ATTORNEY WORK PRODUCT — PRIVILEGED AND CONFIDENTIAL
CALDERWOOD & HARKNESS LLP intellectual Property Group
INTERNAL MEMORANDUM
FROM: Omar Siddiqui, Lead Associate, intellectual Property Group
TO: Alan Ngo, Responsible Partner, intellectual Property Group
CC: Christy Bauer, Working Associate; Marcus Delgado, Billing Partner
DATE: September 19, 2022
RE: Claim Construction discovery Plan — Optiwave Semiconductor Corp. v. Lumos Photonics, Inc., Case No. 5:22-cv-03417-EJD (N.D. Cal.)
MATTER NO.: 1014-00003
This memorandum constitutes attorney work product prepared in anticipation of litigation and is protected by the attorney-client privilege, the work product doctrine, and all other applicable privileges. It is intended solely for the use of the individuals to whom it is addressed. Any unauthorized review, use, disclosure, or distribution is strictly prohibited.
Right-click to update Table of Contents
I. Introduction and Purpose
This memorandum sets forth the discovery plan for the claim construction phase of Optiwave Semiconductor Corp. v. Lumos Photonics, Inc., Case No. 5:22-cv-03417-EJD, pending before the Honorable Edward J. Torvelis in the United States District Court for the Northern District of California, San Jose Division. The purpose of this plan is to organize and prioritize the discovery activities necessary to develop and support our client Optiwave Semiconductor Corp.'s ("OWS") claim construction positions in advance of the Markman hearing currently scheduled for January 30, 2023.
By way of procedural background, the Court held the Initial Case Management Conference on August 15, 2022, at which time it set the Markman hearing date and established the case schedule. Consistent with the Court's order and the Patent Local Rules for the Northern District of California, the parties exchanged their initial disclosures under Federal Rule of Civil Procedure 26(a) on August 22, 2022. On September 12, 2022, both sides exchanged their preliminary claim construction contentions pursuant to Patent Local Rule 4-1, which require each party to identify the claim terms it believes require construction and to set forth its proposed constructions.
Several critical deadlines now govern the near-term trajectory of this matter. The responsive claim construction contentions are due on October 3, 2022, pursuant to Patent Local Rule 4-2. The parties must file a Joint Claim Construction and Prehearing Statement no later than October 14, 2022, in accordance with Patent Local Rule 4-3. All claim-construction-related discovery must be completed by October 31, 2022, the Court-imposed close of claim construction discovery. These compressed timelines demand that our team act with considerable urgency in identifying discovery targets, preparing our inventor witnesses, completing prosecution history analyses, and assembling extrinsic evidence.
This litigation involves four patents-in-suit, all owned by OWS:
• U.S. Patent No. 10,234,567 (the "'567 Patent");
• U.S. Patent No. 10,891,234 (the "'234 Patent");
• U.S. Patent No. 11,045,678 (the "'678 Patent"); and
• U.S. Patent No. 11,312,890 (the "'890 Patent").
The Complaint alleges that Lumos Photonics, Inc. ("Lumos") infringes each of these patents through its manufacture, sale, offer for sale, and importation of two accused products: the LumosBridge™ 800G Optical Engine and the LumosEdge™ 400ZR+ Pluggable Transceiver. These products are commercially significant entries in the high-speed optical interconnect market and represent a substantial share of Lumos's photonics revenue.
Opposing counsel in this matter is the firm of Brannick Howell & Stein LLP ("BHS"), which is led by Victoria "Tori" Ashworth, Lead Partner, and Daniel Kowalski, Associate Team Lead. BHS has also engaged Dr. Raymond Ng as Of Counsel and Technical Consultant. Dr. Ng's involvement is particularly noteworthy and is discussed further below in the context of expert evidence and extrinsic evidence preparation.
This memorandum is organized as follows: Section II identifies the key disputed claim terms that will drive claim construction discovery; Section III outlines the deposition plan, including both our affirmative deposition strategy and our defensive preparation for depositions noticed by Lumos; Section IV details the prosecution history review for each patent-in-suit; Section V addresses the identification and compilation of extrinsic evidence; Section VI analyzes the prior art references relevant to claim construction; Section VII discusses preparation for the Markman hearing technical tutorial; Section VIII coordinates claim construction discovery with the concurrent fact discovery process; Section IX addresses budget and resource allocation; Section X provides a comprehensive schedule of key deadlines; Section XI summarizes task assignments; and Section XII identifies open issues and action items requiring prompt resolution.
II. Key Disputed Claim Terms Driving discovery
Based on the preliminary claim construction contentions exchanged on September 12, 2022, and our analysis of the patents-in-suit and the accused products, we have identified four claim terms that will be most consequential to the claim construction phase and that should receive the highest priority in discovery planning. While the Joint Claim Construction and Prehearing Statement will ultimately identify eleven disputed terms across all four patents, the four terms analyzed below are the ones most likely to determine the outcome of the Markman hearing and, in turn, the viability of OWS's infringement theories against Lumos's accused products.
A. '567 Patent — "Adiabatically Coupled"
Independent Claim 1 of the '567 Patent recites, in relevant part, "a tapered silicon waveguide having a tip width of less than 200 nanometers adiabatically coupled to a silicon nitride waveguide layer." The term "adiabatically coupled" is the critical dispute with respect to this patent and will significantly affect whether Lumos's optical engine architecture falls within the scope of the asserted claims.
OWS's Proposed Construction. OWS proposes that "adiabatically coupled" means "coupled such that optical mode transitions occur without significant power loss." This construction is consistent with the ordinary and customary meaning of the term as understood by a person of ordinary skill in the art ("POSITA") in the field of integrated photonics. Adiabatic coupling, as described in the specification at column 4, lines 12–35, refers to the gradual transfer of optical energy between two waveguide structures such that the fundamental optical mode evolves smoothly from one waveguide to the other without exciting higher-order modes or radiation modes, thereby minimizing insertion loss.
