Enhanced processing and hardware architecture for detecting and correcting real-time product substrates

Abstract

Embodiments disclosed herein include a processing tool for semiconductor processing. In embodiments, the processing tool includes a chamber and a plurality of witness sensors integrated with the chamber. In embodiments, the processing tool further includes a drift detection module. In embodiments, data from the plurality of witness sensors is provided to the drift detection module as input data. In embodiments, the processing tool further includes a dashboard for displaying output data from the drift detection module.

Description

[0001] Cross-references to related applications

[0002] This application claims priority to U.S. non-provisional application No. 17 / 075,321, filed October 20, 2020, the entire contents of which are incorporated herein by reference. Technical Field

[0003] Embodiments of this disclosure relate to the field of semiconductor processing, and more particularly to processing tool architectures capable of real-time processing parameter drift detection and / or real-time correction to mitigate processing parameter drift. Background Technology

[0004] As semiconductor devices continue to shrink in feature size, semiconductor wafer processing has become increasingly complex. A given process can include many different processing parameters (i.e., knobs), which can be individually controlled to achieve desired results on the wafer. For example, desired results on the wafer could refer to feature profiles, layer thickness, layer chemical composition, etc. As the number of knobs increases, the theoretical processing space available for tuning and optimization becomes extremely large.

[0005] Furthermore, once the final processing formulation has been developed, chamber drift during multiple iterations of processing for different wafers can cause variations in the results on the wafer. Chamber drift can result from erosion of consumable parts of the chamber, component degradation (e.g., sensors, lamps, etc.), byproduct film deposition on the surface, and so on. Therefore, additional tuning is required even after extensive formulation development processes. Summary of the Invention

[0006] The embodiments disclosed herein include a processing tool for semiconductor processing. In one embodiment, the processing tool includes a chamber and a plurality of witness sensors integrated with the chamber. In one embodiment, the processing tool further includes a drift detection module. In one embodiment, data from the plurality of witness sensors is provided to the drift detection module as input data. In one embodiment, the processing tool further includes a dashboard for displaying output data from the drift detection module.

[0007] Implementations may also include processing tools comprising physical tools. In one embodiment, the physical tools include control loop sensors and witness sensors. In one embodiment, the processing tool may further include a drift detection module. In one embodiment, the drift detection module receives control loop sensor data and witness sensor data as input. In one embodiment, the drift detection module outputs processing parameter data indicating whether one or more processing parameters have drifted.

[0008] Implementations may also include a processing tool comprising physical instruments. In one embodiment, the physical instrument includes a chamber and a cartridge for allowing one or more processing gases to flow into the chamber from a plurality of gas sources. In one embodiment, the physical instrument further includes: a mass flow controller for each of the plurality of gas sources, a mass flow meter located between the gas source and the cartridge, a first pressure gauge located between the mass flow meter and the cartridge, a second pressure gauge fluidly coupled to the chamber, and an outlet line coupled to the chamber. In one embodiment, the processing tool further includes a drift detection module. In one embodiment, the drift detection module receives data as input from one or more of the mass flow controller, the mass flow meter, the first pressure gauge, and the second pressure gauge, and wherein the drift detection module outputs processing parameter data. Attached Figure Description

[0009] Figure 1A This is a schematic block diagram of a processing tool including a drift detection module according to one embodiment.

[0010] Figure 1B It is a schematic block diagram of a processing tool including a drift detection module and a correction module according to one embodiment.

[0011] Figure 1C It is a schematic block diagram of a processing tool including a drift detection module, a correction module and a prediction module according to one embodiment.

[0012] Figure 2 This is a schematic diagram of a processing tool according to one embodiment, the processing tool including multiple witness sensors for notifying one or more of a drift detection module, a correction module, and a prediction module.

[0013] Figure 3 This is a block diagram of a processing tool according to one embodiment, which uses witness sensors and a hybrid data model to notify one or more of a drift detection module, a correction module, and a prediction module.

[0014] Figure 4A It is a dashboard presented to the user of a processing tool according to one embodiment, which provides information about one or more processing parameters of the processing tool.

[0015] Figure 4B According to one implementation method Figure 4A The statistical process control chart for one of the processing parameters in the process.

