1014-00003 · Optiwave Semiconductor Corp · Claim Construction

memo-rebuttal-prosecution-history-estoppel.docx DOCUMENT Litigation

13 / 83 ← → ✕

PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT

This memorandum is protected by the attorney-client privilege and the work product doctrine. It is intended solely for use by attorneys and authorized personnel of Drummond & Fenwick LLP and Optiwave Semiconductor Corp. in connection with the above-referenced litigation. Any disclosure, copying, distribution, or use of this memorandum by any person other than the intended recipients is strictly prohibited.

INTERNAL MEMORANDUM

TO: Alan Ngo, Responsible Partner, intellectual Property Group Omar Siddiqui, Lead Associate, intellectual Property Group

FROM: Christy Bauer, Associate, intellectual Property Group

DATE: December 12, 2022

RE: Rebuttal of Lumos Photonics' Prosecution History Estoppel Arguments — U.S. Patent Nos. 10,234,567 ("'567 Patent") and 11,045,678 ("'678 Patent")

MATTER: Optiwave Semiconductor Corp. v. Lumos Photonics, Inc., Case No. 5:22-cv-03417-EJD (N.D. Cal.)

C&H MATTER NO.: 1014-00003

CLIENT: Optiwave Semiconductor Corp. ("OWS" or "Optiwave")

I. EXECUTIVE SUMMARY

This memorandum analyzes and rebuts the prosecution history estoppel ("PHE") arguments raised by defendant Lumos Photonics, Inc. ("Lumos") with respect to two of the four patents asserted in this action. Lumos first raised these arguments in its Preliminary Claim Construction Contentions, served September 12, 2022, and further developed them in its Responsive Claim Construction Contentions, served October 3, 2022. We anticipate that PHE will be a central theme of Lumos's Responsive Claim Construction Brief, which is due December 19, 2022. This memorandum is intended to inform both our Reply Claim Construction Brief (due January 9, 2023) and OWS's overall infringement strategy in advance of the Markman hearing currently scheduled for January 30, 2023, before the Honorable Edward J. Torvelis.

Lumos contends that OWS is estopped from relying on the doctrine of equivalents ("DOE") for two key claim limitations: (1) "adiabatically coupled" in independent Claim 1 of the '567 Patent, and (2) "monolithically formed" in independent Claim 1 of the '678 Patent. In both instances, Lumos asserts that these limitations were added by amendment during prosecution to overcome prior art rejections under 35 U.S.C. § 102 and § 103, and that OWS has therefore surrendered the territory between the original, broader claim scope and the amended, narrower claim scope. Lumos argues that this surrender precludes OWS from invoking DOE to capture Lumos's accused LumosBridge 800G transceiver platform, which employs a segmented taper coupler design (relevant to the '567 Patent) and micro-transfer-printed III-V gain elements (relevant to the '678 Patent).

After a thorough review of the prosecution histories for both patents-in-suit, the relevant case law, and the technical literature, my conclusions are as follows:

PHE applies to both terms under the framework established by the Supreme Court in *Festo Corp. v. Fenmore & Co.*, 535 U.S. 722 (2002). Both "adiabatically coupled" and "monolithically formed" were added by amendment during prosecution in direct response to prior art rejections, creating a presumption that OWS surrendered the territory between the original and amended claims. This presumption is, however, rebuttable. OWS can overcome it by establishing one of the three Festo exceptions for each patent, though the strength of the rebuttal varies significantly between the two patents.

For the '567 Patent, the rebuttal position is moderate to strong. The most promising path is the tangential relation exception. The amendment adding "adiabatically coupled" was made to distinguish the Tanaka prior art (U.S. Patent No. 9,456,012), which disclosed butt-coupling and directional coupling — coupling mechanisms that are fundamentally different in kind from any form of taper-based coupling. Because the prior art that prompted the amendment (butt-coupling) is qualitatively far afield from the accused equivalent (segmented taper coupling), the distinction between a continuous taper and a segmented taper bears, at most, only a tangential relation to the reason for the amendment. There is no explicit prosecution disclaimer in the '567 Patent file history that would independently bar DOE.

For the '678 Patent, the rebuttal position is weak. The amendment adding "monolithically formed" was accompanied by the applicant's explicit prosecution statement that "the present invention is distinguished from heterogeneous integration approaches" — a categorical disclaimer that creates both amendment-based estoppel and independent argument-based estoppel. Because micro-transfer printing is generally classified within the field as a species of heterogeneous integration, this disclaimer is likely dispositive. The tangential relation exception is difficult to sustain, the unforeseeable equivalent exception is undermined by prior academic literature, and the inexpressible substitute exception faces an unsympathetic factual record. The '678 Patent's DOE theory is substantially weaker than the '567 Patent's, and the team should be candid about these risks in internal strategy discussions.

The sections that follow provide the detailed legal framework, patent-by-patent analysis, comparative assessment, and recommendations for claim construction briefing and Markman hearing preparation.

II. LEGAL FRAMEWORK — PROSECUTION HISTORY ESTOPPEL

A. The *Festo* Doctrine

The doctrine of prosecution history estoppel limits a patentee's ability to invoke the doctrine of equivalents when the patentee has narrowed the scope of a patent claim during prosecution. The Supreme Court's decision in Festo Corp. v. Fenmore & Co., 535 U.S. 722 (2002), provides the controlling framework.

In Festo, the Supreme Court held that when a patent claim is narrowed by amendment during prosecution for a reason "related to patentability," a presumption arises that the patentee has surrendered the territory between the original claim and the amended claim, and the patentee is barred from asserting DOE to recapture the surrendered subject matter. Id. at 740–41. The Court rejected both the "complete bar" rule (which would have categorically precluded DOE whenever a narrowing amendment was made) and the "flexible bar" rule (which would have left the scope of estoppel to case-by-case determination without any presumption). Instead, the Court adopted a middle ground: a rebuttable presumption of surrender.

The presumption is rebuttable, and the patentee may overcome it by establishing one of three exceptions:

1. The Tangential Relation Exception. The amendment's rationale bore no more than a tangential relation to the equivalent in question. Under this exception, if the reason for the narrowing amendment was only peripherally or incidentally related to the alleged equivalent, the patentee should not be presumed to have surrendered that equivalent. Festo, 535 U.S. at 740–41. The Federal Circuit has explained that this exception applies when "the amendment was not directly relevant to the alleged equivalent," and the patentee can show that the rationale for the amendment was focused on a different aspect of the prior art. See Honeywell Int'l, Inc. v. Hamilton Sundstrand Corp., 370 F.3d 1131, 1143 (Fed. Cir. 2004) (en banc).

2. The Unforeseeable Equivalent Exception. The equivalent was unforeseeable at the time of the amendment. This exception recognizes that a patentee cannot be expected to have anticipated every possible equivalent that might later emerge, particularly in rapidly evolving technological fields. The Federal Circuit has held that an equivalent is "unforeseeable" when it was not "known in the relevant art" at the time of the amendment. Festo Corp. v. Fenmore & Co. (Festo on remand), 344 F.3d 1359, 1369 (Fed. Cir. 2003). The court has further clarified that the inquiry focuses on whether a person of ordinary skill in the art would have known of the equivalent, not merely whether the equivalent existed in the abstract. See id.

3. The Inexpressible Substitute Exception. There was some other reason suggesting that the patentee could not reasonably be expected to have described the insubstantial substitute in question. Festo, 535 U.S. at 740–41. This exception is a residual catch-all that accounts for situations in which the state of claim drafting art, or other practical considerations, prevented the patentee from crafting language that would have expressly encompassed the equivalent. The Federal Circuit has construed this exception narrowly, recognizing that patentees are generally expected to be capable of drafting claims with reasonable precision. See Honeywell, 370 F.3d at 1143–44.

The Federal Circuit's subsequent decisions have further elaborated on the application of these exceptions. In Honeywell International, Inc. v. Hamilton Sundstrand Corp., 370 F.3d 1131 (Fed. Cir. 2004) (en banc), the Federal Circuit addressed several important aspects of the Festo framework, including the allocation of the burden of proof and the proper scope of inquiry under each exception. The court confirmed that the Festo presumption applies to any narrowing amendment made for reasons of patentability, regardless of whether the amendment was made in response to a prior art rejection, an enablement rejection, or any other patentability-related rejection. Id. at 1139–40. In Festo on remand, the Federal Circuit applied the Supreme Court's framework and provided guidance on how courts should evaluate the tangential relation exception, emphasizing the need to compare the rationale for the amendment with the relationship between the claimed feature and the accused equivalent. 344 F.3d at 1369–70.

Additional Federal Circuit precedent informs the analysis here. In Schwarz Pharma, Inc. v. Paddock Laboratories, Inc., 504 F.3d 1371 (Fed. Cir. 2007), the court found that the tangential relation exception was met where the amendment was directed at overcoming a prior art reference disclosing a fundamentally different formulation technology. In Intervet Inc. v. Merial Ltd., 617 F.3d 1282 (Fed. Cir. 2010), the court addressed the scope of surrender in the biotechnology context, holding that the patentee had failed to rebut the Festo presumption where the amendment directly addressed the type of equivalent at issue. These cases illustrate the fact-intensive nature of the Festo rebuttal inquiry and the importance of careful analysis of the prosecution record.

B. Amendment-Based vs. Argument-Based Estoppel

It is essential to distinguish between the two forms of prosecution history estoppel, as both are potentially relevant to the patents at issue.

Amendment-based estoppel arises when a patentee narrows the scope of a claim by formal amendment during prosecution. This is the form of estoppel addressed by the Festo framework, which creates a rebuttable presumption of surrender. Both the '567 Patent and the '678 Patent involve formal claim amendments — the addition of "adiabatically coupled" and "monolithically formed," respectively — and are therefore subject to amendment-based estoppel analysis.

Argument-based estoppel arises when, during prosecution, the patentee makes arguments to the patent examiner that clearly and unmistakably disclaim certain subject matter from the scope of the claims, even without a corresponding formal amendment. See Ridgeway & Co. v. Conceptronic, Inc., 90 F.3d 1576, 1583 (Fed. Cir. 1996); Omega Eng'g, Inc. v. Raytek Corp., 334 F.3d 1314, 1323–24 (Fed. Cir. 2003). Argument-based estoppel operates independently of the Festo amendment framework; it does not require a formal narrowing amendment and does not involve the three Festo exceptions. Instead, it asks whether the patentee made a "clear and unmistakable" disclaimer during prosecution that limits the scope of the claims. If such a disclaimer is found, it constrains both literal claim scope and DOE scope.

Both patents involve formal claim amendments, making amendment-based estoppel applicable to each. However, argument-based estoppel is an additional and independent concern for the '678 Patent, where the applicant's prosecution remarks included the explicit statement that "the present invention is distinguished from heterogeneous integration approaches." This statement constitutes a clear and unmistakable prosecution disclaimer that operates independently of the amendment-based estoppel analysis and is not subject to the Festo rebuttal exceptions. As discussed in Section IV.D below, this argument-based estoppel may be the most significant obstacle to OWS's DOE theory for the '678 Patent.

C. Scope of Surrender

Under both Festo and the prior Supreme Court decision in Warner-Jenkinson Co. v. Harborline Enterprises Davis Chemical Co., 520 U.S. 17 (1997), the scope of surrender is determined by the territory between the original claim language and the amended claim language. The narrowing amendment defines the boundaries of the estoppel: the patentee is presumed to have surrendered all subject matter that falls within the scope of the original claim but outside the scope of the amended claim. Warner-Jenkinson, 520 U.S. at 30–33.

For the Festo presumption to apply, the amendment must have been made for a "substantial reason related to patentability." Warner-Jenkinson, 520 U.S. at 33; Festo, 535 U.S. at 736. Amendments made for purely procedural or cosmetic reasons — such as correcting typographical errors, improving clarity without changing claim scope, or complying with formality requirements — do not trigger the presumption. However, amendments made in response to prior art rejections under 35 U.S.C. §§ 102 or 103 are paradigmatic examples of amendments made for reasons related to patentability. Festo, 535 U.S. at 736.

