pith. machine review for the scientific record. sign in

arxiv: 2604.23916 · v1 · submitted 2026-04-27 · 🌌 astro-ph.HE

Recognition: unknown

GRRMHD Simulations of State Transitions in Non-Jetted Tidal Disruption Events

Brandon Curd , Safira Heridia , Aviyel Ahiyya , Richard Anantua
Authors on Pith no claims yet

Pith reviewed 2026-05-08 02:19 UTC · model grok-4.3

classification 🌌 astro-ph.HE
keywords tidal disruption eventsthermal instabilityaccretion disk collapseX-ray spectrablack hole spinGRRMHD simulationsstate transitionscooling envelope model
0
0 comments X

The pith

GRRMHD simulations show TDE disks become thermally unstable in 17-46 days, causing X-ray luminosity to drop by nearly 100 times.

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

The paper performs general relativistic radiation magnetohydrodynamics simulations of magnetized tori adapted from the late, near-Eddington phase of the cooling envelope model for a solar-mass star disrupted by a ten-million-solar-mass black hole. It establishes that these disks grow thermally unstable after 17 to 46 days, with the timescale set by the black hole spin. Before collapse the thermal spectra develop a soft X-ray excess, after which the X-ray luminosity falls by almost two orders of magnitude. The blackbody radius and temperature track the spin, and the overall spectral evolution matches the observed non-jetted TDE AT2021ehb. The work therefore supplies a disk-physics account of state transitions in events that lack jets.

Core claim

In the cooling envelope model the circularized debris reaches a shallow density profile and near-Eddington accretion only after several months. The GRRMHD runs of the corresponding magnetized tori show that the disk becomes thermally unstable within 17.1-46.5 days. Collapse is preceded by a soft X-ray excess in the thermal spectrum and is followed by a nearly two-order-of-magnitude decline in X-ray luminosity. Blackbody radius and temperature evolve in a spin-dependent manner, and the simulated spectra and soft X-ray luminosities reproduce those of AT2021ehb.

What carries the argument

General relativistic radiation magnetohydrodynamics simulations of magnetized tori taken from the near-Eddington stage of the cooling envelope model, used to evolve the thermal stability and radiative spectra of the accretion flow.

If this is right

  • X-ray luminosity falls by nearly two orders of magnitude once the disk collapses.
  • A soft X-ray excess appears in the spectrum immediately before the instability develops.
  • Blackbody radius and temperature change in a manner that depends on black hole spin.
  • The resulting spectra and luminosities are similar to those measured for AT2021ehb.

Where Pith is reading between the lines

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

  • Thermal instability in the disk itself may be the dominant cause of state transitions in non-jetted TDEs.
  • Multi-epoch observations of many TDEs could test whether measured transition times scale with black hole spin as the models predict.
  • The same instability may appear in other accretion flows that develop shallow density profiles and strong radiative cooling.

Load-bearing premise

The chosen initial magnetized torus faithfully represents the late-stage debris structure from the cooling envelope model without a full self-consistent calculation from the original stellar disruption.

What would settle it

An observed non-jetted TDE that maintains steady high X-ray luminosity without a soft excess or sharp drop for more than 50 days after reaching near-Eddington accretion rates.

Figures

Figures reproduced from arXiv: 2604.23916 by Aviyel Ahiyya, Brandon Curd, Richard Anantua, Safira Heridia.

