US 8569800

I'll run searches for the patent number and any 2026 CAFC activity tied to it.

Search results and caveats first

What I could and could not verify. My searches did not return any USPTO Patent Public Search record, PACER docket, or CAFC 2026 docket entry for US 8,569,800. A search for Furukawa Electric "8569800" returned zero results, and general searches for the number surfaced only unrelated documents. I therefore cannot confirm or deny any 2026 Federal Circuit activity involving this patent. If such an appeal exists, it would have to be confirmed against PACER/CAFC docket search and the USPTO Patent Center, which I cannot query authoritatively here. Treat the absence of litigation findings below as "not found," not as "none exists."

Important false-positive warning. Several searches returned a different "'980 patent" — a security-monitoring/video-surveillance patent involved in IPR2016-01041 (cited against Monroe '183). That is not US 8,569,800 and has nothing to do with this case. Do not conflate the two.

The bibliographic and claim data below is drawn from the authoritative full text supplied (Google Patents rendering of US8569800B2, fetched 2026-09-24), which I treat as controlling.

---

US 8,569,800 B2 — bibliographic data

Priority chain (per the "Cross-Reference" paragraph): JP 2010-083172 (filed 2010-03-31), which itself is based on JP 2009-087361 (filed 2009-03-31); and U.S. Provisional 61/406,399 (filed 2010-10-25).

Family: sibling application US 13/075,843 → US 8,450,782 B2 ("Field effect transistor, method of manufacturing field effect transistor, and method of forming groove"); JP 2010258441 A and JP 5697012 B2.

Cited prior art (per the record): WO 2003/071607 A1 (Furukawa, "GaN FIELD-EFFECT TRANSISTOR"); US 7,038,253 B2 (same family, normally-off GaN FET); US 2009/0242938 A1 (Niiyama); US 2011/0241088 A1 (Furukawa sibling). Non-patent literature: D. Visalli et al., "AlGaN/GaN/AlGaN Double Heterostructures on 4 inch Si Substrates…", Solid State Device Meetings 2008, No. G-1-2.

---

Abstract (as published)

A field effect transistor includes: a buffer layer that is formed on a substrate; a high resistance layer or a foundation layer that is formed on the buffer layer; a carbon-containing carrier concentration controlling layer that is formed on the high resistance layer or the foundation layer; a carrier traveling layer that is formed on the carrier concentration controlling layer; a carrier supplying layer that is formed on the carrier traveling layer; a recess that is formed from the carrier supplying layer up to a predetermined depth; source/drain electrodes that are formed on the carrier supplying layer with the recess intervening therebetween; a gate insulating film that is formed on the carrier supplying layer so as to cover the recess; and a gate electrode that is formed on the gate insulating film in the recess.

---

Plain-language overview of the independent claims

Claim 1 — the "recess into the high-resistance/foundation layer" embodiment (MOSHEMT-type device).

A nitride-semiconductor field-effect transistor built as a vertical stack: substrate → buffer layer → a "high resistance layer or foundation layer" (e.g., Mg- or carbon-doped GaN) → a carbon-containing carrier concentration controlling layer → carrier traveling layer (e.g., undoped GaN) → carrier supplying layer (e.g., AlGaN). A recess is etched from the carrier supplying layer down into the high resistance/foundation layer — i.e., the groove goes deep enough to pass through the carbon-containing layer, splitting it into left and right segments (104a/104b in FIG. 1). Source and drain electrodes sit on the carrier supplying layer on either side of that recess; a gate insulating film covers the recess; and the gate electrode sits on the insulating film within the recess (MOS gate). The final limitation requires that the carrier traveling layer contain a two-dimensional electron gas (2DEG) created by the bandgap difference between the traveling layer and the supplying layer.

The technical point of the claim is the carbon-doped "carrier concentration controlling layer": its carbon compensates carriers in the lower tail of the 2DEG distribution, lowering total 2DEG concentration (helping RESURF balance / withstand voltage) while leaving carrier density near the AlGaN/GaN interface nearly unchanged (preserving low on-resistance).

Claim 11 — the second independent claim (recess at least through supplying + traveling layers, with a numeric carbon range).

Same stack and same 2DEG requirement as claim 1, but with two differences: (a) the recess is formed at least in the carrier supplying layer and the carrier traveling layer (it need not necessarily reach the high resistance/foundation layer), and (b) the claim affirmatively recites that the carbon concentration of the carrier concentration controlling layer is from 5×10¹⁷ cm⁻³ to 5×10¹⁹ cm⁻³. That numeric range is what is described in the specification and in the working example (carbon stepped down from 5×10¹⁹ to 5×10¹⁷ cm⁻³ across a 20 nm layer).

Dependent claim structure:

• Claims 2–10 depend from claim 1; claim 21 also depends from claim 1. Notable dependents: carbon range 5×10¹⁷–5×10¹⁹ cm⁻³ (cl. 2); 2DEG concentration 1×10¹²–1×10¹³ cm⁻² (cl. 3); high-resistance/foundation layer of Mg- or C-containing GaN (cl. 4); undoped GaN traveling layer 2–500 nm (cl. 5); AlGaN supplying layer with Al ratio 0.01–0.99 and thickness 1–50 nm (cl. 6); an added heterojunction forming layer between buffer and high-resistance/foundation layer with a different bandgap (cl. 7; AlGaN with Al ratio 0.01–0.5 in cl. 8, the "double heterostructure" embodiment of FIG. 5); carbon concentration monotonically decreasing from the high-resistance/foundation side toward the traveling-layer side (cl. 9); carbon in the traveling layer below 5×10¹⁷ cm⁻³ (cl. 10); and 2DEG generated near the traveling/supplying interface (cl. 21).
• Claims 12–20 depend from claim 11 and mirror the same limitations (12–18, 20), with claim 19 adding that the recess extends downward from the carrier supplying layer into the high resistance/foundation layer.

---

Uncertainty / verification notes

• CAFC 2026 activity: Not found in any search I ran. I have no authoritative information on any 2026 appeal, IPR appeal, or district-court case involving US 8,569,800. Do not report any such proceeding without independent confirmation from PACER, the CAFC docket, or USPTO Patent Center.
• The patent itself (title, numbers, dates, claims) is well supported by the authoritative full text; the "Active" status and 2011/2013/2031 dates are as stated in that source. Maintenance-fee entries (2017 4th year, 2021 8th year, 2025 12th year) are as listed.
• Numeric/identifier values are reproduced as written in the source, without correction; note the source itself contains apparent typographical errors (e.g., "PIE" for RIE, "hetero unction," "nigh resistance layer," "u-GaN," "104 h/105 h" typo) — these are OCR/typing artifacts in the rendering, not claim language.

I'll search for litigation involving US patent 8569800 across patent litigation sources.

The initial results reference different patents (e.g., Deere's 6,052,980). Let me refine my search to target the specific patent number.

Let me try more targeted litigation databases and the sibling/family patents.

Litigation Search Results for U.S. Patent No. 8,569,800

Bottom line: I found no litigation — past or pending — specifically asserting U.S. Patent No. 8,569,800. Based on my searches of patent litigation sources, there is no record of any district court, ITC, or Federal Circuit case in which this patent has been asserted as a patent‑in‑suit.