[0016] Figure 5A block diagram of an exemplary computer system according to one embodiment of the present disclosure is shown. Detailed Implementation

[0017] This document describes a processing tool architecture capable of real-time processing parameter drift detection and / or real-time correction to mitigate processing parameter drift. In the following description, numerous specific details are set forth to provide a thorough understanding of embodiments of this disclosure. It will be apparent to those skilled in the art that embodiments of this disclosure can be practiced without these specific details. In other instances, well-known methods, such as integrated circuit manufacturing, have not been described in detail so as not to unnecessarily obscure embodiments of this disclosure. Furthermore, it should be understood that the various embodiments illustrated in the accompanying drawings are illustrative representations and are not necessarily drawn to scale.

[0018] As described above, drift during multiple iterations of processing in a processing tool is a common problem in the semiconductor manufacturing industry. Therefore, the embodiments disclosed herein include processing tools that include a drift detection module. In one embodiment, the drift detection module uses machine learning and / or hybrid models to detect when one or more processing parameters have drifted. Once drift is identified, the tool operator can adjust one or more tool settings to mitigate the drift. In another embodiment, the processing tool may further include a correction module. The correction module may utilize machine learning and / or hybrid models to generate a control effort to mitigate processing drift. That is, when drift occurs, the processing tool can automatically correct itself, rather than relying on a tool operator for correction. In yet another embodiment, a prediction module may be included in the processing tool. The prediction module may utilize machine learning and / or hybrid models to predict processing parameter drift before it occurs. In such embodiments, the prediction module may provide control to the processing tool to prevent drift before it can occur.

[0019] Now for reference Figure 1A The illustration shows a schematic diagram of a processing tool 100 according to one embodiment. In one embodiment, the processing tool 100 may include a detection module 161. The detection module 161 is used to identify drift conditions within the processing tool 100. In one embodiment, the detection module 161 may include a chamber 105. The chamber 105 may be any chamber for processing a substrate (such as, but not limited to, a wafer). For example, the wafer may have any suitable shape factor (e.g., 300 mm, 450 mm, etc.). The wafer may be a semiconductor wafer, such as a silicon wafer, or a III-V semiconductor material. In other embodiments, the substrate may have a shape factor different from that of a standard wafer.

[0020] Chamber 105 can be adapted to different types of processing operations. For example, chamber 105 can be a lamp-based chamber 105, a heater-based chamber 105, or a plasma-based chamber 105. In one embodiment, chamber 105 may include chamber hardware 140. Hardware 140 may include, but is not limited to, the chamber itself, gas lines, valves, exhaust, lamps, a base, an RF source, etc. In one embodiment, hardware 140 may also include a control loop sensor. The control loop sensor can be used to control the processing conditions within the chamber. For example, the control loop sensor can be used to set a desired pressure within chamber 105.

[0021] In one embodiment, chamber 105 may further include a witness sensor 145. The witness sensor is located outside the control loop. Therefore, the witness sensor can be used to monitor the control loop sensors. When the control loop sensors drift, even if the control loop sensors do not indicate any change in processing conditions, the change in the output of the witness sensor 145 can be identified to alert the processing engineer to the drift condition.

[0022] In one implementation, the detection module 161 may include a detection software and algorithm block 120, referred to for simplicity as "detection block 120". Detection block 120 may include software and / or algorithms that utilize outputs from the witness sensor 145 and / or the control loop sensor to determine whether the processing tool 100 is experiencing drift. For example, the outputs from the witness sensor 145 and / or the control loop sensor may be compared with a process fingerprint. If there is a difference between the expected value from the process fingerprint and the actual value obtained from the witness sensor and / or the control loop sensor, it can be determined that processing drift has occurred.

[0023] In one implementation, detection block 120 may use machine learning algorithms and / or hybrid models to generate a processing fingerprint. The hybrid model may include statistical and physical models. In one implementation, a statistical model may be generated by implementing a physical design of experiment (DoE) and using interpolation to provide an extended multidimensional processing space model. In one implementation, a physical model may be generated using real-world physical and chemical relationships. For example, the physical and chemical equations of various interactions within the processing chamber may be used to build a physical model. The combination of statistical and physical models allows the hybrid model to be a multidimensional model capable of mapping various tool settings to predicted outcomes (i.e., predicted processing parameters). The predicted processing parameters for a given set of tool settings can be considered a processing fingerprint, with the outputs from witness sensor 145 and / or control loop sensors compared to this process fingerprint.

