JWST catches star vaporizing the hottest rocky exoplanets

For planets, just like in real estate, the most important property they can possess is “location.” If you want to form a rocky planet, that’s no problem if you’re in the inner part of your stellar system: interior to the analogue of the asteroid belt in our own Solar System. Further out, beyond those frost lines, ices and other volatiles can dominate, giving rise to solid-surfaced worlds that are more akin to giant iceballs than to planets like Mercury, Venus, Earth, or Mars. However, if you venture too close to your parent star, it won’t just be ices and volatiles that get vaporized, but even the layers of the planet itself: atmosphere, crust, mantle, and even the core in the most extreme scenarios.

While “hot Jupiter” exoplanets were among the first planets beyond our Solar System ever discovered, only a few have been caught close enough to their parent stars that their atmospheres are in the process of evaporating. Even more rarely, planets have been caught being completely swallowed by their parent stars. But rocky planets, too, can have their contents evaporated away if they’re heated sufficiently: to conditions where even the heavy elements composing their solid surfaces and interiors can be vaporized and literally blown away. With high enough temperatures, a small, terrestrial-sized planet won’t last long. But for the first time, we’ve caught one definitively vaporizing, and more will surely soon follow. Here’s the story.

This artist’s impression of a hot exoplanet showcases the difference in temperature and brightness between the daytime side and the nighttime side. The illustration of nighttime clouds occurs when vaporized volatiles, during the day, become transported over to the night side and condense. It’s a very ambitious endeavor to try to detect exoplanet atmospheres around small, rocky worlds, and exoplanets that get too hot can have their surfaces and interior vaporize and disintegrate as well.

When you think about the planets of the Solar System, which one do you think is the densest of all? Believe it or not, it’s planet Earth: with an average density (i.e., its mass divided by its volume) of 5.51 g/cm³. However, this doesn’t mean that Earth is made out of the heaviest elements, on average, of any planet; it isn’t. Mercury is actually composed, percentagewise, of more metal than Earth is, and it has fewer low-density volatiles. It’s only because of Earth’s much greater mass that our planet’s atoms, especially in the interior, undergo gravitational compression, and so take up less volume than Mercury’s atoms. If gravitational compression weren’t a part of Earth at all, our density would be much lower: about 4.4 g/cm³.

In other words, based on composition alone, Mercury ought to be the densest planet. But because it doesn’t experience gravitational compression, even though it’s made out of heavier elements, Mercury’s density is a little bit less than that of Earth: at a density of 5.43 g/cm³, or about 1.5% less dense than planet Earth.

Why is Mercury’s composition, on average, made of heavier stuff than Earth is?

To understand that, we have to look at a cutaway of the worlds of our Solar System, and in particular, to notice how much of each world is composed of core material, versus mantle material, versus crustal and atmospheric material.

This cutaway view of the four terrestrial planets (plus Earth’s moon) shows the relative sizes of the cores, mantles, and crusts of these five worlds. Despite the fact that the Earth is only 5% larger in diameter than Venus, it has more mass than Mercury, Venus, Mars, and the Moon combined. If you could pass through the Earth’s interior as a projectile that didn’t interact electromagnetically with the Earth, you would see your trajectory change slightly as you transitioned across one internal layer to another.

Earth is what we think of as the prototype for a rocky planet, and for good reason: we’re here, and it’s the most familiar rocky planet of all. Inside, Earth’s core makes up about 55% of our planet’s internal radius, with the mantle making up almost all of the rest. It’s only the uppermost layer of our mantle that’s bound to the crust, forming Earth’s lithosphere, that represents perhaps the outermost ~4% of our planet’s radius, plus another ~2% or so in atmosphere, if we include that.

But Mercury tells a very different story. When we examine Mercury’s interior, we find that a whopping 85% of its internal radius is made up of its core, with only the outermost ~14% of its radius composing its mantle. Additionally, Mercury’s crust is very thin, making up just ~1% of its volume, and is made largely of solid silicates. In terms of volume, Earth’s core is only 17% of our planet’s volume while Mercury’s core is 57% of its volume.

It’s suggestive, but not conclusive proof, of a scenario where Mercury, being in such close proximity to the Sun throughout the history of the Solar System, didn’t just have its atmosphere, would-be oceans, and ices vaporized away, but had a significant portion of its mantle removed by the Sun as well.

When it comes to the large, non-gaseous worlds of the Solar System, Mercury has by far the largest metallic core relative to its size. However, it’s Earth that’s the densest of all these worlds, with no other major body comparing in density, owing to the added factor of gravitational compression. Unlike Venus, Earth, and Mars, Mercury has no separate crustal layer to speak of.

Of course, Mercury is far from the most extreme planet that we know of in terms of temperature and proximity to its parent star; it’s merely the most extreme such example in our own Solar System. Elsewhere in the galaxy, planets have been observed orbiting their parent stars with surface temperatures that rise into the thousands of degrees, with most of those ultra-hot planets having large volumes to them: gas giant-like worlds.