Figure 1
Figure 1. Figure 1: A zoomed in view of the accretion flow and funnel of model m7a0-M22 at ∆t = 0, 11.4, 22.8, and 31.4 days (increasing from left to right). The colors show the gas density (top), gas temperature (middle), and gas to radiation pressure ratio (bottom). Streamlines indicate the fluid velocity. The ISCO radius is indicated as the white circle. The disk height visibly decreases over time. We also note that the ga… view at source ↗
Figure 2
Figure 2. Figure 2: Here we show the mass accretion rate (top, 1st column), luminosity (top, 2nd column), total efficiency (top, 3rd column), mass outflow rate at 500rg (top, 4th column), density scale height (bottom, 1st column), density weighted disk temperature (bottom, 2nd column), mean temperature of the corona (bottom, 3rd column), and mean Compton cooling rate (bottom, 4th column). In the top left panel, we show the Ed… view at source ↗
Figure 3
Figure 3. Figure 3: Here we show the mass accretion rate without smoothing view at source ↗
Figure 5
Figure 5. Figure 5: Here we show the accretion disk after thermal collapse for each model. The colors show the gas density, streamlines indicate the fluid velocity, and the ISCO radius is indicated as the white circle. The disk is observed to truncate near the ISCO in each model. are shown in view at source ↗
Figure 6
Figure 6. Figure 6: Here we show the radiation temperature (colors), radially integrated photosphere location (yellow line), and ISCO (white circle) for model m7a0-M22. The top panels shows a zoomed out view at 11.4 days (left) and 31.4 days (right). Note that different ranges for the coordinates are used to emphasize the temperature at the outer photosphere surface. The bottom panels show the same times but zoomed in to high… view at source ↗
Figure 7
Figure 7. Figure 7: Here we show the spectrum for each BH spin before disk collapse (∆t = 5.7 days, top) and after disk collapse (∆t = ∆tfinal, bottom). Note θ = 0◦ (viewer at +z) is shown as the solid lines while θ = 90◦ (viewer at +x) is shown as the dashed lines. Viewing angle effects lead to obscuration of X-ray photons when the disk is viewed edge-on (θ = 90◦). Note that the a• = −0.9 model has the weakest soft X-ray emi… view at source ↗
Figure 9
Figure 9. Figure 9: Here we show the spectral properties for all models for an observer at θ = 0◦. The a• = 0.9 model shows the brightest soft X-ray emission, but all models have a similar value for αOSX prior to disk collapse. All models show rapid OUV and X-ray variability prior to collapse. We indicate LOUV/LX = 1 (dashed black line) and LOUV/LX = 0.1 (dash-dotted black line) in the bottom panel for ease of comparison view at source ↗
Figure 10
Figure 10. Figure 10: Here we zoom in on the band luminosities for all models in the last 5 days for an observer at θ = 0◦. luminosity exiting the emission surface (dLr or dLϑ) directly from the koral data to normalize the distribution. Since it is possible for observers viewing the system from near edge on angles to receive emission from parts of the disk where the funnel boundary faces them, it is necessary to per￾form the e… view at source ↗
Figure 11
Figure 11. Figure 11: Here we show the blackbody fits for radius (top) and temperature (middle) in the optical/UV band (green) and soft X￾ray band (blue). The viewing angle is θ = 0◦. We show the range of values obtained with our luminosity constraints for the optical/UV fits with the green shaded region. The X-ray blackbody radius appears to approach the ISCO radius (black dashed line) as the disk collapses. In the bottom pan… view at source ↗
Figure 12
Figure 12. Figure 12: Here we show the blackbody fits for a• = −0.9 (top) and a• = 0.9 (bottom). The viewing angle is θ = 0◦. radius and temperature illustrate that X-ray fluctuations arise due to rapid variation in the photosphere scale and tem￾perature. Furthermore, the temperature of the optical/UV photosphere is nearly constant. In fact, the temperature is more representative of the koral data as the disk begins to collaps… view at source ↗
Figure 13
Figure 13. Figure 13: Here we show the properties of the blackbody fits for the soft X-ray as a function of viewing angle. The ISCO radius is indicated with a dashed horizontal line for each value of BH spin. Note that subscript ’b’ and subscript ’a’ define quantities as before and after disk collapse. More precisely, quantities with subscript ’b’ are averaged over 0 < ∆t ≤ ∆tfinal − 5 days and subscript ’a’ are the maximum ra… view at source ↗
read the original abstract

Circularization of the stream material into a debris cloud during tidal disruption events (TDEs) was recently demonstrated in one of the most accurate long duration TDE simulations to-date. The cooling envelope model (CEM) provides a description of the circularized debris cloud and its emission over time well beyond circularization across different disruption parameters. In the CEM, sub-Eddington accretion rates occur early in TDEs and the debris has a shallow density profile of roughly $\rho \propto r^{-1}$, with Eddington accretion only being achieved after several months. To explore the late stages of the CEM, we perform general relativistic radiation magnetohydrodynamics (GRRMHD) simulations of magnetized tori adapted from the near Eddington phase of the CEM for a $1M_\odot$ star disrupted around a $10^7 M_\odot$ black hole (BH). We find that the disk becomes thermally unstable within 17.1-46.5 days depending on the spin of the BH. Thermal spectra show a soft X-ray excess prior to collapse, with a nearly two order of magnitude decline in X-ray luminosity upon disk collapse. Furthermore, the evolution of the blackbody radius and temperature of our models are correlated with the spin of the black hole. The spectral properties and soft X-ray luminosity in our models are similar to the TDE AT2021ehb, which is a non-jetted TDE with late X-rays and a state transition after $\approx 271$ days.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper performs GRRMHD simulations of magnetized tori adapted from the near-Eddington phase of the cooling envelope model (CEM) for a 1 M⊙ star disrupted by a 10^7 M⊙ black hole. It reports that the disk undergoes thermal instability and collapse within 17.1–46.5 days (depending on BH spin), producing a soft X-ray excess prior to collapse followed by a nearly two-order-of-magnitude drop in X-ray luminosity. The blackbody radius and temperature evolution correlate with spin, and the spectral properties plus soft X-ray luminosities are stated to be similar to the observed non-jetted TDE AT2021ehb (which shows a state transition at ~271 days).

