{"id":"e0c8d107-daba-44de-8270-4f96bd4535bc","arxiv_id":"1908.07526","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A 35-supernova spectropolarimetric sample reveals a linear pSiII-vSiII correlation and a tentative polarization dichotomy between Chandrasekhar and sub-Chandrasekhar explosion channels.","lead":"This study analyzes archival spectropolarimetry of 35 Type Ia supernovae and measures the polarized light from the silicon line in each explosion. It reports a new correlation between silicon line polarization and expansion velocity, which offers a fresh way to test explosion models.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The v–p relation in Eq. 11 may be a cadence artifact: p_max is the maximum over a 12-day window with no common-epoch correction, and Sect. 5.2 admits no common evolution pattern exists.","rationale":"The paper is an unusually complete observational study: 35 SNe, 127 epochs, uniform reduction, and the authors are transparent about the absence of a common polarization-epoch pattern in Sect. 5.2. That transparency is exactly what makes the construction of Eq. 11 vulnerable. The quantity on the left of Eq. 11 is defined as the maximum over a 12-day window; for the single-epoch SNe and other sparsely sampled objects, this maximum is only a lower bound on the true peak. If the epoch of peak polarization varies from SN to SN, as Fig. 10 indeed shows with peaks spanning about -10 to 0 d, and if sampling cadence is nonuniform, then p_max mixes objects observed near their true peak with objects observed far from it. The regression in Fig. 13 then conflates a physical correlation with a selection effect whenever sampling pattern correlates with velocity. Table 3's sub-window correlations reduce but do not eliminate the problem, because they still use the max-over-window statistic and the same heterogeneous cadences. The velocity measurements are also derived from polynomial fits with varying baselines, but that affects the independent variable and is less likely to create a spurious trend of the observed strength. The proposed test, interpolating p to a common -5 d epoch using each SN's own fitted evolution, directly targets the comparability assumption and would distinguish a physical v-p relation from a sampling artifact. If the relation survives that test, the central claim is considerably stronger; if not, the conclusion must be softened to a correlation that depends on monitoring depth. This is the same load-bearing assumption the reader identified, so the reader's conditional verdict remains appropriate without further change.","tokens_in":35813,"tokens_out":5560,"duration_ms":505276,"concrete_test":"Replace the -11 to +1 d maximum in Fig. 13 with the polarization evaluated at -5 d, obtained by fitting the same quadratic form p(t)=a*(t-tmax)^2+pmax used in Sect. 4.5 to each SN's own epochs and interpolating to -5 d. If the Pearson rho of Eq. 11 falls below about 0.5, or the slope moves outside the quoted 1-sigma range, the window-maximum definition is not a common-epoch observable and the central claim needs to be re-framed as cadence-dependent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline correlation, Eq. 11 in Sect. 5.3 and Fig. 13, regresses vSiII,-5 against p_max, defined as the maximum Si II 6355 polarization over a 12-day window (-11 to +1 d relative to B-max). Sect. 5.2 explicitly states that pSiII(t) has no common evolution pattern and therefore the authors do not correct to a common epoch. Under those conditions, p_max is not a comparable observable across objects: an SN with a single epoch in the window yields a lower bound on its true maximum, while a 12-epoch SN is observed near its peak. If cadence density or the placement of the single epoch correlates with vSiII,-5, the fitted slope 6.40e-5 per cent per km/s and rho=0.80 could arise from sampling rather than from an intrinsic kinematics-asymmetry relation. The robustness tests in Table 3 use the same maximum-within-window definition on sub-windows, so they narrow the epoch range but do not remove the bias. Because the central claim is precisely that the two quantities are linearly related, the heterogeneous-window maximum is the weakest load-bearing link.