US 5143854

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

Identification (matches the number you gave exactly)

Note on my sourcing: I could not reach USPTO PatentCenter/Patent Public Search or the CAFC docket system directly; the searches returned Google Patents, Espacenet, and Justia mirrors. The claim text below comes from the Espacenet claims record for US5143854A, because the authoritative full text you supplied was truncated in the description section (it ends mid-sentence at "the substrate is preferably as thin as po…") and never reached the claims.

Abstract (as published)

"Polypeptide arrays can be synthesized on a substrate by attaching photoremovable groups to the surface of a substrate, exposing selected regions of the substrate to light to activate those regions, attaching an amino acid monomer with a photoremovable group to the activated regions, and repeating the steps of activation and attachment until polypeptides of the desired length and sequences are synthesized. The resulting array can be used to determine which peptides on the array can bind to a receptor."

This is the founding "VLSIPS" (very large scale immobilized polymer synthesis) patent — the light-directed, photolithographic microarray synthesis disclosure. The spec's own summary of the invention matches: linker molecules with photoremovable protective groups, selective irradiation through masks, coupling of protected amino acids, repetition, then exposure to a labeled receptor and location-resolved detection.

Independent claim — plain-language overview

Claim 1 (the only independent claim I could confirm; all of claims 2–12 that I retrieved depend from claim 1). It is a method of identifying a polypeptide that binds a receptor, and it recites synthesis and screening as one sequential method:

• Step (a): Shine light through a pattern (light and dark areas) onto a first area of a substrate surface — but not a second area — to strip a photoremovable protecting group off polypeptides already immobilized on that surface in the first area.
• Step (b): Contact the whole surface (both areas at once) with a first amino acid carrying its own photoremovable protecting group. It couples only where the surface was deprotected — i.e., in the first area, not the second.
• Step (c): Shine light through another pattern, this time hitting at least part of the first area and at least part of the second area, deprotecting both.
• Step (d): Again contact the whole surface simultaneously with a second amino acid, coupling in both partially illuminated zones.
• Step (e): Repeat the irradiation + coupling cycle until a matrix array of at least 100 different polypeptides is built on the surface, each different polypeptide occupying an area of less than 0.1 cm². Crucially, the sequence and the location of each polypeptide are determined by the mask patterns used and the amino acids delivered — i.e., no separate tagging or positional encoding is needed.
• Step (f): Contact the array with a receptor and determine which polypeptides specifically bind it.

In short: build a spatially addressable peptide library by light-directed synthesis, then screen it by binding.

Dependent claims I retrieved:

• 2 — amino acids may be L-, D-, or synthetic amino acids.
• 3 — receptor is an antibody.
• 4 — substrate is Langmuir-Blodgett film, glass, Ge, Si, (poly)tetrafluoroethylene, polystyrene, GaAs, GaP, silicon oxide, silicon nitride, or combinations.
• 5 — photoremovable group is NVOC (6-nitroveratryloxycarbonyl), 2-nitrobenzyloxycarbonyl, dimethoxybenzyloxycarbonyl, 5-bromo-7-nitroindolinyl, o-hydroxy-α-methylcinnamoyl, 2-oxymethylene anthraquinone, or mixtures.
• 6 — each different polypeptide occupies about 1 µm² to 10,000 µm².
• 7 — irradiation performed while the substrate contacts a wash solution.
• 8 — receptor bears a radioactive or fluorescent marker, and the identifying step is detecting marker location.
• 9 — step (c) uses a mask placed adjacent to the substrate with transparent/opaque regions, illuminated so as to cover half of what was illuminated and half of what was not illuminated in step (a) (the binary "half-and-half" masking scheme).
• 10 — repeat to make 10³ different peptides in 10³ regions.
• 11 — repeat to make 10⁶ different peptides in 10⁶ regions.
• 12 — identifying step further involves contacting the receptor with a second, marked receptor (the retrieved text is truncated here in the source).

Uncertainty I want to flag explicitly: the claims listing available to me was cut off at claim 12, so I cannot confirm the total number of claims or whether any further independent claim exists. I have not seen claim text beyond claim 12, and I am not going to reconstruct it from memory. If a complete count matters for your purpose, pull the patent's claims directly from USPTO Patent Public Search or the granted-PDF on Espacenet.

Litigation / CAFC check for this number

• Google Patents flags the family as "Family has litigation / First worldwide family litigation filed," linking to Darts-IP family 27001838 — so the family has litigation history. Google Patents does not name the cases on the page I retrieved.
• I found no CAFC 2026 docket or appeal tied to US 5,143,854. That is consistent with the record: the patent expired 2009-09-01, so a 2026 appeal on this patent itself would be anomalous. What the 2026 CAFC activity in my results concerned were unrelated patents (e.g., an E.D. Tex. case involving U.S. 9,215,613 and U.S. 8,589,541, and a Delaware § 101 case on U.S. 7,640,517).
• I also note the historical context that surfaced in the search results: the patent drew a $2.2M NIH Human Genome grant announcement for Affymax's VLSIPS chip work, with the article noting the patent issued the previous month and had been applied to peptide and nucleotide arrays; and that the 1991 Science paper and this patent were recognized with an IPO Distinguished Inventor Award (1993). I did not verify specific infringement suits or interferences involving the '854 patent from primary sources, so I am deliberately not naming defendants or case captions.

Number-collision warning (per your "do not auto-correct" rule)

Searching the bare string "5143854" also returns JP 5143854 B2 (Japanese Patent No. 5143854, registered 2012-11-30, published 2013-02-13, owned by 静岡商工会議所 / Shizuoka Chamber of Commerce and Industry, titled 「植物生臭調節作用を示す新規化合物及びそれらの製造方法」 — novel compounds with plant-growth-regulating activity). That is an entirely unrelated patent that merely shares the digits 5143854. Everything above refers only to US 5,143,854 (US5143854A). Likewise, US 5,143,854 should not be confused with its family members, which are separate patent numbers: US 5,424,186, US 5,405,783, US 5,445,934, US 5,510,270, US 5,527,681, US 5,744,305, US 5,753,788, US 5,770,456, US 5,800,992, US 5,874,101, US 5,889,165, US 5,902,723, US 6,225,625, and others listed on the family page.

Bottom line

US 5,143,854 is the 1992 Pirrung/Read/Fodor/Stryer Affymax patent generally credited as the foundational light-directed (photolithographic, mask-based) solid-phase synthesis patent behind Affymetrix GeneChip-style microarrays. Its one claim I can verify, claim 1, is a hybrid synthesis-plus-screening method claim requiring a ≥100-member matrix array of distinct polypeptides in sub-0.1 cm² features whose sequences and positions are set by the light patterns. The patent is expired, and I found no 2026 CAFC activity on it.

I'll research litigation involving US 5,143,854 using the sources you specified.

Let me search more specifically for this patent number in litigation databases.

Let me check litigation-specific databases and the patent's Google Patents litigation link.

Summary of Findings — Litigation Involving US 5,143,854

Bottom line: I was unable to verify any specific U.S. district court case in which US 5,143,854 was actually asserted as a patent-in-suit. I want to state that plainly rather than present adjacent litigation as if it were '854 litigation. What I can report is (a) a verified indicator of family-level litigation, and (b) the related Affymetrix litigations I checked and in which the '854 was not identified as an asserted patent. Details and caveats below.

---

1. The patent at issue (for identification discipline)

Source: https://patents.google.com/patent/US5143854/en

---

2. Direct litigation findings for US 5,143,854

The Google Patents record carries a "Family has litigation" marker linking to Darts-IP family 27001838. That marker is family-level, not patent-level: the '854 sits at the head of a very large family (WO 90/15070; the '934, '305, '992, '101, '456, '992-line continuations; EP 0476014 / EP 0619321 / EP 0902034 / EP 0953835; CA 2054706; AU 651795; etc.). A family-level flag therefore does not establish that the '854 itself was ever asserted. I could not open the Darts-IP record behind that flag within my search budget, so I cannot attribute it to a specific case.

