US 5785726

I'll search for the patent and any related litigation.

Let me check specifically for litigation or CAFC dockets involving this patent.

US Patent 5,785,726 — Summary

Caveat on the date: Your prompt states the current date is April 26, 2026, but several search results returned documents dated June–July 2026 (e.g., a Fed. Cir. Rule 36 judgment dated 5/14/2026 and a July 8, 2026 Notice of Docketing). I therefore treated the patent record itself as authoritative and flag that my "as of" coverage may be slightly out of step with your stated date.

Bibliographic data (as literally recorded on the face of the record)

Source: https://patents.google.com/patent/[US5785726A](/patent/US5785726A)/en ; claims text corroborated at http://www.everypatent.com/comp/pat5785726.html

Abstract (verbatim)

"A method of making glasses with manufacturing systems employing platinum or molybdenum melting, fining delivery, or forming vessel, wherein the partial pressure of hydrogen outside the vessel is controlled or manipulated, relative to the partial pressure of hydrogen inside the platinum structure, to minimize surface blistering in the glass in contact with the platinum or molybdenum."

Plain-language overview of the independent claims

There are two independent claims — claim 1 and claim 25. (Claims 23 and 24 depend from claim 1, not independent.)

Claim 1 — A method of forming an oxide glass in a manufacturing process that uses, in a melting, fining, delivery, or forming vessel, a metal chosen from platinum, molybdenum, palladium, rhodium, and their alloys, where there is an interface between the vessel and the glass. The method comprises providing a partial pressure of hydrogen outside the vessel in an amount sufficient to prevent formation of oxygen blisters in the glass region adjacent to the vessel/glass interface. In plain terms: hold the hydrogen level on the outside of the precious-metal (or Mo) vessel high enough that you don't get blistering at the metal/glass contact surface.

Claim 25 — A method of forming an oxide glass in the same kind of metal-vessel process, comprising controlling the partial pressure of hydrogen outside the vessel relative to the partial pressure of hydrogen inside the vessel so that the outside partial pressure is sufficient to prevent hydrogen migration through the vessel wall and out of the glass. In plain terms: balance the inside vs. outside hydrogen partial pressures so no hydrogen escapes the melt through the vessel wall — i.e., make the metal vessel behave as if chemically inert.

Notable drafting observation (a factual, not legal, point): Claim 1 recites "the method comprising: providing a partial pressure…," yet most dependent claims (2, 3, 5–12) refer to "said controlling step." The antecedent basis for "controlling" does not appear in issued claim 1, which suggests the claim was amended in prosecution without conforming the dependents. Claim 25, by contrast, does use "controlling." This is an inconsistency on the face of the issued patent and would be a natural claim-construction battleground.

Representative dependents for context: enclosure around the vessel with humidity/dew-point control (claims 4–8); electrochemical monitoring of vessel potential/DC current against a reference electrode, with opposite control logic for platinum (claim 11) vs. molybdenum (claim 12); downdraw/fusion sheet forming (claims 13, 16, 17); low-arsenic batch selection, e.g., <0.2 mole % As₂O₃ and β-OH below about 0.5/mm (claims 14, 15); alternative fining agents Sb₂O₃, CeO₂, SnO₂, Fe₂O₃, halides, sulfates (claims 18–22); and an aluminoborosilicate glass composition window (claim 23), alkali-free with strain point >630 °C and CTE 32–46 × 10⁻⁷/°C (claim 24).

Litigation / CAFC 2026 docket search

• The Google Patents record carries a boilerplate "Family has litigation" flag with a Darts-IP link (family ID 24961556; https://patents.darts-ip.com/?family=24961556). This is a commercial-database flag, not an authoritative litigation listing, and the linked content is behind a license; I could not retrieve it.
• I found no CAFC 2026 docket, appeal, IPR, or district-court case specifically naming US 5,785,726. Targeted searches for "5785726"/"5,785,726" returned only patent-family citations, prosecution-history/IDS references, and unrelated patents that happen to share the last three digits.
• The 2026 Federal Circuit materials surfaced by search (e.g., Spero v. Mercedes-Benz USA, Nos. 2025-1470/1472, Rule 36 affirmance 5/14/2026; Contour IP Holding v. GoPro, No. 2026-2012) are unrelated to this patent and I am not treating them as relevant.
• Structural reason for low likelihood of a 2026 CAFC appeal: the patent expired 2016-10-28, and it is not listed as subject to any IPR/PGR in the record I retrieved.

Uncertainty statement: I do not have authoritative confirmation of whether, or in which courts, US 5,785,726 was ever asserted in litigation, nor any 2026 appellate docket. The Darts-IP "has litigation" flag is a lead, not a verified fact. Anyone needing a definitive answer should query the USPTO PatentCenter/PTAB (https://patentcenter.uspto.gov/), PACER/RECAP (CourtListener), and the Federal Circuit docket directly for the exact string "5,785,726."

Why this patent matters (brief context)

The patent is part of Corning's arsenic-free AMLCD-substrate glass work: it claims the atmospheric/hydrogen-partial-pressure lever for suppressing electrochemical oxygen blisters at platinum/Rh-Pt and molybdenum delivery systems, which enabled high-strain-point, low-As₂O₃ glasses such as Code 1737-type compositions formed by fusion downdraw. Its disclosure is heavily cited by later Corning, Schott, AGC, and NEG filings (75 "cited by" entries in the Google Patents record).

