{"id":"0069e0e9-1664-4bb8-b100-dbc7098cfe02","arxiv_id":"2506.19900","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Spiral waves driven by an eccentric companion crossing an AGN disk produce accretion flares that can explain changing-look AGN with black hole perturbers and quasi-periodic eruptions with stellar perturbers.","lead":"An orbiting star or smaller black hole ploughing through the gas disk around a supermassive black hole can trigger short bursts of accretion, according to this paper's simulations. That would explain changing-look active galaxies (via black hole companions) and quasi-periodic X-ray eruptions (via stars) with one mechanism.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The source-specific conclusions rely on an unvalidated ~100-fold extrapolation of α^{-1}(H/R)^{-2} from the simulated thick disk (α≈0.08, H/R≈0.12) to canonical thin-disk values; the paper flags this sensitivity but still uses it to infer companion masses in §5.2.","rationale":"Reading the paper in good faith, the hydrodynamic model is a genuine contribution: the simulations demonstrate a transient one-armed spiral wave and an accretion-rate enhancement that roughly follows Eq. 10 over q=0.05–0.5 in the thick-disk regime, and the paper is appropriately careful to state where luminosity may not track accretion rate. The decisive question is whether the same dimensionless factor α^{-1}(H/R)^{-2} that normalizes the flare in the simulated disk can be carried over to real thin AGN disks, whose α and H/R differ by roughly two orders of magnitude. This is the reader's weakest assumption and I agree. The source-specific numbers in §5.2—q≈0.07 for ZTF18aahiqfi, q≈0.04 for iPTF16bco, q≈0.08 for ZTF18aaidlyq, and q≈9×10^-3 for AT2019qiz—all use the canonical thin-disk factor. Changing K to the simulation-calibrated value shifts q by a factor ≈10. Thus the inference that CL AGN require minor-merger IMBHs and the inference that QPE IMBH companions merge in <40 yr are not yet robust. A direct check is to simulate thinner disks and test the scaling. If the scaling fails, the paper's qualitative mechanism may still stand but the claimed mass regimes and star/IMBH division for QPEs would need revision. The appropriate verdict therefore remains CONDITIONAL, unchanged from the reader.","tokens_in":24418,"tokens_out":19088,"duration_ms":211931,"concrete_test":"Run the same q={0.1,0.3} encounter setup with the initial pressure lowered so the control disk has H/R≈0.04 or 0.03, measure α in the unperturbed control via Eq. 27, and compare the measured (δ−1)/(4q^2/9) with α^{-1}(H/R)^{-2} for each case. If the ratio changes by more than the simulation scatter rather than staying fixed, then Eq. 10's normalization is not transferable to thin disks; the Section 5.2 companion masses must be re-derived using the thin-disk factor, and the CL/QPE mass estimates are not yet constrained by this model.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The core quantitative claim—that CL AGN require black-hole companions with q∼10^-2 and that QPEs are best explained by stars—rests on inverting Eqs. 10/11, in which the flare amplitude is δ−1 = f α^{-1}(H/R)^{-2}. The simulations that validate f=(4/9)q^2 operate at α≈0.08, H/R≈0.12, giving t_visc/t_orb≈868. The source applications in §5.2 adopt α≈10^-2, H/R≈1/30, giving ≈9×10^4—a factor ~100 larger. Because q ∝ [α(H/R)^2]^{1/2}, the inferred companion mass scales as K^{-1/2}; for ZTF18aahiqfi (δ≈200), K=9×10^4 gives q≈0.07, while the simulation-calibrated K=868 gives q≈0.7, turning a minor merger into a comparable-mass binary. The §5.2/6 text acknowledges the 'highly uncertain' disk properties, but the exclusion of stellar CL-AGN perturbers and the claim that IMBH QPE companions merge in <40 yr both depend on the numerical value of K (and on the direct luminosity-to-Ṁ mapping). The stellar exclusion is more robust, because (R*/rp)^2 is already tiny at the CL-AGN rp values, but the BH mass scale and the QPE IMBH viability are not. No simulation or observation has yet tested Eq. 10 at H/R≪0.1, so this is the least secure load-bearing step.