What I searched and what I checked

• Google Patents page for US8569800 (the authoritative record provided).
• Unified Patents litigation portal references (portal.unifiedpatents.com/litigation).
• General web searches for the patent number in combination with litigation, infringement complaint, and its assignee (Furukawa Electric Co., Ltd.).
• Searches keyed to the patent's family and technology area (the sibling/related patent US 8,450,782 in the same family, titled "Field effect transistor, method of manufacturing field effect transistor, and method of forming groove").

Important disambiguation (per your strict rule)

Several results returned references to a "'980 patent" — but these are NOT U.S. 8,569,800, and I am not treating them as such. They involve a completely different patent:

• Deere & Company v. Bush Hog, LLC, et al., No. 3:09-cv-00095 (S.D. Iowa), and the appeal Deere & Co. v. Bush Hog, LLC, 703 F.3d 1349 (Fed. Cir. 2012) / appeal No. 14-1403. The "'980 patent" there is U.S. Patent No. 6,052,980 (rotary mower cutting deck) — unrelated to U.S. 8,569,800.

I mention this only to flag that the near-identical shorthand ("'980") is a false positive, not a hit on your patent.

About U.S. 8,569,800

• Title: Field effect transistor (GaN-based HEMT/MOSHEMT nitride semiconductor device)
• Assignee: Furukawa Electric Co., Ltd.
• Priority date: 2009-03-31; filed: 2011-03-31; granted: 2013-10-29
• The Google Patents record shows only patent "Cited By" references (e.g., US9324808, US20160118488A1, US20220130990A1) — these are later patents that cite US 8,569,800 as prior art; citation is not litigation.

Caveats on confidence

I could not confirm a definitive "no litigation" from a dedicated litigation database (e.g., Docket Alarm, RPX Insight, Lex Machina) because my searches of those sources did not return a case list keyed to this specific number, and I was not able to complete additional targeted queries. My conclusion is therefore:

• No litigation found in the sources I could access.
• If Furukawa Electric (or a successor/assignee) has ever enforced this patent, or if it has been the subject of a declaratory-judgment or IPR/PGR challenge, such a case did not surface in these searches and I cannot confirm it exists.

If you need a definitive certified answer, I'd recommend running the patent number directly through Unified Patents' litigation portal APIs (https://portal.unifiedpatents.com/litigation/caselist?patents=8569800), RPX Insight, Lex Machina, or PACER's party/patent index — none of which I was able to query exhaustively here.

I'll verify the structured-data default against live sources before drafting.

PTAB Proceedings — U.S. Patent No. 8,569,800 ("Field effect transistor," Furukawa Electric Co., Ltd.)

Verification note: The authoritative source for this analysis is the structured "PTAB proceedings on file" block (USPTO Open Data Portal ingest), which returns zero AIA trial proceedings for US 8,569,800. I ran live web searches over PTAB petition/decision repositories and general web indices to catch anything the ODP ingest missed; those searches surfaced no IPR, PGR, or CBM naming this patent. One search returned a "856 patent" IPR discussion, but that is iRhythm's U.S. Patent No. 9,xxx,856 — a different patent entirely, and I have not imported it. I also could not complete a dedicated ex parte reexamination lookup (a non-AIA, non-PTAB vehicle that would not appear in the ODP AIA-trial field); treat that as an open item rather than a cleared one.

---

Proceedings overview

Total AIA trial proceedings on file: 0 — 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denied. The defensive posture this gives a defendant is therefore not "hardened patent" and not "claims are dead": it is "completely unchallenged at the Board — all 21 claims are live, untested, and available to be asserted, and the full prior-art universe remains open to you." That cuts both ways: nothing is canceled against you, but nothing is estopped in your favor, either.

No proceedings to report

There is no ### IPR####-##### — {Petitioner} v. Furukawa Electric Co., Ltd. entry to write, because no petition has ever been filed and no post-grant trial has ever been instituted against this patent. I will not manufacture a proceeding number, panel, or Final Written Decision to fill the template.

What the public record does show about this patent's exposure:

---

Strategic summary

Claim status: everything is UNTESTED. There are 21 claims, in two independent-claim families — claim 1 (recess extending from the carrier supplying layer into the high resistance/foundation layer; two-dimensional electron gas at the carrier-traveling/carrier-supplying heterojunction) and claim 11 (recess formed at least in the carrier supplying and carrier traveling layers, plus the express numeric limitation that the carbon concentration of the carrier concentration controlling layer is 5×10¹⁷ cm⁻³ to 5×10¹⁹ cm⁻³). Claims 2–10 depend from claim 1; claims 12–21 depend from claim 11; claim 21 is a dependant of claim 1. No claim has been canceled, confirmed, or even construed by the Board. Contrast this with the scenario the task template anticipates — there is no FWD to link, because none exists.

Estoppel landscape is empty, which helps you. Because no IPR/PGR was ever instituted, 35 U.S.C. § 315(e)(2) estoppel is not triggered against anyone on this patent. No petitioner, and no privy of any petitioner, is barred from raising any § 102 or § 103 ground. Practically, a defendant today may build an invalidity case on any prior art it can find, including art already of record (WO 2003/071607, US 7,038,253, US 2009/0242938, the Visalli double-heterostructure paper), and may run that case in both the district court and the PTAB without estoppel risk from a prior PTAB proceeding on this patent. Conversely, you cannot point to a prior petitioner's win to shorten the fight — there is none.

Pattern signals are null. No petitioner has filed once, let alone multiple times; there is no defensive aggregator (Unified Patents, RPX, etc.) in the chain; the patent owner has never had a PTAB ruling to appeal and therefore no Federal Circuit appeal exists. Furukawa is a large operating company, not a litigious NPE, and the public record shows it monetizing power-GaN IP through license/sub-license arrangements rather than a campaign of infringement suits on this patent. One caution for a would-be petitioner: the asserted priority date sits at 2009-03-31, while the U.S. application was not filed until 2011-03-31, with the intervening benefit chain running through JP 2010-083172 (2010-03-31) and provisional 61/406,399 (2010-10-25). Whether the 2009-03-31 date is fully perfected for every claim — and thus whether 2009–2010 art qualifies as § 102(a)(1)/pre-AIA § 102(b) art or instead qualifies as § 102(a)(2)-type art — is a priority question worth developing, and it has never been litigated because no one has ever challenged the patent.

