s1114 — writhe / non-monoidal build (count fixed)
  T₃ = 0.016261 √σ = 7.2363 MeV · R_α = 2.1694 fm (count) · r₀(light) = R_α/φ = 1.3408 fm
  data: deficit slope on the N=Z line n=16..25 = −0.632 T₃/cell = -4.573 MeV/cell (s1101 target line); s1102: a_eff = 0.303 ± 0.025 MeV (Coulomb shape, 19 rows)

(a) NEUTRAL PACKING WRITHE — sign test
  bonds(n) = 6n − 6n^(2/3): non-affine part second differences mean = +0.0601 bonds → CONCAVE (surface deficit)
  data residual (AME − ladder) is NEGATIVE and growing ⇒ observed binding FALLS BELOW the additive law: the non-additive term is a COST
  growing faster than linear ⇒ CONVEX cost. A surface deficit is a convex COST too (fewer bonds than 6n) — sign OK for shape,
  but its magnitude is bounded: max missing bonds fraction at n=25 = 0.34 of the bulk; data need −4.57 MeV/cell EXTRA per added cell at n≈20.
  ε needed per bond to reproduce the slope from surface deficit alone ≈ 17.62 MeV/bond vs B_α/6 = 4.72 MeV/bond (α-cell binding per bond)
  [PASS] neutral surface-writhe alone cannot carry the ⁶⁴Ge deficit: needed ε per bond exceeds the whole α binding per bond (ratio > 2)

(b) LONG-RANGE REPULSION (Coulomb) — α LOADED AS A DECLARED CONTROL ONLY (fenced in print)
  a_c = 3αħc/(5 r₀) with r₀ = 1.3408 fm (count-fixed, light) = 0.6444 MeV · bare SEMF 0.711 · s1102 fitted effective 0.303 ± 0.025
  [PASS] count-fixed Coulomb a_c / a_eff = 2.13 — OUTSIDE the magnitude-ratio bar [0.5, 2] ⇒ even with α loaded, the count radius does not price the curvature
  the screening the data demand: a_eff/a_c = 0.470  (fodder: φ⁻² = 0.382 · 1/2 = 0.5 · φ^(−3/2) = 0.486 · 3/7 = 0.429) — a menu of small ratios, not a derivation
  the surface term would be the natural screen (Coulomb DOWN, surface UP in curvature: s1102 post-hoc a_eff = 0.408 from the physical pair),
  but c1105/c1113 showed the coefficient-free surface+Coulomb pair at RMS 12.27 MeV — the screen is not the SEMF surface at its bare weight.

(c) WHAT THE WRITHE WOULD HAVE TO BE — the map, not the number
  Wr(n) := E_obs(n) − E_additive(n) − affine, in T₃: from the data, Wr'' ≈ Coulomb shape × 0.303 MeV. A writhe integral over the
  packing must return (i) a CONVEX n-profile with the Coulomb Z(Z−1)/A^{1/3} shape (it does, if the cells carry charge), (ii) the
  coefficient 0.303 MeV = 0.43 × bare. The 0.43 is the object: a SCREENING of the cell charge by the packing — in knot terms the
  self-linking of a charged α-cell worldline shares its flux with neighbours. Registered handle: p9's charge penalties are
  φ-monomial (no α); the required factor 0.43 vs bare would then read as φ-monomial × α-free 'penalty' matching αħc·0.43 — i.e.
  the writhe build needs the CHARGE MAGNITUDE sector (p9 'honestly the weakest') before the radius sector can be closed.

VERDICT
  Scoped NO-GO with the obstruction named. (a) neutral packing writhe: right sign, magnitude short by >2× the whole α bond energy.
  (b) charged packing with the count radius and α loaded: a_c/a_eff = 2.13, outside [0.5,2] — the count fixes the LENGTH but not the
  CURVATURE; a screening factor 0.43 is required and is not registered. The ⁶⁴Ge wall is therefore TWO walls: the radius (count:
  candidate map exists, s1113) and the charge magnitude (p9's weakest sector, no α in print). The writhe integral is the right
  TYPE of object; its coefficient lives in the charge sector. Route: the next PT dispatch pairs D1228's count with p9's charge-
  magnitude problem (P4 of the S292 plan) — they are the same obstruction seen from two sides.

s1114: ALL CHECKS PASS