[0024] In one implementation, detection block 120 may output values ​​121 for detecting drift in key processing parameters of the processing performed by processing tool 100. For example, output values ​​121 may include differences between the outputs of the processed fingerprint and the witness sensor and / or control loop sensor. In one implementation, output values ​​121 may be provided to dashboard 165. Dashboard 165 can be quickly viewed by processing engineers to determine whether drift is occurring within the user interface of processing tool 100. Dashboard 165 will be described in more detail below.

[0025] Now for reference Figure 1B The illustration shows a schematic block diagram of a processing tool 100 according to an additional embodiment. The processing tool 100 may include a detection module 161, which is similar to... Figure 1A The system includes a detection module 161. In addition to the detection module 161, a calibration module 162 is provided. The calibration module 162 may include the detection module 161 and calibration software and algorithm block 120. B Abbreviated as "Correction Block 120" B "Correction Block 120" B It can be used in the following ways with detection block 120 A Similar to: Correction block 120 B Includes machine learning algorithms and / or hybrid models. Correction block 120 B Will come from detection block 120 A The output value 121 is used as input, and the control force is output to the chamber controller 122. The control force can be determined by querying a multidimensional model to find the mitigation force caused by the detection block 120. A The tool settings for detecting drift. Therefore, the drift of the processing tool 100 can be corrected without the intervention of the processing engineer. However, from the correction block 120 B The data can also be fed into the dashboard 165 to provide a visual indication of changes being implemented to the tool settings in the processing tool 100.

[0026] Now for reference Figure 1C The illustration shows a schematic diagram of a processing tool 100 according to an additional embodiment. Similar to... Figure 1B In the illustrated embodiment, the processing tool 100 may include a detection module 161 and a correction module 162. The processing tool 100 may further include a prediction module 163. The prediction module 163 may include a continuous learning system 123 and a prediction algorithm 120. C And self-calibration software and / or algorithms 124.

[0027] In one embodiment, the continuous learning system 123 receives input from sensors 145 of chamber 105. In another embodiment, the continuous learning system 123 may also receive input data from the output value 121 of the drift detection module 161. The continuous learning system 123 includes machine learning or artificial intelligence for classifying the predicted types of drift to occur. For example, the continuous learning system 123 may classify whether the predicted drift is the result of a drift condition of one or more of the pumps, lamps, or other changes in the hardware 140 of chamber 105. The continuous learning system 123 learns over time how chamber 105 responds to changes in sensor readings 145. That is, data is continuously fed into the continuous learning system 123 to develop robust models capable of predicting drift in one or more systems of chamber 105.

[0028] In one implementation, the continuous learning system 123 can provide output data that is fed into the prediction algorithm block 120. C The output data may include data identifying which (or which) systems are predicted to drift in the processing tool. In one implementation, the prediction algorithm block 120 C This can then be used to identify how long it will take for one or more processing operations within chamber 105 to exceed a threshold level. The threshold level includes values ​​above (or below) a setpoint, exceeding which would cause a processing operation to exceed specifications. Therefore, in addition to providing a timeframe until the drift exceeds a given threshold, it is also possible to predict the drift before it occurs and exceeds a given processing specification.

[0029] In one implementation, prediction algorithm block 120 C The output can then be provided to the self-calibration algorithm 124. The self-calibration algorithm 124 can provide control signals to the chamber control block 122. The chamber control block 122 provides adjustments to the hardware 140 to prevent predicted drift. In some embodiments, the self-calibration algorithm 124 may alternatively provide signals to the calibration block 120. B Provides output and allows correction of block 120. B Generate the necessary control signals to be sent to chamber control block 122 to accommodate the predicted drift.

