A lava planet that should be bare rock reflects 74% of its starlight, Webb finds

The James Webb Space Telescope has found that a rocky planet covered in molten rock, roasting so close to its star that a year there lasts less than half an Earth day, reflects light almost as well as Venus. Webb followed the planet around almost its whole orbit and saw it send back most of the starlight falling on it instead of soaking it up.

That should not happen. Lava is dark, and a small rocky world under that much radiation was expected to have lost its air long ago, leaving a black, scorched ball. A bright planet means something sits above the magma: a thick atmosphere, probably carrying reflective clouds made from rock that has boiled off the surface.

That makes TOI-561 b one of the strongest cases yet of a rocky planet beyond the solar system that has kept an atmosphere. JWST has checked a string of small rocky worlds, most of them around red dwarf stars, and found mostly bare rock. Here the planet with the least right to an atmosphere seems to be the one that kept it.

The study, led by Samuel Boucher of the Université de Montréal with Lisa Dang of the University of Waterloo as project lead, appears in The Astronomical Journal. “Our discovery challenges current theories,” said co-author Anjali Piette of the University of Birmingham. “We would expect a heavily irradiated rocky planet like TOI-561 b to have lost its atmosphere billions of years ago.”

How Webb weighed the light of a planet it cannot see

Webb cannot photograph TOI-561 b, which is lost in the glare of its star about 280 light-years away in the constellation Sextans. What it can do is measure the combined infrared glow of star and planet with extreme precision and watch that glow change as the planet turns.

The planet is almost certainly tidally locked, with one face in permanent daylight and the other in permanent night. As it orbits, Webb sees more of the hot dayside, then more of the night side. When the planet slips behind the star, its light vanishes, and the size of that drop gives the dayside glow directly. This rise and fall across an orbit is called a phase curve.

The team pointed Webb’s NIRSpec instrument at the system for about 37 hours without a break, covering nearly four orbits, at infrared wavelengths between roughly 3 and 5 micrometres. The planet’s dayside adds only a few dozen parts per million to the star’s light. To isolate it, the researchers also modelled granulation, the slow boiling pattern on the star’s surface that makes its brightness flicker.

“This technique, known as a phase curve, revealed that the planet reflects a significant amount of light from its star, and that heat is being redistributed around the planet instead of only on the sunlit side,” Dang said. “Together, these point to the presence of an atmosphere.”

Why the numbers rule out bare rock

The argument is an energy budget. A dark bare rock at this distance, absorbing everything and moving no heat to its night side, should have a dayside near 3,200 kelvin, about 2,900 degrees Celsius. Webb measured about 2,050 kelvin on the dayside, close to 1,780 degrees Celsius, and roughly 1,090 kelvin on the night side, around 820 degrees Celsius. Bare rock has no way to warm its permanent night.

The night side is still not warm enough to account for the missing energy, so most of it must be reflected before it is absorbed. From that balance the team derives a Bond albedo, the share of incoming starlight a planet sends back to space, of about 0.74, give or take roughly a tenth. Venus, wrapped in sulfuric acid clouds, sits at about 0.77, Earth about 0.3 and the Moon about 0.1. Laboratory tests show lava is a poor reflector, so whatever is doing the reflecting is not the ground.

A computer model of the planet’s climate backed this up. The team tested atmospheres of water vapour, water with carbon dioxide, and nearly pure oxygen at surface pressures up to 10 bars, ten times Earth’s. A dark bare rock gave a phase curve far larger than the one observed, while bright atmospheres of 1 to 10 bars came closest, whatever the gas mix. The authors suspect the real atmosphere is thicker still, since the planet’s low density points to a deep envelope and a thin layer would not survive billions of years of this radiation.

Clouds made of the stuff of sand

The likely reflectors are clouds of silicate minerals, the family that includes quartz and ordinary sand. The dayside is hot enough to vaporise rock from the magma ocean, and where the air cools, near the edge of the dayside, that vapour can condense into grains of silicon dioxide and magnesium silicate, which scatter light very efficiently at the wavelengths Webb used.

