{"id":"a5b61722-1e23-40bd-99fe-b740c5486087","arxiv_id":"2507.00302","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Z-sources show decreasing X-ray polarization from the horizontal to the normal branch, then an increase in the flaring branch, with state-dependent polarization angle rotations in Sco X-1 and GX 349+2.","lead":"This paper re-analyzes IXPE X-ray polarization data for six neutron star X-ray binaries, splitting each observation by the source's spectral state along the Z-track. It finds that polarization drops in the normal branch and rises again in the flaring branch, with some sources also showing a rotation of the polarization angle.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The most fragile link is the GX 349+2 FB GTI: two of three FB intervals were selected only by IXPE flux resemblance, not by NuSTAR; if they are misassigned, the claimed PA rotation and FB increase vanish.","rationale":"The reader's weakest assumption correctly identifies branch-assignment fragility. I focus on GX 349+2 because it is the cleanest case: the paper explicitly admits that two of the three FB intervals lack independent confirmation, and GX 349+2 is one of only two sources driving the PA-rotation result. A single re-analysis restricted to the NuSTAR-confirmed FB interval can discriminate between a real branch change and a selection artifact. The marginal significance of the Sco X-1 increase (1.8σ) and Cyg X-2 (2.3σ) means the overall NB→FB trend is fragile even with correct branch assignments; however, the branch-assignment issue is more fundamental because it could reverse the conclusion for GX 349+2 entirely. The abstract's 'clearly significant' overstates the 90%-confidence-level measurements; this should be corrected in revision, but it does not by itself change the verdict from CONDITIONAL. The proposed test is computationally cheap and directly targets the least secure data selection.","tokens_in":24357,"tokens_out":8298,"duration_ms":82569,"concrete_test":"Isolate the NuSTAR-confirmed final FB interval of GX 349+2 (the one identified in Fig. 2f) and recompute the 2–8 keV PD and PA using only that interval; exclude the two early FB intervals selected by IXPE flux resemblance. If PD and PA become consistent with the NB values (PD=1.0±0.3%, PA=38°±8°), the FB increase and PA rotation for GX 349+2 are artifacts of the GTI selection. If the FB PD remains >2% and PA stays near 7°, the selection is not the cause. As a cross-check, also compute a continuous PD versus IXPE hard-color trend across the NB-FB transition to see whether the polarization changes smoothly or jumps only at the chosen GTI boundaries.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—PD decreases HB→NB and then rises NB→FB, with a PA rotation for Sco X-1 and GX 349+2—depends entirely on the GTIs that define each branch. For GX 349+2 (Sect. 5.6), only the final FB interval is confirmed by simultaneous NuSTAR; the earlier FB intervals were chosen because the IXPE flux 'resembled' that final interval. If those early intervals actually sample the NB or a different spectral state, the combined FB PD (2.1±0.6%) and PA (7°±8°) are biased, and the claimed 30° PA rotation relative to the NB (38°±8°) could be a selection artifact rather than a physical branch change. GX 349+2 is one of only two sources with a claimed PA rotation, so this is directly load-bearing. Additionally, the Sco X-1 NB vs FB PD difference (0.5±0.28) is only 1.8σ, and the PA rotation is at 90% confidence, so the word 'clearly significant' in the abstract is stronger than the measurements support. Without the GX 349+2 ad hoc FB intervals, the evidence for the FB rise and PA rotation is substantially weakened.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a homogeneous, model-independent reanalysis of all IXPE observations of six Z-sources (Cyg X-2, XTE J1701-462, GX 5-1, Sco X-1, GX 340+0, GX 349+2) using the PCUBE algorithm in ixpeobssim. The authors combine IXPE with simultaneous NuSTAR/NICER data to assign each time interval to a branch of the Z-track (HB, NB, or FB) and then measure the 2-8 keV polarization degree and angle per branch and as a function of energy. Their main observational claims are: (i) PD is highest in the HB and decreases from HB to NB; (ii) PD rises again from NB to FB in at least some sources; (iii) Sco X-1 and GX 349+2 show PA rotations between NB and FB; and (iv) PD generally increases with energy, with ~20-30 degree PA rotations in some sources. The manuscript also includes a new analysis of the second IXPE observation of Cyg X-2 and provides a useful