{"id":"a25444d1-2220-4fb0-a95b-af998c9540c8","arxiv_id":"1908.05931","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"S III effective collision strengths from a 2019 B-spline R-matrix calculation are argued to be underestimated by up to a factor of two for allowed transitions, based on new distorted-wave FAC calculations.","lead":"A single astrophysicist re-computes collision strengths for the sulfur ion S III using a different code, and argues that a recent large R-matrix calculation underestimates the rates by up to a factor of two. If true, temperature and density diagnostics for gas clouds and stars that rely on this ion would need updated atomic data.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central claim rests on unvalidated FAC accuracy: if FAC's background collision strengths for S III are themselves high by a comparable factor for these allowed transitions, the factor-of-two gap would not prove Tayal et al. wrong.","rationale":"The reader's weakest_assumption correctly identifies the load-bearing point: the paper's entire case against Tayal et al. rests on the reliability of FAC to about 20% for strong allowed transitions, yet the author explicitly acknowledges his wavefunctions are simpler and that resonances are omitted. My reading of the full text confirms this is the decisive weakness. The paper offers plausible indirect evidence: the FAC values are smooth and higher, and the BSR values show suspicious humps for transitions to the highest levels, which is consistent with a coarse energy mesh. Those observations are worth taking seriously. However, the central quantitative claim—underestimation by up to a factor of two at all temperatures—requires that FAC be a trustworthy ground truth, and that the only significant source of discrepancy is partial-wave truncation in the BSR calculation. Neither is established. The paper does not publish numerical FAC data, does not report Tayal et al.'s Ω, and does not perform a convergence test on J in either calculation. Given the high cost of a full BSR run, a conditional verdict is appropriate: the concern is real and actionable, but not proven. My stress-test does not move the reader's verdict; it reinforces it. The recommended concrete test—an independent converged close-coupling calculation with a much larger Jmax—would directly settle whether the partial-wave mechanism can explain the claimed factor of two.","tokens_in":8776,"tokens_out":4713,"duration_ms":48980,"concrete_test":"Perform a converged B-spline R-matrix (or ICFT) calculation for S III using the same 18 configurations, with Jmax increased to at least 60 and a top-up from J=40, and compare the background (resonance-averaged) collision strengths for transitions 1–25, 2–26, and 5–31 with FAC at the same energies. If the close-coupling background agrees with FAC to within 20%, the paper's attribution of the gap to J≤23.5 truncation is supported; if it agrees with Tayal et al. instead, the central claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central assertion—that Tayal et al.'s effective collision strengths are underestimated by up to a factor of two, at all temperatures, for strong allowed transitions—depends on treating the FAC distorted-wave results as an accurate benchmark. The author's own final paragraph of Section 3 concedes that 'our wavefunctions are simple' and that resonances are omitted, and asserts without numerical support that these limitations 'should not affect the subsequent values of Ω and Υ by more than 20%.' No convergence test for the partial-wave sum in the FAC calculation is shown, and the FAC background is never checked against an independent close-coupling or Coulomb-Born calculation for S III. The paper also compares only Υ, not Ω, and Tayal et al.'s Ω are not reported, so the discrepancy in Υ could in part reflect differences in target energies and f-values, or in the thermal averaging of resonances, rather than the hypothesized truncation at J≤23.5. If FAC overestimates the allowed-transition background by a comparable amount, the factor-of-two gap does not demonstrate that Tayal et al. are wrong.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper assesses the S III atomic data reported by Tayal et al. (2019, ApJS 242, 9), who used MCHF for energy levels and A-values and B-spline R-matrix (BSR) for effective collision strengths. Using the Flexible Atomic Code (FAC), a distorted-wave code, the author performs an independent calculation with the same 18 configurations, compares transition energies and oscillator strengths, and then compares effective collision strengths for 18 representative transitions. The central claim is that Tayal et al.'s Upsilon values are underestimated by up to a factor of two at nearly all temperatures, especially for strong allowed transitions, and that the cause is an insufficient range of partial waves (J <= 23.5) together with an inadequate top-up procedure in the threshold region. The paper