The team tested the harshest case, a planet that reflects nothing and so has the hottest possible dayside. Clouds still formed on the morning side, enough on their own for a Bond albedo of about 0.24. That suggests a feedback: clouds cool the dayside, cooler air lets more clouds form, and the planet brightens further. By turning light away, the clouds also cut the energy that drives gas off into space, acting as a sunscreen for the atmosphere itself. A similar effect has been proposed for the hot Neptune LTT 9779 b, one of the most reflective exoplanets known.

What it doesn’t settle

Nobody knows what the atmosphere is made of. It could be a dense carbon dioxide blanket, a much hotter Venus, or rich in sulfur dioxide like the gases venting from Jupiter’s volcanic moon Io. The climate models fit water-rich and oxygen-rich mixes about equally well, and the authors decline to rank them.

The two NIRSpec detectors also disagree. One shows the hottest spot shifted west of where the models predict, the other a night side colder than any model reproduces. The team says this could be leftover stellar flicker, known drifts in one detector, or real differences in clouds or gases with wavelength. The temperatures rest on a narrow infrared window that misses the peak of the planet’s glow, so they are best estimates.

The deepest problem is why the atmosphere exists at all. Escape theory says a super-Earth this irradiated should be stripped bare. The authors list four ways out: the planet was born gas-rich, as a mini-Neptune; heavy secondary atmospheres leak far more slowly than light hydrogen ones; the planet formed farther out and migrated inward late, after its star’s fiercest youth; or gases locked inside the molten interior are seeping back out, the magma ocean acting as a planet-wide volcano. More than one may be true.

A rocky world with too little rock

TOI-561 b is the innermost of four known planets around a star slightly smaller and cooler than the Sun and about 10 billion years old, twice the Sun’s age. The planet is about 1.4 times Earth’s width and 2.2 times its mass, so it is less dense than Earth and than most planets on such tight orbits, an early hint of a deep layer of lighter material. It orbits about 35 times closer to its star than Mercury does to the Sun, in about 10.7 hours, and receives some 4,700 times the starlight Earth gets.

Common questions about TOI-561 b

What is TOI-561 b?

A rocky super-Earth about 280 light-years away, roughly 1.4 times Earth’s width. It orbits its star in about 10.7 hours, so close that its surface is thought to be a global ocean of magma.

Does TOI-561 b have an atmosphere?

The evidence says yes. Webb measured a dayside far cooler than bare rock should be and a warm night side, which together require a thick atmosphere that reflects light and moves heat around the planet.

Can a lava planet have clouds?

Probably. On TOI-561 b, rock vapour from the magma ocean can condense into silicon dioxide and magnesium silicate clouds where the dayside air cools, and those clouds are the leading explanation for its Bond albedo of about 0.74.

Could anything live on TOI-561 b?

No. With a dayside near 1,800 degrees Celsius and a surface of molten rock, the planet is far outside any range where life as we know it could survive. Its value is what it shows about how rocky planets keep or lose their air.

What comes next

The same Webb data include a transit spectrum, starlight filtered through the atmosphere as the planet crossed its star, which the team will analyse in an upcoming paper to probe the atmosphere’s thickness, make-up and clouds. They are also modelling how the interior and atmosphere evolved together to see which escape route fits.

Webb collected the phase curve over two days in May 2024 under observing program 3860. A first analysis of the dayside, led by Johanna Teske of the Carnegie Institution for Science, was announced in December 2025 and found the planet too cool for bare rock. The phase-curve paper was published online on 29 September 2026, and the University of Birmingham announced the result on 26 October 2026.

The decisive test is a phase curve with Webb’s MIRI mid-infrared instrument, which would pin down dayside and nightside temperatures independently, suffer less from stellar flicker and cover the 15-micrometre band where carbon dioxide leaves a strong mark. “Observing the planet at longer wavelengths with JWST could give us insights and help explain how a planet this hot has held on to an atmosphere for billions of years,” Dang said.

Reference: Boucher et al., “A Highly Reflective Atmosphere on the Lava World TOI-561b Revealed by JWST/NIRSpec Phase Curves,” The Astronomical Journal, 2026. DOI: 10.3847/1538-3881/ae9763

Tags: astronomy, James Webb Space Telescope, exoplanets, exoplanet atmospheres, lava planet, super-Earth