`Trials@uspto.gov
`571-272-7822 Entered: December 20, 2022
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`__________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`_______
`
`APPLIED MATERIALS, INC.,
`Petitioner
`v.
`OCEAN SEMICONDUCTOR LLC,
`Patent Owner.
`__________
`
`IPR2021-01342 (Patent 6,968,248)
`IPR2021-01344 (Patent 6,907,305)
`__________
`
`Record of Oral Hearing
`Held: November 15, 2022
`__________
`
`Before MIRIAM L. QUINN, JOHN D. HAMANN, and DAVID D. COTTA,
`Administrative Patent Judges.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`APPEARANCES:
`
`ON BEHALF OF THE PETITIONER:
`
`ERIC A. KRAUSE, ESQ.
`of: Axinn, Veltrop & Harkrider LLP
`55 Second Street
`San Francisco, CA 94105
`(415) 490-1491
`ekrause@axinn.com
`
`ON BEHALF OF THE PATENT OWNER:
`
`ALAN WRIGHT, ESQ.
`Devlin Law Firm LLC
`1526 Gilpin Avenue
`Wilmington, DE 19806
`(302) 449-9010
`awright@devlinlawfirm.com
`
`The above-entitled matter came on for hearing Tuesday,
`
`November 15, 2022, commencing at 11:21 a.m. CST at the USPTO Texas
`Regional Office, 207 South Houston Street, Suite 159, Dallas, Texas.
`
`
`
`
`
`
`2
`
`
`
`1
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`P-R-O-C-E-E-D-I-N-G-S
`
`11:21 a.m.
`JUDGE QUINN: We are convening this afternoon for the session
`
`concerning IPRs 2021-1342 and IPR 2021-1344. We already did an
`introduction this morning and gave the instructions this morning. Those are
`still relevant for this session. So to the extent again, if you have objections,
`we don't do speaking objections, you reserve them for your argument time.
`
`And with that, Petitioner, you are first, and you have a total of 30
`minutes, and how much time do you want to reserve?
`
`MR. KRAUSE: I'd like to reserve 15 minutes, Your Honor.
`
`JUDGE QUINN: All right, you may proceed.
`MR. KRAUSE: May it please the honorable panel, I'm Eric Krause
`for petitioner of Applied Materials. You can jump to slide 3. On slide
`3 we have Claim 1 from each of the two patents at issue in this
`combined proceeding. On the left, Claim 1 of the 305 patent, and on
`the right, Claim 1 of the 248 patent, with the differences between the
`claims underlined in red. As patent owner's expert testified, the minor
`differences between the two claims do not impact the obviousness
`determination.
`We can jump back to slide 2. The two patents at issue in this
`proceeding are related to each other by a continuation application, and
`share the same specification dating to April of 2002. Now in 2002,
`the motivation in the semi-conductor industry was increasingly to
`integrate and automate. Houses were developing more advanced
`technologies to increase efficiency and to deal with increasing and
`
`
`
`
`
`
`3
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`growing complexity in products as the semiconductor industry
`progressed to smaller and smaller geometries.
`Fabrication facilities would use a manufacturing execution
`system, or MES, to control tools, but the MES has to be updated and
`personnel were used to provide those updates, introducing potential
`errors and delay into the system. And so as the overall drive in the
`industry was to increase profitability, that was driving an increased
`desire to automate and integrate.
`We jump to slide 8. These motivations would drive a POSA to
`the Shulze reference, assigned to petitioner-Applied Materials.
`Schulze discloses adding an automated monitoring and assessment
`system 107, circled here and color-filled in in blue, separate from the
`MES 102. Schulze discloses integrating that automated monitoring
`and assessment system be of the bus105, which also connects both of
`those components to the various equipment 115 or tools in the fab.
`Messages sent from the tools would be received by the
`monitoring and assessment system, which would maintain a state for
`the overall fab at all times. This would be a resource for the MES to
`query the monitoring and assessment system to see the state of the fab
`to help determine what to do next and how to control the tools. Some
`of the messages sent by the tools to the automated monitoring and
`assessment system would trigger certain events.
`For example, if a tool gave a message that it was going offline
`suddenly, then the status could be updated for that tool by the
`automated monitoring system, so know that that tool is now
`
`
`
`
`
`
`4
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`experiencing unscheduled down time. This would help the MES not
`send lots to that tool for processing because that tool is offline.
