New Constraints on DMS and DMDS in the Atmosphere of K2-18 b from JWST MIRI

Brief

Madhusudhan et al. (2025) report the first mid-infrared transmission spectrum of a habitable-zone exoplanet, K2-18 b, obtained with JWST MIRI LRS (6–12 μm) during a single 5.85-hour transit on April 25–26, 2024. The observed spectrum is inconsistent with a featureless baseline at 3.4-σ significance, and the spectral features are not reproduced by most molecules expected in K2-18 b's H₂-rich atmosphere except DMS and DMDS, yielding a 3-σ independent detection of at least one of the two at abundance ≳ 10 ppmv. Two independent reduction pipelines (JExoRES and JexoPipe) agree within 1-σ across all data points. The degeneracy between DMS and DMDS remains unresolved and additional observations are required, as are improved experimental cross-sections for both molecules.

Metadata

Category

Search

Venue

The Astrophysical Journal Letters

Type

Peer-reviewed

Year

2025

Authors

Nikku Madhusudhan, Savvas Constantinou, Mans Holmberg, Subhajit Sarkar, Anjali A. A. Piette, Julianne I. Moses

DOI

10.3847/2041-8213/adc1c8

arXiv

2504.12267

Access

Open access

Length

14.7 M

Programs

JWST GO Program 2722

Instruments

JWST MIRI LRS

Tags

biosignature, exoplanet-atmosphere, astrobiology, hycean-world, SETI, transmission-spectroscopy

Key points

• The MIRI LRS spectrum (6–12 μm) is inconsistent with a featureless spectrum at 3.4-σ significance, ruling out a flat transmission baseline.p.1
• DMS and/or DMDS detected at 3-σ significance with abundance ≳ 10 ppmv for at least one of the two molecules, an independent line of evidence from a different instrument and wavelength range.p.1
• The prior JWST NIRISS/NIRSpec DMS inference was ≲ 2-σ and offset-dependent, ranging from 2.4-σ with no detector offsets to below 1-σ with two offsets applied.p.3
• K2-18 b has mass 8.63 ± 1.35 M⊕ and radius 2.61 ± 0.09 R⊕, orbiting an M2.5V star in its habitable zone; prior JWST near-IR spectra detected CH₄ at 5-σ and CO₂ at 3-σ.p.2
• A habitable liquid water ocean on K2-18 b requires Bond albedo A_B > 0.5–0.6; the planet's dayside albedo has not been directly measured.p.2
• The observation used 5,095 integrations at 25 groups per integration over 5.85 hours (2.68 hours in-transit); white light curve residuals were 471 ppm, 1.25× the photon+read noise floor.p.3
• DMS mixing ratios as high as 10⁻² are physically plausible for biogenic fluxes exceeding ~20× Earth levels of sulfur-based biosignature gases.p.3
• Both DMS and DMDS have overlapping spectral features between 6.8–8 μm, with broader features near 9–10 μm (DMS) and 10–11 μm (DMDS), enabling partial discrimination in the MIRI band.p.6

Verbatim

• “The spectrum shows distinct features and is inconsistent with a featureless spectrum at 3.4- σ significance compared to our canonical model.”

  p.1

• “We find that the spectrum cannot be explained by most molecules predicted for K2-18 b with the exception of DMS and dimethyl disulfide (DMDS), also a potential biosignature gas.”

  p.1

• “We report new independent evidence for DMS and/or DMDS in the atmosphere at 3- σ significance, with high abundance ( ≳ 10 ppmv) of at least one of the two molecules.”

  p.1

• “DMS mixing ratios as high as 10 − 2 are possible in K2- 18 b for high biogenic fluxes of sulfur-based biosignature gases above ∼ 20 × Earth levels (Tsai et al. 2024).”

  p.3

Most interesting

• This is the first mid-infrared transmission spectrum ever obtained for any habitable-zone exoplanet, opening a complementary spectral window to the near-IR data that produced the original DMS hint.
• The MIRI LRS band (6–12 μm) was chosen precisely because it avoids the detector-offset problem that made the prior DMS inference unstable; it also captures stronger, less-overlapping DMS/DMDS features than the 3.3–4.3 μm region.
• Mini-Neptune and gas-dwarf interior models are ruled out by the retrieved atmospheric abundances, particularly the low NH₃ and high CO₂/CO ratio, leaving the hycean world as the best-fit scenario.
• The paper explicitly flags the lack of accurate laboratory cross-sections for DMS and DMDS as a blocking uncertainty, meaning the detection significance itself depends on cross-section quality that has not yet been experimentally validated at relevant pressures and temperatures.
• Biogenic CH₄ is also on the table: the CH₄–CO₂ pair in a shallow H₂-rich atmosphere is described as more easily explained by an inhabited hycean scenario than an uninhabited one, per independent photochemical modeling (Wogan et al. 2024; Cooke & Madhusudhan 2024).
• Two fully independent reduction pipelines (JExoRES, JexoPipe) were run in parallel with different background subtraction, outlier rejection, and limb-darkening strategies; all resulting spectra agree within 1-σ, substantially hardening the robustness claim.