US 4846552

I'll verify the patent number and check for any 2026 litigation activity before summarizing.

Let me run one more targeted check for litigation/docket activity and any maintenance-fee status.

Verification note on scope

I was able to confirm US 4,846,552 across Google Patents, PubChem's patent record, and the official patent PDF, and I searched for litigation/docket activity. I do not have a live USPTO Patent Center (PAIR) or an authenticated CAFC docket-feed tool in this environment, so everything below is grounded in the sources I could actually retrieve (URLs cited). Where those sources conflict with each other, I flag it rather than silently harmonizing.

US 4,846,552 — Bibliographic Summary

Sources: https://patents.google.com/patent/US4846552/en ; https://pubchem.ncbi.nlm.nih.gov/patent/US-4846552-A ; http://patentimages.storage.googleapis.com/d4/83/7e/2fc7ed3adeb094/US4846552.pdf

Abstract (as published)

"The high diffraction efficiency regime of binary gratings occurs at periodicities on the order of a wavelength in order to produce grating which diffract radiation with efficiencies of greater than 90%, a fabrication procedure is disclosed which uses halographic and very large scale integration techniques which allow fine control over the periodicity and depth parameters to produce binary planar optical elements having a ratio of λ/T greater than one (where λ equals the wavelength of an illuminating wavefront, and T equals the grating periodicity). Additionally, the disclosed process produces high optical quality diffractive elements with phase precision of as high as λ/100. These diffractive elements include laser beam multiplexers, beam profile shapers, and binary lenses which are lossless optical transfer functions."

Plain-language overview of the single independent claim (Claim 1)

Claim 1 is a process claim for making a high-efficiency reflective binary planar optical element — i.e., a grating whose period T is on the order of the illumination wavelength λ. It recites the following ordered steps:

1. Make the mask pattern. Etch a lithographic pattern into a positive-resist layer coating a first substrate, using an electron-beam pattern generator at feature sizes as small as 0.25 µm and positional accuracy of 0.05 µm.
2. Finish the mask by chromium etching. The chromium etch produces a mask with transparent areas = the troughs of the rectangular grooves and darkened (chromium-coated) areas = the walls. The chromium sits only on the darkened areas; it blocks illumination from above, while the transparent areas pass it. The mask pattern corresponds to a reflective grating of periodicity T, where T approaches λ.
3. Prep the second (product) substrate. Coat it with a thick resist layer, then an aluminum layer, then a photoresist layer. The substrate must be quartz, silicon, plastic, or glass.
4. Contact print. Place the finished mask over the resist-coated second substrate and expose to incoherent ultraviolet light to print the pattern into the photoresist.
5. Wet etch the pattern onto the aluminum layer (aluminum acts as the etch barrier/hard mask).
6. Reactive-ion etch troughs and walls into the thick resist layer, defining the rectangular grooves on the second substrate.
7. Metallize. Deposit a reflective coating over the rectangular grooves (in the thick resist, the aluminum layer, and the second substrate) to form the reflective grating. The reflective coating must be one of: chromium, gold, aluminum, copper, nickel, or silver.

In short: e-beam-written mask → chromium etch → multi-layer resist/aluminum resist stack → UV contact print → wet etch aluminum → reactive-ion etch thick resist → reflective metal coat. The spec's design regime for high efficiency is duty cycle a/T ≈ 0.5 and depth ratio h/T ≈ 0.25, with λ/T greater than about 0.9, producing >90% diffraction efficiency and wavefront precision near λ/100.

Litigation / docket check

• My searches surfaced no CAFC 2026 decision, appeal, or district-court case involving 4,846,552. This is a negative search result, not proof that no litigation ever existed — I lacked a live, authenticated docket tool (PACER/CAFC feed), so I cannot certify the absence of suits.
• Consistent with no enforcement activity: the patent is a U.S. Government-owned (Air Force) patent, carries the standard government-interest statement, and is listed as expired (anticipated expiration 2006-07-11), which is well past any enforcement window.
• The patent is, however, heavily cited — Google Patents shows roughly 99–174 later publications citing it, spanning diffractive-optics, binary-optics, echelle-grating, and even Affymetrix microarray-art references. It was also incorporated by reference in later patents (e.g., US 5,151,917 cites "U.S. Pat. Nos. 4,846,552 and 4,895,790 ... incorporated herein by reference").

Points of uncertainty / anomalies (stated explicitly)

1. Priority date conflict. The patent's own text says the parent application, Ser. No. 852,587, was filed Apr. 16, 1986, and Google Patents lists the priority date as 1986-04-16. One secondary aggregator (Unified Patents) lists 1986-04-15. I treat the patent's own text as authoritative: 1986-04-16.
2. Assignee naming. Google Patents' "current assignee" field says "United States Department of the Air Force," while the printed patent and PubChem say "The United States of America as represented by the Secretary of the Air Force." These describe the same owner; the printed patent wording is the literal one.
3. Internal spec inconsistency on depth ratio. The detailed description contains a stray "a ratio of λ/h which equals about 0.25," whereas the figures and the rest of the spec describe h/T = 0.25. This appears to be a typographical error in the original text; I did not correct it in the quotation.
4. Claim count. There is exactly one claim, and it is independent — there are no dependent claims to summarize.
5. No verified USPTO status record. The "Expired – Fee Related" and 2006-07-11 expiration entries come from Google Patents' legal-status field, which Google itself labels an assumption rather than a legal conclusion. I could not independently confirm via a live USPTO Patent Center record with the tools available here.

