US 9190502

I'll verify this against live sources, focusing on the specific patent number and any 2026 Federal Circuit dockets.

Let me check the second 2026 CAFC docket number and confirm the appeal status.

US Patent 9,190,502 — Analysis

Note on identifiers: All numbers below are transcribed literally as they appear in the source record. "9,190,502" / "9190502" / "US9190502B2" / "'502 patent" all refer to the same document.

1. Bibliographic Data (high confidence — from the patent document itself)

Continuity chain (per the patent's own CROSS-REFERENCE section): 14/515,584 is a continuation of 13/854,319 (filed Apr. 1, 2013, abandoned), which is a continuation of 11/622,496 (filed Jan. 12, 2007, now U.S. Pat. No. 8,421,195), which is a divisional of 10/934,915 (filed Sep. 3, 2004, abandoned). All incorporated by reference.

Ownership/perfection history (from the record):

• 2015-05-13: Assignment from Rao to GREENTHREAD, LLC (recorded).
• 2019-04-30: Security interest to PATENT CAPITAL FUNDING 2018 – SERIES 1-A, LLC.
• 2019-05-07: Security interest to EC HOLDINGS SERIES LLC – SERIES A.
• 8th-year maintenance fee paid (large entity) 2023-03-16 — consistent with the patent having been in force throughout the enforcement campaign, but now expired.

2. Abstract (verbatim, abridged)

"Most semiconductor devices manufactured today, have uniform dopant concentration, either in the lateral or vertical device active (and isolation) regions. By grading the dopant concentration, the performance in various semiconductor devices can be significantly improved… increase in frequency of operation for digital logic, various power MOSFET and IGBT ICS, improvement in refresh time for DRAM's, decrease in programming time for nonvolatile memory, better visual quality including pixel resolution and color sensitivity for imaging ICs, better sensitivity for varactors in tunable filters, higher drive capabilities for JFET's, and a host of other applications."

3. Plain-Language Overview of the Independent Claims

The patent has exactly two independent claims, and they are structural mirror-images of each other. Both recite the same four structural elements:

1. A surface layer;
2. A substrate;
3. An active region containing a source and a drain, sitting on one face of the surface layer;
4. A single drift layer sandwiched between the other face of the surface layer and the substrate, in which the dopant concentration is graded so as to create a first static, unidirectional electric drift field;
5. Plus at least one well region located inside that single drift layer, which is also graded so as to create a second static, unidirectional electric drift field.

Claim 1 — minority carriers swept upward (substrate → surface layer). The two graded-doped regions must produce fields that push minority carriers from the substrate toward the surface. The specification describes this direction as useful, e.g., for accelerating carriers toward the surface when programming non-volatile memory.

Claim 7 — minority carriers swept downward (surface layer → substrate). Identical structure, but the field direction is reversed: carriers are driven from the surface layer down into the substrate. The specification frames this as the "sub-terrain" sweep that removes spurious minority carriers (generated by clock switching in logic/memory, or by photon absorption in imagers) away from sensitive surface nodes — improving DRAM refresh, SRAM soft-error immunity, and CMOS image-sensor pixel/color resolution.

Two features in particular recur in the litigation and IPR record and are the crux of the claims:

• "single drift layer" (numerical/structural limitation — the layer must be singular);
• "graded concentration of dopants generating a … static unidirectional electric drift field to aid the movement of minority carriers" (a functional limitation tied to the direction of carrier movement, which Petitioners have attacked as a mere result of any downward-sloping concentration gradient).

Dependent claims are purely application-designation claims:

• Claims 2 and 8 — the two fields are adapted to their respective dopant gradings to aid carrier movement in the respective active regions.
• Claims 3/9 — device is a CPU; 4/10 — a DRAM; 5/11 — a flash memory; 6/12 — an image sensor.

No method claims and no means-plus-function claims are present.

4. Litigation and PTAB Record for the '502 Patent

The '502 patent is the earliest-issued member of a large Greenthread family that has been the subject of an extensive campaign. The following proceedings are tied specifically to 9,190,502 in the record:

PTAB (inter partes review):

The Board's final written decisions turned on prior art (notably Kawagoe and Payne) and on the construction of "graded concentration of dopants … to aid the movement of minority carriers." Public briefing reflects that the Board sided with petitioners on the ground that a downward-sloping concentration gradient inherently creates the claimed unidirectional drift field (citing ParkeVision for the proposition that the device need only be capable of moving carriers, not shown actually moving them).

District Court complaints asserting the '502 patent (from the patent page's litigation record): W.D. Tex. 6:22-cv-00105 and 6:22-cv-01293; D. Del. 1:23-cv-00326, 1:23-cv-00333, 1:23-cv-00579; E.D. Tex. 2:23-cv-00157 (TI), 2:23-cv-00179, 2:23-cv-00212 (OmniVision); and W.D. Tex. 1:23-cv-00369 (Cirrus Logic).

CAFC 2026 dockets (as returned by search):

5. Uncertainty / Caveats — please read

I am flagging these explicitly rather than papering over them, because several items in the live record are inconsistent and I do not have authoritative confirmation:

1. Date inconsistency. Per your stated current date (April 26, 2026), a CAFC notice of appeal filed 05/20/2026 (case 26-1847) would be in the future. The docket aggregators (exparte.com, RPX) report that filing date. One of the two dates is wrong; I cannot resolve which from the available sources. Treat the 26-1847 filing date as unverified.
2. 26-1652 caption mismatch. The search result identifies 26-1652 as Greenthread v. Cirrus Logic but attributes it to originating case IPR2024-00773, whose petitioner is listed as Texas Instruments on the patent page. Those two facts do not line up (Cirrus Logic is a petitioner in IPR2024-00018, not -00773). I cannot confirm which is correct. Do not rely on the originating-case attribution without checking the PACER docket directly.
3. Appeal outcomes are "—" / pending. Both 26-1652 and 26-1847 show no decision and no assigned panel. No merits ruling on the '502 patent claims from the Federal Circuit is available as of the sources reviewed.
4. Assignee of record vs. current owner. The patent document lists GREENTHREAD, LLC as applicant/assignee, and the recorded security interests (Patent Capital Funding 2018 – Series 1-A, LLC; EC Holdings Series LLC – Series A) are liens, not ownership transfers. Current beneficial ownership is not something I can establish from these records.
5. Patent expired. Anticipated expiration 2024-09-03 and status "Expired – Lifetime." A patent that has expired cannot be infringed going forward, which is relevant context for how the pending 2026 appeals would matter (past damages and the IPR/validity record, rather than ongoing injunctive relief).
6. I did not independently retrieve a USPTO PatentCenter or PACER primary document in this session; the summary above is drawn from the full patent text supplied to me plus the search results cited. Where the two could differ, the supplied patent text controls per your instruction.

6. Family Context (same specification, for cross-reference only)

Continuations sharing the '502 specification and priority date 2004-09-03: US 8,421,195; US 9,647,070; US 9,190,502; US 10,510,842; US 10,734,481; US 11,121,222; US 11,316,014; and publication US 2022/0246725 A1. All list G. R. Mohan Rao as sole inventor and Greenthread, LLC as assignee.

I'll search for litigation involving US patent 9190502 specifically.

Let me dig deeper into the specific district court cases and their status.

Known Litigation Involving U.S. Patent No. 9,190,502 ("the '502 Patent")

Scope note: I searched for the literal identifier 9,190,502 (issued Nov. 17, 2015; "Semiconductor Devices with Graded Dopant Regions"; inventor G.R. Mohan Rao; assignee Greenthread, LLC). Greenthread has asserted the '502 Patent only as part of a six-patent family ('195, '502, '842, '481, '222, '014), so the same cases recur across the family. I have segregated (a) cases I could confirm plead the '502, (b) cases where '502 assertion is probable but unconfirmed, and (c) PTAB/CAFC proceedings involving the '502.

1. District Court Litigation — '502 Confirmed as Asserted

2. Additional Greenthread District Court Cases (patent-in-suit set includes the '502 family; '502 assertion not fully verified by the sources retrieved)

Greenthread's docket records show the following concurrent cases. I flag these because I could not confirm from the retrieved documents that the '502 was specifically pleaded in each:

These entries appear in the litigation data associated with the '502 patent on Google Patents (https://patents.google.com/patent/[US9190502](/patent/US9190502)/en), but I was unable to verify the asserting patent(s) and party identities within the retrieved source material. Treat as unverified.

