{"id":"79a7e360-ae68-428f-85d9-cb3d5f62e887","arxiv_id":"2608.01249","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A revised ADAF-collapse model, with the transition radius estimated from optical flux changes, yields cooling timescales comparable to the observed sub-year turn-on timescales of two changing-look AGNs.","lead":"This paper tries to explain why two supermassive black holes switched on in less than a year. It revises an earlier model and finds that the newly calculated cooling timescales can match those fast transitions, but only if an unmeasured parameter is allowed to vary.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The inferred R_tr assumes the dim-state 5100 Å flux is purely outer thin disk; unmodeled host or ADAF emission shifts R_tr and t_cool substantially, and Table 1's errors exclude this systematic.","rationale":"The reader's weakest_assumption identifies the same load-bearing point: the flux ratio L'_5100/L_5100 is mapped entirely to a change in R_tr under the assumption that the inner ADAF emits no 5100 Å flux and that host-galaxy starlight is negligible. My concern sharpens this: the direction of any contamination is unfavorable. Because t_cool ∝ R_tr^{3/2}, even a small host or ADAF contribution increases the inferred R_tr and lengthens t_cool, widening the gap with the observed upper limits. The paper is honest about several limitations: it states in Section 1 that both observed transition timescales are upper limits, and the Table 1 note concedes that model-parameter uncertainties are not included. It also fairly presents the magnetic-field model as failing to meet the timescales. These admissions support moderate confidence, but they do not resolve the flux-decomposition issue, which is the most load-bearing unquantified assumption. The proposed spectral decomposition is a feasible, concrete check with existing data: if the host/ADAF contribution is negligible, the concern is retired and the 'comparable' claim stands as a factor-of-~1.4 post-diction; if not, the central claim is weakened. Since the reader already reached CONDITIONAL and this concern reinforces the same condition rather than overturning the paper, the verdict is unchanged.","tokens_in":11374,"tokens_out":8662,"duration_ms":85071,"concrete_test":"Perform a two-epoch spectral decomposition of the 5100 Å continuum for SDSS J0225+0030 and SDSS J1723+5504 using the existing SDSS/BOSS spectra: fit a power-law AGN continuum, an Fe II pseudo-continuum, and a host-galaxy template, optionally adding an ADAF SED component, to obtain AGN-only 5100 Å flux ratios. Recompute R_tr and t_cool from Eqs. (1)-(2) with those ratios. If the best-fit host or ADAF contribution in the dim state exceeds about 5% of the 5100 Å flux, check whether the revised t_cool values remain within the claimed factor of ~1.4 of the upper limits; if not, the central agreement is an artifact of the assumed pure-disk decomposition.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Eqs. (1)-(2), the dim-state 5100 Å flux is modeled as the outer-disk integral from R_tr to Rout only; no ADAF contribution and no host-galaxy term appear. The central claim t_cool ~ t_tran depends on R_tr through t_cool ∝ R_tr^{3/2}, so any contamination in the observed L'_5100/L_5100 ratio propagates nonlinearly into the comparison. If host starlight or ADAF emission contributes a fraction h of the dim-state 5100 Å flux, the intrinsic disk flux ratio is lower than the observed 74.1% and 69.2%, so the true R_tr is larger than 14.89 Rs and 30.48 Rs and t_cool is longer. This systematic acts in the direction that worsens the already marginal agreement: the constant-mdot t_cool values, 342 and 197 days, exceed the observed upper limits of 254 and 142 days. The quoted uncertainties (Table 1, ±1.20 and ±4.91 Rs) are only formal spectral-fitting errors, as the note states, and exclude this flux-decomposition error. Because both observed timescales are upper limits, the variable-mdot parameter range in Figure 2 is effectively absorbing the gap without an independent measurement of mdot'/mdot. Thus the quantitative match is not yet settled.