Anticipated Lumos Construction. Based on Lumos's preliminary contentions, we anticipate that Lumos will advocate for a narrower reading of "adiabatically coupled" that requires a gradual, continuous taper profile that prevents excitation of higher-order modes throughout the entire transition region. This narrower construction would potentially exclude segmented or stepped taper designs from the scope of the claims.
discovery Focus. discovery on this term must address several key areas. First, we need to develop a thorough understanding of the inventors' conception and use of the term "adiabatically coupled" at the time of filing, which was March 14, 2017. The prosecution history of the '567 Patent is particularly significant here because the "adiabatically coupled" limitation was added during prosecution to overcome a prior art rejection based on Tanaka et al. (U.S. Patent No. 9,456,012, issued September 6, 2016). This amendment creates a prosecution history estoppel risk that must be carefully analyzed. The precise scope of the amendment and the applicant's arguments distinguishing the Tanaka reference will be critical to determining the breadth of the term "adiabatically coupled" as it appears in the issued claims.
A key complicating factor is that Hideo Tanaka is both a named inventor on the '567 Patent and a co-inventor of the cited Tanaka prior art patent (U.S. Patent No. 9,456,012). This dual role creates a sensitive deposition dynamic. Lumos will likely attempt to use Tanaka's testimony to support prosecution history estoppel arguments, and any statements Tanaka makes regarding the differences between the '567 Patent invention and his own prior art disclosure could have significant consequences for the scope of the asserted claims. This issue is addressed in greater detail in Section III below.
B. '234 Patent — "Graded Buffer Layer Having a Continuously Varying Germanium Concentration"
The '234 Patent covers a germanium-on-silicon photodetector architecture and is asserted through Claims 1, 3, 5, 8, and 11. The disputed term "graded buffer layer having a continuously varying germanium concentration" appears in each of the asserted independent claims and goes to the heart of the infringement dispute.
OWS's Proposed Construction. OWS proposes that this term means "a buffer layer in which the germanium concentration changes across the layer thickness." This construction is deliberately broad and is intended to encompass both truly continuous gradients (in which the germanium atomic fraction varies smoothly and without interruption from the silicon substrate interface to the germanium absorber layer) and step-graded approaches (in which the germanium concentration increases in discrete compositional steps across multiple sub-layers).
Anticipated Lumos Construction. Lumos is expected to argue that "continuously varying germanium concentration" requires a monotonically increasing germanium atomic fraction without step changes from the silicon substrate interface to the absorber layer. Under this narrower construction, Lumos's accused design—which uses a step-graded SiGe buffer with four discrete composition steps at germanium fractions of 0.25, 0.50, 0.75, and 1.0—would fall outside the scope of the claims.
discovery Focus. The claim construction fight on this term will center on two categories of evidence. First, we must develop the understanding of the named inventors, Dr. Amara Osei and Dr. Kevin Fong, regarding the meaning of "continuously varying" as used in the claims. Both inventors must be prepared to testify on whether a POSITA would understand "continuously varying" to encompass step-graded buffers or only truly continuous gradients. Second, the specification of the '234 Patent is critical because it describes both "continuously graded" and "step-graded" buffer layer embodiments. However, only the "continuously varying" language appears in the claims themselves. This divergence between the specification and the claim language will be a central battleground. We must be prepared to argue that the broader specification language supports a broad construction of "continuously varying," while Lumos will argue that the choice to use different language in the claims—"continuously varying" rather than "step-graded"—was deliberate and exclusionary. The prosecution history also requires careful review. The '234 Patent was allowed on first office action after an examiner interview, and we must obtain and analyze the examiner interview summary to determine whether any representations were made during that interview that bear on the distinction between continuously graded and step-graded approaches.
C. '678 Patent — "Integrated Wavelength-Division Multiplexing Element Monolithically Formed on a Single Silicon Die"
The '678 Patent covers a WDM-enabled silicon photonic integrated circuit and is asserted through Claims 1, 2, 6, 9, and 14. The disputed term "integrated wavelength-division multiplexing element monolithically formed on a single silicon die" is the pivotal claim construction issue for this patent.
OWS's Proposed Construction. OWS proposes that this term means "a WDM component fabricated as part of the same semiconductor fabrication process on a single die." This construction focuses on the fabrication process rather than the integration methodology and is intended to capture the core inventive concept of integrating WDM functionality directly onto the silicon photonic die rather than using separate, discrete WDM components.
Anticipated Lumos Construction. Lumos is expected to argue that "monolithically formed" excludes hybrid integration, flip-chip bonding, or any heterogeneous integration technique in which components made from different materials or fabricated in separate process flows are assembled onto the silicon die. Under this construction, Lumos would argue that its accused LumosBridge™ 800G platform—which uses micro-transfer-printed III-V gain elements bonded to the silicon photonic integrated circuit ("PIC")—does not infringe because the integration of III-V components constitutes heterogeneous rather than monolithic integration.
discovery Focus. The prosecution history of the '678 Patent is the most significant source of evidence on this term. The limitation "monolithically formed" was added by amendment during prosecution to overcome a prior art rejection based on references disclosing hybrid integration approaches. The applicant's remarks during prosecution explicitly stated: "the present invention is distinguished from heterogeneous integration approaches." This statement, made in the context of narrowing the claims to overcome prior art, creates a significant prosecution history estoppel risk. Lumos will argue that this is a clear and unambiguous surrender of claim scope covering any form of heterogeneous integration, including the micro-transfer-printing technique used in its accused products. We must develop a record, through inventor testimony and extrinsic evidence, that the applicant's remarks were directed to the specific prior art references at issue (which involved traditional hybrid integration techniques such as wire bonding and edge coupling of separately packaged components) and did not constitute a blanket surrender of all heterogeneous integration methodologies. The named inventors on this patent are Dr. Priya Ramanathan, Dr. Amara Osei, and Liwei Zhang, and the testimony of these individuals may be relevant.
D. '890 Patent — "Pre-Emphasis Equalization Stage Electrically Co-Integrated with the Modulator on a Common Substrate"
The '890 Patent covers a modulator driver architecture for silicon photonic Mach-Zehnder modulators and is asserted through Claims 1, 5, 8, and 13. The disputed term "pre-emphasis equalization stage electrically co-integrated with the modulator on a common substrate" raises issues that closely parallel the "monolithically formed" dispute in the '678 Patent.
