US 6858881

I'll search for authoritative information on this patent, including any 2026 litigation activity.

Let me verify litigation/Court of Appeals activity with additional searches.

US Patent 6,858,881 — Verification Summary

Data source note: I retrieved this via open web search of the Google Patents record for the number exactly as given (US6858881 / US6858881B2), which mirrors the USPTO bibliographic record. I could not retrieve the USPTO Patent Center page or an assignment/legal-status API directly, and my search of Court of Appeals for the Federal Circuit (CAFC) 2026 dockets returned no matches for this patent. All numeric identifiers below are reproduced literally as they appear.

Bibliographic data

Abstract (as issued)

"A semiconductor chip has a substrate that is in the form of a parallelepiped whose side surfaces are shaped as tilted parallelograms. Such a semiconductor chip has a high output efficiency and a homogeneous thermal load due to having at least two side surfaces that are provided with an acute angle and are in the form of parallelograms."

Plain-language overview of the independent claims

Claim 1 (apparatus — radiation-emitting semiconductor chip).
The chip has (a) an active layer, and (b) several side surfaces that bound the lateral extent of that active layer. The core requirement: at least two of those side surfaces meet an adjacent face at an acute angle, are tilted (i.e., not perpendicular to the base), and are parallelogram-shaped. In other words, the chip body is a slanted/skewed parallelepiped rather than a rectangular cuboid, with the tilted facets acting to steer light out more efficiently while keeping a constant cross-section.

Claim 6 (method — producing such a chip).
The method has three steps: (1) provide a substrate; (2) form an active layer with sides above one surface of the substrate; and (3) separate the substrate together with the active layer to form individual chips, where the separation is performed so that the resulting chips have multiple side surfaces bounding the active layer's lateral extent, with at least two of those side surfaces at an acute angle, tilted, and parallelogram-shaped. The specification's disclosed implementation is sawing with an oblique saw blade; claim 11 (dependent) and claim 12 capture the "separating surface running obliquely" and "oblique saw blade" variants, while claim 4 and claim 9 cap the acute angle at less than 80°, and claims 5/10 list the substrate materials (sapphire, gallium nitride, silicon carbide, zinc oxide, diamond, quartz glass).

Context from the specification (not repeated from claim text)

The stated advance over the prior art — principally JP 10-326910 (a truncated-pyramid LED), together with the cited US 5,087,949 and EP 0405757 (Hewlett-Packard) and US 6,229,160 — is addressing the varying cross-section of pyramidal chips, which made heat dissipation harder and raised local resistance. The tilted-parallelogram design keeps a uniform cross-section top-to-bottom (uniform heat dissipation, uniform current path, low forward voltage) while ray-tracing (FIG. 4/FIG. 5) shows output efficiency rising from ~25% at a 90° vertical base angle to ~50% at ~50°. The primary non-patent reference cited is Song, J.L. et al., "Efficiency Improvement in Light-Emitting Diodes Based on Geometrically Deformed Chips," SPIE Conference on LEDs, San Jose, 1999.

Docket / litigation status — what I could and could not confirm

• CAFC 2026 dockets: No search result linked patent 6,858,881 to any Federal Circuit appeal in 2026, and no 2026 docket entry was returned. I cannot confirm the absence of any such proceeding from open search alone — that is a negative finding, not authoritative proof, and a PACER/CAFC docket lookup by patent number would be needed to state it conclusively.
• USPTO post-grant proceedings: No PTAB (IPR/PGR), ex parte reexamination, or other contested proceeding record for this patent surfaced. The only post-grant document I could confirm is the 2005 certificate of correction.
• Caveat on search hits: Searches for the bare number "6858881" also returned unrelated matches that are not this patent — including JP Patent No. 6858881 (Mitsubishi Electric, a circuit-breaker case), and a Brazilian corporate registration number 6858881. These were disregarded as distinct identifiers.

Uncertainty statement: The bibliographic fields, claim text, and abstract above are directly quoted/derived from the authoritative full patent document and are high-confidence. The current-owner line (AMS OSRAM International GmbH) is the listing's own tentative value and should be treated as unverified; and the "no litigation" conclusion is bounded by what open search can establish.

I'll search litigation databases for US patent 6858881.

The initial results were polluted by unrelated patents (e.g., OxyContin's 8,337,888, Exergen's 6,292,685). Let me run more targeted searches.

Bottom line

Based on the searches I was able to run, I found no litigation — past or present — that specifically asserts or involves U.S. Patent No. 6,858,881. I want to be explicit about that rather than present a fabricated case list. Below I explain what I searched, what I did and did not find, and the important caveats.

What I searched

I ran targeted queries against general web indexes (which surface Google Patents, Justia, Dockets.Justia, CourtListener/RECAP, insight.rpxcorp.com, the CAFC opinions site, and the Unified Patents portal) using the exact number "6858881" and "6,858,881," combined with terms like "litigation," "infringement," "lawsuit," "OSRAM," and "LED."

Critical caveat: my tool access was cut off before I could complete a full sweep (for example, a direct query against portal.unifiedpatents.com/litigation and PACER returned only training-style/portal-index pages, not a case list). So this should be read as "no litigation found," not as a definitive negative clearance. A proper PACER/Docket Navigator/Unified Portal search should be run to confirm.

Why the results were noisy (important for interpreting them)

Searches for the bare string returned many unrelated matches — for example:

• OxyContin / Purdue v. Amneal, which concerns U.S. Patent No. 8,337,888 ("the '888 patent") — a pharmaceutical patent, not '881.
• Exergen v. Thermomedics/Sanomedics, which concerns U.S. Patent No. 6,292,685 ("the '685 patent") — a thermometer patent.
• Technical LED IP, LLC v. Osram Sylvania, which concerns U.S. Patent No. RE41,685 ("the '685 patent").
• A Japanese patent JP 6858881 B2 (Mitsubishi Electric, a circuit breaker) — a different patent entirely, coincidentally sharing the number 6858881.

None of these involve US 6,858,881, and I flag them only so they aren't mistaken for hits. Per your instruction, I have not auto-corrected any of these identifiers.

