{"id":"a3ca6d9f-9154-49ca-a329-9d79168007c6","arxiv_id":"1908.04969","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"Reflection spectra from thick (Polish donut) accretion disks, when analyzed with thin-disk models, bias black hole spin upward and can create or mask apparent deviations from the Kerr metric.","lead":"By simulating X-ray reflection from thick accretion disks and fitting it with the standard thin-disk model, the authors find that black hole spin is systematically overestimated. The results warn that some published spin measurements and tests of Einstein's gravity from high-accretion-rate sources may carry hidden biases.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Spin-overestimate conclusion is drawn from a single (q=9, Rout=40M, i=35) corner of the Polish-donut parameter space; the paper's own line-shape figures show this corner is not representative.","rationale":"The reader's weakest_assumption targets fidelity of the Polish-donut model to real thick disks; that is real but acknowledged and not fatal. My concern is narrower and internal: even within the Polish-donut family, the fit study samples one point—q=9, Rout=40M, i=35—chosen in part to compensate for the model's artificial outer-radius limitation. The paper's own Figures 3 and 4 demonstrate strong sensitivity to q and i, including disappearance of the redshifted tail at i=60. Since the spin bias is driven by that tail, the 'always' in Section 5 is not supported. The core demonstration that a plausible thick-disk model can produce acceptable thin-disk fits with spurious spin and alpha13 is valid, and the ray-tracing pipeline is independently checked against relline to better than 0.1% (Figure 1). The paper is appropriately cautious overall. I therefore recommend keeping the acceptance but conditioning the strong spin-overestimate claim on the simulated parameter regime, pending a q=3 check. This is a scope restriction, not a rejection of the paper's main contribution.","tokens_in":15327,"tokens_out":9494,"duration_ms":108300,"concrete_test":"Repeat simulations A–D with the same Polish-donut geometry, NICER setup, and relxill_nk fitting pipeline, but replace the emissivity index q=9 with q=3 while keeping all other inputs fixed. If the recovered a* is not systematically above the input value, the 'always overestimated' conclusion must be restricted to steep-emissivity configurations; if the bias persists, the concern is settled against. An additional run at i=60 with q=9 would isolate the role of inner-disk obscuration.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The existence proof—thick-disk spectra can mimic thin-disk fits with reduced chi2 near 1 and biased parameters—is solid for the configurations simulated. The load-bearing weakness is the generalization from these configurations to the statement (Section 5, item iii) that the spin parameter is 'always and significantly overestimated' when thick Polish-donut reflection spectra are fitted with thin-disk models. All fits in Tables 1 and 2 use emissivity index q=9, outer radius Rout=40M, and inclination i=35. The q=9 choice is not neutral: Section 3 says it is adopted specifically to suppress outer-disk contributions because larger Polish donuts obscure the inner region, and the outer radius is 'unnaturally low for real accretion disks.' Figures 3 and 4 show q=3 line shapes are 'remarkably different' and that at i=60 the strongly redshifted low-energy tail—the very feature that drives the spin overestimate—is missing. So with q=3 or at high inclination, the inferred spin bias could be reduced, reversed, or replaced by parameter degeneracies. The paper's hedged conclusion ('can be significantly affected') survives, but the stronger quantitative claim is conditional on an unrepresentative slice of the model space.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies relativistic reflection features from geometrically thick accretion disks modeled with the Polish donut model. The authors implement a ray-tracing convolution for thick disks, compute iron line shapes for a range of spins, inclinations, and emissivity indices, and then simulate NICER observations of full reflection spectra in Kerr and Johannsen spacetimes. The simulated spectra are fitted with the thin-disk relativistic reflection model relxill_nk. The paper reports that thick-disk reflection spectra can be well fitted by thin-disk models with reduced chi-squared near unity, while the recovered spin and deformation parameter alpha13 are often biased. The authors conclude that reflection spectroscopy of sources accreting at or above the Eddington rate may carry significant