OWS's Proposed Construction. OWS proposes that this term means "an equalization circuit located on the same substrate as the modulator." This construction focuses on the physical location of the equalization circuit relative to the modulator without imposing specific requirements on the fabrication process.
Anticipated Lumos Construction. Lumos is expected to argue that "co-integrated" means fabricated in the same semiconductor process, sharing doping profiles and process layers, without bond wires, interposers, or package-level routing. Under this construction, the term would exclude architectures in which a separately fabricated driver die is assembled onto or adjacent to the modulator die using any form of packaging or assembly technology.
discovery Focus. The prosecution history of the '890 Patent is again the primary source of evidence. The "co-integrated" limitation was added during prosecution to distinguish separately fabricated and bonded driver circuits described in the prior art. Two office actions were issued during prosecution, and the amendment arguments must be reviewed carefully to assess the scope of any resulting estoppel. This analysis is directly relevant because Lumos's accused LumosEdge™ 400ZR+ Pluggable Transceiver uses a separate CMOS driver die that is flip-chip bonded to the silicon photonic modulator die via copper pillar micro-bumps—a design that falls squarely between OWS's broad proposed construction and Lumos's narrow proposed construction. The named inventors on the '890 Patent are Dr. Kevin Fong and Rajesh Iyer, and the testimony of both will be relevant.
E. Additional Disputed Terms
The Joint Claim Construction Statement, due October 14, 2022, will identify a total of eleven disputed claim terms across all four patents-in-suit. In addition to the four terms discussed above, the parties are expected to dispute construction of terms relating to waveguide geometry, photodetector responsivity thresholds, modulator bandwidth specifications, and other secondary claim elements. While these remaining seven terms warrant attention and must be addressed in our contentions and briefing, they are not expected to be case-dispositive to the same degree as the four primary terms identified above. Accordingly, our discovery plan prioritizes the four critical terms, with the remaining terms to be addressed as needed within the available discovery window. As the responsive contentions and Joint Claim Construction Statement take shape, we may need to reassess these priorities, and I will advise promptly if any of the secondary terms emerge as more significant than currently anticipated.
III. Deposition Plan
A. Depositions of OWS Named Inventors (Preparing Our Witnesses)
Three OWS named inventors are likely to be the principal witnesses for claim construction purposes. We must prepare each of them thoroughly for both our affirmative use of their testimony and for defense against depositions noticed by Lumos. The following outlines our preparation plan for each inventor witness.
1. Dr. Kevin Fong — CTO & Co-Founder
Dr. Kevin Fong is the Chief Technology Officer and Co-Founder of OWS. He is a named inventor on three of the four patents-in-suit: the '567 Patent, the '234 Patent, and the '890 Patent. His role as the technical leader of OWS and his deep involvement in the inventive work underlying three of the four asserted patents makes him the single most important witness in the claim construction phase.
Dr. Fong's deposition should be scheduled for mid-to-late October 2022, well before the October 31 close of claim construction discovery. This timing will allow us to complete preparation and address any follow-up issues before the discovery cutoff.
The following topics must be covered during Dr. Fong's preparation:
• '567 Patent — "Adiabatically Coupled." Dr. Fong must be prepared to testify regarding his understanding of the term "adiabatically coupled" in the context of the '567 Patent and the state of the art at the time of filing (March 14, 2017). This includes his understanding of the physical phenomenon of adiabatic mode coupling, the manner in which the invention disclosed in the '567 Patent achieves adiabatic coupling through the tapered silicon waveguide structure, and how this approach differs from non-adiabatic coupling techniques. Critically, Dr. Fong must also be prepared to discuss the distinction between the '567 Patent's tapered mode converter and the prior art, particularly the Tanaka et al. patent (U.S. Patent No. 9,456,012). Given that the "adiabatically coupled" limitation was added during prosecution specifically to overcome the Tanaka reference, we must prepare Dr. Fong to explain the narrowing amendment without conceding more scope than necessary. He should be coached to frame the amendment as reflecting a distinction in the nature of the coupling mechanism—adiabatic versus non-adiabatic—rather than as a limitation on the specific taper geometry.
• '234 Patent — "Continuously Varying Germanium Concentration." Dr. Fong must be prepared to testify regarding his understanding of "continuously varying" as used in the claims of the '234 Patent. This is a particularly delicate area because the specification describes both "continuously graded" and "step-graded" buffer layer embodiments, yet the claims use only the "continuously varying" language. Dr. Fong should be prepared to explain that, from his perspective as an inventor, "continuously varying" was intended to capture the general concept of a graded buffer layer in which the germanium composition changes across the layer thickness, rather than to exclude step-graded approaches. However, we must also prepare him for aggressive cross-examination on why, if both approaches were intended to be covered, the claims did not use broader language such as "graded" without the modifier "continuously varying."
• '890 Patent — "Co-Integrated." Dr. Fong must be prepared to testify regarding his understanding of "co-integrated" as used in the claims of the '890 Patent. He should be able to articulate the meaning of co-integration in the context of modulator driver architectures, including the distinction between co-integration (where the driver circuitry and the modulator are on the same substrate) and discrete approaches (where the driver is a separate chip connected via bond wires, interposers, or flip-chip bonding). The prosecution history narrows this term somewhat, and Dr. Fong must be prepared to navigate the estoppel issues.
• General Background. Dr. Fong should be prepared to testify regarding his co-founding role at OWS, his technical background (including his Ph.D. in Electrical Engineering with a focus on photonics and his prior role as a principal engineer at a major semiconductor foundry), and his qualifications to speak to the state of the art in silicon photonics during the relevant time period.
I recommend two preparation sessions of approximately two to three hours each. The first session should focus on prosecution history walkthrough and the specific claim terms, while the second should focus on anticipated cross-examination and difficult areas. Christy Bauer should prepare the first draft of the deposition outline for Dr. Fong's testimony.
