{"id":"98f857aa-a5b5-4d59-973c-7009ae1df87f","arxiv_id":"2507.06289","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":10,"one_line_summary":"The Fe K line of GX 340+0 shows a dual-peaked structure with residual narrow emission features at the ~5% level that RELXILLNS reflection modeling alone does not reproduce.","lead":"High-resolution XRISM/Resolve spectra of the neutron star X-ray binary GX 340+0 reveal a dual-peaked iron line that a standard reflection model cannot fully describe. Narrow residual features near 6.7 and 6.97 keV hint at ionized plasma or the need for upgraded reflection models.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Narrow 6.7/6.97 keV residuals are shown visually but never tested for significance against the RELXILLNS model; without a delta-stat test the central 'residuals remain' claim is not established.","rationale":"The reader's weakest assumption is that the model and calibration are accurate enough that the residuals are real features. My concern is more specific: even granting the model, the paper does not demonstrate that the residuals are statistically significant after the reflection model is applied. The visual '~5%' inset and the Table 3 Gaussian detections in a continuum-only model do not establish a residual against RELXILLNS. The APEC test is the only quantitative attempt, and it fails BIC, which undermines the plasma interpretation but also highlights that the residual significance is marginal. If a simple two-Gaussian addition to Model 1b is not significant, the central claim reduces to 'the current model leaves small, possibly insignificant residuals,' which is much weaker and would not justify the suggested need for ionized plasma or model updates. The proposed concrete test directly settles this: it is a standard ΔC/ΔBIC comparison using the same data and model, with no new assumptions. I therefore keep the reader's CONDITIONAL verdict but sharpen the condition: the paper should either provide a significant delta-stat for the narrow residuals or explicitly report them as upper limits. This is a good-faith reading of the paper, which is otherwise careful about calibration checks (pixel 30, event loss, AFe=5) and appropriately cautious about the plasma interpretation.","tokens_in":23714,"tokens_out":5766,"duration_ms":72095,"concrete_test":"Refit the joint XRISM/NuSTAR/NICER spectra with Model 1b (Table 4) plus two narrow Gaussian lines with energies fixed at 6.7 and 6.97 keV, widths free (or fixed to the Table 3 values). Compute ΔC and ΔBIC relative to Model 1b. If the two lines do not improve C by at least ~9–16 (for 4 extra free parameters, or ~9 for 2 if widths are fixed), the residuals are not statistically significant and the central claim should be softened. For robustness, repeat the test on the XRISM/Resolve data alone to confirm the features are not driven by cross-calibration. Also report the residual significance after replacing RELXILLNS with the highest-density available table (e.g., a 10^20–10^22 cm^-3 extension if usable) to test whether the 6.7 keV feature disappears with a more physical disk density.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is that, after fitting RELXILLNS, narrow emission features remain near 6.7 and 6.97 keV at the ~5% level (§3, Fig. 3 insets). This is the basis for invoking ionized plasma or future higher-density models. However, the paper never quantifies the significance of these residuals against Model 1b. The Gaussian parameters in Table 3 are measured in the continuum-plus-three-Gaussian model (Model 1a), not after the reflection model has already removed the broad line. The only statistical test with the reflection model is the APEC addition: ΔC = 50.42 for 7 dof, yet ΔBIC < 1, i.e., not preferred. Furthermore, the 6.97 keV residual persists even after APEC. Given that RELXILLNS is used at its density ceiling (log ne = 19), with spin fixed to 0 and τ fixed to 10, and that higher-density reflection tables produce stronger Fe XXV at ~6.7 keV (Ding et al. 2024), the residuals could be artifacts of the model's parameter limits rather than real spectral features. Because the paper reports only a visual assessment of these residuals, the load-bearing premise that they are real and physically meaningful is not yet established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a 152 ks XRISM/Resolve observation of the Z source GX 340+0, with simultaneous NuSTAR and NICER coverage, split into normal-branch, soft-apex, and flaring-branch spectra. The authors model the broadband continuum with two alternative thermal-Comptonization prescriptions, then add the public RELXILLNS reflection model tailored for thermal illumination of a neutron-star disk. They