SIGRAV Graduate School in Contemporary Relativity and Gravitational Physics Laura Ferrarese Rutgers University Lecture 6: The Future

Lecture Outline  A Recap of the Observational Status of SBH Research  Open Questions: What Remains to be Done 1. Building the Local Sample 2. The Redshift Evolution of SBH Scaling Relations 3. The Structure of the Broad Line Region 4. Black Holes and Globular Clusters 5. Binary Supermassive Black Holes.

Primary Methods: Phenomenon:BL Lac Objects Quiescent Galaxies Type 2 AGNs Type 1 AGNs Fundamental Empirical Relationships: Secondary Mass Indicators: Low-z AGNs Estimating SBH Masses Fundamental plane:  e, r e   *  M BH Stellar, gas dynamics Megamasers2-d RM 1-d RM M BH –  * AGN M BH –  * [O III ] line width V   *  M BH High-z AGNs Broad-line width V & size scaling with luminosity R  L 0.7  M BH

1. SBHs are fundamental constituents of galaxies: the local SBH mass density is equal to what is needed to explain the energetics of high redshift QSOs (Merritt & Ferrarese 2001; Ferrarese 2002; Yu & Tremaine 2002). 2. The existence of tight relations between SBHs masses and the large scale properties of their host galaxies suggests that the formation and evolution of SBHs and their hosts must go hand in hand. Understanding how SBHs form might help us to understand how galaxies form/evolve (or viceversa). What We Have Learned, and Open Questions 1. How small or how large can SBHs be? 2. What is the exact characterization of the SBH-galaxy connections? Self regulating SBH formation links the mass of SBHs to that of the dark matter halo in which they reside. Such models have a built in lower limit to the SBH mass that can be created (e.g. Silk & Rees 1998, Haehnelt, Natarajan & Rees 1998). 3. Can we obtain dynamical evidence for the existence of IBHs? An increasing amount of evidence supports the existence of “intermediate mass black holes” ( M), which could be the seeds for nuclear SBHs (Ebisuzaki et al. 2001; Portegies Zwart & McMillan 2002; Miller & Hamilton 2002) 4. Do binary black holes exist? The formation of binary black holes as a consequence of mergers can have dramatic consequences, from driving the morphology and dynamics of the host core (Milosavljevic & Merritt 2001), to destroying nuclear dark matter halo cusps (Merritt et al. 2002) The persistence of the M  relation in the face of mergers probes the cooling/heating feedback of the ISM (e.g. Kauffmann & Haehnelt 2000).

SPIRALS ELLIPTICALS LENTICULARS Biases and Systematics Systematics in the M  relation (or any other SBH scaling relation!) have not been fully investigated:  Slope, zero point & scatter  10 9 M  regimes  Dependence on Hubble type  Dependence on galaxy environment  Cosmic evolution  Reliability of SBH mass measurements

Biases in the Current Mass Estimates  In the current sample there is only one galaxy for which a SBH mass estimate has been obtained using two independent methods (IC1459; Verdoes Kleijn et al. 2000, Cappellari et al. 2002)  M BH (stars) = (2.6  1.1)  10 9 M  (using 3I modeling of HST/STIS data with N 0 /N c =2.0)  M BH (gas) = (0.4  1.0)  10 9 M  (depending on the assumptions made for the gas velocity field)  There is only one galaxy for which the same data has been analyzed by two different teams using the same method (3-I modeling using different codes - M32, van der Marel et al. 1998, ApJ, 493, 613; Valluri et al. 2003, astro-ph/ ):  M BH (vdM) = (3.4  0.7)  10 6 M   M BH (Valluri) = (1  6)  10 6 M   Finally, some of the data being analyzed might not be adequate in terms of spatial resolution and signal-to-noise ration

Resolving the Sphere of Influence Kormendy & Gebhardt 2001, Gebhardt et al  The quality of the data might very well influence the characterization of the scaling relations (remember Magorrian et al. 1998…)