Relevant OSRAM LED litigation I found — none of which names 6,858,881

Because the '881 patent is an OSRAM Opto Semiconductors LED chip-shaping patent, I looked specifically at OSRAM's LED patent campaigns. I found substantial OSRAM LED litigation, but U.S. 6,858,881 does not appear as an asserted patent in any of it:

• Osram's Section 337 ITC investigation (inv. based on a June 2004 complaint; see USITC Pub. 3944). The asserted patents were U.S. 6,066,861; 6,277,301; 6,613,247; 6,245,259; 6,592,780; 6,576,930; 6,376,902; 6,469,321; 6,573,580; and 6,716,673 — not 6,858,881.
• Osram GmbH v. LG Electronics, Inc., et al., No. 4:11-cv-02699 (N.D. Cal. 2011) — asserted patent 6,812,500.
• LG Electronics, Inc. v. OSRAM GmbH, et al., No. 1:11-cv-00660 (D. Del. 2011).
• Bluestone Innovations Texas LLC v. OSRAM GmbH, et al., No. 4:12-cv-00731 (N.D. Cal.) — plaintiff asserting its own patents against OSRAM.
• Bench Walk Lighting LLC v. OSRAM Sylvania Inc., et al., No. 1:2018-cv-01732 (D. Del., filed Nov. 2, 2018) — plaintiff asserting its own LED patents (a separate Bench Walk/related action cited U.S. 7,239,080 and others).
• Lexington Luminance LLC v. Osram Sylvania Inc., 972 F.3d 1324 (Fed. Cir. 2020) — plaintiff's patent.
• Document Security Systems, Inc. v. OSRAM GmbH, et al., No. 2:17-cv-05184 (C.D. Cal.).
• Jam Strait, Inc. v. Osram Sylvania, Inc. (E.D. La.) — U.S. 8,517,583.
• Church of Scientology / Citizen Electronics v. OSRAM (D.D.C.) — declaratory judgment over the '861, '259, '301, '930, '780, '247, and '500 patents.

In every one of these, the patent at issue is a different number from 6,858,881. I did not find 6,858,881 named in any complaint, ITC complaint, or docket.

What the record does show about the '881 patent

• It is a continuation of PCT/DE01/01952 (published as WO 2002/003477 A1), claiming priority to DE 100 32 838.5 (filed July 6, 2000). U.S. filing date Jan. 6, 2003; issued Feb. 22, 2005.
• It has a division: U.S. 10/949,915 → US 6,972,212 B2 ("Method for producing a radiation-emitting semiconductor chip").
• The Google Patents record lists only patent-literature citations in "Cited By" (e.g., US 2004/0262625 A1; US 2006/0151798 A1) — these are later patent documents citing '881 as prior art, not assertions of '881 against anyone.
• Legal status: Expired – Lifetime; anticipated expiration 2021-05-22. The patent term has run, which is consistent with my not finding active assertion, though it does not by itself rule out past suits.

My honest conclusion

No litigation involving U.S. Patent No. 6,858,881 is known to me based on the available search results. I will not supply a plaintiff/defendant/jurisdiction/case-number table, because doing so would require me to fabricate case data I did not find — which I won't do.

If you need a definitive answer, the most reliable next steps would be:

1. A direct search on PACER (all district courts) and CourtListener/RECAP for "6,858,881."
2. The Unified Patents litigation portal and a Docket Navigator patent-number search.
3. USITC EDIS for any Section 337 complaint naming '881.

I can run any of these if you re-enable searching, and I'll report exact case numbers, parties, filing dates, and outcomes for any true hits.

I'll verify the structured data against public sources before finalizing.

The initial results are false positives (a different "'685 Patent" in camera/IPR2024-00923 matters). Let me disambiguate.

Confirmed: IPR2024-00923 concerns U.S. 10,873,685 (Sony v. Optimum Imaging) — unrelated. Let me run a final targeted sweep.

Proceedings overview

Zero AIA trial proceedings are on file for US 6,858,881. The count is 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denied — the USPTO Open Data Portal "PTAB proceedings on file" block returns no IPR, PGR, or CBM for this patent, and my independent web sweep (PTAB petition documents on ptacts.uspto.gov, IPR docket aggregators, and Google Patents' litigation/proceeding tab at https://patents.google.com/patent/US6858881/en) surfaced no petition, institution decision, FWD, or appeal naming this patent. That means there is no IPR-driven estoppel, no claim-cancellation record, and no PTAB-honed claim construction to hand a defendant. What a defendant has instead is a bare, never-PTAB-tested patent — and one that is already expired: the Google Patents record shows "Anticipated expiration" on 2021-05-22, with the last maintenance-fee event (12-year fee) on 2016-08-16 and no subsequent fee payment. Anybody asserting this patent today is asserting an expired right.

---

No proceedings to report

I want to be blunt about the negative finding rather than pad this section with invented numbers. Specifically:

• No IPR/PGR/CBM docket number exists for US 6,858,881. Nothing to list, no petitioner to name, no FWD to quote.
• Search trap worth flagging: the phrase "the '685 Patent" is heavily used in PTAB papers for U.S. Patent No. 10,873,685 (Sony Corp. v. Optimum Imaging Technologies LLC, IPR2024-00923, filed 2024-05-20, terminated-settled 2025-06-05 before Judges Baer, Quinn, Cass, and Horvath; panel addressed obviousness over Iwasawa and Enomoto, Fintiv factors, with parallel Canon/Olympus litigations). That is a digital-camera-aberration patent owned by Optimum Imaging — a different patent entirely. Likewise, search hits referencing "claims 1–30 of the '685 Patent" and Sony's Watanabe/Takane grounds are all US 10,873,685. Do not let anyone paste those into a brief about 6,858,881.
• Search trap #2: Osram's own litigation history involves many patents but not this one. The ITC § 337-TA-512 action (Certain Light-Emitting Diodes, Inv. No. 337-TA-512) asserted U.S. 6,066,861, 6,245,259, 6,277,301, 6,376,902, 6,469,321, 6,573,580, 6,576,930, 6,592,780, 6,613,247 and 6,716,673 — not 6,858,881 (see https://www.usitc.gov/intellectual_property/documents/pub3944.pdf). The 2019 Osram v. Lighting Science Group Delaware suit asserted 6,848,819, 7,045,956, 7,227,191, 7,642,734, 7,957,146, 7,994,519, 9,241,392 and 9,698,282 — again not this patent.
• Caveat: ODP ingest can lag newly filed papers, and my search is a point-in-time sweep as of 2026-10-01. I found nothing, and the structured ODP block is the canonical source, so the correct conclusion is "no PTAB activity on file." I state that with high confidence because the absence is corroborated by two independent sources (ODP + web).

