{"id":"4d1f871c-af5f-43a0-836b-29d74f37be39","arxiv_id":"2501.00774","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A new non-LTE correction grid for neutral sulfur lines in A-K stars, based on updated atomic collision data, gives consistent sulfur abundances across multiple spectral lines in 13 test stars.","lead":"Astronomers calculated a new grid of corrections for sulfur lines in A-K type stars, accounting for departures from local thermodynamic equilibrium. The grid updates an earlier model with newer atomic collision data and extends the usable range to hotter A-type stars and infrared H-band lines.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Validation stars all have log g ≥ 2.32, so the grid's largest NLTE corrections (log g = 0–1) are tested only by extrapolation; a low-gravity mismatch would invalidate the grid's core regime.","rationale":"The reader identified the solar-calibrated gf values as the weakest assumption; that is a valid concern, and the paper does use the IR-triplet lines (whose gf are not adjusted) as an independent anchor for the Sun. However, the more load-bearing gap for the grid's stated purpose is the absence of any validation at log g < 2.32, where NLTE corrections are largest and where hydrogen collisions dominate the departure coefficients. Even if the gf transfer is perfect, an error in the hydrogen collision rates would corrupt the corrections for low-gravity stars without being visible in the current sample. The paper's own comparison with Takeda et al. (2005) at log g = 2 and 4 shows the model differences are already ~0.1–0.2 dex at log g = 2; the unvalidated log g = 0–1 region could be substantially worse. The proposed test is a single, observationally feasible check that directly probes this regime. Because the reader's verdict is already CONDITIONAL, and my concern reinforces (rather than overturns) that conditionality, I recommend no change to the verdict. Credit is given for the genuine multi-line consistency in the 13 stars, the un-tuned IR triplet solar fit, and the use of modern quantum-mechanical collision data; the issue is purely the extrapolation to the grid's extremes.","tokens_in":37117,"tokens_out":8800,"duration_ms":88182,"concrete_test":"Obtain high-resolution (R ≥ 50,000), high-S/N spectra of a K giant with log g ≲ 1.5 (e.g., α Boo or a similar bright giant) covering the S I lines at λλ 6052, 8694, 9212, 10455 (and ideally the H-band multiplet at λ15400). Derive sulfur abundances from each multiplet using the paper's recommended line parameters and the published NLTE corrections (or a direct MULTI calculation at the star's measured parameters). If the IR-triplet abundance agrees with the optical-multiplet abundance within ~0.1 dex, the low-gravity extrapolation is supported. A disagreement of more than ~0.2 dex would indicate that the hydrogen collision rates or the gf calibration do not transfer to the regime where the grid corrections are largest, weakening the grid's applicability to giants and supergiants.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The published grid is claimed for effective temperatures 4000–10000 K and log g 0–5 (Section 6), but every one of the 13 test stars has log g ≥ 2.32 (Table 2). The strongest NLTE effects occur at low gravity: for log g = 0 the IR-triplet corrections in Figs. 7–11 reach magnitudes of order −1 dex or larger, whereas the most NLTE-sensitive test star (HD 195295, log g = 2.32) shows corrections only down to −0.83 dex. In a cool low-gravity atmosphere the level populations are controlled chiefly by inelastic H-impact collisions (Belyaev & Voronov 2020), which dominate the statistical equilibrium precisely where the grid corrections are largest. The solar calibration of the optical log gf values (Section 3) cannot test this regime because the Sun has log g = 4.44 and hydrogen collisions are far less influential there. Thus the central claim of model adequacy across the grid rests on an extrapolation from dwarfs and moderate giants to the very stars (supergiants, low-gravity giants) where the correction grid is most needed. The paper provides no observed spectra of a star with log g below 2, so a failure in the low-gravity regime would go undetected while affecting exactly the grid's most extreme entries.