may be performed by the user device 120. The method 400E may be performed by the prediction system 110 (e.g., server machine 180). The method 400F may be performed by the prediction system 110 (e.g., prediction server 112). In some embodiments, a non-transitory storage medium stores instructions that, when executed by a processing device (e.g., a processing device of the prediction system 110, a processing device of the server machine 180, a processing device of the prediction server 112, a processing device of the user device 120, etc.), cause the processing device to perform one or more of the methods 400A-F.
[0081] For ease of explanation, methods 400A-F are illustrated and described as a series of operations. However, operations in accordance with the present disclosure may be performed in various orders and / or simultaneously, as well as with other operations not shown and described herein. Moreover, not all illustrated operations may be performed to implement methods 400A-F in accordance with the disclosed subject matter. Furthermore, those skilled in the art will understand and appreciate that methods 400A-F may also be represented as a series of interrelated states via a state diagram or events.
[0082] FIG. 4A is a flow diagram of a method 400A for generating a dataset for a model for generating predicted data (eg, predicted data 168 of FIG. 1) according to certain embodiments.
[0083] Referring to FIG. 4A, in some embodiments, at block 401, processing logic performing method 400A initializes a training set T to be an empty set.
[0084] At block 402, processing logic generates a first data input (e.g., a first training input, a first validation input), which may include sensor data (e.g., sensor data 262 of FIG. 2), etc. In some embodiments, this first data input may include a first set of features for a type of data, and the second data input may include a second set of features for the type of data (e.g., as described with respect to FIG. 3).
[0085] At block 403, processing logic generates a first target output for one or more of these data inputs (e.g., a first data input). In some embodiments, this first target output is an indication of a verification of a maintenance step associated with the input sensor data. In some embodiments, the target output may be related to a predicted performance of a product produced using a manufacturing equipment associated with the input sensor data.
[0086] At block 404, processing logic may generate mapping data indicating an input / output mapping. This input / output mapping (or mapping data) may relate to a data input (e.g., one or more of the data inputs described herein), a target output for the data input, and an association between the data input and the target output.
[0087] In some embodiments, at block 405, processing logic adds the mapping data generated at block 404 to the dataset T.
[0088] At block 406, processing logic branches based on whether the dataset T is sufficient for at least one of training, validating, and / or testing the model 190 of FIG. 1. If so, execution proceeds to block 407; if not, execution returns to block 402. It should be noted that in some embodiments, whether the dataset T is sufficient may be determined simply based on the number of inputs in the dataset, and in some embodiments based on the number of inputs in the dataset that are mapped to outputs, and in some other implementations, whether the dataset T is sufficient may be determined based on one or more other criteria (e.g., measures of diversity of data examples, accuracy, etc.) in addition to or instead of the number of inputs.
[0089] At block 407, processing logic provides dataset T (e.g., to server machine 180 of FIG. 1 ) for training, validating, and / or testing model 190. In some embodiments, dataset T is a training set, and dataset T is provided to training engine 182 of server machine 180 to perform training. In some embodiments, dataset T is a validation set, and dataset T is provided to validation engine 184 of server machine 180 to perform validation. In some embodiments, dataset T is a test set, and dataset T is provided to testing engine 186 of server machine 180 to perform testing. The datasets generated by method 400A may then be used to receive as input data from sensors that indicate the performance of a maintenance step, and to generate as output data that is indicative of an outcome of the performance of the maintenance step (e.g., quality of maintenance, quality of product produced, etc.).
[0090] FIG. 4B is a method 400B for a server that facilitates satisfactorily performing a maintenance process, according to some embodiments.
[0091] 4B, at block 410 of method 400B, processing logic of the server initiates a maintenance process. The maintenance process is executed on the manufacturing equipment. The maintenance process includes several maintenance steps or operations. The maintenance steps / operations are related to one or more components of the manufacturing equipment. In some embodiments, initiating the maintenance process includes sending data to a user device instructing the user device to instruct a user to initiate the maintenance process or steps. In some embodiments, initiating the maintenance process includes sending instructions to the manufacturing equipment to update parameters of the maintenance process. In some embodiments, the server initiates the maintenance event in response to a trigger, such as a certain number of hours the manufacturing equipment has been in production, a certain number of products produced, etc. In some embodiments, the server initiates the maintenance event in response to an alert from the user device.
