index

The KK spectrum computation established that the Poincaré homology sphere $S^3/\mathrm{2I}$ has the deepest spectral protection gap among ADE orbifolds—56× enhancement of the KK mass gap, 15 forbidden levels, genus $g = h/2$. The ADE classification¹ permits other branches: $E_7$ (binary octahedral), $E_6$ (binary tetrahedral), and the $D$-series (binary dihedral), each with its own orbifold, semigroup, and spectral gap. The question is whether these branches can coexist on the same 6D substrate, and if so, what energy cost the domain walls between them carry. If domain wall energies are infinite, branches are globally exclusive and $E_8$ selection is a boundary condition. If finite, coexistence is possible and the stability hierarchy determines relative populations.

This post computes the spectral mismatch between all pairs of the four principal ADE branches ($E_8$, $E_7$, $E_6$, $D_8$), establishes that all domain wall energies are finite, and shows that $E_8$ selection is a dynamical stability outcome—primordial and irreversible.

1. Gap Sets and Structural Ordering

The semigroup selection rule determines the forbidden KK levels for each branch from the Klein invariant degrees² alone. The numerical semigroup³ gap sets for the four branches are:

The gap sets satisfy a strict inclusion hierarchy:

$$\mathrm{gap}(E_6) \subset \mathrm{gap}(E_7) \subset \mathrm{gap}(E_8)$$ $$\mathrm{gap}(D_8) \subset \mathrm{gap}(E_7) \subset \mathrm{gap}(E_8)$$

Verified: $\{1,2,5\} \subset \{1,2,3,5,7,11\} \subset \{1,2,3,4,5,7,8,9,11,13,14,17,19,23,29\}$, and $\{1,3,5\} \subset \{1,2,3,5,7,11\} \subset \{1,2,3,4,5,7,8,9,11,13,14,17,19,23,29\}$. The $E_6$ and $D_8$ gap sets are not nested in either direction: $E_6$ uniquely forbids $l = 2$, $D_8$ uniquely forbids $l = 3$.

$E_8$ is the unique maximal element under gap-set inclusion in the sparse regime ($l \leq 29 = h_{E_8} - 1$). This ordering uses no energetic modeling, no KK multiplicities, and no wall tension functional. The invariant ring structure alone induces it. The $A$-family (binary cyclic groups) has empty gap sets—the halved Klein degrees always include 1, generating all of $\mathbb{Z}_{\geq 0}$—and contributes no structural deletion.

The gap-depth hierarchy has a character-theoretic origin in the Chebyshev polynomials of the second kind. The KK multiplicity at spin $l$ on $S^3/\Gamma$ is determined by a character sum over conjugacy classes of $\Gamma$, each evaluated at the Chebyshev polynomial $U_{2l}(\cos(\theta_g/2))$, where $\theta_g$ is the rotation angle of the group element. A spin $l$ falls in the gap set precisely when destructive interference among these Chebyshev amplitudes produces exact cancellation. The depth of each branch’s gap set therefore depends on the algebraic nature of the cosines at its characteristic half-angles. For binary tetrahedral ($\mathrm{2T}$, $E_6$), the half-angles are $\pi/3$ and $\pi/2$, giving $\cos(\pi/3) = 1/2$ and $\cos(\pi/2) = 0$. All even-index Chebyshev values $U_{2l}$ at these arguments are integers cycling through $\{0, \pm 1\}$, and the resulting character sums cancel at only 3 spins. For binary octahedral ($\mathrm{2O}$, $E_7$), the additional half-angle $\pi/4$ introduces $\cos(\pi/4) = \sqrt{2}/2$, but the irrationality cancels at even indices: $U_{2l}(\sqrt{2}/2) = \sin((2l+1)\pi/4)/\sin(\pi/4)$, and since $(2l+1)$ is always odd, $\sin((2l+1)\pi/4) \in \{\pm\sqrt{2}/2\}$, so $U_{2l}(\sqrt{2}/2) \in \{\pm 1\}$ exactly. Six gaps result. For binary dihedral ($\mathrm{BD}_6$, $D_8$), the half-angle $\pi/6$ gives $\cos(\pi/6) = \sqrt{3}/2$, and $U_{2l}(\sqrt{3}/2) = 2\sin((2l+1)\pi/6) \in \{\pm 1, \pm 2\}$—again all integer. Three gaps. For binary icosahedral ($\mathrm{2I}$, $E_8$), the half-angle $\pi/5$ gives $\cos(\pi/5) = \varphi/2$. The even-index values $U_{2l}(\varphi/2) = \sin((2l+1)\pi/5)/\sin(\pi/5)$ cycle through $\{0, \pm 1, \pm\varphi\}$ with period 5 in $l$. The golden ratio $\varphi$ enters the character formula directly at even indices—the only branch where this occurs. Because $\varphi$ is the worst-approximable irrational⁴ (Hurwitz 1891), the resulting destructive interference in the character sum is maximally effective: integer and irrational amplitudes cannot achieve the partial cancellations that would produce near-zero but nonzero multiplicities. They either cancel exactly or miss by an irreducible irrational margin. This algebraic incompatibility between $\varphi$ and $\mathbb{Z}$ produces the deepest gap structure—15 forbidden spins—among all ADE branches.

