Host galaxy and orientation differences between different types of AGN
Anamaria Gkini, Manolis Plionis, Maria Chira, Elias Koulouridis
AAstronomy & Astrophysics manuscript no. aanda © ESO 2021January 6, 2021
Host galaxy and orientation differences between different types ofAGN
Anamaria Gkini , , Manolis Plionis , , Maria Chira , and Elias Koulouridis Department of Astrophysics, Astronomy & Mechanics, Faculty of Physics, National and Kapodistrian University of Athens,Panepistimiopolis Zografou, 15784, Greece Department of Physics, Aristotle University of Thessaloniki, Thessaloniki 54124, Greece Institute for Astronomy & Astrophysics, Space Applications & Remote Sensing, National Observatory of Athens, GR-15236 PalaiaPenteli, GreeceJanuary 6, 2021
ABSTRACT
Aims.
The main purpose of this study is to investigate aspects regarding the validity of the AGN unification paradigm (UP). Inparticular we focus on the AGN host galaxies, which according to the UP should show no systematic di ff erences depending on theAGN classification. Methods.
For the purpose of this study we use (a) the spectroscopic SDSS (Sloan Digital Sky Survey) DR14 catalogue, in order toselect and classify AGNs using emission line diagnostics, up to a redshift of z = .
2, and (b) the Galaxy Zoo Project catalogue, whichclassifies SDSS galaxies in two broad Hubble types, spirals and ellipticals.
Results.
We find that the fraction of type-1 Seyfert nuclei (Sy1) hosted in elliptical galaxies is significantly larger than the corre-sponding fraction of any other AGN type, while there is a gradient of increasing Spiral-hosts from Sy1 to Liner, type-2 Seyferts (Sy2)and Composite nuclei. These findings cannot be interpreted within the standard Unification Paradigm, but possibly by a co-evolutionscheme for supermassive black holes (SMBH) and galactic bulges.Furthermore, for the case of spiral host galaxies we find the Sy1 population to be strongly skewed towards face-on configurations,while the corresponding Sy2 population range in all host-galaxy orientation configurations, having a similar, but not identical, orien-tation distribution with star-forming galaxies (SF). These results also cannot be interpreted by the standard Unification Paradigm, butpoint towards a significant contribution of the galactic disk to the obscuration of the nuclear region. This is also consistent with theobserved preference of Sy1 nuclei to be hosted by ellipticals, ie., the dusty disk of spiral hosts contributes to the obscuration of thebroad line region (BLR) and thus relatively more ellipticals are expected to appear hosting Sy1 nuclei.
Key words. galaxies: Seyfert - galaxies: active - galaxies: nuclei - galaxies: evolution - galaxies: bulges - galaxies: statistics
1. Introduction
The unification model of Active Galactic Nuclei (AGN) explainsthe wide variety of features discerned in di ff erent classes of AGNin terms of the anisotropic geometry of the black hole’s immedi-ate surroundings (e.g. Antonucci 1993; Urry & Padovani 1995;Netzer 2015)). The accreting supermassive black hole (SMBH)is fed via a disc, while a thick and obscuring torus of gas anddust surrounds the disc absorbing a fraction of the emitted radi-ation. Depending on the orientation of the torus with respect tothe observer’s line of sight, the view towards parts of the galac-tic nucleus may be obstructed, giving rise to di ff erences in theobserved brightness and spectra of the AGN.An important implication of the Unification Paradigm is that,since the observed di ff erences between AGN types are attributedsolely to orientation e ff ects, the host galaxies of AGN should beintrinsically the same (e.g. Netzer 2015; Hickox & Alexander2018). Thus, despite their success in explaining a range of AGNobserved features, such as the absence of broad-line features inthe spectra of Sy2 galaxies, it has become clear that the simplestmodels of unification are inconsistent with observations and cannot explain aspects such as the lack of BLR in many Sy2 AGNin polarised spectra (Tran 2001, 2003) and the lack of Sy1’s inclusters (Martínez et al. 2008). Specifically, according to stud-ies investigating the di ff erences between Sy1’s and Sy2’s (e.g. Koulouridis et al. 2006; Rigby et al. 2006; Martínez-Sansigreet al. 2006a; Lacy et al. 2007; Villarroel et al. 2012; Netzer2015; Bornancini & García Lambas 2018; Zou et al. 2019; Bor-nancini & García Lambas 2020) and the connection with theirhost galaxies at redshifts 0 . < z < . ff erent optical, mid-IR, X-rayand morphological properties and also reside in statistically dif-ferent environments (Bornancini & García Lambas 2020). More-over, it is found that the neighbours of Sy2 AGN are more star-forming and bluer than Sy1 AGN (see also Koulouridis et al.2013) and also that Sy2 hosts are surrounded by a larger num-ber of dwarf galaxies (Villarroel et al. 2012). Additionally, themorphology of Sy1 galaxies show no indications of close inter-actions which means either that they rarely merge (Koulouridis2014), or that they are extremely short-lived AGNs (Villarroelet al. 2012). Bornancini & García Lambas (2018) find at highredshifts (0 . < z < .
