Jet-accretion system in the nearby mJy Radio Galaxies
Paola Grandi, Eleonora Torresi, Duccio Macconi, Bia Boccardi, Alessandro Capetti
DD RAFT VERSION F EBRUARY
18, 2021Typeset using L A TEX preprint style in AASTeX63
Jet-accretion system in the nearby mJy Radio Galaxies P AOLA G RANDI , E LEONORA T ORRESI , D UCCIO M ACCONI ,
2, 1 B IA B OCCARDI , AND A LESSANDRO C APETTI INAF-Osservatorio di Astrofisica e Fisica dello Spazio, Via Gobetti 101, I-40129, Bologna, Italy Dipartimento di Fisica e Astronomia, Universitá degli Studi di Bologna, via Gobetti 93/2, I-40129 Bologna, Italy Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69, D-53121 Bonn, Germany INAF-Osservatorio Astrofisico di Torino, Strada Osservatorio 20, I-10025, Pino Torinese, Italy (Revised December 22, 2020; Accepted February 15, 2021)
ABSTRACTIt is generally thought that FRII Radio Galaxies host thin optically thick disks, while FRIs arepowered by Advected Dominated Accretion Flows. The sources with an efficient engine areoptically classified as High Excitation Radio Galaxies (HERGs) and those with an inefficientmotor as Low Excitation Radio Galaxies (LERGs). Recently, the study of Radio Galaxies downto mJy fluxes has cast serious doubts on the LERG-FRI and HERG-FRII correspondence,revealing that many LERGs show FRII radio morphologies.The FR catalogs recently compiled by Capetti et al. (2017a,b) and Baldi et al. (2018) haveallowed us to explore this issue in the local ( z ≤ . ) mJy Universe. Our statistical studyshows that the majority of nearby mJy objects are in a late stage of their life. FRII-LERGsappear more similar to the old FRI-LERGs than to the young FRII-HERGs. FRII-LERGsmay be aged HERGs that, exhausted the cold fuel, have changed their accretion regime or aseparate LERG class particularly efficient in launching jets. Exploiting the empirical relationswhich convert L [OIII] and L . into accretion power and jet kinetic power, respectively, weobserved that LERGs with similar masses and accretion rates seem to expel jets of differentpower. We speculate that intrinsic differences related to the black hole properties (spin andmagnetic field at its horizon) can determine the observed spread in jet luminosity. In this view,FRII-LERGs should have the fastest spinning black holes and/or the most intense magneticfluxes. On the contrary, compact LERGs (i.e. FR0s) should host extremely slow black holesand/or weak magnetic fields. Keywords: galaxies: active-galaxies: galaxies:-galaxies:jet INTRODUCTIONRadio Galaxies (RGs) are historically divided in core-brightened FR I and bright edge-brightened FR II(Fanaroff & Riley 1974), on the basis of their extended radio morphology that approximately changes at acritical power P . ∼ × W Hz − . Ledlow & Owen (1996) refined the classification showing that Corresponding author: Paola [email protected] a r X i v : . [ a s t r o - ph . H E ] F e b G RANDI ET AL .the break between FRIs and FRIIs is a strong function of the host galaxy absolute magnitude ( M R ). As thehost galaxy luminosity traces the black hole mass (Magorrian et al. 1998) and the radio power is proportionalto the accretion luminosity (Willott et al. 1999), the FRI-FRII separation was later re-interpreted in terms ofaccretion rates (Ghisellini & Celotti 2001). The less powerful radio galaxies (i.e. FRIs) host an inefficient hotthick flow, while the more powerful sources (i.e. FRIIs) have an efficiently accreting cold disk. In support ofthis interpretation, Marchesini et al. (2004) found an accretion rate gap between FRIs and FRIIs, suggestiveof a different accretion regime. From the optical point of view, radio galaxies are split into High ExcitationRadio Galaxies (HERGs) and Low Excitation Radio Galaxies (LERGs) (Jackson & Rawlings 1997), withLERGs characterized by [OIII] equivalent width < Å and/or [OII]/[OIII] ratios > . More recently,Buttiglione et al. (2010) proposed a combination of emission lines, the excitation index (EI ), to distinguishthe classes: LERGs have EI < . and HERGs EI > . .As FRIIs are generally associated to HERGs and FRIs to LERGs, it is almost natural to consider thenuclear engine as the main driver of the FRI-FRII dichotomy. However, this one-to-one correspondence(FRI-LERGs versus FRII-HERGs), based on the study of powerful sources with Jy flux densities, is probablya simplification.For example, 24 FRII sources in the 3CR sample (Buttiglione et al. 2010) lack high excitation emissionlines and are classified as LERGs. Similarly, Tadhunter et al. (1998) studying the 2Jy sample (Wall &Peacock 1985) found that 23% of the FRIIs are Weak Line Radio Galaxies (WLRGs), i.e. objects withEW [OIII] < Å. As discussed by Tadhunter (2016), WLRGs generally correspond to LERGs, although theclassification criteria are slightly different. Moreover, some FRIs have efficient accretion disks (i.e. theyare FRI-HERGs). 3C 120, with broad and intense optical lines, a prominent UV bump, and an iron line inthe X-ray spectrum (Torresi 2012), is a typical example. The difficulty in reconciling accretion mode andkpc radio morphology has become more evident in recent years when large-area radio (NVSS, FIRST) andoptical (SDSS and 6dFGRS) spectroscopic surveys have allowed expanding the study of radio galaxies downto mJy fluxes (see Heckman & Best (2014) for a review). Several studies show that radio galaxies with FRIImorphologies preferentially host low efficient accretion flows (i.e. they are classified as LERGs) at low fluxdensities (Capetti et al. 2017a,b; Miraghaei & Best 2017). Finally, a recent analysis of low luminosity radiogalaxies observed by LOFAR Mingo et al. (2019) has questioned the FRI/FRII break based on the radiopower. At low fluxes, any association between morphology and radio luminosity seems to disappear. If radiogalaxies of similar radio morphology (radio power) can come into different "accretion flavors", new scenarioshave to be considered. The accretion rate could not be the driving parameter and something else related tothe black hole could play a major role in launching the jet (Ghisellini et al. 2014). The environment couldbe also important, as radio, optical and X-ray studies (Croston et al. 2005; Gawro´nski et al. 2006; Crostonet al. 2008; Gendre et al. 2013; Ineson et al. 2017; Croston et al. 2018; Mingo et al. 2019; Macconi et al.2020) seem to suggest. Finally, we could be observing different phases that AGN pass through their life.For example, a recent X-ray analysis of 3C radio galaxies (Macconi et al. 2020) has shown that FRII-LERGnuclei have less cold gas, i.e. smaller column densities (N H ) than FRII-HERGs. A possible suggestion is thata transition occurs from a thin disk to a thick flow in FRIIs when the cold fuel has been depleted. Incidentally,this leads to speculate that FRIs could switch from LERG to HERG if a sudden replenishment of fresh coldgas occurs, maybe due to a galaxy merger (see, for example, Garofalo & Singh (2019)). However, mostof the results (and controversial interpretations) are based on the study of very bright (Jy) radio sources EI = Log ([ OIII ] /H β ) − / Log ([ N II ] /H α ) + Log ([ SII ] /H α ) + Log ([ OI ] /H α )) ADIO G ALAXIES IN THE NEARBY M J Y U NIVERSE Table 1.
