Reasoning:
- Survival modeling: a strangelet of rapidity y in the lunar frame survives stopping with P = exp[-4.85(1 + A^(1/3)/3)^2 (cosh y - 1) A/Z^2] (geometric dissociation cross-section, ionization-only stopping in silicon); for A=20, Z=-1 this is exp[-350(cosh y - 1)] — only near-rest strangelets survive, the same physics that limits what RHIC itself could stop (RHIC-side survival probabilities ~1e-2).
- With the physically motivated rapidity form dPi/dy ∝ y^a exp(-by) near the kinematic limit (inclusive hadronic processes fall as powers of rapidity at phase-space edges; exponential suppression at central rapidity would violate the emergence of a rapidity plateau), the net lunar-vs-RHIC suppression ranges from an enhancement of 3 (a=1) to 3e-6 (a=4). Adopting 1e-6: exposures of 1e22 p (Fe-Fe AGS), 2e16 p (Fe-Fe RHIC), 1e17 p (Au-Fe AGS), 2e11 p (Au-Fe RHIC) → bounds p < 2e-11, 1e-5, 2e-6, 1 for cases I-IV.
- Case selection: because production physics favors lower energies and iron is a perfectly good heavy nucleus, case I is the one to take seriously → p < 2e-11, where p (= 2e11 x the per-collision, survival-weighted probability) is already the total probability over RHIC’s whole lifetime. Only by insisting on exact RHIC circumstances plus the delta-function-at-central-rapidity model of Dar-De Rujula-Heinz (for which no theoretical motivation exists and against which heavy-ion phenomenology argues) does the lunar limit evaporate — and those same assumptions do nothing to make production at RHIC more likely.
- Metastability does not reopen this loophole: a strangelet made in the lunar rest frame has exactly as long to react as one made at RHIC. The paper’s honest bottom line: with worst-case assumptions that bend if not break the laws of physics, no totally satisfactory purely empirical limit exists — which is why the theoretical production/charge/stability arguments, not the cosmic-ray bounds, are called the central case for safety.
Step 6 verdict
Verdict: approved (checked). Load-bearing step: the lunar exposure only bounds p if produced strangelets survive stopping, and the survival exponential exp[-350(cosh y - 1)] plus a power-law (not exponentially suppressed) tail near central rapidity leaves effective exposure ~1e22 x p even after a punishing 1e-6 suppression factor - hence p < 2e-11 for case I. The single surviving escape (production only in Au-Au, only at RHIC energies, only exactly at CM rest - the DDH conjunction) is internalized in the statement and correctly marked physically unmotivated, so it does not undercut the stated conditional conclusion. Case-I selection follows from the A-58 energy-scaling premise. Anthropic selection does not bite (a Moon-converting strangelet leaves Earth observers). The metastability loophole is correctly shown closed for this particular bound (equal reaction time in lunar and RHIC rest frames). Valid conditional on the stopping/survival model.