Seyfert galaxies in the local Universe (z ≤ 0.1): the average X-ray spectrum as seen by BeppoSAX
aa r X i v : . [ a s t r o - ph ] J a n Astronomy & Astrophysics manuscript no. dadina2 November 2, 2018(DOI: will be inserted by hand later)
Seyfert galaxies in the local Universe (z ≤ BeppoSAX
Mauro Dadina , INAF/IASF-Bo, via Gobetti 101, 40129 Bologna, Italy Dipartimento di Astronomia dell’Universit`a degli Studi di Bologna, via Ranzani 1, 40127 Bologna, ItalyReceived date/ Accepted date
Abstract.
The
BeppoSAX archive is currently the largest reservoir of high sensitivity simultaneous soft and hard-X raydata of Seyfert galaxies. From this database all the Seyfert galaxies (105 objects of which 43 are type I and 62are type II) with redshift lower than 0.1 have been selected and analyzed in a homogeneous way (Dadina 2007).Taking advantage of the broad-band coverage of the
BeppoSAX
MECS and PDS instruments ( ∼ ∼ ∼
290 keV), and therelative amount of reflection (R ∼ α energy centroid) are similar in type Iand type II objects while the absorbing column and the iron line equivalent width significantly differ between thetwo classes of active galactic nuclei with type II objects displaying larger columns (N H ∼ × and 6.1 × cm − for type I and II objects respectively) and equivalent width (EW ∼
220 and 690 eV for type I and II sourcesrespectively). Confirming previous results, the narrow FeK α line is consistent, in Seyfert 2, with being producedin the same matter responsible for the observed obscuration. These results, thus, support the basic picture ofthe unified model. Moreover, the presence of a X-ray Baldwin effect in Seyfert 1 has been here measured usingthe 20-100 keV luminosity (EW ∝ L(20-100) − . ± . ). Finally, the possible presence of a correlation between thephoton index and the amount of reflection is confirmed thus indicating thermal Comptonization as the most likelyorigin of the high energy emission for the active galactic nuclei included in the original sample. Key words.
X-rays: galaxies – galaxies: Seyfert: – galaxies: active
1. Introduction
The high energy emission from active galactic nuclei(AGN) is thought to come from the innermost regions ofaccretting systems that are centered around super-massiveblack-holes (SMBH). For this reason, X-rays are expectedto be tracers of the physical conditions experienced bymatter before disappearing into SMBH. Moreover, thanksto their high penetrating power, energetic photons, es-caping from the nuclear zones, test the matter locatedbetween their source and the observer. Thus, they offerpowerful diagnostics to understand the geometry and thephysical conditions of the matter surrounding the SMBH.The broad-band of
BeppoSAX offered for the very firsttime the opportunity to measure with a remarkable sen-sitivity, the spectral shape of AGN in the 0.1-200 keVrange. This potential has been previously exploited tostudy in details a number of sample selected in differentmanners (see for example Maiolino et al. 1998; Perola et al. 2002). These studies were fundamental in making impor-tant steps forward in the comprehension of the emittingmechanism at work in the production of X-rays (Perola etal. 2002) and to partially unveil the geometry of the coldmatter surroundings the central SMBH (Maiolino et al.1998; Bassani et al. 1999; Risaliti et al. 1999).The
BeppoSAX database full potential, however, wasnever exploited before. In a previous paper, the entire cat-alog of the Seyfert galaxies at z ≤ BeppoSAX archive has been presented (Dadina 2007).This sample was selected starting from the catalog ofSeyfert galaxies contained in the V´eron-Cetty & V´eron(2006) sample of AGN and contains 13 radio-loud objectsand 7 narrow-line Seyfert 1.The spectral analysis was performed in the 2-100 keVband whenever possible and the data were fit with a set oftemplate models to obtain a homogeneous dataset. Herethe X-ray data so collected are statistically inspected inorder to infer what are the average characteristics of the
Mauro Dadina: Average X-ray properties of Seyferts observed by
BeppoSAX nearby Seyfert galaxies contained in this sample in the2-100 keV band. Finally, present dataset is used to per-form simple tests on the unified scheme (UM) for the AGN(Antonucci 1993) and on the emission mechanism actingin the core of the Seyfert galaxies. More detailed anal-ysis on this latter topic will be presented in another pa-per (Petrucci, Dadina & Landi, submitted) where detailedthermal Comptonization models (Poutanen & Svensson1996, Haardt & Maraschi 1993 ) will be used to fit theBeppoSAX data with the main scope to study the depen-dence of the spectral properties in the “two phase” sce-nario (Haardt 1991, Haardt & Maraschi 1993) assumingdifferent geometries of the corona.
