The Mid-Infrared Continua of Seyfert Galaxies
Rajesh. P. Deo, Gordon. T. Richards, D. M. Crenshaw, S. B. Kraemer
aa r X i v : . [ a s t r o - ph . C O ] O c t To appear in ApJ November 2009 Issue 1
Preprint typeset using L A TEX style emulateapj v. 08/22/09
THE MID-INFRARED CONTINUA OF SEYFERT GALAXIES
Rajesh. P. Deo , Gordon. T. Richards , D. M. Crenshaw , S. B. Kraemer (Received 2009 January 7; Accepted 2009 September 3) To appear in ApJ November 2009 Issue 1
ABSTRACTAn analysis of archival mid-infrared (mid-IR) spectra of Seyfert galaxies from the
Spitzer Space Tele-scope observations is presented. We characterize the nature of the mid-IR active nuclear continuum bysubtracting a template starburst spectrum from the Seyfert spectra. The long wavelength part of thespectrum contains a strong contribution from the starburst-heated cool dust; this is used to effectivelyseparate starburst-dominated Seyferts from those dominated by the active nuclear continuum. Withinthe latter category, the strength of the active nuclear continuum drops rapidly beyond ∼ µ m. Onaverage, type 2 Seyferts have weaker short-wavelength active nuclear continua as compared to type1 Seyferts. Type 2 Seyferts can be divided into two types, those with strong poly-cyclic aromatichydrocarbon (PAH) bands and those without. The latter type show polarized broad emission linesin their optical spectra. The PAH-dominated type 2 Seyferts and Seyfert 1.8/1.9s show very similarmid-IR spectra. However, after the subtraction of the starburst component, there is a striking sim-ilarity in the active nuclear continuum of all Seyfert optical types. PAH-dominated Seyfert 2s andSeyfert 1.8/1.9s tend to show weak active nuclear continua in general. A few type 2 Seyferts withweak/absent PAH bands show a bump in the spectrum between 15 and 20 µ m. We suggest that thisbump is the peak of a warm ( ∼
200 K) blackbody dust emission, which becomes clearly visible whenthe short-wavelength continuum is weaker. This warm blackbody emission is also observed in otherSeyfert optical sub-types, suggesting a common origin in these active galactic nuclei.
Subject headings: galaxies: active, galaxies: Seyfert, infrared: galaxies INTRODUCTION
The complete mid-infrared (mid-IR) spectra of Seyfertgalaxies (a class of active galactic nuclei (AGN): Seyfert1943) have been available only in recent past (e.g.,Sturm et al. 2002; Verma et al. 2005; Weedman et al.2005; Deo et al. 2007). Many
Spitzer spectra of nearbySeyfert galaxies show a strong contribution from star-forming features in the form of poly-cyclic aromatichydrocarbon (PAH) bands (e.g., Clavel et al. 2000;Buchanan et al. 2006; Tommasin et al. 2008). A num-ber of these features, such as the 7.7 µ m and the 17 µ mPAH complex, are strongly blended with each other,complicating the estimation of the underlying active nu-clear continuum. Further, the mid-IR opacity includesa strong contribution from the silicate bands at 10 and18 µ m. Thus, estimating the intrinsic active nuclear con-tinua in the mid-IR is a non-trivial task, with the primaryhurdle being the subtraction of the starburst component.Previous studies have hinted that the continuum inthe 1–8 µ m range is non-stellar in origin and is likely aresult of thermal emission from dust heated close to sub-limation temperature by the optical/ultra-violet contin-uum from the central source (Edelson & Malkan 1986;Alonso-Herrero et al. 2001; Imanishi & Alonso-Herrero2004; Mushotzky et al. 2008). In a multi-wavelengthphotometric study of spectral energy distributions Department of Physics, Drexel University, 3141, ChestnutStreet, Philadelphia, PA 19104-2816, USA; [email protected] [email protected] Department of Physics and Astronomy, Georgia State Univer-sity, Atlanta, GA 30303, USA; [email protected] Catholic University of America, and the Exploration of theUniverse Division, NASA’s Goddard Space Flight Center, Code667, Greenbelt, MD 20771, USA; [email protected] (SED) of predominantly radio-quiet Sloan Digital SkySurvey (SDSS) quasars, Richards et al. (2006) andGallagher et al. (2007) noted that the 1–8 µ m spectralindex ( α ν ) is strongly anti-correlated with infrared lu-minosity in type 1 quasars. The more luminous quasarshave flatter 1–8 µ m slopes. A linear correlation betweenthe optical continuum luminosity and the infrared lumi-nosity suggests that the observed bump around ∼ . µ min the SED is driven by the dust re-emission. For ex-ample, the 2 . µ m bump is clearly visible in the near-IR spectrum of Mrk 1239 (Rodr´ıguez-Ardila & Mazzalay2006).Due to the presence of strong PAH bands in the 5–8 µ m range in many Seyfert spectra, direct measurementof the active nuclear continuum in this region is only pos-sible in sources with very weak or absent PAH features.An alternative is to subtract the starburst contributionusing a template starburst spectrum and then study theresidual continuum. We take this simple approach inthis paper. In Section 2, we describe our archival sampleand the data analysis techniques. We discuss the ob-served mid-IR spectra in Section 3. In Section 4, we usesimple continuum diagnostics that allows us to classifySeyfert mid-IR spectra into PAH-dominated and AGN-dominated groups and understand continuum properties.In Section 5, we discuss the starburst contribution inSeyfert spectra and the resulting continuum shapes af-ter subtraction of the starburst template spectrum. InSection 6, we summarize our results. SAMPLE SELECTION AND DATA ANALYSIS
We consider a sample of Seyfert galaxies derivedfrom the
Spitzer Space Telescope archives. This sam-ple is listed in Table 1 with redshifts obtained from the Deo et al.NASA/IPAC Extragalactic Database (NED) along withthe Spitzer archive numbers (AORKEY). We primarilyuse archival InfraRed Spectrograph (IRS; Houck et al.2004) spectra extracted from programs 3069 (PI: J. Gal-limore, mapping mode spectra ), 3374 (PI: S. Kraemer,staring mode spectra) and Weedman et al. (2005, ; Mrk3, low-resolution staring mode spectrum). Deo et al.(2007) presented single slit extractions of spectra fromabove mentioned programs for about half of the currentsample. During the initial work on that paper, CUBISMsoftware was not yet available (Smith et al. 2007a). Inthis paper, we expand that sample and consider the com-plete IRS mapping-mode datasets from program 3069along with other archival datasets now available. Thissample was chosen due to its large size and good samplingof different Seyfert optical types. We also test if the useof complete mapping-mode spectra leads to changes inthe short wavelength continuum. After the initial anal-ysis, we noted that our sample lacked type 2 Seyfertswith detected polarized broad emission lines. So, weadded 16 Seyfert 2s selected from the compilation of Tran(2003) for which low resolution IRS spectra are availablefrom the archive. These objects come from a numberof different programs and their AOR numbers are listedin Table 1. Further, we obtained optical Seyfert classi-fications from the compilation of V´eron-Cetty & V´eron(2006, VCV, hereafter) and host galaxy minor-to-majoraxis ratios ( b/a ) from NED. These are listed in Table 1.This paper contains a sample of 109 Seyfert galaxies withlow-resolution IRS spectra.The mapping-mode observations were reduced as be-low. From the low-resolution basic calibrated data(BCD) spectral images (for short-low, SL and long-low,LL modules), we built three-dimensional spectral cubes(two spatial ( x , y ) and one wavelength ( z ) dimension) us-ing the CUBISM software package (Smith et al. 2007a).We wrote Interactive Data Language (IDL) routines toscript internal CUBISM routines to automate the reduc-tion of spectral cubes. The pixel backtracking facilityavailable within CUBISM was used to clean the cubesof bad pixels. One-dimensional spectra were extractedfrom these cubes using a common rectangular aperturefor all four modules. These extraction apertures (see Ta-ble 1, Column 9) were selected to encompass as small aregion as possible in the LL1 module without introduc-ing pixel-aliasing effects in the extracted one-dimensionalLL1 spectra. We note an increase in the absolute SLflux density values as compared to single slit extractions.This is expected due to the different pixel scales of theSL and LL modules and the larger aperture area as com-pared to a single slit. The extracted one-dimensionalspectra were then imported into the SMART softwarepackage (Higdon et al. 2004) to normalize the separatemodules. In most cases no multiplicative shifts werenecessary due to matched aperture extractions. Whencorrections were necessary, the correction factor was al-ways less than 10%. The ends of the spectral orders wereclipped to remove data points affected by the reducedspectral response. The spectra were then re-binned to acommon wavelength grid.The staring-mode spectra were extracted as below. We See: http://ssc.spitzer.caltech.edu/documents/SOM/ started with BCD products obtained from the
Spitzer archive. We median-combined multiple data collectionevent (DCE) image files into one image per module, or-der and nod-position. We differenced images from op-posite orders to remove the sky background and thenextracted spectra using the tapered-column extractionoption within SMART. For cases, when differencing op-posite orders was not possible, we used opposite nod posi-tions. All the image and spectrum extraction operationswere carried out inside SMART. The extracted spectrawere cleaned by removing deviant data points and nor-malizing the spectra from different modules to form thecomplete mid-IR spectrum. Figure 1 shows the completemid-IR spectra grouped according to their Seyfert typesfrom Table 1.We measured the mid-IR continuum at 5 .
5, 10, 14 .