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper revisits two rapidly changing-look AGNs, SDSS J0225+0030 and SDSS J1723+5504, whose observed transition timescales are shorter than one year and which were outliers in the ADAF-collapse model of Li & Cao (2025, LC2025). The authors first show that a thin disk with a large-scale poloidal magnetic field cannot reproduce the observed timescales, then revise the LC2025 model by estimating the ADAF/outer-disk transition radius R_tr from the observed ratio of dim-state to bright-state 5100 Å flux using standard thin-disk flux integrals. With a constant mass accretion rate, the inferred R_tr values give cooling timescales of 342 days and 197 days for the two objects, compared with observed upper limits of 254 days and 142 days. Allowing the dim-state accretion rate to be lower than the bright-state value gives mdot'/mdot ranges of about 74%--91% and yields R_tr values that make the cooling timescale equal to the observed transition timescale. The revised model is then applied to the full LC2025 sample.","tokens_in":11541,"tokens_out":6149,"duration_ms":54506,"significance":"If the result holds, the paper would offer a plausible explanation for the short transition timescales in two previously problematic CL AGNs and would strengthen the LC2025 cooling-collapse scenario. The approach is transparent: the flux-integral calculation is simple and reproducible, the magnetic-field appendix provides a concrete negative result, and the application to the full sample is a useful consistency check. The main novelty is using the observed 5100 Å flux variation to locate the transition radius rather than relying on a temperature criterion. However, the quantitative match is not yet a free prediction: it is completed by allowing the unconstrained ratio mdot'/mdot to vary, and the dominant systematic in the flux decomposition is not estimated. The central claim is therefore plausible but currently under-supported.","major_comments":[{"comment":"The central claim that the cooling timescale is 'comparable to' the observed transition timescale is not established as a prediction. In the constant-mdot case, the model gives t_cool = 342 days and 197 days, which exceed the observed upper limits of 254 days and 142 days listed in Table 1. Agreement is achieved only by allowing mdot'/mdot to vary, and the quoted ranges are obtained by first setting t_cool = t_tran, which selects R_tr = 12.21 R_s and 24.5 R_s from the degeneracy curve in Figure 2, and then reading off mdot'/mdot. Because mdot'/mdot is not independently measured, the reported agreement is a consistency condition rather than a test of the model. The paper should either be reframed as an allowed parameter-space consistency check, with clear language, or provide an independent observational constraint on mdot'/mdot.","section":"§3, Figures 1–2 and Table 1"},{"comment":"The inferred R_tr rests entirely on the assumption that the dim-state 5100 Å flux is produced by the outer thin disk, with no contribution from the inner ADAF and no host-galaxy starlight. Since t_cool ∝ R_tr^{3/2}, any contamination in the observed L'_5100/L_5100 ratio propagates nonlinearly into the predicted timescale. The note to Table 1 correctly states that the quoted uncertainties are only formal spectral-fitting errors, but the dominant systematic is the flux decomposition and it is not estimated. If even roughly ten percent of the dim-state flux is host starlight or ADAF emission, the intrinsic disk-flux ratio is lower than 74.1% and 69.2%, R_tr becomes larger, and the already marginal constant-mdot agreement worsens. The authors should quantify this systematic using the available spectra or justify its neglect with a specific argument.","section":"§2, Eqs. (1)–(2), Table 1 note"},{"comment":"The observed transition timescales for both objects are upper limits, as the text itself notes: SDSS J1723+5504 has only two epochs and SDSS J0225+0030 has an intermediate observation that still gives an upper limit. The analysis nevertheless treats t_tran as exact when setting t_cool = t_tran and deriving the allowed mdot'/mdot ranges. If the true transition timescale is shorter than the quoted upper limit, the required R_tr is smaller and the inferred mdot'/mdot range shifts. The comparison should be formulated as inequalities, t_cool ≤ t_tran,upper, and the allowed parameter space should be presented as such rather than as a single curve crossing.","section":"§1 and Table 1, col. (7)"}],"minor_comments":[{"comment":"The title contains 'SD SS J0225+0030' with a spurious space; it should read 'SDSS J0225+0030' consistently throughout.","section":"Title and Abstract"},{"comment":"The text uses 'Angstrom' instead of the proper symbol 'Å', and Eq. (4) should state the dimensions of the numerical constant explicitly, since the constant is dimensional.","section":"§3, after Eq. (4)"},{"comment":"It is not clear whether the full-sample comparison in Figure 3 uses the constant-mdot or variable-mdot variant, and whether the two target objects are plotted with the tuned mdot'/mdot values or with the constant-mdot values. Please clarify which model variant produced the points in Figure 