US 9638634

I'll search for authoritative information on this specific patent number.

Now let me check for litigation and CAFC 2026 docket activity.

Let me make one more targeted check for any litigation or PTAB proceedings.

Search scope and confidence

I searched for the specific alphanumeric identifier 9638634 in patent databases (Google Patents, which mirrors USPTO bibliographic data; Justia; FreePatentsOnline; patentimages PDFs) and for Court of Appeals for the Federal Circuit (CAFC) 2026 activity.

Limitation on my results: my tools return public web content, not a direct authenticated query of USPTO PatentCenter/PAIR or PACER/CAFC docket systems. I therefore cannot claim a definitive "no docket" result — only that no CAFC 2026 appeal, PTAB proceeding, or district-court litigation involving US 9,638,634 appeared in any search I ran. Treat that as "none found," not "none exists."

Important identifier warning: the digit string 9638634 also appears as an unrelated WO 9638634 A1 PCT publication (Huesker Synthetic GmbH & Co., published 05-12-1996; family includes EP 0828898, US 6429153). That is a different document. Per your instruction, I did not substitute or merge it with US 9,638,634 — the two must not be conflated.

US 9,638,634 B2 — Bibliographic data

Abstract (verbatim)

"A multidimensional spectrometer provides broadband white light pump and probe pulses to allow acquisition of multidimensional spectral information over a wide range commensurate with solar-related chemical processes."

Plain-language overview of the independent claims

Claim 1 (apparatus): A multidimensional (2D/3D+) spectrometer built from five elements:

1. a sample volume holding a sample;
2. a light detector near the sample;
3. at least one white-light source producing pulses with a substantially continuous bandwidth of at least 500 nm;
4. a first optical system that sends those pulses through the sample as probe pulses to the detector;
5. a second optical system that takes the same source pulses and splits them into at least first and second pump pulses with a controllable time separation, and directs them through the sample;
   plus an electronic computer system that (a) reads out the detector and (b) drives the second optical system to vary that time separation over many laser shots so that a multidimensional spectrograph can be constructed.

Core inventive concept: use broadband "white light" for both pump and probe, so no center-frequency tuning (e.g., no NOPA) is needed across a very wide spectral range — targeting solar/photosynthesis-type measurements spanning the visible and near-IR.

Claim 14 (method): Takes the same spectrometer and claims the measurement sequence:

• (a) excite the sample with the first and second pump pulses at a given time separation;
• (b) afterward, probe with the probe pulse and measure the detector signal;
• (c) repeat (a)–(b) across a range of time separations;
• (d) process the collected measurements in the computer to produce a two-dimensional spectrum.

Dependent claims add: substantially constant center frequency (2); laser + spectrally-broadening crystal generating white light without modulation (3); bandwidth ≥700 nm (4, 15), ≥900 nm (5), ≥1000 nm (16); spectral range including 1000 nm (6, 17) and 400–1400 nm (7, 18); bandwidth ≥1½ octaves (8, 19); probe and pump spectra substantially identical to the source (9, 20); white-light source alternatives (beam-splitter-then-two-broadening-elements, or single-broadening-element-then-beam-splitter) (10); YAG crystal broadening element (11); a birefringent-crystal/polarization-selective-wedge/polarizer pulse-splitting arrangement computer-controlled (12); and a controllable pulse delay element (13).

Provenance and drafting points worth noting

• Sources used: Google Patents (https://patents.google.com/patent/US9638634), Justia (https://patents.justia.com/patent/9638634), FreePatentsOnline (https://www.freepatentsonline.com/y2016/0018323.html), and the WARF technology page (WARF ID P140311US02, https://www.warf.org/technologies/summary/P140311US02/). Zanni's university CV lists US 9,638,634 as "(Licensed)."
• Claim 1 wording changed between publication and grant: the pre-grant version (Justia/FPO for US 2016/0018323) refers to the second optical system breaking light into "first and second probe pulses," whereas granted claim 1 reads "first and second pump pulses." The granted text is the operative one.
• Internal inconsistency in the specification (persisting in the granted patent): the "Detailed Description" labels the pump branch as "first optical system 16" and the probe branch as "second optical system 18," while the Summary and claim 1 use the opposite assignment (first = probe, second = pump). FIG. 1's narrative ("first optical system 16 and second optical system 18 to develop probe and pump pulses respectively") conflicts with the later paragraphs. This is a description/claim-mapping mismatch that could matter in any §112 or claim-construction dispute.
• Typos retained verbatim in the granted text (not auto-corrected): claim 9 recites "the probe and PUMP pulses"; claim 12 recites "the second optical path pros ides a bifringent crystal"; the spec says "bitringent" and "position-selective wedge" where the summary says "polarization-selective wedge."
• Funding-statement discrepancy between sources: the granted patent recites FA9550-12-1-0063 (USAF/AFOSR) and NSF award 1121288; the Justia rendering of the application instead says FA9550-12-1-0063 was "awarded by the National Science Foundation" — the award number and the agency attribution do not match between renderings.
• Cited prior art of note: US 4,512,660 (picosecond broadband CARS with continuum probe), US 5,479,256 (transient grating spectroscopy), US 2009/0161092 (Zanni, "Multidimensional spectrometer"), US 2010/0110426 (NIST), and non-patent citations Krebs et al., New J. Phys. 15 (2013); Mikhailovsky, Basics of femtosecond laser spectroscopy; and Bourquin et al., Optics Letters 28(17):1588–1590 (2003).
• Related family: WARF also holds US 9,052,239 ("Multidimensional spectrometer," also Zanni) and US 9,568,366 / US 7,760,342, which appear in the patent's similar-documents list.

Explicit uncertainty

I have no authoritative information establishing any 2026 CAFC docket, appeal, IPR/PGR, or infringement suit involving US 9,638,634. Any statement that litigation exists (or does not) should be verified directly in USPTO PatentCenter, the PTAB API, and PACER/CAFC CM-ECF before being relied upon.

I'll search for litigation involving US patent 9638634 across patent litigation databases.

The initial results reference an unrelated WO9638634. Let me search more specifically for the US patent and its owner.

No litigation has appeared yet. Let me check litigation-specific databases (Unified Patents, RPX, CourtListener) and any PTAB challenges.