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents an updated non-LTE atomic model for neutral sulfur and a grid of non-LTE abundance corrections for S I lines in the visible and infrared, including the H-band, for effective temperatures 4000-10000 K, log g 0-5, and [Fe/H] 0 to -2. The model incorporates quantum-mechanical inelastic collision rates with electrons (ADAS) and hydrogen (Belyaev & Voronov 2020), adds S II and S III levels to handle higher temperatures, and is tested by fitting S I and S II lines in 13 stars with well-determined parameters. The authors refine the wavelengths and oscillator strengths of several S I multiplets using the solar spectrum, propose a recommended line list, and compare their correction grid with earlier non-LTE calculations.","tokens_in":37503,"tokens_out":7888,"duration_ms":78161,"significance":"If the results hold, the model and correction grid provide a modern, practical tool for sulfur abundance determinations in A-K stars, with genuinely improved input physics. The strongest evidence is the solar fit of the 10455-10459 Å IR triplet using laboratory-measured oscillator strengths at (S/H)=7.16, and the S I/S II agreement in the two A-type stars. The multi-line, multi-star consistency test is a meaningful relative test of the non-LTE calculations. The electronic correction grid is a useful community resource. However, the absolute scale of the visible and H-band analyses is tied to the meteoritic solar abundance by construction, and the low-gravity portion of the grid is not covered by any of the 13 test stars; both points limit the strength of the validation as stated.","major_comments":[{"comment":"The oscillator strengths for multiplets 6, 8, and 10 and for the H-band lines are adjusted so that solar synthetic profiles match the adopted meteoritic sulfur abundance (log gf ZB +0.088 for multiplet 6, +0.075 for multiplet 8, +0.113 for multiplet 10, and BQZ -0.12 for the 15400 Å multiplets). Consequently, the agreement of these lines with the solar spectrum is partly built in and cannot serve as an independent validation of the absolute abundance scale; only the 10455-10459 Å triplet, with oscillator strengths measured by Zerne et al. (1997), provides an independent solar test. The 13-star consistency check then uses the solar-calibrated gf values, so it tests the transferability of the calibration and the non-LTE corrections, but not the absolute zero point. The authors should make this distinction explicit, quote uncertainties for the fitted gf offsets, and demonstrate quantitatively that the offsets lie within the mutual scatter of the theoretical sources (ZB, BQZ, DH), as the text currently suggests but does not document.","section":"Section 3, Table 1, Section 5"},{"comment":"The grid is claimed for log g 0-5, but all 13 test stars have log g greater than or equal to 2.32 (Table 2; HD 195295 has log g=2.32). The largest non-LTE corrections occur precisely in the low-gravity regime (log g=0-1), where the corrections for the IR triplet reach magnitudes of order -1 dex, whereas the most non-LTE-sensitive test star shows corrections only down to about -0.8 dex. The solar calibration in Section 3 cannot test this regime because the Sun has log g=4.44 and H-impact collisions are far less influential there. Thus the most extreme entries of the correction grid are validated only by extrapolation from dwarfs and moderate giants. The authors should either analyze spectra of low-gravity giants or supergiants with log g below about 2, or explicitly state that grid entries below log g approximately 2 are extrapolations that require individual non-LTE calculations.","section":"Section 6, Table 2, Figs. 7-11"},{"comment":"The quoted uncertainties are only the line-to-line scatter of the fitted abundances. Systematic errors from the adopted stellar parameters (Jofré et al. 2015; Lyubimkov et al. 2010), from the roughly 0.1-0.3 dex uncertainty in the oscillator-strength sources, and from continuum placement are not propagated. Since the paper's central claim is that lines with very different non-LTE corrections yield similar abundances, a quantitative sensitivity analysis is needed to show that the scatter among lines and stars is compatible within the full error budget. For example, the authors should report abundance variations for perturbations of Delta Teff=+-100 K, Delta log g=+-0.2, Delta [Fe/H]=+-0.1, and Delta Vt=+-0.2 km/s, and for choosing ZB, BQZ, or DH gf values rather than their adopted offsets.","section":"Section 4.1, Table 2"}],"minor_comments":[{"comment":"The statement that sulfur lines in all test stars are fitted with similar abundances is stronger than the data support: for HD 84937 the 6757 Å line has an equivalent width below 1 mÅ and only a depression is visible, and several cells in Table 2 are empty because of blends, weakness, or missing spectral coverage. Please qualify the claim accordingly.","section":"Abstract, Section 4.1"},{"comment":"The example grid table contains many blank entries without a legend; please specify whether blanks mean 'correction not computed because equivalent width is below 5 mÅ' or 'outside the computed range', so that users can interpret the electronic grid correctly.","section":"Section 6, Table 4"},{"comment":"Adding a column with the source and estimated uncertainty of each adopted log gf would make the