Warm Absorber Energetics in Broad Line Radio Galaxies
aa r X i v : . [ a s t r o - ph . H E ] A ug Mon. Not. R. Astron. Soc. , 000–000 (0000) Printed 15 November 2018 (MN L A TEX style file v1.4)
Warm Absorber Energetics in Broad Line Radio Galaxies
E. Torresi , , P. Grandi , E. Costantini , G.G.C. Palumbo Istituto di Astrofisica Spaziale e Fisica Cosmica-Bologna, INAF, via Gobetti 101, I-40129 Bologna Dipartimento di Astronomia, Universit`a di Bologna, via Ranzani 1, I-40127 Bologna, Italy SRON, Netherlands Institute for Space Research, Sorbonnelaan 2, 3584 CA Utrecht, The Netherlands
Accepted 2011 August 24. Received in original form 2011 March 7
ABSTRACT
We review the soft X–ray properties of 3C 390.3, 3C 120, 3C 382 and 3C 445, the onlyBroad Line Radio Galaxies (BLRG) for which good quality gratings data are currentlyavailable. The
XMM–Newton /RGS data of 3C 390.3 and 3C 120 were re–analyzedsearching for warm absorbers, already discovered in 3C 382 and 3C 445. We confirmthe absence of ionized absorption features in 3C 120, but find signatures of outflowinggas (v out ∼ km s − ) in 3C 390.3. Its warm absorber (Log ξ ∼ − ,N H ∼ cm − ), similar to that observed in 3C 382, is probably placed in theNarrow Line Regions. Its gas content is slower and less dense than the accretion diskwind discovered in 3C 445. Independently from the location of the warm gas, theoutflowing masses ( ˙ M out ) of BLRGs are significantly (but improbably) predominantwith respect to the accretion masses ( ˙ M acc ), suggesting a clumpy configuration of thewarm absorber. However, even assuming overestimated values of ˙ M out , the kineticluminosity of the outflow ( ˙ E out ) is well below 1% of the kinetic power of the jet (P jet ).Thus, the jet remains the major driver of the radio–loud AGN feedback at least onpc scale and beyond. The warm absorber parameters (N H , ξ ) of BLRGs span similarrange of values of type 1 radio–quiet AGNs. However, when the mass outflow rateof BLRGs and Seyfert 1s is plotted as a function of the radio–loudness, R=Log [ ν L ν (5 GHz ) /L (2 − keV ) ], the mass outflow rate seems to increase with radio power. Key words: galaxies:active – X-rays:galaxies – galaxies: general– galaxies: RadioGalaxies– techniques: spectroscopic
In the last few years high–resolution X–ray spectroscopy hasmade progress in the exploration of the circumnuclear en-vironment of radio–loud (RL) AGNs. Studies performed onobscured RL sources (Grandi et al. 2007 hereafter G07; Sam-bruna et al. 2007; Reeves et al. 2010 hereafter R10; Piconcelliet al. 2008; Torresi et al. 2009; Evans et al. 2010) revealedphotoionized emitting–line gas as responsible for the soft–excess, similarly to radio–quiet (RQ) Seyfert 2 galaxies. Ifabsorption and emission are processes occurring in the sameplasma, as suggested for Seyfert galaxies (Kinkhabwala et al.2002), it is natural to assume that warm absorbers (WA),characterizing at least 50% of Seyfert 1 spectra, shouldbe present also in unobscured Broad Line Radio Galaxies(BLRG). With the term warm absorber we intend ionizedoutflowing gas in our line–of–sight that produces narrow ab-sorption lines of several elements from C to Fe in the softX–ray spectrum. Generally, these structures are blueshiftedwith moderate velocities, v ∼ − (Kaastra etal. 2000; Kaspi et al. 2002; Crenshaw et al. 2003), but can reach also v > km s − (Pounds et al. 2003; Reeves et al.2003, 2010; Braito et al. 2011) when the wind originates di-rectly from the disk. The detection of WAs in BLRGs wasexpected to be more difficult than in Seyfert 1s, becauseof the jet. If it is closer to the line–of–sight, the Doppler–boosted, non–thermal radiation could mask the absorptionfeatures.Hints of WAs have been observed in the past in a handfulof RL sources: 3C 382 and 3C 390.3 with ASCA (Reynolds1997); 3C 351 with
ROSAT
PSPC (Nicastro et al. 1999);4C+74.26 with the
XMM–Newton
EPIC cameras (Ballan-tyne 2005).The presence of absorption features in RL AGNs was con-firmed by the high–resolution X–ray analysis of 3C 382 (Tor-resi et al. 2010; Reeves et al. 2009). Successively R10 re-ported the presence of a warm absorber also in 3C 445. Sincethe gas has non–negligible outflowing velocities, it could con-tribute in transferring momentum to the environment in ad-dition to the jet. However to what extent is the wind impor-tance in the energetic budget of powerful RL AGNs is still c (cid:13) E.Torresi et al. an open question. In order to shed some light on this issue,we collect high–resolution data from literature and explorethe
XMM–Newton
Reflection Grating Spectrometer (RGS)archive. The main goal is to enlarge the sample of RL sourceswith clear detection of warm absorbers. Finally, we attempta RL/RQ comparison of the ionization and kinematic prop-erties of the X–ray absorbing gas.The paper is organized as follows: in Section 2 the very smallsample of BLRGs considered throughout the work is de-scribed; the RGS spectral analysis performed on 3C 390.3and 3C 120 together with the results are reported in Sec-tion 3. In Section 4 we discuss the physical and energeti-cal properties of BLRG WAs and attempt a first compari-son between RL and RQ outflows. The main results of thiswork are summarized in Section 5. The cosmological valuesH =71 km s − Mpc − , Ω m =0.27, Ω Λ =0.73 (Komatsu et al.2009) are assumed throughout. The sample of BLRGs consists of four sources: 3C 390.3,3C 120, 3C 382 and 3C 445. The main properties of eachobject are reported in Table 1, where the redshifts, the jetinclination angles, the black hole masses, the ionizing lumi-nosities between 1—1000 Ry, and the radio luminosities at151 MHz are listed.3C 390.3 (z=0.0561; Hewitt & Burbidge 1991) is a clas-sical double–lobed FRII Radio Galaxy (Pearson & Read-head 1988). It is one of the closest radio sources whose coreexhibits superluminal motion in the pc–scale jet (Alef etal. 1996). From the apparent velocity of 3.5c and the coredominance, Giovannini et al. (2001) estimated the jet incli-nation angle 30 ◦ < θ < ◦ with β ∼ ⋆ . Double–peaked emission lines characterize the optical and UV spec-tra of the source (Eracleous & Halpern 1994; Zheng 1996;Wamsteker et al. 1997), while the UV bump is weak or evenabsent (Wamsteker et al. 1997). 3C 390.3 is known to bevariable in the X–ray band, with variations in both soft andhard band on a timescale of weeks to months (Leighly &O’Brien 1997; Gliozzi et al. 2003). All previous X–ray tele-scopes observed this source. While in the hard X–ray bandthere is a general agreement on the presence of an iron lineand reflection hump (Grandi et al. 1999; Sambruna et al.2009), the modeling of the soft X–ray band is controversial. Einstein–
IPC (Kruper et al. 1990) and
BeppoSAX (Grandiet al. 1999) required an excess of column density.
