`
`
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`____________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`____________
`
`APPLE INC.,
`Petitioner,
`
`v.
`
`COREPHOTONICS, LTD.,
`Patent Owner.
`____________
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`____________
`
`
`PATENT OWNER’S SUR-REPLY
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`TABLE OF CONTENTS
`
`Corephotonics’ Arguments Are Not Foreclosed by Limitations in
`
`The Konno and Mercado Patents Do Not Show that the Proposed
`
`INTRODUCTION .................................................................. 1
`I.
`II. GROUND 2 – OGINO IN VIEW OF BAREAU ....................... 1
`A. Apple Fails to Demonstrate that the Proposed Design Is
`Manufacturable Using Any Technique ............................................ 1
`B.
`Other Claims or by Unsuccessful Arguments in Another IPR ........ 3
`C.
`Design Is “Useful” .......................................................................... 5
`D. Apple Fails to Show the Preferred Embodiments Are Not
`“Finished” or Manufacturable ....................................................... 10
`E.
`Apple’s New Untimely Design Should Be Disregarded ............... 11
`III. GROUND 3 – OGINO IN VIEW OF BAREAU AND
`KINGSLAKE ....................................................................... 13
`A. Apple Provides No Articulable Motivation for Changing the
`Image-Side Surface from Concave to Convex .............................. 13
`B.
`Dr. Sasián’s Zemax Analysis Remains Unreliable ........................ 16
`IV. GROUND 4 – CHEN IN VIEW OF IWASAKI AND BEICH . 18
`The Proposed Combination Is Not Manufacturable ..................... 18
`A.
`B.
`the ’897 Patent and Unreliable ...................................................... 22
`CONCLUSION .................................................................... 30
`
`Dr. Sasián’s Calculations Are Contrary to the Disclosures of
`
`V.
`
`
`
`i
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`Cases
`
`TABLE OF AUTHORITIES
`
`Ariosa Diagnostics v. Verinata Health, Inc.,
`805 F.3d 1359 (Fed. Cir. 2015) ................................................................ 13
`
`General Plastic Industrial Co., Ltd. v. Canon Kabushiki Kaisha,
`IPR2016-01357, Paper 16 (PTAB Nov. 14, 2016) .................................. 13
`
`Intelligent Bio-Sys., Inc. v. Illumina Cambridge Ltd.,
`821 F.3d 1359 (Fed. Cir. 2016) ................................................................ 12
`
`Wasica Fin. GmbH v. Contl. Automotive Sys., Inc.,
`853 F.3d 1272 (Fed. Cir. 2017) ................................................................ 12
`
`Statutes
`
`
`
`35 U.S.C. § 312 ............................................................................................ 12
`
`
`
`
`
`
`ii
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`PATENT OWNER’S EXHIBIT LIST
`
`Description
`Declaration of Tom D. Milster, Ph.D.
`Curriculum Vitae of Tom D. Milster, Ph.D.
`Deposition transcript of José Sasián, January 22, 2021
`José Sasián, Introduction to Lens Design (2019)
`Peter Clark, “Mobile platform optical design,” Proc. SPIE
`9293, International Optical Design Conference 2017,
`92931M (17 December 2014)
`Symmons and Schaub, Field Guide to Molded Optics
`(2016)
`G. Beall, “By Design: Part design 106 – Corner radiuses,”
`Plastics Today (199)
`Handbook of Optics, 2nd ed., vol. 2 (1995)
`Declaration of José Sasián in IPR2019-00030
`Declaration of Marc A. Fenster in Support of Motion to
`Appear Pro Hac Vice
`Declaration of James S. Tsuei in Support of Motion to Ap-
`pear Pro Hac Vice
`Deposition transcript of José Sasián, May 28, 2021
`
`Exhibit No
`2001
`2002
`2003
`2004
`2005
`
`2006
`
`2007
`
`2008
`2009
`2010
`
`2011
`
`2012
`
`iii
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`I.
`
`INTRODUCTION
`
`As set forth below, Apple’s reply fails to rebut the core arguments made
`
`in Corephotonics’ response. Moreover, the reply is rife with technical misun-
`
`derstandings, with unreliable or conclusory analysis, and with untimely new
`
`arguments. Grounds 2–4 of this IPR should be rejected.
