A Jamming–Unjamming Mechanism for Rapid Continental Break-up

This white paper is the convergence node of the Vacuum-Physics research program. Where the foundational paper derives physics from a single premise — the vacuum as a jammed elastic solid — this one asks a focused, falsifiable question downstream of that premise: is a granular jamming → unjamming → liquefaction → suction → sliding mechanism physically capable of producing a rapid (kyr-scale) phase of continental break-up? The opening of the Atlantic is used as the test case. The work is presented as a disciplined, falsifiable framework, not a confirmed claim: it formalises one mechanism and separates, throughout, claims, observable signatures, tests, and falsifiers, so that any pre-registered criterion that is not met sends the corresponding conclusion to HOLD (insufficient evidence) or STOP (a physics No-Go). “Self-consistent and reproducible” is stated explicitly to be a weaker property than “externally confirmed,” and the load-bearing claims are held open until external evidence arrives.

The engine is read in one direction only, and the inverse ordering is forbidden. A trigger — Pacific-side uplift driving an antipodal tensile rupture — nucleates a transient void, which is not a literal vacuum but a pressure deficit: a region of reduced confining pressure produced by dilatancy and the unjamming of the basal material. The deficit exerts a suction that pulls the plate from the front rather than pushing it from behind (claim C2), and the resulting acceleration is sustained only if the basal resistance can collapse. The central physical content of the paper is therefore the resistance-collapse step, and the central honesty of the paper is that it does not import that step as a fault-mechanics assumption but derives it from the program's own granular physics.

That derivation is the paper's load-bearing link to the rest of the program. The basal shear zone is treated as a granular gouge in one of two states — jammed (a load-bearing contact backbone with a finite yield stress) or unjammed and liquefied (the backbone dissolved, viscous flow) — with the mean contact coordination number as the control variable. Frictional heating drives the gouge toward isostaticity, where the relaxed shear modulus and the yield stress vanish while the bulk modulus stays finite, so the load-bearing network is lost before any melting. The same marginal, just-rigid state and its unjamming flow are the subject of the program's configured-continuum paper; the rule form used here is derived from first-principles soft-disk jamming simulations and reproduces the classical and spine results — a jamming fraction near 0.84, an isostatic coordination of six in three dimensions, and a relaxed shear modulus that scales linearly to zero with the excess contact number. The outcome is an effective friction coefficient of order 2×10⁻³ reached at a temperature rise of only tens of kelvin — about one percent of the energy a melting account would require — and a bistable, velocity-weakening, stick–slip rheology in which an apparent slow rate is the time-average of brief, fast slips. The chapter on this mechanism is the place to start, because every later quantitative claim depends on it.

The rheology is not merely permissive; it is constraining. Because the liquefied state is self-consistent only while shear heating outruns conduction, there is a critical sliding velocity below which the liquefied branch cannot exist, and that velocity exceeds plate-tectonic creep rates by roughly eight orders of magnitude. The system therefore has no stable slow-liquefied branch: it is stuck or it is fast. A liquefied zone of millimetre thickness re-jams by conduction in of order ten seconds once slip stops, so the fast phase is intrinsically self-limiting rather than open-ended. This is the honest replacement for any appeal to “slow residual viscous creep,” and it converts the rapid-motion picture from a qualitative story into a falsifiable physical regime with definite signatures in basal shear zones — gouge fluidization, injection and hydrofracture textures, and grain-intact but fabric-destroyed material, with frictional melt expected only on segments that failed to unjam.

Feasibility is then locked into numbers rather than asserted. A unified energy ledger separates three stresses that would otherwise read as a contradiction: the suction pressure deficit as energy input, the liquefied friction as heat, and the mean-rate drive as the residual kinematic term. The heat turns out to be only a few percent of the input work; conduction alone is honestly stated to be insufficient to remove that heat over a kyr event, and the residual is converted into a falsifiable advective hydrothermal-flux prediction rather than swept aside. Four items are carried, deliberately and quantitatively, as bounded HOLDs — not hidden, because hiding them would be the failure mode this program is built to avoid.