2. Hideo Tanaka — Former OWS Engineer (2016–2019)
Hideo Tanaka is a named inventor on the '567 Patent only. His testimony presents the most complex and sensitive issues in the claim construction phase because he is also the inventor of the Tanaka prior art patent (U.S. Patent No. 9,456,012, issued September 6, 2016), which was cited by Lumos against the '567 Patent under 35 U.S.C. § 102(a)(1) and which prompted the narrowing amendment that added the "adiabatically coupled" limitation to the claims.
Tanaka's prior art patent was invented before he joined OWS. He subsequently joined OWS in 2016 and contributed to the '567 Patent invention while employed at OWS, before departing the company in 2019. His current employment and affiliation are not yet confirmed. Because Tanaka is no longer an OWS employee, we cannot compel his cooperation, and we must determine whether he will appear voluntarily for preparation and deposition or whether a subpoena will be required.
Lumos will almost certainly seek to depose Tanaka and will use his testimony to support its prosecution history estoppel arguments. Specifically, Lumos will argue that the '567 Patent claims were narrowed during prosecution to add the "adiabatically coupled" limitation to overcome Tanaka's own prior art disclosure, and that Tanaka—as the inventor of both the prior art and the narrowed claims—is uniquely positioned to testify about the scope of the distinction between the two inventions. If Tanaka testifies that his prior art patent also disclosed adiabatic coupling, or that the distinction between the '567 Patent and his prior art is narrower than OWS's proposed construction suggests, such testimony could be devastating.
Tanaka's deposition should be scheduled for mid-October 2022. The following topics must be covered during preparation:
• Contributions to the '567 Patent. Tanaka must be prepared to distinguish his contributions to the '567 Patent invention from his earlier, independent work that gave rise to U.S. Patent No. 9,456,012. We must establish a clear narrative regarding the inventive contributions that each patent represents and the technical distinctions between them.
• Understanding of "Adiabatically Coupled." Tanaka must be prepared to testify regarding his understanding of the term "adiabatically coupled" and how the coupling mechanism disclosed in the '567 Patent differs from the coupling structures disclosed in his prior art patent. This testimony should be consistent with OWS's proposed construction of the term.
• Prosecution History Amendments. Tanaka must be prepared to address the prosecution history amendments that added the "adiabatically coupled" limitation. We must determine the extent of his involvement, if any, in providing technical input to patent counsel during prosecution. If Tanaka was not involved in the prosecution (which is the more favorable scenario), we can minimize his exposure on estoppel issues. If he was involved, we must understand what input he provided and whether any of it is inconsistent with OWS's current claim construction positions.
• Tenure at OWS and Current Affiliation. We must confirm Tanaka's dates of employment at OWS (2016–2019) and his current affiliation and location. This information is necessary to arrange either a voluntary appearance or subpoena service, and to assess whether Tanaka has any current adverse interests (for example, if he is now employed by Lumos or a Lumos competitor).
I will prepare the deposition outline for Tanaka's testimony. We must coordinate with Ted Marchetti (OWS Vice President of intellectual Property) and Sandra Ellison (OWS General Counsel) on the logistics of contacting Tanaka. This coordination should begin immediately given the compressed timeline.
3. Dr. Amara Osei — Principal Engineer
Dr. Amara Osei is a Principal Engineer at OWS and a named inventor on the '234 Patent and the '678 Patent. Her deposition should be scheduled for late October 2022.
Dr. Osei's preparation must cover the following topics:
• '234 Patent — "Continuously Varying Germanium Concentration." This is the critical term for the '234 Patent. Dr. Osei must be prepared to testify regarding her understanding of "graded buffer layer having a continuously varying germanium concentration." She must address whether, in her view as an inventor and as a POSITA in the field of germanium-on-silicon photodetectors, the term "continuously varying" would be understood to encompass step-graded buffer layers (in which the germanium concentration increases in discrete steps) or only truly continuous gradients (in which the germanium concentration varies smoothly and without interruption). Her testimony on this point should be consistent with OWS's proposed construction.
• '678 Patent — "Monolithically Formed on a Single Silicon Die." Dr. Osei must be prepared to testify regarding her understanding of "monolithically formed on a single silicon die" as used in the '678 Patent. This includes the distinction between monolithic integration (in which all components are fabricated on the same substrate using a single process flow) and heterogeneous integration (in which components fabricated on different substrates or in different process flows are assembled together). She must be prepared to address the state of the art in silicon photonic integration circa 2017–2018, including the various integration techniques that were known and used in the industry at the time of filing (June 22, 2018).
• State of the Art. Dr. Osei should be prepared to provide context on the state of the art in germanium-on-silicon photodetectors and silicon photonic integration during the relevant time period of 2017–2018. This background testimony may be useful both for claim construction purposes and for preparing the technical tutorial.
Christy Bauer should prepare the deposition outline for Dr. Osei's testimony.
B. Anticipated Depositions by Lumos (Defending Our Witnesses)
We fully expect Lumos, through BHS, to notice depositions of Dr. Fong, Tanaka, and Dr. Osei for claim construction purposes. This expectation is based on the central role these witnesses play as named inventors on the patents-in-suit and the significance of their testimony to the disputed claim terms. We should coordinate the scheduling of these depositions to minimize disruption to our witnesses' professional responsibilities and to align the timing with our own preparation schedule. Where possible, we should attempt to schedule our affirmative preparation sessions before the Lumos-noticed depositions, so that our witnesses are fully prepared before sitting for examination by opposing counsel.
In addition to the three named inventor witnesses, it is also possible that Lumos may seek to depose Ted Marchetti, OWS's Vice President of intellectual Property. Mr. Marchetti is a registered patent attorney and a former USPTO Patent Examiner who served in Art Unit 2828. If Marchetti was involved in the prosecution of any of the patents-in-suit—either as the prosecuting attorney or as an internal advisor to outside prosecution counsel—his testimony could implicate attorney-client privilege and work product protections. We must evaluate the scope of Marchetti's involvement in prosecution and prepare privilege and work product objections as appropriate. If Lumos notices Marchetti's deposition, we should be prepared to move for a protective order to the extent that the deposition seeks to invade privileged communications regarding prosecution strategy, claim drafting decisions, or legal advice concerning the scope of the claims.
