On the 10-micron silicate feature in Active Galactic Nuclei

The 10-micron silicate feature observed with Spitzer in active galactic nuclei (AGN) reveals some puzzling behavior. It (1) has been detected in emission in type 2 sources, (2) shows broad, flat-topped emission peaks shifted toward long wavelengths in several type 1 sources, and (3) is not seen in deep absorption in any source observed so far. We solve all three puzzles with our clumpy dust radiative transfer formalism. (1) We present the spectral energy distribution (SED) of SST1721+6012, the first type 2 quasar observed to show a clear 10-mic silicate feature in emission. We constructed a large database of clumpy torus models and performed extensive fitting of the observed SED, constraining several of the torus parameters. We find that the source bolometric luminosity is ~3*10^12 L_sun. Our modeling suggests that <35% of objects with tori sharing characteristics and geometry similar to the best fit would have their central engines obscured. This relatively low obscuration probability can explain the clear appearance of the 10-mic emission feature in SST1721+6012 together with its rarity among other QSO2. (2) We also fitted the SED of PG1211+143, one of the first type 1 QSOs with a 10-mic silicate feature in emission. Among similar sources, this QSO appears to display an unusually broadened feature whose peak is shifted toward longer wavelengths. Although this led to suggestions of non-standard dust chemistry in these sources, our analysis fits such SEDs with standard galactic dust; the apparent peak shifts arise from radiative transfer effects. (3) We find that the distribution of silicate feature strengths among clumpy torus models closely resembles the observed distribution, and the feature never occurs deeply absorbed. (abridged)