Data capture and transformation for data analytics and machine learning supporting substrate manufacturing systems

Abstract

A data collection system for semiconductor manufacturing includes: T substrate processing tools, each of the T substrate processing tools includes: N processing chambers, each of the N processing chambers includes a processing chamber controller, the processing chamber controller is configured to receive a plurality of different types of data during operation of a corresponding processing chamber in the N processing chambers, wherein the plurality of different types of data have different formats, wherein the processing chamber controller is further configured to format the plurality of different types of data into formatted data, and wherein T and N are integers; and a data diagnostic service computer, which is configured to: receive the formatted data and store it in a shared document by category, the shared document having a tabular data structure, the tabular data structure including rows with context data; and in response to a request, generate an output document containing a subset of the data from the shared document.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] The present disclosure is a PCT international application of U.S. Patent Application No. 62 / 805,454 filed on February 14, 2019. The entire contents of the above application are incorporated herein by reference. Technical Field

[0003] The present disclosure relates to substrate processing systems and, more particularly, to data capture and transformation for supporting data analytics and machine learning for substrate manufacturing systems. Background Art

[0004] The background description provided here is for the purpose of generally presenting the context of the present disclosure. The work of the presently designated inventors is neither explicitly nor implicitly admitted to be prior art against the present disclosure to the extent that it is described in this background section and in aspects of the specification that were not determined to be prior art at the time the application was filed.

[0005] Semiconductor manufacturers use one or more substrate processing tools to perform deposition, etching, cleaning, and / or other substrate processing during the manufacture of semiconductor wafers. Each of these substrate processing tools may include multiple processing chambers that perform the same type of processing (e.g., deposition, etching, or cleaning) or different processing (e.g., a series or sequence of processing).

[0006] The processing chambers in a substrate processing tool typically perform the same work repeatedly on multiple substrates. The processing chambers operate based on a recipe that defines process parameters. For example, a recipe defines the sequence, operating temperature, pressure, gas chemistry, plasma usage, time period for each operation or sub-operation, and / or other parameters. When manufacturing variations and / or defects occur, semiconductor manufacturers need to determine the root cause. Summary of the invention

[0007] In one feature, a data collection system for semiconductor manufacturing includes: T substrate processing tools, each of the T substrate processing tools includes: N processing chambers, each of the N processing chambers includes a processing chamber controller, the processing chamber controller is configured to receive a plurality of different types of data during operation of a corresponding processing chamber in the N processing chambers, wherein the plurality of different types of data have different formats, wherein the processing chamber controller is further configured to format the plurality of different types of data into formatted data, and wherein T and N are integers; and a data diagnostic service computer, which is configured to: receive the formatted data and store it by category in a shared document, the shared document having a tabular data structure, the tabular data structure including rows with context data; and in response to a request, generate an output document from the shared document including a subset of the data.

[0008] In further features, the context data includes a material identification field and a time field; and the plurality of different types of data include: at least three or more data types selected from a group consisting of recipes, hardware calibration, process fine-tuning, optical emission spectroscopy, high-speed data, and variable tracking.

[0009] In further features, the data diagnostic service computer is configured to generate the output document based on at least one of: a subset of the categories; a subset of the material identification fields; and a subset of the time fields.

[0010] In further features, the shared documents have a three-level hierarchy including documents, groups, and categories.

[0011] In further features, the shared document has a Technical Data Management Scheme (TDMS) format.

[0012] In further features, data is stored in the shared document in binary format.

[0013] In further features, the machine learning computer is configured to communicate with the data diagnostic service computer and perform machine learning on the data in the output document.

[0014] In further features, the processing chamber controller includes: a data generation object configured to transmit the multiple different types of data; a data scheduler configured to receive the multiple different types of data from the data generation object; a plurality of data slots, wherein each of the multiple data slots is configured to receive a corresponding type of data among the multiple different types of data from the data scheduler; a plurality of data formatters, wherein each of the multiple data formatters is configured to receive a corresponding type of data among the multiple different types of data from the data slots and output the formatted data back to a corresponding data slot among the multiple data slots; and a data interface manager configured to transfer the formatted data from the multiple data slots to the data diagnostic service computer.

[0015] In further features, the data diagnostic service computer includes: a data manager configured to receive the formatted data from the chamber controller; and a data storage configured to store the shared document and control access to the shared document.

[0016] In further features, a data adapter is configured to receive the request and generate the output document.

[0017] In further features, the data adapter is further configured to adapt the output document to one of a plurality of data transports.

[0018] In further features, the plurality of data transmitters include at least one of a file transfer protocol (FTP) transmitter, a machine learning transmitter, and an archive transmitter.