C. Depositions of Lumos Witnesses
We must evaluate whether to notice any depositions of Lumos personnel for claim construction purposes. Two potential deposition targets have been identified:
• Dr. Nikolai Petrov, Vice President of Engineering, Photonics Division. Dr. Petrov is responsible for the engineering team that developed the accused LumosBridge™ 800G and LumosEdge™ 400ZR+ products. His testimony could be relevant to understanding the technical implementation details of the accused products in the context of the disputed claim terms—for example, whether Lumos characterizes its micro-transfer-printing process as a form of "monolithic" or "heterogeneous" integration, or how Lumos describes the coupling mechanism in its optical engine.
• Dr. Mei-Lin Chen, Chief Technology Officer. Dr. Chen is Lumos's CTO and holds fourteen patents in her own name. She received her Ph.D. in Applied Physics from MIT. Her testimony could be relevant to the state of the art and Lumos's understanding of the disputed claim terms from the perspective of a POSITA. Dr. Chen's academic credentials and publication record may make her a credible witness on the ordinary meaning of technical terms, but her testimony could also support OWS if she acknowledges broad usage of the disputed terms in the field.
Recommendation. At this stage, I recommend against noticing depositions of Lumos witnesses for claim construction purposes. The Markman hearing is principally an exercise in interpreting intrinsic evidence—the claim language, specification, and prosecution history—supplemented by extrinsic evidence such as expert declarations, technical dictionaries, and treatises. A defensive posture that relies primarily on intrinsic evidence, bolstered by our expert Dr. Whitfield's declaration, is the most appropriate strategy for the Markman hearing. Deposing Lumos's engineers at this stage carries the risk of generating testimony that could be used by Lumos to support narrow constructions. We should defer Lumos depositions to the fact discovery phase unless a clear strategic need emerges from the responsive claim construction contentions. I will reassess this recommendation after reviewing Lumos's responsive contentions, which are due October 3, 2022.
IV. Prosecution History Review
A. Scope of Review
A thorough review of the complete file histories for all four patents-in-suit is essential to the claim construction phase. The prosecution history is the most important intrinsic evidence source after the claims and specification themselves, and under Phillips v. AWH Corp., 415 F.3d 1303 (Fed. Cir. 2005) (en banc), it constitutes part of the public record of claim scope upon which both the patentee and the public are entitled to rely. Moreover, under Festo Corp. v. Fenmore & Co., 535 U.S. 722 (2002), any narrowing amendments made during prosecution create a presumption that the patentee surrendered the territory between the original claim scope and the amended claim scope, a presumption that can only be overcome in limited circumstances. Several of the patents-in-suit underwent narrowing amendments, making the prosecution history review particularly critical.
The focus areas for each patent are detailed below.
B. '567 Patent (Filed March 14, 2017; Issued March 19, 2019)
The '567 Patent underwent two office action rejections during prosecution before the claims were allowed. The central event in the prosecution history is the narrowing of the claims to include the "adiabatically coupled" limitation in order to overcome the Tanaka et al. prior art reference (U.S. Patent No. 9,456,012, issued September 6, 2016). This amendment was made in response to a rejection of the original claims under 35 U.S.C. § 102(a)(1), which found that the Tanaka reference anticipated the claimed waveguide coupler structure.
We must compile all applicant arguments and amendments that were made to distinguish the Tanaka reference. This compilation should include the full text of each office action, each applicant response, each set of claim amendments, and any examiner interview summaries. We must also identify any arguments made in the applicant's remarks that characterize the scope of the "adiabatically coupled" limitation, particularly any statements that could be construed as narrowing the term beyond OWS's current proposed construction.
The prosecution history estoppel analysis under Festo is critical for this patent. Lumos will argue that OWS's narrowing amendment during prosecution—adding the "adiabatically coupled" limitation to overcome Tanaka—estops OWS from asserting infringement under the doctrine of equivalents against Lumos's segmented (non-continuous) taper design. If the Court adopts a narrow construction of "adiabatically coupled" that excludes segmented tapers from the literal scope of the claims, the doctrine of equivalents may be OWS's only remaining theory of infringement under this patent, and the estoppel analysis will determine whether that theory is viable.
Christy Bauer should prepare a detailed prosecution history timeline and estoppel analysis memorandum as a separate deliverable. This memorandum should address each of the three Festo exceptions to the presumption of prosecution history estoppel: (1) whether the equivalent was unforeseeable at the time of the narrowing amendment; (2) whether the rationale underlying the narrowing amendment bears only a tangential relation to the equivalent in question; and (3) whether there is some other reason suggesting that the patentee could not reasonably have been expected to have drafted a claim that literally encompassed the equivalent.
C. '234 Patent (Filed September 8, 2017; Issued January 14, 2021)
The '234 Patent has a relatively clean prosecution history. It was allowed on the first office action after an examiner interview. No narrowing amendments were required, and the claims as issued are substantially similar to the claims as originally filed. This reduces the prosecution history estoppel risk but increases the importance of specification-based claim construction arguments—because there is no significant prosecution history to constrain the construction, the specification and its description of various embodiments will play a more prominent role.
The most significant task for this patent is to obtain and review the examiner interview summary. Examiner interview summaries are part of the public file history and are available through the USPTO's Patent Application Information Retrieval (PAIR) system. We must determine whether the applicant made any representations during the examiner interview that distinguish step-graded buffer layers from continuously graded buffer layers, or that otherwise bear on the construction of "continuously varying germanium concentration." If the applicant made no such representations during the interview, that silence may be helpful in arguing that the term was not intended to exclude step-graded approaches.
Additionally, we must carefully analyze the specification's treatment of "continuously graded" and "step-graded" embodiments. The fact that the specification describes both approaches but the claims use only "continuously varying" language is a double-edged sword. On one hand, we can argue under Phillips that the specification's disclosure of step-graded embodiments supports a broad construction of "continuously varying" that encompasses both approaches. On the other hand, Lumos will invoke the doctrine of claim differentiation and the principle that the patentee is presumed to have chosen its claim language deliberately, arguing that the use of "continuously varying" rather than "graded" or "varying" reflects an intentional choice to limit the claims to truly continuous gradients.
