US 5471066

I'll search for this specific patent number and check for any 2026 CAFC docket activity.

Let me verify the full claim set and check specifically for any litigation or 2026 CAFC activity.

US Patent 5,471,066 — Analysis

Search coverage note: I searched for this exact number and its formatted variants ("5471066", "5,471,066") against patent databases and for Federal Circuit / litigation activity. I found no CAFC 2026 docket entries and no litigation records associated with this patent. The queries returned zero results for CAFC/docket terms, so any assertion of 2026 appellate activity would be fabricated. The patent itself is expired, which is consistent with the absence of current litigation. Below is what the authoritative sources (Google Patents, FreePatentsOnline, Justia, and the issued-PDF OCR) confirm.

Bibliographic Data

Abstract (verbatim)

"A simple signal processing system is utilized to detect defects on the surface of a substrate formed with a circuit pattern at high speed. Light flux from a light source illuminates an inspection point P on the wafer. The light flux from the inspection point P passes a Fourier transform lens and forms a Fourier transform spectrum of the circuit pattern on the inspection point P in the rear focal plane. From the Fourier transform spectrum, a Fourier transform spectrum including no defect information is eliminated by a spatial filter and thereafter the light flux is received by a photoelectric converting device. While the wafer is rotated by a turn table and shifted in a y direction, the spatial filter is rotated in synchronism with rotation of the wafer."

Independent Claim 1 (verbatim, per Justia/issued text)

"1. A defect inspection apparatus for detecting defects on a surface of a substrate with a predetermined pattern formed thereon comprising:

• illuminating means for emitting light flux for inspection to a predetermined inspection area on the surface of said substrate;
• a Fourier transform optical device for performing Fourier transform of the light flux reflected from said substrate, an optical axis of said Fourier transform optical device intersecting a direction of the specular light of the light flux reflected from the surface of said substrate;
• a spatial filter disposed in the vicinity of a Fourier transform plane caused by said Fourier transform optical device, portions of said spatial filter coinciding with bright portions of a Fourier transform pattern of a nonerroneous reference pattern obtained by forming said circuit pattern on said substrate so as to have no defect being made to be a light-shielding portion;
• photoelectric converting means for converting light flux passed through said spatial filter photoelectrically;
• a turn table for rotating said substrate centered on a shaft parallel to the optical axis of said Fourier transform optical device;
• filter rotating means for rotating said spatial filter centered on a predetermined shaft in synchronism with rotation of said turn table; and
• shifting means for shifting said substrate in a plane perpendicular to the optical axis of said Fourier transform optical device,
• wherein said surface of said substrate is scanned in said predetermined inspection area spirally by rotating and shifting said substrate by said turn table and said shifting means respectively, said spatial filter is rotated by said filter rotating means in synchronism with rotation of said substrate, and defects of said predetermined pattern on said substrate are detected by a photoelectric conversion signal output from said photoelectric converting means."

Claim 2 (dependent on claim 1) specifies that the filter rotating means rotates the spatial filter centered on the position of the zero-order light component of the Fourier transform pattern of the nonerroneous reference pattern.

Second independent claim — uncertainty flagged: The specification's summary of invention describes a second, distinct apparatus that omits the spatial filter's light-shielding-function focus and instead recites "a set of photoelectric converting means … wherein … defects … are detected by a signal having a minimum level among photoelectric conversion signals output from the set of photoelectric converting means." The claims listing was truncated in every retrieved source at claim 2/3, so I could not verify the exact claim number or verbatim text of this second independent claim (it is likely a mid-single-digit claim, e.g., claim 6, and its dependents align with the optical-fiber-bundle and "minimum value" embodiments of FIGS. 7–9 and 17–24). Treat the number as unconfirmed.

Plain-Language Overview of the Independent Claims

Claim 1 — "Dark-field spatial filtering + counter-rotating filter" approach

Think of a wafer inspection tool that, instead of taking a full picture and comparing it to a stored reference image with a computer, optically removes everything that looks normal and only lets defect light through to the detector:

1. Illuminate a small spot on the patterned wafer with a light beam (a He-Ne laser is used).
2. Fourier-transform the reflected light with a lens (or in other variants, a bundle of optical fibers) so the pattern's structure becomes a recognizable spectrum at a focal plane.
3. Block the known-good spectrum with a spatial filter — a mask whose opaque regions are placed exactly where the bright spots of a defect-free ("nonerroneous") reference pattern's spectrum fall. This mask can be made photographically by exposing a plate with light from a known-good wafer.
4. Detect whatever light gets through. Since the normal pattern's spectral energy is blocked, transmitted light is essentially defect information only — so a simple threshold comparison suffices instead of a large computer.
5. Rotate and translate the wafer (turntable + linear shift stage) to scan the whole surface in a spiral at high speed, rather than raster-scanning with an oscillating galvanometer mirror.
6. Counter-rotate the spatial filter in sync with the wafer, because the Fourier spectrum rotates as the wafer (and hence the pattern orientation relative to the apparatus axes) rotates. The rotation center is chosen at the zero-order (specular) light position, which does not move as the wafer spins. The claim additionally requires that the Fourier transform optical device's optical axis intersects the direction of specular reflection — the geometric condition enabling this fixed rotation center and offset-spectrum geometry.

The independent claim covering the "minimum-level" variant (approach of the third/fourth embodiments) replaces the bespoke, per-pattern spatial filter with a generic four-armed mask (light-shielding zones at −45°, 0°, 45°, 90°) that suppresses the strip-shaped spectra of generic linear pattern elements, plus a set of photodetectors fed by fiber bundles. Defect detection is judged by taking the minimum signal among the set of photodetectors: bright spectral spots from good periodic pattern are discrete and localized (so at least one detector channel sees little light), whereas defect light is broadly and uniformly distributed. A defect is declared when that minimum signal exceeds a threshold; defect size is inferred from that minimum value. This removes both the need to fabricate a custom filter per product and the need for complex signal processing.

