UNNS Substrate Research Programme · Core Results · 2026

Voyager Reveals a Structural Boundary
in the Heliosphere

For the first time, a physical boundary crossing is detected directly from structural data alone — without physical labels, without parameter fitting, without domain-specific models. Using Voyager 1 and Voyager 2 trajectories, we show that the heliopause is not merely a physical surface. It is a measurable geometric event in realizability space. This work introduces a computable operator that identifies phase transitions with no prior knowledge of where the boundary is.
3 500 STRUC-PERC-I Evaluations t* = 2012 · No physical label used Structural Boundary Operator · Formally specified and validated W = 512 / 1024 / 2048 · All confirm ISM = Distinct Structural Basin
§1 · The Space

Where this happens — the realizability manifold

First, we need to understand the space we are working in.

Every physical measurement can be sorted into a structural sequence — a ladder. The UNNS Substrate provides a formal geometric space, the admissibility manifold ℳadm, where all such ladders live. Different physical regimes occupy different regions of this space.

When a system evolves in time, it traces a path through ℳadm — a structural trajectory. Physical boundaries between regimes appear as geometric boundaries in ℳadm, and those boundaries leave measurable signatures in the data.

A 3D visualization of the UNNS admissibility manifold showing different structural regimes (FULL, GIANT, HARD) and the trajectory of Voyager 1 crossing from the heliosheath into the interstellar medium.
Figure 1 — The Realizability Atlas and dominant regimes. Physical systems occupy regions within ℳadm. Voyager 1 follows a trajectory from a heliosheath structural regime into an interstellar one, crossing a geometric boundary. Voyager 2, in contrast, remains confined to the heliosheath region. The atlas is assembled from local charts — the geometry is system-dependent, not global. The geometry here is coordinate-based and empirical; an invariant formulation remains an open problem.

What ℳadm means in plain terms

Think of it as a map of structural states. Every physical observable, at every moment, maps to a point on this map. Different regimes cluster in different regions. A boundary crossing means moving from one region to another — and that movement has a geometric signature detectable from the data alone, regardless of what is being measured.

§2 · The Mechanism

What a structural transition actually looks like

Then, what does a transition actually look like?

When a physical system approaches a realizability boundary, three things happen in sequence — predictably, regardless of domain. Together they form the Boundary Signature Triplet:

Signal 1 κ_conn reaches minimum Maximum boundary proximity Signal 2 Excursions cluster near t* GIANT / TAIL windows concentrate Signal 3 Post-crossing plateau Distinct non-overlapping basin

This triplet appears wherever a boundary crossing exists and is absent in control trajectories without a physical transition. It is diagnostic, not generic — which means it can serve as the basis for a formal detection operator.

A layered geometric illustration showing how a system evolves as it approaches a structural boundary, transitioning from stable organization to fragmentation, with a predicted crossing point.
Figure 2 — The structural signature of a boundary crossing (conceptual). As a system approaches a realizability boundary, connectivity weakens, structural fluctuations cluster, and a new regime stabilizes after the crossing. This transition is detectable before it happens, using the structural estimator t*.
The Structural Boundary Operator — centerpiece result
A computable map from trajectory data to boundary epoch
op[γ]  =  arg mint κ̄conn(t)
subject to: unique minimum  ·  excursions localized near t*  ·  pre/post coordinate ranges disjoint

Returns the boundary epoch from structural connectivity alone. Returns undefined when no boundary is detected — it does not always produce output.

No training data No parameter fitting No physical labels Specified before evaluation Identical across domains Explicit failure modes

The critical property

The operator is not fitted to data. Its admissibility conditions are fully specified before any data is seen and applied identically across all datasets. The result t* = 2012 follows from the Voyager data — not from tuning the method to match it.

📘 Formal foundation

The complete theoretical framework — including six conditional theorems, the Structural Boundary Operator, and the full multi-scale robustness analysis — is developed in the primary manuscript.

Read the full manuscript →
§3 · The Detection

What this looks like in real Voyager data

And how does that signature appear in real physical data?

Voyager 1 crossed the heliopause in August 2012. The 48-second MAG dataset spans 2011–2017, covering heliosheath approach, crossing epoch, and five years of interstellar medium. 3,500 STRUC-PERC-I evaluations (500 windows per annual epoch) produce a structural trajectory of the magnetic field magnitude |B| through ℳadm.

Total runs
3 500
500 windows × 7 epochs
FULL conformance
97.4%
3 409 / 3 500 windows
GIANT excursions
61
69% concentrated in 2011–2012
κ jump at crossing
×2.1
14 686 → 31 057
Structural t*
2012
κ minimum · all 3 scales
ISM 2013 & 2017
100%
FULL · zero excursions
Annual mean κconn — Voyager 1 MAG |B| · 2011–2017 (W=1024 baseline)
HELIOSHEATH CROSSING INTERSTELLAR MEDIUM 0 10k 20k 30k 40k t* = 2012 ▼ ×2.1 jump into ISM 2011 2012 2013 2014 2015 2016 2017
● κconn annual mean (W=1024) ● t* = 2012 · operator output ● ISM basin entry (2013)
A trajectory crossing a boundary in the admissibility manifold, with time-series plots showing changes in structural class, connectivity margin, and fluctuations.
Figure 3 — What a real boundary crossing looks like in data. Three concurrent signals mark the crossing: connectivity margin drops to its minimum in 2012, structural fluctuations (GIANT/TAIL windows) concentrate in 2011–2012, and a stable post-crossing regime begins in 2013. These are the exact signals the boundary operator detects.
The operator returns t* = 2012
Applied to the Voyager 1 MAG corpus, the Structural Boundary Operator identifies 2012 as the crossing epoch from connectivity data alone. This aligns with the known physical heliopause crossing (August 2012) — but the operator did not use that information. The transition is not inferred. It is structurally detected.
§4 · The Confirmation

Before and after — the physical validation

Here is the actual physical confirmation.