3.","section":"Table 2 and Figure 3"},{"comment":"The Figure 4 note mentions revised-model R_tr values of 10.83 R_s and 21.52 R_s, but these do not appear in Table 1; please define them and explain how they are obtained.","section":"Figure 4 note"},{"comment":"The statement that the magnetic-field model timescale 'remains significantly longer than one year' is vague; the appendix gives values of order 10^6 days for strong magnetic fields, which is many orders of magnitude longer than one year and should be stated precisely.","section":"Abstract and §1"},{"comment":"The inequality '74.13% < mdot'/mdot < 90.99%' uses 74.13%, the observed flux ratio, as the lower bound; since the lower limit is set by R_tr = 3 R_s, this should be stated as equal to the observed ratio, 74.1%, or the provenance should be clarified.","section":"§3, parameter range"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the manuscript addresses a genuine puzzle and the model idea is worth pursuing, but the central quantitative claim is currently a consistency check with a free parameter and an unquantified systematic. I would be willing to review a revision that reframes the claim as an allowed-parameter consistency test, adds a flux-decomposition systematic estimate, and reports the comparison as inequalities against the observed upper limits."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a straightforward follow-up to Li & Cao (2025) with one genuinely new ingredient: instead of guessing the ADAF-to-disk transition radius from a temperature threshold, they infer it from the observed ratio of 5100 Angstrom fluxes between the dim and bright states. That is a cleaner method, and it shrinks the predicted cooling timescales enough that the two outliers in the LC2025 sample - SDSS J0225+0030 and SDSS J1723+5504 - go from excluded to borderline. In the constant-mdot case the prediction is 342 vs 254 days and 197 vs 142 days, off by a factor of about 1.4. The authors then show that letting the dim-state mass accretion rate be 74-91% of the bright-state value makes the agreement exact. They are open about the fact that this ratio is not measured, and they present both as comparable, not as a clean prediction. I find that honest. What is actually new: the R_tr re-estimation from the optical flux ratio, the physical insight that high-mass, low-mdot disks are cool enough for the 5100 Angstrom flux to be very sensitive to the inner disk edge, and a clean null result for the large-scale magnetic field mechanism. The magnetic-field calculation is useful, and the full-sample sanity check is reasonable, including the conscious decision to exclude objects whose continua decline as lines appear. Where the soft spots are: the stress-test worry is real. The entire inference of R_tr assumes the dim-state 5100 Angstrom light is produced only by the outer thin disk. Any host-galaxy starlight or ADAF emission in that band would make the true disk-flux ratio lower than the observed 74% and 69%, which pushes R_tr upward and t_cool longer - toward the wrong side of the already marginal constant-mdot agreement. The quoted errors in Table 1 are purely spectral-fitting formalities; the table note says so, and the model-parameter systematics are not included. The variable-mdot branch essentially selects R_tr so that t_cool equals the observed upper limit and then reads off the implied mdot ratio, which is not a strong test. None of these flaws is fatal, and the paper is transparent about the upper-limit nature of both observed timescales, but they cap the significance. Who gets value from this: people actively working on CL AGN disk models, and anyone comparing theoretical timescale arguments to sparse monitoring data. It does not settle the mechanism, but it rescues the two hardest cases from immediate exclusion and gives a clear recipe to falsify the model with better photometric decomposition or repeated spectroscopy. Recommendation: send it to peer review. I would ask the referee to push for a robustness test with a plausible host-galaxy contribution added to the dim-state flux, and for a discussion of how much ADAF emission at 5100 Angstrom would matter. As it stands, this is a legitimate, modest step forward, and the authors' willingness to state their own limitations deserves a fair read.","headline":"A transparent, incremental revision of the authors' own ADAF-cooling model that turns two outlier CL AGNs into borderline successes, but the exact match relies on an unmeasured accretion-rate ratio and a clean flux decomposition that is not robustly tested.","tokens_in":893,"tokens_out":981,"would_cite":false,"duration_ms":44351,"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":"The