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes that extreme AGN variability (changing-look AGN and quasi-periodic eruptions) is produced by an eccentric companion—either a star or a moderately massive black hole—crossing the accretion disk of the central SMBH. The passage excites a transient one-armed spiral wave that drives mass inflow on a few orbital timescales, and the flare amplitude is derived as δ = f α^{-1}(H/R)^{-2} + 1, with f = (4/9)q^2 for a point-mass companion (Bondi radius) and f = (R_*/r_p)^2 for a stellar companion. The scaling is tested with a suite of 3D Athena++ simulations for q = 0.05–0.5, e = 0.93, i = 90°, using a control run to measure α ≈ 0.08 and H/R ≈ 0.12; the simulations reproduce the predicted δ ∝ q^2 dependence (Figure 4). The framework is then applied to three ZTF CL AGN and to QPE sources, leading the authors to conclude that CL AGN require eccentric SMBH companions with q ∼ 10^{-2} (stars excluded) and that QPEs are best explained by main-sequence or stripped stars, with IMBH companions disfavored because of short gravitational-wave merger timescales.","tokens_in":24837,"tokens_out":13203,"duration_ms":142763,"significance":"If correct, this would be a valuable step toward a predictive model connecting disk-embedded companions to both CL AGN and QPEs. The paper's strengths are that the central scaling is derived from a simple physical picture and tested against global simulations rather than fitted to the source light curves, that the predicted flare duration of a few orbital timescales is concrete and falsifiable, and that the source constraints are presented explicitly rather than hidden in a black box. The main limitation is that the α and H/R values adopted for the source inference differ by roughly two orders of magnitude in the combination α^{-1}(H/R)^{-2} from the values that calibrate the simulations, and the paper's quantitative companion-mass claims rest on that extrapolation.","major_comments":[{"comment":"The source-specific companion masses are computed with α = 10^{-2} and H/R = 1/30, giving K ≡ α^{-1}(H/R)^{-2} ≈ 9 × 10^4, whereas the simulations that validate Eq. (10) are calibrated at α ≈ 0.08 and H/R ≈ 0.12 (Section 4.1), giving K ≈ 868. Since q ∝ K^{-1/2}, the inferred q is a factor of about 10 larger when the simulation-calibrated K is used: for ZTF18aahiqfi, Eq. (31) gives q ≈ 0.07, but the simulation-calibrated value is q ≈ 0.7, turning a minor merger into a near-equal-mass binary; similar factor-of-ten shifts apply to iPTF 16bco, ZTF18aaidlyq, and AT2019qiz. The caveat in Section 6 that disk properties are 'highly uncertain' does not quantify this factor-of-ten shift. The authors should either justify the thin-disk extrapolation of the α^{-1}(H/R)^{-2} normalization with a concrete physical argument, or present the source inferences as explicit functions of K that include the simulation-calibrated value, and revise the abstract and summary claims about q ∼ 10^{-2} accordingly.","section":"§5.2, Eqs. (10)–(11), (31)–(34)"},{"comment":"The numerical validation covers only point-mass black hole perturbers; no simulation includes an extended star. The QPE conclusions—that main-sequence or stripped stars can reproduce the observed amplitudes and that low-mass or stripped stars are required for many sources—rest on the geometric f = (R_*/r_p)^2 scaling of Eq. (16), which is not tested by the simulations. At minimum, the paper should state explicitly that Eq. (16) is an untested geometrical ansatz for the star-disk interaction; ideally, it should add a stellar-companion simulation or cite a dedicated study that validates the area-ratio prescription in a similar disk.","section":"§4.3 vs. §5.2.2 and §5.3, Eqs. (14)–(19)"},{"comment":"The sentence 'For all QPE sources shown in Figure 8, t_merger ≲ 40 years. This remains true even in the conservative case where the mass of the perturber is computed using α^{-1}(H/R)^{-2} = 10^6' is quantitatively incorrect. For q ≪ 1, Eq. (35) gives t_merger ∝ q^{-1}, while Eq. (11) gives q ∝ [α(H/R)^2]^{1/2} = K^{-1/2}; therefore increasing K from 9 × 10^4 to 10^6 increases the merger time by a factor of about √11.1 ≈ 3.3, so the bound becomes roughly 130 years rather than 40 years. The conclusion that IMBHs are short-lived may still hold, but the stated bound should be corrected and the range of t_merger under the disk-parameter uncertainty should be reported.","section":"§5.3.1, Eq. (35)"}],"minor_comments":[{"comment":"The phrase 'for a point mass perturber (for a stellar perturber)' is confusing; the simulations only test the point-mass case, and the sentence should be rewritten to say so directly.","section":"§4.3, final paragraph"},{"comment":"The footnote notes that a prograde orbit with i = 15° can increase δ by up to a factor of about 500. Since the source applications in Section 5.2 all assume i = 90°, this inclination dependence should be mentioned in Section 5.2 as an additional systematic uncertainty on the inferred companion parameters.","section":"§2.2, footnote 1"},{"comment":"The reference list contains duplicate entries for Franchini et al. (2023), Linial & Metzger (2023), and Miniutti et al. (2019); these should be merged.","section":"References"},{"comment":"The paper correctly notes that the luminosity-to-accretion-rate mapping is conditional, but Section 5.2 then uses bolometric luminosity ratios directly as δ. A brief quantitative caveat about how radiative efficiency or anisotropic emission could change the inferred δ would help the reader judge the source constraints.","section":"§5, first paragraph"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the journal's scope and the numerical/analytical core is basically sound. The main concern is the factor-of-ten sensitivity of the source applications to the assumed disk parameters, which is fixable by revising the presentation and quantifying the extrapolation. I did not find evidence of duplicate publication or problematic citation behavior; the model is sufficiently distinct from the authors' earlier related work (Huang et al. 2025) to stand alone."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this paper contains a real, new numerical result: global 3D Athena++ simulations of an eccentric, inclined point-mass perturber crossing an AGN disk. The simulations show a transient one-armed spiral wave and an accretion flare whose amplitude follows (4/9)q^2 alpha^-1(H/R)^-2+1 across q = 0.05-0.5, using alpha and H/R measured from the control run. The Bondi-radius picture holds up in the simulations, and the clean separation between stellar and black-hole perturbers is useful. That part is solid, and the paper deserves credit for it.\n\nThe soft spot is in the application to specific sources. The inferred companion masses for CL AGN and the rejection of IMBHs for QPEs both rest on extrapolating the scaling from the simulated thick disk (alpha ~ 0.08, H/R ~ 0.12, viscous-to-orbital time ~ 870) to canonical thin-disk values (alpha ~ 0.01, H/R ~ 1/30, time ~ 9e4), a factor of about 100 that no simulation or observation has tested. Because q scales as the inverse square root of that factor, ZTF18aahiqfi's delta ~ 200 gives q ~ 0.07 at the canonical value but q ~ 0.7 at the simulation-calibrated value. That is the difference between a minor merger and a comparable-mass binary, so the paper's claim that CL AGN require \"lighter SMBH companions\" is only true under one particular (untested) disk model. The paper does flag that the disk properties are highly uncertain, which is honest, but the abstract and summary do not carry that caveat, and the point estimates in Section 5.2 are presented more firmly than the evidence allows.\n\nSome parts of the source application are more robust. The exclusion of stellar perturbers for CL AGN does not depend on the viscous factor, because rp is hundreds of Schwarzschild radii and (R*/rp)^2 is tiny even for red giants. The QPE stellar interpretation is also plausible. The IMBH exclusion for QPEs, however, depends on the same unvalidated K, and the direct mapping from observed luminosity ratio to Mdot ratio is an additional assumption. These are addressable weaknesses: rerun at least one simulation with H/R ~ 0.03, or present the source fits as a function of K rather than as single masses.