{"id":"5002aebd-cd64-454e-b3d8-939b165761d7","arxiv_id":"2501.13174","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Contemporaneous UV and optical spectra of three changing-look quasars show that variable dust extinction cannot explain their state transitions, favoring an intrinsic drop in accretion power.","lead":"New ultraviolet and optical spectra of three changing-look quasars, caught in their dim state, were compared with archival bright-state spectra. The data rule out dust blocking as the cause of the state changes and instead point to a drop in the central engine's brightness.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The dust-extinction rejection is robust, but the positive LLAGN/accretion-flow interpretation rests on qualitative SED comparison with SNR~2 UV data and is not quantitatively established.","rationale":"After reading the paper carefully, I find the dust-extinction test sound. The cross-method A_V inconsistencies are not merely self-consistency failures; the ordering of inferred effective extinctions (H-alpha dimming more than the 5100 Angstrom continuum) is impossible for any foreground dust law, so the negative claim is robust even if the high state is reddened. My main concern is the positive interpretation: the LLAGN-like SED and accretion-flow-structure conclusion relies on visual similarity and low-SNR UV data, with quantitative SED fitting explicitly deferred. This is a genuine missing support, but it does not break the paper's central negative claim; it limits how strongly the positive claim can be stated. The reader's CONDITIONAL verdict already captures this, so I recommend no change to the verdict.","tokens_in":26569,"tokens_out":17381,"duration_ms":180785,"concrete_test":"Fit the binned low-state UV-optical SEDs of all three quasars with (a) an ADAF + truncated thin disk model from Nemmen et al. (2014) and (b) a standard thin-disk model with foreground reddening, using the tabulated binned fluxes and realistic errors that include calibration offsets between the HST gratings. Compare models with a goodness-of-fit statistic such as BIC. If model (b) is preferred, or model (a) cannot reproduce the UV measurements within 2 sigma, then the LLAGN interpretation fails and the 'change in accretion flow structure' claim should be weakened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim has two parts. The first, that variable dust obscuration alone cannot explain the state changes, is well supported: for all three quasars, the effective extinction required by the broad H-alpha flux drop exceeds that required by the 5100 Angstrom continuum drop (e.g., J1011: A_V = 4.4 from H-alpha versus 2.6 from the continuum; Table 5), which violates the wavelength ordering of any foreground dust screen. This conclusion does not depend on the high state being unreddened, because the comparison is differential and adding a constant high-state extinction cannot fix the ordering. The load-bearing weakness is in the second part. The claim that the low-state SEDs are 'reminiscent of LLAGN' and that the transitions reflect a change in accretion flow structure rests on a qualitative visual comparison with LLAGN templates (Section 5.2, Figures 8-10), not on a quantitative fit. The HST UV continuum has SNR ~2 in the G140L band (Section 3.1), the G230L spectra show no discernible emission lines (Section 3.5), and Section 7 explicitly defers detailed SED modelling to future work. If the apparent UV turnover and lack of a UV bump are not statistically significant, the SEDs could be consistent with other models (e.g., a reddened standard disk or a low-luminosity thin disk), leaving the accretion-flow-structure interpretation unsupported. This does not threaten the dust rejection, but it is the least secure pillar of the abstract's conclusion.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports contemporaneous HST/COS ultraviolet and HET/LRS2 optical spectroscopy of three changing-look quasars (J1011, J1021, J2336) obtained in their low states, together with archival SDSS high- and low-state spectra. The authors decompose the spectra, measure broad H-alpha, H-beta, and Ly-alpha fluxes and 5100 Angstrom continuum levels, and use the high-to-low state changes to derive V-band extinctions under four different assumptions. Because the resulting A_V values disagree strongly among methods for all three objects, they conclude that variable dust obscuration cannot be the sole cause of the changing-look transitions. They also construct low-state UV-to-X-ray SEDs and note the absence of the canonical UV bump, a peak at log nu ~ 15-15.2, and qualitative resemblance to LLAGN SEDs, which they interpret as evidence for a transition to a radiatively inefficient accretion flow driven by a reduced Eddington ratio.","tokens_in":26844,"tokens_out":4634,"duration_ms":49320,"significance":"If the dust-extinction rejection holds, this is a valuable contribution: it is the first test of the variable-obscuration hypothesis for CLQs using contemporaneous rest-frame UV and optical spectra, and the multi-method A_V comparison is a well-designed differential test. The paper also provides useful new low-state measurements of three previously identified CLQs, including UV line fluxes that are rarely available in the low state. The bootstrap-resampled fitting and the explicit exploration of different extinction/attenuation laws are commendable. The second part of the claim, concerning accretion-flow structure, is