This point is directly relevant here: both the "adiabatically coupled" amendment in the '567 Patent and the "monolithically formed" amendment in the '678 Patent were made in direct response to prior art rejections — a § 102(a)(1) anticipation rejection for the '567 Patent and a § 103 obviousness rejection based on hybrid integration references for the '678 Patent. There is no dispute that both amendments were made for reasons related to patentability, and Lumos has so argued. Accordingly, the threshold requirement for the Festo presumption is satisfied for both patents, and the analysis turns to whether OWS can rebut the presumption under one of the three Festo exceptions.

D. Burden of Proof

Under the Festo framework, the burden of establishing prosecution history estoppel is allocated as follows. The accused infringer (Lumos) bears the initial burden of establishing that a narrowing amendment was made during prosecution for a reason related to patentability. Festo, 535 U.S. at 740. Once this threshold showing is made — as it has been here for both patents — the burden shifts to the patentee (OWS) to rebut the presumption of surrender by demonstrating that one of the three Festo exceptions applies. Id.; Honeywell, 370 F.3d at 1140.

The Federal Circuit has emphasized that this burden is a meaningful one. The patentee must come forward with evidence and argument sufficient to persuade the court that the presumption of surrender has been overcome. Conclusory assertions or bare allegations are insufficient. See Honeywell, 370 F.3d at 1141–42. At the same time, the inquiry is ultimately a question of law, to be decided by the court, informed by the prosecution history and the relevant technical context. Id. at 1140.

In the context of claim construction briefing and the Markman hearing, PHE arguments are properly addressed as part of the court's overall claim construction analysis, as the scope of estoppel informs the permissible scope of DOE. See Cybor Corp. v. FAS Technologies, Inc., 138 F.3d 1448, 1456 (Fed. Cir. 1998) (en banc). While the ultimate determination of infringement — whether literal or under DOE — is a question for trial, the legal predicates for DOE (including the availability of the defense of prosecution history estoppel) are appropriately resolved at the claim construction stage.

III. ANALYSIS — '567 PATENT (U.S. PATENT NO. 10,234,567)

A. Prosecution History Background

The '567 Patent is entitled "Integrated Waveguide Coupler with Tapered Mode Converter for Silicon Photonics Platforms." It was filed on March 14, 2017, and issued on March 19, 2019, after approximately two years of prosecution. The named inventors are Dr. Kevin Fong, Dr. Priya Ramanathan, and Hideo Tanaka, all of whom were employed by OWS during the relevant period. The patent is generally directed to an integrated optical coupler design for silicon photonics platforms that achieves low-loss fiber-to-chip coupling through a tapered waveguide mode converter.

The prosecution history of the '567 Patent is set forth in detail in Appendix A. For purposes of this memorandum, the critical events are summarized as follows.

The '567 Patent underwent two office action rejections during prosecution. In the first office action, dated approximately August 2017, the examiner rejected the original independent Claim 1 — which did not include any limitation specifying the type of coupling mechanism — as anticipated under 35 U.S.C. § 102(a)(1) by U.S. Patent No. 9,456,012 (Tanaka et al., issued September 6, 2016) (hereinafter "the Tanaka reference" or "U.S. 9,456,012"). The Tanaka reference disclosed a waveguide coupler for silicon photonics that employed butt-coupling and directional coupling approaches — both of which involve abrupt or discrete optical transitions between waveguide sections, rather than the gradual modal transition characteristic of adiabatic coupling.

An irony worth noting: U.S. 9,456,012 was co-invented by Hideo Tanaka prior to his joining OWS (Tanaka was at a different institution from approximately 2014 to 2016 when U.S. 9,456,012 was developed). Tanaka subsequently joined OWS and is a named inventor on the '567 Patent. While this fact does not alter the legal analysis, it underscores the technical continuity between the prior art and the claimed invention and may be worth addressing proactively if Lumos attempts to suggest that OWS was intimately familiar with the Tanaka reference's scope at the time of the amendment.

In response to the first office action rejection, the applicant amended independent Claim 1 to add the limitation "adiabatically coupled to a silicon nitride waveguide layer." Specifically, the claim was narrowed from a generic recitation of a waveguide coupler to recite "a tapered silicon waveguide having a tip width of less than 200 nanometers adiabatically coupled to a silicon nitride waveguide layer." The amendment was accompanied by applicant's remarks, dated approximately January 2018, arguing that the Tanaka prior art (U.S. 9,456,012) did not disclose adiabatic coupling and instead used "abrupt butt-coupling or directional coupling approaches" that resulted in "significant mode mismatch and insertion loss at the coupling interface." The applicant's remarks emphasized that adiabatic coupling, by contrast, achieves "a gradual, low-loss transfer of optical power between waveguide sections through a continuously varying effective index," and that this distinction rendered the amended claims patentable over the Tanaka reference.

Following this amendment and a subsequent second office action (which raised a new § 103 obviousness rejection that was overcome without further amendment to the claim language at issue here), the examiner allowed the claims and the patent issued on March 19, 2019.

Critically, the prosecution history does not contain any statements by the applicant that could be characterized as a categorical disclaimer of any broad class of coupling technologies. The applicant's remarks were focused specifically on distinguishing the Tanaka reference's butt-coupling and directional coupling approaches. The applicant did not state, for example, that "the present invention is distinguished from all non-adiabatic coupling approaches" or make any similarly sweeping disclaimer. This is a significant factual distinction from the '678 Patent prosecution history, discussed in Section IV below.

B. Lumos's PHE Argument

In its Preliminary Claim Construction Contentions (September 12, 2022) and Responsive Claim Construction Contentions (October 3, 2022), Lumos advances the following PHE argument for the '567 Patent.

Lumos argues that by adding "adiabatically coupled" to overcome the Tanaka prior art, OWS surrendered all equivalents of "adiabatically coupled" that fall within the territory between the original broad claim (which specified no coupling type) and the amended claim (which specified adiabatic coupling). Lumos contends that this surrendered territory encompasses all non-adiabatic coupling approaches, including Lumos's accused segmented taper design.

Lumos's accused LumosBridge 800G transceiver product uses a segmented (non-continuous) taper design with discrete coupling sections. In this design, rather than employing a single, continuously tapered waveguide section (as in a classic adiabatic taper coupler), the LumosBridge 800G uses a series of discrete waveguide segments of progressively varying widths. Lumos characterizes this design as non-adiabatic coupling, arguing that the discontinuities between segments prevent the "gradual, continuous" mode transition that defines adiabatic coupling.

Lumos's proposed construction of "adiabatically coupled" reinforces this argument. Lumos proposes that the term be construed as "coupled through a gradual, continuous taper that prevents excitation of higher-order modes throughout the entire transition region." This construction is deliberately narrow: by requiring a "gradual, continuous taper" and by specifying the prevention of higher-order mode excitation "throughout the entire transition region," Lumos seeks to exclude its segmented taper design from the literal scope of the claim. Lumos then argues that PHE independently bars DOE for any design that falls outside this narrow construction, because OWS surrendered such territory when it amended the claims to add "adiabatically coupled."

The net effect of Lumos's argument, if accepted, would be to insulate the LumosBridge 800G from infringement of the '567 Patent under both literal infringement (via the narrow construction of "adiabatically coupled") and DOE (via PHE).

C. Rebuttal Arguments

We have identified arguments under each of the three Festo exceptions. As discussed below, the tangential relation exception provides the strongest basis for rebuttal, while the unforeseeable equivalent and inexpressible substitute exceptions provide supplementary support of varying strength.

1. Tangential Relation Exception. The tangential relation exception is our strongest argument for rebutting the Festo presumption with respect to the '567 Patent. The argument proceeds as follows.

The amendment adding "adiabatically coupled" was made to distinguish the Tanaka prior art (U.S. 9,456,012), which disclosed butt-coupling and directional coupling. These coupling mechanisms are fundamentally different in kind from adiabatic coupling. Butt-coupling involves the abrupt juxtaposition of two waveguide facets, with no taper or gradual transition; optical power is transferred across the interface by direct end-fire coupling. Directional coupling involves the evanescent transfer of optical power between two parallel, closely spaced waveguides over a defined coupling length. Neither of these approaches involves any form of taper — whether continuous or segmented — to achieve mode conversion.

By contrast, the accused equivalent (Lumos's segmented taper design) is a species of taper-based coupling. Like the claimed "adiabatic" coupling, the segmented taper achieves mode conversion through a progression of waveguide sections with varying widths, guiding the optical mode from one waveguide geometry to another. The key distinction between a continuous adiabatic taper and a segmented taper lies in the geometry of the transition (continuous vs. discretized), not in the fundamental coupling mechanism (taper-based mode conversion vs. butt-coupling or directional coupling).

The critical question under the tangential relation exception is whether the reason for the amendment was directed at the same distinction as the distinction between the claimed limitation and the accused equivalent. Here, the answer is no:

• The reason for the amendment was to distinguish butt-coupling and directional coupling (Tanaka prior art) from taper-based adiabatic coupling.

• The distinction between the claimed limitation and the accused equivalent is the distinction between a continuous taper and a segmented taper — both of which are forms of taper-based coupling.

The amendment was directed at the macro-level distinction between taper-based coupling (the claimed invention) and non-taper-based coupling (the prior art). The distinction between continuous and segmented tapers is a micro-level distinction within the family of taper-based coupling approaches. The rationale for the amendment — distinguishing butt-coupling — bears, at most, only a tangential relation to the question of whether a segmented taper is equivalent to a continuous taper.

This argument is well-supported by the prosecution history record. The applicant's remarks in response to the first office action focused exclusively on the absence of any taper-based coupling in the Tanaka reference. The remarks did not discuss, or even mention, taper geometry — continuous, segmented, or otherwise. There is no indication in the prosecution history that the examiner or the applicant considered the distinction between continuous and segmented tapers to be relevant to the patentability analysis. The amendment was not made to distinguish segmented tapers from continuous tapers; it was made to distinguish butt-coupling from adiabatic coupling.

The Federal Circuit's guidance supports the application of the tangential relation exception in these circumstances. In Schwarz Pharma, Inc. v. Paddock Laboratories, Inc., 504 F.3d 1371 (Fed. Cir. 2007), the court applied the tangential relation exception where the amendment was directed at a fundamentally different aspect of the prior art than the alleged equivalent. In Insituform Technologies, Inc. v. CAT Contracting, Inc., 385 F.3d 1360 (Fed. Cir. 2004), the court found the tangential relation exception was met where the reason for the amendment was "only tangentially related to the equivalent in question." Here, the technological distance between butt-coupling (the prior art that prompted the amendment) and segmented taper coupling (the accused equivalent) is substantial, supporting application of the tangential relation exception.

Strength assessment: Moderate to strong. The tangential relation argument is most compelling when the prior art that prompted the amendment is far afield from the accused equivalent, which is the case here. The fundamental distinction between butt-coupling (no taper, no mode conversion through gradual transition) and segmented taper coupling (taper-based, mode conversion through progressively varying waveguide geometry) is qualitatively significant. There is a credible argument that no reasonable patent drafter, amending claims to distinguish butt-coupling, would have understood that she was also surrendering segmented taper designs. This exception should be presented as the primary rebuttal argument in the Reply Claim Construction Brief.

2. Unforeseeable Equivalent Exception. The unforeseeable equivalent exception provides supplementary, though weaker, support for rebutting the Festo presumption.

The argument is that at the time of the amendment — approximately late 2017 to early 2018 — segmented taper couplers for silicon photonics fiber-to-chip coupling were not yet established in the literature as a commercially viable alternative to continuous adiabatic tapers. While the concept of segmented waveguide structures existed in the academic literature, the specific application of segmented taper couplers as a practical substitute for continuous adiabatic taper couplers in commercial silicon photonics platforms had not yet been demonstrated. Lumos's LumosBridge 800G product, which employs a segmented taper coupler, did not ship commercially until Q1 2022 — more than four years after the amendment.

However, this argument carries a significant risk. The Chen & Liu journal article ("Segmented Taper Coupler Design for Low-Loss Fiber-to-Chip Coupling," Journal of Photonic Integration, Vol. 14, No. 3, pp. 217–229 (2016)) was published in 2016, before the '567 Patent's filing date of March 14, 2017, and well before the amendment was made in early 2018. The Chen & Liu article introduces the concept of segmented taper couplers and provides simulation-based results demonstrating their theoretical viability. While the article is primarily theoretical and does not demonstrate commercial implementation, it does establish that the concept of segmented taper couplers was "in the literature" at the time of the amendment.