Possible earlier case (unverified): Greenthread, LLC v. Samsung Electronics Co., Ltd., et al., No. 2:19-cv-00147-JRG (E.D. Tex.) — a claim-construction order (Dkt. 67, Apr. 20, 2020) is repeatedly cited by Greenthread as adopting its constructions. Samsung filed IPR2020-00289 and IPR2020-00290 (both settled), which are conventionally paired with the '195 and '502 patents. I could not confirm the '502 was asserted in the Samsung suit from the retrieved documents.

3. PTAB (Inter Partes Review) Proceedings on the '502 Patent

Key development affecting all of the above: On Apr. 24, 2025, Acting USPTO Director Coke Morgan Stewart vacated PTAB final written decisions and remanded related Greenthread IPRs (IPR2023-01242/-01243/-01244), holding the Board abused its discretion by denying Greenthread privity discovery into petitioner–Intel relationships and erred in refusing to consider district-court claim-construction orders (legal authority, not "evidence"). This supplier–customer/§315(b) time-bar issue is the through-line in the MPS, TI, and onsemi '502 proceedings.

4. Court of Appeals for the Federal Circuit

Two Greenthread-related appeals appear in the docket data with 2026 docket numbers:

• CAFC No. 26-1652
• CAFC No. 26-1847

Both are recorded as "US case filed in Court of Appeals for the Federal Circuit" against the Greenthread family. I could not confirm from the retrieved sources which specific patent or which underlying PTAB/district-court judgment each appeal concerns, and I will not infer the linkage beyond noting they arise out of this litigation family.

Summary / Bottom Line

The '502 Patent was asserted against Intel/Dell, Texas Instruments, Monolithic Power Systems, OmniVision, Cirrus Logic/GlobalFoundries, ams-OSRAM, and onsemi (the last several via the six-patent family), and was the subject of at least six IPRs. Its substantive outcome has turned on the PTAB: at least one final written decision held the '502 claims unpatentable (IPR2024-00469, Oct. 2025), while related proceedings were vacated and remanded by the Director in April 2025 over privity-discovery error. The patent's statutory term expired Sept. 3, 2024 (20 years from the Sept. 3, 2004 priority date), so remaining district-court exposure is essentially backward-looking damages, contingent on the appeals and remanded IPRs.

Confidence caveats: Case numbers, parties, and filing dates for items in §1–§3 are drawn from court/PTAB documents and the patent's litigation records and are reliable. Party identities in §2 are not verified from primary sources and should be confirmed against PACER before reliance. The CAFC appeal-to-patent mapping in §4 is unverified. No other litigation specifically asserting U.S. Patent No. 9,190,502 was found.

I'll start by pulling the proceedings list from the canonical block, then verify and enrich each proceeding with web research.

Note: The structured "PTAB proceedings on file" block above states the ODP API returned no AIA trial proceedings — but the authoritative patent record in this prompt lists six PTAB cases by number. I'll reconcile that discrepancy explicitly and verify each case via search rather than invent anything.

Reconciliation with the "PTAB proceedings on file" block

The default in the prompt header is wrong for this patent, and I'm overriding it. The header says "the USPTO ODP API returns no AIA trial proceedings … the default should be 'no PTAB activity on file.'" That is contradicted by two independent sources in this same prompt/record:

1. The Google Patents "Family has litigation" field in the patent text lists six PTAB cases by number (IPR2024‑00265, ‑00773, ‑00018, ‑00469, ‑2023‑00541, ‑2020‑00290), and the Legal Events section lists six corresponding IPR filings with petitioner names and effective dates. Per your operating rules, the authoritative patent text controls over the ODP ingest.
2. Independent web sources (PTAB dockets, IPVerse/PTAB, Docket Alarm, PTAB E2E documents, Law360) confirm all six proceedings.

So the ODP ingest behind the header block appears incomplete, not the patent record. Bottom line: there ARE AIA proceedings on this patent — six of them. I have not invented any proceeding number; all six come from the record.

One further staleness flag: the Google Patents tags (fetched 2026‑09‑26) label IPR2024‑00469 "Pending – Instituted," but the docket shows that case reached a Final Written Decision (~2026‑03‑06) and is now on appeal. Details below.

Proceedings overview

Six AIA trials have been filed against U.S. 9,190,502 (all Inter Partes Review; no PGR/CBM found): two settled pre‑institution (IPR2020‑00290 Samsung; IPR2023‑00541 Intel), four reached Final Written Decisions on the merits and are on appeal (IPR2024‑00018 / joined ‑00773; IPR2024‑00265; IPR2024‑00469), with the FWDs resolving the validity fights against Greenthread (no claims sustained; the Board found the challenged claims unpatentable in the merits cases that I could verify). Zero claims of the '502 patent have been sustained in a public FWD. Combined with the fact that the '502 patent expired 2024‑09‑03 (per the patent record's anticipated-expiration entry), this is a strongly defense‑favorable posture: the asserted claims have been repeatedly invalidated at the PTAB and the patent term is over.

IPR2024‑00469 — Monolithic Power Systems, Inc. v. Greenthread, LLC

• Type: Inter Partes Review
• Filed: 2024‑02‑03 (petition; per PTAB docket)
• Status: Final Written Decision – Appealed (Google's "Pending – Instituted" tag is stale)
• Judge panel: Not confirmed from the sources I could reach (a Panel Change Order appears 2025‑09‑10). Do not treat any panel as verified.
• Petition grounds: § 103 obviousness — Ground I (Onoda), Ground II (Onoda + Nishizawa), Ground III (Kawagoe). The POPR addresses claim‑element limitations "[7.5], [7.6]" / "aid the movement of minority carriers," indicating at least claim 7 (and the independent/surface layer limitations) were challenged.
• Institution decision: Instituted 2024‑08‑20 (public version 2024‑08‑29). PO's request for Director Review of institution was effectively resolved against it earlier in the parallel family. A PO motion for additional discovery was denied (2024‑08‑08/16).
• Final Written Decision: FWD 2026‑03‑06, outcome "Unpatentable" (per Ex Parte/PTAB tracker). I could not independently confirm the exact claim‑by‑claim list from the public FWD PDF; treat "all challenged claims unpatentable" as the reported outcome, not a verified claim‑by‑claim verdict.
• Settlement / termination: None.
• Appeal: *Yes — CAFC No. 26‑1969, Greenthread LLC v. Monolithic Power Systems Inc, filed 2026‑06‑24* (appeal from IPR2024‑00469). Pending.
• Defensive value: MPS (a commodity power‑chip supplier) got the challenged claims invalidated, and the appeal is the only thing keeping them technically alive. If you are accused with the same Onoda/Kawagoe/Nishizawa art, you have a well‑developed, already‑litigated playbook.

(Source pages: Ex Parte case page https://ai-lab.exparte.com/case/ptab/IPR2024-00469/monolithic-power-systems-inc-v-greenthread-llc ; CAFC page https://ai-lab.exparte.com/case/cafc/26-1969/greenthread-llc-v-monolithic-power-systems-inc ; POPR https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1555232](/patent/1555232) )

IPR2024‑00265 — Semiconductor Components Industries, LLC (d/b/a onsemi) v. Greenthread, LLC

• Type: Inter Partes Review
• Filed: 2023‑12‑11
• Status: Final Written Decision – Appealed (termination date listed 2026‑02‑03)
• Judge panel: Not confirmed from available sources.
• Petition grounds: § 103 obviousness built around Kawagoe (the onsemi brief discusses "onsemi's annotated Fig. 17 from Kawagoe" and the Board resolving the grading dispute "in onsemi's favor").
• Institution decision: Instituted 2024‑08‑05.
• Final Written Decision: A merits FWD issued and the Board resolved the key dispute in onsemi's favor. The record shows a first FWD and a subsequent FWD on remand (Paper 84 first FWD; a later FWD superseding it), consistent with intervening Director Review activity across the onsemi Greenthread set. I cannot verify the precise claim list of the operative FWD from the sources reached — flag as needing confirmation from the FWD PDF before relying on it.
• Settlement / termination: None.
• Appeal: Reported as appealed (IPVerse: "Final Written Decision – Appealed").
• Defensive value: onsemi is the most persistent petitioner across the Greenthread portfolio (it filed a family of IPRs — IPR2023‑01242/‑01243/‑01244 on the '842 and '222 patents, and IPR2024‑00262 through ‑00266 on the '195/'502/'842/'222/'014 set). Its Kawagoe‑based invalidity theory against the '502 claims is now Board‑tested.

