(12) United States Patent
`Stoddard et al.
`
`USOO658774.4B1
`(10) Patent No.:
`US 6,587,744 B1
`(45) Date of Patent:
`Jul. 1, 2003
`
`2- Y----
`
`(54) RUN-TO-RUN CONTROLLER FOR USE IN
`MICROELECTRONIC FABRICATION
`
`(75) Inventors: Kevin D. Stoddard, Phoenix, AZ (US);
`ity E.siz SS).
`OnStanlinoS Saka IIS, Uhandler,
`(US)
`(73) ASSignee: Brooks Automation, Inc., Chelmsford,
`MA (US)
`Subject to any disclaimer, the term of this
`patent is extended or adjusted under 35
`U.S.C. 154(b) by 266 days.
`
`(*) Notice:
`
`(21) Appl. No.: 09/599,357
`(22) Filed:
`Jun. 20, 2000
`
`(56)
`
`Related U.S. Application Data
`(60) Provisional application No. 60/168.984, filed on Dec. 3,
`1999, and provisional application No. 60/140,434, filed on
`Jun. 22, 1999.
`(51) Int. Cl. ................................................ G06F 19/00
`(52) U.S. Cl. ............................. 700/121; 700/7, 700/17;
`700/45; 700/95; 700/29; 700/30; 700/31;
`438/5; 438/14; 438/17; 714/51
`(58) Field of Search .............................. 700,29, 30, 31,
`700/121, 95, 7, 17, 45; 438/5, 14, 17: 714/51
`s 1 as s --
`as
`s u is -u-
`s
`References Cited
`U.S. PATENT DOCUMENTS
`is
`4,700,311. A 10/1987 Tributsch et al. ...........
`4819,176 A 4/1989 Ahmed et al... 364468
`5,196,997 A 3/1993 Kurtzberg et al. .......... 364/152
`5,283,746 A 2/1994 Cummings et al. ......... 364/468
`5,347.460 A 9/1994 Gifford et al. .......
`... 364/468
`5,408.405 A 4/1995 Mozumder et al. ......... 364/151
`5.492.440 A 2/1996 Spaan et al. .................. 409/80
`5,495,417 A 2/1996 Fuduka et al. .............. 364/468
`5,526.293 A 6/1996 Mozumder et al. ......... 364/578
`5,546.312 A 8/1996 Mozumder et al. .... 364/468.03
`5,571,582 A 11/1996 Katoh ....................... 428/35.5
`5,654,895 A 8/1997 Bach et al. ................. 364/482
`
`
`
`56 Claims, 10 Drawing Sheets
`
`aO C al. ....
`
`- - -
`
`5,658.423 A 8/1997 Angell et al. .................. 438/9
`5,661,669 A 8/1997 Mozumder et al. ........... 702/84
`5,710,700 A * 1/1998 Kurtzberg et al. .....
`... 700/29
`5. A : SE Warsal - - - - - - - -
`- - - 7.
`5,926,690 A * 7/1999 Toprac et al. ................. 438/17
`5.993,043 A * 11/1999 Fujii .......................... 700/121
`6,110,214. A
`8/2000 Klimasauskas ................ 703/2
`6,148.239 A * 11/2000 Funk et al. ......
`... 700/1
`6,197,604 B1 -
`3/2001 Miller et al. .
`... 438/5
`6,263,255 B1
`7/2001 Tan et al. ................... 700/121
`6,424,876 B1
`7/2002 Cusson et al. ................ 700/51
`6,446,022 B1
`9/2002 Coss et al. .................. 700/109
`sk -
`cited by examiner
`Primary Examiner John Follansbee
`ASSistant Examiner Thomas Pham
`(74) Attorney, Agent, or Firm-Perman & Green, LLP
`57
`ABSTRACT
`(57)
`A automated run-to-run controller for controlling manufac
`turing processes comprises Set of processing tools, a set of
`metrology tools for taking metrology measurements from
`the processing tools, and a SuperVising Station for managing
`and controlling the processing tools. The Supervising station
`comprises an interface for receiving metrology data from the
`metrology tools and a number of variable parameter tables,
`one for each of the processing tools, collectively associated
`with a manufacturing process recipe. The Supervising Station
`also includes one or more internal models which relate
`received metrology data to one or more variables for a
`processing tool, and which can modify variables Stored in
`the variable parameter table to control the process tools
`using feedback and/or feed-forward control algorithms.
`Feed-forward control algorithms may, in certain
`embodiments, be used to adjust process targets for closed
`loop feedback control. The Supervising Station may have a
`user interface by which different feedback or feed-forward
`model formats (single or multi-variate) may be interactively
`Selected based upon experimental or predicted behavior of
`the System, and may also permit users to utilize their own
`models for run-to-run control.
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 1 of 23
`
`

