A path to radio-loudness through gas-poor galaxy mergers and the role of retrograde accretion
***FULL TITLE**ASP Conference Series, Vol. **VOLUME**, c (cid:13) **YEAR OF PUBLICATION****NAMES OF EDITORS** A path to radio–loudness through gas–poor galaxymergers and the role of retrograde accretion
M. Dotti, M. Colpi, L. Maraschi, A. Perego, M. Volonteri Max Planck Institute for Astrophysics, Karl-Schwarzschild-Str. 1,85741 Garching, Germany Dipartimento di Fisica, Universit`a degli Studi di Milano-Bicocca,Piazza Della Scienza 3, 20126 Milano, Italy INAF, Osservatorio Astronomico di Brera, Via Brera 28, 20123Milano, Italy Department of Physics, University of Basel, Klingerbergstr. 82, 4046Basel, Switzerland Department of Astronomy, University of Michigan, Ann Arbor, MI48109, USA
Abstract.
In this note, we explore a pathway to radio–loudness under the hy-pothesis that retrograde accretion onto giant ( M BH ∼ M (cid:12) ) spinning ( a ∼ > . gas–poor galaxymergers progenitors of giant (missing–light) ellipticals. The occurrence of ret-rograde accretion enters as unifying element that may account for the radio–loudness/galaxy morphology dichotomy observed in AGN. Thanks to a wealth of observational and theoretical studies of AGN from radioto gamma-ray frequencies, radio-loud AGN are presently thought to containrelativistic jets which propagate from the nucleus out to kpc and, in some cases,Mpc scales. The power of jets required to at least fuel their radio lobes is huge,up to 10 erg s − corresponding to an energy release of ∼ erg over a timeof 10 yr. In radio-quiet AGN, the radio emission is instead weak and largescale radio jets are absent. Although the boundary between these two classes isblurred, strong jets (in terms of kinetic relative to accretion power) exist onlyin a minority of AGN.Optical studies of nearby galaxy nuclei (Capetti & Balmaverde 2006) haveshown that all radio-loud AGN invariably reside in giant core (or missing–light ;Kormendy et al. 2009) ellipticals, i.e. in early-type galaxies whose nuclearsurface brightness profile shows a deficit in star-light with respect to the outerprofile. By contrast, normal, less massive coreless (or extra–light ; Kormendyet al. 2009) ellipticals all host radio-quiet AGN (Capetti & Balmaverde 2006),suggesting that a morphology-related dichotomy exists among ellipticals , likelydetermined by galaxy evolution. The dichotomy seems also to emerge from asample of radio-loud and radio-quiet QSOs at redshift z < .
2, the former resid-ing in giant ellipticals that show signs of interaction, and the latter residing in1 a r X i v : . [ a s t r o - ph . H E ] O c t Dotti, Colpi, Maraschi, Perego, Volonteri merging gas-rich galaxies of intermediate mass (Wolf & Sheinis 2008). Evidenceexists that the most powerful radio and gamma–loud AGN are associated to theheaviest supermassive BHs in the universe, with M BH ∼ > M (cid:12) (Ghisellini et al.2009a,b) indicating that the BH mass can be a key parameter for radio-loudness.From a theoretical point of view, the launching of jets from accreting BHshave been extensively studied in the past three decades, following the seminalpaper by Blandford and Znajek (BZ hereafter; 1978). The BZ process exploitsenergy extraction from a rapidly rotating black hole via a purely electromag-netic interaction of a large scale magnetic field which threads the rotating eventhorizon. In this framework, jets are produced at the expense of the rotationalenergy of the black hole E rot = M BH [1 − − / (1 + √ − a ) / ] c , where a is thedimensionless spin parameter related to the BH spin J BH = ˆl ( GM a/c ) point-ing along ˆl (for a maximally rotating [ a = 1] Kerr BH of 10 M (cid:12) , E rot = 6 × erg). This may suffice to power a jet and to explain why AGN with comparableoptical luminosities can be either radio-loud (large a ) or radio-quiet (small a , as E rot decays ∝ M BH a / a →