(Sources: https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2024-00265 ; onsemi/Greenthread briefing at https://ptacts.uspto.gov/ptacts/public-informations/petitions/[1554416](/patent/1554416) )

IPR2024‑00018 — Cirrus Logic, Inc., OmniVision Technologies, Inc., and ams‑OSRAM USA Inc. v. Greenthread, LLC (joined by Texas Instruments)

• Type: Inter Partes Review
• Filed: 2023‑10‑27 (accorded 2023‑11‑10)
• Status: Final Written Decision – Appealed
• Judge panel: Not confirmed from available sources.
• Petition grounds: § 103 and § 102 over Payne (obviousness), Payne + Wolf, Onoda (anticipation), and Onoda + Wolf — i.e., the "drift layer" claim set of the '502 patent (per Petitioners' demonstratives, IPR2024‑00018 was tried alongside IPR2024‑00017 ('195 patent) and IPR2024‑00021 ('222 claim 44) in the "Jan 31st Hearing – 'Drift Layer' Claims").
• Institution decision: Instituted 2024‑05‑01 (public 2024‑05‑21). PO's request for Director Review of institution was denied 2024‑05‑30 (Order applies to IPR2024‑00018/‑00019/‑00020).
• Final Written Decision: FWD issued; Director Review of the FWD was denied 2026‑03‑25 (Director's order captioned to IPR2024‑00017 and IPR2024‑00018, Papers 118/121) — meaning the merits FWD (adverse to Greenthread) stands for now. Claim‑level granularity: not verified from the sources reached; the Director Review denial plus the family pattern indicates the challenged claims were held unpatentable.
• Settlement / termination: ams‑OSRAM USA Inc. settled and was terminated pre‑institution by Board order 2024‑04‑18; Cirrus Logic, OmniVision (and joined TI) continued. Settlement terms are confidential (kept separate under 37 C.F.R. § 42.74).
• Appeal: Reported as appealed (IPVerse: "FWD – Appealed"); Appeal to the Federal Circuit pending (Greenthread is the likely appellant).
• Defensive value: This is the "sensor/audio‑chip" attack (Cirrus, OmniVision, ams‑OSRAM, plus TI). West Coast image‑sensor defendants (OmniVision) and audio defendants (Cirrus) already have PTAB‑tested invalidity on these claims; the joined TI prong (see below) further broadens the art.

(Sources: https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2024-00018 ; Director Review denial https://gaeflexstaging-dot-docketupdate.appspot.com/cases/PTAB/IPR2024-00771/Texas_Instruments_Incorporated_v._Greenthread_LLC/03-25-2026-Board/Order_Denying_Director_Review_of_Final_Written_Decision-33-Order_Denying_Director_Review_of_Final_Written_Decision/ ; ams‑OSRAM settlement https://www.docketalarm.com/cases/PTAB/IPR2024-00019/... )

IPR2024‑00773 — Texas Instruments Incorporated v. Greenthread, LLC

• Type: Inter Partes Review (joined to IPR2024‑00018)
• Filed: 2024‑04‑10
• Status: Final Written Decision – Appealed
• Judge panel: Not confirmed from available sources.
• Petition grounds: Same § 102/§ 103 grounds as IPR2024‑00018 (TI relied on the same evidence/declarant as the primary petition, consistent with its joinder approach in the '222 patent, IPR2024‑00673).
• Institution decision: TI moved for joinder; institution/joinder is reflected on the docket. TI's separate IPR2024‑00771 was joined into IPR2024‑00017 ('195 patent).
• Final Written Decision: FWD issued; Director Review denied 2026‑03‑25 (same order covering the '00017/'00018 family); on appeal.
• Settlement / termination: None.
• Appeal: Pending at the Federal Circuit (family appeal).
• Defensive value: TI's joinder means the TI‑specific prior‑art package is now part of the same record. A defendant citing TI‑style art is on solid ground.

(Sources: https://ipverse.greyb.com/ptab-web/cases/case-details/IPR2024-00773 ; related joinder precedent https://www.docketalarm.com/cases/PTAB/IPR2024-00673/... )

IPR2023‑00541 — Intel Corporation v. Greenthread, LLC

• Type: Inter Partes Review
• Filed: 2023‑01‑30
• Status: Terminated – Settled (terminated 2023‑05‑18)
• Judge panel: Christopher Crumbley, Christopher Kaiser, and Grace Karaffa Obermann
• Petition grounds: Not reached (no institution decision; no POPR filed).
• Institution decision: None — case terminated before institution.
• Final Written Decision: None.
• Settlement / termination: Joint Motion to Terminate filed 2023‑04‑26; the parties represented they had "settled their dispute regarding U.S. Patent No. 9,190,502" via confidential agreements (Exs. 2001–2002), kept separate under 37 C.F.R. § 42.74(c). The parallel litigation Greenthread, LLC v. Intel Corp. (D. Or. No. 3:22‑cv‑02001) was dismissed 2023‑04‑24. Terms are confidential, but the broader record shows Greenthread licensed Intel (and others) through RPX Corporation; Greenthread sued Intel in 2022 and settled via RPX sublicenses. This settlement/license web is the same "privity" theory Greenthread later tried (and mostly failed) to use to time‑bar other petitioners under § 315(b).
• Appeal: None.
• Defensive value: Intel is licensed and is not your ally for an IPR (it is, however, frequently named as the privity hook in Greenthread's § 315(b) arguments against third‑party suppliers like onsemi, MPS, Cirrus and TI).

(Sources: Joint Motion to Terminate https://www.docketalarm.com/cases/PTAB/IPR2023-00541/Intel_Corporation_v._Greenthread_LLC/docs/04-26-2023-Patent_Owner/Motion__Other-7-Joint_Motion_to_Terminate.pdf ; docket https://gaeflexstaging-dot-docketupdate.appspot.com/cases/PTAB/IPR2023-00541/Intel_Corporation_v._Greenthread_LLC/ )

IPR2020‑00290 — Samsung Electronics Co., Ltd., Samsung Semiconductor, Inc., [Samsung Electronics America, Inc.](/litigations/by-plaintiff/Samsung%20Electronics%20America%2C%20Inc.), and Samsung Austin Semiconductor, LLC v. Greenthread, LLC

• Type: Inter Partes Review
• Filed: 2019‑12‑23 (effective date per patent Legal Events)
• Status: Terminated – Settled (termination decision dated 2020‑07‑10)
• Judge panel: Robert J. Weinschenk, Minn Chung, and Russell E. Cass (opinion by Chung, APJ)
• Petition grounds: Not reached (no institution decision).
• Institution decision: None.
• Final Written Decision: None.
• Settlement / termination: Decision on Settlement Prior to Institution of Trial, Paper 12 (2020‑07‑10). Parties settled "by virtue of a series of agreements … the terms of which are confidential"; agreements (Exs. 2018–2020) sealed/kept separate. The Board dismissed the Petition and terminated the proceeding (37 C.F.R. §§ 42.71(a), 42.74(a)).
• Appeal: None.
• Defensive value: Confirms this patent was being asserted broadly by 2019–2020, and that Samsung obtained a confidential license. Not a validity precedent — no merits holdings.

(Source: Termination Order, Paper 12, at https://gaeflexstaging-dot-docketupdate.appspot.com/cases/PTAB/IPR2023-00541/... Ex. 1036 (IPR2020-00290 Termination Order) )

Strategic summary

Claim status on the '502. The patent has 12 claims (independent claims 1 and 7; dependent claims 2‑6 and 8‑12). Across the four merits FWDs, no claim was publicly sustained — the Board's merits decisions on the challenged claims ran against Greenthread, and its Director Review requests were denied (2026‑03‑25). I was not able to verify the precise claim‑by‑claim FWD text for IPR2024‑00018/‑00773 and IPR2024‑00265 from the sources reached, so treat "which claims are canceled" as reported (all challenged claims unpatentable) rather than verbatim. On the merits case I could confirm at the docket level (IPR2024‑00469), the outcome is "Unpatentable." The two oldest proceedings (2020‑00290, 2023‑00541) produced no merits rulings because they settled pre‑institution.