revised model attributes the rapid turn-ons of SDSS J0225+0030 and SDSS J1723+5504 to the collapse of an inner hot accretion flow, with cooling times comparable to the observed 254-day and 142-day transitions.","keywords":["changing-look AGN","accretion disk","advection-dominated accretion flow","transition radius","cooling timescale","super massive black hole","viscous timescale","SDSS J0225+0030"],"falsifier":"A high-cadence spectroscopic campaign that caught either object completing its transition in well under 254 or 142 days rest frame, or a spectral decomposition showing that the dim-state 5100 Å continuum contains a substantial component from the inner hot flow or host stars rather than only the outer thin disk, would settle the claim against the model.","tokens_in":11002,"feed_emoji":"🔭","tokens_out":6029,"duration_ms":54473,"temperature":0.7,"pith_summary":"This paper addresses two changing-look active galactic nuclei that switched on in less than a year, faster than standard disk models can easily produce. It argues that the earlier ADAF-collapse explanation failed for these two only because the transition radius was overestimated, and that using the observed 5100 Å flux change to locate that radius makes the predicted ADAF cooling times comparable to the observed 254-day and 142-day transitions. Because these black holes are massive (about $10^9$ solar masses) and accrete at only about one percent of Eddington, the disk is cool enough that a modest inward move of the hot inner flow can halve the optical flux. If right, the result removes the two outlier objects from the sample and strengthens the case that rapid changing-look events are accretion-mode transitions rather than obscuration or tidal disruptions.","feed_headline":"Rapid AGN turn-ons blamed on collapsing hot inner disk","feed_subtitle":"Cooling times for the inner accretion flow now match the observed 254-day and 142-day transitions.","key_machinery":"The load-bearing object is the transition radius $R_{\\rm tr}$ between the inner ADAF and the outer thin disk, because the ADAF cooling time scales as $t_{\\rm cool}\\propto \\Omega_K^{-1}\\propto R_{\\rm tr}^{3/2}$. The paper replaces the earlier temperature-based estimate with a flux-based one: equations (1)--(3) compute the 5100 Å flux from a multicolor thin disk with an inner edge at $R_{\\rm tr}$, and the observed ratio $L'_{5100}/L_{5100}$ fixes $R_{\\rm tr}$ under the assumption that the inner ADAF contributes no 5100 Å light. A second curve relates the flux ratio to the accretion-rate ratio $\\dot m'/\\dot m$, showing the degeneracy between a smaller $R_{\\rm tr}$ and a lower dim-state accretion rate.","core_discovery":"The paper claims that the sub-year turn-on times of SDSS J0225+0030 and SDSS J1723+5504 can be explained by the collapse of an inner advection-dominated accretion flow (ADAF) into a thin disk, provided the transition radius $R_{\\rm tr}$ is inferred from the observed 5100 Å flux ratio between the dim and bright states rather than from a rough temperature criterion. With the flux ratio alone and a constant mass accretion rate, the inferred transition radii drop from $78.1\\,R_s$ to $14.89\\,R_s$ and from $158.4\\,R_s$ to $30.48\\,R_s$ for the two objects, shortening the ADAF cooling times from thousands of days to $342$ and $197$ days. Allowing the accretion rate to drop in the dim state pushes the required $R_{\\rm tr}$ even smaller, so the cooling time can be made shorter than the observed upper limits of $254$ and $142$ days. The paper also shows that a large-scale poloidal magnetic field, though it shortens thin-disk viscous timescales by about three orders of magnitude, still leaves them far longer than one year for these objects.","pith_inferences":["The same flux-ratio method could be applied to other changing-look AGNs with multi-epoch spectra, converting each measured 5100 Å flux change into a predicted transition timescale and giving a ready-made test of the ADAF-collapse scenario across a larger sample.","Because the observed transition timescales are upper limits and the constant-accretion-rate predictions already exceed them, the model's success leans on allowing the accretion rate to drop in the dim state; a decisive check would be independent dim-state accretion-rate indicators, such as X-ray or UV luminosity, falling in the inferred $\\dot m'/\\dot m$ ranges.","If the inner ADAF or host starlight contributes a measurable fraction of the dim-state 5100 Å continuum, the inferred $R_{\\rm tr}$ would move outward and the predicted cooling time would grow, so high-quality spectral decomposition of the dim-state continuum