find that the reflection model accounts for the broad Fe K line but leaves narrow residuals near 6.7 keV and 6.97 keV at roughly the 5% level. An APEC ionized-plasma component improves the 6.7 keV residual visually, but is not preferred by BIC; the authors conclude that the remaining structure may indicate ionized plasma in the system or, alternatively, limitations in the current reflection model, and they point to upcoming higher-density reflection tables as a potential resolution.","tokens_in":23852,"tokens_out":3934,"duration_ms":44578,"significance":"If the residual features are real, this is a valuable, first high-resolution demonstration of structure in the Fe K complex of GX 340+0 in the normal-to-flaring branches, and it would motivate extensions of thermal reflection models beyond their current density ceiling. The paper has notable strengths: the multi-mission joint fitting is careful, with explicit checks of the XRISM pixel-30 gain jump, event-loss GTIs, pixel cross-talk, and cross-calibration; the modeling uses public, pre-existing codes (RELXILLNS, APEC, THCOMP), so there is no circularity in the sense of fitting residuals with a model calibrated on those residuals; and the authors report BIC comparisons and state when an added component is not statistically preferred. The APEC addition is honestly described as having weak BIC support, and a robustness check with iron abundance fixed at 5x solar is included. The central weakness, discussed below, is that the key claim of persistent narrow residuals after reflection modeling is supported only visually and is not subjected to a significance test against Model 1b.","major_comments":[{"comment":"The central claim that narrow emission features remain near 6.7 keV and 6.97 keV after applying RELXILLNS is established only by visual inspection of ratio insets (Fig. 3b and 3d). The Gaussian line parameters in Table 3 are measured for Model 1a, i.e., the continuum model plus three Gaussians, not for the residuals after subtracting the Model 1b reflection model, so they cannot be used to quantify the significance or strength of the residuals against the reflection model. The paper should add a quantitative test, for example adding narrow Gaussians at 6.7 and 6.97 keV to Model 1b and reporting the delta-C-statistic, delta-BIC, and error range on the line normalizations, or equivalent steppar-based significances. Without this, the abstract's statement that narrow features remain at the ~5% level is not statistically supported.","section":"§3, Fig. 3 insets and Table 3"},{"comment":"The 6.7 keV residual may be a consequence of using RELXILLNS at its parameter boundary rather than evidence for an additional spectral component. The disk density is fixed at the model ceiling log(ne/cm^-3)=19, spin is fixed at a=0, and Comptonization optical depth at tau=10. As the paper itself notes in §4, higher-density reflection tables produce stronger Fe XXV emission near 6.7 keV (Ding et al. 2024). The manuscript should either test an available higher-density model or explicitly discuss the degeneracy between the residual feature and the model's density ceiling; otherwise the claim that the reflection model is incomplete is not cleanly separated from the claim that the model is being used outside its valid parameter range.","section":"§3, RELXILLNS setup and §4, Ding et al. (2024)"},{"comment":"The ionized-plasma interpretation is presented in the abstract and discussion as a plausible explanation for the residuals, but the statistical evidence is weak: the APEC addition improves C-stat by 50.42 for 7 degrees of freedom yet gives delta-BIC < 1, which the paper itself describes as weak to no evidence. Furthermore, a residual near 6.97 keV remains even after adding APEC (Fig. 4b inset). The discussion then derives a plasma density from the APEC normalization, which assumes the validity of the APEC geometry and is therefore conditional on a component that is not statistically required. The conclusions should be reworded to present ionized plasma as one speculative possibility among others, rather than as a favored explanation.","section":"§3, Model 1c and §4"}],"minor_comments":[{"comment":"The phrase \"differed to a follow up analysis\" appears to be a typo for \"deferred to a follow-up analysis.\"","section":"Footnote 1"},{"comment":"The instrument tool names \"NICERL 2\" and \"NICERL 3\" should be formatted consistently as \"NICERL2\" and \"NICERL3\" (or with the intended product names) to avoid ambiguity.","section":"§2.3"},{"comment":"The sentence beginning \"Trying to model the apparent dual-peaked emission line ... as a Gaussian absorption line ... is unable to account for the observed line profile\" is grammatically awkward; consider rephrasing to \"A Gaussian absorption line ... is unable to account for the observed dual-peaked emission profile.