{"id":"40f5da7c-f2c5-44b0-91fa-185b0421f905","arxiv_id":"2502.07654","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A map-space stacking analysis of Planck PR4 maps finds a CMB polarization rotation angle of about 0.46 to 0.48 degrees, dominated by instrument calibration uncertainty and consistent with no parity-violating cosmic birefringence.","lead":"A map-based stacking analysis of the latest Planck data release measures a rotation of the cosmic microwave background polarization of about half a degree, with an uncertainty dominated by instrument calibration. The result is consistent with no cosmic birefringence, and is useful mainly as a cross-check and as a map for where foreground systematics may hide.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Foreground EB/TB contamination, evidenced by the paper's own T-vs-E and mask-dependent shifts, is not included in the error budget and may bias the central β estimates.","rationale":"The reader's weakest assumption identified foreground contamination as the load-bearing premise, and my reading agrees. The paper itself flags 'hints of a foreground systematic (north versus south hemispheres) or uncontrolled miscalibration effect (T peaks versus E peaks)' in the abstract and Section 4.2, and the mask-dependent shifts in Figs. 8–10 provide internal evidence that the stacked profiles respond to foregrounds, particularly in the T peaks. Since the isotropic miscalibration angle affects all peak classes equally, the observed T–E discrepancy must arise from another systematic, most plausibly residual Galactic TB/EB. The quoted systematic uncertainty of ±0.28° covers only polarimeter calibration, so the central values carry an unquantified foreground bias risk. The proposed foreground-only stacking test would directly measure this bias using the same public maps and pipeline, settling whether the concern lands. The 300 end-to-end simulations are not sufficient because they do not inject parity-violating foreground correlations. Thus the verdict remains CONDITIONAL as the reader set it; no change is needed.","tokens_in":19819,"tokens_out":10630,"duration_ms":96285,"concrete_test":"Using the same peak catalogs, masks, and stacking pipeline, compute the Ur profiles around T and E peaks from the Commander foreground-only Q/U maps (dust and synchrotron components). Fit a spurious β_fore for each peak class and mask. If |β_fore| is larger than the T–E offset (~0.2°) or the statistical error (0.04°), then the CMB-map estimates are biased and the central total-rotation values must be corrected or an additional foreground systematic must be added to the error budget. As a cross-check, repeat with the 353 GHz and 30 GHz Planck maps, which are dust- and synchrotron-dominated, respectively.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result—β ≈ 0.46°–0.48° as a map-space estimate of total polarization rotation—rests on the assumption that the stacked Ur profiles around T and E peaks are dominated by rotated CMB E modes, with no significant Galactic foreground contribution. The paper's own mask tests (Figs. 8–10) show that T-peak results shift by ~0.2°–0.4° between hemispheres and between dust/synchrotron masks, and the four extrema splits are mutually inconsistent at about the 1–2σ level (e.g., SEVEM All T = 0.63°±0.10° vs All E = 0.43°±0.04°). These variations cannot be caused by isotropic miscalibration (which would affect all cuts equally) and are attributed by the authors to a foreground TB/EB systematic, as stated in the abstract and Section 5. Yet the quoted systematic uncertainty (±0.28°) includes only polarimeter miscalibration, not a foreground-bias term. If the foreground-induced bias is comparable to the T–E offset (~0.2°), then the central values are not robust estimates of the total rotation, and the claim that the results are 'fairly robust against different spatial data cuts' is unsupported. The 300 NPIPE end-to-end simulations do not relieve this concern because they are generated without a parity-violating foreground component, so they cannot calibrate this bias.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper applies a map-space peak-stacking estimator to the Planck PR4 (NPIPE) SEVEM and Commander CMB maps to measure the isotropic cosmic birefringence angle β and an anisotropic (dipolar) component. Around local extrema of the temperature and E-mode maps, the authors construct radial Q_r and U_r profiles and fit the β that best matches the U_r signal produced by rotation of the CMB E and TE correlations. They report β = 0.46° ± 0.04°(stat.) ± 0.28°(syst.) for SEVEM and β = 0.48° ± 0.04° ± 0.28° for Commander. The systematic is dominated by Planck polarimeter miscalibration, so the results are consistent with zero cosmological birefringence. They also search for a dipole in β and find no significant signal. The pipeline is tested on 300 end-to-end NPIPE simulations, 100 injected-rotation simulations, white-noise and cosmic-variance-limited simulations, recovering the input 0.3° rotation in the idealised cases.","tokens_in":20054,"tokens_out":8697,"duration_ms":81996,"significance":"If the central values are