1) that Sy2’s have more abundant neigh-bors as well as, that Sy1 hosts are preferably elliptical or com-pact galaxies, while Sy2 hosts present a broader Hubble-typedistribution.Since, observations do not fully comply with the predictionsof the Unification Paradigm, alternative or complementary fac-tors a ff ecting the observed AGN types should be sought in order Article number, page 1 of 8 a r X i v : . [ a s t r o - ph . GA ] J a n & A proofs: manuscript no. aanda to explain the AGN variety. Indeed, Koulouridis et al. (2006)and Jiang et al. (2016) studying the environments of Sy1 andSy2 galaxies at low redshifts found that both AGN classes havesimilar clustering properties, but at scales smaller than 100 kpc
Sy2’s have significantly more neighbours than Sy1’s. Jiang et al.(2016) also found significant di ff erences in the infrared color dis-tributions of the host galaxies of the two AGN types. Further-more, some studies claim that, not only the torus, but also thedust in the galactic disk may have non-negligible contribution tothe optical obscuration of nuclei (e.g. Maiolino & Rieke 1995;Matt 2000; Lagos et al. 2011; Netzer 2015; Bornancini & GarcíaLambas 2018; Zou et al. 2019). Malizia et al. (2020) using thehard X-ray selected AGN sample detected by INTEGRAL / IBIS,have shown that material located in the host galaxy on scales ofhundreds of parsec, while not aligned with the absorbing torus,can be extended enough to hide the Broad Line Region (BLR)of some Sy1’s causing their misclassification as Sy2 objects andgiving rise to the deficiency of around 24% of Sy1’s in edge-ongalaxies.According to Koulouridis et al. (2006) and Krongold et al.(2002), in the context of a time-evolution scenario, in some casesthe interaction between gas-rich galaxies ignites starburst activ-ity while large amounts of gas and dust obscure the central nu-clear region at this stage. As the starburst dies o ff , the remainingmolecular gas and dust forms a torus around the disk and, even-tually the AGN will attenuate the obscuring medium. Namely,this model proposes an AGN- evolutionary sequence going fromstarburst to type-2 and finally to type-1 Seyfert galaxies (e.g.Springel et al. 2005; Hopkins et al. 2006; Koulouridis et al. 2013;Yang et al. 2019).Studies in polarised light to Sy2 galaxies, have shown thatin many low-luminous AGN the dusty torus is absent while theBLR is also not detected (e.g. Elitzur & Shlosman 2006; Perlmanet al. 2007; Trump et al. 2011; Koulouridis 2014; Hernández-Ibarra et al. 2016). These results are consistent with those ofTrump et al. (2011), who support that above a specific accretionrate (( L / L Edd (cid:38) . L / L Edd (cid:46) .
01) theBLR becomes non-detectable but also the obscuring torus tendsto become weaker or disappears (Elitzur et al. 2014).The above findings could be incorporated within the evolu-tionary scheme. If the accretion-rate dependent scenario is valid,one would expect that AGN could lose their torus or / and theirBLR at the end of the AGN duty cycle, as the accretion ratedrops below a critical value (Elitzur et al. 2014; Elitzur & Ho2009; Koulouridis 2014). This implies that firstly, AGN can ap-pear as type-1, after the quenching of the star-forming activity bythe AGN feedback and the disappearance of the torus (Krongoldet al. 2002) and secondly, these type-1 will evolve to true type-2AGN due to the elimination of the BLR, based for example onthe wind-disk scenario of Elitzur & Ho (2009).In summary, the plethora of results of many relevant stud-ies clearly indicate that the viewing angle alone cannot fully ac-count for the di ff erent AGN types. It is within this ideology thatthe current study lies, investigating the orientation properties ofspiral hosts as well as the Hubble type distribution of di ff erenttypes of AGN.After the presentation of the data used in section 2, the mainpart of our analysis is organised as follows; In section (3.1) westudy the morphology frequency distribution of host galaxies fordi ff erent AGN-types comparing to that of the non-active star-forming sample, after statistically matching their respective red-shift distribution in order to suppress possible evolutionary ef-fects. In the section (3.2) we study the frequency distribution of spiral host galaxy orientations ( b / a ), for the Sy1 and Sy2 sub-samples, comparing with that of non-active star-forming galax-ies (which we use as a control sample). We use only galaxieswith high "spirality" probability > .