`JUDGE QUINN: And did you say that in Schulze the MES
`also controlled whether to send a new lot or not to a machine that has
`been detected as down?
`MR. KRAUSE: I think it would depend on the particular
`design of the fab and organization of the fab, but certainly the MES is,
`generally speaking, the MES is used to control the ongoings in the
`fab. And so the MES would be making a recommendation or a
`determination of what to do next.
`JUDGE QUINN: But Schulze does not control the conveyor
`system or anything like that. I didn't see any of that in Schulze.
`MR. KRAUSE: No, Schulze doesn't get as specific into
`specific machinery being controlled. In fact, Schulze is assessment
`and monitoring. So Schulze is receiving messages and status, and it's
`maintaining a status. So the Schulze system will track whether the
`lots have arrived at a tool and are being loaded, whether they're
`finished in the tool and need to be unloaded, not necessarily where
`they need to go next, not a decision, but more like a status, right?
`This lot is here. So Schulze would track that.
`JUDGE QUINN: Got it.
`MR. KRAUSE: And with that realtime tracking info, a POSA
`would be motivated to further use that information to make things like
`scheduling decisions or decisions about what to do next, and that
`would bring the POSA to Gupta, which discloses a realtime event-
`
`
`
`
`
`
`5
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`driven scheduler. A POSA would therefore be motivated to integrate
`the Gupta scheduler into the automated monitoring and assessment
`system of Schulze, shown here in 105.
`We could jump to slide 10. The patent owner raises several
`disputes as to whether the combination of Schulze in view of Gupta
`renders obvious the asserted claims, and they're summarized a bit
`here. But before we jump into those, I wanted to note what is not
`highlighted in the claim, the parts of the claim that are not really at
`dispute in the papers, and that is that there is an automatic detection of
`an occurrence of an event. There's automatically notifying that the
`occurrence has occurred, and that the art discloses reactively
`scheduling an action responsive to the detection of that event.
`I'd like to jump first into the preamble dispute. We could go to
`slide 11. The preamble regarding the automated manufacturing
`environment. And on slide 12, we know that a preamble is not
`limiting if it simply informs of what the intent of the claim is to be
`used for, intended use or purpose. And that is exactly what the
`preambles of these claims state here. It's a method that is intended for
`scheduling in an automated manufacturing environment.
`It's petitioner's position that the preambles are not limiting.
`However, as shown on slide 13, even if they are determined to be
`limiting, both Schulze and Gupta in isolation and as a combination,
`disclose an automated manufacturing environment. Schulze is
`directed to a semiconductor fabrication system and it discloses an
`automated monitoring and assessment system as we just discussed.
`
`
`
`
`
`
`6
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`And Gupta is directed to a scheduling system in a frontend
`manufacturing facilitator for integrated circuits. That frontend is
`disclosed by Gupta to be highly automated.
`JUDGE QUINN: Are you taking the position that for the words
`automated, manufacturing, that that is MES by itself?
`MR. KRAUSE: It's possible. I'm not sure that we've discussed
`what the minimum requirement would be for an automated
`manufacturing environment as intended by the 248 and 305 patents.
`I'm not sure that it's clear. It's possible that a MES might be sufficient
`to be an automated manufacturing environment. MES's were known,
`they did have some automatic features in that they would be able to
`control tools, and I think they could be used to some extent to make
`decisions and implement the next process step; that sort of thing,
`follow the process. So there is some level of automation there.
`There's certainly a lot of testimony in the record as to the
`varying degrees of automation in the industry, which is still true
`today. And so I'm not sure that it's necessarily clear what the
`minimum amount of automated manufacturing is required if the
`preamble were limiting, but it seems to Petitioner that under any
`reasonable construction, Schulze and Gupta as a combination
`disclosed that on it.
`We can return to slide 10. The next dispute is regarding the
`integrated automated process flow, and as I did before, I just want to
`take a moment and recognize that the limitation that we're talking
`about is detecting an occurrence of an event, where that event had to
`
`
`
`
`
`
`7
`
`
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`occur in a process flow. I should also note that the words integrated,
`automated, do not appear in the 305 patent claim. Let's jump to slide
`16.
`
`The patent owner's argument is that the automated integrated
`process flow requires an AMHS based on figure 2, which disclosed
`AMHS software components, 280 highlighted here. Now, figure 2 is
`just a portion of figure 1, which is itself an exemplary process flow.