---

Strategic summary

Claim status: all 12 claims are UNTESTED in any AIA trial — none canceled, none sustained by the Board, none construed by the Board. Independent claim 1 (radiation-emitting chip with ≥2 side surfaces at an acute angle forming parallelograms) and independent claim 6 (method of making the same by separating the substrate and active layer so as to produce such side surfaces) stand exactly as they issued on 2005-02-22, plus dependent claims 2–5 and 7–12. The only amendment-style wrinkle in the file is a certificate of correction dated 2005-05-31, which is clerical, not substantive. There is also an independent prosecution-history fact worth knowing: the patent was granted over a § 102/§ 103 record that included US 5,087,949 (Hewlett-Packard, "Light-emitting diode with diagonal faces"), EP 0 405 757 A2, JP 10-326910 A, JP H04-78174 A, US 5,869,848, US 5,990,497, US 5,864,171, GB 2,326,023 A, US 6,229,160 B1 (Lumileds, "Light extraction from a semiconductor light-emitting device via chip shaping"), EP 0 961 328 A2, and the non-patent literature Song, J.L. et al., "Efficiency Improvement in Light-Emitting Diodes Based on Geometrically Deformed Chips," SPIE Conference on Light-Emitting Diodes, San Jose, 1999. That body of art — especially Song and the HP diagonal-faces patents — is the natural starting point for any invalidity theory, because the examiner already considered and distinguished it, which makes it § 325(d) risk but also makes it the most defensible technical ground.

Estoppel landscape: there is none — § 315(e)(2) estoppel requires a petitioner who reached a final written decision, and no one did. For a defendant being asserted against today, every prior-art ground under § 102/§ 103 and § 112 remains available. There is no bar, no "reasonably could have raised" shadow, and no privity chain to worry about. Practically, though, the more important point is that the patent is expired (2021-05-22, per the ODP/Google Patents anticipated-expiration entry). A defendant's first move should not be an IPR — the Board would likely deny institution under § 325(d) and/or discretionary denial, and there is no injunctive relief to enjoin going forward. The first move should be a damages-period and standing analysis: expired patents can still support back damages for the six-year lookback under § 286, but only for infringement occurring on or before expiration.

Pattern signals: none of the usual ones apply. No serial petitioner (no petitioner at all). No PTAB appeals by the patent owner, because there is nothing to appeal — the only Federal Circuit-facing event in this family would be the original prosecution, and I found no CAFC docket. No defensive aggregator (Unified Patents, RPX, Open Invention Network) appears in the chain; the current assignee is Ams Osram International GmbH (formerly Osram Opto Semiconductors GmbH), a large operating entity, not a troll or an NPE — which materially lowers the probability that this specific patent ever becomes an assertion vehicle. Note the family context: the sibling division US 6,972,212 ("Method for producing a radiation-emitting semiconductor chip," filed 2004-09-24, issued 2005-12-06) is also expired and also has no PTAB activity. Foreign counterparts EP 1 297 579 A1 (Withdrawn), JP 2004-503094 A (Withdrawn), WO 2002/003477 A1 (Ceased), DE 100 32 838 B4 (Expired – Fee Related), and TW I229459 B (IP Right Cessation) are all dead or abandoned — a family whose prosecution history shows the owner largely walked away from the non-US filings.

The meta-signal: the absence of PTAB activity here is not the "well-asserted patents eventually attract IPRs" signal, because this patent was apparently never asserted in a US district court litigation or ITC action that I can find. It is a 2000-priority LED chip-shaping patent, expired in 2021, that sat quietly in a large portfolio. The absence of proceedings therefore carries little information about claim strength — it reflects non-assertion, not survival.

---

Recommended next steps

1. Confirm the expiration and the maintenance-fee history directly. The Google Patents legal-status timeline (https://patents.google.com/patent/US6858881/en) shows "2021-05-22 Anticipated expiration" and "Expired - Lifetime," with the last fee payment logged 2016-08-16 (12-year). Pull the USPTO Patent Center transaction history to confirm no 16-year or later fee, no reissue, and no ex parte reexamination. If the patent is expired and clean, any demand letter citing it should be met with a § 286 damages-window analysis rather than a validity fight.
2. Do not file an IPR. With no prior IPR and no prior FWD, the Board has no estoppel hook to apply — but § 325(d) cuts hard against you here because the examiner distinguished US 5,087,949, US 6,229,160, GB 2,326,023 A, EP 0 405 757 A2 and the Song SPIE paper during prosecution. Save the art for district court, where the Phillips standard and a lay jury are more favorable to a chip-geometry obviousness story.
3. If someone does assert it, check whether they briefed you with the wrong patent. The single most likely error you will encounter is citation to Sony Corp. v. Optimum Imaging Technologies LLC, IPR2024-00923 (P.T.A.B.), or to the related Nikon/Panasonic/Olympus/Fujifilm petitions (e.g., IPR2024-01373, on U.S. 7,612,805) — all of which concern the other "'685 Patent," U.S. 10,873,685, in the Optimum Imaging camera family. None of that is probative for U.S. 6,858,881. Link the correct record: https://patents.google.com/patent/US6858881/en.
4. If an active proceeding does surface after my sweep, the milestones to calendar are the § 314(b) three-month institution deadline from the petitioner's § 313 preliminary-response date (or four months from service of the petition if no POPR), the § 316(a)(11) one-year FWD deadline from institution, and the oral-hearing notice roughly two to three months before the FWD. I found no such proceeding, so treat any claim of one as requiring a docket number before you rely on it.

Bottom line: there is nothing on file. No proceeding number to cite, no FWD to quote, no cancellation to lever. The real defensive asset here is not a PTAB record — it is the expiration date and the absence of any assertion history.

I'll start by verifying the assignment records and the corporate history around this patent.

Let me check for any litigation or NPE assertion activity involving this patent or its family.

Let me try to pull the actual USPTO assignment records directly.

I've verified what I can from the authoritative patent record and cross-checks. Two important caveats up front: the USPTO Assignment Center's correspondent-of-record field is not reproduced in the data I can access for this patent, and my searches did not surface it. I will not invent it.

Inventors

Pattern notes: No unusual departure pattern is evidenced. All four executed the assignment on the same two days (2003-02-06 to 2003-02-07), i.e. roughly one month after the 2003-01-06 U.S. continuation filing and ~2.5 years after the 2000-07-06 German priority date (DE 10032838.5). That timing is consistent with a routine continuation-filing formalities execution, not with a pre-fire-sale inventor exodus. Nothing in the record shows any inventor leaving OSRAM within 12 months, and I found no evidence either way on that point beyond the absence of an assignment out of OSRAM.

Original assignee

OSRAM Opto Semiconductors GmbH, Regensburg, Germany — named as assignee on the issued patent (recorded as "OSRAM OPTO SEMICONDUCTOR GMBH" on reel 015059/0194 and "OSRAM OPTO SEMICONDUCTORS, GMBH" on reel 016018/0123 — a one-letter entity-name discrepancy between the two recordings of the same instrument).