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper compiles archival VLT/FORS spectropolarimetry for 35 Type Ia supernovae at 127 epochs, measures the Si II λ6355 line polarization after wavelet-based continuum subtraction, and analyzes its time evolution, its relation to Si II velocity and Δm15, q–u loop behavior, and comparisons with DDT, double-detonation, and violent-merger model predictions. The central new claim is Eq. (11): a linear relation between the maximum Si II polarization in a −11 to +1 day window and the Si II velocity at −5 days, with slope (6.40 ± 1.28)×10⁻⁵ %/(km/s), intercept −0.484 ± 0.147, and Pearson ρ = 0.80. The paper also reports that subluminous and transitional objects lie below the Δm15–pSiII relation, finds a polarization dichotomy between objects above and below the Polin et al. double-detonation line in the Δm15–vSiII plane, documents q–u loop evolution in several well-sampled objects, and derives an upper limit on the intrinsic continuum polarization.","tokens_in":36099,"tokens_out":5734,"duration_ms":57016,"significance":"If Eq. (11) is robust, it is a new and interesting observational constraint connecting ejecta kinematics to the geometry of the Si II line-forming region, and it sharpens the comparison among SN Ia explosion models. The paper's strengths include the largest systematically reduced sample of SN Ia spectropolarimetry to date, explicit treatment of polarization bias correction and binning choices, a reproduction of the Δm15–pSiII relation with an independent and larger sample, new quantitative q–u loop area measurements, and quantitative model comparisons. The main weakness is that the headline v–p relation uses a maximum-within-window polarization without correcting for heterogeneous observing cadence, and the paper itself states that no common time-evolution pattern exists; this is load-bearing for the central claim. The two-population polarization dichotomy in Fig. 15b is also partly a corollary of Eq. (11), so it is not independent evidence for two explosion channels as currently presented.","major_comments":[{"comment":"The central claim uses p_max, the maximum Si II polarization in a −11 to +1 day window, as a comparable quantity across objects whose cadence varies from one epoch to twelve. As the authors state in Sect. 5.2 that no common time-evolution pattern exists and therefore no correction to a common epoch is applied, p_max is not a comparable observable: a single-epoch object gives a lower limit on its true peak, while a twelve-epoch object is likely observed near its peak. If cadence density or epoch placement correlates with vSiII,−5, the fitted slope and ρ = 0.80 in Eq. (11) could be partly a sampling artifact. The sub-window tests in Table 3 use the same maximum-within-window definition and therefore do not remove this bias. I request a quantitative test: either fitting per-object p(t) curves (e.g., the nested-sampling fits introduced in Sect. 4.5) and using the fitted peak, restricting to objects with at least three pre-maximum epochs, or injecting synthetic p(t) curves at the actual observed cadences to estimate the bias in the slope and correlation coefficient.","section":"Section 5.3, Eq. (11), Fig. 13, and Table 3"},{"comment":"The claimed polarization dichotomy between 'Chandrasekhar-like' and 'sub-Chandrasekhar' groups is not independent evidence for two explosion channels, because the group assignment is made in the Δm15–vSiII plane and pSiII is strongly correlated with vSiII via Eq. (11). Objects above the Polin et al. line have systematically higher velocities, and Eq. (11) therefore predicts that they have higher polarization; Fig. 15b is largely a corollary of Eq. (11). To support the two-population claim, the polarization distributions should be compared after removing the Eq. (11) trend, or on a sample matched in vSiII. Without such a test, the dichotomy does not add independent support for the two-explosion-mechanism interpretation.","section":"Section 5.3.2, Fig. 15b"},{"comment":"The velocity vSiII,−5 is derived from low-order polynomial fits to heterogeneous flux spectra, often extrapolated to −5 days when no spectrum is available near that epoch, but the reported Pearson p-values and the linear least-squares fit in Table 3 treat the velocity as error-free. The x-uncertainties are not propagated into the slope, intercept, or correlation coefficient. Please propagate the velocity fit uncertainties, for example by bootstrap resampling of the velocity fits or by using an orthogonal regression, and show that the slope and ρ in Eq. (11) remain statistically significant.","section":"Section 4.4 and Eq. (11)"},{"comment":"The Δm15–pSiII reproduction uses the maximum polarization in a −10 to 0 day window with no epoch correction, for the same reason stated in Sect. 5.2. Because the observed maximum in this window is a lower limit for sparsely sampled objects, the reported ρ = 0.63 (p = 0.005) should be presented as a qualitative confirmation of the Wang et al. relation rather than a quantitative reproduction. Please state this limitation explicitly in the text near Fig. 12 and in the conclusions, and consider a robustness test restricted to objects with at least two epochs in the