Pith. sign in

REVIEW 2 major objections 6 minor 5 cited by

The paper argues that the absence of variability in little red dots is not a mystery to be explained away but a predicted signature of super-Eddington accretion: standard sub-Eddington AGN variability models, applied to the same observing c

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Super-Eddington accretion models can explain why little red dots show almost no variability, whereas standard sub-Eddington AGN variability models predict changes that should already have been seen.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection A useful, transparent model comparison that makes the super-Eddington explanation for LRD non-variability plausible, though the optical PSD extrapolation is a real weakness. the 2 major comments →

arxiv 2509.03571 v1 pith:NIYET2CF submitted 2025-09-03 astro-ph.GA

Do Little Red Dots Vary?

Amy Secunda , Rachel S. Somerville , Yan-Fei Jiang , Jenny E. Greene , Lukas J. Furtak , Adi Zitrin This is my paper
classification astro-ph.GA
keywords little red dotssuper-Eddington accretionAGN variabilitydamped random walkradiation MHD simulationsJWSThigh-redshift AGNbroad line variability
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks why little red dots — the compact, red, V-shaped-spectrum objects JWST finds at high redshift, thought to be active galactic nuclei — barely change in brightness even when observed several times. The authors generated mock light curves with two models: the standard empirical 'damped random walk' description of variability calibrated on lower-redshift, slowly accreting (sub-Eddington) AGN, and a model built from radiation magnetohydrodynamic simulations of disks accreting a few times faster than the Eddington limit, the rate at which radiation pressure would balance gravity. With the sub-Eddington model, a third to a half of already-observed LRDs should have shown clear variability; none did. With the super-Eddington model, fewer than 1% of simulated light curves show detectable changes at current cadences, matching the observations. If the paper is right, the stillness of LRDs is expected for super-Eddington accretion rather than evidence against black-hole power, and two ongoing JWST campaigns, NEXUS and TWINKLE, can decide between the two models.

Core claim

The paper's central claim is that moderately super-Eddington accretion can easily account for the observed lack of variability in little red dots, while standard sub-Eddington AGN variability models cannot. For the sub-Eddington case the authors adopt the empirical damped-random-walk prescription from lower-redshift AGN studies and generate thousands of mock light curves sampled at the cadences of real observations; these predict detectable (>3σ) magnitude changes in 33–47% of the LRDs already studied, yet none of those objects show such changes. For the super-Eddington case they build a power-law power spectral density by extrapolating radiation MHD simulations from the inner disk (soft X-r

What carries the argument

The load-bearing object is the power spectral density (PSD) of the disk's light curves. For the super-Eddington case, the authors fit power laws PSD ∝ C₀ ν⁻ᵝ to radiation MHD simulations of the inner disk — the soft X-ray region at ~10 r_g (β = 2.4) and the UV region at <200 r_g (β = 2.6) — and logarithmically extrapolate the fitted slope and amplitude to the optical photosphere at 10^4 r_g, obtaining β = 3.0 and C₀ = 5×10⁻⁹. That steep, low-amplitude PSD places essentially all optical variability at timescales of years or longer, far beyond the month-scale baselines of current and NEXUS observations. The contrasting machinery is the damped random walk (PSD ∝ ν⁻² above a damping timescale),

Load-bearing premise

The entire super-Eddington prediction rests on one long extrapolation: the variability scaling measured between 10 and 200 r_g in the simulations is assumed to hold unchanged out to the optical emitting region at 10^4 r_g, a step the paper itself flags in Section 4 as the model's main uncertainty.

What would settle it

The NEXUS campaign results: the paper's super-Eddington model predicts that only ~12% of mock light curves reach Δm > 0.21 mag over its three-year baseline, while the sub-Eddington model predicts 97–100% do. If NEXUS detects strong continuum variability in a large majority of LRDs, the super-Eddington model fails; if nearly all stay quiet, the sub-Eddington model fails. A second, more immediate check is the paper's own alternative scaling (constant variability amplitude with timescales growing as r^1.5), which predicts magnitude changes two orders of magnitude smaller still and could be tested

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If LRDs are moderately super-Eddington, their lack of variability is expected, and the AGN interpretation does not need exotic non-AGN alternatives to explain the quiet light curves.
  • The NEXUS campaign will discriminate: large magnitude changes (Δm > 1) in many LRDs would support sub-Eddington models, while near-universal quiescence, with only ~12% of sources above the 3σ threshold, supports super-Eddington accretion.
  • TWINKLE can detect broad emission line variability even when the continuum stays constant, provided soft X-ray emission from the inner disk reaches the broad line region — an indirect probe of X-rays that are not directly detected.
  • Rest-frame multi-year baselines, such as ten years of observer-frame monitoring or lensed LRDs with long time delays, should eventually reveal super-Eddington variability because the steep PSD predicts variability on few-year timescales.
  • The rare LRDs that do vary (for instance the two sources with correlated multi-band variability in the largest compiled sample) are consistent with the super-Eddington model's predicted small fraction of large excursions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • A consequence the paper leaves implicit: variability-selected searches for high-redshift AGN are biased toward sub-Eddington accretors, so a predominantly super-Eddington population would be systematically underrepresented in such surveys.
  • The inner-to-outer disk PSD extrapolation, if validated by NEXUS, could serve as a template for predicting variability in other X-ray-weak, gas-enshrouded AGN scenarios such as black-hole-star and quasi-star models.
  • A testable extension available right now: stacking all multi-epoch JWST photometry of LRDs to check whether the two-epoch magnitude-difference distribution is concentrated below ~0.1 mag with a thin large-variation tail, exactly as the super-Eddington model predicts.
  • If TWINKLE sees line variability in objects where NEXUS sees no continuum variability, that combination would constrain the irradiation geometry — soft X-rays reaching the broad line region while the disk photosphere stays shielded — a joint constraint no single observation currently provides.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 6 minor