Sample selection criteria: F . GHz > mJySample z Optical ExtensionClass (kpc)FR0cat < . LERG < . FRIcat < . LERG > sFRIcat < . LERG > and < FRIIcat < . LERG/HERG > which make up only a small fraction of the total radio galaxy population. In order to shed light on these openquestions, the jet-accretion system is explored through the study of local faint (mJy) radio galaxies takingadvantage of the recently compiled FR catalogs FRcat (Capetti et al. 2017a,b; Baldi et al. 2018) that includesources well characterized both in the radio and optical bands.A cosmology with H = 67 km s − Mpc − , Ω m = 0 . , and Ω Λ = 0 . (Planck Collaboration et al. 2014)is assumed in this paper. THE FRCAT SAMPLESThe FR0 Baldi et al. (2018), FRI (Capetti et al. 2017a) and FRII (Capetti et al. 2017b) catalogs include108, 219 and 122 radio galaxies, respectively. They are part of a large sample assembled by Best & Heckman(2012) (hereafter B12 sample), cross-correlating the seventh data release of the Sloan Digital Sky Survey(SDSS) with the NRAO (National Radio Astronomy Observatory) VLA (Very Large Array) Sky Survey(NVSS) and the Faint Images of the Radio Sky at Twenty centimeters (FIRST) survey. For all the AGN inthe sample, Best & Heckman (2012) provided an optical classification (LERG/HERG) assuming differentcriteria (Buttiglione et al. 2010; Kewley et al. 2006; Cid Fernandes et al. 2010) depending on the signal tonoise ratio of emission lines (e.g. H α , H β , OIII, OI, NII, SII) revealed in the SDSS spectra.The FRI and FRII catalogs are limited to local radio galaxies (maximum distance z=0.15) with an NVSSflux density larger than 5 mJy and a (one-side) extension of at least 30 kpc . The radio classification wasperformed by a visual inspection of the FIRST images. If a radio galaxy showed a higher surface brightnessnear the core (edge-darkness), it was defined as FR Type I. On the contrary, if it appeared brighter at the end(edge-brightened), the associated radio class was FR Type II. An additional sample of 14 small FRIs (sFR)was also included in the FRIcat. It consists of sources located at z ≤ . and with a radio extension between10 and 30 kpc. As stressed by Capetti et al. (2017b), the FRI and FRII catalogs are statistically complete at alevel of ∼ in the optical range and have a flux limit of ∼ mJy at 1.4 GHz.The FR0 catalog (Baldi et al. 2018) consists of FIRST compact radio galaxies with a minimum fluxdensity of 5 mJy at 1.4 GHz, at redshift ≤ . (i.e. with a maximal radio extension of 2.5 kpc), all opticallyclassified as LERGs (see Baldi et al. (2019) for a review on this class of objects). Four compact sources withHERG properties were also revealed but not included. A summary of the selection criteria is reported inTable 1.While in the 3CR catalog more than 40% are powerful radio galaxies with an efficient accretion disk,in the FR catalogs Radio Galaxies with high excitation emission lines are a minority ∼ (see Figure 1).Interestingly enough, Miraghaei & Best (2017) provided the radio classification of another BH12 subsample
1" corresponds to 2.72 kpc at z=0.15. G RANDI ET AL . Figure 1.
Fraction of FRI and FRII radio galaxies with z < . in the 3CR ( left pie ) and in the FRcat ( right pie )catalogs. The number of LERGs increases going down to mJy fluxes. adopting slightly different selection criteria. Despite the different approaches, they confirm the predominanceof LERGs in the FRII population.2.1. SDSS observables and derivated quantities
Thanks to the MPA-JHU DR7 release of spectrum measurements , we could estimate black hole massand radiative luminosity for each source of the FR catalogs. The velocity dispersion ( σ ∗ ) was converted intoBH mass using the relation Log ( M BH /M (cid:12) ) = 8 .
32 + 5 . × Log ( σ ∗ / km s − ) (McConnell & Ma 2013)and the [OIII] λ L acc ) usingthe multiplicative factor 3500, L acc = 3500 × L [OIII] (Heckman et al. 2004).Other important DR7 quantities, useful to characterize the galaxies hosting different FR classes, are thestellar mass and the calcium break D n n Checking the FR0 sample
The selection and classification of large samples of objects necessarily imply the inclusion of a smallfraction of spurious sources. As noted by Best & Heckman (2012), this is not a problem if the number ofsources is large (of the order of several hundreds or more), but it could have an impact on smaller samples.This is particularly true for the FR0s that are not resolved in the FIRST survey and could be misclassified.Low radio flux density sources without any resolved jet structure could hide a weak BL Lac nucleus or aradio-quiet LINER with intense star-formation.We sought for possible spurious sources exploring the WISE color-color diagram. Three infrared bands,w1(3.4 µ m), w2(4.6 µ m), and w3(12 µ m) were considered. The color w3-w2 was plotted versus the color w2-w1. It is known that different sources occupy different regions of the plot with redder objects characterized ADIO G ALAXIES IN THE NEARBY M J Y U NIVERSE Figure 2.