I'll research litigation involving US Patent 5785726. Let me search multiple sources.

Let me dig deeper into litigation-specific sources.

Litigation Search: US Patent 5,785,726

Bottom line: I found no litigation that specifically asserts or challenges US 5,785,726. The searches did not surface any district court case, ITC investigation, or PTAB/IPR proceeding naming the '726 patent as the patent-in-suit or the challenged patent. However, I must flag an important caveat and some limits on this finding.

Patent identification (as listed literally)

• Patent: US 5,785,726 — "Method of reducing bubbles at the vessel/glass interface in a glass manufacturing system"
• Inventors: William G. Dorfeld; David M. Lineman
• Assignee: Corning Incorporated
• Application: US 08/736,848, filed 1996-10-28; granted 1998-07-28
• Status: Expired – Lifetime (anticipated expiration 2016-10-28)
• Family (per Google Patents): WO1998018731A2, EP0951453B1, DE69730350T2, JP4683674B2, KR100512538B1, CA2269843A1

What the sources showed

• The Google Patents record contains a "Family has litigation" flag linking to a Darts-IP entry (https://patents.darts-ip.com/?family=24961556). This indicates that some member of this patent family has been involved in litigation somewhere in the world. I could not retrieve the underlying Darts-IP record, so I cannot confirm which family member, which case, or which jurisdiction — and the litigation may relate to a sibling foreign family member rather than to US 5,785,726 itself. I am explicitly uncertain here and do not want to overstate it.
• Hits on the number "5785726" in my searches were prosecution/prior-art citations, not litigation. For example, the number appears in the file history of Corning's later "Fining of Boroalumino Silicate Glasses" application (US 11/478,493) and in the prior-art listings of other Corning glass applications. These are Office Action/IDS references, not lawsuits.
• Related Corning glass-patent litigation exists but does not involve the '726 patent. Corning has been active in display-glass litigation (e.g., the ITC § 337-TA-1433 glass-substrates matter and related PTAB proceedings involving Corning's EAGLE XG patents such as US 8,627,684 and the '491 patent, and the Wilson Wolf cell-culture disputes). None of the records I retrieved assert or challenge US 5,785,726.

Caveats on completeness

This finding rests on open web search only. I did not obtain direct, queryable results from PACER, Lex Machina, Darts-IP, or the Unified Patents litigation portal, and I was not able to open the Darts-IP family page. Given the "family has litigation" flag, the most likely explanations are:

1. The litigation flagged by Darts-IP concerns a foreign sibling of the '726 patent (EP/JP/KR/CA/WO family members), not the US patent; or
2. The flag reflects a broader Darts-IP family grouping that includes other Corning glass patents implicated in litigation.

To reach a definitive "no litigation" conclusion for US 5,785,726 itself, the number should be run against:

• Darts-IP (family 24961556) — to resolve the flag;
• PACER / Lex Machina / Unified Patents — exact-number search for "5,785,726" (and "5785726") in complaint and docket text;
• PTAB (PTAB E2E / PTABIS) — for any IPR/PGR/CBM/Derivation naming the '726 patent.

I cannot state with high confidence that no litigation exists; I can only report that none was found in the accessible sources.

I'll start by verifying the structured ODP data against live sources, since IPR filings can surface outside the ingested feed.

The first pass surfaced only unrelated patents (e.g., Apple's '578, IPVX's 5,572,576 — different patents). Let me run more targeted queries on the exact patent number and on Corning enforcement.

No hits on the '726 patent itself. Let me do two final confirmation searches — one on the inventors/patent title against PTAB, and one on litigation.

PTAB / AIA Trial Analysis — US 5,785,726

Method note: I began from the canonical ODP structured block ("PTAB proceedings on file: none"), then ran independent web searches to catch anything the ODP feed may not have ingested. I reached my tool-call limit and could not retrieve a docket-confirmed negative; the conclusions below rest on the ODP structured result plus the absence of any hit in targeted searches. Treat the "zero proceedings" finding as high-confidence but not court-certified.

---

Proceedings overview

There are zero AIA trial proceedings on US 5,785,726 — no IPRs, no PGRs, no CBMs, in any status (active, claims invalidated, claims sustained, settled, or institution-denied). The breakdown is therefore 0 / 0 / 0 / 0 / 0, and there is no PTAB estoppel and no claim-level invalidation to hand a defendant. The bottom line is not "the patent is hardened by surviving IPRs" and not "claims 1–5 were canceled" — it is that no one has ever tested this patent at the PTAB. The dominant defensive fact for a defendant today is not a PTAB outcome but the statutory clock: the patent's anticipated expiration was 2016-10-28 (status: Expired – Lifetime), meaning any live exposure is limited to past damages within the § 286 six-year lookback window and any pre-expiration willfulness posture.

---

Per-proceeding detail

No proceedings exist, so there are no proceeding entries to populate. Because the requested per-proceeding template (petitioner, panel, grounds, FWD, appeal) would be entirely fabricated, I am leaving it blank rather than filling it in.