2. Mean X-ray Properties
Scope of this section is to determine and study the meanX-ray spectral properties of the sample and to use thesevalues to test the UM model for AGN (Antonucci 1993).The key parameters are the ones that describe the con-tinuum and the absorption properties. In the frameworkof the UM models for AGN, the continuum shape is ex-pected to be independent from the orientation angle underwhich the source is observed. Thus, no difference shouldbe measured in the parameters describing the continuumbetween type I and II objects. On the contrary, the absorb-ing column intervening in Seyfert 2 should be the princi-pal discriminator between the two classes of objects. Thusmeasuring the mean X-ray properties means to test thebasic assumptions of the UM.
The origins of the X-ray photons from AGN are thought tobe due to Comptonization of optical-UV radiation, comingfrom the accretion disk and Comptonized by the e − in thehot corona that sandwiches the disk (Haardt & Maraschi1991; Haardt 1993, Poutanen & Svensson 1996, Czernyet al. 2003). The mechanism is assumed to be, at leastat the zero-th order, very similar in each Seyfert galaxy.Under this hypothesis, the differences between the X-rayspectra of different objects are supposed to be mainly dueto two kind of factors: i) the time-dependent state of theemitting source; ii) the intervening matter that imprintson the emerging spectrum the features typical of its phys-ical state. In such a scenario, the observations of manysources can be regarded as the long-term monitoring of asingle source. On the other hand, it is also true that thecontrary has some comparison in the literature: e.g. timesparse observations of a single source in different statescan resemble observations of objects with completely dif-ferent spectral properties. This is the case, for example, ofthe narrow line Seyfert 1 NGC 4051 that displayed vari-ations in flux/luminosity by a factor of ∼
100 (Guainazziet al 1998) associated to strong variations of the spec-tral shape (Γ ∼ H measured in each sin-gle observation were treated separately.In between these two cases, are the properties of theemission FeK α line. BeppoSAX had a too low sensitiv-ity to detect the relativistically broadened component ofthis feature in a large number of sources. Thus it turned-out that only the narrow line has been detected in thevast majority of the objects included in the original cata-log. This component is supposed to originate far from theSMBH, at least at ∼ α line were observed to vary with its line energy centroidchanges between ∼ ∼ α line were, thus, treated as variable ones in order to testits origin, and not averaged when objects were observedmore than once.Finally, it is worth noting here that in a number ofcases, it was necessary to deal with censored data (see ta-ble 1). To manage these data properly, the ASURV soft- auro Dadina: Average X-ray properties of Seyferts observed by BeppoSAX ware (Feigelson & Nelson 1985; Isobe et al. 1986) hasbeen used. In particular, to establish if the distributionsof parameters of type I and type II objects were drawnfrom different parent populations, the Peto & Prenticegeneralized Wilcoxon test (Feigelson & Nelson 1985) hasbeen used while to calculate the mean values consider-ing also the censored data the Kaplan-Meyer estimatorhas been used. To establish the presence of correlationsbetween different quantities, both the Spearman’s ρ andthe generalized Kendall’s τ methods were applied. Thelinear regressions were calculated using the Bukley-Jamesand Schmitt‘s methods. In the following, two quantitiesare considered as drawn from different parent populationwhen the probability of false rejection of the null hypoth-esis (same parent population) is P null ≤ Table 1.