---

3. Affymetrix litigations I checked where the '854 was NOT the (or a) patent-in-suit

I examined these because they are the obvious candidates, and in each the sources identify the asserted patents — and '854 is not among them. I am reporting them as excluded, not as '854 litigation:

Sources consulted include: Incyte investor disclosures (investor.incyte.com), Illumina 2007 Form 10‑K (stocklight.com), the N.D. Cal. joint case management statement in C 06‑05958 JW (cases.justia.com), the Germeshausen/IPMall "Microarray Battle" newsletter, and Foley Hoag's Drug Discovery News reprint (Nov. 2007).

---

4. Similar-number results I affirmatively excluded (per your instruction)

Because you asked me not to return results for similar numbers, I note these look-alikes that surfaced and were not treated as the patent in question:

• A PTAB invalidity filing discussing "the '514 Patent" asserted against Dell and Samsung and listing a UC3854 datasheet (TI, 6/98) and a Sorensen power-supply manual — this is a power-management patent, plainly not Pirrung et al. (ptacts.uspto.gov petitions/1548865).
• JP 5143854 B2 (issued 2013-02-13) — a Japanese patent to the Shizuoka Chamber of Commerce on plant growth-regulator compounds. A JP number, unrelated to US 5,143,854.
• Numerous patents merely citing US 5,143,854 (e.g., US 5,858,659; US 6,646,243; US 2006/0259251) — citation, not litigation.
• Ex parte Seul appeal citation of "Pirrung (US 5,143,854)" as prior art — prosecution citation, not litigation.

---

5. Material caveats you should weigh

1. I could not fully complete the research. My search run terminated before I could query the Unified Patents litigation caselist directly, or run PACER/CAFC docket lookups for the '854. My negative finding is therefore "not verified," not "confirmed none."
2. The practical window matters. The '854 expired by its own terms on 2009-09-01 (anticipated expiration, per Google Patents). Assertion campaigns after that date would be damages-only and are less likely to be indexed prominently.
3. The likely reason '854 is hard to find in litigation: Affymetrix's commercial assertion portfolio in the U.S. was built on the apparatus and method claims of sibling patents ('934, '305, '992, '243, '432, '531, '365, '716) rather than the '854's synthesis method claims. That is consistent with every case I reviewed, but it is an inference, not proof that '854 was never asserted.
4. Foreign/family litigation and oppositions are a separate question — e.g., the European members of this family (EP 0 476 014; EP 0 619 321) were the subject of EPO opposition-type proceedings, and there was U.K. microarray litigation involving Oxford Gene Technology and Affymetrix in the 1990s–2000s. I could not verify in this run whether US 5,143,854 itself was pleaded in any of those, and I am flagging it as unverified rather than asserting it.

---

6. Recommended next steps to close the gap

If you need a defensible "no litigation" or a complete list, I would run, in this order:

1. Unified Patents litigation portal, filtered to patent number 5,143,854: https://portal.unifiedpatents.com/litigation/caselist
2. Darts-IP family 27001838 (the record behind the Google Patents litigation flag), to see which family member was asserted and where: https://patents.darts-ip.com/?family=27001838
3. PACER / CourtListener RECAP full-text search on "5,143,854" — this catches assertion in complaints and joint claim construction statements, not just docket metadata.
4. CAFC docket search on the same string, plus the '854's reissue/reexam and any appeal docketed from an N.D. Cal. or D. Del. case.
5. Westlaw/Lexis or RPX Insight patent-to-case mapping, which is what actually resolves family-level flags into case-level records.

If you want, I can run items 1 and 3 now and give you a case-by-case table with plaintiff, defendant, jurisdiction, case number, filing date, and status — which is the format you asked for and which I cannot fully populate from the sources retrieved so far.

I'll verify the structured-data signal against public sources before writing anything.

The one "IPR2022-00311" hit is for a different patent (8,477,514 — also nicknamed "the '514 patent"). Let me check PTAB dockets directly.

No PTAB hits yet. Two more targeted checks — reexamination/appeal records and any assertion history — to make sure I'm not missing a non-ODP-indexed proceeding.

Proceedings overview

Zero AIA trial proceedings have ever been filed against US 5,143,854 — the USPTO Open Data Portal structured block returns no IPR, PGR, or CBM for this patent, and independent web searches of PTAB/P-TACTS material, Darts-IP-linked family records, and Federal Circuit reports surface no petition, institution decision, or Final Written Decision naming this patent. The defensive bottom line for a defendant is therefore not "the patent is hardened" and not "the claims are canceled," but something more useful: the patent expired on 2009-09-01, roughly three years before the AIA trial regime even opened, so it never entered the IPR era at all — a demand letter asserting live infringement of the '854 patent today is unsupportable on its face.

---

Proceedings reviewed — negative findings

There are no proceedings to detail. What follows is the record of what I checked and what I found, so the absence is documented rather than assumed.

Near-miss to flag — do NOT confuse

IPR2022-00311 is captioned against "the '514 patent," but the patent at issue is US 8,477,514 B2 (Carmichael IP, PLLC counsel of record), a power-management patent. It is unrelated to US 5,143,854. This is the single most likely source of a false positive if someone runs a loose "514" search, and I am flagging it explicitly so it does not contaminate a freedom-to-operate or validity memo. The FWD in that case held claims 1–20 of 8,477,514 unpatentable — that holding says nothing about Pirrung et al.

Adjacent non-AIA proceeding involving the same technology family (unconfirmed as to this patent)

There is an interference appeal involving Affymetrix personnel (Goldberg, Diggelman, Hubbell, McGall, Ngo, Morris, Yamamoto, Tan) that was affirmed by the Federal Circuit on 2008-07-15 in Appeal No. 2007-1519, from Interference No. 105,422 (nonprecedential, Fed. Cir. R. 36 disposition). See http://www.worldlii.org/us/cases/federal/USCAFED/2008/208.html. I have not confirmed that US 5,143,854 was an involved patent in that interference, and I will not assert it was. It is noted only so a defendant's counsel knows there is interference-era activity in the Affymetrix microarray portfolio that predates the AIA and is not an AIA trial. This is a lead to verify, not a finding.

---

Strategic summary

Claim status: every claim is UNTESTED, and the issue is now academic. No claim of 5,143,854 has been canceled, confirmed, or construed in any AIA trial, because none was ever filed. There is no narrowing, no certificate, no surviving-claim list to work from — the patent stands exactly as issued on 1992-09-01, subject only to ex parte prosecution/reexamination history that is outside the scope of this task and that I have not verified. The controlling fact is temporal: per the Google Patents bibliographic record, the patent's anticipated expiration was 2009-09-01, and its status is "Expired - Lifetime" (https://patents.google.com/patent/US5143854/en). The AIA IPR regime opened 2012-09-16; PGR would have been categorically unavailable (it reaches only first-inventor-to-file patents, and this patent's priority runs to 1989-06-07); CBM was likewise unavailable (not a financial-services covered business method, and the "technological invention" exclusion would apply). So the reason there are no IPRs is structural, not vindication — a would-be petitioner had no live patent to attack.

Estoppel landscape: there is none, and that cuts entirely for the defendant. With zero IPRs and zero PGRs, no petitioner (or privy) is bound by § 315(e)(2) or § 325(e)(2). Nothing is off the table by estoppel. That means every prior-art ground — including art that would have been "reasonably could have raised" in a hypothetical IPR — remains fully available in district court, before the PTAB on a different patent in the family, or in an ex parte reexamination. Practically, though, the expiring nature of the patent makes the invalidity case the less important defense: 35 U.S.C. § 286 caps pre-filing damages at six years, so a complaint filed today could reach back only to roughly 2020 — eleven years after the patent's 2009-09-01 expiration. Absent an extraordinary tolling theory, there is no window in which recoverable infringement could have occurred. That is a Rule 12(c)/summary-judgment point, not an IPR point.

