Pdf/07/aa4031-05.pdf.url

A&A 447, 97–112 (2006) The host galaxy/AGN connection
in nearby early-type galaxies
,


Is there a miniature radio-galaxy in every "core" galaxy?
B. Balmaverde1 and A. Capetti2 1 Universitá di Torino, Via Giuria 1, 10125, Torino, Italy 2 INAF - Osservatorio Astronomico di Torino, Strada Osservatorio 20, 10025 Pino Torinese, Italy Received 11 August 2005 / Accepted 20 September 2005 This is the second of a series of three papers exploring the connection between the multiwavelength properties of AGN in nearby early-typegalaxies and the characteristics of their hosts. We selected two samples with 5 GHz VLA radio flux measurements down to 1 mJy, reachinglevels of radio luminosity as low as 1036 erg s−1. In Paper I we presented a study of the surface brightness profiles for the 65 objects withavailable archival HST images out of the 116 radio-detected galaxies. We classified early-type galaxies into "core" and "power-law" galaxies,discriminating on the basis of the slope of their nuclear brightness profiles, following the Nukers scheme. Here we focus on the 29 core galaxies(hereafter CoreG).
We used HST and Chandra data to isolate their optical and X-ray nuclear emission. The CoreG invariably host radio-loud nuclei, with anaverage radio-loudness parameter of Log R = L5 GHz/LB ∼ 3.6. The optical and X-ray nuclear luminosities correlate with the radio-core power,smoothly extending the analogous correlations already found for low luminosity radio-galaxies (LLRG) toward even lower power, by a factorof ∼1000, covering a combined range of 6 orders of magnitude. This supports the interpretation of a common non-thermal origin of the nuclearemission also for CoreG. The luminosities of the nuclear sources, most likely dominated by jet emission, set firm upper limits, as low asL/LEdd ∼ 10−9 in both the optical and X-ray band, on any emission from the accretion process.
The similarity of CoreG and LLRG when considering the distributions host galaxies luminosities and black hole masses, as well as of thesurface brightness profiles, indicates that they are drawn from the same population of early-type galaxies. LLRG represent only the tip of theiceberg associated with (relatively) high activity levels, with CoreG forming the bulk of the population.
We do not find any relationship between radio-power and black hole mass. A minimum black hole mass of MBH = 108 M is apparentlyassociated with the radio-loud nuclei in both CoreG and LLRG, but this effect must be tested on a sample of less luminous galaxies, likely tohost smaller black holes.
In the unifying model for BL Lacs and radio-galaxies, CoreG likely represent the counterparts of the large population of low luminosity BL Lacnow emerging from the surveys at low radio flux limits. This suggests the presence of relativistic jets also in these quasi-quiescent early-type"core" galaxies.
Key words. galaxies: active – galaxies: bulges – galaxies: nuclei – galaxies: elliptical and lenticular, cD – galaxies: jets –
galaxies: BL Lacertae objects: general
which to explore the classical issue of the connection betweenhost galaxies and AGN.
The recent developments in our understanding of the nuclearregions of nearby galaxies provide us with a new framework in All evidence now points to the idea that most galaxies host a supermassive black hole (SMBH) in their centers (e.g.
Kormendy & Richstone 1995) and that its mass is closely Based on observations obtained at the Space Telescope Science Institute, which is operated by the Association of Universities linked to the host galaxies properties, such as the stellar veloc- for Research in Astronomy, Incorporated, under NASA contract ity dispersion (Ferrarese & Merritt 2000; Gebhardt et al. 2000).
NAS 5-26555.
This is clearly indicative of a coevolution of the galaxy/SMBH

Appendix A and B are only available in electronic form at system and it also provides us with indirect, but robust, SMBH mass estimates for large sample of objects. Furthermore, B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? the innermost structure of nearby galaxies have been revealed to separate these early-type galaxies into core and power-law by HST imaging, showing the ubiquitous presence of singu- galaxies following the Nukers scheme, rather than on the tra- lar starlight distributions with surface brightness diverging as ditional morphological classification (i.e. into E and S0 galax- Σ(r) ∼ r−γ with γ > 0 (e.g. Lauer et al. 1995). The distribution ies). Here we focus on the sub-sample formed by the 29 "core"of cusp slopes (Faber et al. 1997) is bimodal, with a paucity of objects with 0.3 < γ < 0.5. Galaxies can then be separated on We adopt a Hubble constant H0 = 75 km s−1 Mpc−1.
the basis of their brightness profiles in the two classes of "core"(γ ≤ 0.3) and "power-law" (γ ≥ 0.5) galaxies, in close corre-spondence to the revision of the Hubble sequence proposed by 2. A critical analysis of the classification as core
Kormendy & Bender (1996).
But despite these fundamental breakthroughs we still lack a clear picture of the precise relationship between AGN and In Paper I we adopted the classification into power-law and host galaxies. For example, while spiral galaxies preferentially core galaxies following the scheme proposed by Lauer et al.
harbour radio-quiet AGN, early-type galaxies host both radio- (1995). We then separated early-type galaxies on the basis of loud and radio-quiet AGN. Similarly, radio-loud AGN are gen- the slope of their nuclear brightness profiles obtained using the erally associated with the most massive SMBH as there is a Nukers law (i.e. a double power-law with innermost slope γ) median shift between the radio-quiet and radio-loud distribu- defining as core-galaxies all objects with γ ≤ 0.3. Since this tion, but both distributions are broad and overlap considerably strategy has been subsequently challenged by Graham et al.
(e.g. Dunlop et al. 2003).
(2003), who introduced a different definition of core-galaxies, In this framework, in two senses early-type galaxies appear it is clearly important to assess whether the identification of to be the critical class of objects, where the transition between an object as a core galaxy is dependent on the fitting scheme the two profiles classes occurs (i.e. in which core and power- law galaxies coexist) and in which they can host either radio- Graham et al. argued that a Sérsic model (Sérsic 1968) loud and radio-quiet AGN. We thus started a comprehensive provides a better characterization of the brightness profiles of study of a sample of early-type galaxies (see below for the early-type galaxies. In particular they pointed out that, among sample definition) to explore the connection between the mul- other issues, i) the values of the Nukers law parameters depend tiwavelength properties of AGN and the characteristics of their on the radial region used for the fit; ii) the Nukers fit is unable hosts. Since the "Nuker" classification can only be obtained to reproduce the large scale behaviour of early-type galaxies when the nuclear region, potentially associated with a shal- and, most importantly for our purposes; iii) the identification low cusp, can be well resolved, such a study must be limited of a core galaxy from a Nuker fit might not be recovered by a to nearby galaxies. The most compact cores will be barely re- Sérsic fit. Conversely, they were able to fit power-law galaxies solved at a distance of 40 Mpc (where 10 pc subtend 0.05) even (in the Nukers scheme), as well as dwarf ellipticals (Graham & in the HST images. Furthermore, high quality radio-images are Guzmán 2003), with a single Sérsic law over the whole range required for an initial selection of AGN candidates.
of radii. They also suggested a new definition of core-galaxy We then examined two samples of nearby objects for which as the class of objects showing a light deficit toward the center radio observations combining relatively high resolution, high with respect to the Sérsic law (Trujillo et al. 2004).
frequency and sensitivity are available, in order to minimize In this context, we discuss in detail here the behaviour of the contribution from radio emission not related to the galaxy's the most critical objects, i.e. the two core galaxies for which nucleus and confusion from background sources. More specifi- the Nuker law returns the smallest values for the break radius, cally we focus on the samples of early-type galaxies studied by namely UGC 7760 and UGC 7797 for which rb = 0.49 and Wrobel & Heeschen (1991) and Sadler et al. (1989) both ob- rb = 0.21 respectively. We fit both objects with a Sérsic law.
served with the VLA at 5 GHz with a flux limit of ∼1 mJy. The The final fits, shown in Fig. 1, were obtained iteratively, fitting two samples were selected with a very similar strategy. Wrobel the external regions while flagging the innermost points out to (1991) extracted a northern sample of galaxies from the CfA a radius at which the residual from the Sérsic law exceeded redshift survey (Huchra et al. 1983) satisfying the following a threshold of 5%. The Sérsic law in general provides a re- markably good fit to the outer regions, with typical residuals of 1950 ≥ 0, (2) photometric magnitude B ≤ 14; (3) heliocentric velocity ≤3000 km s−1, and (4) morphologi- ∼1%, but a substantial central light deficit is clearly present in cal Hubble type T ≤ −1, for a total number of 216 galaxies.