EXOSAT claimed the presence of a soft excess (Ghosh & Sondarara-japerumal 1991) while Reynolds (1997) found hints of warmabsorption in the
ASCA data, successively confirmed by thedetection of an absorption edge at 0.65 keV (Sambruna et al.1999). Recently Sambruna et al. (2009) observed an emissionline in the RGS spectrum associated with O
VII forbiddenline possibly produced in the NLR.3C 120 (z=0.033; Burbidge 1967) is classified as an FRIexhibiting a one–sided jet (Seielstad et al. 1979; Walker et al.1987; Harris et al. 2004). The apparent transverse velocity ⋆ β =v/c is the bulk velocity in units of the speed of light (Urry& Padovani 1995). of the jet v app =5.3c, as obtained by the VLBA observation(Lister et al. 2009), implies an upper limit on the inclina-tion angle of 21 ◦ . The optical spectrum of 3C 120 is typicalof Seyfert 1 galaxies, with strong and broad emission lines,quite unusual for FRI radio sources. Reverberation mappingconstrains the black hole mass to be 5.5 +3 . − . × M ⊙ (Pe-terson et al. 2004). At UV wavelengths, 3C 120 has a typicalAGN spectrum with a strong blue bump and strong emissionlines signature of a standard optically–thick geometrically–thin accretion disk (Maraschi et al. 1991). The hard X–rayspectrum is characterized by a slightly broadened iron line(EW ∼
100 eV) at 6.4 keV, a weak ionized line at 6.9 keV(Yaqoob & Padmanabhan 2004; Kataoka et al. 2007) andCompton reflection Ω/2 π ∼ Chandra and
XMM–Newton . In the
Chandra
High Energy Transmission Gratings(HETG) spectrum an O
VIII Ly α absorption line, blueshiftedby ∼ -5500 km s − , was observed (McKernan et al. 2003).This feature was not revealed in the XMM–Newton /RGSobservation (Ogle et al. 2005), that on the contrary shows aslightly redshifted O
VIII Ly α emission structure.3C 382 (z=0.0579; Marzke et al. 1996) is a FRII lobe-dominated Radio Galaxy showing a long jet (1.68’ from thecore) and two radio lobes, with a total extension of 3’ (Blacket al. 1992). In the optical–UV and X–ray regimes there arehints of no strong jet contamination. The optical spectrumshows broad lines (FWZI > − ) which are variableon a timescale of months to years. Yee and Oke (1981) sug-gested the presence of the UV bump from the accretion disk.In the X–ray band 3C 382 is a bright source (F − keV ∼ × − erg cm − s − ). It can be well fitted with a singlepower law between 2–10 keV, but shows a strong excess atlower energies (Prieto 2000; Grandi et al. 2001). In part thisis probably related to an extended emission (0.2–2.4 keV) re-vealed by ROSAT/HRI and
Chandra (Prieto 2000; Gliozziet al. 2007). The presence of a slow highly ionized outflow inthis source was attested by Torresi et al. (2010) using
XMM–Newton /RGS data, and confirmed by the
Chandra /HETGspectrum (Reeves et al. 2009).3C 445 (z=0.05623) is a powerful FRII Radio Galaxy,classified as a BLRG because of its broad and intense Balmerlines in the optical spectrum. Near–IR observations show asubstantial reddening of E B − V =1 mag and the radio–to–IRSED indicates a predominance of dust emission with a negli-gible contribution of synchrotron photons. The existence ofabsorbing material is also supported by X–ray data show-ing a strong depletion of the continuum photons below afew keV. R10 suggests that the nuclear view could be ob-scured by an outflowing and clumpy accretion disk windwith high column density (N H ∼ cm − ) and velocity(v=10 km s − ). In spite of its optical classification, 3C 445seems more similar to Seyfert 2s than Seyfert 1s (G07; Sam-bruna et al. 2007). In agreement with that, the soft X–rayexcess can be fitted with a mix of emission lines and scat-tered continuum produced by photoionized gas (G07; R10).Indeed, the presence of the WA in this source was deducedby the detection of a high energy absorption feature around7 keV.Since WAs have been already ascertained in 3C 382 and c (cid:13) , 000–000 arm Absorber Energetics in Broad Line Radio Galaxies Table 1.
Summary of the BLRG properties.z i LogM
BH a
LogL ionb
LogL
MHzc (degrees) (M ⊙ ) (erg s − ) (W Hz − sr − )
3C 445 ≤ d
3C 390.3 e
3C 382 e
3C 120 ≤ f ion directly measured from the proper SED of each source, except for 3C 445taken from R10. L ion is the ionizing luminosity between 1–1000 Ry( ≡ MHz extrapolated from the power at 178 MHz of Hardcastleet al. (1998) and rescaled to our ΛCDM cosmology. For 3C 120 werefer to Arshakian et al. (2010).(d) Estimate from the radio jet/counterjet ratio (G07).(e) Estimate from the radio band (Giovannini et al. 2001).(f) Upper limit on the inclination angle obtained by using the larger apparent transversevelocity v=5.3c (Gomez et al. 2001).
3C 445, here we re–analyze the RGS data of 3C 390.3 and3C 120.