\n\nThe minority: the simulations go down to q = 0.05, so the extrapolation to q ~ 0.01 is an extrapolation, but a mild one. The paper's own limitation statement in Section 6 acknowledges the sensitivity to disk properties, which I read as a sign the authors know where the weak joint is.\n\nWho should read this: anyone working on QPEs, changing-look AGN, or disk-embedded binary dynamics. It is worth a serious referee, not a desk reject, and the core mechanism will survive. The source-specific conclusions need to be reframed as conditional on the viscous-to-orbital time ratio. I would send it to review, and I would push hard on the thin-disk extrapolation in the referee report.","headline":"The paper's simulation-validated q^2 scaling is a genuine step forward, but the source-specific companion masses rely on an untested factor-of-100 extrapolation to thin-disk parameters and are not as secure as the abstract suggests.","tokens_in":25367,"tokens_out":2872,"would_cite":true,"duration_ms":31335,"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":"An orbiting companion crossing the accretion disk of a supermassive black hole can drive the rapid, extreme flares seen in changing-look AGN and quasi-periodic eruptions.","keywords":["changing-look AGN","quasi-periodic eruptions","accretion disk perturbations","spiral density waves","supermassive black hole binaries","intermediate-mass black holes","tidal disruption","Bondi radius"],"falsifier":"Compute the required companion mass for ZTF18aahiqfi using the disk parameters actually measured in the simulations ($\\alpha=0.08$, $H/R=0.12$) rather than the canonical thin values ($\\alpha=10^{-2}$, $H/R=1/30$); Equation 12 then gives $q\\approx0.7$ instead of $q\\approx0.07$, turning the claimed minor merger into a comparable-mass binary. A direct measurement of the pre-outburst disk's thickness and viscosity in any real CL AGN would determine which regime is physical.","tokens_in":24263,"feed_emoji":"🌌","tokens_out":15092,"duration_ms":134201,"temperature":0.7,"pith_summary":"This paper proposes one mechanism for two extreme kinds of variability in galactic nuclei: changing-look AGN, which switch their accretion state on timescales of months to years, and quasi-periodic eruptions, which recur every hours to days. The mechanism is a companion—either a star or a moderately massive black hole—on a highly eccentric orbit that crosses the accretion disk of the central supermassive black hole once per orbit. The crossing excites a transient spiral density wave that carries mass inward on roughly the orbital timescale instead of the much longer viscous timescale, producing a burst of accretion whose amplitude is set by the companion's sphere of influence: its Bondi radius for a black hole, its physical radius for a star. Using analytic scalings and three-dimensional hydrodynamic simulations, the paper shows that changing-look AGN require black hole companions with mass ratio $q \\sim 10^{-2}$ and eccentricity $e \\gtrsim 0.8$, while most quasi-periodic eruptions require low-mass main-sequence or stripped stars. If correct, the model explains the otherwise puzzling speed and amplitude of these transients and turns them into probes of otherwise unseen companions around supermassive black holes.","feed_headline":"One mechanism unifies changing-look AGN and quasi-periodic eruptions","feed_subtitle":"The same disk-crossing mechanism produces both month-long changing-look flares and hour-long QPEs","key_machinery":"The engine is the transient one-armed spiral density wave excited by a companion plunging through the disk, together with the identity that fixes its effect: $\\delta = f\\,\\alpha^{-1}(H/R)^{-2} + 1$, where $f$ is the fraction of the enclosed disk mass inside the companion's sphere of influence and $\\alpha^{-1}(H/R)^{-2}$ is the factor by which the viscous time exceeds the orbital time at the crossing radius. For a black hole, $f = (4/9)q^2$ because the sphere of influence is the Bondi radius, the radius at which the companion's gravity dominates the relative kinetic energy of the gas. For a star, $f = (R_*/r_p)^2$ because the star's