What the searches did return (flagged, not PTAB)

• Other, unrelated "578" patents. Search noise included Apple/PTAB matters on a different '578 patent (gesture/UI claims, e.g., claims 1, 15, 27) and IPVX Patent Holdings' U.S. 5,572,576 (telephone answering device, IPR2013-00216). Neither is US 5,785,726. Do not conflate them.
• Corning as a frequent PTAB party generally. Corning appears repeatedly in IPR dockets — e.g., Corning v. DSM IP Assets (IPR2013-00044 through -00050), Panduit v. Corning Optical Communications (IPR2016-01703/01709, IPR2017-00528, institution denied), and Caihong Display Devices challenging Corning's US 8,642,491. These are Corning patents other than the '726, and none is the subject patent here.
• A possible citation error worth noting. US 7,032,412 (Corning) describes "U.S. Pat. No. 5,758,726" as disclosing "a process for reducing bubbles … controll[ing] the partial pressure of hydrogen outside the vessel relative to … inside." The description maps exactly onto US 5,785,726, so this is very likely a typo for 5,785,726 in that later patent. I flag it because if a defendant runs a keyword search for the subject matter, they may see the "5,758,726" string and mistake it for a separate patent or reference. It is not a proceeding and, per the operating rules, I am not auto-correcting the number — just flagging the discrepancy.
• Google Patents shows a "Family has litigation / First worldwide family litigation filed" flag linking to a Darts-IP family page (family ID 24961556). That flag indicates the family has been in litigation somewhere, but the source page did not disclose the parties, forum, or whether the '726 itself (as opposed to a sibling/continuation) was asserted. I could not substantiate specifics, so I am not asserting any litigation details.

---

Strategic summary

Claim status: all 25 claims are UNTESTED and intact. Claims 1–25 have never been construed by the PTAB, never canceled, never confirmed. That cuts both ways for a defendant: there is no PTAB roadmap of winning art, but there is also no estoppel binding anyone and no panel has blessed the claims' validity. The only claim-level guidance available is the intrinsic/prosecution record — notably that claim 1 (as issued) recites "providing a partial pressure of hydrogen outside the vessel … sufficient to prevent formation of oxygen blisters," while dependent claims 2–12 recite "said controlling step" — a phrase with no antecedent in claim 1. That is the kind of § 112(b) / indefiniteness and written-description vulnerability a challenger would probe, along with the process-parameter claim format (§ 101/§ 112 issues) and the "sufficient to prevent blister formation" result-oriented language. None of this has been adjudicated.

Estoppel landscape: none. Because no IPR/PGR was ever filed, § 315(e)(2) estoppel does not attach to anyone. A defendant is free to raise any prior-art ground — §§ 102, 103, 112 — in district court or in a fresh IPR, with no risk that a prior petitioner's institution or FWD bars them. The corollary: there is also no prior petitioner's art to free-ride on; you would be building the invalidity case from scratch.

Pattern signals: none. No repeat petitioner (there are no petitioners), no patentee PTAB-appeal activity (the patent owner, Corning, has no FWD to appeal here), and no defensive aggregator (Unified Patents, RPX, etc.) in the chain — nothing to indicate the patent has drawn organized validity attacks. For a patent that was, per its own specification, central to Corning's arsenic-free fusion-draw platform (Code 1737 / Eagle-class glass chemistry), the total absence of PTAB activity is itself informative: the patent appears to have been enforced (if at all) through trade-secret/contract routes rather than patent litigation, so no defendant ever had the standing-plus-incentive combination to file an IPR.

---

Recommended next steps

1. Confirm the negative on the record sources before relying on it. Pull the PTAB E2E / USPTO Patent Center "Proceedings" tab for patent 5,785,726 directly, and cross-check Docket Alarm / Lex Machina / Unified Patents' portal. My search budget ran out before I could retrieve a docket-confirmed zero, so verify rather than assume.
2. Lead with expiration, not validity. The anticipated expiration was 2016-10-28, roughly 9 years ago as of 2026-10-01. If you are facing a demand letter or complaint citing this patent, the first question is whether the asserted acts are within the § 286 six-year damages lookback (i.e., on or after ~2020-10-01). Accused conduct after 2016-10-28 cannot infringe an expired patent; pre-expiration conduct outside the lookback is time-barred. This likely disposes of the matter without any PTAB filing at all.
3. If a live, in-window damages theory survives, treat invalidity as an open field. With no § 315(e)(2) estoppel and no PTAB history, an IPR (if you can still satisfy the § 315(b) one-year service bar) or an § 112 challenge in court faces no prior-adjudication baggage. Focus on the claim 1 / claim 2 antecedent-basis defect ("said controlling step") and the functional "sufficient to prevent" language, and note that the '726 is highly cited by Corning's own later patents (e.g., US 7,032,412; US 8,127,572), which supply an admissions-rich intrinsic record.
4. Do not cite the "5,758,726" string as this patent. It appears to be a typo in US 7,032,412; the correct subject number is 5,785,726. Flag it so your team's prior-art and family maps stay clean.

Sources reviewed: USPTO Open Data Portal (structured "PTAB proceedings on file" block — none); Google Patents, US5785726A; USPTO/PTAB document repository (PTACTS) for Corning and unrelated '578 matters; CourtListener dockets for Corning litigation. No PTAB Final Written Decision, institution decision, or Federal Circuit opinion on US 5,785,726 was located, and none is asserted here.