both objects. This indicates that both objects can be classified Sadler et al. (1989) selected a similar southern sample of 116 E as core-galaxies in the Graham et al. scheme.
and S0 with −45 ≤ δ ≤ −32. The only difference between the Using the brightness profiles for the core-galaxies for two samples is that Sadler et al. did not impose a distance limit.
which we obtained Nuker fits in Paper I (14 additional objects) Nonetheless, the threshold in optical magnitude effectively lim- we obtained similar results. Very satisfactory fits can be ob- its the sample to a recession velocity of ∼6000 km s−1.
tained with a Sérsic law on the external regions of these galax- In Capetti & Balmaverde (2005, hereafter Paper I), we fo- ies, but they all show an even clearer central light deficit, as cused on the 116 galaxies detected in these VLA surveys to expected given the presence of well resolved shallow cores.
boost the fraction of AGN with respect to a purely optically We conclude that, for the galaxies of our sample, the objects selected sample. We used archival HST observations, available classified as core-galaxies in the Nuker scheme are recovered for 65 objects, to study their surface brightness profiles and as such with the Graham et al. definition.
B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Fig. 1. Sérsic fits for the two core galaxies of our sample with the smallest values for the core radius. A substantial central light deficit is clearly
present in both objects, conforming to the "core" classification in the Graham et al. scheme.
3. Basic data and nuclear luminosities
this is expected in the presence of a nuclear point source, sincethe convolution with the Point Spread Function produces a The basic data for the selected galaxies, namely the reces- smooth decrease of the slope toward the center when only a sion velocity (corrected for Local Group infall onto Virgo), diffuse galactic component is present. We therefore preferred the K band magnitude from the Two Micron All Sky Survey to adopt a "local" approach to identify nuclear sources, based (2MASS), the galactic extinction and the total and core radio on the characteristic up-turn they cause in the nuclear bright- fluxes were given in Paper I.
ness profile.
In the following three subsections, we derive and discuss More specifically, we evaluated the derivative of the bright- the measurements for the nuclear sources in the optical, X-ray ness profile in a log-log representation for the sources of our and radio bands.
sample. In order to increase the stability of the slope measure-ment this has been estimated by combining the brightness mea- 3.1. Optical nuclei sured over two adjacent points on each side of the radius of in-terest, yielding a second order accuracy. We then look for the The detection and measurement of an optical nuclear source at presence of a nuclear up-turn in the derivative requiring for a the center of a galaxy is a challenging task particularly when nuclear detection a difference larger than 3σ from the slopes at it represents only a small contribution with respect to the host the local minimum and maximum. This a rather conservative emission, as is likely often to be the case of the weakly active definition since the region over which the up-turn is detected galaxies making up our sample.
extends over several pixels while we only consider the peak- Different approaches have been employed in the literature.
The most widely used method is to fit the overall brightness To illustrate this we focus on three cases. In the HST image profile of a galaxy with an empirical functional form and to de- as well as in the brightness profile of IC 4296 a nucleus clearly fine a galaxy as "nucleated" when it shows a light excess in its stands out against the underlying background and the central central region with respect to the model (e.g. Lauer et al. 2004; steepening at about r = 0.15 is highly significant. NGC 4373 Ravindranath et al. 2001). The drawback of this "global" ap- is the detection with the least significance of our sample, in proach is that it assumes that the model can be extrapolated in- which the presence of a nucleus is uncertain from just the vi- wards from the radial domain over which the fit was performed.
sual inspection of the optical image, but the derivative of the Furthermore, the measurement and identification of the nuclear brightness profile reveals the effect of the point source with an component are coupled with the behaviour of the brightness increase of 0.013 ± 0.004 from r = 0.1 and r = 0.07. Instead profile at all radii and with the specific choice of an analytic in NGC 1316 we do not have evidence for any compact point form. Although this is not a significant issue for bright point source, both in the image and in the brightness profile deriva- sources, it is particularly worrisome for the faint nuclei we are tive, and it is considered as a non-detection.
dealing with. Nonetheless, Rest et al. (2001) pointed out that in Adopting this strategy in 18 out of 29 objects we identify general nuclear light excesses are associated with a steepening an optical nucleus, with a percentage of ∼60% of the total sam- of the profile as the HST resolution limit is approached. Indeed ple. In seven objects we did not find any upturn and these are B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Table 1. (1) Optical name (2) Chandra observational identification number, (3) exposure time [ks], (4) reference for the X-ray analysis (see
below for the list), (5) instrument and filter of the HST observation, (6) optical nuclear flux [erg cm−2 s−1].
Chandra data summary WFPC2/F814 W (1) This work, (2) Balmaverde & Capetti (2005), (3) Ho et al. (2001), (4) Loewenstein et al. (2001), (5) Filho et al. (2004), (6) Randall et al.
(2004), (7) Terashima & Wilson (2003), (8) Kim & Fabbiano (2003), (9) Satyapal et al. (2004), (10) Evans et al. (2004), (11) Fabbiano et al.
(2003), (12) Pellegrini et al. (2003).
considered upper limits. Note that this is again a conservative the up-turn and as the background region a circumnuclear an- approach, since a nuclear source can still be present but its in- nulus, 0.1 in width. For the undetected nuclei we set as upper tensity might not be sufficient to compensate the downward limits the light excess with respect to the starlight background trend of the derivative sets by the host galaxy.
within a circular aperture 0.1 in diameter. Then we use the In the remaining 4 objects the central regions have a com- PHOTFLAM and EXPTIME keyword in the image header to plex structure and no estimate of the optical nucleus intensity convert the total counts to fluxes. Errors on the measurements can be obtained. In two cases (UGC 8745 and UGC 9723) the of the optical nuclei are dominated by the uncertainty in the central regions are completely hidden by a kpc scale edge-on behaviour of the host's profile, while the statistical and abso- disk, while in NGC 3557 the study of its nuclear regions is lute calibration errors amount to less than 10%. The very pres- hampered by the presence of a circumnuclear dusty disk. In ence of the nucleus prevents us from determining accurately the UGC 9655, the innermost region (r < 0.1) has a lower bright- host contribution within the central aperture. Our strategy is to ness than its surrounding; since only a single band image is remove the background measured as close as possible to the available we cannot assess if this is due to dust absorption or to nucleus, i.e. effectively we adopted a constant starlight distri- a genuine central brightness minimum as in the cases discussed bution in the innermost regions. An alternative approach would by Lauer et al. (2002).
be to extrapolate the profile with a constant slope instead. Our We measured the nuclear luminosity with the task definition of nuclear sources (an increase in the profile's deriva- RADPROF in IRAF, choosing as the extraction region a cir- tive) implicitly requires that the observed profile lies above this cle centered on the nucleus with radius set at the location of extrapolation, but the resulting flux is reduced by at most a B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Table 2. Core galaxies data: (1) UGC name, (2) intrinsic nuclear X-ray luminosity (2–10 keV) [erg s−1], (3) nuclear optical luminosity (8140 Å)
[erg s−1] corrected for absorption using the galactic extinction values in Paper I, (4) nuclear radio luminosity (5GHz) [erg s−1] derived from
Paper I, (5) total radio luminosity (5GHz) [erg s−1] derived from Paper I, (6) Hα+[NII] line luminosity [erg s−1] from Ho et al. (1997) or
a Phillips et al. (1986), (7) total K band galaxy's absolute magnitude from 2MASS, (8) logarithm of black hole mass in solar unity from b
Marconi et al. (2003) or derived using the velocity dispertion.
Log ν Lcore Log ν Ltot Log LHα+[NII] Fig. 2. Brightness profile and its derivative for three objects of the sample, namely IC 4296, NGC 4373 and NGC 1316. The first two galaxies
show, at decreasing significance level, the characteristic up-turn in the profile associated with a nuclear source. This is not seen in the third
object which is then considered as a non-detection.
B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? factor of 2 (with respect to the case of constant background)for the nuclei with the smallest contrast against the galaxy. Aswill become clear in the next sections, errors of this magnitudeonly have a marginal impact on our conclusions. The resultingfluxes are reported in Table 1. We finally derived all the lumi-nosities referred to 8140 Å (see Table 2), after correcting forthe Galactic extinction as tabulated in Paper I and adopting anoptical spectral index1 αo = 1.