If you need a certified litigation history or a current USPTO status extract, that would require direct PACER/Patent Center access, which I do not have in this session.

I'll search for litigation involving US patent 4846552.

Let me search more specifically on patent litigation databases.

Let me try a couple more targeted searches, including the Unified Patents litigation portal and Google Patents litigation section.

Litigation Search Results: US Patent 4,846,552

Bottom Line

No litigation involving U.S. Patent No. 4,846,552 ("Method of fabricating high efficiency binary planar optical elements") was found. Based on the searches run against litigation-covering sources (Unified Patents litigation content, Court of Appeals for the Federal Circuit materials, and general district-court/PACER-indexed material surfaced through web search), I found zero infringement suits, declaratory judgment actions, ITC investigations, or PTAB proceedings naming this patent.

Because no cases exist, I cannot provide the requested fields (plaintiff, defendant, jurisdiction, case number, filing date, outcome) — there is nothing to populate. I want to be explicit that this is a negative finding from the sources I could reach, not a certified "no litigation ever" conclusion, per your strict rule against fabricating results.

Why This Result Is Credible (Patent Profile)

The patent's characteristics make third-party or even government-initiated litigation very unlikely:

Key reasons litigation is unlikely/absent:

1. Government ownership. The patent is assigned to the U.S. Air Force and carries the standard government-interest statement ("may be manufactured and used by or for the Government ... without the payment of any royalty"). Federal ownership strongly discourages private enforcement actions.
2. A single, narrowly drafted process claim. The sole claim is a method claim directed to a specific fabrication sequence (E-beam mask generation → chromium etch → contact printing → wet etch → reactive-ion etch → reflective metallization). A one-claim process patent of this type is a weak and unattractive assertion vehicle.
3. Long-expired status. The patent expired around July 11, 2006 (confirmed by the Google Patents legal-status entries). Any past enforcement window closed roughly two decades ago, which is consistent with the absence of active cases.

Caveat on Sources

I attempted searches aimed at the specified litigation sources (including Unified Patents litigation listings, Federal Circuit/CAFC materials, and PACER-referenced content). The search results returned:

• The patent's own bibliographic record (Google Patents, PubChem) — no litigation tab entries surfaced;
• Multiple unrelated patents and cases sharing the digits "552" (e.g., U.S. Patent 5,411,552 in Edwards Lifesciences v. Medtronic CoreValve; U.S. Patent 7,454,002-related matters) — these are different patents and were not counted; and
• Unrelated business filings (e.g., a Florida LLC's FEI number "45‑4846552" and Portuguese/Lithuanian registry items) — these are not patent litigation and are irrelevant.

Per your instruction to interpret identifiers literally, I did not treat any similarly numbered patent or any coincidental numeric string as this patent.

If you need a certified negative or a formal litigation history certificate, I'd recommend confirming against:

• The Unified Patents litigation portal case list (https://portal.unifiedpatents.com/litigation/caselist), and
• PACER / the Court of Appeals for the Federal Circuit docket,
  using the exact string "4,846,552." I could not complete a fully exhaustive check of every district-court docket within the allowed search steps, so state this as a good-faith negative from available sources rather than an absolute guarantee.

I'll verify the structured "no proceedings on file" result against public sources before drafting.

Let me run a couple more targeted checks before concluding.

Proceedings overview

Zero AIA trial proceedings are on file for US 4,846,552 — the structured USPTO ODP block returns none, and my web searches surfaced none — so the correct bottom line for a defendant is: there is no PTAB record to litigate over, and the reason is not that the patent is "hardened" but that it went expired (2006-07-11) roughly six years before AIA trials even became available (2012-09-16), which is why no petitioner ever had a practical reason to file. This is a negative result, and I state it as such: it is corroborated by two independent sources (the ODP ingest and my searches), but I do not have live PTAB E2E / ODP API access in this session, so I cannot certify it as an authenticated docket extract.

Because there are no proceedings, the per-proceeding template below has no entries to fill. I've substituted a "why the record is empty" section with the same evidentiary rigor, plus decoy cases you should not mistake for this patent.

Why the PTAB record is empty (and why that is the whole story)

Two of the three AIA vehicle types are legally impossible here regardless of anyone's intent:

• PGR — available only for patents with an effective filing date on or after 2013-03-16. This patent's chain runs to 1986-04-16. Categorically unavailable.
• CBM — requires a claim to a method/apparatus for a financial product or service used in the practice, administration, or management of a financial product or service (and not a "technological invention"). A binary-grating fabrication process fails the eligibility gate.
• IPR — the only theoretically available vehicle (a patent can be IPR'd even after expiry for past-damages purposes). None was ever filed. Given expiry in 2006 and a government owner that never asserted, that is unsurprising rather than a signal about claim strength.