Caveat: this analysis reflects the sources I could reach within the available search budget. It is not a substitute for a docket-confirmed PACER/PTAB E2E check, and I have not verified the classification of any family litigation referenced by the Google Patents flag.

I'll research this patent's assignment record before writing anything up.

I researched the assignment chain via Google Patents legal events (which mirror the USPTO assignment records), the EPO/PubChem family data, IPR filings citing the patent, and general web sources. Below is the reconstruction with an explicit note about what I could and could not verify.

Important limitation up front: I could not query the live USPTO Assignment Center (assignmentcenter.uspto.gov) directly from this environment — it is a JavaScript/API-gated database that does not expose records as fetchable text. The findings below are based on Google Patents' "Legal Events" tab (which reproduces the USPTO assignment reel/frame entries), the EPO Register, and secondary sources. No post-issuance assignment (reel/frame) surfaced in any of them. Treat reel/frame numbers as "not surfaced in accessible sources" rather than "does not exist."

Inventors

• The USPTO assignment record itself lists both as assignors to CORNING INCORPORATED, confirming they were Corning employees at filing (invention assigned to employer).
• No unusual pattern. Neither inventor appears to have departed Corning around filing. Dorfeld in particular is a prolific Corning inventor continuing through the arsenic‑free‑glass family (US 6,128,924). There is no 12‑month exodus signature that typically precedes a portfolio fire‑sale.

Original assignee

• Entity on the issued patent: Corning Incorporated, SP‑TI‑03‑1 / 1 Riverfront Plaza, Corning, NY 14831.
• Product embodying the claims: Yes — the patent's method is practiced in Corning's fusion/overflow downdraw sheet‑glass lines for display substrates (the spec expressly references Corning Code 7059 and Code 1737 glass, and the sibling arsenic‑free work that became EAGLE). Corning ships these products commercially.
• Primary line of business: Specialty glass and ceramics; the relevant segment is display‑glass substrates for LCD/OLED panels.
• Current status: Operating, publicly traded (NYSE: GLW). Not acquired, not dissolved, not in bankruptcy. It remains the assignee of record.
• Patent status: Expired – Lifetime, anticipated expiration 2016‑10‑28 (20 years from filing, no maintenance delinquency indicated).

Assignment timeline

Only one assignment is of record in the accessible sources — the original inventor‑to‑employer assignment. No subsequent conveyance (change of name, merger, security agreement, or transfer) surfaced.

• 1996‑10‑28 (executed) / recorded 1996‑10‑28 — Reel/Frame not surfaced in accessible sources (Google Patents legal events / EPO Register; the raw reel/frame is only exposed on the Assignment Center UI I could not query).
  • Conveyance: Assignment (inventor‑to‑employer, "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: Dorfeld, William G.; Lineman, David M.
  • Assignee: CORNING INCORPORATED
  • Correspondent: not surfaced — the recording here is the standard post‑filing employer assignment; no attorney correspondent name is exposed by the indexed sources.
  • Context: Original employment assignment at filing — not a reorg, not a sale.

No further assignments (Acacia, Marathon, IV, Wi‑LAN, etc. — or any operating buyer) appear in the chain. Per your instruction that "many patents have no recorded post-issuance assignments — that is itself a finding," this patent is in exactly that category: the original assignee, Corning, still owns it.

Timeline diagram

timeline
    title Ownership of US 5785726
    1996 : Inventors Dorfeld and Lineman file
         : Assigned to Corning Incorporated
    1998 : Patent 5785726 issues
    2016 : Patent expires anticipated
         : No later assignments of record

NPE / troll-pattern signals

1. Shell-entity transfer — not present. No transfer to any IP/Licensing/Holdings/Ventures LLC of record; the sole assignee is the operating parent Corning Incorporated.
2. Known asserter in the chain — not present. No Acacia, Marathon, IV, Wi‑LAN, Conversant/Mosaid, Pendrell, Round Rock, Spangenberg, or other listed NPE appears anywhere in the record.
3. Repeat correspondent across the chain — not present / unclear. Only one assignment is recorded and no attorney correspondent is exposed in the indexed sources, so recurrence cannot be assessed. (Note: Thomas R. Beall, Reg. No. 40,424, Corning Incorporated, SP‑TI‑03‑1, Corning, NY 14831 appears as the prosecuting correspondent on a later Corning application citing this patent — that is prosecution counsel, not an assignment recording correspondent, and is not an NPE tell.)
4. Cascading transfers — not present. Zero consecutive assignments; nothing to cascade.
5. Pre-litigation transfer — not present. No assignment within 6 months of any suit, because no post‑issuance assignment exists. Google Patents shows a Darts‑ip "family has litigation" link for the family, but I could not confirm that US 5,785,726 itself was ever asserted, and there is no transfer that would indicate venue‑shopping.
6. Bankruptcy fire‑sale — not present. Corning has not filed Chapter 7/11; no patent‑sale proceedings involving this asset.
7. Privateering — not present. No transfer to an NPE asserting on Corning's behalf.
8. Defensive aggregator (anti‑NPE) — not present. Chain does not terminate at RPX, AST, LOT, Unified, or OIN; it terminates at Corning.

Verdict

Insufficient data — chain is a single original assignment; no NPE activity of record.