7, 20and 30 µ m, averaging within a window of 1 µ m, on therest-frame spectra. These specific spectral regions wherechosen to minimize the contribution from emission linesor PAH bands, thus primarily sampling the mid-IR con-tinuum. We also measured the optical depth at 9 . µ m.The silicate strength commonly quoted in Spitzer stud-ies is S . = ln( f λ /f C ), which translates to − τ . in ourcase. We follow the convention that negative opticaldepth implies a silicate emission feature. The peaks ofthese emission features typically occur around 10 . µ m.We present optical depth as measured at 9 . µ m for suchsources. The continuum and optical depth measurementsare given in Table 2. We also measured the equivalentwidths and fluxes of PAH bands and the narrow emis-sion lines. In this paper, we restrict ourselves only tocontinuum and optical depth measurements.In all the figures ahead, Seyfert 1s (including 1.2s and1.5s) are represented by filled circles (types S1, S1.0,S1.2, and S1.5 from VCV). Seyfert 2s with broad emis-sion lines detected in their polarized optical spectra (typeS1h in VCV; and called hidden broad-line region, HBLR,in Tran 2003) are represented by filled diamond sym-bols. Seyfert 2s with undetected broad emission linesin their polarized optical spectra (type S2 from VCV,and called non-HBLR in Tran 2003) are represented withopen squares. Seyfert 1.8/1.9s are represented with opentriangles (S1.8 and S1.9 from VCV). Galaxies classi-fied as LINER (type S3 from VCV) are represented bycross symbols. Each figure includes legends describingall Seyfert optical sub-types mentioned here. THE MID-IR SPECTRA
Complete mid-IR spectra of the sample are presentedin Figure 1. The top left panel shows type 1 Seyfertspectra. The top right panel shows type 2 Seyferts withbroad-line region (BLR) detected in polarized light (e.g.,Antonucci & Miller 1985; Tran 2003). The bottom leftpanel shows Seyfert 1.8/1.9s; these spectra are domi-nated by starburst-related PAH emission. These nucleiare likely weak emitters with a dominant contributionfrom circum-nuclear starbursts (Deo et al. 2007). Thebottom right panel shows type 2 Seyferts with dominantstarburst contribution (strong PAH bands). Clearly themid-IR spectra of Seyfert 1.8/1.9s are very similar tothese type 2 Seyferts. Polarized broad emission lineshave not yet been detected in these PAH-dominated type2 Seyferts. This suggests that the starburst contributionfrom the host galaxy dominates over the active nuclearhe Mid-Infrared Continua of Seyfert Galaxies 3continuum in these sources. One of the goals of this studyis to understand if these starburst-dominated nuclei havesimilar mid-IR continuum properties as starburst-weaktype 1 and type 2 nuclei.Examination of the spectra in Figure 1 shows thaton average, the continuum is similar between all fourclasses. PAH-weak type 2 spectra (top right) showweaker and steeper short-wavelength continuum thanPAH-weak type 1 spectra (top left) which are flatter.This is a direct evidence that short-wavelength mid-IRemission is absorbed to a certain degree in type 2 objectsby the intervening dust torus.Previous research work (e.g., Edelson & Malkan1986; Rodriguez Espinosa et al. 1996; Klaas et al. 2001;Schweitzer et al. 2006) has suggested that the mid-IRSED is likely composed of three thermal components:hot ( ∼ ∼
200 K) and cold ( ∼
60 K) dustemissions. This is reflected in the
IRAS
25 and 60 µ mphotometry with the active nuclear SED being “warm”compared to the “cool” starbursts.Almost all spectra show a curious emission peak/bumpbetween 15 and 20 µ m. This emission should not beconfused with the silicate 18 µ m peak (which is alwaysmuch weaker than the 10 µ m peak) or the 17 µ m PAHcomplex. The feature is broad and the peak wavelengthvaries between 15 and 20 µ m. This bump is observedmost clearly in the top right panel of Figure 1 (S1hsources), as the short-wavelength continuum is weak inthese sources. The presence of this feature in many otherspectra (some with strong PAH contributions also) sug-gests a common origin in most Seyfert sub-types.In Figure 2, we subtract the spectrum of Mrk 335 (aSeyfert 1.2) from the starburst-subtracted spectrum ofMrk 766 (a Seyfert 1.5). The Mrk 335 spectrum is apower-law-like spectrum with no hint of the 15–20 µ mbump. The residual Mrk 766 spectrum shows the 15–20 µ m bump. We compare the residual spectrum tothe Mrk 3 (type 2, S1h) spectrum. Both spectra arealmost identical. This suggests that the 15–20 µ m bumpis only visible when the power-law-like hot dust com-ponent is being absorbed, as in type 2 sources. Thus,it seems that the continuum shape of the mid-IR AGNspectrum depends on which of the three hot, warm, andcold components is brighter than others. In type 1 spec-tra, the hot component dominates; in type 2 spectra, thewarm component dominates; while starburst-dominatedobjects show an excess of cold dust emission. THE OBSERVED MID-IR CONTINUUM
Figure 3 shows a plot of spectral indices for all galaxiesin the sample. Here, the spectral index is defined as α λ − λ = log ( f λ /f λ ) / log ( λ /λ ) (1)where f λ is the flux density in W cm − µ m − and wave-lengths are in µ m. The dotted lines at α λ (5 . .
7) = − . α λ (20–30) = − . See Section 5 ahead for details of the starburst subtractionprocess. The wavelength index α λ and the frequency index α ν are re-lated by α ν = − ( α λ + 2). Seyfert 1s are in the bottom left quadrant; and (3)PAH-dominated Seyferts (some non-HBLR Seyfert 2sand Seyfert 1.8/1.9s) lie mostly in the right quadrants.Positive values of α λ (20–30) indicate red and steep long-wavelength continua and a stronger contribution fromcold ( T ∼ T ∼ T ∼
200 K) dominating at 15–20 µ m, and (3) a “cool”component ( T ∼ α λ (5 . .
7) suggest a dominance of warm( T ∼
200 K) thermal components and relatively steepshort-wavelength continua in F λ . Objects that will fallin this quadrant will show enhanced 15–20 µ m bumpin the mid-IR spectra. This emission bump is likelydue to a warm thermal component peaking at thesewavelengths. We would like to clarify that this is notthe 18 µ m silicate feature, but rather a thermal modi-fied blackbody (dust) emission underneath it, the short-wavelength side of which is clearly visible only when theshort-wavelength continuum is weak. No object with the15–20 µ m bump as the dominant feature in the mid-IRspectrum shows 10 µ m in emission; this suggests that theshort-wavelength continuum due to the hot componentand any associated silicate emission features is being ab-sorbed.As the emission of the hot dust component increases,the spectra begin to show a power-law-like flattening atshort wavelengths: α λ (5 . .
7) becomes more negativeand the dominance of the warm thermal continuum de-creases. This power law is likely the Rayleigh–Jeans tailof the hot component with dust close to its sublimationtemperature. At any point along this type 2 to type 1 se-quence, adding a cold dust component to the spectrumshifts the spectrum toward positive (steep) α λ (20–30).NGC 7469, a Seyfert 1.5 with strong starburst contri-bution is identified in Figure 3. Beyond this point inthe right quadrants, we see that there are no Seyfert 1sand the region with α λ (20–30) & − . . µ m include a contribu-tion from nearby 6 . µ m PAH complex.Figure 4 compares the apparent silicate optical depthat 9 . µ m ( τ . ) with α λ (5 . . τ . and α λ (5 . . τ . ≤ .
4, Deo et al.we find a Spearman rank correlation, ρ = 0 .
46 with P null = 1 . × − . This suggests that as the incli-nation of observer’s line of sight changes from pole-onfor type 1 objects to edge-on for type 2 objects, the in-ner hot dust is obscured, in agreement with the unifiedmodels (Antonucci 1993). This result is in agreementwith trends noted by Hao et al. (2007) between contin-uum colors and silicate strength.Measurement of optical depth requires defining a con-tinuum which is a subjective process, leading to largeerrors in measured values of optical depths. We generatethe continuum by fitting a spline curve through certainpivot points. As mentioned above, pivots at 5.5, 14.7, 20and 30 µ m are used. Apart from these pivots, we alsouse additional points around 8 and 12 µ m to define theblue and red ends of the silicate absorption feature. Theprocess is very similar to the one presented in Figure 2 ofSpoon et al. (2007) and should produce identical resultsin most cases. As a test, in Figure 5, we present a com-parison between the 11.3 µ m PAH equivalent width andthe optical depth. The absence of any correlation in thefigure suggests that we are not biasing our continuumplacements when strong PAH bands are present in thespectra. Thus, our optical depth measurements shouldbe fairly accurate in most cases.A few objects in Figure 4 do not follow this trend andshow strong silicate optical depths τ . > .
5. In all suchobjects, the strong silicate absorption is a result of ab-sorption due to dust in the host galaxy rather than inthe immediate vicinity of the central source. This con-clusion is supported by objects that show very red long-wavelength continua (NGC 3079 and Mrk 938). To in-vestigate the dependence of τ . on the inclination of thehost galaxy, we plotted τ . against the b/a of the hostgalaxy. This comparison is shown in Figure 6, whichindicates that for b/a < .
5, dust in the host galaxydisk can contribute significantly to the observed silicateabsorption and the long-wavelength continuum. A fewhigh-inclination galaxies (e.g., IC4329A) do not show sili-cate absorption, but rather silicate emission. We checked
Hubble Space Telescope images of IC4329A and note thatthe narrow-line region of this object is clearly visiblein the F533N (narrow-band optical filter around [O
III ] λ α λ (5 . .
7) with τ . for ob-jects with b/a > . τ . ) versus α λ (5 . . µ m) plot from Figure 4. On the left-hand side, as the α λ (5 . .
7) increases from − . .
0, the short wavelength continuum is graduallysuppressed, and we progress from strong type 1 source(bottom left) dominating the short-wavelength mid-IR tostrong type 2 source (top left) peaking between 15 and20 µ m. The left-hand side panels show the true behaviorof mid-IR active nuclear continua without contributionfrom external host galaxy features. As the apparent op-tical depth increases above 0.5, the spectra show strongercontribution from dust in the host galaxy, and enhancedPAH features (top right and bottom right in Figure 4).NGC 1194 (left, middle box in Figure 7) lacks these star-burst features, but shows one of the strongest silicate ab-sorption features at both 10 and 18 µ m. Figures 6 and 7validate the observation that many Seyfert galaxies withhighly inclined host galaxy disks are classified as type 2or type 1.8/1.9s (Keel 1980; Maiolino & Rieke 1995).These results imply that to study effects of dust in thetorus, we need to look at sources with − . . τ . . . α λ (20–30) . − .
