The Geometry of Coupling:
Deriving the Fine Structure Constant
(α
−1
≈ 137) from Lattice Dilution Factors in
a K = 12 Vacuum
Raghu Kulkarni
SSMTheory Group, IDrive Inc., Calabasas, CA 91302, USA
raghu@idrive.com
March 8, 2026
Abstract
The fine structure constant (α = 1/137) determines the strength of the electro-
magnetic interaction. Within the Standard Model, its value remains an empirical
input. We propose that α
−1
operates as a geometric invariant of a discrete vacuum
lattice. Using the Selection-Stitch Model (SSM), the vacuum functions as a satu-
rated Face-Centered Cubic (K = 12) tensor network. We derive the inverse coupling
as the product of three geometric dilution factors: (i) a topological selection factor
(P
topo
= 1/2) from the bipartite/non-bipartite face partition of the cuboctahedron,
previously established in the context of fermion doubling evasion; (ii) an algebraic
projection factor (P
alg
= 1/4) from the vector subspace of the Cl(1, 3) spacetime
algebra, derived from the Cosserat Lagrangian of the SSM vacuum; and (iii) a reso-
nant coherence factor (P
res
= 1/17) from the discrete harmonic quantization of the
face-diagonal coherence path. Adding the unitary self-node correction (+1) yields
α
−1
= 2 × 4 × 17 + 1 = 137, matching the integer part of the observed value. The
+0.036 residual (0.026%) is attributed to higher-order lattice corrections.
1 Introduction
The fine structure constant characterizes the strength of the electromagnetic interaction.
Its low-energy value is approximately 1/137.035999 [1, 2]. Early theoretical attempts to
derive the integer 137 by Eddington [3] and Wyler [4] constructed isolated algebraic ratios.
These mathematical models lacked a consistent physical framework capable of predicting
other observables.
In discrete physics, coupling constants arise from geometric probability: the likelihood
that a random fluctuation in the vacuum medium aligns with a specific propagation
channel. We extend the Selection-Stitch Model (SSM) [5] to the electromagnetic sector.
The SSM models the vacuum as a discrete tensor network with cuboctahedral topology
(K = 12). The integer 137 follows from three independent geometric properties of this
1
unit cell. Each dilution factor connects to a result established independently in companion
papers, providing structural cross-validation beyond the single value of α.
2 The Geometric Dilution Framework
We define electromagnetic coupling as a geometric efficiency: the fraction of vacuum
fluctuations that successfully couple to a photon channel. In a discrete lattice, an elec-
tromagnetic interaction requires a fluctuation at a node (fermion vertex) to align with a
link (boson propagator). The geometry of the lattice dilutes this alignment. The inverse
coupling α
−1
counts the total number of vacuum channels per electromagnetic channel,
plus a unitary correction for self-interaction:
α
−1
= 1/P
total
+ 1 (1)
The total geometric probability P
total
operates as the product of three independent dilution
factors:
P
total
= P
topo
× P
alg
× P
res
(2)
Each factor addresses a distinct physical constraint: (i) Does the local topology permit
flux screening? (ii) Does the algebra permit vector boson propagation? (iii) Does the
discrete scale permit coherent wave propagation? We derive each factor directly from the
geometry of the K = 12 cuboctahedron.
3 Derivation of the Three Dilution Factors
3.1 Factor 1: Topological Selection (P
topo
= 1/2)
A fermion resides on a vertex of the cuboctahedral coordination shell. Interactions op-
erate locally: the node samples only the faces meeting at its vertex. Each vertex of the
cuboctahedron is shared by exactly 4 faces: 2 triangular (C
3
symmetry) and 2 square (C
4
symmetry) [6].
The triangular faces are non-bipartite because they contain odd-length cycles. Our
companion analysis of fermion doubling on the FCC lattice establishes that non-bipartite
substructures frustrate alternating charge assignments [7]. They cannot support the
charge-screening oscillations required for electromagnetic flux propagation. These tri-
angular faces instead confine flux, governing the strong interaction. The square faces are
bipartite (even cycles). They support charge alternation and permit electromagnetic flux
screening.
