Foundation II: The Chiral Universe and the Three Fossils of the Bounce — Baryogenesis, the Geometric Dark Sector, and Cosmic Alignment within an Einstein-Cartan Framework, Extended Version

Foundation II develops the three observable fossil signatures of the primordial spin-torsion bounce, giving geometric accounts of the cosmological dark sector — no new particles, no ad hoc scalar fields. The three fossils. Chiral baryogenesis (t ~ 10−11 s): η ≈ 5.9×10−10, within 3.5% of the Planck-inferred value, from a 5:1 ratio set by 20:4 torsion degrees of freedom. Semi-derived: two of its three stages follow from first principles, the third is calibrated on the Foundation I background. Geometric dark matter (t ~ 1 s): χ˜2med = 0.80 against 1.55 for NFW, on 171 of the 175 SPARC galaxies, with two geometric parameters per galaxy and no free dark-sector parameter. Cosmic alignment (t ≈ 343,000 yr): a Torsion Vortex Ring of radius 322 Mpc, matching the Big Ring and the Giant Arc. Topological zoology. Micro-Knots (M = 3.74×1024 kg, the horizon mass at the electroweak crossing), Macro-Knots (~105 M☉, Little-Red-Dot seeds), Chiral Walls (void boundaries) and the Torsion Fluid (dark-energy condensate). ⚠️ The 1D sector is carried as an open problem (PO-F2-8): torsion strings have no established Lagrangian foundation in the present field content, Sμ being a real axial vector with no internal phase. Version notes — PREPRINT v4 (15 September 2026) Eleven corrections. No result is withdrawn; two are weakened, and three touch a published value. Correction of record — DESI DR2 significance. Earlier versions stated 4.9σ for the DESI DR2 preference for dynamical dark energy. That value appears nowhere in arXiv:2503.14738, which reports 3.1σ (BAO + CMB) and 2.8–4.2σ with supernovae. The 4.9σ was the distance of the ΛCDM point from the DESI central values in the marginalised plane — computed here, never published by DESI. The distance is kept where it belongs, now explicitly distinguished from the reported significance. Corrected in the text and in the figure legend. Correction of record — Macro-Knot population. The collision-parameter section quoted Nmacro ~ 103 per halo, stated without derivation and inconsistent by two to three orders of magnitude with the fmacro ~ 10−7–10−6 derived elsewhere in the same paper. The derived value stands: Nmacro = O(1–10), and ΓM(0) scales down to ≤ 2.5×10−29 s−1. Correction of record — dynamical friction. The same section claimed Macro-Knots sink to galactic centres in ≲1 Gyr. Chandrasekhar friction gives 5,851 Gyr at 10 kpc for 105 M☉. Macro-Knots found in galactic nuclei are there because they formed there, not because they migrated. Correction of record — Israel repulsion, dimensional. The acceleration was written z̈ = (8πGσ2D/3)z, applying the Friedmann coefficient to a surface density: m2s−2 instead of m s−2. The effective volume density of a wall of causal thickness ct is ρwall = σ2D/(ct) — a conversion the paper already used a few lines below. Corrected in all eight occurrences. No numerical result depended on the missing factor. Correction of record — χ2 comparison. Four passages compared χ˜2med = 0.80 to 1.31, which is the NFW weighted reduced χ2 — a different statistic. The median-to-median comparison is 0.80 versus 1.55, and it is more favourable to the framework than the figure previously quoted. The table itself always carried both columns correctly. Correction of record — recombination age. 365,000 yr → 343,000 yr, evaluated at the Planck decoupling redshift z* = 1089.9 and reproducible by script. Neither the earlier 365,000 nor the ΛCDM 380,000 could be traced to the stated parameters; the recomputed ΛCDM value, 371,000 yr, matches Planck's own ~372,000. Reopened — PO-F2-6 (spectral status of the axial sector). An earlier conclusion declared it closed, on the strength of the Proca form being the only consistent kinetic structure for a propagating torsion. That closure is withdrawn: the cited work reports the result rather than deriving it, and selecting one form among two candidates is a different question from determining what the reduction of the full action produces. Revised bound — CMB quadrupole. The bound on bounce ellipticity rested on Qrms = 6 μK, untraceable to any standard convention. With the standard Qrms = √(5C2/4π) ≈ 9.5 μK, the bound becomes ε ≤ 1.2×10−7 instead of 7.8×10−8 — weaker. The earlier figure was conservative, not favourable. Clarification — why the mass spectrum is bimodal. A new paragraph states what was previously implicit: the two populations come from two distinct mechanisms, not two ends of one distribution. Poisson coagulation (λ ≥ 36) yields Mtyp ~ 7×10−5 M☉ with a tail exhausted near 2×10−4; Macro-Knots come instead from the Jeans mass at the Great-Annihilation horizon. Reaching 105 M☉ by coagulation would need 5×1010 successive mergers against ⟨N⟩ = 36. ⭐ The gap between the two populations is a range neither mechanism can populate — and the framework predicts it empty. Falsifiable: a primordial population between ~10−3 and ~104 M☉ would contradict it. Clarification — Great Annihilation temperatures. The text described 1.02 MeV as “the mass-energy equivalence of an electron”; it is that of an e+e− pair, 2mec2. A clarification now distinguishes the kinematic threshold (2mec2), the epoch marker usually quoted in the literature (mec2 = 0.511 MeV), and complete freeze-out near 16 keV. This work uses the window T ~ 0.16–1.6 MeV throughout: the Great Annihilation is a process, not an instant. Narrowed attributions. Fabbri: “reports”, not “demonstrates”. Carroll & Field: “akin to”, not “exactly the construction of”. Kibble–Zurek: the citation covers the mechanism only; the geometric chiral-splitting is this framework's own addition. Notation: _μK, which read as microkelvin but meant Micro-Knot, renamed _MK in 54 places. Figures regenerated. Three carried values already corrected in the text: the recombination age in the chronological roadmap and in the mass-ratio plot, and the 4.9σ in the (w0, wa) legend. Compilation verified: 156 pages · 0 undefined references · 0 dimensional flags. Part of the Einstein–Cartan Framework corpus: PIT Letter (10.5281/zenodo.19798923), Foundation I (10.5281/zenodo.19577447), Foundation III — Cosmological Physics (10.5281/zenodo.22359982). Reproducibility code: github.com/pfichant/spin-torsion-cosmology (CC-BY-4.0).