Astronomers Identify New Class of Planet with Sulfur-Rich Atmosphere

An exoplanet with a global ocean of molten rock and an atmosphere laced with rotten-egg gas does not fit any known category. Its discoverers say the taxonomy of planets needs a new entry.

The Discovery

In March 2026, a team led by Dr. Harrison Nicholls of the University of Oxford published a paper in Nature Astronomy announcing that the super-Earth L 98-59 d — a world roughly 1.6 times Earth's diameter, orbiting an M3 red dwarf star 35 light-years away in the constellation Volans — belongs to a type of planet no one had described before [1][2]. The planet had been known since 2019, when NASA's Transiting Exoplanet Survey Satellite (TESS) first spotted three terrestrial-sized worlds transiting L 98-59 [3]. But its true nature only came into focus after the James Webb Space Telescope (JWST) trained its NIRSpec G395H instrument on the system during Cycle 1, capturing a transmission spectrum across a single transit [4].

What NIRSpec revealed was unexpected. The spectrum showed clear signatures of hydrogen sulfide (H₂S) and sulfur dioxide (SO₂) in a hydrogen-dominated atmosphere [4][5]. For a planet roughly five billion years old, retaining that cocktail of sulfur gases is unusual — such volatile compounds typically dissipate over those timescales [6]. Something was replenishing them.

A Planet Unlike Any Other

Astronomers have long sorted small exoplanets into two broad bins. A world of L 98-59 d's size and low density would ordinarily be classified as either a rocky "gas dwarf" wrapped in hydrogen, or a water-rich "hycean" world with deep oceans and ice [1][2]. L 98-59 d fits neither.

The research team used coupled atmosphere-interior evolutionary models to simulate the planet's history from birth to the present day — a span of nearly five billion years. Their conclusion: the JWST data is best explained by a world with a chemically reducing mantle containing greater than 1.8% sulfur by mass, sustaining a permanent global magma ocean that extends thousands of kilometers beneath the surface [5][7]. That magma ocean acts as a reservoir, continuously outgassing sulfur into the atmosphere, where ultraviolet radiation from the host star drives photochemical production of SO₂ [5].

The surface temperature exceeds 1,500°C [6]. The density is unusually low for a planet of this size. The researchers propose that L 98-59 d likely formed with far more volatile material than a typical rocky planet, initially resembling a larger sub-Neptune before shrinking over billions of years as it cooled and lost portions of its primordial atmosphere [2][7].

"This discovery suggests that the categories astronomers currently use to describe small planets may be too simple," Nicholls stated [1].

Source: NASA Exoplanet Archive

Data as of May 1, 2026CSV

What Makes the Atmosphere Distinctive

The atmospheric chemistry of L 98-59 d separates it from other known planet classes in specific ways. Sub-Neptunes typically retain thick hydrogen-helium envelopes with trace water vapor and methane. Super-Earths, where atmospheres have been detected, tend toward carbon dioxide or nitrogen-dominated compositions similar to Venus or Mars. Gas giants like HD 189733 b, where JWST detected hydrogen sulfide in 2024, carry H₂S as one trace component in a massive hydrogen-helium envelope [8].

L 98-59 d is different. It has a hydrogen-rich atmosphere — but one in which heavy sulfur molecules play a defining structural role, sustained by continuous exchange with a molten interior [5]. The in situ photochemical production of SO₂ in an H₂ background is inconsistent with both gas-dwarf and water-world scenarios, according to the Nature Astronomy paper [5]. The >1.8 mass% early sulfur and hydrogen content that the models require is substantially higher than what standard formation models predict for a rocky world at this orbital distance.

The JWST observations were obtained using NIRSpec's G395H grating, which covers wavelengths from roughly 2.87 to 5.14 micrometers — a range where H₂S and SO₂ have strong absorption features [4]. Earlier Hubble Space Telescope observations had hinted at an atmosphere, but lacked the spectral resolution to identify specific molecules [3].

Challenging Formation Models

The existence of L 98-59 d as described poses a problem for the core accretion theory of planet formation — the dominant model explaining how planets build up from dust and gas in protoplanetary disks. Core accretion predicts that rocky planets forming close to their host star, inside the "frost line" where volatiles like water and sulfur compounds remain gaseous, should end up volatile-poor [9].

L 98-59 d orbits close enough to its red dwarf to receive Venus-like levels of stellar radiation (4–25 times Earth's insolation) [10]. Under standard core accretion assumptions, a planet at that distance should not have retained the quantity of sulfur the models now require. The research team's simulations suggest that a protoplanetary disk unusually rich in volatile compounds could produce a world that is molten from surface to core — a scenario not predicted by existing formation models [1][2].

Jo Barstow, a planetary scientist not involved in the study, noted the possibility of extreme volcanic heating comparable to what occurs on Jupiter's moon Io, which could contribute to the planet's sustained magma ocean through tidal interactions with neighboring planets [6]. The L 98-59 system contains at least five planets, confirmed through a combination of TESS transits and radial velocity measurements, with the gravitational interplay between them potentially generating significant tidal heating [3][10].

The Skeptical Case

Not everyone is convinced L 98-59 d represents a genuinely new planetary class. The central question is whether one planet can define a category, or whether it is simply an outlier within the existing super-Earth classification that happens to have an unusual atmospheric composition.

Julien de Wit, a planetary scientist at MIT, has argued that the field should move "beyond boxes" of discrete planet types and instead work toward "establishing a continuum of types" that captures how worlds change and evolve [6][11]. Under a continuum model, L 98-59 d might occupy one end of a spectrum rather than requiring its own label.