3.2. X-ray nuclei For the measurements of the X-ray nuclei we concentrate onlyon the Chandra measurements, as this telescope provides thebest combination of sensitivity and resolution necessary to de-tect the faint nuclei expected in these weakly active galaxies.
Data for 21 core galaxies are available in the Chandra publicarchive.
When available, we used the results of the analysis of the X-ray data from the literature. We find estimates of the lumi-nosities of the nuclear sources (usually defined as the detectionof a high energy power-law component) based on Chandra datafor 16 objects of our sample (12 of which are detections and 4 Fig. 3. Radio core flux density for CoreG obtained at 5 GHz with VLA
are upper limits) which we rescaled to our adopted distance and (used in this work) compared to higher resolution data from Slee et al.
(1994); Nagar et al. (2002); Filho et al. (2002); Krajnovi´c & Jaffe converted to the 2–10 keV band, using the published power law (2002); Jones & Wehrle (1997). The dotted line is the bisectrix of the index. In Table 1 we give a summary of the available Chandra data, while references and details on the X-ray observationsand analysis are presented in Appendix A.
We also considered the Chandra archival data for the 5 un- Sadler et al. (1989), performed with the VLA at 5 GHz with published objects, namely UGC 5902, UGC 6297, UGC 7203, a resolution of ∼5. Although these represent the most uni- NGC 3557 and NGC 5419. We analyzed these observations us- form and comprehensive studies of radio emission in early-type ing the Chandra data analysis CIAO v3.0.2, with the CALDB galaxies, they do not always have a resolution sufficient to sepa- version 2.25, using the same strategy as in Balmaverde & rate the core emission from any extended structure. Sadler et al.
Capetti (2005). We reprocessed all the data from level 1 to (1989) argued that at decreasing radio luminosity there is a cor- level 2, subtracting the bad pixels, applying ACIS CTI cor- responding increase of the fractional contribution of the radio rection, coordinates and pha randomization. We searched for background flares and excluded some period of bad aspect.
To verify whether the VLA data overestimate the core flux, We then extracted the spectrum in a circle region centered we searched the literature for radio core measurements ob- on the nucleus with a radius of 2 and we take the background tained at higher resolution (and/or higher frequency) than our in an annulus of 4. We grouped the spectrum to have at least data. This would improve the estimate of the core flux den- 10 counts per bin and applied Poisson statistics.
sity, avoiding the contribution of extended emission or spuri- For two objects (NGC 3557 and NGC 5419) we obtain a ous sources to the nuclear flux as well as revealing any radio detection of a nuclear power-law source by fitting the spectrum structure. Better measurements, from VLBI data or from higher using an absorbed power-law plus a thermal model, with the frequency/resolution VLA data, are available for most CoreG hydrogen column density fixed at the Galactic value. Details of (23 out of 29) and compact cores were detected in all but 2 ob- the results are given in Appendix A. For the remaining 3 galax- jects. The radio core fluxes are taken from Nagar et al. (2002) ies we set an upper limit to any nuclear emission, with the (15 GHz VLA data and 5 GHz VLBI data), Filho et al. (2002) conservative hypothesis that all flux that we measure is non- and Krajnovi´c & Jaffe (2002) (8.4 GHz at the VLA), Jones & thermal. We then fit the spectrum with an absorbed (to the Wehrle (1997) (8.4 GHz VLBI data) and Slee et al. (1994) (PTI galactic value) power law model with photon index Γ = 2.
5 GHz interpolated data).
The X-ray luminosities for all objects are given in Table 2.
In Fig. 3 we compare the radio core flux density used in our analysis against observations made at higher resolu-tion. Overall there is a substantial agreement between the two 3.3. Radio nuclei datasets, with a median difference of only ∼0.25 dex (a factor The radio data available for all objects of our sample are 1.6), with only two objects substantially offset (by a factor drawn from the surveys by Wrobel & Heeschen (1991) and of ∼10). However, since these data are highly inhomogeneousand given the general agreement with the 5 GHz VLA measure- 1 We define the spectral index α with the spectrum in the form ments, we prefer to retain the values of Wrobel & Heeschen and Fν ∝ ν−α.
Sadler et al. Nonetheless, we always checked that using these B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Fig. 4. Radio core luminosity for the early-type galaxies with a "core" profile versus the optical (left) and X-ray (right) nuclear luminosities.
higher resolution core fluxes our main results are not signifi- Table 3. Correlations summary.
cantly affected (see Appendix B for a specific example).
4. The multiwavelength properties of nuclei
of core galaxies
Having collected the multiwavelength information for the nu- clei of our core galaxies we can compare the emission in the different bands. First of all, we can estimate the ratio between the radio, optical and X-ray luminosities: the median values are Log (νLr/νLo) ∼ −1.5 (equivalent to a standard radio-loudnessparameter Log R ∼ 3.6)2 and Log RX = Log (νLr/LX) ∼ −1.3, the same behaviour of the stronger radio galaxies, extending it both with a dispersion of ∼0.5 dex. These ratios are clearly in- downward by 3 orders of magnitude in radio-core luminosity dicative of a radio-loud nature for these nuclei when compared as they reach levels as low as L to both the traditional separation into radio-loud and radio- r ∼ 1036 erg s−1.
We estimated the best linear fit for the combined quiet AGN (Log R = 1, e.g. Kellermann et al. 1994), as well CoreG/LLRG sample in both the L as with the radio-loudness threshold introduced by Terashima r vs. Lo and Lr vs. LX planes.
The best fits were derived as the bisectrix of the linear fits us- & Wilson (2003) based on the X-ray to radio luminosity ratio ing the two quantities as independent variables following the (Log RX = −4.5). Furthermore, the nuclear luminosities in all suggestion by Isobe et al. (1990) that this is preferable for three bands are clearly correlated (see Fig. 4 and Table 3 for problems needing symmetrical treatment of the variables. The a summary of the results of the statistical analysis): the gener- presence of upper limits in the independent variable suggests alized (including the presence of upper limits) Spearman rank that we could take advantage of the methods of survival anal- correlation coefficient ρ is 0.63 and 0.89 for Lr vs. Lo and Lr ysis proposed by e.g. Schmitt (1985). However, the drawbacks vs. LX respectively, with probabilities that the correlations are discussed by Sadler et al. (1989) and, in our specific case, the not present of only 0.002 and 0.0001.
non-random distribution of upper limits, argue against this ap- Both results are reminiscent of what is observed for the proach. We therefore preferred to exclude upper limits from the radio-loud nuclei of low luminosity radio-galaxies (LLRG).
linear regression analysis. Nonetheless, a posteriori, 1) the ob- Chiaberge et al. (1999) and Balmaverde & Capetti (2005) re- jects with an undetected nuclear component in the optical or ported on similar multiwavelength luminosity trends for the X-ray are consistent with the correlation defined by the detec- sample of LLRG formed by the 3C sources with FR I mor- tions only; 2) the application of the Schmidt methods provides phology. The connection between the CoreG and LLRG be- correlation slopes that agree, within the errors, with our esti- comes more evident if we add LLRG in the diagnostic planes (see Fig. 5 and Table 4). The early-type core galaxies follow We obtained (indicating the Pearson correlation coefficient 2 R = L with r and slope with m) r 5 GHz/LB. As in Sect. 3 we transformed the optical fluxes to ro = 0.90 and mro = 0.89 ± 0.07, rrx = the B band adopting an optical spectral index αo = 1.
0.89 mrx = 1.02 ± 0.10 for the radio/optical and radio/X-ray B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Fig. 5. Comparison of radio and optical (left) and X-ray (right) nuclear luminosity for the sample of core-galaxies (filled circles) and for the
reference 3C/FR I sample of low luminosity radio-galaxies (empty circles). The three sources in common are marked with a filled square. The
solid lines reproduce the best linear fits.
correlations respectively. The slopes and normalizations their other properties, such as the structure of the host, black derived for CoreG, LLRG and the combined CoreG+LLRG hole mass, radio-morphology and optical spectra.
sample (see Table 3) are consistent within the errors and this Our sample was selected to include only early-type galax- indicates that there is no significant change in the behaviour ies with a core profile, e.g. with an asymptotic slope (toward the between the two samples. Only the dispersion is slightly larger nucleus) of their surface brightness profiles γ < 0.3. Recently for the CoreG nuclei being a factor of ∼4 rather than ∼2 for the de Ruiter et al. (2005) showed, from the analysis of a com- LLRG sample alone.
bined sample of B2 and 3C sources, that they are all hosted Chiaberge et al. (1999) first reported the presence of a by early-type galaxies and that the presence of a flat core is a correlation between radio and optical emission in the LLRG characteristic of the host galaxies of all nearby radio-galaxies.
and they concluded that this is most likely due to a com- A strong similarity between CoreG and LLRG emerges mon non-thermal jet origin for the radio and optical cores.
when comparing the mass of their supermassive black holes.
