{"id":"5e71aedb-d2a4-4ddb-84ec-d912317b1462","arxiv_id":"2505.22186","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Three JWST eclipses show LHS 1140c is a low-albedo bare rock with T_day = 561±44 K, and a new pixel-level GP method improves eclipse-depth recovery.","lead":"Using three JWST eclipses at 15 microns, the team found that the rocky super-Earth LHS 1140c emits like a bare rock with almost no atmosphere, with a dayside temperature of 561±44 K. They also introduced a new pixel-by-pixel Gaussian process analysis method that can recover eclipse depths more precisely when individual detector pixels misbehave.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The >3σ exclusion of pure H2O atmospheres is conditional on the Koll (2022) heat-redistribution scaling; under f=2/3 the 1-bar H2O model is within 1σ, so the abstract's H2O claim is not a robust bare-rock constraint.","rationale":"The reader and I converge on the heat-redistribution dependence as the weak point. I do not treat the 10.3±3.2% BT-Settl flux offset as load-bearing for the central claim: propagating a larger stellar flux through fp/f* lowers the model eclipse depths in Figure 15, which would make the observed shallow eclipse more, not less, discrepant with thick atmospheres, and the brightness temperature is computed from the measured absolute flux rather than from BT-Settl. The paper has independent support for its main detection: aperture photometry, the pixel GP approach, and a separate pipeline agree, and the pixel method is validated on simulations. The CO2-based exclusions survive the f=2/3 test, so the bare-rock conclusion is not in jeopardy. The H2O exclusion, however, is a headline claim that the authors themselves show vanishes under a different redistribution assumption. That requires an abstract and Section 5 caveat, not rejection; the reader's CONDITIONAL verdict remains appropriate.","tokens_in":46364,"tokens_out":8651,"duration_ms":99717,"concrete_test":"Run a 3D GCM (e.g., ROCKE-3D or ExoFMS) for a 1-bar pure-H2O atmosphere on LHS 1140c using the same stellar spectrum, instellation, rotation, and surface albedo assumptions, and compute the GCM-predicted f and 15 µm eclipse depth with HELIOS-like radiative post-processing. Compare that depth to the joint-fit d=273±43 ppm under the same white-noise plus GP uncertainty model. If the GCM-predicted depth is within 1σ of 273 ppm, the abstract's >3σ H2O exclusion is not supported and the H2O claim must carry the redistribution caveat; if it is still >3σ discrepant, the exclusion becomes robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing assumption sits in Section 5: the pure-H2O and H2O/CO2 forward models are run with 1D HELIOS and the analytic f-factor from Koll (2022), and the text states that the 1-bar pure-H2O exclusion at 3.1σ is 'completely dependent on our heat redistribution model.' The same paragraph says that fixing f=2/3 brings all pure-H2O models up to 1 bar within 1σ of the data. The abstract nevertheless reports pure-H2O atmospheres with surface pressure ≥1 bar as ruled out at >3σ, and no pure-H2O model above 1 bar is actually computed, so '≥1 bar' overstates the grid. The CO2 and N2/CO2 constraints are much less sensitive to the redistribution choice, so the broader 'low-albedo bare rock or optically thin atmosphere' conclusion survives. What does not survive is the H2O-specific exclusion: it is a prediction of one analytic redistribution scaling, not an independent observational constraint, and that scaling is not validated in this paper for a 422 K, 1-bar H2O atmosphere.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents 15 μm MIRI/F1500W eclipse photometry of the super-Earth LHS 1140c from three eclipses, develops and validates a new pixel-level Gaussian process fitting method, and compares the measured eclipse depth to bare-rock and atmospheric forward models. The authors report a robust eclipse detection around 250–275 ppm across aperture photometry, the new pixel-fit, and an independent reduction pipeline, with a dayside brightness temperature of 561±44 K. They interpret this as consistent with a low-albedo bare rock and use HELIOS forward models to exclude a broad range of CO2-bearing and other atmospheres. The paper also documents a possible correlation between MIRI detector settling and the previously used filter, and releases the pixel-fitting code as part of the open-source luas package.","tokens_in":46737,"tokens_out":6665,"duration_ms":80632,"significance":"If accepted, the eclipse measurement would be among the most constraining 15 μm eclipse observations of a cool rocky exoplanet, and the pixel-level GP method is a genuinely useful methodological contribution with realistic simulations and an open-source implementation. The cross-checks between independent reductions, multiple systematics models, and varying reduction choices are a clear strength: Tables 4–7 and G.1–G.2 show the bare-rock interpretation is robust to these choices. However, the headline atmospheric exclusion of pure H2O is conditional on a specific heat-redistribution