[0030] In addition to providing predictive control over performance drift on the substrate, the prediction module 163 can also provide predictive maintenance for the hardware 140 in chamber 105. For example, hardware 140 may include pumps, lamps, etc. Predictive maintenance can be used to identify when hardware 140 will fail or when the performance of hardware 140 will degrade beyond a given threshold. In the case of predictive maintenance, the continuous learning system 123 can analyze sensor data and its relationship to failures of hardware 140 in chamber 105. That is, the continuous learning system 123 can identify patterns in the sensor data 145 that correspond to a failure or degradation of a piece of hardware 140. For example, in the case of pump failure, one or more sensors in sensor 145 may exhibit offsets beyond their typical range, and the pump may subsequently fail some time after the offset is detected.

[0031] After a model for the relationship between sensor data offset and hardware 140 faults has been developed for the continuous learning system, prediction algorithm block 120... C The identified offset can be searched for in the data from sensor 145. When a specific offset has been detected, the prediction algorithm block 120... C The processing engineer can be provided with an indication that a hardware 140 failure will occur within a known timeframe. This allows the processing engineer to initiate corrective maintenance to replace or repair the anticipated failure hardware 140 before it occurs. Consequently, waste material is reduced because no substrate will be processed in the chamber 105 containing the faulty hardware.

[0032] In some implementations, hardware calibration may not require replacement or repair. For example, in-situ chamber cleaning may be sufficient to prevent hardware failure or damage. In such cases, the self-calibration software and algorithm 124 can provide control signals to the chamber control 122 to initiate the necessary maintenance.

[0033] Now for reference Figure 2 The diagram illustrates a processing tool 200 according to one embodiment. The processing tool 200 is illustrated above and can be referenced. Figures 1A to 1C The hardware components utilized in one or more processing tools 100 are described. In the illustrated embodiment, processing tool 200 is described as a lamp-based chamber for free radical oxidation treatment. However, it should be understood that processing tool 200 is exemplary in nature, and the embodiments disclosed herein can be adapted for use in conjunction with other processing tools, such as, but not limited to, heater-based chambers or plasma-based chambers. Those skilled in the art will recognize that the placement, number, and type of sensors can be modified to track desired processing parameters for various types of processing tools.

[0034] In one embodiment, the processing tool 200 includes a chamber 205. The chamber 205 may be a chamber adapted to provide subatmospheric pressure, in which a substrate (e.g., a semiconductor wafer) is processed. In one embodiment, the chamber 205 may be sized to accommodate a single substrate or multiple substrates. Semiconductor substrates suitable for processing in the chamber 205 may include silicon substrates or any other semiconductor substrate. Other substrates (such as glass substrates) may also be processed in the chamber 205.

[0035] In one embodiment, the gas distribution network supplies gas (e.g., gas 1, gas 2, gas n, etc.) from one or more gas sources to cylinder 210. In a particular embodiment, the gas source may include one or more of oxygen, hydrogen, and nitrogen. Although in Figure 2 Figure A illustrates three gas sources; however, it should be understood that embodiments may include one or more gas sources. Cylinder 210 may include an inlet for receiving gas from line 211 and an outlet for distributing gas into chamber 210. In the illustrated embodiment, cylinder 210 is illustrated as supplying gas into chamber 210 from one side. However, it should be understood that cylinder 210 may optionally supply gas into chamber from above or below. In some embodiments, cylinder 210 may also be referred to as a nozzle, particularly when the processing tool is a plasma generating tool.

[0036] In one embodiment, the flow rate of each processed gas in the processed gases can be controlled by a separate mass flow controller (MFC) 203. In one embodiment, the MFC 203 may be part of a control loop sensor group. The MFC 203 controls the flow rate of the gas entering the input line 211. In one embodiment, a mass flow meter (MFM) 212 is located upstream of the cylinder 210. The MFM 212 allows measurement of the actual flow rate from the source gas. A pressure gauge 213 is also included upstream of the cylinder 210. The pressure gauge 213 allows measurement of the pressure in the input line 211. The MFM 212 and pressure gauge 213 can be considered witness sensors because they are outside the control loop.

[0037] In one embodiment, a chamber pressure gauge 217 may be provided to measure the pressure in chamber 205. Chamber pressure gauge 217 may be part of a control loop sensor group. In one embodiment, additional witness sensors are provided along the discharge line 215 of the processing tool 200. The additional sensors may include a leak detection sensor 216 and additional pressure gauges 218 and 219. Leak detection sensor 216 may include an optical emission spectrometry (OES) device to measure oxygen leaking into chamber 205. Pressure gauges 218 and 219 may be located upstream and downstream of throttle valve 214, respectively.