D. '678 Patent (Filed June 22, 2018; Issued June 29, 2021)
The '678 Patent has the most extensive prosecution history of the four patents-in-suit, having been subject to three office actions before allowance. The central event in the prosecution history is the addition of the "monolithically formed" limitation by amendment to overcome prior art rejections based on references disclosing hybrid integration approaches.
The applicant's remarks during prosecution are particularly significant. In responding to the prior art rejection, the applicant explicitly stated: "the present invention is distinguished from heterogeneous integration approaches." This statement constitutes a clear and unambiguous characterization of the scope of the claims, and Lumos will argue that it operates as a prosecution history disclaimer—that is, that the applicant disclaimed any claim scope that covers heterogeneous integration techniques. If the Court agrees, the accused LumosBridge™ 800G platform, which uses micro-transfer-printed III-V gain elements bonded to the silicon PIC, may fall outside the scope of the claims regardless of how "monolithically formed" is ultimately construed.
We must gather all three office action responses, all amendments, and all applicant remarks for thorough analysis. Our goal should be to develop arguments that the applicant's statement about heterogeneous integration was directed specifically to the prior art references at issue—which may have involved traditional hybrid integration techniques qualitatively different from micro-transfer-printing—rather than constituting a blanket surrender of all forms of heterogeneous integration. The success of this argument will depend on the specific prior art references that prompted the rejection and the precision (or imprecision) of the applicant's remarks.
E. '890 Patent (Filed November 15, 2018; Issued April 26, 2022)
The '890 Patent underwent two office actions during prosecution. The "co-integrated" limitation was added by amendment to distinguish driver circuits that are separately fabricated and bonded to the modulator, as described in the prior art. The amendment arguments must be reviewed to assess the scope of any resulting estoppel. The estoppel analysis will closely parallel the analysis for the '678 Patent's "monolithically formed" limitation, and the two patents should be analyzed in tandem to ensure consistency in our claim construction positions across the portfolio.
A procedural note: the '890 Patent issued very recently, on April 26, 2022, just weeks before the Complaint was filed on May 16, 2022. We must confirm that the complete file wrapper is available from the USPTO through the PAIR system, as there can sometimes be a delay in the availability of the full prosecution record for recently issued patents. If the file wrapper is not yet fully available, we should request a certified copy of the complete file history from the USPTO.
F. Assignments and Timeline
The prosecution history review assignments are as follows:
• Omar Siddiqui: Lead review of the '567 Patent and '890 Patent prosecution histories. These patents present the most complex estoppel issues and involve the most sensitive deposition witnesses.
• Christy Bauer: Lead review of the '234 Patent and '678 Patent prosecution histories. The '234 Patent review will focus on the examiner interview summary and specification-based arguments, while the '678 Patent review will focus on the applicant's remarks regarding heterogeneous integration.
All prosecution history summaries should be completed by October 10, 2022. However, because the responsive claim construction contentions are due earlier—on October 3, 2022—the summaries should be substantially drafted before that date in order to inform the responsive contentions. Additionally, the summaries will feed directly into the Joint Claim Construction Statement, which is due October 14, 2022, and which must identify all disputed terms and each party's proposed constructions and supporting intrinsic evidence.
V. Extrinsic Evidence Identification
While intrinsic evidence—the claims, specification, and prosecution history—is the primary source for claim construction, extrinsic evidence plays an important supplementary role, particularly when intrinsic evidence alone does not resolve the meaning of a disputed claim term. Under Phillips, 415 F.3d at 1317–18, courts may consider extrinsic evidence such as expert testimony, inventor testimony, dictionaries, and learned treatises, though such evidence may not be used to contradict the claim meaning established by the intrinsic record. The following categories of extrinsic evidence should be identified and compiled in support of OWS's claim construction positions.
A. Technical Dictionaries and Treatises
For each of the four primary disputed terms, we must identify and compile relevant definitions from standard reference works in the silicon photonics, integrated optics, and semiconductor fabrication fields. The specific terms and their corresponding reference categories are as follows:
• "Adiabatically coupled." We should consult photonics and optical engineering dictionaries, including the Encyclopedia of Laser Physics and Technology, the Handbook of Optics (published by the Optical Society of America), and textbooks on integrated optics and mode coupling theory such as Integrated Photonics: Fundamentals by Pollock and Lipson, and Photonic Devices by Liu. The goal is to establish that the term "adiabatic" in the context of optical waveguide coupling refers broadly to a coupling mechanism that preserves the fundamental mode and minimizes power loss, without requiring any specific taper geometry.
• "Continuously varying." We should consult materials science references on graded buffer layers and semiconductor epitaxy textbooks, including Kasper and Lyutovich's Properties of Silicon Germanium and SiGe:Carbon and textbooks on molecular beam epitaxy and chemical vapor deposition processes. The goal is to establish that "continuously varying" in the context of compositional gradients in semiconductor layers is understood in the field to encompass gradual changes in composition, including step-graded approaches with sufficiently fine compositional steps.
• "Monolithically formed" / "monolithic integration." We should consult semiconductor fabrication references that define and distinguish monolithic, hybrid, and heterogeneous integration, including the International Technology Roadmap for Semiconductors (ITRS) and standard textbooks on VLSI fabrication. The goal is to establish that "monolithic" integration, while often contrasted with hybrid integration, is a term whose precise boundaries depend on context and that may encompass certain forms of integration that involve multiple materials but are completed on a single die.
• "Co-integrated" / "co-integration." We should consult circuit design and semiconductor process integration references, including textbooks on mixed-signal IC design and system-on-chip architectures. The goal is to establish that "co-integration" refers broadly to the placement of multiple functional circuits on a common substrate, without necessarily requiring that all circuits share identical doping profiles and process layers.
Christy Bauer should compile an initial list of candidate references for each term by September 30, 2022. This list should include full bibliographic citations and, where readily available, excerpts of the relevant definitions or discussions.
B. Technical Publications and Journal Articles
In addition to reference works, we must identify peer-reviewed publications from the relevant time period—approximately 2016 through 2019, corresponding to the filing dates of the patents-in-suit—that use or define the disputed claim terms. These publications will serve as evidence of how a POSITA understood the terms at the time the patents were filed. Publications authored by prominent researchers in the silicon photonics field, including publications by the inventors themselves, are particularly relevant.