parameterization, and the abstract states this claim more strongly than the analysis supports, which is a load-bearing issue for the paper's central interpretation.","major_comments":[{"comment":"The abstract's statement that pure H2O atmospheres with surface pressure ≥1 bar are ruled out at >3σ is not supported by the body of the paper. In Section 5 the authors state that the 1 bar pure H2O model is excluded at 3.1σ only when heat redistribution is described by the analytic f-factor of Koll (2022), and that fixing f=2/3 brings all pure H2O models up to 1 bar within 1σ of the data. No pure H2O model with surface pressure above 1 bar is actually computed, so the '≥1 bar' wording overstates the grid tested. The H2O exclusion is a conditional prediction of one redistribution parameterization rather than an independent observational constraint, and the abstract (and the conclusions, which repeat it) should either be reworded to state the conditionality explicitly or drop the H2O claim.","section":"Section 5, pure H2O paragraph; Abstract"},{"comment":"The absolute flux calibration in Section 4.7 shows that the observed F1500W flux exceeds the BT-Settl prediction by 10.3±3.2%, and the text acknowledges that this may shift the forward-model eclipse depths by roughly 10%. However, the exclusion significances quoted in Section 5 and Table 8 are not recomputed under this systematic, even though the stellar model enters both the incident stellar flux and the predicted planet-to-star flux ratio. The stated direction of the effect is conservative for the CO2 constraints, but for the already model-dependent H2O exclusion the quoted 3.1σ does not include this 10% systematic. The authors should either propagate this offset through the model comparisons or explicitly quantify the largest plausible shift in each quoted significance.","section":"Sections 4.7, 5, and Table 8"}],"minor_comments":[{"comment":"The statement that 'an independent analysis' detects the eclipse at >5σ is not true for all variants in Table 6: the optimal-extraction L+E+GP model gives 235±70 ppm (3.4σ) and L+GP gives 254±53 ppm (4.8σ), although several classic-extraction variants do exceed 5σ. Please specify which reduction and detrending combination is being cited.","section":"Abstract and Table 6"},{"comment":"The text says that 'observations which both used the same previous filter tend to have quite consistent slopes' and then two paragraphs later says 'the two observations which previously used the F560W filter show inconsistent settling.' This apparent contradiction should be resolved, for example by explicitly noting that consistency holds for most filters but not for F560W.","section":"Section 4.5"},{"comment":"The sentence introducing the absolute flux calibration says stellar model inaccuracies could affect the eclipse-spectra models 'in two di fferent ways', but the two ways are not explicitly enumerated. Please spell them out for the reader.","section":"Section 4.7"},{"comment":"The Morrison et al. (2023) reference appears twice with different journal abbreviations (PASP 135, 075004 and PASA 135, 075004); these should be merged into a single correct citation.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope and the core bare-rock conclusion appears robust. The main issue is that the abstract overstates the H2O exclusion, which the authors themselves admit is completely dependent on the Koll (2022) heat-redistribution model. This is fixable with rewording and possibly minor additional modeling, but it is a central claim and should not be left as is."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a solid, careful paper, and the part that is most worth your time is not the astrophysical conclusion. The first 15 µm eclipse detection of LHS 1140c is a real measurement, and it is consistent across three independent analyses (aperture photometry, the new pixel-level GP fit, and a secondary pipeline) at better than 5σ. The inferred brightness temperature of 561±44 K fits a low-albedo bare rock with negligible heat redistribution, and the stronger exclusions (pure CO2 ≥10 mbar, CO2/N2 mixtures at 1 bar) survive changes in the assumed heat redistribution. That is a useful result for the cosmic-shoreline debate.\n\nThe genuinely new thing here is the pixel-level Gaussian process method. The authors jointly fit individual pixel light curves with a shared eclipse model, PSF shifts, and a kernel that separates pixel and time covariances, so the exact GP likelihood is tractable. They validate it on 200 simulations in three noise scenarios, and it outperforms aperture extraction when a single pixel carries systematics, while matching it otherwise. The paper ships code (luas) and data, and it uses the method to weight away from a cosmic-ray persistence effect. That is a reusable contribution for JWST time-series work.