Significance. If the reported instability is robust, the work supplies a concrete numerical pathway from the CEM's late-stage debris to observed TDE state transitions, including spin-dependent timing and the characteristic soft X-ray excess plus luminosity drop. This strengthens the physical interpretation of events like AT2021ehb and supplies falsifiable predictions for blackbody evolution versus spin.

major comments (3)
  1. [Numerical Setup] Numerical Setup section: No grid resolution, convergence tests, or quantitative error estimates are reported for the GRRMHD runs. Because the central result is the spin-dependent thermal instability onset time (17.1–46.5 days) and subsequent collapse, the absence of these diagnostics leaves open the possibility that the instability is sensitive to numerical dissipation or resolution choices rather than being a physical outcome of the ideal MHD + radiative cooling setup.
  2. [Initial Conditions] Initial Conditions / §3: The tori are described as 'adapted from' the CEM near-Eddington phase with the ρ ∝ r^{-1} profile. The manuscript must detail the precise mapping procedure for density, angular momentum, entropy, and embedded magnetic field, and demonstrate that this adaptation preserves the thermal instability threshold; otherwise the reported spin dependence and collapse times cannot be confidently attributed to the CEM debris rather than to the adaptation choices.
  3. [Results] Results / Comparison to AT2021ehb: The simulated collapse occurs 17.1–46.5 days after the start of the near-Eddington phase, while the observed transition in AT2021ehb is at ~271 days. The paper should explicitly map the simulation start time onto the observational timeline (accounting for the months of sub-Eddington evolution in the CEM) and quantify whether the soft X-ray luminosity and spectral shape match at corresponding epochs; without this, the claimed similarity remains qualitative and the timescale discrepancy is unaddressed.
minor comments (2)
  1. [Abstract] Abstract: 'nearly two order of magnitude decline' should read 'nearly two orders of magnitude decline'.
  2. [Figures] Figure captions and axis labels should explicitly state the time origin (days after simulation start or after disruption) and the spin values used for each curve to improve readability.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their thorough and constructive comments on our manuscript. We have carefully considered each point and provide detailed responses below. We believe these revisions will strengthen the paper and address the concerns raised.

read point-by-point responses

  1. Referee: [Numerical Setup] Numerical Setup section: No grid resolution, convergence tests, or quantitative error estimates are reported for the GRRMHD runs. Because the central result is the spin-dependent thermal instability onset time (17.1–46.5 days) and subsequent collapse, the absence of these diagnostics leaves open the possibility that the instability is sensitive to numerical dissipation or resolution choices rather than being a physical outcome of the ideal MHD + radiative cooling setup.

    Authors: We acknowledge the importance of demonstrating numerical robustness for our key results on the thermal instability. In the revised manuscript, we will add details on the grid resolution employed in the GRRMHD simulations, including the number of zones in each dimension. Additionally, we will include convergence tests by running lower-resolution simulations and comparing the onset times of the instability and collapse. Quantitative error estimates on the collapse times will be provided based on these tests. This will confirm that the reported spin-dependent times (17.1–46.5 days) are physical and not artifacts of numerical choices. revision: yes

  2. Referee: [Initial Conditions] Initial Conditions / §3: The tori are described as 'adapted from' the CEM near-Eddington phase with the ρ ∝ r^{-1} profile. The manuscript must detail the precise mapping procedure for density, angular momentum, entropy, and embedded magnetic field, and demonstrate that this adaptation preserves the thermal instability threshold; otherwise the reported spin dependence and collapse times cannot be confidently attributed to the CEM debris rather than to the adaptation choices.

    Authors: We agree that a more precise description of the initial condition adaptation is necessary. In the revised version, we will expand §3 to detail the exact mapping procedure, specifying how the density profile (ρ ∝ r^{-1}), specific angular momentum distribution, entropy, and the initial magnetic field configuration (including the plasma beta) are extracted and interpolated from the CEM model. We will also include a comparison of key quantities such as the thermal instability criterion (e.g., the cooling function and heating balance) between the original CEM and our adapted tori to show that the threshold is preserved. This will strengthen the attribution of the results to the physical CEM debris evolution. revision: yes

  3. Referee: [Results] Results / Comparison to AT2021ehb: The simulated collapse occurs 17.1–46.5 days after the start of the near-Eddington phase, while the observed transition in AT2021ehb is at ~271 days. The paper should explicitly map the simulation start time onto the observational timeline (accounting for the months of sub-Eddington evolution in the CEM) and quantify whether the soft X-ray luminosity and spectral shape match at corresponding epochs; without this, the claimed similarity remains qualitative and the timescale discrepancy is unaddressed.

