UNNS Laboratory · March 2026 · STRUC-I v1.0.4 · α-Sweep Study

Does the Fine-Structure Constant Govern Structure?

A cross-domain investigation into whether varying α — the constant that sets the strength of electromagnetism — breaks the ordering stability of physical systems. The answer is surprising, and it points toward something deeper than a parameter.
Atoms · H / He / Na / Li CMB · Planck 2018 Cosmology · DESI Geoid · Earth / Moon / Mars Nuclear · 15 Isotopes · ENSDF 1,270+ Evaluations · 0 Clean Violations Principle of Structural Alignment
Instrument: STRUC-I v1.0.4 α range: 0.80 → 1.20 · 17 values Domains: 5 physical domains Outcome: Admissibility persistent · Optimality domain-dependent

What This Study Found

If you change the fine-structure constant α — the dimensionless number ~1/137 that governs the strength of electromagnetism — do physical systems become structurally unstable? Does their internal ordering collapse?

Across five physical domains, more than 1,270 structural evaluations, and a ±20% variation of α, the answer is: no clean structural breakdown occurs anywhere. Admissibility — the condition that a system's inversion pressure stays within its own vulnerability budget — persists. But something more subtle happens. The optimal configuration — the α value at which structural pressure is minimised — is entirely domain-dependent. In gravitational fields, it is exactly α = 1.00. In nuclear spectra, it is nowhere.

α does not decide whether structure exists. It decides where structure sits relative to instability.

🎯 The Question Behind the Experiment

The fine-structure constant α ≈ 1/137 is one of the most precisely measured quantities in physics. It governs the strength of the electromagnetic interaction — the force behind atomic spectra, chemical bonding, and light. What it does not obviously govern is the structural organisation of ordered sequences.

The UNNS Substrate framework analyses physical systems as ordered ladders — sequences of energy levels, harmonic coefficients, acoustic peaks — and asks: when perturbed, does the ordering survive? The central inequality is:

The Admissibility Inequality
inv(Pε; L) ≤ ν(Vε(L))

Inversions produced by perturbation ≤ vulnerability capacity of the gap structure.

When this holds: the system is admissible — structurally stable. When it fails: the ordering collapses under perturbation.

The question this study asks is new: does varying α break this inequality? Does changing the electromagnetic coupling — by as much as 20% in either direction — push any physical system into structural collapse?

Preregistered Falsification Criterion

Before any evaluation ran, the falsifier was registered: any clean violation of the admissibility inequality at any tested α in any domain would falsify the persistence hypothesis. A clean violation means sustained breach comparable to the synthetic adversarial baseline (Aκ ≈ 0.52, ρ > 1 at multiple κ-steps) — not a near-boundary excursion.

After 1,270+ evaluations across five domains: the falsifier was never triggered.

⚙️ The Instrument: STRUC-I v1.0.4

Every evaluation in this study was run through Chamber STRUC-I v1.0.4, the UNNS program's primary admissibility instrument. STRUC-I was designed as a falsification engine, not a confirmation tool. It tests whether the admissibility inequality holds across a κ-sweep (κ ∈ [0.01, 1.0], 40 log-spaced steps), with M = 2,000 Monte Carlo runs per κ and perturbation scale ε = κ · median(gaps).

RAW DATA spectrum / field α-deformed ladder GAP STRUCTURE vulnerability graph ν(Vε(L)) PERTURBATION M = 2,000 MC runs κ ∈ [0.01, 1.0] 40 log steps INEQUALITY inv ≤ ν ? ρ = inv / ν OUTPUT ρ̄, ρmax, Aκ regime · state κ-profile percolation threshold κ* = 0.554
STRUC-I v1.0.4 evaluation pipeline. Each α-deformed ladder enters the same preregistered protocol.

The α-deformation is applied via a proxy-deformation protocol — not a trivial rescaling (which would be structurally invisible), but a differential deformation that applies different exponents to the smooth background and the fine-structure residual of each system. This makes α genuinely structurally active, confirmed by normalised gap MAD values well above numerical noise before any chamber run.

Domains tested
5
Atoms · CMB · Cosm · Geoid · Nuclear
α values swept
17
0.80 → 1.20, refined 0.95–1.05
Unique ladders
1,270+
all domains combined
Clean violations
0
at any α in any domain
κ* confirmed
0.554
CMB TT / TE / EE independently
MC runs / step
2,000
× 40 κ-steps per ladder

Open in Fullscreen!