US 5,785,726 was assigned once, by inventors William G. Dorfeld and David M. Lineman to Corning Incorporated at filing (executed/recorded 1996‑10‑28, per the Google Patents legal‑events rendering of the USPTO record), and no post‑issuance assignment of any kind appears in the accessible sources. The patent is a Corning‑originated, Corning‑owned operating‑company asset that expired 2016‑10‑28; there is no shell LLC, no known asserter, no repeat assignment correspondent, and no cascading or pre‑litigation transfer to support any NPE reading. Because only the original assignment exists (and the raw reel/frame and correspondent were not retrievable from the sources I could reach), the correct call is insufficient data to find any NPE pattern, not merely "no NPE"—though on the evidence available, all eight NPE signals are not present.

Verification links:

• USPTO Assignment Center (search patent 5785726): https://assignmentcenter.uspto.gov/ (also https://assignment.uspto.gov/patent/index.html)
• Google Patents legal events: https://patents.google.com/patent/[US5785726A](/patent/US5785726A)/en
• EPO Register / family (US 6,128,924 arsenic‑free sibling): https://patents.google.com/patent/[US6128924A](/patent/US6128924A)/en

Two caveats I want to be explicit about: (1) I could not read the Assignment Center UI directly, so I cannot rule out a maintenance‑fee‑related record or a late‑recorded assignment that Google/EPO did not index — re‑run the search on assignmentcenter.uspto.gov to be certain. (2) I did not fabricate a reel/frame; if the Assignment Center returns one for the 1996‑10‑28 employer assignment, it should be dropped into the timeline entry above.

I'll research the prior art cited against US 5,785,726, starting with the USPTO record and the patent's front-page references.

Let me pull the patent's front-page "References Cited" section directly from the patent image and the USPTO record.

Let me retrieve the actual "References Cited" list from the patent front page and the USPTO full-text record.

Important sourcing limitation (please read first)

I was not able to retrieve the front-page "References Cited" list — i.e., the U.S. Patent Documents and Foreign Patent Documents that the examiner actually cited against US 5,785,726. Specifically:

• The authoritative Google Patents full text you provided omits the "References Cited / Patent Citations" block entirely (it jumps from the "Images" section straight to "Classifications").
• My web searches for the examiner-cited list returned the patent's "Cited By" (forward-citation) table and third-party citations of '726, but not the '726 front-page references.
• I will not fabricate an examiner-cited list. Where I list references below, I label precisely what each is.

Also, per the operating rule, I have not auto-corrected any identifier. Where a source cites "U.S. Pat. No. 5,758,726" I report it literally and flag it rather than silently fixing it to 5,785,726.

Note on the date: your prompt says April 26, 2026; this environment's clock says 2026-10-01. I treat the patent record itself as authoritative and flag the mismatch.

---

Two different "citation" sets — do not conflate them

The task asks about "each patent citation for 5785726" as prior art. Only the backward set qualifies. The forward set is irrelevant to § 102 against '726 (though useful for infringement/validity-over-time context).

---

References of record I can document (applicant-cited in the '726 specification)

These are the references the '726 specification itself cites. I could not confirm they also appear on the front page, but they are the only '726 references I can verify verbatim.

On reference 5: the '726 text calls it out only as a co-filed application incorporated by reference. I am not confident which issued patent it became and will not guess — do not rely on any issue number for 08/742,610 without checking PatentCenter.

Discrepancy flag (literal): US 7,032,412 states "U.S. Pat. No. 5,758,726 discloses a process for reducing bubbles in glass sheets formed in systems utilizing platinum-containing vessel." Taken literally, 5,758,726 ≠ 5,785,726. The described subject matter matches '726, so this appears to be a citation error somewhere, but per the literal-interpretation rule I record it as written and do not correct it.

---

§ 102 anticipation analysis (the actual question)

Legal frame applied: Under 35 U.S.C. § 102, a single reference anticipates only if it discloses every element of the claim, arranged as in the claim. Also critical here: a dependent claim includes all limitations of the claim(s) from which it depends — so no dependent claim can be § 102-anticipated by a reference that does not anticipate independent claim 1 (or 25). Constructing anticipation of a dependent claim from a reference that teaches only the added limitation is legally wrong.

Because the core of both independent claims is the hydrogen-partial-pressure control step — (claim 1) "providing a partial pressure of hydrogen outside the vessel … sufficient to prevent formation of oxygen blisters … adjacent said vessel/glass interface"; (claim 25) "controlling the partial pressure of hydrogen outside the vessel relative to … inside the vessel so that [it] is sufficient to prevent hydrogen migration through the vessel" — the analysis turns on whether any single reference discloses that step together with the precious-metal/Mo vessel.

Bottom line: On the references of record I can verify, none anticipates claims 1 or 25, and therefore none anticipates any dependent claim. Each reference maps to a sub-element of a dependent claim (process type, glass composition, fining agent), and would matter only as § 103 obviousness material in combination — and even then only if some other art supplied the hydrogen-partial-pressure control that is the heart of the invention. I found no such single anticipatory reference in the accessible record.