\"","section":"§3"},{"comment":"The text \"the reflection convolution RFXCONV ... is an interpolate between two reflection codes\" should read \"is an interpolation between two reflection codes.\"","section":"§3, RFXCONV sentence"},{"comment":"The note defines the Gaussian normalization K as total photons/cm^2/s in the line, but the table does not report equivalent widths; since the text quotes equivalent widths of 15-70 eV, 10-19 eV, and 2.0-3.7 eV, it would be helpful to state explicitly in the table note that these were computed from the Table 3 parameters and the continuum model.","section":"§3, Table 3 note"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the journal's scope and the data reduction is genuinely careful. My main concern is that the headline claim of residual narrow features is not backed by a statistical test against the reflection model; this is fixable within a revision. I do not see a circularity problem: RELXILLNS is a public, pre-existing model and the residuals are not used to recalibrate it. The paper's own honest reporting of the weak BIC for APEC makes the revision straightforward."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a solid, useful observation paper. The genuinely new thing is the XRISM/Resolve view of the Fe K line in GX 340+0 across the normal to flaring branch — it clearly shows a dual-peaked line, and the reflection model RELXILLNS reproduces the broad component but leaves narrow residuals near 6.7 and 6.97 keV. The data work is careful: they check the pixel-30 energy jump, event loss, cross-talk, try two continuum descriptions, and the residual pattern survives both. That part earns its keep.\n\nThe soft spot is exactly where the stress-test note lands: the residual claim is never tested for significance against the reflection model. The Gaussian parameters in Table 3 come from the continuum-plus-three-Gaussians fit, not from after RELXILLNS is subtracted. So when they say \"narrow features remain at the ~5% level,\" it is an eyeball claim from the insets. That is load-bearing if the point is that current reflection models fail. They do try adding APEC, and it is not BIC-preferred (ΔBIC < 1) — they are honest about that — but then the discussion says the plasma \"improves constraints\" on Rin and i, which overstates a non-significant addition. The 6.97 keV residual persists even with APEC, so the interpretation is incomplete either way. The model sits at its density ceiling (log ne = 19) with spin and optical depth fixed, and AFe ~10x solar, so we cannot fully rule out that the residuals are artifacts of pushing RELXILLNS to its limits. The AFe=5 test is reassuring for Rin and i, but not for the narrow features.\n\nProportion: these are not fatal flaws. The observation is new, the residual pattern is visible and continuum-independent, and the paper flags its own uncertainties. But the central \"residuals exist\" claim deserves a proper Δ-stat or similar before it is sold as a target for the next reflection tables. A referee should ask for that.\n\nWho is this for: people working on NS reflection and XRISM line spectroscopy. It is a genuine dataset and a fair test of RELXILLNS, so a serious journal should send it out. I would engage with it, but the paper needs a revision that quantifies the residual significance and pulls back the language about the plasma \"improving constraints.\"","headline":"First high-res XRISM look at the Fe line in GX 340+0 shows a dual-peaked profile that RELXILLNS leaves with ~5% narrow residuals, but the residual significance is never quantified and the APEC interpretation is statistically weak.","tokens_in":24658,"tokens_out":3256,"would_cite":true,"duration_ms":37867,"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":"High-resolution spectra of GX 340+0 show iron-line structure that the standard neutron-star reflection model cannot fully explain, with ~5% residuals at 6.7 and 6.97 keV.","keywords":["X-ray binaries","neutron stars","accretion disks","iron K line","reflection spectroscopy","GX 340+0","XRISM","RELXILLNS"],"falsifier":"Refit the same three-mission spectra once RELXILLNS is extended to disk densities up to $10^{22}\\,\\mathrm{cm}^{-3}$ with updated atomic data (the extension described by Ding et al. 2024). If the $\\sim 5\\%$ residuals at 6.7 and 6.97 keV disappear without adding any plasma component, the paper's central claim that the standard model is incomplete would be falsified; if they persist, the claim is supported.","tokens_in":23364,"feed_emoji":"🔭","tokens_out":7864,"duration_ms":70073,"temperature":0.7,"pith_summary":"The