Pattern signals: the absence is itself diagnostic. Well-asserted, economically live patents attract IPRs, and the Affymetrix microarray portfolio was heavily litigated — but the asserted patents in the major campaigns (the 2004–2007 Delaware action against Illumina over BeadArray products, and the earlier Incyte, Hyseq, and Oxford Gene Technology disputes) were later family members and continuations — US 5,445,934, US 5,800,992, and the '432/'531/'365/'243 group — not the '854 patent itself. Google Patents flags the family as having litigation (Darts-IP family 27001838), which is family-level, not patent-level. My searches surfaced no defensive aggregator (Unified Patents, RPX, etc.) in the chain and no serial petitioner against this patent. The reasonable inference is that 5,143,854 functioned as the foundational priority document for the GeneChip family rather than as a standalone enforcement asset, and that by the time the AIA trial regime matured, the patent had already been off the books for three years.

---

Recommended next steps

1. If you are a defendant holding a demand letter on 5,143,854, lead with expiration, not invalidity. Confirm the expiration date against the USPTO Patent Center file wrapper (https://patents.google.com/patent/US5143854/en links out to USPTO PatentCenter) — for a patent filed 1990-03-07 and issued 1992-09-01, the pre-URAA term is 17 years from issue, i.e., 2009-09-01. Then apply § 286: any claim for damages must be premised on acts within six years of the complaint, which cannot overlap a patent that expired in 2009.
2. Do not cite the IPR2022-00311 FWD as to this patent. That decision invalidated claims 1–20 of US 8,477,514, an unrelated power-management patent. If opposing counsel or a search vendor has conflated them, correct it in writing; a brief built on that conflation is sanction-bait.
3. Verify the interference lead before relying on it. If Interference No. 105,422 (appealed as Fed. Cir. 2007-1519, affirmed 2008-07-15) touched this patent, the involved-patent identification in the Board's interference file is worth pulling. I could not confirm the link; treat it as unverified. The CAFC disposition is at http://www.worldlii.org/us/cases/federal/USCAFED/2008/208.html.
4. Confirm the negative with a direct registry query. Run the patent number through P-TACTS (https://ptacts.uspto.gov/ptacts/ui/home) and the USPTO PTAB decisions search (https://www.uspto.gov/patents/ptab/decisions) as of the date you file anything. If either returns a proceeding that ODP has not yet indexed, the analysis above must be revised — but the pre-AIA timing of this patent makes a newly surfaced AIA trial against it highly unlikely.
5. If you are the patent owner (or a licensee), the honest assessment is that the '854 patent is not an assertable asset today. Its value is genealogical — it is the 1989 priority root of the GeneChip family — and any monetization theory must run through unexpired continuations, not through this patent.

Confidence and limits. High confidence that no AIA trial proceeding exists against US 5,143,854: the ODP canonical list says none, and three independent search angles produced no petition, institution decision, FWD, or appeal. High confidence in the 2009-09-01 expiration date from the Google Patents legal-status block. I did not verify the patent's full ex parte reexamination history, its complete district-court assertion record as an asserted (rather than cited) patent, or whether Interference No. 105,422 involved this patent; those remain open items. No claim-level PTAB outcomes exist to quote, so none are quoted — any claim-number disposition you see attributed to an IPR on the '854 patent is a conflation with US 8,477,514.

I'll start by pulling the ownership-chain data available from public sources, then cross-check names, dates, and any litigation/NPE signals.

Good — I found a concrete reel/frame for the GE Capital security agreement. Let me chase the remaining reel/frame data and the litigation record.

Ownership Chain Reconstruction — US 5,143,854

Patent: Large scale photolithographic solid phase synthesis of polypeptides and receptor binding screening thereof
Appl. No. 07/492,462 · filed 1990-03-07 · earliest priority 1989-06-07 (US 07/362,901) · issued 1992-09-01 · expired 2009-09-01 (17-yr term from grant, pre-URAA) · legal status per Google Patents: Expired – Lifetime

Data-quality caveat, stated up front. I could not open the USPTO Assignment Center detail pages in this session. The two sources I did retrieve — the Google Patents legal-events snapshot and Espacenet INPADOC legal-status records — do not expose the recording correspondent (attorney/firm). Reel/frame numbers appear only on INPADOC records, and only two of them were retrievable (noted below, with the sibling patent each was verified against). Everything else in the reel/frame and correspondent fields is marked not retrievable. I have not filled gaps by inference.

---

Inventors

Pattern notes:

• Someone left the assignee inside 12 months — but only one of four. Pirrung relocated to Duke University "at the beginning of 1990," ~7 months after the 1989-06-07 priority filing. This is a single-inventor academic relocation, not the "all inventors bolt" pattern that precedes a portfolio fire-sale. Fodor and Read stayed with the group and moved into the Affymetrix spin-off; Stryer remained at Stanford.
• No institutional-co-ownership gap. All four inventors assigned to Affymax N.V. on the same 1990-04-23 recording date — including the Stanford professor. There is no university co-assignee, no Bayh-Dole interest, and no separate institutional assignment in the chain. That is why the chain of title is clean and short.
• Post-departure addresses corroborate the split: the later EP publication (EP 0 902 034 A3) lists Pirrung at Durham, NC 27707, Read at Palo Alto, CA, Fodor at Palo Alto, CA, Stryer at Stanford, CA.
• The four inventors shared the 2006 European Inventor of the Year award (SME category) for this work — evidence the invention was recognized as an operating-company R&D output, not a paper patent.

---

Original assignee

Entity on the issued patent: Affymax Technologies N.V. (recorded as "AFFYMAX TECH NV [NL]"; registered address Willemstad, Curaçao, Netherlands Antilles — the Netherlands Antilles was dissolved in 2010, so the domicile no longer exists as a jurisdiction).

• Shipped a product embodying the claims? No — not Affymax itself. Affymax was a peptide/drug-discovery R&D house. Its own first proof-of-concept was an animated polymeric membrane bearing the Affymax logo imaged by UV through a drilled shadow mask — a demo used to raise money, not a product. Commercialization happened through the subsidiary: Affymetrix operations began 1991–92, first commercial product 1994 (HIV genotyping), GeneChip system commercialized from 1996. By 2001 Affymetrix was shipping 5,000–10,000 arrays/month.
• Primary line of business: Affymax — combinatorial chemistry / drug discovery (library synthesis, receptor-binding screening), i.e. exactly the two subject areas of the patent.
• Current status: Acquired, then merged away. Affymax was purchased by Glaxo in March 1995 for ~US$500M (including Affymax's then-65% stake in Affymetrix), per the 1998 Exp. Opin. Ther. Patents company profile. Affymax N.V. / Affymax Technologies N.V. stop appearing in this patent's record after the 1997/1998 transfer. Note the name collision: a different, later Affymax, Inc. (the post-2001 peginesatide/Omontys developer, spun out of the Glaxo-era institute) is not an assignee of this patent; do not attribute that entity's later financial history to this chain.
• Successor operator: Affymetrix, Inc. (NASDAQ: AFFX) — GeneChip microarrays; acquired by Thermo Fisher Scientific on 2016-03-31 for ~US$1.3B, brands retired from 2017. That acquisition post-dates the patent's 2009 expiry, which is why no assignment for it appears in this patent's record.

---

Assignment timeline

Recording dates below are as reported in the Google Patents legal-events snapshot; where a distinct execution/effective date was retrievable it is given separately. Conveyance types are quoted literally from the record.

Not retrievable — recorded 1990-04-23 — Reel/frame not retrievable

• Conveyance: Assignment of assignors interest (four separate records, one per inventor)
• Assignors: PIRRUNG, MICHAEL C. (separate record) · STRYER, LUBERT (separate record) · FODOR, STEPHEN P. A. (separate record) · READ, J. LEIGHTON (separate record)
• Assignee: AFFYMAX N.V., a Netherlands corp.
• Correspondent: not retrievable — I could not access the Assignment Center detail page that carries the recording attorney.
• Context: Original inventor-to-company assignment — the patent's root of title; executed in connection with the June 1989 filing campaign.

Not retrievable — recorded 1990-08-03 — Reel/frame not retrievable

• Conveyance: Assignment of assignors interest
• Assignor: AFFYMAX N.V., a corp. of the Netherlands
• Assignee: AFFYMAX TECHNOLOGIES N.V.
• Correspondent: not retrievable
• Context: Internal reorg / intra-group shift of the IP into the group's Netherlands-Antilles holding entity (Willemstad, Curaçao), which is the entity that then prosecuted and holds the family.