considerably less secure, as it rests on qualitative SED comparison rather than quantitative modeling; the paper itself defers detailed SED fitting to future work.","major_comments":[{"comment":"The abstract and Section 6 state that the low-state SEDs are 'reminiscent of' LLAGN and that the transition 'may result from a change in accretion flow structure caused by a reduced Eddington ratio,' and Section 7 lists this as a main conclusion. However, the evidence presented is a qualitative visual comparison with the Nemmen et al. (2014) ADAF+truncated-disk model fits, with no quantitative fitting, no goodness-of-fit statistics, and no comparison against alternative models such as a reddened standard disk or a low-luminosity thin disk. Section 7 explicitly states that 'detailed SED modelling and fit LLAGN-type SEDs to the observed data' remains future work. Because the accretion-flow conclusion is a load-bearing part of the paper's central claim, it needs either to be backed by quantitative model fitting or to be explicitly downgraded to a hypothesis rather than presented as a finding.","section":"Section 5.2, Figures 8-10; Section 6; Section 7"},{"comment":"The UV data used to define the low-state SED shapes have very low signal-to-noise: Section 3.1 reports a mean continuum SNR of approximately 2 in the G140L band at 1350 Angstroms, and Section 3.5 states that no emission lines are discernible in the G230L spectra except C IV in J1021. The SED bins in Figure 8 carry error bars that are the standard deviation of pixel values within each bin. The claims that the SEDs 'lack the canonical UV bump' and peak near log nu ~ 15-15.2 therefore need a statistical justification, for example a test that the observed UV-optical continuum is inconsistent with a single power law or with a standard thin-disk SED, or at least a propagation of the SNR into the binned fluxes. Without such a test, the apparent UV turnover and the absence of a UV bump may not be significant, weakening the LLAGN interpretation.","section":"Section 3.1, Section 3.5, Section 5.2, Figures 8-9"},{"comment":"The derivation of A_V from the emission lines in Table 5 assumes that the high-state H-alpha/H-beta ratio and the high-state broad-line fluxes are the unreddened intrinsic values. This is stated explicitly in Section 5.1. If the high state is itself reddened, the absolute A_V values in Table 5 shift, and the 'presence of Ly-alpha alone' argument is priors-dependent through the assumed intrinsic Ly-alpha/H-alpha = 2-5 range. The qualitative conclusion that dust alone cannot explain the transitions is nevertheless likely robust, because the different methods then measure (A_V_low - A_V_high) and the cross-method disagreement persists. The manuscript would be stronger if it stated this differential interpretation explicitly and demonstrated, even with a simple algebraic example, that a constant high-state extinction cannot reconcile the different A_V values. As written, the text invites the reader to think the central conclusion depends on the unreddened-high-state assumption.","section":"Section 5.1, Table 5"}],"minor_comments":[{"comment":"The title contains 'T rio' with a stray space; please correct to 'Trio'.","section":"Abstract/Title"},{"comment":"The COS entrance aperture is described as having a 'redius' of 1.25 arcseconds; this should be 'radius'.","section":"Section 2.3"},{"comment":"The caption contains 'spectruum'; it should read 'spectrum'.","section":"Figure 7 caption"},{"comment":"The summary says the SEDs are 'remniscent' of LLAGN; the correct spelling is 'reminiscent'.","section":"Section 7"},{"comment":"The text refers to 'telluric Na I D emission'; telluric features of this kind are absorption, not emission, so the wording should be adjusted.","section":"Section 4.2"}],"recommendation":"major_revision","confidential_remarks":"The dust-extinction test is well executed and is likely to be a useful contribution to the CLQ literature. The main risk to the paper is that the abstract presents the LLAGN/accretion-flow interpretation as a result when the evidence is only qualitative, and the UV SNR is marginal. In revision, the authors should either add a quantitative SED analysis or soften the conclusions to match what the data can support. The paper fits the journal's scope and the central dust-rejection claim is defensible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this paper does a solid job of killing the simple dust-obscuration explanation for three changing-look quasars, using a nicely designed consistency test on new contemporaneous UV and optical spectra. The claim that the low states resemble LLAGN is more suggestive than established.\n\nWhat's new: they obtained HST COS UV spectra and HET optical spectra of three CLQs in their low states, close in time, and combined these with archival high/low-state SDSS spectra. That contemporaneous UV+optical low-state coverage is the missing ingredient in earlier tests, and they use it well. The spectral decomposition with pPXF, including quasar power-law, Fe II, and Balmer continuum templates, is careful, and they propagate uncertainties via bootstrap.\n\nThe core result is the A_V inconsistency. For each object, they derive extinction from broad H-alpha flux, H-alpha/H-beta ratio, Ly-alpha/H-alpha, and the 5100 A continuum. These disagree wildly (e.g., J1011 gives A_V=4.4 from H-alpha versus 2.6 from the continuum; Ly-alpha/H-alpha gives ~0.6). That ordering is exactly backwards for a foreground dust screen, since shorter wavelengths should suffer more extinction. The stress-test note is right that this conclusion survives even if the high state is reddened, because the comparison is differential. So the negative claim is solid.