`The 248 patent tells us at column 5, lines 3 through 8, that the process
`flow 100 shown in figure 1 and further discussed in figure 2, is a
`semiconductor fab process flow, but it explicitly states that the
`invention may be applied to other types of manufacturing processes.
`As the law tells us, it's improper to read limitations from an
`embodiment into a claim when it is apparent that the claims were not
`intended to be so limited, and the patent tells us that right here. In any
`event, as shown on slide 17, Schulze discloses software that
`automatically deals with materials handling. For example, on
`paragraph 214, shown here, the process state is monitored using a
`status variable identification parameter that monitors process state,
`including whether materials are being processed or loaded. In
`addition, Gupta discloses considering whether to proceed with a
`partial load or wait for more lots to arrive so that a full load can be
`processed. And that's disclosed in Gupta at column 7, lines 36
`through 44; and column 15, lines 49 through 54; and discussed in our
`petition on pages 29 to 30.
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`
`
`
`
`
`8
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`I'd like to jump back to slide 5. I'd like to discuss the last
`dispute raised concerning the software scheduling agent. Here in
`column 6, the 248 patent describes it as a software that reactively
`schedules on behalf of a manufacturing domain entity, and in
`prosecution the applicant made this event clearer. We have a lot cited
`here because this portion is discussed a lot in our briefing, but the
`words, the bullets, these are applicants, as disclosed on pages 165 to
`166 of the file history. Here they make clear that, as the phrase
`implies, it's got to be in software, it's got to represent a respective
`manufacturing domain, and it has to reactively schedule.
`We can jump to slide 22. This is the greater discussion. On the
`left side here we have the greater discussion the applicant raised
`during prosecution when it gave this list of characteristics of the
`software scheduling agent. Applicant was distinguishing the Kline
`reference, which is indicated in the first highlighted area, operates
`across the whole fab. Applicant strongly stated there is no support in
`the, for the proposition that a scheduling agent represent more than
`one manufacturing domain entity at any given time. Again, they
`stress too, no support that the agent could cover a whole subsystem.
`In the district court litigations that are going on in parallel, to
`which Petitioner is not a party, the district court evaluated the same
`argument patent owner is raising here and rejected it, stressing that the
`patentees made clear that the specification could not support any
`relationship that includes more than one manufacturing domain entity.
` Petitioner submits therefore that any reasonable construction of a
`
`
`
`
`
`
`9
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`software scheduling agent would not exclude a software scheduling
`agent that exhibited the three characteristics we saw earlier.
`And if we can go to slide 19. We see here in Gupta it checks all
`three boxes. Gupta is implemented in software. The list code for it is
`attached to the patent as Appendix A. Gupta's event driven,
`specifically disclosed that a determination of what comes next is made
`whenever certain events take place, and finally, the scheduling system
`calculates what each machine will do next. So it's limited to a specific
`machine in the process flow. You can jump to slide 25. Another
`argument raised by Patent Owner is that the software scheduling agent
`should be limited to a resource scheduling agent. As shown here on
`figure 3, there are highlighted four different examples of software
`scheduling agents. One of them, 320, is the resource scheduling
`agent, but the others are Machine, Lot, and PM or Predictive
`Maintenance scheduling agents.
`JUDGE QUINN: Let me ask a question about that. As I read
`the arguments by Patent Owner, I thought the argument was that at a
`minimum the scheduling agents must have all four of these particular
`agents identified in the specification, not that a software agent must be
`a resource agent.
`MR. KRAUSE: If that's true it boils down a little bit to the
`same thing because the art here, there's no dispute that Gupta discloses
`Machine, Lot, and PM scheduling. Patent Owner conceded that POR
`17, it's also the title of their slide 27, what their disputing then is the
`software scheduling agent needs to also include resource scheduling
`
`
`
`
`
`
`10
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`agent. But here, as shown in figure 3, they're all four of them separate
`boxes, separate scheduling agents. That's also how they're described
`in the spec as shown on the right. There may be different types of
`scheduling agents depending on the implementation.
`These are the different types of agents. I haven't seen any
`support in the 305 or 248 specifications for a software scheduling
`agent, a single agent providing all of these functions. What I've seen
`here is that there are four different examples of software scheduling
`agents, and one of them is a resource scheduling agent.
