THE NECESSITY CHAIN
A First-Principles Derivation of Electromagnetic Unification
Kelly B. Heaton / Coherence Research Collaboration, 2026
FULL PAPERS (all open access)
Foundational series — Thread Frame coordinate system
Recursive Geometry of Atomic Spectra (2025)
https://doi.org/10.5281/zenodo.17188444
Information-Theoretic Confirmation of the alpha-Affine Thread Frame (2025)
https://doi.org/10.5281/zenodo.17335815
Hydrogenic Alignment as a Coordinate Principle for Atomic Spectra (2025)
https://doi.org/10.5281/zenodo.18167643
Geometric unification series
Determinacy Under Quotient Representations (2026)
https://doi.org/10.5281/zenodo.18868210
Electromagnetic Closure and the Fine-Structure Constant: A Geometric Derivation (2026)
https://doi.org/10.5281/zenodo.19157339
The Encounter of Two Primitives and One Flaw: Geometric Unification at the Electromagnetic Scale (2026)
https://doi.org/10.5281/zenodo.19194036
Vacuum Impedance as the Organizing Principle of Atomic Spectra (2026)
https://doi.org/10.5281/zenodo.19164224
The Proton Mass from Atomic Spectra: A Geometric Derivation (2026)
https://doi.org/10.5281/zenodo.19355588
HOW TO READ THIS DOCUMENT
This document leads with two things standard physics cannot explain, then shows that a single suppressed coordinate resolves both. The geometry is presented as the solution to a pre-established problem, not as a proposed representation. What is claimed is that the framework generates novel predictions confirmed by independent data, and that these predictions have not appeared elsewhere in the literature and are not generated by standard quantum mechanics.
The most attack-resistant single result is Prediction 2 in the Empirical Confirmation section: the electron rest mass is recovered to -2.8 plus or minus 2.7 ppm from the location of a structural absence in atomic spectra — not from any spectral line, not from any fit, and not predicted by standard QM.
Each step closes before the next opens. An objection to step N is only valid if it does not require rejecting steps 1 through N-1.
TWO UNSOLVED PROBLEMS
Problem 1: The fine-structure constant
The constant alpha (approximately 1/137.036) has resisted first-principles derivation since its introduction by Sommerfeld in 1916. Eddington attempted it in 1946. Wyler produced a numerically close formula in 1969 from the conformal group of Maxwell's equations, but Robertson (1971) showed it required an arbitrary radius setting with no physical justification. Singh (2022) obtained an asymptotic value from the exceptional Jordan algebra, but the derivation requires a fitted electroweak symmetry-breaking scale. No derivation has produced alpha from zero free parameters without importing external physical structure. The problem is documented as open in Jentschura and Nandori (2014).
Problem 2: The IE/4 boundary
There is a structural forbidden zone in the transition spectrum of every atomic ion, centered at exactly one quarter of the ion's ionization energy. The nuclear charge Z cancels exactly from the boundary condition, so the gap is not ion-specific: it falls at the same dimensionless position in every ion's spectrum regardless of atomic number, electron configuration, or ionization stage. Standard quantum mechanics does not predict this boundary. It has not appeared in the spectroscopic literature. It is confirmed across 71 ions with zero falsifications.
Both problems have the same origin. Identifying it is the next step.
THE SUPPRESSED COORDINATE
Maxwell's equations in vacuum are governed by two independent electromagnetic invariants: the propagation speed c (the product of the vacuum constants) and the vacuum impedance Z0 (their ratio). Together they uniquely recover both vacuum constants. The compression from (epsilon-0, mu-0) to c-squared retains the product and discards the ratio. This is a non-injective map: infinitely many pairs of vacuum constants share the same propagation speed while partitioning field energy differently. Any observable sensitive to the electric/magnetic energy partition is indeterminate from c alone. This is proved formally in the Determinacy paper.
Atomic physics has used c extensively. Z0 as an organizing principle of atomic structure has not been systematically investigated. Both unsolved problems arise from working in the compressed representation. Restoring Z0 resolves them.
STEP 1. TWO PRIMITIVES AND ONE ASYMMETRY
Any electromagnetic vacuum requires exactly two irreducible properties. The first is the capacity to contain and bound a field: to support a structured interior, a surface, a stored charge. This is epsilon-0. The second is the capacity to transmit and pass a field: to remain open, to offer no resistance to flux. This is mu-0. Remove either and there is no electromagnetic process.
The SI units make this concrete. Epsilon-0 in farads per meter is capacitance per unit length — the capacity to hold separated charge across a boundary. Mu-0 in henries per meter is inductance per unit length — the capacity to sustain flux that passes through without termination. These are not analogies to structural and non-structural primitives; they are those primitives, dimensionally.
