When a molecule breaks apart under ionization, the standard picture is ballistic: bonds break, fragments fly away with the excess energy, and the products are determined by which bonds were weakest. The fragments don't look back.
Note (2026-11-14; read against the paper's full text): the essay says the oxygen-oxygen bond forms not through a direct rearrangement pathway but through the roaming, and that the standard picture is ballistic. The paper reports two pathways to the C+ plus O2+ products. One is a bond rearrangement through a triangular intermediate, reached by structural deformation; the other is the roaming pathway, in which a frustrated C-O cleavage leaves an oxygen atom without enough energy to escape and it wanders over a flat region of the potential energy surface. The measured kinetic energy release spectrum of that channel has two dominant peaks (about 6.4 and 7.6 eV) that the paper fits with a branching ratio of about 55 to 45. The experiment is electron-impact double ionization at 200 eV (chosen to match free electrons in Mars's atmosphere), with the ionization mechanism determined from the projectile energy loss; the roaming details come from potential energy surface and trajectory calculations. The essay's account of the landscape deciding and of a traveler who missed an exit is the essay's own.
Ma, Hao, Zhou, Xue, Zeng, Li, Wang, and Ren (arXiv:2509.05626) find something stranger in doubly-ionized CO₂. When a C-O bond attempts to break but the oxygen atom doesn't receive enough energy to escape, it doesn't simply fall back. Instead, it roams — wandering through flat regions of the potential energy surface, exploring varied configurations, before eventually encountering and bonding with the other oxygen atom to form O₂⁺.
The product is C⁺ + O₂⁺ — a molecular oxygen ion that didn't exist in the parent molecule. The oxygen-oxygen bond forms not through a direct rearrangement pathway but through a frustrated dissociation that turns into an extended search. The roaming atom acts like a traveler who missed their exit and discovers a new destination while lost.
The structural insight: the flat potential energy landscape between dissociation and recombination is not empty space. It's a navigation surface where the outcome depends on which configuration the roaming atom happens to visit. The same initial conditions (double ionization of CO₂) can produce either standard fragmentation or this roaming pathway, depending on whether the first bond-breaking succeeds or fails. The product diversity comes not from energy differences but from a topological feature of the potential surface — a flat region that permits exploration.
The atom doesn't know where it's going. The landscape decides.