General characteristics of the data analyzed inthis work. The number of detections and censored data arereported for the interesting parameters for the whole sam-ple of objects (columns 2 and 3), for the Seyfert 1 galaxies(columns 4 and 5), and for the Seyfert 2 objects (column6 and 7). The 90% confidence interval limits were usedfor censored data and the detected values were defined ifdetermined with a 99% confidence level (Dadina 2007)
Parameter Tot. Sample Seyfert 1 Seyfert 2Det. Cens. Det. Cens. Det. Cens.N H
83 80 31 53 52 27EW 129 7 66 4 63 3R 68 18 46 9 22 9Ec 33 51 27 26 6 25
The X-ray continua of the sources have been modeled us-ing a cut-off power-law, that describes the primary emis-sion from the hot corona plus a reflection component(
P EXRAV model in Xspec, Magdziarz & Zdziarski 1995)to account for the contribution expected to be due to thedisk. It is worth noting, however, that an additional re-flection component could rise from the torus (Ghisellini,Haardt & Matt 1994) and that the disentangling betweenthe two reflection component is impossible with the qual-ity of the available data. Wherever its origin, the reflectionhas been assumed to be due to cold matter. The inter-esting parameters are the photon index (Γ), the relativeamount of reflection (R), and the high-energy cut-off (Ec). In table 2 the results for the whole set of observationsand for the two classes of Seyferts are reported as well asthe probability that Seyfert 1 and 2 are drawn from thesame parent populations. The histograms of the distribu-tions of the interesting parameters for the entire sampleof objects (first column) and for type I (second column)and type II (third column) are reported in figure 1.
Table 2.
Mean spectral properties. Col. I: Spectral pa-rameter; Col. II: Seyfert 1 mean value; Col.III: Seyfert 2mean value; Col. IV: Probability that Seyfert 1 and Seyfert2 are drawn from the same parent populations.
Parameter Tot. Seyfert 1 Seyfert 2 P null
Γ 1.84 ± ± ± ± ± ± † ±
24 230 ±
22 376 ±
42 5%N H ‡ ± ± ± ≤ † in units of eV; ‡ in units of 10 cm − As previously said, the UM for AGN (Antonucci 1993)predicts that Γ, R, and Ec are observables independentfrom the inclination angle, thus the two classes of Seyfertgalaxies should display very similar characteristics. This isconfirmed by the analysis of the present sample. In partic-ular, there are no hints that the distributions of photon in-dex Γ for the two types of Seyfert are drawn from differentparent populations (P null ∼ ∼ ≤
1. Type II objects that show so hard X-ray spec-tra are supposed to be Compton-thick sources for which,in the 2-10 keV band, only the reflected/flat spectrum isobservable. This is the case for NGC 2273 which displaysthe harder X-ray spectrum. This source was first classi-fied as a Compton-thick object by Maiolino et al. (1998).The Seyfert 1s with flattest spectra are NGC 4151 that isknown to have a hard spectrum with complex and variableabsorption (De Rosa et al. 2006) and Mrk 231 (classifiedas a type I AGN by Farrah et al. 2003). The latter sourceshows a very hard X-ray spectrum with Γ ∼ XM M - N ewton and
BeppoSAX data, these authors speculatedthat the spectrum of the source below ∼
10 keV is reflec-tion dominated, thus presenting a case that can be hardlyreconciled with the UM of AGN. Moreover, Mrk 231 is
Mauro Dadina: Average X-ray properties of Seyferts observed by
BeppoSAX
Fig. 1.