Summary: No Known Litigation for U.S. Patent No. 9,638,634

Based on my searches, I found no record of any litigation — district court, ITC, CAFC, or PTAB — involving U.S. Patent No. 9,638,634. To be explicit and avoid the sort of false-positive that "similar numbers" can create: U.S. 9,638,634 is a completely different patent from the many unrelated references to "WO 9638634" (a 1996 PCT publication to Huesker Synthetic GmbH) that surface in a naive keyword search. Those are not the same identifier and I have excluded them.

The patent, as identified (for the record)

Source: https://patents.google.com/patent/US9638634/en

What I searched and what I found

1. Google Patents family/litigation data for US 9,638,634 — the page lists no litigation events, no "Litigation" section entries, and only three forward-citing documents (all unrelated: CN105515726B, CN910995B, DE112021006351T5). https://patents.google.com/patent/US9638634/en
2. WARF's own technology page for the invention (WARF ref. P140311US02) — commercialization/technology-transfer listing only; no enforcement activity noted. https://www.warf.org/technologies/summary/P140311US02/
3. Inventor's CV (Martin T. Zanni, UW–Madison, updated Aug. 2024) — lists patent 9,638,634 simply as "(Licensed)," i.e., a licensing/monetization status, not an assertion/enforcement status. https://zanni.chem.wisc.edu/.../zanni_cv_8_13_2024_no_funding.pdf
4. Keyword searches across court-opinion and docket text (e.g., CourtListener/recap, govinfo USCOURTS opinions, District of Delaware, EDTX docket text) pairing "9638634" / "9,638,634" with litigation terms ("complaint," "infringement," "IPR," WARF/Zanni) — no hits matching this patent.

Important caveats

• Scope of my search is limited. I was not able to fully query the dedicated litigation trackers directly in this session; my Unified Patents, RPX, and PACER-oriented queries were cut off before returning results, and PACER/Unified/RPX dockets are the definitive sources. My conclusion is based on Google Patents, WARF's materials, the inventor's CV, and general web/docket-text searching.
• Verification is not exhaustive. A patent can be asserted in a complaint without the number appearing in indexed full text, and recently filed cases may not yet be indexed. Absence of evidence here is not a guarantee of non‑litigation.
• What the evidence affirmatively suggests: This appears to be a university research-instrument patent (a multidimensional optical spectrometer) that WARF has licensed (consistent with the inventor's CV notation and the existence of a spin-off, PhaseTech Spectroscopy, Inc., in which the inventor holds an interest, noted in a 2019 J. Phys. Chem. A article). This is the profile of a technology-transferred patent rather than a litigated one.

Bottom line

For U.S. Patent No. 9,638,634, I have no known case to report — no plaintiff, defendant, jurisdiction, case number, filing date, or outcome, because I found no litigation. Before relying on this for a freedom-to-operate, enforcement, or diligence opinion, I would recommend a direct check of PACER (all district courts), Docket Navigator, Unified Patents' litigation portal (https://portal.unifiedpatents.com/litigation/caselist), and USPTO PTAB/Assignment records, since those are the authoritative sources I could not fully exhaust here.

I'll start by checking the structured data and then search for any PTAB activity the ODP index may have missed.

The first searches returned mostly noise from unrelated "634" patents. Let me target the patent by title, inventor, and assignee.

Still no PTAB hits. Let me run targeted docket-style searches and check for any litigation that might have spawned challenges.

PTAB Proceedings — U.S. Patent No. 9,638,634

Search note (read first): The structured "PTAB proceedings on file" block returns an empty set for U.S. 9,638,634. I ran web searches against PTAB E2E/PTABCTS, Docket Alarm, CourtListener, and general web indexes for the patent number, the title, the inventor (Martin T. Zanni / Martin Thomas Zanni), and the assignee (Wisconsin Alumni Research Foundation). I found no AIA trial proceeding of any kind. What the searches did return was a large volume of false positives from unrelated patents sharing the "634" fragment — flagged below so you don't get misled.

Proceedings overview

Total AIA trial proceedings on U.S. 9,638,634: zero (0). Breakdown by status: 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denied. The patent has never been the subject of an IPR, PGR, or CBM petition, so there is no claim-level post-grant outcome to report and no claim of the '634 patent has been canceled or narrowed by the PTAB. The defensive posture this gives a defendant is the opposite of the usual "PTAB-hardened" story: the patent is untested post-grant, which means (a) the full claim set — claims 1–20 — is live and available to the patent owner, and (b) an IPR is still an entirely open strategic option, not foreclosed by prior petitioner estoppel or by a prior adverse FWD.

No proceedings to enumerate

There is no IPR20xx-xxxxx, PGR20xx-xxxxx, or CBM20xx-xxxxx number to populate the per-proceeding template with. Any proceeding number I wrote here would be fabricated. Confirming the negative is the deliverable.

Confirmed posture of the patent (grounding the negative)

False positives to disregard (each is a different patent that happens to end in "634," confirmed from the search results themselves):

• U.S. 8,938,634 — HashiCorp, Inc. v. Invincible IP LLC, IPR2022-00565 / IPR2022-00566 (data-center power management). Unrelated family.
• U.S. 7,543,634 — challenged by Weatherford entities in IPR2016-00597 / IPR2016-01505 / IPR2016-01514 (oilfield "Rapid Completions" family). Unrelated.
• U.S. 7,718,634 — Hoffmann-La Roche v. Apotex (ibandronate dosing), Fed. Cir. 2014. Unrelated.
• WO 9638634 A1 — a 1996 Huesker Synthetic GmbH geotextile PCT publication. The literal string "9638634" appears repeatedly in patent-search-result snippets because of this document; it is not this patent.

Strategic summary

1. Claim status — all claims UNTESTED. Claims 1–20 of U.S. 9,638,634 retain their original, as-granted scope. Nothing has been canceled, disclaimed by certificate, or narrowed through reissue or reexamination as far as the available record shows. For a defendant receiving a demand letter or complaint asserting this patent, the asserted claims are whatever the patent owner chose to assert — the patent owner has not been forced by any PTAB proceeding to retreat to dependent claims, and no claim has been held unpatentable. Independent claim 1 (broadband "white light" source with a substantially continuous bandwidth of at least 500 nm, a probe optical system, a pump optical system producing first and second pump pulses with controllable time separation, and a computer that sweeps that separation to build a multidimensional spectrograph) is the centerpiece claim and it is intact. Note also the very broad bandwidth laddering in the dependent claims — 700 nm (claim 4 / claim 15), 900 nm (claim 5), 1,000 nm (claim 16), 400–1400 nm including 1000 nm (claims 7 and 18), and "no less than one and one half octaves" (claims 8 and 19). Those numerical ranges are the natural, and largely the only, narrowing hooks.