taken as robust map-space estimates of the total polarization rotation in PR4 maps, they confirm earlier harmonic-space results and provide an independent cross-check in a different systematic regime. The simulation validation is a genuine strength: the injected-rotation recovery in the cosmic-variance-limited case (0.300°±0.003°) and in the white-noise case (0.302°±0.022°) demonstrates the estimator's internal consistency, and the use of realistic NPIPE end-to-end simulations is appropriate for statistical calibration. The dipole-null result is a useful addition to the anisotropic-birefringence literature. The principal limitation is that the quoted systematic budget does not include a foreground-induced parity-violating component, despite the paper's own mask and extrema-split tests suggesting such an effect, and the data error bar relies on approximations whose impact on the real data is not fully demonstrated.","major_comments":[{"comment":"The quoted systematic error, ±0.28°, is attributed entirely to polarimeter miscalibration (§4.1). However, Table 1 shows an offset of about 0.20° between the SEVEM All-T and All-E values (0.63°±0.10° vs 0.43°±0.04°), and Figs. 8–10 show T-peak β varying by roughly 0.2°–0.4° between hemispheres and between dust/synchrotron masks. These variations cannot be produced by an isotropic miscalibration, and the 300 NPIPE simulations contain no parity-violating foreground component, so they cannot calibrate the bias. The paper itself interprets the variations as hints of foreground TB/EB or uncontrolled miscalibration (§§4.2, 5). Since a foreground bias of this order would shift the central β values, the claim that the results are 'fairly robust against different spatial data cuts' is not supported unless either an explicit foreground-bias systematic term is added to the error budget or the robustness claim is restricted to E peaks.","section":"Abstract; §4.2; Table 1; Figs. 8–10"},{"comment":"The estimator assumes that the pixel noise is diagonal and equal for all pixels in a profile and ignores pixel-to-pixel and peak-to-peak correlations. The data uncertainty is then computed from Eq. (3.15), an inverse-variance weighted scatter that would underestimate the error if the peaks are correlated. The authors state that Eq. (3.15) agrees with the width of the 300 simulation histograms, but those simulations do not include parity-violating foregrounds; agreement there does not validate the data error if foregrounds add correlated variance. Because the reported statistical error is only ±0.04° and the method is used to assess 1–2σ consistency among data cuts, a direct check (e.g., jackknife over independent patches, or half-ring noise estimates) is needed to verify the error bar on the real data.","section":"§3, Eqs. (3.12)–(3.15)"}],"minor_comments":[{"comment":"The phrase 'the T Band EB correlations' should read 'the TB and EB correlations'.","section":"§2.2"},{"comment":"The word 'anisotopic' in the first sentence should be 'anisotropic'.","section":"§4.3"},{"comment":"The sentence justifying the diagonal covariance says the pixels are 'all relatively close' and therefore have similar noise levels; this wording is confusing because close pixels are more strongly correlated, not less, and the intended statement about similar noise levels should be separated from the neglect of correlations.","section":"§3, after Eq. (3.15)"},{"comment":"The caption states that the data are inconsistent with zero birefringence at more than 2σ, but the 0.28° miscalibration systematic is not shown in the figure; the caption should state explicitly that this significance is statistical only.","section":"§4.1, Fig. 7"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of JCAP and is honest about its self-identified limitations. The main issue is that the abstract and conclusions present the results as 'fairly robust' even though the mask and extrema-split tests indicate a foreground or miscalibration effect of the same order as the T–E offset; this needs to be reconciled in a revision. If the authors add an explicit foreground-bias systematic or restrict the robustness claim appropriately, the paper would be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is the first map-space peak-stacking measurement of cosmic birefringence on the Planck PR4/NPIPE maps, giving β ≈ 0.46°–0.48° for the total polarization rotation, with the 0.28° polarimeter-miscalibration systematic dominating the 0.04° statistical precision. The paper is careful, honest about its own limitations, and does not overclaim — it states plainly that the result is consistent with no parity violation once calibration uncertainty is included.\n\nWhat is actually new is modest but real: the NPIPE data release with a mature stacking method, per-peak bias weighting (the appendix shows this stabilizes mask splits), wider mask comparisons, and a null dipole constraint. The strongest part is the pipeline validation. Cosmic-variance-limited runs recover an injected 0.3° as 0.300°±0.003°, white-noise runs recover 0.302°±0.022°, and the 300 end-to-end NPIPE simulations confirm the uncertainty estimates. That is genuine evidence that the estimator works and the error treatment is not badly off.