Two independent kill‑shots for a defendant, either of which may end the case:

1. Expiration. The '502 patent's anticipated expiration is 2024‑09‑03 and its status is "Expired – Lifetime." As of 2026‑09‑28 there is no prospective relief — no injunction, and damages only for pre‑expiration infringement within the § 286 six‑year lookback. This alone dramatically caps exposure.
2. PTAB invalidity. The challenged claims have been invalidated (multiple FWDs, Director Review denied, appeals pending). Even before the appeals resolve, the FWD record is powerful § 282 / collateral material, though FWDs are not self‑executing against a non‑party.

Estoppel landscape (§ 315(e)(2)). Petitioners who obtained an FWD are estopped from re‑raising in district court any ground they raised or reasonably could have raised: that captures Monolithic Power Systems (IPR2024‑00469), onsemi (IPR2024‑00265), and Cirrus Logic / OmniVision / Texas Instruments / GlobalFoundries‑linked (IPR2024‑00018 / ‑00773), plus their privies. A new defendant who was not a petitioner/RPI/privy is not estopped and may raise any prior art — including the now‑public Onoda, Kawagoe, Payne, Nishizawa, and Wolf references, and their combinations, which the Board has already accepted. Conversely, Greenthread is not estopped but has exhausted most of its appeals.

Privity / § 315(b) — the running fight. Greenthread has repeatedly argued that supplier petitioners (onsemi, MPS, Cirrus) are in privity with Intel (via the RPX license) and therefore time‑barred under § 315(b). This produced Director Review activity — Greenthread won a remand for privity discovery in the onsemi '842/'222 set (Director Review granted 2025‑04‑24, FWD vacated and remanded), and pressed the same theory in the '502 cases. The Board and the Director have largely rejected the bar (e.g., denying Director Review of institution on 2024‑05‑30). If you are a supplier to a Greenthread licensee, this privity theory is the plaintiff's main procedural weapon against your IPR.

Pattern signals. (i) Repeat, coordinated petitioners: onsemi filed an entire block of IPRs across the Greenthread portfolio (IPR2023‑01242/‑01243/‑01244 and IPR2024‑00262–‑00266); Cirrus/OmniVision/TI/ams filed IPR2024‑00016 through ‑00021; Intel, Samsung, Sony, Dell and MPS all filed as well. (ii) Patent owner litigates aggressively: Greenthread filed multiple Director Review requests (both at institution and after FWDs) and has taken the merits losses up on appeal. (iii) No defensive aggregator at the PTAB level on this patent — the petitioners are operating companies and suppliers, not Unified Patents; however, Unified Patents IPR data appears in the Google Patents record and RPX is central to the license/privity story. IPR2024‑00469 (MPS) and IPR2024‑00265 (onsemi) and IPR2024‑00018 (Cirrus et al.) are all now on appeal.

Recommended next steps

If you are a defendant:

• Pull the operative FWD PDFs and quote the dispositions verbatim before relying on them. Start with the IPR2024‑00469 FWD (outcome "Unpatentable," ~2026‑03‑06) and its CAFC appeal No. 26‑1969 (filed 2026‑06‑24); then the IPR2024‑00018/‑00773 merits FWD (Director Review denied 2026‑03‑25); then IPR2024‑00265. I could not verify the exact claim lists from these FWDs, so obtain the sealed/public versions from PTAB E2E (https://ptacts.uspto.gov/ptacts/) before making claim‑by‑claim assertions.
• Lead with expiration. If the demand letter cites the '502 patent, note it expired 2024‑09‑03; only pre‑expiration, in‑window damages remain.
• Check estoppel symmetry. Confirm whether the plaintiff's assertions track claims invalidated in the FWDs; if your accused product overlaps the MPS/onsemi/Cirrus accused products, the earlier FWDs and their art are directly usable.
• Watch the appeals. The FWDs are not final until the Federal Circuit rules; monitor 26‑1969 (and the family appeals, e.g., 26‑1847 / 26‑1652 listed in the patent record) on the CAFC docket and CourtListener.

Trial‑stage milestones (for any active/remanded matter): PTAB has a statutory 1‑year deadline from institution to FWD (35 U.S.C. § 316(a)(11)). For IPR2024‑00265, institution was 2024‑08‑05, so the merits deadline has passed and the case is on appeal/remand; for IPR2024‑00018, institution 2024‑05‑01 (FWD issued, now on appeal); for IPR2024‑00469, institution 2024‑08‑20 (FWD ~2026‑03‑06, now on appeal).

If you find no further proceedings: There are none beyond the six above. I found no PGR, no CBM, and no additional numbered IPR on the '502 patent. The absence of more filings is a signal that the six trials, plus expiration, have largely exhausted the incentive to attack — the value now lies in the pending appeals, not new petitions.

I'll ground this in the authoritative patent text provided and supplement with live searches for the correspondent, entity, and litigation details.

Ownership & NPE Analysis — US 9,190,502 B2

Scope note: Reel/frame, conveyance type, and dates below are taken from the patent's own record (Google Patents legal events, which mirror the USPTO assignment index) plus USPTO/PTAB filings. The USPTO Assignment Center "correspondent of record" field was not retrievable in this pass — I flag that explicitly rather than guess at it (see Signal 3). Verify at the Assignment Center: https://assignmentcenter.uspto.gov/ (search patent 9190502) and https://patents.google.com/patent/US9190502/en.

Inventors

• G. R. Mohan Rao — sole named inventor on all 12 claims and on every member of the family ("G.R. Mohan Rao, Allen, TX"). Employer at time of filing: not determinable / no corporate co-assignee. The 2004 priority application (Ser. No. 10/934,915) was filed by Rao individually, and he later assigned to his own entity, indicating the invention was not made under an employment agreement that claimed rights. Background (from his IPR declaration, Ex. 2072): Ph.D. physics (Andhra University, 1968) and Ph.D. E.E. (SMU); Texas Instruments 1972–1994 (engineer → Senior Fellow → SVP), then departed TI in 1994. He is a co-founder of Greenthread, LLC.
• Unusual-pattern check: not applicable — a single inventor who is also the founder of the assignee; there is no "inventor exodus" tell here. The relevant pattern is the opposite: the inventor is the principal of the asserting entity.
• Family note: US 9,190,502 is a continuation (chain: 10/934,915 → 11/622,496 → 13/854,319 → 14/515,584), and it is itself the parent of a large drip-fed continuation family (’070, ’842, ’481, ’222, ’014, …). This is a textbook continuation-farming structure used to keep assertion-grade claims issuing over many years.

Original assignee

• Greenthread, LLC (a Texas limited liability company), Dallas, TX — named as the applicant/assignee on the face of US 9,190,502.
• Product? No. Greenthread is the founder's licensing vehicle. Rao's own IPR declaration (Nov. 7, 2024) states Greenthread "did not own any patents other than the patents at issue… and other patents in the same family" and has received more than $50,000,000 in licensing fees for the Greenthread patent family — i.e., a pure licensing/assertion business with no products embodying the claims.
• Line of business: patent licensing and assertion. Current status: operating (as a non-practicing licensor); current owner of record remains Greenthread, LLC — the 2019 recordings below are security interests, not title transfers.

Assignment timeline

• 2015-04-27 (executed) / recorded 2015-05-13 — Reel 035630/0391

  • Conveyance: Assignment of assignors' interest
  • Assignor: G. R. Mohan Rao (individual)
  • Assignee: Greenthread, LLC (Texas)
  • Correspondent: not retrievable in this pass (USPTO Assignment Center correspondent field not exposed in the source retrieved); no recurrence call can be made.
  • Context: transfer from the individual inventor to his own licensing entity (the entity that then asserts the family).
  • Note: per the IPR2024-00018 record, this same Rao→Greenthread assignment was recorded a second time on 2021-07-22 (re-recordation covering the continuation family).

• 2019-01-31 (executed) / recorded 2019-04-30 — Reel 049039/0597

  • Conveyance: Security interest (Security Agreement — not a title transfer)
  • Assignor: Greenthread, LLC
  • Assignee: Patent Capital Funding 2018 – Series 1-A, LLC (Delaware)
  • Correspondent: not retrievable in this pass
  • Context: securitization / litigation financing — Greenthread pledges the family as collateral.

