{"id":"01e368c3-84cb-42c9-bab9-de3323a7e20a","arxiv_id":"2608.13470","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":6,"one_line_summary":"No metastable helium absorption is detected in four JWST NIRISS transits of LHS 1140 b, ruling out the 2024 WINERED-detection amplitude with odds ratios of 300:1 or more.","lead":"Four JWST transits of the temperate planet LHS 1140 b show no helium absorption, with the data preferring a flat spectrum in every visit. The results strongly rule out a helium signal as strong as a 2024 ground-based detection, but only at epochs at least nine months away from it.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The fixed 18-Å He line width in the model comparison and injection-recovery is the main unresolved assumption; broad He lines at the JWST epochs would not be constrained by the quoted odds and limits, though the ruled-out 2024-equivalent signal is robust.","rationale":"The reader's verdict of ACCEPT with high confidence is well supported by the paper's careful epoch-scoping and appropriate statistical methods. Our stress-test focused on the single assumption that could change the interpretation of the null result: the fixed 18 Å FWHM used in both the free-amplitude model and the injection-recovery tests. This is indeed the weakest assumption, exactly as the reader identified. However, it does not undermine the primary claim, because the paper explicitly defines the ruled-out signal as having the width observed by Cherubim et al. (2026) in 2024. The residual risk is confined to the secondary generalization \"disfavour the presence of He absorption in general\", which is already hedged. The paper's non-contemporaneity caveat is appropriately prominent, and the numerical inconsistencies (550 vs 500 ppm, etc.) are minor typographical issues. The test we propose would settle whether the width assumption is benign: if free-width fits still favour flat and broad-line injections remain constrained, the conclusion is robust; if not, the paper's language in the abstract and Section 3 should be narrowed. In either case, the central, carefully-worded claim about the 2024-equivalent signal is secure, so the ACCEPT verdict stands.","tokens_in":13773,"tokens_out":17870,"duration_ms":166998,"concrete_test":"Re-analyse each of the four transmission spectra in Section 3 with the Gaussian FWHM as a free parameter, e.g., a log-uniform prior over 5–100 Å, and compute the Bayes factor versus the flat continuum. Also run the injection-recovery sweep with injected FWHMs of 10, 30, and 50 Å. If the data still favour flat for all widths and the 100:1 upper limits on integrated equivalent width (or on amplitude at each width) remain within ~50% of the quoted values, the fixed-width assumption is benign; if broad lines are substantially unconstrained, the \"disfavour in general\" statement should be explicitly restricted to narrow lines.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim has two parts: (1) ruling out a He signal equivalent to the 2024 WINERED detection, and (2) disfavouring He absorption in general. The first part is self-consistent because the 18 Å FWHM is defined as the Cherubim line convolved to SOSS resolution (Section 3). The second part, however, rests on the free-amplitude Gaussian model that also fixes the FWHM to 18 Å, and the injection-recovery tests only inject and recover lines with that same width. If the actual He line at any JWST epoch were significantly broader (e.g., a 40 Å profile from a different outflow geometry), the peak amplitude per pixel would be lower, and the odds ratios and upper limits quoted in Section 3 and Table 2 would not directly constrain it. The paper itself acknowledges the width assumption in Section 5 (\"assuming its width is comparable to that reported by Cherubim et al. 2026\"), so this is a known limitation rather than an internal contradiction. The concern is load-bearing only for the generalization \"disfavour ... in general\"; the specifically-scoped \"equivalent to the 2024 detection\" claim is not threatened.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes four archival JWST NIRISS/SOSS transits of LHS 1140 b, spanning December 2023 to July 2026, to search for metastable helium (1.083 µm) absorption. The authors reduce the data with a custom pipeline, fit white-light and pixel-level light curves, and construct transmission spectra. They perform Bayesian model comparison between a flat continuum, a continuum plus a free-amplitude Gaussian He line (with FWHM fixed to 18 Å, corresponding to the 2024 WINERED detection convolved to SOSS resolution), and a continuum plus a Gaussian with both amplitude and width fixed to that detection. They find that all four visits disfavor the presence of He absorption over a flat continuum, with log-Bayes factors of 1.36–2.45 (odds 3.9:1–11.6:1), and strongly rule out the 2024-equivalent signal, with log-Bayes factors of 5.69–11.36 (odds ~300:1–8.6×10^4:1). Injection-recovery tests yield upper limits on the He line amplitude of 350–550 ppm at SOSS resolution for the four visits. In a separate analysis, the authors fit the out-of-transit stellar spectra from these visits plus one additional planet-c transit, using multi-component NewEra models, and find consistent photospheric