The asymmetry follows directly from Maxwell's equations. Gauss's law for electricity states that electric field lines begin and end on sources. Gauss's law for magnetism states that magnetic field lines never terminate. Therefore only the electric sector can host a localized defect. The defect must live in epsilon-0.
STEP 2. ONE SINGULARITY, ONE COMPLETION
On the real line, the map 1/z terminates at z = 0 and has no orientation. A vacuum must propagate without termination. The real line fails.
The complex plane gives orientation: the map 1/z reverses the argument of z, which is the origin of the perpendicularity of electric and magnetic fields in Maxwell's curl equations. But the process still terminates at z = 0.
The unique minimal non-terminal completion is the Riemann sphere, obtained by adding a single point at infinity. This gives z = 0 an image, makes the space compact and oriented, and leaves the map well-defined everywhere. No other completion is minimal. The map 1/z is also the unique meromorphic bijection of the sphere with a single simple pole at z = 0 — maps with higher-order poles produce multiple-sheeted covers, not bijections.
STEP 3. ONE LOGARITHMIC FIELD
The Riemann sphere is a two-dimensional manifold. The Green's function of the two-dimensional Laplacian is logarithmic. By the Poincare-Hopf theorem, the Euler characteristic of the sphere (equal to 2) requires any smooth directed field on it to have singularities whose indices sum to 2. The minimal realization is two poles of index +1: one source and one sink. No simpler realization satisfies the topology.
The unique rotationally symmetric potential between these poles is the logarithmic Green's function of the Laplacian. No other potential satisfies both isotropy and Laplace's equation in this geometry. The logarithmic field is forced.
STEP 4. ONE ATTRACTOR, ONE DEPARTURE, ONE ALPHA
The logarithmic field has a self-similar scaling structure. The fixed point of the scaling recursion x = 1 + 1/x is the golden ratio, phi. The Markov spectrum theorem establishes that phi is the unique irrational maximally resistant to rational approximation at every scale simultaneously. Hurwitz (1891) proved it is the sole number satisfying this property uniformly. Every other irrational loses self-similar structure at some scale. Phi is the unique coherent attractor.
Rational winding is excluded: if the winding number were rational, every orbit would close exactly, no transition would occur, no photon would be emitted. Matter exists and radiates; therefore the winding is irrational and the attractor is phi.
The irremovable defect at the source pole forces a departure from phi. The geometry of steps 1 through 3 places exactly three dimensioned quantities on the table: Z0 (the vacuum impedance ratio), e (the minimum charge of the defect), and h (the unique scale-invariant of the self-similar field, identified in Step 6). Any dimensionless departure must be a monomial in these three quantities. Solving the dimensional constraints produces a unique one-parameter family: (Z0 times e-squared divided by h) to the power a, which equals (Z0 divided by R_K) to the power a, where R_K is h/e-squared. The topological minimum selects a = 1, and the minimum winding condition introduces a factor of 2. The result is:
alpha = Z0 / (2 R_K) = e-squared / (2 epsilon-0 h c), approximately 1/137.036.
No alternative combination is dimensionless, scale-invariant, and topologically minimal simultaneously. This resolves Problem 1.
STEP 5. ORBITAL IMPEDANCE AND THE IE/4 BOUNDARY
A nucleus produces a Coulomb field closing at orbital velocity v = alpha times c divided by n. The field impedance of this orbital, derived from Coulomb's law and the Biot-Savart law, is Z0 times n divided by alpha. The nuclear charge and orbital radius cancel exactly from the ratio of electric to magnetic field amplitude. No free parameters. For all bound orbitals, the impedance far exceeds the vacuum impedance and the reflection coefficient approaches 1: the orbital stores energy because the impedance mismatch is severe. No quantum postulate about forbidden radiation is required.
The natural depth coordinate kappa places the impedance structure on a ruler with tick marks at powers of alpha. On this ruler the impedance boundary falls at universal depth kappa-q approximately 0.2818 in every ion's spectrum, corresponding to transition energy IE/4. Nuclear charge cancels exactly.
No bound-to-bound transition can carry exactly IE/4. The Diophantine proof is exhaustive: setting the transition energy equal to IE/4 requires an integer solution to an equation that has none for any combination of initial and final quantum numbers. IE/4 is a structural forbidden zone — not dynamical suppression but geometric absence. This resolves Problem 2.
