Preprint: Strong helium outflow detected on sub‑Neptune GJ 3090b could strip its gas in ~300 million years

A new arXiv preprint reports one of the strongest helium signals yet seen from a sub-Neptune, arguing that the nearby planet GJ 3090b is shedding its atmosphere in a vigorous outflow and could lose its hydrogen-helium envelope on a timescale of roughly 300 million years. If that interpretation holds up, it would sharply shorten earlier estimates for how long the planet could keep its gas.

The study, “WINERED Detects a Strong Atmospheric Outflow on the Sub-Neptune GJ 3090b,” by Vissapragada and colleagues, was posted as arXiv:2609.14704 and says it was submitted to The Astrophysical Journal Letters. It has not yet been peer reviewed. The team observed two consecutive transits of GJ 3090b on Sept. 22 and Sept. 25, 2024, using the WINERED high-resolution near-infrared spectrograph on the Magellan II, or Clay, 6.5-meter telescope at Las Campanas Observatory in Chile. They report peak helium absorption during transit of 2.033 ± 0.086% on the first night and 1.619 ± 0.122% on the second, with quoted detection significances of 23.6 sigma and 13.3 sigma.

The signal comes from metastable helium at 10,833 angstroms, a spectral line that has become a key way to probe escaping exoplanet atmospheres. Unlike many transit measurements that mainly sample lower atmospheric layers, this helium feature traces gas high above the planet, where it can reveal material flowing away into space. In the new paper, the fitted helium line had a full width at half maximum of about 25 kilometers per second on both nights, with no significant Doppler shift. The authors say that broad profile is more consistent with photoevaporation — gas driven off by stellar high-energy radiation — than with a slower core-powered wind. Their abstract says, “The signal is variable in amplitude and deeper than anticipated by the JWST/NIRISS observations of Ahrer et al. (2025).”

GJ 3090b is the kind of planet astronomers care deeply about in atmospheric-loss studies: a short-period sub-Neptune, the most common type of close-in planet known. It circles an M2V star every 2.85 days at a distance of about 22 parsecs from Earth. Recent measurements put the planet at 4.52 ± 0.47 Earth masses and 2.18 ± 0.06 Earth radii. The discussion in the preprint adopts a system age of about 1 billion years, placing the planet in what the authors describe as an “adolescent” phase.

The two transit observations were not identical. The first was affected by a stellar flare near egress, while the second was not, and the paper says the helium signal differed between the nights at about 2.8 sigma. The authors’ hydrodynamic modeling points to mass-loss rates on the order of 10^11 grams per second for models that match the helium data. Using that rate and a small hydrogen-helium envelope fraction, they estimate an atmospheric-loss timescale of about 300 million years.

That is far shorter than the roughly 50 billion years reported by Ahrer and colleagues in 2025 from JWST/NIRISS observations, which found a much smaller helium signal of 434 ± 79 parts per million and a largely muted transmission spectrum. The new paper says the difference is not necessarily a direct contradiction. Spectral resolution can dilute narrow helium features in lower-resolution data, and the signal may also vary over time. Earlier VLT/CRIRES+ observations found no robust molecular detections in the lower atmosphere, a result the new study says could still fit a picture in which the lower atmosphere is obscured by aerosols while the upper atmosphere drives a strong hydrogen-helium outflow.

The authors also argue the atmosphere’s bulk metallicity is likely less than about 100 times solar, because more metal-rich models struggle to reproduce the strong helium absorption. Their main point, though, is evolutionary: GJ 3090b may not be in a long-term steady state, but in an active transitional stage of atmospheric loss. As the preprint puts it, “This adolescent (~1 Gyr) sub-Neptune is currently in a transformative phase of photoevaporative evolution.” Peer review and follow-up observations will determine whether that shorter atmospheric lifetime stands.