[0019] In further features, the host server includes a document selector configured to select R portions of data from R output documents of R of the T substrate processing tools, where R is an integer greater than one and less than or equal to T; and the machine learning computer is configured to combine the data in the R output documents and perform machine learning on them.

[0020] In further features, a data collection method for semiconductor manufacturing includes: receiving multiple different types of data during operation of multiple processing chambers among N processing chambers of T substrate processing tools, where T and N are integers greater than zero, and wherein the multiple different types of data have different formats; formatting the multiple different types of data into formatted data; storing the formatted data by category in a shared document, the shared document having a tabular data structure, the tabular data structure including rows with context data; and generating an output document including a subset of the data from the shared document in response to a request.

[0021] In further features, the context data includes a material identification field and a time field; and the plurality of different types of data include: at least three or more data types selected from a group consisting of recipes, hardware calibration, process fine-tuning, optical emission spectroscopy, high-speed data, and variable tracking.

[0022] In further features, generating the output document comprises generating the output document based on at least one of: a subset of the categories; a subset of the material identification fields; and a subset of the time fields.

[0023] In further features, the shared documents have a three-level hierarchy including documents, groups, and categories.

[0024] In further features, the shared document has a Technical Data Management Scheme (TDMS) format.

[0025] In further features, data is stored in the shared document in binary format.

[0026] In further features, the data collection method further comprises: performing machine learning on the data in the output document.

[0027] Further scope of applicability of the present disclosure will become apparent from the detailed description, claims and drawings.The detailed description and specific examples are intended for purposes of illustration only and are not intended to limit the scope of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The present disclosure will be more fully understood from the detailed description and accompanying drawings, in which:

[0029] Figure 1 is a functional block diagram of an example of a substrate processing tool including one or more processing chambers;

[0030] Figure 2 is a functional block diagram of an example of a substrate processing chamber connected to a data diagnostic service computer according to the present disclosure;

[0031] Figure 3 is a functional block diagram of an example of a data collection architecture for a substrate manufacturer according to the present disclosure;

[0032] Figures 4A to 4G A diagram showing an example of data handling;

[0033] Figure 5 is a functional block diagram of an example of data flow from a source to a data diagnostic service computer according to the present disclosure;

[0034] Figure 6Functional block diagrams for examples of data handling and formatting;

[0035] Figure 7 A flowchart showing an example of a method for data handling and formatting;

[0036] Figure 8 is a functional block diagram of an example of a system for updating variables and generating tracking data; and

[0037] Fig. 9 A flowchart of an example of a method for data handling is provided.

[0038] Among the drawings, reference numerals may be repeated to identify similar and / or identical elements. DETAILED DESCRIPTION

[0039] A large amount of data, such as structured data, is generated during the operation of substrate processing tools. For example, to mass-produce integrated circuits, semiconductor manufacturers may configure arrays of substrate processing tools that are configured to manufacture semiconductor wafers. Since production uniformity is very important for reducing non-uniformities and / or defects, it is important to ensure that the same steps performed by each of the process chambers in each of the substrate processing tools are performed in the same manner to produce the same results. Substrate manufacturing requires precise control of process parameters. One method of identifying the root cause of defects or non-uniformities is to compare the process conditions of different substrates produced in the same process chamber and / or different process chambers.

[0040] The processing chambers typically contain sensors to sense chamber parameters, such as temperature distribution of electrostatic chucks, optical emission spectroscopy (OES) data, recipes, pressure, etc. The data generated is typically in a variety of different file formats, such as binary, binary MOSS, ASCII, XML, TDMS, etc. In addition, there is a large amount of data generated by each processing chamber. For example, each processing chamber can generate about 2MB of data per second during operation. When problems such as defects or non-uniformities occur in current systems, it is very difficult to capture the data, reformat the data into a usable format, and analyze the data due to the way it is collected and stored.

[0041] The systems and methods according to the present disclosure provide a data collection architecture for receiving, formatting, and organizing data generated by one or more substrate processing tools. The architecture enables data to be collected across multiple processing chambers, substrate processing tools, and / or locations so that it can be more easily analyzed. In the following description, examples of substrate processing tools and processing chambers are first described to provide context for the subsequent architecture description.

[0042] During processing, substrates are typically transported to substrate processing tools in a cassette. A transfer robot loads the substrate into a transfer station via a load lock. A vacuum robot moves the substrate from the cassette to the load lock. The vacuum robot moves the substrate from the load lock to one of the processing chambers for processing. After processing in a processing chamber, the vacuum robot moves the substrate to another processing chamber or back to the cassette via a load lock.

[0043] When in the processing chamber, the substrate is subjected to processing. For example, the temperature of the substrate can be controlled by a temperature-controlled electrostatic chuck (ESC). The pressure in the processing chamber can be controlled. One or more gas mixtures including one or more precursors, inert gases, etching gases, or other gas mixtures are introduced into the processing chamber. In some processes, a plasma can be triggered in the processing chamber or supplied by a remote plasma source to initiate a chemical reaction. Other processes may not involve the use of a plasma. An RF bias can be supplied to the substrate support to control the ion energy.