Summary. The paper investigates whether the observed lack of variability in JWST-discovered little red dots (LRDs) can be explained by super-Eddington accretion. The authors generate mock light curves using two models: (i) a Damped Random Walk (DRW) model calibrated to lower-redshift sub-Eddington AGN (Burke et al. 2021, 2023) and (ii) a novel super-Eddington model in which the optical PSD is constructed by extrapolating radiation-MHD simulation results from the inner disk (Jiang et al. 2019, 2025). Comparing with existing observations (T24, KH24), they find that the DRW model predicts detectable variability in 33-47% of mock light curves, while none is observed; the super-Eddington model predicts <1% detections for current KH24-like cadences and 12% for the planned NEXUS survey. They also predict that the TWINKLE campaign will detect broad-line variability if a soft X-ray driver reaches the broad-line region. The paper concludes that super-Eddington accretion can naturally account for the lack of observed continuum variability in LRDs.

Significance. If the central prediction can be made robust, this paper offers a timely and physically motivated resolution to one of the key puzzles of LRD research: the absence of variability despite AGN-like spectra and compactness. Its strengths are the use of state-of-the-art radiation-MHD simulations to construct a variability model, the explicit comparison with published observational limits, and the provision of falsifiable predictions for the ongoing NEXUS and TWINKLE campaigns. The main weakness is that the super-Eddington optical PSD rests on a two-decade radial extrapolation from only two simulation anchors; the claimed non-detection fractions depend sensitively on this extrapolation. The qualitative conclusion that super-Eddington disks are less variable than DRW models is likely robust, but the quantitative consistency with current non-detections needs additional support.

major comments (2)
  1. [Section 2.2.2, Figure 1] The super-Eddington optical PSD is constructed by logarithmic extrapolation of beta and C0 from two inner-disk simulations (<10 r_g and 50-200 r_g) to 10^4 r_g. The two anchors already differ in slope (beta=2.4 vs 2.6) and normalization (C0=1e-6 vs 1e-7), so the extrapolated values (beta=3.0, C0=5e-9) have no direct support in the optical region. Section 4 acknowledges this but the alternative test (constant amplitude, tau proportional to r^3/2) only reduces variability further and does not bound the opposite direction, where a steeper or higher-normalized PSD at 10^4 r_g could raise the predicted detection fractions. Since the claimed <1% (KH24) and 12% (NEXUS) detection rates are controlled by these two numbers, the central quantitative claim is not yet robust. I recommend either obtaining an outer-disk anchor (e.g., from simulations at intermediate radii or from empirical constraints
  2. [Section 2.1, Eq. (1)] The DRW detection fractions depend on the assumed Eddington ratio L/L_Edd = 0.1 used to compute Mi from black hole masses. LRD Eddington ratios are uncertain and may be biased high if the objects are super-Eddington; the Burke et al. relation has a non-negligible dependence of SF_infinity on Mi (coefficient C=0.131). The paper should show how the predicted detection fractions (33-47% for T24/KH24) change when this assumption is varied over a plausible range (e.g., 0.01 to 1, or higher). The qualitative contrast with the super-Eddington model would likely survive, but the quantitative rejection of sub-Eddington models is only as strong as this assumption.
minor comments (6)
  1. [Title] The title contains a typo: 'V ary' should be 'Vary'.
  2. [Section 2.2.2] The term 'structure function' is used for the PSD normalization C0, but C0 is not the structure function; please clarify the connection or use consistent terminology.
  3. [Section 2.2.2] The implementation of the low-frequency break is not fully described. Setting the light-curve length to the damping timescale in the Timmer & Koenig (1995) method effectively imposes a low-frequency cutoff, but this is not equivalent to a bending-power-law PSD with a break; please specify the exact procedure used.
  4. [Figure 2 caption] It would be helpful to explicitly state the assumed 3-sigma detection limit (Delta m = 0.21 mag) in the caption, and to differentiate it from the higher T24 limit shown as the thin gray line.
  5. [Section 3.1.1] The sentence 'both studies do not observe significant variability for any LRDs' is slightly ambiguous; T24 reported a mean Delta m that is not significant, but individual objects may show larger changes. Please clarify.
  6. [Methods (global)] For reproducibility, please provide a table of the input parameters used for the mock light curves (masses, redshifts, observing cadences, number of realizations) and consider releasing the analysis code or a data repository.