2. Observational Data
For the purposes of the current study we use galaxy cataloguesextracted from the Sloan Digital Sky Survey (SDSS) DR14(Abolfathi et al. 2018) in five bands (u,g,r,i,z) with a magni-tude limit for the spectroscopic sample of m r = .
77 in the r-band, in order to have a homogeneous magnitude cuto ff over thelargest possible SDSS area. Our catalogues consist of 3 .
800 Sy1,56 .
846 Sy2, 120 .
025 Composite, 107 .
034 Liners and 263 . z < .
2. The morphologycharacterization of the galaxies is based on the Galaxy ZooProject (Lintott et al. 2008), a crowd sourced astronomy projectwhich asks citizens to characterize galaxies as spirals or ellip-ticals and to determine the rotation direction of spirals by in-specting SDSS galaxy images (Raddick et al. 2007). In orderto ensure trustworthy classification of our galaxies, we performa quality cut on our catalogue, rejecting galaxies with Ha rest-frame equivalent width (EW) < ffi ciently large subsample.The Sy1 sample comprises all galaxies with a Balmer linewidth of σ greater than 500 km / sec (FWHM > / s). Note,that in the SDSS database all such sources are catalogued ashaving a σ =
500 km / sec, while this is actually a lower limit. Vi-sual inspection of a large number of spectra validated that thesesources are bonafide broad line Seyferts. All spectra with anemission line having σ >
200 km / sec are also characterized asbroadline in the SDSS database. We have visually reviewed allspectra in our spiral galaxy sample that fall in this category (450sources) and we have concluded that they contain very few thatcould be unambiguously classified as broad-line AGNs. There-fore, to reduce noise in our analysis we choose to exclude fromour Sy1 sample all spectra with Balmer lines having σ < / sec.After the above procedures our final sample of secure objectsconsists of 1.378 Sy1, 7.498 Sy2, 26.544 Composite, 1.926 Linerand 203.298 SF galaxies.The classification between di ff erent type of narrow-lineAGNs and SF galaxies has been performed utilizing the BPTdiagram classification method of Baldwin et al. (1981).It is important to note that we will use the SF galaxies, be-ing spirals in their vast majority, as our control sample since,their disk orientations cover the whole range of viewing angleswith respect to the line of sight. Additionally, they are mainlynon-AGN galaxies, while, even in the case of hosting an AGNin their center, the star-formation dominates the emission, lead-ing to spectra characteristic of non-AGN galaxies (Siebenmor-gen et al. 2015).Before proceeding we felt necessary to investigate the levelof consistency between the morphology classification of theGalaxy Zoo project and our own assessment via inspection ofthe host galaxy image. To this end, we select a small subsampleof images and spectra of galaxies of the highest "spirality" (prob-ability of being a spiral) and "ellipticity" (probability of being anelliptical). In detail, we select 50 nearby galaxies, 5 spirals and5 ellipticals of every AGN type, with redshift z < . Article number, page 2 of 8. Gkini, M. Plionis, M. Chira & E. Koulouridis: Host galaxy and orientation di ff erences between di ff erent types ofAGN Galaxy Zoo morphology classification is reliable for the major-ity of our galaxies, we found some cases in which the assessmentof the citizens participating in the Galaxy Zoo project seems tobe dubious. Our analysis showed that we can trust the character-ization of spirals up to z ∼ . z ∼ . ffi cult to distinguishat higher redshifts, where a galaxy image with weak spirality canbe interpreted as being an elliptical.