\n\nThe soft spots, in proportion. The main one is the one the authors themselves flag: the T-peak results move around by roughly 0.2°–0.4° between hemispheres and between dust or synchrotron masks, and the four extrema splits disagree at the 1–2σ level. They attribute this to foreground TB/EB or uncontrolled miscalibration, yet no foreground-bias term enters the quoted ±0.28° systematic. So the central values are best read as total-rotation estimates, not cosmological ones. The abstract says essentially this, but the phrase 'fairly robust against different spatial data cuts' overstates what the figures show, at least for T peaks. The per-peak fits also overlap heavily (each peak covers roughly 20 square degrees), so the ~10^5 peaks are far from independent; the simulation scatter covers this, but it is worth remembering. Neglecting pixel-pixel correlations and assuming CBB=0 are standard simplifications, acknowledged by the authors, and minor here. No code is released, which limits immediate reproducibility.\n\nThe central argument holds up as a total-rotation measurement. This is a cross-check paper, not a detection paper, and it does not pretend otherwise.\n\nThe audience is the cosmic birefringence subfield and anyone concerned with Planck polarization systematics. It deserves a serious referee: the pipeline is validated, the writeup is unusually transparent, and the same group produced much of the prior map-space work. I would send it to review, expecting revisions to sharpen the foreground-bias discussion rather than to fix a broken analysis.","headline":"Solid, well-caveated map-space cross-check of Planck PR4 birefringence; new PR4 numbers, honest about the calibration wall, but the foreground-bias term is missing from the error budget.","tokens_in":20664,"tokens_out":3302,"would_cite":true,"duration_ms":27697,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Map-space stacking of Planck PR4 maps finds a total CMB polarization rotation between 0.46 and 0.48 degrees.","keywords":["cosmic birefringence","CMB polarization","Planck PR4","NPIPE","map-space peak stacking","parity violation","E/B mixing","polarimeter calibration"],"falsifier":"A calibration measurement that pins the absolute polarization angle of the Planck detectors to better than about 0.1° would decide the matter: if the residual rotation after subtracting that calibration is still near 0.4°, the excess is not polarimeter miscalibration. Alternatively, a frequency-resolved stacking over the 100, 143, 217, and 353 GHz channels showing β varying with frequency beyond the miscalibration uncertainty would demonstrate foreground EB contamination.","tokens_in":19584,"feed_emoji":"🌌","tokens_out":6126,"duration_ms":58723,"temperature":0.7,"pith_summary":"This paper measures cosmic birefringence, a rotation of the CMB polarization plane caused by a hypothetical parity-violating field, directly in map space using Planck's reprocessed PR4 (NPIPE) data. Stacking polarization around temperature and E-mode extrema, the authors find a total rotation angle of 0.46° ± 0.04°(stat.) ± 0.28°(syst.) for SEVEM maps and 0.48° ± 0.04°(stat.) ± 0.28°(syst.) for Commander maps. These values agree with earlier harmonic-space estimates and, because the dominant systematic is the calibration uncertainty of Planck's polarimeters, they are also compatible with no parity violation. The authors additionally find no evidence of a birefringence dipole. The wider point is that a map-space method offers a cross-check that can expose spatially varying foreground or calibration systematics that power-spectrum analyses might hide.","feed_headline":"Map-space stacking finds 0.46-0.48 deg CMB polarization rotation","feed_subtitle":"Peak-stacking check confirms earlier Planck estimates; calibration uncertainty still allows zero rotation.","key_machinery":"The machinery is the radial/tangential Stokes decomposition around local extrema: transforming Stokes Q and U into Qr = -Q cos(2φ) - U sin(2φ) and Ur = Q sin(2φ) - U cos(2φ) turns the local E and B patterns into separate profiles. Around a temperature or E-mode peak, the expected Ur profile is proportional to sin(2β) times C_TE or sin(4β) times C_EE, so it vanishes in the absence of rotation. Per-peak bias parameters from peak theory weight each extremum according to its height, and linear least-squares fits yield a β per peak that can be averaged over the sky or split by mask.","core_discovery":"On the paper's own terms, the central discovery is that the peak-stacked Ur profiles around temperature and E-mode extrema in the Planck PR4 maps are consistent with an isotropic polarization rotation of about 0.46 to 0.48 degrees, slightly higher than previously published estimates, with the excess plausibly arising from the fact that no correction was attempted for the polarimeter miscalibration angle. The same profiles show no significant large-scale directional dependence: the fitted birefringence dipole is consistent with zero. The variations that do appear, notably between temperature and E peaks and between northern and