The topological dilution factor evaluates the fraction of local faces that support elec-
tromagnetic coupling:
P
topo
=
N
square
N
total
=
2
4
=
1
2
(3)
This relies on the identical non-bipartite/bipartite distinction that lifts fermion doublers
to E ∼ 5/a on the FCC lattice. This selection acts as a direct consequence of the lattice
geometry utilized to solve the fermion doubling problem.
2
3.2 Factor 2: Algebraic Projection (P
alg
= 1/4)
The vacuum state space is defined by the spacetime Clifford algebra Cl(1, 3). The selection
of the (1 + 3)-dimensional algebra, rather than the purely spatial Cl(3, 0) algebra, derives
from the dynamical structure of the SSM vacuum. Each SSM lattice node carries both
translational (u, 3 components) and rotational (θ, 1 independent torsional mode) degrees
of freedom. These kinematics are governed by the Chiral Cosserat Lagrangian [8]:
L =
1
2
(∂
t
u)
2
+
1
2
(∂
t
θ)
2
−
c
2
2
(∇u)
2
−
c
2
2
(∇θ)
2
+ Ω(u
˙
θ − θ ˙u) (4)
The (u, θ) pair constitutes a (3+1)-dimensional dynamical system at each node, generating
the Cl(1, 3) algebra with total dimension 2
4
= 16. The 16 basis elements decompose
exactly as: 1 scalar, 4 vectors, 6 bivectors, 4 trivectors, and 1 pseudoscalar [9].
The photon acts as a vector boson (A
µ
) and can only occupy the 4-dimensional vector
subspace. The algebraic dilution factor is therefore:
P
alg
=
dim(Vector)
dim(Cl(1, 3))
=
4
16
=
1
4
(5)
Using the spatial algebra Cl(3, 0) with dim = 8 would yield P
alg
= 3/8, generating a
bare inverse coupling of α
−1
bare
= 91. This strictly rules out a static spatial model. The
dynamical (3 + 1) structure of the Cosserat Lagrangian exclusively determines the correct
physical algebra.
3.3 Factor 3: Resonant Coherence (P
res
= 1/17)
A wave packet propagating through a discrete lattice maintains coherence only if its
wavelength matches the integer harmonics of the grid. The relevant coherence path in a
cubic crystal is the face diagonal of the unit cell [10]. This diagonal represents the path of
least resistance for a shear wave, characterizing the transverse mode of the electromagnetic
field.
For the K = 12 cuboctahedral cell, the face diagonal possesses a length of:
L
diag
= K ×
√
2 = 12
√
2 ≈ 16.97... (6)
On a discrete lattice, the coherence path must consist of an integer number of lattice
spacings. A fractional resonance length is unphysical; a wave cannot terminate at a
point between discrete nodes. The physical constraint requires the coherence path to
accommodate at least full wavelengths. Because 16 lattice spacings fall short of the face
diagonal (16 < 16.97), 16 wavelengths would leave a gap of 0.97 lattice spacings at the end
of the path. This geometric gap destroys the required standing wave boundary condition.
The minimum integer satisfying the coherence requirement evaluates to:
λ = ⌈16.97⌉ = 17 (7)
This ceiling function operates as the standard quantization condition for standing waves
on a finite discrete lattice. The mode number must be the smallest integer ≥ the geometric
path length. The resonant dilution factor evaluates to:
P
res
=
1
λ
=
1
17
(8)
3
4 Assembly and the Self-Node Correction
Substituting the three independent geometric dilution factors into Eq. 1:
α
−1
bare
=
1
(1/2)(1/4)(1/17)
= 2 × 4 ×17 = 136 (9)
This value counts the 136 geometric vacuum channels surrounding the specific interaction
vertex. The physical observable must also formally account for the vertex node itself.
