{"id":"655267a0-4ac9-4cff-bbb9-72f772bea6f2","arxiv_id":"2506.19193","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"The UV-to-IR transmission spectrum of the hot Jupiter KELT-7b requires a H- abundance about seven orders of magnitude above equilibrium chemistry, implying photochemical disequilibrium, and also reveals bright patches on the host star.","lead":"Astronomers measured how starlight filters through the atmosphere of the hot gas giant KELT-7b, from ultraviolet to infrared. They found that a rare form of hydrogen, the negative hydrogen ion, is far more abundant than standard chemistry predicts, pointing to an atmosphere thrown out of equilibrium by starlight.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The H- abundance claim depends on a phenomenological bright-inhomogeneity model that the paper itself calls physically infeasible for KELT-7; adding the TESS point shifts log H- by >1 dex, so the disequilibrium enhancement is not yet robust.","rationale":"I read the central claim as the high-confidence H- detection at log H- ~ -5 with Bayes factors of 17,000 and 6,000. The strongest evidence for the paper is the dual independent reduction, dual retrieval codes, and the fact that H- is still preferred when the UVIS data are removed (Appendix A). However, the quantitative abundance and the 'seven orders above equilibrium' statement are not robust to reasonable model choices. The paper's own Table 1 shows a 1.3 dex shift when the TESS point is included, and Section 4.4 excludes TESS by arguing its baseline is uncertain. That is a legitimate treatment, but it means the headline abundance is not a stable observable. The stellar contamination model is the natural culprit: the blue slope that drives the bright-inhomogeneity component is the same spectral region where H- and stellar heterogeneity are hardest to separate. The paper's Section 4.5 admission that faculae are physically infeasible for this star means the contamination component is not physically validated. This is a correctness risk rather than a circularity issue because both retrieval codes share the same 3-parameter prescription, so agreement between codes does not test this assumption. The proposed reanalysis with the TESS point and an offset prior directly tests whether the preferred exclusion of the most precise continuum datum changes the central claim. If the H- posterior is stable, the conditional accept stands; if not, the paper should report the H- abundance as degenerate with the stellar model and lower the significance claim. The abstract's overstatement about an asymmetric water feature is an additional editorial problem but is not the load-bearing issue for the central H- claim.","tokens_in":31198,"tokens_out":9056,"duration_ms":107291,"concrete_test":"Re-run both NEMESISPY and POSEIDON retrievals including the TESS point, but give the TESS-UVIS offset a wide Gaussian prior (e.g., N(0, 100 ppm) or a uniform +/-200 ppm) instead of fixing it to zero, and record the marginalized log H- posterior and the Bayes factor relative to a no-H- model. If log H- remains within about 0.3 dex of the no-TESS preferred value (-4.94 to -5.32) and the Bayes factor stays above ~100, the concern is answered. If the posterior shifts toward the with-TESS value (-5.8 to -6.2) or becomes bimodal, the H- abundance is not robust to the excluded data point, and the claimed >1 dex disequilibrium enhancement should be downgraded.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that H- is present at log10(H-) ~ -5, about seven orders of magnitude above equilibrium, with 4.6-4.8 sigma significance. That claim is computed inside a retrieval that simultaneously fits a 3-parameter stellar-inhomogeneity model (T_phot, Delta T_het, f_het; Sections 4.1-4.2). The paper itself states in Section 4.5 that true faculae are physically infeasible for the fast-rotating star KELT-7, so the bright patches are a phenomenological catch-all. The data require a blueward decrease in transit depth, and this catch-all provides it; the H- abundance is inferred partly through continuum opacity over the same UV-optical region. The fragility of this separation is visible in the authors' own TESS experiment: adding one precise 0.8 micron photometric point shifts log H- from -4.94 to -6.20 in NEMESISPY (Table 1, Figure 6) and lowers f_het from 0.09 to 0.05, yet the point is excluded because of a possible baseline offset. Appendix A similarly shows that removing the UVIS data moves the POSEIDON H- abundance from -5.32 to -6.07. The four-point spread (-4.94 to -6.20) brackets rather than pins down the claimed 'seven orders above equilibrium' enhancement. Until the stellar contamination model is validated or the TESS point is included with a defensible offset treatment, the H- abundance and its significance are conditional on an unverified model component.