5. A key property that dis-tinguishes active nuclear continua is the similarity of α λ (20–30) for both type 1 and type 2 Seyferts (see Fig-ure 1), and the fact that the 20–30 µ m continua are bluerthan the star formation dominated very red continua.For type 1.5 and type 2 Seyferts (S1h), the dominant15–20 µ m feature suggests that the dust structures re-sponsible for these features are warm ( T ∼
200 K), andthe spectral index in the 20–30 µ m is decided by theslope of the Rayleigh–Jean’s tail of this warm compo-nent. As can be seen from the top left panels in Figure 7,the short-wavelength continua are significantly weaker intype 2 sources as compared to type 1 sources (bottomleft). In the 20–30 µ m region of the spectrum, the differ-ences are smaller in agreement with previous observationby Buchanan et al. (2006). Non-HBLR Seyferts and Starburst Contributions
Based on the spectro-polarimetry survey of the CfAand the 12 µ m sample of Seyfert galaxies, Tran (2001,2003) suggested that there are two different populationsof Seyfert galaxies: (1) type 2 Seyferts that host a HBLR(S1h in Table 1), and hence are intrinsically identicalto a type 1 Seyfert; and (2) type 2 Seyferts that showno signatures of the BLR in spectro-polarimetry (non-HBLR) and have optical narrow-line ratios weaker thanthe HBLRs on the BPT diagram (Veilleux & Osterbrock1987; Baldwin et al. 1981). Hence, they are closer tothe sequence of star-forming galaxies on the BPT dia-gram. The term “non-HBLR” does not imply a lackof the BLR, but that it is not detected in the observa-tions that lead to their classification as S2. For example,Mrk 573 was originally classified as a non-HBLR. It hasbeen shown to contain a HBLR by Nagao et al. (2004)with higher signal-to-noise observations. Further workby Lumsden et al. (2004) and Haas et al. (2007) suggeststhat non-HBLR objects are weak AGN with a dominanthost galaxy component. Their mid-IR spectra shouldthen be dominated by star-forming features within thehost galaxies, which is what we find here.Five out of the 16 HBLR Seyferts that we added toour original sample show the enhanced emission bumparound 15–20 µ m and the rest show power-law like mid-IR continua similar to most Seyfert 1s. One out of 16shows strong PAH bands, and all of them have mod-he Mid-Infrared Continua of Seyfert Galaxies 5erately strong 10 µ m silicate absorption features. Thus,our comparison here provides qualitative agreement withthe picture of non-HBLRs as weak AGN with a strongstarburst component. Since we have complete mid-IRspectra, we can quantify starburst and AGN contribu-tions over the whole IRS spectral range. STARBURST CONTRIBUTION IN SEYFERT GALAXIES
Many studies have explored the question of sep-arating starburst contribution from AGN contri-bution using different diagnostics. Genzel et al.(1998) compared the strength of 7 . µ m PAH to[O IV ] 25 . µ m/[Ne II ] 12 . µ m emission line ratio, tostudy the AGN contribution in ultra-luminous infraredgalaxies (ULIRGs). Laurent et al. (2000) used spectraltemplates of H II regions, photo-dissociation regions andAGN to construct diagnostic diagrams. Sturm et al.(2002) suggested use of emission lines and continuumdiagnostics. Mel´endez et al. (2008b) used estimate of[Ne II ] from pure AGN sources to constrain AGNand starburst fraction in a sample of Seyfert galaxies.Nardini et al. (2008) used templates for starburst andAGN components in the 5–8 µ m region to estimate rel-ative strengths in ULIRGs.In Deo et al. (2007), we noted that the equivalentwidth of 6 . µ m PAH increases as α λ (20–30 µ m) be-came more positive, which suggested that this relationwas driven by the starburst content of the active galaxy.In Figure 3, we find that most Seyfert 1.8/1.9 and Seyfert2 galaxies dominated by PAH emission have positive α λ (20–30 µ m). To measure the contribution of starburstfeatures to individual galaxies on this diagram, we usethe average starburst galaxy spectrum from Brandl et al.(2006). We assume that the PAH inter-band ratios inthis average spectrum are representative of typical PAHinter-band ratios in starburst galaxies. We are awarethat PAH inter-band ratios can be different within dif-ferent star-forming template spectra (e.g., Peeters et al.2004; Smith et al. 2007b), and depend particularly onthe strength of the starburst ionization field. Further,Sturm et al. (2006) also show that PAH ratios can bedifferent in LINER spectra. We find similar variationsin our Seyfert spectra also. We note that the strength ofthe 6.2, 7.7, and the 8.6 µ m PAH complexes vary muchmore than others. As a first-order comparison, assuminga fixed PAH template is acceptable. We also assume thatthe contribution of cool dust to the long-wavelength con-tinuum in this average spectrum is typical of star-forminggalaxies. This simple subtraction of a scaled starbursttemplate assumes that any obscuration in the object isentirely due to the torus and not due to the surround-ing starburst. We normalize the template starburst spec-trum, so that the peak flux density of the 6.22 PAH bandis unity. Then, we scale and subtract this template bytrial and error from the Seyfert spectra. We require thatthe 17 µ m PAH complex and the 11 . µ m PAH bandsbe cleanly subtracted. Figure 8 shows examples of suchsubtractions.Out of 107 sources for which we performed the star-burst subtraction, 50 show power-law-like continuumover the whole IRS range with weak silicate emission fea-tures, after the subtraction. Twenty four objects showsilicate absorption at 10 µ m. Seven objects show strongsilicate emission features with f µ m > f µ m and an underlying power-law-like continuum. We find 26 objectswhere the 15–20 µ m emission is much more prominentthan the 10 µ m emission.The majority of Seyfert galaxies in this sample showsimilar PAH inter-band ratios as in the starburst tem-plate spectrum. However, in a few cases like Mrk 477 inFigure 8, we find that the PAH 6 . . µ m bandsare over-subtracted. This indicates that PAH inter-bandratios in AGN spectra are not always similar to thosein starburst galaxies. Assuming that the 11 . µ m and17 µ m bands represent the actual strength of the star-burst in these sources, we find a deficit of emission at6 . . µ m PAH bands in some Seyfert galaxies. Onthe other hand, in sources clearly dominated by starburstcontribution such as Mrk 938 and NGC 3079, we find anexcess of emission at 6 . . µ m, after PAH bands at11 . µ m are cleanly subtracted. We note that theover-subtraction of PAH 6 . . µ m bands occurs inSeyfert galaxies with weak short-wavelength mid-IR con-tinua. These variations in PAH ratios are interesting, asrelating the change in PAH inter-band ratios with the de-creasing intensity of the interstellar radiation field overtime, as the starburst fades, will provide crucial con-straints on the time required for the active nucleus tobecome recognizable after the starburst episode.The starburst subtraction process essentially yieldsthe AGN spectrum devoid of any star-forming features.These starburst-subtracted spectra are shown in Fig-ure 9. It is instructive to compare this figure with Fig-ure 1. Note the striking similarity of mid-IR active nu-clear continua between all Seyfert optical subtypes. Type1 spectra are flatter at short wavelengths than type 2and type 1.8/1.9 spectra. Emission lines are generallyweaker in PAH-dominated Seyfert 2s and Seyfert 1.8/1.9s(see Figure 1) suggesting weaker active nuclear continua(Deo et al. 2007; Mel´endez et al. 2008a).We measure the contribution from the starburst sub-tracted AGN spectrum at continuum wavelengths of 5.5,8, 10, 14.7, 20, and 30 µ m. In Figure 10, we plot theluminosity density at 5.5 and 20 µ m from the active nu-cleus and the starburst component. Starburst contri-butions at 10, 1, 0.1, 0.01, and 0.001 times the activenucleus contribution are also shown with thin lines. Thestarburst-to-active nuclear continuum ratios are given inTable 3 for reference. The galaxies with strong PAHbands (almost all Seyfert 1.8/1.9s and some Seyfert 2swith undetected polarized broad lines) indeed have star-burst luminosities almost as much as their active nucleuscontribution, but not larger at this wavelength. The sam-ple as a whole is weighted toward significant contribution( ∼ factor of 10) from the active nucleus component at5.5 µ m. Figure 10 (right) also shows similar comparisonat 20 µ m. This figure shows that the AGN contribu-tion decreases rapidly at longer wavelengths. Around ∼ µ m, the starburst and the active nuclear contribu-tion are similar. This validates our use of α λ (20–30 µ m)in Figure 3 to separate objects dominated by starburstcontribution.We construct spectral indices, as previously describedin Section 3, from these starburst-subtracted spectra.In Section 3, we could not use the 8 µ m flux densityto construct α λ (5 . . µ m) due to the contribution ofthe 7 . . µ m PAH bands. With the starburst-subtracted spectra this is now possible. A compari- Deo et al.son of α λ (5 . . µ m) with α λ (20-30 µ m) shows thatthe long-wavelength continua are now flatter after sub-traction of the starburst component (see Figure 11,top). We find that h α λ (5 . µ m) i = − . ± . h α λ (20-30 µ m) i = − . ± .
10. For comparison,the average values for these quantities in Figure 3 are − . ± .
06 and − . ± .
10, respectively. There isclearly a large change in α λ (20-30 µ m), confirming ourconclusion again from last paragraph. The main ef-fect of subtracting the starburst template spectrum isto shift the PAH-dominated Seyfert 1.8/1.9s and Seyfert2s to the left. Note that there are no Seyferts with α λ (20-30 µ m) > − . . µ m. Figure 11 (top) showsthat there are at least two types of dust distributionsin the active nuclear region, one that generates theshort-wavelength continua and other that generates thelong-wavelength continua. The major difference betweenthese two dust distributions is likely to be their meantemperatures. By subtracting the starburst and associ-ated cool dust contribution, we have essentially removedthe starburst component that contributes most to thevariety of Seyfert spectra (Buchanan et al. 2006).A comparison of α λ (5 . µ m) against luminosity den-sity at 5 . µ m (Figure 11, bottom) shows that Seyfert1s have h α λ (5 . µ m) i = − . ± .
08, while Seyfert2s (type S1h, S1.8, S1.9 and S2) have h α λ (5 . µ m) i = − . ± .