[0044] Reference now Figure 1 , which shows an exemplary substrate processing tool 100. The substrate processing tool 100 includes a plurality of processing chambers 104-1, 104-2, ..., and 104-M (collectively referred to as processing chambers 104) (where M is an integer greater than one). By way of example only, each of the processing chambers 104 may be configured to perform one or more types of substrate processing. In other words, a substrate may be loaded into one of the processing chambers 104, processed, and then moved to one or more other of the processing chambers 104 (if at least one performs a different process) and / or removed from the substrate processing tool 100 (if all perform the same process).

[0045] The substrate to be processed is loaded into the substrate processing tool 100 via a port of a loading station of an atmosphere to vacuum (ATV) transfer module 108. In some examples, the ATV transfer module 108 includes an equipment front end module (EFEM). The substrate is then transferred to one or more of the processing chambers 104. For example, the transfer robot 112 is configured to transfer the substrate from the loading station 116 to an air lock (or load lock) 120. The vacuum transfer robot 124 of the vacuum transfer module 128 is configured to transfer the substrate from the load lock 120 to each processing chamber 104.

[0046] A chamber controller 130, a transport controller 134, and / or a system controller 138 may be provided. The transport controller 134 controls the robots 112 and 124, actuators and sensors associated with the transport of substrates to and from the processing tool 100. The chamber controller 130 controls the operation of the processing chamber 104. Generally, the chamber controller 130 monitors sensors 135 (e.g., temperature sensors, pressure sensors, position sensors, etc.) and controls actuators 136 (e.g., robots, ports, heaters, gas delivery systems, ESCs, RF generators, etc.). The chamber controller 130 associated with the processing chamber 104 typically follows a recipe that specifies the timing of steps, process gases to be supplied, temperatures, pressures, RF power, etc.

[0047] Reference now Figure 2 , an exemplary substrate processing system 200 for processing a substrate is shown to illustrate examples of the types of data generated during operation of a processing chamber. Although a particular type of processing chamber is shown, other types of processing chambers may be used. The substrate processing system 200 includes a processing chamber 202, which surrounds the other components of the substrate processing system 200 and contains an RF plasma. The substrate processing system 200 includes an upper electrode 204 and a substrate support, such as an electrostatic chuck (ESC) 206. During operation, a substrate 208 is disposed on the ESC 206.

[0048] By way of example only, the upper electrode 204 may include a showerhead 209 that introduces and distributes process gases, such as deposition precursors, etching gases, carrier gases, etc. The ESC 206 includes a conductive bottom plate 210 that serves as a lower electrode. The bottom plate 210 supports a heating plate 212, which may correspond to a ceramic multi-zone heating plate. A thermal resistance layer 214 may be disposed between the heating plate 212 and the bottom plate 210. The bottom plate 210 may include one or more coolant channels 216 for flowing a coolant through the bottom plate 210.

[0049] The RF generation system 220 generates and outputs an RF voltage to one of the upper electrode 204 and the lower electrode (e.g., the bottom plate 210 of the ESC 206). The other of the upper electrode 204 and the bottom plate 210 may be DC grounded, AC grounded, or floating. By way of example only, the RF generation system 220 may include an RF voltage generator 222 that generates an RF voltage that is fed to the upper electrode 204 or the bottom plate 210 through a matching and distribution network 224. In other examples, the plasma may be generated inductively or remotely.

[0050] The gas delivery system 230 includes one or more gas sources 232-1, 232-2, ..., and 232-N (collectively referred to as gas sources 232), where N is an integer greater than zero. The gas source 232 supplies one or more deposition precursors, etching gases, carrier gases, etc. Vaporized precursors can also be used. The gas source 232 is connected to the manifold 238 by valves 234-1, 234-2, ..., and 234-N (collectively referred to as valves 234) and mass flow controllers 236-1, 236-2, ..., and 236-N (collectively referred to as mass flow controllers 236). The output of the manifold 238 is fed to the processing chamber 202. By way of example only, the output of the manifold 238 is fed to the showerhead 209.

[0051] An optical emission spectroscopy (OES) sensor 239 may be disposed near a window 240 disposed on a chamber surface 241. The OES sensor 239 selectively generates OES data. A temperature controller 242 may be connected to a plurality of thermal control assemblies (TCEs) 244 disposed in the heating plate 212. For example, the TCEs 244 may include, but are not limited to, corresponding large TCEs corresponding to corresponding regions in the multi-zone heating plate, and / or micro-TCE arrays arranged in multiple regions of the entire multi-zone heating plate. The temperature controller 242 may be used to control the plurality of TCEs 244 to control the temperature of the ESC 206 and the substrate 208.