We must also cross-reference the prior art cited by Lumos in its preliminary contentions and identify any publications that bear on the meaning of the disputed terms. Two references warrant specific attention:
• Chen & Liu, "Segmented Taper Coupler Design for Low-Loss Fiber-to-Chip Coupling," *Journal of Photonic Integration*, Vol. 14, No. 3, pp. 217–229 (2016). This publication was cited against the '567 Patent under 35 U.S.C. § 103 and discusses segmented taper coupler designs. Its discussion of segmented versus continuous taper profiles is directly relevant to the construction of "adiabatically coupled." If Chen & Liu characterize segmented taper couplers as a form of adiabatic coupling, this would support OWS's broader construction.
• Iyer & Patel, "Co-Design of Modulator Drivers and Silicon Photonic Mach-Zehnder Structures," *IEEE Transactions on Circuits and Systems*, Vol. 65, No. 8, pp. 2701–2713 (2018). This publication was cited against the '890 Patent under 35 U.S.C. § 103 and discusses the co-design of modulator drivers and silicon photonic Mach-Zehnder structures. Notably, the "Iyer" in this reference is Rajesh Iyer, who is a circuit designer at OWS and a named inventor on the '890 Patent. We must confirm whether this publication describes the same invention disclosed in the '890 Patent or a distinct work. If the publication constitutes a prior art publication of the patented invention, it could create inventorship or prior art exposure issues that must be assessed. In any event, the publication's discussion of co-design and co-integration terminology is relevant to the construction of the "co-integrated" term.
C. Expert Declarations
OWS's Expert — Dr. Harold Whitfield. Dr. Harold Whitfield of the Whitfield Technical Consulting Group, Inc. has been retained as OWS's technical expert for the claim construction phase. Dr. Whitfield's retention agreement was executed in August 2022. His office is located at 530 Lytton Avenue, Suite 210, Palo Alto, California 94301.
Dr. Whitfield is Professor Emeritus of Electrical and Computer Engineering at the University of California, Santa Barbara, where he conducted research in integrated photonics for over thirty years. He holds a Ph.D. in Electrical Engineering (photonics) from the University of California, Berkeley. He has authored or co-authored over two hundred publications in the field of integrated photonics and is widely recognized as a leading authority in silicon photonic device design and fabrication.
Dr. Whitfield will provide a declaration in support of OWS's claim construction positions. The declaration should address the understanding of a POSITA for each of the eleven disputed claim terms identified in the Joint Claim Construction Statement. For the four primary disputed terms identified in Section II above, the declaration should provide detailed analysis of the intrinsic evidence (claims, specification, and prosecution history) and extrinsic evidence (technical dictionaries, treatises, and journal articles) supporting OWS's proposed constructions. For the remaining seven terms, the declaration may be more concise but should still set forth Dr. Whitfield's opinions and their bases.
Dr. Whitfield is also designated to provide a technical tutorial for the Court at the Markman hearing. The tutorial is discussed in greater detail in Section VII below.
A scope memorandum was provided to Dr. Whitfield in August 2022 (see Matter File: Memorandum to Dr. Whitfield re: Scope of Technical Expert Engagement, August 2022). The timeline for the declaration is as follows: the initial draft should be prepared by November 15, 2022, to allow adequate time for review, revision, and finalization before the final declaration accompanies OWS's Opening Claim Construction Brief, which is due December 5, 2022.
Lumos's Expert — Dr. Raymond Ng. Lumos has engaged Dr. Raymond Ng as a technical consultant. Dr. Ng serves as Of Counsel to BHS and is a former professor of photonics at the University of Colorado Boulder. We expect Dr. Ng to provide a competing declaration in support of Lumos's claim construction positions and/or to serve as Lumos's technical tutorial witness at the Markman hearing. We must research Dr. Ng's publication record, any prior expert testimony he has provided in patent cases, and any prior positions he has taken on the disputed claim terms or related technical concepts. This research will help us anticipate Dr. Ng's likely opinions and prepare a targeted rebuttal in Dr. Whitfield's declaration or in our Reply Claim Construction Brief.
D. Industry Standards and Specifications
We should identify any relevant industry standards and specifications that bear on the meaning of the disputed claim terms. Potential sources include:
• IEEE standards relating to optical communications and silicon photonic device design;
• OIF (Optical Internetworking Forum) specifications for high-speed optical interconnects, coherent transceivers, and pluggable transceiver form factors;
• MSA (Multi-Source Agreement) specifications for transceiver modules, including the QSFP-DD and OSFP MSA specifications relevant to the 400ZR+ transceiver form factor used by the accused LumosEdge product.
These standards and specifications may contain definitions or technical descriptions that inform the POSITA's understanding of monolithic integration, WDM on silicon, and transceiver architectures. They may also provide context for understanding the commercial and technical significance of the accused products within the broader industry landscape.
VI. Prior Art Analysis (Claim Construction Relevance)
While prior art bears primarily on invalidity—an issue that will be fully developed during fact discovery, expert discovery, and summary judgment—certain prior art references are relevant to claim construction insofar as they inform the POSITA's understanding of the disputed claim terms. Under Phillips, courts may consider how a term is used in prior art references to illuminate the meaning the term would have conveyed to a POSITA. Additionally, prior art references that were cited and distinguished during prosecution are intrinsic evidence to the extent they form part of the prosecution history record.
The following key prior art references should be analyzed for their claim construction relevance:
• Tanaka et al. (U.S. Patent No. 9,456,012, issued September 6, 2016). This reference was cited against the '567 Patent under 35 U.S.C. § 102(a)(1), and the claims of the '567 Patent were narrowed to overcome it by adding the "adiabatically coupled" limitation. The Tanaka reference's disclosure of waveguide coupler structures directly illuminates the meaning of "adiabatically coupled" because the scope of this term was defined, at least in part, by the need to distinguish the Tanaka reference during prosecution. We must analyze the specific coupler structures disclosed in Tanaka—including the taper geometry, mode coupling mechanism, and loss characteristics—to determine whether the Tanaka reference describes adiabatic or non-adiabatic coupling, and how the '567 Patent's coupling mechanism differs.
