Active galactic nucleus torus models and the puzzling infrared spectrum of IRAS F10214+4724

We present a revised model for the infrared emission of the hyperluminous infrared galaxy IRAS F10214+4724 which takes into account recent photometric data from Spitzer and Herschel that sample the peak of its spectral energy distribution. We first present and discuss a grid of smooth active galactic nucleus (AGN) torus models computed with the method of Efstathiou & Rowan-Robinson and demonstrate that the combination of these models and the starburst models of Efstathiou and coworkers, while able to give an excellent fit to the average spectrum of Seyfert 2s and spectra of individual type 2 quasars measured by Spitzer, fails to match the spectral energy distribution of IRAS F10214+4724. This is mainly due to the fact that the nuSnu distribution of the galaxy falls very steeply with increasing frequency (a characteristic that is usually indicative of heavy absorption by dust) but shows a silicate feature in emission. Such emission features are not expected in sources with optical/near-infrared type 2 AGN spectral signatures. The Herschel data show that there is more power emitted in the rest-frame 20-50um wavelength range compared with the model presented by Efstathiou which assumes three components of emission: an edge-on torus, clouds (at a temperature of 610 and 200K) that are associated with the narrow-line region (NLR) and a highly obscured starburst that dominates in the submillimetre. We present a revised version of that model that assumes an additional component of emission which we associate with NLR clouds at a temperature of 100K. The 100K dust component could also be explained by a highly obscured hot starburst. The model suggests that the NLR of IRAS F10214+4724 has an unusually high covering factor (>= 17%) or more likely the magnification of the emission from the NLR clouds is significantly higher than that of the emission from the torus.