3C 390.3 was observed by
XMM–Newton /RGS (den Herderet al. 2001) on 2004 October 8–9 for a total exposure of 50ks and on October 17 for 20 ks.3C 120 was pointed twice, on 2002 September 6 for 12 ksand on 2003 August 26 for 130 ks. In this paper we consideronly the second and longer observation.The RGS1 and RGS2 spectra were extracted using the
SAS (v. 9.0.0) task rgsproc , which combines the event listsfrom all
RGS CCD , produces source and background spec-tra using a region spatially offset from that containing thesource, and generates response matrices. The resulting spec-tra were analyzed using the fitting package
SPEX (v.2.0)(Kaastra, Mewe and Nieuwenhuijzen 1996). The Galac-tic absorption was modeled with the
SPEX HOT compo-nent. For all spectral models Solar elemental abundanceswere adopted (Anders & Grevesse 1989). For each sourcethe proper line–of–sight Galactic column density was con-sidered, N H =3.5 × cm − and N H =1.1 × cm − for3C 390.3 and 3C 120, respectively (Kalberla et al. 2005).The absorption/emission features were searched for follow-ing two steps:(i) a phenomenological approach, consisting in the inspec-tion of the (data–model/error) residuals after having fittedthe continuum;(ii) a physical approach, fitting the absorption featureswith the xabs model in SPEX . xabs calculates the transmis-sion through a gas layer. Free parameters in this model arethe outflow velocity ( v out ), the total hydrogen column den- sity ( N H ) and the ionization parameter ξ † . In the modelthe column densities of different ions are linked through anionization balance, which is precalculated using CLOUDY (Ferland et al. 1998). The ionization balance is dependenton the spectral energy distribution (SED) of the source. TheUV/X–ray SEDs for both 3C 120 and 3C 390.3 were con-structed using both the EPIC–pn (Str¨uder et al. 2001) andthe optical monitor (OM; Mason et al. 2001) and are shownin Fig. 1. For 3C 390.3 we considered only the longest ob-servation performed on Oct. 8. The data were reduced usingthe
SAS (v. 9.0.0) with standard procedures. The light curveover 10 keV was extracted to check high background peri-ods. The source and the background spectra were extractedfrom circular regions of 35 ′′ radius. Backgrounds were takenfrom a region within the same CCD of the targets and notcontaminated by the sources. The response matrices werecreated using the
SAS commands
RMFGEN and
ARFGEN .Events outside the 0.4–10 keV band were discarded in thepn spectra of both sources (Guainazzi 2010). We produceda rough representation of the continuum using a simple ab-sorbed power law for 3C 390.3 (Γ ∼ = 2 .
04, Γ = 1 .
75 with a break at 2.5 keV) for3C 120. For the purpose of constructing the SED, both con-tinua were then unabsorbed, to obtain the true ionizing X–ray flux. For the optical/UV part of the SED we used theOM measurements for both 3C 120 (V and UVW1 filters)and 3C 390.3 (U, UVW1, UVM2 and UVW2 filters). Thephotometric analysis was performed using the standard pro-cedure within the
XMM–Newton
SAS (v. 9.0.0). The opticalfluxes have been dereddened using the extinction curves ofCardelli et al. (1989), knowing the optical extinction A V . † ξ = Ln e R , L is the 1-1000 Rydberg (Ry) source ionizing lumi-nosity (corresponding to 13.6 eV–13.6 keV), n e is the electrondensity of the gas and R is the distance of the gas from the cen-tral source.c (cid:13) , 000–000 E.Torresi et al.
Figure 1.
Spectral energy distributions (SED) of 3C 390.3 and3C 120, constructed from the
XMM–Newton data in the optical–UV–to–X–ray band, and from the standard AGN radio–IR con-tinuum included in CLOUDY.
This has been calculated following Bohlin et al. (1978) for-mula: N H ∼ A V ∗ . × cm − (for R V = 3 . CLOUDY (Mathews &Ferland 1987).
At first the continuum of both observations was mod-eled with a power law (Γ ∼ .
9) plus a neutral absorber( N H ∼ . × cm − ) in addition to the Galactic one(C/d.o.f.=583/406). Although during the second pointing(Oct. 17) the source flux was about 14% lower, the spectralparameters are completely consistent. A careful inspectionof the residuals revealed photon deficits in the regions ofNe X (12.134 ˚A ), Fe XX (12.864 ˚A ), O VIII Ly α (18.969 ˚A )and N VI (24.898 ˚A ). An example of absorbed structures isshown in Fig. 2. In order to confirm the WA detection andconstrain physical properties of the WA, the xabs componentwas added to the continuum model. The column density ofthe ionized absorber N H , the ionization parameter ξ andthe velocity of the gas were let free to vary. The velocitydispersion between different blend components is fixed tothe default value v=100 km s − . The covering fraction pa-rameter (f cov ) is fixed to the default value equal to 1. Thefit improves with the addition of the warm absorber withrespect to the power law alone, i.e. ∆C=18 and 11 for a de-crease by 3 in the number of degrees of freedom, correspond- Figure 2.
3C 390.3 residuals in a zoomed region around NeX,FeXX lines (observed–frame) after fitting the first observation(Oct. 8) with a power law plus a neutral absorber in addition tothe Galactic one.
Figure 3.
SPEX best–fit modeling of the RGS spectrum of3C 390.3. The most prominent absorption lines are labelled. ing to a significance > ∼ ξ ≈ − ) and column density( N H ≈ . × cm − ) seems to better reproduce theshape of the most prominent features. However, the statisti-cal improvement of the fit is not significant. Finally, we notepositive residuals around 23.5 ˚A (observed–frame) in bothdata sets (Fig. 4). In agreement with Sambruna et al. (2009),a narrow gaussian component at the theoretical wavelengthof the O VII forbidden line (Table 3) is a good parametriza-tion of this emission feature.
The soft X–ray continuum of 3C 120 is very complex. AnRGS fit with a simple power law absorbed by two neu-tral absorbers is not a satisfying representation of the data(C/d.o.f.=803/328). The spectral shape of the continuumseems to be curved, as suggested by negative residuals c (cid:13) , 000–000 arm Absorber Energetics in Broad Line Radio Galaxies Table 2.
