Seyfert Galaxies in the Local Universe: Analysis of Spitzer Spectra of a Complete Sample
S. Tommasin, L. Spinoglio, K. Hainline, M.A. Malkan, H. Smith
aa r X i v : . [ a s t r o - ph ] J u l Mem. S.A.It. Vol. 79, 1 c (cid:13) SAIt 2008
Memorie della
Seyfert Galaxies in the Lo al Universe: Analysisof Spitzer Spe tra of a Complete Sample
S. Tommasin , L. Spinoglio , K. Hainline M. A. Malkan and H. Smith Istituto di Fisica dello Spazio Interplanetario, INAF, Via Fosso del Cavaliere100, I-00133 Roma, Italy e-mail:
[email protected],[email protected] Astronomy Division, University of California, Los Angeles, CA 90095-1547, USA e-mail: [email protected], [email protected] Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138e-mail: [email protected]
Abstract.
The Spitzer high resolution spectra of 72 Seyfert galaxies from the 12 µ m GalaxySample are presented and discussed. The presence of starburst components in these galaxiescan be quantified by powerful mid-IR diagnostics tools (i.e. 11.25 µ m PAH feature equiv-alent width and the H emission line intensity), as well as the AGN dominance can bemeasured by specific fine structure line ratios (e.g. [NeV] / [NeII], [NeV] / [SiII], etc.). Thetwo types of Seyfert galaxies do not show any statistical di ff erence in our diagnostic tools.However, the Seyfert 2’s showing hidden Broad Line Regions in spectro-polarimetric ob-servations have on average an higher AGN dominance, a weaker star formation componentand a warmer [60 - 25] spectral index than those without broad emission lines. Key words.
Galaxies: Active - Galaxies: Starbursts - Infrared: Galaxies
1. Introduction
From the observational evidence that the mainenergy-generating mechanisms in galaxies areblack hole accretion and star formation andthat Starbursts and AGNs may be linked inan evolutionary sequence, we present mid-infrared spectroscopy of a complete flux-limited sample of Seyfert galaxies with theaim to derive the bi-variate AGN and StarFormation luminosity functions in the LocalUniverse. Mid-infrared spectra can provide acensus of the dominance of the two processesat zero redshift and it will be the basis for any
Send o ff print requests to : S. Tommasin future comparison with the history of the en-ergy production mechanisms along galaxy evo-lution as derived from future cosmological sur-veys. Because most of the local active galax-ies contain both an AGN and a starburst com-ponent, often obscured by dust, infrared spec-troscopy has to be used to separate the twoemission processes. Many spectroscopic indi-cators such as the ratios of high to low ion-ization lines, e.g. [NeV] / [NeII], [NeV] / [SiII],[OIV] / [NeII], [NeIII] / [NeII] are directly linkedto the AGN dominance, while others are indi-cators of the Star Formation dominance, e.g.the PAH 11.25 µ m, the molecular hydrogen ro-tational lines and some nebular lines (e.g. from[NeII] and [SIII]) equivalent widths. Tommasin et al.: Spitzer spectra of Seyfert galaxies
2. The 12 µ m Sample Dust absorbs the continuum at short wave-lengths and re-emit it in the FIR, howeverthere is a spectral interval (7-12 µ m) at whichthe absorption of the original continuum isbalanced by the thermal emission. The 12 µ msample (hereafter 12MGS, Rush, Malkan &Spinoglio 1993, hereafter RMS) is an IRAS12 µ m-selected all-sky survey flux-limited to0.22Jy. From the 12MGS, a relatively unbi-ased sample of active galaxies in the localUniverse has been extracted using optical spec-troscopic classification. It is a complete sam-ple in bolometric flux, from the experimentalevidence that the flux at 12 µ m is about 1 / Fig. 1.
Example of an high resolution IRS
Spitzer spectrum, of IRAS15091-2107. The PAH featuresat 11.25 µ m and the H rotational lines can be seen,as well as the bright mid-IR fine-structure lines orig-inated from either the AGN and from stellar evolu-tion processes. Spitzer
IRS spectra
About half (53 objects) of the active galaxiesof the 12MGS have been observed by SpitzerIRS at high resolution, in a GT program by G.Fazio and the CfA team in collaboration withus. The data reduction of the first 30 sourcesobserved and their analysis have been reportedin Tommasin et al. (2008a, hereafter paper I)and a three-component photoionization modelhas been computed to separate the emission of28 sources in three emission line components:Seyfert NLR, starburst and ”coronal line re-gion” (Hainline et al. 2007).The other half of the sample is in theSpitzer public archive and its reduction andanalysis are in progress. A second article withother 42 spectra is in preparation (Tommasinet al. 2008b, in prep., hereafter paper II) anda subsequent article will present the overallcomparison of data with models (Hainline etal. 2008, in prep.). Fig.1 shows, as an exam-ple of the data quality, the IRS spectrum ofIRAS15091-2107 (paper I). We anticipate heresome of the new results of paper II dealing with72 objects and therefore with a larger statisticsthan in paper I.The ratio of lines of the same element atdi ff erent ionizing potential gives the galaxyradiation field: [NeV]14.32 µ m / [NeII]12.81 µ mand [NeIII]15.56 µ m / [NeII]12.81 µ m areamong the prime AGN tracers (Spinoglio &Malkan 1992, Sturm et al. 2002), while theintensity of the features of the PolyciclycAromatic Hydrocarbons are inversely propor-tional to the AGN activity, as it was suggestedfrom the analysis of ISO spectra of a sampleof ULIRGs (Genzel et al. 1998). We find thesame trend for the Seyfert galaxies of oursample: the PAH equivalent width decreasesas the [NeV]14.32 µ m / [NeII]12.81 µ m ratioincreases (originating [NeII] at 12.81 µ m inthe SF regions) (Fig.2). Besides the di ff erencebetween type 1’s and type 2’s, in this figurewe also distinguish among the Seyfert 2’swith a measured hidden broad line region(h-BLR) and those without (pure-Seyfert 2’s)and those without polarization measurements.The classification of the Seyfert type 2’s inh-BLR and pure-Seyfert 2’s, as well as the ommasin et al.: Spitzer spectra of Seyfert galaxies 3 Fig. 2.