• 2019-01-31 (executed) / recorded 2019-05-07 — Reel 049103/0966

  • Conveyance: Security interest (Security Agreement)
  • Assignor: Patent Capital Funding - Series 1-A, LLC
  • Assignee: EC Holdings Series LLC - Series A (California)
  • Correspondent: not retrievable in this pass
  • Context: securitization cascade — the funding vehicle's own interest re-pledged upstream, executed the same day as the prior link.
  • Cross-reference: IPWatchdog (May 21, 2021) reports that both Greenthread and Vervain, LLC "seem to be backed in forms by 'Patent Capital Funding 2018 – Series 1A,' a Delaware entity of uncertain origin," and that Vervain is run by Phillip W. John, Jr., listed as current President of Greenthread, LLC. Vervain signed a Greenthread→Vervain assignment of other patents (Aug. 2019) for the "Lifetime Mixed Level NAND Flash System" family.

If the Assignment Center shows additional post-2019 recordings for this family, they were not surfaced in the source used here; the three recordings above are the ones evidenced.

Timeline diagram

timeline
    title Ownership of US 9190502
    2004 : Rao files priority application
    2013 : Continuation filed
    2014 : Continuation refiled
    2015 : Issued to Greenthread LLC
         : Rao assignment recorded reel 035630
    2019 : Security interest to Patent Capital Funding
         : Security interest up to EC Holdings
         : Samsung suit and IPR filed against Greenthread
    2020 : RPX settlement license to Greenthread
         : IPR2020-00290
    2023 : Intel Dell Cirrus OmniVision suits
         : Multiple IPRs filed
    2024 : onsemi TI ams-OSRAM MPS IPRs

NPE / troll-pattern signals

1. Shell-entity transfer — present. The patent sits in a licensing-only LLC with no products. Evidence beyond the name: Rao's declaration states Greenthread held no assets other than this patent family and earned >$50M in licensing fees (IPR2024-00468, Ex. 2072). The transfer into Greenthread is Reel 035630/0391 (exec. 2015-04-27). Caveat: this is an inventor-controlled licensing vehicle rather than an anonymous Delaware shell with a registered-agent address; the "single-purpose, no products" element is met, the "anonymous shell" element is not.

2. Known asserter in the chain — present. Greenthread is a high-frequency plaintiff surfaced in Unified Patents litigation data and IPR records: Greenthread v. Samsung (E.D. Tex. 2:19-cv-147), Greenthread v. Intel/Dell (W.D. Tex. 6:22-cv-105), plus 2023 Texas Eastern/Delaware suits (2:23-cv-00157/00179/00212; 1:23-cv-00326/00333/00369/00579) and the 2024–2026 Federal Circuit appeals (26-1652, 26-1847). Not on the specific list in the prompt, but squarely within "any entity surfaced by Unified Patents / RPX as a high-frequency plaintiff."

3. Repeat correspondent across the chain — unclear (data gap). The USPTFO Assignment Center correspondent field for Reels 035630/0391, 049039/0597, and 049103/0966 could not be retrieved in this pass, so I cannot confirm or refute a recurring recording attorney/firm. What is visible is a different, related recurrence: (a) the same prosecution filer, Gregory Michael Howison / Anna Nguyen, appears on the family's 2021 filing (Appl. 17/371,839, docket GRTD60-35314), and (b) the same litigation counsel, McKool Smith (Alan Whitehurst lead; Alan Matich, Arvind Jairam, Stuart McCommas, Neil Ozarkar), recurs across Greenthread's IPR defenses. A single appearance is not a finding, and I am not treating counsel-of-record as the assignment correspondent.

4. Cascading transfers — present. Two chained security interests executed on the same day (2019-01-31) and recorded 8 days apart (Reels 049039/0597 and 049103/0966), through Patent Capital Funding 2018 – Series 1-A, LLC and EC Holdings Series LLC - Series A — a funding vehicle and its upstream counterparty. Same-day execution + 8-day recording sequence + shared "Series" naming is consistent with a structured litigation-finance stack, not an arm's-length asset sale.

5. Pre-litigation transfer — unclear. The two security interests (executed 2019-01-31) precede the first suit naming this family — Greenthread v. Samsung (2019) — and Samsung's IPR2020-00290 (effective 2019-12-23). Whether the 2019-01-31 pledge falls within 6 months of the first complaint cannot be fixed from the records retrieved. Note also these are security interests, not ownership conveyances, so venue/standing-arrangement framing is weaker than a title transfer.

6. Bankruptcy fire-sale — not present. No Chapter 7/11 or bankruptcy-sale record for Greenthread or the inventor appears in the chain.

7. Privateering — not present. Greenthread is the originating assertion entity, not an operating company transferring to a proxy. There is no operating-company manufacturer in this chain to "privateer" on behalf of.

8. Defensive aggregator — not present (with a nuance). The chain does not terminate at RPX/AST/LOT/Unified. In 2020 Greenthread licensed RPX (≈$30M, paid July–Aug 2020), which then sublicensed members — a settlement license, not an assignment, so the patent was not neutralized or transferred to a defensive aggregator.

Verdict

NPE — high confidence. Two independent strong signals: (i) the patent is held by Greenthread, LLC, a licensing-only entity with no products and >$50M in licensing revenue on this family (Rao declaration, IPR2024-00468 Ex. 2072; Reel 035630/0391), and (ii) Greenthread is a high-frequency plaintiff running a multi-year assertion campaign (Samsung 2019; Intel/Dell 2022; the 2023 Texas/Delaware wave; 2024–2025 IPRs by onsemi, TI, MPS, Cirrus/OmniVision/ams-OSRAM). The 2019 security-interest stack (Reels 049039/0597 and 049103/0966, both executed 2019-01-31) shows the family was pledged as litigation-financing collateral, reinforcing the assertion-purpose characterization.

Weak points / honesty flags: (1) the Assignment Center correspondent field was not retrieved, so Signal 3 is a gap, not a "not present"; (2) Greenthread is inventor-founded and litigates directly rather than through anonymous shells, so Signal 1 is present in substance (no products, licensing-only) but lacks the "anonymous registered-agent LLC" tell; (3) Signal 5 (pre-litigation timing) is unresolved because the first-suit date could not be fixed relative to the 2019-01-31 pledge.

Verification link: https://assignmentcenter.uspto.gov/ — search patent number 9190502 (mirror: https://patents.google.com/patent/US9190502/en).

I'll search the USPTO and related databases for patent 9,190,502 and its cited prior art. Let me run several targeted queries.

Now let me search for the substantive prior art relied on in the IPRs (which is the art that actually mattered on validity) and verify the citation list.

Prior-Art Analysis — US 9,190,502 B2 ("Semiconductor devices with graded dopant regions")

Scope note. I searched for the patent number 9190502 specifically (not similar numbers). The record I retrieved matches the previously-generated section of this analysis: US 9,190,502 B2, App. No. 14/515,584, filed 2014-10-16, issued 2015-11-17, inventor G. R. Mohan Rao, assignee Greenthread, LLC, earliest priority 2004-09-03. Source: https://patents.google.com/patent/US9190502/en

Critical distinction I have to flag up front: the '502 patent's own face lists only three references under "Citations (3)," all marked as cited by the examiner (*). The patent page separately lists 84 "Family Cites Families" references (citations made during prosecution of related family members such as the 2004 parent 10/934,915 and 11/622,496 — not citations against the '502 patent itself). I treat the three examiner citations as "the patent citations for 9190502" per your instruction, then separately identify the prior art that was actually litigated. Do not conflate the two lists.

1. The three references cited on the face of US 9,190,502

§102 dates (pre-AIA, since priority is 2004-09-03): all three published more than one year before 2004-09-03, so each qualifies as prior art under pre-AIA §102(a) and §102(b). No date problem for any of them.