[0092] At block 412, the processing logic of the server provides instructions to the user device directing it to perform a maintenance step of the maintenance process. In some embodiments, the server may further provide instructions related to the maintenance step for output from the user device to the user. The instructions may include information regarding the execution of the maintenance step, such as step-by-step maintenance instructions, pictures, video, audio, etc. In some embodiments, some or all of the maintenance steps (e.g., flushing the chamber with gas, controlling the heating elements, etc.) may be performed without user input. In some embodiments, the server may instead provide instructions directly to the manufacturing equipment during steps of the maintenance process that are not performed in response to user input. In some embodiments, providing instructions may include providing settings for a connected tool, such as setting the torque of a smart wrench.
[0093] At block 414, the processing logic of the server receives data resulting from the execution of the maintenance step. In some embodiments, the received data includes sensor data from sensors associated with the manufacturing equipment. In some embodiments, the received data is provided by a user via a user device. In some embodiments, the received data includes image or video data provided by the user device. For example, the user device may include a camera. The user may point the camera of the user device at one or more areas of the manufacturing equipment and generate images and / or videos of those areas. In some embodiments, the user device provides instructions to the user regarding which areas to image. In some embodiments, the received data includes data provided by a connected tool (e.g., a connected smart torque wrench, which provides a measure of the torque achieved by the tool). In some embodiments, the server sends instructions to the connected tool that cause the connected tool to perform the maintenance operation according to specified parameters. For example, the server may send instructions indicating a maximum torque to apply, and the torque wrench may not apply more than the indicated maximum torque.
[0094] At block 416, the processing logic of the server performs operations to determine whether the performance of the maintenance step was satisfactory. The particular metrics, models, thresholds, etc. used to make this determination will vary between maintenance processes, maintenance events, and maintenance steps. In some embodiments, the processing logic of the server may compare a data value indicating the performance of a maintenance step to a threshold, and (for example) determine that the maintenance step was not performed satisfactory if the data value is greater than the threshold (e.g., the maintenance step may include pumping a manufacturing chamber to vacuum, and a pressure sensor reading may be compared to a threshold to determine whether the chamber achieved an appropriate level of vacuum).
[0095] In some embodiments, the processing logic of the server may provide data (possibly pre-processed by the user device, server, etc.) directing the execution of a maintenance step to a physics-based model for validation. As an example, a physics-based digital twin model may be constructed that describes gas flow rates in a fabrication chamber. Pressure sensor data may be collected when the fabrication chamber is vented to atmospheric pressure. The sensor data may then be compared to simulated pressure data from a physics-based model that models venting the fabrication chamber in the same way (e.g., using the same valve). Differences between the measured and simulated sensor data may indicate component drift, aging or failure, sensor drift or failure, etc. In some embodiments, statistical data (e.g., the average of several simulated pressure data, how many standard deviations from that average to the measured sensor data points, etc.) may be used. In some embodiments, the physics-based model may predict which components, failures, etc. contribute to differences between the simulated and measured sensor data. Many types of sensors, many different maintenance operations, many types of physics-based models, etc. are possible within the scope of this disclosure.
[0096] In some embodiments, the processing logic of the server may provide data indicative of the performance of the maintenance step to a trained machine learning model for validation. The trained machine learning model may receive as input the data indicative of the performance of the maintenance step and may generate as output a validation indicating whether the performance of the maintenance step is satisfactory or not. As an example, a user may provide an image of a component associated with a maintenance operation, e.g., using a camera of a user device (e.g., user device 120 of FIG. 1). In some embodiments, the user may be instructed (e.g., via a display of user device 120) to take, record, or the like, an image, video, or the like of a component of the manufacturing equipment. In some embodiments, pictures of the component may be taken before and after maintenance. The one or more images may be provided as input to a trained machine learning model, which is configured to output data (e.g., validation data) indicative of the quality of the performance of one or more maintenance operations associated with the component for which the picture was taken.