Photon index Γ ( first row ), R ( second row ), Ec ( third row ), and N H ( fourth row ) in units of cm − distributions for thewhole dataset ( left column ), for type I objects ( center column ), and for type II objects ( right column ). an ultra-luminous infra-red galaxy. These sources displaystrong starburst activity that can dominate the total X-ray luminosity of the galaxy (see Franceschini et al. 2003and Ptak et al. 2003 for details on this topic) althoughin Mrk 231 at least 60% of the observed 0.5-10 keV fluxseems to be due to the AGN component (Braito et al.2004).Less conclusive results are obtained for R and Ec. Inboth cases, the probability of false rejection of the null hy-pothesis (the two distributions are drawn from same par-ent populations, in accordance with the UM predictions),is P null ∼ ∼ null ∼ ∼ cm − ) and some Seyfert 2have low columns (down to 10 cm − ). These are not newresults: the high column of NGC 4151 ( ∼ × , De Rosaet al. 2006) is well known. The highest column measuredin a Seyfert 1 is detected during the June 9, 1998 obser-vation of NGC 4051. During this observation the sourceappeared “switched-off” and only a pure reflection compo-nent was measured (Guainazzi et al. 1998). The spectralfit in Dadina (2007) degenerated between two-solutions,one in which the source was purely reflection-dominated(R ≥
7) and a second one in which a direct componentwas visible but highly absorbed. The latter scenario wasslightly preferred by a pure statistical point of view whenthe 2-200 keV band is condidered, and, for homogeneity,entered in the catalog (Dadina 2007). Nonetheless, whenthe entire
BeppoSAX band (0.1-200 keV) is considered,the reflection scheme is preferred (Guainazzi et al. 1998).Finally, a number of objects show upper-limits of theorder of ∼ cm − to the absorbing column. This is a auro Dadina: Average X-ray properties of Seyferts observed by BeppoSAX Fig. 2.
Distribution of the Fek α emission line energy centroid ( first row ) and of its EW ( second row ). Distributions for the wholesample of observations ( left panel ), for the type I objects ( center panel ) and for type II objects ( right panel ). selection effect induced by the energy band considered us-ing only MECS and PDS data ( ∼ cm − )peaks in fact at E ∼ H below 10 cm − . This is most probably responsible ofthe high value obtain for the average N H in type I objects.
3. Probing the origin of the FeK α emission line The FeK α line is produced by reprocessing the primary X-ray emission in matter surrounding the source of hard pho-tons. In the framework of the UM, the origin of this com-ponent can be placed in a number of region s such as theaccretion disk, the dusty torus, and the broad-line regions(even if this last hypothesis is disfavored by the recentresults obtained with the XM M - N ewton and
Chandra observatories and presented in Nandra 2006). If the lineoriginates in the disk close to the SMBH, relativistic ef-fects that broaden the resulting line are expected. For thevast majority of the sources included in the original sam-ple, only narrow component of such feature were detectedand only in a few cases broad emission lines (e.g. IC 4329a)or relativistically blurred features were detected (for ex-ample in MCG-6-30-15). Thus, the results presented hereare essentially based on the measured properties of thenarrow features.As shown in figure 2 (first row), the line energy cen-troid is peaked at ∼ ∼
200 eV FWHM, Boella et al.1997), the results obtained here are in agreement with theline being mainly produced in cold or nearly cold matter(ionization state below FeXVII), i.e. in both type I and IIobjects, by matter in the same physical state. However, awell known difference between the two classes of objects isthe EW of the narrow FeK α line in type II objects, whichshows stronger features than type I objects (see secondrow of figure 2 and table 3, Bassani et al 1999, Risaliti et al. 1999, Cappi et al. 2006 and Panessa et al. 2006).As shown in the central panel of figure 2, the Seyfert 1peak at EW F eKα ∼ α line (above 300-400 eV). The large EWtail of the Seyfert 1 distribution is composed mainly of ob-jects in which the broad components of the FeK α line aredetected such as in MCG-6-30-15, Mrk 841 and IC 4329a.The Seyfert 1 with largest EW is NGC 4051 during theJune 9, 1998 observation when its spectrum was due topure reflection (Guainazzi et al. 1998).The larger FeK α EW in Seyfert 2 galaxies is in agree-ment with the UM (Antonucci 1993). If the origin of thiscomponent is indeed located in the dusty torus, than theline EW has to be correlated with the absorber columndensity. This is indeed what is observed also in this sample(see figure 3, left panel). Moreover, the Spearman ρ andKendall’s τ tests indicate that a correlation between theFeK α EW and the N H is highly probable for type II ob-jects (P null ≤ H estimates have been tested by corre-lating it with the model independent indicator offered bythe ratio of the observed fluxes at 2-10 and 20-100 keV re-spectively (center panel of figure 3). The two quantities arestrongly correlated (generalized Spearman ρ and Kendal τ tests give P null ≤ ∼
10 keV while at harder energies the ra-diation pierces the matter for columns ≤ × cm − .Otherwise, the Compton absorption dominates and alsophotons with energy above 10 keV are stopped since theKlein-Nishina regime is reached. Mauro Dadina: Average X-ray properties of Seyferts observed by
BeppoSAX
Fig. 3.