Resolving the Sphere of Influence 4.58  0.52 (Ferrarese 2002) 4.02  0.32 (Tremaine et al. 2002)

1. Addressing the Faint End of the M-  Relation  How far does the M  relation extend?

HST/STIS/0.1” slit 1. Addressing the Faint End of the M-  Relation : M33 M33 is an ideal target: very nearby (850 kpc)  tightest limits on a small BH. very compact nucleus, reaching a stellar mass density of several million solar masses per cubic parsec  ideal conditions for BH formation (?) Very low central velocity dispersion (~ 24 km/s, Kormendy & McClure 1993)  very small black hole Inconspicuous bulge  very small black hole (?) The M33 nucleus probably contains the most luminous ULX in the Local Group (Long et al. 1981), strengthening the connection with the M82 ULXs (Ebisuzaki et al. 2001) Tightest limit from ground based data: M BH < 50,000 M  The M BH  relation predicts 2,600 < M BH < 26,000 M   radius of influence < arcsec  Only an upper limit can be set on the mass. Merritt, Ferrarese & Joseph (2001)

Gebhardt et al. 2001; Merritt, Ferrarese & Joseph 2001 M33 1. Addressing the Faint End of the M-  Relation : M33  The upper limit is still consistent with the M  relation.

1. Addressing the Faint End of the M-  Relation : NGC205  On-going HST-ACS/NICMOS/STIS project (Ferrarese, Merritt, Valluri, Joseph). Of the galaxies for expected to harbor a SBH with M BH < 10 6 M , NGC 205 is the only one for which the sphere of influence can be resolved by HST in a finite amount of time. Andromeda, NGC 205 and M X 2 degrees NGC205 - HST/ACS/HRC - 29X29 arcsec

 Stellar kinematics in giant ellipticals, all of which have low central surface brightness.  Exposure times requirements with HST/STIS to observe an object like M87 are 4  10 6 seconds (to reach a S/N=50 at 8500 Å).  All giant ellipticals must (and have been) studied using gas kinematics. 2. Building the Local Sample: What HST Cannot Do Ferrarese et al. 1994, AJ

 Nearby dwarf systems  Exposure times requirements with HST/STIS to observe an object like NGC147 are 10 6 seconds (to reach a S/N=50 at 8500 Å) 2. Building the Local Sample: What HST Cannot Do NGC 147 (Han et al. 1997, AJ) WFPC2, F555W

2. Building the Local Sample HST 30m 8m FornaxVirgoM101 group  What if we wanted to, say, increase the current sample size to include significantly statistical samples of galaxies belonging to all Hubble types and disparate environments?  Below: SBH masses for all galaxies belonging to the CfA redshift sample (Huchra et al. 1990, ApJS, 72, 433), estimated from the bulge luminosity, as a function of distance

2. Building the Local Sample  The sample thins out even more if S/N requirements are added. The figure shows the requirements for a complete sample of elliptical galaxies (from Faber 1989) observed with S/N=50 in the absorption lines at 8500 Å, in the equivalent of 3 HST orbits. 30m8mHST

2. Building the Local Sample: SBHs in Low Surface Brightness Galaxies (?)  Studies of giant LSB galaxies have found a ~50% incidence of low luminosity AGNs (Sprayberry et al. 1995, Schombert 1998).  This is important for two reasons:  LSBs are not accounted for in studies of SBH demographic, but they could contribute significantly to the local SBH mass function.  LSBs differ from HSBs in morphological appearance, past and current SFR, disk kinematics, mass to light ratios, gas ratios and molecular content. Their SBH mass function is certain to shed light on the mechanisms of formation and evolution of SBHs. For instance, could LSBs be be remnants of the most massive QSOs (Silk & Rees 1999)? Schombert ‘98

2. Building the Local Sample: SBHs in Low Surface Brightness Galaxies (?)  This is a project that could be possibly done with a combination of ground based and HST observations. For instance:  There are 73 LSB galaxies with z < 5000 km/s in the Impey et al. (1996) catalogue. For all: 1) 4m-class telescope: low resolution spectra to determine incidence of nuclear activity. For the AGN subsample: 2) HST/ACS H  images to determine the extent and morphology of the ionized gas in the nuclear region. For the most promising candidates: 3) HST/STIS spectra at 0.1 arcsec spatial resolution to determine the nuclear gas kinematics and constrain the masses of the central SBHs.