0) depending on a (Wilson & Colbert 1995;Sikora et al. 2007).Advanced fully general-relativistic magneto-hydrodynamical simulations ofspinning BHs (McKinney 2005, 2006) predict high collimation of the innerPoynting flow, bulk Lorentz factors of order 10, and jet powers P jet ∝ a , or ∝ a , implying in this last case high jet luminosities only for close to maximallyspinning BHs (Tchekhovskoy et al. 2009). Another essential ingredient deter-mining the BZ power is the value of the magnetic field threading the horizon,which is tied to the accretion rate and disc structure (Ghisellini et al. 2009a).In two recent works, Garofalo (2009a,b) stressed the importance of theplunging region between the innermost stable circular orbit (ISCO ) of the ac-cretion disc and the BH horizon. In this region the magnetic field can be substan-tially amplified from its value at ISCO, particularly if the BH has a retrograde spin vector with respect to the accretion disc angular momentum (we will usehere the convention of negative spin, a <
0, for retrograde accretion). As aconsequence, high jet powers can be extracted even/also from non-maximallyrotating BHs (Garofalo 2009b).Here, we explore the consequences of Garofalo’s conjecture in the aim atconnecting the BH spin and accretion mode to the structure of the underly-ing host galaxy. Moderately high values of the spin parameter a ( ∼ > .
7) andretrograde accretion appear to enter as unifying ingredients to account for theradio-loud/host morphology dichotomy observed, at low redshifts, among ellip-ticals. These considerations go in the direction of the very recent proposal thatmost radio-loud AGN harbor spinning BHs accreting in the retrograde mode(Garofalo et al. 2010). ISCO, expressed in units GM BH /c , is at 6 for a = 0, and at 9 (1) for retrograde accretionwith a = − a = 1). See Sikora et al. (2007) for the discussion on the most general late/early-type versus radio-quiet/loud dichotomy observed in AGN, and Fanidakis et al. (2009) for a different spin depen-dent cosmological model of jet formation in AGN. adio loudness and gas-poor mergersadio loudness and gas-poor mergers
7) andretrograde accretion appear to enter as unifying ingredients to account for theradio-loud/host morphology dichotomy observed, at low redshifts, among ellip-ticals. These considerations go in the direction of the very recent proposal thatmost radio-loud AGN harbor spinning BHs accreting in the retrograde mode(Garofalo et al. 2010). ISCO, expressed in units GM BH /c , is at 6 for a = 0, and at 9 (1) for retrograde accretionwith a = − a = 1). See Sikora et al. (2007) for the discussion on the most general late/early-type versus radio-quiet/loud dichotomy observed in AGN, and Fanidakis et al. (2009) for a different spin depen-dent cosmological model of jet formation in AGN. adio loudness and gas-poor mergersadio loudness and gas-poor mergers Can retrograde accretion onto a spinning BH be established during galaxy evo-lution?
To answer this question in a broader context, we first consider isolated disc galaxies, in which BHs are expected to grow only by accretion. In these galax-ies BH fueling can occur either via multiple uncorrelated episodes of accretion(Moderski et al. 1998), or via secular bar-in-bar instabilities (Shlosman, Frank& Begelman 1989) that remove gradually the gas angular momentum. In thefirst mode, episodes of chaotic accretion result in low BH spins (King & Pringle2006; Volonteri et al. 2007). Prograde and retrograde accretion change both theBH spin modulus a and direction ˆl . Since retrograde orbits carry a larger angularmomentum than prograde orbits, random episodes result in average values of (cid:104) a (cid:105) ∼ < .