2. Spectral Mismatch and Domain Walls

At a domain wall between branches $\Gamma_1$ and $\Gamma_2$, a KK mode at spin $l$ that propagates on one side ($n_l > 0$) but sits in the gap set of the other ($n_l = 0$) becomes evanescent, with decay length $\delta_l = R / \sqrt{4l(l+1)}$. Each evanescent mode contributes boundary energy proportional to its multiplicity times inverse penetration depth:

$$\varepsilon_l \propto m_l \cdot \sqrt{\lambda_l}$$

The hard mismatch set $\mathcal{M}$ consists of all spin values that survive on one side but are forbidden on the other. This set is bounded above by $\max(h_1, h_2)$: above both Coxeter numbers, all modes survive on both sides and contribute no hard mismatch. The hard tension sum therefore terminates after finitely many terms. No regularization is required.

At every $E_8$ wall, the mismatch is entirely one-directional: the column “modes forbidden only in $E_8$” is empty. The partner branch always has modes that $E_8$ forbids; $E_8$ never has modes the partner forbids. This is the set-theoretic consequence of gap-set nesting. The only pair with bidirectional mismatch is $E_6 \leftrightarrow D_8$, where each branch uniquely forbids one mode the other permits.

3. Wall Tension Hierarchy

The dimensionless hard wall tension $T_{\mathrm{hard}} = \sum_{l \in \mathcal{M}} m_l \cdot \sqrt{4l(l+1)}$ was computed for each pair. The physical tension is $\sigma_{\mathrm{wall}} = T_{\mathrm{hard}} / R^3$, where $R$ is the compactification radius.

The hierarchy spans a factor of $\sim$456 between cheapest and most expensive. The binary polyhedral groups satisfy the subgroup chain $\mathrm{2T} \subset \mathrm{2O} \subset \mathrm{2I}$, so the $E_8 \to E_7$ and $E_8 \to E_6$ transitions are symmetry-breaking cascades. The $E_8 \to D_8$ transition is not ($\mathrm{BD}_6$ is not a subgroup of $\mathrm{2I}$). The cheapest wall ($E_6 \leftrightarrow D_8$, $T = 73$) involves groups without a subgroup relation. The wall tension is dominated by the number and eigenvalue weight of mismatched modes, not the algebraic relationship between the groups.

The $E_8 \leftrightarrow E_7$ wall is dominated by its highest mismatched mode: $l = 29$ (the Frobenius number of $E_8$) contributes $\sim$49% of the total tension due to its high multiplicity ($n_{29} = 2$ in $E_7$, giving $m_{29} = 118$).