3. Methodology and Results
This section is organised as follows: in subsection §3.1 we studythe Hubble-type frequency distributions of the host galaxies forthe di ff erent samples: Sy1s, Sy2s, Liners, Composite or star-forming, while in section §3.2 we compare the projected axis-ratio (related to the orientation with respect to the line-of-sight)frequency distributions for the spiral hosts of Sy1’s and Sy2’s. In this section, we seek to reveal if there is any correlation be-tween the AGN-type and the Hubble-type morphology of theirhost galaxies. According to the Unification Paradigm, the di ff er-ent AGN classes are a result of di ff erent viewing angles with re-spect to the orientation of an obscuring torus and thus, the prop-erties of the host galaxies should not show any statistical signifi-cant di ff erences. For this study, we use subsamples of Sy1, Sy2,Liner, Composite and SF galaxies, derived from our SDSS cat-alogue with a redshift limit of z < .
1, in order to have morerobust Hubble-type classification (as discussed in the previoussection). For these subsamples we generate the frequency distri-bution of the Zoo ellipticity- and spirality- probabilities, whichwithin the context of the UP, are expected to be statisticallythe same for all classes of AGN (e.g. Antonucci 1993; Urry &Padovani 1995).In order to be able to compare among the di ff erent proba-bility distributions, avoiding possible evolutionary e ff ects, it isnecessary to take into account any statistically significant dif-ferences between the redshift-distributions of the di ff erent sam-ples, which we present in Fig.1( upper panel ). Due to the di ff er-ent number of objects in the subsamples and in order to revealsystematic trends among the di ff erent subsamples, we normalisethe distributions dividing with the total number of objects in eachsubsample. Using a random sampling procedure we match thenormalized redshift-distributions to a common fractional distri-bution (Fig.1 - lower panel ). Because of the low number of Sy1galaxies and in order to avoid further depleting it, we re-sampleall other activity-types, so that their normalised distributions arematched to that of the Sy1 sample.For the redshift matched subsamples, we present in Fig. 2the "ellipticity" ( upper panel ) and "spirality" ( lower panel ) -probability distributions. The two quantities are complementaryas should be expected, since the sum of the two Zoo-probabilitiesshould be roughly equal to unity. Moreover, we should note thatSF galaxies, with spectra dominated by young stellar popula-tions, are by definition spirals (Hubble 1926) and indeed, as seenin Fig. 2, their "spirality"-distribution (black solid line) peaks athigh-"spirality" probabilities, while their respective "ellipticity"-distribution, roughly complementary, peaks at zero-"ellipticity"probabilities. The SF galaxy sample can thus be used as a controlsample of the Zoo Project "spirality"- and "ellipticity"- distribu-tions of the various AGN host galaxies.In Fig. 2 we see that the di ff erent AGN classes are dis-tributed in the whole range of probabilities indicating a wide Fig. 1.
Upper panel:
Normalised redshift frequency distributions forthe four subsamples of AGN types and SF galaxies, limited to z < . ff erent subsamples are coded with di ff erent colours as denoted inthe key. Lower panel:
Redshift-matched normalized distribution. Theshaded area corresponds to the 1 σ Poisson uncertainty. range of Hubble-type hosts, with the predominance of Spirals.However, the normalized frequency distribution of the di ff erentAGN-types are dissimilar at a statistically significant level (asindicated by the 1 σ Poisson uncertainty) which implies a di ff er-ent Hubble-type distribution for the di ff erent AGN types, a re-sult that contradicts the original Unification Paradigm accordingto which there should be no dependence of the AGN class to thehost galaxy Hubble-type classification (e.g. Antonucci 1993). In-terestingly, the Sy1’s show a peak at both high and low "spiral-ity" probabilities, indicating a relatively higher fraction of Sy1’s,with respect to other AGN types, residing in elliptical hosts.For a more revealing comparison of the previously discussedmorphology di ff erence of the various AGN host galaxies withrespect to SF galaxies, we present in Fig.3 the excess factor bywhich the fractional number of the various AGN types exceedthat corresponding to SF galaxies, for each "spirality" or "ellip-ticity" probability: ∆ (AGN , p) = N i (AGN) / N tot (AGN)N i (SF) / N tot (SF) − , Article number, page 3 of 8 & A proofs: manuscript no. aanda
Fig. 2.
Probability-distributions of "ellipticity" ( upper panel ) and "spi-rality" ( lower panel ) for the Sy1, Sy2, Liner, Composite and SF hostgalaxies, limited to z < .
1. The di ff erent AGN and the SF subsamplesare color coded as in Fig.1 while the shaded area corresponds to the 1 σ Poisson uncertainty.