---

Recommended next steps

If you are a defendant and you hold a demand letter citing this patent: understand that you have no ready-made invalidity judgment to lean on. Nothing is canceled; no FWD exists to quote or link. Your invalidity case is built from scratch, and the practical consequences are:

1. Watch the § 315(b) clock. If you have been served with a complaint alleging infringement of the '800 patent, you have one year from service to file an IPR. That is your only realistic window for a PTAB challenge, and it cannot be revived after it lapses. Fee-paid maintenance through 2025 confirms the patent is in force to 2031-03-31, so there is no expiration-based walkaway defense.
2. If a petition is filed, the trial-stage milestones run on statute. The Board must decide institution within 3 months of the patent owner's preliminary response (or its due date); the statutory 1-year trial deadline from institution then applies, with the FWD plus any appeal to the Federal Circuit. Date the schedule from the petition's filing date once a proceeding number issues — I cannot supply dates for a proceeding that does not exist.
3. Start the prior-art search on the two independent claims separately. Claim 1's recess-into-the-high-resistance-layer geometry is a structural limitation amenable to art-of-record combinations; claim 11's numeric carbon-concentration range (5×10¹⁷–5×10¹⁹ cm⁻³) is the kind of parameter limitation that invites a § 103 obviousness attack using a carbon-doped-GaN buffer reference plus a recessed-gate AlGaN/GaN HEMT reference. The Visalli double-heterostructure NPL cited on the face of the patent is a natural secondary reference.
4. Check the ex parte reexamination gap. I flagged above that I could not verify the reexam status of this patent's family. If an ex parte reexam (or a related-family reexam) is pending or has issued a certificate, that changes the claim scope and the estoppel picture. Confirm this via USPTO Public PAIR / Patent Center and the reexamination certificate database before finalizing any invalidity theory.
5. Absence of PTAB activity is itself the signal here — and it is a mixed one. It is not "the patent has survived two IPRs and is hardened"; there have been zero IPRs. The most defensible reading is that this patent has simply never been worth challenging to anyone yet — it is a 2013-granted power-GaN device patent asserted, if at all, through licensing, and challengers have had no litigation-driven reason to spend the money. If a demand letter has now landed on your desk, you may well be the first party with an incentive to test it. If you do, you will be writing the first page of this patent's PTAB history, and there is no § 315(e) estoppel, no adverse FWD, and no General Plastic multiple-petition problem standing in your way.

Bottom line for the client: No PTAB proceedings on US 8,569,800. All 21 claims are alive and untested; no estoppel attaches to anyone. The patent runs to 2031-03-31 with fees current. Treat this as a live, uncontested patent, not a wounded one — and preserve your § 315(b) one-year IPR window from any date of service.

I'll verify the assignment record through the USPTO Assignment Center and cross-reference sources.

Let me try to locate the recorded assignment document itself to capture the correspondent of record.

Let me attempt to retrieve the specific recorded assignment document for reel 026054/0852.

Ownership & Assignment Analysis — US 8,569,800 B2

Scope note / data limitation: I was able to pull the full issued patent text (authoritative, from Google Patents) and the legal-events record reproduced there. I was not able to retrieve the underlying recorded assignment document for reel 026054/0852 from the USPTO Assignment Center or the legacy-assignments PDF mirror; the record below is drawn from the assignment metadata carried in the Google Patents legal-events block. Where the Assignment Center would surface the correspondent of record, I could not verify that field and I state so rather than guess.

---

Inventors

Pattern check — departures: No abnormal pattern. The "all inventors gone within 12 months" fire-sale precursor is not present; the evidence runs the other way. Kokawa and Iwami are still filing as Furukawa Electric inventors in 2016 applications (US 2016/0225889), and every assignor signed the same-day assignment to the company rather than to a third-party acquirer. Inventors are Japanese residents, consistent with the company's Tokyo research organization.

---

Original assignee

The Furukawa Electric Co., Ltd. (Furukawa Denki Kogyo K.K.), 6-1 Marunouchi 2-chome, Chiyoda-ku, Tokyo, Japan — the assignee named on the face of US 8,569,800 and the assignee in the sole recorded assignment.

• Line of business: Diversified Japanese electrical/industrial manufacturer (TSE-listed); power and optical cable, automotive systems, copper/electronic materials, and a compound-semiconductor program that includes GaN-on-silicon power field-effect transistors — the exact device class claimed here (hybrid MOSHEMT normally-off FET).
• Product embodiment: The specification reports a measured fabricated device (on-resistance ~5 mΩ·cm², breakdown ~1200 V for the carbon-controlled example vs. ~500 V for the control), so the applicant had a working device. Whether a commercial part shipped under these claims is not verifiable from the patent record; I found no evidence of a productized part-number or of the patent being asserted.
• Current status: Operating; no bankruptcy, dissolution, or assignment of this patent to any acquirer or NPE appears of record. Maintenance fees were paid in 2017, 2021 and 2025 by a large-entity owner — consistent with the original corporate assignee retaining title.

---

Assignment timeline

Only one assignment is of record for this patent. No post-issuance transfer, security interest, name change, or release is recorded.

• 2011-03-24 / 2011-03-25 (executed) / recorded 2011-03-31 — Reel 026054/0852
  • Conveyance: Assignment (ASSIGNMENT OF ASSIGNORS' INTEREST)
  • Assignor: Nariaki Ikeda; Takuya Kokawa; Masayuki Iwami; Sadahiro Kato
  • Assignee: FURUKAWA ELECTRIC CO., LTD. (6-1 Marunouchi 2-chome, Chiyoda-ku, Tokyo, Japan)
  • Correspondent: Not retrievable. The Google Patents legal-events block reproduces the reel/frame, conveyance, assignors and signing dates but omits the "correspondence to be mailed to" block that the Assignment Center exposes. I could not open the reel 026054/0852 document image, so I decline to name an attorney. (For context only, and not attributed to this reel: a separate Furukawa Electric recording on reel 055660/0866, recorded 2021-03-21, lists Birch, Stewart, Kolasch & Birch, LLP, Falls Church VA, as correspondent for a different Furukawa application. That is a single data point on a different matter and does not satisfy the recurrence test.)
  • Context: Ordinary inventor-to-employer assignment of rights on filing — routine corporate prosecution, not an acquisition, fire-sale, reorg, or transfer to an asserter.

Related but distinct events (no ownership change):

Family note: sibling application US 13/075,843 → US 8,450,782 (same Furukawa Electric family, priority 2009-03-31). Also Furukawa-owned; no third-party chain.

---

Timeline diagram

timeline
    title Ownership of US 8569800
    2009 : Priority date
    2010 : US provisional filed
    2011 : Application filed
         : Inventors assign to Furukawa Electric
         : Recorded reel 026054 frame 0852
    2013 : Patent issued
    2017 : Maintenance fee year 4 paid
    2021 : Maintenance fee year 8 paid
    2025 : Maintenance fee year 12 paid

---

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The only recorded conveyance is reel 026054/0852 (recorded 2011-03-31), inventor → The Furukawa Electric Co., Ltd. No LLC, no "IP/Holdings/Licensing/Ventures" suffix, no registered-agent address appears anywhere in the chain.

2. Known asserter in the chain — not present. The assignee of record is Furukawa Electric, an operating manufacturer. It matches none of the listed asserters (Acacia, Marathon, IV, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, DGC, Spangenberg entities). I did not find the patent in any Unified Patents or RPX high-frequency-plaintiff dataset during this review; I also found no litigation naming the patent in the sources indexed.

3. Repeat correspondent across the chain — unclear. There is only one link in the chain, so the recurrence test cannot be satisfied by construction, and the correspondent field for reel 026054/0852 was not retrievable. The 2021 Furukawa recording by Birch, Stewart, Kolasch & Birch (reel 055660/0866) concerns a different application and is not evidence about this reel. Recording here.

4. Cascading transfers — not present. One assignment in 15+ years; no consecutive LLC-to-LLC hops, no common-principal addresses (contrast the Acacia/Monarch-style chains visible on reels 044625/0315, 051192/0596, 051238/0718, 053654/0254, 055100/0624, 055101/0608 in unrelated records).

5. Pre-litigation transfer — not present. Nothing to time: no post-issuance assignment exists, and no infringement suit naming US 8,569,800 was identified. The 2011-03-31 recording predates issuance by ~2.6 years and is a filing-time employer assignment.