Executed/effective 1997-02-22 — recorded 1998-02-10 — Reel 008955/0530

• Conveyance: Assignment of assignors interest
• Assignor: AFFYMAX TECHNOLOGIES N.V.
• Assignee: AFFYMETRIX, INC. (California)
• Correspondent: not retrievable
• Reel/frame provenance (flag): 008955/0530 was verified on sibling US 5,510,270 in Espacenet INPADOC (.../inpadoc?CC=US&NR=5510270A...), which shows "ASSIGNMENT OF ASSIGNORS INTEREST; ASSIGNOR: AFFYMAX TECHNOLOGIES N.V.; REEL/FRAME: 008955/0530" with effective date 1997-02-22 and event date 1998-02-10 — identical to the date Google Patents reports for the Affymetrix assignment on '854. The instrument is a portfolio-wide array-family assignment; I flag that I did not visually confirm the reel/frame on '854's own record page.
• Context: Portfolio carve-out/consolidation — the VLSIPS array family is moved out of the offshore Affymax holding entity into the operating subsidiary (Affymetrix, public since June 1996, ~33%-owned by Glaxo Wellcome at end-1996). This is affiliate-to-operator consolidation, not a sale to a third party.

Not retrievable — recorded 1998-12-23 — Reel/frame not retrievable

• Conveyance: Merger (see document for details)
• Assignor: AFFYMETRIX, INC., a corp. of California
• Assignee: AFFYMETRIX INC., a corp. of Delaware
• Correspondent: not retrievable
• Context: Internal reorg — California→Delaware reincorporation only, no change of control. Corroborated by the N.D. Cal. opinion in the Hyseq litigation: shareholders voted in June 1998 to reincorporate in Delaware and "completed this reincorporation in late September of 1998." Not a fire-sale, not a transfer to an asserter.

Executed 2012-06-25 — recorded 2012-06-27 — Reel 028465/0541

• Conveyance: Security Agreement
• Assignor: AFFYMETRIX, INC.
• Assignee: GENERAL ELECTRIC CAPITAL CORPORATION, as Agent (Maryland)
• Correspondent: not retrievable
• Reel/frame provenance (flag): 028465/0541 verified on sibling US 6,308,170 B1 via Espacenet INPADOC (.../inpadoc?CC=US&NR=6308170B1...): "SECURITY AGREEMENT; ASSIGNOR: AFFYMETRIX, INC.; REEL/FRAME: 028465/0541", event date 2012-06-27, effective 2012-06-25. Portfolio-wide IP collateral filing.
• Context: Securitization — IP collateral for Affymetrix's $100,000,000 Credit Agreement dated as of June 25, 2012, among Affymetrix as Borrower, GE Capital as Agent, other credit parties, and Silicon Valley Bank as Syndication Agent, with GE Capital Markets and SVB as Joint Lead Arrangers (credit agreement public in Affymetrix's SEC filings). No title transferred.

Not retrievable — recorded 2015-11-13 — Reel/frame not retrievable

• Conveyance: Release by Secured Party (see document for details)
• Assignor: GENERAL ELECTRIC CAPITAL CORPORATION, as Agent
• Assignee: AFFYMETRIX, INC.
• Correspondent: not retrievable
• Context: Lien release — termination of the 2012 collateral position on repayment/refinancing. Not an acquisition; the encumbrance clears and Affymetrix, Inc. is left as assignee of record.

No further assignments recorded. There is no recorded assignment after 2015-11-13. In particular, the Thermo Fisher Scientific / Affymetrix merger of 2016-03-31 (~$1.3B) does not appear in this patent's assignment record — consistent with the patent having already expired in 2009, so there was no live right to record against it. Succession to Thermo Fisher is therefore a matter of corporate merger fact, not of an assignment recorded against '854.

Litigation (context, not an assignment): Google Patents flags "Family has litigation – First worldwide family litigation filed" (Darts-IP family 27001838). The earliest family suit I could actually verify is Affymetrix, Inc. v. Hyseq, N.D. Cal. Case No. C 98-03192, filed 1998-08-18, asserting family siblings — the retrieved opinion refers to the '305, '934, '992 and '716 patents, not to '854. I could not confirm a suit naming '854 itself from the sources retrieved. Given the 2009-09-01 expiry and the 35 U.S.C. § 286 six-year damages lookback, any assertion of '854 would have had to be filed by roughly late 2015 to recover anything — and no such assignment or assertion appears.

---

Timeline diagram

timeline
    title Ownership of US 5143854
    1989 : Priority application filed Jun 7
    1990 : Four inventors assign to Affymax NV
         : Affymax NV moves rights to Affymax Technologies NV
    1992 : Patent issued Sep 1
    1997 : Assignment to Affymetrix executed
    1998 : Assignment recorded Feb 10
         : Reincorporation from California to Delaware
    2009 : Patent expires Sep 1
    2012 : Security interest granted to GE Capital
    2015 : Security interest released
    2016 : Thermo Fisher acquires Affymetrix

---

NPE / troll-pattern signals

1. Shell-entity transfer — not present.
No assignee in the chain carries an "IP / Patents / Licensing / Holdings / Ventures" suffix, and no transfer runs from an operating company to a licensing-only vehicle. The closest structural analogue — the 1990-08-03 shift from AFFYMAX N.V. to AFFYMAX TECHNOLOGIES N.V. (both Netherlands-Antilles entities, Willemstad, Curaçao) — is an intra-group IP-holding reorganization, not a shell transfer: the group ran real laboratories (Affymax Research Institute, Palo Alto) and the receiving entity was the family's prosecution vehicle on the parallel EP filings (EP 0 902 034 A3 lists Affymax Technologies N.V., Willemstad, Curaçao as applicant). The 1997-02-22 transfer moved the patent out of the offshore holder into the operating company — the inverse of the pattern. Offshore domicile alone is not a finding.

2. Known asserter in the chain — not present.
Neither the current nor any prior assignee of record matches a public NPE list (Acacia, Marathon, Intellectual Ventures, IPNav, Wi-LAN, MOSAID/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, Spangenberg entities). The assignees of record are: Affymax N.V. → Affymax Technologies N.V. → Affymetrix, Inc. (CA) → Affymetrix Inc. (DE) → Affymetrix, Inc., with General Electric Capital Corporation, as Agent (reel 028465/0541) appearing only as collateral agent, which is not an assertion posture — and it was released (2015-11-13). No high-frequency-plaintiff entity ever touched this title.

3. Repeat correspondent across the chain — unclear / not verifiable.
I could not retrieve the recording correspondent for any entry in this chain; the Google Patents legal-events snapshot and the Espacenet INPADOC records expose reel/frame and dates but not the filing attorney. I therefore cannot test for recurrence, and I will not infer a name. For completeness only (and not as a finding): Affymetrix's inventor-side assignments in the 2000s were faxed internally from "Affymetrix Legal" (legacy assignment reel 014171/0604 for appl. 10/389,194) — that is in-house filing activity on unrelated applications, and it post-dates every entry in this chain.

4. Cascading transfers — not present.
The record is the opposite of a cascade: one intra-group transfer (1990-08-03), eight years of stability, one affiliate consolidation (executed 1997-02-22 / recorded 1998-02-10), one pure reincorporation (1998-12-23), then nothing. No chained LLCs, no shared correspondent address that can be identified, no common-principal pattern within any 24-month window.

5. Pre-litigation transfer — not present.
The only transfer proximate to litigation is the Affymax Technologies N.V. → Affymetrix, Inc. assignment (effective 1997-02-22, recorded 1998-02-10, reel 008955/0530), whose recordation falls ~6 months before the 1998-08-18 Hyseq complaint. Three reasons this is not a pre-litigation NPE transfer: (a) the transferee is the operating company that made and sold the products, not a special-purpose asserter; (b) the effective/execution date is ~18 months before the suit, not within 6; and (c) the asserted patents in that suit were family siblings, not '854. It looks like standard affiliate consolidation ahead of/around the public subsidiary's separation from the Glaxo-owned parent, not venue/standing engineering.