also claims that some higher-level transitions show incorrect temperature behavior due to a coarse energy mesh. It concludes by recommending either a fresh calculation or adoption of both datasets by users.","tokens_in":9022,"tokens_out":2681,"duration_ms":28928,"significance":"If the central claim holds, the published S III BSR dataset would be systematically too low for several astrophysically important allowed transitions, and the paper would strengthen the case for atomic-data producers to publish collision strengths, not only effective collision strengths. The paper also provides a useful independent FAC calculation and makes concrete comparisons for specific transitions, which is valuable for benchmarking. However, the assessment is only as strong as the accuracy of the FAC distorted-wave results; the author acknowledges that the FAC wavefunctions are simple and that resonances are omitted, and no independent validation of FAC for this ion is provided. Thus the paper is a credible cautionary assessment rather than a demonstrated correction of Tayal et al.'s data.","major_comments":[{"comment":"The central conclusion that Tayal et al.'s Upsilon values are underestimated by up to a factor of two depends on using the FAC distorted-wave results as a benchmark that is accurate to about 20%. No validation of FAC against an independent close-coupling, Coulomb-Born, or experimental benchmark for S III is shown. The author's only numerical support is the assertion in the final paragraph of Section 3 that limitations in the FAC wavefunctions 'should not affect the subsequent values of Omega and Upsilon by more than 20%', but this statement is not demonstrated and is not enough to rule out a systematic factor-of-two offset in FAC itself. Without such validation, the observed gap does not prove that Tayal et al. are wrong.","section":"Section 2, Figures 3 and 4; Section 3, final paragraph"},{"comment":"The diagnosis that J <= 23.5 is insufficient for convergence is inferred rather than demonstrated. Tayal et al. did not publish Omega values, so the comparison is made only at the level of Upsilon. Differences in Upsilon can arise from the ~10% differences in transition energies and oscillator strengths shown in Table 1, from different resonance treatments and thermal averaging, or from differences in the top-up implementation, in addition to partial-wave truncation. The claim that 'inclusion of insufficient number of partial waves is the reason' is load-bearing but remains a speculation. The author should either obtain Tayal et al.'s Omega data, or demonstrate the partial-wave convergence issue explicitly, e.g., by showing how FAC Omega for these transitions changes as the maximum J is increased.","section":"Section 2, section on partial waves; Section 3, first paragraph"},{"comment":"The assertions about incorrect behavior of Upsilon for higher-level transitions are supported only by visual inspection and by analogy to previous work on other ions. The 'humps' in Tayal et al.'s Upsilon curves are attributed to a coarse energy mesh, but no sensitivity test of this mesh is performed here, and the alternative explanation of pseudo-resonances is not analyzed for these specific transitions. Similarly, the claim that a rise by more than a factor of 50 between 10^3 and 10^6 K is 'neither noted nor expected' lacks a quantitative justification; some optically allowed transitions with high excitation energies can legitimately show steep temperature dependence. This part of the assessment needs a concrete test or a more cautious formulation.","section":"Section 2, Figures 11 and 12; Section 3"}],"minor_comments":[{"comment":"The keyword 'electon impact excitation' contains a typo; it should read 'electron impact excitation'.","section":"Abstract and keywords"},{"comment":"The formula used to compute Upsilon from the FAC Omega values is not given. Including the standard temperature-averaging expression would allow readers to reproduce the calculation and check the energy-grid treatment.","section":"Section 2, after Figure 1"},{"comment":"The table format is hard to read because the notation 'a±b ≡ a×10±b' is not applied consistently across all entries, and some f-values appear without the exponent convention. A cleaner rendering would avoid ambiguity.","section":"Table 1"},{"comment":"The phrase 'at all temperatures' is used in the abstract and several places, but the author later notes that 'for a few the differences decrease towards the higher end of the temperature range'. The wording should be adjusted to reflect this qualification.","section":"Section 2, text near Figures 3 and 4"},{"comment":"The statement that 'the two sets of data can also not be (easily) combined because the level orderings are incompatible' is not explained. Since the paper already uses the same transition indices, it would be useful to state explicitly why combining is impractical.","section":"Section 3, final paragraph"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a single-author assessment that relies heavily on comparisons with the author's own prior publications for validation of FAC. The central claim is technically plausible, but the lack of an independent benchmark for the FAC results on S III makes the conclusion stronger than the evidence. I would advise the editor that the paper could become acceptable if the author either provides a direct validation of FAC accuracy for this ion (e.g., a convergence study or comparison with another independent method) or substantially softens the claimed diagnosis of partial-wave insufficiency."