Circularity Check

0 steps flagged

No significant circularity; the super-Eddington and DRW models are independently calibrated and compared to, not fitted from, the LRD variability data.

full rationale

The paper's derivation chain is self-contained with respect to circularity. The two model families are constructed from independent inputs: (1) the sub-Eddington DRW model uses empirical scaling relations from Burke et al. (2021, 2023) fitted to lower-redshift AGN, and (2) the super-Eddington model uses PSDs directly from radiation MHD simulations (Jiang et al. 2019, 2025). Neither model is fitted to the LRD variability non-detections; the KH24, T24, Zhang et al. (2024), and Furtak et al. (2025) data are used only as comparison points after mock light curves are generated. The optical PSD in Section 2.2.2 is obtained by logarithmic extrapolation from the simulated soft X-ray and UV PSDs, which is an openly stated modeling assumption rather than a fit to the target observable; the paper even acknowledges in Section 4 that this assumes the inner–outer disk scaling remains constant and tests a constant-amplitude alternative. Self-citations to Secunda et al. (2025) and Jiang et al. are used to justify physical inputs (e.g., weak X-ray irradiation and simulation-derived variability), but these are independent simulation results, not restatements of the paper's conclusion, and no uniqueness theorem or forbidden-alternative argument is invoked. The central claim that super-Eddington accretion can explain the lack of LRD variability is a genuine model prediction, not a tautology or a renamed fit. The main weakness is the large radial extrapolation, but that is a scientific uncertainty, not a circularity.

Axiom & Free-Parameter Ledger

4 free parameters · 6 axioms · 0 invented entities

No new physical entities are introduced; super-Eddington accretion is an existing physical regime. The model depends on inherited empirical relations and simulation outputs, plus several domain assumptions about LRD geometry and radiation transport.

free parameters (4)
  • Assumed Eddington ratio for DRW LRD luminosity = 0.1
    Assumed typical value because bolometric luminosities of LRDs are uncertain; affects SF_infinity through Mi in Eq. 1.
  • Super-Eddington optical PSD slope beta_opt = 3.0
    Logarithmically extrapolated from fits to soft X-ray (beta=2.4) and UV (beta=2.6) simulation PSDs in Section 2.2.2.
  • Super-Eddington optical PSD normalization C0_opt = 5e-9
    Extrapolated from C0=1e-6 (soft X-ray) and C0=1e-7 (UV) fits; sets the variability amplitude.
  • Super-Eddington damping timescale = 4e4 days
    Derived from the UV bending timescale (~100 days) scaled by r^(3/2) to the optical radius; caps long-timescale variability in the model.
axioms (6)
  • domain assumption LRD variability, if AGN-like, follows a damped random walk with the Burke et al. (2021, 2023) scaling relations extrapolated to high redshift
    The entire sub-Eddington null hypothesis rests on applying empirical lower-z AGN variability relations to LRDs; this extrapolation is unproven.
  • domain assumption Jiang et al. (2019, 2025) radiation MHD simulations capture the variability of super-Eddington AGN disks
    The super-Eddington model is built entirely on these simulations, with no independent observational calibration.
  • domain assumption Optical variability in super-Eddington disks is driven by local intrinsic disk fluctuations, not X-ray irradiation
    Stated in Section 2.2.1 and justified by Secunda et al. (2025); if X-rays reach the UV-optical region, variability could be higher.
  • domain assumption UV emission from the AGN in super-Eddington LRDs is absorbed or scattered, so no UV variability is observable
    Section 2.2.2 assumes UV is suppressed; if UV is not fully suppressed, the comparison with UV observations changes.
  • ad hoc to paper The PSD scaling fitted between 10 r_g and 50-200 r_g extends logarithmically to 10^4 r_g
    This is the central extrapolation of the super-Eddington model, acknowledged as a caveat in Section 4.
  • domain assumption BLR reprocessing is modeled by a Gaussian response function with width 5-100 days
    Standard in the literature, but the width is uncertain and affects the detectability of line variability.

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Do Little Red Dots Vary?." pith.science (2026). https://pith.science/paper/NIYET2CF

@misc{pith2026250903571,
  author       = {Pith},
  title        = {Pith review of: Do Little Red Dots Vary?},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NIYET2CF}},
  note         = {Machine review of arXiv:2509.03571}
}
Share X Bluesky LinkedIn Reddit HN
abstract

Little red dots (LRDs), high-redshift, compact, red objects with V-shaped spectra, are one of the most exciting and perplexing discoveries made by the James Webb Space Telescope (JWST). While the simplest explanation for LRDs is that they are high redshift active galactic nuclei (AGN), due to their compactness and frequent association with broad line emission, the lack of corresponding X-ray emission and observed variability cast doubt on this picture. Here, we simulate LRD light curves using both traditional models for sub-Eddington AGN variability derived empirically from lower-redshift AGN observations and moderately super-Eddington AGN disk models from radiation magnetohydrodynamic simulations to examine the reason for the lack of variability. We find that even though most LRDs have only been observed 2--4 times in a given waveband, we should still be detecting significantly more variability if traditional sub-Eddington AGN variability models can be applied to LRDs. Instead, our super-Eddington model light curves are consistent with the lack of observed LRD variability. In addition, the ongoing high-cadence {\sc nexus} campaign will detect changes in magnitude, $\Delta m>1$, for traditional sub-Eddington models, but will only observe significant continuum variability for the lowest mass LRDs for our super-Eddington AGN models. Even if LRDs lack continuum variability, we find that the ongoing spectroscopic JWST campaign {\sc twinkle} should observe broad emission line variability as long as soft X-ray irradiation manages to reach the broad line region from the inner disk. Our models show that super-Eddington accretion can easily explain the lack of continuum variability in LRDs.

Figures

Figures reproduced from arXiv: 2509.03571 by Adi Zitrin, Amy Secunda, Jenny E. Greene, Lukas J. Furtak, Rachel S. Somerville, Yan-Fei Jiang.