`Just as a final note, Petitioner submits I suppose, that any
`reasonable construction of software scheduling agent would not limit
`the agent to just being a resource scheduling agent, or having to
`include resource scheduling agent that would exclude the Lot
`scheduling agent 305, Machine scheduling agent 310, and the
`Predictive Maintenance scheduling agent 315. I'll reserve the balance
`of my time.
`JUDGE QUINN: Before you leave, is there a specific – was
`there any argument about the word integrated in the integrated process
`flow?
`
`MR. KRAUSE: I don't recall argument as in a dispute raised by
`Patent Owner or discussion of that term as it appears--
`JUDGE QUINN: Dispute in the record. Do you foresee that
`that is a term that is at issue?
`MR. KRAUSE: If we can go back to slide 8. I think not, Your
`Honor. The combination of Schulze and Gupta is integrated, both
`
`
`
`
`
`
`11
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`with the MES and with the tools as we proposed a POSA would
`combine them to result end. So this is all integrated using a bus 105,
`so to the extent there's a dispute over integration, I'm not sure that I
`recall any, but we believe that this combination discloses that
`integration to the extent it's required.
`JUDGE QUINN: Okay, thank you. Did you state your
`appearance for the record?
`MR. WRIGHT: I will, ma'am. Thank you, Your Honor. May it
`please the honorable board, my name is Alan Wright. I am with the
`Devlin Law Firm representing patent owner Ocean Semiconductor.
`And I'm going to see if I can get my slides are on here, but I'm not
`sure that I can because I've not used your system before.
`UNIDENTIFIED SPEAKER: Did you want them on the—
`MR. WRIGHT: No, if you could just put them up I can just
`scroll. Okay, great, thank you.
`JUDGE QUINN: How much time would you like to reserve?
`MR. WRIGHT: Can I reserve 10 minutes, please.
`JUDGE QUINN: All right, you may proceed.
`MR. WRIGHT: Thank you. So as Petitioner noted, there are a
`lot of similarities between the two patents at issue, and there are some
`fundamental issues that cross both IPRs, and I'll deal with them
`collectively, but if there's any specific questions you have regarding
`the IPRs, one patent versus the other, I'm happy to address that.
`JUDGE QUINN: Are you contending that the preamble is
`limiting?
`
`
`
`
`
`
`12
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`MR. WRIGHT: Yes, Your Honor, we are contending that the
`preamble is limiting. These claims could have set a method for
`scheduling compromising and then laid out the steps of the method,
`but they were specific to a method for reactive scheduling in an
`automated manufacturing environment. And you will see also that of
`the three steps in the method, the first step is also talking about
`automatically detecting an occurrence of a predetermined event in an
`integrated, automated process flow. So again, automated process flow
`within the context of the automated manufacturing environment.
`JUDGE QUINN: Well, I'm looking at your sur-reply and I
`want to know where exactly is your claim construction contention that
`the preamble is limiting?
`MR. WRIGHT: I'm not sure it was in the sur-reply.
`JUDGE QUINN: Was it in the patent owner response?
`MR. WRIGHT: I thought it was. I thought that there was a
`question there about the preamble being limiting.
`JUDGE QUINN: Well, I need to know with certainty, where is
`it in your response, or your sur-reply?
`MR. WRIGHT: Off the top of my head, I don't know where
`that is, Your Honor. I'd have to look at that.
`JUDGE QUINN: Okay, can your colleague find out for us?
`MR. WRIGHT: Sure. So we're looking at the term integrate
`automated process flow first. Under a Phillips claim construction
`analysis, claims are to be given their ordinary meeting as informed by
`the specification, not importing limitations as we've been accused of,
`
`
`
`
`
`
`13
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`but recognizing that the claim terms have meaning. That meaning is
`often illuminated in the specification and it cannot be ignored in a
`Phillips analysis.
`Petitioner wants to read certain limitations out of the claims and
`ignore what the specification states regarding the integrated process
`flow. So specifically, if you look at slide 3, you will see some
`noticeable things about the patent. First of all, the patent's title is an
`agent reactive scheduling in an automated manufacturing
`environment. The abstract echoes that and explains that a method and
`apparatus for scheduling and automated manufacturing environment
`comprising are disclosed. If you look at the specification and what it
`specifically directs you to, it talks about problems in the prior art
`dealing with MES systems, and it talks about problems related to the
`flow of materials within an automated system and the human factor
`that adds certain inefficiencies and problems to the overall efficient
`flow within that integrated process.