6. Bankruptcy fire-sale — not present. No bankruptcy of The Furukawa Electric Co., Ltd. of record; no sale of this patent in any proceeding; the 12th-year maintenance fee was paid in 2025 by a large entity.

7. Privateering — not present. No transfer from Furukawa to any NPE asserting on Furukawa's behalf; the company kept the asset.

8. Defensive aggregator — not present. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates where it began: with the operating original assignee.

Corroborating "still owned by the original assignee" indicator: the 2017/2021/2025 maintenance-fee payments at large-entity rates, with no reassignment entry in between, are the classic signature of uninterrupted original-owner custody.

---

Verdict

Insufficient data (no records beyond the original assignment).

The complete recorded chain of title is a single inventor-to-employer assignment — executed 2011-03-24/25 and recorded 2011-03-31 at reel 026054/0852 — from Ikeda, Kokawa, Iwami and Kato to The Furukawa Electric Co., Ltd. Nothing thereafter has been recorded, and Furukawa Electric has continued paying maintenance fees at large-entity rates through the 12th-year payment on 2025-04-16, so the operating original assignee still holds title. No NPE signal is present on any of the eight tests, and the absence of post-issuance assignments is itself the finding: this is a clean, single-owner, operating-company patent record, not a patent-troll chain. The one gap in the record is the correspondent-of-record field on reel 026054/0852, which I could not retrieve and therefore did not name.

Verification links:

• USPTO Assignment Center (search by patent number 8569800): https://assignmentcenter.uspto.gov/
• Legacy assignment search: https://assignment.uspto.gov/patent/index.html
• Patent and full legal-events record: https://patents.google.com/patent/US8569800/en

Caveat: Because I could not open the reel 026054/0852 document image, the correspondent, the assignee's exact recorded address, and any domestic-representative designation on that cover sheet remain unverified. If you can pull the reel image, that one field would close the only evidentiary gap in this chain.

I'll verify the patent and gather details on each cited reference via search.

Let me gather details on the remaining cited references.

Prior-Art Analysis — US 8,569,800 B2 ("Field effect transistor")

Scope and method

I re-ran the search against the patent number itself (per the instruction not to return similar numbers) and against each cited reference. My verification path was the Google Patents / FreePatentsOnline / USPTO.report mirrors of USPTO data rather than a direct Patent Public Search query — I could not execute an authenticated USPTO Patent Center lookup in this session, so the citation set below is taken from the "Patent Citations," "Non-Patent Citations," and "Family Cites Families" tables of US 8,569,800 B2 as rendered on Google Patents (fetched 2026-09-24), consistent with the earlier-generated bibliographic section.

Two cautions carried forward from the prior section:

1. The "'980 patent" false positives in general web search (e.g., the surveillance patent in IPR2016-01041) are not this patent. Ignore them.
2. I found no USPTO/CAFC/PACER reexamination, IPR, or 2026 appellate activity for US 8,569,800. Nothing below is based on litigation findings.

Applicable law. This patent has a 2009/2010 priority chain and a 2011 US filing date, so pre-AIA 35 U.S.C. § 102 governs. Because foreign priority does not move the § 102(b) critical date, the § 102(b) one-year critical date is approximately 2010-03-31 (one year before the 2011-03-31 US filing).

The novelty-bearing limitation. For every § 102 question below, the dispositive element is the "carrier concentration controlling layer containing carbon" interposed between the high-resistance/foundation layer and the carrier traveling layer (claim 1), and its recited carbon range of 5×10¹⁷–5×10¹⁹ cm⁻³ (claim 11 / claim 2). This layer is the stated point of the invention — carbon compensates the lower tail of the 2DEG distribution to preserve RESURF balance without degrading near-interface carrier density.

---

The cited reference set (as it appears on the face of US 8,569,800)

Only four patent documents and one NPL item are cited on the face of the patent — a notably small set, which is itself informative: the examiner did not cite any carbon-doped-buffer or carrier-compensation art.

---

Reference-by-reference

A. WO 2003/071607 A1 — "GaN FIELD-EFFECT TRANSISTOR"

• Full citation: WO 2003/071607 A1, The Furukawa Electric Co., Ltd., "GaN field-effect transistor," published 2003-08-28 (priority 2002-02-21; PCT filing 2003-02-21). Corresponds to the US national-phase/continuation that issued as US 7,038,253 B2 (ref. B).
• Description: The acknowledged starting point of US 8,569,800. Discloses the "hybrid MOSHEMT" normally-off FET: an undoped GaN carrier traveling layer and an AlGaN carrier supplying layer sequentially stacked on a p-type GaN layer; a 2DEG formed near the traveling/supplying heterojunction; and a recess etched through the carrier supplying and traveling layers near the gate to produce normally-off operation. The p-type layer + carrier traveling layer form a RESURF structure that relieves field concentration.
• § 102 analysis: This is the closest structural reference and the strongest § 102 candidate, but it is not anticipatory of claims 1 or 11. It discloses essentially every element of claim 1's stack except the carbon-containing carrier concentration controlling layer, and while its recess reaches the p-type layer, its disclosure does not recite a discrete carbon-doped controlling layer between the p-layer and the traveling layer. The patent's own specification characterizes WO '607 as the prior approach and identifies the un-balanced RESURF problem (elevated 2DEG concentration collapsing withstand voltage) as the problem the invention solves — i.e., the applicant conceded WO '607 as background rather than as anticipating art. Maps to: claims 1/11 skeleton (substrate, buffer, p-GaN layer, traveling layer, supplying layer, recess, source/drain, gate insulator, gate in recess, 2DEG) — anticipated only if the carbon layer limitation were read out of the claim, which it cannot be. Best characterized as § 103 base art.

B. US 7,038,253 B2 — "GaN-based field effect transistor of a normally-off type"

• Full citation: US 7,038,253 B2, The Furukawa Electric Co., Ltd.; issued 2006-05-02; priority 2002-02-21 (US app. 10/920,203, filed 2004-08-18). Same family as ref. A.
• Description: Normally-off GaN FET with: a channel portion of i-GaN or p-GaN electrically connected to source/drain; first and second electron supply portions of wider-bandgap AlGaN joined to the channel and separated from one another; an insulating layer on the channel surface between the electron supply portions; and a gate electrode on the insulating layer. Fabrication etch-removes the electron supply layer down to the channel layer at the gate region, forming the recess. The p-type impurity may be Mg, Zn, or C, with carrier concentration ~1×10¹⁹ to 5×10¹⁶ cm⁻³. Two-dimensional electron gases are described as disconnected until gate bias is applied.
• § 102 analysis: Stronger § 102 candidate than ref. A because it is an issued US patent qualifying as § 102(b) art (issued 2006, well before the 2010-03-31 critical date) and because it explicitly names carbon among the p-type dopants — which creates a superficial resemblance to the claimed "carbon-containing" layer. However, its carbon is a p-type dopant in the channel/p-GaN layer, not a discrete carrier concentration controlling layer positioned between a high-resistance layer and the carrier traveling layer, and the reference does not attribute carrier-compensation or RESURF-balancing function to carbon. Maps to: claims 1/11 structural skeleton and the "recess through the supplying layer" feature — not anticipatory of claims 1 or 11 for want of the claimed carbon controlling layer, and not anticipatory of claim 4 (which requires the high resistance/foundation layer, not the channel, to contain Mg or C). § 103 base art, and the most likely reference an examiner would combine with a carbon-doped-buffer teaching.