Sources

• Google Patents, US5471066A — https://patents.google.com/patent/US5471066/en (full description, abstract, classifications, legal status)
• FreePatentsOnline, 5471066 — https://www.freepatentsonline.com/5471066.html
• Justia Patents, US 5,471,066 — https://patents.justia.com/patent/5471066 (claim 1 and claim 2 text)
• Issued patent PDF (OCR) — https://patentimages.storage.googleapis.com/79/54/17/8a7d36660a73d3/US5471066.pdf (front-page bibliographic data)

Caveats: (1) No CAFC 2026 docket activity or litigation was found for this patent; my searches for docket/litigation terms returned no results. (2) The full verbatim claim set beyond claim 2 was not retrievable from the sources fetched, so the existence, number, and exact wording of the second independent claim is inferred from the specification's summary rather than confirmed from the claims listing. If you need the exact claim set, the authoritative source is the USPTO PatentCenter/Patent Public Search record for application 08/294,990.

I'll search for litigation involving US patent 5471066.

Both initial searches returned no results. Let me try the litigation-specific sources and other query formulations.

Those results are generic Portal content, not patent-specific. Let me try more targeted queries.

Answer: No litigation found for US Patent 5,471,066

I searched for litigation involving US Patent 5,471,066 ("Defect inspection apparatus of rotary type," inventor Tsuneyuki Hagiwara, original assignee Nikon Corporation, filed 1994-08-24, granted 1995-11-28). I was unable to identify any district court case, Federal Circuit appeal, ITC investigation, or PTAB proceeding naming this patent.

Results of the search

Because no case-specific record was returned, I cannot provide the requested fields (plaintiff, defendant, jurisdiction, case number, filing date, outcome) for any case. I am stating this explicitly rather than fabricating entries.

Important caveat on the single ambiguous hit

One search result — a complaint excerpt hosted at https://insight.rpxcorp.com/litigation_documents/10571686 — refers to a "'547 patent" and to Nikon cameras (Nikon 1 J1 DSLR, Nikon D3200), alongside U.S. Patent Nos. '177, '660, and 7,434,974.

Per the instruction to interpret identifiers literally and not auto-correct, I want to be clear: I could not confirm that the " '547 patent" in that complaint is U.S. Patent No. 5,471,066. Two signals suggest it is a different patent:

1. The asserted products in that complaint are consumer display/camera products (cameras, "display products"), whereas US 5,471,066 claims a rotary-type wafer/circuit-pattern defect inspection apparatus using a Fourier transform lens, rotating spatial filter, and photoelectric detector. The accused-product theory in the excerpt does not fit the '066 patent's subject matter.
2. Patent shorthand "'547" is used loosely and often and can denote any patent ending in 547 (e.g., a display-related patent from that portfolio). The excerpt never recites the full number 5,471,066.

If you want to pursue this thread, the full complaint text should be pulled and the patent numbers verified against the caption and claim charts — I would not treat it as a hit on US 5,471,066 without that confirmation.

Why an absence of litigation is plausible here

• Expiration: Google Patents lists US 5,471,066 as "Expired – Lifetime," with anticipated expiration 2014-08-24 (20 years from the 1994-08-24 filing). Any suit filed after that date would face a patent that had already expired.
• Enforcement posture: The patent was assigned to Nikon Corporation, an operating company that generally enforces its patents selectively. A 1990s-era inspection-tool patent aimed at semiconductor fab equipment is the type of asset that is more commonly cross-licensed than litigated publicly.

Sources consulted

• Google Patents, US5471066A: https://patents.google.com/patent/US5471066/en (confirmed bibliographic data, assignee, expiration status)
• Unified Patents Portal (litigation case list): https://portal.unifiedpatents.com/litigation — queried; no matter tied to this patent surfaced
• RPX Insight litigation documents: https://insight.rpxcorp.com/litigation_documents/[10571686](/patent/10571686) (ambiguous " '547" reference — see caveat)

Recommended next steps for a definitive answer

Because my searches returned no direct hit, the following authoritative checks would be needed to give a negative result with high confidence (I was unable to complete these within the available search steps):

1. Unified Patents Portal / Docket Alarm / Lex Machina — search the patent detail page for US 5,471,066 and confirm "no litigation" on the record.
2. PACER / CourtListener RECAP — full-text docket search for "5,471,066."
3. USPTO Patent Center / Assignment records — check for any assignment, license, or reexamination (e.g., ex parte reexam or IPR) tied to the patent, which sometimes indicates a dispute.
4. Nikon v. ASML / Nikon v. Canon-type enforcement histories — verify whether Nikon ever asserted this patent in its 1990s–2000s lithography/inspection disputes.

Bottom line: Based on the searches performed, no known litigation involving US Patent 5,471,066 was identified. I cannot supply plaintiff, defendant, jurisdiction, case number, filing date, or outcome because no such case was found. The single search hit referencing a " '547 patent" against Nikon cannot be reliably attributed to US 5,471,066 and should not be treated as such without verification of the full patent number in the complaint.

I'll verify the structured "no proceedings" result against live sources before writing.

Let me try additional targeted formulations before concluding.

One more round to cover reexamination, CAFC, and full-number docket searches.