Inspecting Fig.3 it becomes evident that Sy1s show the high-est relative preference for elliptical hosts with respect to Liners,Sy2s and Composites, although all AGN types, at a di ff erent de-gree each, appear in elliptical hosts. We wish to test the hypothesis that the AGN-host galaxy con-tributes to the obscuration of the AGN-emission. To this end wewill study the orientation of spiral galaxies hosting Sy1 and Sy2nuclei, limited to z < .
2. We select only spiral hosts for this testsince the orientation of spirals with respect to the line-of-sightcan be quantified via their projected axis-ratio, ( b / a ). Further-more, in order to reduce noise we require a high Galaxy Zoo"spirality" probability, i.e., p > .
8. We also use the correspond-ing subsample of SF galaxies as a reference (or control) sample.Firstly, we present in Fig.4-5-( upper panel ) the Sy1 and Sy2subsample normalised redshift distributions and compare themwith the corresponding distribution of SF galaxies. As it canbe clearly seen, and also has quantified by the Kolmogorov-
Fig. 3.
The values of the factor ∆ (AGN , p) by which the di ff erent AGNtypes exceed the corresponding fractional number of SF galaxies forthe "ellipticity" ( upper panel ) and "spirality" (( lower panel ) cases. Thedi ff erent AGN and the SF subsamples are color coded as in Fig.1 whilethe shaded area corresponds to the 1 σ Poisson uncertainty.
Smirnov (KS) test (p-value ∼ ∼ − , respectively), thedistributions are significantly di ff erent. For z (cid:46) .
08, the frac-tion of both the AGN populations is lower than that of the SFgalaxies, while for higher values, z (cid:38) .
1, the AGN-fractions aregreater than that of SF galaxies. Whether due to observationalbiases or evolutionary e ff ects, understanding such di ff erences isout of the scope of the current work. However, we need to en-sure that our results will not be a ff ected by such biases and thusfollow the same resampling technique, as in section 3.1, to ob-tain matched redshift distributions which are shown in Fig. 4-5-( lower panel ).We can now proceed to a meaningful comparison of theorientation distribution of Sy1s and Sy2s with respect to theSF case. We expect that, according to the UP, the orientation-distributions of Sy1s and Sy2s, hosted in spiral galaxies must beidentical to that of spiral SF galaxies. In presenting our results,we again normalise the distributions by the total number of ob-jects in each subsample.In Fig.6 we present the comparative plot of the normalisedaxis-ratio ( b / a ) -distributions for the spiral galaxies hosting Sy1nuclei (purple) and for the control sample of SF galaxies. It is Article number, page 4 of 8. Gkini, M. Plionis, M. Chira & E. Koulouridis: Host galaxy and orientation di ff erences between di ff erent types ofAGN Fig. 4.
Upper panel:
Normalised redshift-distribution of spiral Sy1 andSF galaxies.
Lower panel:
Normalised redshift-matched distribution ofSy1’s and the corresponding SF control sample. The Sy1 and the SFsubsamples are color coded as in Fig.1 while the shaded area corre-sponds to the 1 σ Poisson uncertainty. evident that the two distributions are di ff erent, which we alsoconfirm with a KS test (p-value ∼ − ); the b / a - distributionof Sy1 is skewed towards high b / a values, indicating inclinationangles closer to face-on orientations, while the control samplecovers the full range of orientation angles. The distribution ofthe control sample shows a peak at b / a ≈ .
35, whereas theSy1 population peaks at b / a ≈ .
7. Moreover, for b / a < . b / a values close to the "edge-on" orientation.In Fig. 7 we present the respective b / a comparison plot forthe Sy2 and the star-forming subsamples and we find that al-though their distributions are significantly more similar than thecorresponding of Fig.6, the are still statistically di ff erent as con-firmed by a KS test (p-value ∼ − ). In detail, both the type-2Seyferts and the SF galaxies are distributed in the whole range of b / a , with the star-forming peaking at b / a ≈ .
35, while the Sy2appear to have two local maxima, at b / a ≈ .
37 and at b / a ≈ . Fig. 5.
Upper panel:
Normalised redshift-distribution of spiral Sy2 andSF galaxies.
Lower panel:
Normalised redshift-matched distribution ofSy2 and the corresponding SF control sample. The Sy2 and the SF sub-samples are color coded as in Fig.1 while the shaded area correspondsto the 1 σ Poisson uncertainty. b / a ∈ (0 . , . b / a < . b / a > .