⚠️ Decoy / false-positive hits — do not cite these as proceedings on this patent

My searches returned several records that match only on the string "552" and are unrelated to US 4,846,552. Flagging them so they don't get attributed to this patent in any filing:

• IPR2025-01531, Apotex, Inc. v. Ipsen Biopharm Ltd. — concerns U.S. Patent No. 11,344,552, not 4,846,552. (Patent Owner's Preliminary Response available at https://paragraphfour.com/wp-content/uploads/2024/04/ipr25-1531PR.pdf.) Different patent, different technology (liposomal irinotecan), different decade.
• IPR2016-00782 (Samsung Electronics) — concerns U.S. Patent No. 6,784,552 ("Structure Having Reduced Lateral Spacer Erosion"). Different patent.
• IPR2024-01346 / IPR2024-01347, BMW of North America v. Foras Technologies — concerns the '958 patent; surfaced only incidentally. No relation.
• In re Swanson (CourtListener, https://www.courtlistener.com/opinion/[209660](/patent/209660)/in-re-swanson/) — this is the Hybritech/Abbott immunoassay line of cases (Swanson as a different inventor of record). Gary J. Swanson, the co-inventor here, is not the appellant of record in that matter as far as these results show. Do not treat this as a Federal Circuit appeal of 4,846,552.
• The Unified Patents portal hit showing priority date 1986-04-15 against a URL ending US-5895422-A is a mismatched aggregator record (it carries the '552 title/assignee but a different patent URL and a one-day-off priority date). This is the same anomaly flagged in the earlier bibliographic section (uncertainty #1). The patent's own text and Google Patents both say 1986-04-16; treat 1986-04-16 as authoritative and treat that portal entry as a data error, not a priority-date change.

Strategic summary

Claim status. There is no FWD, no certificate, and no estoppel on this record, so no claim has been canceled by the Board and none has been sustained by the Board — the only claim, claim 1, is untested at the PTAB and reached expiry on 2006-07-11 with its scope unaltered by any post-issuance proceeding. In practice, "untested" is not a defensive weakness for you; the operative fact is not claim validity but claim term. A process claim whose 17-year term ran out in 2006 cannot be infringed by activity occurring after that date, and a demand letter invoking it today would be asserting an expired patent.

Estoppel landscape. There is none to map. Section 315(e)(2) estoppel attaches only to a petitioner (and its RPOs/privies) after a final written decision. With zero petitions, zero institutions, and zero FWDs, no party is estopped, and conversely there are no "still-available" prior-art grounds that a prior petitioner has foreclosed — because there is no prior petitioner. The § 102/§ 103 art you'd find (including the very references already of-record: US 3,328,192; US 3,337,945; US 3,814,498; US 4,305,666; US 4,360,586; US 4,401,367; US 4,678,919; US 4,712,013; JP 58-106828; JP 58-125827; and the Beasley 1970 / Veldkamp SPIE 1983 / Wilson 1976 non-patent literature) remains available for an invalidity defense in litigation, but you are unlikely to need it.

Pattern signals. No repeat petitioner (there are no petitioners). No patent-owner appeal (nothing to appeal). No defensive aggregator in the chain — Unified Patents appears only as an incidental citing data source, not as a petitioner; do not characterize it as having filed anything against this patent. The government-ownership fact from the earlier section is the structural explanation for the empty record: an Air Force-owned, never-asserted patent attracts no IPR filings.

Recommended next steps

1. If you have received a demand letter citing US 4,846,552: the dispositive response is term, not validity. The patent issued 1989-07-11 as a pre-URAA patent with a 17-year term, and Google Patents lists an anticipated expiration of 2006-07-11 with legal status Expired – Fee Related (https://patents.google.com/patent/US4846552/en). Accused activity occurring after that date is not actionable. Quote that record directly rather than paraphrasing it, and treat "Expired – Fee Related" as "lapsed for failure to pay maintenance fees" unless and until you pull a live USPTO Patent Center maintenance-event history to confirm whether the lapse was fee-driven, term-driven, or both — I could not pull that live record in this session, and neither the earlier section nor this one should be treated as an authenticated USPTO status certificate.

2. There is no PTAB track to pursue. Because no petition was ever filed and none can now be meaningfully filed against an expired, government-owned, 1986-priority patent, an IPR-based defense is a non-starter — not because the patent "survived" anything, but because there is no live estate to challenge. If opposing counsel claims PTAB proceedings exist on this patent, ask them to produce the proceeding number; my searches found none, and any number they offer is likely one of the "552" decoys listed above.

3. Verify at the source before filing anything. The two authoritative free portals are USPTO PTAB E2E (https://ptacts.uspto.gov/ptacts/) and the USPTO Open Data Portal API (for the canonical proceeding list). For any appellate history, the Federal Circuit's docket and CourtListener are the right references — but here there is nothing to look up, since an FWD is a prerequisite to a PTAB appeal and no FWD exists. I recommend a same-day E2E/ODP re-query before you rely on the negative result in a brief, because a very recent filing (if any) may not yet be in the ingest.