• Primary line of business: optoelectronic semiconductors — LEDs, high-power laser diodes, infrared emitters, optical sensors. Wholly owned subsidiary of OSRAM GmbH; founded 1999 as an OSRAM/Infineon joint venture; OSRAM bought out Infineon's remaining 49% in August 2001.
• Product embodying the claims: Yes. The claims cover a chip-shape geometry (tilted parallelogram side surfaces giving a uniform cross-section). OSRAM shipped side-emitting/volume-emitter LED chips in this family; the specification's own model is an SiC-substrate GaN LED. This is a mainstream shipped product line, not a paper asset.
• Current status: Operating. In 2021 OSRAM Opto Semiconductors was integrated into ams OSRAM International GmbH (ams OSRAM Group, SIX-listed, ~EUR 4.8bn 2022 revenue), which Google Patents lists as the current assignee. The predecessor was not dissolved or bankrupt — it was folded into a larger listed group.

Assignment timeline

Recorded assignments appearing on the face of US 6858881 (Google Patents legal-events section, which mirrors the USPTO assignment record):

• 2003-02-06 / 2003-02-07 (executed) / recorded 2004-03-11 — Reel 015059/0194

  • Conveyance: Assignment (assignment of assignors' interest)
  • Assignor: Eisert, Dominik; Härle, Volker; Kühn, Frank; and others
  • Assignee: OSRAM OPTO SEMICONDUCTOR GMBH (Germany)
  • Correspondent: Not reproduced in the accessible record — I could not verify the correspondent of record and will not guess.
  • Context: Ordinary inventor-to-employer assignment executed in connection with the U.S. continuation filing; no consideration of a third party.

• 2003-02-06 / 2003-02-07 (executed) / recorded 2004-11-26 — Reel 016018/0123

  • Conveyance: Assignment
  • Assignor: Eisert, Dominik; Härle, Volker; Kühn, Frank; and others
  • Assignee: OSRAM OPTO SEMICONDUCTORS, GMBH (Germany)
  • Correspondent: Not reproduced in the accessible record — unverified.
  • Context: Second recording of the identical instrument (same signing dates, same assignors). Treated as a duplicate/corrective recording, not a distinct transfer. The only difference between the two entries is the assignee-name spelling.

• No post-issuance assignments are recorded. There is no recorded transfer from OSRAM Opto Semiconductors GmbH to ams OSRAM International GmbH against this patent number. The 2021 change is a corporate integration reflected in Google Patents' "Current Assignee" field but not as a reel/frame on this patent. Likewise, no security agreement, license, merger, or release is recorded.

• Also in the family (not a recorded assignment, but relevant to chain reconstruction): continuation-in-part US 10/949,915, filed 2004-09-24 as a division of this application, issued as US 6,972,212 B2 ("Method for producing a radiation-emitting semiconductor chip") and remains with the same OSRAM/ams OSRAM line.

If your Assignment Center query returns nothing under the patent number, that is consistent with the above: the only recorded links are the two duplicate inventor→OSRAM recordings. There is no post-issuance chain to trace.

Timeline diagram

timeline
    title Ownership of US 6858881
    2000 : German priority filing DE10032838
    2001 : PCT application filed
    2003 : US continuation filed
         : Inventors sign assignment to OSRAM
    2004 : Assignment recorded Reel 015059/0194
         : Duplicate recorded Reel 016018/0123
    2005 : US 6858881 issued to OSRAM
    2021 : Patent expired

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The patent never leaves the operating group. The chain is inventors → OSRAM Opto Semiconductors GmbH (reel 015059/0194, reel 016018/0123) and stops. No "IP / Holdings / Licensing / Ventures" entity appears, and no registered-agent-service address is associated with this patent.

2. Known asserter in the chain — not present. Neither OSRAM Opto Semiconductors GmbH nor ams OSRAM International GmbH appears on the Acacia / Marathon / IV / IPNav / Wi-LAN / Conversant / Vringo / Pendrell / Round Rock / Spangenberg lists or on Unified Patents / RPX high-frequency-plaintiff rosters.

3. Repeat correspondent across the chain — not present (and indeterminate). The two recordings share signing dates and assignors, so any common correspondent would be expected — but the correspondent field is not available in the data I can access, and I did not locate it via search. I am therefore recording this as unverified rather than as a finding. This is the one item worth a manual pull from Assignment Center.

4. Cascading transfers — not present. Only one substantive transfer exists (inventor → employer), plus a duplicate recording 8.5 months later. Nothing resembling chained LLC transfers inside 24 months.

5. Pre-litigation transfer — not present. No infringement suit naming US 6858881 was found in my searches. The only "685 patent" litigation hits returned were unrelated (MIT v. Harman, Exergen v. Sanomedics) and involve different patent numbers.

6. Bankruptcy fire-sale — not present. No Chapter 7/11 proceeding involving OSRAM Opto Semiconductors GmbH or ams OSRAM; the 2021 event was a corporate integration into a listed parent, not a distress sale. (Contrast the genuine fire-sale model: Kodak, Nortel, Polaroid.)

7. Privateering — not present. No transfer to an assertion vehicle; the group continued to make and sell the accused-eligible product itself.

8. Defensive aggregator — not present. The chain terminates at an operating manufacturer, not at RPX, AST, LOT, Unified Patents, or OIN. The patent has not been neutralized by aggregation — it simply expired.

Status note: Anticipated expiration 2021-05-22; Google Patents status "Expired – Lifetime." Maintenance fees were paid at the 4-year (2008-07-15), 8-year (2012-07-06) and 12-year (2016-08-16) intervals, with a certificate of correction on 2005-05-31. A patent kept in force for its full term by a manufacturer is inconsistent with the abandonment-on-the-vine pattern typical of NPE inventory.

Verdict

Operating-company assertion — with an explicit caveat that only the "operating company" prong is established.

Justification: the complete recorded chain is two duplicate inventor-to-employer assignments (reel 015059/0194, recorded 2004-03-11, and reel 016018/0123, recorded 2004-11-26) running to OSRAM Opto Semiconductors GmbH, with no post-issuance transfer, no security interest, and no LLC intermediary anywhere. The current owner of the family is ams OSRAM International GmbH, a large publicly listed optoelectronics manufacturer that ships the claimed chip geometry and paid maintenance fees through full term (expiration 2021-05-22). All eight NPE signals are negative or unverified, so this is not an NPE matter; note, however, that no infringement assertion is on record either, so the "assertion" half of this bin is unevidenced — the practical takeaway for the tracker is "operating-company-held, non-asserted, expired."

The one open item is signal 3: pull the correspondent of record for reel 015059/0194 and reel 016018/0123 manually, since it is not in the accessible record and is the field your workflow depends on most. Verify at the USPTO Assignment Center: https://assignmentcenter.uspto.gov/ (search patent number 6858881), mirrored at https://assignment.uspto.gov/patent/index.html.