07. A Kolmogorov–Smirnov test of the twodistributions gives a probability of 4 . × − of null hy-pothesis that the two samples are drawn from the samedistribution. The value of α λ (5 . µ m) for Seyfert 1s isin agreement with estimates derived for type 1 quasars(Gallagher et al. 2007; Haas et al. 2003). SUMMARY
An analysis of archival
Spitzer Space Telescope mid-IR spectra of Seyfert galaxies is presented. We focus onunderstanding the intrinsic shape of the active nuclearcontinuum in the mid-IR region and how it relates toother properties of the source such as the 10 µ m silicateoptical depth. We assumed a template spectrum for thestarburst component, and subtracted it from the Seyfertspectra to separate the active nuclear contribution fromthe circum-nuclear starburst contribution. Our primaryconclusions from this study are as follows:1. Seyfert spectra are classified effectively betweenAGN- and starburst-dominated categories basedon the spectral indices, α λ (5 . . µ m) and α λ (20–30 µ m) (see Figure 3). Seyferts domi-nated by the AGN contribution have flatter spec-tra with α λ (5 . . µ m) ∼ − .
13. The addedstarburst contribution from the host galaxy inthe large
Spitzer aperture (see Table 1) makes α λ (20 . . µ m) more positive or steeper. Akey property that distinguishes AGN-dominatedspectra is that the 20–30 µ m continuum is flat-ter ( α λ ∼ −
2) than the very steep 20–30 µ mcontinuum of starburst-dominated objects ( α λ ∼ µ m continuum is likely formedfrom the Rayleigh–Jeans tail of the “warm” dustcomponent. It is likely that there are multi-ple “warm” components with different tempera- tures. Further, Type 2 Seyferts with polarizedbroad emission lines in their optical spectra (typeS1h) show α λ (5 . . µ m) ∼ − .
49 much dif-ferent than average type 1 Seyferts that show α λ (5 . . µ m) ∼ − .
13. Note the steeper andweak short-wavelength continuum in Figure 9, ascompared to type 1 Seyferts. This is a direct evi-dence for presence of the dust torus that blocks ourview of the hot dust closer in.2. After starburst subtraction, Seyfert 1.8/1.9s andSeyfert 2s with strong PAH features in their spectrashow similar active nuclear continuum as Seyfert2s with weak/absent PAH features in their mid-IR spectra and polarized broad emission lines intheir optical spectra (see Figure 9). This suggestspresence of similar quantities and/or properties ofdusty material around the central accretion disk inthese type 2 sources. Tran (2003) had proposedexistence of two types of Seyfert 2s: the HBLRand the non-HBLRs. We compared spectral in-dices in the 5–8 µ m region after starburst subtrac-tion and find that both the HBLR and non-HBLRshow similar spectral indices (Figure 11, bottom),suggesting similarity rather than differences be-tween the two classifications. While, non-HBLRstend to show stronger starburst contribution ascompared to their AGN contribution, the converse(that starburst-dominated systems lack BLR signa-tures) is not necessarily true. The additional hostgalaxy contribution likely complicates the identi-fication of BLR in these systems. As we showin Figure 3, simple continuum indices effectivelyseparate AGN-dominated Seyferts from starburst-dominated Seyferts in the mid-IR.3. Deo et al. (2007) showed that a large part of thesilicate absorption in some Seyfert galaxies comesfrom starburst-heated cold dust in the host galaxyrather than the dust torus. This conclusion wasbased on the fact that only highly inclined galaxiesshowed large silicate optical depths. Here, we putthat result on a better statistical basis by present-ing a correlation between the b/a and the measuredoptical depth for this sample of 109 sources (seeFigure 6). All objects with significant optical depthare highly inclined and are also classified as Seyfert2s or Seyfert 1.8/1.9s. This confirms previous re-sults by Keel (1980) and Maiolino & Rieke (1995),and highlights the importance of considering thehost galaxy contribution in concealing AGN.4. On average, Seyfert galaxies dominated by theAGN continuum tend to show weak silicate ab-sorption ( τ . . . α λ (5 . . µ m)) and the apparentsilicate optical depth τ . (Figure 4) may to be cor-related in AGN-dominated objects. Seyfert opti-cal types form a continuous sequence of increas-ing optical depth along this correlation from type1s, type 1.8/1.9s, to type 2s with HBLRs. Thisvalidates the general inclination dependence inher-ent in AGN models noted before in the mid-IR byHao et al. (2007). But, as can be noted in Fig-ure 4, there is not a strict relationship between thehe Mid-Infrared Continua of Seyfert Galaxies 7strength of silicate features and the optical spectraltype.5. Figure 11 (top) shows that there are at least twotypes of dust distributions in the active nuclearregion, one that generates the short-wavelengthcontinua and another that generates the long-wavelength continua. The major difference be-tween these two dust distributions is likely to betheir mean temperatures. By subtracting the star-burst and associated cool dust contribution, wehave essentially removed the starburst componentthat contributes most to the variety in Seyfertspectra (Buchanan et al. 2006). In Figure 2, wedemonstrate the existence of this “warm” compo-nent which dominates the long-wavelength contin-uum, by separating it from the hot dust componentin Mrk 766. This simple template subtraction ex-ercise provides proof that Seyfert spectra are pri-marily thermal in nature and composed of at leastthree thermal components with T ∼ ∼ µ m in type 1 Seyfert spectra is a result ofthis warm component being brighter at ∼ µ mthan the Rayleigh–Jeans tail of the hot component.The above-mentioned similarity of continua beyond ∼ µ m in AGN-dominated sources is also due tothis warm dust component being present in almostall observed AGN spectra.We would like to thank Nadia Zakamska for insight-ful comments on an early draft of this paper. G.T.R.acknowledges support from an Alfred P. Sloan ResearchFellowship. This work is based on archival data obtainedwith the Spitzer Space Telescope , which is operated bythe Jet Propulsion Laboratory, California Institute ofTechnology under a contract with NASA. This researchhas made use of the NASA/IPAC Extragalactic Database(NED) which is operated by the Jet Propulsion Lab-oratory, California Institute of Technology, under con-tract with the National Aeronautics and Space Admin-istration. This research has also made use of NASA’sAstrophysics Data System Bibliographic Services. TheIRS was a collaborative venture between Cornell Univer-sity and Ball Aerospace Corporation funded by NASAthrough the Jet Propulsion Laboratory and Ames Re-search Center. SMART was developed at Cornell Univer-sity and is available through the Spitzer Science Centerat Caltech.
Facility: Spitzer
REFERENCESAlonso-Herrero, A., Quillen, A. C., Simpson, C., Efstathiou, A.,& Ward, M. J. 2001, AJ, 121, 1369Antonucci, R. 1993, ARA&A, 31, 473Antonucci, R. R. J., & Miller, J. S. 1985, ApJ, 297, 621Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5Brandl, B. R., et al. 2006, ApJ, 653, 1129Buchanan, C. L., Gallimore, J. F., O’Dea, C. P., Baum, S. A.,Axon, D. J., Robinson, A., Elitzur, M., & Elvis, M. 2006, AJ,132, 401Clavel, J., et al. 2000, A&A, 357, 839Deo, R. P., Crenshaw, D. M., Kraemer, S. B., Dietrich, M.,Elitzur, M., Teplitz, H., & Turner, T. J. 2007, ApJ, 671, 124Edelson, R. A., & Malkan, M. A. 1986, ApJ, 308, 59Gallagher, S. C., Richards, G. T., Lacy, M., Hines, D. C., Elitzur,M., & Storrie-Lombardi, L. J. 2007, ApJ, 661, 30Genzel, R., et al. 1998, ApJ, 498, 579Haas, M., Siebenmorgen, R., Pantin, E., Horst, H., Smette, A.,K¨aufl, H.-U., Lagage, P.-O., & Chini, R. 2007, A&A, 473, 369Haas, M., et al. 2003, A&A, 402, 87Hao, L., Weedman, D. W., Spoon, H. W. W., Marshall, J. A.,Levenson, N. A., Elitzur, M., & Houck, J. R. 2007, ApJ, 655,L77Higdon, S. J. U., et al. 2004, PASP, 116, 975Houck, J. R., et al. 2004, ApJS, 154, 18Imanishi, M., & Alonso-Herrero, A. 2004, ApJ, 614, 122Keel, W. C. 1980, AJ, 85, 198Klaas, U., et al. 2001, A&A, 379, 823Laurent, O., Mirabel, I. F., Charmandaris, V., Gallais, P.,Madden, S. C., Sauvage, M., Vigroux, L., & Cesarsky, C. 2000,A&A, 359, 887Lumsden, S. L., Alexander, D. M., & Hough, J. H. 2004,MNRAS, 348, 1451Maiolino, R., & Rieke, G. H. 1995, ApJ, 454, 95Mel´endez, M., et al. 2008a, ApJ, 682, 94Mel´endez, M., Kraemer, S. B., Schmitt, H. R., Crenshaw, D. M.,Deo, R. P., Mushotzky, R. F., & Bruhweiler, F. C. 2008b, ApJ,689, 95 Mushotzky, R. F., Winter, L. M., McIntosh, D. H., & Tueller, J.2008, ArXiv e-prints:0807.4695MNagao, T., Kawabata, K. S., Murayama, T., Ohyama, Y.,Taniguchi, Y., Sumiya, R., & Sasaki, S. S. 2004, AJ, 128, 109Nardini, E., Risaliti, G., Salvati, M., Sani, E., Imanishi, M.,Marconi, A., & Maiolino, R. 2008, MNRAS, 385, L130Peeters, E., Spoon, H. W. W., & Tielens, A. G. G. M. 2004, ApJ,613, 986Richards, G. T., et al. 2006, ApJS, 166, 470Rodr´ıguez-Ardila, A., & Mazzalay, X. 2006, MNRAS, 367, L57Rodriguez Espinosa, J. M., Perez Garcia, A. M., Lemke, D., &Meisenheimer, K. 1996, A&A, 315, L129Schweitzer, M., et al. 2006, ApJ, 649, 79Seyfert, C. K. 1943, ApJ, 97, 28Smith, J. D. T., et al. 2007a, PASP, 119, 1133Smith, J. D. T., et al. 2007b, ApJ, 656, 770Spoon, H. W. W., Marshall, J. A., Houck, J. R., Elitzur, M., Hao,L., Armus, L., Brandl, B. R., & Charmandaris, V. 2007, ApJ,654, L49Sturm, E., Lutz, D., Verma, A., Netzer, H., Sternberg, A.,Moorwood, A. F. M., Oliva, E., & Genzel, R. 2002, A&A, 393,821Sturm, E., et al. 2006, ApJ, 653, L13Tommasin, S., Spinoglio, L., Malkan, M. A., Smith, H.,Gonz´alez-Alfonso, E., & Charmandaris, V. 2008, ApJ, 676, 836Tran, H. D. 2001, ApJ, 554, L19—. 2003, ApJ, 583, 632Veilleux, S., & Osterbrock, D. E. 1987, ApJS, 63, 295Verma, A., Charmandaris, V., Klaas, U., Lutz, D., & Haas, M.2005, Space Science Reviews, 119, 355V´eron-Cetty, M.-P., & V´eron, P. 2006, A&A, 455, 773Weedman, D. W., et al. 2005, ApJ, 633, 706
Deo et al.