[0052] The temperature controller 242 may communicate with the coolant assembly 246 to control the coolant flowing through the passage 216. For example, the coolant assembly 246 may include a coolant pump and a reservoir. The temperature controller 242 operates the coolant assembly 246 to selectively flow the coolant through the passage 216 to cool the ESC 206.

[0053] A valve 250 and a pump 252 may be used to evacuate reactants from the process chamber 202. A controller 260 may be used to control components of the substrate processing system 200. A robot 270 may be used to transfer substrates to and from the ESC 206. For example, the robot 270 may transfer substrates between the ESC 206 and a load lock 272. Although shown as a separate controller, the temperature controller 242 may be implemented within the controller 260.

[0054] As can be appreciated, during the processing and transfer of substrates, a large amount of discrete data and streaming data are generated by substrate processing tools. Event data can be used to determine the location and residence time in various components. For example, the residence time in a module or FOUP may cause processing differences between substrates. The system log records system level data. Additional data is recorded during substrate transfer. Each of the processing chambers also records data during substrate processing. However, the recorded data includes different data types, sampling rates and / or formats. Changing the data format can cause problems in historical data storage and retrieval. In addition, some data are only stored regionally at the processing chamber, while other data are only stored at the wafer factory level.

[0055] Data is typically streamed from the tool to the host in the form of information at a fixed frequency. Data is typically not transferred based on substrate. Data is transferred over time. Data is typically collected in files at a fixed frequency or file size. Data is typically collected continuously and without limits. In some systems, data is collected during processing of the initial and final substrates at the beginning of the recipe and at the end of the recipe, respectively, but not for intermediate substrates.

[0056] Semiconductor manufacturers cannot easily access all of the data in all of the above forms using existing fab data collection methods. Therefore, semiconductor manufacturers cannot gain comprehensive insights from various data segments to improve productivity, diagnose problems, and / or make process improvements. In summary, semiconductor manufacturers cannot access fab data in a usable format and organization. Semiconductor manufacturers currently cannot access certain data. Many data files are incomplete in their context information and / or lack consistency in timestamps or naming for the same subsystems.

[0057] For deep understanding, semiconductor manufacturers may wish to merge data from different documents with different formats. The contextual data added to the data files is inconsistent, which requires a lot of data wrangling to correct. Some data is not of sufficient quality to be trusted.

[0058] Systems and methods according to the present disclosure capture, transmit, store, and / or convert data generated during substrate processing to facilitate subsequent analysis using big data tools and / or machine learning.

[0059] In some examples, the data has a common structured file format that supports encryption and forward / reverse compatibility. In some examples, the data format is consistent with a Technical Data Management Scheme (TDMS) document format. In some examples, the data format is accessible and can be read by a customer using a key (e.g., an application protocol interface (API) key). The data context and collection are consistent across different data types to enable use without cleaning, reformatting, and / or other work.

[0060] In some examples, data is collected from the moment when the substrate leaves the box until the substrate returns to the box. In some examples, the data can be stored in a single file. In some examples, the file format is self-consistent. In some examples, data is collected at different frequencies based on data type and context. In some examples, encrypted TDMS is used to format the data. In some examples, data adapters are used to support legacy data and backward compatibility, to incorporate new data types and support host communication / streaming data. Additional details will be further described below.

[0061] Reference now Figure 3 , a wafer fab data collection system 300 for a semiconductor processing system is shown. The wafer fab data collection system 300 includes N substrate processing tools 320-1, 320-2, ..., and 320-T (where T is an integer) (collectively referred to as substrate processing tools 320). Each of the substrate processing tools 320-1, 320-2, ..., and 320-T includes one or more chamber controllers 340-1, 340-2, ..., and 340-T (collectively referred to as chamber controllers 340) as described above to control the operation of the processing chamber. Examples of data collected by the processing chamber include substrate or wafer data logs, automatic preventive maintenance, high-speed data, optical emission spectroscopy (OES), tracking data, OES snap shots, susceptor temperature maps and other data, calibration files, equipment constants, sensor data, and / or other data.

[0062] Each of the substrate processing tools 320-1, 320-2, ..., and 320-T may include a transport controller 344-1, 344-2, ..., and 344-N (collectively referred to as transport controllers 344) to control dynamic alignment and store calibration files, platform tracking data logs, equipment constants, transfer module jobs, and / or other data. Dynamic alignment involves the position of the substrate relative to other chamber components (such as the center of a pedestal, edge ring, or other object).