`Integrated is different than automated. Integrated is referring to
`the fact that you have a software scheduling agent that's actually
`integrated with a number of processes in this automated process flow.
` It's not limited to a single machine. It's a system that is integrated and
`is designed to take the human inefficiencies out, and to automate the
`process so that the materials flowing through this automated process
`flow do so efficiently.
`And so in figure 1, there is a general overview of the process
`flow, but that is not specific to any particular type of manufacturing.
`
`
`
`
`
`
`14
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`We're not pointing to figures 1 and 2 and saying that the claims are
`directed to automatic process flow in a semiconductor manufacturing
`facility, it's integrated automated process flow for any manufacturing,
`whatever the final product would be.
`The idea is, is that whatever the process flow is, whatever
`you're making, you're trying to take the human element out of it and
`you're trying to put a software scheduling agent or agents. There are
`different types that are used and recognized by one of ordinary skill in
`the art that apply to different types of tooling or actual material
`handling. But the idea is to take the human element out of it and to
`automate it, to integrate it, put these materials together and have the
`software service agent act as a very efficient decision maker in
`determining where that process flow needs to be corrected if there's an
`event that's detected that says that there's a problem.
`Take a look at slide 13 and you'll see that the cover of the
`patent, figure – excuse me, I'm sorry, the specification says although
`MES systems are sufficient for tracking watts in machines, such
`systems cover several deficiencies, the most obvious of which is their
`passive nature, and lack of any advanced scheduling and inability to
`support highly automated factory operations. So what you have with
`an MES system is you have, and this is essentially what Schulze
`discloses, is you have a system that tracks a tools status. How is this
`tool behaving?
`Ideally, from a profitability standpoint you would love to run
`that tool 24/7 and have nothing but great products come out of it.
`
`
`
`
`
`
`15
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`That doesn't happen in the real world. When mistakes happen or
`problems arise an MES system can tell you how the system is utilized.
` You can look at it, you can store it in a database. Process engineers
`can look at it, they can say we noticed every time we do a load or an
`unload there's an inefficiency involved; we're aware of that.
`But Schulze doesn't say if you run into a problem on this
`machine what do you do with that problem? Schulze does not make a
`realtime determination, as the claim software service agents do, to say
`we have a problem here, now we need to fix it. That's why you will
`note that in figure 2 of our patents, you will see that the MES system,
`the user interference software, and the computer itself, that essentially
`is what the Schulze system is. The 248 and the 305 patents have the
`additional software service agent, which I will discuss in a few
`minutes, and there are other components in here that are listed that can
`be part of this automated integrated system.
`There is an automated material handling system shown in this
`picture. That particular element is perhaps not required in every
`situation of an automated process flow, but it likely is because, again,
`you're trying to take the human element out, and therefore the
`automated material handling system would likely be part of any
`integrated automated system. It just isn't expressly listed in the
`claims.
`JUDGE QUINN: We're missing a lot of words here and I'm
`concerned that we're going into this automated process flow,
`integrated system environment, and there are very specific claim
`
`
`
`
`
`
`16
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`limitations here. There is the automated manufacturing environment
`in the preamble. We started talking about that, and now you're going
`into the integrated automated process flow. And then you're somehow
`saying that MES, by itself, could be an automated manufacturing
`system if it did all this other stuff.
`MR. WRIGHT: It could be part of an automated, integrated
`control system.
`JUDGE QUINN: Okay. And so what is the integrated
`automated process flow that is not part of the computing system?
`MR. WRIGHT: All right, so the computing system and the
`MES are part of the system that considers how this tool is being used.
` The software service scheduling agents will take that information,
`and based on that information they can, as the claims say, reactively
`schedule an action responsive to the detection of that predetermined
`event.
`The integration is the fact that you have more than one tool, if
`you will, in this manufacturing flow, and therefore you have to
`integrate these together. The idea is, I'll use the semiconductor
`manufacturing as an example, you may have etching, you may have a
`deposition, you may have the handling in between. You may have
`some other measure to make sure whether the etching and the
`deposition operated properly. And that has to be measured. All of
`those can be integrated together, and the software service agent can
`consider that information and make a decision based on it. Schulze
`does not have that.