6. Bankruptcy fire-sale — not present.
No assignee or assignor in this chain filed for bankruptcy protection in connection with the patent. Affymax was acquired by Glaxo (March 1995, ~$500M) — a solvent trade sale, not a Chapter 7/11 liquidation. Affymetrix was acquired by Thermo Fisher (2016-03-31, ~$1.3B) — likewise solvent. The patent had been expired for ~7 years before the Affymetrix sale, so it could not have been an asset of value in that or any later transaction. Name-collision warning: any "Affymax bankruptcy" hit refers to a different, later company (the post-2001 peginesatide developer) that is not in this chain; I did not verify that entity's proceedings here.

7. Privateering — not present.
Affymetrix was a famously aggressive licensor (≈10% of income from license fees and royalties by 2000) and a serial litigant against competitors, but it asserted in its own name — it did not vest patents in an NPE to assert on its behalf. No interposed third-party assertion vehicle appears anywhere in this record, and no SEC filing or third-party coverage retrieved describes one. This is vertical integration of assertion, not privateering.

8. Defensive aggregator (anti-NPE) — not present.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. It terminates at an operating company (Affymetrix, Inc., of record; now within Thermo Fisher Scientific), and — decisively — the patent expired 2009-09-01 and is flagged "Expired – Lifetime." It was neutralized by the calendar, not by an aggregator.

---

Verdict

Operating-company assertion.

The chain never leaves the Affymax → Affymetrix operating group: four inventor assignments to Affymax N.V. (recorded 1990-04-23), an intra-group shift to Affymax Technologies N.V. (1990-08-03), the family's consolidation into the operating subsidiary Affymetrix, Inc. (effective 1997-02-22, recorded 1998-02-10, reel 008955/0530), and a bare California→Delaware reincorporation (1998-12-23). The only non-operating entity to appear is General Electric Capital Corporation, as Agent, and only as collateral agent under a security agreement (executed 2012-06-25 / recorded 2012-06-27, reel 028465/0541), which was released (2015-11-13) — securitization is not assertion. The patent expired 2009-09-01, no assignment has been recorded since 2015, and Affymetrix commercialized the invention (GeneChip) and litigated the family itself (e.g., Affymetrix v. Hyseq, N.D. Cal. C 98-03192, filed 1998-08-18) — so this is an operating-company patent that was asserted by its owner and is now expired. Caveats I am not papering over: (i) reel/frame numbers for four of the six entries and the recording correspondent for all six are not retrievable from the sources I could reach, so signal #3 is genuinely unverifiable; (ii) I could not confirm that a suit ever named '854 specifically; (iii) the 2016 Thermo Fisher acquisition is not a recorded assignment on this patent.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent number 5143854) · also indexed at https://assignment.uspto.gov/patent/index.html
Supporting records used: https://patents.google.com/patent/US5143854/en (legal events) · Espacenet INPADOC for US 5,510,270 (reel 008955/0530) · Espacenet INPADOC for US 6,308,170 (reel 028465/0541) · Thermo Fisher close announcement, 2016-03-31: https://thermofisher.mediaroom.com/2016-03-31-Thermo-Fisher-Scientific-Completes-Acquisition-of-Affymetrix-Following-Approval-of-Transaction-by-Affymetrix-Stockholders

I'll research this systematically. Let me start with searches on the patent and its cited references.

Let me get the specific "References Cited" section for US5143854.

Let me try to retrieve the actual front-page reference list from the patent document itself.

1. Patent identification (verified)

Identifier caution (operating rule: no auto‑correction). Search results returned an unrelated JP 5143854 B2 (2013‑02‑13, "novel compounds showing plant odor‑regulating action," https://patentimages.storage.googleapis.com/51/65/fa/6031ef9d47948c/JP5143854B2.pdf). That is not the patent in question. Likewise, US5143854 (US) and EP 0 476 014 B1 / WO 90/15070 are family members of the same priority (PCT/NL1990/000081) and are therefore not prior art against US 5,143,854.

Important procedural note on my retrieval. The Google Patents text supplied in the prompt is truncated after the classification list and contains no "References Cited" section. Using the search tool I could not reach USPTO PatentCenter/PAIR directly, so I grounded the reference list on mirrors of the USPTO front page (Google Patents, Espacenet, EveryPatent). I retrieved the "Other References" (non‑patent) list in full and the patents cited in the specification text. I could not retrieve the complete face‑page "U.S. Patent Documents" (56) list, and I will not reconstruct or guess U.S. patent numbers for it. All § 102 conclusions below are limited to the verified record.

---

2. The verified prior‑art record of US 5,143,854

2a. Non‑patent references on the face page ("Other References")

Source: http://www.everypatent.com/comp/pat5143854.html

2b. References cited in the body of the specification (verified from the patent text supplied)

• U.S. Pat. No. 4,719,615 (Feyrer et al.) — cited for "translating the substrate under a modulated laser or diode light source." Relevant to the light‑patterning steps of claim 1 and to maskless exposure. (Issue date not verifiable from the records retrieved in this session.)
• Copending U.S. application Ser. No. 404,920 — catalytic polypeptides; a co‑pending application, not published prior art.
• Sze, VLSI Technology (McGraw‑Hill, 1983) and Mead & Conway, Introduction to VLSI Systems (Addison‑Wesley, 1980) — photolithographic masking/alignment background.
• Stryer, Biochemistry, 3rd ed., 1988 — amino‑acid nomenclature only.
• Merrifield (1963), Patchornik (1970), Amit (1974), Flanders (1977) — as tabulated above.

---

3. § 102 analysis, claim by claim

Claim 1 is the operative independent claim (13 claims total). Its elements are:

• (E1) selective irradiation of a first area of a substrate surface bearing immobilized polypeptides having a photoremovable protective group, leaving a second area unirradiated;
• (E2) simultaneous contact of both areas with a first amino acid bearing a photoremovable protective group, coupling only in the first area;
• (E3) a further light/dark pattern covering part of both areas; (E4) simultaneous contact with a second amino acid;
• (E5) repetition to yield a matrix array of ≥100 different polypeptides, each in an area <0.1 cm², sequence and location defined by the light patterns;
• (E6) contact with a receptor and identification of the specifically binding polypeptides.

Anticipation under 35 U.S.C. § 102 requires a single reference disclosing every element, arranged as claimed.

Merrifield 1963 (§102(b) art). Discloses E2/E4 chemistry (stepwise coupling of α‑protected amino acids to a solid support) and is the SPPS foundation the specification itself concedes. It discloses no patterned light, no photoremovable group, no array, and no receptor screening → anticipates no claim. Use: § 103 motivation/background.

Patchornik 1970 and Amit 1974 (§102(b) art). These are the two references that supply the chemical core of E1's deprotection mechanism: photolabile (o‑nitrobenzyl / NVOC) protecting groups and their photochemical removal, including on peptides. Amit 1974 in particular bears directly on claim 5 (NVOC, nitrobenzyloxycarbonyl, etc.) and on the "photoremovable protective group" limitations of claim 1. Neither discloses spatial patterning by a light/dark mask, an array of ≥100 different sequences in sub‑0.1 cm² areas, or receptor screening → each, alone, anticipates no claim; the combination with Merrifield is the classic § 103 route to the chemistry of claim 1.

Geysen et al. 1986/1987/1988 (the most relevant prior art). These references disclose the purpose and readout of the invention:

• synthesis of a large multiplicity of distinct peptides at predefined, known solid‑phase locations (the multipin/microtitre‑plate format), and
• contacting that collection with antibody receptors and identifying which peptides bind — including "replacement sets" for epitope mapping, which is E6 in substance.

They are the strongest references against the concept of an addressable peptide library screened by a receptor, and they are the reason the claims were drafted around the light‑directed, photolithographic spatial addressing and the ≥100 sequences in <0.1 cm² density limitation (E5). Geysen's addressing is mechanical (physical pins/wells), not photochemical; there is no photoremovable protecting group, no mask, and no two‑dimensional array defined by light patterns. Because every dependent claim (2–13) carries all of claim 1's limitations, Geysen anticipates no claim as issued; its proper role is as the primary § 103 reference combined with Patchornik/Amit (photolabile groups) and McGillis/Flanders (masking/alignment).