physical radius is the cross-section. The simulations confirm that the wave is transient, that the disk becomes eccentric and then returns to circular within about an orbital period, and that the flare delivers 90 percent of its excess mass in $t_{90} \\approx 3\\,\\mathrm{t_{orb}}$, which validates using the pericenter orbital time as the flare duration.","core_discovery":"The paper's central claim is a scaling law for the accretion flare produced when an eccentric companion crosses a thin disk: the flare amplitude is $\\delta = (4/9)q^2 \\alpha^{-1}(H/R)^{-2} + 1$ for a black hole companion of mass ratio $q$, and $\\delta = (R_*/r_p)^2 \\alpha^{-1}(H/R)^{-2} + 1$ for a star of radius $R_*$ crossing at pericenter $r_p$, where $\\alpha$ is the disk viscosity parameter and $H/R$ its scale height ratio. The Bondi radius sets the disturbance size for a point mass and the stellar radius for a star, and the relative amplitude is just the ratio of the viscous time to the orbital time at the crossing radius times the fraction of disk mass within that disturbance. The paper calibrates these scalings against global hydrodynamic simulations with mass ratios $q = 0.05$ to $0.5$, which reproduce the predicted amplitudes and show the flare lasts a few orbital periods. Applied to individual sources, the observed transition timescale fixes the pericenter distance and the luminosity contrast fixes the companion mass: changing-look AGN are best matched by moderately massive black hole companions with $q \\sim 10^{-2}$ on orbits with $e \\gtrsim 0.8$, whereas quasi-periodic eruptions are best matched by stars, often low-mass main-sequence or stripped stars, because IMBH companions would merge by gravitational-wave emission in under about forty years.","pith_inferences":["If the scaling survives at lower mass ratios, the same mechanism should appear in any tilted or eccentric binary disk system, and the predicted $t_{90}\\sim$ a few orbital times envelope could be searched for in existing X-ray light curves of other repeating transients.","A multi-epoch light curve of a repeating QPE could identify the companion type on its own: a stellar perturber's amplitude should scale as $(R_*/r_p)^2$ and vary if the pericenter precesses, while a black hole perturber's amplitude is independent of $r_p$.","The companion masses are sensitive to the assumed disk transport factor; a thin-disk simulation with $\\alpha\\approx10^{-2}$ and $H/R\\approx1/30$ would test the extrapolation directly, since the present runs only calibrate $\\alpha\\approx0.08$ and $H/R\\approx0.12$.","Because the paper sets 'turn-off' CL AGN aside, a symmetric mechanism that shuts off accretion rather than turning it on would be a natural complement; if turn-off events show the same fast transitions, they may be the same perturbation acting on a disk close to a stability boundary."],"forward_implications":["A measured transition timescale directly fixes the pericenter distance through $t_{\\rm orb}(r_p)$, so CL AGN and QPE light curves become geometric probes of the companion's orbit.","Because most CL AGN have not repeated within observational baselines of about twenty years, the model requires $e \\gtrsim 0.8$; sources that do repeat should show the orbital period as their recurrence timescale.","Stellar companions cannot account for CL AGN flares, since even red giants at the required pericenter distances yield amplitudes far below the observed values, so bright turn-on events become evidence for hidden black hole companions with $q \\sim 10^{-2}$.","QPE sources with short eruption durations require pericenters near the tidal radius, which selects low-mass main-sequence stars and stripped-envelope stars; IMBH companions are ruled out because their merger timescales are shorter than about 40 years.","If the disk returns to its pre-flare state after a few orbital times, the model predicts repeat flares with the orbit, so long-term monitoring should reveal either recurrence on that timescale or a persistent change of state."],"supporting_citations":[{"why":"Supplies the numerical solver used for the global disk simulations that calibrate