window.","section":"Section 5.2, Fig. 12"}],"minor_comments":[{"comment":"The header says the slope α is 'in %/mag', but the independent variable vSiII is in km/s; the correct unit is %/(km/s) or % per km/s.","section":"Table 3 header"},{"comment":"The text refers to 'SN 20011iv' in the sentence listing the six low-reddening objects; this should be SN 2011iv.","section":"Section 5.6"},{"comment":"The text says the GG435 filter has a cut-off at ∼435 µm; this should be ∼435 nm.","section":"Section 3.1"},{"comment":"The criterion for manually choosing the lower and upper wavelength edges of the Si II line is not described quantitatively; please state how the edges were selected and whether the choice was checked for sensitivity to bin size.","section":"Section 4.5"},{"comment":"The caption would benefit from a sentence stating exactly how the orange and green diamond groups were assigned (i.e., by eye relative to the dashed Polin et al. line), since the group assignment is used in the quantitative polarization comparison in panel (b).","section":"Figure 15 caption"}],"recommendation":"major_revision","confidential_remarks":"The dataset is valuable and the analysis is largely careful, but the headline v–p relation in Eq. (11) currently rests on a maximum-within-window polarization that is not corrected for heterogeneous cadence, and the paper itself states that no common p(t) pattern exists. The authors can likely address this with a Monte Carlo cadence-bias test or a per-object fitted-peak analysis; if those tests support the relation, the paper would be a solid MNRAS contribution. I do not see evidence of citation or novelty problems; the comparison to Polin et al. and Bulla et al. is appropriate. My main concern is that the two-population interpretation in Sect. 5.3.2 is presented more strongly than the data warrant, given the coupling through Eq. (11)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear —,\n\nThis is worth a serious referee. The paper gives the largest uniform VLT/FORS spectropolarimetric sample of SNe Ia (35 objects, 127 epochs) and reports a genuinely new empirical relation: maximum Si II line polarization in the -11 to +1 day window against Si II velocity at -5 days (Eq. 11, rho=0.80). I checked the cited prior work; Maund et al. (2010b) linked polarization to velocity evolution rate, not to absolute velocity at a fixed epoch. So the v-p claim is new, and the sample and uniform reduction make it a useful reference for the subfield. The reproduction of the Delta m15-p relation with larger scatter, the subluminous/transitional offsets, the q-u loop study, and the upper limit on intrinsic continuum polarization are all solid contributions.\n\nThe soft spot is the cadence issue, and it is real. The paper's own Sect. 5.2 says there is no common polarization evolution pattern, so they do not correct to a common epoch. That makes 'maximum polarization between -11 and +1 days' a different quantity for a one-epoch SN and a twelve-epoch SN. If sparse epoch placement correlates with vSiII,-5, the slope in Eq. 11 could be biased. The robustness tests in Table 3 use the same maximum-within-window definition on narrower windows, so they reduce but do not remove the bias. I would not call it fatal: the relation persists in pre-max-only windows with a similar slope, and it oddly reappears at 10-20 days past maximum, which is not an obvious cadence artifact. But the authors should test this explicitly, e.g., by repeating the fit with only well-sampled SNe or by modelling individual p(t) curves and extracting an epoch-corrected peak.\n\nThe second issue is that the claimed two-population polarization dichotomy is not independent evidence for two explosion mechanisms. The groups are defined by position relative to the Polin et al. line in the Delta m15-v plane, and the higher-polarization group has higher velocity; given the v-p relation, higher p is expected. The authors call the dichotomy tentative, but the wording in the abstract and around Fig. 15 overstates it.