4. Thermal Stability and the Selection Epoch

A domain wall is stable against thermal hole nucleation when the wall energy exceeds the available thermal energy at the thermal wavelength: $\sigma_{\mathrm{wall}} > T_{\mathrm{cosmo}}^3$. The critical temperature for each wall is

$$T_{\mathrm{crit}} = \frac{(T_{\mathrm{hard}})^{1/3}}{R}.$$

Below $T_{\mathrm{crit}}$, the wall is stable; above it, thermal fluctuations dissolve it. The dissolution ordering preserves the wall tension hierarchy: cheapest walls dissolve first, $E_8$ walls dissolve last.

At GUT-scale compactification ($R = 10^4 \, l_{\mathrm{Pl}}$):

where the stability ratio $\Omega = \sigma_{\mathrm{wall}} / T_{\mathrm{cosmo}}^3$, with $\Omega > 1$ indicating wall stability. At the GUT scale, the cheapest wall ($E_6 \leftrightarrow D_8$) has $\Omega = 0.13$—marginally unstable. The $E_8 \leftrightarrow D_8$ wall has $\Omega = 61$—stable by a factor of 60. A selection window exists between $\sim 5 \times 10^{15}$ and $\sim 4 \times 10^{16}$ GeV where cheap walls dissolve while $E_8$ walls persist.

As the universe cools through the branch selection epoch, walls dissolve in order:

1. $E_6 \leftrightarrow D_8$ dissolves at $T > 5.1 \times 10^{15}$ GeV
2. $E_7 \leftrightarrow D_8$ dissolves at $T > 1.1 \times 10^{16}$ GeV
3. $E_7 \leftrightarrow E_6$ dissolves at $T > 1.1 \times 10^{16}$ GeV
4. $E_8 \leftrightarrow E_7$ dissolves at $T > 3.0 \times 10^{16}$ GeV
5. $E_8 \leftrightarrow E_6$ dissolves at $T > 3.9 \times 10^{16}$ GeV
6. $E_8 \leftrightarrow D_8$ dissolves at $T > 3.9 \times 10^{16}$ GeV

$E_8$ walls form first and are the last to become unstable. $E_8$ is the last branch standing.

The dissolution ordering has a statistical-mechanical origin. The KK partition function on each orbifold $S^3/\Gamma$ is $Z_\Gamma(T) = \sum_l m_l \, e^{-\sqrt{\lambda_l}/(TR)}$, where the sum runs only over modes surviving the quotient. Branches with more gap-set deletions have fewer thermally active modes at any temperature below the compactification scale. $E_8$ deletes 15 of the first 30 levels and pushes the spectral floor to $\lambda_1 = 168$; $E_6$ deletes 3 and has $\lambda_1 = 48$; $D_8$ deletes 3 and has $\lambda_1 = 24$. As the universe cools below the compactification scale, branches with lower spectral floors activate thermal excitations first—each activated mode contributes $\sim T^4$ to the effective potential density, raising $V_{\mathrm{eff}}$ above the value of branches whose modes remain frozen. The branch that activates last has the lowest effective potential at every temperature in the cooling sequence. Gap-set nesting ($\mathrm{gap}(E_6) \subset \mathrm{gap}(E_7) \subset \mathrm{gap}(E_8)$) guarantees that $E_8$ activates last at every level: any mode frozen in a smaller gap set is necessarily frozen in $E_8$, but $E_8$ freezes additional modes that the smaller sets permit. The wall dissolution hierarchy is the domain-wall manifestation of this partition function ordering—the same physics viewed from the boundary rather than the bulk.

5. Frozen Tunneling

In the thin-wall approximation, the Euclidean bounce action for tunneling from a false vacuum to a true vacuum separated by a domain wall of tension $\sigma$ with vacuum energy splitting $\varepsilon$ is⁵

$$B = \frac{27\pi^2 \sigma^4}{2\varepsilon^3}.$$

For the ADE branches, $\sigma = T_{\mathrm{hard}} / R^3$ and $\varepsilon = \Delta\mathrm{drive} / R^4$. Substituting:

$$B = \frac{27\pi^2 \, T_{\mathrm{hard}}^4}{2 \, \Delta\mathrm{drive}^3}.$$

The $R$ dependence cancels exactly. The bounce action is a pure dimensionless number determined by the spectral data ($T_{\mathrm{hard}}$ and $\Delta\mathrm{drive}$). Selection operates identically at any compactification scale.

