Spectroscopic constraints on the properties of dust in active galactic nuclei

The lack of a significant silicate or silicon carbide emission feature in bright active galactic nuclei (AGNs) is used to constrain models in which the IR continuum is emitted by dust. We consider two models for the dust, a graphite + silicate grain mixture and a graphite + silicon carbide mixture. The optical properties of the grains are calculated using Mie theory, the Rayleigh-Gans approximation and geometric optics, for grains in the 0.005-10 μm size range, over the 1000 μm-1 Å wavelength range. We use these grain models to calculate the emission of optically thin and of optically thick dust, with various grain compositions incorporating both absorption and scattering in the detailed radiative transfer. We find that ∼1:1 mixtures of graphite + silicate grains of a ≤ 3 μm in any configuration which is optically thin at 10 μm produce a very strong emission feature and are dearly ruled out. Optically thin dust must either be depleted of silicates by at least a factor of 5, or be composed mostly of grains as large as 10 μm. Dust with a large optical depth at 10 μm produces a significantly weaker emission feature, but its amplitude is still larger than the observational limits in most objects. This feature is washed out if the dust composition or grain size distribution are somewhat modified, or possibly if another heat source (e.g., stars) exists at a large optical depth inside the clouds. Similar results are obtained for a graphite + silicon carbide mixture. The different solutions to the absence of a silicate or silicon carbide emission feature can be further constrained using high S/N IR spectroscopy at 10 μm, X-ray spectroscopy, near-IR variability, and by looking for high ionization lines, or molecular lines from the associated gas. The constraints on the dust configuration and composition imply the following: (1) the observed broad emission lines and continuum are unlikely to be noticeably reddened in most object. (2) Dust cannot exist in the broad-line region clouds if their distance from the continuum source is smaller than 0.2L^1/2₄₆ pc as recently indicated in a few objects. (3) Dust can exist in the narrow-line region clouds, and will have a strong effect on the narrow line flux if the ionization parameter U ≥ 0.01. (4) If the IR emission originates in clouds which are optically thick at 10 μm, then U ≥ 0.1 at the cloud surface. We finally note that physical and dynamical arguments lead to similar constraints: small grains in an optically thin dust configuration are likely to be destroyed on a short time scale, while very large grains and grains in an optically thick cloud, both of which do not produce a pronounced emission feature, or reddening, are likely to survive longer.