`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`
`
`
`NVIDIA CORPORATION,
`Petitioner,
`
`v.
`
`INVENSAS CORPORATION,
`Patent Owner.
`
`IPR2020-006021
`U.S. Patent No. 6,849,946
`
`
`
`PATENT OWNER’S SUR-REPLY
`
`
`1 Semiconductor Components Industries, LLC d/b/a ON Semiconductor, the
`petitioner in IPR2020-01735, is joined as a party to this proceeding.
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`
`I.
`II.
`
`III.
`
`B.
`
`C.
`
`B.
`
`TABLE OF CONTENTS
`Introduction ..................................................................................................... 1
`The Relevant Configuration, Relevant Solution, and Intermediate
`Width Limitation Are Not Disclosed in Any of the References. .................... 2
`Petitioner Advances Several Erroneous Arguments in an Attempt to
`Avoid or Reduce Its Burden of Proof. ............................................................ 4
`A.
`Petitioner’s Claim Construction Abrogates the Clear Meaning
`of the Claims. ....................................................................................... 4
`Petitioner Uses Abstraction to Mask Material Differences
`Between the Challenged Claims and the Prior Art. ............................. 8
`Petitioner Seeks to Establish Limitations Missing from the Prior
`Art by Arguing That They “Could” Be Present. .................................. 9
`IV. Petitioner Improperly Mischaracterizes and Recasts the ’946 Patent. ......... 14
`V.
`Petitioner Improperly Mischaracterizes and Recasts the Prior Art. ............. 17
`A.
`The Prior Art Does Not Instruct a POSA to Insert Dummy
`Metal in “Every” Open Area. ............................................................. 17
`The Prior Art Does Not Disclose or Suggest the Relevant
`Configuration, Relevant Solution, or Intermediate Width
`Limitation. .......................................................................................... 18
`1.
`Jaso ........................................................................................... 18
`2.
`Stine1998 ................................................................................. 21
`VI. A POSA Would Not Combine Jaso with Stine1998 as Proposed by the
`Petition. ......................................................................................................... 34
`VII. Pramanik1998 Does Not Support Obviousness. .......................................... 37
`VIII. The Intermediate Width Limitation Is Nonobvious. .................................... 39
`IX. Conclusion .................................................................................................... 40
`
`
`
`
`
`i
`
`
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`UPDATED LIST OF EXHIBITS
`
`
`Exhibit No. Description
`2001
`NVIDIA Answer to Complaint 19-cv-861 (July 1, 2019)
`
`2002
`
`2003
`
`2004
`
`2005
`
`2006
`
`2007
`
`2008
`
`2009
`
`2010
`
`2011
`
`2012
`
`2013
`
`2014
`
`2015
`
`2016
`
`2017
`
`Invensas Complaint 19-cv-861 (May 8, 2019)
`
`August 23, 2019 Scheduling Conference Transcript 19-cv-861
`(Aug. 13, 2019)
`Maly et al., Memory Chip for 24-Port Global Register File (1991)
`(“Maly1991”)
`Scheduling Order 19-cv-861 (Aug. 29, 2019)
`
`Invalidity Contentions Ex B2 - 946 Jaso 19-cv-861 (Mar. 21, 2020)
`
`Invalidity Contentions Ex B3 - 946 Stine 1998 19-cv-861 (Mar. 21,
`2020)
`Certain Semiconductor Devices, Semiconductor Device Packages,
`and Products Containing Same, Inv. No. 337-TA-1010, Initial
`Determination (Jun. 30, 2017)
`Sanudin et al., Structure Design Challenge in Nano-CMOS Device
`(“Sanudin”)
`Tarek Austin PHV Declaration
`
`Declaration of Dr. John Givens, Ph.D.
`
`Curriculum Vitae of Dr. John Givens, Ph.D.
