Explaining changing-look AGN with state transition triggered by rapid mass accretion rate drop

TL;DR

This paper models the spectral evolution of the changing-look AGN Mrk 1018, linking its state transition to rapid accretion rate drops and disc instabilities, revealing faster-than-expected changes due to radiation and magnetic pressures.

Contribution

It introduces a model connecting spectral changes in Mrk 1018 to state transitions driven by accretion rate drops, emphasizing the role of radiation and magnetic pressures in faster disc evolution.

Findings

Spectral evolution correlates with Eddington ratio changes.

Soft X-ray excess drop causes disappearance of broad line region.

Disc instability and pressure effects accelerate state transitions.

Abstract

We model the broadband (optical/UV and X-ray) continuum spectrum of the "changing-look" active galactic nucleus (AGN) Mrk 1018 as it fades from Seyfert 1 to 1.9 in $\sim 8$ years. The brightest spectrum, with Eddington ratio $L/L_{\rm Edd}\sim 0.08$, has a typical type 1 AGN continuum, with a strong "soft X-ray excess" spanning between the UV and soft X-rays. The dimmest spectrum, at $L/L_{\rm Edd} \sim 0.006$, is very different in shape as well as luminosity, with the soft excess dropping by much more than the hard X-rays. The soft X-ray excess produces most of the ionizing photons, so its dramatic drop leads to the disappearance of the broad line region, driving the "changing-look" phenomena. This spectral hardening appears similar to the soft-to-hard state transition in black hole binaries at $L/L_{\rm Edd} \sim 0.02$, where the inner disc evaporates into an advection dominated accretion flow, while the overall drop in luminosity appears consistent with the Hydrogen ionization disc instability. Nonetheless, both processes happen much faster in Mrk 1018 than predicted by disc theory. We critically examine scaling from galactic binary systems, and show that a major difference is that radiation pressure should be much more important in AGNs, so that the sound speed is much faster than expected from the gas temperature. Including magnetic pressure to stabilize the disc shortens the timescales even further. We suggest that all changing-look AGNs are similarly associated with the state transition at $L/L_{\rm Edd} \sim$ a few percent.