3. The Redshift Evolution of the M BH  Relation  The most distant objects with a direct SBH mass measurement are at 100 Mpc. However, studying the M BH  relation as a function of redshift would tell us about the incidence of merging and accretion on the evolution of SBHs and their host galaxies.  Even using an 8m diffraction limited telescope, resolved stellar/gas dynamics can probe 10 9 M  SBHs only up to a few hundred Mpc away. However, if we restrict ourselves to AGNs, we do not have to rely on resolved kinematics.  Is targeting AGNs feasible? Is it a good idea?  AGN samples are quite large. For instance, in the Veron-Cetty & Veron (2001) catalogue there are: 13 Seyfert 1s with 1<z< northern QSOs with 1.0<z< northern QSOs with 2.0<z<3.0  In the local Universe, quiescent galaxies and AGNs obey the same scaling relation as far as the supermassive black holes at the the center are concerned. But is this true at high redshift? (there is no answer!)

3. The Redshift Evolution of the M BH  Relation  “Quick and dirty” approach :  obtain AGN spectra to 1) estimate the BLR radius from the monochromatic luminosity at 5100 Å; and 2) the virial velocity from the width of Balmer lines.  The SBH mass follows by combining the BLR size (from the continuum luminosity) with the line width.  In applying this method to high redshift AGNs, we make two tacit assumptions:  the R BLR -L 5100 relation is valid at high redshifts  The R BLR -L 5100 extends at large luminosities. From Kaspi et al. (2000)

3. The Redshift Evolution of the M BH  Relation  “Quick and dirty” approach :  The velocity dispersion can be difficult to estimate at high redshifts. It has been proposed that the [OIII]5007 FWHM can act as a surrogate for  (Nelson & Whittle 1986, ApJ, 465, 96). From Nelson & Whittle (1986) Left: exactly this procedure applied to a sample of 107 radio quite QSO and Seyfert 1s (Boroson 2003). Notice the large scatter.

3. The Redshift Evolution of the M BH  Relation  The only attempt at applying this method at high redshift was made by the Shields et al. (2002). Although the authors conclude that “the MBH  relationship is not a strong function of redshift… The black hole bulge mass relationship is roughly obeyed at a time when much of the growth of present day black holes lay in the future”, notice how:  only very large SBH masses are probed at high redshift. Therefore the study contains no information about the slope of the relation, although it does suggest that the relation is pinned at the very high mass end.  This is to be expected: the largest SBHs must not have changed much since a redshift 3. M BH =10 7  3  10 8 M  M BH =10 9  M  From Shields et al. 2002

 There is a more direct but much more involved approach to this issue:  Reverberation mapping for selected objects: needs monitoring at 4d to 3m intervals (depending on luminosity/expected SBH mass) over periods of 1y to 15y. Can be carried out with 4m and 8m class telescopes.  Stellar velocity dispersion in the AGN host: direct measurements have been obtained for QSOs up to z=0.3 using conventional 4m class telescope (Hughes et al. 2000) measurements at z=1 can be obtained with an 8m class telescope equipped with AO in the near IR. measurements at z=2 are NGST material.  A project on this scale would likely require international collaboration over at least a decade, and is not currently being undertaken.  Furthermore, the reliability of reverberation mapping as a mass estimator ultimately lies in probing the the structure of BLR. Are we doing anything on this front? (no, but we should be!) 3. The Redshift Evolution of the M BH  Relation