`
`Transcript of the Deposition of Dr. Duane Boning, Ph.D. (Dec. 15,
`2020)
`Transcript of the Deposition of Dr. Duane Boning, Ph.D. (Dec. 18,
`2020)
`Pramanik et al., Integration of CMP into Deep Sub-Micron
`Multilevel Metallization Circuits (1997)
`Stine, A General Methodology for Assessing and Characterizing
`Variation in Semiconductor Manufacturing (1997) (“StineThesis”)
`Divecha et al., Effect of Fine-Line Density and Pitch on
`Interconnect ILD Thickness Variation in Oxide CMP Processes
`(1997)
`
`
`
`ii
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`2018
`Ouma et al., Wafer-Scale Modeling of Pattern Effect in Oxide
`Chemical Mechanical Polishing (1997)
`Divecha et al., A Novel Statistical Metrology Framework for
`Identifying Sources of Variation in Oxide Chemical-Mechanical
`Polishing (1998)
`Stine et al., Rapid Characterization and Modeling of Pattern-
`Dependent Variation in Chemical-Mechanical Polishing (1998)
`Park et al., Pattern and Process Dependencies in Copper
`Damascene Chemical Mechanical Polishing Processes (1998)
`Pan et al., Copper CMP and Process Control (1999)
`
`2019
`
`2020
`
`2021
`
`2022
`
`2023
`
`2024
`
`2025
`
`2026
`
`2027
`
`2028
`
`2029
`
`2030
`
`2031
`
`2032
`
`2033
`
`2034
`
`2035
`
`
`
`
`
`Park et al., Electrical Characterization of Copper Chemical
`Mechanical Polishing
`Hymes et al., Determination of the Planarization Distance for
`Copper CMP Process (1999)
`Park et al., Multi-Level Pattern Effects in Copper CMP (1999)
`
`Park et al., Overview of Methods for Characterization of Pattern
`Dependencies in Copper CMP (2000)
`Tugbawa et al., Modeling of Pattern Dependencies in Multi-Step
`Copper Chemical Mechanical Polishing Processes (2001)
`Ouma et al., Characterization and Modeling of Oxide Chemical-
`Mechanical Polishing Using Planarization Length and Pattern
`Density Concepts (2002)
`US Patent 6,187,680
`
`US Patent 6,376,353
`
`US Patent 6,239,494
`
`Disputed Terms for Construction, 1:19-cv-00861-RGA, Dkt. 144-
`1 (May 6, 2020).
`US Patent 6,730,982
`
`Transcript of April 27, 2021 PTAB Conference Call
`
`Transcript of the Deposition of Dr. Duane Boning, Ph.D. (May 3,
`2021)
`
`iii
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`Introduction
`I.
`Petitioner’s Reply and the supporting Boning Declaration do not meaningfully
`
`engage with — let alone rebut — the dispositive deficiencies in the Petition. Instead,
`
`Petitioner returns to the misdirection and obfuscation that it employed in the Petition,
`
`this time committing itself to increasingly extreme mischaracterizations of the
`
`challenged claims and prior art. Petitioner also repeatedly misrepresents PO’s
`
`arguments, at one point even injecting new and confusing terminology and falsely
`
`attributing it to PO.
`
`The Petition was so deficient that Petitioner submitted a 95-page expert
`
`declaration, replete with eight new exhibits, to try to bolster its 35-page Reply.2 But
`
`the declaration does little more than cloud the issues and sow more confusion. PO
`
`respectfully submits that when the record is viewed in its entirety, there is only one
`
`possible conclusion: each of the Grounds asserted in the Petition fails. The Board
`
`should find all of the challenged claims patentable.
`
`
`2 PO maintains its objections to the scope, content, and length of the Boning
`
`Declaration, as it cannot address all of Dr. Boning’s arguments/evidence in the 6,350
`
`words it has been allotted for this brief. Ex. 2034, 42:5–8; Ex. 1042, 522:16–527:4;
`
`Paper 43. PO will be prepared to discuss any unaddressed issues at the hearing.
`
`
`
`1
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`II. The Relevant Configuration, Relevant Solution, and Intermediate
`Width Limitation Are Not Disclosed in Any of the References.