`

`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 2 of 23
`
`

`

`U.S. Patent
`
`Jul. 1, 2003
`
`Sheet 2 of 10
`
`US 6,587,744 B1
`
`
`
`
`
`
`
`E |
`
`©NISSEOONG
`
`TOO |
`
`5) NISSIEOO?]d
`TIOOL
`
`5) NISSE OO??d
`
`
`TOOL
`
`5) NISSE OO?!d
`
`
`T?OOL
`
`Xº)OTONLEW
`,\0OTONLEW
`/\{0OTONLEW
`/\{0OTONLEW
`
`TOOL
`
`TOOL
`
`TOOL
`
`TOOL
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 3 of 23
`
`

`

`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 4 of 23
`
`

`

`U.S. Patent
`
`Jul. 1, 2003
`
`Sheet 4 of 10
`
`US 6,587,744 B1
`
`2OO
`
`15O
`
`OO
`
`OOO
`
`95O
`
`RUN
`FIG. 3
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 5 of 23
`
`

`

`U.S. Patent
`
`Jul. 1, 2003
`
`Sheet 5 of 10
`
`US 6,587,744 B1
`
`
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 6 of 23
`
`

`

`U.S. Patent
`
`Jul. 1, 2003
`
`Sheet 6 of 10
`
`US 6,587,744 B1
`
`35A
`PROC. L.
`TOOL 1
`
`35B
`PROC.
`TOOL 2
`
`36A
`
`4OA
`
`WORKPIECE
`
`36B
`
`WORKPIECE
`
`METROLOGY
`TOOL 1
`
`METROLOGY
`
`O
`o
`O
`
`
`
`B 40
`
`PROC.
`TOOL N
`
`WORKPIECE
`
`METROLOGY
`TOOL N
`
`35N
`
`56
`N
`
`4ON
`
`37A
`
`VPT
`(TOOL 1)
`
`VPT
`(TOOL 2)
`
`O
`O
`O
`
`37B
`
`37N
`
`VPT
`TOOL N)
`(
`
`ESW
`
`FSW
`
`-----
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 7 of 23
`
`

`

`U.S. Patent
`
`Jul. 1, 2003
`
`Sheet 7 of 10
`
`US 6,587,744 B1
`
`40
`
`METROLOGY
`TOOL
`
`1 OO
`
`
`
`95
`
`105
`
`METROLOGY
`ESW
`
`MES
`HOST
`
`APC/CIM
`FRAMEWORK
`
`MANUAL
`ENTRY
`
`(CORBA)
`
`(CORBA)
`
`- - - - - - -
`
`VPT
`
`37
`
`METROLOGY
`MAPS
`
`2O5
`
`CONTROL
`POINT
`DEFINITION
`
`207
`
`FEEDFORWARD
`CONTROL
`ALGORTHMS
`
`21 O
`
`TO
`F.G. 6B
`
`TO
`TO
`FIG. 6B
`F.G. 6B
`F.G. 6A
`
`TO
`F.G. 6B
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 8 of 23
`
`

