`issues for advanced process control
`
`Cite as: Journal of Vacuum Science & Technology A 19, 1241 (2001); https://doi.org/10.1116/1.1380225
`Submitted: 19 September 2000 . Accepted: 30 April 2001 . Published Online: 13 July 2001
`
`Richard J. Markle, and Elfido Coss
`
`ARTICLES YOU MAY BE INTERESTED IN
`
`The application of in situ monitor of extremely rarefied particle clouds grown thermally above
`wafers by using laser light scattering method to the development of the mass-production
`condition of the tungsten thermal chemical vapor deposition
`Journal of Vacuum Science & Technology A 19, 1248 (2001); https://doi.org/10.1116/1.1365127
`
`Journal of Vacuum Science & Technology A 19, 1241 (2001); https://doi.org/10.1116/1.1380225
`
`19, 1241
`
`© 2001 American Vacuum Society.
`
`Applied Materials, Inc. Ex. 1015
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 1 of 8
`
`
`
`Data requirements and communication issues for advanced process
`control
`Richard J. Marklea) and Elfido Coss, Jr.
`Advanced Micro Devices, Incorporated, 5204 E. Ben White Boulevard, Mailstop 568, Austin, Texas 78741
`~Received 19 September 2000; accepted 30 April 2001!
`
`Data streams and communication issues are the most critical areas for successful advanced process
`control ~APC! programs. These areas are vital for both APC run-to-run controllers and fault
`detection and classification ~FDC! systems used for high volume manufacturing applications in the
`semiconductor industry. All APC systems rely on data streams to make their process changes, to
`keep the process on target and in control, and to otherwise signal a need for engineering
`involvement to make similar corrective actions. The access to, communication of, and reliability and
`integrity of these data streams are essential to all APC programs. APC run-to-run controllers use the
`data to make changes in the process. FDC systems focus on predicting pending equipment- or
`process- related problems or detecting them quickly when they occur. The inability to access the
`needed data stream can prohibit the use of APC run-to-run controllers or FDC systems on critical
`process operations. Worse yet, the use of unreliable or corrupted data can cause undesirable
`consequences. In order to better capitalize on the improvements demonstrated with APC run-to-run
`controllers and FDC systems, end users have often had to create their own communication and data
`processing methods. The first decade of the 21st century will place increased demands on process
`and metrology equipment manufacturers, APC software and hardware suppliers, and APC
`programmers. Improvements in these areas through the use of industry standards and best known
`methods could greatly accelerate the APC field. Wafer-to-wafer and within wafer process control
`could be essential for 300 mm wafer processing; large flat panel processing will also need these
`improvements. We will discuss examples that Advanced Micro Devices experienced in Fast atom
`beam-25 within the past year. The case studies relate to complex, but necessary, methods to get the
`data we need for a FDC system and the role of metrology data on APC run-to-run controllers. Data
`and communication requirements for the next three to five years will also be discussed. The
`increased demands on the current process and metrology systems will increase as we begin to use
`new and alternative technologies to support more advanced APC strategies. © 2001 American
`@DOI: 10.1116/1.1380225#
`Vacuum Society.
`
`I. INTRODUCTION
`
`As the cost of developing new semiconductor devices and
`building new facilities, fast atom beams ~Fabs!, for their
`manufacture continues to increase, companies must continue
`to make every possible die yield. Making a die yield includes
`more than just making it electrically functional—it must be
`salable, preferably at the premium performance targeted so it
`provides the highest revenue.
`Time-to-market is increasingly critical for the profitability
`of a company. Gone are the days of leisurely designing the
`next device. Today, markets expect and demand the latest
`products and services in ‘‘Internet
`time.’’ The need to
`quickly ramp production is more critical to meeting the fast
`shifts and changes in today’s market. State-of-the-art Fabs
`now must face the daunting challenge of maintaining ever
`more stringent controls for hundreds of manufacturing opera-
`tions. Meanwhile, they must develop the means to rapidly
`and flexibly respond to the market demand and allow major
`changes in devices and processes to be developed, demon-
`
`a!Electronic mail: rick.markle@amd.com
`
`strated, and implemented in high volume manufacturing in a
`short period.
