@@ -152,7 +152,7 @@ explicitly **unmeasured**.
 | palette256 = one PQ subspace's 256 centroids | **[G]** | `nsm_word.rs`: CAM codebook = **6 subspaces × 256 centroids**; `cam_codes.bin` = N words × 6 bytes (lance‑graph PR #477) |
 | palette256 ↔ indexed‑color codec palette | **[G]** | HEVC‑SCC + VVC ship an indexed‑palette mode (the codec's own ≤‑256‑ish codebook for screen content) |
 | palette256 ↔ OLED subpixel emission (layout) | **[S]** | OLED PenTile RGBG layout is a palette‑on‑a‑lattice for *perceived* resolution; the subpixel *layout* is shape‑match only and already covered by Morton tiling. Kept [S] but **superseded by the exciton leg below** as the OLED→substrate connection worth taking. |
-| OLED **excitons** ↔ **irrational bundling under bias** | **[H — promoted from prior [S], 2026‑06‑08]** | **Operator: "isn't OLED excitons also some irrational bundling?"** — yes, in the sense the substrate cares about. Measured (web sources, see §11): (1) the rational 1:3 singlet:triplet ratio is quantum spin statistics, **but** under bias the formation ratio *deviates* — *"the generation efficiency of singlets scales with the bias, whereas that of triplets is nearly bias‑independent"* (PMC4614446), giving a **continuously‑variable non‑integer S:T ratio per operating point**. (2) Singlet/triplet binding energies differ irreducibly: *"~0.5 eV (singlet) vs up to 1.5 eV (triplet)"* (noctiluca / ScienceDirect S0927796X22000286). So OLED ships, under bias, **the same discrete coprime aperiodicity D‑BGZ17 names**: non‑commensurate spin populations + non‑commensurate binding energies, broken from the rational integer ratio by a continuous control parameter. The shape‑match is structural, not analogical: a substrate that wants aperiodicity *at every quantized level* (D‑QUANTGATE) has a literal physical precedent in OLED exciton populations *exactly because* they refuse to sit on integer ratios under operation. The recommendation from the prior version ("do not build on") **is withdrawn for the exciton leg specifically** — it remains for the *subpixel layout* leg above. |
+| OLED **excitons** ↔ **irrational bundling under bias** | **[S — REVERTED from [H], 2026‑06‑09 by the 5+3 agent review]** (mirror of the DISCOVERY‑MAP D‑EXCITON revert, PR #48) | The 2026‑06‑08 `[H]` promotion below this line does **not** survive review — kept for the record, struck in substance. **(1) Category error (theorem‑checker):** a bias‑shifted *scalar* ratio and two energy *scalars* have no period to be aperiodic about — aperiodicity is a property of an infinite tiling/sequence, not of points in ℝ; "the same discrete coprime aperiodicity D‑BGZ17 names" is MERE‑RHYME: D‑BGZ17's beat‑272 comes from two coprime *integer cycles* in an address lattice (CRT), and OLED physics offers no integer cycles and no lattice. **(2) Number mis‑label (arxiv‑grounder):** the "~1.5 eV triplet *binding* energy" is the **T1 excited‑state energy mis‑labelled** — literature triplet E_b is sub‑eV (≈0.3–0.6 eV, arXiv 2311.03927); corrected, the ratio is order ~1, not 3, and the motivating "incommensurable" coincidence dissolves. **Kept (`[G]`‑real, held apart from the dead analogy):** the exactly‑rational 1:3 (SU(2) microstate count — definitional); the bias‑dependence observation (PMC4614446 — real but material‑specific, carries no aperiodicity); the ~100% IQE harvest physics (Baldo/Adachi; Uoyama 2012). **"Do not build on" is RESTORED** for this leg. The one untested survivor lives as **D‑LOSSCHAN `[S]`** (loss‑channel suppression ↔ cascade early‑exit, Stern‑Volmer form) in DISCOVERY‑MAP §2.7. |
 
 **The convergence number is 256 = 2⁸ = one byte.** PQ centroids, codec
 palette indices, attention weight buckets (§4), and Binary16K lane
@@ -494,14 +494,18 @@ Ordered by leverage (highest first):
   by Z‑order/Hilbert row ordering. Lance (the substrate's columnar
   instance) — fragment/page random‑access columnar format, chosen over
   classic Parquet row‑groups for random tile access.
-- OLED exciton physics (§3 promoted leg): *"Direct monitoring of bias‑
-  dependent variations in the exciton formation ratio of working organic
-  light emitting diodes,"* PMC4614446 (singlets scale with bias; triplets
-  nearly bias‑independent → continuous non‑integer S:T ratio). *"Excitons
-  in OLEDs Explained,"* noctiluca.eu (singlet ≈ 0.5 eV vs triplet up to
-  1.5 eV binding energies). *"Exciplexes in OLEDs: Principles and
-  promises,"* ScienceDirect S0927796X22000286 (binding‑energy and
-  formation‑ratio review).
+- OLED exciton physics (§3 leg — **REVERTED to [S], 2026‑06‑09**):
+  PMC4614446 (bias‑dependent S:T formation ratio — real but
+  material‑specific; a scalar, carries no aperiodicity). The
+  "~1.5 eV triplet *binding* energy" previously cited from
+  noctiluca.eu / ScienceDirect S0927796X22000286 was the **T1
+  excited‑state energy mis‑labelled** — literature triplet E_b is
+  sub‑eV (≈0.3–0.6 eV, **arXiv 2311.03927**); a 1.5 eV triplet E_b
+  would make organic photovoltaics energetically impossible. Kept
+  `[G]`‑real: the rational 1:3 (SU(2), definitional) and the ~100% IQE
+  harvest physics (Baldo/Adachi *J. Appl. Phys.* 90:5048, 2001;
+  Uoyama/Adachi *Nature* 492:234, 2012). Survivor: D‑LOSSCHAN `[S]`
+  (DISCOVERY‑MAP §2.7).
 
 ---