C. US 2009/0242938 A1 — "Field effect transistor" (Niiyama)

• Full citation: US 2009/0242938 A1, Yuki Niiyama (Furukawa Electric Co., Ltd.), published 2009-10-01; filed 2008-03-31; priority JP 2008-094030 (filed 2008-03-31).
• Description: A normally-off III-nitride FET with a GaN electron running layer on a substrate and an electron supplying layer of Al_xGa_(1-x)N, aimed at simultaneously achieving high mobility and high breakdown voltage. Discusses a GaN-based MOSFET threshold-voltage control and the SiO₂/GaN interface. Examiner-cited (marked *).
• § 102 analysis: Because its effective filing date (2008-03-31) precedes the 2009-03-31 priority date of US 8,569,800, it qualifies as pre-AIA § 102(e) art (and, having published 2009-10-01, also § 102(b) art relative to the 2011 US filing). It is relevant background for the high-mobility/high-breakdown-voltage objective and the MOSFET gate-stack, but it discloses no carbon-containing carrier concentration controlling layer, no p-GaN high-resistance layer plus carbon layer combination, and no recess extending into such a layer. Not anticipatory of claims 1 or 11. Relevant as § 103 context and as evidence of the state of the art.

D. US 2011/0241088 A1 — "Field effect transistor, method of manufacturing field effect transistor, and method of forming groove" (US 8,450,782 B2)

• Full citation: US 2011/0241088 A1 (granted as US 8,450,782 B2), Furukawa Electric Co., Ltd., published 2011-10-06; filed 2011-03-30; priority 2009-03-31. Examiner-cited (*).
• Description: Directed to recess/groove formation — a two-stage dry-etch process in which an altered layer is removed at higher bias power, then the groove is formed at lower bias power, to control recess depth reproducibly. This is the same-family sibling of US 8,569,800 (shared 2009-03-31 priority; both are listed under Family ID 43318955).
• § 102 analysis: Because it shares the same earliest priority chain, it does not qualify as § 102(e) prior art against US 8,569,800, and it published after the 2011-03-31 filing date, so it is not § 102(a)/(b) art either. Its real relevance is non-§102: (i) potential obviousness-type double patenting given the common priority and overlapping inventors/assignee, and (ii) it is cited for the recess-forming method context. It does not anticipate any claim of US 8,569,800.

E. D. Visalli et al. (NPL) — "AlGaN/GaN/AlGaN Double Heterostructures on 4 inch Si Substrates…"

• Full citation: D. Visalli, M. Van Hove, J. Derluyn, S. Degroote, M. Leys, K. Cheng, M. Germain, G. Borghs, "AlGaN/GaN/AlGaN Double Heterostructures on 4 inch Si Substrates for High Breakdown Voltage Field-Effect Transistors with low On-Resistance," in Proc. 40th Int'l Conf. on Solid State Devices and Materials (SSDM 2008), No. G-1-2, pp. 148–149 (Ibaraki, Japan, Sept. 23, 2008). A journal counterpart appeared in Jpn. J. Appl. Phys. 48, 04C101 (published online 2009-04-20).
• Description: AlGaN/GaN/AlGaN double heterostructure field-effect transistors (DHFETs) on 4-inch Si substrates, using an AlGaN buffer/barrier to confine the 2DEG and suppress electron spillover, achieving high breakdown voltage with low on-resistance (e.g., ~830 V at 8 µm L_GD; R_on ~6.2 Ω·mm). It is cited in the patent's Background as the origin of the "double hetero structure" approach.
• § 102 analysis: Qualifies as a § 102(b) printed publication (Sept. 2008, before the critical date). It is relevant only to the double-heterostructure subject matter of dependent claims 7/8 and 16/17 (the "heterojunction forming layer" of different bandgap energy; AlGaN with Al ratio 0.01–0.5). It is not anticipatory of claims 7/16 either, because those claims depend from claims 1/11 and thus incorporate the carbon-containing carrier concentration controlling layer that Visalli does not describe. Visalli also does not disclose the p-GaN/Mg high-resistance layer or the MOS gate-in-recess structure. § 103 art supporting the double-heterostructure limitation.

---

§ 102 bottom line

No reference cited on the face of US 8,569,800 anticipates independent claim 1 or claim 11 under 35 U.S.C. § 102. The record is unambiguous on this point: every reference is missing the carbon-containing carrier concentration controlling layer — the single element that distinguishes the claims from the hybrid-MOSHEMT and DHFET prior art.

Practical exposure. Any invalidity theory would run through § 103, not § 102: e.g., WO '607 / US 7,038,253 (hybrid MOSHEMT skeleton with recess + p-GaN RESURF) combined with a carbon-doped-GaN-buffer or GaN:C compensation teaching, optionally plus Visalli for the double-heterostructure dependent claims — with secondary considerations (the measured jump from ~500 V to ~1200 V breakdown at unchanged ~5 mΩ·cm² R_on) cutting against obviousness. The absence of any carbon-doped-buffer reference from the cited set is the most conspicuous gap in the face-of-the-patent prior art.

---

Supplementary: documents cited in family members (not on the face of US 8,569,800)

These appear in the "Family Cites Families" table and were cited during prosecution of related applications; they are listed for completeness and are generally § 102(b) art (pre-2009 publication or pre-2010 grant):

Of these, the Furukawa "semiconductor element" documents (JP 2008205221 A, JP 5064824 B2) and US 9,048,302 B2 are the most likely to bear on the buffer/high-resistance-layer and recess-sidewall aspects, but none is cited on the face of US 8,569,800 and none is known to disclose the carbon controlling layer.

---

Caveats

• Verify against primary USPTO data. I could not execute a direct USPTO Patent Public Search / Patent Center query here. The citation list, dates, and examiner-citation flags are reproduced from the Google Patents rendering of US 8,569,800 B2 and cross-checked against FreePatentsOnline and USPTO.report mirrors. Before relying on this for a legal conclusion, confirm the "References Cited" list in the issued patent PDF (US8569800.pdf) and the file-wrapper citations.
• § 102(e) dependency. My treatment of US 2009/0242938 A1 as § 102(e) art depends on its effective filing date (2008-03-31) preceding the earliest priority date claimed. If the priority chain for US 8,569,800 were construed differently (e.g., only the 2010-03-31 JP filing relied upon), the reference's status shifts, though it would remain § 102(a)/(b)-type art.
• No anticipation found is an analytical conclusion, not a legal opinion. This is a technical mapping of cited references to claim language, not an invalidity opinion.
• Identifiers reproduced literally. No auto-correction has been applied to patent or application numbers; note again that the source text contains apparent rendering typos ("PIE" for RIE, "hetero unction," "nigh resistance layer," etc.), which are artifacts of the rendering, not claim language.

I'll ground the analysis in the prior-art references of record and verify what they teach before building combinations.

US 2009/0242938 A1 (Niiyama) is a significant hit. Let me pull its full disclosure and the remaining family-cited references.

Let me verify the Niiyama grant counterpart, the Furukawa family references, and the general carbon-doped-GaN art.

Obviousness Analysis — U.S. Patent No. 8,569,800 B2 under 35 U.S.C. § 103

---

0. Scope notes and flags before the analysis

Two housekeeping flags.