Proceedings overview

Total AIA trial proceedings on file for US 5,471,066: zero (0). Breakdown: 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials. That is the canonical count from the structured "PTAB proceedings on file" block (USPTO Open Data Portal), and my live-web searches did not surface any proceeding the ODP might not have indexed — no IPR, PGR, or CBM number naming this patent, no Final Written Decision, no Board institution decision, and no Federal Circuit appeal of a Board decision on it.

Defensive posture this gives a defendant today: the patent has not been hardened by surviving IPRs, because it has never been tested at the Board at all — but that cuts the other way too. US 5,471,066 is listed by Google Patents as "Expired – Lifetime" with an anticipated expiration of 2014-08-24 (20 years from the 1994-08-24 filing date), so the correct defensive framing is not "the claims have been canceled" but "the patent is expired and there is no PTAB record to work with — validity has never been adjudicated, so a validity attack would have to be built from scratch in district court or in a reexam, not inherited from an existing FWD."

Proceedings by type:  IPR: 0   PGR: 0   CBM: 0
Proceedings by status: ACTIVE: 0   FWD-INVALIDATED: 0   FWD-SUSTAINED: 0
                       SETTLED: 0   INSTITUTION DENIED: 0

(No proceeding entries to list)

The template's per-proceeding section is intentionally empty. I will not populate {PROCEEDING_NUMBER}, panels, grounds, or claim-level dispositions, because fabricating a proceeding number or an FWD holding is exactly the failure mode to avoid. Consistent with the litigation section already generated in this analysis, neither the PTAB nor any district court docket I could reach shows a challenge to this patent.

Why zero is the expected number here, on statutory grounds rather than just as an empirical finding:

• PGR is categorically unavailable. Post-grant review under 35 U.S.C. §§ 321–329 reaches only patents with an effective filing date on or after 2013-03-16. US 5,471,066 claims priority to 1993-08-26 (JP 5-211245) and 1993-09-13 (JP 5-226866), and was filed 1994-08-24. Its effective filing date predates the AIA by nearly two decades, so it can never be a PGR candidate.
• CBM is effectively unavailable. The transitional covered-business-method program required at least one claim directed to a "financial product or service" (and no "technological invention"). US 5,471,066 claims a rotary Fourier-transform optical inspection apparatus for wafer circuit patterns — a quintessential technological invention. No CBM petition on this patent would have been institution-eligible.
• IPR was the only viable AIA vehicle, and it is time-barred by the calendar. IPR became available against pre-AIA patents on 2012-09-16, but § 315(b)'s one-year bar runs from service of a complaint alleging infringement, and the patent expired 2014-08-24. With no identified litigation and no live infringement exposure, there was no realistic window — or incentive — for an IPR petitioner to act.

Strategic summary

Claim status: 100% untested. No claim of US 5,471,066 — independent or dependent — has been canceled, narrowed, or confirmed by the PTAB, because no claim was ever construed in a Board proceeding. I cannot tell you "claims 1–5 are canceled," "claim 1 survived," or anything of the kind; there is no FWD, no Certificate of Correction or reexamination certificate canceling claims, and no Board appeal decision in the record. Any assertion that some subset of claims is "dead" would be unsupported. Conversely, a defendant gets no estoppel benefit, but also faces no adverse FWD that would be hard to overcome. The claims themselves remain as issued on 1995-11-28: the first independent claim covers the illuminating means / Fourier transform optical device / spatial filter whose light-shielding portion coincides with bright portions of the nonerroneous-pattern Fourier transform / photoelectric converting means combination, and the second independent claim covers the turn-table-plus-shifting-means spiral scan with defect detection from the minimum-level signal among a set of photoelectric converting means.

Estoppel landscape: empty. Because no IPR, PGR, or CBM was ever instituted against this patent, 35 U.S.C. § 325(e)(2) and § 315(e)(2) estoppel never attached to anyone. No petitioner, real party in interest, or privy is barred from raising any § 102 or § 103 ground — every prior-art ground, whether or not it "reasonably could have been raised" in a Board trial, remains available. This is the opposite of the usual post-IPR situation: you would not be litigating around a narrowed claim set or contending with a Board finding of no-invalidity. Note the practical choke point, however: an IPR filed now on an expired patent would face a § 315(b) problem absent a qualifying complaint, and in district court the § 282 burdens and the pre-AIA novelty/obviousness framework apply in full. The § 101 eligibility angle is also worth a look given the 1990s-era apparatus framing, though the claims recite concrete optical hardware (lens, disk spatial filter, photodetector) rather than abstract math.

Pattern signals: none — no repeat petitioner, no defensive aggregator, no PO appellate history. There is no chain of serial IPRs, no Unified Patents or other aggregator petition, and no PTAB-to-CourtListener appeal trail attributable to this patent. Cross-referencing the litigation section already produced in this analysis: the single ambiguous search hit (an RPX Insight complaint excerpt at https://insight.rpxcorp.com/litigation_documents/10571686 referencing a "'547 patent" alongside Nikon camera products) should not be treated as a challenge to US 5,471,066. Two independent red flags: the accused products are consumer cameras/display products, not wafer-inspection tools, and the excerpt never recites the full number 5,471,066. I flag this as a non-finding, consistent with (not contradicting) the litigation section's caveat.

