No Helium Detected in LHS 1140 b from Four JWST NIRISS/SOSS Transits

Abstract: In the effort to determine which low-mass exoplanets have atmospheres, LHS 1140 b remains one of the most favorable targets. Its large size (5.6 $\rm M_{\oplus}$ and 1.7 $\rm R_{\oplus}$) and relatively long orbital period (24.7 days) imply an atmosphere may be likely, and notably, recent interior models favor either a hydrogen-dominated "mini-Neptune" or a "water world" over a true terrestrial planet. Another possibility is that it has a helium-rich atmosphere. This hypothesis is supported by recent ground-based observations that detected the metastable helium triplet during transit. These observations indicated there may be current helium escape from the planet's upper atmosphere, yet the signal was not detected during a subsequent observation, suggesting time-variable escape. Here we present four observations of LHS 1140 b with JWST NIRISS/SOSS, which covers the metastable helium triplet, obtained between 2023 and 2026. These observations span the epoch of the ground-based measurements, and although none were contemporaneous with the ground-based transits, all four are sensitive to helium absorption at the previously reported level. However, we detect no helium absorption in any visit. We reject the best-fit ground-based model at $>3σ$ in each visit, and find no clear trend in mass-loss with time. Our results suggest the reported ground-based detection may be spurious, although variability cannot be excluded if detectable helium absorption occurs in $\lesssim50\%$ of transits. The nature of LHS 1140 b thus remains a mystery until future transmission and emission analyses are complete.

• The paper finds no metastable helium absorption in four JWST NIRISS/SOSS transits, jointly rejecting the reported ground-based signal at 9.9σ and each individual model at more than 3σ.
• The analysis combines column-level light curves from two reduction pipelines and places 2σ helium-depth limits of 0.02–0.04% at SOSS resolution, equivalent to roughly 0.41–0.61% at WINERED resolution.
• The results indicate that the 2024 helium detection was non-persistent and possibly spurious, while leaving time-variable escape, a hydrogen-rich mini-Neptune, and a water-world interpretation unresolved.

Context and motivation

LHS 1140 b is among the most favorable low-mass exoplanets for atmospheric characterization: with a mass of $5.60\pm0.19\,M_{\oplus}$ and radius of $1.730\pm0.025\,R_{\oplus}$ orbiting a 3096 K M dwarf every 24.7 days at an equilibrium temperature of only 226 K, it is the largest and least XUV-irradiated temperate small planet known. Interior modeling favors either a hydrogen-dominated mini-Neptune or a water world (9–19% water by mass) over a terrestrial composition. A third hypothesis—a helium-dominated atmosphere produced by diffusive fractionation during hydrodynamic escape—gained support from a ground-based detection of the metastable helium triplet at 1.0833 $\mu$m using WINERED on Clay/Magellan II in September 2024, which reported a 1.24% absorption depth and implied ongoing escape from a helium-rich upper atmosphere. A subsequent ground-based nondetection in 2025 was attributed to time-variable escape.

This paper tests that interpretation using four transits observed with JWST NIRISS/SOSS between December 2023 and July 2026: two archival visits from DD 6543 and two new visits from GO 7073. Although none of the SOSS visits are contemporaneous with the ground-based observations, all four are sensitive to helium absorption at the level reported by the ground-based detection, which should translate to roughly 600 ppm at SOSS pixel resolution after convolution.

Observations and data reduction

All four visits used the SUBSTRIP256 subarray with NISRAPID readout; Visits 1–2 used three groups per integration (949 integrations) while Visits 3–4 used five groups (547 integrations). The analysis focuses on the narrow 1.07–1.095 $\mu$m range containing the helium triplet, deferring panchromatic analysis to future work. Because the triplet is narrower than a single SOSS pixel, light curves were extracted at the column level to maximize signal-to-noise.

The primary reduction uses FIREFLy, including group-level $1/f$ correction, scaled STScI background subtraction (with separate scalings on either side of the background step near column 750), conservative temporal cleaning ($\sigma>10$), and box extraction with a 36-pixel aperture. For Visit 1, which suffered blind pointing, the trace offset was re-derived empirically as $\Delta x=-167$, $\Delta y=-14$ pixels using the three well-pointed visits, improving on the earlier cross-correlation-based solution. The two new visits were independently reduced with exoTEDRF; both pipelines agree within $1\sigma$ for nearly all columns, and neither shows any hint of helium absorption, confirming the result is not reduction-dependent. Residual wavelength-solution discrepancies between pipelines are consistent with known PWCPOS-dependent scatter of up to ~1/3 pixel in the pastasoss calibration.

