{"id":"7bace8dd-2293-44b9-979a-2383dbe1c344","arxiv_id":"2607.16140","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":15,"one_line_summary":"The Comptonized component in Z-type neutron-star X-ray binaries is polarized at 3–6% (up to ~6% in the horizontal branch), exceeding spreading-layer model predictions, while disk polarization stays below ~4% with a non-perpendicular angle in several sources.","lead":"Using IXPE, NICER, and NuSTAR data, this paper measures X-ray polarization separately for the disk, the Comptonized hot plasma, and the reflected component in six Z-type neutron-star binaries, branch by branch along their color–color diagrams. The hot Comptonized component is polarized at 3–6%, higher than standard spreading-layer models predict, and the disk's polarization angle often differs from the simple perpendicular relation.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Component-resolved PD/PA in Table 2 are conditional on a fixed reflection polarization prior and frozen spectral decomposition; without a sensitivity analysis, the claimed high Comptonized PD and non-orthogonal disk PA are not established.","rationale":"The reader's weakest assumption correctly identifies the fixed reflection polarization priors and the frozen spectral decomposition as the key vulnerability. My read of §4 and Table 2 confirms that the component polarizations are conditional quantities. Since the paper's stated goal is to compare against spreading-layer predictions, the lack of a sensitivity test for the reflection PD (which the authors themselves note can be up to 20%) is the single most load-bearing gap. The reader's CONDITIONAL verdict is appropriate: the analysis is competent and the total-polarization trend is supported by the companion model-independent work, but the component-level interpretation needs a quantitative robustness check before the headline can be accepted at face value. I do not see a flaw requiring rejection; I would keep the CONDITIONAL verdict pending the authors' response. The disk misalignment sub-claim is weaker because it relies on only the free-PA rows, but this is secondary to the high-Comptonization claim.","tokens_in":28113,"tokens_out":6751,"duration_ms":56843,"concrete_test":"Re-fit the IXPE Stokes Q/U spectra for the two highest-PD branches (XTE J1701–462 HB and Cyg X-2 HB) with relxillNS PD fixed to 0%, 10%, and 20% and with PA_refl decoupled from PA_comp (including an orthogonal option), simultaneously allowing the reflection normalization and thcomp covering fraction to vary within their 90% confidence ranges. Compute the 90% confidence interval on the thcomp*bbodyrad PD in each variant. If the PD lower edge stays above ~3% in all variants, the high-PD claim is robust; if it falls below ~2% in any plausible variant, the abstract's comparison with spreading-layer models must be relaxed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline result—thcomp*bbodyrad PD ~3–6% in HB/NB, above spreading-layer predictions—depends on the component-level polarization decomposition in Table 2. In §4 the spectral model is fixed to its photometric best fit and polconst is applied per component, with relxillNS PD frozen at 10% and its PA tied to thcomp for every source and branch; in several rows the disk PA is also frozen at PA_Comp+90 (Table 2 notes). The paper itself acknowledges the Comptonization/reflection degeneracy (§4: 'difficult to estimate the PD and PA for each component'; §5: 'strong degeneracy'), but no sensitivity analysis is presented. The concern is quantitative: reflection carries 5–20% of the 2–8 keV flux, so if the true reflection PD is 20% rather than 10% (the paper cites up to 20% for some geometries), the inferred Comptonized PD moves down by roughly 1–3 percentage points for the cases with larger reflection fractions—enough to cross the ~2% spreading-layer threshold (e.g., Cyg X-2 HB). Conversely, the PD would be higher if reflection were unpolarized, so the direction of the bias is prior-dependent. The PA-misalignment claim is also only supported by the subset of rows with a free disk PA; frozen perpendicular rows cannot test it. Thus the central claim is conditional on untested priors.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents the first branch-resolved spectropolarimetric analysis of a sample of six Z-type neutron star low-mass X-ray binaries observed with IXPE, NICER, and NuSTAR. The authors fit a common spectral model — TBabs*(diskbb+thcomp*bbodyrad), plus relxillNS reflection for all sources except GX 5–1 — to each branch of the Z-track, then fix the spectral parameters and apply polconst component-by-component to the IXPE Stokes spectra. They report a Comptonized-component PD of roughly 3–6% in the HB/NB, a disk PD generally below 3%, and a disk PA that is often not perpendicular to the Comptonization PA. They interpret the high Comptonized PD as evidence that the Comptonizing region is not a simple spreading layer, and they compare the disk PD with plane-parallel atmosphere predictions. They also find no correlation between polarization and inclination or reflection