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents the first UV-optical transmission spectrum of the hot Jupiter KELT-7b, obtained with HST WFC3/UVIS G280 from 0.2 to 0.8 micron, and combines it with previously published WFC3/IR G141 and Spitzer photometry to form a 0.2-5.4 micron dataset. Two independent reduction pipelines (lluvia and hazelnut) produce broadly consistent spectra. Atmospheric retrievals with two independent codes, NEMESISPY and POSEIDON, find that the spectrum is best explained by H- bound-free opacity with log10 H- abundances of -4.94 and -5.32, respectively, with Bayesian evidences of 17,000 and 6,000, together with bright stellar inhomogeneities covering about 9% of the stellar disk at roughly 300 K above the photosphere. The retrievals do not detect H2O, find tentative CO2 driven entirely by the two Spitzer points, and forward equilibrium chemistry models cannot reproduce the retrieved H- abundance, which the authors interpret as evidence for disequilibrium. The paper also tests the impact of including a TESS photometric point and finds that it shifts the H- abundance, but the authors exclude that point because of a suspected baseline offset.","tokens_in":31561,"tokens_out":7239,"duration_ms":77865,"significance":"If the H- detection holds, this paper would add KELT-7b to the small set of hot Jupiters with strong evidence for H- opacity and would strengthen the case that UV-optical observations are essential for revealing disequilibrium chemistry in the upper atmospheres of hot Jupiters around F-type stars. The study has clear strengths: two independent reduction pipelines, two independent retrieval codes, explicit Bayesian model comparison, a forward equilibrium chemistry check, and a transparent appendix showing the impact of removing the UVIS data. The two codes agree on the H- abundance and on the stellar contamination parameters, and the forward models provide an independent falsifiable test. However, the central claim is conditional on the adopted three-parameter stellar contamination model, which the authors themselves describe as a phenomenological catch-all because true faculae are physically infeasible for the fast-rotating, radiative-envelope host star KELT-7. The sensitivity of the H- abundance to the TESS point and to the presence or absence of UVIS data means that the quoted abundance is not yet robust at the level implied by the abstract.","major_comments":[{"comment":"The central H- abundance is highly sensitive to the treatment of the TESS photometric point. Including TESS shifts the NEMESISPY log H- from -4.94 to -6.20 and the POSEIDON log H- from -5.32 to -5.79 (Table 1), while the text in Section 4.4 reports -5.50 for POSEIDON, a discrepancy that should be reconciled. The authors exclude TESS because of a possible baseline offset, but this exclusion is itself a modeling choice. Because the TESS point directly samples the optical continuum where H- and stellar contamination are degenerate, the paper should treat the TESS/no-TESS difference as a systematic uncertainty, report a combined abundance range, or explicitly present a retrieval with a free TESS offset as a sensitivity test even if it weakens the constraint.","section":"Section 4.3, Table 1, and Section 4.4"},{"comment":"Both retrieval codes adopt the same three-parameter stellar contamination prescription (photosphere temperature, heterogeneity temperature difference, and covering fraction, from Rathcke et al. 2021), and Section 4.5 states that true stellar faculae are physically infeasible for the fast-rotating star KELT-7. The bright inhomogeneities are therefore a phenomenological catch-all. Since the blue-UV slope can be absorbed either by H- continuum opacity or by the stellar contamination parameters, the agreement between NEMESISPY and POSEIDON does not validate the separation between these two contributions. The authors should test an alternative parameterization of the stellar contamination, for example a different spectral template for the inhomogeneities or a prior allowing both cooler and hotter regions simultaneously, and show whether the H- posterior and the claimed Bayes factors are stable under such changes.","section":"Section 4.1, 4.2, and 4.5"},{"comment":"The forward equilibrium chemistry comparison is a useful falsifiable test, but it inherits the stellar contamination and cloud assumptions from the free retrieval. The forward models are not fit to the data, and they are anchored to the maximum-likelihood retrieved T-P profile and to the retrieved stellar