[0038] In one embodiment, pressure gauges 213, 217, 218, and 219 may have operating ranges suitable for providing typical pressures at their locations within the processing equipment. For example, pressure gauge 213 may operate at a pressure range higher than the pressure ranges of the other pressure gauges 217, 218, and 219. Similarly, pressure gauge 218 may operate at a pressure range higher than the pressure range of pressure gauge 219. In a particular embodiment, pressure gauge 213 may operate in a range including 1,000 T, pressure gauge 217 may operate in a range including 20 T, pressure gauge 218 may operate in a range including 100 T, and pressure gauge 219 may operate in a range including 10 T.

[0039] In one implementation, witness sensors (e.g., 212, 213, 216, 218, and 219) can be used to provide monitoring of chamber drift. For example, during use of the processing tool 200, control loop sensors (e.g., 203 and 217) may become inaccurate. Thus, the readings of control loop sensors 203 and 217 may remain constant, while results on the wafer (e.g., film deposition rate) change. In this case, the output of the witness sensors will change to indicate that the chamber has drifted.

[0040] In another embodiment, a witness sensor can be used to implement a virtual sensor in chamber 205. A virtual sensor can refer to a sensor that provides a computationally generated output, unlike a direct reading of a physical value (as in the case of a physical sensor). Therefore, virtual sensors are powerful for determining states within processing tool 200 that are difficult or impossible to measure with known physical sensors.

[0041] In one embodiment, a virtual sensor can be used to determine the flow rate of the process gas at the outlet of cylinder 210. Calculating the flow rate at cylinder 210 is a valuable metric that can be used to control the deposition rate and / or deposition uniformity of the film on the wafer. In a particular embodiment, the flow rate at cylinder 210 can be calculated using a Bernoulli equation with variables provided by using the outputs of MFM 212, pressure gauge 213, pressure gauge 217, and the known geometry of cylinder 210. While examples of flow rates at cylinders are provided, it should be understood that virtual sensor calculations can be used to determine other unknowns within the processing tool 200. For example, virtual sensor implementations can be used to determine unknowns such as, but not limited to, the gas composition at different locations within the chamber, the deposition rate on the wafer, the pressure on the wafer, and the film composition on the wafer.

[0042] In one embodiment, one or more temperature sensors 207 are disposed within chamber 205. For example, temperature sensor 207 may be a thermocouple or the like. In one embodiment, temperature sensor 207 may be disposed on a reflector plate (not shown) of the chamber. In some embodiments, temperature sensor 207 may be considered a witness sensor; that is, temperature sensor 207 may be located outside the control loop.

[0043] Temperature sensor 207 can provide additional known variables to enable a wider implementation of virtual sensors. In one embodiment, temperature sensor 207 can also be used to determine when a steady state has been reached in chamber 205. This is particularly beneficial when the processing tool 200 is brought up from a cold state, such as after a maintenance event. For example, the output of temperature sensor 207 in combination with one or more pressure gauges 213, 217, 218, and 219, as well as the angle of throttle valve 214, can be monitored, and the chamber can be ready for use when the various sensors reach a steady state. In one embodiment, monitoring when the chamber reaches a steady state is useful because it eliminates wafer scrap or rework typically experienced due to the first wafer effect in the processing tool.

[0044] Now for reference Figure 3 The figure illustrates a schematic diagram of a processing tool 300 according to an embodiment. As shown, an algorithm server 320 can be integrated with the processing tool 300. For example, as indicated by the arrow, the algorithm server 320 can be communicatively coupled to a front-end server 360 via a network connection. However, in other embodiments, the algorithm server 320 can be external to the processing tool 300. For example, the algorithm server 320 can be communicatively coupled to the processing tool 300 via an external network or the like.