2. Estoppel landscape — a blank slate, which favors the defense. Because no petitioner has ever filed, § 315(e)(2) estoppel attaches to nobody. There is no prior-art ground that is off the table, and no petitioner/privy relationship to worry about importing estoppel into a district-court case. Contrariwise, there is also no benefit to inherit: there is no earlier petitioner's institution decision, no PTAB claim construction, no expert declaration, and no FWD reasoning that a current defendant could adopt or distinguish. Every ground — § 102 anticipation, § 103 obviousness, and § 112 written description / indefiniteness — remains fully available to a defendant who is willing to bear the cost of building the record from scratch. The single hard constraint is § 315(b): if suit has been served, the one-year IPR-filing clock is running, and it does not restart.

3. Pattern signals. There is no pattern, because there is no activity: no repeat petitioner, no serial petitioning by competing manufacturers, no WARF appeal to the Federal Circuit from a PTAB decision (there is no decision to appeal), and no defensive aggregator such as Unified Patents in the chain. Coupled with the fact that WARF has kept the patent alive through two maintenance-fee cycles and lists it publicly on its tech-transfer page (WARF reference P140311US02, https://www.warf.org/technologies/summary/P140311US02/), the picture is of a university-held, licensing-stage patent that has not yet drawn a validity fight — not of a troll-ejected, IPR-tested asset.

4. One timing consideration that cuts against filing now. U.S. 9,638,634 issued 2017-05-02 and has therefore been in force for over nine years. As of the March 2025 guidance from USPTO Acting Director Stewart (and the follow-on discretionary-denial decisions applying it — e.g., Dabico Airport Solutions Inc. v. AXA Power Aps, IPR2025-00408, Paper 21), the Board's institution practice expressly weighs the patent owner's "settled expectations," including the length of time the claims have been in force, with six years treated as a de facto threshold for "strong" settled expectations. A petition filed today would have to overcome that discretionary hurdle, in addition to Fintiv considerations if there is a parallel district-court case on a near-term trial schedule. That does not make an IPR unavailable — it makes the petition's discretionary section as important as its merits section.

Recommended next steps

• If you are a defendant: state plainly in your invalidity contentions and any board-facing submission that U.S. 9,638,634 has no PTAB proceedings on file — no FWD, no certificate cancelling claims, and no § 315(e)(2) estoppel attaching to any party. Because no FWD exists, there is no opinion to link and no disposition to quote; do not cite a PTAB decision for this patent in a brief. The supportable citation is the patent's own Google Patents record, https://patents.google.com/patent/US9638634/en, and the PTAB's public docket at https://ptab.uspto.gov/ (E2E) / https://ptacts.uspto.gov/ptacts/ for the empty result set.
• Prior art development is unconstrained. With no estoppel and no prior PTAB claim-construction record, your § 102/§ 103 attack is limited only by the art itself. Given the patent's claim 1 breadth and the claim 3 / claim 10–11 narrowing to laser-plus-spectrally-broadening-crystal generation (YAG), the highest-yield targets are likely (i) supercontinuum/white-light generation literature and (ii) the birefringent delay-line pump-pulse-pair art the patent itself cites — D. Brida, C. Manzoni, G. Cerullo, "Phase-locked pulses for two-dimensional spectroscopy by a birefringent delay line," Optics Letters 37, 3027 (2012-08-01) — plus the applicant's own admitted prior-art applications (US 2006/0063188, US 2009/0161092, US 2012/0236305), all incorporated by reference on the face of the patent. The two non-patent citations of record are also worth mining: Krebs et al., "Two-dimensional Fourier transform spectroscopy in the ultraviolet with sub-20 fs pump pulses and 250–720 nm supercontinuum probe," New Journal of Physics 15 (2013), and S. Bourquin et al., Optics Letters 28(17), 1588–1590 (2003).
• Mind the one-year clock. If a complaint asserting the '634 patent has been served on you, calendar the § 315(b) bar date immediately; it is the binding deadline, not the discretionary-denial environment.
• Do not rely on the "634" search hits. Before any internal stakeholder is told "this patent was already invalidated at the PTAB," confirm the number: the IPR2022-00565/00566 HashiCorp decision and the Weatherford IPRs concern U.S. 8,938,634 and U.S. 7,543,634, neither of which is this patent.
• Confidence statement. I am confident in the negative as of the ingested ODP data plus my searches. I cannot rule out a petition filed in the last few weeks that has not yet been indexed by ODP, PTAB E2E, or third-party dockets. Verify with a direct E2E proceeding-number and patent-number query before filing anything that depends on the absence of prior PTAB art or admissions.

I'll verify the assignment record against live sources before constructing the timeline.

Assignment & Ownership Analysis — US 9,638,634 B2

Scope note on sources. The USPTO Assignment Center (https://assignmentcenter.uspto.gov/) is a JavaScript application that I could not query directly in this session; the assignment list below is reconstructed from the Google Patents "Legal Events" tab, which mirrors the recorded USPTO assignment data, cross-checked against the patent text and public records. I retrieved one assignment entry and no reel images, so the correspondent of record (the attorney/agent who filed the recording) is not verifiable from what I could access — I have flagged that rather than guessing.

Inventors

• Single-inventor patent. No co-inventor departures to analyze — the "all inventors left within 12 months" fire-sale tell is inapplicable.
• Federally funded, Bayh-Dole invention. The specification states support from USAF/AFOSR grant FA9550-12-1-0063 and NSF grant 1121288. This is the standard legal predicate for the inventor's obligation to assign to the university's IP arm, and explains the assignment record below.
• Inventor remains affiliated with UW–Madison and is active in the field (elected to the National Academy of Sciences; 2024 ACS Fellow). No inventor-attrition pattern.
• Context beyond the chain (not an assignment fact): Zanni co-founded PhaseTech Spectroscopy, Inc. (Madison, WI), which commercialized some of his spectrometer technology. I found no recorded assignment of US 9,638,634 to PhaseTech; any relationship would run through a WARF license, which is not an ownership transfer.