Recently Balmaverde & Capetti (2005) extended the analysis When no direct measurement (taken from the compilation by to the X-ray cores; the nuclear X-ray luminosity also corre- Marconi & Hunt 2003) was available, we estimated MBH us- lates with those of the radio cores and with a much smaller dis- ing the relationship with the stellar velocity dispersion (taken persion (∼0.3 dex) when compared to similar trends found for from the LEDA database) in the form given by Tremaine et al.
other classes of AGN (see e.g. Falcke et al. 1995), again point- (2002). The distributions of MBH (see Fig. 6) of the two sam- ing to a common origin for the emission in the three bands.
ples are almost indistinguishable3, as they have median values Furthermore, the broad band spectral indices of the 3C/FR I of Log MBH = 8.54 and Log MBH = 8.70, for CoreG and cores are very similar to those measured in BL Lacs objects LLRG respectively, and they also cover the same range, with (for which a jet origin is well established) in accord with the most objects with Log MBH = 8−9.5.
FR I/BL Lacs unified model (we will return to this issue in Further indications of the nature of CoreG cores and their connection with LLRG come from the emission lines in The core galaxies of our sample thus appear to smoothly their optical spectra. LLRG are characterized as a class by extend the results obtained for LLRG to much lower radio lu- their LINER spectra (e.g. Lewis et al. 2003) and this is the minosity, expanding the multiwavelenght nuclear correlations case also for the CoreG of our sample. In the NED database, to a total of 6 orders of magnitude. This strongly argues in although about half of the CoreG do not have a spectral clas- favour of a jet origin for the nuclear emission also in the core sification, 13 objects are classified as LINERs4. The only galaxies and that they simply represent the scaled down ver-sions of these already low luminosity AGN.
3 The probability that the two samples are drawn from the same parent distribution is 0.32, according to the Kolmogorov-Smirnoff test.
4.1. Core galaxies vs. low luminosity radio-galaxies 4 This result provides further support to the suggestion by Chiaberge et al. (2005) that a dual population is associated with galax- The results presented above indicate that the nuclei of the ies with a LINER spectrum, being formed by both radio-quiet and by CoreG show a very similar behaviour to those of LLRG. Here radio-loud objects. The CoreG are part of this latter sub-population of we explore in more detail how CoreG and LLRG compare in radio-loud LINER.
B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Table 4. Radio galaxies of the 3C/FR I sample data: (1) name, (2) intrinsic nuclear X-ray luminosity (2–10 keV) [erg s−1], (3) nuclear optical
luminosity (8140 Å) [erg s−1], (4) nuclear radio luminosity (5 GHz) [erg s−1] and (5) total radio luminosity (178MHz) [erg s−1] from Chiaberge
et al. 1999, (6) Hα+[NII] line luminosity from Capetti et al. 2005 [erg s−1], (7) total K band galaxy's absolute magnitude from 2MASS, (8)
logarithm of black hole mass in solar unity derived using the velocity dispertion or from a Marconi et al. (2003).
Log ν Lcore Log ν Ltot Log LHα+[NII] exception is UGC 7203, with a Seyfert spectrum, but its diag- and 3C 274), while in the Southern sample we have the well nostic line ratios are borderline with those of LINERs (Ho et al.
studied radio-galaxies NGC 1316 (Fornax A), a FR II source, 1997). Concerning the emission line luminosity, Capetti et al.
NGC 5128 (Cen A) and IC 4296. A literature search shows that (2005) found a tight relationship between radio core and line at least another 11 sources have extended radio-structure in- luminosity studying a group of LLRG formed by the 3C/FR I dicative of a collimated outflow, although in several cases this complemented by the sample of 21 radio-bright (Fr > 150 can only be seen in high resolution VLBI images, such as the mJy) UGC galaxies defined by Noel-Storr et al. (2003). Line mas scale double-lobes in UGC 7760 or the one-sided jet of luminosity for our CoreG clearly follow the same trend defined UGC 7386 (Nagar et al. 2002; Falcke et al. 2000).
by LLRG, although with a substantially larger dispersion, not Conversely, hosts of 3C/FR I radio-sources are on average unexpected given their low line luminosity and the non unifor- more luminous than core-galaxies (see Fig. 6, left panel) al- mity of the data used for this analysis.
though there is a substantial overlap between the two groups: Considering the radio structure, several objects of our the median values are MK = −24.8 and MK = −25.7 for CoreG CoreG sample have a radio-morphology with well developed and 3C/FR I respectively, with a KS probability of only 0.003 jets and lobes: UGC 7360, UGC 7494 and UGC 7654 are of being drawn from the same population. This reflects the FR I radio-galaxies part of the 3C sample (3C 270, 3C 272.1 well known trend, already noted by Auriemma et al. (1977), B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Fig. 6. Distributions for CoreG (shaded histograms) and for LLRG (empty histogram) of (left panel) black hole mass MBH and (right panel)
absolute magnitude MK . The LLRG histograms have been re-normalized multiplying by a factor 29/19 for MBH and 29/17 for MK respectively,
i.e. the number of objects in the two samples for which estimates of these parameters are available.
for which a brighter galaxy has a higher probability of beinga stronger radio emitter, and which is present also in our sam-ple (Paper I). The selection of relatively bright radio sources,such as the 3C/FR I, corresponds to a bias toward more lumi-nous galaxies. Indeed, within our sample, imposing a thresh-old in total radio-luminosity of Ltot > 1039 erg/s,5 the low endfor LLRG, decreases the median magnitude to –25.1, in closeragreement with the 3C/FR I value.
We conclude that the properties of our low radio luminos- ity CoreG show a remarkable similarity to those of classicalLLRG, in particular, they share the presence of a flat core intheir host's brightness profiles, they have the same distributionin black hole masses, as well as analogous properties concern-ing their optical emission lines and radio-morphology. Theseresults indicate that core galaxies and LLRG can be consid-ered, from these different point of view, as being drawn fromthe same population of early-type galaxies. They can only beseparated on the basis of their different level of nuclear activ-ity, with the LLRG forming the tip of the iceberg of (relatively)high luminosity objects. Furthermore, the emission processesassociated to their activity scale almost linearly over 6 orders Fig. 7. Emission line vs. radio core luminosity for CoreG galaxies
of magnitude in all bands for which data are available.
(filled circles) and for the LLRG 3C/FR I sample (empty circles), fromCapetti et al. (2005).
5. Black hole mass and radio luminosity
The issue of the relationship between the black hole mass andthe radio-luminosity has been discussed by several authors, tak- hole estimates, that the radio-luminosity tightly correlates with ing advantage of the recent possibility to measure (or at least the black hole mass, with a logarithmic index of ∼2.5. This re- estimate) M sult was subsequently challenged, by e.g. Ho (2002). We here BH. Franceschini et al. (1998) pioneered this field showing, from a compilation of objects with available black re-explore this issue limiting ourselves to the sample of coreearly-type galaxies; while this substantially restricts the acces- 5 The 5 GHz luminosity was converted to 178 MHz for consistency sible range in MBH and it applies only to radio-loud nuclei, it with the 3C/FR I values adopting a spectral index of 0.7.
has the substantial advantage of performing the analysis on a B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? sample) we do include galaxies with expected black holemasses MBH < 108 M. This represents a severe bias againstthe inclusion of galaxies with low values of MBH, regardless oftheir radio emission. The lack of low black hole mass LLRGseems to favour the reality of this effect, as they are not directlyselected imposing an optical threshold; however, the alreadydiscussed statistical trend linking radio and optical luminos-ity might represent a more subtle bias leading to the same ef-fect. The existence of a minimum black hole mass to produce aradio-loud nucleus must be properly tested extending the anal-ysis to a sample of less luminous galaxies, likely to harbourless massive black holes.