[0045] In one embodiment, algorithm server 320 may include one or more of detection block 161, correction block 162, and prediction block 163. That is, algorithm server 320 may include machine learning and / or hybrid models for detecting, correcting, and / or predicting drift in processing tool 300. In the illustrated embodiment, algorithm server 320 is illustrated as hosting a hybrid model. The hybrid model may include physical model 327 and statistical model 325. Statistical model 325 and physical model 327 may be communicatively coupled to database 330, which stores input data (e.g., sensor data, model data, metrological data, etc.) used to construct and / or update statistical model 325 and physical model 327. In one embodiment, statistical model 325 may be generated by implementing a physical DoE and using interpolation to provide an extended process space model. In one embodiment, physical model 327 may be generated using real-world physical and chemical relationships. For example, physical and chemical equations for various interactions within a processing chamber may be used to build the physical model.

[0046] In one implementation, processing tool 300 may include a front-end server 360, a tool control server 350, and tool hardware 340. The front-end server 360 may include a dashboard 365 for the algorithm server 320. The dashboard 365 provides a user interface for processing engineers to leverage data modeling to perform various operations, such as drift monitoring, drift correction, and drift prediction.

[0047] The tool control server 350 may include an intelligent monitoring and control block 355. The intelligent monitoring and control block 355 may include modules for providing diagnostics and other monitoring of the processing tool 300. These modules may include, but are not limited to, health checks, sensor drift, fault recovery, and leak detection. The intelligent monitoring and control block 355 may receive data as input from various sensors implemented in the tool hardware. Sensors may include standard sensors 347 typically present in the semiconductor manufacturing tool 300 to allow operation of the tool 300. For example, sensor 347 may include control loop sensors, such as those described above. Sensors may also include witness sensors 345 added to the tool 300. Witness sensors 345 provide additional information necessary to build a highly detailed data model. For example, witness sensors may include physical sensors and / or virtual sensors. As described above, virtual sensors may utilize data obtained from two or more physical sensors and use computation to provide additional sensor data that cannot be obtained from physical sensors alone. In a particular example, a virtual sensor may utilize upstream and downstream pressure sensors to calculate the flow rate through a portion of the processing tool, such as a gas cylinder. Typically, witness sensors can include any type of sensor, such as, but not limited to, pressure sensors, temperature sensors, and gas concentration sensors. In one embodiment, the intelligent monitoring and control block 355 can provide data that is used by the algorithm server 320. In other embodiments, output data from various witness sensors 345 can be directly provided to the algorithm server 320.

[0048] Now for reference Figure 4A The figure illustrates a schematic diagram of a dashboard 465 according to one embodiment. As shown, the dashboard 465 may include multiple different parameters 467 of the processing tool. 1-n These parameters can be monitored using one or more of a detection block, a correction block, and a prediction block. Parameter 467 displayed on dashboard 465 can be a key parameter of the process operation being monitored. For example, in a lamp-based chamber, key parameter 467 may include one or more of gas flow rate, pressure, temperature, deposition characteristics, coating amount on the chamber wall, and leak detection. In the case of a heater-based chamber, key parameter 467 may include one or more of pressure, temperature, deposition characteristics, coating amount on the chamber wall, and free radical density. In the case of a plasma-based chamber, key parameter 467 may include gas flow rate, pressure, plasma density, leak detection, temperature, RF parameters, and coating amount on the chamber wall.

[0049] In one implementation, parameter 467 can provide a visual state indicator. For example, green, yellow, or red dots can be displayed to provide a general indication of the state of a given parameter. That is, a green dot can indicate that the process is operating without drift, a yellow dot can indicate that the process is moving away from the desired operating range, and a red dot can indicate that the process has drifted beyond a predetermined range. However, it should be understood that any visual mechanism (with any desired resolution) can be used to provide a simple indication of drift.

[0050] Dashboard 465 can also provide additional information about a given parameter 467. For example, a processing engineer can click on a parameter in parameter 467 to bring up additional information. Figure 4B This is an illustration of a statistical process control (SPC) chart 468 with given parameters 467. Each point on the SPC chart 468 can be a representation of a wafer or substrate that has been processed by a processing tool. In one embodiment, the chart illustrates the most recently processed wafers (e.g., the past 100 wafers). However, it should be understood that the history shown in the SPC chart 468 can include any number of wafers. The ability to track each wafer in the wafer array provides improved drift monitoring compared to existing methods. For example, existing methods rely on physical metrology to populate the SPC chart. This requires additional resources and time and is typically not feasible for each wafer or substrate.