I'll research the prior art cited in US 6,858,881, verifying each reference against live sources.

Let me verify the remaining key references.

Let me verify the remaining cited references relevant to claim analysis.

Prior-Art Analysis — US 6,858,881 ("6858881")

Scope note. I searched for the number exactly as given (6858881 → US 6,858,881 B2, "Radiation-emitting semiconductor chip, and method for producing the semiconductor chip," OSRAM). I did not treat the unrelated JP 6,858,881 (Mitsubishi circuit-breaker) or the Brazilian registration number as this patent. Sources used: Google Patents record for US6858881 (which mirrors the USPTO bibliographic record) and the individual cited references. I could not reach USPTO Patent Center directly, so bibliographic fields are drawn from the Google Patents mirror; the cited-reference list itself is taken verbatim from the patent's own "Patent Citations (13)" table.

Critical-date / §102 framework. US 6,858,881 claims priority to 2000-07-06 (DE 10032838.5; PCT/DE2001/001952 filed 2001-05-22; US continuation filed 2003-01-06). For the §102 analysis below, the operative reference date is the 2000-07-06 priority. A reference is only an anticipation candidate under §102 if it predates that date and discloses every element of a claim, arranged as claimed. I flag each reference accordingly. (Note: this is a pre-AIA patent; §102(a)/(b)/(e) apply.)

---

1. The 13 cited references (as listed by the patent)

---

2. Tier-1 references — highest anticipation potential

(a) JP H10-326910 A — Song-Jae Lee (the closest art)

Full citation: Japanese Unexamined Patent Application JP H10-326910 A; appl. JP 10-129000; filed 1998-05-12; published 1998-12-08; inventor/applicant Song-Jae Lee (KR 245192 B1 family member). Source: J-GLOBAL record (jglobal.jst.go.jp, JGLOBAL_ID 200903088543710571).

Brief description: The abstract states the LED's body is "a triangular-prism shape, or a quadrilateral-prism shape having at least one acute angle" ("三角柱型または少なくとも一つの鋭角を有する四辺柱型のチップ"), in which total internal reflection inside the chip is suppressed so photons generated in the active region are maximally emitted. The claim set recites a body containing a crystalline layer with an active region, having many wall faces, a top face and a bottom face, with a triangular cross-section in the direction parallel to the top/bottom faces.

§102 relevance — the single most relevant reference.

• A prism has a constant cross-section and, by definition, parallelogram lateral faces. A "quadrilateral-prism having at least one acute angle" therefore discloses a body whose side surfaces are parallelograms meeting at an acute angle — i.e., squarely within the language of claim 1 ("at least two of said side surfaces … disposed at an acute angle, being tilted and forming parallelograms"), and it inherently has the uniform cross-section the patent touts.
• This is the reference the patent itself names and tries to distinguish in its Background of the Invention. The reference accordingly is the strongest §102 anticipation candidate for claim 1 (and, to the extent it teaches separating prism-shaped chips, for the method of claim 6).
• Flagged discrepancy (important): the patent's Background characterizes JP H10-326910 as a "truncated pyramid" with a "diamond-shaped" base (i.e., varying cross-section), whereas the reference's own abstract describes a prism (constant cross-section) with a triangular or acute-angle quadrilateral cross-section. These two characterizations are not the same thing, and the distinction matters: a truncated pyramid has trapezoidal angled faces (varying cross-section), while a prism has parallelogram faces (constant cross-section). If the reference in fact discloses an acute-angle prism, the patent's stated point of novelty (constant cross-section + parallelogram faces) is directly met and claim 1 is anticipated. This contradiction between the patent's characterization and the reference's abstract should be resolved against the actual JP text before relying on it.

(b) US 5,087,949 A (Hewlett-Packard / Haitz) and EP 0 405 757 A2 (same family)

Full citations & dates: US 5,087,949 A, priority 1989-06-27, published 1992-02-11, Hewlett-Packard (inventor Roland H. Haitz). EP 0 405 757 A2, appl. 90305903.8, filed 1990-05-31, priority 1989-06-27 (US 372,275), published 1991-01-02. (Google Patents / EPO publication server.)

Brief description: A transparent LED formed from a body with a flat back face, side faces perpendicular to the back face, surmounted by a truncated rectangular pyramid with diagonal faces, the LED truncation being a front face parallel to the back face. Each diagonal face is tilted relative to the side face at an angle of roughly twice the critical angle for total internal reflection; results of 1.5–2× light-output improvement are reported. The EP counterpart (same disclosure, Haitz) adds that the diagonal faces are made "by sawing V-shaped grooves in the front face of a wafer."

§102 relevance.

• These references disclose tilted faces meeting the base at an acute angle on a radiation-emitting chip — the general "angled-facet" concept of claim 1. However, a truncated pyramid's diagonal faces are trapezoids, not parallelograms, and the cross-section varies from base to top (this is precisely the deficiency the patent criticizes). They therefore do not appear to disclose every element of claim 1, and are best treated as §103 obviousness references, not true anticipations.
• EP 0 405 757 A2's V-groove sawing is materially relevant to the method side — claims 6, 11 and 12 (separating along an oblique surface / oblique blade), though the reference saws V-grooves to form facets on a pyramid, not to separate oblique parallelepiped chips.

(c) US 6,229,160 B1 (Lumileds) — with its HP/Lumileds siblings DE 198 07 758 A1 and GB 2 326 023 A

Full citations & dates: US 6,229,160 B1, appl. 08/868,009, filed 1997-06-03, priority 1997-06-03, published 2001-05-08, Lumileds Lighting U.S., LLC (inventors Krames, Kish, Tan). DE 198 07 758 A1 (1997-06-03 / 1998-12-10) and GB 2 326 023 A (1997-06-03 / 1998-12-09) are Hewlett-Packard counterparts of the same chip-shaping case. (US 6,229,160 is the parent of the continuation US 6,570,190 B2, "LED having angled sides for increased side light extraction.")

Brief description: A method of designing/fabricating LEDs in which the side surfaces (non-parallel to the epitaxial layers) are formed at preferred oblique angles relative to vertical to improve light extraction; the oblique angle is selected to increase light escaping through the (truncated-pyramid) frustum. The '160 disclosure explicitly discusses and criticizes U.S. 5,087,949 and describes a truncated pyramid / shaped sidewall ("oblique angle"), with claims directed to oblique angle <50° and area-yield limits.

§102 relevance.