paper tries to establish that the current public reflection model for neutron-star accretion disks, RELXILLNS, cannot fully describe the iron K line of the Z source GX 340+0 when viewed at the high energy resolution of XRISM/Resolve. Simultaneous NICER and NuSTAR spectra were included so the broad continuum and the line could be modeled consistently across the normal, soft-apex, and flaring branches. The model captures the broad relativistic line profile, but narrow emission features remain near 6.7 keV and 6.97 keV at roughly the $5\\%$ level. If those features are real, they imply additional physics in the system, such as an ionized plasma or an updated reflection model with higher disk densities.","feed_headline":"XRISM sees iron-line structure that reflection models miss","feed_subtitle":"Residuals at 6.7 and 6.97 keV hint at ionized plasma or higher-density disk models.","key_machinery":"The load-bearing tool is RELXILLNS, a flavor of the reflection code RELXILL built for thermal illumination of the accretion disk by a neutron star: it combines angle-dependent reflection spectra computed with a blackbody illuminating spectrum with the RELLINE ray-tracing code so that the curved-space-time emissivity and Doppler/gravitational broadening of the Fe line are captured self-consistently. Its disk density is capped at $\\log(n_e/\\mathrm{cm}^{-3})=19$, a ceiling the paper argues is below the expected density ($>10^{20}\\,\\mathrm{cm}^{-3}$) of the inner disk of an accreting neutron star. The other key ingredient is the energy resolution of the XRISM/Resolve microcalorimeter, which for the first time resolves structure inside the Fe line complex of this source in the normal-to-flaring branches.","core_discovery":"The discovery is that the high-resolution Fe K line of GX 340+0 in the normal-to-flaring branches is dual-peaked and more structured than moderate-resolution spectra had shown. Modeling the reprocessed line with RELXILLNS, the reflection model tailored for thermal illumination of the disk by the neutron-star surface or boundary layer, reproduces the broad components but leaves narrow emission residuals at 6.7 keV and 6.97 keV at roughly the $\\sim 5\\%$ level. Adding a single-temperature APEC ionized plasma, as motivated by the Cyg-like Z source Cygnus X-2, reduces the 6.7 keV residual but not the 6.97 keV one, and is not statistically preferred. The paper concludes that the structure likely calls for either a multi-temperature or photoionized plasma in the system or for the upcoming extension of the reflection model to higher disk densities (up to $10^{22}\\,\\mathrm{cm}^{-3}$) with updated atomic data, which predicts stronger Fe XXV emission near 6.7 keV.","pith_inferences":["A decisive test: when the higher-density RELXILLNS tables become available, the same joint fits should be rerun. If the 6.7 and 6.97 keV residuals vanish without a plasma component, the paper's suggestion that additional physics is needed would be overturned; if they persist, the ionized-plasma interpretation is strengthened.","The small pixel-30 energy-scale jump (3.5 eV at 5.9 keV) cannot by itself create residuals separated by hundreds of eV, but a dedicated calibration check excluding pixel 30 would confirm that the ~5% features are not an instrument artifact.","The same ratio-residual technique could be applied to measure how the Fe line structure evolves along the full Z track (horizontal, normal, flaring) once longer XRISM coverage exists, testing whether the residual features strengthen with accretion state."],"forward_implications":["The inner disk is slightly truncated outside the innermost stable circular orbit ($R_{\\rm in} \\sim 1.5$-$1.8\\,R_{\\rm ISCO}$) at an inclination of about 39 degrees, consistent with moderate-resolution studies of the horizontal branch.","The near-10-times-solar iron abundance inferred by the reflection fit is likely a symptom of the model's density ceiling; fixing the abundance at 5 times solar worsens the fit but does not change the radius or inclination conclusions.","The narrow residual at 6.7 keV is consistent with an ionized plasma in the system, with an inferred density around $10^{15}\\,\\mathrm{cm}^{-3}$, the order expected for an extended accretion-disk corona in a Cyg-like Z source.","A disk-seeded Comptonization continuum (THCOMP on DISKBB) is strongly preferred over a blackbody-seeded one by the Bayesian Information Criterion, and only the disk-seeded model gives radii consistent across the continuum and reflection components.","Updated reflection models with densities up to $10^{22}\\,\\mathrm{cm}^{-3}$ and new atomic data predict broader Fe lines and stronger Fe XXV emission near 6.7 keV, making GX 340+0 