`The number and complexity of the arguments submitted by the parties in this
`
`proceeding must not be allowed to obscure a central flaw in the grounds asserted in
`
`the Petition (the “Grounds”): not a single one of the references upon which the
`
`Petition’s Grounds rest (the “References”) discloses the Relevant Configuration,
`
`the Relevant Solution, or the Intermediate Width Limitation, all of which are
`
`required by each of the challenged claims. Petitioner has failed to rebut this critical
`
`fact in its Reply because it is impossible to do so.
`
`As the POR explained, the “Relevant Configuration” consists of a relatively
`
`wide conductive line, separated only by a dielectric region that contains no
`
`conductive material, from a series of relatively narrow conductive lines. The
`
`“Relevant Solution” consists of a plurality of laterally spaced “dummy”
`
`trenches/conductors in the dielectric region in the middle of the Relevant
`
`Configuration. And the “Intermediate Width Limitation” requires that at least one
`
`of the dummy trenches/conductors be narrower than the relatively wide conductive
`
`line, and wider than at least one of the relatively narrow conductive lines. POR 1,
`
`7–8, 32. The Relevant Configuration, Relevant Solution, and Intermediate Width
`
`Limitation are required by and embodied in the five limitations 16.a–16.e of
`
`challenged independent claim 16, from which all other challenged claims depend.
`
`
`
`2
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`The POR explains that the Relevant Configuration, Relevant Solution, and
`
`Intermediate Width Limitation — and, thus, claim limitations 16.a–16.e — are not
`
`disclosed in Jaso, Stine1998, or Pramanik1998. POR 1–3, 33–56, 67–73. It
`
`necessarily follows, then, that any combination of these references does not disclose
`
`the limitations either. Id., 56–61, 67–73. This is accordingly not the prototypical
`
`obviousness challenge — in which one prior art reference discloses some of the
`
`limitations, and one or more other references disclose the remainder. Instead, this is
`
`a challenge in which five of the six limitations of the challenged independent claim
`
`are missing from every one of the References.
`
`Notably, Petitioner does not contend that the Relevant Configuration,
`
`Relevant Solution, and Intermediate Width Limitation are disclosed in Jaso,
`
`Stine1998, Pramanik1998, or any combination thereof. Instead, Petitioner shifts to
`
`discussing something that it refers to as the “so-called ‘Claimed Configuration’” —
`
`a term that, though not used in the POR, Petitioner implies was coined by PO.
`
`Petitioner then proceeds to recast arguments PO made regarding the “Relevant
`
`Configuration” as having been made regarding the “Claimed Configuration” — a
`
`term Petitioner invented and left undefined. Reply 1–3, 7, 20, 22, 28–30, 35.
`
`Petitioner even quotes a sentence from the POR and changes “Relevant
`
`Configuration” in the original text to “Claimed Configuration” — without
`
`
`
`3
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`identifying its revision, and without citing the POR so that the reader can check the
`
`accuracy of the purported quote. Compare id., 2, with POR 58–59.
`
`But the point remains: five of the six limitations from the challenged
`
`independent claim are not disclosed in any of the References, and Petitioner’s
`
`arguments should be viewed against this backdrop.
`
`III. Petitioner Advances Several Erroneous Arguments in an Attempt to
`Avoid or Reduce Its Burden of Proof.
`Petitioner’s Claim Construction Abrogates the Clear Meaning of
`A.
`the Claims.
`As the POR explains, Petitioner and Dr. Boning conducted their analyses
`
`using an erroneous construction for challenged claim 16 — one in which any
`
`additional structures interposed among the structures referenced in the claim may be
`
`ignored. POR 17–20, 30. Such a construction is so broad that the claim would read
`
`on a hypothetical embodiment in which a series of narrow lines sits on one edge of
`
`a chip and a relatively wide line sits on the opposite edge — so long as there are
`
`dummy conductors commingled with other structures somewhere in the middle of
`
`the chip. Id. The POR demonstrated that this construction has no basis and
`
`contradicts a POSA’s understanding. Id.3
`
`
`3 Petitioner’s claim that there is no “intrinsic evidence to support PO’s
`
`construction” (Reply 33) ignores the POR. POR 17–18. And Petitioner’s Omega
`
`
`
`4
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`But Petitioner and Dr. Boning have now doubled down, reaffirming their view
`
`that claim 16 covers this hypothetical. Reply 33; Ex. 1052, ¶10. They also claim
`
`that the hypothetical is “not meaningful,” which proves PO’s point: Petitioner’s
`
`construction stretches the claim to such an implausible degree that it reads onto an
`
`embodiment that no POSA would consider to be within the claim’s proper scope.