the flare amplitude scaling.","marker":"Stone et al. 2020"},{"why":"Provides the method for measuring the effective viscosity parameter $\\alpha$ from the accretion rate and surface density.","marker":"Ju et al. 2016"},{"why":"Supplies the linear spiral-wave dispersion relation used to fit the shock pitch angle in the simulations.","marker":"Huang et al. 2025"},{"why":"Gives the canonical thin-disk values $\\alpha\\simeq10^{-2}$ and $H/R\\simeq1/30$ used to convert flare amplitudes into companion masses.","marker":"King et al. 2007"},{"why":"Supplies the changing-look AGN sources with their transition timescales, luminosity contrasts, and black hole masses.","marker":"Frederick et al. 2019"},{"why":"Supplies the AT2019qiz parameters and the QPE sample used to compare stellar and IMBH perturber models.","marker":"Nicholl et al. 2024"},{"why":"Provides the main-sequence stellar radius-mass relation used to compute stellar flare amplitudes and allowed stellar masses.","marker":"Tout et al. 1996"},{"why":"Supplies the gravitational-wave merger timescale formula used to rule out IMBH companions for QPE sources.","marker":"Zwick et al. 2020"},{"why":"Supplies the iPTF16bco example used to estimate the pericenter distance and eccentricity for a changing-look AGN.","marker":"Gezari et al. 2017"},{"why":"Introduces GSN 069 and the QPE phenomenon used as the worked example for QPE orbital parameters.","marker":"Miniutti et al. 2019"}],"fun_headline_variants":["One mechanism drives changing-look AGN and QPEs","Companions crossing AGN disks explain extreme flares","Eccentric disk crossings power AGN variability","Black holes and stars stir AGN disks into flares","Unified origin for changing-look AGN and eruptions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that real thin AGN disks have the transport factor $\\alpha^{-1}(H/R)^{-2}\\approx9\\times10^4$ used to convert observed luminosity contrasts into companion masses—the simulations only calibrate a thicker, more viscous disk where the factor is about 870—and that observed luminosity contrast tracks the accretion-rate contrast; the paper itself notes this sensitivity, and if the simulation-calibrated values were used instead, the inferred mass ratio for ZTF18aahiqfi would rise from $q\\approx0.07$ to $q\\approx0.7$.","fun_headline_variants_meta":{"raw":{"variants":["One mechanism drives changing-look AGN and QPEs","Companions crossing AGN disks explain extreme flares","Eccentric disk crossings power AGN variability","Black holes and stars stir AGN disks into flares","Unified origin for changing-look AGN and eruptions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000204,"raw_usage":{"total_tokens":1461,"prompt_tokens":1085,"completion_tokens":376,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":701,"completion_tokens_details":{"reasoning_tokens":298}},"tokens_in":701,"tokens_out":376,"duration_ms":4102,"temperature":1.0,"reasoning_tokens":298,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T18:25:24.330396+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the required companion mass for ZTF18aahiqfi using the disk parameters actually measured in the simulations ($\\alpha=0.08$, $H/R=0.12$) rather than the canonical thin values ($\\alpha=10^{-2}$, $H/R=1/30$); Equation 12 then gives $q\\approx0.7$ instead of $q\\approx0.07$, turning the claimed minor merger into a comparable-mass binary. A direct measurement of the pre-outburst disk's thickness and viscosity in any real CL AGN would determine which regime is physical.","supporting_citations":[{"cited_title":"J., et al","cited_arxiv_id":null,"evidence_quote":"Supplies the changing-look AGN sources with their transition timescales, luminosity contrasts, and black hole masses."},{"cited_title":"A., Pols, O","cited_arxiv_id":null,"evidence_quote":"Provides the main-sequence stellar radius-mass relation used to compute stellar flare amplitudes and allowed stellar masses."},{"cited_title":"B., et al","cited_arxiv_id":null,"evidence_quote":"Supplies the iPTF16bco example used to estimate the pericenter distance and eccentricity for a changing-look AGN."}],"review_version":2}