The AMUSE surveys: Down-sizing in black hole accretion Elena Gallo | University of Michigan B. Miller & K. Gultekin (U. of Michigan), T. Treu & R. Antonucci (UCSB), J.-H. Woo (Seoul U.) 12 Years of Science with Chandra, 218th AAS

Black holes and bulge relations Empirical correlation between super-massive black hole mass and bulge properties (central stellar velocity dispersion, luminosity, Sersic index, etc) suggest deep evolutionary link However, poorly constrained at the low mass end Gultekin+09

Evolution of quasar LF Quasar density peaked near redshift z~2 Lower-luminosity quasars peak at lower redshifts (“downsizing”) Quasars accrete/radiate at 0.01-1 Eddington, but only for relatively short lifetimes (~108 yr) Ciotti & Ostriker 07

Low-level SMBH activity Post-quasar phase: highly sub-Eddington, radiatively inefficient accretion (Sgr A*: Lx/LEdd~1e-9) Simulations indicate “radio mode” feedback at low redshift is required to quench cooling and inhibit star formation (e.g., to match observed galaxy colors) Croton+06

X-rays: AGN vs. ‘inactive’ galaxies X-rays from inactive galaxies: ROSAT effectively sensitive down to 1e40 erg/sec for nearby galaxies Chandra bridges the gap between active (>1E-2 L_Edd) and (formally) inactive galaxies Sub-arcsec angular resolution is key Pellegrini 05

The AMUSE surveys: science goals Two Large Chandra Programs (~1Msec) to bridge the gap between AGN and formally inactive galaxies. Specifically, to: Investigate highly sub-Eddington (Lx/Ledd<1e-5) activity in local super-massive BHs Quantify active fraction as a function of BH/host mass, corrected for ‘Eddington incompleteness’ Constrain the local BH occupation fraction Control for environmental effects on the nuclear activity (cluster members vs. isolated galaxies)

AMUSE-Virgo: the survey Targets 100 Virgo early type galaxies which compose the HST ACS Virgo Cluster Survey (VCS, Cote’ et al 2004): 84 new targets with Chandra ACIS-S (454 ksec Cycle 8, PI: Treu) + 16 archival (>1 Msec) Complete down to Ledd for a 3 Msun object 57 new targets with Spitzer MIPS (9.5 hr) + 43 archival Archival HST ACS g- & z-band (100 orbits) Nuclear black hole mass distribution

Black holes, star clusters & LMXBs 1. Diffuse gas emission modeled and removed: search for nuclear BH limited to hard X-rays (>2keV) 2. Astrometry matched to SDSS, positional accuracy: 0.2-0.5 arcsec 3. Contamination from Low-Mass X-ray Binaries (LMXBs) addressed quantitatively based on X-ray luminosity function of LMXBs: Gilfanov04) in the absence of nuclear star Sivakoff+ 2007 in the presence of a nuclear cluster

AMUSE-Virgo:Nuclear X-ray census 32/100 show a nuclear X-ray source 51/100 show a massive nuclear star cluster (HST) 6/100 show both a nuclear X-ray source and a star cluster After accounting for LMXB contamination: 24-34% of the galaxies host an active super- massive black hole (95% C.L.) Next: ACTIVE FRACTION as a function of host stellar mass and MBH

AMUSE-Virgo: Active fraction Active fraction raises with host stellar mass Mostly due to ‘Eddington incompleteness’: Dealing with `Eddington-limited’ sub-samples results in no evidence that the fraction of active black holes depends on host mass (see Ho+97 Kaufmann+03, Decarli+07, Seth+08,10) (Gallo+10)

AMUSE-Virgo: Down-sizing Assume: Log( LX,38 )=A + B log( M BH,8) Bayesian code handles limits, errors; Uncertainties: 0.44 dex on MBH to be 0.44 dex, 30% on Lx ; include intrinsic scatter Priors: MBH mass function power- law index -1, rotational invariance for slope Most likely parameter value taken to be median of 10000 random draws from posterior distribution DOWN-SIZING: Lx/Ledd increases with decreasing BH mass ‹LX / MBH ›MBH-0.62 log LX log (MBH)

AMUSE-Virgo: Summary 32/100 nuclear X-ray sources ; 51/100 nuclear clusters ; 6/100 hybrids Bona fide active black holes (after LMXB contamination assessment): between 24-34% host an accreting black hole. Strong lower limit to occupation fraction in the local universe. AVERAGE Lx/Ledd DECREASES WITH INCREASING BLACK HOLE MASS (Gallo+08,10) PLUS : Spizter MIPS results: absorption, dust reprocessing etc. (Leipski+ Paper III, submitted) Off-nuclear X-ray source population (Katolik+ Paper IV, in prep.)

Influence of environment Relative to their cluster counterparts, local field early- type galaxies tend to: Have a lower incidence of past major mergers Face reduced ram pressure stripping, but outflows less confined Generally have more cold gas (e.g., higher HI content) along with younger stellar populations

AMUSE-Field Chandra snapshots of 75 field spheroids (PI Gallo, Cycle 11, 479 ks), plus 25 archival within 30 Mpc Sensitive down to the Eddington limit for a 6 Msun BH Sample drawn from HyperLeda catalog, with following criteria: Spheroidal, MB<-13 v<1800 km/s and d<30 Mpc b>294o or b<230o (no Virgo or Fornax) l>30o or l<-30o (avoids absorption from Galactic plane)

AMUSE-Field: Active Fraction 34/75 X-ray nuclei Detection fraction increases with Mstar Assuming LMXB contamination as for Virgo (red), ~40% X-ray active fraction (filled histogram, upper) Comparable to Virgo; (dashed lines, lower) B.Miller+ in prep.

AMUSE-Field: Stacking Stacked 0.3-7 keV images of snapshot observations lacking nuclear X-ray detection (after removing off-nuclear X-ray sources) 291 ks of net exposure 31 net counts within 5’’ aperture (2.8 sigma) log(Lx) = 37.7, consistent with LMXB origin B.Miller+ in prep.

Virgo vs Field: Eddington ratios Lx/LEdd versus MBH and Mstar for field (blue, purple) and Virgo (red) Lx/LEdd declines with increasing MBH, Mstar Field sample includes relatively more low MBH, Mstar objects: all comparisons based on weighted subsamples consistent with Virgo

Virgo vs Field: Down-sizing Assume: Log( LX,38 )=A + B log( M BH,8) Bayesian code handles limits, errors; Uncertainties: 0.44 dex on MBH to be 0.44 dex, 30% on Lx ; include intrinsic scatter Priors: MBH mass function power-law index -1, rotational invariance Most likely parameter value taken to be median of 10000 random draws from posterior distribution DOWN-SIZING IN BOTH SAMPLES: Lx/Ledd increases with decreasing BH/host mass B.Miller+ in prep.

Virgo vs Field: Down-sizing Slopes consistent between the field and Virgo samples; also holds for detections only, as well as for the weighted subsamples Intercept slightly higher for the field sample B.Miller+ in prep.

The AMUSE-surveys Two Large Chandra Programs (Cycle 8 & 11) to investigate the rate of highly sub-Eddington accretion onto nuclear super- massive BHs in nearby early type galaxies: 200 objects (100 in Virgo, 100 field, D<30 Mpc) Overall active fraction: 30% (Virgo), 40% (Field); does NOT increase with host mass Evidence for DOWN-SIZING: ‘smaller’ BHs shine closer to their Eddington limit The large-scale environment does not appear to influence strongly activity in local spheroids; Field galaxies show modestly enhanced nuclear X-ray emission, perhaps due to greater access to fuel supply