4. In this scenario, retrograde accretion events are common, but the ac-creting BH is always slowly spinning, so no powerful, long–lived collimated jetsare produced. In the second scenario, the BH grows through coherent accretion.Even if a ∼ a ∼ √
6, Bardeen 1970). In this scenario, the BH spin J BH aligns with the angular momentum of the accretion disc. The BH remains inthe radio-quiet mode if a ∼ < .
9, as retrograde accretion is unlikely to establish,given the coherence of the flow. A strong jet in a disc galaxy could be triggeredby prograde accretion only if a > . ∼
30 times dimmer than that produced by retrograde accretiononto a maximally spinning BH (Garofalo 2009b).
The situation is different in the case of major mergers , i.e. mergers betweengalaxies of comparable masses. Galaxy mergers are commonly divided in twoclasses, gas rich (or wet ) mergers where the fraction of cold gas is large ( ∼ > dry ) mergers where gas is a small fractionof the stellar mass and has no dynamical effect in the merger. During a wet merger, the tidal field between the two disc galaxies drives largeamounts of gas (up to 50% of the total gas mass) toward the centres of the twointeracting galaxies (Mayer et al. 2007). When the two nuclei later merge in asingle structure, the gas settles into a dense, selfgravitating circumnuclear discin which the BH relative orbit becomes circular and corotating until the BHsform a Keplerian binary (Dotti et al. 2009). In this phase, lasting ∼ < yr,the BHs accrete in a coherent manner at a rate sufficient to align their spins,initially oriented at random, to the angular momentum of the nuclear disc (Liu2004; Bogdanovic, Reynolds & Miller 2007; Dotti et al. 2010): in response tothe Bardeen–Petterson warping of the small–scale accretion discs grown aroundeach BH, total angular momentum conservation imposes fast ( ∼ < Dotti, Colpi, Maraschi, Perego, Volonteri prograde until coalescence, with no major changes in the BH spin orientation.Under these circumstances the BH remnant retains the spin direction of theparent BH spins, both oriented parallel to the angular momentum of their or-bit: the post-coalescence BH may thus acquire a large spin a ∼ > . fast alignment and spin coherence relative to the flow, prograde accretion continues even after BH coalescence and so no powerful long–lived jetsare produced, according to our working hypothesis, unless the BH remnant isclose to be maximally rotating. We can also follow the evolution of a wet merger by looking at the propertiesof the galaxy remnant. Gas rich galaxy mergers are expected to form eitherspheroids with residual rotation or discs, depending on their initial gas content(Robertson et al. 2006), encounter geometry and the presence of cold gaseousstreams flowing onto the evolving remnant (Governato et al. 2009). The stellarnucleus of the relic galaxy can in principle be perturbed by the massive BHbinary. The binary hardens ejecting stars via three-body scatterings (Merritt &Milosavljievic 2004) creating a stellar core . The post-coalescence BH can furtherheat the stellar nucleus if it experiences a strong gravitational–wave inducedrecoil ( ∼ < − ; Lousto & Zlochower 2009 and references therein): inits oscillations back to the galactic centre the BH deposits its kinetic energyexpanding further the core (Boylan–Kolchin et al. 2004; Gualandris & Merritt2008). However, in a gas–rich merger, these effects are strongly suppressed:the stellar deficit by the BH binary is weak since the binary hardens mainly viagaseous torques (see Colpi & Dotti 2009 for a review) and extra–light is producedfollowing the formation of new stars inside the circumnuclear disc (Hopkins et al.2009a). Furthermore, spin-orbit alignment from coherent accretion implies smallgravitational kicks for the relic BH, limiting the effect of dynamical heating of thenucleus. As a consequence, we predict that wet mergers result mostly in radio– quiet AGN hosted by coreless, extra–light elliptical or spiral galaxies . Figure 1illustrates schematically the evolution of the two BHs and of the remnant galaxyin a gas–rich merger.