Left panel : Log(EW
F eKα ) vs. Log(N H ). As expected, the sources are divided in two families: the ones that follow theexpectations if the FeK α line is produced in the absorption matter and the candidate Compton-thick ones that display lowabsorption and large EW. The solid line indicates the prediction by Makishima (1986). Center panel : Log(F − keV /F − keV )vs. Log(N H ). The two quantities are correlated as expected since the 2-10 keV band is strongly affected by the absorption whilethe 20-100 band is almost free from absorption. The solid line is the best fit obtained with linear regression methods Right Panel :Log(EW
F eKα ) vs. Log(F − keV /F − keV ). The two quantities are strongly correlated ( P null ≤ α line in emission is indeed produced by the same matter responsible for the absorption. Thesolid line is the linear regression obtained using Bukley-James method (Isobe et al. 1986). Table 3.
Mean properties of the FeK α emission line. Col.I: spectral parameter; Col. II mean value for the wholesample; Col. III: mean value for Seyfert 1; Col. IV: meanvalue for Seyfert 2. Col. V: Probability that the parame-ters of type I and type II objects are drawn from the sameparent population. Parameter Tot. Seyfert 1 † Seyfert 2 † P null E F eKα † ± ± ± F eKα ‡ ±
67 222 ±
33 693 ± ≤ † in units of keV; ‡ in units of eV As stated above, when the EW of the FeK α emissionline is tested against the measured N H (left panel of Figure3) a result in good agreement with what predicted by the-oretical models is obtained (Makishima 1986; Leahy &Creighton 1993). The majority of the sources, in fact, be-have as expected if the line is produced by the absorb-ing matter that depress the direct continuum (Makishima1986). All the known Compton-Thick sources are locatedat low N H and high EW, in accordance with previous re-sults (Bassani et al. 1999, Risaliti et al. 1999 ). As anadditional test, the EW of the FeK α line has been plottedversus the F − keV /F − keV ratio. As expected (seeright panel of figure 3), a good correlation ( P null ≤ ρ and Kendall τ tests)is obtained. These results thus confirm that the properties of theFeK α line agree with the expectations of the UM forAGN. Nonetheless, this is not the only information wehave about the Iron line. In recent papers (Iwasawa &Taniguchi 1993; Page et al. 2004, Grandi et al. 2006;Bianchi et al. 2007) it has been claimed that a X-ray“Baldwin effect” (or Iwasawa-Taniguchi effect) is presentin AGN when the FeK α intensity is probed against the2-10 keV luminosity. Here this effect is tested consideringfor the first time both the 2-10 keV and the 20-100 keVluminosities.A strong correlation ( P null ≤ ρ and generalized Kendal τ test) is found when the EW ofthe FeK α line is plotted both against the observed 2-10keV (panel (a) of figure 4) and 20-100 keV (panel (c) of fig-ure 4) keV luminosities. The nature of these correlations,however, is not straightforward, especially for the 2-10 keVluminosity. In this energy band the effect of the absorberis very important. As previously demonstrated, the EW ofthe Iron line correlates with N H , but stronger absorbingcolumns imply lower fluxes. Moreover, when the relationbetween the FeK α line EW and the 2-10 keV flux is in-vestigated (panel (b) of figure 4), it is found that the twoquantities are correlated (P null ≤ α line must be strong enough. Since the originalsample is limited to the local Universe, this correlation influx acts, at least partially, also in the EW vs L − keV relation. auro Dadina: Average X-ray properties of Seyferts observed by BeppoSAX Fig. 4.