STEP 6. THE HARMONIC LIGHT LAW
The vacuum field potential satisfies Laplace's equation (Step 3). The energy of a bound state at depth kappa is E0 times alpha to the power kappa. Planck's relation gives frequency as energy divided by h. Therefore the logarithm of frequency equals a constant minus the field potential. Since the field potential satisfies Laplace's equation, so does the logarithm of frequency:
Laplacian of (log nu) = 0.
This is the Harmonic Light Law. It is derived, not assumed.
The scaling invariant h emerges as the unique constant ratio between the energy ladder and the frequency ladder constructed independently from NIST data. The Thread Law — log frequency equals a constant plus beta times kappa, with slope beta equal to log alpha — is the unique spherically symmetric solution of the Harmonic Light Law. The orbital circuit admittance is the same equation in the voltage domain. Two instruments, one law.
WHAT IS GENUINELY NEW
The following results are not in the existing literature and must be engaged on their own terms.
1. Z0 as an organizing principle of atomic spectra. The formula alpha = Z0/2R_K is a known SI identity. That Z0 functions as the vacuum coordinate from which atomic spectral structure is derivable — with nuclear charge cancellation giving a universal boundary across the periodic table — is not known.
2. The IE/4 forbidden zone. No spectroscopic or quantum chemistry publication identifies a universal structural absence at IE/4 across multi-electron ions with exact nuclear charge cancellation.
3. Recovery of electron mass from a structural absence. The identification of the electron rest mass as recoverable from the geometric center of the forbidden zone, at ppm precision, is not in the literature.
4. The nuclear magic analog. The correspondence between the atomic IE/4 boundary and depletion below S_n/4 in doubly magic nuclei (4 of 4 confirmed, 0 of 20 non-magic) is not in the literature.
5. First-principles derivation of alpha. Derivation of alpha = Z0/2R_K from the geometry of the reciprocal singularity with zero free parameters is the first such derivation. Prior attempts since 1916 are catalogued in the fine-structure paper.
6. Information-theoretic confirmation of the Thread slope. The slope beta = log alpha has been confirmed by Minimum Description Length analysis across six independent spectral domains with probability less than 10 to the power -6 of occurring by chance. This confirmation does not assume the geometric framework; it asks only which slope minimizes the description length of photon data, and recovers log alpha.
7. Derivation of the proton mass from atomic spectra. The proton condenses at Fibonacci winding step j = 5, the first step at which the convergence error toward the golden ratio falls below alpha. From this geometry: m_p times c-squared equals 4 to the 13th power times E0 times the factor (2157 minus 869 times the square root of 5) divided by 208, giving 938.273 MeV, accurate to 0.61 ppm. The only empirical input is the hydrogen ionisation energy E0 from NIST. No coupling constant, quark mass, nuclear parameter, or fitted constant enters. The correction factor is a closed algebraic expression in Fibonacci integers and the square root of 5 alone. This result is not in the literature.
EMPIRICAL CONFIRMATION: ZERO FALSIFICATIONS
The framework makes predictions standard QM does not make in this form. Each was derived before the data was searched. None has been falsified.
Prediction 1: Universal IE/4 gap with nuclear charge cancellation.
Tested against the NIST Atomic Spectra Database across 80 ions spanning Z = 1 to 82, all four subshell families, and multiple ionization stages. Result: 71 confirmed, 0 falsified. The remaining 9 have catalogs that do not bracket IE/4 from both sides and are inconclusive, not failures. In 42 of 71 confirmed ions the gap ranks in the top 10% by width. Gap width stratifies by subshell in the predicted direction: d-block narrowest (median 8.4 meV), f-block intermediate (32.8 meV), s/p-block widest (189.5 meV).
Prediction 2: Electron mass from a geometric absence.
The gap center should recover the electron rest mass after reduced-mass correction. Recovery is from a structural absence — not a spectral line. Standard QM predicts no such recovery. Tested on five precision-tier ions with sub-0.1 meV gap resolution. The reduced-mass correction is not fitted; it is derivable from CODATA constants. Result: -2.8 plus or minus 2.7 ppm (standard error of the mean, n = 5).
Prediction 3: Uniqueness of IE/4.
IE/3 returns -250,133 plus or minus 1.1 ppm; IE/5 returns +250,129 plus or minus 4.6 ppm — exact agreement with algebraic predictions for wrong-fraction evaluations. IE/4 is the unique forbidden boundary.
Prediction 4: Nuclear analog.
Doubly magic nuclei (oxygen-16, calcium-48, tin-132, lead-208) should show complete depletion below S_n/4; non-doubly-magic nuclei should not. Result: 4 of 4 doubly magic confirmed; 0 of 20 non-magic show any gap. Zero overlap. The nuclear confirmation emerged as an independent test after the atomic result.