`
`II. GROUND 2 – OGINO IN VIEW OF BAREAU
`
`A. Apple Fails to Demonstrate that the Proposed Design Is Manu-
`facturable Using Any Technique
`
`As explained in Corephotonics’ response, Apple’s combination of Ogino
`
`with Bareau rests on a lens design that cannot be manufactured. A lens that
`
`cannot be made cannot satisfy the limitations of the ’897 patent claims to a
`
`“lens assembly” comprising “a plurality of lens elements.” (Ex. 1001 at 8:22,
`
`9:26–27.) And even if it could, no POSITA would be motivated to design such
`
`an impossible lens. Apple’s reply suggests this argument “seems to rely on an
`
`implicit requirement of large-scale injection plastic molding.” (Paper 14 at 9.)
`
`Apple’s own expert has opined that Ogino’s lens would “preferably be made
`
`of plastic via injection molding.” (IPR2019-00030, Ex. 2009, Sasián Decl. at
`
`69.)
`
`1
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`But Apple’s reply profoundly misses the point made in Corephotonics’
`
`response. The point of Corephotonics’ response is that Apple’s proposed de-
`
`sign, with its edge thickness of 0.0394 mm, edge slope of 58.86°, center-to-
`
`edge ratio of 15.238, lack of oversizing, and sharp corners, cannot be success-
`
`fully made using any technique for lens manufacture, whether it be injection
`
`molding of plastic (Ex. 2001, ¶¶ 60, 62, 77–78, 103–108, 112, 117, 121), in-
`
`jection molding of glass (id., ¶¶ 60, 63, 103–108, 112, 117, 119, 120), grinding
`
`or polishing of glass (id., ¶¶ 60, 63, 104–107, 110, 117, 119, 120), diamond
`
`turning (id., ¶¶ 104, 107, 117, 120, 121), or any other technology (id., ¶¶ 106–
`
`107, 117).
`
`To rebut this evidence, it is not enough to say that Ogino is not limited to
`
`injection molded plastic, as Apple does, or to argue that these claims should
`
`not be limited to specific center-to-edge thickness ratios taught in the ’897
`
`patent specification. Rather, presented with such detailed evidence that the
`
`proposed lens is contrary to the teachings of Dr. Sasián’s textbook, references
`
`that Apple itself relies on in its obviousness grounds, and the recognized limits
`
`of fabrication using any technique, it was incumbent on Apple to identify some
`
`technique that could be used to make the proposed lens, even in limited num-
`
`bers and at great cost. Apple does not even attempt to do so, instead attacking
`
`2
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`strawmen such as that Corephotonics argues the ’897 patent is limited to plas-
`
`tic injection molding or to mass production. (See Ex. 2012, Sasián Depo.at
`
`121:7–123:21 (when asked if he had pointed to any technique other than in-
`
`jection molding of plastic that could be used to manufacture these lenses, Dr.
`
`Sasián could only point to a discussion of glass polishing used to make the
`
`spherical glass lenses cited in a different Apple IPR, which he acknowledged
`
`is not “the most appropriate technique” for making aspheric surfaces like
`
`those in Ogino or Chen, aside from a type of glass “correcting plate” used in
`
`large telescopes.))
`
`B. Corephotonics’ Arguments Are Not Foreclosed by Limitations
`in Other Claims or by Unsuccessful Arguments in Another IPR
`
`Apple suggests that it is irrelevant whether the proposed lens could actu-
`
`ally be made, because the claims challenged under Ground 2 do not have a
`
`limitation for the L11/L1e ratio. (Paper 14 at 10–11.) That might be a pursua-
`
`sive argument if Apple were presenting a lens with L11/L1e of 4.0 or even
`
`5.0, but Apple’s lens has a ratio of 15.238! And this ratio is just one of several
`
`features, including edge thickness, edge slope, lack of oversizing, and sharp
`
`corners that make the proposed lens design profoundly unmanufacturable.