🌐 Five Domains, One Sweep

The same α-sweep was applied to five qualitatively different physical domains. Each domain uses a different proxy-deformation operator, physically motivated by how α enters that system's gap structure. The results reveal three response archetypes — and one striking cross-domain contrast.

0.0 0.2 0.4 0.6 0.8 1.0 ρ̄ (structural pressure) 0.80 0.90 1.00 1.10 1.20 ← α (fine-structure constant, normalised) α = 1.00 DESI H CMB He lev. Na gaps Earth ρ̄=0.057 ⁴⁸Ca gaps DESI (Type I) He (Type II) Na (Type III, max) Earth (Type III, min) ⁴⁸Ca (frustrated)
Schematic ρ̄(α) profiles for representative ladders across five domains. Earth coeffmag (yellow) collapses to ρ̄ = 0.057 uniquely at α = 1.00. ⁴⁸Ca (red dashed) maintains elevated pressure at all α — structural frustration. DESI (cyan) is flat to four decimal places.

Three Archetypes of α-Response

Type I · Metric Activity Only

α is structurally invisible

Ladder geometry unchanged. ρ̄ flat to 4 d.p. State and regime locked regardless of α.

Hydrogen levels/transitions: Δρ̄ < 0.001 over full sweep. Perfect empirical grounding of the STRUC-I Invariance Proposition.

Lithium levels: Boundary-Stabilized at every α from 0.80 to 1.20. Δρ̄ = 0.007.

DESI cosmology: Δρ̄ = 0.0001 — the flattest α-response in the corpus.

INACTIVE
Type II · Sub-threshold Activity

α deforms but does not break

Measurable ρ̄ variation. No state transition. Aκ = 1.000 throughout.

Helium levels: monotone ρ̄ decrease as α rises (0.398 → 0.364). First directional trend in any atomic ladder.

CMB (TT/TE/EE): Δρ̄ = 0.015–0.040. Stable Structure throughout all 17 α values. α is encoded but not controlling.

Nuclear levels: Stable Structure at all α for all 14 completed isotopes.

WEAK / DEFORMING
Type III · Threshold-Crossing

α changes structural state

State transitions occur. Near-boundary excursions (Aκ < 1.000) appear. The physical α plays a structurally privileged role.

Geoid fields (Earth/Moon/Mars): ρ̄ collapses uniquely at α = 1.00. Near-Critical Structure at α = 0.80 (Earth degpow: Aκ,min = 0.619).

Sodium gaps: Boundary-Stabilized exclusively at α = 1.00. Weak Persistence everywhere else — the physical α is a pressure maximum.

Nuclear gaps (⁴⁸Ca, ¹⁵⁰Nd): persistent near-boundary excursions at every tested α.

THRESHOLD-CROSSING

🌍 The Geoid Result — The Strongest Signal

Of all the results in this study, the geoid finding is the most structurally informative. Planetary gravitational fields — represented as spherical harmonic expansions of Earth (L=720), Moon (L=300), and Mars (L=85) — behave in a manner that no other domain reproduces.

0.00 0.25 0.50 0.75 1.00 NEAR-BOUNDARY ZONE (α < 1.00) NEAR-BOUNDARY (α > 1.00) α = 1.00 0.80 0.90 1.00 1.10 1.20 ρ̄ = 0.057 Earth coeffmag ρ̄ ADMISSIBLE
Earth coeffmag ρ̄ vs α. The gravitational field collapses to its structural minimum (ρ̄ = 0.057) uniquely at α = 1.00. At every other tested value, near-boundary excursions appear and ρ̄ rises above 0.93.

The result is exact across all three bodies and all non-table ladder types without exception. At α = 1.00, every geoid ladder achieves its structural pressure minimum and maintains full admissibility. At any other α, violations appear and ρ̄ rises dramatically.

The Degree-Power Table — A Structural Invariant

The degree-power table representation (degree-averaged harmonic power) returns ρ̄ = 0.0177 ± 0.0001 at every tested α for all three bodies. This is independent of the body's mass, size, and geological history — and independent of α. It is the most precisely reproduced structural constant in the corpus.