`
`POR 20.
`
`Petitioner's extreme position eliminates its burden of proving that the prior art
`
`discloses the Relevant Configuration and Relevant Solution. Under Petitioner’s
`
`overbroad construction, it need only show a wide conductive line somewhere, a set
`
`of narrow conductive lines somewhere else, and a plurality of dummy conductors
`
`anywhere in between — and it can disregard all intervening structures. This is
`
`precisely the approach that Dr. Boning has taken on Figure 8 of Stine1998, where
`
`he asserts (incorrectly) that there is a wide conductor near the chip periphery, narrow
`
`
`case supports PO, because it acknowledges that a “negative limitation” is justified if
`
`the specification “reveals [an] express intent to confer on the claim language the …
`
`meaning imparted by th[e] negative limitation.” Omega Eng’g, Inc, v. Raytek Corp.,
`
`334 F.3d 1314, 1323 (Fed. Cir. 2003). The cases cited in the POR are in accord.
`
`
`
`
`
`5
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`lines somewhere in the core, and “open … or sparse” space in between into which
`
`dummies might be inserted — while ignoring everything else. E.g., Ex. 1002, ¶112.
`
`The correct construction, however, is the one advanced by PO. And notably,
`
`Dr. Boning never attempted to demonstrate that the Relevant Configuration and
`
`Relevant Solution — understood correctly — are in Stine1998, Jaso, or
`
`Pramanik1998. Instead, he simply ignored intervening structures. This use of an
`
`incorrect claim construction renders the entire Petition meritless.
`
`Petitioner asserts the Petition “do[es] not rely on intervening trenches.” Reply
`
`31. But the issue is not whether Petitioner’s analysis relied upon intervening
`
`trenches, but rather whether it ignored the presence of intervening trenches (it did).
`
`And Petitioner’s suggestion that its analysis would not change if the correct
`
`construction were used is untrue. For example, Dr. Boning annotated the below
`
`excerpt of Figure 8 from Stine1998 to create an image resembling Figure 7 of the
`
`’946 Patent, and then cited this as support for rearranging Jaso’s figures 3A–9B (Ex.
`
`1002, ¶¶176, 180). But Dr. Boning and Dr. Givens agree that there are and/or may
`
`be intervening structures in the light blue, green, and dark blue regions — structures
`
`that are not resolved in Figure 8. POR 55; Ex. 2011, ¶¶192–94; Ex. 2035, 25:12–20
`
`(green region), 27:14–28:4 (green), 34:22–35:4 (green), 40:17–41:3 (dark blue),
`
`49:1–50:18 (dark blue), 56:22–57:16 (light blue); Ex. 2014, 286:21–288:16 (dark
`
`
`
`6
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`blue). If those intervening features were included in Dr. Boning’s image, it would
`
`no longer look anything like Figure 7 from the ’946 Patent. Ex. 2011, ¶¶192–94.
`
`
`
`
`
`Pet. 60 (annotated).
`
`Ex. 1001, Fig. 7.
`
`
`
`7
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`Petitioner Uses Abstraction to Mask Material Differences
`B.
`Between the Challenged Claims and the Prior Art.
`Petitioner frequently describes the claims and prior art at a level of abstraction
`
`that masks their material differences. For example, Petitioner argues that “the prior
`
`art addresses the same problem as the claims — improving the planarity of
`
`semiconductor layers subject to polishing.” Reply 1.
`
`This generalization obliterates the context and details that a POSA would
`
`immediately understand upon reading the ’946 Patent and each of the References.
`
`Indeed, Petitioner appears to be describing the patent and prior art from the
`
`perspective of a layperson, and in a manner so generalized as to sweep in references
`
`addressing a broad range of issues.