consolidated table of branch-resolved polarimetric results.","tokens_in":24679,"tokens_out":6340,"duration_ms":66672,"significance":"If the branch-resolved trends are correct, the work would establish that Z-source polarization is not monotonic along the Z-track and that the flaring branch is re-polarized, a non-trivial constraint on corona geometry and on the mechanism producing high polarization. The systematic use of one model-independent pipeline (PCUBE) across all sources, the explicit use of IXPE, NuSTAR, and NICER state indicators, and the inclusion of previously unanalyzed data (Cyg X-2 Obs. II) are strengths. The main results are nevertheless conditional on the branch classification GTIs, several of which are not independently verified, and the key PA-rotation claims rest on marginal significances. With additional robustness tests and appropriately softened language, this would be a useful reference paper for the field.","major_comments":[{"comment":"The FB GTI for GX 349+2 is not independent of the polarimetric result it is used to derive. As stated in Sect. 5.6, only the last FB interval is identified with simultaneous NuSTAR data, while the earlier FB intervals at the beginning of the IXPE observation were selected because the IXPE flux behavior 'resembled' that final interval; Table 1 shows no simultaneous NuSTAR/NICER coverage at that early epoch. The combined FB PD (2.1±0.6%) and PA (7°±8°), and hence the claimed ~30° PA rotation relative to the NB (38°±8°), therefore depend on intervals selected by a criterion that could simply select a different flaring-like flux state. Please validate the early-FB intervals with an independent state indicator (e.g., the IXPE hardness-color track or a comparison with the known GX 349+2 Z-track morphology) and show the NB-vs-FB comparison with those intervals excluded.","section":"§5.6, Fig. 1f"},{"comment":"The HB assignment of GX 340+0 Obs. I is not verified by any simultaneous high-energy monitor: Table 1 lists no NICER or NuSTAR observation overlapping the IXPE exposure, and Sect. 4 says that the HB identification is made 'from IXPE data' alone (Fig. 3e). This HB measurement (PD=4.2±0.4%, PA=37°±3°) is then used as one of the pillars of the claimed HB-to-NB decrease. The authors should either provide an external check of the branch assignment (e.g., monitoring from Swift/BAT or AstroSat if available) or explicitly mark this result as provisional and assess how a misclassification would affect the claimed HB-NB-FB pattern.","section":"§5.5, Table 1"},{"comment":"The abstract's 'clearly significant' NB-to-FB increase for Sco X-1 is not supported by the quoted 1-sigma errors. With NB PD=0.8±0.2% and FB PD=1.3±0.2%, the difference is only 1.8 sigma, and the text itself says the PD is consistent within the errors at the 90% confidence level; the PA rotation (1°±8° to 17°±5°) is also quoted at the 90% confidence level. Please either rephrase the abstract and conclusions to match the actual significance or provide a quantitative significance statement (e.g., confidence from a proper 2D contour comparison) for the increase and rotation.","section":"§5.4 and Abstract"},{"comment":"The segmentation of the second IXPE observation into S1-S3 intervals is introduced in Sect. 4 as ad hoc time boundaries 'similarly to Zhao et al. (2024),' and the resulting polarization values are then used to infer a roughly 60° PA difference between the pre-FB and post-FB NB intervals (59°±9° vs -57°±8°). Because these intervals are not defined by a quantitative criterion, the claimed cancellation that yields the NB upper limit and the quoted PA difference may be sensitive to the chosen boundaries. Please provide the actual GTIs or a reproducible segmentation criterion, and test the sensitivity of the results to reasonable shifts of the boundaries.","section":"§5.2, XTE J1701-462 Obs. II"}],"minor_comments":[{"comment":"The source name 'XTE J1071-461' in the table header should read 'XTE J1701-462'; the same typo appears in the table body.","section":"Table 1"},{"comment":"The sentence 'The normalized Stokes parameters computed with PCUBE are shown in Fig. 2e' should refer to Fig. 4e, because Fig. 4 shows the Stokes parameters.","section":"§5.5"},{"comment":"The linear fit for GX 349+2 is drawn without reporting a p-value; please either report it or state why it is not meaningful.","section":"§5.6, Fig. 6i"},{"comment":"The paper mixes 1-sigma errors in the text, 90% errors in Fig. 7, and 99% upper limits in Table 2; please state the confidence level consistently in the table, text, and figure captions.","section":"Table 2 and Fig. 7"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope for A&A. The self-citations to the authors' corona models in Sect. 6 are not circular because the central polarimetric results are model-independent. The main risk is that several branch GTIs are partly selected using the same temporal behavior that the paper then interprets physically, especially for GX 349+2; this should be addressed with robustness tests before publication. The abstract also overstates the Sco X-1 significance relative to the 90% confidence level quoted in the text."