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: Aggarwal has run his own FAC distorted-wave calculation for S III and claims Tayal et al.'s BSR effective collision strengths are underestimated by up to a factor of two for strong allowed transitions, at all temperatures. That is the headline. The paper does something genuinely useful: it compares 18 representative transitions, focuses on the strongest electric-dipole lines where DW should be most reliable, and makes a concrete, testable complaint about the energy mesh—0.0001 Ryd below thresholds, 0.2 Ryd above. The humps and steep rises in Tayal et al.'s higher-level transitions look real, and that criticism stands independently of the factor-of-two claim. The paper also makes a fair procedural point: without Omega, outside users cannot audit the BSR results.\n\nThe soft spots, in proportion: the central claim is not demonstrated to the claimed precision. The FAC results are never checked against an independent close-coupling or Coulomb-Born calculation for S III, and no partial-wave convergence test for FAC is shown. The author admits his wavefunctions are simple and that resonances are omitted, and then asserts—without numerical support—that this should not affect Omega and Upsilon by more than 20%. That 20% tolerance is doing a lot of work. MCHF and FAC energies differ by up to about 10%, and for weak transitions the f-values can differ by orders of magnitude; for the strong allowed transitions they agree within ~20%, so the comparison is not unreasonable, but it does not justify a factor-of-two accusation unless the FAC background is itself accurate to better than that. Since Tayal et al.'s Omega are unavailable, part of the discrepancy could be in the resonance averaging rather than the J truncation. Also, the paper leans heavily on the author's own prior validation papers; that's not disqualifying, but the reader should weigh it.\n\nWho is this for? Anyone who uses S III line ratios in H II regions, planetaries, or stellar atmospheres, and anyone maintaining atomic databases. It is not the last word—it is a warning notice with a plausible mechanism. A serious referee would send it out, ask for the numerical data, a partial-wave convergence check, and a direct comparison with an independent R-matrix or Coulomb-Born calculation for at least a few allowed transitions. As is, I would not cite it as a correction, but I would want it on the table when deciding which S III dataset to trust.","headline":"A credible red flag against Tayal et al.'s S III collision strengths, but the case rests on an unvalidated FAC benchmark, so it lands as 'worth a serious look' rather than a settled correction.","tokens_in":9490,"tokens_out":1797,"would_cite":false,"duration_ms":20521,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["34.80.Dp"],"model":"deepseek-v4-flash","headline":"The paper argues that the recently published effective collision strengths for S III are systematically underestimated by up to a factor of two, mainly because the close-coupling calculation included too few partial waves and an…","keywords":["electron impact excitation","S III","effective collision strengths","collision strengths","B-spline R-matrix","distorted-wave approximation","atomic data assessment","partial wave convergence"],"falsifier":"Recompute the same 198-level close-coupling model with the partial-wave expansion extended well beyond $J = 23.5$, a top-up applied consistently inside the threshold region, and a finer energy mesh above thresholds, then compare $\\Upsilon$ for the six strongest transitions (1-25, 2-26, 3-27, 4-29, 4-30, and 5-31 in the paper's numbering) with both datasets. If the revised values stay near the published ones instead of rising toward the distorted-wave values, the paper's diagnosis is wrong; a direct experimental measurement of the excitation cross section for one strong line would also settle which calculation is closer.","tokens_in":8552,"feed_emoji":"⚛️","tokens_out":13512,"duration_ms":114262,"temperature":0.7,"pith_summary":"This paper assesses a recently published large-scale set of effective collision strengths for the astrophysically important ion S III, produced with a B-spline R-matrix close-coupling calculation. Using an independent distorted-wave calculation, it argues that the published values are systematically