`Advanced process control ~APC! supplies the capabilities
`needed to run today’s best Fabs. APC provides run-to-run
`~RTR! and fault detection and classification ~FDC! as tools
`to control a Fab. APC RTR controllers change the metrology
`or process recipe used on the wafer to better control the
`outcome. FDC systems, on the other hand, do not attempt to
`change the recipe. Instead, these systems predict the need for
`human intervention, such as preventative maintenance, or to
`rapidly detect a catastrophic failure and automatically shut
`down the associated equipment to prevent further processing.
`Ready access to process and metrology equipment data
`~recipe parameters, tool-state parameters, and metrology re-
`sults! is essential for successful APC RTR and FDC pro-
`grams.
`As wafer size increases, wafer cost increases. Each indi-
`vidual wafer may need to be treated as though it was its own
`process lot. Larger wafers will put an increased demand on
`the Fab for better and faster RTR and FDC control systems.
`This article demonstrates the need for easier access to equip-
`
`1241
`
`J. Vac. Sci. Technol. A 19(cid:132)4(cid:133), Jul(cid:213)Aug 2001
`
`0734-2101(cid:213)2001(cid:213)19(cid:132)4(cid:133)(cid:213)1241(cid:213)7(cid:213)$18.00
`
`©2001 American Vacuum Society
`
`1241
`
`Applied Materials, Inc. Ex. 1015
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 2 of 8
`
`
`
`1242
`
`R. J. Markle and E. Coss, Jr.: Data requirements and communication issues for APC
`
`1242
`
`ment data and more reliable data in order to meet the needs
`of today’s state-of-the-art Fabs using APC.
`
`II. TYPES OF DATA REQUIRED
`A variety of data is required for APC RTR controllers and
`FDC systems. Data comes from both the process and metrol-
`ogy equipment and is either used as collected or processed
`using automated algorithms to calculate desired values. All
`data can be categorized into three main areas: wafer state,
`process state, or tool state.
`
`A. Wafer-state data
`
`Wafer-state information is acquired by measuring a vari-
`ety of properties and characteristics on the wafer itself. Fun-
`damentally, wafer-state measurements are the most accurate
`reflection of the probable performance of the chips that are
`on that wafer. Typical metrology operations provide data on
`several properties including:
`~1! film thickness and refractive properties;
`~2! CD width;
`~3! chemical stoichiometries;
`~4! visual defects; and
`~5! topography.
`
`B. Process-state data
`
`Process-state information is acquired by measuring prop-
`erties and characteristics of the environment in which the
`wafer is being processed. Process-state measurements are an
`indirect measurement of what the wafer-state may eventually
`be. The process-state can be a very useful tool to keep the
`process in control and on target since it can be measured
`during the actual process instead of after several process
`steps are completed.
`The need for in-line metrology measurements was driven
`by the need to recognize and correct process changes prior to
`finding them at sort and test. Similarly, the need to monitor
`critical process-states during the processing itself is driven
`by the need to better predict the subsequent wafer-state mea-
`surements. Typical measurements provide data on several
`properties including:
`~1! chemistries present during a given process step; and
`~2! absolute or relative changes in chemistries present dur-
`ing a given process step.
`
`C. Tool-state data
`
`Tool-state information is acquired by measuring the func-
`tional performance of the tools. These measurements provide
`only an indirect indication of what the wafer state may even-
`tually be. They can be, however, one of the most important
`methods to single out specific tools operating off target. Ac-
`cessing, monitoring, and controlling critical tool-state data is
`vital for process and metrology tools alike to avoid their use
`when conditions indicate misprocessing is likely. Some of
`the most common tool-state parameters measured and moni-
`tored include:
`
`J. Vac. Sci. Technol. A, Vol. 19, No. 4, Jul(cid:213)Aug 2001
`
`~1! temperature;
`~2! pressure;
`~3! fluid flow;
`~4! time;
`~5! rf power; and
`~6! component status ~open, closed, partially opened, on,
`off, etc.!.
`
`Tool-state parameters contained in recipes are the most
`common parameters modified by APC RTR controllers. In
`this respect, access to the tool state parameters is important,
`but more important is the ability to modify them remotely.
`Most RTR controllers used in high volume manufacturing
`perform recipe modifications between runs of lots containing
`multiple wafers. Some RTR controller, for example, epitax-
`ial growth or deposition, may perform recipe modifications
`on a wafer-by-wafer basis. With larger wafers, however, the
`process recipe may need to be modified in between the indi-
`vidual wafers within a lot or during the processing of a single
`wafer as the norm.