Voyager 2 (plasma data, 2007–2018) approached the heliopause but never crossed it in this dataset. Voyager 1 (magnetic field, 2011–2017) crossed it and recorded five years of post-crossing ISM. The two trajectories are not redundant — they are complementary projections of the same boundary object in ℳadm.

Side-by-side comparison of Voyager 2 and Voyager 1 trajectories, showing pre-crossing instability in the heliosheath and post-crossing stability in the interstellar medium.
Figure 4 — Before and after the boundary. Voyager 2 (left) shows pre-crossing behaviour in the heliosheath: regime mixing and boundary-adjacent excursions approaching the heliopause. Voyager 1 (right) shows the post-crossing ISM: stable FULL structure, zero excursions in 2013 and 2017, higher structural connectivity throughout. This contrast validates the structural boundary interpretation. The transition is not inferred — it is structurally detected.

Two-trajectory structural tomography

Voyager 2 resolves the approach side: pre-heliopause softening, FULL→GIANT onset in 2017–2018. Voyager 1 resolves the crossing and post-crossing ISM basin. Together they reconstruct the heliopause boundary from two independent structural perspectives — the first two-sided geometric reconstruction of any physical boundary in the UNNS programme.

Open in Fullscreen!

Open in Fullscreen!

Results Summary

Key Findings

Finding 1

First Structural Boundary Detection

The heliopause is detected at t* = 2012 from connectivity data alone — no physical label, no domain model, no prior knowledge of the crossing location.

Finding 2

Scale-Invariant Result

The same epoch is identified at window sizes 512, 1024, and 2048 samples. The ordering of κconn is preserved across a fourfold scale variation.

Finding 3

ISM as Distinct Structural Basin

Post-crossing κ ranges (27k–36k) are completely disjoint from pre-crossing ranges (14k–18k). No overlap at any tested scale.

Finding 4

Universal Boundary Triplet

The three-signal signature appears at every realized boundary and is absent in negative controls. Diagnostic, not generic.

Finding 5

Two-Trajectory Tomography

Voyager 2 (approach) and Voyager 1 (crossing + ISM) are complementary projections of the same heliopause boundary object in ℳadm.

Finding 6

Operator with Failure Modes

The operator is proved correct under regularity conditions and explicitly returns undefined when admissibility conditions fail — no spurious outputs.

Significance

Why this matters

Physical phase transitions have always been characterized by changes in measurable quantities: temperature, pressure, density. This work takes a different approach. It shows that phase transitions can be detected as geometric events in a structural space, without knowing which physical quantities change or where the boundary is.

  • Excursions are not noise. GIANT and TAIL structural windows near the boundary are signals of geometric proximity. They cluster because the structure of ℳadm requires it.
  • The ISM has a structural identity. Not just a physical location — it is a distinct, measurable basin in ℳadm, separated from the heliosheath by a geometric boundary.
  • The approach generalizes. The same operator, applied identically, has been observed across atomic spectra, cosmological structure, and any domain once embedded in ℳadm.
One-line takeaway
A boundary in physics has been detected without using domain-specific physical models. The transition is not inferred — it is structurally detected.
Scientific Rigor

What makes this result reliable

Scale robustness — all three window sizes confirm t* = 2012

ScaleWindow durationκ minimumt* = 2012?Excursion peakBasin separation
W = 512~6.8 h5 872✓ 2011–12✓ No overlap
W = 1024~13.7 h · baseline14 686✓ 2011–12✓ No overlap
W = 2048~27.3 h47 562✓ 2011–12✓ No overlap

Why this eliminates the segmentation objection

Absolute κ values change with window size — expected and unimportant. The ordering of annual means does not change. The operator depends only on the ordering. Therefore t* = 2012 reflects an intrinsic property of the trajectory, not a property of how the data was segmented.

Evaluations
3 500
Full corpus · 7 years
Scale variants
3
W=512 · 1024 · 2048 · all confirm
Negative controls
2
Monotonic + stable ISM · triplet absent
Theorems proved
6
Conditional · formally grounded

Open in Fullscreen!

Implication

What this establishes

This result changes how phase transitions can be identified. The heliopause crossing is not inferred from physical interpretation, but detected directly from structural organization. A boundary in physics appears as a geometric event in ℳadm, and its location is recoverable from data alone.

  • Boundary detection becomes algorithmic — the transition epoch is obtained from a pre-specified operator, not from post hoc interpretation.
  • Structural signals replace heuristic indicators — connectivity minima and excursion clustering act as direct markers of boundary proximity.
  • Regimes acquire geometric meaning — the heliosheath and ISM are not just physical regions, but distinct structural basins with non-overlapping coordinates.
  • The method is falsifiable — if the signature triplet is absent, the operator returns no boundary.

The decisive test

Apply the Structural Boundary Operator to an independent dataset with a known phase transition and show that it returns the correct epoch without any modification. A single successful replication would establish boundary detection as a domain-independent structural procedure.

Resources & References

UNNS Substrate Research Programme · unns.tech · April 2026 · STRUC-PERC-I v2.5.0 · DLCP 2025