There are also observational uncertainties. The atmospheric characterization rests on a single JWST transit observation [4]. A molten magma world best fits the data collected so far, but scientists acknowledge that uncertainties remain — particularly regarding how much sulfur is actually in the atmosphere versus how much the models infer from indirect evidence [6][11]. Stellar contamination from the M-dwarf host star is a known challenge in extracting clean atmospheric signals from transmission spectroscopy of planets orbiting such stars [10].

Settling the debate will require additional transits with JWST to improve the signal-to-noise ratio, along with independent confirmation from complementary instruments. Multiple JWST programs are already allocated to the L 98-59 system, including GTO1201, GTO1224, GO2512, GO3942, GO4098 for transmission spectroscopy and GO3730 for eclipse photometry [10].

The Telescopes, the Timeline, the Funding

The discovery of L 98-59 d's atmosphere involved a chain of missions spanning nearly a decade. TESS, funded by NASA's Science Mission Directorate, first identified the transiting planets in 2019 [3]. The Hubble Space Telescope provided early atmospheric hints. JWST delivered the definitive spectroscopic data during its Cycle 1 observations in 2024 [4][5].

The Nature Astronomy study was a collaboration between researchers at the University of Oxford, the University of Groningen, the University of Leeds, and ETH Zurich [1][2]. While specific grant amounts for this study have not been publicly disclosed, JWST itself represents a $10 billion investment by NASA, ESA, and the Canadian Space Agency, and observing time on the telescope is allocated through competitive peer review.

Looking ahead, the researchers plan to apply their evolutionary simulation framework to data from two upcoming European Space Agency missions: Ariel (expected launch 2029), which will survey roughly 1,000 exoplanet atmospheres, and PLATO (expected launch 2026-2027), which will search for Earth-like planets around Sun-like stars [1][7]. Machine learning methods are being developed to scale the analysis across larger planetary populations.

Source: NASA JWST Program

Data as of May 1, 2026CSV

Habitability: A Sulfur Question

L 98-59 d itself is not habitable by any conventional definition. Its surface is molten rock at over 1,500°C, and it orbits too close to its star for liquid water to exist [6][7]. However, the L 98-59 system contains at least one planet — L 98-59 f, confirmed in 2025 — that does orbit within the star's habitable zone [10][12].

The broader question is what the existence of sulfur-rich atmospheres means for habitability assessments across the galaxy. On Earth, sulfur compounds sustain entire ecosystems. Chemosynthetic bacteria near hydrothermal vents use hydrogen sulfide as an energy source, powering food chains independent of sunlight [13]. Giant tubeworms in the deep ocean rely on symbiotic bacteria that metabolize H₂S, using specialized hemoglobin systems to safely transport the otherwise toxic gas [13].

Research has shown that sulfur-metabolizing microbes can increase atmospheric hydrogen sulfide levels by nearly an order of magnitude over what non-biological processes produce alone [14]. In principle, this makes H₂S a potential biosignature — elevated levels in a habitable-zone planet's atmosphere could indicate biological sulfur cycling. However, volcanism can also produce large quantities of hydrogen sulfide, making it difficult to distinguish biological from geological sources without additional context [14].

If sulfur-rich planets turn out to be common around M-dwarf stars — the most abundant stellar type in the galaxy and a primary target for biosignature searches — the implications for astrobiology are significant. A background of geological H₂S production could complicate efforts to identify biological sulfur signatures, requiring more sophisticated atmospheric models to disentangle the two.

How Common Are These Worlds?

L 98-59 d orbits an M3 red dwarf, the most common type of star in the Milky Way. M-dwarfs account for roughly 70% of all stars in the galaxy [15]. The L 98-59 system is a benchmark for studying atmospheric processes around such stars, with three transiting terrestrial-sized planets receiving Venus-like radiation levels [10].

The researchers describe L 98-59 d as "the first recognized member" of a broader population of gas-rich sulfurous planets sustaining long-lived magma oceans [1][5]. No specific population estimate has been published yet — with only one confirmed example, statistical extrapolation is premature. However, the team has suggested that sulfur-dominated worlds could constitute a "substantial category of planets" once more are identified [6].

The prevalence question hinges on stellar type. M-dwarfs are more likely to host close-in rocky planets subject to the intense UV radiation and tidal forces that the L 98-59 d models require to sustain a magma ocean [10]. G-type stars like the Sun, with wider habitable zones and lower UV flux, may be less likely to produce the specific conditions needed for this planetary type. This creates an observational bias: the planets most amenable to JWST atmospheric characterization (those transiting small, nearby M-dwarfs) are also the ones most likely to belong to this new class.

Source: OpenAlex

Data as of Jan 1, 2026CSV

The volume of research on exoplanet atmospheres has surged in the JWST era, with publications peaking at 2,719 papers in 2024. That research infrastructure is now being directed at characterizing more small planets around M-dwarfs, increasing the odds that additional sulfur worlds will be identified.

What Comes Next

The classification debate is unlikely to be resolved by argument alone. More data will decide it. If JWST follow-up observations confirm the sulfur-rich atmosphere with higher confidence, and if similar signatures appear in other small planets orbiting M-dwarfs, the case for a new planetary class strengthens. If L 98-59 d remains a singular oddity, it may be absorbed into existing categories as an extreme outlier.

The Ariel mission, specifically designed for large-scale exoplanet atmosphere surveys, will be the first opportunity to search for sulfur worlds systematically across hundreds of targets [1]. Until then, L 98-59 d stands as a single, provocative data point — a world that smells of rotten eggs and refuses to be classified.