Panel (a) : Log(EW
F eK α ) vs. L − ,observed . Panel (b) : Log(EW
F eK α ) vs. F − ,observed . Panel (c) : Log(EW
F eK α ) vs.L − ,observed . Panel (d) : The same of panel (c) but only for Seyfert 1 objects.
Panel (e) : Log(EW
F eK α ) vs. F − ,observed .Solid lines in panel (a), (b), and (c) are the linear regressions obtained for the whole sample of observations. The dashed linein panels (c) and (e) is the linear regression obtained considering only type I objects. As visible in panel (b) the less scatteredrelation is obtained considering the 2-10 keV observed flux. The relation is linear, as expected if the correlation is due to selectioneffects, i.e.considering that in faint objects only large EW were detectable by the MECS instruments on-board BeppoSAX . This N H effect should be negligible when the 20-100keV band is considered. In fact, in this case, one expectsto find a correlation between the observed 20-100 keV lu-minosity and the EW of the FeK α line only in the mostextreme cases, i.e. for the “pure Compton thick” objects.Apart from NGC 1068, these sources are too weak to bedetected by the PDS, thus unable to drive the relation ob-served in plot (c) of figure 4. Moreover, panel (e) of figure 4indicates that the EW of the iron line is not related to the20-100 keV flux thus excluding that the “Baldwin effect”measured using the 20-100 keV band is due to instrumen-tal selection effect as it happens for the F − keV . On theother hand, in the 20-100 keV band the reflection-hump at ∼ α line is due to the same matter responsible ofthe reflection, one should expect that the EW of the Ironline should increases as the reflection component augmentsthe 20-100 keV flux (i.e. the Iron line EW should correlatewith the 20-100 keV flux). Thus, the net contribution ofthe reflection component should act in the opposite direc-tion to that observed (i.e. larger iron line EW at smaller 20-100 keV flux). If the two classes of Seyfert galaxies areanalyzed separately it is obtained that a strong correlationis found for Seyfert 1 (P null ≤ ρ and generalized Kendal τ tests, panel (d) of figure 4) whileno correlation is evident for type II objects (P null ∼ − keV is considered and it hasthe following relation: Log(EW)=-(0.22 ± × Log(L − )+11.91 ± The slope of the relation found in this work is in agree-ment with what previously obtained by Page et al. (2004)(EW ∝ (L − ) − . ± . ) using a sample containing bothradio-quiet and radio-loud objects. Present result is alsoconsistent with what found by Zhou & Wang (2005) (whoused a sample containing both radio-quiet and radio-loudobjects founding EW ∝ (L − ) − . ± . ) and Bianchi etal. (2007) (who used only radio-quiet objects obtaining Mauro Dadina: Average X-ray properties of Seyferts observed by
BeppoSAX
Fig. 5.