Lumos will almost certainly cite the Chen & Liu article to argue that segmented tapers were foreseeable at the time of the amendment. The Federal Circuit's standard asks whether the equivalent was "known in the relevant art," and the existence of a peer-reviewed journal article describing the equivalent concept is strong evidence that it was known. See Festo on remand, 344 F.3d at 1369. While there is a distinction between a theoretical simulation study and a commercially viable product, the Federal Circuit has not adopted a "commercial viability" threshold for foreseeability; the inquiry focuses on whether a person of ordinary skill in the art would have been aware of the equivalent as a concept. A person of ordinary skill in the art of silicon photonics in late 2017 would likely have been aware of the Chen & Liu article and the concept of segmented tapers.

Strength assessment: Weak to moderate. The existence of the Chen & Liu 2016 article substantially weakens this argument. While we can argue that there is a meaningful difference between a theoretical simulation study and a known, commercially viable alternative, this distinction may not be sufficient to satisfy the Festo unforeseeability standard. A court may find that the concept of segmented taper couplers was foreseeable, even if the specific commercial implementation used by Lumos was not yet available.

Recommendation: Do not lead with this exception. Use it as a supporting argument to bolster the tangential relation exception, emphasizing that the combination of tangential relation and partial unforeseeability strengthens the overall rebuttal. However, this exception should not bear the primary weight of the argument.

3. Inexpressible Substitute Exception. The inexpressible substitute exception provides additional, moderate support for the rebuttal.

The argument is that the applicant could not reasonably have been expected to draft claim language during prosecution that would expressly encompass both continuous and segmented tapers while still distinguishing the Tanaka butt-coupling prior art. At the time of the amendment, the term "adiabatically coupled" was the accepted term of art in the silicon photonics field to capture taper-based low-loss coupling as a class, encompassing various taper geometries. The distinction between various taper geometries — continuous, segmented, multi-section, inverse — was not well-defined in the art at the time, and there was no established terminology or claim-drafting convention that would have enabled the applicant to draft a single claim limitation that expressly included both continuous and segmented tapers while excluding butt-coupling and directional coupling.

Had the applicant attempted to draft broader claim language, such as "coupled through a taper-based mode conversion mechanism," this language might not have been sufficiently distinguishing from the Tanaka prior art, which the examiner had already found anticipating. The applicant was caught between the need to distinguish the Tanaka reference (requiring specificity) and the desire to maintain broad claim scope (requiring generality). The term "adiabatically coupled" represented the best available term of art to achieve this balance.

Strength assessment: Moderate. This argument has some force, particularly in light of the evolving terminology in the silicon photonics field, but it depends on the Court's assessment of how a skilled patent practitioner would have understood the ability to draft narrower or more inclusive claim language in 2017–2018. The Court may be skeptical of the argument that a skilled patent drafter could not have anticipated the need for broader language — for example, by claiming "taper-based coupling" or "coupling through a waveguide taper" rather than "adiabatically coupled." However, the argument is supported by the fact that these broader formulations might not have overcome the Tanaka rejection.

D. Impact Assessment for '567 Patent DOE Theory

Even if the Festo presumption is successfully rebutted — that is, even if the Court finds that one of the three Festo exceptions applies and PHE does not bar DOE for the "adiabatically coupled" limitation — OWS still must prove that Lumos's segmented taper design is equivalent to the claimed "adiabatic" coupling under the applicable test for equivalence.

The Federal Circuit recognizes two formulations of the equivalence test: the "function-way-result" test and the "insubstantial differences" test. Under the function-way-result test, an element of the accused device is equivalent to a claimed limitation if it performs substantially the same function, in substantially the same way, to achieve substantially the same result. Under the insubstantial differences test, equivalence is found if the difference between the claimed limitation and the accused element is insubstantial. See Warner-Jenkinson, 520 U.S. at 39–40.

OWS's technical expert, Dr. Harold Whitfield of the Whitfield Technical Consulting Group, Inc. (Palo Alto, California), will need to opine that Lumos's segmented taper performs substantially the same function (low-loss optical mode transition between waveguide sections), in substantially the same way (gradual mode conversion through progressively varying waveguide geometry), to achieve substantially the same result (efficient fiber-to-chip coupling with minimal insertion loss). This opinion is technically supportable: segmented tapers achieve mode conversion through a series of waveguide sections that progressively transform the optical mode, and the underlying physics — gradual effective index variation — is fundamentally similar to that of a continuous adiabatic taper. The key difference is geometric (discretized vs. continuous), not functional.

However, this opinion is subject to challenge by Lumos's technical consultant, Dr. Raymond Ng, who serves as Of Counsel to Brannick Howell & Stein LLP and has extensive expertise in silicon photonics waveguide design. Dr. Ng may argue that the "way" in which a segmented taper achieves mode conversion is fundamentally different from a continuous adiabatic taper — specifically, that the discrete segments introduce mode coupling and scattering effects that are absent in a continuous taper, and that the segmented taper design requires different optimization approaches. We should anticipate this argument and ensure that Dr. Whitfield's analysis addresses it directly.

It is also important to note that the literal infringement theory for the '567 Patent depends critically on the Court's construction of "adiabatically coupled" at the Markman hearing scheduled for January 30, 2023. OWS has proposed a broader construction of "adiabatically coupled" as "coupled such that optical mode transitions occur without significant power loss." If the Court adopts this construction, it may encompass Lumos's segmented taper design within the literal scope of the claim, and DOE — and the associated PHE issues — may be unnecessary. The claim construction briefing strategy should therefore prioritize achieving the broader literal construction as the primary path, with DOE as a fallback.

IV. ANALYSIS — '678 PATENT (U.S. PATENT NO. 11,045,678)

A. Prosecution History Background

The '678 Patent is entitled "Multi-Channel Optical Transceiver with Integrated Wavelength-Division Multiplexing on Silicon." It was filed on June 22, 2018, and issued on June 29, 2021, after approximately three years of prosecution. The named inventors are Dr. Priya Ramanathan, Dr. Amara Osei, and Liwei Zhang, all of whom were employed by OWS during the relevant period. The patent is generally directed to a multi-channel optical transceiver that integrates wavelength-division multiplexing (WDM) components on a silicon photonic integrated circuit.

The prosecution history of the '678 Patent is set forth in detail in Appendix B. The prosecution was more complex than the '567 Patent, involving three office action rejections over a period of approximately three years.

The critical amendment occurred during prosecution in response to the examiner's prior art rejection, which was based on references describing hybrid integration (heterogeneous integration) approaches to optical transceiver design. The examiner rejected the claims as obvious under 35 U.S.C. § 103, citing a combination of references that, taken together, disclosed a multi-channel WDM transceiver assembled through hybrid integration techniques — i.e., the separate fabrication of optical gain elements and silicon photonic components, followed by assembly through flip-chip bonding, die attach, or similar hybrid packaging methods.

In response, the applicant amended independent Claim 1 to add the limitation "monolithically formed on a single silicon die." This amendment narrowed the claim from a generic recitation of an integrated WDM transceiver to one that is specifically "monolithically formed" — i.e., fabricated within a single, unified manufacturing process on a single silicon substrate, as distinguished from devices assembled from separately fabricated components.

Critically, the amendment was accompanied by applicant's prosecution remarks that included the following explicit statement:

"The present invention is distinguished from heterogeneous integration approaches, such as those disclosed in the cited references, in which separately fabricated optical gain elements are bonded or attached to a silicon photonic substrate. The WDM components of the claimed transceiver are monolithically formed within the silicon die during front-end-of-line and back-end-of-line processing, thereby achieving superior alignment precision, thermal uniformity, and manufacturing scalability that cannot be achieved through hybrid or heterogeneous integration."

This statement is a clear, unambiguous prosecution disclaimer. It explicitly distinguishes the claimed invention from "heterogeneous integration approaches" as a category, not merely from the specific prior art references cited in the rejection. The language is categorical — "distinguished from heterogeneous integration approaches" — and is not limited to the particular embodiments disclosed in the cited references.

This prosecution statement creates two independent bases for estoppel: (1) amendment-based estoppel under Festo, arising from the formal narrowing amendment adding "monolithically formed"; and (2) argument-based estoppel under Vitronics and its Federal Circuit progeny, arising from the explicit prosecution disclaimer of "heterogeneous integration approaches." As discussed in Section IV.D below, this dual basis for estoppel significantly compounds the difficulty of OWS's PHE rebuttal for the '678 Patent.

B. Lumos's PHE Argument

Lumos argues that by amending the claims to add "monolithically formed on a single silicon die" and expressly disclaiming heterogeneous integration, OWS has surrendered all equivalents that encompass any form of heterogeneous bonding, integration, or assembly of separately fabricated components — including micro-transfer printing of III-V gain elements.

Lumos's accused 800G transceiver platform uses micro-transfer-printed III-V semiconductor gain elements bonded to the silicon photonic integrated circuit (PIC). In this process, III-V semiconductor elements (such as indium phosphide-based gain blocks) are fabricated on a separate donor substrate, then transferred onto the silicon PIC using a micro-transfer printing process that uses an elastomeric stamp to pick up and place micron-scale devices with high precision. The transferred III-V elements are then permanently bonded to the silicon PIC through a subsequent bonding and processing step.

Lumos characterizes its micro-transfer printing process as a form of heterogeneous integration — the III-V gain elements are fabricated separately from the silicon PIC and are subsequently integrated through a bonding process. Lumos argues that this falls squarely within the "heterogeneous integration approaches" that OWS expressly disclaimed during prosecution.

Lumos's proposed construction of "monolithically formed" reinforces this argument. Lumos proposes the following construction: "a WDM component formed during front-end-of-line processing within the same silicon substrate, excluding hybrid integration, flip-chip bonding, or any heterogeneous integration technique." This construction incorporates the prosecution disclaimer directly into the claim construction, effectively reading the applicant's prosecution remarks as a limitation on the claim scope.

Lumos contends that OWS cannot use DOE to capture Lumos's heterogeneously integrated design because: (1) the amendment adding "monolithically formed" triggers the Festo presumption of surrender, and (2) the explicit prosecution disclaimer of "heterogeneous integration approaches" independently bars DOE through argument-based estoppel. Lumos argues that no Festo exception can overcome the categorical prosecution disclaimer.

C. Rebuttal Arguments

We have analyzed the three Festo exceptions as applied to the '678 Patent. As discussed below, the rebuttal position is considerably weaker than for the '567 Patent.

1. Tangential Relation Exception. The tangential relation argument for the '678 Patent proceeds as follows: the prior art references that prompted the "monolithically formed" amendment were directed to macro-scale hybrid integration approaches — specifically, techniques involving the attachment of separately fabricated optical modules or gain elements to a silicon substrate through conventional assembly techniques such as flip-chip bonding, wire bonding, or die attach. These prior art approaches involve the assembly of macroscopic components (on the order of millimeters in scale) using relatively imprecise alignment techniques. By contrast, Lumos's micro-transfer printing process involves the transfer of micron-scale III-V semiconductor elements with sub-micron alignment precision, followed by wafer-level bonding and processing steps that integrate the transferred elements into the silicon PIC in a manner that approaches monolithic integration in terms of alignment precision, thermal coupling, and process integration.

The argument is that micro-transfer printing represents a fundamentally different technology from the macro-scale hybrid integration approaches disclosed in the cited prior art references, and that the distinction between conventional hybrid integration (prior art) and micro-transfer printing (Lumos's design) bears only a tangential relation to the reason for the amendment (distinguishing conventional hybrid integration from monolithic integration).

However, this argument faces a significant obstacle: the applicant's prosecution remarks did not disclaim only the specific hybrid integration techniques disclosed in the cited references. Instead, the applicant disclaimed "heterogeneous integration approaches" as a category. Micro-transfer printing is generally classified within the silicon photonics literature as a form of heterogeneous integration — it involves the integration of III-V semiconductor materials (which cannot be grown on silicon) onto a silicon substrate through a bonding process. While micro-transfer printing differs in scale, precision, and process integration from conventional flip-chip bonding, it is nonetheless a heterogeneous integration technique in the taxonomic sense.

