Supermassive black holes from OASIS and SAURON integral-field kinematics
M. Cappellari, R. Bacon, R. L. Davies, P. T. de Zeeuw, E. Emsellem, J. Falcon-Barroso, D. Krajnovic, H. Kuntschner, R. M. McDermid, R. F. Peletier, M. Sarzi, R. C. E. van den Bosch, G. van de Ven
aa r X i v : . [ a s t r o - ph ] S e p Formation and Evolution of Galaxy BulgesProceedings IAU Symposium No. 245, 2007M. Bureau, E. Athanassoula, and B. Barbuy, eds. c (cid:13) Supermassive black holes from OASIS andSAURON integral-field kinematics
Michele Cappellari, R. Bacon, Roger L. Davies, P. T. de Zeeuw, Eric Emsellem, Jes´us Falc´on-Barroso, Davor Krajnovi´c, HaraldKuntschner, Richard M. McDermid, Reynier F. Peletier, MarcSarzi, Remco C. E. van den Bosch, Glenn van de Ven Sub-Department of Astrophysics, University of Oxford, England Univ. de Lyon 1, CRAL Observ. de Lyon; CNRS, Ecole Normale Sup´erieure de Lyon, France Sterrewacht Leiden, Leiden University, Leiden, The Netherlands European Space and Technology Centre, Noordwijk, The Netherlands Space Telescope European Coordinating Facility, ESO, Garching, Germany Kapteyn Astronomical Institute, Groningen, The Netherlands Centre for Astrophysics Research, University of Hertfordshire, England Institute for Advanced Study, Princeton, USA
Abstract.
Supermassive black holes are a key element in our understanding of how galaxiesform. Most of the progress in this very active field of research is based on just ∼
30 determi-nations of black hole mass, accumulated over the past decade. We illustrate how integral-fieldspectroscopy, and in particular our
OASIS modeling effort, can help improve the current situation.
Keywords. black hole physics, galaxies: elliptical and lenticular, galaxies: kinematics and dy-namics, galaxies: nuclei
1. Supermassive black holes and galaxy evolution
Fifteen years ago the existence of supermassive black holes (BHs) in galaxy nuclei wasconsidered an interesting possibility which had to be demonstrated. Nowadays BHs areregarded as the crucial ingredient for our understanding of how galaxies form. Key to thisparadigm shift was the launch in 1990 of the Hubble Space Telescope (HST). It all startedwith the realisation that the mass of the BH is correlated to other global characteristicsof the host galaxy as a whole. Initially a correlation M BH − L was found between the massof the BH and the luminosity of the host-galaxy stellar spheroid (Kormendy & Richstone1995; Magorrian et al. 1998). In 1997 the installation of the STIS long-slit spectrographon HST allowed the spatially-resolved kinematical observations to probe inside the radiusof the subarcsecond BH sphere of influence R BH ≡ GM BH /σ in nearby galaxies ( σ being the velocity dispersion of the galaxy stars). The increased accuracy in the M BH determinations contributed to the discovery of the much tighter M BH − σ correlation(Gebhardt et al. 2000; Ferrarese & Merritt 2000).Similar correlations were found between the M BH and respectively the galaxy concen-tration (Graham et al. 2001), the dark-halo mass (Ferrarese 2002; Pizzella et al. 2005),the galaxy mass (Marconi & Hunt 2003; H¨aring & Rix 2004) and the stars gravitationalbinding energy (Aller & Richstone 2007). The existence of these correlations is broadlyconsistent with a scenario in which the BH regulates the galaxy formation, during thehierarchical galaxy merging, by shutting off the conversion of gas into stars via a feedbackmechanism due to its powerful jet (Silk & Rees 1998; Di Matteo et al. 2005).1 Cappellari et al. Figure 1.
Updated M BH − σ relation. Open squares are values from the literature; open circlesare literature values, plotted against the luminosity-weighted σ within 1 R e determined from SAURON data (see Cappellari et al. 2006); filled circles are our new M BH OASIS determinations,against the
SAURON σ . For reference, the solid line is the relation of Tremaine et al. (2002), whilethe dotted one is from Ferrarese & Ford (2005).