1. Date inconsistency. This task states the current date as April 26, 2026. The system context and the authoritative patent fetch timestamp I am working from both read 2026-09-24 / 2026-09-29. I proceed on the facts as they are (nothing below turns on which of these is correct), but the discrepancy is noted rather than silently reconciled.
2. Cross-reference to my prior sections — an important development, not a contradiction. In my earlier "Patent summary" I listed US 2009/0242938 A1 (Niiyama) as one of four cited prior-art references, without substance. Having now pulled that reference's disclosure, Niiyama is by far the most consequential § 103 reference in this record — substantially more so than the WO 2003/071607 prior art the specification itself debates. Its granted counterpart is US 8,072,002 B2 (Furukawa Electric). Nothing in my earlier sections is contradicted; they were simply incomplete on this point.

What I am treating as prior art, and its statutory status. The '800 patent's application (US 13/076,980) was filed 2011-03-31, i.e., before the AIA first-to-file changeover. Pre-AIA § 103(a) therefore governs, and pre-AIA § 102 categories apply. That matters specifically for Niiyama: its publication date (2009-10-01) postdates the '800's earliest priority date (2009-03-31), so it is not § 102(a)/(b) art by publication. It is, however, available as pre-AIA § 102(e) art as of its U.S. filing (App. 12/382,941, March 2009 — BEFORE 2009-03-31), and its JP priority (JP 2008-094030, 2008-03-31) independently predates. I flag that if the actual 12/382,941 filing date were after 2009-03-31 the § 102(e) route would close — this should be confirmed against the face of US 8,072,002 B2 before relying on it in a formal proceeding. I could not verify the exact filing date in my searches; I infer March 2009 from the serial series and the 18-month publication window.

---

1. Legal framework applied

Graham v. John Deere Co., 383 U.S. 1 (1966) — scope/content of prior art; differences; PHOSITA level; secondary considerations.

KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) — a combination is obvious where (a) the prior art elements were known, (b) combined according to known methods, (c) to yield predictable results; motivation may come from the nature of the problem, design incentives/market forces, or the interconnected teachings of the references themselves, and need not be found in the references' own words. A finite list of identified, predictable solutions renders the selection obvious.

In re Peterson / In re Woodruff / Titanium Metals — where a claimed numerical range overlaps or abuts an express prior-art range, prima facie obviousness attaches unless the applicant shows criticality. See In re Peterson, 315 F.3d 1325 (Fed. Cir. 2003).

Nexus requirement for secondary considerations — an unexpected result must be measured against the closest prior art, not against a convenient non-analogous control. In re De Blauwe, 736 F.2d 699 (Fed. Cir. 1984).

Level of ordinary skill in the art (PHOSITA). I would define it as: a B.S. or M.S. in electrical engineering, applied physics, or materials science, plus 3–5 years of hands-on MOCVD growth and fabrication of III-nitride (AlGaN/GaN) power devices, including recessed-gate MOSHEMT/HEMT process integration. This is a mature-by-2009 art: AlGaN/GaN HEMTs, p-GaN-gated MOSHEMTs, carbon-compensated GaN buffers, and recessed-gate normally-off designs were all in the literature by 2007–2008.

---

2. Isolating the point of novelty: claim 1 over the admitted prior art

The specification's own "Description of the Related Art" (building on my earlier claims overview) concedes the entire device architecture:

"This field effect transistor has a HEMT structure in which a carrier traveling layer formed of an undoped GaN semiconductor and a carrier supplying layer formed of an AlGaN semiconductor are sequentially stacked on a p-type layer formed of a p-type GaN semiconductor... The field effect transistor forms a recess by etching the carrier supplying layer and the carrier traveling layer only in the vicinity of a gate... the p-type layer and the carrier traveling layer have a RESURF structure."

That concession maps WO 2003/071607 A1 / US 7,038,253 B2 onto essentially every structural element of claim 1:

The single structural delta of claim 1 is the interposition of a discrete, carbon-containing "carrier concentration controlling layer" between the high-resistance (p-type) layer and the carrier traveling layer. The specification defends this on exactly one ground: carbon in that sub-layer compensates the low-energy tail of the 2DEG distribution, lowering total sheet carrier density (preserving RESURF balance → high BV) while leaving carrier density at the AlGaN/GaN interface essentially unchanged (preserving low Ron).

That is the entire inventive contribution. Every obviousness question therefore reduces to: was placing a carbon-containing acceptor sub-layer between a p-GaN layer and an undoped GaN channel an obvious design choice by March 2009?

---

3. The references bearing on that question

3.1 US 2009/0242938 A1 (Niiyama) → US 8,072,002 B2, Furukawa Electric — the critical reference

This is a Furukawa Electric application disclosing a III-nitride normally-off MOSFET that teaches the disputed feature expressly:

• Claim 1 (granted US 8,072,002 B2): electron running layer of GaN; electron supplying layer of AlₓGa₁₋ₓN (0.01 ≤ x ≤ 0.4) "having a band gap energy different from that of the electron running layer and being divided with a recess region having a depth reaching the electron running layer"; source/drain "formed on the electron supplying layer with the recess region in between"; "a gate insulating film layer formed on the electron supplying layer for covering a surface of the electron running layer in the recess region"; "a gate electrode formed on the gate insulating film layer in the recess region."
• Claim 4 (granted): "said electron running layer comprises a lower part layer and an upper part layer formed on the lower part layer and having a density of an acceptor different from that of the lower part layer, said recess region having the depth reaching to the lower part layer."
• Published ¶[0012]: the electron running layer contains "any one of Mg, Be, Zn and C as an acceptor."
• Published ¶[0013]: acceptor density "not lower than 1×10¹⁵ cm⁻³ but not higher than 5×10¹⁷ cm⁻³."
• Published ¶[0014]: the lower/upper split with the recess reaching the lower layer.
• Published ¶[0063]: the rationale — "the threshold voltage... is designed to be determined according to the density of the acceptor ions... regarding the lower part layer 203a. On the contrary, there is influenced on the sheet carrier density of the two dimensional electron gas by the density of the acceptor ions... regarding the upper part layer 203b and 203c. Therefore... by performing an optimization independently for the addition density of Mg regarding the lower part layer and the upper part layer, it becomes able to control independently... the withstand voltage thereof and the threshold voltage thereof."

Why this is devastating. Niiyama's "upper part layer of the electron running layer, having a different acceptor density, positioned between the high-acceptor (high-resistance) lower layer and the 2DEG-containing region, with the recess passing through it into the lower layer" is the claimed carrier concentration controlling layer in substance. The only differences are: (i) the '800 renames the upper split-off portion a separate "carrier concentration controlling layer" and adds a discrete "carrier traveling layer" above it; and (ii) Niiyama's worked example emphasizes Mg rather than carbon (while expressly listing C in the alternative).

Note also that Niiyama's paragraph [0063] independently states the very principle the '800 patent asserts as its point of novelty: that acceptor density in a sub-layer below the channel controls the 2DEG sheet charge and hence withstand voltage, decoupled from the threshold-determining layer below it. A PHOSITA reading Niiyama in 2009 would expect precisely the '800's result.

