K2-18b has produced some of the most tantalizing atmospheric data ever collected from a planet outside our solar system, but the case for life there remains far from settled. JWST observations confirmed methane and carbon dioxide in the planet’s hydrogen-rich atmosphere, a combination consistent with a world that could harbor a liquid water ocean beneath its thick gas envelope. A faint hint of dimethyl sulfide, a molecule produced by marine life on Earth, made international headlines in 2023, yet independent reanalyses and new data have since called even that tentative detection into serious question. What we have is a genuinely intriguing planet whose atmosphere we can now study in real detail, not proof of biology.
What JWST Actually Found in K2-18b’s Atmosphere
K2-18b orbits a red dwarf star about 124 light-years from Earth. It weighs roughly eight times as much as our planet, placing it in a size category called sub-Neptunes, larger than Earth but smaller than Neptune, a type of world with no counterpart in our own solar system. Earlier measurements using the HARPS spectrograph pegged its mass at about eight Earth masses and its density at roughly 3.3 grams per cubic centimeter, loose enough to suggest either a rocky core wrapped in a substantial gas envelope or a world whose mass is more than half water.1Astronomy & Astrophysics. Characterization of the K2-18 multi-planetary system with HARPS Before JWST, observations with the Hubble Space Telescope had already detected water vapor in its atmosphere, making it one of the first habitable-zone planets of its size class with a confirmed atmospheric signature.2Nature Astronomy. Water vapour in the atmosphere of the habitable-zone eight-Earth-mass planet K2-18 b
The JWST data, collected using the NIRISS and NIRSpec instruments across wavelengths from 0.9 to 5.2 micrometers, delivered something more specific. The transmission spectrum showed strong signatures of methane and carbon dioxide, at roughly five-sigma and three-sigma confidence respectively, with each gas present at mixing ratios of about one percent in a hydrogen-dominated atmosphere.3The Astrophysical Journal Letters. Carbon-bearing Molecules in a Possible Hycean Atmosphere Those are robust detections by astronomy standards. The study that reported them also noted a tentative signal from dimethyl sulfide, or DMS, a sulfur-containing organic molecule. On Earth, DMS is overwhelmingly produced by phytoplankton in the ocean, which is why it caught everyone’s attention. But “tentative” is doing a lot of work in that sentence, as later sections will make clear.
The Hycean World Idea
K2-18b’s atmospheric composition lines up with predictions for a theoretical class of planet called a Hycean world: hydrogen-rich atmosphere sitting on top of a global liquid water ocean. The concept is appealing because it dramatically expands the kinds of planets that could support conditions friendly to life. Unlike a rocky Earth-like planet, a Hycean world does not need to be exactly the right size or distance from its star; the thick hydrogen blanket could maintain warm surface temperatures across a wider range of orbits.
The ocean on such a world, if it exists, would be unlike anything familiar. Modeling work suggests that Hycean ocean depths could range from roughly 20 kilometers on a dense, iron-rich planet to around 1,000 kilometers on a lighter, water-dominated one, depending on the planet’s mass, water content, and surface temperature.4Monthly Notices of the Royal Astronomical Society. On the ocean conditions of Hycean worlds For comparison, Earth’s deepest ocean trench bottoms out at about 11 kilometers. Even a shallow Hycean ocean would be a vast, high-pressure environment completely unlike a terrestrial sea.
There is a catch, though. A hydrogen-helium atmosphere acts as a powerful greenhouse blanket. Calculations show that for a planet receiving solar-like amounts of starlight, as little as ten bars of hydrogen-helium atmosphere could drive surface temperatures past the supercritical point of water, meaning the “ocean” would cease to be a liquid in any meaningful sense.5The Astrophysical Journal. The Runaway Greenhouse Effect on Hycean Worlds K2-18b receives less radiation than Earth does from the Sun, which helps, but there is a narrow window of atmospheric thickness that keeps the surface warm enough for liquid water without overshooting into a runaway greenhouse. Whether K2-18b sits inside that window is one of the open questions.