fraction.","tokens_in":28621,"tokens_out":7925,"duration_ms":71251,"significance":"If the component-level polarization values are robust, this is a valuable step: it is the first uniform, branch-resolved spectropolarimetric study of Z-sources, it includes a reflection component in the polarimetric decomposition, and it provides a direct observational test of spreading-layer/boundary-layer geometries. The paper also makes useful connections to ADC systems and winds. However, the central quantitative claims rest on priors for the reflection polarization and on only a subset of rows with free disk polarization angles. The paper itself acknowledges the Comptonization/reflection degeneracy and the difficulty of estimating component PD/PA. The significance of the results therefore depends on sensitivity checks that are not currently presented.","major_comments":[{"comment":"The component-resolved PD/PA values in Table 2 are derived by fixing the spectral model and assigning polarization with polconst under three priors: reflection PD is fixed to 10%, reflection PA is tied to the Comptonization PA for every source and branch, and in several rows the disk PA is frozen to Comp+90 (Cyg X-2 HB/FB, XTE J1701–462 FB, Sco X-1 FB, GX 340+0 NB/FB). The central claim in the abstract and §4 — that the Comptonized PD is 3–6% and significantly above spreading-layer expectations of ~2% — is sensitive to these priors. Reflection contributes 5–32% of the 2–8 keV flux in the Appendix tables; e.g., Cyg X-2 HB has N_relxillNS/N_Tot = 17.8% and Table 2 gives Comp PD = 4.2±0.9. If the true reflection PD were 20% (a value the paper itself cites from Matt 1993 and Podgorný et al. 2025) with PA aligned to the Comptonization, the inferred Comptonized PD would shift downward by sever","section":"§4, Table 2, Appendix A"},{"comment":"The claim that the disk PA is 'significantly misaligned and not perpendicular' to the Comptonization PA is only testable in rows where the disk PA was left free. In several rows the disk PA is frozen to Comp+90, so those rows cannot provide evidence for misalignment. Among the free rows, some are actually consistent with perpendicular orientation within the quoted 90% errors: for example, Cyg X-2 NB has disk PA = 54±15 and Comp PA = −42±7, giving ΔPA≈96°, consistent with 90° within errors. Moreover, the reflection PA is tied to the Comptonization PA in §4 under the assumption of an axisymmetric configuration, but the paper later invokes a possible break in axial symmetry to explain the absence of an inclination trend and the non-orthogonal PAs. In a non-axisymmetric geometry, the reflection PA would not in general be locked to the Comptonization PA. The present treatment therefore preclu","section":"§4, Table 2, §5"}],"minor_comments":[{"comment":"The header 'XTE J1071–461' is a typo; the source is XTE J1701–462 throughout the text.","section":"Table 1"},{"comment":"There is a repeated word in 'also simultaneously observed the the source'; please correct.","section":"§3.5"},{"comment":"The notation '=PA Comp +90' and bracketed values such as '[10]' should be defined explicitly in the table notes as frozen priors, not as measured quantities. This will avoid confusion about which entries are constraints and which are assumptions.","section":"§4, Table 2"},{"comment":"The gray-filter correction E^{−ΔΓ} for Sco X-1 is mentioned but ΔΓ is not defined or described in terms of how it was constrained. Please clarify whether this factor is applied to the model I, Q, and U consistently and what value of ΔΓ was used.","section":"§3.4"},{"comment":"The legend states that empty markers correspond to 'values frozen during the fits,' but it is not always clear whether the PD or the PA was frozen. Please make the legend more explicit, e.g., 'PD frozen' vs 'PA frozen'.","section":"Figures 2 and 4"},{"comment":"The uncertainties in Table 2 are statistical only, since the spectral parameters are fixed to their best-fit values. The paper should state explicitly that the quoted errors do not include covariance with the spectral parameters, especially given the Comptonization/reflection degeneracy.","section":"§4"}],"recommendation":"major_revision","confidential_remarks":"The paper is potentially important and the data analysis is largely standard, but the headline component-level claims are not yet robust to the stated priors. I do not see this as a rejection: the sensitivity analyses needed are straightforward and can be performed with existing data. A revised version that adds reflection-PD sensitivity runs, frees or tests the disk-PA constraints, and restricts the PA-misalignment claim to the actually testable rows would be publishable in my view."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nQuick take: this is the first branch-resolved spectropolarimetric decomposition of a full Z-source sample with a uniform spectral model that actually includes reflection in the polarization fit. That is a genuine step forward, and the paper is honest about its assumptions. But the headline numbers — Comptonized PD ~3–6% and disk PA not perpendicular — are conditional on fixed polarization priors that the authors do not stress-test. The values are likely in the right ballpark, but the claim that they are “significantly higher than spreading-layer predictions” needs a sensitivity analysis before it should be treated as robust.