Any α ≠ 1.00

  • ρ̄ > 0.93 (Earth coeffmag)
  • Near-boundary excursions present
  • Near-Critical Structure at α = 0.80 / 1.20
  • Aκ,min = 0.619 (worst case)
  • State: Boundary-Stabilized / Near-Critical

α = 1.00

  • ρ̄ = 0.057 (Earth coeffmag)
  • Aκ = 1.000 throughout
  • Stable Structure across all bodies
  • Universal across Earth, Moon, Mars
  • Unique — no other α reproduces this

The Operator Is Not the Source of the Signal

A critical point: the αl harmonic deformation operator has no built-in preference for α = 1.00. It is monotonic in α for each (l, m) harmonic — no minimum is structurally encoded by the operator design. The optimum at α = 1.00 arises from the interaction between the operator and the real empirical harmonic coefficient distribution of each gravitational field. The operator is the lens; the data is the source.

⚛️ The Nuclear Contrast — Structural Frustration

The nuclear domain provides the clearest counterpoint to the geoid result. Fifteen isotopes spanning the nuclear chart from ²⁴Mg (light sd-shell) to ²³⁸U (heavy actinide rotator) were evaluated across 17 α values using a spin-weighted proxy deformation. The finding is unambiguous.

No Universal α Optimum in Nuclear Spectra

Of 26 complete nuclear ladder groups, α = 1.00 is the ρ̄ minimum in zero cases. The α minimising structural pressure is nucleus-dependent, scattered across the full range [0.80, 1.20], with no correlation to shell closure, deformation class, or mass number.

0.80 0.90 1.00 1.10 1.20 α = 1.00 ← NOT special ²⁴Mg lev ²⁴Mg gap ²⁸Si lev ²⁸Si gap ⁴⁸Ca lev ⁴⁸Ca gap ⁵⁶Fe lev ⁵⁶Fe gap ⁹⁰Zr lev ⁹⁰Zr gap ¹⁵⁰Nd gap* ¹⁵²Sm gap ¹⁶⁶Er gap ¹⁷⁴Yb gap ²³⁸U gap levels α_min gaps α_min * persistent violations
Position of ρ̄ minimum (α_min) for each nuclear ladder group. Points are scattered across the full α range with no clustering near α = 1.00. Compare to the geoid result where all 12 groups locate their minimum at exactly α = 1.00.

Two nuclear gap ladders — ⁴⁸Ca (doubly-magic Z=20, N=28) and ¹⁵⁰Nd (N=90 transitional) — show persistent near-boundary excursions at every tested α. No value of α in [0.80, 1.20] resolves their structural excess pressure. The doubly-magic shell closure and the N=90 shape-phase transition both create gap architectures that are structurally stressed regardless of the coupling strength.

This is what the study calls structural frustration: competing internal constraints (shell closures, deformation, spin-orbit coupling) prevent global alignment between the α-induced deformation and the ordering geometry of the level sequence. No fine-tuning of α can dissolve the tension.

📄

Full Manuscript: A Structural Principle for the Fine-Structure Constant

The complete formal treatment — including the Principle of Structural Alignment, Theorem 10.1 (Admissibility Persistence), Theorem 10.2 (Structural Alignment Condition), Definition 10.1 (Ordering Symmetry), the Two Universality Classes, cross-domain synthesis tables, and all proofs — is available as a 22-page LaTeX-typeset paper.

⚖️ The Principle of Structural Alignment

The cross-domain results converge on a structural principle — not a dynamic law, not a field equation, but a condition on how physical constants interact with ordering geometry.

Principle of Structural Alignment · March 2026
The fine-structure constant acts as a structure-selective deformation operator. It minimises structural pressure only in systems whose intrinsic ordering geometry is aligned with the symmetry of its induced deformation — that is, only when the α-induced deformation acts within the same equivalence class of gap-structure transformations as the intrinsic ordering dynamics of the system. In all tested physical systems, it does not produce clean admissibility breakdowns within the tested range.

This principle rests on a deeper separation:

ADMISSIBILITY inv(Pε;L) ≤ ν(Vε(L)) Governed by: gap architecture α-independent structural constraint Universal · Persistent · Not broken by α Does structure EXIST? OPTIMALITY α* = argmin ρ̄(L(α)) Governed by: alignment class Geoid: α* = 1.00 (universal) Nuclear: α* = scattered (frustrated) Where does structure sit vs instability? separated
Admissibility (whether structure exists) is governed by gap architecture — universal and α-independent. Optimality (where structure sits) is governed by alignment class — domain-dependent.

