On the BL Lacertae objects/radio quasars and the FRI/II dichotomy
aa r X i v : . [ a s t r o - ph . GA ] F e b R ECEIVED
ECEMBER
ACCEPTED
EBRUARY Preprint typeset using L A TEX style emulateapj v. 08/22/09
ON THE BL LACERTAE OBJECTS/RADIO QUASARS AND THE FRI/II DICHOTOMY Y A -D I X U , X INWU C AO , Q INGWEN W U Physics Department, Shanghai Jiao Tong University,800 Dongchuan Road, Min Hang, Shanghai, 200240, China Key Laboratory for Research in Galaxies and Cosmology, Shanghai Astronomical Observatory, Chinese Academy of Sciences, 80 Nandan Road, Shanghai200030, China and International Center for Astrophysics, Korean Astronomy and Space Science Institute, Daejeon 305348, Republic of KoreanEmail: [email protected], [email protected],[email protected].
Received 2008 December 17; accepted 2009 February 12
ABSTRACTIn the frame of unification schemes for radio-loud active galactic nuclei (AGNs), FR I radio galaxies arebelieved to be BL Lacertae (BL Lac) objects with the relativistic jet misaligned to our line of sight, and FRII radio galaxies correspond to misaligned radio quasars. The Ledlow-Owen dividing line for FR I/FR IIdichotomy in the optical absolute magnitude of host galaxy–radio luminosity ( M R – L Rad ) plane can be translatedto the line in the black hole mass–jet power ( M bh – Q jet ) plane by using two empirical relations: Q jet – L Rad and M bh – M R . We use a sample of radio quasars and BL Lac objects with measured black hole masses to explorethe relation of the jet power with black hole mass, in which the jet power is estimated from the extended radioemission. It is found that the BL Lac objects are clearly separated from radio quasars by the Ledlow & OwenFR I/II dividing line in the M bh – Q jet plane. This strongly supports the unification schemes for FR I/BL Lacobject and FR II/radio quasar. We find that the Eddington ratios L bol / L Edd of BL Lac objects are systematicallylower than those of radio quasars in the sample with a rough division at L bol / L Edd ∼ .
01, and the distributionof Eddington ratios of BL Lac objects/quasars exhibits a bimodal nature, which imply that the accretion modeof BL Lac objects may be different from that of radio quasars.
Subject headings: black hole physics —galaxies: active—galaxies: nuclei—quasars: emission lines—BL Lac-ertae objects: general INTRODUCTION
FR I radio galaxies (defined by edge-darkened radio struc-ture) have lower radio power than FR II galaxies (de-fined by edge-brightened radio structure due to compact jet-terminating hot spots) (Fanaroff & Riley 1974). Relativis-tic jets are observed in many radio-loud active galactic nu-clei (AGNs). In the frame of unification schemes of radio-loud AGNs, FR I radio galaxies are believed to be misalignedBL Lacertae (BL Lac) objects, and FR II radio galaxies cor-respond to misaligned radio quasars (see Urry & Padovani1995, for a review). Most BL Lac objects have featureless op-tical and ultraviolet continuum spectra, and only a small frac-tion of BL Lac objects show very weak broad emission lines,while quasars usually have strong broad-line emission. Thebroad emission lines of quasars are produced by distant gasclouds in broad-line regions (BLR), which are photo-ionizedby the optical/UV continua radiated from the accretion diskssurrounding massive black holes. The difference of the broad-line emission between radio-loud quasars and BL Lac objectsmay be attributed to their different central engines (e.g., Cav-aliere & D’Elia 2002; Cao 2002, 2003).The unified scheme of BL Lac objects and FR I radio galax-ies have been extensively explored by many previous authorswith different approaches, such as the comparisons of thespectral energy distributions (SEDs) in different wavebands(e.g., Owen et al. 1996; Capetti et al. 2000; Bai & Lee 2001),the radio morphology, the radio luminosity functions (LFs)(e.g., Padovani & Urry 1991; Kollgaard et al. 1992; Laurent-Muehleisen et al. 1993) and the optical line emission (e.g.,Marchã et al. 2005). Padovani & Urry (1992) derived the ra-dio LFs of flat spectrum radio quasars (FSRQs) and FR IIgalaxies from a sample of radio-loud AGNs. They consid-ered a two-component model, in which the total luminosity is the sum of an unbeamed part and a beamed jet luminos-ity. The beamed LFs of FR II radio galaxies are consistentwith the observed LFs of FSRQs and steep spectrum radioquasars (SSRQs), which strengthens the unification of FR