Why the Biosignature Claims Are Disputed
The original 2023 analysis of K2-18b’s near-infrared JWST spectrum reported methane and carbon dioxide with confidence, but the DMS signal was always described as tentative. Independent teams have since reexamined the same data and come away with a bleaker assessment. One comprehensive reanalysis found no reliable evidence for either carbon dioxide or DMS in the transmission spectrum. Depending on which combination of the NIRISS and NIRSpec datasets was used, the statistical support for COâ‚‚ topped out at about 2.3 sigma, and DMS at roughly 2.2 sigma, levels described as “weak evidence at best” on standard scales of statistical confidence.6The Astronomical Journal. A Comprehensive Reanalysis of K2-18 b’s JWST NIRISS+NIRSpec Transmission Spectrum In particle physics, a detection claim requires five sigma. In exoplanet atmospheric science, thresholds are more relaxed, but two sigma is not enough to bet on.
The situation got more complicated when JWST’s mid-infrared instrument, MIRI, was brought in to look for DMS and a related molecule, DMDS, at longer wavelengths where they absorb more strongly. Those newer observations yielded only marginal signals. None of the sulfur molecules exceeded three sigma in model preference, and all dropped below about two sigma unless analysts imposed a strong haze model that is itself debatable.7The Astrophysical Journal Letters. New Constraints on DMS and DMDS in the Atmosphere of K2-18 b from JWST MIRI A separate study examining the MIRI data found that the results changed substantially depending on arbitrary choices in how the raw data were reduced, such as which starting pixel was used to bin the spectral channels. Some choices gave a slight preference for DMS; others pointed to ethyl chloride instead; still others gave flat, uninformative posteriors.8The Astronomical Journal. K2-18b Does Not Meet the Standards of Evidence for Life When the answer shifts depending on where you start counting pixels, the instrument is telling you it cannot clearly see what you are looking for.
The methane detection remains solid. Nobody disputes that K2-18b’s atmosphere is rich in methane. But methane alone does not point to biology; it is abundant on worlds with no life, including the gas giants in our own solar system. The whole excitement was about DMS, and the data are not cooperating.
DMS Does Not Require Life
Even if DMS were eventually confirmed in K2-18b’s atmosphere at robust confidence levels, it would not prove that something is alive there. On Earth, DMS is primarily a biological product: marine phytoplankton release a precursor compound, and bacteria and chemical processes in seawater convert it to DMS, which then enters the atmosphere.9PubMed Central. Evidence that dimethyl sulfide facilitates a tritrophic mutualism between marine primary producers and top predators That biological origin is what made DMS seem like a promising biosignature. But recent work has shown that the molecule can form through purely nonbiological chemistry.
Laboratory experiments simulating hydrogen-rich atmospheres containing hydrogen sulfide, methane, and nitrogen, conditions plausible for K2-18b, produced DMS and several other organosulfur gases through photochemistry alone. The mixing ratios generated in these experiments were substantial, above a tenth of a part per million, meaning that ultraviolet light hitting an atmosphere with the right starting ingredients can produce DMS in detectable quantities without any living organisms.10The Astrophysical Journal Letters. Abiotic Production of Dimethyl Sulfide, Carbonyl Sulfide, and Other Organosulfur Gases via Photochemistry: Implications for Biosignatures and Metabolic Potential
Even more striking, DMS has now been detected in the interstellar medium itself, far from any planet, in a molecular cloud near the center of the Milky Way. Radio telescope surveys found its spectral fingerprint in a region where star formation, not biology, drives the chemistry. The discovery provides direct evidence that DMS can be produced efficiently through abiotic processes involving grain-surface and gas-phase reactions in cold, dense environments.11The Astrophysical Journal Letters. On the Abiotic Origin of Dimethyl Sulfide: Discovery of Dimethyl Sulfide in the Interstellar Medium The researchers themselves noted that this finding casts doubt on using DMS as a reliable biomarker for exoplanets.12arXiv. On the abiotic origin of dimethyl sulfide: discovery of DMS in the Interstellar Medium Finding the molecule in a gas cloud between the stars is about as far from biology as you can get.
The Mini-Neptune Alternative
There is a more fundamental question lurking behind the biosignature debate: does K2-18b even have a surface? The Hycean world interpretation assumes a relatively thin hydrogen atmosphere above a liquid water ocean, but K2-18b could instead be a gas-rich mini-Neptune with no defined solid or liquid surface at all, just a thick atmosphere that gets progressively denser and hotter with depth. Photochemical and climate modeling has shown that the JWST observations, including the methane and carbon dioxide detections, can be explained by either scenario.13The Astrophysical Journal Letters. JWST Observations of K2-18b Can Be Explained by a Gas-rich Mini-Neptune with No Habitable Surface
In the mini-Neptune picture, methane and carbon dioxide are produced by the usual thermochemistry and photochemistry expected in a deep, warm hydrogen envelope. No ocean is needed. No surface biology is implied. The atmosphere’s composition would look similar from the outside because a transmission spectrum only probes the uppermost layers, the sliver of atmosphere that starlight passes through during a transit. What lies deeper is essentially invisible to this technique. Telling the two models apart requires either detecting a molecule that could only exist with an ocean underneath (which has not happened) or constraining the atmosphere’s total depth and temperature profile well enough to rule one scenario out (which current data cannot do).