`

`FROM
`F.G. 6A
`
`FROM
`FIG. 6A
`
`22O
`
`
`
`MODEL/
`FEEDBACK
`CONTROL -- CONTROLLER
`ALGORITHMS
`PLUGINT
`
`PROCESS
`TARGETS
`(
`
`
`
`
`
`
`
`|| PROCESS
`TARGET
`ADJUSTMENT
`
`?
`
`U.S. Patent
`
`Jul. 1, 2003
`
`Sheet 8 of 10
`
`US 6,587,744 B1
`
`FROM
`FIG. 6A
`
`215
`VPT
`- - - - - - VARIABLES
`(erpass FEEDBACKY ADJUSTMENT
`CONTROL
`
`
`
`21
`2
`
`225
`
`DOWNLOAD
`
`as
`
`EW-C - - - - - - - - - - - - -
`
`35
`
`PROCESS
`TOOL
`
`36
`
`F.G. 6B
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 9 of 23
`
`

`

`U.S. Patent
`
`Jul. 1, 2003
`
`Sheet 9 of 10
`
`US 6,587,744 B1
`
`|| ||
`
`
`
`N
`
`
`
`
`
`
`
`
`
`
`
`
`
`s
`|| ||
`|| ||
`
`e
`
`
`
`
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 10 of 23
`
`