Figure 1
Figure 1. Figure 1: Left panel: The frequency-binned variability PSD of the UV light curves emitted from < 200 rg for three different athena++ simulations from Jiang et al. (2025) labeled by their mass accretion rate. Right panel: The frequency-binned variability PSD of the soft X-ray (in yellow, emitted from ∼ 10 rg) and UV (in purple, emitted from < 200 rg) light curves from the simulations in Jiang et al. (2019) and Jiang … view at source ↗
Figure 2
Figure 2. Figure 2: The mean maximum difference in magnitude, ∆m = max(m) − min(m), between all epochs for 1000 simulated observer-frame F115-band (F356-band) light curves in pink (green) mock observed with different previous and ongoing LRD observation cadences. Closed circles represent ⟨∆m⟩ for our sub-Eddington DRW model light curves and triangles represent ⟨∆m⟩ for our super-Eddington model light curves. Error bars show t… view at source ↗
Figure 3
Figure 3. Figure 3: The distribution of the percent change in broad line flux for 300 mock twinkle light curves made assum￾ing three different driving continuum light curves. The pink distribution is for an X-ray driving light curve simulated us￾ing empirical models for lower redshift sub-Eddington AGN. The yellow and purple distributions are for soft X-ray and UV driving light curves, respectively, simulated using our best-f… view at source ↗
Figure 4
Figure 4. Figure 4: The maximum difference in magnitude, ∆m = max(m) − min(m), between all epochs as a function of mass for 1000 simulated observer-frame F356W-band light curves. Filled circles show the mean ∆m for mock sub￾Eddington DRW model nexus light curves. The triangles (squares) show the mean ∆m we estimate for our super￾Eddington model nexus (KH24) light curves using the scal￾ing SF∞ ∝ M−0.15 from Equation 1 for a co… view at source ↗
Figure 5
Figure 5. Figure 5: Example rest-frame optical light curves for a 108 M⊙ AGN at z = 6 generated using our sub-Eddington DRW model (top panel) and our super-Eddington model (bottom panel). We also show the observed times for an example KH24 and nexus light curve starting at t = 0 for this mock light curve as the blue and orange dots, respectively. The sub-Eddington light curve has significantly more variability over this rest-… view at source ↗
Figure 6
Figure 6. Figure 6: The fractional change in flux as a function of rest-frame time, F(t)/⟨F(t)⟩, for three example light curves for our three different models for reprocessed broad line emission for a 108 M⊙, z = 5.3 AGN. The difference between the three models is the underlying driving light curve which from left to right is a sub-Eddington model hard X-ray driver, super-Eddington model soft X-ray driver, and super-Eddington… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. A new sample of Little Red Dots at $z<0.45$ in DESI DR1: Broad Balmer lines, low ionization spectrum and no variability

    astro-ph.GA 2026-05 conditional novelty 7.0

    Eight low-redshift Little Red Dots identified in DESI DR1 exhibit broad Balmer lines, steep decrements, compact shapes, and negligible variability, with a number density roughly 10,000 times lower than at z>4.

  2. ATLAS. II. Extremely High Incidence of Balmer Line Absorption with Predominant Blueshifts in LRDs: Statistical Insights through Comparison with Type 1 AGNs

    astro-ph.GA 2026-07 conditional novelty 6.0

    Balmer-line absorption occurs in ~35% (14/40) of JWST little-red-dot AGNs, roughly 850x the rate in SDSS type-1 AGNs, with mostly slow blueshifted absorber velocities.

  3. NEXUS: A Search for Nuclear Variability with the First Two JWST NIRCam Epochs

    astro-ph.GA 2025-09 conditional novelty 6.0

    Using two JWST NIRCam epochs, difference imaging finds 465 nuclear variable sources and sets tight F444W variability upper limits of 3 to 10 percent for ten Little Red Dots.

  4. You can't see me: Super-Eddington growth hindering X-ray detection in high-z broad-line active galactic nuclei

    astro-ph.GA 2026-02 conditional novelty 5.0

    High-redshift JWST broad-line AGNs may be low-mass black holes accreting far above Eddington, whose steep, over-cooled coronal spectra explain their X-ray non-detections and make them appear overmassive.

  5. Unveil the nature of JWST-AGN and Little Red Dots with SKAO continuum surveys

    astro-ph.GA 2026-06 unverdicted novelty 3.0

    SKAO continuum surveys will detect radio emission from JWST AGN and LRDs and distinguish between Compton-thick absorption, intrinsically weak accretion, and dense gas cocoon scenarios.

Reference graph

Works this paper leans on

81 extracted references · 3 canonical work pages · cited by 5 Pith papers

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter doi edition editor eprint howpublished institution journal key month number organization pages publisher school series title misctitle type volume year version url label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts ...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION format.url url empty "" new.block "" url * "" * if FUNCTION format.eprint eprint empty "" archivePrefix empty "" archivePrefix "arXiv" = new.block " " eprint * " " * new.block " " eprint * " " * if if if FUNCTION format.doi doi empty "" " " doi * " " * if FUNCTION format.pid doi empty eprint empty ur...