`
`Apple argues that Corephotonics’ argument is foreclosed by an argument
`
`it made against a POSITA combining Ogino with Beich in IPR2019-00030.
`
`3
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`What Apple neglects to mention is that Apple argued the opposite position in
`
`that IPR and that the Board rejected this argument from Corephotonics, ex-
`
`pressly finding that:
`
`Beich teaches that a lens designer should have a basic under-
`standing of the manufacturing process, which would include the
`limits for fabricating lens elements in an optical lens assem-
`bly. . . . We are persuaded that the information conveyed in Beich
`would have been understood, known, and applied by the ordinar-
`ily skilled lens designer . . . .
`
`(IPR2019-00030, Paper 32 at 18.) Moreover, Apple argues in this very IPR
`
`that a POSITA would have been motivated to apply Beich’s teachings con-
`
`cerning center-to-edge thickness ratio when modifying Ogino. (Paper 2 at 47–
`
`52.) Apple cannot have it both ways, successfully invalidating one Corepho-
`
`tonics patent on the grounds that a lens designer would consider manufactur-
`
`ability and seeking to do the same for some claims in Ground 3 of this very
`
`IPR, but then arguing Corephotonics cannot make arguments based upon
`
`manufacturability with respect to Ground 2.
`
`Moreover, even if Apple were to somehow change course, and argue a
`
`POSITA would not have been concerned with teachings such as those in
`
`Beich, Corephotonics manufacturability arguments for Ground 2 also rest on
`
`4
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`Dr. Sasián’s textbook, Exhibit 2004. (See Ex. 2001, Milster Decl., ¶¶ 106, 110,
`
`114, 118.) This textbook is titled “Introduction to Lens Design” (Ex. 2004),
`
`which strongly suggests that its teachings on manufacturability are within the
`
`ken of a POSITA lens designer. As Dr. Sasián acknowledged, this textbook
`
`teaches that both glass and plastic lenses must avoid sharp corners to prevent
`
`chipping and other damage. (Ex. 2012, Sasián Depo.at 119:4–120:4.)
`
`C. The Konno and Mercado Patents Do Not Show that the Pro-
`posed Design Is “Useful”
`
`Apple suggests that lens manufacturability is not necessary in the art be-
`
`cause of two published patents with purportedly “useful” lenses that purport-
`
`edly fail to meet the requirements of manufacturability. (Paper 14 at 14–15.)
`
`But Apple’s argument misunderstands these lens designs and the crucial way
`
`they differ from Apple’s proposed lens. While these patents show drawings of
`
`first lenses that appear to have narrow edges and sharp corners, a careful look
`
`at the ray traces shows that no light rays that form the image actually pass
`
`through the part of the first lenses close to these edges:
`
`5
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`(Ex. 1035, Fig. 11, excerpt enlarged; see Ex. 2012, Sasián Depo.at 111:18–
`
`
`
`112:19.)
`
`
`
`(Ex. 1036, Fig. 13, excerpt enlarged; see Ex. 2012, Sasián Depo.at 116:14–
`
`
`
`117:16.)
`
`6
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`In other words, the outer portions of these lens elements are outside of
`
`the “clear apertures” where the lens shape matters to formation of the lens.
`
`(See, Ex. 2001, Milster Decl., ¶¶ 101, 102; Ex. 2012, Sasián Depo.at 109:14–
`
`23, 111:11–17, 112:11–19.) While lens design software may have drawn the
`
`lenses with these sharp edges, a POSITA would understand that these sharp
`
`edges are not necessary to the function of the lens and would not actually be
`
`present. This is confirmed by Dr. Sasián’s own Zemax analysis of precisely
`
`the same lens (unmodified) from the Konno patent, which showed the first
`
`lens ending at the edge of the clear aperture and having a thick edge, not the
`
`sharp edge shown in the Konno figure Apple points to:
`
`7
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`
`
`(IPR2020-00906, Ex. 1021, Sasián Decl. at 31; see Ex. 2012, Sasián Depo. at
`
`108:17–109:13 (confirming this IPR2020-00906 declaration addresses the
`
`same Konno design as cited in the present IPR).) Dr. Sasián acknowledged
`
`this during his deposition. (Ex. 2012, Sasián Depo.at 113:7–114:9.) As Dr.