TABLE 1Observation Summary, Redshifts, Optical Seyfert Types andExtraction Apertures.
Galaxy Redshift Seyfert b/a
Spitzer
Obs. Extraction Rectangle ParsecName Type AOR Mode R.A. (deg.) Dec. (deg.) Aperture ( ′′ ) per ′′ (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)3C 226 0.817700 S1h 1.00 11298560 S · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · he Mid-Infrared Continua of Seyfert Galaxies 9 TABLE 1 — Continued
Galaxy Redshift Seyfert b/a
Spitzer
Obs. Extraction Rectangle ParsecName Type AOR Mode R.A. (deg.) Dec. (deg.) Aperture ( ′′ ) per ′′ (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)NGC 4922 NED01 0.023860 S2 0.79 12477952 M 195.349342 29.317079 25.406x43.701 497.72NGC 4941 0.00370 S2 0.53 12471552 M 196.052465 -5.542940 25.400x47.925 76.01NGC 4968 0.00986 S2 0.47 12464128 M 196.781667 -23.674650 20.352x27.507 203.52NGC 5005 0.00316 S3b 0.48 12475648 M 197.741665 37.061589 18.499x24.887 64.89NGC 513 0.01954 S1.9 0.43 12467712 M 21.112954 33.797022 20.350x28.780 406.29NGC 5135 0.013693 S2 0.67 12445696 M 201.439919 -29.831243 18.495x27.497 283.46NGC 5256 NED01 0.027600 S2 0.75 12459264 M 204.581404 48.278533 20.353x30.119 577.35NGC 526A 0.01910 S1.9 0.53 12454912 M 20.987415 -35.061446 25.401x47.921 397.01NGC 5347 0.00779 S2 0.76 12481792 M 208.329504 33.492829 14.798x22.273 160.54NGC 5506 0.006181 S1i 0.24 12453888 M 213.317594 -3.205489 18.499x24.891 127.22NGC 5548 0.01717 S1.5 0.93 12481024 S · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · Note . — Col. (1): Galaxy name; Col. (2): Redshift, obtained from NASA Extragalactic Database (NED); Col. (3): Seyfert type from V´eron-Cetty & V´eron (2006)—S1:Seyfert 1 optical spectrum; S1h: broad polarized Balmer lines detected; S1i: broad Paschen lines observed in the infrared; S1n: narrow-line Seyfert 1; S1.0, S1.2, S1.5,S1.8, and S1.9: intermediate Seyfert galaxies (Note: In this paper, we consider S1.0 to S1.5 to be Seyfert 1s in all analysis.); S2: Seyfert 2 spectrum; S3: LINER; S3b:LINER with broad Balmer lines; S3h: LINER with broad polarized Balmer lines detected; H2: nuclear HII region; Col. (4): Axial ratio, minor-to-major axis ratio ofhost galaxy obtained from NED, for host galaxies where b/a was not available we assumed it to be 1; Col. (5):
Spitzer archive Astronomical Observation Request (AOR)number; Col. (6): Observing mode for
Spitzer/IRS spectrum: S: staring mode, M: mapping mode; Col. (7 and 8): Extraction rectangle for mapping-mode spectra: R.A.and Dec. of center-point in degrees; Col. (9): Extraction aperture in arc-seconds. Col. (10): Radial extent in parsecs per arc-second of the extraction aperture for thegalaxy.
TABLE 2Continuum Flux Densities and Apparent Optical Depth at . µ m . Galaxy Name Continuum Flux Density (Jy) Optical Depth5.5 µ m 10 µ m 14.7 µ m 20 µ m 30 µ m 9.7 µ m(1) (2) (3) (4) (5) (6) (7)3C 226 3.82E-03 4.73E-03 1.57E-02 2.03E-02 · · · · · · · · · -0.1688(1.50E-03) (1.19E-03) (1.08E-03)3C 321 1.19E-02 3.28E-02 1.78E-01 2.71E-01 5.15E-01 0.6349(1.42E-03) (5.12E-03) (1.53E-02) (1.37E-02) (2.16E-02)CGCG381-051 8.73E-02 1.24E-01 1.93E-01 3.62E-01 5.51E-01 -0.1339(6.20E-02) (1.13E-02) (6.00E-03) (1.65E-02) (8.89E-03)ESO 12-G21 6.02E-02 1.14E-01 1.40E-01 1.64E-01 2.55E-01 -0.1241(1.45E-02) (7.53E-03) (5.46E-03) (6.74E-03) (3.20E-03)ESO 33-G2 8.98E-02 1.89E-01 3.11E-01 3.56E-01 3.16E-01 -0.0620(1.84E-02) (1.69E-02) (7.48E-03) (9.98E-03) (4.86E-03)FSC 09104 5.98E-02 1.35E-01 2.75E-01 3.70E-01 · · · · · · · · · · · · (3.19E-02) (3.37E-02) (6.16E-02)IRAS 05189-2524 2.04E-01 3.95E-01 1.16E+00 2.01E+00 5.91E+00 0.3367(1.35E-02) (3.12E-02) (1.91E-02) (1.29E-01) (1.32E-01)IRASF 01475-0740 4.46E-02 1.61E-01 2.61E-01 4.37E-01 4.81E-01 -0.1827(1.10E-02) (1.57E-02) (1.03E-02) (2.00E-02) (7.02E-03)IRASF 03450+0055 1.39E-01 2.65E-01 3.47E-01 4.57E-01 3.87E-01 -0.1850(8.77E-03) (1.48E-02) (5.59E-03) (1.54E-02) (1.22E-02)IRASF 04385-0828 2.27E-01 2.38E-01 8.02E-01 1.07E+00 1.46E+00 0.7856(1.45E-02) (1.85E-02) (2.88E-02) (2.24E-02) (2.50E-02)IRASF 15480-0344 4.29E-02 1.53E-01 2.99E-01 4.22E-01 4.92E-01 -0.0815(9.36E-03) (2.09E-02) (1.86E-02) (5.74E-03) (8.65E-03)MCG+0-29-23 6.69E-02 1.01E-01 2.48E-01 3.83E-01 8.54E-01 0.0629(2.04E-02) (6.78E-03) (6.52E-03) (8.24E-03) (1.75E-02)MCG-03-34-064 1.75E-01 5.04E-01 1.39E+00 2.05E+00 2.95E+00 0.2911(2.53E-02) (6.26E-02) (5.68E-02) (1.39E-02) (5.14E-02)MCG-2-33-34 2.74E-02 6.44E-02 1.12E-01 1.59E-01 2.31E-01 -0.0590(6.77E-03) (1.87E-02) (1.90E-02) (3.64E-03) (6.59E-03)MCG-2-40-4 2.02E-01 3.27E-01 5.45E-01 6.78E-01 9.53E-01 0.1155(1.89E-02) (1.62E-02) (1.93E-02) (1.34E-02) (1.19E-02)MCG-2-8-39 3.12E-02 1.21E-01 2.53E-01 2.92E-01 2.22E-01 -0.0427(8.95E-03) (1.09E-02) (7.18E-03) (1.61E-02) (3.33E-03)MCG-3-34-63 1.13E-02 1.61E-02 2.31E-02 4.08E-02 1.01E-01 -0.0420(8.81E-03) (5.04E-03) (3.59E-03) (6.41E-03) (3.51E-03)MCG-3-58-7 1.47E-01 2.78E-01 4.76E-01 7.01E-01 9.38E-01 -0.0072(8.75E-03) (1.13E-02) (1.21E-02) (2.02E-02) (1.45E-02)MCG-5-13-17 5.27E-02 1.11E-01 2.10E-01 3.19E-01 4.19E-01 0.1221(1.13E-02) (9.16E-03) (4.28E-03) (4.38E-03) (5.80E-03)MCG-6-30-15 1.80E-01 3.34E-01 4.80E-01 6.45E-01 6.10E-01 -0.0771(1.34E-02) (2.20E-02) (8.65E-03) (1.05E-02) (6.27E-03)Mrk 1239 4.57E-01 7.29E-01 8.85E-01 1.03E+00 9.14E-01 -0.0910(1.60E-02) (2.24E-02) (2.70E-02) (3.31E-02) (1.42E-02)Mrk 3 1.02E-01 2.94E-01 1.25E+00 2.07E+00 2.36E+00 0.3461(1.12E-02) (4.24E-02) (7.25E-02) (2.26E-02) (2.27E-02)Mrk 334 4.81E-02 1.13E-01 2.52E-01 5.71E-01 1.36E+00 0.2496(7.58E-03) (7.10E-03) (9.27E-03) (2.39E-02) (4.01E-02)Mrk 335 1.44E-01 2.03E-01 2.53E-01 2.80E-01 2.62E-01 -0.1258(1.28E-02) (1.28E-02) (9.50E-03) (9.27E-03) (2.92E-03)Mrk 348 1.25E-01 1.88E-01 4.30E-01 5.67E-01 4.95E-01 0.3455(1.41E-02) (1.69E-02) (1.79E-02) (1.51E-02) (5.85E-03)Mrk 463E 2.47E-01 3.30E-01 8.39E-01 1.28E+00 1.54E+00 0.3582(1.42E-02) (3.13E-02) (2.08E-02) (1.53E-02) (1.44E-02)Mrk 471 7.44E-03 1.58E-02 2.69E-02 4.31E-02 9.11E-02 0.1419(1.59E-03) (1.15E-03) (2.13E-03) (3.01E-03) (3.88E-03)Mrk 477 2.44E-02 7.46E-02 2.31E-01 4.21E-01 6.18E-01 0.2195(3.23E-03) (9.39E-03) (1.34E-02) (5.48E-03) (6.51E-03)Mrk 609 1.76E-02 4.52E-02 9.80E-02 1.80E-01 4.35E-01 0.3291(5.24E-03) (2.70E-03) (6.13E-03) (7.76E-03) (9.05E-03)Mrk 6 1.45E-01 2.08E-01 3.72E-01 5.37E-01 4.98E-01 -0.0280(1.02E-02) (1.73E-02) (1.72E-02) (1.80E-02) (8.54E-03)Mrk 622 7.80E-03 3.78E-02 1.06E-01 2.55E-01 5.51E-01 0.1071(1.54E-03) (3.40E-03) (6.61E-03) (8.40E-03) (5.31E-03)Mrk 704 1.91E-01 3.54E-01 5.05E-01 5.22E-01 4.02E-01 -0.0163(2.05E-02) (1.90E-02) (9.12E-03) (1.85E-02) (6.29E-03)Mrk 766 1.49E-01 3.02E-01 6.95E-01 1.02E+00 1.38E+00 0.1533 he Mid-Infrared Continua of Seyfert Galaxies 11 TABLE 2 — Continued