[0097] Verification of satisfactory performance may indicate that the maintained component has been assessed as likely to operate correctly after maintenance. Verification of satisfactory performance may indicate that a user performed the maintenance operation properly. Verification of satisfactory performance may indicate that the sensor data is similar to the sensor data achieved during a successful maintenance process.
[0098] If the performance of the maintenance step is determined to be satisfactory, validation data may be generated (not shown) indicating successful performance of the maintenance step, and the method ends. If the performance of the maintenance step is determined to be unsatisfactory, flow proceeds to block 418. At block 418, processing logic of the server causes corrective action to be taken. In some embodiments, the corrective action may include issuing an alert to a user. In some embodiments, the alert may include instructions to display to a user via a user device. In some embodiments, the alert includes instructions for a user to assist in the maintenance process. In some embodiments, the instructions include instructions to repeat a maintenance step or a portion of a maintenance step. In some embodiments, the instructions suggest troubleshooting operations. In some embodiments, the instructions may include maintenance instructions in written form, pictures, video, audio, etc.
[0099] In some embodiments, this corrective action may include updating the process recipe, hi some embodiments, the processing logic may update the execution of the maintenance step operations, for example to account for chamber drift, component aging, etc.
[0100] Figure 4C illustrates an enhanced method 400C for a server that facilitates satisfactorily performing a maintenance process, according to some embodiments. Many operations of method 400C have similar characteristics to those of method 400B, for example, blocks 410, 412, 414, and 416 of Figure 4B may include similar characteristics to blocks 420, 421, 422, and 423 of Figure 4C, where the maintenance stage and data from the maintenance stage of Figure 4B maps to the first maintenance stage and first data from the first maintenance stage of Figure 4C.
[0101] At block 424, processing logic of the server causes the user device to display instructions to the user to repeat the first maintenance step, which may be a special case of the overall operation shown in FIG. 4B of causing a corrective action to be taken.
[0102] At block 425, the processing logic of the server receives second data from the second execution of the first maintenance step. As discussed with respect to block 414 of FIG. 4B, this data may take many forms. At block 426, the processing logic of the server determines that the second execution of the first maintenance step was satisfactory. This may be determined, for example, by comparing sensor data to a threshold and finding that the sensor data meets at least one threshold criterion, or by providing sensor data (including sensors associated with manufacturing equipment, sensors associated with user devices, sensors associated with connected tools, etc.) to a trained model (e.g., a model based on physics, a machine learning model, etc.) and receiving an output verifying that the maintenance step was performed satisfactorily.
[0103] At block 427, the processing logic of the server provides instructions to the user device directing the user device to perform a second maintenance phase of the maintenance process. The instructions may be provided in consideration of the server receiving validation data indicating that the first maintenance phase was performed satisfactorily.
[0104] At block 428, the processing logic of the server receives third data from the first execution of the second maintenance phase of the maintenance process. This data may take many forms and may be similar to the data received at block 425. At block 429, the processing logic of the server determines that the first execution of the second maintenance phase was satisfactory. At block 430, the processing logic of the server provides instructions to the user device directing it to execute a third maintenance phase of the maintenance process.
[0105] It will be understood that method 400C is representative and that the disclosure is not limited to this arrangement of validation operations or to any particular arrangement of validation operations. For example, a first run of a first maintenance stage may pass validation and a first run of a second maintenance stage may fail validation, and the server may provide corrective action instructions. Additionally, in some operations, the server may provide instructions to a device different from the user device, such as a device that controls manufacturing equipment, connected instruments, etc., and may be included within the scope of the disclosure.
[0106] 4D is a flow diagram illustrating a method 400D for a user device to facilitate execution of a maintenance phase, according to some embodiments. The user device may include a smartphone, tablet, laptop, wearable computing device, etc. The user device may be a dedicated machine, a program or application running on a general-purpose computing device, etc. The user device may be configured to communicate with manufacturing equipment and a server. The user device may communicate over a network.
[0107] At block 440, processing logic of the user device receives instructions in association with the manufacturing equipment directing the user to perform a maintenance step of a maintenance process. The instructions may be accompanied by materials that the user device displays to the user. The materials may include a description of the maintenance step, pictures or videos describing the performance of the maintenance step, etc.