Old View: Constants as Scale-Setters

α sets the strength of EM coupling. It determines energies. A different α means a different spectrum — same law, different numbers.

The structural organisation of those spectra is a consequence of quantum mechanics, not of α directly.

New View: Constants as Structural Operators

α acts in two roles simultaneously: as a scale-setting parameter and as a structural operator on ordered sequences.

In aligned systems (geoid), it selects a unique optimal configuration. In non-aligned systems (nuclear), it modulates pressure without selection.

Two Universality Classes

The results partition physical systems into two fundamental classes — independent of scale and independent of the underlying physical theory:

Class Property Behaviour Empirical instances α* position
Alignment α-operator symmetry matches gap ordering symmetry Unique structural minimum at α = 1.00 Earth, Moon, Mars gravitational fields α* = 1.00 (universal)
Non-alignment Competing internal constraints block global alignment No privileged α; structural frustration Nuclear level spectra (15 isotopes) α* = scattered
Type I (invariant) α metrically active, structurally inactive ρ̄ flat; regime locked Hydrogen, DESI cosmology, lithium levels α* undefined (flat)
Type II (deforming) Measurable Δρ̄; no state transition Sub-threshold activity Helium, CMB (TT/TE/EE), nuclear levels α* varies, no universal

Open in Fullscreen!

💡 What Has Been Gained

This study did not simply run an α-sweep. It exposed a new structural layer of physical description — one that sits beneath the dynamical laws and above the specific configurations they produce.

The Separation That Changes the Question

The traditional question about α is: "Why is α this particular value?" — a question about fine-tuning, anthropic selection, or deeper theory.

The question this study reveals is different: "Which structures does α align with?" That is a structural question, not a dynamical one, and it has a concrete, empirically testable answer.

Five Concrete Gains

Structural robustness
Proven
Admissibility survives ±20% α-variation across all domains
New separation
Clear
Existence vs optimality — controlled by different things
Classification
3 types
Reusable across any domain with ordered sequences
Geoid result
Exact
α = 1.00 is the unique admissible point in gravitational harmonics
Nuclear contrast
Sharp
No universal α minimum — structural frustration confirmed
κ* confirmed
0.554
Percolation threshold in CMB TT/TE/EE source tables

A Structural Layer Between Laws and States

Physics currently has three levels of description: laws (the equations), parameters (the constants), and solutions (the states). This study provides evidence for a fourth:

LAWS Schrödinger equation · Einstein field equations · Standard Model STRUCTURE ← NEW Ordered sequence geometry · Perturbation stability · Admissibility constraint Constants act as operators on this layer — not just as scales PARAMETERS α · ℏ · G · mass ratios · coupling constants STATES Observed configurations · Energy levels · Harmonic fields · Acoustic peaks
A structural layer sits between dynamical laws and physical states. It is expressed in ordered sequences rather than fields, governed by perturbation stability rather than dynamics, and it is where fundamental constants act as operators — not just as scale-setters.

🔗 Relation to Existing Theories

These results do not modify or contradict established physical theories. They operate at a different level of description — complementing rather than competing.

Quantum Mechanics

Predicts energies, not ordering stability

Standard QM tells us how atomic levels depend on α. It does not address whether the ordering structure of those levels is stable under perturbation. This study provides that missing layer.

COMPLEMENTARY
Cosmology

α is encoded in CMB — but does not control structure

CMB recombination physics explains how α shapes peak positions. The present results show this encoding does not alter structural admissibility. α is in the data without controlling the geometry.

REFINES INTERPRETATION
Nuclear Physics

EM corrections exist — α alignment does not

Nuclear models include electromagnetic corrections to level energies. They do not predict a structural optimum in α. The observed absence of alignment is consistent with theory but reveals a structural property it does not capture.

REVEALS NEW PROPERTY
Geophysics

Harmonics explained — α optimality is new

Standard geophysical models explain the spherical harmonic structure of gravitational fields. They do not predict a universal structural optimum at α = 1.00. The harmonic structure may encode constraints not yet captured by conventional formulations.

EXPOSES HIDDEN STRUCTURE

Resources

UNNS Substrate Research Program · March 2026 · Chamber STRUC-I v1.0.4 · α ∈ [0.80, 1.20] · 17 values · 1,270+ evaluations · 5 physical domains · Proxy-deformation protocol · Falsification-first design · 0 clean violations