\n\nThe soft spots are real but localized. The abstract's H2O claim is overstated. Section 5 states that the pure-H2O ≥1 bar exclusion at 3.1σ is 'completely dependent on our heat redistribution model,' and with f=2/3 all pure-H2O models up to 1 bar fall within 1σ. The abstract reports pure-H2O ≥1 bar ruled out at >3σ without that caveat, and the grid does not actually extend above 1 bar. The CO2-based constraints are essentially insensitive to that choice, so the central bare-rock conclusion holds; the H2O sentence should be softened or the grid extended. Second, the absolute flux calibration finds BT-Settl models are ~10% too faint in F1500W, and that offset is not propagated into the forward-model comparisons. The authors flag it and argue it would not change the main conclusion, but a reader has to dig to find it. Minor: the jump-detection threshold was tuned on the same data, though the appendix tests show little sensitivity to that choice.\n\nThis paper deserves a serious referee. I would send it out. The referee should ask for the abstract to match the body, and for the flux-model offset to be propagated or explicitly bounded. The pixel-GP method is the most citable part and should be scrutinized carefully, because the simulations are the main evidence that it works as claimed.","headline":"Solid eclipse measurement and a reusable pixel-level GP method, but the abstract's H2O exclusion overstates what the analysis supports.","tokens_in":47250,"tokens_out":3066,"would_cite":true,"duration_ms":35332,"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":"A deep 15 µm eclipse of the rocky super-Earth LHS 1140c, detected at >5σ by three independent analyses, puts its dayside at 561±44 K, matching a low-albedo bare rock and ruling out a wide range of CO2 and H2O atmospheres at >3σ, while a…","keywords":["exoplanet atmospheres","eclipse photometry","Gaussian process","pixel-level fitting","MIRI","LHS 1140c","heat redistribution","super-Earth"],"falsifier":"A phase-curve observation of LHS 1140c that detects nightside flux or a phase offset, or a rerun of the atmospheric comparison with heat redistribution fixed to zero, would directly test the bare-rock claim.","tokens_in":46188,"feed_emoji":"🔭","tokens_out":7419,"duration_ms":75293,"temperature":0.7,"pith_summary":"This paper claims that three 15 µm eclipses of the rocky super-Earth LHS 1140c, observed with JWST's MIRI imager, show a dayside that is hot, bright, and consistent with a low-albedo bare rock that does not redistribute heat. The recovered dayside brightness temperature is 561±44 K, close to the 537±9 K maximum expected for a zero-albedo rock with no heat redistribution, and inconsistent with full redistribution. Combining this measurement with atmospheric forward models, the paper claims to rule out pure CO2 atmospheres with surface pressures of 10 mbar or more, pure H2O atmospheres of 1 bar or more, and CO2-rich or Earth-like 1 bar atmospheres at better than 3σ. It also introduces a Gaussian process method that fits individual pixel light curves jointly, which on simulated data recovers eclipse depths more precisely and reliably than aperture photometry when systematics contaminate a single pixel. If the bare-rock interpretation is right, LHS 1140c sits on the airless side of the cosmic shoreline and provides a benchmark for atmospheric escape around M dwarfs.","feed_headline":"JWST eclipse shows LHS 1140c is a bare hot rock","feed_subtitle":"Three 15-micron eclipses put the super-Earth's dayside at 561 K and rule out CO2 and water atmospheres.","key_machinery":"The argument is carried by two pieces of machinery. The first is a two-dimensional Gaussian process that joint-fits pixel light curves rather than an aperture sum, with a covariance kernel written as a sum of Kronecker products, $K = K_p \\otimes K_t + \\Sigma_p \\otimes \\Sigma_t$, so the exact likelihood scales as $O(2N_p^3 + 2N_t^3 + N_pN_t(N_p+N_t))$ instead of $O(N_p^3N_t^3)$. Its pixel-side components model flux-conserved anti-correlations between neighbouring pixels, independent pixel systematics, common systematics, interpixel capacitance, and background row/column noise, and its shared time kernel captures PSF evolution; this lets the fit down-weight pixels contaminated by a cosmic-ray persistence effect. The second is the physical interpretation chain: a heat-balance model with a redistribution factor (from Koll 2022) sets the dayside temperature for bare rocks and atmospheres, HELIOS computes emission spectra, and the measured 15 µm eclipse depth is compared to those spectra to exclude atmospheres.","core_discovery":"The central discovery is a deep eclipse: joint fits of the three LHS 1140c eclipses deliver a 15 µm planet-to-star flux ratio of about 270 ppm at >5σ, with aperture photometry, the new pixel Gaussian process fit, and an independent reduction all agreeing. Converting the eclipse depth through an absolute flux calibration gives a dayside brightness temperature of 561±44 K, matching the 537±9 K maximum for a smooth zero-albedo bare rock with no heat redistribution and rejecting the 421±7 K expected from full redistribution. Atmospheric forward models with the analytic heat-redistribution prescription predict that any substantial CO2 or H2O atmosphere would cool the dayside and add 15 µm