`

`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 11 of 23
`
`

`

`US 6,587,744 B1
`
`1
`RUN-TO-RUN CONTROLLER FOR USE IN
`MICROELECTRONIC FABRICATION
`
`CROSS-REFERENCE TO RELATED
`APPLICATIONS
`This application is a Provisional application of U.S.
`Provisional Application Ser. No. 60/140,434, filed on Jun.
`22, 1999, and of U.S. Provisional Application Ser. No.
`60/168,984, filed on Dec. 3, 1999, both of which are hereby
`incorporated by reference as if set forth fully herein.
`
`2
`chart per day, it is estimated that there would be on average
`82 false alarms per day. Due to the sheer magnitude of faults
`that are reported in Such circumstances, only the processes
`with the most significant excursions tend to be maintained.
`In Some cases, however, the opposite is true, and too much
`attention is given to a chart, leading to over-adjustment of
`data points which in turn results in processes “ringing.”
`Additional proceSS Variation can be introduced between
`shifts or individuals as they all try to compensate for each
`other's process adjustments, compounding the problem.
`The present inventors have recognized the foregoing
`problems and have developed an advanced run-to-run con
`troller Suitable for use in a microelectronic fabrication
`facility.
`
`SUMMARY OF THE INVENTION
`The invention provides in one aspect an advanced run
`to-run controller for use in microelectronic fabrication.
`In one embodiment, an advanced run-to-run controller for
`controlling manufacturing processes comprises Set of pro
`cessing tools, a Set of metrology tools for taking metrology
`measurements from the processing tools, and a Supervising
`Station for managing and controlling the processing tools.
`The Supervising Station comprises an interface for receiving
`metrology data from the metrology tools and a number of
`variable parameter tables, one for each of the processing
`tools, collectively associated with a manufacturing process
`recipe. The Supervising Station also includes one or more
`internal models which relate received metrology data to one
`or more variables for a processing tool, and which can
`modify variables stored in the variable parameter table to
`control the process tools using feedback and/or feed-forward
`control algorithms. Feed-forward control algorithms may, in
`certain embodiments, be used to adjust proceSS targets for
`closed loop feedback control.
`In a preferred embodiment, the Supervising Station has a
`user interface by which different feedback or feed-forward
`model formats (single or multi-variate) may be interactively
`Selected based upon experimental or predicted behavior of
`the System. The Supervising Station may also permit users to
`utilize their own models for run-to-run control.
`Further variations, modifications and alternative embodi
`ments are also described herein.
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`FIG. 1A is a schematic block diagram of a hardware
`platform on which the run-to-run controller of the present
`invention may be executed.
`FIG. 1B is a schematic block diagram of a further
`hardware platform on which the run-to-run controller of the
`present invention may be executed.
`FIG. 2 is a diagram illustrating one embodiment of the
`Software components that may be used to execute the
`run-to-run controller of the present invention.
`FIGS. 3 and 4 are graphs illustrating the result of an
`implementation of the run-to-run controller of the present
`invention as applied to an exemplary wet oxide deposition
`proceSS.
`FIG. 5 is a block diagram illustrating the update of
`variable parameter tables (VPTs) associated with the various
`processing tools, for use in run-to-run control processes.
`FIGS. 6A and 6B are block diagrams illustrating a run
`to-run control process flow in accordance with a preferred
`embodiment as described herein.
`FIG. 7 is a diagram showing the contents of a preferred
`variable parameter table (VPT).
`
`15
`
`BACKGROUND OF THE INVENTION
`1) Field of the Invention
`The field of the present invention pertains to microelec
`tronic circuit fabrication and, more particularly, to methods
`and apparatus for controlling microelectronic circuit fabri
`cation processes.
`2) Background
`The quality of microelectronic circuits and/or
`components, Such as those manufactured from a Semicon
`ductor wafer, is directly dependent on the consistency of the
`processes used in its fabrication. More particularly, produc
`tion of Such circuits and/or components requires reproduc
`25
`ible etching, deposition, diffusion, and cleaning processes. A
`failure to maintain control of the processes within defined
`manufacturing tolerances results in decreased yield and
`decreased profitability for a fabrication facility.
`In a typical Scenario, the manufacturing process exhibits
`slow drifts that change the batch-to-batch properties of the
`product. Very often, these effects are due to slight variations
`in the operation of one or more processing tools over the
`time in which the different batches are processed.
`Additionally, in large Scale operations, the same processing
`operation may be executed on a plurality of processing tools
`of the same type to process parallel batches of the product.
`The same processing recipe is generally used to concurrently
`control the operations of the plurality of Similar processing
`tools. However, minor variations in the way in which an
`individual tool responds to the recipe parameters to execute
`the proceSS can drastically affect the resulting product per
`formance when compared with products processed on other
`ones of the Similar processing tools.
`Traditionally, this problem has been handled manually by
`a human operator, using Statistical process control (SPC)
`concepts. More particularly, a human operator monitors the
`product output as the result of the execution of a proceSS
`recipe on a particular tool and tweaks the recipe for Subse
`quent product runs. In many instances, however, the proceSS
`recipes can number in the hundreds. AS Such, monitoring
`and manually adjusting these recipes for process drift can be
`very time consuming, error prone and lacking in accuracy.
`A common methodology for monitoring batch processes
`utilizes X-bar/S or X-bar/r plots in commercial or internally
`developed SPC software packages. Normally, distributed
`proceSS data is typically monitored automatically utilizing a
`set of rules (such as Western Electric) to determine if the
`proceSS is “in-control.” Manual investigation and adjustment
`of the process is necessary once a data point is determined
`to be out of control. A large percentage of these adjustments
`are made to compensate for the run-to-run variations attrib
`uted to process equipment drift. Unfortunately, there are
`many problems using manually adjusted processes based on
`SPC charts. A typical wafer fabrication plant may have about
`2,500 on-line SPC charts. If all of the Western Electric rules
`were used, and if just two new points were added to each
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 12 of 23
`
`