Left panel:
Photon index Γ plotted versus the measured N H (in units of cm ). No significance trend relating these twoquantities is found; center panel: R vs. Ec (in units keV), no correlation is found between these quantities; right panel
R vs. Γ.The two quantities are correlated with a high significance level (P null ≤ EW ∝ (L − ) . ± . ). On the contrary, Jiang, Wang &Wang (2006) found that, excluding the radio-loud AGNfrom a sample similar to the one used by Page et al.(2004), found that EW ∝ (L − ) . ± . . It is worth re-calling here, that the the present sample is composed byboth radio-loud and radio quiet sources (Dadina 2007). Inparticular, it contains 7 radio-loud Seyfert 1, and onlyfor 5 of them the 20-100 and iron line data are avail-able. Nonetheless, the presence of these sources in thesample does not affect the FeK α EW vs. L − rela-tion (EW ∝ (L − ) . ± . excluding radio-loud type I ob-jects).The origin of the X-ray “Baldwin effect” is unclear. Inthe light bending scenario (Miniutti & Fabian 2004) theheight of the source above the accretion disk determinesthe degree of beaming along the equatorial plane of thehigh energy emission. Because of relativistic effects, thecloser the source is to the disk, the greater will be the frac-tion of X-rays beamed in the equatorial plane (i.e. towardsthe disk) and correspondingly lower will be the observedflux. Thus, the EW of the relativistically blurred FeK α line produced in the inner regions of the disk and the EWof the narrow Iron line produced in the outer parts of thedisk would appear enhanced in low-state sources.On the contrary, Page et al. (2004) speculated that thiseffect could indicate that luminous sources are surroundedby dusty tori with lower covering factor thus pointing to-wards a torus origin of most of the narrow FeK α line.The present work supports this view. The FeK α line EWof the Iron line correlates with the observed N H as pre-dicted by theory (Makishima 1988; Lehay & Creighton1993; Ghisellini, Haardt & Matt 1994). Moreover, also thecase of the extremely low state of NGC 4051 (Guainazziet al. 1998) included in this dataset seems to point in thisdirection. The huge EW of the narrow FeK α line recordedin this observation, in fact, is typical of Compton-thicktype II objects, but the line does not show evidence of rel- ativistic broadening due to the contribution of the innerorbits of the accretion disk.
4. The Γ -R relation. In the fitting procedure some parameters may degenerategiven the interdependence among them. This is the case,for example of the photon index with the column densityfor low statistics observations. The same could happen forthe determination of R and Ec, since R introduces in thespectrum a bump peaked between 20-40 keV and decliningat higher energies where the Ec may appear. To check ifthe results presented here are affected by such effects, thecorrelations between these parameters have been studiedand the results are presented in Figure 5.Left panel of Figure 5 shows how, on average, the esti-mate of Γ is not affected by the simultaneous determina-tion of the absorbing column. In fact, no trend is observedbetween Γ and N H . Obviously, this does not imply thatthis is true for each single source included in the origi-nal sample. On the other hand, this is an expected resultsince the broad band of BeppoSAX should reduce thisspurious effect. Similar results are obtained also when theΓ vs. Ec, and R vs. Ec (center panel of Figure 5) rela-tions are investigated. In these cases the Spearman ρ andKendall’s τ tests do not sustain the existence of a relationbetween these quantities (P null ∼ null ≤ R=(4.54 ± × Γ-(7.41 ± It is hard to define if this correlation is the result ofa systematic effect or not since it is possible that thesetwo quantities degenerate in the fitting procedure. Flat auro Dadina: Average X-ray properties of Seyferts observed by
BeppoSAX power-law with small reflection could be described also bysteep power-law plus strong reflection. The total absenceof similar correlations between Γ and Ec and R vs. Ecseems to suggest that this correlation is indeed real.A similar relation was previously found using Ginga and
RXT E data (Zdziarski, Lubinski, & Smith 1999;Gilfanov, Churazov & Revnivtsev 1999). Zdziarski,Lubinski, & Smith (1999) interpreted it as evidence ofthermal Comptonization as origin of X-rays providing thatthe optical-UV seed photons were mainly produced by thesame material responsible for the reflected component. Inthis case, in fact, the cooling rate of the hot corona is di-rectly linked to the power-law slope. But the cooling rateis also related to the angle subtended by the reflector. Thisresult is also in agreement with predictions of models thatconsider mildly relativistic outflows driven by magneticflares (Beloborodov 1999). More in general, Merloni et al.(2006) demonstrated that any geometry in which the hot,X-ray emitting plasma, is photon starved (i.e. geometriesof the accretting systems in which the accretion disk isonly partially covered by the Comptonizing plasma suchas patchy coronae, inner ADAF plus outer disks etc.) willproduce hard X-ray spectra, little soft thermal emissionand weak reflection component. On the other hand, ge-ometry corresponding to a very large covering fraction ofthe cold phase, have strong soft emission, softer spectraand strong reflection fraction (Collin-Souffrin et al. 1996).Thus, moving along the Γ vs. R relation it implies movingfrom lower to higher accretting systems.