  3. [3]

    Frequency-resolved lags in UV/optical continuum reverberation mapping

    thebibliography [1] 20pt to REFERENCES 6pt =0pt -12pt 10pt plus 3pt =0pt =0pt =1pt plus 1pt =0pt =0pt -12pt =13pt plus 1pt =20pt =13pt plus 1pt \@M =10000 =-1.0em =0pt =0pt 0pt =0pt =1.0em @enumiv\@empty 10000 10000 `\.\@m \@noitemerr \@latex@warning Empty `thebibliography' environment \@ifnextchar \@reference \@latexerr Missing key on reference command E...

    Pith/arXiv arXiv 2021

  4. [4]

    B., Casey , C

    Akins , H. B., Casey , C. M., Lambrides , E., et al. 2024, arXiv e-prints, arXiv:2406.10341, 10.48550/arXiv.2406.10341

    Pith/arXiv arXiv doi:10.48550/arxiv.2406.10341 2024

  5. [5]

    T., Bogd \'a n , \'A ., Kov \'a cs , O

    Ananna , T. T., Bogd \'a n , \'A ., Kov \'a cs , O. E., Natarajan , P., & Hickox , R. C. 2024, , 969, L18, 10.3847/2041-8213/ad5669

    doi:10.3847/2041-8213/ad5669 2024

  6. [6]

    G., Kocevski , D

    Barro , G., P \'e rez-Gonz \'a lez , P. G., Kocevski , D. D., et al. 2024, , 963, 128, 10.3847/1538-4357/ad167e

    doi:10.3847/1538-4357/ad167e 2024

  7. [7]

    C., & Dexter , J

    Begelman , M. C., & Dexter , J. 2025, arXiv e-prints, arXiv:2507.09085, 10.48550/arXiv.2507.09085

    doi:10.48550/arxiv.2507.09085 2025

  8. [8]

    2025, arXiv e-prints, arXiv:2501.03309, 10.48550/arXiv.2501.03309

    Bellovary , J. 2025, arXiv e-prints, arXiv:2501.03309, 10.48550/arXiv.2501.03309

    Pith/arXiv arXiv doi:10.48550/arxiv.2501.03309 2025

  9. [9]

    D., & McKee , C

    Blandford , R. D., & McKee , C. F. 1982, , 255, 419, 10.1086/159843

    doi:10.1086/159843 1982

  10. [10]

    J., Shen , Y., Liu , X., et al

    Burke , C. J., Shen , Y., Liu , X., et al. 2023, , 518, 1880, 10.1093/mnras/stac2478

    doi:10.1093/mnras/stac2478 2023

  11. [11]

    J., Shen , Y., Blaes , O., et al

    Burke , C. J., Shen , Y., Blaes , O., et al. 2021, Science, 373, 789, 10.1126/science.abg9933

    doi:10.1126/science.abg9933 2021

  12. [12]

    M., Akins , H

    Casey , C. M., Akins , H. B., Kokorev , V., et al. 2024, , 975, L4, 10.3847/2041-8213/ad7ba7

    doi:10.3847/2041-8213/ad7ba7 2024

  13. [13]

    C., Li , R., & Inayoshi , K

    Chen , C.-H., Ho , L. C., Li , R., & Inayoshi , K. 2025, , 989, L12, 10.3847/2041-8213/adee0a

    doi:10.3847/2041-8213/adee0a 2025

  14. [14]

    2023, , 675, A163, 10.1051/0004-6361/202345844

    Czerny , B., Panda , S., Prince , R., et al. 2023, , 675, A163, 10.1051/0004-6361/202345844

    doi:10.1051/0004-6361/202345844 2023

  15. [15]

    2018, , 856, 6, 10.3847/1538-4357/aaae6b

    Du , P., Zhang , Z.-X., Wang , K., et al. 2018, , 856, 6, 10.3847/1538-4357/aaae6b

    doi:10.3847/1538-4357/aaae6b 2018

  16. [16]

    J., Zitrin , A., Plat , A., et al

    Furtak , L. J., Zitrin , A., Plat , A., et al. 2023, , 952, 142, 10.3847/1538-4357/acdc9d

    doi:10.3847/1538-4357/acdc9d 2023

  17. [17]

    J., Labb \'e , I., Zitrin , A., et al

    Furtak , L. J., Labb \'e , I., Zitrin , A., et al. 2024, , 628, 57, 10.1038/s41586-024-07184-8

    doi:10.1038/s41586-024-07184-8 2024

  18. [18]

    J., Secunda , A

    Furtak , L. J., Secunda , A. R., Greene , J. E., et al. 2025, arXiv e-prints, arXiv:2502.07875, 10.48550/arXiv.2502.07875

    Pith/arXiv arXiv doi:10.48550/arxiv.2502.07875 2025

  19. [19]

    2021, , 508, 3463, 10.1093/mnras/stab2818

    Georgakakis , A., Papadakis , I., & Paolillo , M. 2021, , 508, 3463, 10.1093/mnras/stab2818

    doi:10.1093/mnras/stab2818 2021

  20. [20]

    2012, , 544, A80, 10.1051/0004-6361/201219008

    Gonz \'a lez-Mart \' n , O., & Vaughan , S. 2012, , 544, A80, 10.1051/0004-6361/201219008

    doi:10.1051/0004-6361/201219008 2012

  21. [21]

    E., Labb\' e , I., Goulding , A

    Greene , J. E., Labb\' e , I., Goulding , A. D., et al. 2024, , 964, 39, 10.3847/1538-4357/ad1e5f

    doi:10.3847/1538-4357/ad1e5f 2024

  22. [22]