2. Observational evidences and current limitations
Our understanding of the role of BHs in galaxy evolution is however far from complete.In fact the current observables do not uniquely constrain the models, which depends ona number of assumptions. This is due in part to the relatively small number of secure BHmeasurements: in a decade of high-resolution observations and models only ∼
30 valueshave been obtained (see Ferrarese & Ford 2005, for a recent review). It is remarkablethat so much progress was based on an extrapolation of so few BH measurements, andfor a biased galaxy sample, to the entire galaxy population!An additional complication comes from the fact that the above correlations provide arather indirect test for the models. A complementary and more direct approach to testthe BH formation paradigm consists of looking in nearby galaxy centres for the signaturesof the joint formation of the BH and the galaxy spheroid. Simulations of galaxy mergersshow in fact that, when two galaxies with a BH merge, the distribution of the stellarorbits in the resulting remnant, after the two BHs coalesce, is significantly different fromthe one of the progenitor galaxies. This is due to the ejection of stars passing, alongradial orbits, close to the resulting BH binary. Two observables signatures are expected:(i) the density profile should flatten inside the core radius R C , which is much larger than R BH and (ii) the orbital distribution should be biased towards tangential orbits inside R C (Quinlan & Hernquist 1997; Milosavljevi´c & Merritt 2001).Evidence for the formation of the density core, and its expected relation to the BHmass, was found from photometric observations (Faber et al. 1997; Milosavljevi´c et al.2002). The detection of the orbital signature is more complicated as it requires integral-field spectroscopic observations at high resolution (e.g. Cappellari & McDermid 2005).The orbital distribution, in a non-spherical stationary system, is in fact a function of thethree isolating integrals of motion and requires at least a three-dimensional observablequantity to be constrained. Nonetheless a first attempt at deriving the nuclear stellaranisotropy, for a sample of 12 galaxies, was done using long-slit STIS spectroscopy byGebhardt et al. (2003). They found a tendency for the orbits of the most massive objectsto be tangentially biased. However this appears to be true only well inside R BH and notup to R C as predicted by the simulations. Similar results were recently found for anothercarefully studied galaxy by Houghton et al. (2006) and Gebhardt et al. (2007). lack holes with OASIS Figure 2.
Anisotropy profiles from the dynamical models. In standard spherical coordi-nates σ R is the second moment of the velocity distribution along the radial direction, while σ T = ( σ θ + σ φ ) /
2. Different lines represent measurements at equally-spaced radial sectors in thegalaxy meridional plane, from the equatorial plane to the symmetry axis. The vertical dashedand solid thick line indicate the position of R BH and R C respectively. The values of R C and thelogarithmic nuclear slope γ ′ in the surface brightness were taken from Lauer et al. (2007).
3. The role of
OASIS integral-field spectroscopy
Ground-based and hig-
S/N integral-field observations of the stellar kinematics canovercome the limitations discussed in the previous section, and can be obtained for largegalaxy samples using large telescope mirrors. By tightly constraining the orbital struc-ture, dynamical models fitted to integral-field observations (i) can measure BH masseswith accuracy comparable or better than that obtained with HST spectroscopy, evenwhen R BH is not well resolved (Shapiro et al. 2006); (ii) do not suffer from the degener-acy in the recovery of the orbital distribution.For this we observed with the OASIS integral-field spectrograph a sample of 28 el-liptical and lenticular galaxies (McDermid et al. 2006). The galaxies were selected fromthe
SAURON sample (de Zeeuw et al. 2002), for which the needed large-scale integral-fieldkinematics is also available up to about one half-light radius R e (Emsellem et al. 2004).The OASIS observations complement the
SAURON ones by providing an order of magni-tude increase in the pixels density and a factor of two improvement in the median seeing,resulting in subarcsecond resolution.Here we report some results of the stellar dynamical models for an initial set of eightgalaxies from the
OASIS sample. We constructed the models using our axisymmetric im-plementation (Cappellari et al. 2006) of the orbital-superposition method (Schwarzschild1979), and we combined the
SAURON and