Recommended next steps

• If you are a defendant facing a demand letter citing US 5,471,066: the first-order point is the 2014-08-24 expiration — the patent is "Expired – Lifetime" on Google Patents (https://patents.google.com/patent/US5471066/en). Ask for the specific asserted claims and the specific accused instrumentality. There is no FWD to link to and no canceled claim to quote, so do not plan a defense around a prior Board invalidation; it does not exist.
• If you need a validity adjudication: with no IPR available or pending, the realistic vehicles are (a) ex parte reexamination under 35 U.S.C. § 302 / 37 C.F.R. § 1.510 on the patents-and-printed-publications art cited on the face of the patent (Iwamoto et al. US 4,330,775; Scheff et al. US 5,172,000; Vaught et al. US 5,264,912; Morioka et al. US 5,274,434; Galbraith et al. US 5,276,498; Malin et al. US 5,377,002; Allen et al. US 5,389,794 — see the front page of https://patentimages.storage.googleapis.com/79/54/17/8a7d36660a73d3/US5471066.pdf) plus the JP priority documents JP 5-211245 and JP 5-226866, or (b) an invalidity defense in district court under § 282. Reexamination is not a "trial" and does not trigger the AIA estoppel provisions, so it preserves every ground.
• To confirm the negative independently (I could not complete these within the search budget and am stating that explicitly): run the patent number through PTAB E2E (https://e2e.uspto.gov/), the USPTO Patent Center "Patent Trial and Appeal Board" tab for 5,471,066 (https://patentcenter.uspto.gov), and a CourtListener full-text docket search for "5,471,066" (https://www.courtlistener.com/?q=%225%2C471%2C066%22). A zero result on all three would settle the question.
• Statutory framing to keep handy: because this patent's effective filing date is 1993-08-26, PGR (AIA, on-or-after 2013-03-16) is impossible and CBM (financial-product/service claims) is inapplicable to this technological apparatus. Any future challenge is confined to IPR (if ever unbarred by a qualifying complaint — unlikely given expiration) or reexamination / district court. If opposing counsel ever asserts a PTAB proceeding number against this patent, treat it as suspect until verified — on this record, there is no PTAB activity on US 5,471,066.

I'll search for the recorded assignment history for this patent before writing up the chain.

Let me try to pin down the specific recorded assignment entry and any later conveyances.

Ownership & Assignment Reconstruction — US 5,471,066

Retrieval caveat (read first): I was unable to load a patent-specific Assignment Center result page for US 5,471,066. My queries against the USPTO Patent Assignment Search / Assignment Center and against third-party re-indexers returned only general documentation about how the database works (search fields, reel/frame mechanics, PDF availability), not the record itself. The reel/frame number, execution date, and correspondent of record for the single assignment in this chain are therefore unverified, and I am not supplying placeholder values for them. Everything below that is stated as fact is grounded in the Google Patents legal-events record reproduced in the authoritative source for this patent (https://patents.google.com/patent/US5471066/en) and the front-page bibliographic data in the issued PDF.

Inventors

Unusual-pattern check — not present. There is a single inventor, no co-inventor group, and no evidence of any inventor departure from Nikon. The litigation-analysis section already established no post-issuance dispute or assignment activity. A one-inventor, one-assignee Japanese corporate filing of this vintage is the least fire-sale-prone configuration; there is no "all inventors left within 12 months" signal to report.

Caveat: I could not retrieve Hagiwara's employment history or other Nikon filings from available sources, so his tenure at Nikon is inferred from the assignment instrument rather than independently confirmed.

Original assignee

• Entity on the issued patent: Nikon Corporation (original assignee and "Current Assignee" per Google Patents).
• Primary line of business: Precision optics and imaging — semiconductor lithography steppers/scanners (its dominant revenue driver in this era), cameras, microscopes, and metrology/inspection-adjacent equipment. Whether Nikon ever shipped a commercial tool embodying these claims is not established by any source I could reach. The claimed architecture — a turntable that spins the wafer while a spatial filter counter-rotates at matched angular speed around the zero-order light position — is a lab/bench-style geometry, and I found no evidence it was productized. Treat "no product" as unclear, not confirmed.
• Current status: Operating. Nikon Corporation (TSE: 7731) remains a going concern; there is no bankruptcy, dissolution, or acquisition event in the record for this asset.
• Asset status: Expired – Lifetime, anticipated expiration 2014-08-24. "Expired – Lifetime" (as opposed to lapse for non-payment) indicates all maintenance fees were paid through the full 20-year term — the assignee affirmatively maintained this asset for its entire life.

Assignment timeline

One recorded conveyance only.

• Executed: date not retrieved (unverified) / recorded 1994-08-24 — Reel not retrieved (unverified) / Frame not retrieved (unverified)
  • Conveyance: Assignment (recorded as "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: Hagiwara, Tsuneyuki
  • Assignee: Nikon Corporation
  • Correspondent: not retrieved (unverified) — I cannot name the recording attorney/firm or flag recurrence. This is the one field in the requested schema that I could not populate, and I will not guess it.
  • Context: Standard employee-to-employer pre-filing assignment. Recorded on the same date as the application filing (1994-08-24), which is the classic pattern for an inventor assignment filed together with the application papers.

No further recorded events exist for this patent. The Google Patents legal-events timeline for US 5,471,066 contains only: two foreign-priority claims (JP 5-211245, 1993-08-26; JP 5-226866, 1993-09-13), the 1994-08-24 filing, the single 1994-08-24 Nikon assignment, the 1995-11-28 grant/publication, and the 2014-08-24 anticipated expiration. There is no subsequent assignment, security agreement, release, merger, change-of-name, license, or correction in the record.

This absence is itself the finding: for twenty years, from grant to expiration, the patent remained with the original corporate assignee. It was never securitized or pledged, never transferred into an IP-holding subsidiary, and never sold. It also never appears as the subject of a recorded license or a reexamination/PTAB proceeding (per the litigation section already generated).