[0063] Each of the substrate processing tools 320-1, 320-2, ..., and 320-T may include a tool system controller 348-1, 348-2, ..., and 348-N, respectively (collectively referred to as tool system controllers 348). The tool system controllers 348 record batch history, detailed event logs, batch-based alarms, time-based alarms, tool controller health, part tracking, part history, material scheduling, and other data.

[0064] Each of the substrate processing tools 320-1, 320-2, ..., and 320-T also includes a data diagnostic service computer 350-1, 350-2, ..., and 350-N (collectively referred to as data diagnostic service computers 350), and data storage devices 362-1, 362-2, ..., and 362-N (collectively referred to as data storage devices 362). In other examples, the data diagnostic service computer 350 may be shared by two or more tools, or each tool may include more than one data diagnostic service computer 350. The substrate processing tools 320 are connected to a host server 364 via one or more buses (e.g., a tool data bus or network 364-1 and a streaming data bus or network 364-2).

[0065] In some examples, the host server 364 includes a security module 366 and a data selector module 367. The security module 366 provides security, such as encryption or password protection. The security module 366 uses encryption or passwords to authorize or deny access to data stored by the substrate processing tool 320 and / or to the data selector module 367. The host server 364 also includes a data selector module 365 to enable the user computer 380 to select one or more data categories from one or more substrate processing tools and filter the data using one or more data context fields. In other examples, the security module 366 and / or the data selector module 367 are implemented using separate servers.

[0066] The host server 364 is connected to a machine learning computer 374 and / or one or more user computers 380 via a network 368 (e.g., a WAN or LAN). The machine learning computer 374 can access the data set sent back by the host server 364 for further analysis. In some examples, the machine learning computer 374 includes a security module 375 to control access to the data. The machine learning computer 374 uses one or more data files generated by the data collection system selected by the user to perform machine learning. Since the format of the files from different substrate processing tools is the same, the data can be merged into a single file and analyzed. This enables the same process to be analyzed in multiple machines.

[0067] As can be appreciated, the number T of substrate processing tools 320 is not limited. In addition, the substrate processing tools 320 need not be located in the same facility. In some examples, an equipment manufacturer may be granted access to data stored by multiple semiconductor manufacturers.

[0068] In some examples, the sampling rates of some or all data generating devices (except streaming or HSD devices) may be aligned to a common sampling period, and data may be added to a shared document (described below) based on the sampling period.

[0069] Reference now Figures 4A to 4G , the data collection system according to the present invention implements changes in the way data is transmitted, stored, and / or formatted. Figure 4A In the embodiment of the present invention, data is pre-stored in one or more large substrate data log files 400, which store different types of data in various formats. For example, the substrate data log file 400 may store data including substrate context 410, recipe 414, hardware calibration 418, process fine tuning 422, and one or more variable tracking 425.

[0070] exist Figure 4B In the example, instead of using a large substrate data log file 400 with different types of data and formats, the data collection system according to the present disclosure divides the data into data categories 450-1, 450-2, 450-3, ..., and 450-C (where C is an integer) and adds context data (e.g., material ID, batch ID, process job ID, time, module, etc.). In some examples, the data is stored in a shared document with a tabular data structure (which is similar to a sparsely populated table). The categories of the data are stored in the data structure together with the context data.

[0071] exist Figure 4C In the example, data storage device 370 is used to store data structures. In some examples, data storage device 370 represents a distributed storage system that includes each of the data storage devices 362 associated with the substrate processing tool 320. In other words, as each of the substrate processing tools 320 operates, it generates categories of data and contextual data and populates documents in the corresponding ones of the data storage devices 362. Then, the data selector module 367 enables the user to extract data from one or more of the data storage devices 362. This process can be repeated one or more times to generate new data sets that can be used for further analysis or machine learning without data wrangling or format changes.

[0072] exist Figure 4DIn the example of FIG. 374, as further described below, the semiconductor manufacturer may select a subset of one or more data categories and limit the data set returned by filtering one or more fields in the context data. In other words, the user computer 374 generates one or more requests for wafer fab data by specifying one or more categories and filtering one or more context data fields. This process may be repeated one or more times to generate new data sets for further analysis. Figure 4E In response to a request for data with a material ID, IDs or ID ranges, time ranges, and / or categories, the data diagnostic service computer selects categories limited by material ID and time range from the data storage device 370 for the user (as shown at 480) and generates an output document 482 for further analysis.

[0073] Reference now Figure 4F and 4G , shows an example of a suitable file format or data structure. Figure 4F In , data is appended to a document that contains a linked list. Once data is added to a document, it cannot be deleted or changed. New data is appended to the document. The appended data contains a header and data. The header of the last entry points to the next entry in the linked list. In Figure 4G In , the document contains 3 layers of hierarchy, which include documents, groups, and channel properties. Each item in each layer can contain unique properties, and there is no limit to the number of groups, channels, or channel properties. Groups are similar to tables. Each row represents a time point or an index. Each column represents data. In some examples, the document contains a TDMS document. TDMS documents are binary and are compressed more efficiently than ASCII documents. TDMS documents are written and read faster than other formats. Since the documents are stored in binary, sensitive data is protected. In some examples, streaming data (such as HSD (high speed data up to 1kHz) and / or OES (optical emission spectroscopy)) can be stored in the document or in a separate substrate data log file.