Best–fitting parameters for both 3C 390.3 observations. The observation date, photon index, source rest–frame neutral absorbercolumn density (N H ), column density of the ionized absorber (N H ), ionization parameter (Log ξ ), outflow velocity and ∆C after theaddition of the xabs model to the absorbed power law are reported.Obs. Γ N H N H Log ξ v ∆C(10 cm − ) (10 cm − ) (erg cm s − ) (km s − )Oct. 8 1.89 ± ± +2 . − . +0 . − . <
600 18Oct. 17 1.96 ± ± +2 . − . +0 . − . < Table 3.
3C 390.3 OVII(f) emission line parameters for the two epochs. The wavelength, flux and FWHM together with ∆C after theaddition of the line to the absorbed power law are reported.Obs. Line λ Flux FWHM ∆C(˚A ) (10 − ph cm − s − ) (˚A)Oct. 8 OVII(f) 22.101 0.56 +0 . − . +0 . − . +0 . − . +0 . − . Figure 4.
3C 390.3 residuals for the OVII forbidden line. around 23–26 ˚A that is the region where O I edge is expected.We exclude that this spectral bending can be attributedto a warm absorber. Indeed the xabs model is not statisti-cally required by the data and left the residuals invariant.Following Ogle et al. (2005), the oxygen abundance of thesecond absorber was allowed to vary. The fit greatly im-proves (C/d.o.f.=562/327) and the oxygen abundance dropsto A ( O ) =0.53 ± XVII and O
VIII Ly α (see Table 4), providing a furtherimprovement of the fit (C/d.o.f.=449/323). To ascertain thenature of the emitting–line gas we tested both collisional (i)and photoionized (ii) scenarios.(i) A single temperature collisional model, CIE (CollisionalIonization Equilibrium) in
SPEX , was applied instead of twosingle lines. A thermal component with kT=0.37 +0 . − . keVcan not completely take into account the emission features,infact residuals around the Fe XVII are still present. A second
CIE component gives a poor fit C/d.o.f.=520/323, while a better modeling is provided by a non–equilibrium ionizationjump model (
NEIJ ) in
SPEX (C/d.o.f.=498/323). However,even in this case, there are still positive residuals.(ii) If the emitting–line gas has a photoionized origin, theemission line at 16.892 ˚A could be radiative recombinationcontinuum (RRC) from O
VII slightly redshifted with re-spect to the rest–frame of the galaxy (v ∼ +2000 km s − ).We tested this hypothesis fitting this feature with the RRC model in
SPEX . Since RRCs are narrow and prominentstructures in photoionized plasmas, we assumed a typicalelectron temperature of kT=3 eV. We fit the RRC emissionmeasure of O
VII as a free parameter, however the fit is stillnot satisfying (C/d.o.f.=517/324).We conclude that our proposed models can not reproducethe data better than a single temperature gas plus a gaus-sian line, C/d.o.f.=440/323 (see Table 4), leaving the exactnature of the emitting–line gas still uncertain. However weexclude that one, or two, plasma components in collisionalionization equilibrium, or even in non–equilibrium, are suf-ficient to reproduce the soft–excess in 3C 120.
Table 5 lists, for each source, column density, ionization pa-rameter, outflow velocity and minimum and maximum radiiderived for the WAs of the studied BLRGs. For 3C 390.3 weconsider the parameters obtained from the Oct. 8 observa-tion for which the fit improvement due to the inclusion ofa warm absorber is more significant. For 3C 445, the softemitting gas properties, derived by R10, are also listed. Inorder to establish the location of the emitting/absorbing re-gion, the measured distances of the broad line region (BLR)and the torus are also reported.The minimum distance of the WA (r min ) from the centralengine is measured from: c (cid:13) , 000–000 E.Torresi et al.
Table 4.
3C 120 fit parameters for the two models tested. For Model 1 C/d.o.f.=449/323, for Model 2 C/d.o.f.=440/323.Model 1 Model 2Γ 2.32 ± ± aH +0 . − . ± ( O ) ± ± λ (˚A ) 16.892 16.892Flux b +0 . − . +0 . − . FWHM(˚A ) 0.27 +0 . − . +0 . − . λ (˚A ) 18.969 –Flux a +0 . − . –FWHM(˚A ) 0.13 +0 . − –kT(keV) – 0.37 +0 . − . (a) × cm − .(b) in 10 − ph cm − s − . Figure 5.
3C 120 residuals after fitting the RGS data with apower law plus two neutral absorbers, with the oxygen abundanceof the second absorber free to vary. r min ≥ GMv out (1)where M is the black hole mass and assuming that the out-flow must have a speed (v out ) greater than or equal to theescape velocity.The maximum distance (r max ) can be estimated assumingthat most of the mass of the absorber is concentrated ina thin layer ∆R where ∆R/R ≤
1. The column density is afunction of the density of the material n(R) at ionizationparameter ξ , of its volume filling factor (here assumed equalto 1) and of ∆R: N H ∼ n ( R )∆ RC v (2)This combined with the expression of the ionization param-eter gives∆ RR ∼ ξRN H L ion (3)Therefore if ∆R/R ≤ r max ≤ L ion ξN H (4)The BLR and torus radii are calculated following the pre-scriptions of Ghisellini & Tavecchio (2008), that simply as- sume that typical distances scale as the square root of theionizing disk luminosity: r BLR = 10 L / disk, cm (5) r torus = 2 . × L / disk, cm (6)Table 5 suggests two immediate considerations:(i) depending on the line–of–sight, different features arerevealed (see also Table 1). The absorption lines due tophotoionization and photoexcitation processes are prefer-entially observed in sources seen at small i (i.e. 3C 382and 3C 390.3). On the contrary, the emission lines producedby the inverse processes are dominant in 3C 445, the onlysource with Seyfert 2 characteristics (and presumably withthe larger jet inclination angle). We note that no WA couldbe detected in 3C 120, the BLRG in the sample with thesmallest inclination angle ( i < ◦ ) and the only one with a γ –ray counterpart (Abdo et al. 2010b). However this sourceis quite complex. Indeed the detection of X–ray emissionlines is at odds with the idea that the jet dominates the softX–ray emission;(ii) the location of the ionized outflow is not unique.In 3C 445 the WA was suggested by the detection ofa strong edge around 7 keV. The deduced high velocity(v out ∼ km s − ), high column density (N H > cm − )and low ionization parameter (Log ξ =1.4 erg cm s − ) of thewind indicate a probable origin in the disk (R10). On thecontrary, in 3C 382 and 3C 390.3, the absorption features,signatures of an ionized outflow, were found in the soft partof the grating spectra. The gas has different physical pa-rameters and probably a different origin. Indeed the col-umn densities and ionization parameters vary in the rangeN H =10 − cm − and log ξ =2–3 erg cm s − . Moreover theslower velocities, v out ∼ − km s − , constrain the loca-tion of such gas between the torus and the NLR, favoringthe torus wind scenario (Krolik & Kriss 2001; Blustin et al.2005 hereafter B05). c (cid:13) , 000–000 arm Absorber Energetics in Broad Line Radio Galaxies Warm absorber energetics are summarized in Table 6. Themass accretion rate ˙ M acc (L bol ≈ L acc = η ˙ M c ) was calculatedfor η =0.1.The mass outflow rate estimates how much mass is carriedout of the AGN through the wind, and can be expressed as(B05):˙ M out ∼ . m p L ion v out C v Ω ξ (7)We set the solid angle of the outflow Ω=2.1, using the in-formation that ∼