[0051] SPC chart 468 can be used to visually indicate when a process begins to drift. For example... Figure 4B As shown, these points exhibit an upward trend. This may indicate that the processing parameters are drifting, and the subsequently processed wafers are at risk of exceeding specifications. Similarly, if these points exhibit a downward trend, drift may also be occurring. In some implementations, individual points on the SPC chart can be clicked to provide further details for individual substrates. For example, in some implementations, a chart of consistency data on a given substrate or wafer can be provided.

[0052] Figure 5A graphical representation of a machine is illustrated in the exemplary form of a computer system 500, in which a set of instructions can be executed to cause the machine to perform any one or more of the methods described herein. In alternative embodiments, the machine may be connected (e.g., networked) to other machines in a Local Area Network (LAN), intranet, extranet, or the Internet. The machine may operate as a server or client machine in a client-server network environment, or as a peer machine in a peer (or distributed) network environment. The machine may be a personal computer (PC), tablet PC, set-top box (STB), personal digital assistant (PDA), cellular phone, network device, server, network router, switch, or bridge, or any machine capable of executing a set of instructions (sequentially or otherwise) specifying the actions to be taken by the machine. Furthermore, although only a single machine is illustrated, the term "machine" should also be understood to include any collection of machines (e.g., computers) that individually or jointly execute a set (or more) of instructions to perform any one or more of the methods described herein.

[0053] An exemplary computer system 500 includes a processor 502, main memory 504 (e.g., read-only memory (ROM), flash memory, dynamic random access memory (DRAM, such as synchronous DRAM (SDRAM) or Rambus DRAM (RDRAM)), static memory 506 (e.g., flash memory, static random access memory (SRAM), MRAM), and auxiliary memory 518 (e.g., a data storage device), which communicate with each other via a bus 530.

[0054] Processor 502 represents one or more general-purpose processing devices, such as microprocessors, central processing units, etc. More specifically, processor 502 may be a complex instruction set computing (CISC) microprocessor, a reduced instruction set computing (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, a processor implementing other instruction sets, or a processor implementing a combination of instruction sets. Processor 502 may also be one or more special-purpose processing devices, such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), network processors, etc. Processor 502 is configured to execute processing logic 526 to perform the operations described herein.

[0055] Computer system 500 may also include a network interface device 508. Computer system 500 may also include a video display unit 510 (e.g., a liquid crystal display (LCD), a light-emitting diode display (LED), or a cathode ray tube (CRT)), an alphanumeric input device 512 (e.g., a keyboard), a cursor control device 514 (e.g., a mouse), and a signal generation device 516 (e.g., a speaker).

[0056] Auxiliary storage 518 may include machine-accessible storage medium (or more particularly, computer-readable storage medium) 532, on which one or more sets of instructions (e.g., software 522) embodying any one or more of the methods or functions described herein are stored. During execution by computer system 500, software 522 may also reside wholly or at least partially within main memory 504 and / or processor 502, which also constitute machine-readable storage media. Software 522 may also be transmitted or received on network 520 via network interface device 508.

[0057] Although machine-accessible storage medium 532 is illustrated as a single medium in the exemplary embodiment, the term "machine-readable storage medium" should be understood to include a single medium or multiple media (e.g., a centralized or distributed database, and / or associated caches and servers) that store a set or more sets of instructions. The term "machine-readable storage medium" should also be understood to include any medium capable of storing or encoding a set of instructions for execution by a machine and for enabling the machine to perform any one or more of the methods described in this disclosure. Therefore, the term "machine-readable storage medium" should be understood to include, but is not limited to, solid-state storage and optical and magnetic media.

[0058] According to embodiments of this disclosure, a machine-accessible storage medium has instructions stored thereon that cause a data processing system to perform methods for monitoring, correcting, and / or predicting drift in a processing tool.

[0059] Therefore, methods and apparatus for monitoring, correcting and / or predicting drift in processing tools have been disclosed.

Claims

1. A processing tool, comprising: chamber; A control loop sensor, wherein the control loop sensor is a sensor used to control the processing conditions within the chamber; Multiple witness sensors are integrated with the chamber, wherein the multiple witness sensors are sensors located outside the control loop of the processing conditions of the chamber; A drift detection module, wherein data from the plurality of witness sensors is provided to the drift detection module as input data; as well as A dashboard is used to display output data from the drift detection module. The plurality of witness sensors are configured to monitor the drift of the control loop sensors.