    Authors: We appreciate this point on connecting the simulation timeline to observations. In the revised manuscript, we will add an explicit mapping of the simulation start time to the observational timeline for AT2021ehb. This will account for the several months of sub-Eddington evolution in the CEM prior to reaching the near-Eddington phase where our simulations begin. We will include a discussion and possibly a schematic figure illustrating the total time from disruption to the state transition, showing how the 17.1–46.5 days in simulation correspond to the observed ~271 days. Furthermore, we will quantify the soft X-ray luminosities and spectral shapes at corresponding epochs, providing a more detailed comparison beyond the qualitative similarity stated in the abstract. This will address the timescale discrepancy by clarifying the phase alignment. revision: yes

Circularity Check

0 steps flagged

Simulation dynamics generate instability times and spectral evolution; CEM adaptation is cited prior work but does not reduce claims to self-definition or fitted inputs.

full rationale

The paper's derivation consists of running GRRMHD simulations on initial tori adapted from the CEM, then extracting instability onset (17.1-46.5 days), soft X-ray excess, and luminosity drop as direct outputs of the evolved MHD + cooling equations. These quantities are not predefined by the initial conditions or by any fit to the target observables; the spin dependence and AT2021ehb comparison function as post-simulation diagnostics against external data. The CEM citation supplies the starting density profile but is not invoked as a uniqueness theorem or ansatz that forces the reported collapse behavior. No equation in the provided text equates a 'prediction' to a fitted parameter or renames an input as an output. This yields only minor self-citation load (score 2) without load-bearing circularity in the central claims.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the cooling envelope model providing accurate initial conditions for the near-Eddington torus and on standard assumptions of ideal MHD plus radiative transfer in GRRMHD; no new entities are postulated.

free parameters (1)
  • Black hole spin parameter
    Multiple spin values are explored to demonstrate dependence of instability time and spectral evolution.
axioms (1)
  • domain assumption The cooling envelope model accurately describes the circularized debris at the near-Eddington phase used for initial conditions.
    Initial torus is adapted directly from the CEM without re-deriving the early circularization phase.

pith-pipeline@v0.9.0 · 5581 in / 1390 out tokens · 50896 ms · 2026-05-08T02:19:05.423911+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

57 extracted references · 55 canonical work pages

  1. [1]

    1995, ApJL, 438, L37, doi: 10.1086/187709

    Abramowicz M. A., Chen X., Kato S., Lasota J.-P., Regev O., 1995, @doi [ ] 10.1086/187709 , https://ui.adsabs.harvard.edu/abs/1995ApJ...438L..37A 438, L37

    work page doi:10.1086/187709 1995

  2. [2]

    C., 2025, @doi [ ] 10.3847/1538-4357/adf217 , https://ui.adsabs.harvard.edu/abs/2025ApJ...993...14A 993, 14

    Alush Y., Stone N. C., 2025, @doi [ ] 10.3847/1538-4357/adf217 , https://ui.adsabs.harvard.edu/abs/2025ApJ...993...14A 993, 14

    work page doi:10.3847/1538-4357/adf217 2025

  3. [3]

    A., Narayan R., Sadowski A., 2017, @doi [ ] 10.1093/mnras/stx1345 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.470.2367C 470, 2367

    Chael A. A., Narayan R., S a dowski A., 2017, @doi [ ] 10.1093/mnras/stx1345 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.470.2367C 470, 2367

    work page doi:10.1093/mnras/stx1345 2017

  4. [4]

    Curd B., 2021, @doi [ ] 10.1093/mnras/stab2172 , https://ui.adsabs.harvard.edu/abs/2021MNRAS.507.3207C 507, 3207

    work page doi:10.1093/mnras/stab2172 2021

  5. [5]

    Curd B., Narayan R., 2019, @doi [ ] 10.1093/mnras/sty3134 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.483..565C 483, 565

    work page doi:10.1093/mnras/sty3134 2019

  6. [6]

    Curd B., Narayan R., 2023, @doi [ ] 10.1093/mnras/stac3330 , https://ui.adsabs.harvard.edu/abs/2023MNRAS.518.3441C 518, 3441

    work page doi:10.1093/mnras/stac3330 2023

  7. [7]