33% of the Radio Galaxies belonging tothe 3CR sample with z ≤ v , be-ing unknown, was kept equal to 1, equivalent to assume anupper limit on ˙ M out .The kinetic energy is the power released in the circumnu-clear environment through the outflow and is expressed by:˙ E out = ˙ M out v out out is the blueshift velocity measured for the warmabsorber.P jet is calculated according to the formula of Shankaret al. (2008) adapted from Willott et al. (1999): P jet = 3 × f / L / erg s − (9)L is the observed radio luminosity in units of10 W Hz − sr − at 151 MHz. The factor f accounts forsystematic underestimates of the true jet power. The averagevalue < f > =15 (Hardcastle et al. 2007) supports the pictureof “ heavy ” jets with a dominant protonic component.Looking at Table 6, it appears immediately evident that thekinetic luminosity related to slow outflows is always a neg-ligible fraction ( << jet can also be expressed in terms of accre-tion energy: P jet = η jet ˙ M c (we do not take into account thewind, as its energetic contribution is not important). Then,the ratio between P jet and L bol directly expresses η jet /η .This efficiency ratio is generally <
1, suggesting that ac-cretion power could be preferentially channeled in radiationrather than in jet kinetic power (at least in these sources).Assuming η equal to 0.1 (typical value of standard accretiondisks), η jet ranges between 0.01–0.06. However if accretiondisk winds are energetically important (see Section 4.3) thekinetic (jet + wind) power could compete with the radiativepower. Here we compare the X–ray properties of BLRGs studiedin this work with a sample of type 1 RQ AGNs (Seyfert1s, NLS1s, QSOs) having a good modeling of the WA. Wechose the sample already studied by B05 with the addition ofthe Seyfert 1.2 IC4329a taken from McKernan et al. (2007). Among the sources exhibiting signatures of warm absorp-tion as reported by McKernan et al., IC4329a is the onlyone not in common with B05. Very recently, a possible de-tection of a fast outflow in MR2251–178 has been proposedby Gofford et al. (2011) on the basis of
Suzaku data. Forthis source, belonging to the B05 sample, we however preferto consider only the WA physical parameters derived by the
XMM–Newton /RGS analysis (Kaspi et al. 2004, B05). Forthe sake of consistency, when a source has a multi–phaseoutflow, only the higher phase, similar to BLRGs (Log ξ =2-2.9 erg cm s − ), is considered. It is important to note that,in this case, the estimated mass outflow rates refer only tothe higher phases, neglecting the contribute of mass outflowrates related to lower ionization parameters.However, even considering the averaged warm absorber pa-rameters the final result does not change. The mass outflowrates and the related kinetic powers of RQ objects in Table 4of B05 are rescaled to a volume filling factor equal to 1 in or-der to match our assumption. We keep the small (and of noconsequence) difference between the solid angle of BLRGs(Ω=2.1) and Seyferts/QSOs (Ω=1.6), both estimated on thebasis of similar considerations. Fixing C v =1, the RQ massoutflow rates are obviously shifted to larger values.As shown in Fig. 6 ( upper panel ), all the sources have large(and implausible) mass outflow rates exceeding the massaccretion rates even more than one order of magnitude im-plying, also for RL sources, a non uniform distribution ofthe photoionized gas. We can deduce a BLRG volume fill-ing factor as small as ∼ M out ∼ ˙ M acc with η =0.1).In both RL and RQ AGNs, the kinetic energy associatedto the slow winds is negligible with respect to the radia-tive luminosity, even considering a uniform gas distribu-tion (C v =1) (Fig. 6 lower panel ). There are only threesources, PG 1211+143, PG 0844+349 and 3C 445, for which˙ E out ≥ L bol ( ≈ L acc ). In these objects the winds have ve-locities of several thousands km s − , in spite of the pres-ence/absence of a powerful jet, and are probably directlyconnected to the accretion disk (Pounds et al. 2003a,b).These disk winds are extremely energetic, unless they havesmall covering factors and/or are transient phenomena. In-deed, in BLRGs persistent disk outflows covering large solidangles could even compete with the jets in trasferring mo-mentum to the circumnuclear environment.In order to further investigate the role of the relativisticjet, we explore a possible correlation between the mass out-flow rate and the radio–loudness parameter (R). We adopta radio–loudness: R = Log [ νL ν (5 GHz ) L (2 − keV ) ] (10)as proposed by Terashima & Wilson (2003). For RQ AGNs,we use the 2–10 keV X–ray luminosities as measured byBianchi et al. (2009), while we refer to Torresi et al. (2010)and to this paper for the X–ray luminosities of 3C 382 and3C 390.3, respectively.The 5 GHz luminosities of RL and RQ sources are takenfrom Kellermann et al. (1969) and from Ulvestad et al.