Stellar Boundary Dynamics

Every object across all three physical layers — pre-collapse stellar interior (A), post-collapse brightness response (B), and post-collapse spectral evolution (C) — returned FULL_PERCOLATION with giantRatio = 1.000 and isolated = 0. This constitutes the first UNNS corpus entry for catastrophic stellar boundary events. The complete absence of HARD, GIANT, or TAIL outcomes under core-collapse supernova conditions is consistent with the Margin-Confinement Law prediction that identity-preserving physical evolution cannot persistently cross ∂ℳ_adm, even under extreme forcing, and provides pilot-corpus support for that prediction in the astrophysical regime.

Both pre-supernova radial profiles connect at κ_connect ≤ 2, placing them in the same tier as biological fitness landscapes (κ = 0.42–2.00) and atmospheric wind fields in the 14-domain STRUC-PERC baseline. Zero tail dominance indicates no extreme outlier gaps in the radial structure. The terminal stellar interior, despite its extreme physical conditions (ρ_c ≈ 10¹⁰ g cm⁻³, T_c ≈ 10¹⁰ K), forms a locally homogeneous structural ladder — consistent with the USL prediction that admissible systems occupy the interior of ℳ_adm regardless of physical forcing scale.

Post-collapse brightness trajectories, after compression into phase-gap ladders, form connected vulnerability-graph structures under STRUC-PERC-I. This extends the SN Ia finding from Beyond Fragmentation (where Δ-layer recovery from HARD was demonstrated) to Type II core-collapse events: even without explicit Δ-lifting, the phase-compressed multi-band light curves achieve FULL connectivity. The result supports treating supernova light curves as boundary-response relaxation trajectories within the UNNS / STRUC-PERC workflow.

The SN2012aw spectral line-window ladder (n = 198, 84 spectra across 6 years) is FULL but extreme. Its tail dominance of 0.959 sets the record in this dataset and exceeds any STRUC-I physical domain result (solar flare max ρ̄ = 0.393). Its κ_connect of 3992 places it between the condensed matter upper range (~8100) and the nuclear spectral lower range (~3.8×10⁴), consistent with spectral features having larger gap heterogeneity than photometric features. The extremely high tail dominance is physically interpretable: the Ni/Co decay proxy and late-nebular iron emission lines produce extreme outlier gaps in a line-window sorted by wavelength, driving the tail fraction near unity while the backbone remains connected.

The pre-collapse and post-collapse domains are not identical but they are not strongly separated. The 20 M☉ profile (A2_20M) acts as the structural attractor for the majority of post-collapse light curves: SN1993J (d = 0.233, contact), SN2011dh (d = 0.575), SN2013ej (d = 0.627), and SN2012aw (d = 1.879, still closest to A2 despite being the outlier). This is structurally interpretable: the 20 M☉ progenitor develops a larger, more extended envelope that maps onto the heterogeneous multi-band response trajectories of the Type IIb and Type II objects. The 12 M☉ profile attracts the IIP plateau-dominated objects (SN1999em, SN1987A), consistent with a more compact pre-SN structure.

The centroid separation doubles from A–B (0.584) to B–C (1.318). This is the primary cross-layer finding: the photometric and spectroscopic representations of the same ejecta are structurally distinct in 5D vector space, with the tail dominance shift (+0.740) being the dominant driver. SN1993J behaves coherently across B and C (its matched pair is its nearest C neighbor), validating the contact-case hypothesis from the A–B bridge. SN2012aw does not: its matched C counterpart is less structurally similar to it than C1_SN1993J is, indicating that SN2012aw's photometric and spectroscopic structures are themselves strongly separated — consistent with its persistent anomaly status.

SN2012aw satisfies FULL_PERCOLATION in every phase, confirming the Margin-Confinement Law's prediction that even the most extreme astrophysical systems do not cross ∂ℳ_adm. However, its tail dominance escalation across layers (0 → 0.329 → 0.959), its light-curve coverage (21 photometric bands), its spectral persistence (6 years of spectra), and its BC-bridge incoherence (photometric and spectroscopic representations are themselves structurally distinct) collectively identify it as the strongest Forced Coherent Collapse (FCC) candidate in this corpus — a system that approaches the admissibility boundary under persistent extreme forcing while maintaining the giant component. Its status is consistent with the Beyond Fragmentation Unified Tail Dominance Scaling Law: systems undergoing sustained high-energy dissipation route to high-TD FULL states rather than fragmenting.

Across the 14-domain STRUC-PERC baseline (81 runs, 0 hard violations), the stellar corpus adds 10 runs with the same zero-violation result, extending the program's empirical base to catastrophic stellar boundary events for the first time. The three-layer κ trajectory spans three distinct corpus tiers: Phase A lands in the biological/atmospheric regime (near-immediate connectivity), Phase B in the low condensed-matter regime, and Phase C approaches the CMB/condensed-matter regime. This cross-tier escalation within a single physical object family — where the observational representation, not the underlying physics, governs structural placement — directly corroborates the Phase Mapping finding that representation is the dominant structural variable.

Using the normalization-reviewed v2 structural vectors, the tri-domain bridge classifies the stellar boundary system as A_to_B_contact_with_C_branching. The pre-collapse radial-profile domain and post-collapse light-curve domain remain weakly separated (d = 0.401), while the spectral line-evolution domain is strongly separated from both B (d = 1.287) and A (d = 1.365). This supports a boundary-routing interpretation rather than a single linear A → B → C transition chain. The branching strength is BC − AB = 0.886 and AC − AB = 0.964: the spectral layer branches substantially further from both the pre-collapse and light-curve domains than those two domains are from each other. This establishes the stellar boundary as an admissible cluster-routing event: the system traverses the collapse, preserving admissibility throughout, but the observable post-collapse representations occupy structurally distinct regions of ℳ_adm. Internal percolation (FULL in all 10 runs) and cross-domain equivalence are confirmed to be different properties — a result directly consistent with the Admissible Cluster Geometry manuscript.

The two structural transitions are driven by orthogonal feature dimensions. The A→B crossing (pre-collapse to light curve) is anchored primarily in collapse_onset_radius — the geometric onset coordinate that shifts when moving from radial profiles to temporal brightness trajectories. The B→C crossing (light curve to spectral evolution) is anchored in tail dominance and kappa_connect — the connectivity-depth and gap-outlier dimensions that characterize how heterogeneous the spectral line landscape is. This orthogonality means the boundary event is not a single structural deformation but a sequential activation of different structural variables, each governing one observable transition.

The two object chains represent opposite ends of the boundary routing spectrum. SN1993J achieves near-linear structural inheritance: its A→B distance (0.150) is the corpus minimum, and its full chain is compact with d_AC ≈ d_BC. This is the closest to a genuine A → B → C transition observable in the current dataset. SN2012aw achieves maximal branching: each successive transition amplifies structural displacement, with the B→C step (1.908) nearly 3× SN1993J's B→C step (0.654). Both objects remain FULL_PERCOLATION in all three layers, confirming that this is a difference in structural routing geometry, not in admissibility. The Admissible Cluster Geometry framework anticipates this pattern: internally connected regions can be widely separated in inter-domain bridge space while each maintaining local percolation — consistent with what is observed here.