Verification gap and how to close it

To convert the two unverified fields above into findings, pull the record directly:

• USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search property number 5471066; also try application 08/294,990)
• Legacy index — https://assignment.uspto.gov/patent/index.html

A hit for a single 1994 recording is expected. Note that 1994-dated recordings are well within the Assignment Center's post-August-1980 coverage, so a null result would be surprising and would point to an indexing issue with the property number rather than to an absence of the assignment. (A numeric reel range for 1994 recordings would be an estimate only; I am deliberately not stating one, since it cannot be sourced here and should not be mistaken for a record.)

Timeline diagram

timeline
    title Ownership of US 5471066
    1993 : JP priority filed
         : Second JP priority filed
    1994 : US application filed
         : Assigned to Nikon Corporation
    1995 : Patent issues as US 5471066
    2014 : Patent expires at full term

NPE / troll-pattern signals

Verdict

Insufficient data (no records, or only the original assignment).

The chain consists of a single recorded conveyance — Hagiwara to Nikon Corporation, recorded 1994-08-24 — with no post-issuance ownership events through the patent's 2014-08-24 expiration. That is the literal trigger for this verdict bucket, but the substantive reading is the opposite of a data gap: ownership is effectively certain (Nikon never let go of the asset), and the only genuine insufficiency is evidentiary — the reel/frame, execution date, and correspondent of record were not retrievable from the sources I could reach.

No NPE signal is present in any category. There is no transfer-to-asserter, no shell LLC, no cascading chain, no pre-suit assignment, and no defensive-aggregator termination. The patent is expired and its full maintenance-fee history ("Expired – Lifetime") confirms it was a maintained but never-monetized corporate asset.

Verify the record at the USPTO Assignment Center (property 5471066) — specifically to capture the reel/frame and correspondent for the 1994 recording, which is the only outstanding gap in this reconstruction.

Cross-source note / consistency check: Nothing here contradicts the previously generated summary and litigation sections. The Google Patents per-keyword and legal-event data used above names Nikon as both original and current assignee, which matches those sections; the litigation section's finding of no suits is consistent with a single-link, never-transferred chain. The one item to carry forward is the unresolved second-independent-claim question from the summary section — it is unrelated to ownership, but if the claim set is confirmed later, it does not affect this chain, since all claims share the same single 1994 assignment.

I'll search for the patent and its cited references (prior art). Let me run several targeted queries.

Let me try to retrieve the "Patent Citations" section directly.

Search coverage note (what I could and could not verify)

I ran targeted searches for the prior-art/citation record of US 5,471,066. Two things are important to state up front, because they bound the reliability of everything below:

1. The complete "References Cited" front‑page table for US 5,471,066 was not retrievable in this session. Every query I ran returned the specification body of the '066 patent (SumoBrain and Google Patents both returned the Description text) or returned forward citations ("Cited By") from other patents — but not the examiner's backward-citation table. I will not fabricate entries to fill that table.
2. Everything I state below about § 102 is an analytical prediction, not a reading of full claim charts. I did not retrieve or read the full text of the candidate references in this session. Where I flag confidence, the flag is load-bearing.

Per the operating rules, identifiers are reproduced literally and not auto-corrected. Note in particular that US 5,471,066 is not in the same family as US 5,680,207 (the latter is a Hagiwara/Nikon case whose chain runs through Ser. Nos. 08/638,446 ← 08/527,386 ← 08/405,401 ← 08/117,900 ← 08/095,912 ← 08/045,793 ← 07/990,292), so its reference list cannot be assumed identical to the '066 list.

A. Confirmed forward citations of US 5,471,066 ("Cited By")

These I did confirm from search results — each is a patent that expressly cites US5471066A:

Caveat: forward citations are not prior art against '066. They are useful only (a) as evidence of the field's understanding of the disclosure, and (b) if you were attacking a later patent — which is not the requested task. I list them only because the query returned them and they establish that '066 is treated as the seminal reference in "printed Fourier filtering."

B. Backward citations (prior art) — what I could recover

The searches surfaced a front-page reference list for the related Hagiwara/Nikon case US 5,680,207, not for '066. Given the same inventor, same assignee, and near-identical subject matter (coherent illumination of an inspection point, Fourier/spatial-frequency treatment, segmented photoreceiver), these are the most probable overlap set with the '066 examiner's citations — but I am explicitly flagging this as an inference, not a verified quotation of the '066 front page.

Confirmed to exist (front-page listing on US 5,680,207, Hagiwara/Nikon):

Also confirmed in the same technical lineage and highly relevant as prior art in the field (family/related Nikon filings):

• US 5,629,768 — "Defect inspecting apparatus," which teaches the core '066 idea stated in its own summary: "the spatial frequency spectrum of the circuit pattern on the substrate is intercepted by a light intercepting member… becomes free of spectrum when the circuit pattern… is an errorless circuit pattern… When the circuit pattern… has a defect, a spectral component is created in the spatial frequency spectrum after intercepted." It also adds the inclined-illumination feature: "when the illuminating means applies the illuminating light from an inclined direction… the original Fourier spectrum of the circuit pattern on the substrate is eliminated substantially completely, and the defect information can be detected at a high SN ratio." This is the single most on-point passage I located, because it recites the same "intercept the good-die spectrum, recover only the defect spectrum" mechanism that is the heart of '066 claim 1. (Date/priority not confirmed in-session.)
• US 5,798,831 — Nikon, "Defect inspecting apparatus and defect inspecting method."