contamination parameters. The conclusion that the H- abundance is enhanced by disequilibrium processes is therefore conditional on the phenomenological stellar inhomogeneity model being correct. The paper should state this limitation explicitly, or run the forward models across a range of stellar contamination parameters, to quantify how much of the inferred disequilibrium could be absorbed by an unmodeled stellar surface component.","section":"Section 4.3, forward equilibrium models"}],"minor_comments":[{"comment":"The abstract says the spectrum shows 'an asymmetric water feature in the 1.1-1.7 micron band,' but Section 4.3 concludes 'We do not detect H2O in KELT-7b's atmosphere.' Please rephrase the abstract so that it does not claim a water feature when the retrievals attribute the G141 shape primarily to H- and do not confidently detect H2O.","section":"Abstract"},{"comment":"There is a stray 'red' in the sentence 'which agrees within 4.5 sigma with the T0 obtained with lluvia. red The difference...' that should be removed.","section":"Section 3.2"},{"comment":"The POSEIDON log H- value with TESS is given as -5.50 +/- 0.63 in Section 4.4 but as -5.79 +0.83/-0.71 in Table 1. Please make these consistent.","section":"Table 1 and Section 4.4"},{"comment":"The caption refers to 'NEMESIS' while the text uses 'NEMESISPY'; please unify the naming.","section":"Figure 5 caption"},{"comment":"There is a typo: 'a asymmetric water feature' should be 'an asymmetric water feature.'","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The manuscript reports a valuable new dataset and an honest, transparent analysis, including two pipelines, two retrieval codes, and explicit caveats about the TESS point and the stellar contamination model. My main concern is that the headline H- abundance and its detection significance are conditional on a phenomenological stellar inhomogeneity treatment that the authors themselves argue is not physically motivated for this star. The sensitivity to the TESS point and to the UVIS data is already demonstrated in the paper, so the necessary robustness tests are within reach. I would support publication after a major revision that quantifies this systematic uncertainty and reconciles the text/table inconsistencies."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on arXiv:2506.19193. The genuinely new piece is the 0.2–0.8 μm WFC3/UVIS G280 transmission spectrum of KELT-7b, reduced with two independent pipelines that agree well, combined with previously published G141 and Spitzer data. The paper's main claim is that the combined spectrum requires H− at log10(H−) ≈ −5, roughly seven orders of magnitude above equilibrium, with strong Bayes factors (6,000–17,000) from two independent retrieval codes, and that equilibrium forward models cannot reproduce the spectrum. That is a real result and the work is careful: two reductions, two retrievals, nested-sampling evidence, and appropriately cautious language about CO2.\n\nThe soft spots are real too, and they mostly circle around one issue: the H− abundance is entangled with the stellar contamination model. Both retrievals adopt a 3-parameter phenomenological bright-inhomogeneity prescription (T_phot, ΔT_het, f_het). The paper itself notes in Section 4.5 that true faculae are physically implausible for the fast-rotating KELT-7, so this component is a catch-all. The H− abundance and covering fraction are degenerate with it. The TESS experiment shows this: adding a single precise 0.8 μm point moves log H− from −4.94 to −6.20 in NEMESISPY and cuts f_het from 0.09 to 0.05. The point is excluded because of a possible baseline offset, which is defensible, but the spread across the four retrieval configurations (−4.94 to −6.20) brackets, rather than pins, the claimed enhancement. The detection significance of 4.6–4.8σ is computed inside a model that already includes this unvalidated component, so it is conditional on it. Also, the abstract claims an 'asymmetric water feature' in the 1.1–1.7 μm band, but neither retrieval detects H2O; the paper's own Section 6 says no clear evidence of water. That should be fixed.\n\nNone of this invalidates the core result. The spectrum is new, the dual-pipeline agreement is solid, and the H− preference is consistent between two codes. But the headline abundance and significance should be presented with a clearer caveat about the stellar model and the TESS sensitivity.