`
`Sasián further confirmed, a POSITA would understand that the plastic injec-
`
`tion-molded lenses taught by Konno would actually have a flange for mount-
`
`ing, rather than the sharp edges depicted. (Id. at 115:3–24.) Dr. Sasián
`
`confirmed that the Mercado lens would likewise incorporate a flange and have
`
`8
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`a shape different from that shown outside the clear aperture. (Ex. 2012, Sasián
`
`Depo.at 117:11–118:1.)
`
`The situation with Apple’s proposed lens is far different. As Corephoton-
`
`ics showed in its response, the rays that form the image in this proposed lens
`
`go all the way to the edge of the lens, and must do so to satisfy the claim
`
`elements, meaning that the clear aperture goes to the edge of the lens, and the
`
`lens must have the sharp edges to achieve the f-number of 2.8:
`
`(Ex. 2001, Milster Decl., ¶¶ 96–97.)
`
`
`
`In other words, the sharp edges in the figures from Konno and Mercado
`
`are artifacts of how the lens design software chose to draw the lenses outside
`
`the clear apertures, not features that would be present in a manufactured lens.
`
`9
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`The unmanufacturable sharp corners in Apple’s lenses, by contrast, must be
`
`present in the manufactured lens to satisfy the challenged claim elements.
`
`D. Apple Fails to Show the Preferred Embodiments Are Not “Fin-
`ished” or Manufacturable
`
`Apple tries to show that the Example 1 lens of the ’897 patent is “not a
`
`finished lens, suitable for manufacturing.” (Paper 14 at 17–18.) But it does
`
`not show evidence of any flaw in the lens that would prevent it from being
`
`manufactured using common techniques or from being a useful lens. Apple
`
`alleges that ’897 patent Example 1 “is not desensitized” and “suffers from
`
`serious ghost images that are focused on the image plane.” (Paper 14 at 18.)
`
`But as evidence for this, Apple cites only a single-sentence paragraph from
`
`Dr. Sasián’s reply declaration, with the conclusory statement that these prob-
`
`lems are present. (Ex. 1037, ¶ 22.) Dr. Sasián does not say how he knows or
`
`determined that these problems are present or how significant or “serious”
`
`these problems are. He does not mention whether these problems are present
`
`in the patent’s other examples. He certainly does not show evidence that a
`
`POSITA would consider these problems serious. This opinion amounts to little
`
`more than an ipse dixit that “Corephotonics’ own lens is bad,” and it should
`
`be disregarded in its single-sentence entirety.
`
`
`
`10
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`E. Apple’s New Untimely Design Should Be Disregarded
`
`As a last-ditch fallback, Apple offers a different way of combining Ogino
`
`with Bareau, presenting a lens design never mentioned in the Petition. (Paper
`
`14 at 19–22.)
`
`
`
`
`
`(Paper 14 at 19.) In creating this new lens, Dr. Sasián manually changed the
`
`thickness of the first lens element and performed steps in Zemax that turned
`
`off vignetting and that caused the location of the image plane to change, along
`
`with the conic constant of the first lens surface and 32 higher-order aspheric
`
`terms describing the lens surfaces. (Ex. 2012, Sasián Depo.at 113:7–114:9,
`
`133:5–134:2; Ex. 1037 at 40.) In short, this is a new lens design and a new
`
`purported combination of Ogino with Bareau.
`
`11
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`A complex lens design can have numerous problems that are not apparent
`
`to a lay reader and cannot be demonstrated without expert testimony, as evi-
`
`denced by the 25 pages of expert testimony that Corephotonics’ expert de-
`
`voted to Ground 2 as presented in the petition. (Ex. 2001 at 36–61.) The rules
`
`of IPRs, directed as they are to ensuring proceedings finish within the statu-
`
`tory deadlines, do not permit Corephotonics to marshal such evidence in re-
`
`sponse to a new design presented in reply. Accordingly, Apple’s reply was an
`
`opportunity for it to attempt to show that Corephotonics’ criticisms of the
`
`Ground 2 design presented in the petition were invalid, not an opportunity to
`
`use Corephotonics’ response as a roadmap for creating a new invalidity theory
`
`that Corephotonics is prevented by the rules from rebutting. See Intelligent
`
`Bio-Sys., Inc. v. Illumina Cambridge Ltd., 821 F.3d 1359, 1369 (Fed. Cir.