    C., Roth N., Ramirez-Ruiz E., Miller M

    Dai L., McKinney J. C., Roth N., Ramirez-Ruiz E., Miller M. C., 2018, @doi [ ] 10.3847/2041-8213/aab429 , https://ui.adsabs.harvard.edu/abs/2018ApJ...859L..20D 859, L20

    work page doi:10.3847/2041-8213/aab429 2018

  8. [8]

    C., 2021, @doi [ ] 10.3847/2041-8213/ac2608 , https://ui.adsabs.harvard.edu/abs/2021ApJ...919L..20D 919, L20

    Dexter J., Scepi N., Begelman M. C., 2021, @doi [ ] 10.3847/2041-8213/ac2608 , https://ui.adsabs.harvard.edu/abs/2021ApJ...919L..20D 919, L20

    work page doi:10.3847/2041-8213/ac2608 2021

  9. [9]

    arXiv:2404.10036

    Eftekhari T., et al., 2024, @doi [arXiv e-prints] 10.48550/arXiv.2404.10036 , https://ui.adsabs.harvard.edu/abs/2024arXiv240410036E p. arXiv:2404.10036

    work page doi:10.48550/arxiv.2404.10036 2024

  10. [10]

    R., & Kochanek, C

    Evans C. R., Kochanek C. S., 1989, @doi [ ] 10.1086/185567 , https://ui.adsabs.harvard.edu/abs/1989ApJ...346L..13E 346, L13

    work page doi:10.1086/185567 1989

  11. [11]

    F., McKinney, J

    Gammie C. F., McKinney J. C., T \'o th G., 2003, @doi [ ] 10.1086/374594 , https://ui.adsabs.harvard.edu/abs/2003ApJ...589..444G 589, 444

    work page doi:10.1086/374594 2003

  12. [12]

    Gezari S., 2021, @doi [ ] 10.1146/annurev-astro-111720-030029 , https://ui.adsabs.harvard.edu/abs/2021ARA&A..59...21G 59, 21

    work page doi:10.1146/annurev-astro-111720-030029 2021

  13. [13]

    Golightly E. C. A., Nixon C. J., Coughlin E. R., 2019, @doi [ ] 10.3847/2041-8213/ab380d , https://ui.adsabs.harvard.edu/abs/2019ApJ...882L..26G 882, L26

    work page doi:10.3847/2041-8213/ab380d 2019

  14. [14]

    Guillochon J., Ramirez-Ruiz E., 2013, @doi [ ] 10.1088/0004-637X/767/1/25 , https://ui.adsabs.harvard.edu/abs/2013ApJ...767...25G 767, 25

    work page doi:10.1088/0004-637x/767/1/25 2013

  15. [15]

    B., Tokayer Y

    Guolo M., Gezari S., Yao Y., van Velzen S., Hammerstein E., Cenko S. B., Tokayer Y. M., 2024, @doi [ ] 10.3847/1538-4357/ad2f9f , https://ui.adsabs.harvard.edu/abs/2024ApJ...966..160G 966, 160

    work page doi:10.3847/1538-4357/ad2f9f 2024

  16. [16]

    Habibi A., Abbassi S., 2019, @doi [ ] 10.3847/1538-4357/ab5793 , https://ui.adsabs.harvard.edu/abs/2019ApJ...887..256H 887, 256

    work page doi:10.3847/1538-4357/ab5793 2019

  17. [17]

    Possible power source of Seyfert galaxies and QSOs

    Hills J. G., 1975, @doi [ ] 10.1038/254295a0 , https://ui.adsabs.harvard.edu/abs/1975Natur.254..295H 254, 295

    work page doi:10.1038/254295a0 1975

  18. [18]

    Huang J., Feng H., Gu W.-M., Wu W.-B., 2023, @doi [ ] 10.3847/1538-4357/ace71e , https://ui.adsabs.harvard.edu/abs/2023ApJ...954..150H 954, 150

    work page doi:10.3847/1538-4357/ace71e 2023

  19. [19]

    Kato Y., Mineshige S., Shibata K., 2004, @doi [ ] 10.1086/381234 , https://ui.adsabs.harvard.edu/abs/2004ApJ...605..307K 605, 307

    work page doi:10.1086/381234 2004

  20. [20]

    Lan c ov \'a D., et al., 2019, @doi [ ] 10.3847/2041-8213/ab48f5 , https://ui.adsabs.harvard.edu/abs/2019ApJ...884L..37L 884, L37

    work page doi:10.3847/2041-8213/ab48f5 2019

  21. [21]