Galaxy Name Continuum Flux Density (Jy) Optical Depth5.5 µ m 10 µ m 14.7 µ m 20 µ m 30 µ m 9.7 µ m(1) (2) (3) (4) (5) (6) (7)(1.62E-02) (2.41E-02) (2.73E-02) (8.78E-03) (2.56E-02)Mrk 79 1.77E-01 2.98E-01 4.66E-01 5.98E-01 6.91E-01 -0.0141(1.58E-02) (1.47E-02) (1.78E-02) (1.04E-02) (9.23E-03)Mrk 817 1.23E-01 2.91E-01 4.79E-01 8.08E-01 1.13E+00 -0.0577(2.39E-02) (2.17E-02) (2.37E-02) (3.47E-02) (1.31E-02)Mrk 883 5.80E-03 1.67E-02 4.49E-02 1.18E-01 3.03E-01 0.1741(1.26E-03) (2.41E-03) (1.62E-03) (7.17E-03) (7.93E-03)Mrk 9 1.04E-01 1.92E-01 2.62E-01 3.42E-01 3.91E-01 0.0281(1.92E-02) (1.21E-02) (6.06E-03) (7.71E-03) (7.69E-03)Mrk 938 1.15E-01 9.93E-02 4.99E-01 1.08E+00 4.02E+00 1.0488(3.16E-02) (1.64E-02) (1.48E-02) (6.44E-02) (1.36E-01)NGC 1056 6.46E-02 1.11E-01 1.54E-01 2.33E-01 5.45E-01 -0.0326(2.09E-02) (1.07E-02) (1.90E-02) (6.18E-03) (1.29E-02)NGC 1125 4.71E-02 5.65E-02 2.72E-01 4.95E-01 1.08E+00 0.9306(1.18E-02) (1.26E-02) (1.79E-02) (1.23E-02) (2.18E-02)NGC 1143/4 7.79E-02 8.75E-02 1.85E-01 · · · · · · · · · · · · · · · (2.88E-02) (3.15E-02) (3.96E-02)NGC 3511 6.37E-02 8.39E-02 1.37E-01 1.72E-01 3.46E-01 0.1253(2.21E-02) (1.46E-02) (1.44E-02) (8.91E-03) (1.43E-02)NGC 3516 1.97E-01 3.11E-01 4.80E-01 6.97E-01 7.77E-01 0.0365(1.27E-02) (2.37E-02) (1.45E-02) (7.83E-03) (1.28E-02)NGC 3660 · · · · · · · · · -0.3803(1.46E-02) (4.21E-02) (1.79E-02)NGC 4593 2.22E-01 4.25E-01 5.43E-01 6.64E-01 8.13E-01 -0.0721(1.42E-02) (1.50E-02) (1.30E-02) (1.47E-02) (1.18E-02)NGC 4594 2.43E-01 1.36E-01 8.21E-02 · · · · · · -0.3486(1.74E-02) (7.20E-03) (8.00E-03)NGC 4602 4.78E-02 8.32E-02 1.20E-01 1.75E-01 3.65E-01 -0.0254(1.57E-02) (1.19E-02) (1.05E-02) (1.05E-02) (1.30E-02)NGC 4922 NED01 1.61E-01 3.40E-01 8.56E-01 · · · · · · TABLE 2 — Continued
Galaxy Name Continuum Flux Density (Jy) Optical Depth5.5 µ m 10 µ m 14.7 µ m 20 µ m 30 µ m 9.7 µ m(1) (2) (3) (4) (5) (6) (7)(3.98E-02) (5.47E-02) (6.60E-02)NGC 4941 · · · · · · · · · (1.13E-02) (2.11E-02) (1.30E-02)NGC 4968 1.03E-01 2.49E-01 6.21E-01 9.32E-01 1.02E+00 0.3205(2.13E-02) (3.09E-02) (2.95E-02) (2.00E-02) (1.77E-02)NGC 5005 1.81E-01 1.59E-01 2.25E-01 2.85E-01 9.08E-01 0.6118(1.65E-02) (3.35E-02) (1.23E-02) (1.42E-02) (3.66E-02)NGC 513 4.97E-02 9.63E-02 1.36E-01 1.98E-01 3.30E-01 0.1337(1.68E-02) (1.06E-02) (7.21E-03) (1.00E-02) (1.17E-02)NGC 5135 1.51E-01 2.62E-01 6.70E-01 1.31E+00 3.03E+00 0.3971(3.08E-02) (3.16E-02) (3.61E-02) (3.52E-02) (8.45E-02)NGC 5256 NED01 5.85E-02 7.62E-02 2.56E-01 4.95E-01 1.37E+00 0.5591(1.47E-02) (1.69E-02) (2.22E-02) (2.58E-02) (2.67E-02)NGC 526A · · · · · · · · · (1.35E-02) (1.23E-02) (1.23E-02)NGC 5347 6.19E-02 2.10E-01 5.22E-01 7.86E-01 7.91E-01 0.0025(8.71E-03) (1.92E-02) (1.22E-02) (1.78E-02) (1.03E-02)NGC 5506 9.33E-01 6.88E-01 2.18E+00 2.86E+00 4.05E+00 0.7640(2.17E-02) (1.06E-01) (6.61E-02) (5.59E-02) (6.38E-02)NGC 5929 1.35E-02 2.08E-02 2.89E-02 5.84E-02 1.13E-01 0.0243(7.60E-03) (7.60E-03) (2.66E-03) (3.46E-03) (3.62E-03)NGC 5953 1.20E-01 1.87E-01 3.16E-01 5.14E-01 1.27E+00 0.0941(3.57E-02) (1.81E-02) (3.25E-02) (1.47E-02) (3.37E-02)NGC 6810 2.02E-01 4.13E-01 9.59E-01 2.13E+00 3.33E+00 0.2068(4.37E-02) (2.65E-02) (2.85E-02) (2.86E-02) (7.18E-02)NGC 6860 1.64E-01 2.48E-01 3.25E-01 3.87E-01 3.36E-01 -0.0012(1.28E-02) (5.12E-03) (1.26E-02) (1.16E-02) (4.73E-03)NGC 6890 8.02E-02 1.49E-01 2.62E-01 3.98E-01 5.79E-01 -0.0208(1.55E-02) (1.10E-02) (8.16E-03) (1.08E-02) (9.52E-03)NGC 7130 9.29E-02 2.05E-01 5.38E-01 1.08E+00 2.47E+00 0.2710(1.67E-02) (1.62E-02) (2.21E-02) (3.30E-02) (6.89E-02)NGC 7172 · · · · · · · · · (3.05E-02) (2.41E-02) (2.94E-02)NGC 7213 1.79E-01 4.14E-01 4.87E-01 7.33E-01 5.77E-01 -0.2470(1.62E-02) (2.80E-02) (1.59E-02) (2.00E-02) (1.68E-02)NGC 7314 · · · · · · · · · (2.10E-02) (1.46E-02) (7.01E-03)NGC 7469 3.31E-01 7.67E-01 1.65E+00 3.28E+00 6.11E+00 0.1045(4.74E-02) (4.89E-02) (4.61E-02) (1.04E-01) (1.67E-01)NGC 7496 4.71E-02 1.37E-01 3.77E-01 8.96E-01 2.14E+00 0.1744(1.41E-02) (1.53E-02) (1.85E-02) (2.54E-02) (5.02E-02)NGC 7582 6.34E-01 5.79E-01 1.97E+00 3.52E+00 9.33E+00 0.8284(5.33E-02) (7.27E-02) (8.92E-02) (1.22E-01) (2.74E-01)NGC 7590 4.44E-02 6.55E-02 8.54E-02 1.20E-01 2.46E-01 -0.0658(1.32E-02) (8.10E-03) (6.82E-03) (5.80E-03) (1.17E-02)NGC 7603 2.39E-01 3.31E-01 2.98E-01 3.06E-01 3.12E-01 -0.0630(1.25E-02) (6.28E-03) (1.41E-02) (1.11E-02) (6.83E-03)NGC 7674 1.89E-01 3.89E-01 9.24E-01 1.37E+00 1.83E+00 0.2811(2.60E-02) (3.44E-02) (3.37E-02) (2.31E-02) (1.57E-02)NGC 788 3.01E-02 8.23E-02 2.16E-01 3.01E-01 3.47E-01 0.1234(3.30E-03) (1.05E-02) (7.87E-03) (1.34E-03) (3.59E-03)NGC 931 2.22E-01 3.99E-01 6.56E-01 8.79E-01 9.29E-01 0.0166(1.81E-02) (2.04E-02) (2.67E-02) (1.40E-02) (9.02E-03)SDSS J1039+6430 6.74E-03 1.33E-02 2.72E-02 · · · · · · -0.0103(2.00E-04) (1.17E-03) (7.75E-04)SDSS J1641+3858 3.75E-03 2.13E-02 · · · · · · · · · -0.0473(1.02E-03) (2.20E-03)TOL1238-364 1.18E-01 3.36E-01 9.20E-01 1.67E+00 2.16E+00 0.2025(1.90E-02) (3.82E-02) (3.69E-02) (1.58E-02) (3.93E-02)UGC 11680 NED01 3.62E-02 8.87E-02 1.33E-01 1.80E-01 2.11E-01 0.0946(6.71E-03) (1.54E-02) (7.37E-03) (1.12E-02) (6.44E-03)UGC 12138 2.43E-02 5.56E-02 1.11E-01 1.79E-01 3.03E-01 0.2383(2.07E-03) (5.07E-03) (5.79E-03) (3.68E-03) (1.09E-02)UGC 7064 3.06E-02 8.51E-02 1.52E-01 2.31E-01 3.28E-01 0.1045(1.36E-02) (6.15E-03) (6.79E-03) (1.16E-02) (5.85E-03)UM 146 6.76E-03 1.42E-02 3.22E-02 5.29E-02 9.72E-02 0.2348(8.04E-04) (1.05E-03) (1.62E-03) (3.55E-03) (5.13E-04)WIR-IRAS 23060+0505 1.01E-01 1.58E-01 3.31E-01 4.53E-01 7.51E-01 0.3170(6.50E-03) (9.47E-03) (4.01E-03) (1.27E-02) (1.46E-02) Note . — Col. (1): Galaxy name; Col. (2-6): Continuum flux density as measured on the spectrum at 5.5, 10., 14.7, 20, and 30 µ m in Jansky; Col. (7): Observedoptical depth (apparent) at 9.7 µ m; 1 σ errors are given in parenthesis. he Mid-Infrared Continua of Seyfert Galaxies 13 TABLE 3Starburst-to-AGN Flux Density Ratios