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take on arXiv:2507.00302. It's a genuinely useful paper: the first uniform, model-independent, branch-resolved IXPE analysis across all six Z-sources, and it digs out real new numbers—FB polarization for Cyg X-2, GX 5-1 and GX 349+2, and a PA rotation for GX 349+2. The pipeline is standard (ixpeobssim PCUBE), the measurements are model-independent, and the comparison to the existing literature is honest. I believe the central trend—PD falls from HB to NB and then rises again in the FB—is real, but it is not as solid as the abstract implies.\n\nSoft spots, in order of seriousness. First, the branch assignments. For GX 349+2, two of the three FB intervals were selected solely because the IXPE flux looked like the one FB interval that NuSTAR confirmed. If those intervals are actually NB, the FB PD and the claimed ~30° PA rotation are selection artifacts. The paper says this openly, which is good, but it is load-bearing for the PA-rotation claim. Second, the significance language. The abstract says the NB-to-FB increase is 'clearly significant' for Cyg X-2 and Sco X-1; the Sco X-1 difference is 0.5±0.28 (about 1.8σ), and the PA rotation for Sco X-1 is only at 90% confidence. The abstract also puts the GX 349+2 rotation 'at the 90% confidence level', while Section 6 says 99%. Those need to be reconciled. Third, the GX 340+0 Obs I HB assignment is not simultaneous with NuSTAR/NICER and is inferred from IXPE colors alone; that's minor given the 12σ HB detection and the fact that the HB→NB decrease is already established elsewhere.\n\nThe model comparisons in Section 6 are interpretative, and the self-citations to Gnarini et al. are fine—they are earlier work, not the source of the measurements. Circularity burden is low.\n\nBottom line: this deserves a serious referee. It is a solid observational reference paper; the field will cite it. I would want the authors to fix the branch-selection description for GX 349+2, soften the abstract to match the actual confidence, and list the Sco X-1 significance correctly. I would accept it with minor/moderate revision.","headline":"Useful uniform X-ray polarization reanalysis of six Z-sources; the FB rise and PA rotation are interesting but rest on partly subjective branch assignments, so the abstract's confidence claims need toning down before publication.","tokens_in":25218,"tokens_out":3114,"would_cite":true,"duration_ms":31311,"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":"This paper argues that X-ray polarization of Z-source neutron-star binaries is non-monotonic along their Z-track: it drops from the horizontal branch to the normal branch and then rises again in the flaring branch, with angle rotations in…","keywords":["X-ray polarization","neutron star low-mass X-ray binaries","Z-sources","accretion disk corona","Comptonization","hardness-intensity diagram","IXPE","polarization angle rotation"],"falsifier":"For one source with a claimed normal-to-flaring increase (Cyg X-2 or GX 349+2), redo the time-resolved analysis with branch boundaries set by an independent method, such as a hidden Markov model on the hardness-intensity track or strictly simultaneous NuSTAR/NICER colors, and test whether the flaring-branch polarization degree still exceeds the normal-branch value at 3$\\sigma$; if the excess or the angle rotation disappears, the reported trend is an artifact of the chosen good time intervals.","tokens_in":24190,"feed_emoji":"🛰️","tokens_out":10684,"duration_ms":106721,"temperature":0.7,"pith_summary":"This paper gathers every IXPE observation of a Z-source neutron star low-mass X-ray binary made to date and re-reduces them with a single, model-independent, time-resolved polarimetry pipeline. It aims to establish how the X-ray polarization degree and angle change as each source moves along the three branches of its Z-shaped hardness-color track: horizontal, normal, and flaring. It confirms the earlier result that polarization is highest on the horizontal branch and drops on the