Original assignee

Wisconsin Alumni Research Foundation (WARF) — 614 Walnut Street, Madison, WI 53707-7365.

• Business: WARF is the nonprofit technology-transfer organization for UW–Madison (formed 1925). It does not ship products embodying the claims; its model is patent prosecution plus licensing to commercial partners (historically including the Steenbock vitamin-D patents, warfarin, MRI, and human embryonic stem cells). It is best characterized as a non-practicing licensor, not an operating company.
• Status: Operating and financially healthy. No bankruptcy, dissolution, or acquisition. It is the "current assignee" of record on the face of the patent.
• Classification nuance: WARF appears on at least one published NPE taxonomy (Ropes & Gray's IP Magazine piece and the ITC's "Category 1" NPE definition — "universities, research institutions/entities, start-ups, individual inventors, manufacturers whose products do not practice the patents") because it does not practice its own patents. It is not on any of the enumerated PAE lists (Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, etc.).
• Assertion posture (general, not this patent): WARF has asserted other patents (e.g., the '752 patent against Apple, where its complaint alleged Apple refused to consider licensing proposals). I found no litigation naming US 9,638,634.

Assignment timeline

One recorded assignment exists. There are no post-issuance transfers, no security interests, no licenses recorded as conveyances, and no release/correction filings.

• 2016-10-13 (executed) / recorded 2016-12-14 — Reel 040939 / Frame 0894
  • Conveyance: Assignment — "ASSIGNMENT OF ASSIGNORS INTEREST"
  • Assignor: Zanni, Martin (Martin Thomas Zanni, individual inventor)
  • Assignee: Wisconsin Alumni Research Foundation (WISCONSIN ALUMNI RESEARCH FOUNDATION, Wisconsin)
  • Correspondent: ⚠️ Not determinable from the sources I could retrieve. The Google Patents legal-event record exposes owner name, free-format text, reel/frame, and effective date only; the Assignment Center cover sheet (which carries the correspondent attorney/agent and address) was not retrievable in this session. I will not name a correspondent without the reel image.
  • Context: Original inventor-to-institution assignment — the routine Bayh-Dole title transfer from a UW–Madison faculty inventor to the university's IP foundation. Executed ~15 months after the 2015-07-21 non-provisional filing (priority to provisional 62/026,949, filed 2014-07-21) and ~5 months before the 2017-05-02 grant. University dockets frequently record inventor assignments after filing; this is within normal practice and is not a red flag.

Cross-reference (do not confuse with this patent's chain): a different Zanni patent, publication US 2009/0161092 (issued as US 7,760,342 / US 9,052,239 family), carries Reel 020846/0027 with effective date 2008-01-11, same assignor/assignee pattern (Zanni → WARF). That is a separate record and is cited here only to show the recurring inventor→WARF pattern, not as a link in the US 9,638,634 chain.

Post-grant legal events (non-ownership): maintenance fees paid 2020-10-29 (4th yr) and 2024-10-16 (8th yr), both as a small entity — consistent with a nonprofit university foundation, not an NPE volume filer. Adjusted expiration listed as 2035-11-10.

Timeline diagram

timeline
    title Ownership of US 9638634
    2014 : Provisional application filed
    2015 : Non-provisional filed by Zanni
    2016 : Assigned to WARF reel 040939 0894
    2017 : Patent issued May 2
    2020 : 4th year maintenance fee paid
    2024 : 8th year maintenance fee paid

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. The only recorded assignee is WARF, a 1925-chartered Wisconsin nonprofit with a public Madison, WI address and a $3B+ endowment-backed operation. No "IP / Holdings / Ventures" LLC appears in the chain, and no registered-agent service address is present. The chain is one link long.

2. Known asserter in the chain — NOT PRESENT for any enumerated PAE list; UNCLEAR only under the broadest taxonomy. No assignee matches Acacia, Marathon Patent Group, Intellectual Ventures, IPNav, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or an Erich Spangenberg entity. WARF does appear as a Category 1 NPE under the ITC classification cited in Ropes & Gray's NPE survey, but that category expressly lumps in universities and research institutions and is not evidence of troll behavior. No Unified Patents or RPX high-frequency-plaintiff listing for WARF on this patent.

3. Repeat correspondent across the chain — UNCLEAR / NOT ASSESSABLE. With a single recorded link, recurrence is structurally impossible, and the correspondent of record (Reel 040939/0894 cover sheet) was not exposed in the retrievable record. This is a data gap, not a clean bill of health — the one-hop chain would only become interesting if the reel image showed a correspondent shared with WARF's monetization counsel on other dockets.

4. Cascading transfers — NOT PRESENT. One transfer, in one direction, to the original institutional owner. No chained LLCs, no transfers <24 months apart, no shared correspondent addresses or principals.

5. Pre-litigation transfer — NOT PRESENT. No infringement suit naming US 9,638,634 was found (Darts-IP / PACER-style litigation searches returned only unrelated matters, e.g., Warsaw Orthopedic v. NuVasive, which merely shares a patent-number-adjacent hit). The 2016-10-13 assignment predates grant and is not litigation-driven.

6. Bankruptcy fire-sale — NOT PRESENT. No Chapter 7/11 proceeding involving WARF or Zanni; WARF is solvent and has made all maintenance-fee payments.

7. Privateering — NOT PRESENT (no evidence). WARF licenses to industry broadly, including to a company co-founded by the inventor (PhaseTech Spectroscopy). Nothing indicates WARF transferred title to an NPE to assert against a competitor on WARF's behalf; title has never left WARF.

8. Defensive aggregator — NOT PRESENT. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. The patent has not been neutralized by defensive aggregation.

Verdict

Insufficient data — the recorded chain consists of only the original assignment (Zanni → WARF, executed 2016-10-13, recorded 2016-12-14, Reel 040939/Frame 0894), with no post-issuance transfers, no security interests, and no assertion activity on this patent.

Two sentences of justification: The single recorded conveyance is the routine Bayh-Dole inventor-to-university assignment (federally funded via USAF/AFOSR FA9550-12-1-0063 and NSF 1121288), and WARF — a solvent nonprofit licensor that does not practice its own patents — remains the sole owner of record through the 2017-05-02 grant. WARF's appearance on a Category 1 (university/research) NPE taxonomy is not, standing alone, a troll pattern, and there is no shell LLC, no repeat correspondent evidence, no cascading transfer, and no litigation on US 9,638,634 to support a higher-confidence NPE call.