IIgalaxies and radio quasars (see Padovani & Urry 1992, forthe details). Similar analyses were carried out on the relationbetween FR I galaxies and BL Lac objects (Padovani & Urry1991; Urry & Padovani 1995), which is also consistent withthe unification of FR Is and BL Lac objects. Even though themain observational features of different types of radio-loudAGNs can be successfully explained in the frame of the uni-fication schemes, some authors have found observations indi-cating that the unification may be more complex than usuallyportrayed in these schemes (e.g., Marchã et al. 2005; Landt& Bignall 2008). Landt & Bignall (2008) found that a con-siderable number of BL Lac objects can be identified with therelativistically beamed counterparts of FR II radio galaxies ina sample of BL Lac objects selected from the Deep X-ray Ra-dio Blazar Survey (DXRBS).Ledlow & Owen (1996) found that FR I and FR II ra-dio galaxies can be clearly divided in the host galaxy opticalluminosity–radio luminosity ( M R – L Rad ) plane, by a dividingline showing that radio power is proportional to the optical lu-minosity of the host galaxy. What causes the FR I/FR II divi-sion is still unclear, and there are two categories of models toexplain it: (1) the morphological differences being caused bythe interaction of the jets with the ambient medium of differ-ent physical properties (e.g., Gopal-Krishna & Wiita 2000);and/or (2) different intrinsic nuclear properties of accretionand jet formation processes (e.g., Baum et al. 1995; Bicknell1995; Reynolds et al. 1996; Ghisellini & Celotti 2001; March-esini et al. 2004; Hardcastle et al. 2007). Ghisellini & Celotti(2001) used the optical luminosity of the host galaxy to esti- Xu Y.D. et al.mate the central black hole masses of FR I/FR II radio galax-ies, and the bolometric luminosity is estimated from the radiopower of jets in FR I/FR II galaxies. They suggested that mostFR I radio galaxies are accreting at lower rates compared withFR IIs, which could correspond to different accretion modesin FR I and FR II radio galaxies. If the black hole is spinningrapidly, the rotational energy of the black hole is expectedto be transferred to the jets by the magnetic fields thread-ing the holes, namely, the Blandford-Znajek (BZ) mecha-nism (Blandford & Znajek 1977). The jet can also be accel-erated by the large-scale fields threading the rotating accre-tion disk (i.e., the Blandford-Payne (BP) mechanism, Bland-ford & Payne 1982). Cao & Rawlings (2004) found that theBZ mechanism for rapidly spinning black holes surroundedby advection dominated accretion flows (ADAFs) (Narayan& Yi 1995) provides insufficient power to explain the jets insome 3CR FR I radio galaxies. Wu & Cao (2008) calculatedthe maximal jet power available from ADAFs around Kerrblack holes as a function of black hole mass with an hybridjet formation model (i.e., BP+BZ mechanism). They foundthat it can roughly reproduce the dividing line of the Ledlow-Owen relation for FR I/FR II dichotomy in the black holemass–jet power ( M bh – Q jet ) plane with the mass accretion rate ˙ M ∼ . ˙ M Edd , if the black hole spin parameter a ∼ . - . H = 70 km s - Mpc - , Ω M = 0 . Ω Λ = 0 . THE SAMPLE
The host galaxies of 132 BL Lac objects have been ob-served with the
Hubble Space Telescope
WFPC2 by Urryet al. (2000), among which there are 48 sources with mea-sured redshifts and extended radio emission. We add addi-tional 18 BL Lac objects compiled in the work of Wu et al.(2008) to the Urry et al. (2000)’s sample, which leads to 66 BLLac objects (including 28 low-energy-peaked BL Lac objects(LBLs) and 38 high-energy-peaked BL Lac objects (HBLs))with measured redshifts and extended radio emission data forour present investigation. We search the literature for theemission line data of these sources, and find 44 sources in-cluding 23 LBLs and 21 HBLs. We use the luminosity of nar-row line [O II ] at 3727 to estimate the bolometric luminosity.For the sources in which the emission line data of [O II ] be-ing unavailable, we estimate the [O II ] luminosity using othernarrow emission lines.In order to compare the difference between BL Lac ob-jects and radio quasars, we need a sample of radio quasars.In this work, we adopt the sample of radio quasars compiledby Liu et al. (2006), which is selected from the 1 Jy, S4, andS5 radio source catalogs. Their sample consists of 146 ra- dio quasars including 79 FSRQs (with α - < .