This self-node correction (+1) operates as the same counting principle used consistently
throughout the SSM framework:
SSM Result Neighbor Count Self-Node Total Physical Quantity
Hubble tension 12 +1 13/12 expansion boost
Weinberg angle 12 +1 sin
2
θ
W
= 3/13
Proton mass 12(K) +1 (K + 1)K
2
− cK = 1836
Fine structure 136 +1 α
−1
= 137
Table 1: The self-node correction (+1) appears across four independent SSM derivations.
In each case, the physical quantity evaluates K (or the bare channel count) plus the
central node, yielding K + 1 (or bare + 1).
Applying this geometric correction:
α
−1
= 136 + 1 = 137 (10)
5 The +0.036 Residual
The derived integer 137 differs from the experimental value 137.035999 by exactly +0.036
(0.026%). This residual lies beyond the scope of the bare integer-geometric framework.
In continuum QED, the running of α from its low-energy value arises from vacuum polar-
ization loops. The SSM integer result quantifies the bare topological channel count. The
fractional correction likely arises from next-order lattice effects, such as polycrystalline
grain boundary scattering or anharmonic bond corrections, evaluated at the lattice scale.
We strictly note the residual without attributing it to a singular macroscopic mechanism.
6 Cross-Validation with Independent SSM Results
Unlike historical mathematical numerology [3,4], this derivation provides structural cross-
validation. Each topological factor connects directly to an independently derived SSM
mechanism:
• P
topo
= 1/2: The bipartite/non-bipartite face partition represents the same geomet-
ric property that lifts all 15 fermion doublers to E ∼ 5/a on the FCC lattice [7].
The square faces that carry electromagnetic flux operate as the same faces whose
bipartite structure permits single-species chiral fermions.
4
• P
alg
= 1/4: The Cl(1, 3) algebra follows explicitly from the Cosserat Lagrangian [8].
This Lagrangian independently derives the Schr¨odinger equation and the speed of
light (c = 4v
lattice
). The factor 4 in the speed of light and the factor 4 in P
alg
= 4/16
both originate from the identical structure tensor
P
ˆn
µ
ˆn
ν
= 4δ
µν
.
• P
res
= 1/17: The integer 17 = ⌈K
√
2⌉ acts as a rigid geometric property of the
cuboctahedral unit cell. The observed lepton mass ratios m
µ
/m
e
≈ 206.8 and
m
τ
/m
e
= 3477 align structurally with 17K = 204 and 17
2
K = 3468, presenting ab-
solute errors of only 1.4% and 0.3% respectively. The appearance of the integer 17
within the lepton spectrum requires further investigation. Additionally, the proton-
to-electron mass ratio itself factorizes as 1836 = 12 ×9 ×17, where the path length
through the cuboctahedral cage is 153 = 9 × 17 [13]. The same face-diagonal har-
monic λ = 17 that governs electromagnetic coherence also determines the proton’s
structural path length.
• The +1 correction: The self-node counting principle (K → K + 1 = 13) appears
consistently across four independent SSM frameworks: the Hubble expansion boost,
the Weinberg angle, the proton mass equation, and the fine structure constant. This
rigid consistency argues against numerical coincidence.
7 Falsifiability
This derivation remains falsifiable across three distinct observational fronts:
1. If the Hubble tension is resolved to a value significantly different from the 13/12 pre-
diction (H
local
/H
CMB
= 1.0833), the +1 self-node correction loses its independent
geometric support.
2. If a fourth generation of charged fermions is discovered, the spatial basis of the
lattice mapping (3 directions → 3 generations) is violated.
3. If next-generation lepton mass measurements deviate structurally from the λ = 17
harmonic pattern (204, 3468), the resonant coherence factor loses its phenomeno-
logical cross-validation.