---

What I could not verify — and how to close the gap

1. Examiner-cited front-page references (the list actually relied on in prosecution) — not retrieved. The definitive list is on the '726 front page and in the file wrapper. Retrieve via:

• USPTO PatentCenter (application 08/736,848) — "References Cited" and full IFW.
• The '726 PDF (patentimages) front page, block "(56) References Cited."
• EP 0951453 B1 and WO 98/18731 A2 search reports, which may show the same art with examiner relevance categories (X/Y/A).

1. Whether any reference was applied as an X (anticipatory) reference — unknown without the file wrapper. The '726 claims survived as granted, suggesting no anticipatory art was of record.
2. Litigation flag — the Darts-IP family flag (family 24961556) remains unresolved; not a prior-art issue but relevant to validity challenges.

Uncertainty statement: I can state with confidence only that (a) the examiner-cited list is absent from the material available to me, and (b) the five applicant-cited references above do not, individually, anticipate claims 1 or 25 or any dependent claim. I cannot state that no anticipatory prior art exists — only that none is shown in the references I could verify. Any definitive § 102 table must be built from the front-page "(56) References Cited" and the file wrapper.

I'll start by grounding this analysis in the actual prior-art record for this patent, then build the § 103 case.

Good — that surfaced the key mechanistic art. Now let me pull the actual front-page citations of the '726 patent and any pre-1996 art on hydrogen permeation / electrochemical blistering.

§ 103 Obviousness Analysis — US 5,785,726

Method note / limitation up front. The prompt directs me to use "the results from the Prior Art section of this page." The page I was able to retrieve (via the fetched Google Patents record at https://patents.google.com/patent/[US5785726A](/patent/US5785726A)/en) shows the "Cited By (75)" list and the "Families Citing (15)" list — all forward citations — but does not expose a backward "References Cited / (56)" list. I therefore could not verify the examiner's actual cited art. My searches also ran out before I could pull that (56) list or the pre-1996 humidity-control art directly. I flag this because it materially bounds the analysis below: the combinations I build rest on (i) references cited inside the '726 specification, (ii) the pre-1996 glass-melt electrochemistry literature I could identify (largely via secondary citation), and (iii) post-1996 documents that are not prior art but are admissible as evidence of the background and as admissions. I have not invented any reference numbers; where I could not confirm a document's date, I say so.

(Also, consistent with the earlier sections: the "Current Date: April 26, 2026" in this task conflicts with later documents appearing in search results. I treat the patent record as authoritative and do not repeat the prior sections' content.)

---

1. Legal framework

The controlling test is Graham v. John Deere Co., 383 U.S. 1 (1966) — scope/content of the prior art, differences between the claims and the art, the level of ordinary skill, and objective indicia — as glossed by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). KSR is the decisive lever here for three reasons relevant to this patent:

1. A known problem + a finite number of identified, predictable solutions → obvious. Where the art recognized the problem (here, noble-metal-induced blistering/seeds) and the field had a small menu of known levers, "a person of ordinary skill has good reason to pursue the known options within his or her technical grasp." KSR, 550 U.S. at 421.
2. "Obvious to try" survives where the prior art gives a direction and there is a design need/ market pressure. Id. at 420–22.
3. Result-effective variables. Optimizing a process parameter (here, external hydrogen/water partial pressure or dew point) to achieve a known result (fewer blisters) is routine and is not saved by reciting the result ("sufficient to prevent formation of oxygen blisters"). In re Boesch, 617 F.2d 272 (CCPA 1980); In re Antonie, 559 F.2d 618 (CCPA 1977).

---

2. Level of ordinary skill

A POSITA here would be a glass scientist/ceramic or materials engineer with a B.S. plus ~3–5 years in high-temperature glass melting, fining, and delivery — including working familiarity with (a) platinum/Rh-Pt system design for fusion downdraw, (b) glass-melt electrochemistry (the Nernst equation, oxygen-ion-conducting reference electrodes), and (c) the standard fining-agent/β-OH toolkit. That skill level matters: the claim-1/25 solution is an atmosphere fix, but the diagnostic needed to aim it (measuring pO₂ at the Pt/melt interface) was squarely within this skill by 1996.

---

3. Construction of the operative limitations

---

4. Prior-art landscape

4.1 Pre-1996 art usable for § 103

4.2 Post-1996 documents — not prior art, but highly probative as background/admissions

---

5. The core teaching available before 1996

Combining §4.1 with the post-1996 confirmations, the following was known art by 1996-10-28:

1. The phenomenon and its mechanism. Glass melts contain water that partially dissociates to H₂ and O₂; Pt and Pt alloys are permeable to H₂; H₂ loss through the wall leaves an oxygen-enriched layer at the glass/Pt interface that nucleates O₂ bubbles. (DE 199 55 827 recites this as "long known"; the '726 specification itself recites the same mechanism and then says "While not wishing to be bound by theory…" — i.e., the applicants did not claim to have discovered the phenomenon, only to have exploited it.)
2. The driving force is the H₂ partial-pressure difference across the wall. This is elementary permeation physics; the '726 specification states it outright.
3. External pH₂ is set by the atmosphere's dew point and pO₂ via H₂O ⇌ H₂ + ½O₂ (Keq = pH₂·pO₂^½ / pH₂O). Corning's later patents describe this as the pre-existing understanding.
4. Diagnostics existed — Pt/Rh reference-electrode cells and zirconia solid electrolytes measuring pO₂ in glass melts by the Nernst equation (Plumat; Besson; Kapoor & Frohberg; Csaki & Dietzel; DE 4,324,922 A1; confirmed in maturity by NEG PA-06).
5. The complete menu of levers was known: dry the melt (low β-OH, halide drying agents); electrochemically counter the O₂ evolution (counter-potential/electron injection/Mo electrode); change/coat the vessel metal; and control the vessel's surrounding atmosphere.