\n\nOverall: the paper is solid observational work, the central relation is plausible but not airtight, and the limitations are partly acknowledged. A serious referee should engage with it and ask for cadence-robustness tests plus a clearer statement that the dichotomy is a corollary of v-p. For me, that is a revise-and-resubmit, not a reject.","headline":"Largest uniform SN Ia spectropolarimetric sample to date, with a new but not airtight velocity–polarization relation; worth serious peer review.","tokens_in":36709,"tokens_out":3600,"would_cite":true,"duration_ms":213201,"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":"The paper claims a linear relation between Si II line polarization and ejecta velocity in Type Ia supernovae.","keywords":["Type Ia supernovae","spectropolarimetry","Si II line polarization","ejecta asymmetry","delayed detonation","double detonation","q-u loops","supernova explosion models"],"falsifier":"Re-measure the relation with a homogeneous data set where every supernova is observed at the same epochs relative to B-max, for example daily from $-10$ to $+1$ days, and take the polarization at a fixed epoch such as $-5$ days. If the slope against $v_{\\mathrm{Si\\,II},-5}$ vanishes or the Pearson coefficient drops well below 0.8, the reported relation is an artifact of sparse and uneven sampling rather than a physical connection.","tokens_in":35621,"feed_emoji":"💫","tokens_out":5846,"duration_ms":53637,"temperature":0.7,"pith_summary":"The paper compiles archival VLT/FORS spectropolarimetry of 35 Type Ia supernovae, 127 epochs in total, and measures the polarization of the Si II $\\lambda 6355$ line after removing continuum polarization. It claims a statistically significant linear relation between the maximum Si II line polarization in the window $-11$ to $+1$ days and the Si II expansion velocity at $-5$ days: $p_{\\mathrm{Si\\,II}} = (6.40\\times10^{-5} \\pm 1.28\\times10^{-5})\\,v_{\\mathrm{Si\\,II},-5} - 0.484 \\pm 0.147$ percent, with Pearson $\\rho = 0.80$. If true, ejecta kinematics and geometric asymmetry are tightly connected, giving a new constraint on SN Ia explosion models. The paper also reproduces the known $\\Delta m_{15}$--$p_{\\mathrm{Si\\,II}}$ relation with larger scatter, shows subluminous and transitional objects fall below it, reports a polarization dichotomy between Chandrasekhar and sub-Chandrasekhar candidates, and finds $q$--$u$ loop evolution suggesting the silicon distribution gets clumpier with depth.","feed_headline":"Si II polarization tracks ejecta speed in Type Ia supernovae","feed_subtitle":"35 supernovae link pre-maximum line polarization to expansion velocity, constraining explosion asymmetry.","key_machinery":"The central object is the peak polarization of the Si II $\\lambda 6355$ absorption line measured on continuum-subtracted Stokes $q$ and $u$ spectra. Continuum polarization is removed by an \\`a trous wavelet decomposition of the ordinary and extraordinary beams into continuum and line scales, followed by vector subtraction; peak values are read after bias correction from spectra binned at 100 \\AA, with 25 and 50 \\AA bins used for line-complex and $q$--$u$ analyses. Velocities are measured from the absorption minimum of Si II $\\lambda 6355$ and interpolated to $-5$ days with a polynomial fit. The relation itself is the load-bearing element: it is tested across multiple epoch windows, holds only before maximum ($\\rho \\approx 0.8$ pre-peak, $\\rho \\approx 0.4$ post-peak), and is compared with synthetic polarization from delayed-detonation, double-detonation, and violent-merger simulations.","core_discovery":"On the paper's own terms, the peak linear polarization of the Si II $\\lambda 6355$ line, measured with 100 \\AA binning and corrected for polarization bias, peaks a few days before B-band maximum at levels from about 0.1 to 1.7 per cent. Across the 23 objects with at least one epoch in $-11$ to $+1$ days, the maximum polarization is linearly related to the Si II blueshift velocity at $-5$ days with slope $(6.40\\times10^{-5} \\pm 1.28\\times10^{-5})$ per cent per km s$^{-1}$, intercept $-0.484 \\pm 0.147$ per cent, and $\\rho = 0.80$; two outliers, SN 2004dt and SN 2003eh, are excluded. The interpretation is that an off-center delayed detonation naturally produces both higher silicon velocities and stronger line polarization, because the detonation front's offset controls how much silicon-rich material sits aspherically above the photosphere. Comparisons in the $\\Delta m_{15}$--$v_{\\mathrm{Si\\,II}}$ plane split the sample into a low-polarization Chandrasekhar-like cluster and higher-polarization candidates above the sub-Chandrasekhar double-detonation predictions, tentatively supporting two explosion channels.","pith_inferences":["A dense, homogeneous high-cadence campaign that measures polarization at a common epoch such as $-5$ days for every object could decide whether the reported slope is a physical relation or a sampling-cadence artifact.","If the relation holds with a common epoch, the scatter around the fitted line may carry viewing-angle information, since polarization is orientation-dependent while velocity is roughly not.","The loop-area evolution could be turned into a quantitative test: three-dimensional explosion models that predict silicon plume structure as a function of depth can be compared with the measured loop areas and their time evolution.","The apparent