absorption; the observed depth rules these out, leaving the planet best described as a low-albedo airless rock, possibly with an optically thin or spectrally bland residual atmosphere.","pith_inferences":["If the ~10% stellar-model flux offset is real, the forward-model eclipse depths should be shifted shallower, which would make the atmospheric exclusions even stronger rather than weaker; the brightness temperature itself does not depend on stellar models.","The same anti-correlated pixel systematics found here may explain the excess scatter seen in one-pixel-wide spectroscopic extractions, and the pixel-GP framework could be extended to spectra by sharing the eclipse model only across pixels at the same wavelength.","The eclipse-time constraint that LHS 1140c's secondary eclipse occurs 2.8±0.9 minutes early implies a small non-zero eccentricity; if confirmed by more eclipses, it would be a rare eccentricity constraint for a tidally locked rocky planet.","A direct test of the bare-rock claim is to observe LHS 1140c in a second MIRI band such as F1800W or a shorter wavelength, where CO2 and SO2 have different absorption strengths; the current models predict the same bare-rock depth in those bands."],"forward_implications":["LHS 1140c joins TRAPPIST-1b and c as rocky planets whose 15 µm eclipses point to airless or nearly airless surfaces, strengthening the picture that low-mass M-dwarf planets lose their atmospheres.","A broad class of secondary atmospheres—thick CO2, H2O, SO2/CO2 mixtures, and Earth-like 1 bar N2/O2 with >100 ppm CO2—is excluded, so future characterization can focus on tenuous or optically thin atmospheres.","The new pixel-level Gaussian process method can be applied to other MIRI time series, including LRS spectroscopy, where correlated noise between neighbouring pixels is currently ignored in standard analyses.","The settling-ramp trend with the previously used MIRI filter suggests a simple operational change—switching the filter before target acquisition—could reduce the amount of data needing to be discarded from future observations."],"supporting_citations":[{"why":"Supplies the planetary masses, radii, orbital parameters, and priors on period and eccentricity used in the joint eclipse fits.","marker":"Cadieux et al. 2024a"},{"why":"Supplies priors on transit time, system scale, radius ratio, and impact parameter from NIRISS/SOSS, and the context for the TESS/Spitzer transit-depth discrepancy.","marker":"Cadieux et al. 2024b"},{"why":"Provides the analytic f-factor parameterization that converts surface pressure and composition into heat redistribution for the atmospheric forward models.","marker":"Koll 2022"},{"why":"Provides the BT-Settl stellar spectra used as input to the forward models and as the reference for the absolute flux calibration.","marker":"Allard et al. 2011"},{"why":"Implements the exact two-dimensional Gaussian process with Kronecker-product covariance that the new pixel-fitting method is built on.","marker":"Fortune et al. 2024"},{"why":"Supplies the transit and eclipse model used as the mean function for all eclipse fits.","marker":"Mandel & Agol 2002"},{"why":"Provides expected interpixel capacitance values and the first-exposure effect context used in building the pixel kernel and verifying the fitted IPC parameters.","marker":"Morrison et al. 2023"},{"why":"Provides the F1500W absolute flux calibration factor used to convert the eclipse depth into a brightness temperature.","marker":"Gordon et al. 2025"}],"fun_headline_variants":["LHS 1140c: JWST eclipse shows bare rock","Super-Earth LHS 1140c is a hot bare rock","JWST: LHS 1140c rules out CO2 and H2O atmospheres","LHS 1140c dayside at 561 K rules out thick atmospheres"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The pure H2O and mixed-atmosphere exclusions depend on the analytic heat-redistribution formula relating surface pressure to redistribution; if that formula is replaced by assuming no redistribution, 1-bar pure H2O models become consistent, and the comparison does not propagate the observed ~10% excess of stellar flux over the BT-Settl models used.","fun_headline_variants_meta":{"raw":{"variants":["LHS 1140c: JWST eclipse shows bare rock","Super-Earth LHS 1140c is a hot bare rock","JWST: LHS 1140c rules out CO2 and H2O atmospheres","LHS 1140c dayside at 561 K rules out thick atmospheres"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000682,"raw_usage":{"total_tokens":3181,"prompt_tokens":1113,"completion_tokens":2068,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":729,"completion_tokens_details":{"reasoning_tokens":1982}},"tokens_in":729,"tokens_out":2068,"duration_ms":15914,"temperature":1.0,"reasoning_tokens":1982,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T13:14:07.621233+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A phase-curve observation of LHS 1140c that detects nightside flux or a phase offset, or a rerun of the atmospheric comparison with heat redistribution fixed to zero, would directly test the bare-rock claim.","supporting_citations":[],"review_version":1}