3.2 WO 2003/071607 A1 / US 7,038,253 B2 (Furukawa) — the architecture reference

Supplies every element of claim 1 except the carbon-containing intermediate layer; supplies the p-type high-resistance layer, the undoped-GaN carrier traveling layer, the AlGaN carrier supplying layer, the recess geometry, the MOS gate, and the RESURF/2DEG teaching.

3.3 D. Visalli et al., Solid State Device Meetings 2008, No. G-1-2 — the double-heterostructure reference

"AlGaN/GaN/AlGaN Double Heterostructures on 4 inch Si Substrates for High Breakdown Voltage Field-Effect Transistors with low On-Resistance." Teaches an AlGaN layer disposed below the GaN channel to form a double heterojunction, improving breakdown voltage at low Ron — the "heterojunction forming layer" of claims 7–8 and 16–17.

3.4 Commonly-owned family references — § 103(c) caveat

US 2011/0241088 A1 / US 8,450,782 B2 (the sibling, same 2009-03-31 priority, same assignee) appears in the record as an examiner-cited reference. Do not build a § 103 combination on it. Pre-AIA § 103(c) disqualifies commonly-owned § 102(e)/(f)/(g) art from use in a § 103 rejection; here the sibling shares both the priority date and the assignee (Furukawa), so it is properly excluded (its real significance is a double-patenting / statutory-type question, not § 103).

Unverified but potentially available art (from the family-citation list — I have not verified their disclosures and flag them as leads only): JP 2008205221 A (Furukawa, "Semiconductor element," pub. 2008-09-04, priority 2007-02-20); JP 5348364 B2 (Sanken, heterojunction field-effect semiconductor device, priority 2007-08-27); JP 4761319 B2 (Sharp, nitride semiconductor device, priority 2008-02-19); JP 5400266 B2 (Panasonic, "Field effect transistor," 2006); JP 2000068498 A (NTT, insulating nitride film); JP 5064824 B2 and JP 3960957 B2 (Furukawa); and JP 2005317684 A (dry-etching method, potentially pertinent to the recess etch). Several of these are high-resistance-GaN/carbon-compensation art and would strengthen alternative grounds if their texts confirm it.

---

4. Ground-by-ground obviousness analysis

GROUND 1 — Claim 1 (and claims 9, 19, 21): WO '607/US 7,038,253 in view of Niiyama

Rejection. Claim 1 is obvious over the hybrid-MOSHEMT art of WO '607/US 7,038,253 in view of Niiyama (US 2009/0242938 A1; US 8,072,002 B2).

Element mapping. Section 2 above maps WO '607 onto every element but one. Niiyama supplies the missing element — a discrete acceptor-containing sub-layer between the p-GaN high-resistance layer and the 2DEG region — and also supplies the recess geometry (recess "reaching to the lower part layer," Niiyama claim 4) that the '800 uses to isolate that layer into left/right segments.

Motivation to combine — from the references themselves (KSR prong 1). The motivation is not supplied by me; Niiyama states it in terms (¶[0063]): placing an acceptor-density-differentiated sub-layer beneath the channel lets the designer independently optimize withstand voltage (via the sub-layer's acceptor density → 2DEG sheet density) and threshold voltage (via the underlying layer). The '800 patent's stated problem — "it becomes difficult to keep a balance of the carriers in the RESURF structure and thus difficult to maintain the withstand voltage" when 2DEG concentration is raised — is the identical problem Niiyama addresses. A PHOSITA confronting the '800's problem would look directly at Niiyama's solution.

Motivation from the art's common ownership and lineage (KSR prong 1). Both WO '607 (the '800's own admitted starting point) and Niiyama are Furukawa Electric work. Niiyama's Example section itself cites WO 2003/071607 as "Patent Document 1" — i.e., Niiyama was expressly developed as an improvement on the very reference the '800 uses as its baseline. That is the strongest possible "interconnected teachings" showing: the artisan is being handed two commonly-owned references, one of which explicitly builds on the other.

Motivation from design incentive (KSR prong 2). The '800 specification itself sets out the trade-off as an engineering dial: "In order to increase the two-dimensional electron gas and maintain a low on-resistance, it is only necessary that the thickness or the Al composition ratio of the carrier supplying layers... is increased. Furthermore, in order to decrease the entire carrier concentration of the two-dimensional electron gas, it is only necessary that the carbon concentration in the carrier concentration controlling layers... is increased." A parameter identified as tunable via a monotone trade-off is, under KSR, a design choice within the ordinary skill level.

Reasonable expectation of success. Niiyama's own device achieved both high mobility and high withstand voltage by exactly this mechanism; the '800 asserts no new physics, only a new layer position. Success was predictable.

Anticipation of counterargument — "Niiyama's layer is the channel." Applicant will argue Niiyama conflates the acceptor layer and the channel, whereas the '800 separates them. This fails: (a) claim 1 itself requires only a "carrier concentration controlling layer" on the high-resistance layer and a "carrier traveling layer" on the controlling layer — it does not require the controlling layer to be non-conductive; (b) Niiyama's upper part layer 203b/203c is itself the 2DEG-bearing region, so its displacement upward by a further undoped GaN layer is nothing more than WO '607's admitted undoped-GaN carrier traveling layer stacked on top of Niiyama's acceptor sub-layer — a rearrangement of old elements with predictable results.

Sub-claims riding on Ground 1:

• Claim 9 ("carbon concentration decreased in a monotonous manner from the high-resistance/foundation side toward the carrier traveling side"): Niiyama's lower-layer/upper-layer acceptor-density differentiation (claim 4; ¶[0063]) teaches graded acceptor concentration across the depth of the layer stack; grading a dopant profile is a routine, well-known epitaxial technique. Obvious.
• Claim 19 ("recess extends downward from the carrier supplying layer into the high resistance or foundation layer"): directly taught by Niiyama claim 4 / ¶[0014] — "said recess region having the depth reaching to the lower part layer." Obvious.
• Claim 21 ("2DEG generated near the interface between the carrier traveling layer and the carrier supplying layer"): inherent to the AlGaN/GaN heterojunction; taught by both WO '607 and Niiyama (¶[0026], FIG. 9). Obvious.

GROUND 2 — Claims 1, 2, 10, 11, 20: selection of carbon as the acceptor species

Rejection. Claims 1 and 11 (and dependents 2, 10, 20) are obvious over WO '607 + Niiyama for the additional reason that Niiyama expressly names carbon.

Niiyama ¶[0012] (and granted claim 2): the acceptor is "one of Mg, Be, Zn and C." That is a finite, four-element list of identified, predictable solutions — the paradigm KSR case. Selecting C from that list, to be introduced during MOCVD growth of the GaN sub-layer, was a routine choice: as the '800 specification itself concedes, "carbon can be added through auto doping of the carbon contained in an organo-metallic element that is a raw material while adjusting growing conditions such as the vacuum degree within the apparatus and the growing speed." Carbon's role as a deep-level acceptor that compensates donors and renders GaN semi-insulating was notorious in the III-nitride art by 2009 (carbon-doped high-resistance GaN buffers were standard practice to suppress vertical leakage and raise breakdown voltage). A PHOSITA would expect carbon to perform the compensating function Niiyama assigns to the sub-layer's acceptor, with success reasonably expected.