4. Do not over-read the negative. The absence of IPRs on this patent is explained by expiry (2006-07-11) + government ownership + AIA timing, not by the claim being unusually strong. That distinction matters if you're writing this up: "the patent has never been challenged and has hardened" would be the wrong framing; "the patent expired roughly six years before AIA trials existed and was government-owned and never asserted" is the accurate framing.

Scope-of-verification note: The canonical list in the structured "PTAB proceedings on file" block is my primary source for the count of zero, and my web searches corroborate it. I do not have live, authenticated access to PTAB E2E, the ODP API, PACER, or a CAFC docket feed in this session. Accordingly, "no PTAB activity on file" is a documented negative result, not a certified absence. No proceeding numbers were invented here, and no claim-level outcome is asserted for a proceeding that does not exist.

I'll attempt to pull the actual assignment record before writing anything. Let me query the available search surface.

Let me run a few more targeted probes on assignment records and the inventors' affiliations.

Let me make a final couple of targeted attempts to locate any recorded assignment (including the inventor→Government confirmatory assignment) and check inventor employment.

Verification limitation (read first)

I could not open a live session against USPTO Assignment Center (https://assignmentcenter.uspto.gov/) or its legacy mirror (https://assignment.uspto.gov/patent/index.html) with the tools in this environment — I have a web-search surface only, no authenticated/rendered page fetch for that portal. I therefore reconstructed the chain from the sources I could retrieve, and I flag every point where I am inferring rather than reading a reel/frame. No reel/frame numbers appear below, because I did not retrieve any — I will not invent them.

What the accessible sources do show: Google Patents' legal-events table for US 4,846,552 contains no assignment/conveyance rows at all. It lists only: application filed (1988-02-09), priority to (1988-02-09), application granted (1989-07-11), publication (1989-07-11), anticipated expiration (2006-07-11). PubChem's record likewise carries only the face-of-patent assignee. No secondary aggregator surfaced a post-issuance transfer. For a U.S. Government–owned patent this is expected, but note the caveat in the timeline section below: Google's legal-events feed is not authoritative for the original inventor→Government recordation.

Inventors

Unusual patterns — assessment: none of the fire-sale precursor kind. Both inventors were FFRDC/Lincoln Laboratory staff, not Air Force civil servants. That is the single most important structural fact in this chain: it is why the assignee is "the Secretary of the Air Force" rather than MIT. Title vested in the Government via the inventors' assignment (the employment/FFRDC regime and an Executive Order 10096 determination), which is a mechanically different path from the E.O. 9424 "confirmatory license" path used when a contractor such as MIT retains title and merely licenses the Government — e.g., the MIT→Air Force Materiel Command E.O. 9424 confirmatory license recorded for US 6,156,435 (see https://legacy-assignments.uspto.gov/assignments/assignment-pat-066709-0686.pdf).

There is no evidence either inventor departed the original assignee within 12 months of filing, and the subsequent record points the other way: both continued publishing in the binary-optics program and appear on later, related patents in the same technology (Swanson on US 4,933,649 and US 6,417,967; Veldkamp on a multi-patent MIT-assigned set). No "inventors leave → portfolio dumped" signature.

Original assignee

On the issued patent: "The United States of America as represented by the Secretary of the Air Force." (Google Patents' "current assignee" field renders this as "United States Department of the Air Force" — same owner, looser wording.)

Assignment timeline

Finding: no post-issuance assignment is recorded in any source I could retrieve, and I could not retrieve the original inventor→Government recordation from Assignment Center. Stated plainly:

• Recorded post-issuance conveyances — NONE found. Google Patents' legal events for US 4,846,552 show no Assignment, Security Agreement, Merger, Change of Name, License, or Release entries. PubChem shows only the issuing assignee. No security-interest or merger lien surface exists because there was never a private owner to encumber the asset.
• The one ownership event that must exist by operation of the record: an inventor → U.S. Government (Secretary of the Air Force) assignment, executed between the 1986-04-16 priority filing and the 1989-07-11 issue, on the E.O. 10096 government-title track described above. I did not retrieve a reel/frame, execution date, recording date, or correspondent for it, and I decline to guess one. If you need that entry, it must be pulled directly at https://assignmentcenter.uspto.gov/ (search → patent number 4846552); it will be a pre-1990 reel (Assignment Center's coverage begins August 1980, so a 1986–1989 recording is in range and should appear).
• Nothing after 1989. No transfer-to-asserter, no securitization, no defensive-aggregation purchase, no change of name.