    C., 2014, @doi [ ] 10.1088/0004-637X/786/1/6 , https://ui.adsabs.harvard.edu/abs/2014ApJ...786....6L 786, 6

    Li S.-L., Begelman M. C., 2014, @doi [ ] 10.1088/0004-637X/786/1/6 , https://ui.adsabs.harvard.edu/abs/2014ApJ...786....6L 786, 6

    work page doi:10.1088/0004-637x/786/1/6 2014

  22. [23]

    Liska M. T. P., Musoke G., Tchekhovskoy A., Porth O., Beloborodov A. M., 2022b, @doi [ ] 10.3847/2041-8213/ac84db , https://ui.adsabs.harvard.edu/abs/2022ApJ...935L...1L 935, L1

    work page doi:10.3847/2041-8213/ac84db 2041

  23. [24]

    Liska M. T. P., Kaaz N., Chatterjee K., Emami R., Musoke G., 2024, @doi [ ] 10.3847/1538-4357/ad344a , https://ui.adsabs.harvard.edu/abs/2024ApJ...966...47L 966, 47

    work page doi:10.3847/1538-4357/ad344a 2024

  24. [25]

    R., Guillochon J., Ramirez-Ruiz E., 2017, @doi [ ] 10.1051/0004-6361/201630092 , https://ui.adsabs.harvard.edu/abs/2017A&A...600A.124M 600, A124

    Mainetti D., Lupi A., Campana S., Colpi M., Coughlin E. R., Guillochon J., Ramirez-Ruiz E., 2017, @doi [ ] 10.1051/0004-6361/201630092 , https://ui.adsabs.harvard.edu/abs/2017A&A...600A.124M 600, A124

    work page doi:10.1051/0004-6361/201630092 2017

  25. [26]

    2014, Monthly Notices of the Royal Astronomical Society, 441, 3177, doi: 10.1093/mnras/stu762

    McKinney J. C., Tchekhovskoy A., Sadowski A., Narayan R., 2014, @doi [ ] 10.1093/mnras/stu762 , https://ui.adsabs.harvard.edu/abs/2014MNRAS.441.3177M 441, 3177

    work page doi:10.1093/mnras/stu762 2014

  26. [27]

    D., 2022, @doi [ ] 10.3847/2041-8213/ac90ba , https://ui.adsabs.harvard.edu/abs/2022ApJ...937L..12M 937, L12

    Metzger B. D., 2022, @doi [ ] 10.3847/2041-8213/ac90ba , https://ui.adsabs.harvard.edu/abs/2022ApJ...937L..12M 937, L12

    work page doi:10.3847/2041-8213/ac90ba 2022

  27. [28]

    R., Huang X., Davis S

    Meza M. R., Huang X., Davis S. W., Jiang Y.-F., 2025, @doi [arXiv e-prints] 10.48550/arXiv.2506.00109 , https://ui.adsabs.harvard.edu/abs/2025arXiv250600109M p. arXiv:2506.00109

    work page doi:10.48550/arxiv.2506.00109 2025

  28. [29]

    V., & Abramowicz, M

    Narayan R., Igumenshchev I. V., Abramowicz M. A., 2003, @doi [ ] 10.1093/pasj/55.6.L69 , https://ui.adsabs.harvard.edu/abs/2003PASJ...55L..69N 55, L69

    work page doi:10.1093/pasj/55.6.l69 2003

  29. [30]

    Navarro Navarro N., Piran T., 2025, @doi [ ] 10.3847/1538-4357/ad96b7 , https://ui.adsabs.harvard.edu/abs/2025ApJ...983..177N 983, 177

    work page doi:10.3847/1538-4357/ad96b7 2025

  30. [31]

    D., Thorne K

    Novikov I. D., Thorne K. S., 1973, in Black Holes (Les Astres Occlus). pp 343--450

    1973

  31. [32]

    R., et al., 2015, @doi [ ] 10.1088/0004-637X/805/1/68 , https://ui.adsabs.harvard.edu/abs/2015ApJ...805...68P 805, 68

    Pasham D. R., et al., 2015, @doi [ ] 10.1088/0004-637X/805/1/68 , https://ui.adsabs.harvard.edu/abs/2015ApJ...805...68P 805, 68

    work page doi:10.1088/0004-637x/805/1/68 2015

  32. [33]

    S., 1989, in Morris M., ed., Vol

    Phinney E. S., 1989, in Morris M., ed., Vol. 136, The Center of the Galaxy. p. 543

    1989

  33. [34]