C. § 102 analysis against the '066 claims

Because I could not pull the exact '066 reference table, the safest useful output is a claim-element mapping so that whichever references are on the '066 front page can be slotted in. Claim 1's limitations, as previously established, are: (a) illumination to a spot; (b) Fourier transform optical device, axis intersecting the specular-reflection direction; (c) spatial filter at the Fourier plane with light-shielding portions coincident with bright portions of the non-erroneous reference pattern's transform; (d) photoelectric conversion of transmitted light; (e) turntable rotating the substrate about an axis parallel to the Fourier optics axis; (f) filter rotating means rotating the spatial filter about a predetermined shaft synchronously with the turntable; (g) shifting means moving the substrate transverse to the optics axis; (h) spiral scan of the surface. The second independent claim (number unconfirmed; the "minimum-level among a set of photodetectors" variant) instead requires (c′) a generic multi-arm spatial filter and (d′) a set of photodetectors, with detection by the minimum signal.

Bottom line on anticipation: On the material I could actually retrieve, no single reference of record was shown to disclose all of claim 1's limitations simultaneously — critical to any finding against claim 1 are (i) filter rotation synchronized with turntable rotation, and (ii) the illumination-spot spiral scan via rotation + translation. The one reference whose summary squarely recites the "intercept the errorless spectrum" mechanism, US 5,629,768, appears to lack the synchronized-rotating-filter + rotating-turntable combination, which is the '066 novelty crux as summarized in its own specification. That is a § 112/§ 103-style gap, not a § 102 clean hit, on the record available.

D. Sources

• SumoBrain, US5471066 — https://www.sumobrain.com/patents/us/Defect-inspection-apparatus-rotary-type/[5471066](/patent/5471066).html (specification body)
• Google Patents, US5471066A — https://patents.google.com/patent/US5471066 (specification body, "Definitions")
• Google Patents, US20070121111A1 — https://patents.google.com/patent/US20070121111A1/en (confirms US5471066A as a cited reference)
• Google Patents, US20120147364 — https://patents.google.com/patent/US20120147364 (confirms US5471066A as a cited reference)
• Google Patents, US8031931 — https://patents.google.com/patent/US8031931 (confirms US5471066A as a cited reference)
• Issued PDF, US5680207 (Hagiwara/Nikon) — https://patentimages.storage.googleapis.com/94/88/e1/60348c1dffe772/US5680207.pdf (front-page reference list: 4,610,541; 4,740,708; 4,999,510; 5,363,187; 4,556,903)
• USPTO.report, US5629768 "Defect inspecting apparatus" — https://uspto.report/patent/grant/[5629768](/patent/5629768) (spatial-frequency-spectrum interception passage)

E. Caveats and the authoritative next step

1. The exact "References Cited" table on the US 5,471,066 front page was not retrieved. The list in § B carries an explicit inference flag; do not treat it as the verified '066 citation list.
2. No § 102 rejection or allowance record was reviewed (no file-wrapper/PTAB data retrieved), so I cannot report which references the examiner actually applied. Whether any reference did anticipate claim 1 as issued is unknowable from the sources I obtained.
3. US 5,363,187's effective prior-art date must be checked before asserting § 102(e); common inventorship with '066 does not disqualify it under pre-AIA § 102(e), but its own filing/priority date controls.
4. Authoritative verification path: USPTO Patent Public Search / PatentCenter record for application 08/294,990, and the granted-PDF front page at https://patentimages.storage.googleapis.com/79/54/17/8a7d36660a73d3/US5471066.pdf, are the only sources that will settle (a) the precise reference list and (b) the number/text of the second independent claim that remains unconfirmed from the earlier section — and which is required before any § 102 mapping of the "minimum-level detector set" claim can be finalized.

I'll first try to retrieve the actual prior-art references cited on this patent's face, since the "Prior Art section" wasn't included in the text I was given.

Let me attempt to retrieve the patent's cited prior art list (front-page references) from other databases.

I found the cited prior-art list on the issued front page (OCR). Let me verify each cited reference's identity and subject matter.

Obviousness Analysis — US 5,471,066 under 35 U.S.C. § 103

0. Scoping note on "the Prior Art section of this page"

The authoritative page text supplied for this task contains no "References Cited" / prior-art table — the Google Patents fetch stops at the description and classifications. I therefore recovered the patent's face citations from the issued-document OCR and from citing-patent pages. What I recovered (OCR, partially truncated):

Source: issued PDF OCR — https://patentimages.storage.googleapis.com/79/54/17/8a7d36660a73d3/US5471066.pdf

Caveats: (a) The OCR excerpt was truncated; the foreign-patent-document line and any "Other Publications" line were not recovered, so the list below may be incomplete. (b) The two JP priority applications (JPH0763689A / JPH0783840A) are applicant's own priority documents and are not prior art. (c) 5,377,002 (12/1994) and 5,389,794 (2/1995) issued after the 1994-08-24 U.S. filing date; they can only be art under § 102(e)/§ 103 with pre-filing effective dates, and I could not verify those dates. (d) Minor metadata inconsistency to flag: the parent analysis used a fetch date of 2026-09-16 while this task states the current date as 2026-04-26. I treat it as immaterial but note it per the no-auto-correction rule.

I was unable to retrieve the text of 5,264,912, 5,274,434, 5,377,002, or 5,389,794 within the available search budget. I will not assert what they disclose.

1. What must be shown

Claim 1 (confirmed, per Justia) requires, in substance:

1. Illuminating means → spot on patterned substrate;
2. A Fourier-transform optical device whose optical axis intersects the direction of specularly reflected light;
3. A spatial filter at the Fourier plane whose opaque portions coincide with bright portions of the Fourier transform of a non-erroneous reference pattern;
4. Photoelectric converting means for the transmitted flux;
5. A turntable rotating the substrate about a shaft parallel to the Fourier-transform optical axis;
6. Filter rotating means rotating the spatial filter about a predetermined shaft in synchronism with the turntable; and
7. Shifting means translating the substrate in a plane perpendicular to that axis, giving a spiral scan.