Because there is no post-issuance chain to itemize, the per-assignment bullet format requested (conveyance / assignor / assignee / correspondent / context) reduces to a single line, and I mark its unknowns explicitly:

• Unknown date (between 1986-04-16 and 1989-07-11) / recorded unknown — Reel unknown/unknown
  • Conveyance: Assignment of assignors' interest (inferred from the face-of-patent assignee; not independently verified against a reel)
  • Assignor: Wilfrid B. Veldkamp; Gary J. Swanson (individually)
  • Assignee: The United States of America as represented by the Secretary of the Air Force
  • Correspondent: not retrieved. No repeat-player attorney can be identified from this chain. (For context on what an Air Force correspondent looks like in this dataset, the 2018-vintage recordations route through Air Force Patent Counsel; no comparable record was retrievable here.)
  • Context: Government title vesting under Executive Order 10096 — an administrative ownership determination accompanying an FFRDC employee invention, not a commercial acquisition or fire-sale.

Timeline diagram

timeline
    title Ownership of US 4846552
    1986 : Parent application filed Apr 16
    1988 : Continuation filed Feb 9
         : Inventor assignment to US Government
    1989 : Patent issued Jul 11
         : Assignee is Secretary of the Air Force
    2006 : Anticipated expiration Jul 11

NPE / troll-pattern signals

Corroborating context for the "never sold" conclusion: the patent is heavily cited (Google's "Cited By" list runs roughly 99–174 later publications, spanning diffractive optics, echelle gratings, and even Affymetrix microarray art), and it was expressly incorporated by reference elsewhere (e.g., US 5,151,917 names "U.S. Pat. Nos. 4,846,552 and 4,895,790 … incorporated herein by reference"). High citation value sitting in Government hands with no recorded sale is consistent with a non-transferred asset, and is the opposite of the profile an aggregator would leave behind.

Verdict

Insufficient data — and, more precisely, only the original assignment.

Justification: the ownership chain terminates at the original assignee, The United States of America as represented by the Secretary of the Air Force, and I found no recorded post-issuance conveyance in Google Patents' legal-events table, PubChem, Unified Patents, or Patent Leaderboard; the only underlying transaction is the E.O. 10096 inventor→Government vesting that produced the face-of-patent assignee, for which I could not retrieve a reel/frame, execution date, or correspondent. Every NPE signal in the checklist returns not present or not assessable — there is no shell-entity hop, no known asserter, no cascade, no pre-suit transfer, and the sole owner is a sovereign agency whose asset expired 2006-07-11 — so no NPE or privateering verdict is supportable, and the "defensive aggregator" verdict also fails because no aggregator is in the chain. Verify directly at https://assignmentcenter.uspto.gov/ by searching patent number 4846552; the pre-1989 Government assignment, if recorded, will be the first and only entry.

Cross-references and one flagged inconsistency with the earlier summary

• The prior summary's legal-status call (Expired – Fee Related, anticipated expiration 2006-07-11) and its assignee-wording conflict (Google's "United States Department of the Air Force" vs. the printed "Secretary of the Air Force") are consistent with everything found here and are not contradicted.
• Flagged, unchanged from the prior section: Unified Patents lists this patent's priority date as 1986-04-15, whereas the patent's own continuation statement and Google Patents both say 1986-04-16. I continue to treat 1986-04-16 as authoritative. This does not affect the assignment analysis, since no ownership event is dated to the priority date.
• Correction to emphasis, not to fact: the prior summary characterized the patent as "heavily cited." That remains true and is now directly relevant to the verdict — a heavily cited asset that stayed in Government hands is strong corroboration that no NPE acquisition occurred.

I'll verify the cited references against searchable records before analyzing them. Let me run targeted searches on the patent and its key citations.

Good data on the key fabrication references. Let me pull the remaining citations.

Prior-Art Analysis — US 4,846,552

Scope and source note (read first)

• Literal identifier honored. This analysis covers only US Patent No. 4,846,552 (US4846552A). Per your standing rule, I did not treat any similarly numbered patent (e.g., 5,411,552; 7,454,002) as this patent.
• Database access caveat. I do not have a live, authenticated USPTO Patent Public Search / PatentCenter session in this environment, and I hit my retrieval-step ceiling partway through. The authoritative source for the citation list is the patent text itself (its "Patent Citations (11)" and "Non-Patent Citations" sections). Where I was able to independently confirm a reference against Google Patents / PubChem / FreePatentsOnline (which mirror USPTO data), I say so; otherwise I mark the description as unverified this session.
• Date caveat. Your task header states April 26, 2026; the operating environment date is 2026-09-30. This does not affect the analysis (the patent expired 2006-07-11).

1. The claim under attack

4,846,552 has exactly one claim, and it is independent — no dependent claims. It is a process claim with these limitations (L1–L7):

Because there is a single claim, the only §102 question is whether any one reference discloses all of L1–L7 arranged as claimed.

2. The 11 patent citations (examiner-cited)

3. The non-patent citations (3 unique)

Note: Google Patents/PubChem list these entries twice (once as "cited by examiner," once de-duplicated), which is why PubChem's summary header reads "9 Citations" while enumerating 11 patents + 3 NPL. The unique set is 11 patents + 3 NPL.