8 Conclusion
We derived α
−1
= 137 purely from the topology of the K = 12 cuboctahedral vacuum unit
cell. The three evaluated dilution factors—topological (1/2), algebraic (1/4), and reso-
nant (1/17)—each connect to an independently established SSM result: fermion doubling
evasion, the Cosserat Lagrangian, and the face-diagonal coherence path. The self-node
correction (+1) operates as the identical counting principle deployed in four other SSM
derivations. The mathematical factorization 136 = 2
3
×17 is unique among integers near
137. No other integer in this local range factors cleanly into the geometric harmonics
(K, λ) of the cuboctahedron. This rigidly suggests the electromagnetic coupling constant
acts as a fundamental geometric invariant of the discrete vacuum.
5
A Self-Contained SSM Summary
A.1. K = 12 Lattice Saturation. The FCC lattice represents the unique geometric
solution to the Kepler conjecture [11]. Each node possesses 12 nearest-neighbor bonds.
The coordination polyhedron acts as the cuboctahedron (V = 12, E = 24, F = 14: 8
triangles + 6 squares).
A.2. Fermion Doubling Evasion. The FCC lattice is non-bipartite due to its
triangular cycles. This violates the Z
2
sublattice symmetry required by the Nielsen-
Ninomiya theorem, elevating all 15 potential doublers to E ∼ 5/a without breaking chiral
symmetry [7].
A.3. The Cosserat Lagrangian. Each SSM node contains translational (u) and
rotational (θ) degrees of freedom. The chiral coupling Ω(u
˙
θ − θ ˙u) generates the complex
Schr¨odinger equation via the complexification mapping ψ = u+iθ [8]. The structure tensor
eigenvalue
P
n
µ
n
ν
= 4δ
µν
yields the geometric light speed renormalization c = 4v
lattice
.
A.4. The Self-Node Correction. In the SSM, macroscopic physical quantities
count the full cuboctahedral cluster: K neighbors + 1 center = K + 1 = 13. This unitary
principle produces the Hubble boost (13/12), the Weinberg angle (3/13), the proton mass
formula ((K + 1)K
2
− cK), and the fine structure constant (136 + 1).
References
[1] L. Morel et al., “Determination of the fine-structure constant with an accuracy of 81
parts per trillion,” Nature 588, 61 (2020).
[2] R. H. Parker et al., “Measurement of the fine-structure constant as a test of the
Standard Model,” Science 360, 191 (2018).
[3] A. S. Eddington, Fundamental Theory, Cambridge Univ. Press (1946).
[4] A. Wyler, “L’espace sym´etrique du groupe des ´equations de Maxwell,” C. R. Acad.
Sci. Paris 269, 743 (1969).
[5] R. Kulkarni, “Constructive Verification of K = 12 Lattice Saturation,” Zenodo:
10.5281/zenodo.18294925 (In Review) (2026).
[6] H. S. M. Coxeter, Regular Polytopes, Dover Publications (1973).
[7] R. Kulkarni, “Fermion Chirality from Non-Bipartite Topology: Resolving the Dou-
bling Problem via Lattice Saturation,” Zenodo: 10.5281/zenodo.18410364 (In Re-
view) (2026).
[8] R. Kulkarni, “Geometric Emergence of Spacetime Scales: Deriving the Speed of
Light, the Planck Scale, and the Two-Step Mass Limit of Quantum Decoherence,”
Zenodo: 10.5281/zenodo.18752809 (In Review) (2026).
[9] P. Lounesto, Clifford Algebras and Spinors, Cambridge Univ. Press (2001).
[10] D. Hull and D. J. Bacon, Introduction to Dislocations, 5th ed., Butterworth-
Heinemann (2011).
[11] T. C. Hales, “A proof of the Kepler conjecture,” Annals Math. 162, 1065 (2005).
6
[12] R. Kulkarni, “Late-Universe Dynamics from Vacuum Geometry: Unifying Dark
Energy and the Hubble Tension via Holographic Phase Transitions,” Zenodo:
10.5281/zenodo.18753874 (In Review) (2026).
[13] R. Kulkarni, “A Topological Ansatz for the Proton-to-Electron Mass Ratio,” Zenodo:
10.5281/zenodo.18253326 (In Review) (2026).
7