`

`US 6,587,744 B1
`
`15
`
`35
`
`40
`
`25
`
`3
`FIG. 8 is a process flow diagram illustrating the run-to-run
`control process of FIG. 6 from an alternative perspective.
`DETAILED DESCRIPTION OF PREFERRED
`EMBODIMENTS
`An advanced run-to-run controller (ARRC) system is set
`forth herein that provides a formal methodology recipe
`adjustment to compensate for gradual process drift and/or
`upstream process variation. This functionality assists in
`Significantly reducing the engineering time required for
`proceSS maintenance and adjustment.
`One embodiment of a processing platform architecture
`that may be used to implement the disclosed ARRC system
`is set forth in FIG. 1A. In the illustrated embodiment, the
`platform architecture, shown generally at 20, is comprised of
`a Fabrication Supervisor Workstation (FSW) 25, one or
`more Equipment Supervisor Workstations (ESW)30, one or
`more processing tools 35, and one or more metrology tools
`40. The metrology tools 40 may either be in-situ or ex-situ
`in nature.
`The FSW 25 monitors and controls the overall operation
`of the microelectronic fabrication facility. One or more
`operators may monitor the operation of all or a Substantial
`portion of the tools used throughout the fabrication facility.
`Based on the operations monitored at the FSW 25, the
`operator may control the tool Sets and, further, direct the
`processing recipes that are to be used by one or more tool
`Sets in the fabrication of the product.
`An equipment tool set is shown generally at 45 of FIG.
`1A. The equipment tool Set 45 includes one or more pro
`cessing tools 35 that are connected for bilateral communi
`cation with a common ESW 30. Processing tools 35 are
`generally of the same type. For example, all of the proceSS
`ing tools 35 may be furnaces. However, it will be recognized
`that processing tools 35 may include different tool types
`which be grouped based upon the type of processes that are
`to be executed upon the workpieces to fabricate the end
`products.
`The ESW 3.0 is preferably configured to accept processing
`recipes from the FSW 25 and to direct processing tools 35
`in the execution of the processing recipes. In those instances
`in which processing tools 35 are of the same tool type, the
`FSW 25 may provide a single processing recipe to the ESW
`30, which recipe is to be concurrently used by all of the
`45
`processing tools 35 for parallel batch processing.
`Alternatively, the ESW 3.0 may receive different recipes for
`one or more of the processing tools 35, in which case
`processing tools 35 may be tools of the same type or of
`different types. Processing tools 35 are subject to inter-tool
`deviations of the execution of a Single recipe on more than
`one tool as well as run-to-run, intra-tool deviations in the
`execution of a single recipe over time. Accordingly, the
`ESW 30 includes a Variable Parameter Table (VPT) asso
`ciated with each of the processing tools 35, as illustrated, for
`example, in FIG. 5. The VPT37 includes the parameters that
`are used in the execution of a processing recipe by a given
`processing tool. The parameters of the VPT37 are based on
`the particular characteristics of the associated tool 35 and, as
`such, will frequently differ between the tools of the same
`type. In general, each fabrication proceSS is comprised of
`one or more recipes (one recipe per process step), and each
`recipe will involve one or more processing tools 35, each
`processing tool 35 having a VPT 37 for all of the process
`recipes.
`The parameters of the VPT37 are calculated and updated
`based on metrology data for the particular process imple
`
`50
`
`55
`
`60
`
`65
`
`4
`mented by the associated processing tool. To this end, FIG.
`1A illustrates the use of one or more metrology measure