Diagnostic diagrams of radio sources with z ≤ . of the sample of Best & Heckman (2012). LERGs arepurple open circles, HERGs purple squares and Star-forming galaxies (SF) orange stars. FR0s from Baldi et al. (2018)are marked by black points. Left panel – WISE color-color diagram. LERGs are in the lower right corner, while the fewHERGs are shifted towards redder colours. SF on the left are spread on a wide zone. FR0s mainly fall in the LERGzone. Those with w3-w2 > and w2-w1 > . (open black squares) are excluded by the sample as possible spuriousobjects. Right panel – D n L . GHz /M ∗ diagnostic plane. The majority of FR0s are clustered in the AGNregion with D n ∼ . The sources with D n < . are rejected as possible Star-forming galaxies or BL Lacs.. by higher values of w3-w2 and w1-w2. Elliptical galaxies are expected to have colors near zero whilestar-forming galaxies (SF) are very red in both w3-w2 and w2-w1. Radio quiet and radio loud AGN withefficient accretion disks and dusty screens are in between (see Fig. 12 of Wright et al. (2010)).Figure 2 ( left panel ) shows all the objects of the B12 sample with z ≤ . together with FR0s markedas black points. LERGs are in the elliptical region, star-forming galaxies mainly above w3-w2 > andthe few HERGs in the AGN area. As expected there is overlapping between FR0s and LERGs, although ahandful of compact radio sources are shifted to redder colors (w3-w2 > and w2-w1 > . ). Some of themare also clearly separated from the bulk of the FR0 population in the D n L . /M ∗ diagram(Fig. 2– right panel ), one of the diagnostic plots proposed by Best & Heckman (2012) to divide AGN fromstar-forming galaxies.Taking a conservative approach, we then decided to exclude from our statistical analysis those FR0swith redder WISE colours (w3-w2 > and w2-w1 > . ). These have a not negligible probability to be starforming galaxies. Also, we also excluded FR0s with D n < . . As pointed out by Capetti & Raiteri(2015), a small amplitude of the Å break could be a signature not only of young stars but also of ajet. The non-thermal radiation can indeed dilute the optical continuum reducing the D n M (cid:12) (typical of radio-quiet AGN), and a highprobability to reside in spiral galaxies (Huertas-Company et al. 2011; Tempel et al. 2017).At the end of this selection, the "clean" sample consisted of 99 FR0s with only 9 rejected sources (lessthan 1% of the sample). G RANDI ET AL . COMPARISON AMONG THE DIFFERENT CLASSESTable 2 reports median, average, and relative standard deviation of all the studied quantities. A comparisonamong the different classes was performed by applying a Kolmogorov-Smirnov (KS) test. We conservativelyassumed that two data sets are different if the Kolmogorov-Smirnov probability is less than × − . Inother words, we reject the null hypothesis that the two data sets are drawn from the same distribution at aconfidence level > σ . The KS results are in Table 3. In Figure 3 the most interesting histograms are shown.As expected, FRII-HERGs and FRIs are distinct populations. FRII-HERGs have smaller black holes, alarger accretion rate expressed in terms of L acc /L Edd and more stellar activity. They are younger systems.More interesting is that our analysis shows that FRII-HERGs and FRII-LERGs are also different. FRIIs withan inefficient engine have more massive black holes and a more evolved stellar population (Table 2 and Fig.3), i.e. are more similar to FRIs. Indeed, FRI and FRII-LERG classes are almost completely overlapped inthe histograms of Fig. 3.The nuclear properties of mJy LERG sources, independently of their radio morphology, are very similar, atodds with the trend observed in Jy radio galaxies. As shown by Macconi et al. (2020), the FRII-LERGs of the3C sample have indeed accretion rates ( L acc /L Edd ) generally lower than FRII-HERGs but higher than FRIs.At z ≤ . no significant difference is observed between small and extended FRIs. FR0s have accretionrates slightly higher than FRIs (see Tab 2). Although potentially interesting, we do not further speculate onthis result, as still missed outliers in the FR0 sample can not be definitively excluded. However, recent stellaractivity is not observed in any classes, suggesting that all the LERGs in the Local Universe are in a late stageof evolution. Figure 3.
Eddington-scaled radiative luminosity ( left panel ) and calcium break ( right panel ) histograms. FRII-LERGsare similar to FRIs. Both live in old galaxies and are powered by hot accretion flows. FRII-HERGs are more radiativelyefficient and show signs of star-forming activity.4.
JET POWER VERSUS ACCRETION POWERIn this section we explore the jet-accretion link in mJy radio galaxies to find possible intrinsic differencesin their nuclear engine. Following Shankar et al. (2008), we define the jet and the radiation power as:
ADIO G ALAXIES IN THE NEARBY M J Y U NIVERSE Table 2.
FRCat Average Properties
Class median average std N. objects Class median average std N.objects ( z < . ) Log (L . GHz ) (erg s − ) ( z < . ) Log( L . GHz ) (erg s − )FRI 40.32 40.34 0.34 219 FR0 38.87 38.96 0.36 99FRII-LERG 40.77 40.75 0.49 108 FRI 40.12 40.13 0.40 9FRII-HERG 41.37 41.26 0.55 14 small FRI 39.52 39.60 0.34 14Log( L OIII ) (erg s − ) Log( L OIII ) (erg s − )FRI 39.87 39.86 0.23 107 FR0 39.60 39.58 0.27 98FRII-LERG 39.91 39.95 0.39 66 FRI 39.51 39.50 0.17 9FRII-HERG 41.42 41.30 0.52 14 small FRI 39.39 39.37 0.15 12Velocity Dispersion σ ∗ (km s − ) Velocity Dispersion σ ∗ (km s − )FRI 254 256 35 218 FR0 231 237 38 97FRII-LERG 246 243 39 102 FRI 270 276 31 9FRII-HERG 209 204 23 12 small FRI 253 261 32 14Stellar Mass Log( M ∗ ) (M (cid:12) ) Stellar Mass Log( M ∗ ) (M (cid:12) )FRI 11.38 11.38 0.19 214 FR0 11.15 11.13 0.22 97FRII-LERG 11.33 11.33 0.31 107 FRI 11.32 11.34 0.13 9FRII-HERG 11.09 11.08 0.25 14 small FRI 11.22 11.23 0.17 14Calcium break (D n n (cid:12) ) Log(BH) (M (cid:12) )FRI 8.91 8.91 0.39 218 FR0 8.67 8.71 0.39 97FRII-LERG 8.83 8.77 0.41 102 FRI 9.06 9.10 0.27 9FRII-HERG 8.43 8.36 0.28 12 small FRI 8.89 8.95 0.29 14Log( L acc /L Edd ) Log( L acc /L Edd )FRI -4.03 -4.03 0.40 106 FR0 -3.99 -4.04 0.47 96FRII-LERG -3.85 -3.74 0.55 62 FRI -4.38 -4.50 0.32 12FRII-HERG -1.84 -2.1 0.57 12 small FRI -4.38 -4.50 0.32 12 G RANDI ET AL . Table 3.