and heterogeneity parameters across all five epochs. The paper concludes that no He escape is detected and places strict limits that complement the ground-based WINERED results.","tokens_in":14135,"tokens_out":8377,"duration_ms":75149,"significance":"If the results stand, this work provides valuable multi-epoch upper limits on He escape from LHS 1140 b, directly complementing the 2024 WINERED detection and 2025 non-detection by Cherubim et al. It demonstrates that JWST/NIRISS SOSS, despite its low spectral resolution, can constrain a narrow He line nearly as well as ground-based high-resolution spectroscopy, leveraging its photometric stability. The analysis is technically careful: it uses standard data-reduction practices, nested sampling for model comparison, injection-recovery sensitivity tests, and explicit checks of systematics (e.g., spot crossings, a potential flare, and a serendipitous planet-c transit). The stellar spectral fits, while secondary, provide a useful consistency check over a 2.5-year baseline. The main caveats—that the He line width is fixed to the 2024 WINERED value and that the four JWST transits are not contemporaneous with the 2024 detection—are acknowledged in the text. The central claim that the 2024-equivalent He signal is ruled out is robust.","major_comments":[{"comment":"The free-amplitude Gaussian He model fixes the line width to 18 Å but leaves the amplitude free; however, the prior on that amplitude is never specified in the text or in Table 2. The Bayesian evidence (and hence the reported odds ratios of 3.9–11.6:1 for the disfavoring claim) depends directly on this prior's width and shape. Please state the prior explicitly and, ideally, test the sensitivity of the Bayes factors to a broader or narrower prior range. Without this, the model comparison is not fully reproducible.","section":"§3, Table 2"},{"comment":"The quoted upper limits on the He amplitude are internally inconsistent between Sections 3 and 5. Section 3 reports limits of 350, 550, 420, and 380 ppm for Visits 1–4, with corresponding WINERED-resolution amplitudes of 0.72%, 1.15%, 0.88%, and 0.79%. Section 5 reports limits of 350, 500, 420, and 380 ppm, with corresponding amplitudes 0.68%, 1.07%, 0.81%, and 0.74%. At least one of these sets contains errors; the authors should correct the numbers and verify the ppm-to-percent conversion.","section":"§3 and §5"},{"comment":"The claim that the data 'disfavour the presence of He absorption in general' and the 'strict limits' quoted in the abstract and summary are conditioned on the assumed 18 Å line width. The text acknowledges this in Section 5, but the abstract and the final summary present the limits without that qualifier. Since a broader He line (e.g., from a different outflow geometry) would have a lower peak amplitude per resolution element and would be less constrained by the present analysis, I recommend stating explicitly in the abstract and conclusions that the limits apply to He lines with a width comparable to the 2024 WINERED detection.","section":"§3 and abstract"}],"minor_comments":[{"comment":"There is a typo: 'herterogeneity' should be 'heterogeneity' in the abstract and in the corresponding sentence in Section 4.","section":"Abstract"},{"comment":"The phrase 'the revisement of its mass' should read 'the revision of its mass'.","section":"§1"},{"comment":"The legend label 'He (WINERED Amp)' is not defined in the caption; please define it as the model with both amplitude and width fixed to the 2024 WINERED detection.","section":"Figure 2 caption"},{"comment":"The caption says 'Parameter Prior Range' but the prior distributions for some parameters (e.g., u1, u2) are not fully described; the table says 'Prior Range' but the entries are only ranges, not the exact prior families (e.g., uniform vs Gaussian). Please clarify.","section":"Table 1"},{"comment":"When describing the potential flare in Visit 4, the text says 'there is no clear structure in the Hα light curves' but does not state whether the Hα data were examined in the same way as the white light; a brief clarification of the diagnostic used would be helpful.","section":"§2.1"}],"recommendation":"minor_revision","confidential_remarks":"The paper is well within the scope of ApJL and the central result—that the 2024-equivalent He signal is strongly ruled out—is robust. The main issues are local: an unspecified prior for the free-amplitude He model, an internal inconsistency in the reported upper limits between Sections 3 and 5, and a need to qualify the 'strict limits' language with the line-width assumption. These are easily fixable. I recommend minor revision rather than acceptance as-is because of the numerical inconsistency, which must be corrected before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Clean null result that does what it says on the tin. Four archival JWST NIRISS transits of LHS 1140 b, no metastable He at 1.083 µm in any of them. The flat continuum is preferred over a free-amplitude He line by odds of 3.9–11.6:1, and the 2024 WINERED-equivalent signal (589 ppm at 18 Å) is ruled out with odds of 300–8.6e4:1. Injection-recovery places sensitivity limits of 350–550 ppm at SOSS resolution, which when scaled to WINERED resolution is comparable to the 0.6% upper limit from the 2025 WINERED non-detection. That's genuinely useful.