`ment units 40 that receive workpieces from one or more
`respective tools 35 and measure the physical characteristics
`of the workpieces processed by the processing tools 35. In
`the illustrated embodiment, a plurality of metrology mea
`surement units 40a-40d are employed, each metrology
`measurement unit being respectively associated with one of
`the processing tools. However, it will be recognized that
`Such a one-to-one correspondence is not absolutely neces
`Sary. Rather, depending on the particular processing tools
`utilized, one or more of the processing tools 35 may use a
`Single metrology measurement unit in order to economize on
`capital costs and Space.
`In operation, as illustrated in FIG. 5, microelectronic
`Workpieces 36 are transported from each processing tool 35
`to a corresponding metrology measurement unit 40. This
`transportation, illustrated at lines 50, may include an auto
`matic or manual transportation of the workpieces 36. Each
`metrology measurement unit 40 is designed to measure one
`or more physical and/or electrical characteristics of the
`Workpiece 36 processed by the associated processing tool
`35. The measurement data is then made available to the
`ESW 30 along, for example, a communication bus 55 or the
`like. Once the measurement data has been provided by the
`metrology measurement unit 40, the ESW30 may update the
`VPT 37 for the particular tool 35 that processed the work
`pieces 36 measured by the metrology measurement unit 40.
`FIG. 1B illustrates a further system architecture that may
`be used in those instances in which the at least two ESWs
`30a and 30b are employed. The first ESW30a is preferably
`asSociated with one or more processing tools 35 while a
`second ESW 30b is preferably used to control one or more
`metrology measurement units 40. The second ESW 30b
`communicates the metrology data to the first ESW30a along
`a communication bus 60 or the like. The metrology data
`received by ESW 3.0a is then used to calculate and/or adjust
`the parameters of the VPTS 37 associated with the process
`ing tools 35a–35d.
`The Advanced Run-to-Run Controller (ARRC) system
`preferably provides a formal methodology for recipe adjust
`ment to compensate for gradual process drift (feedback
`control) and upstream process variation (feed-forward
`control). Additionally, other control modes of operation may
`be employed Such as, for example, combined feedback/feed
`forward control and adjustable feedback control, as illus
`trated in FIG. 6, described later herein. In each instance, the
`functionality significantly reduces engineering time required
`for proceSS maintenance and adjustment.
`In the case of feedback control, the ARRC system pref
`erably provides automatic adjustment of the recipe through
`the parameters all of the VPT (e.g., VPT37 shown in FIG.
`5) based on the measured process results from the current
`process. This automatic adjustment is accomplished in part
`by modeling the process using measurements from past
`experience, first principles, or a design of experiment. With
`Such a model, a controller can make intelligent decisions as
`to what variables to change in the recipe through the VPT to
`maintain the desired process target.
`In the case of feed-forward control, the ARRC system can
`use metrology measurements from a previous process Step to
`make adjustments to either the proceSS target or a process
`variable (recipe parameter) to correct for problems upstream
`in the processing Sequence. This is accomplished by empiri
`cally modeling the relationship between either two process
`measurements or a proceSS measurement from a past process
`and a recipe parameter in the current process.
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 13 of 23
`
`

