AGN SED Model: Structure, Dust, and Radiative Transfer

• AGN SED Models are frameworks that simulate the multi-wavelength emission of active galactic nuclei by integrating detailed dust geometry, accretion disk physics, and radiative transfer.
• They employ a two-phase dust torus—with a smooth disk and clumpy components—to replicate observed silicate features and emission anisotropy over infrared wavelengths.
• These models use extensive SED libraries and advanced Monte Carlo methods to infer intrinsic AGN properties and guide studies on black hole growth and galaxy evolution.

Active Galactic Nuclei (AGN) Spectral Energy Distribution (SED) models describe the multi-wavelength emission from accreting supermassive black holes and their immediate dusty and gaseous environments. Modern SED models address the geometry and physics of the circumnuclear dust, the emission properties of accretion disks, and the radiative transfer of photons in complex, inhomogeneous media. These frameworks underpin the interpretation of AGN spectral diagnostics and provide critical constraints for AGN surveys, black hole growth paradigms, and galaxy–AGN co-evolution.

1. Physical Structure and Parameterization of the SED Model

The canonical AGN SED model is anchored by a luminous central engine emitting a scale-invariant "big blue bump" (BBB), typically modeled as isotropic with $L_{\rm AGN}=10^{11}\,L_\odot$ for scale setting. The immediate circumnuclear environment is represented by a toroidal distribution of dust, parametrized as a two-phase medium: a geometrically thick, homogeneous disk (smooth component) and a population of optically thick, randomly distributed clumps (clumpy component). The key geometric and physical parameters are:

• Inner torus radius $r_{\rm in}$: Set by sublimation of silicate/carbonaceous dust; $$r_{\rm in} = 3, 5.1, 7.7, 10, 15.5 \times 10^{17}\,$$cm, scaling as $r_{\rm in}\propto\sqrt{L_{\rm AGN}}$.
• Outer torus radius $r_{\rm out}$: Fixed as $r_{\rm out}=170\,r_{\rm in}$.
• Viewing angle $\theta$: Defined from the polar direction; nine discrete bins from 19°–86°.
• Homogeneous disk: Vertical Gaussian with scale height $h(r)=r_{\rm out}/8\,(2r/r_{\rm out})^{1.125}$; optical depth $\tau_V^{\rm mid}=0, 30, 100, 300, 1000$.
• Clumpy component: Volume filling factor $\eta=1.5\%, 7.7\%, 38.5\%, 77.7\%$; individual cloud optical depth $\tau_V^{\rm cl}=0, 4.5, 13.5, 45$; random spatial distribution.
• Clump size and placement: Uniform-density cubes of $d=0.3$–$3$ pc; positions randomized in spherical coordinates using uniform deviates.

This two-phase scheme comprehensively samples the observed phenomenology of AGN IR emission and silicate features, accommodating both edge-on (Type 2) and face-on (Type 1) lines of sight (Siebenmorgen et al., 2015).

2. Dust Grain Physics and Its Impact on the SED

The two-phase torus SED model employs "fluffy aggregate" dust grains—mixed silicate and amorphous carbon mixtures with high internal vacuum volume (50%). The grain size distribution is $n(a)\propto a^{-3.5}$ for $16 < a < 260$ nm. These aggregates have key radiative properties distinguishing them from standard diffuse ISM grains:

• Elevated far-IR/submm absorption: $\kappa_{1\,\rm mm}$ up to $\times12$ relative to ISM dust.
• Reduced optical scattering.
• Broadened and shifted silicate feature: The canonical 10.4 μm feature (ISM) is broader, shallower, and peaks at $\sim$11.5 μm for the aggregate model.

Consequences include a shift of the SED peak to longer wavelengths for identical dust geometry and enhanced isotropization of emission at $\sim$200 μm (compared to $\sim$100 μm for ISM dust). Edge-on geometries may entirely erase classic silicate absorption features (Siebenmorgen et al., 2015).

3. Radiative Transfer Methodology

The full three-dimensional, monochromatic radiative transfer equation is solved for the two-phase torus:

$$\frac{dI_\nu}{ds} = -\kappa_\nu\,\rho\,I_\nu + \eta_\nu^{\rm em} + \int p(\cos\Theta)\,I_\nu(\Omega')\,d\Omega'$$

where $\kappa_\nu$ is mass absorption coefficient, $\eta_\nu^{\rm em} = \kappa_\nu\,\rho\,B_\nu(T)$ the dust emission, and $p(\cos\Theta)$ the Henyey–Greenstein scattering phase function. The solution employs a Bjorkman–Wood (2001) energy packet-based Monte Carlo method with immediate temperature updating. Each SED is synthesized using typically $2\times10^5$ photon packets per frequency bin over 256 frequency bins, with the emergent SED either obtained via direct photon counting in a fixed number of viewing angles or via ray-tracing. Special handling for optically thin cells (Lucy 1999) ensures computational efficiency and accuracy (Siebenmorgen et al., 2015).