systematic uncertainties, especially for tests of the Kerr hypothesis.","tokens_in":15573,"tokens_out":10032,"duration_ms":103425,"significance":"If the result holds, this is an important cautionary result for X-ray reflection spectroscopy. The paper benefits from a clearly described ray-tracing implementation that is validated against relline in the thin-disk limit to better than 0.1% (Fig. 1), realistic NICER response simulations, and transparent reporting of best-fit parameters and reduced chi-squared values in Tables 1 and 2. The existence proof that a thick-disk spectrum can masquerade as a thin-disk spectrum with a biased spin and a non-vanishing deformation parameter is a valuable contribution with direct implications for spin measurements of super-Eddington sources and for tests of the Kerr metric. The authors also openly acknowledge several limitations of the Polish donut model, including its idealized emissivity profile and the unnaturally small outer radii required to keep the inner disk visible, which helps the reader judge the scope of the conclusions.","major_comments":[{"comment":"The claim that 'the spin parameter a* is always and significantly overestimated' is not supported by the parameter coverage presented. All six simulations use the same emissivity index q = 9, the same outer radius Rout = 40 M, and the same inclination i = 35 deg, so the result is established only for one corner of the Polish-donut parameter space. This is not merely a cosmetic point: Section 3 and Figs. 3-4 show that the strongly redshifted low-energy tail, which is the spectral feature that drives the spin overestimate, is absent at i = 60 deg and is strongly modified when q = 3 because of the larger contribution from the outer disk. I ask that the statement be either explicitly restricted to the simulated configurations or tested against additional configurations with varying q, i, and Rout.","section":"Section 5, item (iii); Tables 1 and 2"},{"comment":"The magnitude of the reported bias may depend on the adopted emissivity law and disk size. The emissivity index q = 9 is chosen to suppress the contributions of the outer part of the disk, and Section 3 states that Rout = 40 M is 'unnaturally low for real accretion disks' and that larger Polish donuts obscure the inner region. Realistic thick disks, for example those from numerical simulations, may have different emissivity profiles and radial extents, so the quantitative spin and alpha13 biases presented in Tables 1 and 2 could change. The paper's final conclusion ('can be significantly affected') is appropriately hedged, but the quantitative bullet in item (iii) should be reconciled with this limitation or the analysis extended.","section":"Section 4, simulation setup; Section 3"}],"minor_comments":[{"comment":"The caption contains the typo 'Comparision' and the disk model is spelled inconsistently as 'polish donut' in several places; please use 'Polish donut' consistently.","section":"Fig. 5 caption and text"},{"comment":"The metric component in the footnote and the line element should be typeset as g_{tr} rather than 'gt r', and the notation for the metric components could be made explicit.","section":"Section 2, Eq. (1)"},{"comment":"The four curves in Fig. 5 are distinguished by line style, but two of them are blue; using distinct colors for all four curves would improve readability.","section":"Section 4.1, Fig. 5"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe thing to know about this paper is simple: if you fit reflection spectra from a geometrically thick (Polish donut) disk with the standard thin-disk model, you can get a statistically acceptable fit and systematically wrong parameters. The paper demonstrates this with full NICER-simulated spectra, not just iron lines. The direction of the spin bias makes physical sense: the thick disk's inner edge sits inside the ISCO, producing strongly redshifted photons that mimic a smaller ISCO and therefore a higher spin. The same mechanism can either create or mask apparent non-Kerr deviations, depending on the input spin.\n\nWhat is actually new is the full reflection calculation for Polish donut disks and the explicit fitting exercise against a thin-disk model. The ray-tracing is validated against relline to better than 0.1% for thin disks, which gives confidence in the numerical machinery. The simulations are honest and carefully executed, and the MYTorus comparison is a sensible check on one of the model assumptions.