• EP 3,234,567 A1 (published August 2, 2017; applicant: Nordlicht Photonische GmbH). This European patent application was cited against the '234 Patent under 35 U.S.C. § 102(a)(2). Its disclosure of germanium buffer layer fabrication techniques is relevant to the meaning of "continuously varying germanium concentration." We must analyze how this reference describes the germanium concentration profile in its buffer layers and whether it uses "continuously varying," "step-graded," or other terminology to characterize the composition gradient.
• Chen & Liu (2016). As discussed in Section V.B above, this publication's discussion of segmented versus continuous taper coupling is relevant to the understanding of "adiabatically coupled." The publication was cited against the '567 Patent under 35 U.S.C. § 103 in combination with the Tanaka reference.
• Iyer & Patel (2018). As discussed in Section V.B above, this publication's discussion of co-design of modulator drivers and silicon photonic structures is relevant to the understanding of "co-integrated." The publication was cited against the '890 Patent under 35 U.S.C. § 103. As noted, the Rajesh Iyer who co-authored this publication is the same Rajesh Iyer who is a named inventor on the '890 Patent, and this relationship requires further investigation.
The prior art analysis for claim construction purposes should be completed in conjunction with the prosecution history review and should be incorporated into the prosecution history summaries due October 10, 2022.
VII. Technical Tutorial Preparation
The Court has indicated that it will permit each side to present a technical tutorial at the outset of the Markman hearing on January 30, 2023. Technical tutorials are a common feature of Markman hearings in patent cases involving complex technology, and they provide the Court with the scientific and engineering background necessary to understand the claim construction disputes. The tutorial is not an advocacy exercise; it should present the underlying technology in a fair, accurate, and accessible manner. However, the framing and emphasis of the tutorial presentation can subtly favor one side's construction over the other's, and the tutorial should be prepared with that reality in mind.
Dr. Harold Whitfield has been designated as OWS's technical tutorial witness. The tutorial presentation should cover the following topics:
1. Silicon Photonics Fundamentals. An introduction to silicon photonic integrated circuits, including the basic building blocks: optical waveguides, directional couplers, photodetectors, and modulators. This section should explain why silicon is used as a platform material (compatibility with CMOS fabrication, cost advantages, scalability) and the challenges associated with integrating active optical functions on silicon.
2. Adiabatic vs. Non-Adiabatic Mode Transitions in Waveguide Couplers. An explanation of the physics of optical mode coupling between waveguides, including the concept of adiabatic evolution of the optical mode along a tapered transition region. This section should explain the conditions under which a transition is considered "adiabatic" (gradual changes that allow the fundamental mode to evolve without exciting higher-order modes or radiation modes) and the practical consequences of non-adiabatic transitions (power loss, crosstalk). The presentation should explain the physics without prejudicing the claim construction dispute by defining the term too narrowly or too broadly.
3. Germanium-on-Silicon Photodetector Architectures. An explanation of why germanium is integrated onto silicon for infrared photodetection (germanium's narrower bandgap allows absorption at telecommunications wavelengths), including the lattice mismatch challenge and the role of buffer layers in managing strain and defect density. This section should describe both continuously graded and step-graded buffer layer approaches, explaining the advantages and tradeoffs of each, without taking a position on whether one is a subset of the other.
4. Monolithic vs. Heterogeneous Integration Techniques in Silicon Photonics. An explanation of the various approaches to integrating non-silicon components (such as III-V semiconductor gain elements) onto silicon photonic platforms. This section should describe monolithic integration (fabrication of all components on a single substrate in a single process flow), hybrid integration (assembly of separately fabricated components using wire bonding, edge coupling, or other packaging techniques), and heterogeneous integration (bonding of dissimilar materials to the silicon substrate, including wafer bonding, die bonding, and micro-transfer-printing). The distinction between these approaches is central to the '678 Patent claim construction dispute and must be presented accurately.
5. Mach-Zehnder Modulator Driver Architectures. An explanation of how silicon photonic Mach-Zehnder modulators work and the role of the driver circuit in providing the electrical signals that modulate the optical signal. This section should describe co-integrated approaches (where the driver circuitry is fabricated on the same substrate as the modulator) and multi-chip approaches (where the driver is a separate chip connected to the modulator via bond wires, interposers, or flip-chip bonding), and should explain the advantages of co-integration (reduced parasitic capacitance, improved signal integrity, smaller form factor) and the challenges (process complexity, performance tradeoffs between CMOS and photonic device optimization).
The timeline for the tutorial presentation is as follows. The first draft of the tutorial slides should be completed by December 15, 2022, to allow sufficient time for review and revision. The slides should be finalized by January 15, 2023, two weeks before the Markman hearing. This timeline must be coordinated with Dr. Whitfield's schedule. His consulting firm, the Whitfield Technical Consulting Group, Inc., is located at 530 Lytton Avenue, Suite 210, Palo Alto, California 94301, and we should arrange in-person or video conference working sessions to develop the tutorial content collaboratively.
VIII. Coordination with Fact discovery
The claim construction phase is proceeding concurrently with the early stages of fact discovery, and certain discovery activities will overlap and inform both tracks.
Fact discovery Timeline. Lumos served its first set of document requests (ninety-eight requests) on November 7, 2022. OWS's first set of document requests to Lumos (one hundred twelve requests) is to be served on November 21, 2022. OWS's first set of interrogatories to Lumos is also to be served on November 21, 2022.
Overlap Areas. Several categories of fact discovery are relevant to both claim construction and the broader infringement analysis:
• Technical Design Documents for Accused Products. Documents relating to the design, engineering, and manufacturing of the accused LumosBridge™ 800G Optical Engine and LumosEdge™ 400ZR+ Pluggable Transceiver are relevant to both claim construction (to understand the accused designs in the context of the disputed claim terms) and to the ultimate infringement analysis. For claim construction purposes, understanding the specific technical implementation of the accused products will help us frame our proposed constructions i