6. Constraints on the radiative manifestation
of the accretion process
Taking advantage of the estimates of black hole mass we canconvert the measurements of the nuclear luminosities to unitsof the Eddington luminosity. All CoreG nuclei are associ-ated with a low fraction of L Fig. 8. Nuclear radio-luminosity vs. black hole mass M
Edd, being confined to the range galaxies (filled circles) and for the LLRG 3C/FR I sample (empty L/LEdd ∼ 10−6−10−9 in both the X-ray and optical bands (with only one X-ray exception), see Fig. 9. Furthermore, as dis-cussed in Sect. 4, the tight correlations between radio, opti-cal and X-ray nuclear luminosities extending across LLRG and complete sample with well defined selection criteria and cov- CoreG strongly argue in favour of a jet origin for the nuclear ering a large range of radio-luminosity.
emission also in the core galaxies. If this is indeed the case, the The comparison of the radio-core luminosity with the black observed nuclear emission does not originate in the accretion hole mass is presented in Fig. 8. Apparently, a dependence of Lr process and the values reported above should be considered as on MBH is present, although with a substantial scatter. However, upper limits.
the radio flux limit of the samples exclude objects with lower Our results add to the already vast literature reporting emis- radio luminosity, potentially populating the lower part of the sion corresponding to a very low Eddington fraction associ- LR vs. MBH plane. Furthermore, the inclusion of LLRG (which, ated with accretion onto supermassive black holes. These re- as discussed above, represent the high activity end of the early- sults prompted the idea that in these objects accretion occurs type population) radically changes the picture, as they populate not only at a low rate but also at a low radiative efficiency, the whole upper portion of this plane. This indicates that a very such as in the Advection Dominated Accretion Flows (ADAF, large range (at least 4 orders of magnitude) of radio-power can Narayan & Yi 1995) in which most of the gravitational energy correspond to a given M BH . This is a clear indication that, not of the accreting gas is advected into the black hole before it unexpectedly, parameters other than the black hole mass play can be dissipated radiatively, thus reducing the efficiency of a fundamental role in determining the radio luminosity of a the process with respect to the standard models of geometri- cally thin, optically thick, accretion disks. The ADAF mod- More notable is the lack of sources with MBH < 108 M els have been rather successful in modeling the observed nu- (with only one exception). The effects produced by our se- clear spectrum in several galaxies, such as e.g. the Galactic lection criteria must be considered before any conclusion can Center and NGC 4258 (Narayan & Yi 1995; Lasota et al. 1996).
be drawn. In particular the correlation between the black hole Conversely, ADAF models substantially over-predict the ob- mass and the spheroidal galactic component, combined with served emission in the nuclei of nearby bright elliptical galax- the limiting magnitude, translates into a threshold in the ac- ies (Di Matteo et al. 2000; Loewenstein et al. 2001).
cessible range of black hole masses. Using the limit in ap- This suggested the possibility that a substantial fraction of parent magnitude of our sample (mB < 14), an average color the mass included within the Bondi's accretion radius (Bondi of B − K = 4.25 (Mannucci et al. 2001) and the best fit 1952) might not actually reach the central object, thus fur- to the relationship between MBH and MK from Marconi & ther reducing the radiative emission from the accretion pro- Hunt (2003) we obtain that at distances larger than 20 Mpc cess with respect to the ADAF models. This may be the (corresponding to 7/8 of the volume covered in the Wrobel's case in the presence of an outflow (Advection Dominated Inflow/Outflow Solutions, or ADIOS, Blandford & Begelman With respect to previous studies we report the nuclear radio emis- sion only, instead of the total radio luminosity. However, since the 1999) or strong convection (Convection Dominated Accretion fraction of extended emission grows with radio luminosity, using the Flows, or CDAF, Quataert & Gruzinov 2000) in which most total power would just move the LLRG upward, further increasing gas circulates in convection eddies rather than accreting onto the black hole.
B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Fig. 9. Distributions of the nuclear luminosities measured as fraction of the Eddington luminosity in the X-ray (left) and optical (right) bands.
Unfortunately, in the case of the galaxies under investiga- luminosity extension of LLRG. The CoreG nuclei appear to be tion, the comparison of the theoretical predictions with the ob- the scaled down versions of those of LLRG when their multi- servations so as to get constraints on the properties of the ac- wavelength nuclear properties are considered. Thus here we are cretion process is quite difficult. This is due to the presence sampling a new regime for radio-galaxies in terms of nuclear of different competing models, all of these with several free power and it is important to explore the implications of this re- parameters, and to the observational data, in particular to the sult for the model unifying BL Lac objects and radiogalaxies.
scarce multiwavelength coverage of the nuclear emission mea- Unification models ascribe the differences between the ob- surements which prevents us from deriving a detailed Spectral served properties of different classes of AGN to the anisotropy Energy Distribution of these objects. As discussed above, this of the nuclear radiation (see e.g. Antonucci 1993; Urry & is more complicated for our radio-loud nuclei in which the Padovani 1995, for reviews). In particular, for low luminos- emission is most likely dominated by the non-thermal radia- ity radio-loud objects, it is believed that BL Lac objects are tion from their jets.
the pole-on counterparts of radio-galaxies, i.e. their emission Nonetheless, Pellegrini (2005) recently studied in detail a is dominated by the radiation from the inner regions of a rel- sample of nearby galaxies for which the Chandra observations ativistic jet seen at a small angle from its axis which is thus provide an estimate of the temperature and density of the gas strongly amplified by relativistic Doppler beaming. In FR I, in the nuclear regions, thus enabling one to derive the expected whose jets are observed at larger angles with respect to the Bondi accretion rate, ˙ MB. It is interesting to note that the esti- line of sight, the nuclear component is strongly de-amplified.
MB for three sources common to both samples with Contrary to other classes of AGN there is growing evidence the lowest X-ray luminosity (namely NGC 1399, UGC 7629 that obscuration does not play a significant role in these objects (AKA NGC 4472) and UGC 7898 (AKA NGC 4649)) are rel- (Henkel et al. 1998; Chiaberge et al. 1999; Donato et al. 2004; atively large, ˙ MEdd = 10−2−10−4, while their X-ray lumi- Balmaverde & Capetti 2005).
nosities are LX/LEdd = 10−8−10−10 (see her Fig. 3). These lu- Balmaverde & Capetti (2005) found that there is a close minosities are between 3 and 5 orders of magnitude lower than similarity of the broad band spectral indices between LLRG expected from an ADAF model, and they should be considered and the sub-class of the BL Lacs, the Low energy peaked only as upper limits. These results argue in favour of an ef- Padovani & Giommi 1995), in agreement fective accretion rate substantially smaller than expected in the with the unified model7. We performed the same comparison8 case of spherical accretion, suggesting that an important role isplayed by mass loss due to an outflow or by convection.
7 The small offsets between the two classes can be quantitatively accounted for by the effects of beaming since Doppler beaming notonly affects the angular pattern of the jet emission, but it also causes 7. CoreG and the BL Lacs/LLRG unifying model
a shift in frequency of the spectral energy distribution (see Chiabergeet al. 2000; Trussoni et al. 2003).
In Sect. 4.1 we presented evidence that "core" galaxies and 8 We used the standard definition of spectral indices, measured be- LLRG are drawn from the same population of early-type galax- tween 5 GHz, 5500 Å and 1 keV. Optical fluxes have been converted ies. They can only be separated on the basis of their differ- from 8140 Å to 5500 Å using a local slope of α = 1; 1 keV fluxes are ent level of nuclear activity, with CoreG representing the low directly derived from the spectral fit.
B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? available radio-maps do not allow us to separate core and ex-tended radio-emission and Lext must be considered as an upperlimit. This suggests that the CoreG represent the counterpartsof the large low luminosity population of BL Lac of LBL typewhich is now emerging from the low radio flux limit surveyssuch as the DXRBS (Landt et al. 2001). Clearly, this still re-quires measurements of the extended radio-luminosity of theselow power BL Lac. A ramification of this possible extension ofthe unified model toward lower luminosities would be the pres-ence of relativistic jets also in our sample of quasi-quiescentearly-type galaxies, as this is a prerequisite to produce a sub-stantial dependence of the luminosity on the viewing angle.