`

`S
`FIG. 2 illustrates an exemplary software framework for
`the exchange of metrology information and calculation/
`updating of the parameters of the VPT tables37 (see FIG. 5).
`The ARRC system illustrated in FIG. 2 is preferably
`designed to take full advantage of a communication frame
`work using a standardized interface, such as CORBA. This
`framework allows for the exchange of metrology informa
`tion between ESWS with or without an FSW Since the
`metrology requests preferably comprise CORBA objects.
`All metrology information may then be Stored locally on the
`ESW associated with the processing tool and can be
`accessed via the communication framework by any other
`ESW in the case of feed-forward control. In Such instances,
`metrology is not stored on the FSW, thus avoiding a single
`point of failure.
`Further details will now be set forth concerning metrology
`acquisition, Storage and maintenance, after which will be
`described further details concerning run-to-run control using
`the metrology data.
`The acquisition of metrology data (i.e., process
`measurements) can be difficult for a variety of reasons. For
`example, the particular metrology measurement unit may
`not have external communication. In Such instances, manual
`entry of the proceSS results must be undertaken. Manual
`entry, however, is tedious, highly Susceptible to data entry
`errors, and not the preferred manner of obtaining metrology
`information in the System. Another obstacle with imple
`menting Such a System is the acquisition of the proceSS
`measurements in a timely manner. It may be anywhere from
`an hour to several days before the results from the latest run
`are obtained or are otherwise made available to the ESW.
`Since each fabrication facility Stores and analyzes the pro
`cess results in a different way, it can be quite challenging to
`provide a Standard interface to acquire this information
`without writing Special code for each customer. The Soft
`ware architecture illustrated in FIG. 2 is designed to over
`come or mitigate the above obstacles. A description of each
`of the functional modules, as they relate to metrology
`acquisition, Storage, and maintenance, follows below.
`With reference to FIG. 2, a Metrology Broker 70 is used
`to manage all of the metrology acquisition requests provided
`by an ARRC Controller 75. Each metrology acquisition
`request from the ARRC Controller 75 is associated with a
`metrology map defining the method of acquiring the metrol
`ogy results. Once the metrology information is acquired, it
`is stored in a Metrology Database 85 along with the Date,
`Time, Tool, MiniSpec, Lot ID and Run Number. The
`requesting ARRC Controller 75 will also be notified of the
`acquisition of the metrology results when it occurs.
`The metrology map is the vehicle that allows the user to
`define the method of acquiring the process measurements as
`well as the format in which they are presented. The user can
`define the number of wafers and Sites (process measurement
`locations) and define more specific names for the metrology
`points.
`Various automated methods of obtaining process mea
`Surement results may be used and defined in the metrology
`map. With reference to FIG. 2, the automated methods
`include, for example, the following:
`GEM Interface 90-A standardized GEM interface may
`be provided to transfer the proceSS measurements into a high
`performance database and to provide the measurements to
`the ARRC system. This method generally requires that the
`users at the fabrication facility write custom code to inter
`face their process measurement database to a Supervisory
`workstation, Such as an ESW.
`
`15
`
`25
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`US 6,587,744 B1
`
`6
`CIM Framework (CORBA Interface) 95 This interface
`is provided for customers with proceSS measurement tools or
`a SPC database that is compliant with Sematech's APC
`Framework. For the CIM Framework 95, the ESW 30
`subscribes to a CORBA object to automatically obtain the
`process measurements as they are measured.
`ESW Metrology Tool Interface 100 This interface is
`provided for users at fabrication facilities who wish to
`connect an ESW directly to the metrology tool for the
`purpose of run-to-run control.
`SMC Result- This interface allows linkage to a Statisti
`cal Machine Control application which provides automated
`fault detection using the equipment real-time measurements.
`Calculations from this application may generally be pro
`vided to the ARRC system as measurements from real-time
`SCSOS.
`In addition to the foregoing automatic metrological acqui
`Sition methods, the ARRC System may also employ a
`Manual Metrology Entry interface 105. Here, a user inter
`face is provided to manually enter the proceSS measurements
`of the metrology tool for fabrication facilities that do not
`have a centralized SPC database or that have process mea
`Surement tools without external communication capabilities.
`Although not generally the most efficient manner of acquir
`ing metrology data, this functionality is especially useful for
`users who wish to validate the functionality of the ARRC
`System without committing resources to code a GEM or
`CORBA compliant interface to establish a link with the
`metrology tools or SPC database.
`Preferably, the Manual Metrology Entry interface 105
`allows the user to Select an open metrology request and then
`enter new data, preferably in a Unix(R) environment. The user
`interface window preferably contains a scrollable list of
`open metrology requests and displays the following infor
`mation about each request: date, time, tool, lot ID, and recipe
`name. Because there may be a multitude of open metrology
`requests, the following three fields may be used to narrow
`the Search: tool, recipe name and lot ID. Any name or
`portion of a name followed by an asterisk can be entered to
`automatically filter the available selections. After all metrol
`ogy values have been entered for a particular request, the
`ARRC Controller 75 will be notified of the acquisition of the
`data and can process the data accordingly.
`Metrology Database 85 is preferably in the form of a
`proprietary, flat file database, although other database Struc
`tures (e.g., relational databases) may be used instead. The
`Metrology Database 85 is used to store and maintain the
`process measurement results. These values are generally not
`stored in the standard database of the ESW 30 since often it
`is necessary to Store measurement information on the order
`of ten times longer than the real-time proceSS data Stored in
`the standard ESW database. An abundance of process mea
`surements should preferably be made available for statistical
`analysis of the process and for allowing robust modeling of
`the process. Separate database clean up and maintenance
`operations may also be provided for the Metrology Database
`85.
`Each ESW 3.0 may comprise an equipment Supervisory
`workstation of the type such as available from SEMY
`Engineering, Inc., of Phoenix, Ariz. Integrating run-to-run
`functionality in Such a WorkStation may require the addition
`of several hooks to the existing ESW Software. Implemen
`tation is preferably achieved by associating run-to-run con
`trol methods 110 to values defined in the VPT. This asso
`ciation process allows the user to define adjustable variables
`in the recipe for each individual tool.
`
`Applied Materials, Inc. Ex. 1008
`Applied v. Ocean, IPR Patent No. 6,836,691
`Page 14 of 23
`
`