underestimated, by up to a factor of two and at almost all temperatures, for many strong allowed transitions and some forbidden ones. The cause, the paper argues, is an insufficient range of partial waves (quantum angular-momentum contributions up to $J \\leq 23.5$) together with a top-up correction that does not fix the threshold region, leaving the underlying collision strengths unconverged exactly where they matter most. It also claims that the temperature behaviour of $\\Upsilon$ for transitions involving high-lying levels is wrong because of a coarse energy mesh above threshold. If correct, plasma models that adopt the published dataset would underpredict emission in strong S III lines and could bias density or temperature diagnostics.","feed_headline":"Published S III collision strengths are up to 2x too low","feed_subtitle":"Independent distorted-wave check finds published S III collision strengths low at all temperatures.","key_machinery":"The load-bearing comparison is between the distorted-wave calculation and the B-spline R-matrix dataset for strong electric-dipole transitions, where the collision strength behaves as $\\Omega_{ij} \\sim 4 \\omega_i (f / \\Delta E_{ij}) \\ln E$; because both calculations produce comparable transition energies and oscillator strengths, the factor-of-two gap in the averaged quantity $\\Upsilon$ is attributed to the R-matrix calculation's partial-wave cutoff $J \\leq 23.5$ and to its top-up procedure, defined as the correction for neglected higher angular momenta. The paper argues that performing the top-up only above threshold leaves the threshold-region values unconverged, so the underestimate appears at all temperatures. A secondary mechanism is the energy mesh: a fine $0.0001$ Ryd mesh below thresholds and a coarse $0.2$ Ryd mesh above them, which the paper uses to explain the humps and overly steep rises in $\\Upsilon$ for transitions involving high-lying levels.","core_discovery":"The paper's central claim is that the published effective collision strengths for S III are not accurate at the claimed level of about twenty percent: for a representative set of eighteen transitions, the published values fall below an independent calculation by up to a factor of two, at nearly all temperatures, and most clearly for the strong electric-dipole transitions, where resonances are not a complicating factor. Since the two calculations agree on the transition energies and oscillator strengths, the paper locates the discrepancy inside the collision calculation: a partial-wave expansion cut off at $J \\leq 23.5$, with a top-up procedure applied only above threshold, leaves the background collision strength underestimated in the threshold region, and this propagates to all temperatures. For transitions involving the highest levels, the paper reports that the published $\\Upsilon$ curves show humps and steep rises by factors of fifty or more, and attributes this to the coarse $0.2$ Ryd energy mesh used above thresholds rather than to resonances. The paper recommends fresh or improved calculations, and in the meantime advises users to compare both datasets because the level orderings of the two calculations are incompatible.","pith_inferences":["This suggests a general cross-check: for any large R-matrix dataset that reports effective collision strengths but not the underlying collision strengths, a background distorted-wave calculation restricted to strong allowed transitions can flag possible partial-wave-convergence problems, because those transitions depend mainly on oscillator strength and transition energy.","A sharper causal version of the paper's claim is testable by re-running the close-coupling calculation with more partial waves and a finer mesh; if the strong transitions fail to rise toward the independent values, the underestimate would have to come from the wavefunctions or the close-coupling expansion rather than from the partial-wave cutoff.","For forbidden transitions the two datasets are not symmetric alternatives: the independent calculation omits resonances, so for those transitions the true effective collision strengths could be higher than both published values, and neither set alone would bracket the correct answer."],"forward_implications":["If the published effective collision strengths are low by up to a factor of two, modeled intensities of strong S III emission lines in H II regions, planetary atmospheres, and stellar plasmas would be correspondingly underpredicted.","Temperature and density diagnostics built from S III line ratios involving the affected transitions could be biased, because the underestimate occurs at all temperatures rather than in one regime.","The identified mechanism implies that future close-coupling calculations for S III must include a substantially larger partial-wave range and a top-up that is valid in the threshold region before their $\\Upsilon$ values can be trusted.","The reported humps and steep rises for