Prediction 5: Thread Law slope across all spectral domains.
The Thread Law predicts slope beta = log alpha in every spectral domain. Confirmed across six independent domains (atomic, solar, stellar, molecular) by two independent methods: direct Thread Frame analysis across 30 or more ions, and Minimum Description Length analysis across six independent datasets without assuming the geometric framework. All six MDL datasets converge to a slope statistically indistinguishable from log alpha, with probability less than 10 to the power -6 of occurring by chance.
Prediction 6: Proton mass from Fibonacci condensation geometry.
The H-mode winding converges toward the golden ratio through successive Fibonacci approximations. The proton condenses at step j = 5, where the convergence error first falls below alpha, at winding depth F7 = 13. The formula, derived from the destructive interference between the j = 5 and j = 6 half-steps and expressed entirely in Fibonacci integers and the square root of 5:
m_p times c-squared = 4 to the 13th power times E0 times (2157 minus 869 root 5) divided by 208 = 938.272657 MeV, residual +0.61 ppm.
Inputs: hydrogen ionisation energy from NIST, Fibonacci integers F5 = 5, F6 = 8, F7 = 13, F8 = 21, and the Cassini identity. No value of alpha, pi, e, or phi enters. The same geometry predicts the proton reduced Compton wavelength (0.2103 fm, agreement to 0.6 ppm) and the exact 13-step contraction of the Bohr radius to the nuclear scale.
All data are public (NIST, NSO/Kitt Peak, ELODIE, ExoMol). All analysis scripts are open source.
THE LOCK
The four fundamental constants are related within this framework in two distinct senses, which we distinguish explicitly.
Tier 1: Algebraic lock (fully established by this chain)
alpha = Z0 / (2 R_K), from the defect geometry of Step 4.
h = E(kappa) / nu(kappa), the unique scale-invariant of Step 6.
m_e times c-squared = 8G / alpha-squared, where G = E0/4, the geometric floor below which no closure mode exists.
a0 = hbar times c times alpha divided by (8G), from the same geometry.
Once the geometry of Steps 1 through 6 is granted, these four relations are algebraically forced. Knowing any one determines all others. No relation contains a free parameter.
Tier 2: First-principles derivation (partially established)
Tier 1 is not the same as deriving the electron mass and Planck's constant from pure geometry with zero spectroscopic input. In this chain, h is identified as the unique scale-invariant of two independently observed ladders, and the electron mass is recovered from the spectroscopic gap center. Both are derived within the framework rather than fitted to it, but both retain empirical contact. The full field-geometric derivation of the electron mass from geometry alone is the principal outstanding step. It does not undermine the lock; it is what the lock is waiting for.
A framework that locks four constants with zero free parameters and recovers electron mass to ppm from a structural absence is not weakened by acknowledging what it has not yet fully derived. It is made more credible.
SUMMARY
Two problems resolved:
Fine-structure constant: first-principles derivation, zero free parameters.
IE/4 gap: structural necessity, 71 of 80 ions confirmed, 0 falsified.
Most attack-resistant results:
Electron mass: -2.8 plus or minus 2.7 ppm from a structural absence.
Uniqueness: IE/3 at -250,133 ppm; IE/5 at +250,129 ppm.
MDL: slope recovers log alpha across 6 domains, p less than 10 to the -6, no geometric assumption required.
Proton mass: 938.273 MeV, +0.61 ppm, from E0 and integer arithmetic alone, zero free parameters.
Standard QM predicts none of these.
Empirical record:
Free parameters: Zero
New postulates: Zero
Empirical inputs: One (matter exists)
Ions confirmed: 71 of 80 tested
Ions falsified: Zero
Nuclear confirmation: 4 of 4 doubly magic, 0 of 20 mid-shell
Spectral domains confirmed: Six
Proton mass: 938.273 MeV, +0.61 ppm
All papers are open access. Links at the top of this page.
The vacuum is not a featureless stage. It is a two-primitive process with one irremovable defect. The defect generates the Riemann sphere, the logarithmic field, the golden ratio attractor, and the fine-structure constant. The atom is the particular closure that a given nuclear charge sustains against this geometry. The photon is the Z0 relationship propagating between closures. The electron mass is the floor below which no closure can form. The proton is the first condensate of the Fibonacci winding — the structure that forms at the depth where the convergence record of the vacuum geometry first becomes indistinguishable from its own irremovable flaw. The Harmonic Light Law governs photon frequency throughout, confirmed by independent information-theoretic analysis across six spectral domains without assuming any geometric framework.
Everything else is consequence.