4. §102 anticipation analysis

Bottom line: no single cited reference anticipates Claim 1 under 35 U.S.C. § 102. Claim 1 is a seven-step combination process (e-beam mask → chromium-etch finished mask → laminated thick-resist/Al/photoresist substrate → incoherent-UV contact print → wet Al etch → reactive-ion etch of thick resist → reflective metal coat). Each cited reference supplies, at most, one or two isolated limitations, and several are directed to an entirely different field (heterodyne detection; waveguide devices). Anticipation requires all claimed elements in a single reference; none reaches that bar. These are properly characterized as background/§103 references, not anticipatory art.

Date/statute notes that matter:

• §102(a)/(b) art (public before the 4/16/1986 priority date): items 1–9 above (all published 1967–1983) and N1, N3. N2 (Veldkamp 1983) is also the inventors' own publication more than one year before filing, so it is squarely §102(b) art.
• §102(e) art (granted after the priority date, but filed before it):
  • US 4,678,919 — filed 1983-01-31, granted 1987-07-07 (post-priority) → §102(e) as of its 1983 filing date.
  • US 4,712,013 — filed 1984-09-29, granted 1987-12-08 (post-priority) → §102(e) as of its 1984 filing date.
    Both still fail §102 because each discloses only the e-beam/charged-particle patterning step (L1) and nothing of L2–L7.

Closest single reference — US 4,401,367 (UTC): It reaches the same result-type (a reflective IR diffraction grating built by a masked etch followed by deposition of an Au/Cr reflective layer) but does so by photoelectrochemical etching of a semiconductor mask layer, not by e-beam-written chromium mask + UV contact print + wet Al etch + reactive-ion etch of a thick resist. The specific process steps of Claim 1 are absent, so it does not anticipate. (It is, however, the reference I would flag as the primary §103 starting point.)

5. §103 perspective (for completeness)

Although you asked about §102, the honest technical assessment is that the examiner's citations were almost certainly assembled for obviousness, not anticipation. The two natural combinations are:

• Grating-fabrication art (UTC '367 / Bell '498 / Bausch & Lomb '945) for the reflective-grating structure and metallization, combined with
• E-beam pattern-generation art (Fujitsu '919, Toshiba '013, JP '828, JP '827, Beasley, Wilson) for the mask-writing step,
• plus Asahi Kasei '582 / MIT '586 for the photosensitive-element / fine-period lithography aspects.

The novelty of Claim 1 — and the reason the patent issued over these references — lies in the specific ordered micro-fabrication sequence for a binary (two-level) reflective element operating in the λ/T ≈ 1 (EM-resonance) regime, which none of the cited references teaches as a combination.

6. Discrepancies / anomalies to flag

1. Citation-list count inconsistency. Google Patents displays 11 patent citations and lists the NPL entries twice; PubChem's summary says "9 Citations." I treat the patent's own enumeration as authoritative: 11 patents + 3 unique NPL.
2. Two references are post-priority-date grants (US 4,678,919; US 4,712,013). They are prior art, if at all, only under §102(e) as of their filing dates — not as printed publications before 4/16/1986. Any anticipation theory resting on these two must be framed under §102(e).
3. N2 is the inventors' own prior work (Veldkamp et al., SPIE 1983). Its citation is material — it is §102(b) art against the 1986 filing — but it is the applicants' own publication, so it cannot support an "invented-by-another" §102(a)/(g) theory.
4. Verification status. I independently confirmed items 1, 2, 3, and 7 this session via Google Patents / FreePatentsOnline / the USPTO-derived PDF. I did not get to verify items 4, 5, 6 (US 4,360,582; US 4,360,586; US 4,305,666) against their full texts before reaching the retrieval limit; their descriptions above are drawn from the patent's own citation titles and are marked unverified. If you need certified abstracts for those three, that requires a follow-up pull from USPTO Patent Public Search or the patent images server.

If you need a formally certified prior-art/IDS transcript, it should come from USPTO Patent Public Search (patft/ppubs) or the file wrapper in PatentCenter for application 07/157,087 — I cannot certify that from the sources reachable here.

I'll ground the reference disclosures before analyzing. Let me pull what I can on the key cited references.

Obviousness Analysis — US 4,846,552 under 35 U.S.C. § 103

0. Scope, framework, and evidence limits (read first)

Governing law. This is a pre-AIA patent. Application 07/157,087 was filed 1988-02-09 as a continuation of Ser. No. 852,587 (filed 1986-04-16, now abandoned). Effective filing date for § 103 purposes is therefore 1986-04-16, and the pre-AIA version of § 103(a) applies: the question is whether "the differences between the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art." The analysis follows Graham v. John Deere (scope/content of the art; differences; PHOSITA level; secondary considerations) as elaborated by KSR Int'l v. Teleflex, 550 U.S. 398 (2007), which applies to pre-AIA patents and imports the "obvious to try," "predictable results," and "finite number of identified, predictable solutions" rationales into the TSM framework.

Evidence base. Per instructions I used the Prior Art section of this page: 11 cited patent documents (US 3,328,192; US 3,337,945; US 3,814,498; US 4,305,666; US 4,360,582; US 4,360,586; US 4,401,367; US 4,678,919; US 4,712,013; JP 58-106828; JP 58-125827) plus the non-patent citations (Beasley et al.; Veldkamp et al., SPIE vol. 437 (1983); Wilson et al. 1976).

