A lost planet could explain Barnard’s Star’s unusual orbital pattern

Barnard’s Star, the nearest single star to the Sun, hosts four known planets with measured minimum masses below Earth’s. A new study by Tian Yi and colleagues asks whether their tightly packed orbits preserve evidence of a fifth planet—not one waiting to be discovered, but one that disappeared in a merger.

Their computer simulations identify two possible routes to that outcome. Neither demonstrates that the collision happened, but both offer explanations for an otherwise awkward orbital pattern.

An incomplete orbital rhythm

From the star outward, the planets are named d, b, c and e. They complete their orbits in approximately 2.34, 3.15, 4.12 and 6.74 days.

The inner pair’s orbital periods lie near a 4:3 ratio: the inner planet makes roughly four circuits while its neighbor makes three. Planets b and e, despite having c between them, lie relatively near a 2:1 ratio. Planet c does not fit neatly into a sequence of neighboring low-order resonances.

An orbital resonance is more than a numerical resemblance. It involves a repeating relationship between planets’ orbital positions that can organize their gravitational interactions. Three planets can also participate in a shared resonance, even when individual pairs are no longer resonant. Near-integer period ratios alone do not establish that these relationships exist today.

The researchers propose that the system began with five planets linked in a resonant chain. Two intermediate planets, labeled c1 and c2, subsequently merged into the present planet c, breaking some links while leaving traces of others.

Testing four possible histories

The team used REBOUND and its REBOUNDx extension to simulate the planets’ mutual gravity, supplemented with models of tidal dissipation and gas-disk forces. They first assembled resonant chains through simulated migration, then followed their subsequent evolution.

Two scenarios began with only four planets and required no collision. In one, tides dissipated energy and increased orbital period ratios over billions of years. Among 100 simulations, 90% lost the relevant resonant configuration before reaching the observed ratios. The remainder stayed too close to exact resonance to reproduce the measured offsets.

The other four-planet scenario used an outward-moving inner edge of the young star’s gas disk to carry planets outward. None of the 18 tested runs reproduced all three neighboring period ratios. The outermost planet, e, was too low in mass to keep pace effectively with the retreating edge. The authors note that a differently tuned disk-dispersal history might improve agreement, so this result does not exclude every four-planet origin.

Two routes to a merger

The five-planet models were more promising.

In the first, the resonant chain survived the disappearance of the gas disk. Tidal evolution then gradually changed the planets’ relative spacing over billions of years, eventually breaking the resonance and allowing their paths to cross.

Of 75 simulations, nine produced the intended c1–c2 merger. Five—6.7% of the total—left four planets that remained stable for a further 10 million simulated years. Their period ratios generally approached the observations, although the outermost pair tended to end up somewhat too widely separated.

The second route operated while the gas disk still existed. Planet e remained near its inner edge, while planets farther inside the largely cleared cavity still felt gravitational pulls from the exterior gas. These interactions, called Lindblad torques, removed orbital angular momentum and drove inward migration, most strongly affecting c2.

That uneven forcing disrupted the chain. Of 232 runs, 110 produced the intended merger and a four-planet system surviving at least 10,000 years. This roughly 47% merger frequency is not an exact-match rate or a probability of the system’s actual history. Matching the observed configuration also required the disk to disperse near the right moment, stopping further migration.

A possible history, not a recovered event

The results depend on assumed starting chains, uncertain planet masses and simplified tidal and disk prescriptions. Radial-velocity observations measure minimum masses; the true masses depend on orbital inclination. The two merger scenarios also received different post-merger stability checks, limiting direct comparisons of their success rates.

A broader search identified 14 compact four-planet configurations with similar broken-chain patterns. Ten had the third planet as the apparent break. That resembles the proposed disk-driven outcome, but does not establish a shared collision history.

One potential discriminator lies in the inner pair. None of the five long-lived tidal-merger outcomes retained its two-planet 4:3 resonance, whereas 34 of the 110 disk-driven merger outcomes did. Future measurements establishing that resonance could therefore favor the disk-driven route over the stable tidal outcomes sampled here.

Posted to arXiv on November 10, 2026, the preprint supplies testable possibilities—not observational proof of a vanished planet.