2. The processing tool of claim 1, wherein the drift detection module utilizes a machine learning algorithm to process the input data from the plurality of witness sensors.

3. The processing tool of claim 1, wherein the drift detection module utilizes a hybrid model to process the input data from the plurality of witness sensors.

4. The processing tool according to claim 3, wherein the hybrid model includes a physical model and a statistical model.

5. The processing tool of claim 1, wherein the output data from the drift detection module includes statistical process control (SPC) charts.

6. The processing tool according to claim 1, further comprising: Processing correction module.

7. The processing tool according to claim 6, wherein the processing correction module comprises: A correction algorithm, wherein the output data from the drift detection module is fed into the correction algorithm as input, and wherein the output from the correction algorithm is a control force; and A chamber control interface, wherein the control force causes the chamber control interface to change one or more tool settings of the processing tool.

8. The processing tool according to claim 7, further comprising: Processing prediction module.

9. The processing tool according to claim 8, wherein the processing prediction module comprises: Continuous learning system; Prediction algorithm; as well as Self-calibration module.

10. The processing tool of claim 1, wherein the chamber is a lamp-based chamber.

11. The processing tool of claim 10, wherein the output data from the drift detection module includes one or more processing parameters, wherein the one or more processing parameters include one or more of gas flow rate, pressure, temperature, deposition characteristics, coating amount on chamber wall, and leak detection.

12. The processing tool of claim 1, wherein the chamber is a heater-based chamber.

13. The processing tool of claim 12, wherein the output data from the drift detection module includes one or more processing parameters, wherein the one or more processing parameters include one or more of pressure, temperature, deposition characteristics, coating amount on the chamber wall, and free radical density.

14. The processing tool of claim 1, wherein the chamber is a plasma-based chamber.

15. The processing tool of claim 14, wherein the output from the drift detection module includes one or more processing parameters, wherein the one or more processing parameters include one or more of gas flow rate, pressure, plasma density, leak detection, temperature, RF parameters, and coating amount on the chamber wall.

16. A processing tool, comprising: Physical tools, wherein the physical tools include: A control loop sensor, wherein the control loop sensor is a sensor used to control the processing conditions within the chamber; and Witness sensor, wherein the witness sensor is a sensor located outside the control loop of the processing conditions within the chamber and is configured to monitor the drift of the control loop sensor; A drift detection module, wherein the drift detection module receives control loop sensor data and witness sensor data as input, and wherein the drift detection module outputs processing parameter data indicating whether one or more processing parameters have drifted.

17. The processing tool according to claim 16, further comprising: A processing correction module, wherein the processing correction module receives the processing parameter data as input and outputs control force to change one or more tool settings in the tool settings of the physical tool.

18. The processing tool according to claim 17, further comprising: A drift prediction module, wherein the drift prediction module receives control loop sensor data and witness sensor data as input, and wherein the drift prediction module outputs prediction data indicating when the physical tool will operate outside a threshold.

19. A processing tool, comprising: Physical tools, including: chamber; A cylinder for allowing one or more processing gases to flow into the chamber from multiple gas sources; A mass flow controller, the mass flow controller being used for each of the plurality of gas sources; A mass flow meter is located between the gas source and the cylinder. A first pressure gauge is located between the mass flow meter and the cylinder; A second pressure gauge, the second pressure gauge being fluidly coupled to the chamber; and A discharge line, the discharge line being coupled to the chamber; A drift detection module, wherein the drift detection module receives data as input from one or more of the mass flow controller, the mass flow meter, the first pressure gauge, and the second pressure gauge, and wherein the drift detection module outputs processing parameter data, wherein... The mass flow controller and the second pressure gauge are control loop sensors used to control the processing conditions within the chamber. The mass flow meter and the first pressure gauge are witness sensors located outside the control loop of the processing conditions within the chamber, and The witness sensor is configured to monitor the drift of the control loop sensor.

20. The processing tool of claim 19, wherein the drift detection module includes one or both of a hybrid model and a machine learning module, wherein the hybrid model includes a physical model and a statistical model.

Images