directly tests the paper's central mapping.","The reverse process, a turn-off where the thin disk recedes and an ADAF reforms, should obey the same flux-ratio relation with a heating timescale, suggesting that turn-off changing-look AGNs with well-sampled light curves could be used as independent checks of the model."],"forward_implications":["The two fastest turn-on changing-look AGNs no longer require exotic mechanisms; the same ADAF-collapse process that explains the broader sample can account for them once $R_{\\rm tr}$ is measured from the flux change.","High-cadence spectroscopic monitoring of these objects should find transitions that are consistent with the predicted cooling times, and any further shortening of the observed upper limits would tighten the allowed $\\dot m'/\\dot m$ parameter space.","Large black hole mass and low Eddington ratio make an AGN a good candidate for rapid changing-look behavior, because the low disk temperature lets a small inward move of the inner flow produce a large optical flux change.","In the constant-accretion-rate case, the model predicts that the thin disk in the dim state is truncated at about $15\\,R_s$ and $30\\,R_s$ for the two objects, a geometry that could be tested with detailed continuum fitting or reverberation mapping.","A large-scale poloidal magnetic field acting on a thin disk is insufficient to explain sub-year transitions in these two objects, so the ADAF-collapse channel is the preferred explanation."],"supporting_citations":[{"why":"Supplies the original ADAF-collapse model and the cooling-time formula that this paper revises by improving the $R_{\\rm tr}$ estimate.","marker":"Li & Cao 2025"},{"why":"Reports the discovery, black hole mass, and 254-day rest-frame transition timescale of SDSS J0225+0030.","marker":"MacLeod et al. 2016"},{"why":"Reports SDSS J1723+5504 as a changing-look AGN with a 142-day upper-limit transition timescale and supplies its black hole mass and accretion-rate estimate.","marker":"Potts & Villforth 2021"},{"why":"Provides the evidence that dim-state CL AGNs contain an inner ADAF plus outer thin disk and that the accretion rate increases during brightening.","marker":"Ruan et al. 2019"},{"why":"Provides the intermediate-epoch observation of SDSS J0225+0030 and the continuum measurements used for the flux-ratio analysis.","marker":"Panda & Śniegowska 2024"},{"why":"Defines the standard thin-disk viscous timescale that is too long to explain the observed transitions and serves as the baseline the magnetic-field and ADAF models must beat.","marker":"Shakura & Sunyaev 1973"},{"why":"Supplies the magnetized thin-disk equations and radial-velocity formalism used to compute the magnetically shortened viscous timescales in the appendix.","marker":"Li & Begelman 2014"}],"fun_headline_variants":["Sub-year AGN turn-ons traced to collapsing inner disk","Flux reveals hot inner disk collapse behind fast AGNs","Magnetic field can't explain rapid AGN turn-ons; disk collapse can","Rapid AGN flickers solved: inner ADAF collapses to thin disk","A 200-day AGN turn-on explained by revised disk model"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the only thing changing between the dim and bright states is where the hot inner disk begins, with no other source of 5100 Å light and no change in how much gas falls in; if starlight, the inner hot flow, or a changing accretion rate shifts the balance, the predicted timescales change substantially.","fun_headline_variants_meta":{"raw":{"variants":["Sub-year AGN turn-ons traced to collapsing inner disk","Flux reveals hot inner disk collapse behind fast AGNs","Magnetic field can't explain rapid AGN turn-ons; disk collapse can","Rapid AGN flickers solved: inner ADAF collapses to thin disk","A 200-day AGN turn-on explained by revised disk model"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000418,"raw_usage":{"total_tokens":2239,"prompt_tokens":1115,"completion_tokens":1124,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":731,"completion_tokens_details":{"reasoning_tokens":1031}},"tokens_in":731,"tokens_out":1124,"duration_ms":9588,"temperature":1.0,"reasoning_tokens":1031,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T15:10:10.136320+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A high-cadence spectroscopic campaign that caught either object completing its transition in well under 254 or 142 days rest frame, or a spectral decomposition showing that the dim-state 5100 Å continuum contains a substantial component from the inner hot flow or host stars rather than only the outer thin disk, would settle the claim against the model.","supporting_citations":[],"review_version":1}