Strength assessment: Weak. The tangential relation argument is difficult to sustain when the prosecution remarks explicitly disclaimed "heterogeneous integration" as a category, and the accused equivalent (micro-transfer printing) is generally classified within that category. The argument depends on the Court drawing a distinction between micro-transfer printing and "heterogeneous integration" that the technical community generally does not draw. While there may be a factual basis for arguing that micro-transfer printing is sufficiently different from conventional hybrid integration to support the tangential relation exception, the categorical nature of the prosecution disclaimer significantly undermines this argument.

2. Unforeseeable Equivalent Exception. The unforeseeable equivalent argument for the '678 Patent is that micro-transfer printing of III-V gain elements onto silicon photonic substrates was an emerging technology at the time of the amendment (approximately the 2019–2020 timeframe, during prosecution) and was not foreseen as a practical commercial alternative to monolithic integration. Lumos's LumosBridge 800G product, which employs micro-transfer printing, did not ship commercially until Q1 2022, well after the amendment was made. At the time of the amendment, micro-transfer printing for photonics applications was primarily a research-stage technology, and its commercial viability had not been demonstrated.

However, as with the '567 Patent, this argument is weakened by the existence of prior academic publications. Academic publications on micro-transfer printing for photonics applications existed before 2019. Researchers at multiple institutions had published on the use of micro-transfer printing to integrate III-V semiconductor devices onto silicon photonic platforms, and the concept was known to persons of ordinary skill in the art. While the technology was not yet commercially deployed, it was a known research concept that a skilled practitioner would have been aware of at the time of the amendment.

Strength assessment: Weak to moderate. The argument faces the same challenge as the unforeseeability argument for the '567 Patent: the concept was in the literature, even if commercial implementation came later. The Federal Circuit's foreseeability standard focuses on whether the equivalent was "known in the relevant art," and the existence of published research on micro-transfer printing for photonics undermines the argument that the equivalent was unforeseeable.

3. Inexpressible Substitute Exception. The inexpressible substitute argument for the '678 Patent is that the applicant could not reasonably have been expected to draft claim language that would distinguish between monolithic integration and micro-transfer printing while still overcoming the hybrid integration prior art. At the time of the amendment, the terminology for distinguishing among various integration approaches in silicon photonics was evolving, and the boundary between "monolithic" and "heterogeneous" integration was not precisely defined. The applicant's use of "monolithically formed" was a reasonable drafting choice that reflected the state of the art and the accepted terminology.

Strength assessment: Weak. The Court is unlikely to be sympathetic to this argument given the applicant's broad, categorical disclaimer of "heterogeneous integration approaches." If the applicant had intended to leave open the possibility of claiming equivalents involving heterogeneous integration techniques, the applicant could have refrained from making such a sweeping prosecution disclaimer. The inexpressible substitute exception is designed to address situations in which the patentee genuinely could not have drafted claim language to encompass the equivalent — not situations in which the patentee affirmatively disclaimed the category of technology to which the equivalent belongs. The applicant's categorical disclaimer effectively foreclosed the ability to argue that the substitute was "inexpressible."

D. Argument-Based Estoppel — Additional Challenge for '678 Patent

Separate from and independent of the amendment-based estoppel analysis under Festo, the explicit prosecution statement disclaiming "heterogeneous integration approaches" creates argument-based estoppel that poses an additional — and potentially dispositive — obstacle to OWS's DOE theory for the '678 Patent.

Under the Federal Circuit's prosecution disclaimer doctrine, as articulated in Ridgeway & Co. v. Conceptronic, Inc., 90 F.3d 1576 (Fed. Cir. 1996), and its progeny, clear and unmistakable statements made during prosecution that disclaim certain subject matter from the scope of the claims operate as a binding limitation on claim scope. See also Omega Eng'g, Inc. v. Raytek Corp., 334 F.3d 1314, 1323–24 (Fed. Cir. 2003) (prosecution disclaimer applies when the patentee "unambiguously disclaimed a particular interpretation during prosecution"). Prosecution disclaimers limit the scope of claims not only for purposes of literal infringement but also for purposes of DOE. See Pharmacia & Upjohn Co. v. Mylan Pharm., Inc., 170 F.3d 1373, 1379 (Fed. Cir. 1999).

The applicant's statement — "the present invention is distinguished from heterogeneous integration approaches" — is a categorical disclaimer. It does not limit the disclaimer to the specific prior art references cited in the examiner's rejection. It does not qualify the disclaimer by reference to particular types of heterogeneous integration (e.g., flip-chip bonding, die attach). It categorically distinguishes the claimed invention from "heterogeneous integration approaches" as a class.

This is the most significant obstacle for DOE on the '678 Patent. Unlike amendment-based estoppel, argument-based estoppel is not subject to the three Festo rebuttal exceptions. The Festo framework applies only to amendment-based estoppel; argument-based estoppel operates under the separate prosecution disclaimer doctrine, which asks simply whether the patentee made a "clear and unmistakable" disclaimer. If the disclaimer is found, it bars DOE without any opportunity for rebuttal under the Festo exceptions. See Omega Eng'g, 334 F.3d at 1325–26.

The argument-based estoppel may be dispositive regardless of whether the Festo amendment-based estoppel exceptions apply. Even if OWS could successfully argue that one of the Festo exceptions rebuts the presumption of surrender arising from the amendment, the independent argument-based estoppel from the prosecution disclaimer would still bar DOE. OWS would need to overcome both the amendment-based estoppel and the argument-based estoppel to pursue DOE for the '678 Patent, and the argument-based estoppel has no rebuttal mechanism analogous to the Festo exceptions.

Assessment: This is a formidable challenge. The prosecution disclaimer is clear, categorical, and directly on point. Micro-transfer printing is a species of heterogeneous integration, and the disclaimer encompasses it. While we will argue in our Reply Brief that the disclaimer was directed at the specific prior art approaches and should not be read as a blanket exclusion of all heterogeneous integration technologies (see Section VI.A below), this argument faces long odds given the categorical language of the disclaimer.

Recommendation: We should acknowledge this risk candidly in internal strategy discussions. The '678 Patent's DOE theory is substantially weaker than the '567 Patent's, and the team should plan the overall infringement strategy accordingly.

E. Impact Assessment for '678 Patent DOE Theory

The combination of amendment-based estoppel and argument-based estoppel makes DOE extremely difficult for the '678 Patent. The amendment-based estoppel triggers the Festo presumption of surrender, and the rebuttal arguments under the three Festo exceptions are, at best, weak. The argument-based estoppel from the prosecution disclaimer operates independently and may be dispositive, with no rebuttal mechanism available.

The DOE challenge is compounded by the literal infringement challenge. If the Court adopts Lumos's proposed narrow construction of "monolithically formed" — "a WDM component formed during front-end-of-line processing within the same silicon substrate, excluding hybrid integration, flip-chip bonding, or any heterogeneous integration technique" — literal infringement is also precluded. Lumos's micro-transfer-printed III-V gain elements are not fabricated within the same front-end-of-line/back-end-of-line processing sequence as the silicon PIC; they are fabricated on a separate donor substrate and subsequently transferred. Under Lumos's proposed construction, this approach falls outside the scope of "monolithically formed" as a matter of literal claim scope.

The '678 Patent is the most vulnerable of the four asserted patents from an estoppel perspective. While we should continue to argue for a broader construction of "monolithically formed" and preserve the DOE argument for the record, the team should realistically assess the likelihood of success on this patent and consider whether the overall infringement case should be structured to emphasize the other asserted patents, particularly the '567 Patent and the '234 Patent, which present stronger infringement positions.

V. COMPARATIVE ANALYSIS — STRENGTH OF REBUTTAL BY PATENT

A. Summary Table

The following table summarizes the PHE analysis for the two patents at issue:

B. Strategic Implications

Several strategic implications emerge from this comparative analysis.

First, the '567 Patent offers the best opportunity for a successful PHE rebuttal. The tangential relation exception is well-supported by the prosecution record and the fundamental technological distinction between the Tanaka butt-coupling prior art and Lumos's segmented taper design. The absence of any categorical prosecution disclaimer in the '567 Patent file history means that OWS faces only amendment-based estoppel, which is subject to the full Festo rebuttal framework. If OWS can establish that the tangential relation exception applies — which we believe is a credible argument — DOE remains available for the '567 Patent.

Second, the '678 Patent presents the weakest PHE rebuttal position of the two patents analyzed. The compound effect of amendment-based estoppel and a categorical prosecution disclaimer creates a formidable barrier to DOE. The best Festo exception for the '678 Patent (unforeseeability) is itself weak, and even if it were successful, it would not overcome the independent argument-based estoppel arising from the prosecution disclaimer. The '678 Patent should not be viewed as a reliable infringement vehicle if the case depends on DOE.

Third, for claim construction strategy, OWS's primary focus should be on arguing for broader constructions that capture Lumos's accused designs under literal infringement, thereby reducing or eliminating reliance on DOE. This is particularly critical for the '678 Patent, where the DOE path is effectively foreclosed, and the literal infringement path depends on achieving a construction of "monolithically formed" that encompasses Lumos's micro-transfer-printed design. For the '567 Patent, the broader construction of "adiabatically coupled" ("coupled such that optical mode transitions occur without significant power loss") would similarly capture Lumos's segmented taper design within the literal claim scope.

Fourth, if literal infringement is not available under the Court's constructions, the DOE-plus-PHE-rebuttal path is a viable fallback for the '567 Patent but not a reliable path for the '678 Patent. The infringement case should be structured accordingly, with the '567 Patent positioned as the stronger of the two patents for both literal infringement and DOE.

VI. RECOMMENDATIONS FOR CLAIM CONSTRUCTION BRIEFING

A. Reply Brief Strategy (Due January 9, 2023)

Based on the analysis above, the following approach is recommended for the Reply Claim Construction Brief.

For the '567 Patent:

• Lead with the tangential relation exception. Emphasize that Lumos's PHE argument is built on a false premise — namely, that the amendment adding "adiabatically coupled" was directed at distinguishing all non-adiabatic coupling approaches, including segmented tapers. The prosecution record demonstrates otherwise. The amendment was made to distinguish the Tanaka prior art (U.S. 9,456,012), which disclosed butt-coupling and directional coupling — coupling mechanisms that involve no taper of any kind. The examiner's rejection was based on Tanaka's butt-coupling disclosure, and the applicant's responsive remarks were exclusively directed to the absence of any taper-based coupling in Tanaka. The distinction between a continuous taper and a segmented taper was simply not at issue during prosecution.

• Cite the prosecution history record extensively. The Reply Brief should include detailed quotations from the examiner's rejection and the applicant's responsive remarks to demonstrate that the rationale for the amendment was focused on butt-coupling, not on taper geometry. These quotations should be drawn from Appendix A and presented in a manner that makes the tangential relation argument self-evident from the record.

• Present the tangential relation exception as the primary rebuttal argument. Under Festo and Honeywell, the tangential relation exception applies when the rationale for the amendment was only peripherally related to the accused equivalent. Here, the rationale for the amendment (distinguishing butt-coupling) is qualitatively remote from the accused equivalent (segmented taper coupling), supporting the exception.

• Use the unforeseeable equivalent and inexpressible substitute exceptions as secondary, supporting arguments. While neither is independently strong, both contribute to the overall picture of a patentee who could not reasonably have been expected to surrender segmented taper designs when amending claims to overcome butt-coupling prior art.

• Argue in the alternative for a broader literal construction. Even if the Court is not persuaded by the PHE rebuttal, OWS should argue that the Court should adopt a broader literal construction of "adiabatically coupled" — "coupled such that optical mode transitions occur without significant power loss" — that would capture Lumos's segmented taper design within the literal claim scope, thereby mooting the PHE issue. Emphasize that Lumos's proposed narrow construction ("coupled through a gradual, continuous taper that prevents excitation of higher-order modes throughout the entire transition region") imports limitations from the specification and is inconsistent with the plain and ordinary meaning of "adiabatic" in the art, which refers to the absence of significant power loss during mode transition, not to a specific taper geometry.