The PAH 11.25 µ m equivalent width as afunction of the [NeV] / [NeII] line ratio, which mea-sures the AGN strength. Red squares: Seyfert type1; Pink pentagons: Seyfert type 2 with hidden BLR;Cyan exagons: Seyfert type 2 without hidden BLR = pure S2; Blue triangles: Seyfert type 2 with nopolarimetric data; Green circles: Normal / starburstGalaxy. Full symbols: Detection, Open symbolswith arrow: Upper Limit. reclassification of Seyfert galaxies in normalor starburst galaxies, has been taken from Tran(2001, 2003) and Shu et al. (2007) and theformer is based on optical spectro-polarizationobservations. It appears from the figure that al-most all galaxies (23 /
25) having [NeV]14.3 µ mdetections with the | PAH 11.25 µ m | < µ m have a BLR (14 type 1’s, 9 h-BLR),while only two objects are ”pure-Seyfert 2’s”and 4 objects do not have polarimetric data.This is confirmed by a Kolmogorov-Smirnov2-dimensional test that shows that Seyfert 1’sand h-BLR Seyfert 2’s are drawn from thesame population (with P = / [NeII] ratio and therefore aweaker AGN component. This result will befurther discussed in paper II.As already shown in paper I, we presentin Fig.3 the [NeIII]15.5 µ m / [NeII]12.8 µ m lineratio versus the (60 µ m -25 µ m) spectral index,with an higher statistics and, as in the previ-ous figure, the separation between Seyfert 2’s Fig. 3. [NeIII]15.5 µ m / [NeII]12.8 µ m line ratio ver-sus the (60 µ m -25 µ m) spectral index. Symbols as inthe previous figure. Most broad-lined objects havea higher [NeIII] / [NeII] ratio and a warmer spectralindex than objects without a BLR. with hidden broad lines and those without. Theionization in general increases with the flatten-ing of the spectral index. Using a Kolmogorov-Smirnov 2-dimensional test, we find that theSeyfert type 1 and type 2 should be drawnby the same population (P =
4. X-ray Counterparts
Objects obscured at soft X-rays but detectableat hard X-rays are defined Compton-thick, asthey present a very high absorption hydro-gen column density (N H > cm − ) (e.g.Guainazzi 2006). On the other side, objectswith a lower N H ( < cm − ) are calledCompton-thin. A large fraction of the 12MGSSeyferts are Compton-thick, as can be seen in Tommasin et al.: Spitzer spectra of Seyfert galaxies
Fig.4, which plots their IR luminosity versusthe hydrogen column density. On the contrary,hard X-rays are not very e ffi cient in detect-ing Compton-thick objects, especially at highluminosities, as can be seen from the figurewhich includes the AGN samples observed athard X-rays by Integral-IBIS and Swift-BAT(Bassani et al. 2006, Markwart et al. 2005).Therefore one of the advantages of the 12MGSis that it allows a combined X-ray / IR study thatcan help in understanding the role of AGN inthe IR galaxy population and the relationshipbetween the accretion power, the obscurationand the star formation properties of the galax-ies.
Fig. 4. µ m Sample Seyferts cover a wide range ofhydrogen column density, from Compton-thin ob-jects to Compton-thicks, while the X-ray selectedsources cover a range. (N H data from Bassani et al.2006, Markwart et al. 2005 and Shu et al. 2007)
5. Conclusions
In Tommasin et al. (2008a) we found no intrin-sic IR spectroscopic di ff erence between type1 and type 2 Seyferts. We confirm this re-sult, however this study, which increases thestatistics from 30 to 72 objects, indicates thatthose Seyfert type 2 that have an hidden BLR,as seen from spectro-polarimetry, lie preferen- tially in the regions of Seyfert type 1’s, forwhat concerns the AGN dominance and [60- 25] spectral index, while ”pure” Seyfert 2’shave stronger ”star formation” components, asmeasured by the PAH EW.The fact that our sample of galaxies wellcovers the plane luminosity-hydrogen columndensity containing Compton-thick objects evenat high luminosities makes it a suitable sam-ple to study the three parameters governinggalaxy evolution: accretion luminosity (andblack-hole properties), star formation strengthand obscuration. Acknowledgments
This work is based on ob-servations made with the Spitzer Space Telescopewhich is operated by the Jet Propulsion Laboratoryand Caltech under a contract with NASA. ST ac-knowledges support by ASI. We thank GiovanniFazio and the IRAC Team at CfA for contributingGuaranteed Time to obtaining these data.
References
Bassani L. et al. 2006, ApJ, 636, 65,Genzel, R. et al, 1998, ApJ, 498, 579Guainazzi M. 2006, astroph //