\n\nThis paper deserves a serious referee. It is a solid observational contribution for the exoplanet atmosphere community, and the H− vs stellar-contamination degeneracy is a useful methodological lesson. I would send it to review, with a request to fix the abstract and to add a retrieval that marginalizes over or directly explores the stellar-inhomogeneity parameterization. I'd bring it to reading group to discuss the TESS inclusion decision and the model dependence.","headline":"A careful dual-pipeline, dual-retrieval paper whose new G280 spectrum is solid but whose headline H- abundance is heavily model-dependent, and the abstract overstates a water feature the retrievals do not detect.","tokens_in":32304,"tokens_out":2844,"would_cite":true,"duration_ms":28153,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"KELT-7b's atmosphere is flooded with H- ions, seven orders beyond equilibrium.","keywords":["Exoplanet atmospheres","Transmission spectroscopy","Hot Jupiters","KELT-7b","H- opacity","UV-optical spectroscopy","HST WFC3/UVIS G280","Atmospheric retrieval"],"falsifier":"A JWST NIRSpec G395H spectrum from 2.9 to 5.1 microns would settle the claim: if the retrieved H- abundance is correct, the continuum should show the corresponding H- opacity and the 4.3 micron CO2 feature should appear as retrieved; if the blue slope was caused by unmodeled stellar heterogeneity instead, the infrared spectrum would not require H- at that level. Independent Doppler imaging or spot-modulation photometry that finds no bright regions at the retrieved covering fraction would also refute the stellar contamination part and call the H- abundance into question.","tokens_in":30959,"feed_emoji":"🪐","tokens_out":6503,"duration_ms":59767,"temperature":0.7,"pith_summary":"The paper presents a new 0.2-0.8 micron transmission spectrum of the hot Jupiter KELT-7b from HST WFC3/UVIS, combined with previously published near-infrared G141 data and Spitzer photometry. The central claim is that the combined spectrum is best explained by a very high abundance of the negative hydrogen ion H-, around log10 H- = -5, about seven orders of magnitude above the equilibrium chemistry prediction near $10^{-12}$. Two independent retrieval codes agree on this result with Bayesian detection significances above 4.5 $\\sigma$, and the excess H- is attributed to photochemical disequilibrium in the upper atmosphere. The paper also finds that the UV-optical data are needed to constrain H- and to reveal bright stellar surface inhomogeneities that infrared data alone would miss.","feed_headline":"H- ions flood KELT-7b's atmosphere, 10^7 times equilibrium levels","feed_subtitle":"UV-optical spectra reveal hydrogen anions, not clouds, shape this hot Jupiter's atmosphere.","key_machinery":"The load-bearing object is the H- bound-free continuum, the continuous absorption of the negative hydrogen ion across UV-optical wavelengths, which both retrieval codes need to fit the slight blue slope. The second piece of machinery is the 3-parameter stellar contamination model from Rathcke et al. (2021), which represents unocculted bright regions on the stellar disk by a temperature offset and a covering fraction; the paper calls these 'stellar inhomogeneities' rather than faculae because the fast-rotating host star makes true faculae physically infeasible. Retrievals compare models with and without each opacity source via Bayesian evidence, which converts the Bayes factors into the reported detection significances.","core_discovery":"On its own terms, the paper claims that KELT-7b's transmission spectrum from 0.2 to 5.4 microns is dominated by H- bound-free opacity rather than by molecular bands or clouds. Both retrieval codes, NEMESISPY and POSEIDON, return H- abundances of log H- = -4.94 (+0.89/-1.10) and -5.32 (+0.75/-0.94), with Bayes factors of 17,000 (4.8 sigma) and 6,000 (4.6 sigma) relative to models without H-. The spectrum is otherwise featureless longward of 0.4 microns, shows an asymmetric water feature in the G141 bandpass that is not required, and a tentative CO2 signal driven solely by the two Spitzer points. The same retrievals find bright stellar inhomogeneities covering about 9% of the stellar disk at roughly 300 K above the photosphere, a 2.7-sigma detection. Equilibrium chemistry forward models cannot reproduce the observed H- abundance, so the paper interprets the excess as evidence for photochemistry producing atomic H and free electrons in KELT-7b's upper atmosphere.","pith_inferences":["If the H- super-abundance is real, higher-resolution UV spectroscopy of KELT-7b's terminator should show the bound-free continuum with a distinctive wavelength dependence that aligns with the retrieved abundance, while stellar contamination models would predict a steeper blue slope tied to photospheric structure.","A testable extension is to check the stellar inhomogeneity interpretation against independent stellar activity diagnostics: if time-series photometry