The Origins of AGN Obscuration: The Torus as a Dynamical, Unstable Driver of Accretion

Multi-scale simulations have made it possible to follow gas inflows onto massive black holes (BHs) from galactic scales to the accretion disk. When sufficient gas is driven towards the BH, gravitational instabilities generically form lopsided, eccentric disks that propagate inwards. The lopsided stellar disk exerts a strong torque on the gas disk, driving inflows that fuel rapid BH growth. Here, we investigate whether the same gas disk is the ’torus’ invoked to explain obscured AGN. The disk is generically thick and has characteristic ~1-10 pc sizes and masses resembling those required of the torus. The scale heights and obscured fractions of the predicted torii are substantial even in the absence of strong stellar feedback providing the vertical support. Rather, they can be maintained by strong bending modes and warps excited by the inflow-generating instabilities. Other properties commonly attributed to feedback processes may be explained by dynamical effects: misalignment between torus and host galaxy, correlations between local SFR and turbulent gas velocities, and dependence of obscured fractions on AGN luminosity or SFR. We compare the predicted torus properties with observations of gas surface density profiles, kinematics, scale heights, and SFR densities in AGN nuclei, and find that they are consistent. We argue that it is not possible to reproduce these observations and the observed column density (N_H) distribution without a clumpy gas distribution, but allowing for clumping on small scales the predicted N_H distribution is in good agreement with observations from 10^20-27 cm^-2. We examine how N_H scales with galaxy and AGN properties, and find that AGN feedback may be necessary to explain some trends with luminosity and/or redshift. The torus is not merely a bystander or passive fuel source for accretion, but is itself the mechanism driving accretion.

💡 Research Summary

This paper presents a comprehensive study of active galactic nucleus (AGN) obscuration using a suite of multi‑scale “zoom‑in” hydrodynamic simulations that follow gas inflow from galactic (∼100 kpc) down to sub‑parsec scales. Starting from galaxy‑scale merger and bar‑unstable disk simulations (∼50 pc resolution), the authors isolate the central 0.1–1 kpc region that contains the bulk of the gas driven inward (∼10¹⁰ M⊙). They then re‑populate this region at much higher resolution (∼1 pc) and evolve it for several dynamical times, capturing the emergence of strong self‑gravity‑driven instabilities. A second re‑simulation step focuses on the innermost 10–30 pc, achieving a spatial resolution of 0.1 pc and particle masses of ∼100 M⊙.

The key dynamical finding is that, in sufficiently gas‑rich nuclei, the central few tens of parsecs become unstable to the formation of a lopsided (m = 1) eccentric gas+stellar disk. This eccentric pattern exerts a powerful gravitational torque on the surrounding gas, driving rapid angular‑momentum loss and inflow rates up to ∼10 M⊙ yr⁻¹—enough to power luminous quasars. The same gas disk, while in its gas‑rich phase, naturally attains a thick, torus‑like geometry with characteristic radii of 1–10 pc, masses of 10⁶–10⁸ M⊙, and scale‑height‑to‑radius ratios h/R ≈ 0.3–0.5. Crucially, these large scale heights are maintained without invoking strong stellar or AGN feedback; instead, vertical support arises from gravitational bending modes, warps, and twists that are excited by the inflow‑driving eccentric instability.

The simulations also reveal that the torus is intrinsically clumpy. On unresolved sub‑parsec scales the gas fragments into clouds of ∼10³–10⁴ M⊙ with densities 10⁴–10⁶ cm⁻³, consistent with observed molecular cloud properties in nearby AGN. Cloud–cloud collisions generate turbulent velocity dispersions of 30–100 km s⁻¹, reproducing the line widths seen in infrared and millimetre observations. By assuming a simple clumping prescription, the authors compute line‑of‑sight column densities (N_H) and find a distribution spanning 10²⁰–10²⁷ cm⁻² that matches the observed X‑ray absorption statistics of both type 1 and type 2 AGN.

The paper further explores how the obscured fraction depends on AGN luminosity, host galaxy star‑formation rate (SFR), and redshift. While higher luminosities modestly reduce the covering factor—likely due to radiation‑driven clearing—the dominant driver of obscuration is the amount of gas supplied to the nucleus, i.e., the strength of the large‑scale inflow. A positive correlation between SFR and obscured fraction emerges naturally because vigorous star formation enhances the clumpiness of the torus, increasing the probability of high‑N_H sightlines.

Overall, the authors argue that the traditional “torus” is not a passive, static obscuring screen but the very dynamical structure that fuels black‑hole growth. Its thickness, clumpiness, and kinematic properties arise from gravitational instabilities rather than from ad‑hoc feedback prescriptions. This paradigm simultaneously explains a wide range of observational constraints: gas surface‑density profiles, scale heights, turbulent velocities, star‑formation surface densities, and the full column‑density distribution.

The study concludes that a physically motivated, self‑consistent treatment of gas inflow naturally produces a torus‑like obscurer whose properties are dictated by the same processes that drive accretion. Future work incorporating explicit radiative transfer and AGN feedback will test how these mechanisms modify the torus over cosmic time and will provide tighter links between simulated predictions and high‑resolution infrared, sub‑mm, and X‑ray observations.