normal branch, and it reports a new qualitative rise from the normal to the flaring branch, significant for Cyg X-2 and Sco X-1. For Sco X-1 and GX 349+2, the polarization angle also rotates between the normal and flaring branches. If correct, this means the geometry of the Comptonizing corona in these systems changes along the accretion track in a way that ordinary spectral fits do not show.","feed_headline":"Z-source polarization rebounds on the flaring branch","feed_subtitle":"Polarization falls from horizontal to normal branch, then climbs again in flaring branch; two sources rotate their angle.","key_machinery":"The load-bearing object is the Z-track itself: the three-branch path in the hardness-intensity or color-color diagram that defines a Z-source and encodes its accretion state. The authors assign every 200-second interval of each IXPE observation to the horizontal, normal, or flaring branch using simultaneous NuSTAR and NICER light curves and color-color diagrams, build good time intervals for each branch, and then compute the unweighted normalized Stokes parameters $q$ and $u$ in each branch and energy bin with a model-independent binning algorithm. Comparing polarization degree and angle across the three branches and across 1 keV energy bins is the mechanism that carries the argument; no spectral model is required for the polarization measurement.","core_discovery":"On the paper's own terms, the discovery is that the polarization of Z-sources along their Z-track is non-monotonic: after the well-known decrease from the horizontal branch (about 4% in the 2–8 keV band) to the normal branch (about 1–2%), the polarization degree turns upward again in the flaring branch. The rise is significant for Cyg X-2, from $1.6\\pm0.3\\%$ in the normal branch to $4.2\\pm1.1\\%$ in the flaring branch, and for Sco X-1 in the 3–8 keV band, from $0.8\\pm0.2\\%$ to $1.3\\pm0.2\\%$; for GX 5–1 and GX 349+2 the flaring-branch values are higher but consistent at the 90% confidence level, while for XTE J1701–462 and GX 340+0 only upper limits are available. The paper also claims a rotation of the polarization angle between the normal and flaring branches for Sco X-1 (about $15^\\circ$ at 90% confidence) and GX 349+2 (about $30^\\circ$ at 99% confidence), and a general increase of polarization degree with energy in most sources, with some sources showing a sub-$90^\\circ$ rotation of the angle with energy.","pith_inferences":["A testable consequence the paper leaves implicit: if the branch boundaries were re-derived using only IXPE's own colors (or a fully data-driven state classifier), the normal-to-flaring rise should survive; if it does not, the trend is an artifact of the branch assignment.","The ~60-degree polarization-angle difference between the two normal-branch segments of XTE J1701–462's second observation suggests that branch-averaged polarization can hide large intra-branch angle swings, so upper limits built from merged intervals may underestimate real polarization in that branch.","The wind-boosting interpretation for the flaring branch could be checked by looking for simultaneous changes in absorption lines or Doppler shifts in high-resolution spectra during flaring intervals, which the current X-ray data do not test.","The contrasting behavior of Cyg X-2 and Sco X-1, one with a polarization angle aligned with its radio jet and one misaligned, suggests that system inclination or jet direction, not just corona shape, controls the observed angle; a larger sample with known jet orientations could separate these effects."],"forward_implications":["If the rise from the normal to the flaring branch is real, polarization along the Z-track is non-monotonic, so coronal models must reproduce a minimum in the normal branch rather than a monotonic decline from the horizontal branch.","The normal-branch-to-flaring-branch angle rotations in Sco X-1 and GX 349+2 imply that the effective scattering geometry changes orientation between these branches, disfavouring a fixed axisymmetric corona.","The energy-dependent rise in polarization degree and the sub-90-degree angle rotations with energy become quantitative constraints: slab-like coronae can match the horizontal branch but overpredict the normal branch, while spherical configurations are too weakly polarizing for the horizontal and flaring branches.","For XTE J1701–462 and GX 340+0, the flaring branch remains unmeasured, so longer exposures or future missions are needed to decide whether the flaring-branch rise is universal among Z-sources.","Because all six sources were processed with the same