or Doppler imaging finds no bright regions at the retrieved temperature and covering fraction, the H- abundance required by the retrievals would be biased.","The failure of equilibrium chemistry by seven orders of magnitude suggests that photochemical models of KELT-7b's upper atmosphere should predict a specific departure of H- abundance with altitude; comparing the retrieved column to such a profile would test the disequilibrium mechanism.","Since the paper excludes the TESS point from its preferred solution due to baseline uncertainty, a dedicated joint analysis of the TESS transits with a free baseline offset could test whether the stellar inhomogeneity constraints are robust."],"forward_implications":["KELT-7b joins HAT-P-41b and WASP-79b as hot Jupiters around F-type stars with H- enriched upper atmospheres, suggesting EUV-driven photochemistry may be common in this population.","Observations with JWST redward of 1 micron, such as NIRISS SOSS and NIRSpec G395H, should further constrain the H- abundance and confirm or refute the tentative CO2 detection.","The cloudy-to-clear transition in hot Jupiter atmospheres may occur at lower temperatures than previously thought if H-, rather than a cloud deck, is what flattens the UV-optical spectra of these planets.","Including UV-optical coverage in transmission retrievals is necessary to break the degeneracy between clouds and H- and to detect bright stellar inhomogeneities that are invisible in infrared data alone."],"supporting_citations":[{"why":"Provides the framework for H- detection in hot Jupiters around F stars and the stellar inhomogeneity parameterization used here.","marker":"Lewis et al. 2020"},{"why":"Supplies the 3-parameter stellar contamination prescription that both retrieval codes adopt.","marker":"Rathcke et al. 2021"},{"why":"Provided the WFC3/IR G141 spectrum and TESS transit depth that are combined with the new UVIS data.","marker":"Pluriel et al. 2020"},{"why":"Provided the Spitzer 3.6 and 4.5 micron photometry used in the retrievals.","marker":"Baxter et al. 2021"},{"why":"Supplies the H- bound-free opacity cross-sections that are central to the detection.","marker":"John 1988"},{"why":"Predicts H- negligible under equilibrium for T < 2500 K, setting the comparison baseline for the retrieved excess.","marker":"Kitzmann et al. 2018"},{"why":"Provides the photochemical pathway (free electrons plus atomic H) that could elevate H- to the observed levels.","marker":"Lavvas et al. 2014"}],"fun_headline_variants":["Hydrogen anions rule KELT-7b's spectrum, 10^7 times expected","UV data exposes H- flood in hot Jupiter KELT-7b's air","KELT-7b's haze isn't clouds—it's H- ions, say retrievals","H- abundances spike 10^7-fold in KELT-7b's upper atmosphere","KELT-7b's UV spectrum points to H- excess, not clouds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The retrieved H- abundance depends on the assumption that the star's surface can be represented as a single uniform photosphere plus single-temperature bright patches; if the real stellar heterogeneity is more complex, the H- abundance needed to fit the blue slope could be biased.","fun_headline_variants_meta":{"raw":{"variants":["Hydrogen anions rule KELT-7b's spectrum, 10^7 times expected","UV data exposes H- flood in hot Jupiter KELT-7b's air","KELT-7b's haze isn't clouds—it's H- ions, say retrievals","H- abundances spike 10^7-fold in KELT-7b's upper atmosphere","KELT-7b's UV spectrum points to H- excess, not clouds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000737,"raw_usage":{"total_tokens":3413,"prompt_tokens":1183,"completion_tokens":2230,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":799,"completion_tokens_details":{"reasoning_tokens":2111}},"tokens_in":799,"tokens_out":2230,"duration_ms":16677,"temperature":1.0,"reasoning_tokens":2111,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T23:07:41.094556+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A JWST NIRSpec G395H spectrum from 2.9 to 5.1 microns would settle the claim: if the retrieved H- abundance is correct, the continuum should show the corresponding H- opacity and the 4.3 micron CO2 feature should appear as retrieved; if the blue slope was caused by unmodeled stellar heterogeneity instead, the infrared spectrum would not require H- at that level. Independent Doppler imaging or spot-modulation photometry that finds no bright regions at the retrieved covering fraction would also refute the stellar contamination part and call the H- abundance into question.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the photochemical pathway (free electrons plus atomic H) that could elevate H- to the observed levels."}],"review_version":1}