A joint analysis of James Webb Space Telescope observations spanning three instruments has found insufficient evidence for dimethyl sulfide or dimethyl disulfide in the atmosphere of exoplanet K2-18 b, according to a new paper. The result directly challenges a separate MIRI-only study that reported roughly 3-sigma evidence for those sulfur compounds, molecules that on Earth are produced almost exclusively by living organisms. The disagreement between these two studies exposes how sensitive trace-gas searches in alien atmospheres remain to the choice of data reduction pipeline, retrieval model, and wavelength coverage.
Why the DMS debate around K2-18 b matters right now
K2-18 b sits in its star’s habitable zone and has drawn intense scientific attention since Hubble WFC3 data revealed a water-vapor signature in the planet’s atmosphere. That early finding turned the sub-Neptune into a prime target for JWST follow-up. In 2023, JWST NIRISS and NIRSpec observations produced strong detections of methane and carbon dioxide in an H2-rich atmospheric envelope, along with a tentative signal that researchers identified as possibly belonging to DMS. NASA’s own mission summary described that DMS feature as a “possible detection” that would require further validation.
The stakes are straightforward: DMS on Earth comes from marine phytoplankton and other biological sources. Confirming it in K2-18 b’s atmosphere would have been the strongest indirect sign of extraterrestrial biology ever recorded by a telescope. Two new papers now pull the evidence in opposite directions, and the tension between them turns on technical choices that most readers never see.
Competing JWST analyses reach opposite conclusions on sulfur compounds
A MIRI-focused study used mid-infrared transmission spectroscopy as an independent line of evidence, separate from the near-infrared results that first flagged DMS. That paper reported roughly 3-sigma evidence for DMS and/or DMDS in K2-18 b’s atmosphere, a statistical threshold that falls short of the conventional 5-sigma standard for a confirmed detection but is strong enough to warrant serious discussion. The authors argued that specific absorption features in the MIRI wavelength range align more closely with sulfur-bearing molecules than with alternative explanations such as additional methane or carbon monoxide.
A broader joint analysis, however, combined NIRISS, NIRSpec, and MIRI data covering 0.6 to 12 microns and tested multiple data reductions and retrieval approaches. That study concluded there is insufficient evidence for DMS and DMDS once the full wavelength range is modeled together. When the near-infrared and mid-infrared spectra are fit simultaneously, the apparent sulfur features can be mimicked by different combinations of bulk gases, temperature profiles, and cloud properties without invoking trace sulfur compounds.
A third paper took a different angle, conducting a systematic search for trace molecules in K2-18 b’s atmosphere using model-comparison techniques. Instead of asking whether a particular spectrum could be matched by a DMS-bearing atmosphere, it assessed which molecules are statistically required by the data. In that framework, adding DMS or DMDS did not yield a decisive improvement in fit quality relative to simpler models, reinforcing the conclusion that the current spectra do not demand sulfur compounds.
Taken together, these studies form a pattern: the more data and the more retrieval methods researchers fold in, the weaker the sulfur-compound signal becomes. The MIRI-only analysis sees a suggestive bump that could be DMS or DMDS. The joint three-instrument analysis finds that bump becomes ambiguous once the rest of the spectrum is accounted for. The model-comparison survey then shows that the data can be explained without invoking those molecules at all.
One plausible explanation for the instrument-to-instrument disagreement involves atmospheric hazes. If hazes in K2-18 b’s atmosphere grow more opaque at longer wavelengths, they could suppress the apparent strength of DMS absorption features in NIRSpec data while leaving marginal MIRI signals partially intact. That kind of wavelength-dependent opacity would produce exactly the tension researchers are seeing: a hint in mid-infrared data that vanishes when shorter wavelengths are jointly modeled. No study has yet confirmed this mechanism for K2-18 b, but it illustrates how atmospheric modeling assumptions can drive conflicting results from the same telescope.
Another contributor is the treatment of instrumental systematics. Each JWST mode has its own quirks-detector noise behavior, pointing jitter, and subtle calibration offsets. The MIRI-only work and the joint analysis used different reduction pipelines and different strategies for correcting those effects. Small choices about which pixels to flag, which time segments to discard, or how to model stellar variability can shift the inferred depth of spectral features by amounts that matter when chasing parts-per-million signals.
Unresolved questions and what comes next for K2-18 b
Several gaps in the published record keep this debate open. None of the studies have released the full set of joint retrieval posterior distributions or covariance matrices from their combined fits, which would let independent teams verify how much the DMS signal depends on specific model assumptions. Without those details, it is difficult to quantify exactly how strongly the data rule out particular DMS abundances, or how sensitive the conclusions are to choices about clouds, temperature, and chemistry.
The lead authors of the original 2023 carbon-molecule paper have not publicly responded to the new non-detection claims, at least in the formal literature. That silence leaves open questions about how they interpret the latest results and whether they view the original DMS hint as fully withdrawn or simply unconfirmed. In parallel, the different spectral reductions used across the three primary JWST studies have not been compared in a single tabulated format, making it difficult for outside researchers to pinpoint where the analyses diverge at the level of raw light curves and calibrated spectra.
Despite the sulfur controversy, the confirmed atmospheric detections remain solid. Methane and carbon dioxide in an H2-rich envelope are seen consistently across instruments and retrieval frameworks. Those findings are not in dispute and continue to make K2-18 b one of the most chemically characterized exoplanets accessible to JWST. The planet still stands out as a temperate sub-Neptune with a volatile-rich atmosphere, potentially overlying a water-bearing interior, even if no biosignature has yet emerged.
What has changed is the confidence that biology might explain part of that chemistry. Where some early commentators speculated about ocean worlds and plankton-like ecosystems, the newer analyses push the interpretation firmly back into the realm of abiotic atmospheric physics and chemistry. For now, K2-18 b looks more like a testbed for understanding sub-Neptune atmospheres than a leading candidate for inhabited worlds.
For anyone following the search for life beyond Earth, the practical takeaway is that a single JWST instrument operating in a narrow wavelength band cannot yet deliver a reliable biosignature detection for a planet 120 light-years away. Future observations will likely need to combine multiple transit epochs, additional MIRI wavelength channels, and standardized reduction pipelines before any sulfur-compound claim can clear the bar. Coordinated community efforts to reprocess the same raw data through multiple pipelines, and to share retrieval setups in detail, will be essential to avoid over-interpreting marginal features.
The next round of scheduled JWST time on K2-18 b, whenever it is executed, will be watched closely for signs that the current picture can be sharpened. Longer time-series observations could improve signal-to-noise, while expanded wavelength coverage may help disentangle hazes from molecular absorption. In parallel, laboratory spectroscopy of DMS and related molecules at the relevant temperatures and pressures can refine the theoretical templates used in retrievals. Only by combining better data, better models, and greater transparency in analysis methods will astronomers be able to say with confidence whether K2-18 b’s atmosphere harbors exotic sulfur chemistry-or simply reflects the complex, but non-biological, physics of a sub-Neptune world.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
