Changing-Look AGN Powered By Disk Tearing

TL;DR

This work demonstrates that disk tearing of a strongly tilted accretion disk around a rapidly spinning $10^8 M_\odot$ black hole can power changing‑look AGN by driving large continuum and broad‑line variability on months‑to‑years timescales. Using an extremely high‑resolution GRMHD simulation (HAMR) of a tilted disk, ray tracing with RAPTOR, and CLOUDY‑based BLR emission, the authors predict luminosity swings, inner‑disk precession induced modulation, and an intraday quasi‑periodic oscillation tied to radial breathing of the torn inner disk. They show that tearing cycles imprint time‑dependent asymmetries in broad lines via asymmetric BLR illumination, offering a potential smoking gun for disk tearing in CSAGN. The study also provides observational predictions for ULTRASAT and Rubin Observatory, highlighting multi‑band photometry and high‑cadence spectroscopy as key tools, while noting caveats related to radiative transfer, BLR geometry, and jet/corona coupling. Overall, disk tearing offers a physically motivated pathway to the rapid, noncoherent variability seen in CLAGN and connects inner disk dynamics to observable BLR responses.

Abstract

Changing-look active galactic nuclei (CLAGN) feature order-of-magnitude variability in both the continuum and broad line luminosities on months-to-years long timescales, and are currently unexplained. Simulations have demonstrated that rotating black holes sometimes tear apart tilted accretion disks. These tearing events violently restructure the disk on timescales much shorter than a viscous timescale, hinting at a connection to CLAGN. Here, we show that disk tearing can power changing-look events. We report synthetic observations of an extremely high resolution three-dimensional general-relativistic magnetohydrodynamic simulation of a geometrically thin, tilted accretion disk around a rapidly rotating, $10^8\,M_\odot$ black hole. We perform ray-tracing calculations that follow the disk light to both a line of sight camera and to a distribution of cameras in a prescribed torus-like broad line region. The continuum photoionizes the broad line region and we calculate the resulting spectrum. Both the continuum and line luminosities undergo order of magnitude swings on months-to-years long timescales. We find shorter, weeks long variability driven by the geometric precession of the inner disk and an intraday quasi-periodic oscillation driven by radial breathing of the inner disk. When the torn disk precesses, it causes asymmetric illumination of the broad line region, driving time-evolving red-to-blue asymmetries of the broad emission lines that may be a smoking gun for disk tearing. We also make predictions for future photometric observations from ULTRASAT and Vera Rubin Observatory, both of which may play an important role in detecting future changing-look events.

Figures (9)

• Figure 1: Three-dimensional gas density rendering of a torn disk with a cartoon depiction of the broad line region (BLR). Light colors are high density, dark colors are low density. The majority of the emission emanates from the inner disk (yellow squiggles). This emission photoionizes the BLR, which is idealized as a rotating torus of optically thin gas located $\sim45^\circ$ above the outer disk (white region). Each segment of the BLR then emits lines (orange squiggles) that are Doppler shifted due to the Keplerian motion of the BLR gas. Throughout this paper, we assume a line of sight at a $15^\circ$ inclination. Panels (i)-(iii) Film strip of ray-traced images featuring various phases in the evolution of the inner disk. Panel (ii) is shown at the same time as the 3D rendering in the main panel.
• Figure 2: Simulated light curves. Each curve exhibits roughly order-of-magnitude variability on months-to-years timescales. Panel a. Bolometric isotropic-equivalent luminosity at $15^\circ$ inclination and $0^\circ$ azimuth (black). We also show $\dot{M}$ rescaled to a luminosity assuming 10% radiative efficiency. $L_{\rm iso}$ mostly tracks $\dot{M}$, but features some weeks-long periodicity due to the geometric precession of the inner disk. One example is highlighted by the vertical blue lines which correspond to the panels in Fig. \ref{['fig:raptor']}. Vertical green lines correspond to panels in Figs. \ref{['fig:3d']} and \ref{['fig:spectra']}. Panel b. Same as panel a, except we show the luminosity convolved with filters for ULTRASATULTRASAT_2024 and the $u$/$g$/$r$ LSST LSST_2019 bands. Panel c. Line luminosities in H$$, H$$ and He I (5876 Å). Panel d. Equivalent widths of the lines.
• Figure 3: Spacetime diagram of disk midplane density and tilt, which reveals repeated tearing cycles and, sometimes, transient alignment of the inner disk. Panel a. Midplane density, $_{\rm midplane}$. Low-density (purple) regions indicate gaps between sub-disks or a recently-consumed inner disk. We have highlighted two tears with white lines; the outer disk will refill the inner region at a rate $v_r\gtrsim10^{-2}v_{\rm k}$ (white dashed lines) and eventually lead to plunging flows with $v_{\rm r}\gtrsim 10^{-1}v_{\rm k}$ (white dotted lines), which are coincident peaks in the luminosity (Fig. \ref{['fig:lightcurve']}). Panel b. Tilt angle, $T$. Usually, $T\sim60-65^\circ$, but the inner regions sometimes transiently align ($\sim100-200$ days and $\sim450-550$ days). The transiently-aligned phases are coincident with the troughs in the light curves (Fig. \ref{['fig:lightcurve']}. Green vertical lines correspond to panels in Fig. \ref{['fig:3d']} and Fig. \ref{['fig:spacetime']}; blue vertical lines correspond to panels in Fig. \ref{['fig:raptor']}.
• Figure 4: Effective temperature and luminosity per unit radius at time $t-t_0=40\,{\rm days}$. Panel a. The effective temperature of the disk as a function of radius. We also have labeled the radius of the tear, $r_{\rm tear}$. Panel b. The luminosity per unit radius, $\frac{dL}{dr}$ (Eq. \ref{['eq:dLdr']}), as a function of radius. We show curves both for the band-dependent luminosity and ($0.1\%$ of) the bolometric luminosity, where we have assumed that each radius of the disk radiates as a blackbody. Note that theses curves are calculated from the data, not ray-traced, so that there are no viewing angle effects.
• Figure 5: Ray-traced images of the precessing inner disk (corresponds to vertical green lines in Fig. \ref{['fig:lightcurve']}a and Fig. \ref{['fig:spacetime']}b). We can see a $\sim180^\circ$ geometric precession from panel (a) to (d), which super-imposes weeks-long periodicity in the light curve on top of the months-to-years long variability.