\n\nWhat I like: the spectral fits are standard and statistically acceptable, with parameters consistent with prior work. The model-independent total PD from the companion paper anchors the total signal, so the qualitative branch dependence (HB > NB) is secure. The paper also situates itself well against earlier spectropolarimetric studies and explicitly states its choices — reflection PD fixed at 10%, PA tied to Comptonization — and flags the degeneracy between Comptonized and reflected components. The reference list is thorough and appropriate.\n\nWhere it is soft: the component-resolved values in Table 2 are derived by fixing the spectral model and applying polconst per component. That means the reported Comptonized PD is what is left after the assumed reflection polarization is subtracted. Since reflected photons carry 5–20% of the 2–8 keV flux, changing the reflection PD from 10% to 20% can shift the inferred Comptonized PD downward by a few points — enough to cross the ~2% spreading-layer threshold for some branches (Cyg X-2 HB looks like a case in point). Conversely, if reflection were unpolarized, the Comptonized PD would rise. The direction of the bias is prior-dependent. The disk PA misalignment claim also rests on only the subset of rows where the disk PA was left free; in several rows it was frozen perpendicular, so those rows cannot test the claim. The authors acknowledge all of this in prose, but they do not quantify how much the headline conclusions move under alternative priors. That is the main gap.\n\nThe FB results are weak, as the authors admit, and the wind talk is speculative but clearly labelled as such.\n\nBottom line: the paper deserves peer review. The analysis is careful and the dataset is valuable. The referee should ask for a sensitivity analysis (reflection PD/PA, unpolarized reflection, free disk PAs where feasible) and a clearer separation of rows that carry the misalignment claim. After that, the qualitative conclusions will likely stand, but with appropriately softened language.","headline":"Careful, useful branch-resolved spectropolarimetric analysis with a real discovery claim, but the component-level polarization values inherit fixed priors that need a sensitivity test before the headline should be taken at face value.","tokens_in":29137,"tokens_out":3475,"would_cite":true,"duration_ms":31234,"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":"For most Z-type neutron star X-ray binaries, the dominant polarized signal is Comptonized emission, reaching about 6% on the horizontal branch — higher than spreading-layer models predict.","keywords":["X-ray polarimetry","Z-sources","neutron star low-mass X-ray binaries","Comptonization","accretion disk","reflection","spectropolarimetry","IXPE"],"falsifier":"Take a bright Z-source (e.g., Cyg X-2) and leave the reflected component's polarization degree and angle free instead of fixing them at 10% and parallel to the Comptonized angle; if the best fit yields a reflection PD far below 10% or a PA that is not parallel to the Comptonized PA, the headline Comptonized PD values would not be reproduced. A second, weaker test: obtain high-signal branch-resolved data (e.g., with a more sensitive future polarimeter) and check whether the disk PA remains misaligned when it is not frozen to Comptonization+90°.","tokens_in":27958,"feed_emoji":"🔭","tokens_out":6059,"duration_ms":54348,"temperature":0.7,"pith_summary":"Using imaging X-ray polarimetry together with NICER and NuSTAR spectra, the paper builds the first branch-resolved spectropolarimetric picture of six Z-type neutron star X-ray binaries as they move along their color–color diagrams. It tries to establish that Comptonized emission is the main source of both flux and polarization for most sources and branches: the 2–8 keV polarization degree is about 6% on the horizontal branch and falls to 3–4% on the normal branch. Those numbers exceed theoretical expectations for spreading-layer or boundary-layer geometries, so the authors argue the Comptonizing region is not a simple spreading layer. The disk component is less polarized (below 3%) but still above the plane-parallel scattering-atmosphere prediction, and its polarization angle is generally not perpendicular to the Comptonized angle, suggesting a non-axisymmetric system.","feed_headline":"Comptonized X-rays polarized up to 6% in Z-sources","feed_subtitle":"Branch-by-branch spectropolarimetry shows polarization exceeds spreading-layer predictions and disk angles are misaligned.","key_machinery":"The analysis uses a single spectral baseline for all sources — a thermal accretion disk model plus thermal Comptonization of a blackbody seed, with a relativistic reflection model (and occasionally an extra hard tail) added where needed — and fits the IXPE, NICER, and NuSTAR spectra separately for each branch of the color–color diagram. Polarization is then assigned component by