\n\nThe paper is careful: pixel-level spectrophotometric fits, nested sampling, model comparison, injection-recovery. It also fits the out-of-transit stellar spectra across five visits and finds no strong variability, though this part is more exploratory. The author is open about the fact that none of the JWST transits are contemporaneous with the 2024 detection, so the result is a set of strict limits at four other epochs, not a refutation.\n\nMain soft spot is the fixed 18 Å line width, inherited from convolving the Cherubim detection to SOSS resolution. Both the free-amplitude model comparison and the injection-recovery tests assume that width, so 'disfavour the presence of He absorption in general' is a bit too broad. A broader line at a JWST epoch would be less constrained. The paper acknowledges the width assumption in Section 5, so it's a known limitation, but it would be better to soften the Section 3 wording and maybe add a sentence on how limits would change for broader lines. Minor stuff: the Visit 4 flare is cut post hoc; the justification is reasonable, but it would be nice to see a test with it included. And the Visit 2 upper limit is quoted as 550 ppm in Section 3 and 500 ppm in Section 5—one of those is a typo. The stellar spectrum analysis is the weakest section: the three-component model with a hot 3000 K component is hard to believe, but the author acknowledges the caveats and it's not load-bearing.\n\nOverall the central claim holds up. I'd send this to a serious referee and expect acceptance after minor revisions.","headline":"Clean four-epoch null result that robustly rules out the 2024 WINERED helium detection at those epochs; the fixed line-width assumption is the main caveat but the paper acknowledges it.","tokens_in":14610,"tokens_out":4895,"would_cite":true,"duration_ms":44607,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Four JWST transits of LHS 1140 b show no helium absorption","keywords":["Exoplanets","Exoplanet atmospheres","Helium absorption","Atmospheric escape","JWST NIRISS","Transmission spectroscopy","LHS 1140 b","Planetary atmospheres"],"falsifier":"Observe a high-resolution ground-based spectrum of LHS 1140 b during one of the same epochs as these JWST transits (or re-analyze contemporaneous archival data). If a helium line at the roughly 0.7–1.1% level were present while the NIRISS spectrum remains flat at 1.083 µm under the assumed width, the width assumption would be falsified; if both datasets show no helium, the limits are confirmed.","tokens_in":13602,"feed_emoji":"🔭","tokens_out":15111,"duration_ms":121684,"temperature":0.7,"pith_summary":"LHS 1140 b is a habitable-zone planet whose atmosphere is not well constrained; a recent ground-based campaign reported metastable helium absorption in 2024, which would indicate an escaping, helium-rich atmosphere. This paper analyzes four archival JWST NIRISS transits of the planet, spanning December 2023 to July 2026, for the same helium signature. In every visit a flat continuum is preferred over a model with helium absorption, with odds ratios from roughly 4:1 to 12:1 against the helium model, and a helium line with the reported 2024 amplitude and width is rejected with odds ratios from 300:1 up to 86,000:1. Because none of the JWST transits occurred at the same time as the 2024 detection, the work does not refute that detection; instead it places strict upper limits of 350–550 ppm on any helium line at the four observed epochs. The paper also finds the star's photosphere and surface-heterogeneity parameters consistent across five visits spanning about 2.5 years.","feed_headline":"Four JWST transits of LHS 1140 b show no helium absorption","feed_subtitle":"Archival NIRISS data prefer a flat continuum to helium at four epochs; odds against the 2024 signal reach 86,000 to 1.","key_machinery":"The engine of the analysis is the unresolved 1.083 µm metastable helium triplet seen through NIRISS/SOSS at $R\\approx600$, where the line appears in only one or two spectral channels. The paper treats the line as a Gaussian with its width fixed to 18 Å (the 2024 ground-based line convolved to SOSS resolution) and uses nested-sampling evidence to compare a flat continuum, a Gaussian with free amplitude, and a Gaussian with amplitude fixed to the reported 2024 value. Injection-recovery tests then map the Bayes factor as a function of injected amplitude, converting the data's preference into quantitative upper limits. The photometric stability of JWST, rather than spectral resolution, is what gives these limits sensitivity.","core_discovery":"The central claim is that at every JWST epoch observed, the transmission spectrum of LHS 1140 b near 1.083 µm is consistent with a featureless continuum and inconsistent with a strong metastable helium line. The paper models the unresolved helium triplet as a Gaussian fixed at 18 Å full width at half maximum and compares flat, free-amplitude, and fixed-amplitude models with nested-sampling evidence. The flat model wins in all four visits, with log-Bayes factors against free helium of 1.36–2.45, and against a helium line fixed to the 2024 ground-based amplitude of about 589 ppm of 5.69–11.36. Injection-recovery tests