a target for testing those models."],"supporting_citations":[{"why":"Supplies the RELXILLNS reflection model that is the paper's central modelling tool.","marker":"García et al. 2022"},{"why":"Supplies the RELLINE ray-tracing code used within RELXILLNS for relativistic line broadening.","marker":"Dauser 2010"},{"why":"Provides the earlier moderate-resolution Fe line study of GX 340+0 whose line profile the paper compares to.","marker":"D’Aì et al. 2009"},{"why":"Provides the earlier relativistic Fe line measurement and the 6.8 keV absorption-line ansatz the paper tests.","marker":"Cackett et al. 2010"},{"why":"Reported the apparent absorption near 6.9 keV from Chandra that the high-resolution data here re-interpret as emission.","marker":"Miller et al. 2016"},{"why":"Documents the ionized plasma seen in Cygnus X-2 that motivates the APEC component added in this analysis.","marker":"Ludlam et al. 2022"},{"why":"Presents the higher-density (up to $10^{22}\\,\\mathrm{cm}^{-3}$) RELXILL extension whose predicted broader lines and Fe XXV behavior the paper invokes as the likely fix.","marker":"Ding et al. 2024"},{"why":"Supplies the APEC plasma emission code used to model the ionized-plasma component.","marker":"Smith et al. 2001"},{"why":"Establishes the expectation of disk densities above $10^{20}\\,\\mathrm{cm}^{-3}$ that motivates the criticism of the model's density ceiling.","marker":"Shakura & Sunyaev 1973"}],"fun_headline_variants":["XRISM exposes iron-line complexity in GX 340+0 that models miss","Reflection model falls short on GX 340+0's iron-line fine structure","GX 340+0 iron line hints at ionized plasma or denser disk","High-resolution iron line in GX 340+0 shows unexplained residuals","X-ray data reveal iron-line structure beyond current reflection models"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion rests on the assumption that the narrow residuals at 6.7 and 6.97 keV are intrinsic spectral features of GX 340+0 and not artifacts of the reflection model's fixed parameters (disk density pinned at the $10^{19}\\,\\mathrm{cm}^{-3}$ ceiling, spin $a=0$, Comptonization optical depth fixed at $\\tau=10$) or of the small energy-scale jump seen in one XRISM pixel.","fun_headline_variants_meta":{"raw":{"variants":["XRISM exposes iron-line complexity in GX 340+0 that models miss","Reflection model falls short on GX 340+0's iron-line fine structure","GX 340+0 iron line hints at ionized plasma or denser disk","High-resolution iron line in GX 340+0 shows unexplained residuals","X-ray data reveal iron-line structure beyond current reflection models"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000406,"raw_usage":{"total_tokens":2114,"prompt_tokens":955,"completion_tokens":1159,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":571,"completion_tokens_details":{"reasoning_tokens":1058}},"tokens_in":571,"tokens_out":1159,"duration_ms":12258,"temperature":1.0,"reasoning_tokens":1058,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:07:43.114273+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Refit the same three-mission spectra once RELXILLNS is extended to disk densities up to $10^{22}\\,\\mathrm{cm}^{-3}$ with updated atomic data (the extension described by Ding et al. 2024). If the $\\sim 5\\%$ residuals at 6.7 and 6.97 keV disappear without adding any plasma component, the paper's central claim that the standard model is incomplete would be falsified; if they persist, the claim is supported.","supporting_citations":[{"cited_title":"2010, Master’s thesis, Friedrich Alexander University of","cited_arxiv_id":null,"evidence_quote":"Supplies the RELLINE ray-tracing code used within RELXILLNS for relativistic line broadening."},{"cited_title":"M., Miller, J","cited_arxiv_id":null,"evidence_quote":"Provides the earlier relativistic Fe line measurement and the 6.8 keV absorption-line ansatz the paper tests."},{"cited_title":"M., Raymond, J., Cackett, E., Grinberg, V ., & Nowak, M","cited_arxiv_id":null,"evidence_quote":"Reported the apparent absorption near 6.9 keV from Chandra that the high-resolution data here re-interpret as emission."},{"cited_title":"M., Cackett, E","cited_arxiv_id":null,"evidence_quote":"Documents the ionized plasma seen in Cygnus X-2 that motivates the APEC component added in this analysis."},{"cited_title":"A., Kallman, T","cited_arxiv_id":null,"evidence_quote":"Presents the higher-density (up to $10^{22}\\,\\mathrm{cm}^{-3}$) RELXILL extension whose predicted broader lines and Fe XXV behavior the paper invokes as the likely fix."},{"cited_title":"K., Brickhouse, N","cited_arxiv_id":null,"evidence_quote":"Supplies the APEC plasma emission code used to model the ionized-plasma component."}],"review_version":1}