[0108] At block 441, the processing logic of the user device provides data resulting from the performance of the maintenance step to the server. In some embodiments, this data may be indicative of sensor readings of sensors associated with the manufacturing equipment. In some embodiments, this data may be image or video data or other data provided by the user, such as confirmation of the performance of the maintenance step, a written description of the performance of the maintenance step, etc. In some embodiments, this data may be indicative of readings from sensors of connected tools. In some embodiments, some of this data may be provided to the server over the network without being processed by the user device.
[0109] At block 442, the processing logic of the user device receives validation data from the server indicating that the performance of the maintenance step was unsatisfactory. This validation data may take the form of a pass / fail, a grade or rating, or the like. At block 443, the user device causes corrective action to be taken in light of the validation data. In some embodiments, this corrective action includes issuing an alert to the user. In some embodiments, this alert includes instructions for additional maintenance tasks to perform. In some embodiments, this corrective action includes updating a maintenance process recipe, such as adjusting temperatures, pressures, flow rates, etc. associated with manufacturing equipment or adjusting settings associated with connected tools.
[0110] The operations of blocks 444-449, indicated by dashed boxes, may be performed if data processing indicates that the maintenance phase is to be repeated. At block 444, processing logic of the user device provides second data to the server resulting from the second execution of the first maintenance phase. This data may be of the same type / form as the first data resulting from the first execution of the first maintenance phase. At block 445, processing logic of the user device receives second verification data from the server indicating that the second execution of the maintenance phase was satisfactory. This verification data may have many of the same characteristics as the verification data of block 442.
[0111] At block 446, the processing logic of the user device receives a second instruction from the server directing the execution of a second maintenance stage of the maintenance process. The second stage of the maintenance process is different from the first stage, e.g., may include different operations or may involve different components of the manufacturing equipment. At block 447, the processing logic of the user device provides data resulting from the execution of the second maintenance stage of the maintenance process to the server. This data may share many characteristics with the data of block 441.
[0112] At block 448, the user device receives validation data from the server associated with performance of the second maintenance step, the validation data indicating that performance of the second maintenance step was satisfactory, and at block 449, the user device receives instructions to perform a third maintenance step of the maintenance process.
[0113] Similar to the discussion regarding FIG. 4C, method 400D is an example of one set of operations of the present disclosure. For example, the methods of the present disclosure are not limited to a particular sequence of passing and failing validation and may include other operations between the operations shown.
[0114] 4E-F are flow diagrams illustrating methods 400E and 400F relating to machine learning models, according to some embodiments. Some operations of methods 400E and 400F may also be applicable to models based on physical phenomena, where appropriate.
[0115] FIG. 4E illustrates a flow diagram of a method 400E for training a machine learning model. At block 450, processing logic receives sensor data indicative of performance of a maintenance step of a maintenance process. The processing logic may be part of a system for generating predictive data, such as predictive system 110 of FIG. 1. The data may be received as a training data set, such as from dataset generator 172 of FIG. 1. At block 452, processing logic receives data indicative of a quality of performance of a maintenance step of a maintenance process. The received data may include data correlating data indicative of performance (e.g., sensor data) with data indicative of quality of performance (e.g., an indication of quality such as a measurement, grade, or score of a finished product, etc.).
[0116] At block 454, the processing logic causes training of the machine learning model. The training may include providing sensor data indicative of the performance of the maintenance step as a training input. The training may further include providing data indicative of the quality of the performance of the maintenance step as a target output. In that case, the trained machine learning model may be configured to accept as input data indicative of the performance of the maintenance step (e.g., sensor data from sensors associated with manufacturing equipment, sensors associated with a user device, sensors associated with a connected tool, etc.). Training of the machine learning model may be performed by a training engine, such as training engine 182 of server 180 of FIG. 1.
[0117] 4F is a flow diagram of a method 400F of using a trained machine learning model, according to some embodiments. At block 460, processing logic receives sensor data that indicates the performance of a maintenance step of a maintenance process. At block 462, processing logic provides the sensor data to the trained machine learning model. In some embodiments, some processing of the sensor data may have been performed and data indicative of the sensor data may have been received by processing logic and / or provided to the trained model by processing logic.