\n\nThe soft spot is not in the execution but in the scope of the conclusion. Every fit in Tables 1 and 2 uses q=9, Rout=40M, and i=35°. These are not neutral choices. q=9 is adopted specifically to suppress the outer-disk contribution, and Rout=40M is, in the authors' own words, 'unnaturally low for real accretion disks' because larger donuts obscure the inner region. The paper's own line-shape figures show that at i=60°, the strongly redshifted low-energy tail—the very feature that drives the spin overestimate—is missing. So the Section 5 claim that the spin is 'always and significantly overestimated' is stronger than the evidence. The hedged version, that thick-disk spectra can significantly bias spin and Kerr-deformation measurements, is well supported and is the claim that matters for the community.\n\nThis is not a fatal flaw. The paper is transparent about its model limitations, and the existence proof is solid: for at least this corner of parameter space, thin-disk models absorb thick-disk physics into spin and α13. That is a useful warning for anyone applying reflection spectroscopy to super-Eddington AGN. I would send it to a serious referee. The referee should ask for the 'always' to be softened or for additional simulations at other inclinations and emissivity indices.","headline":"Thick-disk reflection spectra can masquerade as thin-disk spectra with biased spin and Kerr-deformation measurements, but the paper's 'always' spin-overestimate claim outruns its single-corner simulations.","tokens_in":16111,"tokens_out":3825,"would_cite":true,"duration_ms":37018,"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":"Thick accretion disks analyzed with thin-disk reflection models yield statistically good fits but systematically overestimate black hole spin and can produce spurious apparent deviations from the Kerr metric.","keywords":["X-ray reflection spectroscopy","thick accretion disks","Polish donut model","black hole spin","Kerr hypothesis","accretion disk geometry","relativistic reflection","iron line"],"falsifier":"Take a super-Eddington AGN whose black hole spin is known from an independent method (such as disk continuum fitting) and analyze its reflection spectrum with a thin-disk model; if the recovered spin matches the independent measurement, the claimed systematic overestimation would be falsified.","tokens_in":15141,"feed_emoji":"🌀","tokens_out":7704,"duration_ms":71668,"temperature":0.7,"pith_summary":"This paper asks what happens when X-ray reflection spectra of geometrically thick accretion disks are analyzed with the standard thin-disk reflection models. Using the Polish donut model to describe a thick, optically thick disk with its inner edge inside the ISCO, the authors simulate reflection spectra and fit them with a thin-disk model. They find that the fits look statistically acceptable (reduced chi-squared near unity) while recovering a significantly overestimated black hole spin parameter a*, and sometimes a spurious non-zero deformation parameter α13 that mimics a deviation from the Kerr metric. The takeaway is that spin measurements and tests of general relativity from reflection spectroscopy may be systematically biased for sources accreting at or above the Eddington rate.","feed_headline":"Thick disks can fake high spins in reflection fits","feed_subtitle":"Polish-donut spectra fitted with thin-disk models yield good fits but biased spin and fake non-Kerr signals.","key_machinery":"The central object is the Polish donut model: an analytic, stationary, axisymmetric, geometrically thick and optically thick disk of perfect fluid with constant specific angular momentum, whose surface has a cusp and whose inner edge lies inside the ISCO. The argument is carried by ray tracing null geodesics backwards from a distant observer's image plane to the disk surface, computing the redshift factor and integrating the reflection intensity (from the rest-frame reflection model) over the disk image. The key mechanism is the degeneracy that lets a thick disk's strongly redshifted inner emission mimic the signature of a higher-spin thin disk with a smaller ISCO.","core_discovery":"The central claim is that a thin-disk reflection model applied to a spectrum from a thick disk will not be rejected by the data, but will return biased parameters: the spin a* is always significantly overestimated, and in Kerr simulations the deformation parameter α13 is usually measured non-zero, meaning the fit sees an apparent deviation from the Kerr geometry. The bias in α13 weakens as the input spin increases, until at a* = 0.95 the fit recovers α13 = 0 but with a spin pegged at the model's maximum. The authors treat this as evidence that for near-extremal spins relativistic effects dominate the spectrum, which also explains why some super-Eddington