Using the spectral drives from From Lattice Projection to Cosmic Expansion ($E_8$: 40.7, $E_7$: 19.5, $E_6$: 11.7, $D_8$: 5.8), all downhill transitions are:

Every bounce action satisfies $B \gg 1$. The smallest is $B \approx 1.8 \times 10^7$ ($D_8 \to E_6$). Zero-temperature quantum tunneling between any pair of ADE branches is completely frozen out. $E_8$ has no downhill channels (it is the true vacuum), and all uphill transitions away from $E_8$ have $B > 10^{14}$.

The mechanism: $B \sim T_{\mathrm{hard}}^4 / \Delta\mathrm{drive}^3$. The wall tensions range from 73 to 33,332. The vacuum energy splittings range from 5.9 to 34.9. The fourth power of the wall tension overwhelms the third power of the splitting. The domain walls are much too expensive relative to the vacuum energy gained by conversion.

At finite temperature, the thermal bounce action is $B_{\mathrm{thermal}} = 16\pi \sigma^3 / (3\varepsilon^2 T)$⁶, which does depend on $R$. At GUT-scale compactification and GUT temperature, the cheapest transition ($D_8 \to E_6$) has $B_{\mathrm{thermal}} \approx 23{,}000$. Thermal activation cannot drive transitions either.

Branch selection is primordial. Once a region of the 6D substrate compactifies to a particular $S^3/\Gamma$, it cannot convert to another branch by bubble nucleation—at any temperature below the compactification scale. Selection occurs during the compactification phase transition itself, when $E_8$ has the highest critical temperature, the deepest spectral gap, and the most distributed gap structure. Once formed, the frozen tunneling rates ensure it is irreversible.

6. The Five-Leg Argument

Five independent properties converge on $E_8$:

Leg 1 has a concrete spectral mechanism: the worst-approximability of $\varphi$ manifests through the Chebyshev periodicity theorem (§1)—$U_{2l}(\varphi/2)$ takes values in $\{0, \pm 1, \pm\varphi\}$ with period 5 in $l$, and these irrational amplitudes create the strongest destructive interference in the character sum, producing the deepest gap structure among all ADE branches. The other three branches have exclusively integer even-index Chebyshev values, limiting their gap depth to 3–6 spins. The full Chebyshev analysis connects the phi-sector of constraint geometry (which identifies $\varphi$ as the recursion eigenvalue) to the KK spectral computation (which computes the resulting gap set): Hurwitz worst-approximability is the algebraic engine driving spectral gap depth.

Legs 1–3 establish why $E_8$ is the spectrally preferred branch. Leg 4 identifies the epoch at which selection occurs. Leg 5 proves the selection is permanent.

Footnotes

1. McKay, J. (1980). Graphs, singularities, and finite groups. In The Santa Cruz Conference on Finite Groups, Proceedings of Symposia in Pure Mathematics, Vol. 37, AMS, 183–186. ↩

2. Klein, F. (1884). Vorlesungen über das Ikosaeder und die Auflösung der Gleichungen vom fünften Grade. Teubner, Leipzig. English translation: Lectures on the Icosahedron and the Solution of Equations of the Fifth Degree, Dover, 1956. ↩

3. Rosales, J. C. & García-Sánchez, P. A. (2009). Numerical Semigroups. Developments in Mathematics, Vol. 20, Springer. ↩

4. Hurwitz, A. (1891). Ueber die angenäherte Darstellung der Irrationalzahlen durch rationale Brüche. Mathematische Annalen, 39(2), 279–284. ↩ ↩²

5. Coleman, S. & De Luccia, F. (1980). Gravitational effects on and of vacuum decay. Physical Review D, 21(12), 3305–3315. ↩

6. Linde, A. D. (1983). Decay of the false vacuum at finite temperature. Nuclear Physics B, 216(2), 421–445. ↩