\n\nThe softer spot is the second half: the low-state SEDs are described as 'reminiscent of LLAGN' and interpreted as evidence for a transition to an ADAF+truncated disk. This rests on a visual comparison with LLAGN templates, not a fit, and the UV continuum has SNR~2. The paper acknowledges this and defers detailed SED modeling to future work, but the abstract leans on it. It's an interesting hypothesis, not yet a finding. A referee should push them to either qualify it more strongly or add a quantitative comparison with extinction-reddened standard disk models.\n\nMinor issues: the sample is three objects, all previously identified by the same group, but the central test is object-level, so that's not a real weakness. The assumption that the high state is unreddened is untested, but doesn't break the dust rejection.\n\nThis paper deserves a serious referee. The data are real, the analysis is careful, and the dust test is a meaningful step for the CLQ mechanism debate. The LLAGN interpretation needs tempering or quantitative backing, but that's a revision-level issue, not a reject.\n\nRecommendation: send to peer review.","headline":"A carefully executed multi-wavelength test convincingly rules out dust as the sole cause of three CLQ transitions; the LLAGN comparison is suggestive but not yet a result.","tokens_in":27415,"tokens_out":4733,"would_cite":true,"duration_ms":46933,"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":"For three changing-look quasars, variable dust obscuration is strongly disfavored; their dramatic fading instead reflects a change in the central engine's intrinsic luminosity.","keywords":["changing-look quasars","active galactic nuclei","dust extinction","ultraviolet spectroscopy","optical spectroscopy","spectral energy distribution","accretion flow","Eddington ratio"],"falsifier":"A high-state ultraviolet spectrum of any of these three quasars that showed an intrinsic Ly-alpha/H-alpha ratio near unity, rather than the assumed 2–5, would bring the ultraviolet-derived extinction into agreement with the optical values and would invalidate the dust-rejection for that object.","tokens_in":26368,"feed_emoji":"🔭","tokens_out":12073,"duration_ms":106961,"temperature":0.7,"pith_summary":"This paper tests the two leading explanations for changing-look quasars, active galaxies whose broad emission lines and continuous spectrum fade dramatically within a few years. Using newly obtained ultraviolet and optical spectra taken while three quasars were in their faint state, together with archival spectra from brighter and fainter epochs, the authors measure the dust extinction implied by four independent spectral tracers. Because those tracers give mutually inconsistent extinction values, they conclude that a single variable dust screen cannot account for the state change and that the transformation instead reflects a change in the intrinsic luminosity of the central engine. The low-state spectra also lack the usual ultraviolet bump and resemble the spectral energy distributions of low-luminosity active galactic nuclei, suggesting a drop in Eddington ratio and a change in the structure of the accretion flow.","feed_headline":"Changing-look quasars' sudden dimming traced to engine, not dust","feed_subtitle":"Contemporaneous UV and optical spectra rule out a single dust screen and point to a fading accretion flow.","key_machinery":"The central mechanism is the dust extinction test: under a fixed extinction law, each independent reddening tracer—the Balmer decrement, the absolute H-alpha and H-beta flux drops, the Ly-alpha/H-alpha ratio, and the 5100 Å continuum dimming—predicts its own V-band extinction, and variable-obscuration is viable only if those predictions agree. The test uses the Cardelli et al. (1989) extinction law with RV = 3.1, and brackets the ultraviolet tracers with a steeper Small Magellanic Cloud law and a greyer starburst attenuation law. The companion piece is the low-state SED comparison: the observed SEDs are matched in shape to advection-dominated accretion flow (ADAF) plus truncated thin disk models for low-luminosity active galactic nuclei (Nemmen et al. 2014), where a hot, radiatively inefficient flow plus a truncated thin disk shifts the peak emission from the far-ultraviolet to the infrared/optical and removes the ultraviolet bump.","core_discovery":"For three changing-look quasars (J1011, J1021, and J2336), the authors combined new low-state ultraviolet and optical spectra with archival spectra from both the high and low states. Taking the high-state broad-line fluxes and the H-alpha/H-beta ratio as the intrinsic, unreddened values, they derived V-band extinction from the Balmer decrement, the drop in broad H-alpha flux, the drop in broad H-beta flux, the Ly-alpha/H-alpha ratio (assuming an intrinsic ratio of 2–5), and the drop in the 5100 Å continuum. In all three quasars these values are mutually inconsistent: for J1011 even the optical line measurements alone disagree, and for J1021 and J2336 the ultraviolet- and continuum-derived values fall well below the optical-line values. The authors therefore conclude that variable dust obscuration alone cannot explain the state changes and that the observed transformation