    J., Peterson , B

    Grier , C. J., Peterson , B. M., Horne , K., et al. 2013, , 764, 47, 10.1088/0004-637X/764/1/47

    doi:10.1088/0004-637x/764/1/47 2013

  23. [23]

    M., et al

    Horne , K., De Rosa , G., Peterson , B. M., et al. 2021, , 907, 76, 10.3847/1538-4357/abce60

    doi:10.3847/1538-4357/abce60 2021

  24. [24]

    2022, , 934, 132, 10.3847/1538-4357/ac75d8

    Hu , H., Inayoshi , K., Haiman , Z., Quataert , E., & Kuiper , R. 2022, , 934, 132, 10.3847/1538-4357/ac75d8

    doi:10.3847/1538-4357/ac75d8 2022

  25. [25]

    E., de Graaff , A., Miller , T

    Hviding , R. E., de Graaff , A., Miller , T. B., et al. 2025, arXiv e-prints, arXiv:2506.05459, 10.48550/arXiv.2506.05459

    doi:10.48550/arxiv.2506.05459 2025

  26. [26]

    2024, arXiv e-prints, arXiv:2412.03653, 10.48550/arXiv.2412.03653

    Inayoshi , K., Kimura , S., & Noda , H. 2024, arXiv e-prints, arXiv:2412.03653, 10.48550/arXiv.2412.03653

    Pith/arXiv arXiv doi:10.48550/arxiv.2412.03653 2024

  27. [27]

    2024, arXiv e-prints, arXiv:2409.07805

    Inayoshi , K., & Maiolino , R. 2024, arXiv e-prints, arXiv:2409.07805. 2409.07805

    Pith/arXiv arXiv 2024

  28. [28]

    K., Prince , R., Panda , S., & Czerny , B

    Jaiswal , V. K., Prince , R., Panda , S., & Czerny , B. 2023, , 670, A147, 10.1051/0004-6361/202244352

    doi:10.1051/0004-6361/202244352 2023

  29. [29]

    2025, arXiv e-prints, arXiv:2501.13082, 10.48550/arXiv.2501.13082

    Ji , X., Maiolino , R., \"U bler , H., et al. 2025, arXiv e-prints, arXiv:2501.13082, 10.48550/arXiv.2501.13082

    doi:10.48550/arxiv.2501.13082 2025

  30. [30]

    2021, , 253, 49, 10.3847/1538-4365/abe303

    Jiang , Y.-F. 2021, , 253, 49, 10.3847/1538-4365/abe303

    doi:10.3847/1538-4365/abe303 2021

  31. [31]

    2025, arXiv e-prints, arXiv:2505.09671

    Jiang , Y.-F., Blaes , O., Kaul , I., & Zhang , L. 2025, arXiv e-prints, arXiv:2505.09671. 2505.09671

    Pith/arXiv arXiv 2025

  32. [32]

    2024, arXiv e-prints, arXiv:2408.16856, 10.48550/arXiv.2408.16856

    Jiang , Y.-F., & Dai , L. 2024, arXiv e-prints, arXiv:2408.16856, 10.48550/arXiv.2408.16856

    Pith/arXiv arXiv doi:10.48550/arxiv.2408.16856 2024

  33. [33]

    M., & Davis , S

    Jiang , Y.-F., Stone , J. M., & Davis , S. W. 2019, , 880, 67, 10.3847/1538-4357/ab29ff

    doi:10.3847/1538-4357/ab29ff 2019

  34. [34]

    C., Bechtold , J., & Siemiginowska , A

    Kelly , B. C., Bechtold , J., & Siemiginowska , A. 2009, , 698, 895, 10.1088/0004-637X/698/1/895

    doi:10.1088/0004-637x/698/1/895 2009

  35. [35]

    Kido , D., Ioka , K., Hotokezaka , K., Inayoshi , K., & Irwin , C. M. 2025, arXiv e-prints, arXiv:2505.06965, 10.48550/arXiv.2505.06965

    doi:10.48550/arxiv.2505.06965 2025

  36. [36]

    R., & Pounds , K

    King , A. R., & Pounds , K. A. 2003, , 345, 657, 10.1046/j.1365-8711.2003.06980.x

    arXiv 2003

  37. [37]

    R., et al

    Kobayashi , H., Ohsuga , K., Takahashi , H. R., et al. 2018, , 70, 22, 10.1093/pasj/psx157

    doi:10.1093/pasj/psx157 2018

  38. [38]

    D., Onoue , M., Inayoshi , K., et al

    Kocevski , D. D., Onoue , M., Inayoshi , K., et al. 2023, , 954, L4, 10.3847/2041-8213/ace5a0

    doi:10.3847/2041-8213/ace5a0 2023

  39. [39]

    D., Finkelstein , S

    Kocevski , D. D., Finkelstein , S. L., Barro , G., et al. 2024, arXiv e-prints, arXiv:2404.03576, 10.48550/arXiv.2404.03576

    Pith/arXiv arXiv doi:10.48550/arxiv.2404.03576 2024

  40. [40]

    2023, , 957, L7, 10.3847/2041-8213/ad037a

    Kokorev , V., Fujimoto , S., Labb\' e , I., et al. 2023, , 957, L7, 10.3847/2041-8213/ad037a

    doi:10.3847/2041-8213/ad037a 2023

  41. [41]

    I., Greene , J

    Kokorev , V., Caputi , K. I., Greene , J. E., et al. 2024, , 968, 38, 10.3847/1538-4357/ad4265

    doi:10.3847/1538-4357/ad4265 2024

  42. [42]