McGillis, VLSI Technology 1983; Sze 1983; Flanders 1977; and U.S. Pat. No. 4,719,615 (Feyrer). These disclose the enabling lithographic technologies that E1 and E3 borrow: producing patterns of light and dark areas with a mask, registering successive masks, and translating a substrate under a modulated laser/diode source. McGillis/Flanders bear on claim 9 (mask with transparent/opaque regions; illumination of half the previously illuminated and half the previously unilluminated substrate), and Feyrer bears on the maskless alternative. All are directed to semiconductor/optics applications, disclose no peptide chemistry, no photoremovable amino‑acid protecting groups, and no receptor screening → no anticipation of any claim; § 103 combination material only.

Applied Biosystems Model 431A manual (15 Aug 1989). Discloses automated, instrument‑controlled peptide coupling — E2/E4 in isolation. Its date is decisive and cuts both ways:

• If the claims are entitled to the 7 June 1989 parent priority date, this manual (and Lieberman 1990 and the two press items) postdates the invention and cannot be § 102(a) or § 102(b) art — only possible § 103 background.
• If any claim is supported only by the 7 March 1990 CIP disclosure, the manual becomes available as § 102(a) art, but it still cannot anticipate claim 1 for want of the light‑directed array elements.

Anonymous press items (Biotechnology News 1 Feb 1990; Chem. & Eng. News 5 Mar 1990) and Fodor et al. (submitted 21 Oct 1990). The two press items fall in the same priority gap noted above and, more importantly, describe the Applicants' own programme (and could raise the Applicants' own § 102(b)/§ 103 "own‑work" issue if the priority date were lost). The Fodor et al. manuscript was submitted after both the priority and the filing dates and is therefore not prior art at all.

References whose content I could not verify (#7, 8, 9, 11, 12, 14, and in part 13, 19). I am not able to describe these accurately from the records retrieved, and I will not assign a § 102 mapping to references whose disclosure I have not read. From the roles they play in the list (synthetic photochemistry, surface/polymer chemistry, immunochemistry), they appear to be § 103 background, but that is an inference, not a verified finding.

---

4. Bottom line

1. No single reference in the verified record anticipates claim 1 of US 5,143,854 under § 102. Each candidate supplies at most one cluster of elements — Merrifield/Patchornik/Amit the solid‑phase photolabile‑group chemistry; Geysen the addressable peptide library plus receptor readout; McGillis/Sze/Flanders/Feyrer the lithographic pattern generation and alignment. Anticipation fails on the missing elements of E1–E5 taken as an ordered whole.
2. The most relevant prior art is the Geysen corpus (Science 235:1184‑1190, 1987; Mol. Immunol. 23:709‑715, 1986; 1986 CIBA Symposium; Molecular Recognition 1988), which teaches synthesizing many different peptides at known positions and screening them against antibody receptors — i.e. the substance of E6 and the library concept. It is squarely a § 103 reference, not an anticipatory one.
3. The second most relevant cluster is Patchornik 1970 + Amit 1974, which supply the photoremovable protecting groups recited in claim 1 and enumerated in claim 5.
4. Third is the lithography cluster (McGillis/VLSI 1983, Sze 1983, Flanders 1977, U.S. 4,719,615), which supplies the mask/alignment/illumination techniques of claims 1 and 9.
5. Two caveats for any real validity work: (a) I could not obtain the complete face‑page (56) U.S. Patent Documents list for US 5,143,854 from the sources reachable here, so my reference set is complete only as to the "Other References" list and in‑specification citations — a full § 102 review should be run against the USPTO PatentCenter/Global Dossier file wrapper (https://patents.google.com/patent/US5143854/en, Global Dossier link on that page); and (b) litigation is recorded for this family (Darts‑IP family 27001838), so invalidity contentions from that litigation would be the authoritative real‑world § 102/§ 103 record.
6. The patent expired 1 September 2009; the analysis above is therefore historical/defensive rather than of current exclusionary significance.

Verification status. Patent bibliographic data, the non‑patent reference list, and the claim text were retrieved and are cited above. Issue dates and subject matter for items #7–9, #11–14 and #19, and the issue date of U.S. 4,719,615, were not verified in this session and are flagged as such rather than asserted.

I'll ground this analysis in the actual prior-art record for US 5,143,854 rather than my memory, since the full text you supplied was truncated before the "References Cited" section. Let me pull the cited references.

Good — I now have the U.S. patent references. Let me get the foreign documents and non-patent literature, plus the prosecution/validity history.

Let me pull the complete cited-reference list and the key non-patent references' content.

Obviousness Analysis — US 5,143,854 under 35 U.S.C. § 103

0. Framework, and honest limits of what I could verify

Governing law. US 5,143,854 was filed 1990‑03‑07, so pre‑AIA § 103(a) applies (the AIA first‑inventor‑to‑file provisions govern applications filed on or after 2013‑03‑16). The controlling obviousness framework is therefore Graham v. John Deere Co., 383 U.S. 1 (1966), as refined by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) — which applies to pre‑AIA § 103 as well. Where the earlier sections of this analysis flagged uncertainty, I carry it forward:

• The claim set I actually verified stops at claim 12, and I could not confirm whether further independent claims exist. A full § 103 analysis is claim‑by‑claim; I can only do claim 1 comprehensively and claims 2–12 provisionally. This is the single biggest limitation on this memo.
• I could not retrieve the file wrapper — i.e., the examiner's actual grounds of rejection and the Reasons for Allowance. The reference list I use below is the applicant‑side IDS list (via the everypatent.com mirror and uspto.report), which is not the same as the examiner's cited art. The everypatent listing I retrieved was truncated at "4,946,942 | Aug., …", and I retrieved no foreign patent documents at all (I flag one that matters in § 7).
• Dates and citation strings below are quoted from those mirrors; where I have not personally read a reference's text, I say so rather than paraphrasing it as if I had.

Critically for a § 103 analysis: the patent's own specification frames the problem in a way that makes the motivation inquiry unusually easy to satisfy. The supplied text concedes the prior art and states the unmet need:

"a tubular reactor system may be used to synthesize a linear polymer on a solid phase support by automated sequential addition of reagents. This method still does not enable the synthesis of a sufficiently large number of polymer sequences for effective economical screening."

"Methods of preparing a plurality of polymer sequences are also known in which a porous container encloses a known quantity of reactive particles… As with other methods known in the art, this method cannot practically be used to synthesize a sufficient variety of polypeptides for effective screening."

The first quoted passage is Merrifield‑style solid‑phase synthesis; the second is Houghten's "tea‑bag" method (US 4,631,211, which is on the patent's own IDS list). So the asserted invention is self‑characterized as an answer to "how do we make more sequences, faster, on known locations." That framing is a gift to a § 103 challenger.

---

1. Decomposition of claim 1 into limitations, and the art that maps onto each

Claim 1 (as reconstructed in the prior section) is a hybrid synthesis‑plus‑screen method claim:

Notably, every limitation has a home in the cited art. The § 103 question is therefore not whether the elements were known, but whether the combination — in particular, substituting light‑directed masked deprotection for mechanical addressing (pins, bags, wells) — was obvious with a reasonable expectation of success.

Partial note on limitation (e): I was not able to verify Geysen's actual pin dimensions or per‑pin surface area. If Geysen's pin tips were on the order of a few mm² each, the "< 0.1 cm²" and even "≥100" numerical limitations may be literally met by the prior art's own hardware — which would convert claim 1 into a pure "old elements, new arrangement" case. I flag this as the highest‑value unverified fact in the whole analysis; it should be checked before relying on it.

---

2. Primary § 103 combination #1 — the "three‑reference" case (Houghten/Geysen + Patchornik + photolithographic masking)

This is the strongest and most natural challenge.

Reference A — Parallel peptide synthesis + receptor screening. Geysen, Meloen & Barteling, PNAS 81:3998‑4002 (1984) ("Use of peptide synthesis to probe viral antigens for epitopes to a resolution of a single amino acid"); Houghten, PNAS 82:5131 (1985) ("General method for the rapid solid‑phase synthesis of large numbers of peptides…"); Geysen J. Immunol. Meth. 102:259‑274 (1987). Geysen's own later work was patented (US 4,833,092, issued 1989‑05‑23 — on the IDS list). Verified via multiple independent citing documents and the patent's own reference list.