Galaxy Name Starburst-to-AGN Flux Density Ratio5.5 µ m 10 µ m 14.7 µ m 20 µ m 30 µ m(1) (2) (3) (4) (5) (6)3C 321 0.0263 0.0400 0.0207 0.0354 0.0640CGCG 381-051 · · · · · · ESO 12-G21 0.3020 0.7335 13.4681 · · · · · ·
ESO 33-G2 0.0291 0.0359 0.0622 0.1534 0.9593IC 4329A 0.0043 0.0082 0.0145 0.0317 0.1481IRAS 05189-2524 0.0355 0.0640 0.0610 0.0925 0.1097IRASF 01475-0740 0.0702 0.0835 0.1556 0.2870 1.7003IRASF 04385-0828 0.0253 0.0887 0.0715 0.1456 0.4536IRASF 15480-0344 0.0679 0.0696 0.1084 0.2052 0.9631MCG+0-29-23 1.4460 11.5036 · · · · · · · · ·
MCG-03-34-064 0.0247 0.0309 0.0323 0.0571 0.1434MCG-2-33-34 0.1739 0.3531 0.6773 2.3617 · · ·
MCG-2-40-4 0.0613 0.1260 0.2427 0.6416 17.9313MCG-3-34-63 0.3026 38.4053 · · · · · · · · ·
MCG-3-58-7 0.0501 0.0891 0.1635 0.3254 1.5569MCG-5-13-17 0.1574 0.3035 0.5114 1.3713 · · ·
MCG-6-30-15 0.0171 0.0288 0.0588 0.1249 0.6378Mrk 1239 0.0138 0.0281 0.0706 0.1727 1.1758Mrk 3 0.0060 0.0077 0.0051 0.0077 0.0233Mrk 334 0.4212 0.6677 1.0960 1.4361 4.5979Mrk 348 0.0395 0.0754 0.0971 0.2081 1.8993Mrk 463E 0.0124 0.0329 0.0364 0.0627 0.1964Mrk 471 0.6422 1.7036 · · · · · · · · ·
Mrk 477 0.0695 0.0817 0.0705 0.1061 0.2781Mrk 6 0.0170 0.0411 0.0668 0.1237 0.6543Mrk 609 1.3144 2.3839 25.1028 · · · · · ·
Mrk 622 0.5186 0.3268 0.3328 0.3661 0.7087Mrk 766 0.0574 0.0914 0.1203 0.2250 0.7952Mrk 79 0.0211 0.0429 0.0809 0.1754 0.7295Mrk 817 0.0417 0.0625 0.1186 0.1911 0.6321Mrk 883 0.5226 0.8595 0.7567 0.6741 1.1640Mrk 9 0.0627 0.0951 0.2400 0.6527 · · ·
Mrk 938 0.7001 · · · · · · · · · · · · · · · · · ·
NGC 1241 0.2527 1.2900 3.9115 · · · · · ·
NGC 1320 0.0667 0.0964 0.1615 0.3454 2.5940NGC 1365 0.8418 3.5882 2280.9476 · · · · · ·
NGC 1386 0.0528 0.1621 0.1460 0.3804 1.4503NGC 2273 0.2300 0.4914 0.5444 1.4474 88.6744NGC 2622 0.1445 0.2242 0.2628 0.5276 13.8732NGC 2639 0.3437 31.3315 · · · · · · · · ·
NGC 2992 0.2337 0.4174 0.6511 3.4637 · · ·
NGC 3079 4.6747 · · · · · · · · · · · ·
NGC 3081 0.0442 0.0485 0.0480 0.0785 0.2137NGC 3227 · · · · · · · · ·
NGC 3511 2.6708 · · · · · · · · · · · ·
NGC 3516 0.0170 0.0389 0.0758 0.1423 0.5974NGC 3660 · · · · · · · · · · · ·
NGC 3786 0.2673 0.6172 1.4630 11.1716 · · ·
NGC 3982 2.6710 9.3812 · · · · · · · · ·
NGC 4051 0.0702 0.1056 0.1872 0.4395 7.1327NGC 4151 0.0129 0.0186 0.0263 0.0497 0.2452NGC 424 0.0102 0.0200 0.0390 0.0885 0.5310NGC 4388 0.1237 0.5162 0.3060 0.5262 2.9754NGC 4501 0.1361 1.2925 · · · · · · · · ·
NGC 4507 0.0242 0.0372 0.0588 0.1080 0.3233NGC 4579 0.0539 0.1936 · · · · · · · · ·
NGC 4593 0.0346 0.0630 0.1552 0.3887 3.2523NGC 4594 0.0094 0.0609 · · · · · · · · ·
NGC 4602 0.8372 2.8868 · · · · · · · · ·
NGC 4922 NED01 0.1114 0.2474 · · · · · · · · ·
NGC 4941 · · · · · · · · ·
NGC 4968 0.1301 0.1651 0.1894 0.3638 4.7134NGC 5005 0.1619 2.3674 · · · · · · · · ·
NGC 513 0.9011 1.1506 · · · · · · · · ·
NGC 5135 0.7242 3.4740 8.6441 · · · · · ·
NGC 5256 NED01 1.1450 65.3742 8.9031 · · · · · ·
NGC 5347 0.0552 0.0598 0.0665 0.1229 0.5838NGC 5506 0.0250 0.1339 0.1109 0.2495 0.9290NGC 5929 0.5391 · · · · · · · · · · · ·
NGC 5953 4.1648 · · · · · · · · · · · ·
NGC 6810 1.0366 1.9786 4.2739 17.9650 · · ·
TABLE 3 — Continued
Galaxy Name Starburst-to-AGN Flux Density Ratio5.5 µ m 10 µ m 14.7 µ m 20 µ m 30 µ m(1) (2) (3) (4) (5) (6)NGC 6860 0.0602 0.1206 0.3297 1.0168 · · · NGC 6890 0.2445 0.4524 0.9600 5.6312 · · ·
NGC 7130 1.0761 2.3340 2.4662 10.1020 · · ·
NGC 7213 0.0530 0.0914 0.2545 0.5246 · · ·
NGC 7314 · · · · · · · · ·
NGC 7469 0.3845 0.5837 0.9244 1.6895 · · ·
NGC 7496 1.3900 1.0345 0.9555 1.2744 3.7034NGC 7582 0.3118 9.1000 2.2330 536.9538 · · ·
NGC 7590 0.6934 · · · · · · · · · · · ·
NGC 7603 0.0458 0.1133 0.4841 3.8650 · · ·
NGC 7674 0.1407 0.2002 0.2614 0.5388 7.2992NGC 788 0.0274 0.0361 0.0380 0.0728 0.2437NGC 931 0.0326 0.0600 0.1093 0.2347 1.4468SDSS J1039+6430 0.0114 0.0188 0.0268 · · · · · ·
TOL 1238-364 0.3450 0.3095 0.3202 0.5216 8.6584UGC 11680 NED01 0.1504 0.1820 0.4536 1.4949 · · ·
UGC 12138 0.1290 0.2047 0.3179 0.6320 3.1993UGC 7064 0.6323 0.4803 1.2807 276.5345 · · ·
UM 146 0.0639 0.1100 0.1527 0.2621 0.5764WIR-IRAS 23060+0505 0.0113 0.0248 0.0338 0.0665 0.1457
Note . — Col. (1): Galaxy name; Col. (2-6): Starburst-to-AGN flux density ratio measured at 5.5, 10, 14.7, 20, and 30 µ m; a missing ratio implies either the spectrumwas incomplete at that wavelength or the AGN contribution was negative after starburst subtraction. A very large ratio indicates very weak AGN contribution, thesetypically occur at or beyond 14.7 µ m in a few galaxies. See Figure 10 for a visual representation of these ratios, and Table 1 for aperture size in parsecs. Even in nearbyAGN Spitzer spectra sample regions of size ∼ he Mid-Infrared Continua of Seyfert Galaxies 15 λ [ µ m]10 -18 -17 -16 -15 λ F λ [ W c m - ] Sey 1.0/1.2/1.5 λ [ µ m]10 -18 -17 -16 -15 λ F λ [ W c m - ] Sey 1h λ [ µ m]10 -18 -17 -16 -15 λ F λ [ W c m - ] Sey 1.8/1.9s λ [ µ m]10 -18 -17 -16 -15 λ F λ [ W c m - ] Sey 2s
Fig. 1.—
Mid-infrared spectra of the entire sample normalized at 14.7 µ m. Top left: Seyfert 1s; top right: Seyfert 2s with evidence ofbroad optical emission lines in polarized light; bottom left: Seyfert 1.8/1.9s spectra dominated by PAH features; bottom right: Seyfert 2spectra dominated by PAH features. The short-wavelength continuum of PAH-weak type 2 Seyferts (top right) is on average weaker andsteeper than type 1 Seyferts (top left). Seyfert 1.8/1.9s and PAH-strong type 2 Seyferts (bottom right) show very similar spectra.