4. SED Library Construction, Parameter Effects, and Diagnostic Features

A comprehensive library of 3600 SEDs spans the plausible AGN parameter space:

• Parameter grid: 5 $r_{\rm in}$ values × 4 $\eta$ values × 4 $\tau_V^{\rm cl}$ × 5 $\tau_V^{\rm mid}$ × 9 viewing angles.

Parameter variations systematically affect the SED:

• Larger inner radius ($r_{\rm in}$): Yields colder dust, shifting SED peak to longer wavelength and increasing far-IR/submm emission.
• Adding a homogeneous disk: Increases face-on near-IR emission and overall submm flux.
• Elevated $\tau_V^{\rm mid}$: Switches the 10 μm silicate feature from emission to absorption, especially in edge-on views.
• Increasing $\eta$ or $\tau_V^{\rm cl}$: Deepens silicate absorption and flattens the near-IR continuum.

The model library encompasses the observed diversity of 10 μm silicate band strengths and peak wavelengths ($-0.3 \lesssim \tau_{\rm Si} \lesssim 1.5$, 9.7–11.9 μm), and matches the range of observed mid-IR colors for known AGN (Siebenmorgen et al., 2015). The silicate feature statistics, in particular, underscore the necessity of both smooth and clumpy phases for replicating observed SED scatter.

5. SED-Fitting and Practical Inference of AGN Properties

SED fitting follows a direct, library-matching approach:

1. Data assembly: Compile observed IR photometry and/or mid-IR spectroscopy.
2. Component separation: Remove non-torus emission if present (e.g., synchrotron or stellar blackbody).
3. Model fitting: For each library SED, evaluate $\chi^2$ (or likelihood) against observed data.
4. Parameter inference: Select the best-fit model; infer $$r_{\rm in}, \eta, \tau_V^{\rm cl}, \tau_V^{\rm mid}, \theta$$.
5. Intrinsic luminosity calculation: Use the anisotropy correction:

$$\epsilon(\theta) = \frac{F_{\rm model}(\theta)}{F_{\rm model}^{\rm no\,dust}(\theta)},\quad L_{\rm AGN} = \frac{9L_{\rm obs}(\theta)}{\epsilon(\theta)}$$

This yields a robust estimate of the intrinsic AGN power even for highly anisotropic or obscured objects.

1. Predictive power: Unmeasured fluxes at other wavelengths are directly predicted from the best-fit SED.
2. Additional starburst component: If a residual far–IR excess remains, a starburst SED is added; typically this component is $\leq 20\%$ of $L_{\rm bol}$.

The full SED library and associated documentation are publicly accessible, enabling rapid observational inferences for diverse AGN populations (Siebenmorgen et al., 2015).

6. Context, Limitations, and Extensions

The two-phase AGN SED model uniquely captures the empirical diversity of the AGN mid-IR through far-IR spectrum in a physically motivated, scale-invariant fashion. Its core strengths are:

• Empirically validated: Consistently fits the SEDs of Seyfert galaxies, luminous quasars, radio galaxies, and ULIRGs with minimal additional components.
• Parameter mapping: Enables the estimation of intrinsic luminosity, obscuration, and torus geometry from IR data alone.
• Physical interpretability: The explicit parameterization allows for modeling of anisotropy and direct translation to physical torus configurations.
• Silicate band flexibility: Replicates the observed spread in silicate feature strength/shape with both clumpy and smooth dust phases.

Constraints and future directions include the identification and resolution of model degeneracies—specifically between cloud filling factor, optical depth, and viewing angle—particularly in heavily obscured objects, and the incorporation of spatially resolved mid-IR observations to independently constrain dust geometry. The library approach is especially suited for surveys and follow-up of large AGN samples where rapid, physically anchored SED fitting is required. Further advances will likely integrate variable dust grain compositions, anisotropic central source emission, and polar dusty wind components.

References: Siebenmorgen, Heymann & Efstathiou (2015) (Siebenmorgen et al., 2015).