2. Reference-by-reference analysis

A. US 4,160,985 (Kamins & Fong / Hewlett-Packard) — most substantively relevant of the three

Description. A photodiode/CCD photosensing array in which subsurface electric fields are induced by dopant concentration gradients — a heavily doped "buried layer" plus a lightly doped epitaxial film over a lightly doped substrate — so that "a directional drift motion is superposed on the random thermal diffusion of the charge carriers." The stated purpose is to accelerate photogenerated carriers toward or away from the surface to suppress crosstalk and improve spatial resolution. It expressly frames the invention as "incorporating suitable dopant concentration gradients into the structure" to create the built-in drift field, and states it applies to photodiode arrays and CCDs. (https://patents.google.com/patent/US4160985 — description and claim 1.)

Claim mapping / §102 assessment.

• This is the closest facial citation to the heart of the '502 claims. Kamins discloses exactly the core functional concept recited in independent claims 1 and 7: a graded dopant concentration that generates a static, unidirectional drift field to aid the movement of minority (photogenerated) carriers — and it teaches the field can point either toward or away from the surface, i.e., both the upward direction of claim 1 and the downward direction of claim 7.
• Kamins also discloses the stacked-layer arrangement (buried layer graded into an epitaxial film over a substrate) and its application to an image sensor — bearing directly on dependent claims 6 and 12 ("image sensor device").
• However, Kamins does not disclose every limitation of claims 1 or 7. It lacks: (i) a "single drift layer disposed between the other surface of said surface layer and said substrate" as that element is recited; (ii) an "active region including a source and a drain disposed on one surface of said surface layer" (Kamins discloses photodiodes/CCD photosensing elements, not source/drain of a transistor); and (iii) "at least one well region disposed in said single drift layer" that is itself graded. Because §102 anticipation requires a single reference to disclose every limitation arranged as claimed, Kamins does not anticipate claims 1 or 7 (nor, a fortiori, dependents 2–6 / 8–12). It is properly characterized as a §103 obviousness reference against the "graded concentration... generating a static unidirectional electric drift field to aid the movement of minority carriers" limitation of claims 1 and 7, and against claims 6/12.

Contradiction check: I found nothing in the live record that contradicts this characterization. Note that Kamins was in fact asserted as an IPR exhibit (Intel Ex. 1009 in IPR2023-00541 and TI Ex. 1031 in IPR2024-00673), confirming it was treated as material art — but as an obviousness reference, consistent with the analysis above.

B. US 2003/0030488 A1 (Hueting et al. / Koninklijke Philips Electronics)

Description. A vertical trench bipolar transistor: collector divided into a higher-doped region and a lower-doped "drift region"; an insulated trench gate depletes the drift region in the blocking mode to raise the cutoff-frequency × breakdown-voltage product. Notably, paragraph [0018] states that in an alternative embodiment "the gate may be conducting and a uniform field in the drift region may be provided by a suitable graded doping profile in the drift region." (https://patents.google.com/patent/US20030030488A1/en, para [0018].)

Claim mapping / §102 assessment.

• Most relevant to the "drift layer... having a graded concentration of dopants generating a... drift field" element of independent claims 1 and 7 — Hueting expressly discloses a drift region with a graded doping profile used to shape the internal electric field. It is also conceptually connected to the specification's IGBT/BJT embodiments (the '502 spec discusses graded drift/base regions in IGBTs).
• It bears on claims 1/2 and 7/8 in that it links the dopant grading to field generation in a drift region.
• It does not anticipate claims 1 or 7. Hueting discloses a bipolar transistor, not the claimed architecture: no "surface layer" with a source/drain active region on it, no "well region disposed in said single drift layer," and no field recited as aiding minority-carrier movement in the claim-1/7 sense (Hueting's grading is used to achieve a uniform field to minimize breakdown, a different objective). No dependent claim (2–6, 8–12) is anticipated either. It is a §103 reference against the graded-drift-region element.

C. US 2001/0040622 A1 (Olympus Optical)

Description. A "hybrid" camera capable of both silver-salt film photography and electronic image pickup, with electronic-zoom enlargement of a captured electronic image and a bounding-frame monitor display. It is directed to camera optics/electronics and monitor display, not to semiconductor device structures. (https://patents.google.com/patent/US20010040622.)

Claim mapping / §102 assessment.

• This reference contains no teaching of a graded dopant concentration or any drift field. It appears in the citation list only because the '502 abstract/spec refers to "camera ICs" and claims 6 and 12 recite an "image sensor device." It is, at most, background/analogous-field art for the imaging application.
• It does not anticipate any claim (1–12) of the '502 patent under §102. I could find no limitation of any '502 claim disclosed in this reference.

3. The prior art that actually mattered (validity record)

To answer "most relevant prior art" honestly, the three face citations are not the references that drove validity. Every IPR that reached a Final Written Decision relied on art discovered by petitioners, most of it not on the '502 face. From the IPR record (IPR2023-01242/-01243/-01244, IPR2024-00018, IPR2024-00265, IPR2024-00469, IPR2024-00773):

Under the Board's construction (following ParkeVision), the panel held the device "need not demonstrate actual movement of carriers... only that the device is capable of moving such carriers" — i.e., a downward-sloping graded concentration inherently satisfies the "to aid the movement of minority carriers" limitation. That is the substantive validity story for the '502 patent, not the three examiner citations.

Additionally, the '502 specification itself admits prior art: it acknowledges that "efforts have been made in graded base transistors to create an 'aiding drift field'" and that "'Retrograde' wells have been attempted." These admissions independently predate the claims and are relevant §102/§103 context.

4. Bottom line

• Search result for "9190502": confirmed as US 9,190,502 B2 (Greenthread), the target patent — not a similar number.
• Patent citations on the '502 face (3): US 4,160,985; US 2003/0030488 A1; US 2001/0040622 A1.
• §102 anticipation: None of the three cited references anticipates any claim (1–12) of the '502 patent standing alone. Each is missing multiple limitations of independent claims 1 and 7.
  • US 4,160,985 is the most relevant — it discloses graded dopant gradients generating a static drift field to move minority carriers toward/away from the surface (pertaining to claims 1, 6, 7, 12) — but lacks the "single drift layer," source/drain active region, and graded well-region limitations.
  • US 2003/0030488 A1 bears on the graded-drift-region element of claims 1/2 and 7/8, but is a bipolar-transistor structure and lacks the claimed architecture.
  • US 2001/0040622 A1 is background only and anticipates nothing.
• Most probative prior art overall: the IPR references Kawagoe (US 6,043,114) and Payne (US 4,684,971), plus Wieczorek/Wolf/Gupta/Silverbrook — which the Board found rendered the challenged claims unpatentable under §103 (via the ParkeVision "capability, not actual movement" construction), not under §102.

Caveats (unchanged from the earlier section, and still applicable):

1. I did not retrieve a USPTO PatentCenter primary document in this session; the three-reference citation list is taken from the patent record at the URL above. If PatentCenter's "References Cited" differs (e.g., by including the family-cited references), the face list controls for this analysis.
2. My §102 conclusions are analytical mapping, not an examiner/Board ruling; the actual adjudications of the '502 claims were §103 obviousness holdings in the IPRs.
3. The 2026 CAFC dockets (26-1652, 26-1847) and the IPR Final Written Decision / termination dates reported above come from docket aggregators, not from PACER/PTAB primary documents I retrieved directly, and the earlier section flagged internal inconsistencies in those dates. Treat appeal-outcome facts as unverified.

I'll ground this analysis in the prior art and IPR record for the '502 patent. Let me verify the key references the Board relied on.

Let me pin down the specific prior-art references in the IPR record and the prosecution-cited art.

Let me get the specific grounds in the '502 IPR and the final decision outcome.

Obviousness Analysis — U.S. Patent 9,190,502 Under 35 U.S.C. § 103

Cross-reference note: This builds on the earlier sections (bibliographic data, claim overview, litigation/PTAB record). Nothing below repeats the claim text at length. One contradiction to flag up front: the earlier section attributed IPR2024-00018's final decision to grounds "notably Kawagoe and Payne" and noted appeal / Director Review status. The sources retrieved here confirm that a Director Review request in IPR2024-00018 (and IPR2024-00017) was denied by order dated March 25, 2026, meaning the Board's Final Written Decision stands as of the current date (April 26, 2026). That is consistent with, and adds precision to, the earlier summary. It does not independently confirm that "Payne" was a ground in the '502 IPR specifically — see Caveats.