---

6. Obviousness analysis, claim by claim

6.1 Claim 1 (independent) — obvious; strongest single ground

Combination: Dockerty '696 and/or '609 (Pt/refractory fusion-draw delivery system, glass/metal interface) + the glass-melt electrochemistry literature (mechanism: H₂ permeation → O₂ blister; driving force = ΔpH₂ across the wall) + known dew-point/humidity control of a gas atmosphere (water-gas equilibrium sets external pH₂).

Articulated reasoning (KSR-compliant):

• Problem recognized in the art: noble-metal-induced bubbling was a long-standing, well-documented yield problem (DE 199 55 827: "seit langem bekannt").
• Why the atmosphere lever specifically: If blistering is driven by a partial-pressure difference across a permeable wall, the POSITA is led directly to manipulate one side of that difference. The melt side was hard to change without reformulating the glass (which the industry was simultaneously trying to do for arsenic-free reasons); the outside of the vessel is freely accessible and its water content is trivially adjustable by a humidifier or a forming-gas blend. This is a textbook "adjust the known process parameter" step.
• Finite, predictable solutions: dry the melt / electrochemical countermeasure / change metal / control the atmosphere. Under KSR, the last option was a known, predictable choice within the artisan's grasp.
• Result-oriented limitation: "sufficient to prevent formation of oxygen blisters" is the result of the condition; it does not distinguish the claim from the prior-art process when the atmosphere is set at any blister-suppressing dew point.

The honest weakness in this ground: the claim requires the artisan to have already accepted the H₂-permeation mechanism as the dominant cause and to have chosen the atmospheric rather than the melt-side or electrochemical lever. If the pre-1996 art predominantly taught the anodic O₂-evolution model and counter-potential remedies (as the Schott family suggests), the direction chosen by claim 1 could be portrayed as a non-obvious departure. This is the battleground. Notably, the '726 specification's own concession that "previous efforts… have been hindered by the production of unacceptable amounts of bubbles" is classic long-felt-need evidence for the patentee.

6.2 Claim 25 (independent) — obvious

Claim 25 is the equilibrium/balance formulation: hold external pH₂ such that there is no net H₂ migration out of the glass. Given the known permeation physics (point 2 above), "zero net flux" is the trivial equilibrium solution and the obvious operating target once one accepts that permeation is the problem. The '726 specification states the objective itself — "the precious metal container would approximate an inert vessel" — i.e., the claim recites the goal, and the goal is the natural set point. In re Boesch (optimizing a result-effective variable to its apparent optimum) applies squarely.

6.3 Claims 2–8 — obvious

• Cl. 2 (adjust external pH₂ "in response to the presence of inclusions in the resultant glass") — routine feedback/routine experimentation.
• Cl. 3 (control water-vapor partial pressure outside relative to inside) — a direct restatement of the water-gas equilibrium already used to set pH₂.
• Cl. 4–8 (enclosure; humidity manipulation; humidify; dehumidify; dew-point control) — conventional industrial atmospheric control: humidifiers, dehumidifiers, dew-point controllers, leak-tight enclosures with slight positive pressure. The '726 specification itself says the enclosure "configuration… is not critical" and can be a room-sized enclosure or a "humidity jacket" — an admission that the hardware is conventional. Later art confirms this was the industry's standard ("traditional enclosure," US 7,473,568).

6.4 Claims 9–12 — obvious; possibly anticipated

• Cl. 9 (monitor vessel electric potential vs. a reference; adjust external pH₂ in response) and cl. 10 (monitor the DC current needed to hold a desired potential) are the potentiostatic/galvanostatic feedback loop built on the pre-1996 glass-melt pO₂ measurement art (Nernst cell with a Pt/Rh or zirconia reference electrode; Plumat; Besson; Kapoor & Frohberg; Csaki & Dietzel; DE 4,324,922 A1). The only "twist" — using a section of the vessel wall itself as the measuring electrode — is a design choice well within skill and is, in fact, later claimed outright by Schott (using the precious-metal wall as the measuring electrode).
• Cl. 11 (Pt: raise external pH₂ on a current decrease, lower it on a current increase) follows mechanically from the known permeation model.
• Cl. 12 (Mo: the opposite logic) follows from molybdenum's known reducing character toward oxide melts — the '726 specification states the Mo rationale expressly (avoid reduction of dissolved SO₃ to SO₂ bubbles). A POSITA would select the control polarity from the vessel metal's known redox behavior. (For a validity challenge, note that claim 12's logic is arguably inoperable as claimed, a § 112 issue rather than a § 103 one — the specification's own logic for Mo is about the melt's redox, not about a current direction.)

Combination for 9–12: pre-1996 glass-melt oxygen-sensor art + the atmosphere-control teaching of §6.1. Motivation: to close the loop, you must measure what you are controlling; the sensor art supplies the instrument.