We did not find any CoreG with spectral properties simi- lar to those of the High energy peaked BL Lac (HBL), eventhough HBL have extended radio-emission values of Lext simi-lar to CoreG. The spectral indices of CoreG imply a differencein both the radio-to-optical and radio-to-X-ray flux ratios ofan average factor of ∼100 with respect to HBL. The same re-sult applies to LLRG, as all have a LBL-type SED, with theonly exception of 3C 264 (Capetti et al. 2000). Our optical se- Fig. 10. Broad band spectral indices, calculated between 5 GHz,
lection criteria did not exclude the parent population of HBL 5500 Å and 1 keV, for core galaxies (filled circles), low luminosity since their host galaxies are early-type sufficiently luminous 3C/FR I radio-galaxies (empty circles), Low energy peaked BL Lacs (stars) and High energy peaked BL Lacs (squares). Solid lines mark R < −22.5, Scarpa et al. 2000) to be included in our sam- the regions within 2σ from the mean α ple. Most likely, the dearth of HBL-like CoreG is induced by ro and αox for BL Lacs drawn from the DXRB and RGB surveys. The dashed lines represent con- the radio threshold. Purely radio selected samples of BL Lacs stant values for the third index, α are known to strongly favour the inclusion of LBL; e.g. in the 1 Jy sample there are only 2 HBL out of 34 objects (Giommi& Padovani 1994).
including CoreG, see Fig. 10. We considered the radio selectedBL Lacs sample derived from the 1Jy catalog (Stickel et al.
8. Summary and conclusions
1991) and the BL Lac sample selected from the Einstein Slewsurvey (Elvis et al. 1992; Perlman et al. 1996). We used the The aim of this series of papers is to explore the classical issue classification into High and Low energy peaked BL Lacs (HBL of the connection between host galaxies and AGN, in the new and LBL respectively), as well as their multiwavelength data light shed by the recent developments in our understanding of given by Fossati et al. (1998). We also report the regions (solid the nuclear regions of nearby galaxies.
lines) of the plane within 2σ from the mean αro and αox for We thus selected a samples of nearby early-type galax- the BL Lacs drawn from the Deep X-Ray Radio Blazar Survey ies comprising 332 objects. We performed an initial selection (DXRBS) and the ROSAT All-Sky Survey-Green Bank Survey of AGN candidates requiring a radio detection above ∼1 mJy (RGB) (Padovani et al. 2003).
leading to a sub-sample of 112 sources. Archival HST images Core galaxies are found to be located in the same region enabled us to classify 51 of them into core and power-law covered by LLRG. This is not surprising since they extend the galaxies on the basis of their nuclear brightness profile. We here behaviour of LLRG in the radio/optical and radio/X-ray planes, focused on the 29 core galaxies.
following Log–Log linear correlations whose slope is close to We used HST and Chandra archival data to isolate their nu- unity, implying only a small dependence of spectral indices on clear emission in the optical and X-ray bands, thus enabling us luminosity. More importantly, they populate the same area in (once combined with the radio data) to study the multiwave- which LBL are found.
length behaviour of their nuclei. The detection rate of nuclear We also compared the spectral indices of the different sources is 18/29 in the optical (62%, increasing to 72% if the groups taking into account the extended radio-luminosity Lext sources affected by large scale dust are not considered) and (see Fig. 11) which does not depend on orientation. This en- 14 in the X-ray, out of the 21 objects with available Chandra ables us to properly relate objects from the same region of data (67%). Our selection criteria required a radio detection in the luminosity function of the parent population. Indeed, the order to select AGN candidates; 26 CoreG are confirmed as strongest evidence in favour of the FR I/BL Lac unifying model genuine active galaxies based on the presence of i) an optical comes from the similarity in the power and morphology of (or X-ray) core; ii) a AGN-like optical spectrum, or iii) radio- the extended radio emission of BL Lacs and FR I (see e.g.
jets, with only 3 exceptions, namely UGC 968, UGC 7898 and Antonucci & Ulvestad 1985; Kollgaard et al. 1992; Murphy et al. 1993).
The most important result of this analysis is that "core" The CoreG reach radio-luminosities ∼100 smaller than galaxies invariably host a radio-loud nucleus. The radio- in LLRG and the 1 Jy LBL. In addition, in 13 CoreG the loudness parameter R for the nuclei in these sources is on B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Fig. 11. Broad band spectral indices vs. extended radio luminosity for core galaxies (filled circles), LLRG (empty circles), LBL (stars) and
HBL (squares). The CoreG luminosity has been extrapolated to 178 MHz adopting a spectral index of 0.7.
average Log R ∼ 3.6, a factor of 400 above the classical thresh- same multiwavelength characteristics despite covering a range old between radio-loud and radio-quiet nuclei. The X-ray data of 6 orders of magnitude in luminosity. Thus LLRG represent provide a completely independent view of their multiwave- the tip of the iceberg of (relatively) high luminosity objects.
length behaviour leading to the same result, i.e. a large X-ray It is unclear what mechanism is driving the level of nuclear deficit, at the same radio luminosity, when compared to radio- activity. As noted above, there is a marginal difference (less quiet nuclei.
than 1 mag) in the host galaxies of CoreG and LLRG; this re- multiwavelength nuclear diagnostic flects the well known (but as yet unexplained) trend for which planes, we found that optical and X-ray nuclear luminosities a brighter galaxy has a higher probability of being a stronger are correlated with the radio-core power, reminiscent of the radio emitter. As described in Paper I, this effect is present also behaviour of low luminosity radio-galaxies. The inclusion of within our sample of CoreG but it cannot be simply described CoreG indeed extends the correlations reported for LLRG as a correlation between Lr and MK.
toward much lower power, by a factor of ∼1000.
We explored if there is a relationship between the black The available radio maps show that in 17 CoreG the ex- hole mass and the radio-luminosity. Again, a very large range tended radio morphology is clearly indicative of a collimated of radio-power corresponds to a given MBH. We do not find any outflow, in the form of either double-lobed structures or jets, relationship between radio-power and black hole mass, clearly although in several cases this can only be seen in high resolu- indicating that parameters other than the black hole mass play tion VLBI images. This finding, combined with the analogy of a fundamental role in determining the radio luminosity of a the nuclear properties, leads us to the conclusion that minia- galaxy. No sources with MBH < 108 M are found. However, ture radio-galaxies are associated with all core galaxies of our this might be due to a bias induced by the sample's selection criteria. The limit in optical magnitude translates into a thresh- The similarity between CoreG and classical low luminos- old of accessible black hole masses. Only by extending this ity radio-galaxies extends to other properties. Recent results study to a sample of less luminous galaxies (harbouring, on av- show that LLRG are always hosted by early-type galaxies with erage, smaller black holes) will it be possible to test the reality a shallow cusp in their nuclear profile, and this is the case, by of a minimum black hole mass to produce a radio-loud nucleus.
definition, for our CoreG. While the distributions of black hole Our data can also be used to set constraints on the radiative masses, MBH, of the two classes are indistinguishable, hosts manifestation of the accretion process. The nuclear luminosi- of 3C/FR I radio-sources are on average slightly more lumi- ties of CoreG correspond, in units of the Eddington luminosity, nous than CoreG but there is a substantial overlap between the to the range L/LEdd ∼ 10−6−10−9 in both the optical and X-ray two groups. CoreG and LLRG also share similar properties bands. In analogy with the scenario proposed for LLRG, the from the point of view of their emission lines, as all sources available data support a common jet origin for the nuclear emis- with available data conform to the definition of a LINER on sion in these observing bands also for CoreG. Thus, the above the basis of the optical line ratios and they follow a common values should be considered as upper limits to the radiative dependence of line luminosity with radio core power. CoreG manifestation of the accretion process, suggesting that accre- and LLRG thus appear to be drawn from the same population tion occurs both at a low accretion level and at a low efficiency.
of early-type "core" galaxies. They host active nuclei with the It is difficult to derive from these results clear constraints on B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? the properties of the accretion flow. In part this is due to the Chiaberge, M., Capetti, A., & Celotti, A. 1999, A&A, 349, 77 limited information on the Spectral Energy Distribution of the Chiaberge, M., Capetti, A., & Macchetto, F. D. 2005, ApJ, 625, 716 CoreG nuclei and by the fact that in these radio-loud nuclei Chiaberge, M., Celotti, A., Capetti, A., & Ghisellini, G. 2000, A&A, the observed emission is most likely dominated by radiation from their jets rather than from the accretion. This is further de Ruiter, H. R., Parma, P., Capetti, A., et al. 2005, A&A, 439, 487Di Matteo, T., Quataert, E., Allen, S. W., Narayan, R., & Fabian, A. C.
complicated by the presence of several competing accretion 2000, MNRAS, 311, 507 models whose predictions of the emitted spectra depend on Donato, D., Sambruna, R. M., & Gliozzi, M. 2004, ApJ, 617, 915 parameters that are not well constrained by the observations.
