NASA's proposed EVE mission would catch exoplanets under 50 million years old and count the sub Neptunes. The result, roughly 100 or about 5, would settle a decade long debate over whether rocky worlds are born dry or stripped of their gas.
The thousands of confirmed exoplanets discovered over the past three decades have a strange feature: they cluster into two size groups, with a noticeable gap between them. Rocky planets dominate below roughly 1.8 Earth radii; gas-rich sub-Neptunes dominate above it. Almost nothing sits in between. Two competing origin stories could explain the gap, and a proposed NASA telescope called the Early eVolution Explorer (EVE) is designed to settle the question by catching planets while they are still under 50 million years old.
The first explanation, often called the "shrinking gas-dwarf" model, starts with rocky cores that form with thick hydrogen and helium envelopes. When these young planets orbit close to active stars, the stars' intense X-ray and ultraviolet radiation boils the gas away, leaving behind bare rocky cores. In this picture, the radius valley marks where photoevaporation ran out of gas to strip.
The second, the "water-world" model, says the gap is compositional rather than erosional. Rocky planets form dry inside the snow line, where water is frozen solid. Beyond it, sub-Neptunes accrete roughly half rock and half water ice, giving them larger radii from the start. The valley sits where the dry-rock and wet-ice populations meet.
EVE would test these by counting sub-Neptunes in young star clusters, where planets are still under 50 million years old. If the gas-dwarf model is right, the concept paper forecasts roughly 100 sub-Neptune detections in the survey fields. If water worlds are the real story, the number drops to about 5, because those planets are smaller and harder to detect with transit photometry.
The instrument design attacks the hardest part of the problem: false positives. Young stars flare constantly, and those flares peak in the ultraviolet. EVE proposes simultaneous NUV, optical, and near-infrared photometry so the UV channel can flag stellar activity while the optical and NIR channels look for real transit dips. Without that step, a transiting planet signal is nearly impossible to distinguish from a flare on a 30-million-year-old star.
The survey plan calls for 30 visits to young star-cluster fields, 30 days each, spread across a 2.5-year prime mission, covering roughly 20,000 newly formed stars. The sample is the point: only about 20 exoplanets younger than 50 million years have been confirmed so far, against a confirmed-exoplanet count of roughly 6,000. EVE would grow the primordial sample by orders of magnitude in one campaign.
The stakes are not academic. The radius valley is one of the cleanest demographic features in the exoplanet population, and how it forms changes how often rocky, Earth-sized, potentially habitable planets actually occur. If most close-in rocky worlds are the stripped cores of sub-Neptunes, then the universe is rich in Earth-size planets that started as gas balls. If the valley is a compositional boundary set at birth, then Earth's true cousins are the dry-rock super-Earths, and the watery sub-Neptunes are a separate, less-habitable population.
EVE is a concept, not an approved mission. The proposal is a preprint by a large author team led by G. Zhou, positioned for NASA's Small Explorer (SMEX) call. The two hypotheses it would test remain genuinely contested. The radius valley itself is a population-level consensus, not an exact constant: real surveys show it scattered around 1.8 Earth radii, with width and depth that still depend on stellar type and orbital period.
The team's most concrete prediction gives the test a built-in cliffhanger. Under the gas-dwarf picture, the primordial sub-Neptune population is large, and EVE should see about 100 transiting examples. Under the water-world picture, those same sub-Neptunes never existed in the form the transit method catches, and the count collapses to roughly 5. Two years of pointed observation, a hundred-to-one ratio, and a decade of argument resolved or sharpened.