high-level transitions imply that a $0.2$ Ryd energy mesh above thresholds is too coarse to produce reliable effective collision strengths for those levels.","Because the two datasets disagree and cannot be trivially combined due to incompatible level orderings, users currently cannot simply average them; fresh or improved calculations are needed."],"supporting_citations":[{"why":"the 198-level MCHF/BSR dataset under assessment; supplies the energies, A-values, and effective collision strengths being compared.","marker":"[1]"},{"why":"describes the Flexible Atomic Code used here to generate the independent distorted-wave collision strengths.","marker":"[8]"},{"why":"shows the background collision strengths from this code compare well with R-matrix results for another ion, supporting its use as a reference.","marker":"[9]"},{"why":"example where including larger ranges of partial waves lifts the whole background collision strength upward.","marker":"[10]"},{"why":"demonstrates the importance of including more partial waves than the assessed calculation appears to have used.","marker":"[11]"},{"why":"shows that an early top-up underestimates collision strengths, the cited mechanism for the S III discrepancy.","marker":"[12]"},{"why":"illustrates slow convergence with partial waves for the type of forbidden transition where the gap appears.","marker":"[13]"},{"why":"attributes the humps in high-level transitions to a coarse energy mesh above thresholds, the explanation adopted here.","marker":"[14]"}],"fun_headline_variants":["Published S III collision strengths are up to 2x too low","S III collision strengths underestimated by up to 2x","New assessment: S III collision strengths up to 2x too low","S III collision strengths: prior results up to 2x too low","S III electron impact data: up to 2x too low"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the simpler, resonance-free calculation used for comparison is itself accurate to within about twenty percent for the strong allowed transitions; if that baseline is off by a comparable amount, the factor-of-two gap would not prove the published values are underestimated.","fun_headline_variants_meta":{"raw":{"variants":["Published S III collision strengths are up to 2x too low","S III collision strengths underestimated by up to 2x","New assessment: S III collision strengths up to 2x too low","S III collision strengths: prior results up to 2x too low","S III electron impact data: up to 2x too low"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001171,"raw_usage":{"total_tokens":4843,"prompt_tokens":948,"completion_tokens":3895,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":564,"completion_tokens_details":{"reasoning_tokens":3804}},"tokens_in":564,"tokens_out":3895,"duration_ms":23878,"temperature":1.0,"reasoning_tokens":3804,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:59:38.504901+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the same 198-level close-coupling model with the partial-wave expansion extended well beyond $J = 23.5$, a top-up applied consistently inside the threshold region, and a finer energy mesh above thresholds, then compare $\\Upsilon$ for the six strongest transitions (1-25, 2-26, 3-27, 4-29, 4-30, and 5-31 in the paper's numbering) with both datasets. If the revised values stay near the published ones instead of rising toward the distorted-wave values, the paper's diagnosis is wrong; a direct experimental measurement of the excitation cross section for one strong line would also settle which calculation is closer.","supporting_citations":[{"cited_title":"S., Zatsarinny, O., & Sossah, A","cited_arxiv_id":null,"evidence_quote":"the 198-level MCHF/BSR dataset under assessment; supplies the energies, A-values, and effective collision strengths being compared."},{"cited_title":"M., & Keenan, F","cited_arxiv_id":null,"evidence_quote":"shows the background collision strengths from this code compare well with R-matrix results for another ion, supporting its use as a reference."},{"cited_title":"M., Keenan, F","cited_arxiv_id":null,"evidence_quote":"example where including larger ranges of partial waves lifts the whole background collision strength upward."},{"cited_title":"M., & Keenan, F","cited_arxiv_id":null,"evidence_quote":"demonstrates the importance of including more partial waves than the assessed calculation appears to have used."},{"cited_title":"M., & Keenan, F","cited_arxiv_id":null,"evidence_quote":"shows that an early top-up underestimates collision strengths, the cited mechanism for the S III discrepancy."},{"cited_title":"M., & Keenan, F","cited_arxiv_id":null,"evidence_quote":"illustrates slow convergence with partial waves for the type of forbidden transition where the gap appears."},{"cited_title":"2017, Phys","cited_arxiv_id":null,"evidence_quote":"attributes the humps in high-level transitions to a coarse energy mesh above thresholds, the explanation adopted here."}],"review_version":1}