Honest limitation, stated up front: I did not have full texts of most cited references. I retrieved and verified substantive text for three: the Veldkamp & Swanson 1983 SPIE paper (full abstract + fabrication section via Semantic Scholar and AbleSci), Beasley et al. (Philips Tech. Review vol. 37, pp. 334–346), and US 3,814,498 (PDF). For the remaining references I rely on titles, dates, and the patent's own characterizations, and I mark inferences as such. Do not treat unmapped disclosure statements below as verified.

Anomaly flagged (do not auto-correct): the patent cites Beasley et al., Philips Technical Review vol. 37, No. 11/12, pp. 334–346 as (1970). The retrieved article carrying those exact page numbers is from volume 37, i.e., 1977 (source PDF). Either date predates 1986, so prior-art status is unaffected, but the printed "1970" appears to be a citation error.

1. The claim decomposed (Claim 1, the sole claim)

Two construction points matter for obviousness:

• "T approaches λ" is broad. The claim does not require λ/T > 1. The specification's design examples use λ/T ≈ 0.90–1.83, and the abstract's "λ/T greater than one" is narrower than the claim. The claim therefore reads on gratings tested by the inventors themselves at λ/T = 0.90.
• "High efficiency" is a statement of intended result in the preamble and, absent claim language tying it to structure or a step, does not add patentable weight.

2. Level of ordinary skill in the art (PHOSITA)

A person of ordinary skill here would hold a degree in optics, electrical engineering, or applied physics with several years' experience in either diffractive/computer-generated holographic optics or semiconductor/VLSI lithographic process engineering — the two fields the patent itself says it bridges. Crucially, the patent concedes the relevant artisan existed: "[t]he great effort expanded by the integrated circuit industry has provided the capability to produce binary phase relief patterns with 0.5 µm periodicities and 0.1 µm accuracies over areas with 20 cm or more diameter," calling these "prerequisites" to the invention. That admission defines the ordinary artisan as someone already equipped with IC mask-making and dry-etch skills.

3. What the prior art actually discloses

4. Combinations that render Claim 1 obvious

Combination A (strongest): Veldkamp & Swanson 1983 in view of Beasley et al.

Mapping. Veldkamp 1983 discloses L1 (binary diffraction gratings/lenses, EM-regime, h/T = 0.25, a/T = 0.5, λ/T spanning 0.90–1.83), L3's chromium mask and the transparent-opaque groove/wall correspondence, L4 in substance (resist-coated substrate with an intermediate protective metallization layer), L6 (contact printing), L7–L8 (two-step transfer, reactive-ion etching), L9 (Cr–Au reflective coating), and L5 (glass, plastic, silicon substrates). Beasley supplies L2 (e-beam pattern generator writing directly into resist for mask-making).

Motivation. (i) The two references are in the same field of endeavor (mask generation for microfabricated devices) and are reasonably pertinent to the problem the inventors state — how to lay down a 0.5 µm-period binary relief pattern with high acuity. (ii) Beasley expressly frames its machine as producing "a mask which can be copied by other means" — precisely the Veldkamp contact-print workflow. (iii) Under KSR, "if a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious." Substituting a higher-resolution e-beam writer for the Mann-3600 optical pattern generator to improve pattern fidelity is exactly that. (iv) Decisively, the patent admits the substitution: "Standard pattern generators in the art use either a laser, electron or ion-beams to trace the pattern in a positive resist on the surface of a glass substrate." An applicant's own admission that electron-beam writers were standard for this exact step is close to dispositive.

Combination B: Combination A + Wilson 1976 and the e-beam patent family (JP 58-106828, JP 58-125827, US 4,678,919, US 4,712,013)

What it adds. Only the specific numbers — 0.25 µm feature size, 0.05 µm positional accuracy.

Motivation. These are result-effective variables (resolution and placement accuracy directly control the phase-quantization error that limits diffraction efficiency) susceptible to routine optimization (In re Boesch; In re Aller). The art gives a finite, predictable menu of e-beam capability, so the claimed values are an "obvious to try" selection (KSR; In re O'Farrell). And again, the patent concedes the point: "Current electron beam technology allows for feature sizes as small as 0.25 µm, and positional accuracy of ±0.05 µm." The claim's numerals are lifted verbatim from the specification's description of the existing art — the strongest possible § 103 admission.

Combination C: Combination A + US 4,401,367 (pattern masking) — for the aluminum hard mask

What it adds. L4's aluminum identity and the wet-etch-stop/hard-mask function.

Motivation. Using a metal as an etch mask over an organic planarity layer is standard IC practice, and the specification itself says so: aluminum "has a relatively low etch rate in nearly all useful reactive gases … and is a better thermal match to the underlying resist than chromium. This is one of many possible processing steps commonly available in the electronics industry." A disclosed, result-driven selection from known materials with the stated rationale is textbook obviousness (KSR "combination of familiar elements according to known methods").