10 15 20 25 30Wavelength [ µ m]24681012 F λ [ - W c m - µ m - ] Mrk 766Mrk 335SubtractionMrk 3, scaled
Fig. 2.—
Subtraction of Mrk 335 (type 1) spectrum from the spectrum of Mrk 766 (type 1, starburst-subtracted, see Section 5). Theresidual spectrum is very similar to the Mrk 3 (type 2) spectrum and shows the prominent 15–20 µ m bump. The spectra are displayed inthe same order at 10 µ m as is indicated in the legend. he Mid-Infrared Continua of Seyfert Galaxies 17 -3 -2 -1 0 1 2 3 α λ (20-30 µ m)-2.5-2.0-1.5-1.0-0.50.00.51.0 α λ ( . - . µ m ) MRK1239 MRK938NGC1194 NGC1386 NGC3079NGC4151 NGC4501 NGC7469NGC7603 MRK622MRK3
Type 2Type 1 Starburst+Host Dust -S1-S1.5 -S2 -S1.8/1.9 -S3 -S1h
Fig. 3.—
Mid-IR continua of Seyfert Galaxies: Seyfert 2s with hidden broad-line regions (S1h) are represented as filled diamond symbols,Seyfert 1s (including 1.2s and 1.5s) are filled circles, Seyfert 1.8/1.9s are open triangles, and Seyfert 2s with undetected polarized broademission lines are open squares. Liners are represented with cross symbols. As α λ (20–30 µ m) increases, spectra contain more starburstcontribution; as α λ (5 . . µ m) varies from 0.5 to -2.0, spectra are more dominated by the hot dust component due to the active nucleus.The arrows indicate how the spectral shape changes from type 2 to type 1 Seyferts (see also Figure 7); and along this sequence, additionof a cold starburst component of increasing strength moves the source position to the right in the figure. The dotted lines show roughdivision between different Seyfert types. See Section 4 for further discussion. -0.5 0.0 0.5 1.0 1.5Apparent Optical Depth ( τ µ m )-2.5-2.0-1.5-1.0-0.50.00.51.0 α λ ( . - . µ m ) MRK938NGC1194NGC1386 NGC3079NGC4151 NGC4501NGC7469NGC7603MRK622 MRK3 -S1-S1.5 -S2 -S1.8/1.9 -S3 -S1h
Fig. 4.—
There is a weak correlation between apparent optical depth ( τ . ) and α λ (5 . . µ m) providing general support for inclinationdependence in AGN models. Some sources show large apparent silicate optical depth ( τ . ∼ τ . & . he Mid-Infrared Continua of Seyfert Galaxies 19 -0.5 0.0 0.5 1.0 1.5Apparent Optical Depth, τ µ m -1.0-0.50.0 PA H . µ m E W NGC7603NGC4151NGC7469NGC2639NGC4501NGC1386 NGC1194NGC3079MRK938 -S3 -S1.8/1.9 -S2 -S1-S1.5 -S1h
Fig. 5.—
There is no correlation between the equivalent width of the 11.3 µ m PAH band and the measured apparent optical depth at9.7 µ m, suggesting that the presence of strong PAH bands does not bias our continuum placement in the short wavelength region of themid-IR spectra, where defining the continuum is highly subjective due to the presence of blended PAH emission bands, and silicate and iceabsorption features. -0.5 0.0 0.5 1.0Apparent Optical Depth ( τ µ m )0.00.20.40.60.81.0 A x i a l R a t i o ( b / a ) IC4329A MRK3MRK622 MRK938NGC1194NGC1386 NGC3079NGC4151 NGC4501NGC7469NGC7603 -S1-S1.5 -S2 -S1.8/1.9 -S3 -S1h
Fig. 6.—
Correlation between the optical depth of 9 . µ m silicate feature and the ratio of minor-to-major axis ( b/a ) of the host galaxy.The b/a values are taken from the NASA/IPAC Extragalactic Database. Objects for which b/a could not be estimated are assumed tohave b/a = 1. he Mid-Infrared Continua of Seyfert Galaxies 21 -0.5 0.0 0.5 1.0 1.5Apparent Optical Depth ( τ µ m )-2.5-2.0-1.5-1.0-0.50.00.51.0 α λ ( . - . µ m ) α ν = - ( α λ + ) V = (18.0 ± τ µ m λ [ µ m]1.01.52.02.5 F λ [ - W c m - µ m - ] Mrk 3, S1h λ [ µ m]0.51.01.52.02.5 F λ [ - W c m - µ m - ] NGC1386, S2 λ [ µ m]1234 F λ [ - W c m - µ m - ] MRK938, S2 λ [ µ m]1234567 F λ [ - W c m - µ m - ] NGC4151, S1.5 λ [ µ m]0.20.40.60.81.01.21.41.6 F λ [ - W c m - µ m - ] NGC1194, S1.9Edge-on λ [ µ m]0.51.01.52.02.5 F λ [ - W c m - µ m - ] NGC7603, S1.5 λ [ µ m]0.20.40.60.81.01.21.4 F λ [ - W c m - µ m - ] NGC4501, S2 λ [ µ m]2468101214 F λ [ - W c m - µ m - ] NGC3079, S2
Fig. 7.—
Effect of host galaxy dust on mid-IR spectra of AGN—central plot shows the variation of α λ (5 . . µ m) with τ . µ m ; leftpanels: spectra of AGN-dominated galaxies; middle panels: increasing cold dust and PAH contribution attributed to star formation in thehost galaxy; right panels: Inclined host galaxies without strong PAH features (NGC 1194) or with strong PAH features (Mrk 938 and NGC3079). Essentially, in all systems with τ . µ m & . . µ m is likely due to cold dust in the host galaxy. Inthe panels, from left to right, star formation contribution increases. Wavelength [ µ m] F l u x D en s i t y [ - W c m - µ m - ] IRAS F01475-0740, S2
NGC 7213, S3b
Mrk 766, S1.5
NGC 4051, S1n
NGC 2992, S1.9
TOL 1238-364, S1h
NGC 4388, S1h
NGC 1125, S2
Mrk 477, S1h
Fig. 8.—
Seyfert spectra after subtraction of starburst components: the dotted line shows the observed spectrum and the solid line showsthe subtraction of an average starburst template from this spectrum. See Figure 1 in Deo et al. (2007) for identification of emission linesand PAH bands in mid-IR spectra. Note the clearly visible silicate features, the power-law-like shape of the 5–8 µ m continuum and the15–20 µ m bump. In most cases, the 7 . µ m [Ne VI ] lines are also cleanly separated. he Mid-Infrared Continua of Seyfert Galaxies 23 λ [ µ m]10 -18 -17 -16 -15 λ F λ [ W c m - ] Sey 1.0/1.2/1.5 λ [ µ m]10 -18 -17 -16 -15 λ F λ [ W c m - ] Sey 1h λ [ µ m]10 -18 -17 -16 -15 λ F λ [ W c m - ] Sey 1.8/1.9s λ [ µ m]10 -18 -17 -16 -15 λ F λ [ W c m - ] Sey 2s
Fig. 9.—
Mid-infrared spectra from Figure 1 after subtraction of a scaled starburst template spectrum. The spectra are normalized at14.7 µ m. Top left: Seyfert 1s; top right: Seyfert 2s with evidence of broad optical emission lines in polarized light; bottom left: Seyfert1.8/1.9s; bottom right: Seyfert 2s with undetected polarized broad emission lines. Compare this figure with Figure 1. On average, afterstarburst subtraction, PAH-dominated Seyfert 1.8/1.9s and Seyfert 2s show similar continuum shapes as type 2 Seyferts with HBLR (topright panel). There is a striking similarity in the active nuclear continuum of all Seyfert types beyond ∼ µ m.
39 40 41 42 43 44 45log L [AGN, erg s -1 µ m -1 ]39404142434445 l og L . [ S t a r bu r s t, e r g s - µ m - ] SB/AGN = 10SB/AGN = 1SB/AGN = 0.1SB/AGN = 0.01SB/AGN = 0.001 MRK938NGC1386NGC3079 NGC4151NGC4501 NGC7469NGC7603MRK622 MRK3
39 40 41 42 43 44 45log L [AGN, erg s -1 µ m -1 ]39404142434445 l og L [ S t a r bu r s t, e r g s - µ m - ] SB/AGN = 10SB/AGN = 1SB/AGN = 0.1SB/AGN = 0.01SB/AGN = 0.001NGC1386 NGC4151NGC7469NGC7603MRK622MRK3
Fig. 10.—
Luminosity density at 5 . µ m (left) and 20 µ m (right) from the starburst and the active nuclear components. Our sampleshows a strong contribution from the active nuclear component, and galaxies with large starburst contributions are preferentially Seyfert1.8/1.9s and some Seyfert 2s with unidentified broad polarized emission lines. The thick red line shows a linear regression fit (bisectormethod) to data points. The diagonal lines show starburst-to-AGN contribution ratio. Symbols for Seyfert types are same as in Figure 3.At longer wavelengths, the average starburst fraction increases and is roughly the same as the AGN contribution at ∼ µ m. he Mid-Infrared Continua of Seyfert Galaxies 25 -4 -3 -2 -1 0 1 α λ (20-30 µ m)-3-2-101 α λ ( . - µ m ) MRK1239 MRK3MRK622 MRK938NGC1194NGC1386NGC3079NGC4151NGC4501NGC7469NGC7603 Avg. in Figure 4Avg. this Figure
41 42 43 44 45log L µ m [AGN, erg s -1 µ m -1 ]-3-2-101 α λ ( . - µ m ) MRK1239MRK3MRK622 MRK938NGC1194NGC1386 NGC3079NGC4151NGC4501 NGC7469 NGC7603Avg. for Seyfert 1sAvg. for Seyfert 2s
Fig. 11.—
Top: Comparison between α λ (5 . . µ m) and α λ (20–30 µ m) after subtraction of starburst component. α λ (20–30 µ m) valuesare more negative than in Figure 3 (dashed lines), indicating flatter continua than those at short wavelength. Bottom: wavelength spectralindex between 5.5 and 8 µ m after starburst subtraction compared to the luminosity density of the AGN component at 5 . µµ