• Claim 10 / Claim 20 ("carrier traveling layer is GaN containing carbon at a concentration lower than 5×10¹⁷ cm⁻³"): this is simply a concentration ceiling, and Niiyama's own upper bound of the acceptor range is 5×10¹⁷ cm⁻³ (¶[0013]). A claim reciting "less than X" where the prior art recites "up to and including X" is an overlapping/abutting range — prima facie obvious under In re Peterson. Independently, the '800 specification states as a design rule that the carrier traveling layer impurity concentration "is preferably 5×10¹⁷ cm⁻³ or less in order to keep the mobility high" — i.e., the patent itself treats this as the ordinary optimum, not an invention.

GROUND 3 — Claims 2, 3, 5, 6, 11, 12, 14, 15: numerical ranges as routine optimization

All of the '800's numeric limitations are, on their face, optimization of parameters the prior art already identified as tunable:

GROUND 4 — Claims 7, 8, 16, 17: the heterojunction forming layer

Rejection. Claims 7–8 and 16–17 are obvious over WO '607 + Niiyama in further view of the Visalli NPL reference.

Visalli teaches an AlGaN/GaN/AlGaN double heterostructure on 4-inch Si yielding high breakdown voltage at low on-resistance. The '800 specification expressly adopts Visalli's teaching as its own second embodiment, stating that the MOSFET 200 "achieves a further high withstand voltage due to the same operation as that described in 'D. Visalli, et al.' because the carrier supplying layers and the heterojunction forming layer form a double hetero structure of AlGaN/GaN."

That is, for claims 7–8 and 16–17 the applicant's own specification supplies the motivation and the expectation of success. A PHOSITA seeking to raise breakdown voltage in the WO '607/Niiyama device would add an AlGaN layer below the GaN — exactly what Visalli prescribes and what the '800 did. Claim 8's Al ratio 0.01–0.5 is a routine range for a lower AlGaN confining layer. Obvious.

GROUND 5 — Claim 18 (monotonic decrease), discussed separately

Claim 18 mirrors claim 9 with respect to claim 11. Same analysis: Niiyama's acceptor-differentiated sub-layers teach a non-uniform acceptor profile across the stack; producing it as a continuous monotonic gradient (as the '800's working example in fact does — "carbon concentration... stepwise decreased from 5×10¹⁹ cm⁻³ to 5×10¹⁷ cm⁻³ from the substrate side toward the surface side") is routine MOCVD practice. Obvious.

---

5. Consolidated motivation-to-combine statement

For all grounds, four independent motivations exist, any one of which suffices under KSR:

1. Same problem, same solution, same assignee. Both WO '607 and Niiyama are Furukawa Electric references directed to normally-off, high-withstand-voltage, low-on-resistance III-nitride FETs. Niiyama's background section expressly identifies WO 2003/071607 as its starting point.
2. The prior art states the claimed function. Niiyama ¶[0063] explains that a sub-layer's acceptor density controls 2DEG sheet carrier density and withstand voltage independently of the threshold-determining layer below — i.e., the '800's entire asserted mechanism.
3. The prior art supplies the exact structural geometry. Niiyama claim 4 / ¶[0014] recites the recess reaching to the lower part layer through the acceptor-differentiated upper layer — the '800's FIG. 1 recess geometry (recess 107 into high-resistance layer 103, isolating layers 104a/104b).
4. The prior art supplies the species. Niiyama ¶[0012] names C explicitly among four acceptors, and carbon auto-doping of MOCVD GaN was routine.

---

6. Anticipated applicant rebuttals and my assessment

Rebuttal A — "Niiyama's acceptor range tops out at 5×10¹⁷; the claim requires up to 5×10¹⁹." Partially effective against claims 2 and 11's upper limit as a standalone ground, but defeated by In re Peterson (abutting range) and by the absence of any asserted criticality. Note the specification itself uses "about" and gives no data isolating 5×10¹⁹ from, say, 1×10¹⁸.

Rebuttal B — "Niiyama teaches away; high acceptor concentrations degrade mobility." The '800 specification makes the same statement ("the impurity concentration such as carbon [in the carrier traveling layer] is preferably 5×10¹⁷ cm⁻³ or less in order to keep the mobility high"). There is therefore no teaching away; both references confine carbon to the layer below the channel. Indeed this coincidence is itself evidence of obviousness — the '800's design rule is Niiyama's.

Rebuttal C — secondary considerations / unexpected results. This is the applicant's only real argument, resting on the working example: 1200 V breakdown at ~5 mΩcm² Ron vs. 500 V at ~5 mΩcm² for the comparison example. Assessment:

• Nexus is weak. The comparison example replaces the carbon layer with undoped GaN ("the carrier concentration controlling layer is replaced with the u-GaN layer with the same thickness") — i.e., it compares against a device with no acceptor sub-layer at all, not against the closest prior art (Niiyama's acceptor-containing sub-layer). Under In re De Blauwe, a comparison against non-closest prior art cannot establish unexpectedness.
• The result was predicted. Niiyama ¶[0063] predicts exactly this: sub-layer acceptor density governs 2DEG sheet density and hence withstand voltage, decoupled from threshold.
• Commercial success / licensing: I found no evidence of such evidence in the record (consistent with my earlier sections, which found no litigation implicating this patent).

Rebuttal D — "the cited references are non-analogous." Unsustainable. All four references are in the same field of endeavor (III-nitride power FETs), address the same problem (normally-off operation with simultaneously high mobility and high breakdown voltage), and two share the assignee.

---

7. Bottom-line conclusions

Overall: The claims of US 8,569,800 appear vulnerable to a § 103 challenge, with claim 1 and claim 11 the least defensible independent claims and claims 6 and 19 the most clearly readable on the prior art. The decisive reference is US 2009/0242938 A1 / US 8,072,002 B2 (Furukawa Electric), which was before the examiner and which teaches both the structural delta (a recess-pierced, acceptor-differentiated sub-layer between a high-acceptor layer and the 2DEG region) and the operative mechanism (independent control of withstand voltage and threshold via sub-layer acceptor density). The applicant's only meaningful counter is the 500 V→1200 V working example, and its nexus is significantly weakened by comparison against a no-acceptor control rather than against Niiyama.

---

8. Confidence and verification caveats

• Confirmed from the authoritative record and from full-text searches: the '800 claim set; the WO '607 / US 7,038,253 hybrid-MOSHEMT architecture; the Niiyama disclosure (¶¶[0011]–[0016], [0063]–[0064], claims 1–4) and its Furukawa Electric assignee; the Visalli NPL citation and the '800's express adoption of Visalli's double-heterostructure teaching; the '800's own working-example figures.
• Not confirmed — verify before formal reliance: (a) the exact U.S. filing date of Niiyama App. 12/382,941, on which its pre-AIA § 102(e) status depends; (b) that US 8,072,002 B2 is the grant of US 2009/0242938 A1 (very strong textual match — identical ¶[0063]/[0064]/FIG. 9 and assignee — but I did not see the application-number linkage stated on one page); (c) the disclosures of the family-cited JP/other references in §3.4, which I flag as leads only and did not verify.
• Litigation and CAFC 2026: unchanged from my earlier sections — no litigation asserting US 8,569,800 was found, and no 2026 CAFC activity was found. The "Monroe '183 / IPR2016-01041" and "Deere v. Bush Hog / 6,052,980" results are false positives on the "'980" shorthand and are not this patent.
• Nothing in this analysis is a legal conclusion or a validity opinion; it is a technical obviousness assessment based on the references of record.