Mergers of gas–poor galaxies (e.g. between two ellipticals) lack massive nucleargas discs. The BHs thus complete their inspiral in a collisionless backgroundof stars. The BH spins are expected to be uncorrelated with the geometry ofthe encounter, so that, as the binary forms, their spins are randomly orientedrelative to the binary orbit. The BHs may capture gas and experience episodesof retrograde accretion until they merge. At coalescence the BH spin changesin modulus and orientation, and memory is lost of the initial spin directions.Berti & Volonteri (2008) predict, for isotropic dry mergers, a broad distributionof final spins centred around (cid:104) a (cid:105) ∼ . giant, missing–light elliptical whose nuclear region has been shaped by the BH (Hopkins et al. 2009b;Kormendy & Bender 2009). Hardening through three–body scatterings withstars excavates a stellar core in the nuclear region, further expanded by BH adio loudness and gas-poor mergersadio loudness and gas-poor mergers
Mergers of gas–poor galaxies (e.g. between two ellipticals) lack massive nucleargas discs. The BHs thus complete their inspiral in a collisionless backgroundof stars. The BH spins are expected to be uncorrelated with the geometry ofthe encounter, so that, as the binary forms, their spins are randomly orientedrelative to the binary orbit. The BHs may capture gas and experience episodesof retrograde accretion until they merge. At coalescence the BH spin changesin modulus and orientation, and memory is lost of the initial spin directions.Berti & Volonteri (2008) predict, for isotropic dry mergers, a broad distributionof final spins centred around (cid:104) a (cid:105) ∼ . giant, missing–light elliptical whose nuclear region has been shaped by the BH (Hopkins et al. 2009b;Kormendy & Bender 2009). Hardening through three–body scatterings withstars excavates a stellar core in the nuclear region, further expanded by BH adio loudness and gas-poor mergersadio loudness and gas-poor mergers Figure 1. Gas–rich mergers between nearly equal mass disc galaxies. Theleft and right branches of the scheme refer to the evolution of the galaxies andthe BHs, respectively. Left: We hypothesize that the merger ends with theformation of a normal coreless elliptical where the central extra–light resultsfor a major star formation episode triggered inside the massive circumnucleardisc. Galaxy evolution models do not exclude the survival or re–growth ofthe disc (Governato et al. 2009) so that a disc galaxy may form at the endof the merger. Right: Observations of normal coreless ellipticals indicatethat they host BHs with M BH < M (cid:12) (Kormendy & Bender 2009). TheBHs that initially inhabit the two galaxies are expected to pair and form abinary corotating with the disc (Dotti et al. 2009). During inspiral underthe action of gas-dynamical torques, gas accretion aligns the BH spins tothe orbit (Dotti et al. 2010) so that the post-coalescence BH has a high( a ∼ > .
7) spin and corotates with the disc. After this phase and according toour working hypothesis the BH is radio-quiet . We notice that a short-livedradio-loud phase could be present, prior to BH coalescence, before disc-orbitand spin-orbit alignment occur.
Dotti, Colpi, Maraschi, Perego, Volonteri heating by the recoil. The final giant rapidly spinning BH later settles in thegalaxy’s centre and starts accreting. In the absence of a rotationally supportednuclear gas structure, half of the accretion events will be retrograde, and theBH enters the radio-loud phase.If the BH is very massive ( M BH > M (cid:12) ), as seen in giant ellipticals(Kormendy & Bender 2009), we argue that its spin direction remains stable through all accretion episodes: because of its large BH mass and the lack ofa massive nuclear disc, the BH angular momentum likely exceeds the angu-lar momentum carried by the accretion disc (King et al. 2005). Indeed, wenote that for giant BHs, the warp radius R warp (delimiting the distance forgravito-magnetic interaction between the BH and the disc) exceeds the outerradius of the accretion disc determined by self-gravity, R out . Following Peregoet al. (2009), for a 10 M (cid:12) BH, the outer radius R out ∼ M − / , f − / R G (where R G is the Schwarzshild radius and f E the Eddington factor) is smallerthan R warp ∼ a / M / , f / R G implying that over the Bardeen-Pettersontimescale ( ∼ <
100 yrs) the entire disc aligns o antialigns (depending on the disc’sinitial orientation) with the BH spin vector that keeps its orientation stable.Thus the BH can sustain a collimated jet over large scales until accretion ceases. We thus expect radio–loud AGN to be hosted in missing–light, core ellipti-cals, remnants of dry mergers . This second scenario is summarized in Figure2.