`
`2016) (“It is of the utmost importance that petitioners in the IPR proceedings
`
`adhere to the requirement that the initial petition identify ‘with particularity’
`
`the ‘evidence that supports the grounds for the challenge to each claim.’ 35
`
`U.S.C. § 312(a)(3).”); Wasica Fin. GmbH v. Contl. Automotive Sys., Inc., 853
`
`F.3d 1272, 1286 (Fed. Cir. 2017) (“Rather than explaining how its original
`
`petition was correct, Continental’s subsequent arguments amount to an en-
`
`tirely new theory of prima facie obviousness absent from the petition. Shifting
`
`12
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`arguments in this fashion is foreclosed by statute, our precedent, and Board
`
`guidelines.” Ariosa Diagnostics v. Verinata Health, Inc., 805 F.3d 1359, 1367
`
`(Fed. Cir. 2015) (holding that “a challenge can fail even if different evidence
`
`and arguments might have led to success”); see General Plastic Industrial
`
`Co., Ltd. v. Canon Kabushiki Kaisha, IPR2016-01357, Paper 16 at 9 (PTAB
`
`Nov. 14, 2016) (recognizing the injustice of petitioner being permitted to wait
`
`until after seeing patent owner’s arguments and then craft a new invalidity
`
`argument in response).
`
`III. GROUND 3 – OGINO IN VIEW OF BAREAU AND
`KINGSLAKE
`
`A. Apple Provides No Articulable Motivation for Changing the
`Image-Side Surface from Concave to Convex
`
`Apple compares Dr. Sasián’s change of the image-side surface of Ogino’s
`
`first lens from concave to convex to the change he made to Ogino’s second
`
`lens from biconcave to meniscus in IPR2018-01140. (Paper 14 at 22–23.) But
`
`the differences are substantial. While Corephotonics argued against Apple’s
`
`motivation to modify the Ogino second lens based upon Chen II in IPR2018-
`
`01140, at least Apple presented a cognizable theory of motivation for the
`
`change. Apple presented a reference, Chen II, that contained a meniscus lens.
`
`(IPR2018-01140, Paper 37, Final Written Decision at 33.) Its expert opined
`
`13
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`that a POSITA looking at Chen II would recognize that its meniscus second
`
`lens provided specific advantages that address specific purported deficiencies
`
`in Ogino’s design. (Id. at 34.) In other words, there was a specific reason,
`
`based upon specific disclosures in Chen II, for changing Ogino to satisfy the
`
`claim limitation in question. In light of these facts, the Board rejected Core-
`
`photonics arguments that were based on Ogino’s teaching of the benefits of
`
`the biconcave second lens. (Id. at 35.) Specifically, the Board argued that a
`
`POSITA could weigh the benefits of the biconcave second lens taught by
`
`Ogino against the benefits of a meniscus second lens purported shown in Chen
`
`II. (Id. at 36.) While Corephotonics disagrees with the conclusion, it was at
`
`least grounded in some way, however tenuous, on a lens that was shown in
`
`Chen II.
`
`Apple would take from this decision by the Board in IPR2018-01140—
`
`that it would have been obvious to change a specific lens surface from concave
`
`to convex based upon Chen II—the conclusion that all changes of lens shape
`
`between convex and concave are per se obvious. The Board made no such
`
`finding.
`
`Here, Apple proposes changing the image side surface of Ogino to con-
`
`vex, but it does not cite to a single reference that has such a convex image side
`
`14
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`surface or to anything in the prior art that would suggest that changing this
`
`surface from concave to convex would provide benefits. For that matter, Ap-
`
`ple does not identify any benefits that actually result from this change, whether
`
`those benefits were taught in the art or not.