6.5 Claims 13, 16, 17 — obvious

Downdraw (cl. 13), sheet-forming downdraw (cl. 16), and fusion (cl. 17) are each expressly disclosed in Dockerty '696/'609, which the '726 specification cites as the very definition of the process. Choosing a known forming process to practice a known glassmaking method is a classic obvious combination.

6.6 Claims 14, 15, 18–22 — obvious

• Cl. 18 (fining agent "Sb₂O₃, CeO₂, SnO₂, Fe₂O₃, halides, sulfates, and mixtures") — this is the standard fining catalogue; the '726 specification recites it as the known set of "fining constituents which are normally less efficient than As₂O₃."
• Cl. 19–20 (<0.1 mol% As₂O₃; essentially As-free) — a result the industry was actively pursuing for environmental reasons; the arsenic-free objective is prior art, and the '726 attributes attainment to the claim-1 atmosphere step.
• Cl. 21 (0.02–1 mol% Sb₂O₃) and cl. 22 (0.1–2 mol% halide) — conventional doses; halides as drying agents were known.
• Cl. 14 (<0.2 mol% As₂O₃) and cl. 15 (β-OH below ~0.5/mm) — reduction of melt water via batch selection and drying agents was known, and β-OH by IR at 2809 nm is a standard measurement. The cl. 14/cl. 15 numeric limits are the sharpest point for the patentee (see §7), but they are result-effective variables optimized to a known end.

6.7 Claims 23–24 — obvious

The aluminoborosilicate window (cl. 23) and the alkali-free/high-strain-point/CTE window (cl. 24) track US 4,824,808 and US 5,374,595, which the '726 specification expressly incorporates and identifies (the spec says the method "may be employed in glasses having compositions within the boundaries set forth above, such as… glasses listed as examples in U.S. Pat. No. 5,374,595"). Because these are method claims whose only method step is the claim-1 atmosphere control, they rise and fall with claim 1.

---

7. The patentee's best non-obviousness arguments — and how they fare

---

8. Bottom line — vulnerability ranking

Net assessment: Every claim except possibly the narrow numeric limits (14, 15) faces a strong KSR-style combination. The single linchpin for the broad independents (1 and 25) is whether pre-1996 art supplied the motivation to move the control lever from the melt side to the vessel's exterior atmosphere. The art I could reach supplies the mechanism, the thermodynamics, the diagnostic, and the hardware — three of the four Graham factors point to obviousness — but the motivation-to-combine prong is where a well-briefed patentee would concentrate, and where the record (the "previous efforts failed" concession) gives them their best material.

---

9. Verification gaps I am explicitly flagging

1. The (56) reference list of US 5,785,726 was not retrievable from the page I was given (only the forward "Cited By" list appeared). The examiner's actual art could materially strengthen or weaken §6. Anyone relying on this should pull the file wrapper at USPTO Patent Center.
2. Dates of the named pre-1996 electrochemistry references (Plumat; Besson; Kapoor & Frohberg; Csaki & Dietzel; Baucke; DE 4,324,922 A1) were reached via secondary citations and are not independently verified. They must be date-confirmed before being used in an invalidity contention.
3. DE 199 55 827 B4, EP 1 101 740 A1, WO 02/044115, US 2005/0050923, US 6,993,936, US 7,473,568, US 7,628,039, US 8,125,572, and JP 2011-068550 A are post-1996 and therefore cannot be § 102/§ 103 prior art against the '726. They are cited here only as (a) admissions about the background, and (b) evidence that the mechanism and the atmosphere-control approach became routine — which reinforces the § 103 case but is not itself the art.
4. The family/foreign counterpart is broader than the US patent. The DE 69730350 T2 translation (https://www.freepatentsonline.com/DE69730350T2.html) recites claims 1–29, including a method of measuring relative hydrogen flux through a glass-making vessel (claims 24–29) that has no US counterpart among the 25 issued US claims. If the foreign family is in scope for your project, note that the measurement claims are separately exposed over the pre-1996 Nernst-cell art.
5. The "5,758,726" string in US 7,032,412 (flagged in the earlier PTAB section) still appears to be a typo for 5,785,726; per the operating rules I have not auto-corrected it, but it should be scrubbed from any prior-art/family map.

Sources: https://patents.google.com/patent/US5785726A/en ; https://patentimages.storage.googleapis.com/46/82/cd/7605c7888c09c1/US5785726.pdf ; https://patentimages.storage.googleapis.com/f6/fc/31/bf87f91d2ab14b/DE19955827B4.pdf ; http://data.epo.org/gpi/EP1101740A1 ; https://patents.justia.com/patent/20040067369 ; https://patentimages.storage.googleapis.com/76/4a/66/70b72d98b07d78/US6993936.pdf ; https://patents.google.com/patent/US7475568 ; https://patents.justia.com/patent/7628039 ; https://patents.google.com/patent/JP2011068550A/en ; https://www.ceramics-silikaty.cz/2014/pdf/2014_04_249.pdf ; https://www.neg.co.jp/uploads/direct_observation.pdf ; https://www.freepatentsonline.com/DE69730350T2.html ; http://www.chinaglassnet.com/magazine/[2015319/12382](https://assignmentcenter.uspto.gov/search/patent/reelFrameDetail?reelFrame=2015319-12382).html