FRCat Kolmogorov Smirnov test results
Classes KS probability z < . Log( L . GHz ) Log ( L OIII ) Log(BH) Log( L acc /L Edd ) Log( M ∗ ) D n FRI vs FRII-LERG < − . × − . × − . × − < − < − < − < − < − < − FRII-LERG vs FRII-HERG . × − < − < − < − . × − < − z < . Log( L . GHz ) Log( L OIII ) Log(BH) Log( L acc /L Edd ) Log( M ∗ ) D n × − . × − < − . × − < − . × − Figure 4. L . GHz /L Edd versus L [ OIII ] /L Edd of FRcat sources compared to the predicted values estimated byequation (2) assuming f = 5 ( left panel ) and f = 20 ( right panel ). Each line in the plots corresponds to a differentvalue of η/(cid:15) . Solid and dotted lines corresponds to M BH = 10 . M (cid:12) and M BH = 10 . M (cid:12) , respectively. P jet = η ˙ M c and L acc = (cid:15) ˙ M c , being η and (cid:15) , the fraction of gravitational energy converted into jet powerand thermal radiation, respectively. Combining the two relations we obtain: P jet = ( η/(cid:15) ) L acc (1)The ( η/(cid:15) ) ratio directly measures the ability of the system to channel gravitational energy into the jet ratherthan to dissipate it in thermal radiation.The radiative power can be related to the [OIII] luminosity through: L acc = L OIII × (see Section2), while the kinetic power, expressed as a function of the radio luminosity ( L rad ), is generally written as: P jet = KL Γ rad . ADIO G ALAXIES IN THE NEARBY M J Y U NIVERSE
9A jet power-radio luminosity relation was at first proposed by Willott et al. (1999): P jet = 4 × f / L . . (2)Here the original relation, that uses the radio luminosity at 151 MHz, is re-scaled to 1.4 GHz adopting aradio spectral index α =0.8 (Heckman & Best 2014). The luminosity is in units of Watt Hz − .As a starting point, Willott et al. (1999) assumed that the jet energy is mainly stored in the lobe and/or utilizedto expand the radio source and considered the radiative losses negligible. They provided a minimal estimationof the internal energy in an equipartition regime (i.e. the internal energy is almost equally distributedbetween magnetic field and relativistic particles) and then divided this quantity by the source age. The f factor (included in the normalization) absorbs all the uncertainties on the physical state of the lobes, such asthe particle composition and their spectral distribution, volume filling factor, possible deviation from theequipartition condition, presence of internal turbulence, and fraction of internal energy lost as work doneduring the lobe expansion (assumed to be ∼ of the internal energy). Among these, the number ofprotons per electron present in the relativistic plasma is the most relevant one. The analysis of Willott et al.(1999), based on FRII and Steep Spectrum Radio Quasars, constrained f between 1 (light jet) and 20 (heavyjet). Later, Daly et al. (2012), investigating 31 FRII radio galaxies with an accurate radio characterization,found a P jet relation in substantial agreement with equation (2) and a value of f ∼ .The extension of these studies to sources in gas-rich environments (mainly FRIs) was viable after thelaunch of the Chandra satellite. The discovery of X-ray cavities around radio lobes suggested a differentapproach to estimate P jet . The jet power could be deduced considering the energy spent by lobes to displacethe surrounding gas and the age of the cavity (see Bîrzan et al. (2008) for details on the age calculation).The required energy (i.e. the enthalpy H) to excavate the medium is the sum of the work done by the lobes(pV) and their thermal energy. The enthalpy is assumed to be H=4pV if the lobe is dominated by relativisticparticles (McNamara & Nulsen 2007). Cavagnolo et al. (2010) studied 16 giant radio galaxies (mainly FRIs)and found P jet ∝ L . ± . . , in quite agreement with the Willott’s relation.Finally, we mention a study of 15 radio galaxies (Merloni & Heinz 2007) where a relation between the jetpower, estimated from X-ray cavities, and the radio core luminosity was explored. The authors reported arelation, P jet ∝ L . core , similar to those deduced by Willott et al. (1999) and Cavagnolo et al. (2010) allowingto extend the study of jet kinetic power to also small and unresolved Radio Galaxies . The assumption here isthat beaming Doppler boosting effects do not amplify the 1.4 GHz radiation. We are quite confident thatthis condition is satisfied by our Radio Galaxies , considering that suspected BL LAC objects have beenexcluded by the FR0 cleaned sample. Incidentally, we note that the core contribution does not significantlyaffect the extended Radio Galaxies . As a check, we obtained a rough estimation of the "extended" 1.4GHz luminosity subtracting the FIRST peak flux (assumed to be a proxy of the core emission) from thetotal NVSS flux. This test was possible for more than 90% of FRI and FRII sources. The luminosities gen-erally changed less than 0.2 dex on logarithmic scale with no significant impact on our study (see next section).Considering the above-mentioned caveats, we will handle the f parameter as an unknown variable in thefollowing. Moreover, when equation (2) is exploited to estimate the η/(cid:15) ratios of FR0s and small FRIs, theirNVSS luminosities will be treated as upper limits.4.1. Predicted luminosities in systems with different accretion efficiency ratios
RANDI ET AL .Relation (2) can be re-written in order to have the 1.4 GHz luminosity as subject: for an [OIII] luminosityas input, it then allows to estimate the expected radio luminosity for any value of f and η/(cid:15) .The predicted L . values, obtained assuming L [OIII] ranging from 39 to 45, are shown in Figure 4for f = 5 ( left panel ) and f =20 ( right panel ). The luminosities are rescaled to the Eddington luminosity(L Edd ), using two different black hole masses ( M BH = 10 . M (cid:12) and M BH = 10 . M (cid:12) ) to cover the massrange observed in the FRcat sources. Each line in the plots corresponds to a different value of η/(cid:15) . Solidand dotted lines correspond to M BH = 10 . M (cid:12) and M BH = 10 . M (cid:12) , respectively. Note that a change ofthe black hole mass has not an important impact on the predicted η/(cid:15) curves. As stressed in the previoussection, the FRI and FRII radio luminosities could be overestimated by ∼ . dex. Considering the intrinsicspread of each class in each plot, this effect is negligible. For small and compact sources the η/(cid:15) valuesshould be considered upper limits, being unconstrained the contribution of the radio core to the total L . GHz luminosity. Comparing the two panels, it appears also evident that a variation of f only translates the η/(cid:15) curves, preserving the relative position of the different classes.As expected, LERGs and HERGs, having different accretion rates (i.e. different L [OIII] /L Edd ), occupydifferent parts of the plot. FR0s, being compact radio sources by definition, populate the lower left corner.However, a more careful inspection of Figure 4 shows that FRIs and HERGs preferentially fall in different η/(cid:15) strips and that LERGs are spread along the y axis. It seems that jet-disk systems in HERGs favoura thermal dissipation of the gravitational power, while jets of different powers can be launched by verysimilar inefficient accretion flows. However, the implicit assumption here is that the normalization (i.e. f ) ofequation (2) is the same for each FR classes.4.2. Eploring the [ η/(cid:15) -f] parameter space of the FRcat sources