`
`III. DATA ACCESS
`
`Currently, access to data can be a complicated and time-
`consuming task for APC systems. One needs to identify the
`best means to access the data, determine the proper data
`collection rate, and then ascertain the data format and con-
`tent.
`The most common means of accessing equipment data
`and information in the semiconductor is through the ~SECSII/
`GEM! interface.1–4 A typical installation of a FDC system
`using the SECSII/GEM communication is presented in Sec.
`III A.
`Increasingly, companies are finding it necessary to access
`additional data not available through the SECSII/GEM interface
`or at sample rates often unachievable through the SECSII/GEM
`interface. A FDC using a custom data access system in de-
`scribed in Sec. III B.
`A. Case Study 1; Typical SECII(cid:213)GEM data access
`Figure 1 is a schematic of Case Study 1 that uses Triant’s
`MODELWARE system to collect and display individual tool
`parameters available through the SECS II interface. The data
`~DC! component
`collector
`from Triant has an active
`‘‘passthrough’’ feature that allows it to request data from the
`tool at the same time that another host is controlling the tool.
`This feature is attractive for factories that already use the
`SECS II port for recipe and tool control and want to add trace
`data collection without disturbing the existing interface. By
`using a network terminal server, the SECS II data can be made
`available to any server. An application server is configured
`to read the terminal server data and convert it back to serial.
`On the application server, the DC component receives and
`transmits the SECS II data. At this point, DC is only acting as
`a SECS II passthrough server.
`With a list of desirable tool variables to collect, DC is
`configured to insert into the SECS II stream requests for tool
`parameters. These requests are usually made as fast as pos-
`
`Applied Materials, Inc. Ex. 1015
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 3 of 8
`
`
`
`1243
`
`R. J. Markle and E. Coss, Jr.: Data requirements and communication issues for APC
`
`1243
`
`FIG. 1. Schematic of a typical system
`using SECS-based data communication
`and collection using Triant Technolo-
`gies, Inc. MODELWARE is shown.
`
`sible to collect trace data from the tool. The data is formatted
`for use with Triant’s MODELWARE system. By using serial
`network terminal servers and an active passthrough, collec-
`tion of any data available via the SECS II port is easily en-
`abled without disrupting production.
`
`B. Case Study 2; Custom data access
`
`Figure 2 is a schematic of Case Study 2. This system also
`uses Triant’s MODELWARE system to collect and display in-
`dividual tool parameters. However, a more complicated, cus-
`tom data access system had to be developed to supply the
`data.
`Serial base SECS II protocol is limited by the RS232 band-
`width. Typically a request for a parameter can be made in
`400–500 ms at 19.2 K baud. Most factories run SECS II at
`9600 baud. That reduces the sampling rate to about 800 ms.
`For faster data rates, a faster protocol and/or medium must
`be used. The high speed message system ~HSMS! standard,
`
`which is SECS II over a TCP I/P connection ~Internet!, can be
`used for much faster data rates. Unfortunately, HSMS is not
`readily available in established Fabs.
`Tools usually have serial or parallel ports available that
`can be used to send trace data. In this case, a serial port was
`allocated for each chamber of a tool. Again, network termi-
`nal servers were used to transport the serial data using the
`network. The tool owner specified a custom binary data for-
`mat that would contain all the available parameters of the
`tool. To ensure the fastest rate possible a minimal protocol
`was established to ensure data integrity. A custom program
`was developed to convert the high speed data into MODEL-
`WARE readable files. This allows us to re-use the entire MOD-
`ELWARE infrastructure and only design a different method of
`collecting the data. The custom program is linked to the SECS
`II host program so that it essentially becomes an extension of
`the SECS II interface.
`The experiences shown in Case Studies 1 and 2 illustrate
`that currently no single approach to accessing tool parameter
`
`FIG. 2. Schematic of a custom data
`collection system and Triant Tech-
`nologies, Inc. MODELWARE are shown.
`
`JVST A - Vacuum, Surfaces, and Films
`
`Applied Materials, Inc. Ex. 1015
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 4 of 8
`
`
`
`1244
`
`R. J. Markle and E. Coss, Jr.: Data requirements and communication issues for APC
`
`1244
`
`FIG. 3. Schematic of shallow trench isolation structure and film stack are
`shown.