`
`This is improper. The challenged claims and patent, like the prior art, must
`
`be read through the eyes of a POSA, who would understand them at a level of detail
`
`far beyond Petitioner’s generalized descriptions. See Medtronic, Inc. v. Cardiac
`
`Pacemakers, Inc., 721 F.2d 1563, 1572 (Fed. Cir. 1983) (“Simplification of a
`
`reference for description purposes cannot … be allowed to cross the line at which it
`
`becomes modification of that reference’s disclosure.”). For example, a POSA would
`
`understand that the ’946 Patent addresses nonplanarities that arise at the end of CMP,
`
`independent of any starting nonplanarities in the metal surface, and that this is a
`
`different problem than that disclosed in Jaso. POR 61–62; Ex. 2011, ¶¶215–18. An
`
`observation that the ’946 Patent and Jaso both attempt to improve planarity, or relate
`
`
`
`8
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`broadly to CMP, or pertain even more generally to semiconductor processing, does
`
`not change this fact — and does not mean that the two patents address the “same
`
`problem.”4
`
`Petitioner must engage with the disclosures at the level of detail that a POSA
`
`would understand them, which it has failed to do.
`
`C.
`
`Petitioner Seeks to Establish Limitations Missing from the Prior
`Art by Arguing That They “Could” Be Present.
`The Petition effectively invites the Board to project into the References a
`
`series of assumed design choices that “could” result in the Relevant Configuration,
`
`Relevant Solution, and Intermediate Width Limitation. Reply 3–4. For example,
`
`suppose that under Ground 1, one were to use Jaso for its express disclosures and
`
`
`4 Petitioner is wrong that Jaso “addresses [the] same planarity problems as the ’946
`
`patent.” Reply 15–17; see POR 61–63; Ex. 2011, ¶¶215–20; Ex. 1042, 553:8–558:3.
`
`Jaso states that high-pattern-factor (HPF) regions experience “erosion of dielectric
`
`and dishing of metal” (Ex. 1004, 3:23–25), a single phenomenon that Jaso (a) refers
`
`to as the “dishing effect” (id., 3:29) and (b) says will result in a region that is “dished”
`
`(id., 6:40–47). The only HPF region contemplated in Jaso is comprised of a series
`
`of densely packed lines (id., 6:34–43, Figs. 10B–10C (region 16a); Figs. 8A–8B).
`
`Jaso thus teaches that regions with densely packed lines suffer from the “dishing
`
`effect,” which encompasses “erosion of dielectric and dishing of metal.”
`
`
`
`9
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`Stine1998 as the vehicle through which to “supply” the Relevant Configuration,
`
`Relevant Solution, and Intermediate Width Limitation required by the claims. In
`
`order to do so, all of the following design choices would need to be assumed and
`
`injected into Figure 8 of Stine1998:
`
`• the features around the perimeter could be bond pad structures;
`
`• the bond pad structures could include a large metal pad (“Pad”);
`
`• the Pad could be electrically connected to other circuitry (which is
`
`necessary to make it the claimed “conductive line”);
`
`• the Pad could have no driver circuitry or other structures on its side
`
`facing the core;
`
`• the white ring around the core could have a dielectric area with no
`
`conductive lines (“Area”), and the Area could be positioned
`
`immediately adjacent to the Pad;
`
`• the Area on its other side could be positioned immediately adjacent to
`
`a series of narrow lines in the core;
`
`• an implementation of Stine1998’s teaching could determine that the
`
`Area falls in a region having a pattern density below the threshold;
`
`• the implementation could result in a plurality of dummies being added
`
`to the Area;
`
`
`
`10
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`• the implementation could result in a plurality of Intermediate Width
`
`dummies being added to the Area.
`
`None of these assumed design choices is actually disclosed in Stine1998, and each
`
`represents just that — a choice from among many different available options. So,
`
`the argument would basically be that one could concatenate a very specific set of
`
`choices for the various parameters identified above, in a way that one could create
`
`an embodiment that resembles the structure claimed in the ’946 Patent. A similar
`
`list of “could” statements can be articulated for each of the other References or
`
`Grounds. See, e.g., Reply 3–4 (injecting a series of assumed design choices into
`
`Jaso’s figures, and falsely alleging that “PO’s expert admitted that the claimed
`
`invention … was an obvious application of Jaso”).
