Astronomers who once celebrated a possible sign of life on the exoplanet K2-18 b are now walking that claim back. Multiple independent reanalyses of James Webb Space Telescope data, using different reduction pipelines and retrieval codes, have found insufficient evidence for dimethyl sulfide, the molecule that briefly made K2-18 b the most talked-about world in exoplanet science. The signal that sparked global headlines about a distant ocean planet appears to be an artifact of how the data were processed, not a fingerprint of biology.
Why the dimethyl sulfide claim on K2-18 b matters now
The original excitement traces to JWST observations covering wavelengths from 0.9 to 5.2 micrometers. That study reported methane at roughly 5 sigma significance with about a 1 percent mixing ratio in a hydrogen-rich atmosphere, along with carbon dioxide and a tentative hint of DMS. NASA described the DMS finding as a possible detection, and the combination of gases was presented as consistent with a “Hycean” world, a planet with a liquid-water ocean beneath a thick hydrogen envelope. That framing turned K2-18 b into a test case for whether JWST could identify biosignatures on sub-Neptune exoplanets.
The stakes are straightforward. If DMS were real, it would represent the first credible detection of a molecule produced almost exclusively by living organisms on a world outside our solar system. If it is not real, the episode becomes a cautionary lesson about how statistical choices during data reduction can manufacture apparent signals. The answer matters for every future biosignature claim JWST will produce, because it shapes how confidently astronomers can interpret faint spectral wiggles as evidence of life rather than quirks of instrumentation or analysis.
Three independent teams find no reliable DMS detection
The strongest challenge came from a joint analysis covering JWST NIRISS, NIRSpec, and MIRI observations spanning 0.6 to 12 micrometers. That study used three independent data-reduction pipelines and two retrieval codes and found insufficient evidence for dimethyl sulfide or related sulfur-bearing molecules in K2-18 b’s atmosphere. The team also showed that other methyl-bearing compounds, such as ethane, fit the same spectral features equally well. In other words, even if a faint absorption feature exists at the relevant wavelengths, it does not uniquely point to a biologically produced gas. The researchers estimated that roughly 25 additional MIRI transits would be needed to reach a 5 sigma detection of DMS, a threshold that would require years of additional telescope time and a major reallocation of JWST’s limited observing schedule.
A separate reanalysis of the near-infrared spectra reinforced those findings. That effort applied 60 different data treatments and ran more than 250 retrievals to probe how robust the original claims were to methodological choices. It confirmed methane at about 4 sigma, lending support to the idea that K2-18 b hosts a hydrogen-rich atmosphere with significant CH4. But it reported no statistically significant evidence for carbon dioxide or DMS. The gap between the original study’s CO2 claim and this reanalysis is notable because CO2 was part of the chemical package used to argue for an ocean world scenario. Without a confident CO2 detection, the case for a temperate, water-rich environment becomes more speculative.
A third line of evidence comes from an independent assessment of the MIRI mid-infrared spectrum itself. That technical analysis, conducted at Oxford, found only about a 2 sigma preference for Gaussian features in the data, well below the bar for a credible detection. A separate MIRI-focused study had initially reported the spectrum was inconsistent with a featureless model at 3.4 sigma and argued that DMS and DMDS were among the few molecules that could explain the shape. But the Oxford analysis suggests even that 3.4 sigma figure may overstate the true significance once different statistical tests and noise treatments are applied, undercutting the idea that any specific molecule is demanded by the data.
A peer-reviewed methodology paper published in Nature Astronomy adds a structural critique. It demonstrates that model comparison choices during atmospheric retrieval can create apparent detections without uniquely identifying a specific gas. The paper uses K2-18 b’s DMS and DMDS comparisons as a case study, showing how Bayesian evidence values shift depending on which molecules are included in the candidate set, how priors are defined, and how the spectra are reduced. In scenarios where many plausible absorbers overlap in wavelength, small changes in assumptions can make one molecule look favored over another even when the data cannot truly distinguish between them.
K2-18 b’s identity as an ocean world is also in question
The DMS debate sits inside a larger argument about what kind of planet K2-18 b actually is. The original interpretation framed it as a Hycean world with a thin hydrogen atmosphere over a global ocean, sitting in the habitable zone of a cool star. That picture relied on combining the tentative DMS signal, methane and carbon dioxide detections, and interior models that allowed for a substantial water layer. It was an appealing narrative: a relatively nearby sub-Neptune that might host a warm ocean under starlight, with a biosignature gas wafting into space.
Subsequent work has pushed back on that storyline. A peer-reviewed study in Nature Astronomy argues that K2-18 b is more likely gas-rich rather than ocean-dominated, with a deep, dense envelope more akin to a mini-Neptune than a super-sized Earth. In that scenario, pressures and temperatures at depth could be extreme, potentially forming high-pressure ices or supercritical fluids instead of a clement liquid-water ocean. If K2-18 b is wrapped in such a massive atmosphere, the chemical conditions that would plausibly produce biogenic DMS become far less likely, and the entire Hycean framework loses its anchor case.
The emerging consensus is that K2-18 b remains an intriguing, methane-bearing sub-Neptune, but not yet a compelling candidate for hosting life as we know it. Methane alone is not a reliable biosignature in a hydrogen-rich atmosphere, because it can be produced efficiently through abiotic processes such as photochemistry and interior outgassing. Without corroborating gases like CO2 at well-measured levels, and without a physically grounded case for a stable ocean, the biological interpretation has little to stand on.
A cautionary template for future biosignature claims
The K2-18 b saga is already reshaping how exoplanet scientists talk about potential biosignatures. One clear lesson is that marginal features near the threshold of detectability are highly vulnerable to analysis choices. Different pipelines, systematics corrections, and retrieval frameworks can shift a signal from “possible” to “not significant” without any new photons being collected. That reality argues for a culture of preemptive skepticism, in which independent teams attempt to reproduce headline-grabbing results before they are widely promoted as signs of life.
Another lesson is the importance of full-spectrum context. The most recent joint analysis that finds no convincing DMS signal spans a broad wavelength range and leverages multiple JWST instruments, enabling cross-checks that were not available to the first, narrower studies. As more comprehensive datasets arrive for other exoplanets, similar multi-instrument approaches will be crucial for separating real atmospheric structure from noise and instrumental quirks.
Finally, the debate underscores the need to tie atmospheric retrievals closely to physically plausible planet interiors. Claims about habitable oceans or biosignature gases should be consistent not only with the spectra but also with what is known about the planet’s mass, radius, and likely composition. For K2-18 b, emerging interior models favoring a gas-rich mini-Neptune make the original Hycean and DMS interpretations harder to sustain. Future targets may fare better, but they will face the same demand: robust, reproducible evidence that survives independent scrutiny and fits within a coherent planetary picture.
For now, K2-18 b stands as a landmark not because it has yielded proof of alien life, but because it has exposed how easily such proof can be claimed and how carefully it must be tested. The retreat from DMS does not diminish the power of JWST; instead, it highlights that the telescope’s exquisite sensitivity must be matched by equally rigorous analysis. As the search for life continues, K2-18 b will be remembered as an early stress test of those standards-and a reminder that in exoplanet science, extraordinary claims demand not just extraordinary data, but extraordinary care.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