`
`data meets all the needs of semiconductor end users. When
`all the data streams are readily available through existing,
`standard communications methods, simpler FDC systems
`can be deployed. When critical and targeted tool parameters
`are not available, more complex and sophisticated FDC sys-
`tems must be developed, often times resulting in a very cus-
`tom system for that particular application. The more complex
`and custom the FDC system is, the more costly it can be and
`the longer it may take to implement.
`
`IV. DATA RELIABILITY
`
`Loss of reliability in the data can result from noise. Noise
`can originate from variation in the process samples, the pro-
`cess and metrology equipment, and in the method of opera-
`tion of the process and metrology equipment. The manner in
`which an APC controller compensates for the effects of nor-
`mal variation in the process and samples is a critical and
`often a company confidential component of the APC control-
`ler. This article does not discuss the noise from normal varia-
`tion in the process but
`instead focuses on metrology
`equipment-related noise.
`The following case studies illustrate how erroneous data
`or the continued use of metrology systems that are not oper-
`ating properly can adversely affect an APC controller. The
`failure to ensure reliable data used in APC could result in an
`automatically ‘‘correctly misprocessing’’ product.
`
`A. Case study 3; Film thickness noise in shallow
`trench isolation etch APC controller
`
`Figure 3 shows a schematic of a typical shallow trench
`isolation ~STI! film stack, structure, and the associated mea-
`surements typically collected so a process engineer can peri-
`odically monitor the process. Essentially, the STI etch pro-
`cess must etch through a layer of silicon nitride (Si3N4), a
`layer of oxide (SiO2), and then down into the silicon ~Si!
`substrate. This trench is later filled with silicon oxide to pro-
`vide greater insulation than that possible by the substrate
`alone. This in turn allows the device designers to greatly
`increase the density of transistors on each chip.
`
`J. Vac. Sci. Technol. A, Vol. 19, No. 4, Jul(cid:213)Aug 2001
`
`FIG. 4. Illustration of a shallow trench isolation advanced process control
`run-to-run controller is shown.
`
`Typically an Etch engineer will manually monitor the STI
`etch process by taking a sample from the production line and
`performing a cross section scanning electron microscopy
`~SEM! to measure the trench depth. The process is also con-
`trolled by monitoring the trench depth in product wafers us-
`ing a profilometer and by monitoring the etch rate of the
`individual films, simplified film stacks, or product.
`Figure 4 shows the basic architecture of an APC control-
`ler that automates the ability to adjust the etch process more
`quickly and more frequently than the manual method. The
`controller responds to the required metrology data input and
`adjusts the etcher accordingly.
`Figure 5 represents the normalized film thickness at the
`typical nine sites measured after the barrier oxide ~SiO2,
`BOX! growth process for four lots. The lots are labeled in
`the order in which they were processed and subsequently
`measured at the BOX.
`Figure 6 represents the normalized film thickness at the
`typical nine sites measured after the silicon nitride ~Si3N4,
`SiN! film deposition for the same four lots. A single site on
`Lot B was measured ten times thicker than all other sites on
`Lot B and the other three lots. This thickness was impossible
`under the process conditions. This ‘‘bad’’ data point repre-
`
`FIG. 5. Normalized BOX film thickness for four lots in Case Study 1 is
`shown.
`
`Applied Materials, Inc. Ex. 1015
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 5 of 8
`
`
`
`1245
`
`R. J. Markle and E. Coss, Jr.: Data requirements and communication issues for APC
`
`1245
`
`FIG. 6. Normalized silicon nitride film thickness for four lots in Case Study
`1 including the single site erroneous measurement is shown.
`
`FIG. 8. Schematic of polysilicon gate structure and film stack is shown.
`
`sents the low probability that any metrology tool, at some
`time, can supply a bad data point. Once again the lots are
`depicted in the order they were processed at this operation.
`The same four lots were then sent on to etch. Lot B was
`etched prior to lots C and B. The SiN film thickness data was
`combined with the actual trench depth measurements to cal-
`culate the etch time required for lots C and D.