This ambiguity is not a failure of JWST; it is an inherent limitation of the transit method. The telescope sees the planet’s atmosphere backlit by its star, but only the thin outermost shell is transparent enough to leave spectral fingerprints. Everything below that is opaque, hidden behind a wall of hydrogen.
Nitrogen Chemistry as a Possible Tiebreaker
If we cannot tell a Hycean ocean world from a mini-Neptune by looking at methane and carbon dioxide alone, what else might help? Nitrogen-bearing molecules are one possibility. Recent modeling of K2-18b’s atmosphere explored what would happen to nitrogen under different assumptions about whether the planet’s nitrogen arrives mostly as molecular nitrogen (Nâ‚‚, as in Earth’s atmosphere) or as ammonia (NH₃).
The results suggest that if ammonia is the dominant nitrogen source, photochemistry can generate a wider variety of detectable nitrogen species, including hydrogen cyanide and methylamine. If nitrogen arrives as Nâ‚‚ instead, photochemistry alone cannot produce enough of those compounds to be measurable. A future detection of ammonia, hydrogen cyanide, or methylamine would therefore point toward ammonia as the nitrogen source, while certain other species like cyanoacetylene and nitric oxide would be produced similarly in both scenarios and therefore cannot distinguish between them.14Monthly Notices of the Royal Astronomical Society. Nitrogen chemistry of hycean worlds on the example of K2-18b This kind of chemical mapping gives future observations a specific set of molecules to hunt for that would actually narrow the interpretation.
Climate on a Tidally Locked Ocean World
K2-18b almost certainly has one face permanently aimed at its star, a consequence of orbiting a small, cool red dwarf at close range. If the planet does have an ocean, the permanent day-night divide raises obvious questions about habitability. Would the dayside boil while the nightside freezes?
General circulation models simulating a Hycean version of K2-18b paint a more moderate picture. The thick hydrogen atmosphere is extremely effective at redistributing heat. In a reference simulation, the temperature difference between the hottest and coldest points on the surface was only about 30 degrees, far less dramatic than you might expect for a world that never rotates relative to its star. The atmosphere develops a broad equatorial jet and a strong pattern of dayside-to-nightside overturning circulation that carries warmth across the planet. Because K2-18b rotates slowly, the jet is not confined to the tropics but extends across much of the globe.15Monthly Notices of the Royal Astronomical Society. General circulation models of Hycean worlds The result is that even the permanent night side stays warm enough for liquid water, at least in these models.
Temperatures higher up in the atmosphere show even less variation from place to place, meaning that the transit spectrum, which samples the upper atmosphere, may not tell us much about surface temperature contrasts. That is another frustrating limitation for observers: the weather at the altitude we can actually see is bland and well-mixed, while the interesting action may be happening far below.
How K2-18b Compares to Other Temperate Sub-Neptunes
K2-18b is not the only temperate sub-Neptune that JWST is studying. TOI-270 d, a somewhat warmer planet orbiting a different red dwarf, has also had its atmosphere probed by the telescope. The two planets share some similarities in their near-infrared spectra, and a comparison of their mid-infrared MIRI data found that the two spectra are statistically consistent with each other within current measurement precision, with a small offset applied.16The Astrophysical Journal Letters. The Mid-infrared Transmission Spectrum of the Temperate Sub-Neptune TOI-270 d Whether that similarity means both planets have comparable atmospheric compositions or simply that current instruments cannot tell them apart is an open question.
Modeling work using the ratio of carbon dioxide to methane as a proxy for interior water content suggests that K2-18b’s interior is heavily enriched in water, with water making up more than half its composition, far exceeding what you would expect from standard assumptions about planetary metallicity. TOI-270 d, by contrast, appears to have roughly a quarter of its interior as water, closer to what conventional models predict.17The Astrophysical Journal Letters. Chemical Mapping of Temperate Sub-Neptune Atmospheres: Constraining the Deep Interior H2O/H2 Ratio from the Atmospheric CO2/CH4 Ratio If those estimates hold up, K2-18b formed in a way that gathered an unusually large amount of ice, which would be consistent with the Hycean scenario, though it does not prove a surface ocean exists today.