AGN in the literature show strong apparent Kerr constraints.","pith_inferences":["A natural extension is to test whether a more realistic emissivity profile, e.g., from radiation-hydrodynamic simulations, changes the magnitude of the spin bias; the paper's q=9 assumption may be optimistic.","The same degeneracy could affect other measurements, such as continuum fitting or the shape of the Compton hump, so the bias may be a general feature of thick-disk geometry, not just of iron-line fitting.","If the interpretation of near-extremal spin sources is correct, then high-spin, tight-Kerr constraints in the literature should be re-examined for super-Eddington objects, possibly with a thick-disk model."],"forward_implications":["Reflection-based spin measurements of super-Eddington sources will tend to overestimate the black hole spin systematically.","Tests of the Kerr hypothesis using thin-disk reflection models on thick-disk sources can yield apparent non-Kerr deviations (non-vanishing α13) even when the black hole is exactly Kerr.","For very high spin, the fitted deformation parameter can return to zero while the spin is overestimated, so near-extremal spins measured this way may not be trustworthy.","Parameters like ionization, iron abundance, and photon index are still recovered correctly under the wrong disk model, so a good fit does not mean the disk geometry is right."],"supporting_citations":[{"why":"Provides the Polish donut model, the analytic thick-disk description used to generate the simulated reflection spectra.","marker":"Kozlowski et al. 1978"},{"why":"Developed the Polish donut solutions for barotropic thick disks in general relativity, the basis of the disk geometry.","marker":"Abramowicz et al. 1978"},{"why":"Describes the thin-disk, non-Kerr reflection model used to fit the simulated spectra.","marker":"Bambi et al. 2017"},{"why":"Introduces the parametric non-Kerr metric with deformation parameters (including α13) used to test deviations from Kerr.","marker":"Johannsen 2013"},{"why":"Provides the rest-frame reflection spectrum that is convolved with the disk geometry to produce the full reflection spectrum.","marker":"García et al. 2013"},{"why":"Used to validate the ray-tracing code for thin disks; supplies the relativistic convolution model that is the thin-disk alternative.","marker":"Dauser et al. 2013"},{"why":"Provides the X-ray instrument response used in the simulated observations.","marker":"Gendreau et al. 2016"}],"fun_headline_variants":["Thick disks bias spin and fake non-Kerr signals","Reflection fits misread thick disks as high spin","Thin-disk models misjudge thick disk reflection","Polish-donut disks trick reflection fits into fake spins","Thick disk spectra fit thin models but with bias"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Polish donut model with constant specific angular momentum and a power-law emissivity index q=9 captures the essential reflection behavior of real super-Eddington thick disks.","fun_headline_variants_meta":{"raw":{"variants":["Thick disks bias spin and fake non-Kerr signals","Reflection fits misread thick disks as high spin","Thin-disk models misjudge thick disk reflection","Polish-donut disks trick reflection fits into fake spins","Thick disk spectra fit thin models but with bias"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000778,"raw_usage":{"total_tokens":3395,"prompt_tokens":858,"completion_tokens":2537,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":474,"completion_tokens_details":{"reasoning_tokens":2458}},"tokens_in":474,"tokens_out":2537,"duration_ms":16585,"temperature":1.0,"reasoning_tokens":2458,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:26:55.115733+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a super-Eddington AGN whose black hole spin is known from an independent method (such as disk continuum fitting) and analyze its reflection spectrum with a thin-disk model; if the recovered spin matches the independent measurement, the claimed systematic overestimation would be falsified.","supporting_citations":[{"cited_title":"1978, A&A, 63, 209","cited_arxiv_id":null,"evidence_quote":"Provides the Polish donut model, the analytic thick-disk description used to generate the simulated reflection spectra."},{"cited_title":"2013, Phys","cited_arxiv_id":null,"evidence_quote":"Introduces the parametric non-Kerr metric with deformation parameters (including α13) used to test deviations from Kerr."},{"cited_title":"C., Arzoumanian, Z., Adkins, P","cited_arxiv_id":null,"evidence_quote":"Provides the X-ray instrument response used in the simulated observations."}],"review_version":1}