4. Discussion
There are various indications for di ff erences in the host galaxiesof AGNs in the literature. Our results in the low-redshift regimeis in agreement with the results of Bornancini & García Lambas(2018), based on higher redshift (0 . < z < .
1) type-1 and type-2 AGNs from the
COSMOS2015 catalogue, who showed thatthe type-1 AGN host galaxies appear more elliptical and com-pact than those of type-2 AGN that span the whole spiral to el-liptical Hubble-type range. In addition, Sorrentino et al. (2006)using data from the Fourth SDSS Data Release (DR4) foundthat 76% of type-1 Seyfert host-galaxies are elliptical, while thecorresponding ratio of Seyfert 2 hosted in early-type galaxiesis 56.8%. Slavcheva-Mihova & Mihov (2011) found that moretype-1 than type-2 AGN prefer elliptical hosts. Similarly, Vil-larroel et al. (2017) and Chen & Hwang (2017) concluded that
Article number, page 5 of 8 & A proofs: manuscript no. aanda
Fig. 6.
Distribution of the projected axial ratio for spiral galaxies hostingSy1s and the corresponding control sample of SF galaxies. The Sy1 andthe SF subsamples are color coded as in Fig.1 while the shaded areacorresponds to the 1 σ Poisson uncertainty.
Fig. 7.
Distribution of projected axial ratios for spiral galaxies hostingtype-2 AGN and the corresponding control SF sample. The Sy2 andthe SF subsamples are color coded as in Fig.1 while the shaded areacorresponds to the 1 σ Poisson uncertainty.
Seyfert 2 nuclei reside more in spiral hosts ( ∼ − ∼ ff erent bulge distributions of Sy1 and Sy2 mightbe related to an evolutionary sequence of AGN activity (e.g.Koulouridis et al. 2013, 2006; Krongold et al. 2002; Tran 2003;Villarroel et al. 2012). These results indicate a possible co-evolution scheme between galaxies and SMBHs (e.g. Hopkinset al. 2006, 2008b,a; Springel et al. 2005). At the initial merg-ing phase, during the enhanced star-forming activity and accre-tion, the AGN is mostly obscured because of the large amountof gas and dust in the circum-nuclear region. However, the AGNwill eventually be revealed after the surrounding material is con-sumed, or expelled by radiation pressure. In parallel to the AGNevolution, the stellar population of the merger will also rapidlyredden and the host will be eventually transformed to a quiescent elliptical galaxy (e.g. Hopkins et al. 2008b,a). AGN feedbackplays an essential role to both AGN and galaxy transformationand probably leads to the observed SMBH-bulge relation (e.g.Ferrarese & Merritt 2000; Gebhardt et al. 2000; Magorrian et al.1998). Although, major mergers and powerful AGNs are far lesscommon in the local Universe, compared to redshift z >
1, thedi ff erence found in the current work could be partly due to theco-evolution of the bulge and AGN activity after such events.Alternatively or in addition, the di ff erence could also be dueto the obscuration of the BLR from the gas- and dust-rich disk ofa spiral galaxy, especially in edge-on systems. This possibility isfurther discussed in the next section. Our analysis also showed that the orientations of the Sy1 and Sy2spiral host galaxies are significantly di ff erent when compared tothe control sample of the spiral SF galaxies. This is also unex-pected within the UP and the interpretation of our results pointstowards two possible scenarios: – additional obscuration of the AGN by dust in the galacticdisk, near the circum-nuclear region, – some level of statistical co-alignment of the plane of the torusand that of the galactic disk.According to the latter scenario Sy1 galaxies, having a "face-on" oriented torus, also have more frequently face-on host galaxyorientations. However, if that were the case, we should corre-spondingly expect a higher fraction of "edge-on" host galaxiesin the Sy2-distribution, with respect to SF galaxies, which is notobserved. Without excluding this scenario, since processes likemerging may induce a rearrangement and facilitate such a co-planarity, we do not have a clear indication to confirm this hy-pothesis (see however Maiolino & Rieke (1995) for a relevantdiscussion).Our results favour the scenario where additional obscura-tion, caused by the host galaxy, might a ff ect the classificationof Seyfert types, regardless of the inclination of the torus, as alsoclaimed by Lagos et al. (2011). Several previous studies havereached similar conclusions like ours. Goulding et al. (2012)having studied a sample of nearby Compton-thick AGNs, con-cluded that the dust of host galaxy and not necessarily the com-pact torus, is the dominant obscurer of the central engine. Lacyet al. (2007) using Spitzer data of six 0 . < z < . SpitzerFirst Look Survey , concluded that the nuclear region could be ef-fectively obscured by dust on large scales, away from the torus.Additionaly, Rigby et al. (2006) using X-ray-selected AGN withspectroscopic redshifts in the