• What it teaches: synthesizing hundreds of distinct peptides on separate solid supports arranged in a coordinate fashion (pins in a microtiter‑plate block; porous bags in a net), and then screening those peptides for binding to a receptor (e.g., antibody) with a labeled read‑out. That is limitations (a)‑(b)‑(d)‑(f) in substance — with mechanical addressing, not light.
• Motivation it supplies: the reference's own stated purpose is the need for large numbers of peptides for epitope mapping and drug discovery. This is the same problem the '854 specification admits was unsolved. Under KSR, "the inventor of [the] subject matter… had the same problem in mind" as the prior art is a classic motivation‑to‑combine finding.

Reference B — Photoremovable protecting groups. Patchornik, Amit & Woodward, J. Am. Chem. Soc. 92:6333‑6335 (1970) ("Photosensitive Protecting Groups"); Amit, Zehavi & Patchornik, J. Org. Chem. 39:192 (1974); Amit et al., "Photosensitive protecting groups — A review," Israel J. Chem. (cited in the descendant patent US 5,744,101's list). All on the '854's own IDS list.

• What it teaches: o‑nitrobenzyl / nitroveratryl carbonyl‑type groups that are removed by light of a selected wavelength, and the use of such groups on amino/functional groups so that a spatially selective photochemical step can reveal reactive functionality. Patchornik's paper — as its title and the subsequent literature attest — expressly contemplated applicability to syntheses carried out on solid supports and to "a photochemical approach to the synthesis of biopolymers."
• Why this is decisive: the '854 does not claim photolabile protecting groups as such; claim 5 (dependent) recites NVOC, 2‑nitrobenzyloxycarbonyl, 5‑bromo‑7‑nitroindolinyl, etc. — all of which are Patchornik/Amit chemistry. Under KSR, reciting the known species of a known genus in a known process is ordinarily obvious. Independent claim 1's novelty, if any, cannot rest on the identity of the protecting group.

Reference C — Masked photolithography at micron scale, with alignment. Sze, VLSI Technology (McGraw‑Hill 1983); Mead & Conway, Introduction to VLSI Systems (Addison‑Wesley 1980); Flanders et al., "A New Interferometric Alignment Technique," Appl. Phys. Lett. 31:426‑428 (1977); plus US 4,719,615 (Feyrer, Jan. 1988) for laser/diode scanning of a translated substrate, and US 4,808,508 (Platzer, Feb. 1989, photoresist), all cited in the '854 or its spec.

• What it teaches: projecting light through a patterned mask onto a sensitized surface to produce a spatially defined pattern of chemical change, with (i) micron/near‑micron resolution, (ii) registration of successive masks via alignment marks, and (iii) the use of complementary masks to cover successive areas. Flanders is cited by the patent itself for aligning successive patterning steps by interferometry — an admission that successive‑mask registration was known art, not invention.
• Motivation it supplies: photolithography is the standard answer to "I need a large number of precisely located, precisely bounded features defined by a pattern." KSR: "if a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious."

The § 103 syllogism. (1) There was a recognized, articulated need for very large numbers of spatially distinct peptides for receptor‑binding screening (Houghten, Geysen; conceded in the '854 spec). (2) The only known way to make a spatially addressable set was to physically separate the synthesis sites — pins or bags — and the known problem with that is scale: pin/bag addressing is limited by mechanical handling, not by chemistry. (3) Photolabile protecting groups had been proposed and demonstrated (Patchornik 1970) precisely so that the deprotection step could be controlled by light rather than by physical contact. (4) Masked photolithography was a mature, off‑the‑shelf way to pattern light on a surface with micron resolution and repeated registration. Substituting a light‑pattern for a mechanical mask/pin pattern as the addressing mechanism is, on its face, "a predictable variation" and arguably "obvious to try" from a finite set of identified solutions. That is a strong prima facie case.

Where I think the case is weakest — and a sophisticated challenger must confront it. The linchpin is the reasonable‑expectation‑of‑success prong, and here the patentee has real ammunition inside the specification. The '854 does not merely shine light through a mask; it recites a substantial body of non‑trivial engineering, including:

• irradiation in contact with a solution containing reagents to scavenge photolysis by‑products (CO₂, nitrosocarbonyls, styrene/indole derivatives) — dependent claim 7 and the spec's discussion of "reagents to prevent the by‑products formed by irradiation from interfering with synthesis";
• a contrast enhancement layer (quinone diazide or transiently bleached material) to sharpen the pattern;
• index‑of‑refraction matching solutions, and turbulent‑flow (ridged‑cavity) reactor design;
• attention to stepwise coupling efficiency across many cycles because "substantially pure" sequence per region is required.

A POSA in 1989 reading Patchornik would know that photolabile groups can be removed by light. What was genuinely uncertain was whether mask‑directed photodeprotection could be iterated 5–20 times on a glass surface with acceptable yield and fidelity to produce substantially pure sequences in ≤ 0.01 cm² regions — i.e., whether the known photochemistry would survive contact with the known engineering constraints. That is the "reasonable expectation of success" battleground, not the "was it known" battleground. Notably, this is a factual dispute, and it is why the answer cannot be given with certainty from the patent text alone.

Note: Fodor et al., "Light‑Directed Spatially‑Addressable Parallel Chemical Synthesis," cited on the IDS as "submitted to Science on Oct. 21, 1990" (later Science 251:767 (1991)), is the inventors' own paper submitted after the 1990‑03‑07 filing. It is not prior art and must be excluded from any § 103 combination; its presence on an IDS list is not a prior‑art admission.

---

3. Primary § 103 combination #2 — "Old elements, new arrangement," with the array itself as the only novel element

If a challenger accepts that light‑directed synthesis is new, claim 1 can still be attacked as an obvious substitution of a known patterning technique into a known synthesis scheme, with the array density argued to be a result‑effective variable.

• Combination: Geysen US 4,833,092 or Houghten US 4,631,211 (parallel synthesis + screening) + Sze / Mead & Conway / Flanders (patterning light through complementary masks) + Patchornik (photolabile groups).
• The "binary masking" argument for claims 1(e) and 9: repeated application of n complementary mask pairs to a common substrate is the standard way to address 2ⁿ regions — the very concept used in masked ROMs and in photolithographic wafer processing by the early 1980s. If a POSA wanted to make "10³ different peptides in 10³ regions" (claim 10) or "10⁶ in 10⁶ regions" (claim 11), the arithmetic of binary masking is the obvious bookkeeping. Caveat: the "2ⁿ regions from n mask pairs" framing is a semiconductor‑art argument that I have not grounded in a specific retrieved passage of Sze or Mead & Conway in this session. It should be pinned to a page cite before being relied on.
• The "result‑effective variable" argument for claims 6, 10, 11: claim 6's "about 1 µm² to 10,000 µm²" and claims 10/11's 10³/10⁶ sequences recite degrees of miniaturization. Absent evidence that the recited ranges produce an unexpected result beyond what the ranges themselves suggest, such numerical limitations are ordinarily obvious (In re Applied Materials, 692 F.3d 1289 (Fed. Cir. 2012); In re Woodruff, 919 F.2d 1575 (Fed. Cir. 1990)**). The counter‑argument — that 10⁶ distinct sequences cannot be made by scaling Geysen's pins, and that the claimed density is only reachable by the light‑directed route — is the patentee's best answer, and it is a real one.

---

4. Claim‑by‑claim § 103 assessment

---

5. Objective indicia (Graham factor 4) that cut against obviousness

Because the prima facie case in § 2 is strong, the patentee's case would rest on secondary considerations. From what I could verify:

1. Industry recognition and immediate uptake. The record I retrieved shows the patent drew a $2.2M NIH Human Genome grant for Affymax's VLSIPS chip work (the article noting the patent had issued the prior month), and the 1991 Science paper and this patent were recognized with an IPO Distinguished Inventor Award (1993). (Earlier section; I did not verify the primary source.)