recommended line list more reproducible and would support the authors' statement that the refined values are consistent with the theoretical scatter.","section":"Section 3, Table 1"},{"comment":"The model description would benefit from a brief justification for including S III and S VI but not S IV and S V, since the ionization balance matters at the 10000 K end of the grid and affects the particle conservation underlying the non-LTE calculations.","section":"Section 2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within scope for a stellar-abundance journal. The main risk is overstatement: the low-gravity regime of the correction grid is validated only by extrapolation, and the solar calibration of the visible-line oscillator strengths makes the absolute abundance scale partly circular. Both issues are fixable by additional analysis or by narrowing the claims, so major revision is appropriate rather than rejection. The adoption of modern inelastic collision data and the independent IR-triplet solar fit are genuine strengths that should be preserved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a genuinely useful update of the sulfur non-LTE correction grid, and the validation is decent for the parameter range actually tested. The weak spot is that the grid is advertised for log g = 0–5 but every validation star sits at log g ≥ 2.32, so the most extreme corrections are pure extrapolation.\n\nWhat's new: the authors replace the Drawin-formula and approximate electron rates with quantum-mechanical rates (Belyaev & Voronov 2020, ADAS), add S II levels to push the model to 10000 K, add H-band lines, and refine wavelengths and gf's for several multiplets. The grid is a real extension of Korotin 2009, not a re-badging. The best evidence is the solar IR-triplet fit: using measured oscillator strengths from Zerne et al., they get 7.16 against the meteoritic 7.15, with no tuning. That is an independent success. The 13-star multi-line consistency is also a meaningful test, though less clean because the optical gf's were solar-tuned.\n\nWhere it's soft: the low-gravity regime is exactly where the corrections are largest (approaching –1 dex or more) and where hydrogen collisions dominate, and there is no observed spectrum below log g = 2.32. The solar calibration cannot cover this because the Sun is a dwarf. This does not invalidate the model, but it means the grid's most extreme entries are unsupported extrapolation. A referee should ask for at least one low-gravity benchmark (e.g., a well-studied supergiant) or a clear statement that the low-log g corrections are provisional. Also, the quoted errors are line-to-line scatter only; systematic uncertainties from gf adjustments and stellar parameters are not quantified. The paper excludes several awkward lines, which is defensible but should be explicit.\n\nWho it's for: stellar abundance practitioners working on A–K stars, especially those with IR spectra. It deserves peer review; with the low-gravity caveat and a proper error budget it can become a standard reference. I would take the grid seriously, with caution at low gravity.","headline":"A useful incremental update to the sulfur non-LTE grid with a better atomic model and H-band coverage, but the low-gravity end is untested extrapolation and the optical gf tuning is partly circular.","tokens_in":37994,"tokens_out":2470,"would_cite":true,"duration_ms":26269,"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 revised neutral sulfur model gives consistent sulfur abundances in 13 stars across 4000 to 10000 K, regardless of how strongly individual lines deviate from LTE.","keywords":["sulfur abundance","non-LTE","stellar atmospheres","A-type stars","K-type stars","oscillator strengths","infrared spectroscopy","atomic model"],"falsifier":"For a metal-poor giant with effective temperature 6000 K, surface gravity $\\log g=2$, and $[\\mathrm{Fe/H}]=-1$, the new model predicts a non-LTE correction of about -0.75 dex for the 9212 Å line, whereas the old model gives -0.85 dex and an earlier independent calculation gives -1.02 dex; a high-resolution, high-S/N spectrum of such a star that yields a sulfur abundance from 9212 Å differing from the 8694 Å or 6757 Å lines by more than the stated uncertainties would contradict the model. A laboratory measurement of the multiplet 6, 8, and 10 oscillator strengths that disagrees with the refined values by more than the quoted adjustments would likewise falsify the solar-calibration assumption.","tokens_in":36944,"feed_emoji":"⭐","tokens_out":12703,"duration_ms":99282,"temperature":0.7,"pith_summary":"The paper argues that a renewed atomic model for neutral sulfur, using quantum-mechanical collision rates and including ionized sulfur levels, can describe sulfur line formation in stellar atmospheres from 4000 to 10000 K. On 13 