Chandra Deep Field South (CDFS) argued that part of the column density that obscures the soft X-rays may come from the galactic disk. Last but not least, Maliziaet al. (2020), using the hard X-ray selected sample of AGN, de-tected by INTEGRAL / IBIS, have shown that material located inthe host galaxy on scales of hundreds of parsecs and not alignedwith the absorbing torus can su ffi ciently hide the BLR of sometype-1 AGN causing their classification as type-2 objects andgiving rise to the deficiency of type 1 in edge-on galaxies. Article number, page 6 of 8. Gkini, M. Plionis, M. Chira & E. Koulouridis: Host galaxy and orientation di ff erences between di ff erent types ofAGN Regarding the small but significant di ff erence in the b / a distribu-tions of the spiral Sy2 and SF galaxies, at low b / a values, a possi-ble explanation could be that in extreme edge-on orientations thedust of the galactic disk can obscure not only the BLR but alsothe NLR region. In an early study, McLeod & Rieke (1995), us-ing samples of optically and soft X-ray selected Seyferts, founda bias against having inclined spiral hosts, while hard X-ray se-lected samples were found unbiased. In addition, Malkan et al.(1998) argued that in Sy2 galaxies irregular structures at largedistances can provide su ffi cient absorbing column density forthe nuclear source. Similar results were presented in Rigby et al.(2006), where they attributed the appearance of X-ray selectedAGN as optically "dull" to galactic absorption.On the other hand, there is a hypothesis that low luminos-ity Seyferts may be diluted by high-luminosity host galaxies, inwhich the continuum can hide the AGN lines. This may be thecase for high- z AGNs, where the source fully falls within thespectroscopic fiber or slit, as demonstrated by Moran & Filip-penko (2002) and later supported by Trump et al. (2009) for asample of high- z dull AGNs in the COSMOS survey. However,our sample is limited to z < . z AGNs. Therefore, we do not consider this scenario as apossible explanation of our results. Neither, a co-alignment ofthe torus with the disk could explain the deficit of Sy2 galaxiesin edge-on systems when compared with the control SF sample.We note finally that the smaller / higher fraction of Sy2 / Sy1with respect to SF galaxies at high b / a values strengthens thehypothesis of a host-galaxy contribution to the obscuration ofthe BLR. Specifically, due to the fact that in face-on galaxies,the gas and the dust of the disk does not intervene between theobserver and the active nuclei, the only obscurer of the BLR isthe torus, which apparently in some cases, it is not su ffi cient tohide the BLR, giving rise to the deficiency of type 2 in face-ongalaxies (e.g. Lacy et al. 2007).
5. Conclusions
The main purpose of the current work was to test aspects of theUnification Paradigm by searching for di ff erences in the proper-ties of the host galaxies of various AGN types. For our purposeswe use (a) the SDSS DR14 spectroscopic galaxy-catalogue, se-lecting subsamples of Sy1 and Sy2, Liner, Composite and SFgalaxies, limited to z < . ff erences- quantified by aKS two-sample test- of the various types of AGN host galaxyHubble-types, with the most significant result being that thefraction of Sy1 galaxies hosted by ellipticals is higher thanthat of any other classes of AGNs. These results can be in-terpreted within a possible co-evolution scenario betweengalaxies and SMBS.2. We also find that the orientation distributions, as revealed bythe disk axis-ratio ( b / a ), of the Sy1 and Sy2 spiral popula-tions show statistically significant di ff erences with respectto the the control sample of star-forming galaxies (whichby definition should cover all possible orientation config- urations), in conflict with the predictions of the Unifica-tion Paradigm. These di ff erences hint towards an e ff ect bywhich the dusty galactic disk has a significant contributionto the obscuration of the broad-line and partially also of thenarrow-line nuclear region. This could also interpret our pre-vious result regarding the host-galaxy Hubble-types of type-1 AGN. Indeed, the fact that we detect more Sy1, than anyother AGN type, in elliptical hosts (which as well known aredeficient of gas and dust) can be explained if the amount ofgalactic dust and gas contributes in the obscuration of thenuclear region and in particular of the BLR. Acknowledgements.