`
`This approach is improper because it isolates each limitation of the challenged
`
`claims and argues that it could be met in the context of the given reference — solely
`
`because the limitation may have been known as one of the many available choices
`
`in the art. But a claim cannot be invalidated by simply proving that each limitation
`
`was separately known in the art. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398, 418–
`
`19 (2007) (“[A] patent composed of several elements is not proved obvious merely
`
`by demonstrating that each of its elements was, independently, known in the prior
`
`art….[C]laimed discoveries almost of necessity will be combinations of what, in
`
`some sense, is already known.”); In re Keller, 642 F.2d 413, 425–26 (C.C.P.A.
`
`
`
`11
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`1981). Instead, Petitioner must demonstrate that the combination of limitations in
`
`each challenged claim is obvious based upon the Grounds. See id. Having failed to
`
`identify five of the six limitations of claim 16 in Jaso, Stine1998, or Pramanik1998,
`
`Petitioner has not carried this burden.
`
` Petitioner cites Unwired Planet for the proposition that prior art that
`
`“sometimes meet[s]” the claim elements can render the claim obvious, even if it
`
`“would not always” satisfy the claim limitations. Reply 2. But Unwired Planet is
`
`inapposite, because the prior art there was shown to meet the claim elements at least
`
`sometimes. See, e.g., Unwired Planet, LLC v. Google Inc., 841 F.3d 995, 1002 (Fed.
`
`Cir. 2016) (“It is enough that the combination would sometimes perform all the
`
`method steps….”).
`
` Here, Petitioner cannot show
`
`that Jaso, Stine1998,
`
`Pramanik1998, or any combination of them ever results in a structure that satisfies
`
`all of the limitations of independent claim 16.
`
`First, as explained above, none of the References discloses the Relevant
`
`Configuration — and it is neither possible nor permissible to inject it into the
`
`References by assuming a particular permutation of design choices to arrive at the
`
`missing claim limitations. The absence of the Relevant Configuration makes it
`
`impossible to conclude that the References, alone or in combination, would
`
`sometimes result in a structure that satisfies claim 16.
`
`
`
`12
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`Second, even assuming the Relevant Configuration were improperly injected
`
`into the References, it would still be impossible to conclude that the References
`
`would sometimes lead to a structure that satisfies claim 16. This is because each of
`
`the References is missing an algorithm, or other instruction, for determining
`
`precisely where to insert dummies. Jaso and Stine1998 merely say to add dummies
`
`somewhere within a region found to have a pattern density below the selected
`
`threshold; they never specify where within that region. And Pramanik1998 says to
`
`add dummies to open areas in a way that has “minimal impact” on the capacitance
`
`and resistance of interconnects, leaving unstated where to add, or not add, dummies.
`
`Ex. 1007, 3. It is therefore impossible to conclude that the References would
`
`sometimes place the dummies into the middle of the Relevant Configuration as
`
`required by claim 16, even assuming (incorrectly) that the Relevant Configuration
`
`were present. See ParkerVision, Inc. v. Qualcomm Inc., 903 F.3d 1354, 1363 (Fed.
`
`Cir. 2018) (affirming nonobviousness where petition failed to provide evidence that
`
`prior art, “even if capable of doing so, actually would output [the claimed
`
`invention]”); see also Unified Patents Inc. v. Intellectual Ventures II, LLC, No.
`
`IPR2016-01404, 2018 WL 4693855, *3 (P.T.A.B. Sept. 27, 2018) (distinguishing
`
`case relied upon in Unwired Planet because that case addressed “a prior art product
`
`that sometimes, but not always, embodies a claimed method….Here, however,
`
`
`
`13
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`Petitioner failed to show that [the asserted prior art] ever embodies the claimed
`
`method.”).
`
`IV. Petitioner Improperly Mischaracterizes and Recasts the ’946 Patent.
`Petitioner’s Reply misrepresents the ’946 Patent’s disclosure, and associated
`
`PO comments, in several ways.