southern hemispheres, are interpreted as hints of foreground systematic effects or an uncontrolled miscalibration, not as evidence for a cosmological signal.","pith_inferences":["Inference: a direct test of the foreground interpretation would be to measure β separately from each Planck polarization channel with the same stacking; if β drifts across 100 to 353 GHz beyond the miscalibration error, Galactic dust or synchrotron EB is contaminating the CMB-only maps.","Inference: the per-peak weighting scheme could be transported to other parity-violating probes, such as stacking on E-mode saddle points or on polarized sources, where the bias parameters behave differently and would provide an independent handle on systematics.","Inference: if the north-south asymmetry in the temperature-peak results is real and tied to the North Galactic Spur, it predicts a measurable TB spectrum in synchrotron-dominated regions at low frequencies that future experiments could directly detect."],"forward_implications":["The map-space and harmonic-space analyses of Planck PR4 data are mutually consistent, strengthening confidence that the roughly half-degree total rotation is not an artifact of a single estimator.","Because the 0.28° polarimeter miscalibration uncertainty dominates the error budget, the measurement cannot by itself distinguish cosmological birefringence from instrument rotation; a cosmological claim would require calibration at better than about 0.1°.","The systematic pattern across data cuts, with higher values in the synchrotron-rich north and differences between T and E peaks, indicates that residual foreground EB/TB or an uncontrolled miscalibration is the likeliest source of the spread.","No birefringence dipole is detected, so on the scales probed by the peak-weighted stacking there is no need for large-scale anisotropic birefringence to explain the PR4 maps."],"supporting_citations":[{"why":"Provides the earlier Planck parity-violation constraints that this paper extends to PR4 and serves as the method baseline.","marker":"[24]"},{"why":"Introduces the direction-dependent map-space birefringence analysis and the dipole question that this paper re-examines.","marker":"[23]"},{"why":"Gives the calibration-independent cosmic birefringence extraction using foreground EB that the PR4 map-space values are compared against.","marker":"[38]"},{"why":"Gives the PR4 harmonic-space birefringence estimates that the map-space results are checked against.","marker":"[39]"},{"why":"Provides the ground-based HFI polarization calibration reference that fixes the 0.28° systematic uncertainty dominating the error budget.","marker":"[25]"},{"why":"Documents the in-flight Crab-based calibration angle used for Planck's polarimeters.","marker":"[26]"},{"why":"Defines the NPIPE/PR4 maps and the end-to-end simulations used as the data and noise realizations.","marker":"[61]"},{"why":"Supplies the peak-bias formalism used to weight temperature and E-mode extrema in the stacking profiles.","marker":"[69]"},{"why":"Introduces the modified Stokes parameters Qr and Ur that turn the stacking into a local E/B measurement.","marker":"[70]"}],"fun_headline_variants":["Planck PR4 maps: 0.46–0.48° CMB polarization rotation, no dipole","Slightly higher CMB rotation angle from Planck PR4, but calibration uncertainty remains","No cosmic birefringence dipole in Planck PR4; isotropic rotation ~0.47°","Map-space Planck PR4: 0.46–0.48° CMB rotation, but systematic bias likely"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that, inside the chosen mask, the stacked Q and U maps contain no significant parity-violating foreground such as dust or synchrotron EB/TB, so any measured Ur profile is entirely rotated CMB signal with zero intrinsic B modes; if that assumption fails, the fitted β is biased.","fun_headline_variants_meta":{"raw":{"variants":["Planck PR4 maps: 0.46–0.48° CMB polarization rotation, no dipole","Slightly higher CMB rotation angle from Planck PR4, but calibration uncertainty remains","No cosmic birefringence dipole in Planck PR4; isotropic rotation ~0.47°","Map-space Planck PR4: 0.46–0.48° CMB rotation, but systematic bias likely"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000635,"raw_usage":{"total_tokens":2945,"prompt_tokens":978,"completion_tokens":1967,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":594,"completion_tokens_details":{"reasoning_tokens":1863}},"tokens_in":594,"tokens_out":1967,"duration_ms":14053,"temperature":1.0,"reasoning_tokens":1863,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T12:00:51.955147+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A calibration measurement that pins the absolute polarization angle of the Planck detectors to better than about 0.1° would decide the matter: if the residual rotation after subtracting that calibration is still near 0.4°, the excess is not polarimeter miscalibration. Alternatively, a frequency-resolved stacking over the 100, 143, 217, and 353 GHz channels showing β varying with frequency beyond the miscalibration uncertainty would demonstrate foreground EB contamination.","supporting_citations":[],"review_version":1}