5) and 67steep-spectrum radio quasars (SSRQs) (with α - > . THE BLACK HOLE MASS AND JET POWER
The relation between black hole mass M bh and host galaxyluminosity L K at K -band (Eq. 1 in McLure & Dunlop 2004)is derived from M bh – M R by using an average color correctionof R - K = 2 . M bh – M R aslog ( M bh / M ⊙ ) = - . ± . M R - . ± . , (1)to estimate the central black hole masses of BL Lac objects inthis sample. For a few BL Lac objects, their black hole massescan also be estimated from their stellar dispersion velocity σ with the empirical M bh – σ relation. It is found that the blackhole masses of three BL Lac objects estimated with M bh – σ relation are roughly consistent with those estimated with Eq.(1) (see Cao 2004, for the details, and references therein).The jet power can be estimated with the relation between jetpower and radio luminosity proposed by Willott et al. (1999), Q jet ≃ × f / L / , (W) , (2)where L ext , is the extended radio luminosity at 151 MHz inunits of 10 W Hz - sr - . Willott et al. (1999) have arguedthat the normalization is uncertain and introduced the factor f (1 ≤ f ≤
20) to account for these uncertainties. This rela-tion was proposed for FR II radio galaxies and quasars. Cao& Rawlings (2004) compared the power of the jet in M87 (atypical FR I radio galaxy) derived with different approaches,and found that Eq. (2) may probably be suitable even for FRIs (see Cao & Rawlings 2004, for the details, and referencestherein). Following Cao (2003), we adopt this relation to esti-mate the power of jets in BL Lac objects, which is believed tobe a good approximation if BL Lacs can be unified with FRIs.For most BL Lac objects, their radio/optical continuumemission is strongly beamed to us due to their relativistic jetsand small viewing angles of the jets with respect to the lineof sight (e.g., Fan & Zhang 2003; Gu et al. 2006). The low-frequency radio emission (e.g. 151 MHz) may still be Dopplerbeamed. We therefore use the extended radio emission de-tected by VLA to estimate the jet power, as adopted in Cao(2003). The observed extended radio emission is K -correctedto 151 MHz in the rest frame of the source assuming α e = 0 . f ν ∝ ν - α e ) (Cassaro et al. 1999).We take the black hole masses of radio quasars from Liuet al. (2006), which are estimated from the broad line widthsof H β , Mg II , or C IV , as well as the line luminosities of theselines (see Liu et al. 2006, for the details). In Liu et al. (2006)’swork, the jet power is estimated from the extended radio emis-sion at 151 MHz with the formula derived by Punsly (2005),which is slightly different from Eq. (2) proposed by Willott etal. (1999). To be self-consistent, we estimate the jet power ofquasars in Liu et al. (2006)’s sample from their extended radioluminosities with Eq. (2), which is the same as the estimatesof jet power for BL Lac objects in this work.For BL Lac objects, the observed optical continuum emis-sion may be dominated by the beamed synchrotron