    L., Mockler B., 2025, @doi [ ] 10.3847/1538-4357/adc729 , https://ui.adsabs.harvard.edu/abs/2025ApJ...985...77P 985, 77

    Piro A. L., Mockler B., 2025, @doi [ ] 10.3847/1538-4357/adc729 , https://ui.adsabs.harvard.edu/abs/2025ApJ...985...77P 985, 77

    work page doi:10.3847/1538-4357/adc729 2025

  34. [35]

    Qiao E., Wu Y., Lin Y., Guo M., Liu J., Guo C., Jin C., Jiang N., 2025, @doi [ ] 10.1093/mnras/staf719 , https://ui.adsabs.harvard.edu/abs/2025MNRAS.539.3473Q 539, 3473

    work page doi:10.1093/mnras/staf719 2025

  35. [36]

    , keywords =

    Rees M. J., 1988, @doi [ ] 10.1038/333523a0 , https://ui.adsabs.harvard.edu/abs/1988Natur.333..523R 333, 523

    work page doi:10.1038/333523a0 1988

  36. [37]

    , archivePrefix = "arXiv", eprint =

    Ressler S. M., Tchekhovskoy A., Quataert E., Gammie C. F., 2017, @doi [ ] 10.1093/mnras/stx364 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.467.3604R 467, 3604

    work page doi:10.1093/mnras/stx364 2017

  37. [38]

    C., Pickrel D., 2025, @doi [ ] 10.3847/1538-4357/adb1c1 , https://ui.adsabs.harvard.edu/abs/2025ApJ...981..144R 981, 144

    Roth N., Anninos P., Fragile P. C., Pickrel D., 2025, @doi [ ] 10.3847/1538-4357/adb1c1 , https://ui.adsabs.harvard.edu/abs/2025ApJ...981..144R 981, 144

    work page doi:10.3847/1538-4357/adb1c1 2025

  38. [39]

    C., Avara M., 2023, @doi [ ] 10.3847/1538-4357/acf5de , https://ui.adsabs.harvard.edu/abs/2023ApJ...957...12R 957, 12

    Ryu T., Krolik J., Piran T., Noble S. C., Avara M., 2023, @doi [ ] 10.3847/1538-4357/acf5de , https://ui.adsabs.harvard.edu/abs/2023ApJ...957...12R 957, 12

    work page doi:10.3847/1538-4357/acf5de 2023

  39. [40]

    S a dowski A., Narayan R., 2015, @doi [ ] 10.1093/mnras/stv2022 , https://ui.adsabs.harvard.edu/abs/2015MNRAS.454.2372S 454, 2372

    work page doi:10.1093/mnras/stv2022 2015

  40. [41]

    S a dowski A., Narayan R., 2016, @doi [ ] 10.1093/mnras/stv2941 , https://ui.adsabs.harvard.edu/abs/2016MNRAS.456.3929S 456, 3929

    work page doi:10.1093/mnras/stv2941 2016

  41. [42]

    S a dowski A., Narayan R., Tchekhovskoy A., Zhu Y., 2013, @doi [ ] 10.1093/mnras/sts632 , https://ui.adsabs.harvard.edu/abs/2013MNRAS.429.3533S 429, 3533

    work page doi:10.1093/mnras/sts632 2013

  42. [43]

    , archivePrefix = "arXiv", eprint =

    S a dowski A., Narayan R., McKinney J. C., Tchekhovskoy A., 2014, @doi [ ] 10.1093/mnras/stt2479 , https://ui.adsabs.harvard.edu/abs/2014MNRAS.439..503S 439, 503

    work page doi:10.1093/mnras/stt2479 2014

  43. [44]

    2015, MNRAS, 447, 49, doi: 10.1093/mnras/stu2387

    S a dowski A., Narayan R., Tchekhovskoy A., Abarca D., Zhu Y., McKinney J. C., 2015, @doi [ ] 10.1093/mnras/stu2387 , https://ui.adsabs.harvard.edu/abs/2015MNRAS.447...49S 447, 49

    work page doi:10.1093/mnras/stu2387 2015

  44. [45]

    S a dowski A., Tejeda E., Gafton E., Rosswog S., Abarca D., 2016, @doi [ ] 10.1093/mnras/stw589 , https://ui.adsabs.harvard.edu/abs/2016MNRAS.458.4250S 458, 4250

    work page doi:10.1093/mnras/stw589 2016

  45. [46]

    , archivePrefix = "arXiv", eprint =

    S a dowski A., Wielgus M., Narayan R., Abarca D., McKinney J. C., Chael A., 2017, @doi [ ] 10.1093/mnras/stw3116 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.466..705S 466, 705

    work page doi:10.1093/mnras/stw3116 2017

  46. [47]