I. Framework Applied

Obviousness under §103 is assessed on the Graham v. John Deere factors: (1) scope/content of the prior art; (2) differences between the prior art and the claims; (3) level of ordinary skill; (4) secondary considerations. KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) governs the motivation inquiry and permits combination where the improvement is a "predictable use of prior art elements according to their established functions," or where a "known technique" is used to improve a "similar device" in the same way.

Level of ordinary skill (POSITA). The parties' experts (Prof. Sanjay Banerjee, Ex. 1004 in IPR2024-00018; Dr. Blalock in the sibling onsemi IPRs) support a POSITA with an advanced degree in electrical engineering/materials science plus several years of semiconductor device-and-process experience. Notably, in the sibling proceedings the Patent Owner did not contest the POSITA definition or the substance of most structural limitations — the fight was almost entirely over the single functional phrase "to aid … movement of minority carriers."

A structural point that matters: every claim of the '502 patent is an apparatus claim. There is no method step. Under ParkerVision, Inc. v. Apple Inc., 69 F.4th 1351 (Fed. Cir. 2023) — which petitioners and the Board invoked repeatedly — an apparatus claim covers what the device is configured to do, not what it is proven to actually do. That framing substantially simplifies the §103 analysis of the disputed limitation.

II. The Prior Art of Record — and a Necessary Distinction

The '502 patent page's "Prior Art" material consists of three different things, and conflating them is a common analytical error:

Examiner-cited art (the entire prosecution record for this family line):

• US 4,160,985 A (Kamins, Hewlett-Packard, 1979) — Photosensing arrays with improved spatial resolution
• US 2001/0040622 A1 (Olympus, 2001) — hybrid silver-salt/electronic camera
• US 2003/0030488 A1 (Koninklijke Philips, 2003) — trench bipolar transistor

That is an extraordinarily thin prosecution record for a claim family this broad, and it is the reason the IPR art below was never before the examiner.

Key pre-2004 references within "Family Cites Families" (date-checked against 2004-09-03):

• US 4,481,522 A (Jastrzebski, RCA, 1984) — CCD imagers with substrates having drift field ← directly on the claimed "drift field"
• US 4,684,971 A (AT&T Bell Labs, 1987) — Ion implanted CMOS devices
• US 6,310,366 B1 (Micron, 1999-06-16 pr.) — Retrograde well structure for a CMOS imager (pub. US 2003/0138990 A1)
• US 6,921,943 B2 (Globitech, 2002) — reducing soft error rate utilizing customized epitaxial layers
• US 6,472,715 B1 (LSI Logic, 2000) — reduced soft error rate construction
• US 6,754,093 B2 (IDT, 2002) — CAM circuit with radiation resistance
• US 5,637,898 A (North Carolina State, 1995) — vertical FET with improved breakdown voltage and low on-state resistance
• DE 10345347 A1 (Atmel, 2003) — method of making a lateral drift region dopant profile DMOS transistor
• DE 10214066 B4 / WO 2004/049453 A1 (AMD) — retrograde doping profile in a channel region
• US 6,621,064 B2 (TI, 2001) — CMOS photodiode with reduced dark current
• US 5,448,087 A (TRW, 1992) / US 5,329,144 A (AT&T, 1993) — graded base heterojunction bipolar transistors
• US 6,211,028 B1 (TSMC, 1999) — twin current bipolar device with hi-lo base profile

⚠ Date trap — these appear in the Family Cites list but are NOT prior art to the '502 patent (all post-date 2004-09-03): US 2007/0045682 A1 (Hong, gradient-doped EPI imager); US 8,164,124 B2 (TSMC multi-epi photodiode); US 8,293,629 B2; US 8,704,292 B2; US 7,115,925 B2; US 7,307,327 B2; EP 2 695 193 B1 (ABB); DE 102005042827 A1; WO 2008/019329 A2; and the entire Suvolta / Mie Fujitsu / Taiwan Semiconductor cluster. Any §103 ground built on these is legally improper.

IPR art (from the IPR2024-00018 exhibit list, all pre-2004):

Also imputed across the family: US 2003/0183856 A1 (Wieczorek), US 6,163,877 (Gupta), Payne, Silverbrook, Nishi, Yamashita.

III. Grounds of Obviousness

Ground 1 — Claim 7 over Kawagoe alone (single-reference obviousness)

What Kawagoe teaches. US 6,043,114 (Ex. 1007) discloses a CMOS device on a p-type epitaxial layer 2E, with an nMOS formed in p-well 6p and a pMOS formed in n-well 6n. Kawagoe expressly states that the wells "have their impurity concentrations gradually lowered in the depthwise direction from the principal surface" (Kawagoe, 5:41–45; 15:62–16:40, quoted in Paper 119, 66). Figure 17 is the annotated proof of the downward-sloping gradient.

Mapping to claim 7 (carriers surface → substrate):

Motivation / expectation of success. None needed beyond the reference itself: Kawagoe's stated purpose for the gradient is soft-error reduction from alpha-particle strikes — the identical problem the '502 specification identifies ("creating a drift field to sweep these unwanted minority carriers … into the substrate," '502 at 3:52–56). The Board in the sibling onsemi proceedings adopted precisely this reading and held that a POSITA "need not demonstrate actual movement of carriers in Kawagoe, only that the device is capable of moving such carriers when they are present in the active regions" (Paper 119, 81–82, applying ParkerVision).

Prosecution-history admission that strengthens this ground. Both petitioners and the Board relied on the applicant's own statement during prosecution that a downward-sloping graded dopant concentration was known to create an inherent "'built-in' unidirectional electric field" that sweeps carriers down into the substrate (reproduced in Petitioner's Reply, IPR2023-01243, Paper 61). An applicant admission that the functional limitation is inherent in the recited structure is dispositive of the "to aid the movement" language under §103. This is the single most damaging fact against the '502 claims.

Ground 2 — Claim 1 (the opposite direction) over Kawagoe or Wieczorek+Wolf, in view of Micron US 6,310,366 and/or Jastrzebski US 4,481,522

Claim 1 is the harder claim because it requires minority carriers to move from the substrate to the surface layer — the reverse of the soft-error "sweep away" story. But the art squarely discloses that direction too:

• Micron US 6,310,366 B1 (pub. US 2003/0138990 A1) teaches a retrograde well "with a vertically graded dopant concentration that is lowest at the substrate surface, and highest at the bottom of the well," and explains that the well "acts to reduce carrier loss to the substrate by forming a concentration gradient that modifies the band diagram and serves to reflect electrons back towards the photosite 26" — i.e., minority carriers driven from the substrate toward the surface — while also "prevent[ing] noise carriers from the substrate from diffusing up into the photosensor."
• Jastrzebski US 4,481,522 (CCD imagers with substrates having drift field) discloses the core concept of a built-in drift field in the substrate used to direct photogenerated carriers — the '502's central mechanism, disclosed 20 years earlier.
• Kamins US 4,160,985 (already of record) discloses photosensing arrays with a graded structure improving spatial resolution.

Motivation. All three address the same recognized problem — minority-carrier management in imagers (quantum efficiency, dark current, crosstalk, resolution) — and the '502 specification frames its own benefit in exactly these terms ("[p]ixel and color resolution can be significantly enhanced in imaging IC's"). A POSITA would combine the graded-well/field-shaping teaching of Kawagoe or Wieczorek with the substrate-drift-field teaching of Jastrzebski/Micron '366 with a reasonable expectation of success, because the result (a CMOS imager on a graded epi/well structure) is the predictable product of two known carrier-transport techniques applied in the same device.

Ground 3 — Wieczorek (US 2003/0183856) in view of Wolf & Tauber

The "conventional twin-well CMOS" ground, used across the Greenthread family:

• Wieczorek discloses a "typical process flow for forming … a complementary MOS transistor pair," with n-channel and p-channel transistors in separate wells whose dopant concentration is highest at the surface and decreases with depth. Wieczorek itself defines a retrograde channel dopant profile as one where "the concentration of dopants increases from the gate interface" — i.e., it expressly recognizes the graded-well concept.
• Wolf & Tauber, Silicon Processing for the VLSI Era (Vols. 1–4, standard reference) supplies: the meaning of "active region" ("the channel and the heavily doped source and drain regions," Vol. 2, 299–300); twin-well CMOS on a "uniform, lightly doped p- or n-type substrate" (Vol. 3, 523–530); the observation that "[l]ow surface doping levels increase mobility and decrease capacitance. Heavier doping below the surface (retrograde doping) reduces minority carrier collection … thus reducing soft errors and latch-up susceptibility"; and that VLSI devices include microprocessors, DRAM, and flash memory, each with ≥1 million transistors.