In order to better investigate the problem, we then decided to explore the [ η/(cid:15) -f] parameter space of eachclass. This time, the η/(cid:15) values were determined via equation (2) utilizing the observed average L [OIII] , L . and L Edd luminosities in Table 2 and running f from 1 to 20. The ( η/(cid:15) -f) pairs that do not satisfy thecondition L acc ≤ L Edd were excluded.In Figure 5-( left panel ), the f and η/(cid:15) permitted values for FRII and FRI radio galaxies at z > . areshown for two different accretion rates. The separation at Log ( L [OIII] /L Edd ) = − . is based on Figure4. As already noted, the efficiency ratio increases from FRII-HERGs to FRII-LERGs if f is kept constant.Different classes could have the same η/(cid:15) ratio only if the normalization of equation (2), i.e. f , is allowed tovary.If the main source of uncertainty included in f is the plasma particle content (Willott et al. 1999), thecondition of equal η/(cid:15) could be reached in FRIs and FRII-HERGs only assuming that jets are lighter in theformer sources. Although not completely rejectable (our understanding of the particle acceleration nearthe black hole is really poor), this hypothesis does not seem to be supported by the observations of radiostructures on kpc scales. The decelerated and less collimated jets seen in FRIs are indeed suggestive ofstrong interactions with the environment and mass loading through mixing in turbulent layers (Perucho 2020).X-ray studies of radio lobes and gaseous environments of FRIs and FRIIs (Ineson et al. 2017; Croston et al.2018) indicate indeed that core-brightened Radio Galaxies contain more protons than edge-brightened RadioGalaxies.Another source of uncertainty, that could be invoked to satisfy the equal η/(cid:15) condition, is the ambientmedium (Cavagnolo et al. 2010). Jets that propagate in a dense environment have to spend more internalenergy pushing away the surrounding gas. A larger corrective factor (thus a larger f ) should then be included ADIO G ALAXIES IN THE NEARBY M J Y U NIVERSE f than HERGs for equal η/(cid:15) ) are preferentiallyfound in groups or clusters when bright (Jy) radio galaxies are considered (Gendre et al. 2013). Moreover, noenvironmental difference between FRIs and FRIIs is observed in the Local mJy Universe (Massaro et al.2019).In summary, it seems unlikely that Radio Galaxies , powered by different accretions, choose the samedissipative channel. It is more plausible that FRIs convert most of their gravitational energy into jets powerand FRII-HERGs into thermal radiation.FRII-LERGs represent a more complex class. They have radio morphologies and particle content (Inesonet al. 2017) similar to FRII-HERGs but habit older galaxies, have more massive black holes and a smalleraccretion rate. In addition, they have the largest η/(cid:15) ratios (Fig. 4) despite their marked similarity with FRIs.As proposed by Macconi et al. (2020), a possibility is that FRII-LERGs are old HERGs that, exhausted theirfuel, have changed the accretion mode. In this case, the values of η/(cid:15) are meaningless, as Equation (2) cannotbe applied anymore being the nuclear region and the extended radio structures temporarily disconnected. Onthe other hand, theoretical studies show that inefficient accretion configurations between an ADAF and acold efficient disk can exist (see Figure 1 of Yuan & Narayan (2014)). If, for some reasons more viscousdissipated energy is transferred into electrons and radiated away, the ADAF accretion flow can enter into amore luminous hot accretion flow regime (Xie & Yuan 2012). If the electron cooling becomes too strong, thematter in accretion collapses in a cold disk or in cold dense clumps embedded in a hot flow (Yuan & Narayan2014). We could then be observing the inverse trend.Another option is that FRII-LERGs are a separate class and not a product of the FRII-HERG evolution.The observed L OIII /L Edd spread of FRII-LERGs (see Fig. 4) could then be simply due to the co-existenceof ADAF configurations with different electron cooling. The high η/(cid:15) ratios are however difficult to explain,unless, for example, more extreme conditions of the black hole properties are assumed for this class (seediscussion below).The [ η/(cid:15) -f] parameter space of Radio Galaxies at z ≤ right panel ).The plot is similar to that observed for sources at higher redshift: the efficiency ratio increases from FR0sto extended FRIs if f is kept fixed. Also in this case equal η/(cid:15) values would require f changing from a classto another one, implying possible different intrinsic (jet content) or external (environment) conditions. FR0s,that are less able to channel energy into the jets ( η/(cid:15) is always less than 1) should expel heavier jets or beembedded in a very dense environment. The first hypothesis is difficult to test (in particular if the radioemission is not extended on large scales). The second one is more intriguing. The idea that a hostile ambientinhibits the jet expansion of small/compact Radio Galaxies is indeed plausible. However, the observations donot support this view. A study based on the galaxy richness around the FRCat sources at z ≤ . does notreveal any FR0 over-density (Capetti et al. 2020). In addition, an X-ray study of the galaxy cluster Abell 795having a FR0 at its center (Ubertosi et al 2021 A & A submitted, Ubertosi Master Thesis ) found gas densityand temperature typical of clusters hosting more extended central radio galaxies.Figures 4 and 5 are suggestive of another viable interpretation. The different radio luminosities observed inLERGs having comparable accretion rates ( L [OIII] /L Edd ) might indicate that similar nuclear engines impart https://amslaurea.unibo.it/21460/ RANDI ET AL ..