Caveat on the one real gap: the Assignment Center cover sheet for Reel 040939/0894 was not retrievable here, so the correspondent of record remains unverified. If you have authenticated Assignment Center access, pull that reel image and check the correspondent attorney against WARF's other dockets — that is the only remaining item that could change this verdict.

Verify at: https://assignmentcenter.uspto.gov/ (search "9638634") and https://assignment.uspto.gov/patent/index.html (same underlying records).

I'll research each cited prior-art reference for US 9,638,634. Let me run targeted searches on the cited patent documents.

Prior-Art Analysis — US 9,638,634 B2 ("Multidimensional white light spectrometer")

0. Scope, method, and an important identifier check

Confirmed target: US 9,638,634 B2, Appl. 14/804,965, Martin Thomas Zanni, Wisconsin Alumni Research Foundation; priority 2014‑07‑21; granted 2017‑05‑02.

Identifier warning (carried forward): the digit string 9638634 also matches an unrelated WO 96/38634 A1 (Huesker Synthetic GmbH, 1996). That is a different document and is excluded from everything below.

Sources for this section: the authoritative full text you supplied (Google Patents rendering of US9638634, https://patents.google.com/patent/US9638634/en) plus live lookups of each cited reference. Live lookups were successful for US 4,512,660, US 5,479,256, US 2008/0125977, US 2009/0122319, and US 2009/0161092. My lookup of US 2010/0110426 was cut off before returning the document text — that one entry is flagged below as provisional and must be verified.

Legal threshold applied. Anticipation under 35 U.S.C. § 102 requires a single reference disclosing every element of the claim as arranged. US 9,638,634 has a post‑AIA priority date (2014‑07‑21), so AIA § 102 governs: § 102(a)(1) (public disclosures) and § 102(a)(2) (U.S. patents/applications "effectively filed" earlier that name another inventor). A reference that misses even one claim element is at most § 103 art.

Honest headline result: none of the six patent citations anticipates independent claim 1 or independent claim 14. The distinguishing element of both independent claims — the white‑light source providing pulses with a substantially continuous bandwidth of at least 500 nm, supplying both the pump and probe pulses, with a computer‑controlled inter‑pump delay scanned to build a multidimensional spectrograph — is absent from every cited reference. The cited art is best characterized as § 103 combination material, with two genuinely material references (US 2009/0161092 and US 2010/0110426).

1. Summary table

2. Reference‑by‑reference analysis

2.1 US 4,512,660 A — Picosecond Broadband CARS Probe Using the Picosecond Continuum

• Full citation: U.S. Patent No. 4,512,660, Goldberg, L.S., "Picosecond broadband cars probe using the picosecond continuum," assignee The United States of America as represented by the Secretary of the Navy. Filed 1983‑04‑14; issued 1985‑04‑23. (https://patents.google.com/patent/US4512660)
• Description: Single‑shot broadband CARS. A mode‑locked Nd:phosphate‑glass laser (1054 nm, ~5 ps, 25 mJ) feeds a KDP second‑harmonic generator and a spectral beam splitter; the fundamental generates a picosecond white‑light continuum in a D₂O liquid cell; the 527 nm second harmonic forms the monochromatic probe pulse ω₁. Probe + continuum are spatially/temporally combined and focused into the sample; the anti‑Stokes emission is spectrally filtered and detected. Claim 12 adds a variable time delay on an excitation pulse.
• § 102 mapping: No anticipation. The reference is a one‑dimensional CARS experiment, not a multidimensional spectrometer; it does not derive pump and probe from the same broadband white‑light pulse, and its continuum is generated in a liquid cell, not a spectrally‑broadening crystal. It therefore does not disclose claim 1's "multidimensional spectrograph" or "controllable time separation between first and second pump pulses," nor claims 14(a)–(d). Even the dependent claims that superficially rhyme fail on details: claim 3 requires "interaction between a laser and a spectrally‑broadening crystal" (D₂O cell ≠ crystal), and claims 4–8 require ≥700/900 nm / 1½‑octave bandwidths, which the reference does not attribute to its continuum. Best use: § 103 art teaching that broadband continua can serve as the probe arm of a nonlinear optical experiment.

2.2 US 5,479,256 A — Transient Grating Spectroscopy

• Full citation: U.S. Patent No. 5,479,256, Ito, T. et al., "Transient grating spectroscopy," assignees Research Development Corp. of Japan and Takashi Ito. JP priority 1992‑12‑04; U.S. filed 1993‑11‑09; issued 1995‑12‑26. (https://patents.google.com/patent/US5479256)
• Description: Transient‑grating (four‑wave‑mixing) spectroscopy for non‑fluorescent chemical intermediates. A femtosecond dye laser (150 fs, 400 µJ at 720 nm) is split into multiple (typically two) excitation beams applied at an angle to write interference fringes in the sample; a white‑light probe — a femtosecond white light produced by focusing the leftover fs pulses into water — is diffracted and the diffracted intensity is tracked versus time to derive a transient spectrum. He‑Ne/Xe‑lamp probes cover the ns–µs window.
• § 102 mapping: No anticipation. The excitation beams are applied simultaneously in the same phase to create a static grating; there is no "controllable time separation between first and second pump pulses" scanned to generate a multidimensional spectrograph, and the white light is used only on the probe side. It therefore does not disclose claim 1's central elements, nor the step sequence of claim 14. Its continuance and adaptation to produce a white‑light probe is meaningful § 103 art against the "probe pulse ... received by the light detector" element and against claim 9/20 only if combined with a broadband pump source. No claim is anticipated.

2.3 US 2008/0125977 A1 — Use of Quantum System Identification and Quantum Control Techniques...

• Full citation: U.S. Patent Application Pub. No. 2008/0125977 A1, Hunter, I.W., Lafontaine, S., Tseng, C.‑H., Somaroo, S., Anquetil, P.A.; assignee Aretais, Inc. Filed 2007‑10‑29 (provisional 60/855,072, filed 2006‑10‑27); published 2008‑05‑29. (https://patents.google.com/patent/US20080125977)
• Description: Quantum‑control/adaptive‑feedback approach to coherent Raman (FAST‑CARS‑style) medical diagnostics. A Ti:sapphire regenerative amplifier pumps two optical parametric amplifiers (OPAs); pump, Stokes, and probe beams pass through delay stages DS1/DS2 and a pulse shaper, are focused into a sample, and the scattered light is spectrally dispersed onto a CCD, with a computer closing a feedback loop. Tuning is explicit ("λ₁ = 712–742 nm, tunable").
• § 102 mapping: No anticipation. This reference teaches the very approach US 9,638,634 positions itself against — tunable pump generation over a range, which the '634 specification says is difficult and is avoided by using white light. It does not disclose a ≥500 nm continuous‑bandwidth source, nor pump and probe of identical broadband spectrum. § 103/background art only; no claim anticipated.