emissionfrom the relativistic jets (e.g., Gu et al. 2006). The narrow-line regions are believed to be photo-ionized by the radiationn the BL Lac objects/radio quasars and the FRI/II dichotomy 3from the accretion disk, and the narrow-line emission can beused to estimate the bolometric luminosity for BL Lac ob-jects. We convert the luminosity of the narrow-line [O II ] tobolometric luminosity using the relation proposed by Willottet al. (1999), L bol = 5 × L [OII] W , (3)for the BL Lac objects in this sample. For the objects whichlack [O II ] line emission data, we convert the luminosities ofother narrow lines ([O III ] or H α +[N II ]) to the luminosity of[O II ] using the ratios suggested by Zirbel & Baum (1995) forFR I galaxies. The narrow-line emission data for the BL Lacobjects are taken from the literature (Sbarufatti et al. 2006;Carangelo et al. 2003; Rector et al. 2000; Rector & Stocke2001; Stickel et al. 1993; Marchã et al. 1996; Morganti et al.1992). We note that Eq. (3) is derived for FR IIs/quasars,while ADAFs may be present in these BL Lac objects. TheSED of an ADAF is significantly different from that of a stan-dard thin disk (e.g., Narayan et al. 1995). Nagao et al. (2002)calculated the emission of narrow-line regions photo-ionizedby two different SED templates respectively, i.e., a standardthin disk SED template with a bump in UV/soft X-ray bandsand a hot ADAF SED template described by a power-law con-tinuum in hard X-ray bands with an exponential cutoff. Theyfound that the narrow-line regions are more efficiently photo-ionized by the ADAF SED template than the standard thindisk case (see the bottom panel of Fig. 5 in Nagao et al. 2002),which implies that the present estimates on the bolometric lu-minosity with Eq. (3) may be over-estimated to some extent(a factor of ∼ - . cm - ).For radio quasars, we estimate their bolometric luminosi-ties from the total broad-line luminosities L BLR calculated byLiu et al. (2006), as the optical continua for most radio-loudquasars may probably be contaminated by the beamed emis-sion from relativistic jets. Liu et al. (2006) derived a tightcorrelation: λ L λ (5100) = 84 . L . β , for the sample of radio-quiet AGNs in Kaspi et al. (2000). Given that the luminosityof the broad-line H β corresponds to ∼ L BLR (seeLiu et al. 2006, and references therein) and using the relation L bol ≃ λ L λ (5100) (Kaspi et al. 2000), the bolometric lumi-nosity can be estimated as L bol ≃ L BLR . THE RESULTS
The division between FR I and FR II radio galaxies isclearly shown by a line in the plane of total radio luminosityand optical luminosity of the host galaxy (Ledlow & Owen1996). The optical luminosity of the host galaxy can be con-verted to black hole mass M bh by using the empirical relation(1), while the jet power Q jet can be estimated from the radioluminosity with relation (2). Thus, the dividing line betweenFR I and II radio galaxies is translated tolog Q jet (ergs s - ) = 1 .
13 log M bh ( M ⊙ ) + . + .