    2024, arXiv e-prints, arXiv:2407.04777, 10.48550/arXiv.2407.04777

    Kokubo , M., & Harikane , Y. 2024, arXiv e-prints, arXiv:2407.04777, 10.48550/arXiv.2407.04777

    doi:10.48550/arxiv.2407.04777 2024

  43. [43]

    E., Matthee , J., et al

    Labb\' e , I., Greene , J. E., Matthee , J., et al. 2024, arXiv e-prints, arXiv:2412.04557, 10.48550/arXiv.2412.04557

    Pith/arXiv arXiv doi:10.48550/arxiv.2412.04557 2024

  44. [44]

    E., Bezanson , R., et al

    Labb\' e , I., Greene , J. E., Bezanson , R., et al. 2025, , 978, 92, 10.3847/1538-4357/ad3551

    doi:10.3847/1538-4357/ad3551 2025

  45. [45]

    2024, arXiv e-prints, arXiv:2409.13047

    Lambrides , E., Garofali , K., Larson , R., et al. 2024, arXiv e-prints, arXiv:2409.13047. 2409.13047

    Pith/arXiv arXiv 2024

  46. [46]

    E., & Ma , Y

    Liu , H., Jiang , Y.-F., Quataert , E., Greene , J. E., & Ma , Y. 2025, arXiv e-prints, arXiv:2507.07190, 10.48550/arXiv.2507.07190

    doi:10.48550/arxiv.2507.07190 2025

  47. [47]

    2024, , 689, A128, 10.1051/0004-6361/202451249

    Lupi , A., Trinca , A., Volonteri , M., Dotti , M., & Mazzucchelli , C. 2024, , 689, A128, 10.1051/0004-6361/202451249

    doi:10.1051/0004-6361/202451249 2024

  48. [48]

    E., Setton , D

    Ma , Y., Greene , J. E., Setton , D. J., et al. 2025, , 981, 191, 10.3847/1538-4357/ada613

    doi:10.3847/1538-4357/ada613 2025

  49. [49]

    L., Ivezi \'c , Z ., Kochanek , C

    MacLeod , C. L., Ivezi \'c , Z ., Kochanek , C. S., et al. 2010, , 721, 1014, 10.1088/0004-637X/721/2/1014

    doi:10.1088/0004-637x/721/2/1014 2010

  50. [50]

    2025, arXiv e-prints, arXiv:2501.09854, 10.48550/arXiv.2501.09854

    Madau , P. 2025, arXiv e-prints, arXiv:2501.09854, 10.48550/arXiv.2501.09854

    doi:10.48550/arxiv.2501.09854 2025

  51. [51]

    2024, , 976, L24, 10.3847/2041-8213/ad90e1

    Madau , P., & Haardt , F. 2024, , 976, L24, 10.3847/2041-8213/ad90e1

    doi:10.3847/2041-8213/ad90e1 2024

  52. [52]

    2025, , 538, 1921, 10.1093/mnras/staf359

    Maiolino , R., Risaliti , G., Signorini , M., et al. 2025, , 538, 1921, 10.1093/mnras/staf359

    doi:10.1093/mnras/staf359 2025

  53. [53]

    P., Brammer , G., et al

    Matthee , J., Naidu , R. P., Brammer , G., et al. 2024, , 963, 129, 10.3847/1538-4357/ad2345

    doi:10.3847/1538-4357/ad2345 2024

  54. [54]

    M., Ar \'e valo , P., Uttley , P., et al

    McHardy , I. M., Ar \'e valo , P., Uttley , P., et al. 2007, , 382, 985, 10.1111/j.1365-2966.2007.12411.x

    arXiv 2007

  55. [55]

    M., Papadakis , I

    McHardy , I. M., Papadakis , I. E., Uttley , P., Page , M. J., & Mason , K. O. 2004, , 348, 783, 10.1111/j.1365-2966.2004.07376.x

    arXiv 2004

  56. [56]

    P., Matthee , J., Katz , H., et al

    Naidu , R. P., Matthee , J., Katz , H., et al. 2025, arXiv e-prints, arXiv:2503.16596, 10.48550/arXiv.2503.16596

    Pith/arXiv arXiv doi:10.48550/arxiv.2503.16596 2025

  57. [57]

    2005, , 628, 368, 10.1086/430728

    Ohsuga , K., Mori , M., Nakamoto , T., & Mineshige , S. 2005, , 628, 368, 10.1086/430728

    doi:10.1086/430728 2005

  58. [58]

    2024, , 976, 96, 10.3847/1538-4357/ad84f7

    Pacucci , F., & Narayan , R. 2024, , 976, 96, 10.3847/1538-4357/ad84f7

    doi:10.3847/1538-4357/ad84f7 2024

  59. [59]

    E., Brandt , W

    Paolillo , M., Papadakis , I. E., Brandt , W. N., et al. 2023, , 673, A68, 10.1051/0004-6361/202245291

    doi:10.1051/0004-6361/202245291 2023

  60. [60]

    E., Brinkmann , W., Negoro , H., & Gliozzi , M

    Papadakis , I. E., Brinkmann , W., Negoro , H., & Gliozzi , M. 2002, , 382, L1, 10.1051/0004-6361:20011763

    doi:10.1051/0004-6361:20011763 2002

  61. [61]