For the '678 Patent:

• Address the prosecution disclaimer directly, but with realistic expectations. The Reply Brief should argue that the applicant's disclaimer was directed to the specific hybrid integration approaches disclosed in the cited prior art references — macro-scale assembly techniques such as flip-chip bonding and die attach — and should not be read as a blanket exclusion of all conceivable heterogeneous integration technologies, including micro-transfer printing, which was an emerging technology not contemplated by the applicant's remarks. This

PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT
This memorandum is protected by the attorney-client privilege and the work product doctrine. It is intended solely for use by attorneys and authorized personnel of Drummond & Fenwick LLP and Optiwave Semiconductor Corp. in connection with the above-referenced litigation. Any disclosure, copying, distribution, or use of this memorandum by any person other than the intended recipients is strictly prohibited.
INTERNAL MEMORANDUM
TO: Alan Ngo, Responsible Partner, intellectual Property Group Omar Siddiqui, Lead Associate, intellectual Property Group
FROM: Christy Bauer, Associate, intellectual Property Group
DATE: December 12, 2022
RE: Rebuttal of Lumos Photonics' Prosecution History Estoppel Arguments — U.S. Patent Nos. 10,234,567 ("'567 Patent") and 11,045,678 ("'678 Patent")
MATTER: Optiwave Semiconductor Corp. v. Lumos Photonics, Inc., Case No. 5:22-cv-03417-EJD (N.D. Cal.)
C&H MATTER NO.: 1014-00003
CLIENT: Optiwave Semiconductor Corp. ("OWS" or "Optiwave")
I. EXECUTIVE SUMMARY
This memorandum analyzes and rebuts the prosecution history estoppel ("PHE") arguments raised by defendant Lumos Photonics, Inc. ("Lumos") with respect to two of the four patents asserted in this action. Lumos first raised these arguments in its Preliminary Claim Construction Contentions, served September 12, 2022, and further developed them in its Responsive Claim Construction Contentions, served October 3, 2022. We anticipate that PHE will be a central theme of Lumos's Responsive Claim Construction Brief, which is due December 19, 2022. This memorandum is intended to inform both our Reply Claim Construction Brief (due January 9, 2023) and OWS's overall infringement strategy in advance of the Markman hearing currently scheduled for January 30, 2023, before the Honorable Edward J. Torvelis.
Lumos contends that OWS is estopped from relying on the doctrine of equivalents ("DOE") for two key claim limitations: (1) "adiabatically coupled" in independent Claim 1 of the '567 Patent, and (2) "monolithically formed" in independent Claim 1 of the '678 Patent. In both instances, Lumos asserts that these limitations were added by amendment during prosecution to overcome prior art rejections under 35 U.S.C. § 102 and § 103, and that OWS has therefore surrendered the territory between the original, broader claim scope and the amended, narrower claim scope. Lumos argues that this surrender precludes OWS from invoking DOE to capture Lumos's accused LumosBridge 800G transceiver platform, which employs a segmented taper coupler design (relevant to the '567 Patent) and micro-transfer-printed III-V gain elements (relevant to the '678 Patent).
After a thorough review of the prosecution histories for both patents-in-suit, the relevant case law, and the technical literature, my conclusions are as follows:
PHE applies to both terms under the framework established by the Supreme Court in *Festo Corp. v. Fenmore & Co.*, 535 U.S. 722 (2002). Both "adiabatically coupled" and "monolithically formed" were added by amendment during prosecution in direct response to prior art rejections, creating a presumption that OWS surrendered the territory between the original and amended claims. This presumption is, however, rebuttable. OWS can overcome it by establishing one of the three Festo exceptions for each patent, though the strength of the rebuttal varies significantly between the two patents.
For the '567 Patent, the rebuttal position is moderate to strong. The most promising path is the tangential relation exception. The amendment adding "adiabatically coupled" was made to distinguish the Tanaka prior art (U.S. Patent No. 9,456,012), which disclosed butt-coupling and directional coupling — coupling mechanisms that are fundamentally different in kind from any form of taper-based coupling. Because the prior art that prompted the amendment (butt-coupling) is qualitatively far afield from the accused equivalent (segmented taper coupling), the distinction between a continuous taper and a segmented taper bears, at most, only a tangential relation to the reason for the amendment. There is no explicit prosecution disclaimer in the '567 Patent file history that would independently bar DOE.
For the '678 Patent, the rebuttal position is weak. The amendment adding "monolithically formed" was accompanied by the applicant's explicit prosecution statement that "the present invention is distinguished from heterogeneous integration approaches" — a categorical disclaimer that creates both amendment-based estoppel and independent argument-based estoppel. Because micro-transfer printing is generally classified within the field as a species of heterogeneous integration, this disclaimer is likely dispositive. The tangential relation exception is difficult to sustain, the unforeseeable equivalent exception is undermined by prior academic literature, and the inexpressible substitute exception faces an unsympathetic factual record. The '678 Patent's DOE theory is substantially weaker than the '567 Patent's, and the team should be candid about these risks in internal strategy discussions.
The sections that follow provide the detailed legal framework, patent-by-patent analysis, comparative assessment, and recommendations for claim construction briefing and Markman hearing preparation.
II. LEGAL FRAMEWORK — PROSECUTION HISTORY ESTOPPEL
A. The *Festo* Doctrine
The doctrine of prosecution history estoppel limits a patentee's ability to invoke the doctrine of equivalents when the patentee has narrowed the scope of a patent claim during prosecution. The Supreme Court's decision in Festo Corp. v. Fenmore & Co., 535 U.S. 722 (2002), provides the controlling framework.
In Festo, the Supreme Court held that when a patent claim is narrowed by amendment during prosecution for a reason "related to patentability," a presumption arises that the patentee has surrendered the territory between the original claim and the amended claim, and the patentee is barred from asserting DOE to recapture the surrendered subject matter. Id. at 740–41. The Court rejected both the "complete bar" rule (which would have categorically precluded DOE whenever a narrowing amendment was made) and the "flexible bar" rule (which would have left the scope of estoppel to case-by-case determination without any presumption). Instead, the Court adopted a middle ground: a rebuttable presumption of surrender.
The presumption is rebuttable, and the patentee may overcome it by establishing one of three exceptions:
1. The Tangential Relation Exception. The amendment's rationale bore no more than a tangential relation to the equivalent in question. Under this exception, if the reason for the narrowing amendment was only peripherally or incidentally related to the alleged equivalent, the patentee should not be presumed to have surrendered that equivalent. Festo, 535 U.S. at 740–41. The Federal Circuit has explained that this exception applies when "the amendment was not directly relevant to the alleged equivalent," and the patentee can show that the rationale for the amendment was focused on a different aspect of the prior art. See Honeywell Int'l, Inc. v. Hamilton Sundstrand Corp., 370 F.3d 1131, 1143 (Fed. Cir. 2004) (en banc).
2. The Unforeseeable Equivalent Exception. The equivalent was unforeseeable at the time of the amendment. This exception recognizes that a patentee cannot be expected to have anticipated every possible equivalent that might later emerge, particularly in rapidly evolving technological fields. The Federal Circuit has held that an equivalent is "unforeseeable" when it was not "known in the relevant art" at the time of the amendment. Festo Corp. v. Fenmore & Co. (Festo on remand), 344 F.3d 1359, 1369 (Fed. Cir. 2003). The court has further clarified that the inquiry focuses on whether a person of ordinary skill in the art would have known of the equivalent, not merely whether the equivalent existed in the abstract. See id.
3. The Inexpressible Substitute Exception. There was some other reason suggesting that the patentee could not reasonably be expected to have described the insubstantial substitute in question. Festo, 535 U.S. at 740–41. This exception is a residual catch-all that accounts for situations in which the state of claim drafting art, or other practical considerations, prevented the patentee from crafting language that would have expressly encompassed the equivalent. The Federal Circuit has construed this exception narrowly, recognizing that patentees are generally expected to be capable of drafting claims with reasonable precision. See Honeywell, 370 F.3d at 1143–44.
The Federal Circuit's subsequent decisions have further elaborated on the application of these exceptions. In Honeywell International, Inc. v. Hamilton Sundstrand Corp., 370 F.3d 1131 (Fed. Cir. 2004) (en banc), the Federal Circuit addressed several important aspects of the Festo framework, including the allocation of the burden of proof and the proper scope of inquiry under each exception. The court confirmed that the Festo presumption applies to any narrowing amendment made for reasons of patentability, regardless of whether the amendment was made in response to a prior art rejection, an enablement rejection, or any other patentability-related rejection. Id. at 1139–40. In Festo on remand, the Federal Circuit applied the Supreme Court's framework and provided guidance on how courts should evaluate the tangential relation exception, emphasizing the need to compare the rationale for the amendment with the relationship between the claimed feature and the accused equivalent. 344 F.3d at 1369–70.
Additional Federal Circuit precedent informs the analysis here. In Schwarz Pharma, Inc. v. Paddock Laboratories, Inc., 504 F.3d 1371 (Fed. Cir. 2007), the court found that the tangential relation exception was met where the amendment was directed at overcoming a prior art reference disclosing a fundamentally different formulation technology. In Intervet Inc. v. Merial Ltd., 617 F.3d 1282 (Fed. Cir. 2010), the court addressed the scope of surrender in the biotechnology context, holding that the patentee had failed to rebut the Festo presumption where the amendment directly addressed the type of equivalent at issue. These cases illustrate the fact-intensive nature of the Festo rebuttal inquiry and the importance of careful analysis of the prosecution record.
B. Amendment-Based vs. Argument-Based Estoppel
It is essential to distinguish between the two forms of prosecution history estoppel, as both are potentially relevant to the patents at issue.
Amendment-based estoppel arises when a patentee narrows the scope of a claim by formal amendment during prosecution. This is the form of estoppel addressed by the Festo framework, which creates a rebuttable presumption of surrender. Both the '567 Patent and the '678 Patent involve formal claim amendments — the addition of "adiabatically coupled" and "monolithically formed," respectively — and are therefore subject to amendment-based estoppel analysis.
Argument-based estoppel arises when, during prosecution, the patentee makes arguments to the patent examiner that clearly and unmistakably disclaim certain subject matter from the scope of the claims, even without a corresponding formal amendment. See Ridgeway & Co. v. Conceptronic, Inc., 90 F.3d 1576, 1583 (Fed. Cir. 1996); Omega Eng'g, Inc. v. Raytek Corp., 334 F.3d 1314, 1323–24 (Fed. Cir. 2003). Argument-based estoppel operates independently of the Festo amendment framework; it does not require a formal narrowing amendment and does not involve the three Festo exceptions. Instead, it asks whether the patentee made a "clear and unmistakable" disclaimer during prosecution that limits the scope of the claims. If such a disclaimer is found, it constrains both literal claim scope and DOE scope.
Both patents involve formal claim amendments, making amendment-based estoppel applicable to each. However, argument-based estoppel is an additional and independent concern for the '678 Patent, where the applicant's prosecution remarks included the explicit statement that "the present invention is distinguished from heterogeneous integration approaches." This statement constitutes a clear and unmistakable prosecution disclaimer that operates independently of the amendment-based estoppel analysis and is not subject to the Festo rebuttal exceptions. As discussed in Section IV.D below, this argument-based estoppel may be the most significant obstacle to OWS's DOE theory for the '678 Patent.
C. Scope of Surrender
Under both Festo and the prior Supreme Court decision in Warner-Jenkinson Co. v. Harborline Enterprises Davis Chemical Co., 520 U.S. 17 (1997), the scope of surrender is determined by the territory between the original claim language and the amended claim language. The narrowing amendment defines the boundaries of the estoppel: the patentee is presumed to have surrendered all subject matter that falls within the scope of the original claim but outside the scope of the amended claim. Warner-Jenkinson, 520 U.S. at 30–33.
For the Festo presumption to apply, the amendment must have been made for a "substantial reason related to patentability." Warner-Jenkinson, 520 U.S. at 33; Festo, 535 U.S. at 736. Amendments made for purely procedural or cosmetic reasons — such as correcting typographical errors, improving clarity without changing claim scope, or complying with formality requirements — do not trigger the presumption. However, amendments made in response to prior art rejections under 35 U.S.C. §§ 102 or 103 are paradigmatic examples of amendments made for reasons related to patentability. Festo, 535 U.S. at 736.
This point is directly relevant here: both the "adiabatically coupled" amendment in the '567 Patent and the "monolithically formed" amendment in the '678 Patent were made in direct response to prior art rejections — a § 102(a)(1) anticipation rejection for the '567 Patent and a § 103 obviousness rejection based on hybrid integration references for the '678 Patent. There is no dispute that both amendments were made for reasons related to patentability, and Lumos has so argued. Accordingly, the threshold requirement for the Festo presumption is satisfied for both patents, and the analysis turns to whether OWS can rebut the presumption under one of the three Festo exceptions.
D. Burden of Proof