[0118] At block 464, the processing logic receives validation data from the trained machine learning model indicating the quality of the performance of the maintenance step. This validation may take the form of pass / fail (e.g., the performance of the maintenance step was satisfactory / unsatisfactory), a grade or rating, etc. In some embodiments, the processing logic may receive additional information related to the performance of the maintenance step. The trained machine learning model may be trained to generate additional information, for example based on sensor data, and the model may suggest corrective actions, specific components to maintain, etc. The operations of block 466, represented by a dashed box, may be performed if the trained machine learning model provides an indication that the performance of the maintenance step was unsatisfactory. At block 466, the processing logic causes the corrective action to be performed. The operations of block 466 may share many characteristics with the operations of block 418 of FIG. 3B.
[0119] FIG. 5 is a diagram illustrating the operation of a system 500 for improving a maintenance process, according to some embodiments. The system 500 includes a maintenance device 502 with various associated maintenance operations shown as flow diagrams, a network 530 (e.g., network 130 of FIG. 1), and a validation module 550 (e.g., one or more models that are part of the predictive system 110 of FIG. 1). The maintenance device 502 may include manufacturing equipment, user devices, connected tools, and the like. The operations shown in relation to the maintenance device 502 are generalized abstractions of maintenance operations, such as preventive maintenance operations, and these operations are included to outline an exemplary method of performing a maintenance process. Each operation shown in FIG. 5 may, for example, include many separate operations, may not appear in this order, may be disjointed, and may be interleaved with operations of other blocks.
[0120] In some embodiments, the maintenance process begins with several operations in maintenance preparation 504. These operations may be directed to preparing the manufacturing equipment for disassembly, e.g., preparing to open the manufacturing chamber to allow access to the internal components. The operations in maintenance preparation 504 may include performing cleaning procedures, purging the interior of the manufacturing chamber, turning off heating and cooling devices, venting the chamber, etc. Data instructing the performance of these operations may be collected from sensors associated with the manufacturing equipment. These sensors may be, for example, on-board sensors such as pressure sensors, temperature sensors, flow rate sensors, etc. In some embodiments, after the performance of a stage of the maintenance process (e.g., after the performance of a subset of the operations of the maintenance process), data instructing the performance of that stage is provided to validation module 550 via network 530. Validation module 550 may determine the quality of the performance of that maintenance stage. The quality of execution may be determined by comparing data directing the execution of the maintenance step to one or more thresholds, providing data directing the execution of the maintenance step to a model based on physical phenomena and receiving an output from the model based on physical phenomena (and optionally comparing the output of the model based on physical phenomena to one or more criteria, such as thresholds), providing data directing the execution of the maintenance step to a trained machine learning model and receiving an output from the trained machine learning model, or the like. A verification method for the step may be selected and adjusted to provide a reliable prediction of the success of the maintenance step. If the verification indicates that the step was performed satisfactorily, the verification module 550 may indicate via the network 530 that the maintenance process may proceed to the next step. In some embodiments, the verification may be performed sequentially for some or all parts of the steps of the maintenance preparation 504. If the verification indicates that the maintenance step was not performed satisfactorily, the verification module 550 may provide an indication of corrective action, such as, for example, a troubleshooting operation, an instruction to repeat the maintenance step, etc.In one embodiment, the stages of maintenance preparation 504 include a dry clean procedure, an oxygen plasma purge procedure, a process chamber isolation procedure (in which the process chamber is taken offline), a purge process, a lid heater deactivation procedure, and a chiller and heat exchanger deactivation procedure, and a chamber vent procedure. After each of these maintenance stages, a specific test may be performed to evaluate whether the maintenance stage was performed successfully based on sensor data collected before and / or during the stage.
[0121] After all maintenance preparation steps are successfully completed, the maintenance operation may then enter a disassembly 506 phase. Disassembly 506 may include removing the fabrication chamber lid, removing parts (e.g., quartz and / or SiC rings), removing components (e.g., chamber liners, process kits, lift pins, shingle rings, etc.), removing screens (e.g., plasma screens), etc. In some embodiments, some operations may not undergo verification, for example, if no sensor data is available and to prevent the verification process from becoming too cumbersome. If the disassembly 506 process is to be verified, data directing the execution of those steps may be provided to the verification module 550 over the network 530, and the verification module 550 may provide the verification data over the network 530, in a manner similar to that described with respect to maintenance preparation 504.