This paper has made use of the data from the SDSS projects.Funding for the Sloan Digital Sky Survey IV has been provided by the Alfred P.Sloan Foundation, the U.S. Department of Energy O ffi / University of Tokyo, the Korean Participation Group, LawrenceBerkeley National Laboratory, Leibniz Institut f¨ur Astrophysik Potsdam (AIP),Max-Planck-Institut f ¨ur Astronomie (MPIA Heidelberg), Max-Planck-Institutf ¨ur Astrophysik (MPA Garching), Max-Planck-Institut f ¨ur ExtraterrestrischePhysik (MPE), National Astronomical Observatories of China, New MexicoState University, New York University, University of Notre Dame, Observat´arioNacional / MCTI, The Ohio State University, Pennsylvania State University,Shanghai Astronomical Observatory, United Kingdom Participation Group, Uni-versidad Nacional Aut ´onoma de M´exico, University of Arizona, University ofColorado Boulder, University of Oxford, University of Portsmouth, University ofUtah, University of Virginia, University of Washington, University of Wisconsin,Vanderbilt University, and Yale University.
References
Abolfathi, B., Aguado, D. S., Aguilar, G., et al. 2018, ApJS, 235, 42Antonucci, R. 1993, ARA&A, 31, 473Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5Bornancini, C. & García Lambas, D. 2018, MNRAS, 479, 2308Bornancini, C. & García Lambas, D. 2020, MNRAS, 494, 1189Chen, Y.-C. & Hwang, C.-Y. 2017, Astrophysics and Space Science, 362Elitzur, M. & Ho, L. C. 2009, ApJ, 701, L91Elitzur, M., Ho, L. C., & Trump, J. R. 2014, MNRAS, 438, 3340Elitzur, M. & Shlosman, I. 2006, ApJ, 648, L101Ferrarese, L. & Merritt, D. 2000, ApJ, 539, L9Gebhardt, K., Bender, R., Bower, G., et al. 2000, ApJ, 539, L13Goulding, A. D., Alexander, D. M., Bauer, F. E., et al. 2012, ApJ, 755, 5Hernández-Ibarra, F. J., Krongold, Y., Dultzin, D., et al. 2016, MNRAS, 459,291Hickox, R. C. & Alexander, D. M. 2018, ARA&A, 56, 625Hopkins, P. F., Cox, T. J., Kereš, D., & Hernquist, L. 2008a, ApJS, 175, 390Hopkins, P. F., Hernquist, L., Cox, T. J., et al. 2006, ApJS, 163, 1Hopkins, P. F., Hernquist, L., Cox, T. J., & Kereš, D. 2008b, ApJS, 175, 356Hornschemeier, A. E., Heckman, T. M., Ptak, A. F., Tremonti, C. A., & Colbert,E. J. M. 2005, AJ, 129, 86Hubble, E. 1926, Contributions from the Mount Wilson Observatory / CarnegieInstitution of Washington, 324, 1Jiang, N., Wang, H., Mo, H., et al. 2016, ApJ, 832, 111Koulouridis, E. 2014, A&A, 570, A72Koulouridis, E., Plionis, M., Chavushyan, V., et al. 2013, A&A, 552, A135Koulouridis, E., Plionis, M., Chavushyan, V., et al. 2006, ApJ, 639, 37Krongold, Y., Dultzin-Hacyan, D., & Marziani, P. 2002, ApJ, 572, 169La Franca, F., Fiore, F., Vignali, C., et al. 2002, ApJ, 570, 100Lacy, M., Sajina, A., Petric, A. O., et al. 2007, ApJ, 669, L61Lagos, C. D. P., Padilla, N. D., Strauss, M. A., Cora, S. A., & Hao, L. 2011,MNRAS, 414, 2148Lintott, C. J., Schawinski, K., Slosar, A., et al. 2008, MNRAS, 389, 1179Magorrian, J., Tremaine, S., Richstone, D., et al. 1998, AJ, 115, 2285Maiolino, R. & Rieke, G. H. 1995, ApJ, 454, 95Malizia, A., Bassani, L., Stephen, J. B., Bazzano, A., & Ubertini, P. 2020, A&A,639, A5
Article number, page 7 of 8 & A proofs: manuscript no. aanda