2.4 US 2009/0122319 A1 — Non‑Uniform Sampling to Extend Dynamic Range of Interferometric Sensors

• Full citation: U.S. Pub. No. 2009/0122319 A1, Rønnekleiv, E. and Waagaard, O.H., assignee Optoplan AS. Filed 2007‑11‑13; published 2009‑05‑14 (granted as US 7,916,303 B2). (https://patents.google.com/patent/US20090122319A1)
• Description: Fiber‑optic interferometric sensor interrogation using deliberately non‑uniform sampling to demodulate fringe rates above the Nyquist limit.
• § 102 mapping: None. Entirely different field (fiber strain/seismic sensing). Nothing about spectrometers, white light, pump/probe pulses, or multidimensional spectra. This is the one citation that carries no asterisk in Google's citation table — i.e., it is not marked "cited by examiner," so it was likely an applicant/third‑party IDS citation of marginal relevance rather than a reference relied upon. No claim anticipated; not even meaningful § 103 art on the merits.

2.5 US 2009/0161092 A1 (= US 7,760,342 B2) — Multidimensional Spectrometer — most material reference

• Full citation: U.S. Pub. No. 2009/0161092 A1, Zanni, M.T. and Damrauer, N.H., "Multidimensional spectrometer"; filed 2007‑12‑21; published 2009‑06‑25; granted as U.S. Patent No. 7,760,342 B2 on 2010‑07‑20. (https://patents.google.com/patent/US20090161092 ; https://patents.google.com/patent/US7760342)
• Description: A multidimensional IR/VIS/UV spectrometer comprising a laser source, a pulse shaper (AOM/SLM/DMD) that converts a single pump pulse into a train of sub‑pulses, a probe pulse, a mostly‑collinear pump–probe geometry, a detector (InSb, HgCdTe, photodiodes, CCD, PMT), and a processor that produces a multidimensional spectrum as a function of the pump–probe delay (and frequency/phase/polarization). It explicitly exploits automatic phase‑locking of sub‑pulses from the shaper.
• § 102 mapping: This is the legally closest reference, and it has a special relationship to the patent:
  • Same technical space, and it is the applicant's own earlier work — it is one of the three documents expressly incorporated by reference in the '634 specification ("...US patent application 2009/0161092 filed Dec. 21, 2007...").
  • Prior‑art status: it names a different inventive entity (Zanni and Damrauer) than the '634 patent (Zanni alone), and it published in 2009 and was effectively filed 2007‑12‑21 — so it is available as AIA § 102(a)(2) art as of 2007‑12‑21.
  • But it does not anticipate. Its source is a narrowband laser pulse that is split and pulse‑shaped; there is no disclosure of a "white light source providing source light pulses having a substantially continuous bandwidth of at least 500 nanometers." Accordingly it does not anticipate claim 1 (missing the broadband white‑light element), claim 14 (same), or the broadband‑dependent dependents 2–9 and 15–20.
  • Where it does bite: it supplies nearly every other element of claim 1 (sample volume, detector, first optical path to probe the sample, a second optical subsystem creating multiple pump pulses with a controllable delay, and a computer controlling the delay over multiple shots to build a multidimensional spectrograph). Combined with a reference showing broadband white‑light generation/use (US 4,512,660, US 5,479,256, US 2010/0110426, or Krebs 2013), it supports a strong § 103 obviousness theory against claims 1 and 14. On its own it cannot anticipate any claim.

2.6 US 2010/0110426 A1 — Method for NRB‑free nonlinear vibrational spectroscopy and microscopy (NIST) — flagged provisional

• Full citation: U.S. Pub. No. 2010/0110426 A1, "Method for nrb‑free nonlinear vibrational spectroscopy and miscroscopy [sic]," assignee The National Institute of Standards and Technology; filed 2008‑11‑03; published 2010‑05‑06. (https://patents.google.com/patent/US20100110426)
• Description (as cited): a method for obtaining nonlinear vibrational (Raman/CARS‑type) spectra free of non‑resonant background. Such methods typically rely on a broadband excitation (e.g., a supercontinuum‑like source) together with phase/amplitude manipulation and heterodyne or phase‑modulated detection to isolate the resonant Raman response from the non‑resonant background.
• ⚠️ Confidence caveat: my retrieval of this reference's full text was cut off in this session. The mapping below is therefore based on the citation title and my general knowledge of the NIST NRB‑free CARS work, and must be verified against the document before being relied on.
• § 102 mapping (provisional): a possible relevance to the broadband‑white‑light source element and to bandwidth‑dependent claims 3–8 and 15–19 (e.g., if it discloses a continuum spanning ≥700/≥900/≥1000 nm or covering ~400–1400 nm). It does not on its face appear to disclose a two‑pulse pump with a computer‑controlled inter‑pump separation used to build a multidimensional spectrograph, so I assess it as § 103 art rather than § 102 anticipation of claim 1/14 — but that conclusion is provisional pending verification.

3. Non‑patent literature cited

(a) Krebs, N. et al., "Two‑dimensional Fourier transform spectroscopy in the ultraviolet with sub‑20 fs pump pulses and 250–720 nm supercontinuum probe," New Journal of Physics 15 (2013), pp. 1–17.

• § 102: No anticipation of claims 1/14. It discloses 2D FT spectroscopy using a broadband supercontinuum probe (≈250–720 nm) — note this is ≈470 nm of bandwidth, below the claimed ≥500 nm — but its pump is a narrowband sub‑20 fs pulse, not white light. It is nevertheless the most cutting‑edge § 103 reference against the broadband‑probe concept and against claims 4–8 (bandwidth) if a ≥500 nm continuum is shown elsewhere. It directly evidences the state of the art the '634 patent sought to extend to both arms.