well-studied A-K stars, the model fits S I lines from different multiplets with essentially the same sulfur abundance even when those lines are affected very differently by departures from local thermodynamic equilibrium. The authors also present a grid of non-LTE corrections for S I lines in the visible, near-infrared, and H-band regions, together with refined wavelengths and oscillator strengths. If the claim holds, sulfur abundances derived from weak optical lines and from strongly non-LTE infrared triplets can be placed on a common scale, which matters for using sulfur as an alpha-element tracer of galactic chemical evolution.","feed_headline":"Sulfur abundances unified across 13 stars with new non-LTE model","feed_subtitle":"Updated atomic model and correction grid cover A-K stars from 4000 to 10000 K, including IR H-band lines.","key_machinery":"The central object is a 64-level neutral sulfur atom augmented by 81 levels of ionized sulfur, the ground level of doubly ionized sulfur, and auxiliary levels treated in LTE, with 775 bound-bound and 146 bound-free transitions. Collisional data come from detailed quantum-mechanical electron-impact rates for the lowest 17 S I levels and hydrogen-impact rates for 40 levels, replacing the earlier approximate formulas. Statistical equilibrium and radiative transfer are solved with a modified non-LTE atmosphere code using opacity distribution functions, and the resulting population ratios are passed to a synthetic-spectrum code to compute line profiles. The correction grid is built by adjusting the sulfur abundance until the non-LTE equivalent width matches the LTE value at each grid point.","core_discovery":"With the updated sulfur model, all observed S I lines in each test star are reproduced with a single sulfur abundance, whether the line forms nearly in LTE or is strongly affected by non-LTE effects. For the two A-type stars, abundances derived from S II lines agree with those from S I lines to within 0.04 dex, which the authors take as evidence that the model is applicable up to at least 10000 K. The accompanying grid of non-LTE corrections covers effective temperatures 4000-10000 K, surface gravities $\\log g$ from 0 to 5, and metallicities $[\\mathrm{Fe/H}]$ from 0 to -2, and it replaces earlier approximate collision data with detailed quantum-mechanical rates.","pith_inferences":["Beyond the paper: the correction grid is immediately testable against large stellar surveys of metal-poor stars, where the disputed [S/Fe] plateau below [Fe/H] = -2 could be re-examined with a homogeneous non-LTE treatment.","Beyond the paper: if future laboratory measurements confirm the adjusted oscillator strengths for multiplets 6, 8, and 10, the solar calibration is independently validated; if laboratory values disagree, part of the abundance consistency may be an artifact of absorbing model error into the line parameters.","Beyond the paper: the model's structure suggests a natural next test--applying it to warmer A-type stars beyond 10000 K where S II lines dominate and checking whether S I and S II abundances continue to agree, which would probe the ionization balance and the coupling to doubly ionized sulfur.","Beyond the paper: the reported non-LTE correction depends on microturbulent velocity for strong lines, so grid users should recompute individual corrections for giants with large microturbulence rather than interpolating blindly."],"forward_implications":["Users can apply the published grid to correct LTE sulfur abundances for S I lines across the full 4000-10000 K range, including H-band lines around 1.5-2.3 microns that earlier grids did not cover.","For dwarf stars, multiplet 8 at 6743-6757 Å and multiplet 10 at 6046-6052 Å have near-zero non-LTE corrections and can be used in LTE analysis, while the 9212-9237 Å and 10455-10459 Å triplets require non-LTE corrections that reach several tenths of a dex.","The refined wavelengths and oscillator strengths for multiplets 6, 8, 10 and the infrared H-band lines provide a recommended S I line list for abundance work.","Agreement between S I and S II abundances in two 9600 K stars extends the validity of the model from cool K stars to A-type stars, so sulfur can be measured in hotter stars with the same atomic model.","For metal-poor giants, the new quantum-mechanical hydrogen collision rates give smaller departures from LTE than the old approximate-formula model, changing derived abundances by roughly 0.1 dex for lines like 9212 Å."],"supporting_citations":[{"why":"Supplies quantum-mechanical inelastic hydrogen collision rates for 40 neutral sulfur levels, replacing an older approximate treatment.","marker":"Belyaev & Voronov 2020"},{"why":"Provides detailed electron-impact excitation rates for the 17 lowest neutral sulfur levels used in the model.","marker":"Summers & O'Mullane 2011"},{"why":"Is the previous sulfur atomic model that this paper updates and to which the new non-LTE corrections are compared.","marker":"Korotin 2009"},{"why":"Earlier non-LTE calculations for neutral sulfur lines whose correction patterns are compared and whose absolute corrections differ from the new model.","marker":"Takeda et al. 2005"},{"why":"Provides the oscillator strengths that serve as the baseline for the refined gf-values of the optical multiplets.","marker":"Zatsarinny & Bartschat 2006"},{"why":"Experimental oscillator strengths for the 10455-10459 Å triplet used to validate the model against the solar spectrum.","marker":"Zerne et al. 1997"},{"why":"Gives the meteoritic sulfur abundance adopted as the solar reference for calibrating line parameters.","marker":"Lodders 2021"},{"why":"Provides the adopted solar chemical abundances used when fitting the solar spectrum and calibrating the sulfur line parameters.","marker":"Asplund et al. 2021"},{"why":"Provides the code used to solve statistical equilibrium and radiative transfer for the non-LTE populations.","marker":"Carlsson 1986"},{"why":"Supplies the model atmospheres and opacity distribution functions used in the calculations.","marker":"Castelli & Kurucz 2003"}],"fun_headline_variants":["Sulfur abundances unified across 13 A-K stars with non-LTE model","One sulfur abundance fits all S I lines in 13 test stars","Non-LTE sulfur grid extends to 10000 K and H-band lines","S I and S II abundances agree to 0.04 dex in A-type stars"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The solar-calibrated oscillator strengths and the adopted literature stellar parameters are assumed to transfer to all 13 program stars, so if the oscillator-strength adjustments absorb model errors or do not transfer, the claimed abundance consistency is partly built in by construction.","fun_headline_variants_meta":{"raw":{"variants":["Sulfur abundances unified across 13 A-K stars with non-LTE model","One sulfur abundance fits all S I lines in 13 test stars","Non-LTE sulfur grid extends to 10000 K and H-band lines","S I and S II abundances agree to 0.04 dex in A-type stars"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000362,"raw_usage":{"total_tokens":1926,"prompt_tokens":887,"completion_tokens":1039,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":503,"completion_tokens_details":{"reasoning_tokens":956}},"tokens_in":503,"tokens_out":1039,"duration_ms":9632,"temperature":1.0,"reasoning_tokens":956,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T22:42:41.670070+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a metal-poor giant with effective temperature 6000 K, surface gravity $\\log g=2$, and $[\\mathrm{Fe/H}]=-1$, the new model predicts a non-LTE correction of about -0.75 dex for the 9212 Å line, whereas the old model gives -0.85 dex and an earlier independent calculation gives -1.02 dex; a high-resolution, high-S/N spectrum of such a star that yields a sulfur abundance from 9212 Å differing from the 8694 Å or 6757 Å lines by more than the stated uncertainties would contradict the model. A laboratory measurement of the multiplet 6, 8, and 10 oscillator strengths that disagrees with the refined values by more than the quoted adjustments would likewise falsify the solar-calibration assumption.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies quantum-mechanical inelastic hydrogen collision rates for 40 neutral sulfur levels, replacing an older approximate treatment."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides detailed electron-impact excitation rates for the 17 lowest neutral sulfur levels used in the model."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Is the previous sulfur atomic model that this paper updates and to which the new non-LTE corrections are compared."},{"cited_title":"2005, PASJ, 57, 751","cited_arxiv_id":null,"evidence_quote":"Earlier non-LTE calculations for neutral sulfur lines whose correction patterns are compared and whose absolute corrections differ from the new model."},{"cited_title":"& Bartschat, K","cited_arxiv_id":null,"evidence_quote":"Provides the oscillator strengths that serve as the baseline for the refined gf-values of the optical multiplets."},{"cited_title":"1997, Phy s","cited_arxiv_id":null,"evidence_quote":"Experimental oscillator strengths for the 10455-10459 Å triplet used to validate the model against the solar spectrum."},{"cited_title":"1986, Uppsala Astronomical Observatory Repor ts, 33","cited_arxiv_id":null,"evidence_quote":"Provides the code used to solve statistical equilibrium and radiative transfer for the non-LTE populations."},{"cited_title":"& Kurucz, R","cited_arxiv_id":null,"evidence_quote":"Supplies the model atmospheres and opacity distribution functions used in the calculations."}],"review_version":1}