Combination D (alternative primary): US 3,814,498 (Bell) + Beasley + the e-beam family + US 3,328,192 / US 3,337,945

For a rejection not dependent on the inventors' own publication, Bell supplies wavelength-scale grating design and fabrication through photolithographic/etch techniques; Soules and Bausch & Lomb supply grating master production and transfer-film replication (the contact-print step); Beasley and the e-beam family supply the writer. Motivation: the well-recognized design incentive to push grating period toward the illumination wavelength to operate in the EM/diffractive regime, combined with the known desire to replicate masters by contact printing for low-cost mass production. This combination is weaker than A–C (Bell's gratings are waveguide-scale and its disclosure of the exact process flow is thinner), but it independently reaches L1, L3, L6–L9.

Combination E: Full "all-elements" rejection

Veldkamp 1983 + Beasley + Wilson + JP 58-106828 + US 4,712,013 + US 4,401,367 + US 3,337,945 + US 3,814,498. Every limitation L1–L9 has a mapped home. Under KSR, where each element was known and the combination yields no more than predictable results, the claim is obvious.

5. Why the combination motive is legally sufficient

• Same field / reasonably pertinent. All references sit in optical-element fabrication or semiconductor mask-making — no field-of-invention gating issue.
• Predictable results. Substituting a higher-resolution writer and a known metal hard mask into a known contact-print/RIE/Cr-Au flow yields a better grating — an expected, not unpredictable, result.
• Finite solutions. The claim's variables (resist/E-beam, hard-mask metal, substrate choice, reflective metal) each resolve to a short, enumerated list — the KSR "finite number of identified, predictable solutions" scenario. The claim's own "selected from a group consisting of" phrasing for the reflective metal (Cr, Au, Al, Cu, Ni, Ag) is itself a catalog of recognized reflective/conductive metals.
• Admissions. The specification repeatedly characterizes the claim's key process elements as prior, standard capabilities (laser/electron/ion writers; 0.25 µm/±0.05 µm e-beam; aluminum as a known hard mask; IC-derived 0.5 µm patterns/0.1 µm accuracy). Admissions in the specification are prior art (In re Nomiya).

6. Secondary considerations (objective indicia)

1. "Unexpected" efficiency (>90%, λ/100 precision). This is the applicant's only potential nexus argument, and it fails on the record: the inventors' own 1983 SPIE paper already reports 94% first-order efficiency at 10.6 µm and diffraction-limited point-spread functions using the same EM-regime design (a/T = 0.5, h/T = 0.25). The result was neither unexpected nor theirs to claim as evidence of nonobviousness over their earlier disclosure.
2. No long-felt-need / failure-of-others evidence of record; the workflow was a straightforward importation of established IC processing into optics, which the 1983 paper had already done optically.
3. No commercial-success nexus — the patent is U.S. Government-owned (Air Force), expired (anticipated 2006-07-11), and no licensing/market evidence is of record.

7. Counterarguments a patentee would raise, and responses

One § 102 note: absent the e-beam limitation, Veldkamp 1983 alone would approach anticipation (L1, L3–L9 substantially disclosed; Veldkamp used a Mann-3600 optical pattern generator, not e-beam). Because the paper does not disclose an electron-beam writer, § 102 fails but § 103 succeeds — which is precisely the combination A analysis above.

8. Conclusion

Claim 1 would have been obvious under § 103. The dominant rejection is Veldkamp & Swanson (SPIE 437, 1983) in view of Beasley et al., optionally further in view of Wilson et al., the JP/US e-beam family (JP 58-106828, JP 58-125827, US 4,678,919, US 4,712,013), US 4,401,367, US 3,337,945, and US 3,814,498. The single non-disclosed element in the inventors' own earlier publication is the electron-beam pattern generator (with 0.25 µm / 0.05 µm performance), and the specification itself concedes that electron-beam pattern generation at those capabilities was standard, current technology. Every other limitation — binary grating in the EM regime, chrome mask, resist/planarizing stack with a metal hard mask, UV contact printing, wet etch, reactive-ion etch, Cr–Au reflective coating with ±45° rotation, and the substrate set — is disclosed in the 1983 paper. Under Graham/KSR, with no available secondary-consideration nexus (the efficiency result was already published by the same inventors in 1983), the claim is a combination of known elements yielding predictable results.

Confidence: High for Combination A (Veldkamp 1983 + Beasley), based on verified text of both references. Moderate for Combinations B–D, because several underlying references were assessed from titles/dates and the patent's own characterizations rather than full text; a rigorous rejection would require quoting the exact passages of US 4,401,367, the JP/US e-beam references, US 3,328,192, US 3,337,945, and US 3,814,498. No fabrication of reference content has been made — where I inferred rather than verified, I said so.

Caveat on the anomaly flagged earlier: if the Beasley citation is in fact 1977 (as the retrieved Philips volume-37 article suggests), nothing changes — it remains prior art — but the discrepancy should be reconciled before any rejection is finalized, along with the specification's internal "λ/h ≈ 0.25" vs. "h/T ≈ 0.25" typo and the tension between the claim's 0.05 µm positional accuracy and the specification's statement that the machine actually used had an "atypically large quantization step size of 0.5 µm."