(b) Mikhailovsky, A., "Basics of femtosecond laser spectroscopy" (Full Document; Santa Barbara, CA).

• § 102: None. General background/tutorial on femtosecond pump–probe methods; cited for context.

(c) Bourquin, S. et al., "High‑speed femtosecond pump‑probe spectroscopy with a smart pixel detector array," Optics Letters 28(17):1588–1590 (2003).

• § 102: None. Discloses high‑speed pump–probe detection with a detector array; relevant to the "light detector" element of claim 1 and to the InGaAs array detector described in the specification. § 103‑support art at most.

4. Bottom line for the § 102 question

1. No cited reference anticipates independent claim 1 or independent claim 14. Each lacks at least the "white‑light source ... substantially continuous bandwidth of at least 500 nanometers" used for both pump and probe. Several lack multidimensional operation entirely.
2. No cited reference anticipates any dependent claim outright on the record available, though US 2010/0110426 (provisional) and Krebs 2013 come closest to the bandwidth‑dependent claims (3–8, 15–19), and US 2009/0161092 comes closest to the architecture of claim 1.
3. The real exposure is § 103, not § 102. The obviousness theory a challenger would build is: US 2009/0161092 (2D spectrometer architecture with computer‑controlled pump‑pulse separation) + US 4,512,660 and/or US 5,479,256 (broadband continuum in a nonlinear spectroscopy experiment) ± US 2010/0110426 / Krebs 2013 (broadband source, NRB suppression) → claims 1 and 14. Whether that combination succeeds turns on (i) whether the art motivates using a single white‑light source for both pump arms (the '634 patent's stated insight is precisely that center‑frequency tuning is thereby eliminated), and (ii) the specific ≥500 nm / ≥700 nm / ≥900 nm / 1½‑octave thresholds, which are numerical limitations the cited art generally does not meet.
4. Provenance note: five of the six patent citations carry Google's "cited by examiner" asterisk (US 4,512,660; US 5,479,256; US 2008/0125977; US 2009/0161092; US 2010/0110426). US 2009/0122319 is the exception (no asterisk) and is likely an IDS‑type citation of marginal technical relevance. Because these references were before the examiner and the claims still issued, the prosecution history would be the best evidence of how (if at all) the examiner distinguished them — worth pulling from USPTO PatentCenter if this analysis is to be relied on.

Verification tasks I could not complete here: (i) full text of US 2010/0110426 (lookup truncated); (ii) the examiner's stated reasons for allowance and any § 103 rejections in the 14/804,965 file wrapper; (iii) whether any of these references were applied rather than merely listed. Until (i)–(iii) are confirmed, treat the US 2010/0110426 entry and the § 103 framing above as analytical assessments rather than established facts.

Obviousness Analysis — U.S. 9,638,634 B2 under 35 U.S.C. § 103

Building on the earlier sections (bibliography, claim map, litigation), this section assumes the granted claim text and reference list already set out and does not repeat them. All claim language below is quoted literally, including the retained typographical errors.

1. Governing standard

Because the application was filed 2015‑07‑21 with priority to provisional 62/026,949 (2014‑07‑21), the AIA version of §§ 102/103 applies. Obviousness is assessed under Graham v. John Deere, 383 U.S. 1 (1966), as refined by KSR Int'l v. Teleflex, 550 U.S. 398 (2007), and MPEP §§ 2141–2143:

1. scope and content of the prior art;
2. differences between the prior art and the claims;
3. level of ordinary skill in the pertinent art;
4. objective indicia of non‑obviousness.

A prima facie case is made where the art teaches or suggests each limitation, there is an articulated motivation to combine with a reasonable expectation of success, and the combination is not directed to non‑analogous art. KSR authorizes the classic rationales: combining known elements according to known methods to yield predictable results; simple substitution of one known element for another; use of a known technique to improve a similar device in the same way; "obvious to try" where there is a finite number of identified, predictable solutions; and design incentives/market forces.

2. Level of ordinary skill (POSITA)

For this field the POSITA would be a person with a Ph.D. in physical chemistry, chemical physics, or ultrafast/optical physics (or equivalent), plus 2–3 years' experience with femtosecond nonlinear optical spectroscopy — i.e., hands‑on familiarity with Ti:sapphire amplifier systems, supercontinuum/"white‑light" continuum generation in bulk media, pump–probe and four‑wave‑mixing (FWM) geometries, optical delay lines, and multi‑dimensional (2D/3D) data acquisition and Fourier processing. This is the level reflected in the references of record (Krebs et al. 2013; Bourquin et al. 2003; Brida et al. 2012), each of which is a graduate‑level ultrafast‑optics contribution.

3. Scope and content of the prior art of record

The page's Citation section supplies six patent references and three non‑patent references; the specification additionally incorporates three patent applications and the Brida TWINS paper by reference.

Disclosure caveat (important). I have the citation metadata and abstracts, not the full texts, of P1–P6 and N1–N3. The characterizations above are inferred from the titles, the patent's own discussion of them, and general knowledge of these well‑known works; they are not verified line‑by‑line quotations. Where a limitation turns on a precise numeric disclosure (e.g., whether P1's continuum exceeds 500 nm), I flag the uncertainty below rather than assert it.

4. Independent Claim 1 — proposed combinations

Claim 1 requires: sample volume + detector; one white‑light source ≥500 nm continuous bandwidth; a first optical system delivering that light as probe pulses; a second optical system splitting the same source light into ≥2 pump pulses with controllable time separation; and a computer reading the detector and sweeping that separation across shots to build a multidimensional spectrograph.

Combination I — Primary: P3 (Zanni '092) + P2 (transient grating) / P4; Secondary: P1 + N1; Splitter: N4 (Brida)

• Limitations (a),(b),(e‑partial),(f): P3 (multidimensional spectrometer: pulse pairs, delay control, detector, computer) and P2 (pump‑pair/probe with a scanned delay) disclose a sample, detector, delay‑controlled pump pulses, and computer‑driven acquisition of a 2D data set. P4 supplies computer‑driven pulse‑sequence control and processing. N3 supplies the array‑detector/fast‑readout element.
• Limitations (c),(d): P1 (laser‑derived picosecond continuum used as a broadband probe in FWM) and N1 (supercontinuum probe in a 2D FT experiment) supply the broadband "white‑light" source and the broadband‑probe concept.
• Limitation (e) — the pump‑pair splitter: N4 (Brida TWINS, incorporated by reference in the patent itself) supplies a device that takes a single broadband input and produces two phase‑locked pulses with a computer‑controlled separation — the exact structure of claim 12.