5. Summary and conclusions
The average properties of Seyfert galaxies in the localUniverse (z ≤ BeppoSAX has been in-vestigated, analyzing the sample of objects presented inDadina (2007). Multiple observations of single objectswere treated independently, i.e. the multiple measure-ments of parameters were not averaged for statistical pur-poses. This method has been chosen since, in the frame-work of the simplest version of UM for AGN (Antonucci,1993) the AGN are thought to be very similar to each-other and only orientation/absorption effects and theactivity-level of the targets could introduce observationaldifferences between different objects. In this scenario, themonitoring of a single source could be reproduced by theobservations of many sources in different states and “viceversa”.
BeppoSAX offered the advantage of a useful X-raybroad energy band. Data studied here fall in the 2-100keV band for a majority of objects. This advantage hasbeen used to investigate the properties of the high-energycontinuum of Seyfert galaxies. As stated in the previouspaper (Dadina, 2007), the basic template was a power-law with a high energy cut-off plus a reflection component(namely PEXRAV model in XSPEC). The results of thisanalysis can be summarized as follow: – the average slope of the power-law is 1.84 ± ± ± – the average value of the relative reflected-to-direct nor-malization parameter R is 1.01 ± ± ± – the high energy cut-off was measured to be Ec=287 ± ±
22 keV for Seyfert 1 and Ec=376 ±
42 keV forSeyferts 2); – as expected and as known from previous works, theabsorbing column is very different in the two classesof objects. On average N H ∼ × cm − type Iand N H ∼ × cm − for type II AGN. The highmean value obtained for Seyfert 1 is caused by a se-lection effect induced by the energy coverage of theMECS+PDS instruments (2-100 keV). – evidence of a X-ray Baldwin effect is found in Seyfert1 galaxies when the EW of the FeK α line is plottedagainst the 20-100 keV luminosity. – a significant correlation has been found between R andΓ.These results are well in agreement with the basic as-sumptions of the UM for AGN (Antonucci 1993). In fact,no differences are measured in the observables that aresupposed to be isotropic while the absorbing column seemsto be the only discriminator between the different types ofSeyfert galaxies. This reflects also in the properties of theFeK α line. No difference is measured in the line centroid(see table 3) between the two classes of Seyfert galaxies.Type II objects, however, display more intense features(EW=222 ±
33 eV for Seyfert 1 and EW=693 ±
195 eV forSeyfert 2). The physical origin of the X-ray “Baldwin ef-fect” here measured for Seyfert 1 using the 20-100 lumi-nosity is unclear. Both light bending (Miniutti & Fabian2004) and torus models (Page et al. 2004) are consistentwith present data even if the strong relation of the FeK α line EW in type II objects with the absorption column in-dicates that the most of the narrow line component shouldbe due to the torus. Finally, the measured Γ-R relationshipis consistent with thermal Comptonization models. Acknowledgements.
I thank G.G.C. Palumbo and M. Cappifor helpful discussion and for careful reading of the previousversions of the manuscript. I also thank the ASDC people fortheir wonderful work in mantaining the
BeppoSAX database.I really thank the referee for her/his helpful comments andsuggestions that contributed to improve the quality of themanuscript. Financial support from ASI is aknowledged.
References
Antonucci, R. 1993, ARA&A, 31, 4730 Mauro Dadina: Average X-ray properties of Seyferts observed by