place 100:1-odds upper limits of 350–550 ppm at the SOSS resolution, corresponding to roughly 0.7–1.1% line depths at high resolution, comparable to the 0.6% upper limit from the 2025 ground-based non-detection. The paper also claims that the absolute out-of-transit stellar spectra are consistent across four planet-b transits and one planet-c transit, with no evidence for changes in photospheric temperature or spot and facula fractions.","pith_inferences":["If helium escape from LHS 1140 b is genuinely variable, the fact that all four JWST epochs, which bracket the 2024 detection in time, show no helium suggests the detectable high-escape state is uncommon; a coordinated ground-and-space campaign observing the same transit would test this directly.","The stellar fits infer a hot component roughly 3000 K above the photosphere, far hotter than the warm heterogeneities previously deduced from transit light source modeling; fitting with spot-specific and facula-specific model spectra rather than temperature perturbations would clarify whether this is physical or a model artifact.","Because the quoted limits depend on the assumed 18 Å line width, an extension that marginalizes over line width in the injection-recovery tests would show how much weaker the constraints become for broader or narrower helium lines."],"forward_implications":["The reported 2024 helium detection and the four JWST non-detections can coexist only if helium escape is time-variable or was absent at the JWST epochs; the paper does not claim to refute the detection.","The 100:1 upper limits of 350–550 ppm at the SOSS resolution correspond to about 0.7–1.1% at a resolving power of roughly 68,000, comparable to the 0.6% upper limit from the 2025 ground-based non-detection.","A flat continuum is the preferred model for all four visits, so the data do not require any helium absorption at the 1.083 µm triplet.","The five visits show consistent photosphere and heterogeneity parameters, so these data do not reveal changes in stellar surface properties across the roughly 2.5-year baseline."],"supporting_citations":[{"why":"Reports the 2024 metastable He detection and the 2025 non-detection; supplies the amplitude and width used in the ruling-out model.","marker":"C. Cherubim et al. (2026)"},{"why":"Determines the stellar and planetary parameters used for the light-curve and spectral models and provided two of the analyzed transits.","marker":"C. Cadieux et al. (2024)"},{"why":"Provides the exoTEDRF data-reduction steps used to extract the NIRISS/SOSS spectra.","marker":"M. Radica et al. (2023)"},{"why":"Shows the unresolved He triplet appears in one or two SOSS channels, motivating the Gaussian fixed-width modeling.","marker":"G. Fu et al. (2022)"},{"why":"Established metastable He as a detectable tracer of exoplanet atmospheric escape.","marker":"J. J. Spake et al. (2018)"},{"why":"Provides the theoretical framework linking metastable He absorption to atmospheric escape.","marker":"A. Oklopčić & C. M. Hirata 2018"},{"why":"Describes the NIRISS/SOSS instrument performance and resolution assumed in the analysis.","marker":"R. Doyon et al. (2023)"},{"why":"Converts Bayes factors into the significance thresholds used for the reported upper limits.","marker":"T. Sellke et al. (2001)"},{"why":"Nested sampling routine used to compute Bayesian evidence for the model comparisons.","marker":"J. S. Speagle (2020)"}],"fun_headline_variants":["JWST rules out helium escape in four LHS 1140 b transits","LHS 1140 b shows no helium across four JWST NIRISS visits","Four JWST transits disfavor helium absorption on LHS 1140 b","Null helium result in LHS 1140 b from four JWST spectra","JWST NIRISS data refute helium detection on LHS 1140 b"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quoted odds ratios and upper limits assume that any real helium absorption would have the same 18 Å width as the 2024 ground-based detection after smoothing to JWST's low resolution; a line with a very different width would evade these limits.","fun_headline_variants_meta":{"raw":{"variants":["JWST rules out helium escape in four LHS 1140 b transits","LHS 1140 b shows no helium across four JWST NIRISS visits","Four JWST transits disfavor helium absorption on LHS 1140 b","Null helium result in LHS 1140 b from four JWST spectra","JWST NIRISS data refute helium detection on LHS 1140 b"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000627,"raw_usage":{"total_tokens":3007,"prompt_tokens":1156,"completion_tokens":1851,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":772,"completion_tokens_details":{"reasoning_tokens":1743}},"tokens_in":772,"tokens_out":1851,"duration_ms":11578,"temperature":1.0,"reasoning_tokens":1743,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:18:00.097357+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Observe a high-resolution ground-based spectrum of LHS 1140 b during one of the same epochs as these JWST transits (or re-analyze contemporaneous archival data). If a helium line at the roughly 0.7–1.1% level were present while the NIRISS spectrum remains flat at 1.083 µm under the assumed width, the width assumption would be falsified; if both datasets show no helium, the limits are confirmed.","supporting_citations":[],"review_version":1}