The first HOLD is the laboratory-to-continental master-scale extrapolation: the material law is scale-invariant as a material property, but its application across the many orders of magnitude from a soft-disk simulation to a continental margin is gated by an explicit set of No-Go limits, and the conclusion is graded HOLD wherever those limits are approached. The second is the energy source — the paper's self-identified most vulnerable link — which is bounded so that only the release of stored internal energy survives as an admissible class, with externally implausible sources excluded by pre-registered limits rather than by preference. The third is an event-window coherence statistic that was deliberately downgraded from a PASS to a HOLD once its engine was found to test the wrong null hypothesis; reporting that downgrade, rather than quietly keeping the PASS, is treated as an integrity requirement. The fourth is the absolute chronology, which is where this paper meets the program's accuracy-limit work.

The chronology HOLD is handled by a deliberate firewall, and it is the paper's clearest link to the cross-chronometer accuracy limit. Magnetic stripes record a product — spreading rate times polarity-chron duration — so the spatial pattern is invariant under a common rescaling of time and cannot, by itself, discriminate a uniform time-compression; a periodic-resonance reading moreover fails on the record's own strongly non-periodic reversal statistics, but that failure vindicates no particular timescale. The absolute timescale is set instead by independent absolute dating. Rather than contest that dating, the paper concedes it at full strength: the high-precision U–Pb break-up ages near 201, 134, and 55 million years rest on igneous baddeleyite and cogenetic zircon, are cross-checked by dual-decay concordance, and are corroborated by biostratigraphy and the bilaterally symmetric ocean-crust age gradient. The companion paper explains the general mechanism by which incorporated foreign older material can bias an age old — and this paper states plainly why that mechanism does not apply to igneous baddeleyite, declining to invoke inheritance, recycling, or contamination to deflate the ages. The claim is therefore strictly physical and chronology-agnostic: the mechanism is capable of a rapid rupture phase if the bounded boundary conditions obtain; whether such a phase occurred is a separate, pre-registered question routed to the falsifiers, and the load-bearing case — rapid initial rupture followed by conventional spreading — does not require the compressed timescale at all.

The predictions themselves are organised so that the procedure cannot quietly rescue the hypothesis. Thirty-five predictions (P1–P35) are pre-registered with explicit gates: seafloor-margin and lubrication signatures sit on the cause axis; freshwater, sea-level budget, drainage-reorganisation, and endorheic-basin signatures sit on the response axis; a joint event-window statistic asks whether the supposedly linked events actually cluster in time; and a negative-control registry checks that the same tests do not “confirm” the mechanism where it should be absent. Two of the most decisive predictions are written as external falsifiers tied to the deep-time ages, so that the chronology that the paper refuses to contest is also the chronology that could most cleanly refute the rapid-opening variants.

The argument is delivered as a reproducibility bundle rather than as prose alone, which is what makes the gates meaningful. About fifteen deterministic NumPy engines compute the load-bearing quantities — the jamming constants, the friction-collapse temperature, the critical velocity, the energy ledger, the stripe time-rescaling degeneracy, and the event-window statistic — and a single-command validator re-runs them and reports thirty-nine internal self-consistency checks. Integrity hashes pin every input file, a machine-readable constraints file records the locked numbers, and a ledger records what was changed and why. Each numeric input is checked against the primary literature and cited, so a reader can separate what is computed from what is assumed and can re-derive any headline figure without trusting the author. This is the operational meaning of “reproducible” in the gating discipline: not that the conclusions are correct, but that they are exactly as strong, and exactly as open, as the bundle says they are.

The whole is held to one discipline: LOCK → Derive → Gate. Inputs are locked, results are derived by rule with no parameter tuning, and outcomes are verified by count and gated as PASS, HOLD, or FAIL. Several reported verdicts are deliberately negative or HOLD; these are integrity features, not omissions. Because the master-scale extrapolation and the absolute chronology remain open, strong statements such as “unique cause” or “decisively proven” are disallowed under the project rules — the contribution is a bounded demonstration of physical possibility, derived from the program's foundational jamming physics and disciplined by its accuracy-limit work, with an explicit and falsifiable path to its own refutation.