`
`Moreover, Apple’s reply and the declaration of Dr. Sasián fail to explain
`
`how Dr. Sasián even arrived at the convex shape. As explained in Corepho-
`
`tonics’ response, the Zemax output in Dr. Sasián’s declaration shows that the
`
`radius of curvature for the image-side surface of the first lens in his modifica-
`
`tion to Ogino was fixed to have a value of -2.94377806, and thus the lens was
`
`fixed to have a convex surface (a negative radius for the image-side surface
`
`meaning it is convex):
`
`
`
`15
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`(Ex. 1003, Sasián Decl. at 111; Ex. 2003, Sasián Depo.at 49:20–50:16; Ex.
`
`2001, Milster Decl., ¶¶ 136–137.) This negative radius of curvature was “most
`
`likely” produced using an earlier Zemax optimization that he did not explain
`
`in his original declaration and that he could not recall any details of during his
`
`deposition. (Ex. 2003, Sasián Depo. at 50:18–53:9.) Dr. Sasián’s reply decla-
`
`ration sheds no further light on how or with what steps he obtained the nega-
`
`tive radius of curvature for the image-side surface of the first lens, beyond the
`
`vague statement that he “experiment[ed] with [sic] to see if a smaller f-number
`
`would also have been attainable for Example 5.” (See Ex. 1037, ¶¶ 31–32.)
`
`B. Dr. Sasián’s Zemax Analysis Remains Unreliable
`
`Dr. Sasián’s reply declaration and deposition raise other questions about
`
`the reliability of his Zemax analysis concerning this modification to Ogino.
`
`For example, Corephotonics noted in its response that Dr. Sasián reported
`
`multiple inconsistent values for the f-number in this modification. (Ex. 1003
`
`at 108 (giving both “2.12” and “2.45” as values for the f-number).) In an ap-
`
`parent attempt to clarify the issue, Dr. Sasián’s reply declaration purports to
`
`provide the true values of f-number (and of the related entrance pupil diameter
`
`ENPD). (Ex. 1037, ¶ 33, Paper 14 at 26.) But rather than provide a screenshot
`
`of Zemax showing the values, as he repeatedly did for other values throughout
`
`16
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`his declarations, he testified that he manually typed these “correct” values into
`
`his reply declaration by hand, as evidenced by the use of a font that Zemax
`
`does not use. (Ex. 2012, Sasián Depo.at 101:1–15.) Even more strangely, Dr.
`
`Sasián testified that when he manually typed the value of f-number 2.45 into
`
`his declaration, he was basing it on his memory of typing that number into
`
`Zemax a year earlier, rather than on any specific number displayed by Zemax.
`
`(Id. at 104:2–25.) While the difference between an f-number of 2.12 and 2.45
`
`may not matter to the ’897 claim limitations, it is relevant to the motivation to
`
`make Dr. Sasián’s modifications, as nothing in the record suggests a motiva-
`
`tion to reduce f-number to 2.12. More generally, the failure to accurately ex-
`
`plain how he did this Zemax work and what his results were further calls into
`
`question his claim that a POSITA would have been motivated to change the
`
`image-side surface from concave to convex.
`
`Nothing about a smaller f-number necessarily requires, or even benefits
`
`from, a convex image-side surface. Apple certainly provides no evidence of a
`
`connection. At most and giving him every benefit of the doubt, Dr. Sasián’s
`
`declarations show that a highly skilled paid expert, with enough experiments
`
`at ways to create a lens with reduced f-number, might arrive at some example
`
`that meets the claim limitation that is otherwise missing from the design. Put
`
`17
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`another way, it shows that a highly skilled lens designer whose job it was to
`
`turn Ogino’s lens into a lens that meets the challenged claims could do so.
`
`This does not show that a POSITA would have changed the surface from
`
`concave to convex or that the POSITA would have seen any motivation to
`
`make that change. Apple points to no such motivation anywhere in the prior
`
`art. Apple’s evidence falls far short of what is required to show obviousness.