[0074] Reference now Figure 5, shows an example of data handling and processing performed by a data diagnostic service computer. Controller 510 (e.g., one of the chamber controllers, tool system controllers, or transport module controllers described above) includes one or more images 512. In some examples, image 512 includes a Smalltalk image, but other types may be used. Image 512 includes a variable data handler 514 and / or a data object handler 518. In some examples, variable data handler 514 and / or data object handler 518 output data via tool network bus 519. Other data (at 530) such as HSD and / or OES is transmitted via data network bus 521 and may include raw streaming data.

[0075] The data diagnostic service computer 350 includes a data manager 540, which receives a data stream from the controller 510. The data manager 540 outputs the data stream to a data storage 544. The data adapter 548 receives a request from the user computer 380. The data adapter 548 outputs the request to the data storage 544 and receives data from the data storage 544. The data adapter 548 adapts the data according to the needs of the destination and outputs the data to one of the multiple transmitters 550. The operation of adapting the data may include further formatting and / or other operations for the specified transmission target. The selected one of these transmitters transmits or transmits the data to the correct target (e.g., machine learning, document library, or FTP). In some examples, the transmitter 550 includes at least one of a machine learning (ML) transmitter 552, a document library transmitter 554, a file transfer protocol (FTP) transmitter (not shown), or other transmitters 558.

[0076] Reference now Figure 6 , image 600 includes data generation objects 610 that generate data that is output to data scheduler 614. In some examples, image 600 includes a Smalltalk image and data generation objects 610 are Smalltalk objects, but other types of images and objects may be used. In some examples, data generation objects 610 generate data related to recipes, process events, hardware calibration, process fine-tuning, etc.

[0077] The data scheduler 614 forwards the data from the data scheduler 614 to one of the plurality of data slots 618 based on the data category. One of the plurality of data slots 618 transmits the data to the data formatter 620, which applies one of the plurality of data formats to the data. In some examples, the data formatter 620 applies a TDMS compatible format to the data.

[0078] The data formatter 620 returns the formatted data to the data slot 618. The data slot 618 transmits the formatted data to the data manager interface 624. The data manager interface 624 converts the formatted data into a string (e.g., Javascript Object Notation (JSON)) and transmits the data to the data manager 540 of the data diagnostic service computer 350.

[0079] Reference now Figure 7 , a method 650 for operating an image 600 is shown. At 660, the data generation object 610 transmits the data to the data scheduler 614. At 664, the data scheduler 614 transmits the data to one of a plurality of data slots 618 based on the data category. At 668, the data slot 618 transmits the data to the data formatter 620. At 672, the data formatter 620 formats the data (using the category-based format) and transmits the data back to the data slot 618. At 676, the data manager interface 624 converts the formatted data and transmits the data to the data manager 540.

[0080] Reference now Figure 8 , image 810 (associated with one of the controllers) includes a variable handler 812 that generates variable updates that are output to a variable handler 820 of image 822. For example, variable updates regarding a process chamber are transmitted. The variable handler 820 generates variable updates (e.g., variable updates regarding a substrate processing tool) that are output to a tracking data collector 824. The tracking data collector 824 outputs sample records (e.g., tool sample records) to a tracking data agent 830. The tracking data agent 830 generates sample reports that are output to the data manager 540 of the data diagnostic service computer 350 in the form of binary arrays. In some examples, the tracking data agent 830 outputs binary data such as arrays via streaming sockets.

[0081] Reference Fig. 9, method 900 is shown. When it is determined at 910 that a processing event occurs, the processing event is sent to a transmission interface at 920. At 924, the transmission interface forwards the processing event to a data scheduler. At 928, the data scheduler determines the category of the input data and transmits the data to a corresponding data slot in the data slot. At 934, the data slot transmits the data (e.g., a Smalltalk object) received from the data scheduler to a corresponding data formatter. At 938, the data formatter formats the data into a managed data object and transmits the data to the data slot. At 950, the data slot transmits the data to a data manager interface. At 954, the data manager interface converts the managed data object into a string (e.g., a JSON string) and transmits the string to the data manager via a connection such as an HTTP connection.