`
`First, Petitioner falsely accuses PO of requiring that the prior art address the
`
`same problem as the ’946 Patent. Reply 8–9. The POR makes clear that only the
`
`Relevant Configuration, Relevant Solution, and Intermediate Width Limitation are
`
`required by the challenged claims, not the Relevant Problems. POR, Sections
`
`VIII.A.1, VIII.B.1. Yet, considering the problems addressed by the prior art, and
`
`comparing them to the problems addressed by the ’946 Patent, are appropriate in an
`
`obviousness analysis, particularly to understand the scope, content, and teachings of
`
`the prior art and to evaluate motivation to combine. See Broadcom Corp. v. Emulex
`
`Corp., 732 F.3d 1325, 1334 (Fed. Cir. 2013) (“[P]rior art references that address
`
`different problems may not, depending on the art and circumstances, support an
`
`inference that the skilled artisan would consult both of them simultaneously.”).
`
`Petitioner’s cases, including Ericsson and Samsung (Reply 8–9), do not suggest
`
`otherwise, but merely confirm that a reference does not necessarily need to address
`
`the same problem to be used in an obviousness challenge.
`
`
`
`14
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`Second, Petitioner claims that PO argued that the ’946 Patent covers situations
`
`in which the starting metal surface is perfectly planar — and then attacks this as
`
`being physically impossible and inconsistent with the patent. Reply 10. This is a
`
`strawman. PO’s point was that the ’946 Patent is concerned about dishing and
`
`erosion that arise at the end of CMP for reasons separate and apart from any
`
`nonplanarities in the starting metal surface. POR 15, 62. In trying to explain and
`
`isolate this separate source of dishing and erosion, PO said that even if —
`
`hypothetically — one could start with a perfectly planar metal surface, the dishing
`
`and erosion described in the ’946 Patent would still occur. Id.; Ex. 2011, ¶¶119,
`
`218, 219 n.31. And this is the problem the ’946 Patent solved.
`
`Third, Petitioner — in an attempt to equate the ’946 Patent with Jaso —
`
`suggests that the ’946 Patent “recognize[s]” pre-CMP nonplanarities in the metal
`
`surface as the “primary,” or perhaps even only, contributor to the dishing and oxide
`
`erosion with which the ’946 Patent is concerned. Reply 10–15. Petitioner supports
`
`this mischaracterization through a quote from the specification that it cuts off mid-
`
`sentence, without even indicating the omission through ellipses. Compare Ex. 1001,
`
`3:6-9, with Reply 11. The portion of the sentence that Petitioner deleted — indicated
`
`with bold underlining here — proves its argument to be baseless: “If the surface
`
`being polished is initially bowed or arcuate (i.e., not planar), the polishing pad will
`
`
`
`15
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`take on the shape of nonplanar regions causing further dishing of the surface
`
`being polished.” Ex. 1001, 3:6–9 (emphasis added).
`
`The reference to starting nonplanarities leading to “further dishing” proves
`
`that there is another, more primary source of dishing that the patent seeks to resolve.
`
`That other source is the dishing that occurs at the end of CMP independent of starting
`
`nonplanarities. Ex. 1042, 558:8–565:21. Moreover, the ’946 Patent does not link
`
`starting nonplanarities to the oxide erosion with which it is concerned, contrary to
`
`Petitioner’s claim. See Ex. 1001, 3:25–44.
`
`Finally, Petitioner incorrectly suggests that the ’946 Patent discloses and
`
`claims a pattern density approach. Reply 14. The ’946 Patent never uses the term
`
`“pattern density.” Additionally, a pattern density approach: (a) focuses across the
`
`entire chip surface; and (b) subdivides that surface into regions, making it possible
`
`to determine the pattern density in each region and compare it against a minimum
`
`threshold. POR 23. None of these concepts is disclosed in the ’946 Patent. Indeed,
`
`there is no instruction on what “region” one would even use — in connection with
`
`the Relevant Configuration — to calculate a pattern density. The ’946 Patent is only
`
`concerned with the Relevant Problems associated with the Relevant Configuration,
`
`and does not address other sources of nonplanarity like variations in pattern density.