In this note, we propose that retrograde accretion onto a massive ( M BH ∼ > M (cid:12) ) post-coalescence spinning BH is conducive to the generation of a pow-erful large-scale jet, in the gas-poor environment of a giant elliptical with core.This is in agreement with the observation that radio–loud AGN are hosted inbright ellipticals that show a stellar core (whenever optical data are available),and with the recent finding on the dichotomy in radio jet orientations in ellipticalgalaxies (Browne & Battye 2010). In less-louder radio-loud AGN, Browne andBattye find a tendency for the axis of the radio emission to align with the minoraxis of the starlight of the host, an oblate rotationally supported elliptical. Bycontrast they find no preferred radio-optical alignment among the radio-louderobjects possibly hosted in triaxial non-rotating ellipticals. This study appears tosupport our evolutionary scheme: coreless (extra–light) rotationally supportedellipticals (Kormendy et al. 2009) that likely form in gas-rich mergers are ex-pected to host a BH accreting in the prograde mode with a high spin alignedwith the angular momentum of the large-scale disc, and so aligned with theminor axis of the starlight. By contrast, core (missing–light) triaxial ellipticalsare expected to host a BH that has a random spin orientation regardless itsaccretion mode (i.e. whether prograde or retrograde).The model rests on Garofalo’s conjecture that around a retrograde BH,magnetic fields threading the horizon reach maximum amplification, resulting ina dramatic enhancement of the jet power even for non–extreme values of a . Theinterval of spin values for operation of this process is still weakly constrained bytheory. Furthermore prograde accretion can also result in large BZ luminosities adio loudness and gas-poor mergersadio loudness and gas-poor mergers
In this note, we propose that retrograde accretion onto a massive ( M BH ∼ > M (cid:12) ) post-coalescence spinning BH is conducive to the generation of a pow-erful large-scale jet, in the gas-poor environment of a giant elliptical with core.This is in agreement with the observation that radio–loud AGN are hosted inbright ellipticals that show a stellar core (whenever optical data are available),and with the recent finding on the dichotomy in radio jet orientations in ellipticalgalaxies (Browne & Battye 2010). In less-louder radio-loud AGN, Browne andBattye find a tendency for the axis of the radio emission to align with the minoraxis of the starlight of the host, an oblate rotationally supported elliptical. Bycontrast they find no preferred radio-optical alignment among the radio-louderobjects possibly hosted in triaxial non-rotating ellipticals. This study appears tosupport our evolutionary scheme: coreless (extra–light) rotationally supportedellipticals (Kormendy et al. 2009) that likely form in gas-rich mergers are ex-pected to host a BH accreting in the prograde mode with a high spin alignedwith the angular momentum of the large-scale disc, and so aligned with theminor axis of the starlight. By contrast, core (missing–light) triaxial ellipticalsare expected to host a BH that has a random spin orientation regardless itsaccretion mode (i.e. whether prograde or retrograde).The model rests on Garofalo’s conjecture that around a retrograde BH,magnetic fields threading the horizon reach maximum amplification, resulting ina dramatic enhancement of the jet power even for non–extreme values of a . Theinterval of spin values for operation of this process is still weakly constrained bytheory. Furthermore prograde accretion can also result in large BZ luminosities adio loudness and gas-poor mergersadio loudness and gas-poor mergers Figure 2. Gas–poor mergers between nearly equal mass galaxies (e.g. be-tween two ellipticals). The left and right branches of the scheme refer to theevolution of the galaxies and the BHs, respectively. Left: We hypothesizethat the merger ends with the formation of a giant missing–light ellipticalwhere a stellar core has been excavated by the BH binary in its hardening byscattering off single stars, and/or by heating from gravitational recoil afterBH coalescence. Observations of giant core ellipticals indicate that they hostvery massive BHs with M BH ∼ > M (cid:12) (Kormendy & Bender 2009). Right:The BHs that initially inhabit the two galaxies are expected to pair, form abinary and eventually coalesce under the action of stellar torques. Prior tocoalescence the two BHs have arbitrary spin moduli and random orientation.In the gas–poor environment the spins of the giant BHs do not align with theorbit and the post-coalescence BH ends with a high spin a ∼ > . retrograde accretion (in 50% of the cases) and become, accordingto our hypothesis a radio–loud AGN. In the remaining 50% cases the BH maybe active and become an AGN. Observations indeed indicate that giant el-lipticals host radio–quiet AGN, and this is a natural outcome of our workingmodel.