[0082] The foregoing description is merely illustrative in nature and is by no means intended to limit the present disclosure, its application or use. The broad teachings of the present disclosure can be implemented in various forms. Therefore, although the present disclosure includes specific examples, the true scope of the present disclosure should not be so limited, because when studying the drawings, the specification and the appended claims, other modifications will become apparent. It should be understood that, without changing the principles of the present disclosure, one or more steps in the method can be performed in different orders (or simultaneously). In addition, although each embodiment is described above as having certain features, any one or more of those features described relative to any embodiment of the present disclosure can be implemented in the features of any other embodiment and / or combined with the features of any other embodiment, even if the combination is not clearly described. In other words, the described embodiments are not mutually exclusive, and the replacement of one or more embodiments with each other remains within the scope of the present disclosure.

[0083] Various terms are used to describe the spatial and functional relationships between elements (e.g., between modules, between circuit elements, between semiconductor layers, etc.), including "connected," "engaged," "coupled," "adjacent," "next to," "on top of," "above," "below," and "disposed." Unless the relationship between a first and a second element is explicitly described as "direct," when such a relationship is described in the above disclosure, the relationship can be a direct relationship, in which there are no other intervening elements between the first and second elements, but can also be an indirect relationship, in which there are one or more intervening elements (spatially or functionally) between the first and second elements. As used herein, the phrase "at least one of A, B, and C" should be interpreted to mean a logical (A or B or C), using a non-exclusive logical OR, and should not be interpreted to mean "at least one of A, at least one of B, and at least one of C."

[0084] In some implementations, the controller is part of a system, which can be part of the above examples. Such a system can include a semiconductor processing device, which includes one or more processing tools, one or more chambers, one or more platforms for processing, and / or specific processing components (wafer pedestals, gas flow systems, etc.). These systems can be integrated with electronic devices for controlling their operations before, during, and after the processing of semiconductor wafers or substrates. The electronic device can be referred to as a "controller", which can control various components or subcomponents of one or more systems. Depending on the processing requirements and / or system type, the controller can be programmed to control any process disclosed herein, including the delivery of process gases, temperature settings (e.g., heating and / or cooling), pressure settings, vacuum settings, power settings, radio frequency (RF) generator settings, RF matching circuit settings, frequency settings, flow rate settings, fluid delivery settings, position and operation settings, wafer transfer in and out tools and other transfer tools and / or load locks connected to or connected to a specific system through an interface.

[0085] In general, a controller can be defined as an electronic device having various integrated circuits, logic, memory, and / or software that receive instructions, issue instructions, control operations, enable cleaning operations, enable endpoint measurements, etc. The integrated circuit can include a chip in the form of firmware that stores program instructions, a digital signal processor (DSP), a chip defined as an application specific integrated circuit (ASIC), and / or one or more microprocessors, or a microcontroller that executes program instructions (e.g., software). The program instructions can be instructions sent to the controller in the form of various individual settings (or program files) that define operating parameters for performing a specific process on or for a semiconductor wafer or system. In some embodiments, the operating parameters can be part of a recipe defined by a process engineer to complete one or more processing steps during the manufacture of one or more (kinds of) layers, materials, metals, oxides, silicon, silicon dioxide, surfaces, circuits, and / or dies of a wafer.

[0086] In some implementations, the controller may be part of or coupled to a computer that is integrated with the system, coupled to the system, otherwise networked to the system, or a combination thereof. For example, the controller may be in the "cloud" or all or part of a wafer fab host system that may allow remote access to wafer processing. The computer may enable remote access to the system to monitor the current progress of a manufacturing operation, check the history of past manufacturing operations, check trends or performance criteria for multiple manufacturing operations, change parameters of a current process, set processing steps to follow the current process, or start a new process. In some examples, a remote computer (e.g., a server) may provide a process recipe to the system via a network (which may include a local network or the Internet). The remote computer may include a user interface that enables input or programming of parameters and / or settings, which are then sent from the remote computer to the system. In some examples, the controller receives instructions in the form of data that specify parameters for each processing step to be performed during one or more operations. It should be understood that the parameters may be specific to the type of process to be performed and the type of tool that the controller is configured to interface with or control. Thus, as described above, the controller may be distributed, for example, by including one or more discrete controllers networked together and working toward a common purpose (e.g., the process and control described herein). An example of a distributed controller for such a purpose is one or more integrated circuits on a chamber communicating with one or more integrated circuits remotely (e.g., at a platform level or as part of a remote computer), which combine to control the process on the chamber.

[0087] Example systems may include, but are not limited to, plasma etch chambers or modules, deposition chambers or modules, spin rinse chambers or modules, metal plating chambers or modules, cleaning chambers or modules, chamfer edge etch chambers or modules, physical vapor deposition (PVD) chambers or modules, chemical vapor deposition (CVD) chambers or modules, atomic layer deposition (ALD) chambers or modules, atomic layer etch (ALE) chambers or modules, ion implantation chambers or modules, track chambers or modules, and any other semiconductor processing system that may be associated with or used in the manufacture and / or preparation of semiconductor wafers.