(1984), respectively. For RQ objects, when the luminosity c (cid:13) , 000–000 E.Torresi et al. at 5 GHz is not available, we extrapolate it from the 1.4GHz luminosity as reported in the NRAO/VLA Sky Surveycatalogue (NVSS; Condon et al. 1998) by assuming α =0(Nagar, Wilson & Falcke 2001). We notice that differentassumptions on the radio spectral slope do not affect theradio–loudness estimation. Indeed even using α =0.8, giventhe dispersion of spectral indexes for AGNs (Kukula et al.1998), the estimated R values are completely consistent withthose obtained for α =0.We consider BLRGs and type 1 AGNs having slow outflows(v out =10 − km s − ) and similar phases. Disk winds are nottaken into account here because we have only three sources.In Fig. 7 the mass outflow rates of both RL and RQ AGNsare plotted as a function of the radio–loudness. This plotsuggests a possible difference between the two classes. In-deed the average ˙ M out value of RQ objects is ∼ ⊙ yr − ,much lower than that of BLRGs that is around 25 M ⊙ yr − .When the Spearman ρ test and the generalized Kendall τ test in the Astronomy Survival Analysis ( ASURV ) pack-age (Feigelson et al. 1985; Isobe et al. 1986; Lavalley et al.1992) are applied to data, a possible positive correlation be-tween ˙ M out and R is suggested. The resulting significanceis α ASURVρ =0.05 and α ASURVτ =0.02, respectively. The cor-relation is assumed to be significant if equal or below 0.05.This trend could simply reflect differences in the gas distri-bution, tending to preferentially clump when the system isless perturbed by the jet, given the dependence of ˙ M out onthe volume filling factor. Alternatively, if the geometry of thegas is similar in both RQ and RL objects, the observed cor-relation would indicate that larger amount of mass escapesfrom the nuclear region when the jet production mechanismis more efficient. We notice that the ionization parameter,the outflow velocity and the ionizing luminosity are not cor-related with R (significance of correlation 0.4, 0.8 and 0.2,respectively, according to the generalized Kendall τ test). In this work we report the detection of a warm absorber inthe BLRG 3C 390.3. This is the third outflow observed after3C 382 and 3C 445. On the contrary 3C 120, the BLRG withthe smallest jet inclination angle ( i ≤ ◦ ) shows a complexsoft X–ray spectrum only modified by a structured Galacticcold gas, without signatures of warm ionized absorption.We discuss the physical and energetical properties of thewarm absorbers found in three BLRGs, 3C 390.3, 3C 382and 3C 445: • depending on the outflow velocities the absorbing gashas different locations: the NLR/torus for slower outflows(3C 390.3 and 3C 382), and the accretion disk for 3C 445.Interestingly, these winds have been detected in differentspectral regions, i.e. the NLR/torus winds in the soft X–rayband (0.3–2 keV), while the accretion disk wind above 6 keV(R10); • the mass outflow rate ( ˙ M out ) of the absorbers is higherthan the mass accretion rate ( ˙ M acc ), implying that we areoverestimating the volume filling factor C v , here assumedequal to 1 but probably much less than 1; • even considering upper limits on ˙ M out , the kinetic lu-minosity associated with the slow outflows is always lowerthan the accretion luminosity and the jet kinetic power. Figure 6.
Upper panel : mass outflow rate ( ˙ M out ) plotted againstthe mass accretion rate ( ˙ M acc ). Lower panel : kinetic luminosityassociated to the outflow ( ˙ E out ) plotted against the accretionluminosity ( L acc ). Type 1 RL AGNs ( red circles ) are the BLRGsconsidered in this work; type 1 RQ AGNs ( black squares ) refer tothe Seyferts, NLS1s, and QSOs belonging to our reference sample. Aware of the scarcity of RL sources with WAs, we attempta first comparison with a sample of type 1 RQ AGNs: • fixing C v =1, the mass outflow rates have implausiblylarge values in both RL and RQ AGNs, suggesting a clumpygas configuration in all AGNs independently of their radiopower; • in both Seyferts/QSOs and BLRGs, the kinetic lumi-nosity related to slow outflows is always negligible with re-spect to the accretion luminosity (and the jet kinetic powerfor RL AGNs); • fast accretion disk winds are observed in AGNs, inde-pendently of the RL or RQ classification (Fig. 8). The asso-ciated kinetic energies appear to be huge ( ˙ E out ≥ L acc ) (un-less these fast outflows have small covering factors and/orare transient phenomena); • although the RL and RQ WA properties are very simi-lar (at least at zeroth order), the mass outflow rate ( ˙ M out )and the radio–loudness parameter (R) seem to be correlated(Fig. 7) indicating a possible effect of a strong radio sourceon the outflowing winds. Considering that ˙ M out depends onthe covering factor, this result could simply indicate a dif- c (cid:13) , 000–000 arm Absorber Energetics in Broad Line Radio Galaxies Table 5.
For each source we report the warm absorber/emitter parameters (LogN H , Log ξ , v out (“+” means the gas is redshifted, “-”is blueshifted)), together with the estimated minimum (r min ) and maximum (r max ) distances of the WA in pc, the distance of the BLR(r BLR ) and of the torus (r torus ) from the central engine, respectively.Log N H Log ξ v out r min r max r BLR r torus (cm − ) (erg cm s − ) (km s − ) (pc) (pc) (pc) (pc)
3C 445 (em) 22.0 3; 1.8 +150 a ;-430 b - < c ; ∼ d
3C 445 (wa) 23.3 1.4 -10
3C 390.3 (wa) 20.7 2.08 < -600 ≥ ≤
450 0.03 0.7
3C 382 (wa) 22.5 2.69 -1000 ≥ ≤
44 0.05 1.5
3C 120 - - - - - 0.03 0.8(a) Redshift velocity (compared to the systemic) of the emitter (R10).(b) Outflow velocity determined by G07.(c) Upper limit estimated from the definition of the ionization parameter of the emitting gasR =L ion / ξ n e (R10).(d) Estimate from the measured line widths of the OVII–OVIII emission (R10). Table 6.
For each BLRG we give the mass outflow rate ( ˙ M out ), the kinetic luminosity of the outflow ( ˙ E out ), the bolometric luminosity(L bol ), the mass accretion rate ( ˙ M acc ) assuming η =0.1, the kinetic power of the jet (P jet ) and the jet extraction efficiency η j =(P jet /L bol ) η . Log ˙ M out Log ˙ E out Log L abol ˙ M acc Log P jet η j (M ⊙ yr − ) (erg s − ) (erg s − ) (M ⊙ yr − ) (erg s − )
3C 445 v b
3C 390.3 v
3C 382 v
3C 120 - - 45.34 0.37 44.71 0.02(a) L bol directly estimated from the total SED of each source, except for 3C 445taken from Marchesini et al. (2004).(b) The value from R10 is rescaled assuming η =0.1. ferent gas distribution in RL and RQ sources, with the WAbeing clumpier in absence of a strong radio source. Alter-natively, if the gas distribution is the same, the correlationcould suggest that powerful jets favor the escape of moremassive (but not necessarily more energetic) winds. ACKNOWLEDGMENTS
The anonymous referee is gratefully acknowledged forthoughtful comments on the manuscript. We warmly thankJelle Kaastra for accurate and constructive comments andsuggestions. We thank Kazushi Iwasawa for valuable anduseful discussions. ET is grateful to Adriano de Rosa forproviding constant technical assistance.