40-cm X-ray telescope 70-cm UV/optical telescope KRONOS (PI B. Peterson) is a proposed NASA/Midex mission which will allow uninterrupted UV/Optical/X- ray observations for as long as 14 days on target. It will allow to distinguish between different simple transfer functions, something that no experiment to date has been able to do! 4. The Structure of the Broad Line Region  Based on the experience accumulated so far, accurate mapping of the BLR requires a number of characteristics:  High time resolution (  0.2 day)  Long duration (several months)  Moderate spectral resolution (  600 km s -1 )  High homogeneity and signal-to-noise (~100)

Ground-based optical image cm Hubble Space Telescope image cm  100 Kronos transfer function Kronos map cm  10, The Structure of the BLR: KRONOS

5. IBH in Globular Cluster: Proper Motion Studies  Confusion in minimized in the U-band (or B-band)  NGST is not ideal  Observing Constraints: Need to measure enough stars to estimate the velocity dispersion accurately (10%): N= Need to resolve N stars within the sphere of influence:N res 2 < r infl 2 Need to have enough S/N to measure position within 0.002% pixel:S/N ~ 150 Need reasonable exposure times / time interval between observations: t =10 orbits,  t =1 year Alcaino et al. 1998, AJ

NGC IBH in Globular Cluster: Proper Motion Studies 30m M  8m 40-1,000 M  HST ,000 M  Mass range estimated for the IBHs in the Antennae and M82 (Fabbiano et al. 2001, Matsumoto et al. 2001)

5. IBH in Globular Cluster: But... Paresce & de Marchi, 2000, ApJ  In this case, going to a larger aperture and having better resolution only helps up to a point! The limit is imposed by the actual number of stars in the critical inner region.  For clusters within 40 kpc, this limit is reached for an 8m class telescope.

5. IBH in Globular Cluster: But... M15 30m 8m HST NGC m 8m HST

5. IBH in Globular Cluster: Proper Motion Studies HST 8m 30m distance = kpc

6. Detecting Binary Black Holes  Milosavljevic & Merritt 2001 (ApJ) have conducted state of the art N-body simulations aimed at following the dynamical evolution of a supermassive black hole binary and its surrounding stellar system following galaxy merging. Rotational Velocity Velocity Dispersion Myr Myr time

6. Detecting Binary Black Holes Separation between the SBHs: 0.18 pc (10 9 M  SBHs) Difference in Rotational Velocity: 95 km/s (10 9 M  SBHs) Difference in Velocity Dispersion: 136 km/s (10 9 M  SBHs) From Milosavljevic & Merritt 2001, ApJ Thick Line: SBH Binary Thin Line: Single SBH

6. Detecting Binary Black Holes HST8m30m

SummarySummary Project Spatial Resol. Aperture FOV Bandpass  Comments Enlarge the sample; probe 10 9 M  SBHs; test biases with type & environment Needed to study systematics, distinguish between “bottom up” or “top down” models for SBH formation; constrain the role of feedback in SBH accretion during merging. 8m few tens of arcsec ,000 Longslit IFU Constrain dynamical models for GC evolution; investigate the connection between GCs, nuclear SBHs and galactic bulge. Same data useful to measure GC distances, constrain ages. ~ 0.1 arcsec HST few tens of arcsec U or B-band N/A High Res. Imaging High Dynamic Range Detecting SBHs in GCs 30m few tens of arcsec > ,000 IFU Constrain dynamical models for galaxy mergers; determine the impact of SBH binaries in the morphological evolution of galaxies; constrain accretion mechanisms. Resolving Binary SBHs Constrain SBHs formation and evolution. Reverberation mapping, KRONOS. Constrain the redshift evolution of SBH scaling relations

Suggested Readings  Thoughts about future work:Ferrarese, L. 2002, in ‘Hubble's Science Legacy: Future Optical-Ultraviolet Astronomy from Space’, ` astroph/ Peterson 2002, in ‘Hubble's Science Legacy: Future Optical-Ultraviolet Astronomy from Space’, astroph/