[0122] The maintenance process may then enter a phase of maintenance 508. Maintenance 508 may include cleaning components (e.g., cleaning an electrostatic chuck), removing and / or disassembling components (e.g., flow valves), replacing and / or installing components (e.g., installing a manometer tree, installing an insulating ring, installing an IED window, installing a viewport, installing a slit valve door, etc.), assembling or reassembling parts or components (e.g., assembling a lid and / or lid assembly, etc.). Operations in this phase may provide data for validation from connected instruments, for example, from a smart tool equipped with sensors to measure characteristics related to its use, such as torque. The maintenance stages included in reassembly and certification 510 may also provide data from a user device, as described above, and may receive validation data over network 530.
[0123] In some embodiments, a user may provide information for the validation of the maintenance 508 operation. The user may provide image or video data via a user device that indicates the execution of the maintenance step. For example, the user may generate pictures via a user device and provide the pictures (e.g., pictures of the electrostatic chuck assembly before and / or after a cleaning procedure) to the validation module 550 via the network 530. The validation module 550 may provide the images to a trained machine learning model configured to output an indication of the quality of the execution of the maintenance step. The validation module 550 may then provide validation data via the network 530 based on the indication of the quality of the execution, similar to the above description. For some maintenance steps (e.g., a cleaning step), image data is generated and used to evaluate whether the maintenance step was successfully executed. For example, a trained machine learning model may identify whether the chamber or part is clean after a cleaning operation. For other maintenance steps, sensor data from a connected tool (e.g., tool readings from a wrench, screwdriver, and / or other tool) is used to evaluate whether the maintenance step was successfully executed. For example, the instrument readings may be compared to a threshold value to determine whether they are within a tolerance of the target instrument reading. For some maintenance steps, a combination of images and instrument readings from the tool are used to determine whether the maintenance step was performed successfully.
[0124] The maintenance process may then enter a qualification 510 phase. Operation of this phase may also provide data for validation from manufacturing equipment sensors to the validation module 550, as described above, and may receive validation data over the network 530. In an embodiment, multiple different qualification phases may be performed. In one qualification phase, the chamber is pumped down to a target pressure (e.g., vacuum). The sensor data may include pressure readings and / or pressure profiles over time that indicate how long it will take to pump the chamber down to the target pressure. In another qualification phase, the cooling device, heat exchanger and / or heating device are activated to heat and / or cool to a target temperature set point. The sensor data may include temperature readings (e.g., indicating how long it will take to reach a target temperature), voltage and / or current readings of heating elements, etc. In another qualification phase, a chamber leak test may be performed. This test may include pumping the process chamber down to a target pressure and then measuring whether the pressure is increasing over time (which may indicate a leak, for example). In another qualification phase, chamber seasoning may be performed. Chamber seasoning may include running one or more processes in the process chamber to change the conditions of the chamber. Such seasoning processes may, for example, form a coating on the chamber parts. After all qualification steps have been successfully performed, the process chamber may be used for production.
[0125] FIG. 6 is a block diagram illustrating a computer system 600 according to certain embodiments. In some embodiments, the 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). The computer system 600 may operate in the capacity of 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. The 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, a wearable computing device, an augmented reality or virtual reality device, or any other 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 a collection of computers that individually or jointly execute a set (or sets) of instructions to perform any one or more of the methods described herein.
[0126] In an additional 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 618, which may communicate with each other via a bus 608.
[0127] 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 other types of instruction sets, or a combination of 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).
[0128] Computer system 600 may further include a network interface device 622 (e.g., coupled to a network 674). Computer system 600 may further 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.
[0129] In some implementations, the data storage device 618 may include a non-transitory computer-readable storage medium 624 (e.g., a non-transitory machine-readable storage medium) having stored thereon instructions 626 encoding one or more of the methods or functions described herein, including instructions encoding the components of FIG. 1 (e.g., the prediction component 114, the model 190, etc.) and instructions for performing the methods described herein.