What would happen, then, if we had a rocky exoplanet, rather than a gas giant one, orbiting extremely close to its parent star?

Obviously, it would lose any atmosphere much more quickly than a gas giant would: there’s less material to start with and a smaller amount of planetary mass to hold onto it gravitationally, and so any atmosphere would quickly vaporize away. But at high enough temperatures, even a world that would otherwise have a solid, rocky surface will melt: possessing a surface of liquid, molten rock. These “lava worlds” aren’t confined to the realm of science fiction, but rather will occur whenever a rocky exoplanet approaches its parent star at a distance of about ~2 million kilometers or less: about 1% of the Earth-Sun distance. At these short distances, surface temperatures on the planet’s surface should reach and exceed a critical threshold of about ~2000 °C: the temperature where rocky materials themselves will vaporize.

The mass, period, and discovery/measurement method used to determine the properties of the first 5000+ (technically, 5005) exoplanets ever discovered. Although there are planets of all sizes and periods, we are presently biased toward larger, heavier planets that orbit smaller stars at shorter orbital distances. The outer planets in most stellar systems remain largely undiscovered, as do Earth-size planets at Earth-like distances around Sun-like stars.

Think about this: here on Earth, we have liquid water oceans, where the liquidity of those oceans requires a thick enough atmosphere to support the existence of water in its liquid phase, where water vapor (i.e., water in its gaseous phase) becomes part of the atmosphere.

What would the analogy be for a heated lava world, with surface temperatures meeting or exceeding that ~2000 °C threshold?

The scenario is similar in some ways, but far more extreme. These small, rocky worlds with lava oceans will still be heated, particularly on the star-facing (i.e., day) side, with such intensity that this heat creates rocky vapor atmospheres: atmospheres that become puffy and diffuse. This is easier for lower-mass rocky planets, like Mercury-mass, but could still happen to larger rocky planets, like Earth-mass or Venus-mass.

When those atmospheres (or atmosphere-like halos) become sparse and diffuse enough, they cool and form dust grains, and so as the planet orbits its parent star, those dust grains form dust tails, similar to how comets form dust tails. Only, in this case, the escaped gas produces tails that orbit both behind the planet and ahead of the planet itself, getting blown away over time by the intense radiation from the parent star.

Photoevaporation is the process by which, when a planet is too close to its parent star or when a star evolves to become very bright, the planetary atmosphere heats up, where stellar emissions can strip particles out of the atmosphere, leading to photoevaporation. For a gas giant planet that orbits a Sun-like star, the act of the star evolving into a red giant can be sufficient to strip even a Jupiter-like or Saturn-like atmosphere away entirely.

We’ve seen this happening in detail for transiting gaseous exoplanets, like the evaporating tail that exoplanet WASP-69b famously exhibits. For gas giant worlds, detecting a transit is easy, as the planet itself is relatively large compared to the size of the parent star, and the transit itself blocks a significant fraction of the parent star’s light. The large dust tail, both the small leading tail and larger trailing tail, can also be seen, as they can block an even greater fraction of the parent star’s light.

It’s more challenging for rocky planets that orbit around stars, however, because of the small size of the rocky planet compared to a giant one. (After all, Earth is just ~9% the radius, and ~0.8% the cross-section, of Jupiter, while Mercury is just ~38% the radius, and ~15% the cross-section, of Earth.)

This explains why detecting Earth-sized planets around Sun-like stars is so challenging, even with the transit method. However, if we look to smaller stars (i.e., low mass, red dwarf stars) and we start searching for these light-obscuring dust tails instead of just a solid planet itself, the search becomes far more interesting and potentially fruitful. After all, detecting “flux dips” in a star’s life over time is how many planet-finding missions have worked, from Kepler to K2 to TESS. It’s data from these missions, and others like them, that can identify candidates for evaporating, rocky exoplanets.

Around the star WASP-69, a “hot Jupiter” exoplanet has its outer layers of atmosphere photoevaporated away, creating a comet-like tail. Even more extreme systems are known to exist, with lower-mass planets orbiting at even closer distances to their parent stars. These rocky exoplanets should evaporate very quickly, and might even have mantles or cores that vaporize, not just their atmospheres.

One fascinating discovery, announced earlier in 2025, is of evaporating rocky exoplanet BD+05 4868 Ab: located just 140 light-years away. The planet is very close in toward its parent star, completing a full revolution every 30.5 hours. Its flux dips are severe, blocking between 1.0-1.5% of its parent star’s light. This isn’t a red dwarf star, either, but a K-class star: stars that are just a little lower in mass and radius than our Sun. The amount of light that’s blocked is too great for a rocky exoplanet alone, but rather indicates a large trailing tail of dust and a small (but existent) leading tail, where the trailing tail is ~9 million kilometers long, and where the leading tail is made primarily of larger dust grains (~10 microns each) while the trailing one is made of smaller (~1 micron) grains.