    M., Ferrarese , L., Gilbert , K

    Peterson , B. M., Ferrarese , L., Gilbert , K. M., et al. 2004, , 613, 682, 10.1086/423269

    doi:10.1086/423269 2004

  62. [62]

    2025, arXiv e-prints, arXiv:2505.09669

    Sacchi , A., & Bogdan , A. 2025, arXiv e-prints, arXiv:2505.09669. 2505.09669

    Pith/arXiv arXiv 2025

  63. [63]

    2015, , 447, 49, 10.1093/mnras/stu2387

    Sadowski , A., Narayan , R., Tchekhovskoy , A., et al. 2015, , 447, 49, 10.1093/mnras/stu2387

    doi:10.1093/mnras/stu2387 2015

  64. [64]

    Secunda , A., Jiang , Y.-F., & Greene , J. E. 2025, , 984, 19, 10.3847/1538-4357/adc25b

    work page doi:10.3847/1538-4357/adc25b 2025

  65. [65]

    J., Greene , J

    Setton , D. J., Greene , J. E., de Graaff , A., et al. 2024, arXiv e-prints, arXiv:2411.03424, 10.48550/arXiv.2411.03424

    Pith/arXiv arXiv doi:10.48550/arxiv.2411.03424 2024

  66. [66]

    J., Greene , J

    Setton , D. J., Greene , J. E., Spilker , J. S., et al. 2025, arXiv e-prints, arXiv:2503.02059, 10.48550/arXiv.2503.02059

    Pith/arXiv arXiv doi:10.48550/arxiv.2503.02059 2025

  67. [67]

    I., & Sunyaev , R

    Shakura , N. I., & Sunyaev , R. A. 1973, , 24, 337

    1973

  68. [68]

    2024, arXiv e-prints, arXiv:2408.12713, 10.48550/arXiv.2408.12713

    Shen , Y., Zhuang , M.-Y., Li , J., et al. 2024, arXiv e-prints, arXiv:2408.12713, 10.48550/arXiv.2408.12713

    Pith/arXiv arXiv doi:10.48550/arxiv.2408.12713 2024

  69. [69]

    J., et al

    Stone , Z., Shen , Y., Burke , C. J., et al. 2022, , 514, 164, 10.1093/mnras/stac1259

    doi:10.1093/mnras/stac1259 2022

  70. [70]

    R., Shen , Y., et al

    Sun , M., Trump , J. R., Shen , Y., et al. 2015, , 811, 42, 10.1088/0004-637X/811/1/42

    work page doi:10.1088/0004-637x/811/1/42 2015

  71. [71]

    L., Fan , X., Wang , F., & Yang , J

    Tee , W. L., Fan , X., Wang , F., & Yang , J. 2024, arXiv e-prints, arXiv:2412.05242, 10.48550/arXiv.2412.05242

    Pith/arXiv arXiv doi:10.48550/arxiv.2412.05242 2024

  72. [72]

    1995, , 300, 707

    Timmer , J., & Koenig , M. 1995, , 300, 707

    work page 1995

  73. [73]

    2025, arXiv e-prints, arXiv:2505.09542, 10.48550/arXiv.2505.09542

    Torralba , A., Matthee , J., Pezzulli , G., et al. 2025, arXiv e-prints, arXiv:2505.09542, 10.48550/arXiv.2505.09542

    doi:10.48550/arxiv.2505.09542 2025

  74. [74]

    L., et al

    Wang , B., de Graaff , A., Davies , R. L., et al. 2024, arXiv e-prints, arXiv:2403.02304, 10.48550/arXiv.2403.02304

    Pith/arXiv arXiv doi:10.48550/arxiv.2403.02304 2024

  75. [75]

    C., Alberts , S., Ji , Z., et al

    Williams , C. C., Alberts , S., Ji , Z., et al. 2024, , 968, 34, 10.3847/1538-4357/ad3f17

    doi:10.3847/1538-4357/ad3f17 2024

  76. [76]

    2018, , 865, L8, 10.3847/2041-8213/aadf8f

    Xiao , M., Du , P., Lu , K.-K., et al. 2018, , 865, L8, 10.3847/2041-8213/aadf8f

    doi:10.3847/2041-8213/aadf8f 2018

  77. [77]

    T., Vogeley , M

    Yu , W., Richards , G. T., Vogeley , M. S., Moreno , J., & Graham , M. J. 2022, , 936, 132, 10.3847/1538-4357/ac8351

    doi:10.3847/1538-4357/ac8351 2022

  78. [78]

    T., et al

    Yue , M., Eilers , A.-C., Ananna , T. T., et al. 2024, arXiv e-prints, arXiv:2404.13290, 10.48550/arXiv.2404.13290

    Pith/arXiv arXiv doi:10.48550/arxiv.2404.13290 2024

  79. [79]

    M., Mullen , P

    Zhang , L., Stone , J. M., Mullen , P. D., et al. 2025, arXiv e-prints, arXiv:2506.02289, 10.48550/arXiv.2506.02289

    Pith/arXiv arXiv doi:10.48550/arxiv.2506.02289 2025

  80. [80]

    Zhang , Z., Jiang , L., Liu , W., & Ho , L. C. 2024, arXiv e-prints, arXiv:2411.02729, 10.48550/arXiv.2411.02729

    Pith/arXiv arXiv doi:10.48550/arxiv.2411.02729 2024

Showing first 80 references.

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.