Claim 2 adds that the filter's rotation center is the zero-order light position.

An unconfirmed second independent claim (number unknown — the claims listing was truncated in every source I reached) covers the "minimum-level" variant: a set of photoelectric converters at the Fourier plane, judged defective when the minimum channel signal exceeds a threshold, with defect size read off that minimum. I treat it as unverified and analyze it conditionally.

2. The primary prior art and what each teaches

US 4,330,775 (Iwamoto et al., 5/18/1982) — "Apparatus for inspecting defects in a periodic pattern" (Tokyo Shibaura Denki; EP 0 028 774 counterpart, JP priority 143228/79). This is the strongest reference and directly addresses elements 1–4. It expressly states: "A photo-electric converter for picking up the light beam component including the defect information transmitted through the filter is located on an image forming plane…", and that the known technique "employs a spatial filter which blocks the transmission of the Fourier transformed images of the normal periodic pattern … while it permits the non-periodic pattern information to be transmitted therethrough." Critically, '775 explicitly identifies the alignment problem: "(1) The spatial filter and a pattern to be examined must be precisely aligned with each other in orientation." Its own Figs. 1–5 show a diffraction pattern "formed by Fourier-transforming the periodic pattern rotated with the rotation of the mesh plate." Source: https://patentimages.storage.googleapis.com/f7/d9/62/1967f5efd899ce/US4330775.pdf

US 4,360,269 — same title, likely the same Iwamoto/Tokyo Shibaura family. Claim 9 recites "shifting means for shifting the periodic pattern in a direction orthogonal to an optical path of the coherent light beam and rotating means for rotating the periodic pattern in a plane normal to the optical path." Claims 7–8 add first/second photoelectric converting means and "size detecting means … for detecting the defect size … by determining the amplitude of the electrical signals." If this family relationship is confirmed, '269 supplies elements 5, 6 (rotation capability) and 7 of the claim plus defect-size-from-amplitude. Source: https://www.sumobrain.com/patents/us/Apparatus-inspecting-defects-in-periodic/[4360269](/patent/4360269).html

US 4,806,774 (Lin et al.) — cited inside '200 as the foundational Fourier-optics wafer inspection system. Illuminates a patterned wafer, Fourier-transforms the reflected light, spatially filters the normal periodic pattern, inverse-transforms and detects the defect light. 102(b) art regardless of citation.

US 5,172,000 (Scheff et al., Tencor, 12/15/1992). Spatial filter for a Fourier-optics defect inspection system: "defect-carrying light rays 52 not blocked by spatial filter 50 strike an inverse Fourier transform lens section 54." Confirms the wafer-level (not just mask/mesh) application of the '775 filter architecture. Source: https://patentimages.storage.googleapis.com/76/97/f2/98ec8a4888997e/US5172000.pdf

US 5,276,498 (Galbraith et al., Tencor, 1/1994; priority 1992-05-12) — "Adaptive spatial filter for surface inspection." Establishes that making the filter adaptively/generically, rather than by a bespoke photographic exposure per product, was known one year before the '066 priority.

Printed publications (all "printed publications" under § 102(b)), cited within the above references:

• Watkins, "Inspection of Integrated Circuit Photomasks with Intensity Spatial Filters," Proc. IEEE, Vol. 57, No. 9, Sept. 1969, pp. 1634–1639.
• Axelrod, Proc. IEEE, April 1972, pp. 447–448.
• Flambolz & Froot, IBM J. Res. Develop., Nov. 1973, pp. 509–518.
• Iwamoto et al., "Rotation-, Shift-, and Magnification-insensitive Periodic-pattern-defects Optical Detection System," Applied Optics, Vol. 19, No. 7, April 1980, pp. 1196–1200 — quoted in the EP 0 028 774 search report.

The last item is the single most damaging piece: the field had already framed rotational misalignment as the problem to be solved in Fourier-filtered defect inspection, two years before '775 issued and thirteen years before the '066 priority.

Admitted prior art in the '066 specification itself: (i) the XY-stage + imaging + pattern-comparison system of Fig. 25, which "needs to perform a complex image processing at high speed … a high-speed computer is required … the signal processing system becomes large"; and (ii) the galvanometer-scanned laser spot system, where "a mirror is oscillated to scan … the inspection speed is limited owing to the upper limit of a scan rate and … it takes a long time to inspect the entire surface." Both are admissions of the problem set and of the state of the art.

3. Grounds of rejection

Ground A — Claim 1 obvious over Iwamoto '775 (or Lin '774 / Scheff '200) in view of '269

Motivation to combine. All references are in the same field of endeavour (coherent-light, Fourier-plane, defect inspection of patterned substrates — wafers, masks, mesh plates), and '775 and '269 are the same apparatus family. The articulated reasoning is: the '066 patent's own admitted problems are (i) throughput capped by galvanometer scan rate and (ii) signal-processing burden. Rotating the specimen on a turntable and translating it radially removes the oscillating-mirror rate ceiling and is the classic way to scan a disk-shaped specimen at high linear velocity — exactly the throughput gain the artisan sought. '269 already teaches rotating and orthogonally shifting the patterned specimen inside a Fourier-filtering defect detector, so the combination is a rearrangement of parts of the same apparatus, not a new architecture.