Dotti, Colpi, Maraschi, Perego, Volonteri (Garofalo 2009a,b). Thus, there should exist a zone of avoidance delimitedby a maximum–negative a − jet and a minimum–positive a +jet , outside which thegeneration of jets is possible. The values of a − jet and a +jet may not be symmetricas non–symmetric is the underlying Kerr spacetime. If retrograde accretion isa necessary condition only for the most powerful jets hosted in ellipticals, thevalue of a − jet that our model requires is a − jet ∼ > − .
7, resulting from isotropic BHcoalescences in dry mergers. If prograde accretion vehicles the production ofjets as well, a value of a +jet close to +0 . a +jet is irrealistic. One would be tempted to require a +jet ∼ +1, according to themodels by McKinney (2005).The rotational energy stored in giant BH ( M BH ∼ > ) may suffice to powera radio–loud AGN at an average level of ∼ erg s − for a lifetime τ jet ∼ yr. However retrograde accretion causes the BH to spin down on a time τ down ∼ min( τ ac , τ jet ). Spin–down by accretion occurs on τ ac ∼ J BH / ( ˙ M net ˜ l ISCO ) ∼ M BH / ˙ M net (where ˙ M net is the net inflow rate onto the BH, and ˜ l ISCO the an-gular momentum per unit mass of a test particle at ISCO). τ ac is close to theSalpeter time for e –folding of the BH mass. The BZ timescale τ jet dependson the magnitude of the magnetic field that threads the horizon and on thethermal/geometrical structure of the accretion disc (Moderski et al. 1998). InGarofalo’s scenario, retrograde accretion is a transitory phase during the lifetimeof a BH (Garofalo et al. 2010) whose duration and recurrence depend on theway the BH is fed whether through continuous or chaotic accretion.Recent observations of extremely bright Blazars with BAT/ Swift at red-shifts z >
Fermi (Ghisellini et al. 2009b) indicate that luminous radio-loud AGN host giant
BHs, and that a prominent accretion disc co-exists witha powerful jet suggesting that extraction of rotational energy occurs throughaccretion (Maraschi 2001). We are tempted to associate the extreme BH mass,inferred from the disc luminosity, to the stability of the BH spin orientationand the jet power to its spin modulus and to accretion in the retrograde mode.Further studies on the stability of retrograde accretion will help in disentaglingthe nature on the AGN radio–loud/quiet dichotomy.
Acknowledgments.
M.V. acknowledges support from NASA award ATPNNX10AC84G.
References
Bardeen J.M., 1970, Nature, 226, 64Berti E. & Volonteri M., 2008, ApJ, 684, 822Blanfdord R.D. & Znajek R.L., 1977, MNRAS, 179, 433Bogdanovic T., Reynolds C.S. & Miller M.C., 2007, ApJ, 661, L147 adio loudness and gas-poor mergersadio loudness and gas-poor mergers