`

`25
`
`7
`Interaction between the processing tools 35 and ESW is a
`controlled by a tool server 115. The tool server 115, in turn,
`interacts with the ARRC Controller 75.
`The ESW 30 may provide visual access to the VPT,
`including a display of parameters, visible minimum and
`maximum limits, maximum change per Step, access levels
`per parameter and parameter type designations with units
`(see FIG. 7). The ESW 30 thus augments the VPT with an
`interface that allows the user to select the method by which
`each variable parameter is calculated/updated. AS Such, the
`user may define a custom tailored run-to-run control algo
`rithm for each parameter if so desired. Each VPT parameter
`is designated with one of the following ARRC adjustment
`methods: feedback only, adjustable feedback, combined
`feedback/feed-forward, feed-forward only, or none. In a
`preferred ARRC system, there are four possible adjustment
`modes available for each adjustment method. They are:
`Automatic-this mode will automatically make changes
`to the variable parameter based on the recommendations of
`the model and controller.
`Manual Verification-this mode will ask the operator to
`approve of the recommended variable parameter changes.
`Manual-this mode will predict the process results from
`the process model and controller without making adjust
`ments to the variable parameter. This mode may be useful to
`test the validity of a model without effecting the process.
`This mode will also allow the operator to approve and make
`the recommended variable parameter changes manually. A
`purpose of this mode of operation is to allow the operator to
`become comfortable with the adjustments made by the
`proceSS model and controller before making the adjustments
`automatically.
`Data Collection-this mode is used to acquire only the
`proceSS measurement data which is required for a feed
`forward process where the upstream process tool is con
`nected to an ESW and does not have a feedback controller
`defined.
`The ARRC Controller 75 interacts with the tool server
`115, the ARRC methods 110, the Metrology Database 85,
`and the Metrology Broker 70. It is a background software
`40
`process that performs off-line processing of the ARRC
`methods 110 after or concurrent with the receipt by the
`Metrology Database 85 of the process measurement results
`by way of the Metrology Broker 70. Off-line processing
`assists in avoiding delay which might be attributable to the
`ARRC calculations or acquisition of proceSS measurements
`prior to downloading the adjusted recipe.
`Further description of the run-to-run control methods will
`now be provided, with occasional reference to the block
`diagrams of FIGS. 6A and 6B. As previously indicated, a
`number of means are available for the acquisition of metrol
`ogy data and the generation of metrology maps. By way of
`example, automated methods of obtaining proceSS measure
`ment results include use of a standardized GEM Interface
`90, a CIM Framework (CORBA) Interface 95, or an ESW
`55
`Metrology Tool Interface 100. Also, the ARRC system may
`also employ a Manual Metrology Entry interface 105. The
`information from these metrology acquisition techniques is
`Supplied to a Set of metrology maps 205, which, as noted, are
`the vehicles which allow the user to define the method of
`acquiring the process measurements, as well as the format in
`which they are presented. Using the metrology maps 205,
`the user can, for example, define the number of waferS and
`Sites (process measurement locations) and define more spe
`cific names for the metrology points.
`Prior to running a process, whether feedback or feed
`forward (or both), the user sets control points for the various
`
`45
`
`50
`
`60
`
`65
`
`US 6,587,744 B1
`
`15
`
`35
`
`8
`process parameters used by the ARRC syst