Active Galactic Nucleus Dusty Torus And Its Infrared Emission

Torus is the key component of active galactic nucleus(AGN)unification.These tori surround the accretion disks(ADs)and broad-line regions(BLRs),and lead to various observational types of AGNs by obcuring ADs and BLRs at different viewing angles,e.g.,optical spectral type-1/2 AGNs,X-ray obscured/unobscured AGNs?Torus is heated by ultraviolent and optical emission from the central region and re-emits the reprocessed energy to infrared band which peaks at 20-70 ?m.Studying the infrared emission of AGNs does not only help us understand the composition and distribution of torus,but also help isolate galaxy infrared emission from AGN contamination so that we could infer the star-formation rate of a galaxy based on their infrared luminosity.There have been several recent works focusing on SEDs of AGNs,but the results are controversial,especially on the far-IR contribution of AGNs.To investigate the origin of the conflicting results,we study the intrinsic SEDs of AGNs at IR bands with 42 z<0.5 optically luminous Palomar-Green(PG)survey quasars through SED decomposition.We collect the mid-IR low-resolution IRS spectra and far-IR band photometric data for each source.Then we build the SEDs by combining the re-sampled IRS spectra with photometric data.We decompose the SEDs(between 6-500 ?m)of the 42 quasars by combining an AGN IR template library from Siebenmorgen et al.(2015)that covers a wide range of the AGN parameter space with three commonly-used galaxy template libraries,and obtain a best-fitting intrinsic AGN SED for each PG quasar.We determine our normalized median AGN SED from the best-fitting results.Our conclusions do not depend on the choice of galaxy template libraries.The far-IR contribution(>100?m)of our median AGN SED is significantly smaller than that of Symeonidis et al.(2016),but roughly consistent with that of Mullaney et al.(2011)and Lyu et al.(2017).This may be a result of the mismatch of S16 using PAH flux to search for best fitting galaxy component in DH02 templates,since the 11.3 ?m PAH line may overlay with 9.7 ?m broad silicate absorption line.There will be some sys-tematic bias when silicate absorption is not subtracted properly.To figure out the PAH flux dependence of galaxy template,we use a combination of power-law continuum with emission line for decomposition to derive the PAH flux for real IRS spectra.We then re-sample the best-fitting templates from the DH02-,DH14-and R09-library-fitted re-sults to obtain the corresponding mock IRS spectra,and derive the PAH flux for each mock spectrum as we do for real IRS spectra.We find significant difference between the results of the three template libraries.While the mock spectra of R09 can reproduce the PAH flux of real IRS spectra,both DH02 and DH14 are systematically biased.Since there is no silicate component in DH02 and DH14,it is not a surprising result.Our results indicate that the way S16 used PAH flux and DH02 template library to decom-pose AGN SED and their far-IR SED is unreliable.Meanwhile,our method does not depend on the choice of galaxy template library,which suggests that our result is more convincing.We also study the correlation between AGN SED color and luminosity,and find that AGN IR SED becomes cooler(smaller mid-IR to far-IR luminosity ratio)with in-creasing luminosity.One possible explanation is that more luminous AGNs tend to have lower dust sublimation temperatures,and thus lower median torus temperatures.To study the morphology of the torus and the dust distribution near the SMBH,we collect the X-ray derived column density(NH),14-195 keV luminosity,optical spectral type and Ha luminosity for 202 Swift-BAT X-ray survey AGNs,aiming to analysis the correlation between X-ray absorption and mid-IR silicate-line absorption(indicating the absorption of dust).We fit their IRS spectra and derive silicate strength for each source.72 out of 202 show an emission feature,and the rest show an absorption line,among which type-1.9 and type-2 AGNs mostly show an absorption feature.Our results also show an anti-correlation between silicate strength and X-ray column density,which is consistent with Shi et al.(2006).This result is expected by the unification model of AGNs,where type-2 AGNs are obscured by dust in the line of sight,resulting in a serious X-ray absorption and silicate absorption line.We also compare silicate-line derived V-band attenuation to the results derived from broad Ha line(Shimizu et al.,2018),and find no correlation between them.This difference may come from:(1)the torus is clumpy;(2)the relation between X-ray luminosity and broad Ha luminosity may not be as tight as expected.Lastly,we study the composition and distributon of dust by comparing our ob-served correlation of silicate optical depth(?Si)and X-ray column density(NHX)to the prediction of models.We calculate the ?Si-NHX distribution for a smooth torus model(Fritz et al.,2006)and a clumpy torus model(Nenkova et al.,2008a).Compared to a smooth torus,our observed result is more consistent with a clumpy torus model.We also find that at NH>1022 cm-2,most cases show an X-ray absorption greatly larger than silicate indicated absorption,which suggests that there is additional dust-free gas that also contributes to X-ray absorption and this gas may come from BLR or disk wind.