`
`
`
`
`
`
`17
`
`
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`25
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`JUDGE QUINN: Well, the petitioner is not relying on Schulze
`alone.
`MR. WRIGHT: No, that's right.
`JUDGE QUINN: It's a combination.
`MR. WRIGHT: I can speak to Gupta. Gupta was earlier in
`time. Gupta deals with a scheduler that is used for a single tool, so
`Gupta has kind of a global scheduling plan that they talk about in
`Gupta. And the idea is you have this global scheduling plan for your
`entire facility and you really don't want to change that. That is a
`master schedule that is not done by a computer necessarily, it's just a
`master schedule for everything in the plant.
`You then have a scheduler that's used on a tool-by-tool basis,
`that looks to see whether or not that given tool is being utilized
`efficiently. And if it isn't, it can make predictions about how perhaps
`that tool can act in a certain time period to make that tool more
`efficient locally. If there's a problem with that tool, it does not send a
`notification to a software service agent or the master scheduler. It
`specifically says do not change the master schedule. You can do some
`localized efficiency scheduling to try to make sure that any particular
`lot that's in that machine –
`JUDGE QUINN: And you're referring to the optimization.
`MR. WRIGHT: The optimization, that's right. It's similar to
`Schulze. It's looking at tool optimization and efficiency, but what it
`isn't doing is, it is not taking an integrated series of tools in a process,
`and it's not reacting to a problem in that process flow other than
`
`
`
`
`
`
`18
`
`
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`recognizing that something has happened and we want to take a look
`at the prediction for that particular tool. But suppose you have a tool
`that does –
`JUDGE QUINN: I don't follow you.
`MR. WRIGHT: Suppose you have a tool that's etching.
`JUDGE QUINN: Why isn't it integrated?
`MR. WRIGHT: Because it's not talking to a software service
`agent whose job, what's that trying to do is determine, okay, we have a
`problem. How do we react to that problem?
`JUDGE QUINN: Who is not talking?
`MR. WRIGHT: The tool itself is not talking. The tool has its
`own scheduler software on the tool, and it's predicting, let's say a day
`in advance, in certain six-minute time periods. Here's what this tool
`intends to do over these next 24 hours. At hour 12 the tool breaks,
`something happens and it can't work anymore. It doesn't send a
`message and notify a software service agent to say, hey I'm down; you
`might want to take the lot I was working on and figure out what to do
`with this because we need to keep this moving. But the tool itself may
`try to finish that lot in that time frame, it may switch to a different lot,
`but it does not send it back to an integrated system.
`JUDGE QUINN: But I think you're arguing this separately
`from the combination, because it's not Gupta alone or Schulze alone.
`MR. WRIGHT: Right, Gupta –
`JUDGE QUINN: It's a combination of the two.
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`9
`
`10
`
`11
`
`12
`
`13
`
`14
`
`15
`
`16
`
`17
`
`18
`
`19
`
`20
`
`21
`
`22
`
`23
`
`24
`
`
`
`
`
`
`19
`
`
`
`
`IPR 2021-01342 (Patent 6,968,248)
`IPR 2021-01344 (Patent 6,907,305)
`
`
`
`MR. WRIGHT: And combined, you still don't have a situation
`where multiple tools are integrated together. The fact that any given
`tool may have automation, or the fact that any given tool may have
`some software on it, that's not a software service agent as that term is
`understood.
`JUDGE QUINN: That's why this is a combination case.
`MR. WRIGHT: Right, but neither one of them teach that.
`Right, you can't say take one that doesn't have a software service
`agent and take another one that doesn't have a software service agent
`and combine them, and now you've got a software service agent.
`Neither of them teach that. You can't say Schulze is looking at
`efficiency of tools and doesn't really even approach scheduling.
`Gupta looks at a scheduler for a single tool, but you can't put that
`together and necessarily get an integrated process flow with multiple
`machines and a software service agent that kind of lords over it and
`determines we have a problem on this machine, we need to fix that,
`where do we go?
`JUDGE QUINN: So I'm confused. Is this the motivation to
`combine argument that you're making right now?
`MR. WRIGHT: No. No. Motivation to combine is separate.
`Even if you combine, you still don't get this invention. You don't
`have the software service agent that is capable of watching a given
`domain entity, whether it's a lot or a series of wafers, or