[0130] The instructions 626 may further reside, completely or partially, within the volatile memory 604 and / or within the processing device 602 during execution of the instructions 626 by the computer system 600, and thus the volatile memory 604 and the processing device 602 may also constitute machine-readable storage media.
[0131] While the computer-readable storage medium 624 is shown as a single medium in the illustrative example, the term "computer-readable storage medium" is intended to include a single medium or multiple media (e.g., centralized or distributed databases and / or associated caches and servers) that store one or more sets of executable instructions. The term "computer-readable storage medium" is further intended to include any tangible medium that can store or encode a set of instructions for execution by a computer that causes the computer to perform one or more of the methods described herein. The term "computer-readable storage medium" is intended to include, but is not limited to, solid-state memory, optical media, and magnetic media.
[0132] The methods, components, and features described herein may be implemented by discrete hardware components or may be integrated into the functionality of other hardware components, such as an ASIC, FPGA, DSP, or similar device. Additionally, the methods, components, and features described herein may be implemented by firmware modules or by functional circuits within a hardware device. Additionally, the methods, components, and features described herein may be implemented in any combination of hardware devices and computer program components, or in a computer program.
[0133] Unless otherwise indicated, terms such as "receive," "perform," "provide," "obtain," "perform," "access," "determine," "add," "use," "train," "initiate," "execute," "update," or other similar terms, refer to operations and processes performed or implemented by a computer system that manipulate and convert data in computer system registers and memory, represented as physical (electronic) quantities, into other data in the computer system memory or registers, or other such information storage, transmission, or display device, also represented as physical quantities. Furthermore, terms such as "first," "second," "third," "fourth," etc., as used herein, are meant as markers to distinguish between different elements and may not have an ordinal meaning based on the numerical designation of those terms.
[0134] The examples described herein further relate to apparatus for carrying out the methods described herein. The apparatus may be specially constructed for carrying out the methods described herein, or may include a general-purpose computer system that is selectively programmed by a computer program stored in the computer system. Such a computer program may be stored in a computer-readable tangible storage medium.
[0135] The methods and illustrative examples described herein are not inherently related to any particular computer or other apparatus. Various 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 the individual functions, routines, subroutines or operations of those methods. Example structures for these various systems are provided in the description above.
[0136] The above description is intended to be illustrative and not limiting. Although the present disclosure has been described with reference to certain illustrative examples and embodiments, it will be recognized that the present disclosure is not limited to the described examples and embodiments. The scope of the present disclosure should be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled.
Claims
1. A system including a server and a manufacturing device, the system comprising: initiating a maintenance process for the manufacturing equipment, the maintenance process including a number of maintenance steps associated with maintaining one or more components of the manufacturing equipment; providing a first instruction to a user device directing the user device to perform a first maintenance step of the plurality of maintenance steps; receiving first data resulting from a first execution of the first maintenance phase; determining that the first performance of the first maintenance phase was unsatisfactory; and causing the user device to display instructions for corrective actions to be taken. A system configured to:
2. The method further includes the step of: receiving a first instruction directing the execution of a first maintenance step of the plurality of maintenance steps; providing data resulting from a first execution of the first maintenance phase to the server; receiving validation data from the server indicating that the first performance of the first maintenance phase was unsatisfactory; and Taking corrective action in light of said validation data The system of claim 1 configured to:
3. The system further comprises: receiving second data resulting from a second performance of the first maintenance phase; determining that the second performance of the first maintenance phase was satisfactory; and providing second instructions to the user device directing the user device to perform a second maintenance step of the plurality of maintenance steps; The system of claim 1 configured to:
4. The system further comprises: receiving second data resulting from a first performance of a second maintenance step of the plurality of maintenance steps; determining that the first performance of the second maintenance phase was satisfactory; and providing second instructions to the user device directing the user device to perform a third maintenance step of the plurality of maintenance steps; The system of claim 1 configured to:
5. the first data resulting from the performance of the first maintenance phase; sensor data from sensors associated with the manufacturing equipment; Image and / or video data, or Data from one or more connected instruments The system of claim 1 , comprising at least one of:
6. 2. The system of claim 1, wherein determining that the first performance of the first maintenance phase was unsatisfactory comprises determining that at least one value associated with the first data does not meet at least one threshold criterion.