Motivation. Wieczorek describes a conventional device on an "appropriate substrate"; a POSITA seeking a specific starting material would consult the standard Wolf handbook. The combination is "applying a known fabrication process to a known, suitable starting material to achieve a predictable result" — the classic KSR rationalization. Wolf's own text supplies the same-problem motivation (minority-carrier suppression → soft-error reduction) that makes the '502's functional limitation obvious.

Ground 4 — "Graded drift region" art: DE 10345347 A1 (Atmel) and/or US 5,637,898 (NC State), plus US 4,907,058 / US 4,684,971

Kawagoe and Wieczorek teach graded wells. The '502 claims additionally require the drift layer itself to be graded. That is squarely disclosed:

• DE 10345347 A1 (Atmel) is titled, in substance, a method of making a lateral drift region dopant profile DMOS transistor — a drift region with a deliberately engineered dopant profile.
• US 5,637,898 (North Carolina State) discloses vertical FETs with a shaped drift/depletion structure achieving improved breakdown voltage and low on-state resistance — field-shaping in a drift region.
• US 4,907,058 and US 4,684,971 supply the how-to-make-it teaching: retrograde/graded doped regions formed by controlled ion implantation — routine fabrication by the 1990s.

Motivation. Grading a drift region to shape the internal field is a known technique directed at the same problem (field/carrier control) as grading a well. A POSITA would have had every reason to apply it to the single epitaxial drift layer of a Kawagoe- or Wieczorek-type CMOS device, with predictable results.

Ground 5 — Objective-indicia / motivation art: Globitech '943, LSI Logic '715, IDT '093, Wang & Agrawal

These references establish that, before the 2004 priority date, "keep minority carriers away from sensitive surface nodes by engineering the epi/well doping" was an established, well-understood objective in SRAM, DRAM, and radiation-hardened design. Wang & Agrawal, Single Event Upset: An Embedded Tutorial (Ex. 1013) is particularly relevant to the '502 specification's own discussion of "Spurious minority carriers … generated by clock switching in digital VLSI logic and memory IC'S." These references supply the motivation that KSR requires and defeat any argument that the patentee's stated problem was unrecognized.

Ground 6 — Bipolar graded-base art as a §103 admission

US 5,448,087 (TRW, graded base doping), US 5,329,144 (AT&T, specific graded base structure), and US 6,211,028 (TSMC, hi-lo base profile) all predate the priority date and disclose dopant gradients creating an aiding drift field.

Notably, the '502 specification concedes this: "Efforts have been made in graded base transistors to create an 'aiding drift field,' to enhance the diffusing minority carrier's speed from emitter to collector." A patentee's admission that the operative principle was known in the art is properly weighed under §103. The only asserted novelty is transplanting that known principle into the well/drift layer of a CMOS (or IGBT) device — a transplant KSR makes obvious where, as here, the field and the problem are the same.

IV. Claim-by-Claim Disposition

On "single drift layer": the singular "single" is a numerosity limitation, not a technical advance. Both Kawagoe's epi layer 2E and the '502 family's own Fig. 3(a) substrate use a single epitaxial layer over bulk. Neither the claims nor the specification articulate any reason why a POSITA would have used more than one; absent teaching otherwise, reciting "one" where the art shows one is not a patentable distinction.

V. Secondary Considerations (Greenthread's Rebuttal) — and Why It Was Insufficient

Greenthread's principal non-obviousness evidence across the family has been licensing: eight major semiconductor companies (including Kioxia, SK Hynix, Intel, Micron, Western Digital, ams-OSRAM) paying >$50 million total, some allegedly approached through RPX without litigation threat, which PO argued is "especially probative" under Transocean Offshore Deepwater Drilling, Inc. v. Maersk Drilling USA, Inc., 699 F.3d 1340 (Fed. Cir. 2012). PO also argued:

• "Graded" ≠ "aided." Greenthread argued the claims require actual carrier movement, not merely a structural gradient.
• Wrong location. Kawagoe's carriers are not in the active region; the reference does not address surface carriers at all.
• Hindsight. Petitioners relied on post-hoc expert testimony and, for some arguments, a post-priority-date reference (Wang).

These arguments were rejected in the Board decisions retrieved: the ParkerVision framing (capability, not actual movement) was adopted; the applicant's own "inherent 'built-in' unidirectional electric field" prosecution statement was used against PO; and licenses entered to settle litigation carry diminished weight (Fox Factory Inc. v. SRAM LLC, IPR2017-00472). I found no evidence of copying, unexpected results, long-felt-but-unmet need, or industry skepticism in the record reviewed.

VI. Bottom Line

On the §103 record, the '502 claims are vulnerable on multiple independent grounds, with a clear hierarchy:

1. Claim 7 falls to Kawagoe alone — one reference, one device, mapped element-by-element, motivated by the identical soft-error problem, and buttressed by the applicant's own inherency admission.
2. Claim 1 falls to Kawagoe (or Wieczorek+Wolf) in view of Micron US 6,310,366 and/or Jastrzebski US 4,481,522, which expressly teach minority carriers moving from the substrate toward the surface via a graded/retrograde dopant field.
3. All dependent claims fall as field-of-use limitations once Wolf & Tauber, Silverbrook, Jastrzebski, and Micron '366 supply the CPU/DRAM/flash/imager applications. Claim 2 and 8 add no separate limitation.
4. Every motivating rationale recognized in KSR is present: same field of endeavor, same disclosed problem, known technique applied to a similar device, predictable result, and a standard handbook (Wolf) a POSITA would consult. No reference teaches away.

Practical significance: the '502 patent expired September 3, 2024 ("Expired – Lifetime," 8th-year fee paid 2023-03-16). Nothing here can support forward-looking injunctive relief; the IPR validity record and any past damages are what remain live in the pending 2026 appeals.

VII. Caveats and Limits on This Analysis

1. I could not retrieve the IPR2024-00018 Petition or its Final Written Decision as primary documents in this session. The ground-by-ground structure above is synthesized from (a) the IPR2024-00018 exhibit list, (b) the Board's Final Written Decisions and briefing in the sibling proceedings IPR2023-01242/-01243/-01244, which share the identical '502 specification and the identical graded dopant concentration to aid carrier movement dispute, and (c) the March 25, 2026 Director Review denial covering IPR2024-00018. The specific grounds, claim scope, and precise reference combinations actually asserted in IPR2024-00018 may differ from those above. Treat this as an analyst's reconstruction, not a quotation of the agency record.
2. "Payne." The earlier section listed Payne among the '502 grounds. Payne appears in the Greenthread demonstratives for the '842/'222 family (IPR2023-0124x), not in the '502 exhibit list I retrieved. I cannot confirm Payne was a ground in IPR2024-00018 — flagging rather than resolving.
3. Kawagoe's gradient direction. Kawagoe's wells are described as having concentration lowered in the depthwise direction from the principal surface, while some retrograde-well art (e.g., Micron '366) describes the opposite polarity (lowest at surface, highest at bottom). The two gradients produce fields with different effects on a given carrier type. Petitioners and the Board treated the Kawagoe gradient as meeting the "aid carrier movement from surface to substrate" limitation; Greenthread disputed this vigorously on both location and physics grounds. Do not treat the mapping as settled engineering fact — it is the precise issue on appeal.
4. Forward citations are not prior art. The "Families Citing this family (69)" list on the page (all Suvolta/Mie Fujitsu/Taiwan Semiconductor/KR/JP post-2004 entries) post-dates the priority date and cannot support any §103 ground.
5. The patent section of the analysis is on firm ground; the litigation-procedural section is not. Any docket entry dated after April 26, 2026 cannot be confirmed as of this analysis date — the earlier section flagged a 05/20/2026 CAFC filing date for case 26-1847 that is inconsistent with the stated current date. I did not retrieve a USPTO PatentCenter or PACER primary document, and where the supplied full patent text conflicts with search results, the patent text controls.
6. Current beneficial ownership is not established by the recorded security interests (Patent Capital Funding 2018 – Series 1-A, LLC; EC Holdings Series LLC – Series A), which are liens, not ownership transfers.