Dunlop, J. S., McLure, R. J., Kukula, M. J., et al. 2003, MNRAS, 340, Nonetheless, in the galaxies with the least luminous nuclei, the estimates of the accretion rate from the literature (derived for Elvis, M., Plummer, D., Schachter, J., & Fabbiano, G. 1992, ApJS, the case of spherical accretion), combined with the very low level of X-ray emission, suggest that an important role is played Evans, D. A., Kraft, R. P., Worrall, D. M., et al. 2004, ApJ, 612, 786 by outflows (or by convection) in order to substantially sup- Fabbiano, G., Elvis, M., Markoff, S., et al. 2003, ApJ, 588, 175 press the amount of gas actually reaching the central object.
Faber, S. M., Tremaine, S., Ajhar, E. A., et al. 1997, AJ, 114, 1771 As reported above, the CoreG can be effectively considered Falcke, H., Malkan, M. A., & Biermann, P. L. 1995, A&A, 298, 375 as miniature radio-galaxies, in terms of nuclear luminosity, thus Falcke, H., Nagar, N. M., Wilson, A. S., & Ulvestad, J. S. 2000, ApJ, we are sampling a new region in terms of luminosity for radio- Ferrarese, L., & Merritt, D. 2000, ApJ, 539, L9 loud AGN. It is interesting to explore the implications of this Filho, M. E., Barthel, P. D., & Ho, L. C. 2002, ApJS, 142, 223 result also for the model unifying BL Lac objects and radio- Filho, M. E., Fraternali, F., Markoff, S., et al. 2004, A&A, 418, 429 galaxies. The broad band spectral indices of CoreG present a Fossati, G., Maraschi, L., Celotti, A., Comastri, A., & Ghisellini, G.
very close similarity to those of Low Energy Peaked BL Lac, 1998, MNRAS, 299, 433 suggesting the extension of the unified models to these lower Franceschini, A., Vercellone, S., & Fabian, A. C. 1998, MNRAS, 297, luminosities. The CoreG might represent the mis-aligned coun- terpart of the large population of low luminosity BL Lac emerg- Gebhardt, K., Bender, R., Bower, G., et al. 2000, ApJ, 539, L13 ing from the recent surveys at low radio flux limits. Clearly, Giommi, P., & Padovani, P. 1994, MNRAS, 268, L51 a more detailed comparison, taking into account e.g. the (as Graham, A. W., Erwin, P., Trujillo, I., & Asensio Ramos, A. 2003, AJ, yet not available) information on the extended radio power and Graham, A. W., & Guzmán, R. 2003, AJ, 125, 2936 morphology, is needed before this result can be confirmed. An Henkel, C., Wang, Y. P., Falcke, H., Wilson, A. S., & Braatz, J. A.
important ramification of this possible extension of the unify- 1998, A&A, 335, 463 ing model toward lower luminosities would be the presence of Ho, L. C. 2002, ApJ, 564, 120 relativistic jets, the essential ingredient of this model, also in Ho, L. C., Filippenko, A. V., & Sargent, W. L. W. 1997, ApJS, 112, our quasi-quiescent early-type galaxies.
In the third paper of the series we will explore the properties Ho, L. C., Feigelson, E. D., Townsley, L. K., et al. 2001, ApJ, 549, of the AGN hosted by galaxies with a power-law brightness Huchra, J., Davis, M., Latham, D., & Tonry, J. 1983, ApJS, 52, 89Isobe, T., Feigelson, E. D., Akritas, M. G., & Babu, G. J. 1990, ApJ, Acknowledgements. This work was partly supported by the Italian MIUR under grant Cofin 2003/2003027534_002. This research has Jones, D. L., & Wehrle, A. E. 1997, ApJ, 484, 186 made use of the NASA/IPAC Extragalactic Database (NED) (which Kellermann, K. I., Sramek, R. A., Schmidt, M., Green, R. F., & is operated by the Jet Propulsion Laboratory, California Institute of Shaffer, D. B. 1994, AJ, 108, 1163 Technology, under contract with the National Aeronautics and Space Kim, D., & Fabbiano, G. 2003, ApJ, 586, 826 Administration), of the NASA/ IPAC Infrared Science Archive (which Kollgaard, R. I., Wardle, J. F. C., Roberts, D. H., & Gabuzda, D. C.
is operated by the Jet Propulsion Laboratory, California Institute of 1992, AJ, 104, 1687 Technology, under contract with the National Aeronautics and Space Kormendy, J., & Bender, R. 1996, ApJ, 464, L119 Administration) and of the LEDA database.
Kormendy, J., & Richstone, D. 1995, ARA&A, 33, 581Kraft, R. P., Forman, W., Jones, C., et al. 2000, ApJ, 531, L9Krajnovi´c, D., & Jaffe, W. 2002, A&A, 390, 423 Landt, H., Padovani, P., Perlman, E. S., et al. 2001, MNRAS, 323, 757 Antonucci, R. 1993, ARA&A, 31, 473 Lasota, J.-P., Abramowicz, M. A., Chen, X., et al. 1996, ApJ, 462, 142 Antonucci, R. R. J., & Ulvestad, J. S. 1985, ApJ, 294, 158 Lauer, T. R., Ajhar, E. A., Byun, Y.-I., et al. 1995, AJ, 110, 2622 Auriemma, C., Perola, G. C., Ekers, R. D., et al. 1977, A&A, 57, 41 Lauer, T. R., Gebhardt, K., Richstone, D., et al. 2002, AJ, 124, 1975 Balmaverde, B., & Capetti, A. 2005, A&A, submitted Lauer, T. R., Faber, S. M., Gebhardt, K., et al. 2004, ArXiv Biller, B. A., Jones, C., Forman, W. R., Kraft, R., & Ensslin, T. 2004, Astrophysics e-prints Lewis, K. T., Eracleous, M., & Sambruna, R. M. 2003, ApJ, 593, 115 Blandford, R. D., & Begelman, M. C. 1999, MNRAS, 303, L1 Loewenstein, M., Mushotzky, R. F., Angelini, L., Arnaud, K. A., & Bondi, H. 1952, MNRAS, 112, 195 Quataert, E. 2001, ApJ, 555, L21 Capetti, A., & Balmaverde, B. 2005, A&A, 440, 73 Maccarone, T. J., Kundu, A., & Zepf, S. E. 2003, ApJ, 586, 814 Capetti, A., Kleijn, G. V., & Chiaberge, M. 2005, A&A, 439, 935 Mannucci, F., Basile, F., Poggianti, B. M., et al. 2001, MNRAS, 326, Capetti, A., Trussoni, E., Celotti, A., Feretti, L., & Chiaberge, M.
2000, MNRAS, 318, 493 Marconi, A., & Hunt, L. K. 2003, ApJ, 589, L21 B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy? Murphy, D. W., Browne, I. W. A., & Perley, R. A. 1993, MNRAS, Scarpa, R., Urry, C. M., Falomo, R., Pesce, J. E., & Treves, A. 2000, Nagar, N. M., Falcke, H., Wilson, A. S., & Ulvestad, J. S. 2002, A&A, Schmitt, J. H. M. M. 1985, ApJ, 293, 178 Sérsic, J.-L. 1968, Atlas de Galaxias Australes (Córdoba: Obs.