[0088] As described above, depending on one or more processing steps to be performed by the tool, the controller can communicate with one or more other tool circuits or modules, other tool components, cluster tools, other tool interfaces, adjacent tools, neighboring tools, tools located throughout the factory, a host computer, another controller, or tools used in material transport to transport wafer containers to and from tool locations and / or load ports in a semiconductor manufacturing facility.

Claims

1. A data collection system for semiconductor manufacturing, comprising: T substrate processing tools, wherein each of the T substrate processing tools comprises: N processing chambers, wherein each of the N processing chambers includes a processing chamber controller configured to receive a plurality of different types of data during operation of a corresponding processing chamber of the N processing chambers, The multiple different types of data have different formats, wherein the chamber controller is further configured to format the plurality of different types of data into formatted data, and wherein T and N are integers; and A data diagnostic service computer configured to: receiving the formatted data and storing it by category in a shared document having a tabular data structure including rows having contextual data; and In response to the request, an output document is generated from the shared document containing the subset of the data.

2. The data collection system according to claim 1, wherein: The context data includes a material identification field and a time field; and The plurality of different types of data include: at least three or more data types selected from the group consisting of recipe, hardware calibration, process fine-tuning, optical emission spectroscopy, high-speed data, and variable tracking.

3. The data collection system of claim 2, wherein the data diagnostic service computer is configured to generate the output document based on at least one of the following: a subset of said categories; a subset of the material identification fields; and A subset of the time field. 4 . The data collection system according to claim 1 , wherein the shared document has a three-level hierarchy including documents, groups, and categories.

5. The data collection system of claim 1, wherein the shared document has a Technical Data Management Scheme (TDMS) format.

6. The data collection system of claim 5, wherein data is stored in the shared document in a binary format.

7. The data collection system of claim 1, further comprising a machine learning computer configured to communicate with the data diagnostic service computer and perform machine learning on the data in the output document.

8. The data collection system of claim 1, wherein the chamber controller comprises: a data generation object configured to transmit the plurality of different types of data; a data scheduler configured to receive the plurality of different types of data from the data generation object; a plurality of data slots, wherein each of the plurality of data slots is configured to receive a corresponding type of data of the plurality of different types of data from the data scheduler; a plurality of data formatters, wherein each of the plurality of data formatters is configured to receive a corresponding type of data of the plurality of different types of data from the data slot and output the formatted data back to a corresponding data slot of the plurality of data slots; as well as A data interface manager is configured to transfer the formatted data from the plurality of data slots to the data diagnostic service computer.

9. The data collection system according to claim 1, wherein the data diagnosis service computer comprises: a data manager configured to receive the formatted data from the chamber controller; and A data storage device is configured to store the shared document and control access to the shared document.

10. The data collection system of claim 9, further comprising: a data adapter configured to receive the request and generate the output document.

11. The data collection system of claim 10, wherein the data adapter is further configured to adapt the output document to one of a plurality of data transports.

12. The data collection system of claim 11, wherein the plurality of data transmitters include at least one of a file transfer protocol (FTP) transmitter, a machine learning transmitter, and an archive transmitter.

13. The data collection system of claim 1, further comprising: a host server comprising a document selector configured to select R portions of data from R output documents of R of the T substrate processing tools, where R is an integer greater than one and less than or equal to T; and A machine learning computer is configured to combine and perform machine learning on the data in the R output documents.

14. A data collection method for semiconductor manufacturing, comprising: receiving a plurality of different types of data during operation of a plurality of process chambers among the N process chambers of the T substrate processing tools, Where T and N are integers greater than zero, and The multiple different types of data have different formats; formatting the plurality of different types of data into formatted data; storing the formatted data by category in a shared document, the shared document having a tabular data structure including rows having context data; and In response to the request, an output document is generated from the shared document containing the subset of the data.

15. The data collection method according to claim 14, wherein: The context data includes a material identification field and a time field; and The plurality of different types of data include: at least three or more data types selected from the group consisting of recipe, hardware calibration, process fine-tuning, optical emission spectroscopy, high-speed data, and variable tracking.

16. The data collection method of claim 15, wherein generating the output document comprises generating the output document based on at least one of the following: a subset of said categories; a subset of the material identification fields; and A subset of the time field.

17. The data collection method according to claim 14, wherein the shared document has a three-level hierarchy, which includes documents, groups and categories.

18. The data collection method of claim 14, wherein the shared document has a Technical Data Management Scheme (TDMS) format.

19. The data collection method according to claim 18, wherein the data is stored in the shared document in a binary format.

20. The data collection method of claim 14, further comprising: performing machine learning on the data in the output document.

Images