A dust condensation instability in AGN atmospheres: failed winds and the broad line region

TL;DR

This work tackles the origin of the AGN broad-line region by identifying a dust-condensation instability in the disc atmosphere that creates an extended, radiation-pressure-supported layer. Through analytic linear stability analyses and 2D hydrodynamic simulations, it demonstrates that isothermal perturbations grow in the presence of density-dependent dust opacity, producing nonlinear dusty fountains that reach speeds up to ~$10^3$ km s$^{-1}$ and then sublimate, causing the material to fall back as a clumpy, failed wind. The results naturally align with the FRADO scenario for the broad-line region, offering a physical mechanism for clumpiness, high velocity dispersion, and variability driven by central radiation and disc geometry. While promising, the study also notes limitations (densities not fully matched to observations, 2D/local geometry, simplified dust physics) and outlines substantial future work to confirm the connection to observed BLR properties and changing-look AGN phenomena.

Abstract

Active galactic nuclei (AGN) are important drivers of galactic evolution; however, the underlying physical processes governing their properties remain uncertain. In particular, the specific cause for the generation of the broad-line region is unclear. There is a region where the underlying accretion disc atmosphere becomes cool enough for dust condensation. Using models of the disc's vertical structure, accounting for dust condensation and irradiation from the central source, we show that their upper atmospheres become extended, dusty, and radiation-pressure-supported. Due to the density--temperature dependence of dust condensation, this extended atmosphere forms as the dust abundance slowly increases with height, resulting in density and temperature scale heights considerably larger than the gas pressure scale height. We show that such an atmospheric structure is linearly unstable. An increase in the gas density raises the dust sublimation temperature, leading to an increased dust abundance, a higher opacity, and hence a net vertical acceleration. Using localised 2D hydrodynamic simulations, we demonstrate the existence of our linear instability. In the non-linear state, the disc atmosphere evolves into ``fountains'' of dusty material that are vertically launched by radiation pressure before being exposed to radiation from the central source, which sublimates the dust and shuts off the radiative acceleration. These dust-free clumps then evolve ballistically, continuing upward before falling back towards the disc under gravity. This clumpy ionized region has velocity dispersions $\gtrsim 1000$ km/s. This instability and our simulations are representative of the Failed Radiatively Accelerated Dusty Outflow (FRADO) model proposed for the AGN broad-line region.

Figures (16)

• Figure 1: The variation of the density and temperature opacity indices ($a$, $b$ -- Equation \ref{['eqn:opac_simple']}) during the dust condensation process at a gas density of $10^{-12}$ g cm$^{-3}$, from dust free (top right) to dusty (top left), as indicated by the arrows. This specific curve is plotted for our dust condensation model described in Section \ref{['sec:opacity']} as the temperature is varied; however, whatever the specific model, the general result is that during dust condensation, $b$ becomes large and negative, and $a$ becomes order of unity.
• Figure 2: The vertical structure of the disc at a radius of $\sim 6.8\times10^{16}$ cm from the central black hole, with an effective temperature of $\sim 1100$ K, and flaring angle 0.003152. In the atmosphere of the disc, above the photosphere ($\tau_z \approx 0.76$), the disc is in a dynamic equilibrium between radiation pressure and vertical gravity as the dust slowly condenses with height. Once radiation from the central source is optically thin, the temperature rapidly rises, and the dust is destroyed, leading to a rapid drop in the density. In the density panel we show two additional density profiles where we increase the flaring angle to 0.02666 (orange) and lower it to 0.001516 (green). Changing the flaring angle changes the amount of radiation from the central source that is absorbed locally, which has a dramatic impact on the size of the radiation pressure supported atmosphere.
• Figure 3: The opacity as a function of temperature and density with isothermal and adiabatic density perturbations shown, isothermal ones have increasing opacity, adiabatic have decreasing opacity.
• Figure 4: The value of $a$ (top), and $a+b(\gamma-1)$ (bottom, for $\gamma=7/5$), the adiabatic term in the small $k$ dispersion relationship. The black solid line is the point where $a+b(\gamma-1)$ equals zero, the dashed line is the position of dust condensation. We see that the dust condensation region is unstable to isothermal perturbations ($a>0$), but stable to adiabatic perturbations ($a+b(\gamma-1)<0)$.
• Figure 5: An example numerical solution to the dispersion relation shown in red solid (unstable) and green dashed (stable), the black solutions show the pure isothermal solutions (solid: unstable, dashed: stable), and the orange shows the pure adiabatic solution. The full solution is well approximated by the isothermal approximation to very short wavelengths.