50 log f , (4)in M bh – Q jet plane (see Wu & Cao 2008, for the details), whichis modified for the cosmology adopted in this paper. In Figure1, we plot the relation between the black hole masses M bh andjet power Q jet for radio quasars and BL Lac objects. It is foundthat BL Lac objects can be roughly separated from quasars bythe FR I/II dividing line.The distributions of Eddington ratios for BL Lac objectsand quasars are plotted in Fig. 2, where only the BL Lacobjects with measured line emission have been included, be-cause the bolometric luminosity is derived from the emission lines for these sources. We estimate the statistical significanceof a possible bimodal distribution of Eddington ratios for BLLac objects and quasars using the KMM algorithm (Ashmanet al. 1994). The distribution for the entire sample is stronglyinconsistent with being unimodal ( P -value < . M bh / M Sun Q j e t / f / ( e r g s s − ) F IG . 1.— The relation between black hole mass M bh and jet power Q jet forthe BL Lac objects and quasars. The open squares and filled squares representFSRQs and SSRQs respectively, while the circles and triangles represent BLLac objects. The filled circles/triangles represent the LBLs/HBLs with mea-sured line emission, while the open circles/triangles the LBLs/HBLs withoutmeasured line emission. The dashed line represents the Ledlow-Owen divid-ing line between FR I and FR II radio galaxies given by Eq. (4). −6 −5 −4 −3 −2 −1 0 1 205101520253035 log( L bol / L Edd ) N ( pe r b i n ) F IG . 2.— The distributions of Eddington ratios ( L bol / L Edd ) for BL Lacobjects (dashed line) and quasars (solid line), respectively.
Xu Y.D. et al. DISCUSSION
Figure 1 shows that the FR I/FR II dividing line givenby Ledlow & Owen (1996) roughly separates the radio-loudquasars from BL Lac objects in the M bh – Q jet plane, whichstrongly supports the FR I/BL Lac objects and FR II/radioquasars unification schemes. This conclusion is independentof the value of the uncertainty factor f in Eq. (2).We find that only a small fraction of LBLs/quasars areabove/below the dividing line, which is similar to the FR I/IIdivision (see for instance the Fig. 1 in Ledlow & Owen 1996).The HBLs have relatively lower jet power than LBLs, andonly one HBL appears above the dividing line. This meansthat the BL Lacs/quasars and the FR I/II divisions may betrue only in a statistical sense. The exceptional sources in the M bh – Q jet plane may provide useful clues to investigations onthe central engines in radio-loud AGNs (e.g., Cao & Rawlings2004; Landt & Bignall 2008).In Fig. 2, we show that the distributions of Eddingtonratios for BL Lacs and quasars in our sample exhibit a bi-modal nature. The BL Lac objects are roughly seperated fromthe quasars at L bol / L Edd ∼ .
01, with most BL Lac objectshaving L bol / L Edd . .
01 and almost all the quasars having L bol / L Edd & .
01. We suggest that this bimodal behavior ofthe distribution may imply different accretion modes in BLLac objects and quasars, and furthermore the transition be-tween the accretion states happens at L bol / L Edd ∼ .
01 ac- cording to Fig. 2. Since this is roughly the critical luminos-ity above which ADAFs are not possible (e.g., Narayan & Yi1995), this suggests that ADAFs are present in BL Lac ob-jects and standard thin disks are in quasars. We note that asimilar explanation is invoked to explain the FR I/II division,in which ADAFs would be present in FR I galaxies while stan-dard thin disks are in FR II galaxies (e.g., Ghisellini & Celotti2001; Wu & Cao 2008). Interestingly enough, Marchesini etal. (2004) found a similar bimodal distribution of Eddingtonratios for a sample of FR I and FR II radio galaxies.As discussed in §3, the bolometric luminosities of BL Lacobjects may be over-estimated, if ADAFs are present in thesesources. This would strengthen the bimodality in the distribu-tion of Eddington ratios of BL Lac objects and quasars.The similarity between the division of BL Lac ob-jects/quasars and FR I/II found in this
Letter strongly sup-ports the unification schemes for FR I/BL Lac object and FRII/radio quasar.We thank the anonymous referee for the helpful com-ments/suggestions. This work is supported by theNSFC (grants 10778621, 10703003, 10773020, 10821302and 10833002), the CAS (grant KJCX2-YW-T03), andthe National Basic Research Program of China (grant2009CB824800).strongly sup-ports the unification schemes for FR I/BL Lac object and FRII/radio quasar.We thank the anonymous referee for the helpful com-ments/suggestions. This work is supported by theNSFC (grants 10778621, 10703003, 10773020, 10821302and 10833002), the CAS (grant KJCX2-YW-T03), andthe National Basic Research Program of China (grant2009CB824800).