Stellar Contamination and the Limits of Transit Spectroscopy
One issue that gets less public attention but looms large in the technical discussion is stellar contamination. K2-18b transits a red dwarf star, and red dwarfs tend to have uneven surfaces: starspots and bright regions that differ in temperature and spectrum from the average photosphere. When the planet passes in front of one of those features, the stellar light that filters through the planet’s atmosphere carries imprints of the star’s own surface inhomogeneity, and those imprints can mimic or distort the molecular signatures astronomers are trying to measure.
Studies of this effect have found that discrepancies between different stellar models can add meaningful noise to JWST transmission spectra of planets around such stars. Under favorable assumptions, where the stellar models accurately capture the star’s true surface properties, stellar contamination falls below the photon noise level for typical transit observations. But when the models are inadequate, the additional uncertainty becomes significant and can affect which molecules appear to be detected and at what confidence.18The Astronomical Journal. Toward Robust Corrections for Stellar Contamination in JWST Exoplanet Transmission Spectra For a marginal detection like DMS, where the signal is already at or below two sigma, even a small amount of uncorrected stellar contamination could be the difference between a tentative detection and noise.
How the K2-18b Story Was Communicated to the Public
The gap between what the K2-18b data actually show and what many people came away believing is a case study in science communication. An analysis of how the 2023 findings were covered found something unusual: the original scientific paper contained more speculative language about life than the press releases written to promote it. The paper described the low-confidence DMS signal as “possible evidence of life” and framed the result as providing “a pathway toward the possible detection of life on an exoplanet with JWST.” The accompanying press releases were actually more cautious, burying the biosignature angle below the headline and attributing speculative claims to the authors rather than presenting them as conclusions. The news coverage, in at least one analyzed case, was more restrained still, reproducing only a single speculative claim.19PubMed Central. Communicating astrobiology and the search for life elsewhere: Speculations and promises of a developing scientific field in newspapers, press releases and papers
That pattern runs against the usual narrative that journalists hype results beyond what scientists claim. In this case, the scientists’ own language was the most enthusiastic in the chain. Part of the reason is that astrobiology is a field where speculative framing is built into the scientific culture: researchers routinely discuss what their findings “could mean” for the search for life, because the search for life is the motivating question behind the entire research program. There is nothing dishonest about it, but it does mean that a reader encountering phrases like “possible evidence of life” needs to understand that “possible” is carrying the entire sentence. The evidence is equally possible to explain without any biology at all, through photochemistry, through an atmosphere with no ocean beneath it, or through the signal being noise in the first place.
What Would Actually Constitute Evidence of Life
The K2-18b story highlights a broader challenge in exoplanet science: no single molecule in a transmission spectrum can prove life exists. DMS was proposed as a biosignature because on Earth it comes predominantly from biology, but the lab experiments and interstellar detections described above have weakened that link. Even a gas that is exclusively biological on Earth might be produced by unfamiliar chemistry on a planet with a completely different atmospheric composition and energy environment.
The more robust approach, and the one most astrobiologists favor, is to look for combinations of gases that would be chemically implausible without a sustained biological source. On Earth, the coexistence of oxygen and methane in the atmosphere is such a combination: the two react with each other and would disappear quickly unless something keeps replenishing both. For a hydrogen-rich atmosphere like K2-18b’s, the equivalent “disequilibrium” signatures are less well understood, because we do not have a catalog of hydrogen-world biospheres to study. Nitrogen chemistry modeling, as mentioned earlier, may eventually help by identifying molecular combinations that cannot be sustained by photochemistry alone. But that work is still in its early stages, and for now, K2-18b’s atmosphere is telling us more about the planet’s chemistry than about whether anything is alive in it.
Future JWST observations will likely accumulate more transits and push the signal-to-noise ratio higher, and next-generation telescopes on the ground and in space may eventually be able to detect reflected light from the planet’s atmosphere, a technique that probes different altitudes and different molecular features. The honest state of play in mid-2025 is that K2-18b is the most interesting candidate Hycean world we have, but the data do not yet distinguish between an ocean world with exotic chemistry and a mini-Neptune with no surface at all, let alone confirm biology.