`
`
`
`16
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`Petitioner Improperly Mischaracterizes and Recasts the Prior Art.
`V.
`A. The Prior Art Does Not Instruct a POSA to Insert Dummy Metal
`in “Every” Open Area.
`Petitioner states that the prior art discloses placing dummies “wherever there
`
`is a sufficiently large open area” or in “every low density or open area, including
`
`every instance of the ‘claimed configuration’….” Reply 1–2. These statements
`
`imply, incorrectly, that the art inserts dummies everywhere.
`
`As the POR explains, Jaso and Stine1998 both use a pattern density approach.
`
`POR 23–24. This entails subdividing the chip into regions and adding dummies into
`
`a region if its pattern density is below the minimum until the minimum is reached.
`
`Id. Stine1998 expressly warns that “blindly adding metal-fill without considering
`
`the impact on capacitance is disastrous….” Ex. 1005, 670. And Jaso’s Figure 8A
`
`shows an “open area” sufficient for a dummy line, that Jaso leaves open because the
`
`region’s pattern density is higher than the minimum:
`
`Ex. 1004, Fig. 8; 5:56–60. Similarly, Pramanik1998 mentions adding dummy metal
`
`to fill open areas, but also cautions that “the techniques used for modifying the metal
`
`
`
`
`
`17
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`pattern should have minimal impact on interconnect properties such as capacitance
`
`and resistance.” Ex. 1007, 3. Thus, none of the References adds dummies to every
`
`open region, to all open areas, or to every instance of the “claimed configuration.”
`
`The Prior Art Does Not Disclose or Suggest the Relevant
`B.
`Configuration, Relevant Solution, or Intermediate Width Limitation.
`While Petitioner and Dr. Boning never expressly state that Jaso or Stine1998
`
`discloses the Relevant Configuration, Relevant Solution, and Intermediate Width
`
`Limitation, they have presented misleading drawings and arguments incorrectly
`
`suggesting this.
`
`Jaso
`1.
`Petitioner’s Reply offers composite figures prepared by Dr. Boning,
`
`purportedly depicting “exemplary application[s]” of Jaso. Reply 4–5. Dr. Boning’s
`
`images are replicated below, after Jaso’s original images:
`
`
`
`
`
`
`
`18
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`
`“b
`
`I?
`
`'5 :2
`
`III;
`
`:5
`
`2r:-
`
`5
`
`3a
`
`a;
`
`FIG. 3A
`
`FIG. SH
`
`FIG. 4A
`
`FIG. 4B
`
`.-
`'J'
`
`-5
`
`I?
`
`3|:
`
`'7;
`
`..‘
`"'
`. In F
`
`ES
`
`'1
`I”...
`
`2:
`
`'5
`
`f
`
`‘.H
`
`FIG. 5A
`
`FIG. 55
`
`Fl
`
`‘3.
`
`FIG. EB
`
`mm] mum
`
`FIG. Ir'h
`
`FIG. ?B
`
`FIG. 3A
`
`FIG. SE
`
`SE.
`
`FIG. 9A
`
`FIG. 9E!
`
`Figures from Jasn
`
`Ex. 1004, Figs. 3A–9B.
`
`
`
`
`
`19
`
`

`

`IPR2020-00602
`U.S. Patent No. 6,849,946
`
`
`
`Pet. 59; Reply 4–5, 21, 32.
`
`
`
`Dr. Boning’s modifications are not based on anything found in Jaso. Jaso
`
`does not: (a) disclose wide conductive lines, or suggest widening some of the
`
`illustrated lines (Ex. 1042, 537:13–538:9, 540:9–541:1); (b) instruct the POSA to
`
`move and rearrange the figures into the precise configuration of the ’946 Patent; or
`
`(c) suggest adding two green dummies, both to the right of the wide blue line, in Dr.
`
`Boning’s Fig. 4A above,5 leaving a large “open area” to the left. So, Dr. Boning
`
`
`5 Jaso includes no algorithm/instruction telling the POSA where to add dummies or
`
`how many to add. The ’946 Patent, by contrast, does include an “algorithm”
`
`instructing the POSA to insert a plurality of dummies into the middle of the Rele