`Figure 7 shows the normalized etch times for the same
`four lots. Lot B was etched prior to lots C and D. The APC
`controller increased the etched time after lot B for two rea-
`sons. First, the etch time for lot B was slightly lower than the
`previous lot, A. Second, the single false reading after nitride
`deposition on Lot B used by the original APC RTR control-
`ler increased the etch time for subsequent lots well beyond
`the typical period.
`Fortunately, the unusually high trench etch time was de-
`tected and investigated. Lots C and D had extensive trench
`depth measurements to ensure they were within specification
`and only two wafers had to be scrapped. A third wafer, a
`single wafer split from lot A prior to etch for other reasons,
`was also etched using the trench etch time calculated after
`Lot B. It was also scrapped bringing the total loss to three
`wafers. Fortunately, only these three wafers were scrapped at
`this early process minimizing loss to essentially the cost of
`
`the starting wafers. The APC controller was subsequently
`modified to filter out such extraneous data. The key message
`here is that the metrology system must provide proper data
`and the APC controller must be able to recognize the inevi-
`table erroneous data point from an associated metrology sys-
`tem.
`Since no metrology system can be absolutely free of a
`random bad data, metrology system suppliers must work dili-
`gently to minimize the frequency of bad data. Users of me-
`trology, of course, must detect and recognize the random bad
`data and often times dismiss data collected inline as an
`anomaly or an outlier. APC controllers, likewise, must also
`be able to recognize and filter out the random bad data. FDC
`systems on the metrology equipment need to be developed to
`help the APC user more easily identify and avoid unreliable
`data. For metrology systems that use models, a goodness of
`fit should be provided with the data or results to help identify
`bad data.
`B. Case Study 4; Critical dimension (cid:132)CD(cid:133) SEM
`stigmation impact on poly CD
`Figure 8 shows a schematic of a typical polysilicon ~Poly!
`gate film stack and the associated measurements typically
`
`FIG. 7. Normalized shallow trench isolation etch time resulting from the
`controller for four lots in Case Study 1 is shown.
`
`FIG. 9. Illustration of a polysilicon gate CD advanced process control RTR
`controller is shown.
`
`JVST A - Vacuum, Surfaces, and Films
`
`Applied Materials, Inc. Ex. 1015
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 6 of 8
`
`
`
`1246
`
`R. J. Markle and E. Coss, Jr.: Data requirements and communication issues for APC
`
`1246
`
`FIG. 10. Normalized polysilicon gate
`CD for RO feature shows good con-
`trol.
`
`collected so a process engineer can periodically monitor the
`process. The width of the semiconductor line, the CD, has a
`direct relationship to the speed of the device. Typically the
`Poly CD is measured with CD SEM.
`Figure 9 shows the basic architecture of an APC control-
`ler that automates the feed back mechanism from the Poly
`CD SEM in-line measurement. Two different product de-
`vices had the same targeted CD. In-line CD SEM measure-
`ments are made on multiple features on the same die on the
`wafer. A scribe line monitor ~SLM! feature located outside
`of the actual device and a ring oscillator ~RO! feature within
`the device are used.
`Figure 10 shows the normalized RO CD data for a single
`device, B, and demonstrates relatively good in-line control.
`Figure 11 shows normalized CD data for two devices, A and
`B, with the same targeted CD. Figure 11 shows a significant
`shift in device B when a compilation of the two SEM fea-
`tures are graphed in contrast to the stability shown in Fig. 10.
`The comparison of the two SEM features is monitored pri-
`marily to watch for photolithography- ~photo-! related stig-
`mation issues. This shift, however, was not due to any
`
`changes or issues in photo on these devices. Instead, a subtle
`change in one of the operating parameters of the CD SEM,
`stigmation, caused the APC RTR controller to change the
`CD.
`Unfortunately, adjusting the stigmation of a CD SEM is
`largely dependent on the particular operator of the metrology
`tool and is very subjective. As in Case Study 3, the shift was
`detected, investigated, and corrected quickly.
`This case study emphasizes the need to carefully monitor
`the metrology systems themselves for changes, shifts, drifts,
`etc. that could adversely affect the data used in APC control-
`lers. Without careful monitoring of the metrology systems,
`the APC controller could automatically ‘‘correctly mispro-
`cess’’ the product.
`Here again, the metrology system supplier must be dili-
`gent in its efforts to minimize system drift and the potential
`adverse effects that can result from subjective adjustments.