    C., 2024, @doi [ ] 10.1038/s41586-023-06875-y , https://ui.adsabs.harvard.edu/abs/2024Natur.625..463S 625, 463

    Steinberg E., Stone N. C., 2024, @doi [ ] 10.1038/s41586-023-06875-y , https://ui.adsabs.harvard.edu/abs/2024Natur.625..463S 625, 463

    work page doi:10.1038/s41586-023-06875-y 2024

  47. [48]

    S., & Dopita, M

    Sutherland R. S., Dopita M. A., 1993, @doi [ ] 10.1086/191823 , https://ui.adsabs.harvard.edu/abs/1993ApJS...88..253S 88, 253

    work page doi:10.1086/191823 1993

  48. [49]

    R., Zubovas, K., & Power, C.\ 2011, , 415, 1, L6

    Tchekhovskoy A., Narayan R., McKinney J. C., 2011, @doi [ ] 10.1111/j.1745-3933.2011.01147.x , https://ui.adsabs.harvard.edu/abs/2011MNRAS.418L..79T 418, L79

    work page doi:10.1111/j.1745-3933.2011.01147.x 2011

  49. [50]

    C., Narayan R., 2012, in Journal of Physics Conference Series

    Tchekhovskoy A., McKinney J. C., Narayan R., 2012, in Journal of Physics Conference Series. p. 012040 ( @eprint arXiv 1202.2864 ), @doi 10.1088/1742-6596/372/1/012040

    work page doi:10.1088/1742-6596/372/1/012040 2012

  50. [51]

    L., Kwan, T

    Thomsen L. L., Kwan T. M., Dai L., Wu S. C., Roth N., Ramirez-Ruiz E., 2022, @doi [ ] 10.3847/2041-8213/ac911f , https://ui.adsabs.harvard.edu/abs/2022ApJ...937L..28T 937, L28

    work page doi:10.3847/2041-8213/ac911f 2022

  51. [52]

    D., Jiang Y.-F., Antoni A., 2025, @doi [arXiv e-prints] 10.48550/arXiv.2512.14810 , https://ui.adsabs.harvard.edu/abs/2025arXiv251214810T p

    Tuna S., Metzger B. D., Jiang Y.-F., Antoni A., 2025, @doi [arXiv e-prints] 10.48550/arXiv.2512.14810 , https://ui.adsabs.harvard.edu/abs/2025arXiv251214810T p. arXiv:2512.14810

    work page doi:10.48550/arxiv.2512.14810 2025

  52. [53]

    R., & Asahina, Y

    Utsumi A., Ohsuga K., Takahashi H. R., Asahina Y., 2022, @doi [ ] 10.3847/1538-4357/ac7eb8 , https://ui.adsabs.harvard.edu/abs/2022ApJ...935...26U 935, 26

    work page doi:10.3847/1538-4357/ac7eb8 2022

  53. [54]

    Wang J.-M., Szuszkiewicz E., Lu F.-J., Zhou Y.-Y., 1999, @doi [ ] 10.1086/307686 , https://ui.adsabs.harvard.edu/abs/1999ApJ...522..839W 522, 839

    work page doi:10.1086/307686 1999

  54. [55]

    Wu W.-B., Gu W.-M., Sun M., 2022, @doi [ ] 10.3847/1538-4357/ac6588 , https://ui.adsabs.harvard.edu/abs/2022ApJ...930..108W 930, 108

    work page doi:10.3847/1538-4357/ac6588 2022

  55. [56]

    Yao Y., et al., 2022, @doi [ ] 10.3847/1538-4357/ac898a , https://ui.adsabs.harvard.edu/abs/2022ApJ...937....8Y 937, 8

    work page doi:10.3847/1538-4357/ac898a 2022

  56. [57]

    A., Berger E., Margutti R., Pooley G

    Zauderer B. A., Berger E., Margutti R., Pooley G. G., Sari R., Soderberg A. M., Brunthaler A., Bietenholz M. F., 2013, @doi [ ] 10.1088/0004-637X/767/2/152 , https://ui.adsabs.harvard.edu/abs/2013ApJ...767..152Z 767, 152

    work page doi:10.1088/0004-637x/767/2/152 2013

  57. [58]

    Zheng S.-M., Yuan F., Gu W.-M., Lu J.-F., 2011, @doi [ ] 10.1088/0004-637X/732/1/52 , https://ui.adsabs.harvard.edu/abs/2011ApJ...732...52Z 732, 52

    work page doi:10.1088/0004-637x/732/1/52 2011