The synchronous filter rotation is nonetheless obvious, and this is the pivotal point:

1. The physics compels it. As the specimen rotates, its Fourier spectrum rotates rigidly about the zero-order point; a stationary blocking mask would lose registration. The '066 specification states this as a fact of nature: "As the angle formed by the reference coordinate system … is changed …, the observable spectrum area 33a is rotated around the position Q …"
2. The problem was expressly recognized. Iwamoto '775 states the alignment-in-orientation requirement as a known drawback, and Iwamoto's own 1980 Applied Optics paper is titled "Rotation-, Shift-, and Magnification-insensitive Periodic-pattern-defects Optical Detection System."
3. The art thereby presents a finite, identified set of predictable solutions to the rotation problem: (a) render the filter rotationally insensitive/omnidirectional — the '269 direction, (b) re-register the filter dynamically — i.e., co-rotate it, or (c) use a rotationally symmetric filter. Under KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007), where "there are a finite number of identified, predictable solutions," one of which is the claimed arrangement, § 103 is satisfied; the artisan had "good reason to pursue the known options within his or her technical grasp," and the result is not "beyond the skill of the ordinary innovator."
4. No unexpected result. Co-rotation produces exactly the expected effect — the blocking zones stay registered on the pattern's spectral orders while defect light continues to pass. Nothing in the specification claims a synergy; to the contrary, the '066 spec concedes a problem introduced by rotation (the "energy modulation" of the detected flux as the non-uniform filter turns, requiring calibration against a reference particle and gain modulation). A limitation that forces the inventor to add a calibration/compensation step argues for routine optimization, not non-obviousness.

Claim 2 is even more clearly obvious: the specification itself treats the rotation center as a forced design choice — "even though the substrate rotates, the position of the zero-order component of the Fourier transform pattern of the pattern on the substrate is not changed. Therefore, the shaft of the spatial filter is preferably set to the position of the zero-order light component." Selecting the physically invariant point as a rotation axis is the definition of an obvious design choice.

Ground B — the second (minimum-level) independent claim, conditionally

Assuming its actual text matches the specification's summary (a set of photoelectric converting means, defect declared when the minimum channel exceeds a threshold, size from the minimum):

• Iwamoto '775 / Lin '774 supply the Fourier plane, the specimen-rotation-created rotating spectrum, and the photoelectric detection; '269 supplies rotation + orthogonal shift.
• Galbraith '498 supplies the generic/adaptive mask — removing the need to fabricate a per-product filter, which is exactly the stated advantage of the third/fourth embodiments' four-armed mask (light-shielding zones at −45°, 0°, 45°, 90° per the '066 description of Fig. 21).
• The '066 specification also states the basis for the minimum-level rule as an objective property of the optics: the good-pattern spectrum is "distributed discretely," the defect spectrum "approximately uniformly." The algorithm is therefore read directly off the known physics of the prior art; taking the minimum of a set of channel outputs to reject a localized (discrete) spectral peak against a uniform background is a routine detection/logic choice, and multi-element detector arrays at a Fourier plane (fiber-bundle-relayed to photomultipliers) were a standard way to exploit high-sensitivity detectors.
• The fiber-bundle "Fourier transform device" (spherical input face at radius R giving image height h = R·sin θ, equivalent to the lens's h = f·sin θ) is a known equivalent of the lens, offered as a mere substitution for the purpose of light-collection efficiency and package miniaturization. Substitution of a known equivalent optical element to obtain its known advantage is obvious.

Ground C — secondary/backup combinations

• '775 + '200 + '498: adds wafer-level (as opposed to mesh/mask) applicability plus adaptive filtering.
• '775 or '774 + Vaught '912 / Morioka '434 / Malin '002 / Allen '794: these were cited by the examiner and may supply the rotary-scan and/or segmented-detection elements, potentially even as § 102 anticipation. I could not verify their disclosures within the search budget; I do not assert what they teach.

4. Secondary considerations

I found no evidence of nexus, unexpected results, copying, licensing, or industry praise tied to the asserted claims. The commercial context (a Nikon wafer-inspection tool) is weak: the patent's advance is a mechanical scan-motion change plus a synchronization expedient, and any success in the market is as attributable to Nikon's overall tool performance as to the co-rotating filter. The need being met (faster inspection with less computation) is the same need the admitted prior art was already attacking, and the '066 specification does not show the rotary approach produced a result the art would not have predicted — it shows a result the art predicted, with a calibration burden the art did not mention.

5. Conclusion

Strength of the strongest counterargument: all of the weight rests on the synchronous-rotation limitation. If a court or examiner accepted that no verified reference discloses a rotating spatial filter, the patentee's best response is that co-rotation was not suggested. That response is weak because (a) the problem was recognized a decade earlier in the very same author's literature, (b) the claim adds no new function, only continued registration, and (c) the specification itself presents rotation as the source of a difficulty (energy modulation) rather than a source of surprise.

Caveats and limitations of this analysis, stated explicitly: (1) I did not obtain the full verbatim claim set; only claims 1–2 were confirmed, so the analysis of the second independent and dependent claims is inferred from the specification's summary and should be re-run against the USPTO PatentCenter record for application 08/294,990. (2) Four of the seven face-cited U.S. references (5,264,912, 5,274,434, 5,377,002, 5,389,794) were not retrievable, and the foreign-reference and "Other Publications" lines of the front page were truncated in OCR; a complete rejection would likely be stronger, not weaker. (3) US 4,330,775 and US 4,360,269 bear the same title and appear to be same-family/same-assignee; I inferred, but did not confirm, that relationship — verify before relying on it. (4) The '066 patent expired 2014-08-24 and, consistent with the parent analysis, no litigation or CAFC activity was found; this § 103 analysis is therefore a historical/validity study rather than a live-dispute opinion.