pipeline and branch definitions, the cross-source comparison is direct and provides the baseline for the companion spectral, model-dependent analysis."],"supporting_citations":[{"why":"Defines the Z-source classification and the horizontal/normal/flaring branch structure that the analysis partitions.","marker":"Hasinger & van der Klis 1989"},{"why":"Reports the first IXPE spectropolarimetric observation of Cyg X-2, the normal-branch baseline this work reanalyzes together with the second IXPE pointing.","marker":"Farinelli et al. 2023"},{"why":"Provides the IXPE detection of XTE J1701-462 in the horizontal branch that this work confirms and extends with segment analysis.","marker":"Cocchi et al. 2023"},{"why":"Supplies the two IXPE observations of GX 5-1 covering the full Z-track, the comparison for the branch-resolved polarization.","marker":"Fabiani et al. 2024"},{"why":"Gives the IXPE Sco X-1 measurement and the gray-filter issue that motivates using the 3-8 keV band for this source.","marker":"La Monaca et al. 2024b"},{"why":"Reports the first IXPE observation of GX 340+0 in the horizontal branch, reused here for the branch comparison.","marker":"La Monaca et al. 2024a"},{"why":"Identifies the flaring-branch time intervals for GX 340+0 from simultaneous AstroSat data, which the paper adopts for its FB upper limit.","marker":"Bhargava et al. 2024b"},{"why":"Provides the three-segment division of the XTE J1701-462 second IXPE observation used to isolate normal and flaring branch intervals.","marker":"Zhao et al. 2024"},{"why":"Provides the ixpeobssim software and the unweighted PCUBE binning algorithm used for all Stokes-parameter measurements.","marker":"Baldini et al. 2022"}],"fun_headline_variants":["Polarization rebounds in Z-source flaring branch","Z-track polarization turns up again in flaring branch","Neutron star polarization rebounds in flaring branch","Polarization dip then rebound along Z-source track","Z-source polarization rises again after normal branch"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The branch labels assigned to each IXPE time interval are assumed correct, but for some sources they rest on non-simultaneous NuSTAR/NICER coverage or ad hoc IXPE flux similarities; if those labels are wrong, the measured branch-dependent polarization trend and angle rotations are spurious.","fun_headline_variants_meta":{"raw":{"variants":["Polarization rebounds in Z-source flaring branch","Z-track polarization turns up again in flaring branch","Neutron star polarization rebounds in flaring branch","Polarization dip then rebound along Z-source track","Z-source polarization rises again after normal branch"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001047,"raw_usage":{"total_tokens":4540,"prompt_tokens":1226,"completion_tokens":3314,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":842,"completion_tokens_details":{"reasoning_tokens":3240}},"tokens_in":842,"tokens_out":3314,"duration_ms":24317,"temperature":1.0,"reasoning_tokens":3240,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:19:59.832892+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For one source with a claimed normal-to-flaring increase (Cyg X-2 or GX 349+2), redo the time-resolved analysis with branch boundaries set by an independent method, such as a hidden Markov model on the hardness-intensity track or strictly simultaneous NuSTAR/NICER colors, and test whether the flaring-branch polarization degree still exceeds the normal-branch value at 3$\\sigma$; if the excess or the angle rotation disappears, the reported trend is an artifact of the chosen good time intervals.","supporting_citations":[{"cited_title":"2023, MNRAS, 519, 3681","cited_arxiv_id":null,"evidence_quote":"Reports the first IXPE spectropolarimetric observation of Cyg X-2, the normal-branch baseline this work reanalyzes together with the second IXPE pointing."},{"cited_title":"2023, A&A, 674, L10","cited_arxiv_id":null,"evidence_quote":"Provides the IXPE detection of XTE J1701-462 in the horizontal branch that this work confirms and extends with segment analysis."},{"cited_title":"2024, A&A, 684, A137","cited_arxiv_id":null,"evidence_quote":"Supplies the two IXPE observations of GX 5-1 covering the full Z-track, the comparison for the branch-resolved polarization."},{"cited_title":"Discovery of Rapid Polarization Angle Variation During the 2022 Outburst of XTE J1701-462","cited_arxiv_id":"2411.11352","evidence_quote":"Provides the three-segment division of the XTE J1701-462 second IXPE observation used to isolate normal and flaring branch intervals."}],"review_version":1}