`
`IV. GROUND 4 – CHEN IN VIEW OF IWASAKI AND BEICH
`
`A. The Proposed Combination Is Not Manufacturable
`
`Corephotonics’ response explained the serious limitations in manufactur-
`
`ing injection-molded lenses that a POSITA would understand make the mod-
`
`ified lens that Apple proposes—with a first lens that is only thousandths of a
`
`millimeter larger than the aperture stop—unmanufacturable. (Paper 12 at 63–
`
`68.) In response, Apple argues that it does not matter whether a lens could
`
`actually be manufactured, because “these manufacturing considerations are
`
`not included in claims 16 or 30” and because “a lens designer would not have
`
`been bound by these specific manufacturing considerations regardless of the
`
`purpose of the lens design.” (Paper 14 at 28.)
`
`This dismissal of manufacturing concerns is inconsistent with the re-
`
`quirements of the claims, with what a POSITA would have understood, and
`
`18
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`with Apple’s own stated motivation for modifying the Chen lens. Apple’s pe-
`
`tition stated that Chen example 1 “would preferably have been manufactured
`
`via injection molding” and gives that fact as a reason for combining with
`
`Beich, a “polymer injection molding reference[].” (Paper 2 at 68.) Apple’s
`
`reply does not dispute that its proposed combination uses injection molded
`
`plastic. Apple’s reply also does not dispute that any of the limitations of man-
`
`ufacturing tolerances for injection molding plastic, such as the tolerances for
`
`diameter, surface displacement, or rounding of corners are real or argue they
`
`are different than what is stated in the patent owner’s response. (See Paper 12
`
`at 66–67.) Apple also does not dispute that a first lens with semi-diameter less
`
`than the stop semi-diameter will lead to highly undesirable light leakage and
`
`a hazy image. (See Paper 12 at 66.)
`
`As a POSITA would recognize, claims 16 and 30 are all about manufac-
`
`turability considerations. The limitation that each claim adds is the require-
`
`ment that the first lens have a ratio L11/L1e less than 3. (Ex. 1001, ’897 patent
`
`at 9:22–25, 10:34–37.) As the ’897 patent specification explains, this require-
`
`ment is entirely about manufacturability. (Ex. 1001, ’897 patent at 2:41–50.)
`
`More generally, the claims of the ’897 patent are directed to “lens as-
`
`19
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`sembl[ies]” comprising a “plurality of lens elements,” i.e. to real physical ob-
`
`jects that can be made in the real world. (E.g., Ex. 1001, ’897 patent at 8:22,
`
`9:26–27. Apple does not suggest—and a POSITA would not understand—that
`
`the ’897 patent claims are directed to non-physical objects that can be simu-
`
`lated in a computer, but not actually made using the techniques known in the
`
`art. A POSITA reading claims 16 and 30 and the related discussion in the ’897
`
`patent specification would understand that they are directed to manufacturable
`
`lenses, not to lenses that happen to meet the L11/L1e ratio requirement but are
`
`completely unmanufacturable for other reasons.
`
`More significantly, Apple’s dismissal of manufacturability is contrary to
`
`the very motivation that its petition relied upon for combining Chen and Iwa-
`
`saki with Beich. As the petition stated, “a POSITA looking to implement op-
`
`tical element specifications using injection molding methods would look to
`
`Beich for guidance on limitations and parameters that affect lens manufactur-
`
`ability.” (Paper 2 at 67.) The very premise of Apple’s obviousness theory for
`
`these claims is that the POSITA is concerned with making the lens manufac-
`
`turable and for that very reason is looking to Beich. To suggest that the
`
`POSITA would not be concerned by the obvious and serious defects in the
`
`20
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`manufacturability of the proposed lens explained in Corephotonics’ response
`
`runs contrary to the core logic of Apple’s obviousness theory.
`
`The most basic problem with manufacturability of the proposed lens—
`
`the problem that the lens semi-diameter is specified as only 0.004 mm larger
`
`than the stop, but that manufacturing tolerance for the lens diameter is
`
`±0.020 mm—is a problem that is apparent from Beich, which recites exactly
`
`that limitation on the precision of the lens diameter in injection molding plas-
`
`tic. (Ex. 1007 at 7.) Based on the teachings of Beich alone, a POSITA would
`
`understand that setting the diameter of the first lens of Chen small enough to
`
`satisfy the Beich Center Thickness to Edge Thickness Ratio teaching will lead
`
`to lenses that leak light in the gap between the lens and the aperture stop, i.e.,
`
`to bad lenses. The POSITA would not be motivated to use a design that would
`
`lead to such defective lenses.