• Directly on point for claim 1 ("at least two side surfaces … disposed at an acute angle, being tilted") and for the method of claim 6 (forming chips with oblique side surfaces). Like the HP pyramid art, however, the '160 sidewalls are the faces of a truncated pyramid/frustum (trapezoids, varying cross-section), not parallelogram side surfaces of a constant-cross-section parallelepiped. §103 relevance is strong; strict §102 anticipation of claim 1 is not made out on the parallelogram limitation.
• DE 198 07 758 A1 and GB 2 326 023 A being same-date counterparts, they add no independent teaching beyond US 6,229,160.

---

3. Tier-2 references — support/obviousness

(d) US 5,990,497 A (Toshiba)

Priority 1996-08-30; granted 1999-11-22; appln. 08/920,003. Abstract: a scribed pellet is produced from a wafer in which a semiconductor layer (incl. a luminous layer) is stacked on a compound-semiconductor substrate inclined at 5°–20° to a (100) surface in the [011] orientation, separated by scribing. Relevance: the off-axis substrate / scribe-and-break separation teaching is pertinent to the production method (claim 6 and its separation sub-features). The "inclined" aspect here is the substrate crystal orientation, not the chip side-surface angle; not an anticipation of claim 1.

(e) US 5,869,848 A (Oki Electric)

Priority 1995-12-14; granted 1999-02-09. "End face light-emitting-type LED… and methods of manufacturing them." Relevance: side/end-face light-emission geometry and element separation methods; background-to-method only. I could not retrieve its full abstract, so treat the specific mapping as unverified.

(f) US 5,864,171 A (Toshiba)

Priority 1995-03-30; granted 1999-01-26. "Semiconductor optoelectric device and method of manufacturing the same." Relevance: optoelectric device with engineered layer/geometry (a Toshiba shaping/processing case). I could not retrieve the abstract; a precise §102 mapping cannot be given. Likely a §103/background reference.

(g) JP H04-78174 A (NEC)

Priority 1990-07-19; published 1992-03-12. "Semiconductor light emitting element." Relevance: early shaped-LED art. I could not verify its detailed disclosure; likely background/§103 only.

(h) EP 0 961 328 A2 (Sumitomo Electric)

Priority 1998-05-28; published 1999-12-01. "Gallium nitride-type semiconductor device." Disclosure concerns GaN/sapphire heteroepitaxy, the non-cleavage of sapphire, the difficulty of dicing sapphire with a mechanical cutter/rolling blade, and the desirability of GaN or SiC substrates with cleavage for chip separation. Relevance: it is cited chiefly for the substrate-material and dicing aspects underpinning claims 5/10 (sapphire, GaN, SiC…) and the separation context of the method claims. Not an anticipation of claim 1.

(i) US 5,341,024 A (LSI Logic)

Priority 1992-07-17; granted 1994-08-23. "Method of increasing the layout efficiency of dies on a wafer, and increasing the ratio of I/O area to active area per die." Relevance: wafer die layout/tessellation geometry — consistent with a non-rectangular (parallelogram) die footprint tiled across a wafer. It is not a radiation-emitting device and does not disclose an active layer, so it cannot anticipate claim 1; it is at most background for the die-shape/manufacture notion in claim 6, and even there only peripherally.

---

4. Reference that is NOT anticipatory prior art (post-dates the patent)

US 2003/0137031 A1 — Tai-Fa Young

Priority 2002-01-23; published 2003-07-24. "Semiconductor device having a die with a rhombic shape." This reference post-dates the 2000-07-06 priority of US 6,858,881 and therefore cannot be §102 prior art against it. Its appearance in the citation table is anomalous for an anticipation analysis (a rhombic/parallelogram die is, however, conceptually aligned with the patented shape). It should be excluded from any §102 ground.

---

5. §102 bottom line — which claims each reference could plausibly anticipate

Most relevant prior art, ranked: (1) JP H10-326910 A (Song-Jae Lee) — the closest art and the only reference whose abstract appears to disclose an acute-angle prism with parallelogram side surfaces (claim 1 / possibly claim 6); (2) US 5,087,949 A & EP 0 405 757 A2 (HP/Haitz) for tilted acute-angle facets and V-groove sawing; (3) US 6,229,160 B1 (Lumileds) and its DE/GB counterparts for oblique side-surface chip shaping.

---

6. Non-patent citation

Song, J.L. et al., "Efficiency Improvement in Light-Emitting Diodes Based on Geometrically Deformed Chips," SPIE Conference on Light-Emitting Diodes, San Jose, 1999 — the patent's sole listed NPL reference. Given the shared author name with JP H10-326910 (Song-Jae Lee), this paper is presumably the printed-publication counterpart of the same "geometrically deformed chip" work and is likewise highly relevant §102/§103 art to claim 1; I did not independently retrieve its full text, so its exact disclosure mapping remains unverified.

---

7. Limitations and uncertainty (stated explicitly)

• Bibliographic fields are per the Google Patents mirror of the USPTO record; I did not reach USPTO Patent Center directly. The current assignee line ("AMS OSRAM International GmbH") and the listed claim count (12; independents 1 and 6) carry the record's own caveats.
• Not independently verified (details could not be retrieved): JP H04-78174 A, US 5,341,024 A, US 5,864,171 A, US 5,869,848 A, GB 2 326 023 A, DE 198 07 758 A1, US 2003/0137031 A1. My characterizations of these rest on their titles/priority data and the patent's own citation table, not on their full texts.
• Verified from live sources: US 5,087,949 A (Google Patents; uspto.report); EP 0 405 757 A2 (EPO publication server); US 6,229,160 B1 and US 6,570,190 B2 (Google Patents / patentimages); JP H10-326910 A abstract (J-GLOBAL); US 5,990,497 A abstract (Unified Patents portal); EP 0 961 328 A2 (Google Patents).
• Key contradiction to resolve: the subject patent's Background calls JP H10-326910 a truncated pyramid with a diamond base, while the JP abstract describes an acute-angle prism. Because the patent's asserted novelty is a constant cross-section with parallelogram side surfaces, this discrepancy is decisive for any claim 1 anticipation question and should be checked against the JP specification before being relied on.
• Nothing here is a legal conclusion. "Anticipates" is used in the hypothetical §102 sense; the granted claims presumably survived these references during prosecution, so the items above are candidates for re-examination/validity analysis, not adjudicated anticipations.

I'll run a few targeted verification searches on the key references before building the § 103 combinations.

Obviousness Analysis — U.S. Patent No. 6,858,881 under 35 U.S.C. § 103

Building on: the Prior-Art section above (reference list, dates, and the flagged Lee "prism vs. truncated pyramid" contradiction). This section does not re-list the references; it combines them. One new, decisive piece of evidence surfaced in verification and is folded in below.