`The metrology tool owners must be equally diligent in moni-
`toring the performance of their tools as they become an in-
`creasing important component of
`the processing itself
`through the use of APC controllers.
`
`FIG. 11. Normalized polysilicon gate
`CD comparing RO and SLM features
`shows shift resulting from metrology
`shift.
`
`J. Vac. Sci. Technol. A, Vol. 19, No. 4, Jul(cid:213)Aug 2001
`
`Applied Materials, Inc. Ex. 1015
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 7 of 8
`
`
`
`1247
`
`R. J. Markle and E. Coss, Jr.: Data requirements and communication issues for APC
`
`1247
`
`V. CONCLUSIONS
`Massive amounts of data are collected in the manufactur-
`ing of semiconductors for a variety of reasons. Ready access
`by factory host systems to process and metrology equipment
`data is essential for APC RTR and FDC systems. As semi-
`conductor manufacturers increasingly turn to APC to better
`control their Fabs in the first ten years of this new century,
`the demands for equipment data communication and data
`integrity will also increase. Larger wafer sizes ~300 mm and
`beyond! and flat panel screens will need to be processed
`essentially as individual lots ~or runs! that need remote modi-
`fication of process parameters either after each individual
`wafer or possibly during the process itself.
`Tool-level control is required to ensure stability of indi-
`vidual equipment but the integration of data from multiple
`equipment is needed to properly control the desired wafer-
`state outcome. The easier the access to equipment data, the
`more versatile and reliable the system is. Semiconductor
`manufacturers can and do create their own systems to
`supplement available data. Cost, however, in terms of time
`and resources is much higher than desired. Equipment sup-
`pliers will be expected to make equipment parameter data
`more readily available.
`Both process and metrology equipment suppliers will
`need to meet new requirements for data and communication
`in the next three to five years to meet the APC needs of
`semiconductor manufacturers. These requirements will in-
`clude capturing more process-and tool-state data as well as
`making them more readily available to the users external to
`the equipment. The most important step is to ensure that all
`critical process- and tool-state parameters are captured. After
`the data is being captured and monitored, they must be com-
`municated to the systems of the user. Users prefer that the
`data be more readily accessible through standard communi-
`cations.
`Additional demands on process and metrology equipment
`will be FDC systems that monitor critical tool parameters.
`
`These critical parameters will be monitored for sudden,
`abrupt, catastrophic changes and failures, as well as more
`subtle drifts and shifts. The FDC system will be used to
`catch failures and predict the need for adjustments, preven-
`tative maintenance, and calibrations to ensure that the data
`integrity is good, too. Means to flag errant data is also critical
`to avoid the improper use of bad data in APC RTR control-
`lers. For example, metrology equipment suppliers should
`provide some data quality monitor or goodness of fit metrics
`for the user to gauge the quality of the data or results re-
`ported.
`Strong participation by process and metrology equipment
`suppliers and users in industry standard efforts will help to
`meet the users requirements over the next three to five years.
`These standard efforts will cover more data capture and
`monitoring of process- and tool-state data, access to data by
`the user, communication methods for user access of the data,
`remote equipment recipe modifications, etc.
`
`ACKNOWLEDGMENTS
`
`The authors would like to thank Naomi Jenkins at Ad-
`vanced Micro Devices, Inc. for preparing the illustrations
`and Dr. W. Jarrett Campbell of KLA-Tencor Corporation,
`Control Solutions Division ~formerly of Advanced Micro
`Devices, Inc.! for his help with the data in Case Study 3.
`
`1SEMI E4, SEMI Equipment Communication Standard 1-Message Trans-
`port ~SECS-I!, Semiconductor Equipment and Materials International,
`Mountain View, CA.
`2SEMI E5, SEMI Equipment Communication Standard 1 - Message Con-
`~SEC-II!, Semiconductor Equipment and Materials International,
`tent
`Mountain View, CA.
`3SEMI E30, SEMI Generic Model for Communications and Control of
`Manufacturing Equipment ~GEM!, Semiconductor Equipment and Mate-
`rials International, Mountain View, CA
`4SEMI E37, SEMI High-Speed SECS Message Services ~HSMS!, Semicon-
`ductor Equipment and Materials International, Mountain View, CA.
`
`JVST A - Vacuum, Surfaces, and Films
`
`Applied Materials, Inc. Ex. 1015
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 8 of 8
`
`