`
`Notably, this situation is different from the situation in IPR2019-00030,
`
`where Apple relied upon the same rules of thumb from Beich. There, Core-
`
`photonics pointed out that other lens elements of the proposed design violated
`
`other rules of thumb from Beich, and argued that meant that a POSITA would
`
`not have looked to Beich’s rules of thumb for the L11/L1e ratio. The Board
`
`rejected that argument, saying that a POSITA would look to supply a value
`
`21
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`not supplied by the Ogino reference, even if other values in Ogino violated
`
`Beich’s rules of thumb. (IPR2019-00030, Paper 32 at 46.) Here, however, the
`
`problem is not that different parameters in Chen violate Beich’s rules. Rather,
`
`the problem is that the very parameter that Apple bases on one of Beich’s
`
`rules, the first lens diameter, leads to a design that is plainly not manufactur-
`
`able under Beich’s rule that specifically addresses lens diameter, i.e. its
`
`“±0.020 mm” lens diameter tolerance. A POSITA looking to Beich’s rules of
`
`thumb would not choose a lens diameter is inconsistent with the very teaching
`
`on tolerances for lens diameter set forth in those rules.
`
`B. Dr. Sasián’s Calculations Are Contrary to the Disclosures of
`the ’897 Patent and Unreliable
`
`Apple also dismisses Corephotonics’ arguments concerning manufactur-
`
`ability on the grounds that the lenses in the ’897 patent purportedly fail to
`
`meet those criteria of manufacturability. But the calculations from Dr. Sasián
`
`that Apple relies on are contrary to the teachings of the ’897 patent and unre-
`
`liable.
`
`The ’897 patent provides three detailed examples of lens assemblies, and
`
`identifies the first lens elements in each of these assemblies as elements 102,
`
`202, and 302, respectively. (Ex. 1001, ’897 patent, Figs. 1A, 2A, 3A.) The
`
`22
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`specification supplies values of the L11/L1e ratio for each of these three ex-
`
`amples. (Ex. 1001, ’897 patent at 2:43–50.) In other words, the lens for the
`
`second example (202) satisfies the limitation of claims 16 and 30, with a value
`
`for L11/L1e of 2.916. The lenses of the first and third examples do not satisfy
`
`this limitation, but rather satisfy different, less demanding ratios that the spec-
`
`ification provides as examples. (Id. at 2:43–48.)
`
`Dr. Sasián’s reply declaration gives very different values for L11/L1e,
`
`saying that the first example is the only one with a ratio less than 3 (equal to
`
`2.99238 according to Dr. Sasián) and that the second example has a ratio of
`
`3.049, i.e. greater than 3. (Ex. 1037, ¶¶ 39–41.) In his deposition, Dr. Sasián
`
`testified that he was unaware that the patent provided values for these ratios
`
`or that his values were different from the patent’s:
`
`Q. So the numbers in the patent are a little different than the num-
`bers that you calculated. In particular, for Example Number 2,
`according to paragraph 40 of your declaration, you calculated an
`L11/L1e ratio of 3.049, whereas the patent says that ratio is
`2.916; would you agree?
`
`A. Could you please tell me the column and the line number on
`the –
`
`23
`
`
`
`Case No. IPR2020-00878
`U.S. Patent No. 10,330,897
`
`Q. In the patent? Yeah, so the number I’m referring to, 2.916,
`appears in Column 2, line 48.
`
`A. Okay. Thank you. Yeah, I see there is a difference.
`
`Q. Prior to the last few minutes, were you aware of this difference
`between the numbers that you gave for the ratio in your declara-
`tion and the number given for the ratio in the patent itself?
`
`A. No, I wasn’t aware of the difference.
`
`(Ex. 2012 at 88:13–89:15.) The fact t