In order to better investigate the problem, we then decided to explore the [ η/(cid:15) -f] parameter space of eachclass. This time, the η/(cid:15) values were determined via equation (2) utilizing the observed average L [OIII] , L . and L Edd luminosities in Table 2 and running f from 1 to 20. The ( η/(cid:15) -f) pairs that do not satisfy thecondition L acc ≤ L Edd were excluded.In Figure 5-( left panel ), the f and η/(cid:15) permitted values for FRII and FRI radio galaxies at z > . areshown for two different accretion rates. The separation at Log ( L [OIII] /L Edd ) = − . is based on Figure4. As already noted, the efficiency ratio increases from FRII-HERGs to FRII-LERGs if f is kept constant.Different classes could have the same η/(cid:15) ratio only if the normalization of equation (2), i.e. f , is allowed tovary.If the main source of uncertainty included in f is the plasma particle content (Willott et al. 1999), thecondition of equal η/(cid:15) could be reached in FRIs and FRII-HERGs only assuming that jets are lighter in theformer sources. Although not completely rejectable (our understanding of the particle acceleration nearthe black hole is really poor), this hypothesis does not seem to be supported by the observations of radiostructures on kpc scales. The decelerated and less collimated jets seen in FRIs are indeed suggestive ofstrong interactions with the environment and mass loading through mixing in turbulent layers (Perucho 2020).X-ray studies of radio lobes and gaseous environments of FRIs and FRIIs (Ineson et al. 2017; Croston et al.2018) indicate indeed that core-brightened Radio Galaxies contain more protons than edge-brightened RadioGalaxies.Another source of uncertainty, that could be invoked to satisfy the equal η/(cid:15) condition, is the ambientmedium (Cavagnolo et al. 2010). Jets that propagate in a dense environment have to spend more internalenergy pushing away the surrounding gas. A larger corrective factor (thus a larger f ) should then be included ADIO G ALAXIES IN THE NEARBY M J Y U NIVERSE f than HERGs for equal η/(cid:15) ) are preferentiallyfound in groups or clusters when bright (Jy) radio galaxies are considered (Gendre et al. 2013). Moreover, noenvironmental difference between FRIs and FRIIs is observed in the Local mJy Universe (Massaro et al.2019).In summary, it seems unlikely that Radio Galaxies , powered by different accretions, choose the samedissipative channel. It is more plausible that FRIs convert most of their gravitational energy into jets powerand FRII-HERGs into thermal radiation.FRII-LERGs represent a more complex class. They have radio morphologies and particle content (Inesonet al. 2017) similar to FRII-HERGs but habit older galaxies, have more massive black holes and a smalleraccretion rate. In addition, they have the largest η/(cid:15) ratios (Fig. 4) despite their marked similarity with FRIs.As proposed by Macconi et al. (2020), a possibility is that FRII-LERGs are old HERGs that, exhausted theirfuel, have changed the accretion mode. In this case, the values of η/(cid:15) are meaningless, as Equation (2) cannotbe applied anymore being the nuclear region and the extended radio structures temporarily disconnected. Onthe other hand, theoretical studies show that inefficient accretion configurations between an ADAF and acold efficient disk can exist (see Figure 1 of Yuan & Narayan (2014)). If, for some reasons more viscousdissipated energy is transferred into electrons and radiated away, the ADAF accretion flow can enter into amore luminous hot accretion flow regime (Xie & Yuan 2012). If the electron cooling becomes too strong, thematter in accretion collapses in a cold disk or in cold dense clumps embedded in a hot flow (Yuan & Narayan2014). We could then be observing the inverse trend.Another option is that FRII-LERGs are a separate class and not a product of the FRII-HERG evolution.The observed L OIII /L Edd spread of FRII-LERGs (see Fig. 4) could then be simply due to the co-existenceof ADAF configurations with different electron cooling. The high η/(cid:15) ratios are however difficult to explain,unless, for example, more extreme conditions of the black hole properties are assumed for this class (seediscussion below).The [ η/(cid:15) -f] parameter space of Radio Galaxies at z ≤ right panel ).The plot is similar to that observed for sources at higher redshift: the efficiency ratio increases from FR0sto extended FRIs if f is kept fixed. Also in this case equal η/(cid:15) values would require f changing from a classto another one, implying possible different intrinsic (jet content) or external (environment) conditions. FR0s,that are less able to channel energy into the jets ( η/(cid:15) is always less than 1) should expel heavier jets or beembedded in a very dense environment. The first hypothesis is difficult to test (in particular if the radioemission is not extended on large scales). The second one is more intriguing. The idea that a hostile ambientinhibits the jet expansion of small/compact Radio Galaxies is indeed plausible. However, the observations donot support this view. A study based on the galaxy richness around the FRCat sources at z ≤ . does notreveal any FR0 over-density (Capetti et al. 2020). In addition, an X-ray study of the galaxy cluster Abell 795having a FR0 at its center (Ubertosi et al 2021 A & A submitted, Ubertosi Master Thesis ) found gas densityand temperature typical of clusters hosting more extended central radio galaxies.Figures 4 and 5 are suggestive of another viable interpretation. The different radio luminosities observed inLERGs having comparable accretion rates ( L [OIII] /L Edd ) might indicate that similar nuclear engines impart https://amslaurea.unibo.it/21460/ RANDI ET AL .. Figure 5.
Left Panel – [ η/(cid:15) - f ] parameter space for FRIs, FRII-LERGs and FRII-HERGs with z > . . The curves areobtained via equation (2) assuming the average L [OIII] and L NVSS luminosities (and standard deviation σ ) in Table2. Dotted lines correspond to 1- σ uncertainty. The separation at Log ( L [OIII] /L Edd ) = − . between HERGs andLERGs is based on Figure 4. The intersections of the horizontal line with the curves mark the f values required tohave a same efficiency ratio ( η/(cid:15) = 0 . in this case) in each class. Right Panel – Same plot for FR0s, small FRIs andFRIs with z ≤ . . The horizontal line corresponds to η/(cid:15) = 0 . . Vertical arrows indicate that the η/(cid:15) values tracedby the curves have to be intended as upper limits. diverse accelerations to the outflowing plasma. One of the most accredited model for the jet production(Blandford & Znajek 1977) links the jet kinetic power to the properties of the black hole, i.e. mass, spin(a) and magnetic field at its horizon ( Φ ): P BZ ∝ Φ a M BH . Since LERGs have similar accretion ratesand black hole masses (Table 2), the vertical displacement of the FRcat [ η/(cid:15) - f ] curves could directly mapdifferent values of a and/or Φ . In this view, FR0s (at least those with Log ( L [OIII] /L Edd ) < − . shouldhave extremely slow black holes and/or weak magnetic fluxes, while FRII-LERGs, assumed not evolvedHERGs, the largest values of a and/or Φ .Finally, we noted that Equation (2) does not take into account mildly relativistic winds that couldcontribute to the total kinetic budget ( P tot = P jet + P winds ).An Advection Dominated Inflow-Outflow Solution (ADIOS) proposed by Blandford & Begelman (1999)predicts the presence of matter outflows that, exceeding the amount of material crossing the black holehorizon, favours a low accretion rate. Magneto-hydrodynamic (MHD) simulations of ADAF (S ˛adowski et al.2013) also predict winds. These are expected to be less energetic than jets unless the black hole spin and/orthe magnetic flux are small . Moreover, Liska et al. (2019) showed that both jets and magnetically drivenwinds can be produced by AGNs with a thin accretion disk (and a fast spinning black hole).On the observational side, several works attest to the existence of outflows in bright radio galaxies. A veryrecent work by Boccardi et al. (2020), exploring the innermost jet profile of several radio-loud AGN usingVLBI data, has confirmed the existence of thick disk-launched layers surrounding the HERG jets and of FR0s could be in this condition and dissipate more of their gravitational power into winds.