polarization dichotomy in the $\\Delta m_{15}$--$v_{\\mathrm{Si\\,II}}$ plane suggests that future large samples should stratify SN Ia analyses by both brightness decline and velocity, rather than by light-curve shape alone."],"forward_implications":["If the velocity--polarization relation is real, a single pre-maximum spectropolarimetric epoch gives a proxy for the ejecta's global asymmetry, not just its line-of-sight velocity.","The relation and the $\\Delta m_{15}$--$p_{\\mathrm{Si\\,II}}$ relation together point to off-center delayed-detonation geometry, where the detonation offset sets both silicon speed and chemical clumpiness.","The polarization dichotomy between objects above and below the sub-Chandrasekhar double-detonation prediction supports two distinct explosion mechanisms, with average polarizations of $0.15 \\pm 0.08$ per cent versus $0.41 \\pm 0.16$ per cent.","Evolving $q$--$u$ loops imply the silicon distribution is not simply layered; it becomes clumpier with depth, a constraint that total-flux spectroscopy cannot provide.","Observed polarization levels match delayed-detonation and double-detonation model predictions, while only SN 2004dt matches the violent-merger predictions."],"supporting_citations":[{"why":"Defines the $\\Delta m_{15}$--$p_{\\mathrm{Si\\,II}}$ relation that this paper reproduces and extends.","marker":"Wang et al. (2007)"},{"why":"Supplies the off-center delayed-detonation framework used to interpret the velocity--polarization relation.","marker":"Höflich et al. (2006)"},{"why":"Provides the sub-Chandrasekhar double-detonation predictions in the $\\Delta m_{15}$--$v_{\\mathrm{Si\\,II}}$ plane used to split the sample.","marker":"Polin et al. (2019)"},{"why":"Predicts violent-merger polarization levels, which only SN 2004dt matches.","marker":"Bulla et al. (2016a)"},{"why":"Predicts delayed-detonation and double-detonation polarization spectra used for comparison with the observations.","marker":"Bulla et al. (2016b)"},{"why":"Provides the SN 2006X spectropolarimetric measurements and interstellar-polarization treatment used as a consistency check.","marker":"Patat et al. (2009)"},{"why":"SNooPy light-curve fitting supplies $T_{\\mathrm{max}}$ and $\\Delta m_{15}$ for objects lacking literature values.","marker":"Burns et al. (2011)"},{"why":"The Open Supernova Catalog supplies the supplementary spectra and light curves used for velocity and decline-rate measurements.","marker":"Guillochon et al. (2017)"}],"fun_headline_variants":["Polarization links supernova ejecta speed to asymmetry","Type Ia line polarization correlates with silicon velocity","Si II polarization reveals two Type Ia explosion paths","Fast silicon means polarized light in Type Ia supernovae","Polarized silicon line velocity relationship separates supernova types"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the maximum Si II polarization measured anywhere in the $-11$ to $+1$ day window is a comparable quantity across the 23 supernovae, even though observing cadence ranges from a single epoch to twelve, and no correction to a common epoch is applied.","fun_headline_variants_meta":{"raw":{"variants":["Polarization links supernova ejecta speed to asymmetry","Type Ia line polarization correlates with silicon velocity","Si II polarization reveals two Type Ia explosion paths","Fast silicon means polarized light in Type Ia supernovae","Polarized silicon line velocity relationship separates supernova types"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000534,"raw_usage":{"total_tokens":2698,"prompt_tokens":1205,"completion_tokens":1493,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":821,"completion_tokens_details":{"reasoning_tokens":1417}},"tokens_in":821,"tokens_out":1493,"duration_ms":11756,"temperature":1.0,"reasoning_tokens":1417,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:04:57.662293+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-measure the relation with a homogeneous data set where every supernova is observed at the same epochs relative to B-max, for example daily from $-10$ to $+1$ days, and take the polarization at a fixed epoch such as $-5$ days. If the slope against $v_{\\mathrm{Si\\,II},-5}$ vanishes or the Pearson coefficient drops well below 0.8, the reported relation is an artifact of sparse and uneven sampling rather than a physical connection.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the $\\Delta m_{15}$--$p_{\\mathrm{Si\\,II}}$ relation that this paper reproduces and extends."},{"cited_title":"R., Wang L., Wheeler J","cited_arxiv_id":null,"evidence_quote":"Provides the SN 2006X spectropolarimetric measurements and interstellar-polarization treatment used as a consistency check."}],"review_version":1}