Under the Festo framework, the burden of establishing prosecution history estoppel is allocated as follows. The accused infringer (Lumos) bears the initial burden of establishing that a narrowing amendment was made during prosecution for a reason related to patentability. Festo, 535 U.S. at 740. Once this threshold showing is made — as it has been here for both patents — the burden shifts to the patentee (OWS) to rebut the presumption of surrender by demonstrating that one of the three Festo exceptions applies. Id.; Honeywell, 370 F.3d at 1140.
The Federal Circuit has emphasized that this burden is a meaningful one. The patentee must come forward with evidence and argument sufficient to persuade the court that the presumption of surrender has been overcome. Conclusory assertions or bare allegations are insufficient. See Honeywell, 370 F.3d at 1141–42. At the same time, the inquiry is ultimately a question of law, to be decided by the court, informed by the prosecution history and the relevant technical context. Id. at 1140.
In the context of claim construction briefing and the Markman hearing, PHE arguments are properly addressed as part of the court's overall claim construction analysis, as the scope of estoppel informs the permissible scope of DOE. See Cybor Corp. v. FAS Technologies, Inc., 138 F.3d 1448, 1456 (Fed. Cir. 1998) (en banc). While the ultimate determination of infringement — whether literal or under DOE — is a question for trial, the legal predicates for DOE (including the availability of the defense of prosecution history estoppel) are appropriately resolved at the claim construction stage.
III. ANALYSIS — '567 PATENT (U.S. PATENT NO. 10,234,567)
A. Prosecution History Background
The '567 Patent is entitled "Integrated Waveguide Coupler with Tapered Mode Converter for Silicon Photonics Platforms." It was filed on March 14, 2017, and issued on March 19, 2019, after approximately two years of prosecution. The named inventors are Dr. Kevin Fong, Dr. Priya Ramanathan, and Hideo Tanaka, all of whom were employed by OWS during the relevant period. The patent is generally directed to an integrated optical coupler design for silicon photonics platforms that achieves low-loss fiber-to-chip coupling through a tapered waveguide mode converter.
The prosecution history of the '567 Patent is set forth in detail in Appendix A. For purposes of this memorandum, the critical events are summarized as follows.
The '567 Patent underwent two office action rejections during prosecution. In the first office action, dated approximately August 2017, the examiner rejected the original independent Claim 1 — which did not include any limitation specifying the type of coupling mechanism — as anticipated under 35 U.S.C. § 102(a)(1) by U.S. Patent No. 9,456,012 (Tanaka et al., issued September 6, 2016) (hereinafter "the Tanaka reference" or "U.S. 9,456,012"). The Tanaka reference disclosed a waveguide coupler for silicon photonics that employed butt-coupling and directional coupling approaches — both of which involve abrupt or discrete optical transitions between waveguide sections, rather than the gradual modal transition characteristic of adiabatic coupling.
An irony worth noting: U.S. 9,456,012 was co-invented by Hideo Tanaka prior to his joining OWS (Tanaka was at a different institution from approximately 2014 to 2016 when U.S. 9,456,012 was developed). Tanaka subsequently joined OWS and is a named inventor on the '567 Patent. While this fact does not alter the legal analysis, it underscores the technical continuity between the prior art and the claimed invention and may be worth addressing proactively if Lumos attempts to suggest that OWS was intimately familiar with the Tanaka reference's scope at the time of the amendment.
In response to the first office action rejection, the applicant amended independent Claim 1 to add the limitation "adiabatically coupled to a silicon nitride waveguide layer." Specifically, the claim was narrowed from a generic recitation of a waveguide coupler to recite "a tapered silicon waveguide having a tip width of less than 200 nanometers adiabatically coupled to a silicon nitride waveguide layer." The amendment was accompanied by applicant's remarks, dated approximately January 2018, arguing that the Tanaka prior art (U.S. 9,456,012) did not disclose adiabatic coupling and instead used "abrupt butt-coupling or directional coupling approaches" that resulted in "significant mode mismatch and insertion loss at the coupling interface." The applicant's remarks emphasized that adiabatic coupling, by contrast, achieves "a gradual, low-loss transfer of optical power between waveguide sections through a continuously varying effective index," and that this distinction rendered the amended claims patentable over the Tanaka reference.
Following this amendment and a subsequent second office action (which raised a new § 103 obviousness rejection that was overcome without further amendment to the claim language at issue here), the examiner allowed the claims and the patent issued on March 19, 2019.
Critically, the prosecution history does not contain any statements by the applicant that could be characterized as a categorical disclaimer of any broad class of coupling technologies. The applicant's remarks were focused specifically on distinguishing the Tanaka reference's butt-coupling and directional coupling approaches. The applicant did not state, for example, that "the present invention is distinguished from all non-adiabatic coupling approaches" or make any similarly sweeping disclaimer. This is a significant factual distinction from the '678 Patent prosecution history, discussed in Section IV below.
B. Lumos's PHE Argument
In its Preliminary Claim Construction Contentions (September 12, 2022) and Responsive Claim Construction Contentions (October 3, 2022), Lumos advances the following PHE argument for the '567 Patent.
Lumos argues that by adding "adiabatically coupled" to overcome the Tanaka prior art, OWS surrendered all equivalents of "adiabatically coupled" that fall within the territory between the original broad claim (which specified no coupling type) and the amended claim (which specified adiabatic coupling). Lumos contends that this surrendered territory encompasses all non-adiabatic coupling approaches, including Lumos's accused segmented taper design.
Lumos's accused LumosBridge 800G transceiver product uses a segmented (non-continuous) taper design with discrete coupling sections. In this design, rather than employing a single, continuously tapered waveguide section (as in a classic adiabatic taper coupler), the LumosBridge 800G uses a series of discrete waveguide segments of progressively varying widths. Lumos characterizes this design as non-adiabatic coupling, arguing that the discontinuities between segments prevent the "gradual, continuous" mode transition that defines adiabatic coupling.
Lumos's proposed construction of "adiabatically coupled" reinforces this argument. Lumos proposes that the term be construed as "coupled through a gradual, continuous taper that prevents excitation of higher-order modes throughout the entire transition region." This construction is deliberately narrow: by requiring a "gradual, continuous taper" and by specifying the prevention of higher-order mode excitation "throughout the entire transition region," Lumos seeks to exclude its segmented taper design from the literal scope of the claim. Lumos then argues that PHE independently bars DOE for any design that falls outside this narrow construction, because OWS surrendered such territory when it amended the claims to add "adiabatically coupled."
The net effect of Lumos's argument, if accepted, would be to insulate the LumosBridge 800G from infringement of the '567 Patent under both literal infringement (via the narrow construction of "adiabatically coupled") and DOE (via PHE).
C. Rebuttal Arguments
We have identified arguments under each of the three Festo exceptions. As discussed below, the tangential relation exception provides the strongest basis for rebuttal, while the unforeseeable equivalent and inexpressible substitute exceptions provide supplementary support of varying strength.
1. Tangential Relation Exception. The tangential relation exception is our strongest argument for rebutting the Festo presumption with respect to the '567 Patent. The argument proceeds as follows.
The amendment adding "adiabatically coupled" was made to distinguish the Tanaka prior art (U.S. 9,456,012), which disclosed butt-coupling and directional coupling. These coupling mechanisms are fundamentally different in kind from adiabatic coupling. Butt-coupling involves the abrupt juxtaposition of two waveguide facets, with no taper or gradual transition; optical power is transferred across the interface by direct end-fire coupling. Directional coupling involves the evanescent transfer of optical power between two parallel, closely spaced waveguides over a defined coupling length. Neither of these approaches involves any form of taper — whether continuous or segmented — to achieve mode conversion.
By contrast, the accused equivalent (Lumos's segmented taper design) is a species of taper-based coupling. Like the claimed "adiabatic" coupling, the segmented taper achieves mode conversion through a progression of waveguide sections with varying widths, guiding the optical mode from one waveguide geometry to another. The key distinction between a continuous adiabatic taper and a segmented taper lies in the geometry of the transition (continuous vs. discretized), not in the fundamental coupling mechanism (taper-based mode conversion vs. butt-coupling or directional coupling).
The critical question under the tangential relation exception is whether the reason for the amendment was directed at the same distinction as the distinction between the claimed limitation and the accused equivalent. Here, the answer is no:
• The reason for the amendment was to distinguish butt-coupling and directional coupling (Tanaka prior art) from taper-based adiabatic coupling.
• The distinction between the claimed limitation and the accused equivalent is the distinction between a continuous taper and a segmented taper — both of which are forms of taper-based coupling.
The amendment was directed at the macro-level distinction between taper-based coupling (the claimed invention) and non-taper-based coupling (the prior art). The distinction between continuous and segmented tapers is a micro-level distinction within the family of taper-based coupling approaches. The rationale for the amendment — distinguishing butt-coupling — bears, at most, only a tangential relation to the question of whether a segmented taper is equivalent to a continuous taper.
This argument is well-supported by the prosecution history record. The applicant's remarks in response to the first office action focused exclusively on the absence of any taper-based coupling in the Tanaka reference. The remarks did not discuss, or even mention, taper geometry — continuous, segmented, or otherwise. There is no indication in the prosecution history that the examiner or the applicant considered the distinction between continuous and segmented tapers to be relevant to the patentability analysis. The amendment was not made to distinguish segmented tapers from continuous tapers; it was made to distinguish butt-coupling from adiabatic coupling.
The Federal Circuit's guidance supports the application of the tangential relation exception in these circumstances. In Schwarz Pharma, Inc. v. Paddock Laboratories, Inc., 504 F.3d 1371 (Fed. Cir. 2007), the court applied the tangential relation exception where the amendment was directed at a fundamentally different aspect of the prior art than the alleged equivalent. In Insituform Technologies, Inc. v. CAT Contracting, Inc., 385 F.3d 1360 (Fed. Cir. 2004), the court found the tangential relation exception was met where the reason for the amendment was "only tangentially related to the equivalent in question." Here, the technological distance between butt-coupling (the prior art that prompted the amendment) and segmented taper coupling (the accused equivalent) is substantial, supporting application of the tangential relation exception.
Strength assessment: Moderate to strong. The tangential relation argument is most compelling when the prior art that prompted the amendment is far afield from the accused equivalent, which is the case here. The fundamental distinction between butt-coupling (no taper, no mode conversion through gradual transition) and segmented taper coupling (taper-based, mode conversion through progressively varying waveguide geometry) is qualitatively significant. There is a credible argument that no reasonable patent drafter, amending claims to distinguish butt-coupling, would have understood that she was also surrendering segmented taper designs. This exception should be presented as the primary rebuttal argument in the Reply Claim Construction Brief.
2. Unforeseeable Equivalent Exception. The unforeseeable equivalent exception provides supplementary, though weaker, support for rebutting the Festo presumption.
The argument is that at the time of the amendment — approximately late 2017 to early 2018 — segmented taper couplers for silicon photonics fiber-to-chip coupling were not yet established in the literature as a commercially viable alternative to continuous adiabatic tapers. While the concept of segmented waveguide structures existed in the academic literature, the specific application of segmented taper couplers as a practical substitute for continuous adiabatic taper couplers in commercial silicon photonics platforms had not yet been demonstrated. Lumos's LumosBridge 800G product, which employs a segmented taper coupler, did not ship commercially until Q1 2022 — more than four years after the amendment.
However, this argument carries a significant risk. The Chen & Liu journal article ("Segmented Taper Coupler Design for Low-Loss Fiber-to-Chip Coupling," Journal of Photonic Integration, Vol. 14, No. 3, pp. 217–229 (2016)) was published in 2016, before the '567 Patent's filing date of March 14, 2017, and well before the amendment was made in early 2018. The Chen & Liu article introduces the concept of segmented taper couplers and provides simulation-based results demonstrating their theoretical viability. While the article is primarily theoretical and does not demonstrate commercial implementation, it does establish that the concept of segmented taper couplers was "in the literature" at the time of the amendment.
Lumos will almost certainly cite the Chen & Liu article to argue that segmented tapers were foreseeable at the time of the amendment. The Federal Circuit's standard asks whether the equivalent was "known in the relevant art," and the existence of a peer-reviewed journal article describing the equivalent concept is strong evidence that it was known. See Festo on remand, 344 F.3d at 1369. While there is a distinction between a theoretical simulation study and a commercially viable product, the Federal Circuit has not adopted a "commercial viability" threshold for foreseeability; the inquiry focuses on whether a person of ordinary skill in the art would have been aware of the equivalent as a concept. A person of ordinary skill in the art of silicon photonics in late 2017 would likely have been aware of the Chen & Liu article and the concept of segmented tapers.