---

0. Threshold matters

• Governing law: pre-AIA § 103 (priority 2000-07-06; filed 2003-01-06), so the obviousness standard is KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007), applied through the Graham v. John Deere, 383 U.S. 1 (1966), factors: scope/content of the prior art, differences from the claims, PHOSITA level, and secondary considerations. MPEP § 2143 supplies the enumerated rationales.
• Prior-art status: the Song/Lee SPIE paper (published 1999) and the HP patents are § 102(b) art (more than one year before the 2000-05-22 critical date, measured from the PCT filing that the '881 continuation relies on); the remaining references are § 102(a)/(e) art.
• Posture caveat: all of the strongest references below were of record before the examiner (citations 1, 2, 8, 9, 10, 11, 12 plus the sole NPL). Any re-examination or PTAB challenge built on them faces § 325(d) discretionary denial; district court is the natural forum. But the obviousness merits are independent of that procedural point, and the merits are strong — because the examiner appears to have misapprehended the teaching of the single most important reference.

---

1. Level of ordinary skill (PHOSITA)

A person with a B.S. in electrical engineering, applied physics, or materials science, plus ~2–5 years in III-V optoelectronic device design and LED chip fabrication/dicing, familiar with Snell's law, total internal reflection and escape-cone geometry, and with wafer sawing/scribing/beveling. The '881 disclosure itself is pitched at exactly this level (it recites standard TIR relations and calls the sawing "already known").

---

2. The dispositive limitation, construed

Claim 1 requires: "at least two of said side surfaces being disposed at an acute angle, being tilted and forming parallelograms." The specification defines its two geometry parameters in FIG. 2/FIG. 3 and the text:

• φ (horizontal base angle) — the in-plane corner angle of the parallelogram cross-section; a "tilted parallelogram with an acute angle φ."
• θ (vertical base angle) — the tilt of the side surfaces from the normal; "tilted parallelograms with an acute angle θ."

So the claim collapses to a single geometric idea: a skewed, constant-cross-section parallelepiped whose side facets are parallelograms with acute corners. Construed most naturally, claim 1 needs (a) parallelogram-shaped side surfaces, and (b) those surfaces tilted at an acute angle (θ<90° and/or φ<90°). Nothing in claim 1 requires an unslanted companion face, a particular substrate, or any specific angle value (that is claim 4). This is the broadest reasonable reading and it is the reading the claimed subject matter demands — which matters, because the closest art meets it head-on.

---

3. Decisive new evidence: the cited NPL actually discloses the claimed geometry

The patent's only non-patent citation is listed as "Song, J.L. et al., Efficiency Improvement in Light-Emitting Diodes Based on Geometrically Deformed Chips, SPIE Conference on Light-Emitting Diodes, San Jose, 1999." The actual paper — Lee, Song Jae & Song, Seok Won, Proc. SPIE 3621, pp. 237–248 (1999), DOI 10.1117/12.344484 (the patent's "Song, J.L." appears to be a garbled form of "S.J. Lee") — is retrieved and its abstract states:

"We propose LEDs based on geometrically deformed chips of which the horizontal cross-section is either rhomboidal or triangular and, in addition, side walls may also be slanted. In such deformed chips, the photon trajectory changes with each reflection off the wall and, as a result, the continued total internal reflections as observed in conventional rectangular cubic chips are suppressed."
— Semantic Scholar / SPIE Proc. 3621:237–248

This is not background art. It is a single-reference disclosure of both claim-1 elements at once: a rhomboidal (i.e., parallelogram, acute-angled) horizontal cross-section = the φ feature, and slanted side walls = the θ feature. A rhomboidal-cross-section chip with slanted side walls is a skewed parallelepiped — exactly the claim-1 body. This resolves the contradiction the Prior-Art section flagged: the patent's Background characterizes JP H10-326910 as a "truncated pyramid with a diamond base," but the Lee/Song work is expressly a deformed (rhomboid/triangular) cross-section with slanted walls, i.e. the constant-cross-section geometry, not a frustum. The patent's characterization of its own closest art appears to be inaccurate, and that is why the case for obviousness (and, a fortiori, anticipation) is strong.

---

4. Ground 1 — Claim 1 obvious over Lee/Song (SPIE 1999), alone or with JP H10-326910

Disclosure mapping (Lee/Song, SPIE 3621:237–248; and its JP sibling H10-326910):

Why obvious, not merely "close": even setting anticipation aside, the Lee/Song paper proposes the very combination — it teaches that deforming the cross-section to rhomboidal and slanting the walls both change the photon trajectory and suppress sustained TIR. A PHOSITA reading it would adopt both teachings together because the paper presents them as the same solution to the same problem (extraction from a cubic chip). Under KSR, "combining two disclosed techniques in the manner claimed" where the reference itself suggests the combination is the paradigm of obviousness. There is no teaching away: nothing in Lee/Song or JP H10-326910 disparages slanted walls or parallelogram cross-sections.

Practical note: because the SPIE paper is the applicant's own cited NPL, the strongest examination-side failure is the apparent failure to appreciate that it teaches the slanted side walls in combination with the rhomboidal cross-section, not merely a deformed top face.

---

5. Ground 2 — Claim 1 obvious over HP '949 / EP 0 405 757 (Haitz) in view of Lee/Song or LSI Logic '024

If a tribunal reads claim 1 to require an unbent, constant-cross-section body and the parallelogram facet, the alternative combination is:

• Primary — US 5,087,949 (Haitz, HP) / EP 0 405 757 A2 (same family): a transparent LED with side faces perpendicular to the back face surmounted by tilted "diagonal faces," tilted "in the order of twice the critical angle," with tested results at 45° and a range of 30°–60° from the side face, yielding 1.5×–2× output. This supplies tilted facets at an acute angle for light extraction.
• Secondary — Lee/Song (SPIE 1999) and, independently, US 5,341,024 (LSI Logic): a parallelogram/rhomboidal die footprint (LSI '024 teaches parallelogram dies to raise layout efficiency and area yield). This supplies parallelogram side surfaces and the constant-cross-section body.