ADIO G ALAXIES IN THE NEARBY M J Y U NIVERSE ∼ . c . The measures of very fast outflows with velocitiesreaching an appreciable fraction of c are technically difficult and probably limited by the transiency of theevent. However, if consolidated, they will attest to the important role of the winds in the energy balance. CONCLUSIONSIn this paper the study of the mJy sources of the FRCat catalogs was performed following two differentapproaches. At first, we performed a statistical analysis of the main observables and compared the averageproperties of the different classes. Then we explored the jet-accretion system exploiting the known relationsthat link L [OIII] and L . to the accretion (thermal) and jet kinetic power, respectively.The main results of our statistical analysis are summarized below:• FRIs compared to FRII-HERGs show more massive black holes, smaller accretion rates (expressed interms of L acc /L Edd ), larger stellar masses, and a more evolved stellar population;• FRII-LERGs are more similar to FRIs than FRII-HERGs;• No significat difference is observed between small FRIs and FRIs at z < . . All the local sources arehosted in massive galaxies with no recent star forming activity and have comparable low accretionrates and black hole masses;• FR0s show M (cid:63) and D n L acc /L Edd ) ratio extend to valueshigher than local ( z ≤ . ) FRIs.These results suggest that, in the mJy Universe, the majority of radio galaxies within z ≤ . are in alate stage of their life. The only exception is represented by the FRII-HERG class which is however poorlypopulated.From a comparison between Jy and mJy FRII-LERGs, it emerges that lower radio flux density sourceshave, on average, FRI-like characteristics, while FRII-LERGs of the 3C sample are more ’active’ withintermediate properties between FRIs and FRII-HERGs (Macconi et al. 2020). This points towards anevolutionary scenario in which FRII-LERGs are aged FRII-HERGs. Once the nuclear cold fuel has beenconsumed, the accretion configuration becomes hot and inefficient while the extended radio structuresconserve traces of the past activity. It has been shown that a wide range of configurations between thickhot flow and thin cold disk is stable. If, for example, a strong and turbulent magnetic field permeates theaccreting matter, MHD instabilities/magnetic reconnections can further heat the electrons that can radiateaway giving origin to more luminous hot accretion flows. Similar accretion configurations could account forthe wide range of Log( L [OIII] /L Edd ) in Fig. 4, and even more for the higher [OIII] luminosities observed inthe 3C FRII-LERGs.We cannot however reject the hypothesis that FRII-LERGs are a separate and independent class with aninefficient accretion regime able to produce extended FRII radio structures. This breaks the correspondencebetween efficient/inefficient accretion and strong/weak jets, making appealing other options directly involvingthe black hole spin and /or the magnetic field at its horizon.To further investigate this possibility we focused on the efficiency ratio parameter ( η/(cid:15) ), that quantifiesthe capability of a source to convert gravitational energy into jet power rather than in thermal radiation. We4 G
RANDI ET AL .exploited the relations P jet = K ( f ) L . radio and L acc = 3500 × L [OIII] that, although empirical and affected byseveral uncertainties (absorbed by the f parameter) allow to directly relate jet kinetic and accretion powers toobserved luminosities. Aware of the intrinsic limitation of this approach, we compared the η/(cid:15) ratios of thedifferent classes considering two main sources of uncertainties: the particle composition of the relativisticplasma (Willott et al. 1999) and the work done by the jets on the surrounding medium (Cavagnolo et al.2010).We observe that:• FRIs and FRII-HERGs have different efficiency ratios. A similar η/(cid:15) in the two classes wouldrequire jet compositions and environment conditions not supported by the observations. In FRIsthe gravitational energy is preferentially channeled into the jets, in FRII-HERGs mainly dissipatedby thermal photons. Although our study does not include sub-relativistic/mildly relativistic matteroutflows, winds probably contribute to the total energy budget. In FRIs, jets launched by the Blandford& Znajek (Blandford & Znajek 1977) process should co-exist with winds produced by the ADAFitself. Outflows of matter are indeed theoretically predicted in the inefficient accretion regimes (seethe ADIOS model) and are also revealed in MHD simulations. In FRII-HERGs both the Blandford &Znajek (Blandford & Znajek 1977) and the Blanford & Payne (Blandford & Payne 1982) mechanismscould then be at work. The former launches jets extracting energy by the spinning black hole, the latterproduces centrifugally driving outflows of matter from a magnetized disc. The recent VLBI study ofinner jet profiles of radio galaxies (Boccardi et al. 2020) strongly supports this scenario.• The wide range of Log( L . /L Edd ) observed in radio galaxies with similar Eddington normalized[OIII] luminosities ( L [OIII] /L Edd ≤ − . ) might indicate that neither the black hole mass nor the rate ofthe mass accretion are the key parameters to explain the class segregation of LERGs. If the differenceoriginates in the nuclear engine, then the spin of the black hole and/or the magnetic field threading itshorizon are fundamental ingredients. Following Blandford & Znajek (1977) the jet propulsion couldbe less potent in FR0s than in FRIs, because the black hole spins are slower and/or the magnetic fieldis weaker. Extending this interpretation to radio sources at z > . , the high Eddington normalizedradio luminosities of FRII-LERGs (assumed as a class on its own) would imply black holes with thefastest spin and/or most intense magnetic field.