XMM–Newton is anESA science mission with instruments and contributions di-rectly funded by ESA Member States and NASA. ET ac-knowledges the support of the Italian Space Agency (con-tract ASI/INAF I/009/10/0 and ASI/GLAST I/017/07/0).
REFERENCES
Abdo, A.A., et al. 2010b, ApJ, 720, 912Alef, W., Wu, S. Y., Preuss, E., Kellermann, K. I., Qiu, Y. H.1996, A&A, 308, 376Anders, E., Grevesse, N. 1989, GeCoA, 53, 197Arshakian, T.G., Torrealba, J., Chavushyan, V.H., Ros, E., et al.2010, A&A, 520, A62Ballantyne, D.R., Fabian, A.C., Iwasawa, K. 2004, MNRAS, 354,839Ballantyne, D.R. 2005, MNRAS, 362, 1183Balmaverde, B., Capetti, A., Grandi, P. 2006, A&A, 451, 35Belsole, E., Worrall, D.M., Hardcastle, M.J., 2006, MNRAS, 366,339Bianchi, S., Guainazzi, M., Matt, G., Fonseca Bonilla, N., Ponti,G. 2009, A&A, 495, 421Black, A.R.S., Baum, S.A., Leahy, J.P., Perley, R.A., Riley, J.M.,Scheuer, P.A.G. 1992, MNRAS, 256, 186Blustin, A.J., Branduardi-Raymont, G., Behar, E., et al. 2002,A&A, 392, 453Blustin, A.J., Page, M.J., Fuerst, S.V., et al. 2005, A&A, 431, 111(B05)Blustin, A.J., Fabian, A.C. 2009, MNRAS, 396, 1732Bohlin, R. C., Savage, B. D., & Drake, J. F. 1978, ApJ, 224, 132c (cid:13) , 000–000 E.Torresi et al.
Figure 7.
Mass outflow rate plotted against the radio loudness(R=Log [ ν L ν (5 GHz ) /L (2 − keV ) ]). Type 1 RL AGNs ( red cir-cles ) are the BLRGs considered in this work; type 1 RQ AGNs( black squares ) refer to the Seyferts, NLS1s, and QSOs belongingto our reference sample. Figure 8.
Histogram showing the velocity distribution of RLand RQ outflows. The separation between NLR/torus winds andaccretion disk winds is evident.Braito, V., Reeves, J.N., Sambruna, R.M., Gofford, J. 2011, MN-RAS, 414, 2739Burbidge, E.M., 1967, ApJ, 149, L51Buttiglione, S., Capetti, A., Celotti, A., Axon, D.J., Chiaberge,M., et al. 2009, A&A, 495, 1033Cardelli, J.A., Klayton, G.C., Mathis, J.S., 1989, ApJ, 345, 254Cleary, K., Lawrence, C.R., Marshall, J.A., Hao, L., Meier, D.2007, ApJ, 660, 117Condon, J.J., Cotton, W.D., Greisen, E.W., et al. 1998, AJ, 115,1963Crenshaw, D.M., Kraemer, S.B., & George, I.M. 2003, ARA&A,41, 117den Herder, J. W., Brinkman, A. C., Kahn, S. M., Branduardi-Raymont, G., et al. 2001, A&A, 365, 7Eracleous, M., & Halpern, J. P. 1994, ApJS, 90, 1Eracleous, M., Sambruna, R.M., Mushotzky, R.F. 2000, ApJ, 537,654Evans, D. A., Reeves, J. N., Hardcastle, M. J., Kraft, R. P., Lee,J. C., Virani, S. N. 2010, ApJ, 710, 859Feigelson, E.D. & Nulsen, P.I. 1985, ApJ, 293, 192Ferland, G. J., Korista, K. T., Verner, D. A., Ferguson, J. W., Kingdon, J. B., & Verner, E. M. 1998, PASP, 110, 761Ghisellini, G. & Tavecchio, F. 2008, MNRAS, 387, 1669Ghosh, K. K., Sondararajaperumal, S. 1991, AJ, 102, 1298Giovannini, G., Cotton, W. D., Feretti, L., Lara, L., Venturi, T.2001, ApJ, 552, 508Gliozzi, M., Sambruna, R.M., Eracleous, M. 2003, ApJ, 584, 176Gliozzi, M., Sambruna, R. M., Eracleous, M., & Yaqoob, T. 2007,ApJ, 664, 88Gomez, J., Marscher, A.P., Alberdi, A., et al. 2001, ApJ, 561,L161Grandi, P., Sambruna, R.M., Maraschi, L., Matt, G., Urry, C.M.,Mushotzky, R.F. 1997, ApJ, 487, 636Grandi, P., Guainazzi, M., Haardt, F., Maraschi, L., Massaro, E.,Matt, G., Piro, L., Urry, C. M. 1999, A&A, 343, 33Grandi, P., Maraschi, L., Urry, C.M., & Matt, G. 2001, ApJ, 556,35Grandi, P., Malaguti, G., & Fiocchi, M. 2006, ApJ, 642, 113Grandi, P., Guainazzi, M., Cappi, M., Ponti, G. 2007, MNRAS,381, L21 (G07)Grevesse, N., & Sauval, A. J. 1998, Space Sci. Rev., 85, 161Guainazzi, M. 2010, “XMM-EPIC status of calibration and dataanalysis”, XMM–SOCCAL–TN-0018Hardcastle, M.J., Alexander, P., Pooley, G.G., & Riley, J.M. 1998,MNRAS, 296, 445Hardcastle, M.J., & Worrall, D.M. 1999, MNRAS, 309, 969Hardcastle, M.J., Evans, D.A., Croston, J.H. 2007, MNRAS, 376,1849Harris, D.E., Mossman, A.E., Walker, R.C. 2004, ApJ, 615, 161Heckman, T.M., Smith, E.P., Baum, S.A. et al. 1986, ApJ, 311,526Hewitt, A., Burbidge, G. 