Narayan, R., & Yi, I. 1995, ApJ, 444, 231 Noel-Storr, J., Baum, S. A., Verdoes Kleijn, G., et al. 2003, ApJS, 148, Slee, O. B., Sadler, E. M., Reynolds, J. E., & Ekers, R. D. 1994, Padovani, P., & Giommi, P. 1995, ApJ, 444, 567 Soldatenkov, D. A., Vikhlinin, A. A., & Pavlinsky, M. N. 2003, Padovani, P., Perlman, E. S., Landt, H., Giommi, P., & Perri, M. 2003, Astron. Lett., 29, 298 Stickel, M., Fried, J. W., Kuehr, H., Padovani, P., & Urry, C. M. 1991, Pellegrini, S. 2005, ApJ, 624, 155 Pellegrini, S., Venturi, T., Comastri, A., et al. 2003, ApJ, 585, 677 Terashima, Y., & Wilson, A. S. 2003, ApJ, 583, 145 Perlman, E. S., Stocke, J. T., Schachter, J. F., et al. 1996, ApJS, 104, Tremaine, S., Gebhardt, K., Bender, R., et al. 2002, ApJ, 574, 740 Trinchieri, G., & Goudfrooij, P. 2002, A&A, 386, 472 Quataert, E., & Gruzinov, A. 2000, ApJ, 539, 809 Trujillo, I., Erwin, P., Asensio Ramos, A., & Graham, A. W. 2004, AJ, Randall, S. W., Sarazin, C. L., & Irwin, J. A. 2004, ApJ, 600, 729 Ravindranath, S., Ho, L. C., Peng, C. Y., Filippenko, A. V., & Sargent, Trussoni, E., Capetti, A., Celotti, A., Chiaberge, M., & Feretti, L.
W. L. W. 2001, AJ, 122, 653 2003, A&A, 403, 889 Rest, A., van den Bosch, F. C., Jaffe, W., et al. 2001, AJ, 121, 2431 Urry, C. M., & Padovani, P. 1995, PASP, 107, 803 Sadler, E. M., Jenkins, C. R., & Kotanyi, C. G. 1989, MNRAS, 240, Wrobel, J. M. 1991, AJ, 101, 127 Wrobel, J. M., & Heeschen, D. S. 1991, AJ, 101, 148 Satyapal, S., Sambruna, R. M., & Dudik, R. P. 2004, A&A, 414, 825 B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy?, Online Material p 1

B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy?, Online Material p 2 Appendix A: Notes on the X-ray observations
of the individual sources
We list references and provide comments for the Chandra dataand X-ray nuclear measurements found in the literature. Wealso give images and spectra for the two newly detected X-raynuclei, as well as a summary of the results of the data analysiswhich can be found in Table A.1.
UGC 7360: this object is part of the 3C/FR I sample of low
luminosity radio-galaxies (3C 270). The Chandra data are pre-
Fig. A.1. Chandra image and spectrum for NGC 3557. The fit and the
sented in Balmaverde & Capetti (2005).
contributions of the two components (thermal and power-law) are also UGC 7386: a total of 310 nuclear counts in 0.2–8 keV band
were extracted from a 2 diameter circle without background
subtraction and converted to X-ray luminosity (2–10 keV) as-
suming an intrinsic power-law spectrum with photon index
Γ = 1.8 and a column density NH = 2 × 1020 cm−2 (Ho et al.
2001).
UGC 7494: this object is part of the 3C/FR I sample of low lu-
minosity radio-galaxies (3C 272.1). The Chandra data are pre-
sented in Balmaverde & Capetti (2005).
UGC 7629: Both Soldatenkov et al. (2003) and Maccarone
et al. (2003) detected emission from the nucleus at energies
below 2.5 keV. Biller et al. (2004) confirmed the presence ofnuclear emission in the 0.3–10 keV band extracting 64 source Fig. A.2. Chandra image and spectrum for NGC 5419. The fit and the
counts in an 1 circle. The spectrum was modeled using a contributions of the two components (thermal and power-law) are also power-law model with photon index Γ = 1.7 and a column plotted. Two datasets were fit simultaneously.
density NH = 1.7 × 1020 cm−2. These results do not contrast the
upper limit given by Loewenstein et al. (2001) who did not find
a nuclear component in the hard X-ray band from 2 to 10 keV.
UGC 7654: this object is part of the 3C/FR I sample of low
luminosity radio-galaxies (3C 274). The Chandra data are pre-
sented in Balmaverde & Capetti (2005).
NGC 1316: Kim & Fabbiano (2003) detected a low luminos-
UGC 7760: Filho et al. (2004) find a bright X-ray nuclear
ity X-ray AGN; the nuclear spectrum is well reproduced by a source,spatially coincident with the radio core position. The power-law model Γ = 1.76 plus a MEKAL model.
spectrum of the nuclear source (1200 net counts) is well fit-ted by a two component model, i.e. a power law (Γ = 1.51) NGC 1399: Loewenstein et al. (2001) did not find a nuclear
plus Raymond-Smith thermal plasma (KT = 0.95 keV).
point source and they give a 3σ upper limit to any nuclear UGC 7878: Loewenstein et al. (2001) give a 3σ upper limit to
X-ray point source converting nuclear counts (28 counts in any nuclear X-ray point source converting the nuclear counts an 1 circle region) to luminosity in (2–10 keV band) assuming (16 counts in an 1 circle region) to luminosity in the 2–10 keV a slope 1.5 power-law spectrum.
band assuming a power-law spectrum with a slope Γ = 1.5.
NGC 4696: Satyapal et al. (2004) extracted 39 events in the
UGC 7898: Soldatenkov et al. (2003) detected nuclear emis-
2–10 keV band with a 2 radius centered on the nucleus and sion at a 3 confidence level only in the range 0.2–0.6 keV with converted to 2–10 keV X-ray luminosity assuming an intrin- a luminosity of 6×1037 erg s−1. Conversely Randall et al. (2004) sic power-law spectrum with Γ = 1.8 and a Galactic column did not find conclusive evidence for a central AGN. They give an upper limit for the X-ray luminosity converting count rate NGC 5128: In Kraft et al. (2000) the nuclear spectrum is mod-
(0.3–10 keV) into the un-absorbed luminosity Lx(0.3–10 keV) eled assuming a power-law spectrum with a photon index of 1.9 using photon index Γ = 1.78. We adopted the latter, more con- and NH = 1023 cm−2. Evans et al. (2004) compared different nuclear observations of Cen A and confirmed that the spectrum UGC 9706: Filho et al. (2004) detected a weak hard x-ray
is well fitted by a heavily absorbed power-law model, consis- nucleus (total counts ) and fitted the spectrum with a power- tent with the previous observation.
law model (Γ = 2.29). Their result is consistent with that ofTrinchieri & Goudfrooij (2002).
IC 1459: Fabbiano et al. (2003) detected an un-absorbed nu-
UGC 9723: Terashima & Wilson (2003) did not detect an
clear X-ray source (∼6500 total counts), with a power law slope X-ray nucleus and suggested that this object is likely to be heavily obscured. They set a nuclear flux upper limit assum- IC 4296: Pellegrini et al. (2003) found point-like and hard
ing the Galactic absorption column density and a power-law X-ray emission, well described by a moderately absorbed model with photon index Γ = 2.
(NH = 1.1 × 1022 cm−2) power-law with Γ = 1.48.
B. Balmaverde and A. Capetti: A miniature radio-galaxy in every "core" galaxy?, Online Material p 3 Table A.1. Summary of the Chandra data analysis for the two newly detected X-ray nuclei.
Observation information Fx,nuc(1 keV) χ2/d.o.f. or PHA bins 1.1+1.0 E − 14 7.5+6.2 E − 14 Fig. B.1. Comparison of radio and optical nuclear luminosity for the
sample of core-galaxies (filled circles) using the high resolution radio
data to measure the radio cores. Empty circles are objects from the
3C/FR I sample of low luminosity radio-galaxies. The dashed line re-
produces the best linear fit, while the solid line shows the fits obtained
with the 5 GHz VLA data, from Fig. 4.
Appendix B: High resolution radio core
In Sect. 3 we presented measurements of the radio core fluxesobtained from observations obtained at higher resolution and/orfrequency with respect to the 5 resolution data available forthe whole sample. Since these data are highly inhomogeneousand given the general agreement with the 5 GHz VLA mea-surements, we prefer to retain the values from the surveysby Wrobel & Heeschen (1991) and Sadler et al. (1989), but,nonetheless, we always checked that using these estimates forthe core fluxes our results are unchanged. Here we give one ex-ample, exploring the influence on the correlation between ra-dio and optical nuclear luminosity: the slope of the best fit isincreased by only 0.02, thus it is essentially unaffected.

Source: http://www.aanda.org/articles/aa/pdf/2006/07/aa4031-05.pdf