2. Commercial success with nexus. The claimed method is the stated basis of Affymetrix's GeneChip products. The patent is cited by ~3,251 documents (Espacenet citing‑documents count) and is classified across three distinct arts (C07K peptide chemistry; G03F photolithography; C40B combinatorial chemistry) — the classification itself documents that this was a cross‑disciplinary combination that the prior art had not made.

3. Deliberate licensing of a core technology. The retrieved record confirms broad licensing under Affymetrix's microarray patents (WO 92/10092 and EP 0 476 014 B1 — the European family member of the '854), with published reports of Affymetrix licensing programs and cross‑licenses with OGT, Agilent, and GE. Evidence that competitors needed licenses is evidence of technical significance.

4. Extensive validity litigation and opposition. The Google Patents "Family has litigation" flag points to Darts‑IP family 27001838. The retrieved record confirms:

   • Affymetrix v. Oxford Gene Technology / OGT v. Affymetrix — U.S. D. Del. jury verdict (Nov. 2000) of infringement of US 5,700,637, plus U.K. High Court (Apr. 2000), Court of Appeal (Nov. 2000, reversal) proceedings.
   • Affymetrix v. Incyte Genomics — interference proceedings against Affymetrix's US 5,744,305 and US 5,800,992, and a reported 2001 ruling "invalidating Affymetrix patent" (I could not verify which Affymetrix patent was invalidated, and I will not name one).
   • EPO oppositions against the '854's European counterpart EP 0 476 014 B1 (multiple opponents, including OGT) and against OGT's EP 0 373 203 / EP 0 386 229.

   Important caveat: these are family‑level and sibling‑patent events. I found no record of a § 103 invalidity holding against US 5,143,854 itself. The litigation does show, however, that sophisticated adversaries with every incentive to find § 103 art did attack the family — and I have no retrieved evidence that any of them succeeded in invalidating the '854 on § 103 grounds. That silence is weak but real evidence of nonobviousness.

5. A claimed technical obstacle (claim 7). The specification's insistence that irradiation occur in contact with a solution containing by‑product scavengers indicates the inventors encountered a real problem (photolysis by‑products poisoning subsequent couplings) that the bare Patchornik teaching would not have flagged. Secondary considerations include "obvious to try" failures; a documented, solved obstacle is the kind of evidence that defeats a KSR "predictable variation" argument.

---

6. Honest assessment: what the § 103 case can and cannot establish

What I can say with reasonable confidence:

• Every element of claim 1 has a home in references the applicant put on its own IDS, including the photolabile protecting groups (Patchornik/Amit) and the successive‑mask alignment technique (Flanders).
• The motivation to combine is unusually well documented because both the prior art (Houghten, Geysen) and the '854 specification itself identify the same unmet need: producing large numbers of spatially distinct peptides for receptor‑binding screening.
• The dependent claims (2–6, 8, 9, 12) are largely vulnerable to a straightforward § 103 attack; several recite chemistry or materials lifted directly from the cited art.

What I cannot establish, and where the case would likely be decided:

• Reasonable expectation of success in early 1989 for iterated, masked photodeprotection of a glass‑immobilized peptide array with sequence fidelity sufficient to give "substantially pure" regions. This is a factual question on which I have no primary record (no expert declarations, no prosecution history, no reexamination file).
• The true content and date of the late‑1989 / early‑1990 references on the IDS (e.g., Pidgeon Anal. Biochem. 1989; Lieberman Science 247:59‑61 (Jan. 5, 1990); Baum, Chem. & Eng. News (Mar. 5, 1990); Biotechnology News (Feb. 1, 1990)). These post‑date the 1989‑06‑07 parent filing and are at best § 102(a)/(e) art relative to the 1990‑03‑07 filing, subject to swearing behind under 37 C.F.R. § 1.131. Whether the '854's claims are entitled to the 1989‑06‑07 date at all (this is a CIP of Ser. No. 07/362,901) determines the entire prior‑art date structure and could materially change the analysis. I could not resolve the CIP's written‑description support for claim 1.
• The examiner's actual § 103 position. I did not retrieve the file wrapper, and I want to be explicit: I do not know whether the examiner ever made a § 103 rejection, what art was applied, or how it was overcome. Any assertion that the '854 "was allowed over" a particular combination would be speculation, and I am not making it.

---

7. A reference that is notably absent from the IDS list I retrieved

For the "array on a solid support" concept — though not for light‑directed synthesis — the relevant art is WO 89/10977 (Southern et al., published 1989‑11‑16), which I surfaced in a retrieved citation table for US 6,028,189 and which corresponds to E.M. Southern's GB 8810400 filing (1988‑05‑03) and to OGT's EP 0 373 203 / EP 0 386 229 family. Its international filing date precedes the '854's priority date, and under pre‑AIA § 102(e) (as amended by the AIPA) an English‑language PCT designating the US can be prior art as of its international filing date.

But three caveats: (i) WO 89/10977 concerns immobilized/deposited oligonucleotide probe arrays, not in‑situ light‑directed peptide synthesis, so it supplies the "array on a substrate" concept but not limitations (a)–(d); (ii) my retrieved IDS list did not include any foreign patent documents at all (the mirror listing was truncated), so I cannot confirm whether it was or was not cited; (iii) the 1989‑11‑16 publication date postdates the 1989‑06‑07 priority date, so its applicability depends entirely on the § 102(e) analysis and on the CIP priority question in § 6. I raise it because it is the one piece of art that a § 103 challenge could pair with the "known regions on a solid support" language, and any serious analysis of this patent must run it down.

---

8. Bottom line

Under § 103, claim 1 of US 5,143,854 is genuinely vulnerable but not clearly invalid, on the record I can verify. The strongest combination is:

Geysen PNAS 81:3998 (1984) and/or Houghten PNAS 82:5131 (1985) and US 4,631,211, in view of Patchornik JACS 92:6333 (1970) / Amit J. Org. Chem. 39:192 (1974), in further view of conventional masked photolithography (Sze, VLSI Technology (1983); Mead & Conway (1980); Flanders Appl. Phys. Lett. 31:426 (1977); Feyrer US 4,719,615).

The motivation is supplied by the shared, expressly stated objective of making large numbers of distinct peptides for receptor‑binding screening; the photolabile‑group element is supplied verbatim by the cited art; and the patterning element is supplied by a mature semiconductor art that the patent itself cites for successive‑mask alignment. The dependent claims are, in my assessment, largely obvious.

The claim‑1 battleground is the "reasonable expectation of success" prong, and the patentee's best evidence is internal to the specification: the inventors had to solve photolysis‑by‑product poisoning (claim 7), contrast loss, and iterative‑coupling fidelity before the concept worked. Under KSR, a combination that yields "predictable results" is obvious; a combination that requires solving unrecognized problems to yield a result the prior art did not predict is not. Which characterization applies here is a question of fact that cannot be resolved from the patent text alone — it requires the prosecution history, contemporaneous expert testimony about the 1989 state of the art, and confirmation of Geysen's actual feature dimensions.

Three things I want on the record as unresolved, because they change the answer: (1) the complete claim set (I verified only through claim 12); (2) the CIP priority question and thus the effective prior‑art date; and (3) the examiner's actual grounds of allowance, which I did not retrieve. Any of the three could flip a close call.

Sources cited: US 5,143,854 full text (https://patents.google.com/patent/US5143854/en); reference list mirror (http://www.everypatent.com/comp/pat5143854.html); US 5,744,101 cited‑references list (https://patentimages.storage.googleapis.com/18/01/0d/9b953bbede9a56/US5744101.pdf); citing‑documents count (https://gr.espacenet.com/publicationDetails/citingDocuments?CC=US&NR=[5143854A](/patent/5143854A)); litigation context (https://www.nature.com/articles/35008240.pdf, https://www.biocentury.com/article/[236774](/patent/236774)/oxford-gene-s-holy-war, http://www.ogt.co.uk/pdfs/2005/18_10_05_DNA_array_patent_upheld_in_Eur.pdf, https://www.medgenetik.de/de/2000/2000_3_337_Troesch.pdf).