component with a constant-polarization multiplicative model (polconst), with the reflected component's polarization degree fixed at 10% and its angle tied to the Comptonized component. This decomposition is the load-bearing device: it turns measured total Stokes parameters into per-component polarization degrees and","core_discovery":"The central claim is that, for most of the sources and branches, the main contribution to the X-ray emission and polarization is due to Comptonization: moving from the horizontal branch to the normal branch, the polarization degree in the 2–8 keV band varies from about 6% to 3–4%, while the flaring branch is loosely constrained. These values are significantly higher than theoretical expectations for typical spreading or boundary layer configurations. The disk polarization is generally lower (below 3%) but still higher than the prediction for an electron scattering-dominated, plane-parallel atmosphere at the corresponding inclination. The polarization angle of the disk appears significantly m","pith_inferences":["If the fixed 10% reflection polarization and the parallel PA assumption are wrong — for example, if the reflection PA follows a warped disk's local normal — the quoted Comptonized polarization degrees would change; a future fit that frees the reflection PA on a bright source would settle this.","The paper's claim that the disk is more polarized than a plane-parallel scattering atmosphere could be tested independently with spectropolarimetric atmosphere models that include absorption, since the measured values are still consistent with those.","A natural extension is to model the Z-track as a continuous sequence of spreading-layer opening angles or covering fractions; if the covering fraction varies systematically along the track, the polarization evolution observed here may be reproduced without invoking an extra wind component.","The strong HB-to-NB contrast suggests that dedicated observations of currently unobserved branches in Sco X-1-like sources could decide whether the two Z-source subclasses differ in polarization or only in sampling."],"forward_implications":["If the high Comptonized polarization degrees (up to about 6–7% in the horizontal branch) are real, standard optically thick spreading-layer geometries, which predict only a few percent, are ruled out for most Z-sources and branches.","The non-orthogonal disk and Comptonized polarization angles imply broken axial symmetry in the inner accretion flow, so inclination alone is insufficient to predict the observed polarization.","Because the disk contribution rises as sources move from the horizontal to the normal branch, the branch-to-branch drop in total polarization can be explained partly by dilution from a weakly and misaligned polarized disk component.","Including the reflected component in the polarization budget matters: even at 5–20% of the photon flux, highly polarized reflected photons can shift the inferred Comptonized polarization if omitted."],"fun_headline_variants":["Z-source X-rays hit 6% polarization via Comptonization","IXPE: Z-sources show up to 6% polarized Comptonized X-rays","Disk polarization misaligned in Z-sources, IXPE reveals","Z-source X-ray polarization exceeds model predictions","Comptonization powers 6% polarization in Z-sources"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The component-level polarization results depend on the assumed spectral decomposition and on fixing the reflected component's polarization degree at 10% with its angle tied to the Comptonized angle; if the degeneracy between Comptonized and reflected emission is resolved differently, or the reflection polarization differs, the reported per-component degrees and angle misalignment would not hold.","fun_headline_variants_meta":{"raw":{"variants":["Z-source X-rays hit 6% polarization via Comptonization","IXPE: Z-sources show up to 6% polarized Comptonized X-rays","Disk polarization misaligned in Z-sources, IXPE reveals","Z-source X-ray polarization exceeds model predictions","Comptonization powers 6% polarization in Z-sources"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000257,"raw_usage":{"total_tokens":1484,"prompt_tokens":881,"completion_tokens":603,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":625,"completion_tokens_details":{"reasoning_tokens":515}},"tokens_in":625,"tokens_out":603,"duration_ms":5239,"temperature":1.0,"reasoning_tokens":515,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T21:14:07.190821+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a bright Z-source (e.g., Cyg X-2) and leave the reflected component's polarization degree and angle free instead of fixing them at 10% and parallel to the Comptonized angle; if the best fit yields a reflection PD far below 10% or a PA that is not parallel to the Comptonized PA, the headline Comptonized PD values would not be reproduced. A second, weaker test: obtain high-signal branch-resolved data (e.g., with a more sensitive future polarimeter) and check whether the disk PA remains misaligned when it is not frozen to Comptonization+90°.","supporting_citations":[],"review_version":1}