Motivation to combine (the heart of the § 103 case):

1. Same field, same problem. All three are LED chip-shaping references directed to increasing extraction efficiency (MPEP § 2143.01(IV); In re Keller). Analogous art, no field-of-invention dispute.
2. The combination is a substitution of known elements with predictable results. Taking the known "tilt the side facets to get light out" technique (HP '949) and applying it to a known constant-cross-section parallelogram die (LSI '024 / Lee) yields a skewed parallelepiped — a merely mechanical rearrangement of known facets. KSR slide rule/tea-kettle reasoning.
3. Explicit design incentive the patent itself concedes. The '881 specification states the only recognized drawback of the HP-type pyramid is its varying cross-section ("harder for heat to be dissipated… current flow is impeded… resistance locally increased"). That concession both (i) identifies the problem and (ii) supplies the motivation to use a constant-cross-section parallelogram body. A reference need not solve the very problem to supply motivation where the applicant's own specification frames the design goal this way.
4. Wafer economics / area yield. LSI '024 teaches parallelogram dies as a layout solution, and Lumileds' own later family (US 7,268,371) frames oblique-walled chips in terms of "area yield." A parallelogram footprint tiles a wafer with minimal kerf loss — a strong, expressly stated incentive the '881 specification echoes when it praises sawing that causes "no loss of active surface area or of the substrate."
5. Finite, predictable parameter space. Choosing the tilt angle and the cross-section angle from a small set of known values (HP's 30–60°; Lumileds' <50°) is "obvious to try" with a reasonable expectation of success — and the '881 FIG. 5 shows the efficiency varies smoothly and monotonically with the angle, confirming the result was predictable, not surprising.

---

6. Ground 3 — Method claims 6–12 obvious over Lee/Song + HP V-groove sawing + known dicing

Claim 6 requires only: provide substrate → form active layer above a surface → separate substrate + active layer so the resulting chips have the tilted-parallelogram side surfaces. This is squarely obvious:

• Lee/Song & JP H10-326910 disclose the deformed-chip product and its fabrication.
• US 5,087,949 / EP 0 405 757 disclose the separation technique: sawing V-shaped grooves with a dicing saw "dressed with a V-shaped face" to form the tilted facets, then cleaving or a second pass — i.e., separating along an oblique surface with an oblique blade (claims 11–12).
• US 5,990,497 (Toshiba) and EP 0 961 328 (Sumitomo) supply the ordinary off-axis substrate/scribe-break and dicing context.

Motivation: the '881 specification concedes it outright — "sawing methods such as these are already known in conjunction with the production of conventional light-emitting diodes… and can essentially be transferred, for which purpose all that is necessary is to modify the inclination of the saw blade." That sentence is an admission that the claimed method is the known dicing process plus an obvious parametric change — the classic case for obviousness. Claim 11's "separating surface running obliquely" and claim 12's "oblique saw blade" are met by EP '405757's V-groove dressing and Lumileds' "beveled sawing."

---

7. Claim-by-claim disposition

No claim adds a limitation that is not either disclosed or a predictable design choice.

---

8. Synthesis of the KSR motivations (consolidated)

---

9. Secondary considerations — none of record, and the "unexpected results" argument fails

• No objective evidence is in the file: no commercial-success, long-felt-need, failure-of-others, or copying evidence, and no nexus evidence tying any commercial result to the claimed geometry (the specification reports OSRAM's own modeled data, not market data).
• The headline "50%" figure undercuts, rather than supports, non-obviousness. The '881 specification presents the 25%→50% output improvement as flowing directly from ray optics ("the incidence angle is less than the critical angle… the spatial angle… is increased"), corroborated by a smooth FIG. 5 curve. A predictable result is the antithesis of an unexpected one. Notably, HP '949 had already reported the same 1.5×–2× doubling nine years earlier from tilted facets — so the magnitude was known and expected.
• Copying / industry uptake is against the patentee. The art the examiner cited (Hewlett-Packard, Lumileds/Krames-Kish-Tan's inverted-truncated-pyramid, Lee/Song) shows the industry moving in the same direction independently. That is evidence of an obvious technical trajectory, not of the patent's inventive contribution.
• Non-assertion/expiration (per the Assignment and PTAB sections: expired 2021-05-22, never asserted): while not itself a Graham factor, the absence of any demonstrated commercial appropriation of the claimed geometry is consistent with the geometry having become ordinary.

---

10. Strongest patentee rebuttals — and why they are weak

1. "Parallelogram" vs. "trapezoid." The patentee will argue HP '949/Lumileds disclose trapezoidal frustum facets, not parallelograms. Response: Lee/Song (the cited NPL) expressly discloses a rhomboidal cross-section and slanted walls — parallelogram facets meeting the limitation without any combination at all.
2. "The references teach a varying cross-section (frustum), which my invention avoids." Response: that is precisely the motivation to substitute a constant-cross-section body; a need identified in the applicant's own Background cannot supply patentability. And Lee/Song's prism is constant-cross-section.
3. "Unexpected 2× efficiency." Response: predictable from TIR geometry and already reported by HP '949.
4. Teaching away. None found in any reference; HP '949 and Lumileds '160 affirmatively teach angling the side surfaces, and Lee/Song affirmatively teach slanting them.
5. § 325(d) (procedural). Real, but procedural: it argues forum, not validity. It will not save the claims in district court.

Residual uncertainty (stated honestly): I have not read the full text of JP H10-326910 or the full SPIE paper; the claim-1 mapping rests on their published abstracts and the specification's own discussion. The authorship discrepancy ("Song, J.L." in the patent vs. S.J. Lee / S.W. Song in the paper) should be confirmed against the SPIE proceedings, and the JP full text should be checked to confirm the "slanted side walls" embodiment is claimed, not merely mentioned. These go to strength of proof, not to the existence of the teaching.

---

11. Bottom line

The claims of US 6,858,881 are, on this record, invalid as obvious under § 103 — and one cited reference may go further and anticipate.

• The single most probative reference is the patent's own cited non-patent literature (Lee & Song, Proc. SPIE 3621:237–248, 1999), which discloses a rhomboidal-cross-section LED chip with slanted side walls — i.e., both claim-1 limitations in one document. This appears to be the teaching the examiner failed to appreciate, and it is why the allowance was likely in error.
• Even without it, the combination of HP '949 / EP 0 405 757 (tilted facets at acute angles; V-groove oblique sawing) with Lee/Song or LSI Logic '024 (parallelogram die/cross-section) renders claims 1–5 obvious, with the motivation supplied by the patent's own conceded problem (varying cross-section) and by ordinary wafer-area-yield and thermal-management design incentives.
• The method claims 6–12 are obvious over the same references, and are admitted obvious by the specification's statement that the known sawing process "can essentially be transferred, for which purpose all that is necessary is to modify the inclination of the saw blade."
• No secondary consideration of record rebuts the prima facie case; the specification's smooth, ray-optics-explained efficiency curve affirmatively supports predictability.

One reconciliation to record for the file: the Prior-Art section flagged a contradiction between the patent's characterization of JP H10-326910 ("truncated pyramid with a diamond base") and the reference's actual "acute-angle prism/deformed chip." The Lee/Song SPIE paper's express "rhomboidal cross-section … side walls may also be slanted" language resolves it in favor of the reference: the art teaches a constant-cross-section skewed chip, which is the very geometry the '881 patent claims. That is the finding on which the § 103 case principally rests.