Strength assessment: Weak to moderate. The existence of the Chen & Liu 2016 article substantially weakens this argument. While we can argue that there is a meaningful difference between a theoretical simulation study and a known, commercially viable alternative, this distinction may not be sufficient to satisfy the Festo unforeseeability standard. A court may find that the concept of segmented taper couplers was foreseeable, even if the specific commercial implementation used by Lumos was not yet available.
Recommendation: Do not lead with this exception. Use it as a supporting argument to bolster the tangential relation exception, emphasizing that the combination of tangential relation and partial unforeseeability strengthens the overall rebuttal. However, this exception should not bear the primary weight of the argument.
3. Inexpressible Substitute Exception. The inexpressible substitute exception provides additional, moderate support for the rebuttal.
The argument is that the applicant could not reasonably have been expected to draft claim language during prosecution that would expressly encompass both continuous and segmented tapers while still distinguishing the Tanaka butt-coupling prior art. At the time of the amendment, the term "adiabatically coupled" was the accepted term of art in the silicon photonics field to capture taper-based low-loss coupling as a class, encompassing various taper geometries. The distinction between various taper geometries — continuous, segmented, multi-section, inverse — was not well-defined in the art at the time, and there was no established terminology or claim-drafting convention that would have enabled the applicant to draft a single claim limitation that expressly included both continuous and segmented tapers while excluding butt-coupling and directional coupling.
Had the applicant attempted to draft broader claim language, such as "coupled through a taper-based mode conversion mechanism," this language might not have been sufficiently distinguishing from the Tanaka prior art, which the examiner had already found anticipating. The applicant was caught between the need to distinguish the Tanaka reference (requiring specificity) and the desire to maintain broad claim scope (requiring generality). The term "adiabatically coupled" represented the best available term of art to achieve this balance.
Strength assessment: Moderate. This argument has some force, particularly in light of the evolving terminology in the silicon photonics field, but it depends on the Court's assessment of how a skilled patent practitioner would have understood the ability to draft narrower or more inclusive claim language in 2017–2018. The Court may be skeptical of the argument that a skilled patent drafter could not have anticipated the need for broader language — for example, by claiming "taper-based coupling" or "coupling through a waveguide taper" rather than "adiabatically coupled." However, the argument is supported by the fact that these broader formulations might not have overcome the Tanaka rejection.
D. Impact Assessment for '567 Patent DOE Theory
Even if the Festo presumption is successfully rebutted — that is, even if the Court finds that one of the three Festo exceptions applies and PHE does not bar DOE for the "adiabatically coupled" limitation — OWS still must prove that Lumos's segmented taper design is equivalent to the claimed "adiabatic" coupling under the applicable test for equivalence.
The Federal Circuit recognizes two formulations of the equivalence test: the "function-way-result" test and the "insubstantial differences" test. Under the function-way-result test, an element of the accused device is equivalent to a claimed limitation if it performs substantially the same function, in substantially the same way, to achieve substantially the same result. Under the insubstantial differences test, equivalence is found if the difference between the claimed limitation and the accused element is insubstantial. See Warner-Jenkinson, 520 U.S. at 39–40.
OWS's technical expert, Dr. Harold Whitfield of the Whitfield Technical Consulting Group, Inc. (Palo Alto, California), will need to opine that Lumos's segmented taper performs substantially the same function (low-loss optical mode transition between waveguide sections), in substantially the same way (gradual mode conversion through progressively varying waveguide geometry), to achieve substantially the same result (efficient fiber-to-chip coupling with minimal insertion loss). This opinion is technically supportable: segmented tapers achieve mode conversion through a series of waveguide sections that progressively transform the optical mode, and the underlying physics — gradual effective index variation — is fundamentally similar to that of a continuous adiabatic taper. The key difference is geometric (discretized vs. continuous), not functional.
However, this opinion is subject to challenge by Lumos's technical consultant, Dr. Raymond Ng, who serves as Of Counsel to Brannick Howell & Stein LLP and has extensive expertise in silicon photonics waveguide design. Dr. Ng may argue that the "way" in which a segmented taper achieves mode conversion is fundamentally different from a continuous adiabatic taper — specifically, that the discrete segments introduce mode coupling and scattering effects that are absent in a continuous taper, and that the segmented taper design requires different optimization approaches. We should anticipate this argument and ensure that Dr. Whitfield's analysis addresses it directly.
It is also important to note that the literal infringement theory for the '567 Patent depends critically on the Court's construction of "adiabatically coupled" at the Markman hearing scheduled for January 30, 2023. OWS has proposed a broader construction of "adiabatically coupled" as "coupled such that optical mode transitions occur without significant power loss." If the Court adopts this construction, it may encompass Lumos's segmented taper design within the literal scope of the claim, and DOE — and the associated PHE issues — may be unnecessary. The claim construction briefing strategy should therefore prioritize achieving the broader literal construction as the primary path, with DOE as a fallback.
IV. ANALYSIS — '678 PATENT (U.S. PATENT NO. 11,045,678)
A. Prosecution History Background
The '678 Patent is entitled "Multi-Channel Optical Transceiver with Integrated Wavelength-Division Multiplexing on Silicon." It was filed on June 22, 2018, and issued on June 29, 2021, after approximately three years of prosecution. The named inventors are Dr. Priya Ramanathan, Dr. Amara Osei, and Liwei Zhang, all of whom were employed by OWS during the relevant period. The patent is generally directed to a multi-channel optical transceiver that integrates wavelength-division multiplexing (WDM) components on a silicon photonic integrated circuit.
The prosecution history of the '678 Patent is set forth in detail in Appendix B. The prosecution was more complex than the '567 Patent, involving three office action rejections over a period of approximately three years.
The critical amendment occurred during prosecution in response to the examiner's prior art rejection, which was based on references describing hybrid integration (heterogeneous integration) approaches to optical transceiver design. The examiner rejected the claims as obvious under 35 U.S.C. § 103, citing a combination of references that, taken together, disclosed a multi-channel WDM transceiver assembled through hybrid integration techniques — i.e., the separate fabrication of optical gain elements and silicon photonic components, followed by assembly through flip-chip bonding, die attach, or similar hybrid packaging methods.
In response, the applicant amended independent Claim 1 to add the limitation "monolithically formed on a single silicon die." This amendment narrowed the claim from a generic recitation of an integrated WDM transceiver to one that is specifically "monolithically formed" — i.e., fabricated within a single, unified manufacturing process on a single silicon substrate, as distinguished from devices assembled from separately fabricated components.
Critically, the amendment was accompanied by applicant's prosecution remarks that included the following explicit statement:
"The present invention is distinguished from heterogeneous integration approaches, such as those disclosed in the cited references, in which separately fabricated optical gain elements are bonded or attached to a silicon photonic substrate. The WDM components of the claimed transceiver are monolithically formed within the silicon die during front-end-of-line and back-end-of-line processing, thereby achieving superior alignment precision, thermal uniformity, and manufacturing scalability that cannot be achieved through hybrid or heterogeneous integration."
This statement is a clear, unambiguous prosecution disclaimer. It explicitly distinguishes the claimed invention from "heterogeneous integration approaches" as a category, not merely from the specific prior art references cited in the rejection. The language is categorical — "distinguished from heterogeneous integration approaches" — and is not limited to the particular embodiments disclosed in the cited references.
This prosecution statement creates two independent bases for estoppel: (1) amendment-based estoppel under Festo, arising from the formal narrowing amendment adding "monolithically formed"; and (2) argument-based estoppel under Vitronics and its Federal Circuit progeny, arising from the explicit prosecution disclaimer of "heterogeneous integration approaches." As discussed in Section IV.D below, this dual basis for estoppel significantly compounds the difficulty of OWS's PHE rebuttal for the '678 Patent.
B. Lumos's PHE Argument
Lumos argues that by amending the claims to add "monolithically formed on a single silicon die" and expressly disclaiming heterogeneous integration, OWS has surrendered all equivalents that encompass any form of heterogeneous bonding, integration, or assembly of separately fabricated components — including micro-transfer printing of III-V gain elements.
Lumos's accused 800G transceiver platform uses micro-transfer-printed III-V semiconductor gain elements bonded to the silicon photonic integrated circuit (PIC). In this process, III-V semiconductor elements (such as indium phosphide-based gain blocks) are fabricated on a separate donor substrate, then transferred onto the silicon PIC using a micro-transfer printing process that uses an elastomeric stamp to pick up and place micron-scale devices with high precision. The transferred III-V elements are then permanently bonded to the silicon PIC through a subsequent bonding and processing step.
Lumos characterizes its micro-transfer printing process as a form of heterogeneous integration — the III-V gain elements are fabricated separately from the silicon PIC and are subsequently integrated through a bonding process. Lumos argues that this falls squarely within the "heterogeneous integration approaches" that OWS expressly disclaimed during prosecution.
Lumos's proposed construction of "monolithically formed" reinforces this argument. Lumos proposes the following construction: "a WDM component formed during front-end-of-line processing within the same silicon substrate, excluding hybrid integration, flip-chip bonding, or any heterogeneous integration technique." This construction incorporates the prosecution disclaimer directly into the claim construction, effectively reading the applicant's prosecution remarks as a limitation on the claim scope.
Lumos contends that OWS cannot use DOE to capture Lumos's heterogeneously integrated design because: (1) the amendment adding "monolithically formed" triggers the Festo presumption of surrender, and (2) the explicit prosecution disclaimer of "heterogeneous integration approaches" independently bars DOE through argument-based estoppel. Lumos argues that no Festo exception can overcome the categorical prosecution disclaimer.
C. Rebuttal Arguments
We have analyzed the three Festo exceptions as applied to the '678 Patent. As discussed below, the rebuttal position is considerably weaker than for the '567 Patent.
1. Tangential Relation Exception. The tangential relation argument for the '678 Patent proceeds as follows: the prior art references that prompted the "monolithically formed" amendment were directed to macro-scale hybrid integration approaches — specifically, techniques involving the attachment of separately fabricated optical modules or gain elements to a silicon substrate through conventional assembly techniques such as flip-chip bonding, wire bonding, or die attach. These prior art approaches involve the assembly of macroscopic components (on the order of millimeters in scale) using relatively imprecise alignment techniques. By contrast, Lumos's micro-transfer printing process involves the transfer of micron-scale III-V semiconductor elements with sub-micron alignment precision, followed by wafer-level bonding and processing steps that integrate the transferred elements into the silicon PIC in a manner that approaches monolithic integration in terms of alignment precision, thermal coupling, and process integration.
The argument is that micro-transfer printing represents a fundamentally different technology from the macro-scale hybrid integration approaches disclosed in the cited prior art references, and that the distinction between conventional hybrid integration (prior art) and micro-transfer printing (Lumos's design) bears only a tangential relation to the reason for the amendment (distinguishing conventional hybrid integration from monolithic integration).
However, this argument faces a significant obstacle: the applicant's prosecution remarks did not disclaim only the specific hybrid integration techniques disclosed in the cited references. Instead, the applicant disclaimed "heterogeneous integration approaches" as a category. Micro-transfer printing is generally classified within the silicon photonics literature as a form of heterogeneous integration — it involves the integration of III-V semiconductor materials (which cannot be grown on silicon) onto a silicon substrate through a bonding process. While micro-transfer printing differs in scale, precision, and process integration from conventional flip-chip bonding, it is nonetheless a heterogeneous integration technique in the taxonomic sense.
Strength assessment: Weak. The tangential relation argument is difficult to sustain when the prosecution remarks explicitly disclaimed "heterogeneous integration" as a category, and the accused equivalent (micro-transfer printing) is generally classified within that category. The argument depends on the Court drawing a distinction between micro-transfer printing and "heterogeneous integration" that the technical community generally does not draw. While there may be a factual basis for arguing that micro-transfer printing is sufficiently different from conventional hybrid integration to support the tangential relation exception, the categorical nature of the prosecution disclaimer significantly undermines this argument.