The volume of dust in the tail tells us how much mass the planet is losing, and it appears to be losing about 1 lunar mass every million years. A planet the size of the Moon or Mercury would completely disintegrate in only a few million years, while a planet even the mass of Venus or Earth would only persist for several tens of millions of years; it would evaporate in under 100 million years. This exoplanet in particular, BD+05 4868 Ab, is an ideal candidate for follow-up transit spectroscopy with JWST, as JWST could reveal whether it’s the atmosphere, crust, mantle, or core that’s vaporizing.

This graph shows the flux dip data for exoplanet BD+05 4868 Ab, at left, along with the asymmetric detail of the flux dips together, at right. The dipping pattern indicates a small leading tail and a large, long, diffuse trailing dust tail.

Even though that’s the clearest example for a rocky exoplanet that definitely has a vapor tail, and one that’s at such a close distance to Earth at just 140 light-years away, we’ve found evidence for disintegrating exoplanets before. One remarkable example is KIC 12557548 b, which was discovered back in 2012 and appears to be an evaporating super-Mercury-like world. Unfortunately, that star is nearly 2000 light-years away, and so performing transit spectroscopy on it, even with JWST, would be a challenge that’s too great for the observatory’s capabilities.

Fortunately, however, there’s a “sweet spot” transiting, low-mass exoplanet that did just recently get observed with JWST: K2-22b. This exoplanet is just 800 light-years away, is very low in mass (comparable to the mass of Mercury or the Moon, rather than Earth), and exhibits widely varying transit depths that range from 0%, where it’s undetectable, to ~1.3%, which means that an enormous amount of the star’s total light (more than 1%) gets blocked. The parent star, K2-22, is on the hot, massive side of red dwarfs, shining at a temperature of ~3800 K, where the planet orbits with a period of only 9 hours: one of the shortest exoplanet orbits known.

When an evaporating exoplanet transits in front of its parent star, as this illustration of evaporating exoplanet KIC 12557548 b shows, it isn’t just the disk of the planet that blocks the parent star’s light, but the tail of dusty debris that follows (or, in some cases, precedes) it.

In April of 2024, a team of observers observed this exoplanet transiting its star with JWST’s MIRI instrument, and then took on the great challenge: trying to identify what the planet’s “vapor” that blocks the star’s light is made out of. What the team found, initially, wasn’t all that surprising: it very clearly wasn’t a “bare core” planet, as compounds like iron, iron oxide, or iron sulfide don’t have the right spectral signatures at all. They then moved on to considering what would happen if this were a planet with an evaporating mantle: a mix of magnesium and silicon compounds, particularly oxides. The match was a lot closer, but still didn’t quite match exactly.

Why didn’t it match?

That mismatch, as first author Nick Tusay put it, was “our embarrassing, forehead-slapping moment,” as they realized they forgot about vapor. If this were a vaporizing exoplanet, there should be species of vapor there: what you’d expect would be magnesium oxide, as well as silicon and silicate oxides. But what they needed to add in to match the spectrum are nitric oxide (NO) and carbon dioxide (CO₂) molecules. Somehow, the results are suggestive that not only is this an exoplanet whose mantle is evaporating right now, but that it still possesses vapor that’s consistent with icy volatiles.

This graph shows the MIRI spectral data for evaporating rocky exoplanet K2-22b, as first presented at January 2025’s American Astronomical Society meeting. The data is consistent with a vaporizing magnesium and silicate mantle signatures augmented by volatile ices nitric oxide and carbon dioxide.

And that’s where we are today: in the early, infant stages of finding out just how exoplanets evaporate when they’re close to their parent stars. It isn’t just their atmospheres that can get blown away, vaporized, and emitted as dust tails, but their interiors as well. Hot enough exoplanets, even rocky exoplanets, can melt, become lava worlds, can have that lava vaporize, and can have that vapor halo get blown away by their parent star’s close proximity. Over time — even relatively short timescales ranging from only a few million to perhaps hundreds of millions of years — hot, rocky exoplanets can evaporate entirely.

This motivates further study of low-mass exoplanets that exhibit dust tails that transit in front of their parent stars, as transit spectroscopy can reveal, particularly for close, bright systems, the atomic, ionic, and molecular constituents of these tails themselves. We can watch, if we make the right observations, exoplanets evaporate in real-time, including potentially atmospheric, mantle, and even planetary core evaporation if we probe the right systems. It was, after all, less than two decades ago that we began discovering our first rocky exoplanets, and used the transit method to successfully identify them. Now, in the JWST era, we’re learning not only what they’re made of, but how the unlucky, hot ones disintegrate before our eyes.