reflects a change in the intrinsic luminosity of the central engine. They further find that the low-state spectral energy distributions peak at longer wavelengths than standard quasar SEDs, lack the ultraviolet bump, and resemble the SEDs of low-luminosity active galactic nuclei, which is consistent with a change from a thin accretion disk to an advection-dominated flow with a truncated disk as the Eddington ratio drops.","pith_inferences":["If the low states truly correspond to advection-dominated accretion, X-ray observations taken during a state transition should show spectral hardening alongside the optical and ultraviolet decline; this is a testable prediction of the paper's interpretation.","The same cross-tracer extinction test could be applied to larger changing-look quasar samples using only optical line ratios plus a continuum measurement, since the paper shows that the 5100 Å continuum tracer alone can expose a dust-screen inconsistency.","The near-identical low-state SED shapes of J1011 and J2336 suggest that at similar low Eddington ratios the accretion flow settles into a similar configuration; a larger sample could test whether SED shape is determined by Eddington ratio alone.","Future high-state ultraviolet spectra of these quasars would directly measure the intrinsic Ly-alpha/H-alpha ratio that the paper had to assume, which would sharpen or overturn the dust rejection for individual objects."],"forward_implications":["For these three quasars, variable dust obscuration alone cannot explain the state change, so the transformation most plausibly reflects an intrinsic change in the quasar's luminosity.","In the low state, the quasars' spectral energy distributions resemble those of low-luminosity active galactic nuclei, implying a drop in Eddington ratio and a possible transition from a thin disk to an advection-dominated flow.","Bolometric corrections calibrated on standard quasars and on low-luminosity active galactic nuclei give mostly consistent Eddington ratios, so existing luminosity estimates remain usable even when a quasar is in its low state.","Two of the three quasars continued to slowly brighten after entering their low state, indicating that the low state is not static and that continued monitoring of confirmed changing-look quasars can reveal further evolution."],"supporting_citations":[{"why":"Supplies the extinction law (with RV = 3.1) that converts every measured flux ratio into an implied AV.","marker":"Cardelli et al. (1989)"},{"why":"Provides the composite quasar spectrum and the adopted intrinsic Ly-alpha/H-alpha ratio of ~2–5 used in the ultraviolet extinction test.","marker":"Vanden Berk et al. (2001)"},{"why":"Presents the ADAF plus truncated thin disk SED models that the low-state SEDs are compared against.","marker":"Nemmen et al. (2014)"},{"why":"Supplies the low-state X-ray fluxes used to compute alpha_ox and the shape-dependent Eddington ratios.","marker":"Ruan et al. (2019)"},{"why":"Reported J1011's changing-look discovery and supplies its black hole mass and high-state properties.","marker":"Runnoe et al. (2016)"},{"why":"Reported J1021's changing-look discovery and supplies its black hole mass and high-state properties.","marker":"MacLeod et al. (2016)"},{"why":"Reported J2336's changing-look discovery and supplies its black hole mass and high-state properties.","marker":"Ruan et al. (2016)"},{"why":"Supplies the radio-quiet and radio-loud composite quasar SEDs whose shapes the low-state SEDs are shown not to match.","marker":"Elvis et al. (1994)"},{"why":"Provides the X-ray-to-optical shape-dependent bolometric correction used for the low-state Eddington ratios.","marker":"Lusso et al. (2010)"}],"fun_headline_variants":["Quasar dimming blamed on engine, not dust, in trio study","Dust ruled out for three changing-look quasars","Changing-look quasars trace to fading engine, not dimming dust","Trio of quasars: intrinsic fade, not dust, causes state shift","UV/optical data rule out dust for three changing-look quasars"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The dust test assumes that the high-state broad-line fluxes and the H-alpha/H-beta ratio are the intrinsic, unreddened values, so that any change in the low state is attributed to dust; if the high state is itself reddened, the derived extinction values shift.","fun_headline_variants_meta":{"raw":{"variants":["Quasar dimming blamed on engine, not dust, in trio study","Dust ruled out for three changing-look quasars","Changing-look quasars trace to fading engine, not dimming dust","Trio of quasars: intrinsic fade, not dust, causes state shift","UV/optical data rule out dust for three changing-look quasars"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000163,"raw_usage":{"total_tokens":1271,"prompt_tokens":1001,"completion_tokens":270,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":617,"completion_tokens_details":{"reasoning_tokens":176}},"tokens_in":617,"tokens_out":270,"duration_ms":4119,"temperature":1.0,"reasoning_tokens":176,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T16:23:28.023075+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A high-state ultraviolet spectrum of any of these three quasars that showed an intrinsic Ly-alpha/H-alpha ratio near unity, rather than the assumed 2–5, would bring the ultraviolet-derived extinction into agreement with the optical values and would invalidate the dust-rejection for that object.","supporting_citations":[],"review_version":1}