1991, ApJS, 75, 297Isobe, T., Feigelson, E.D., Nelson, P.I. 1986, ApJ, 306, 490Kaastra J.S., Mewe R., Nieuwenhuijzen H. 1996, in YamashitaK., Watanabe T., eds, UV and X-Ray Spectroscopy of Astro-physical and Laboratory Plasmas. Universal Academy Press,Tokyo, p. 411Kaastra, J.S., Mewe, R., Liedhal, D.A., Komossa, S., &Brinkman, A.C. 2000, A&A, 354, 83Kalberla, P.M.W., Burton, W.B., Hartmann, Dap, Arnal, E.M.,Bajaja, E., Morras, R., & P¨oppel, W.G.L. 2005, A&A, 440,775Kaspi, S., Brandt, W.N., George, I.M., Netzer, H., Crenshaw,D.M., et al. 2002, ApJ, 574, 643Kaspi, S., Netzer, H., Chelouche, D., George, I.M., Nandra, K.,Turner, T.J. 2004, ApJ, 611, 68Kataoka, J., Reeves, J.N., Iwasawa, K., et al. 2007, PASJ, 59, 279Kellermann, K.I., Pauliny–Toth, I.I.K., Williams, P.J.S. 1969,ApJ, 157, 1Kinkhabwala A., Sako M., Behar E., et al. 2002, ApJ, 575, 732Komatsu, E., et al. 2009, ApJS, 180, 330Krolik, J.H., & Kriss, G.A. 2001, ApJ, 561, 684Krongold, Y., Nicastro, F., Elvis, M., et al. 2007, ApJ, 659, 1022Kruper, J. S., Canizares, C. R., Urry, C. M. 1990, ApJS, 74, 347Lavalley, M.P., Isobe, T., Feigelson, E.D. 1992, Bull. Am. Astron.Soc., 24, 839Leighly, K. M., O’Brien, P. T. 1997, ApJ, 481, 15Lister, M. L., Cohen, M. H., Homan, D. C., Kadler, M., Keller-mann, K. I., Kovalev, Y. Y., Ros, E., Savolainen, T., Zensus,J. A. 2009, AJ, 138, 1874Maraschi, L., Chiappetti, L., Falomo, R., et al. 1991, ApJ, 368,138Marchesini, D., Celotti, A., Ferrarese, L. 2004, MNRAS, 351, 733Marzke, R.O., Huchra, J.P., Geller, M.J. 1996, AJ, 112, 1803Mason, K. O., Breeveld, A., Much, R., et al. 2001, A&A, 365, 36Mathews, W. G., & Ferland, G. J. 1987, ApJ, 323, 456McKernan, B., Yaqoob, T., Mushotzki, R.F., Ian, M.G., Turner,T.J. 2003, ApJ, 598, L83c (cid:13) , 000–000 arm Absorber Energetics in Broad Line Radio Galaxies McKernan, B., Yaqoob, T., Reynolds, C.S. 2007, MNRAS, 379,1359Nagar, N.M., Wilson, A.S., & Falcke, H. 2001, ApJ, 559, L87Nicastro, F., Fiore, F., Perola, G. C., Elvis, M., O’Dea, C.P. 1998,PASP, 110, 493Ogle, P.M., Davis, S.W., Antonucci, R.R.J., et al. 2005, ApJ, 618,139Pearson, T.J., Readhead, A.C.S. 1988, ApJ, 328, 114Peterson, B.M., et al. 2004, ApJ, 613, 682Piconcelli, E., Bianchi, S., Miniutti, G., Fiore, F., Guainazzi, M.,Jimenez-Bailon, E., Matt, G. 2008, A&A, 480, 671Pounds, K.A., King, A.R., Page, K.L., O’Brien, P.T. 2003a, MN-RAS, 346, 1025Pounds, K.A., Reeves, J.N., King, A.R. 2003b, MNRAS, 345, 705Prieto, M.A. 2000, MNRAS, 316, 442Reeves, J.N., O’ Brien, P.T., Ward, M.J. 2003, ApJ, 593, 65Reeves, J. N., Sambruna, R.M., Braito, V., Eracleous, M. 2009,ApJ, 702, 187Reeves, J. N., Gofford, J., Braito, V., Sambruna, R. 2010, ApJ,725, 803 (R10)Reynolds, C. 1997, MNRAS, 286, 513Sambruna R.M., Eracleous M., Mushotzky R.F. 1999, ApJ, 526,60Sambruna, R., Reeves, J. N., Braito, V. 2007, ApJ, 665, 1030Sambruna, R. M., Reeves, J. N., Braito, V., Lewis, K. T., Era-cleous, M., et al. 2009, ApJ, 700, 1473Seielstad, G.A., Cohen, M.H., Linfield, R.P., et al. 1979, ApJ,229, 53Shankar, F., Cavaliere, A., Cirasuolo, M., & Maraschi, L. 2008,ApJ, 676, 131Stru¨der, L., Briel, U., Dennerl, K., et al. 2001, A&A, 365, 18Terashima, Y., & Wilson, A.S. 2003, ApJ, 583, 145Torresi, E., Grandi, P., Guainazzi, M., Palumbo, G.G.C., Ponti,G., Bianchi, S. 2009, A&A, 498, 61Torresi, E., Grandi, P., Longinotti, A. L., Guainazzi, M.,Palumbo, G. G. C., Tombesi, F., Nucita, A. 2010, MNRAS,401L, 10Ulvestad, J.S., Wilson, A.S. 1984, ApJ, 285, 439Urry, C.M., & Padovani, P. 1995, PASP, 107, 803Walker, R.C., Benson, J.M., Unwin, S.C. 1987, ApJ, 316, 546Wamsteker, W., Wang, T.-G., Schartel, N., Vio, R. 1997, MN-RAS, 288, 225Willott, C.J., Rawlings, S., Blundell, K.M., & Lacy, M. 1999,MNRAS, 309, 1017Yee, H.K.C., Oke, J.B. 1981, ApJ, 248, 472Yaqoob, T, Padmanabhan, U. 2004, ApJ, 604, 63Zdziarski, A.A., Grandi, P. 2001, ApJ, 551, 186Zheng, W. 1996, AJ, 111, 1498c (cid:13)000