Nondetection and upper limits

Column-level helium light curves show no excess transit absorption relative to the white light curves in any visit, no pre- or post-transit tails, and structureless residuals in adjacent columns. The resulting $2\sigma$ upper limits on absorption depth, converted to the WINERED resolution for direct comparison, are:

Visit	Depth (SOSS res.)	Depth (WINERED res.)	$1.730\pm0.025\,R_{\oplus}$0 limit (g s$1.730\pm0.025\,R_{\oplus}$1)
1	0.03%	0.45%	$1.730\pm0.025\,R_{\oplus}$2
2	0.02%	0.41%	$1.730\pm0.025\,R_{\oplus}$3
3	0.04%	0.61%	$1.730\pm0.025\,R_{\oplus}$4
4	0.03%	0.44%	$1.730\pm0.025\,R_{\oplus}$5

These limits are comparable to the 0.6% upper limit from the 2025 ground-based nondetection. Each individual SOSS spectrum rejects the convolved best-fit ground-based model at $1.730\pm0.025\,R_{\oplus}$6; combined into a weighted mean transmission spectrum, the rejection reaches 9.9$1.730\pm0.025\,R_{\oplus}$7. This is the central quantitative claim of the paper: if the 2024 signal had persisted, it would have been confidently detected in every SOSS visit.

Mass-loss implications

Using pwinds with all parameters fixed to those inferred by Cherubim et al. (mass-loss rate varied alone, GJ 1132's XUV flux as a proxy for LHS 1140's), the six available observations—two WINERED and four SOSS spanning nearly three years—show no trend in mass loss with time. The single 2024 detection implies $1.730\pm0.025\,R_{\oplus}$8 g s$1.730\pm0.025\,R_{\oplus}$9, whereas all five nondetections constrain $\mu$0–$\mu$1 g s$\mu$2. The 2024 value stands out as an outlier by more than an order of magnitude.

The authors note several caveats that qualify this comparison. Above $\mu$3 g s$\mu$4, hydrogen self-shielding suppresses the metastable population, so the nondetections technically also permit very high mass-loss rates—though maintaining such rates over the system's >5 Gyr age is implausible for so low-mass a planet. More fundamentally, because SOSS does not resolve the line profile, mass-loss rate, wind temperature, H:He ratio, and stellar XUV flux are degenerate, and the EUV flux and upper-atmosphere temperature are themselves poorly characterized. The mass-loss limits should therefore be read as relative comparisons under uniform modeling assumptions rather than strict physical boundaries.

Limitations and open questions

The paper is explicit about what its results do and do not establish. Time variability cannot be fully excluded: if the 2024-level signal occurred during a random fraction $\mu$5 of transits, four SOSS nondetections would occur by chance ~6.3% of the time for $\mu$6, yielding $\mu$7 at $\mu$8. Thus detectable escape could recur in up to roughly half of transits without having been seen. A fifth SOSS transit scheduled for November 2026 will add statistical leverage, but quantifying any variability timescale from one detection and five nondetections remains difficult. The nondetections also do not discriminate between the mini-Neptune and water-world interpretations; resolving LHS 1140 b's nature requires panchromatic transmission analysis and emission photometry from the Rocky Worlds DDT program, neither of which is presented here.

Conclusion

Four JWST NIRISS/SOSS transits of LHS 1140 b show no metastable helium absorption, individually rejecting the reported ground-based detection at $\mu$9 and jointly at $\mu$0. Combined with the 2025 ground-based nondetection, the 2024 signal is now incontrovertibly non-persistent, and if genuinely time-variable it recurs less than half the time—making a spurious origin for the ground-based detection the more likely explanation. The helium-dominated atmosphere hypothesis is thereby weakened, though not eliminated, and the fundamental question of whether LHS 1140 b is a mini-Neptune or a water world remains open pending complete transmission and emission analyses.