Motivation to combine (articulated, KSR‑consistent):

1. Design incentive stated in the art. The '634 background itself identifies the problem: extending multidimensional spectroscopy across the visible/NIR for photosynthesis/solar processes, where "tuning the frequency of pump pulses over a large spectral bandwidth … is difficult." That is a stated design need in the field, supplying motivation to replace the tunable narrowband pump (NOPA) with a broadband source.
2. Simple substitution of a known element. Replacing a tunable narrowband pump with a broadband continuum that already serves as the probe (P1, N1) is the substitution of a known element to obtain the predictable benefit of spectral coverage without tuning.
3. Degenerate pump–probe is old. Using a single laser/continuum to supply both pump and probe (P1's degenerate FWM; standard degenerate pump–probe per N2) makes "one white‑light source feeding both optical systems" a predictable design choice.
4. Reasonable expectation of success. N4 demonstrates that phase‑locked pulse pairs can be produced from a broadband input with controllable delay; N1 demonstrates broadband 2D acquisition; the combination's operation is therefore predictable.

Combination II — Primary: P2 (transient grating) + P1 (broadband CARS/continuum) + N1; Detector/electronics: N3

Same limitations mapped via the pump‑pair/probe architecture plus broadband‑continuum probe, with N3 for the array detector and readout. Motivation: multiplex (broadband) detection of the four‑wave‑mixing signal, a recognized advantage of continuum probes (P1) applied to the pump‑pair/probe geometry (P2/N1).

Combination III (alternative) — Primary: P4 (NIST, NRB‑free FWM) + P3 + P1/N1 + N4

P4 teaches computer‑controlled pulse sequencing in nonlinear (multidimensional) spectroscopy; the same secondaries supply broadband pumping/probing and the phase‑locked splitter.

Likely examiner conclusion on claim 1: A prima facie § 103 case is sustainable, with the honest weak point being the "≥500 nm" magnitude — N1's disclosed supercontinuum is ≈470 nm wide (250–720 nm), and I cannot confirm that P1's picosecond continuum exceeds 500 nm. If no reference is shown to teach a ≥500 nm continuum feeding both branches, the applicant's best argument is that the claimed bandwidth feeding the pump is not taught or suggested.

5. Independent Claim 14 (method)

Claim 14 recites the same apparatus plus steps (a) excite with a pump pair at a first separation, (b) probe and measure, (c) repeat across separations, (d) process to a two‑dimensional spectrum. Steps (a)–(d) are the ordinary 2D data‑acquisition loop of the transient‑grating/multidimensional‑spectrometer art (P2, P3, P4) combined with the broadband source (P1, N1). The KSR "use of a known technique (τ‑scanning + Fourier processing) to improve a similar device" rationale applies directly. Claim 14 rises and falls with claim 1's broadband‑source limitation.

6. Dependent claims

7. Anticipated rebuttals and how they cut

A. "No reference teaches a ≥500 nm continuum feeding both pump and probe." This is the applicant's strongest point. N1 uses a supercontinuum probe (≈470 nm) with narrowband (sub‑20 fs ≈ tens‑of‑nm) pumps; P1's continuum serves as a probe. A rigorous prima facie case therefore depends on (i) P1/Mikhailovsky showing a >500 nm continuum, and/or (ii) an art‑based reason to make the pump broadband. Counter (KSR): the design goal (broad spectral coverage without tuning) is expressly recognized; using the same broadband source for the pump is a predictable substitution; and N4 shows phase‑locked broadband pump pairs are feasible — so the "obvious to try / predictable solution" rationale applies.

B. Teaching away / loss of selectivity. One could argue the art teaches that the pump must be frequency‑selective (a NOPA) and that making it broadband degrades the 2D pump axis. Counter: in 2D spectroscopy the pump axis is recovered by Fourier‑transforming the τ‑scan (the standard broadband‑pulse approach used in 2D IR), so broad pumps are not disfavored; the panel would need actual art statements discouraging broadband pumping.

C. Numeric ranges (claims 4–8, 15–19). Ranges are presumed obvious absent criticality/unexpected results; the applicant would need to show the specific numbers are critical.

D. Objective indicia. No record evidence of unexpected results, long‑felt need, failure of others, or copying appears. The known facts — WARF's licensing of the invention and the inventor's tie to PhaseTech Spectroscopy — are commercialization evidence of weak probative value (no demonstrated nexus between any success and the claimed bandwidth).

E. § 112 / claim‑mapping mismatch (flagged previously). The description assigns "first optical system" to the pump branch while claim 1 assigns it to the probe branch. This is a claim‑construction/§ 112 issue, not a § 103 issue; it does not defeat an obviousness rejection but is worth noting for invalidity exposure.

8. Bottom line

• Claims 3, 9, 10, 12, 13, 20 — clearly vulnerable to § 103. Claim 12 in particular is essentially the Brida TWINS device the applicant itself incorporated by reference; claim 3 is textbook continuum generation (P1/N2).
• Claim 1 and claim 14 — moderately vulnerable. A solid prima facie case exists via Combination I or II (multidimensional‑spectrometer architecture + broadband continuum + TWINS splitter), but it hinges on whether the art teaches or suggests a ≥500 nm continuum applied to the pump; N1's disclosed ≈470 nm supercontinuum and the narrowband pumps of the cited art leave that gap. Expect the examiner to lean on KSR "predictable substitution / obvious to try" reasoning, and the applicant to attack the bandwidth‑feeding‑the‑pump teaching.
• Claims 4–8 and 15–19 — weakest for the patentee but not automatically obvious: they are range claims whose validity turns on criticality, which must be established with art teaching the broader continua or with unexpected‑results evidence.

Caveat. This opinion rests on the citation metadata and abstracts available on the patent page plus general knowledge of these well‑known references; I did not have full‑text claim‑by‑claim disclosures of P1–P6 or N1–N3. Statements about specific bandwidth values (especially P1's continuum span) should be verified against the reference texts before being relied upon in an office action or validity opinion. As previously flagged, the identifier "9638634" also denotes the unrelated WO 9638634 A1 (Huesker Synthetic GmbH, 1996), which is excluded here.