7. determining that the first performance of the first maintenance phase was unsatisfactory; providing data based on the first data to a model based on physical phenomena; and receiving an indication from the physics-based model that a first performance of the first maintenance phase was unsatisfactory; The system of claim 1 , comprising:
8. determining that the first performance of the first maintenance phase was unsatisfactory; providing data based on the first data to a trained machine learning model; and receiving an indication from the trained machine learning model that a first performance of the first maintenance phase was unsatisfactory; The system of claim 1 , comprising:
9. The system of claim 1 , wherein the system is further configured to assign a score describing the success of at least one of a maintenance stage of a maintenance process or the maintenance process.
10. 10. The system of claim 9, wherein the score indicates a likelihood that performance of the maintenance step or process will result in a manufacturing equipment meeting one or more manufacturing equipment qualification standards.
11. initiating, by a server, a maintenance process for the manufacturing equipment, the maintenance process including a plurality of maintenance steps associated with maintaining one or more components of the manufacturing equipment; providing a first instruction to a user device directing the user device to perform a first maintenance step of the plurality of maintenance steps; receiving first data resulting from a first execution of the first maintenance phase; determining that the first performance of the first maintenance phase was unsatisfactory; and Implementing corrective actions The method includes:
12. The corrective action is: Alerting the user; Providing maintenance instructions to users; or Updating the process recipe The method of claim 11 , comprising at least one of:
13. receiving second data resulting from a second performance of the first maintenance phase; determining that the second performance of the first maintenance phase was satisfactory; and providing second instructions to the user device directing the user device to perform a second maintenance step of the plurality of maintenance steps; The method of claim 11 further comprising:
14. receiving second data resulting from a first performance of a second maintenance step of the plurality of maintenance steps; determining that the first performance of the second maintenance phase was satisfactory; and providing second instructions to the user device directing the user device to perform a third maintenance step of the plurality of maintenance steps; The method of claim 11 further comprising:
15. the first data resulting from the performance of the first maintenance phase; sensor data from sensors associated with the manufacturing equipment; Image and / or video data, or Data from one or more connected instruments The method of claim 11 , comprising at least one of:
16. A non-transitory machine-readable storage medium having instructions stored thereon that, when executed, receiving, by a user device, a first instruction directing performance of a first maintenance step of a plurality of maintenance steps of a maintenance process associated with the manufacturing equipment; providing first data resulting from a first execution of the first maintenance phase to a server; receiving first validation data from the server indicating that the first performance of the first maintenance phase was unsatisfactory; and taking corrective action in consideration of said first validation data; A non-transitory machine-readable storage medium that causes a processing device to perform operations including
17. The corrective action is: Alerting the user; Providing maintenance instructions to users; or Updating the process recipe 20. The non-transitory machine-readable storage medium of claim 16, comprising at least one of:
18. providing second data to the server resulting from a second execution of the first maintenance phase; receiving second validation data from the server indicating that the second performance of the first maintenance phase was satisfactory; and receiving a second instruction from the server directing the execution of a second maintenance step of the plurality of maintenance steps; 20. The non-transitory machine-readable storage medium of claim 16, further comprising:
19. the first data resulting from a first execution of the first maintenance phase; sensor data from sensors associated with the manufacturing equipment; Image and / or video data, or Data from one or more connected instruments 20. The non-transitory machine-readable storage medium of claim 16, comprising at least one of:
20. providing second data to the server resulting from a first execution of a second maintenance step of the plurality of maintenance steps; receiving second validation data from the server indicating that the first performance of the second maintenance phase was satisfactory; and receiving a second instruction from the server directing the execution of a third maintenance step of the plurality of maintenance steps; 20. The non-transitory machine-readable storage medium of claim 16, further comprising:
Citation Information
Patent Citations
Substrate processing system
JP2002289506A
Substrate processing device and method for maintenance of substrate processing device
JP2018195607A
Correcting component failures in ion implantation semiconductor manufacturing tools
JP2022523101A
Correcting component failures in ion implant semiconductor manufacturing tool
US20200251360A1