• Assuming typical values of (cid:15) ∼ . for efficient disks and (cid:15) ∼ − - − for ADAFs, an average η viaequation (2) can be derived for each class. In mJy sources the fraction of gravitational power conveyedby the jets is modest, at most 10 % in HERGs (excluding winds) and a few percentages in LERGs(despite their larger η/(cid:15) ratios). FR0s are the more extreme case, with η < a few − .ACKNOWLEDGMENTSWe thank Andrea Merloni for useful discussions and the anonymous referee for his/her constructivecomments. ET acknowledges the financial contribution from the agreement ASI-INAF n. 2017-14-H.O. Thisresearch made use of the NASA/ IPAC Infrared Science Archive and Extragalactic Database (NED), whichare operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with theNational Aeronautics and Space Administration. Part of this work is based on archival data, software onlineservices provided by the ASI Space Science Data Center (SSDC).REFERENCES Baldi, R. D., Capetti, A., & Massaro, F. 2018, A&A,609, A1, doi: 10.1051/0004-6361/201731333 Baldi, R. D., Torresi, E., Migliori, G., & Balmaverde, B.2019, Galaxies, 7, 76, doi: 10.3390/galaxies7030076
ADIO G ALAXIES IN THE NEARBY M J Y U NIVERSE Balogh, M. L., Morris, S. L., Yee, H. K. C., Carlberg,R. G., & Ellingson, E. 1999, ApJ, 527, 54,doi: 10.1086/308056Best, P. N., & Heckman, T. M. 2012, MNRAS, 421,1569, doi: 10.1111/j.1365-2966.2012.20414.xBîrzan, L., McNamara, B. R., Nulsen, P. E. J., Carilli,C. L., & Wise, M. W. 2008, ApJ, 686, 859,doi: 10.1086/591416Blandford, R. D., & Begelman, M. C. 1999, MNRAS,303, L1, doi: 10.1046/j.1365-8711.1999.02358.xBlandford, R. D., & Payne, D. G. 1982, MNRAS, 199,883, doi: 10.1093/mnras/199.4.883Blandford, R. D., & Znajek, R. L. 1977, MNRAS, 179,433Boccardi, B., Perucho, M., Casadio, C., et al. 2020,arXiv e-prints, arXiv:2012.14831.https://arxiv.org/abs/2012.14831Bruzual, G., & Charlot, S. 2003, MNRAS, 344, 1000,doi: 10.1046/j.1365-8711.2003.06897.xButtiglione, S., Capetti, A., Celotti, A., et al. 2010,A&A, 509, A6, doi: 10.1051/0004-6361/200913290Capetti, A., Massaro, F., & Baldi, R. D. 2017a, A&A,598, A49, doi: 10.1051/0004-6361/201629287—. 2017b, A&A, 601, A81,doi: 10.1051/0004-6361/201630247—. 2020, A&A, 633, A161,doi: 10.1051/0004-6361/201935962Capetti, A., & Raiteri, C. M. 2015, A&A, 580, A73,doi: 10.1051/0004-6361/201525890Cavagnolo, K. W., McNamara, B. R., Nulsen, P. E. J.,et al. 2010, ApJ, 720, 1066,doi: 10.1088/0004-637X/720/2/1066Cid Fernandes, R., Stasi´nska, G., Schlickmann, M. S.,et al. 2010, MNRAS, 403, 1036,doi: 10.1111/j.1365-2966.2009.16185.xCroston, J. H., Hardcastle, M. J., & Birkinshaw, M.2005, MNRAS, 357, 279,doi: 10.1111/j.1365-2966.2005.08665.xCroston, J. H., Hardcastle, M. J., Birkinshaw, M.,Worrall, D. M., & Laing, R. A. 2008, MNRAS, 386,1709, doi: 10.1111/j.1365-2966.2008.13162.xCroston, J. H., Ineson, J., & Hardcastle, M. J. 2018,MNRAS, 476, 1614, doi: 10.1093/mnras/sty274Daly, R. A., Sprinkle, T. B., O’Dea, C. P., Kharb, P., &Baum, S. A. 2012, MNRAS, 423, 2498,doi: 10.1111/j.1365-2966.2012.21060.xFanaroff, B. L., & Riley, J. M. 1974, MNRAS, 167, 31PGarofalo, D., & Singh, C. B. 2019, ApJ, 871, 259,doi: 10.3847/1538-4357/aaf056 Gawro´nski, M. P., Marecki, A., Kunert-Bajraszewska,M., & Kus, A. J. 2006, A&A, 447, 63,doi: 10.1051/0004-6361:20053996Gendre, M. A., Best, P. N., Wall, J. V., & Ker, L. M.2013, MNRAS, 430, 3086, doi: 10.1093/mnras/stt116Ghisellini, G., & Celotti, A. 2001, A&A, 379, L1,doi: 10.1051/0004-6361:20011338Ghisellini, G., Tavecchio, F., Maraschi, L., Celotti, A.,& Sbarrato, T. 2014, Nature, 515, 376,doi: 10.1038/nature13856Heckman, T. M., & Best, P. N. 2014, ARAA, 52, 589,doi: 10.1146/annurev-astro-081913-035722Heckman, T. M., Kauffmann, G., Brinchmann, J., et al.2004, ApJ, 613, 109, doi: 10.1086/422872Huertas-Company, M., Aguerri, J. A. L., Bernardi, M.,Mei, S., & Sánchez Almeida, J. 2011, A&A, 525,A157, doi: 10.1051/0004-6361/201015735Ineson, J., Croston, J. H., Hardcastle, M. J., & Mingo, B.2017, MNRAS, 467, 1586,doi: 10.1093/mnras/stx189Jackson, N., & Rawlings, S. 1997, MNRAS, 286, 241,doi: 10.1093/mnras/286.1.241Kewley, L. J., Groves, B., Kauffmann, G., & Heckman,T. 2006, MNRAS, 372, 961,doi: 10.1111/j.1365-2966.2006.10859.xLedlow, M. J., & Owen, F. N. 1996, AJ, 112, 9,doi: 10.1086/117985Liska, M., Tchekhovskoy, A., Ingram, A., & van derKlis, M. 2019, MNRAS, 487, 550,doi: 10.1093/mnras/stz834Macconi, D., Torresi, E., Grandi, P., Boccardi, B., &Vignali, C. 2020, MNRAS, 493, 4355,doi: 10.1093/mnras/staa560Magorrian, J., Tremaine, S., Richstone, D., et al. 1998,AJ, 115, 2285, doi: 10.1086/300353Marchesini, D., Celotti, A., & Ferrarese, L. 2004,MNRAS, 351, 733Massaro, F., Álvarez-Crespo, N., Capetti, A., et al. 2019,ApJS, 240, 20, doi: 10.3847/1538-4365/aaf1c7McConnell, N. J., & Ma, C.-P. 2013, ApJ, 764, 184,doi: 10.1088/0004-637X/764/2/184McNamara, B. R., & Nulsen, P. E. J. 2007, ARA&A, 45,117, doi: 10.1146/annurev.astro.45.051806.110625Mehdipour, M., & Costantini, E. 2019, A&A, 625, A25,doi: 10.1051/0004-6361/201935205Merloni, A., & Heinz, S. 2007, MNRAS, 381, 589,doi: 10.1111/j.1365-2966.2007.12253.xMingo, B., Croston, J. H., Hardcastle, M. J., et al. 2019,MNRAS, 488, 2701, doi: 10.1093/mnras/stz1901
RANDI ET AL ..