2. Unforeseeable Equivalent Exception. The unforeseeable equivalent argument for the '678 Patent is that micro-transfer printing of III-V gain elements onto silicon photonic substrates was an emerging technology at the time of the amendment (approximately the 2019–2020 timeframe, during prosecution) and was not foreseen as a practical commercial alternative to monolithic integration. Lumos's LumosBridge 800G product, which employs micro-transfer printing, did not ship commercially until Q1 2022, well after the amendment was made. At the time of the amendment, micro-transfer printing for photonics applications was primarily a research-stage technology, and its commercial viability had not been demonstrated.
However, as with the '567 Patent, this argument is weakened by the existence of prior academic publications. Academic publications on micro-transfer printing for photonics applications existed before 2019. Researchers at multiple institutions had published on the use of micro-transfer printing to integrate III-V semiconductor devices onto silicon photonic platforms, and the concept was known to persons of ordinary skill in the art. While the technology was not yet commercially deployed, it was a known research concept that a skilled practitioner would have been aware of at the time of the amendment.
Strength assessment: Weak to moderate. The argument faces the same challenge as the unforeseeability argument for the '567 Patent: the concept was in the literature, even if commercial implementation came later. The Federal Circuit's foreseeability standard focuses on whether the equivalent was "known in the relevant art," and the existence of published research on micro-transfer printing for photonics undermines the argument that the equivalent was unforeseeable.
3. Inexpressible Substitute Exception. The inexpressible substitute argument for the '678 Patent is that the applicant could not reasonably have been expected to draft claim language that would distinguish between monolithic integration and micro-transfer printing while still overcoming the hybrid integration prior art. At the time of the amendment, the terminology for distinguishing among various integration approaches in silicon photonics was evolving, and the boundary between "monolithic" and "heterogeneous" integration was not precisely defined. The applicant's use of "monolithically formed" was a reasonable drafting choice that reflected the state of the art and the accepted terminology.
Strength assessment: Weak. The Court is unlikely to be sympathetic to this argument given the applicant's broad, categorical disclaimer of "heterogeneous integration approaches." If the applicant had intended to leave open the possibility of claiming equivalents involving heterogeneous integration techniques, the applicant could have refrained from making such a sweeping prosecution disclaimer. The inexpressible substitute exception is designed to address situations in which the patentee genuinely could not have drafted claim language to encompass the equivalent — not situations in which the patentee affirmatively disclaimed the category of technology to which the equivalent belongs. The applicant's categorical disclaimer effectively foreclosed the ability to argue that the substitute was "inexpressible."
D. Argument-Based Estoppel — Additional Challenge for '678 Patent
Separate from and independent of the amendment-based estoppel analysis under Festo, the explicit prosecution statement disclaiming "heterogeneous integration approaches" creates argument-based estoppel that poses an additional — and potentially dispositive — obstacle to OWS's DOE theory for the '678 Patent.
Under the Federal Circuit's prosecution disclaimer doctrine, as articulated in Ridgeway & Co. v. Conceptronic, Inc., 90 F.3d 1576 (Fed. Cir. 1996), and its progeny, clear and unmistakable statements made during prosecution that disclaim certain subject matter from the scope of the claims operate as a binding limitation on claim scope. See also Omega Eng'g, Inc. v. Raytek Corp., 334 F.3d 1314, 1323–24 (Fed. Cir. 2003) (prosecution disclaimer applies when the patentee "unambiguously disclaimed a particular interpretation during prosecution"). Prosecution disclaimers limit the scope of claims not only for purposes of literal infringement but also for purposes of DOE. See Pharmacia & Upjohn Co. v. Mylan Pharm., Inc., 170 F.3d 1373, 1379 (Fed. Cir. 1999).
The applicant's statement — "the present invention is distinguished from heterogeneous integration approaches" — is a categorical disclaimer. It does not limit the disclaimer to the specific prior art references cited in the examiner's rejection. It does not qualify the disclaimer by reference to particular types of heterogeneous integration (e.g., flip-chip bonding, die attach). It categorically distinguishes the claimed invention from "heterogeneous integration approaches" as a class.
This is the most significant obstacle for DOE on the '678 Patent. Unlike amendment-based estoppel, argument-based estoppel is not subject to the three Festo rebuttal exceptions. The Festo framework applies only to amendment-based estoppel; argument-based estoppel operates under the separate prosecution disclaimer doctrine, which asks simply whether the patentee made a "clear and unmistakable" disclaimer. If the disclaimer is found, it bars DOE without any opportunity for rebuttal under the Festo exceptions. See Omega Eng'g, 334 F.3d at 1325–26.
The argument-based estoppel may be dispositive regardless of whether the Festo amendment-based estoppel exceptions apply. Even if OWS could successfully argue that one of the Festo exceptions rebuts the presumption of surrender arising from the amendment, the independent argument-based estoppel from the prosecution disclaimer would still bar DOE. OWS would need to overcome both the amendment-based estoppel and the argument-based estoppel to pursue DOE for the '678 Patent, and the argument-based estoppel has no rebuttal mechanism analogous to the Festo exceptions.
Assessment: This is a formidable challenge. The prosecution disclaimer is clear, categorical, and directly on point. Micro-transfer printing is a species of heterogeneous integration, and the disclaimer encompasses it. While we will argue in our Reply Brief that the disclaimer was directed at the specific prior art approaches and should not be read as a blanket exclusion of all heterogeneous integration technologies (see Section VI.A below), this argument faces long odds given the categorical language of the disclaimer.
Recommendation: We should acknowledge this risk candidly in internal strategy discussions. The '678 Patent's DOE theory is substantially weaker than the '567 Patent's, and the team should plan the overall infringement strategy accordingly.
E. Impact Assessment for '678 Patent DOE Theory
The combination of amendment-based estoppel and argument-based estoppel makes DOE extremely difficult for the '678 Patent. The amendment-based estoppel triggers the Festo presumption of surrender, and the rebuttal arguments under the three Festo exceptions are, at best, weak. The argument-based estoppel from the prosecution disclaimer operates independently and may be dispositive, with no rebuttal mechanism available.
The DOE challenge is compounded by the literal infringement challenge. If the Court adopts Lumos's proposed narrow construction of "monolithically formed" — "a WDM component formed during front-end-of-line processing within the same silicon substrate, excluding hybrid integration, flip-chip bonding, or any heterogeneous integration technique" — literal infringement is also precluded. Lumos's micro-transfer-printed III-V gain elements are not fabricated within the same front-end-of-line/back-end-of-line processing sequence as the silicon PIC; they are fabricated on a separate donor substrate and subsequently transferred. Under Lumos's proposed construction, this approach falls outside the scope of "monolithically formed" as a matter of literal claim scope.
The '678 Patent is the most vulnerable of the four asserted patents from an estoppel perspective. While we should continue to argue for a broader construction of "monolithically formed" and preserve the DOE argument for the record, the team should realistically assess the likelihood of success on this patent and consider whether the overall infringement case should be structured to emphasize the other asserted patents, particularly the '567 Patent and the '234 Patent, which present stronger infringement positions.
V. COMPARATIVE ANALYSIS — STRENGTH OF REBUTTAL BY PATENT
A. Summary Table
The following table summarizes the PHE analysis for the two patents at issue:
B. Strategic Implications
Several strategic implications emerge from this comparative analysis.
First, the '567 Patent offers the best opportunity for a successful PHE rebuttal. The tangential relation exception is well-supported by the prosecution record and the fundamental technological distinction between the Tanaka butt-coupling prior art and Lumos's segmented taper design. The absence of any categorical prosecution disclaimer in the '567 Patent file history means that OWS faces only amendment-based estoppel, which is subject to the full Festo rebuttal framework. If OWS can establish that the tangential relation exception applies — which we believe is a credible argument — DOE remains available for the '567 Patent.
Second, the '678 Patent presents the weakest PHE rebuttal position of the two patents analyzed. The compound effect of amendment-based estoppel and a categorical prosecution disclaimer creates a formidable barrier to DOE. The best Festo exception for the '678 Patent (unforeseeability) is itself weak, and even if it were successful, it would not overcome the independent argument-based estoppel arising from the prosecution disclaimer. The '678 Patent should not be viewed as a reliable infringement vehicle if the case depends on DOE.
Third, for claim construction strategy, OWS's primary focus should be on arguing for broader constructions that capture Lumos's accused designs under literal infringement, thereby reducing or eliminating reliance on DOE. This is particularly critical for the '678 Patent, where the DOE path is effectively foreclosed, and the literal infringement path depends on achieving a construction of "monolithically formed" that encompasses Lumos's micro-transfer-printed design. For the '567 Patent, the broader construction of "adiabatically coupled" ("coupled such that optical mode transitions occur without significant power loss") would similarly capture Lumos's segmented taper design within the literal claim scope.
Fourth, if literal infringement is not available under the Court's constructions, the DOE-plus-PHE-rebuttal path is a viable fallback for the '567 Patent but not a reliable path for the '678 Patent. The infringement case should be structured accordingly, with the '567 Patent positioned as the stronger of the two patents for both literal infringement and DOE.
VI. RECOMMENDATIONS FOR CLAIM CONSTRUCTION BRIEFING
A. Reply Brief Strategy (Due January 9, 2023)
Based on the analysis above, the following approach is recommended for the Reply Claim Construction Brief.
For the '567 Patent:
• Lead with the tangential relation exception. Emphasize that Lumos's PHE argument is built on a false premise — namely, that the amendment adding "adiabatically coupled" was directed at distinguishing all non-adiabatic coupling approaches, including segmented tapers. The prosecution record demonstrates otherwise. The amendment was made to distinguish the Tanaka prior art (U.S. 9,456,012), which disclosed butt-coupling and directional coupling — coupling mechanisms that involve no taper of any kind. The examiner's rejection was based on Tanaka's butt-coupling disclosure, and the applicant's responsive remarks were exclusively directed to the absence of any taper-based coupling in Tanaka. The distinction between a continuous taper and a segmented taper was simply not at issue during prosecution.
• Cite the prosecution history record extensively. The Reply Brief should include detailed quotations from the examiner's rejection and the applicant's responsive remarks to demonstrate that the rationale for the amendment was focused on butt-coupling, not on taper geometry. These quotations should be drawn from Appendix A and presented in a manner that makes the tangential relation argument self-evident from the record.
• Present the tangential relation exception as the primary rebuttal argument. Under Festo and Honeywell, the tangential relation exception applies when the rationale for the amendment was only peripherally related to the accused equivalent. Here, the rationale for the amendment (distinguishing butt-coupling) is qualitatively remote from the accused equivalent (segmented taper coupling), supporting the exception.
• Use the unforeseeable equivalent and inexpressible substitute exceptions as secondary, supporting arguments. While neither is independently strong, both contribute to the overall picture of a patentee who could not reasonably have been expected to surrender segmented taper designs when amending claims to overcome butt-coupling prior art.
• Argue in the alternative for a broader literal construction. Even if the Court is not persuaded by the PHE rebuttal, OWS should argue that the Court should adopt a broader literal construction of "adiabatically coupled" — "coupled such that optical mode transitions occur without significant power loss" — that would capture Lumos's segmented taper design within the literal claim scope, thereby mooting the PHE issue. Emphasize that Lumos's proposed narrow construction ("coupled through a gradual, continuous taper that prevents excitation of higher-order modes throughout the entire transition region") imports limitations from the specification and is inconsistent with the plain and ordinary meaning of "adiabatic" in the art, which refers to the absence of significant power loss during mode transition, not to a specific taper geometry.
For the '678 Patent:
• Address the prosecution disclaimer directly, but with realistic expectations. The Reply Brief should argue that the applicant's disclaimer was directed to the specific hybrid integration approaches disclosed in the cited prior art references — macro-scale assembly techniques such as flip-chip bonding and die attach — and should not be read as a blanket exclusion of all conceivable heterogeneous integration technologies, including micro-transfer printing, which was an emerging technology not contemplated by the applicant's remarks. This

PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT

INTERNAL MEMORANDUM

C&H MATTER NO.: 1014-00003

I. EXECUTIVE SUMMARY

VI. RECOMMENDATIONS FOR CLAIM CONSTRUCTION BRIEFING

Record

Parties

Dates

Amounts