1. Simulations of galaxy mergers Thorsten Naab MPA, Garching Evolution of galaxies, their central black holes and their large scale environment Potsdam, September, 2010

2. Minor-axis tubes Box orbits Outer major-axis tubes Inner major-axis tubes Boxlets Jesseit et al. 2005/2007 Orbital content of galaxies: The stellar backbone

3. Barnes & Hernquist 1996, Jesseit et al. 2005, Naab et al. 2006, Hofmann et al. 2009, Bois 2010 etc. Influence of gas on the stellar component Gas infall, even in small amounts has a strong influence on the central potential towards axisymmetry Significant to dramatic change in orbital families Strong effect on photometric (i.e. isophotal shape) and kinematic observables (LOSVD) Most early-types require dissipation at assembly Results are general, also valid for dark matter halos

4. Rapid BH growth and starburst at final equal-mass merger (Springel, di Matteo, Johansson, Hopkins, etc) Post merger star formation and BH accretion quenched by AGN feedback, M BH - relation (di Matteo, Springel et al., 2005) Input for numerous semi- analytical, empiricalmodels on BH growth and evolution Black hole formation and feedback

5. Shock heating exceeds PdV heating: du/dt traces shocks Shock induced star formation Barnes 2004, MNRAS, 350 Star formation is more extended and in better agreement with observations Model for NGC4676, The Mice Star formation, heating and feedback

6. (Whitmore et al. 2010) Karl et al. 2010, ApJL The Antennae galaxies

7. o Idealized initial conditions and orbits, large parameter space o No cosmological boundary conditions, i.e. realistic merger histories, gas inflow, gas fractions & internal structure (dry, wet, hot, cold) o Major mergers of minor importance for stellar mass production (tight SFR-mass relation, talk by Lilly) o Major mergers of major importance for black hole growth in models ( e.g. Crotton, Somerville, Hopkins, Bonoli, and many more) o Possibly no causal connection between galaxy growth and BH growth – CLT (Peng et al. 2007, Hirschmann et al. 2010, Jahnke et al. 2010) o Are local disks and their progenitors the progenitors of local elliptical galaxies? Just a minority of low mass early-types (Naab & Ostriker 2009) or more than just a minority (Hopkins et al. 2009 papers) o Significant fraction (40-60%) of the total mass (in dark matter) is accreted smoothly (Genel et al. 2010, ApJ, Wang et al. 2010arXiv1008.5114W) o Limited predictive power for evolutionary models of galaxy formation, e.g. strong size evolution of massive early-type galaxies A few complications…

8. The binary merger-tree o Typical contribution of stellar mergers (>1:4) in massive galaxies since z=2 is 100%

9. Merger-trees from zoom cosmo-hydro-simulations o Typical contribution of stellar mergers (>1:4) in massive galaxies since z=2 is 40% - 50% Hirschmann, Naab, et al., in prep

10. The stellar mass budget o Agreement with observations for low mass galaxies is worse than for high mass galaxies (Wechsler et al., Guo et al. 2010, Moster et al. 2010, Trujillo-Gomez et al. 2010) o IMF? AGN feedback? Stellar mass loss? Star formation driven winds? etc… Oser et al. 2010 submitted

11. o Size evolution for massive early-type galaxies proportional to (1+z) α, α=-1.22 (Franx et al. 2008), -1.48 (Buitrago et al. 2008), -1.17 (Williams et al. 2010) o Mild evolution of 10 11 M ellipticals from 240km/s at z1.6 (240km/s) to 180 km/s at z=0 (Cenarro & Trujillo 2009) from stacked spectra of 11 GMASS ellipticals (Cimatti et al. 2008) o High velocity dispersion of a z=2.168 galaxy – 512 km/s indicates high dynamical mass consistent with mass (2×10 11 M ) and compactness (0.78 kpc) of photometric data o Add large galaxies to the population: faded spirals? o Grow the population by major/minor mergers, expansion and other effects? Minor mergers are favored (Bezanzon et al. 2009, Hopkins et al. 09/10, Naab et al. 2009) Size and dispersion evolution since z2

12. Minor mergers and the virial theorem Dispersion can decrease by factor 2 Naab, Johansson & Ostriker 2009 M f = (1+ )*M i and assume =1, e.g. mass increase by factor two, and varying dispersions… Radius can increase by factor 4 Density can decrease by factor 32 more complex: gas, dark matter, dynamics

13. Size evolution in a high resolution simulation Naab, Johansson & Ostriker 2007/2009 o In-situ stars form a compact high density stellar system o Accreted stars make extended outer system (see e.g. Hopkins et al. 2009 and talk by Tal) o z3: M=5.5*10 10 M ρ eff = 1.6*10 10 M /kpc 3 σ eff = 240 km/s o z0: M=15*10 10 M ρ eff = 1.3*10 9 M /kpc 3 σ eff = 190 km/s o Consistent with accreted mass being responsible for size increase e.g. Dekel, Ocvirk, Keres, Kravtsov, Brooks and more

14. The origin of stars in galaxies o Stellar origin diagrams indicate when and at which radius a star ending up in a present day galaxy was born o In massive galaxies most stars are made at high redshift in-situ in the galaxy and even more ex-situ outside the galaxies virial radius with a low fraction of in-situ formation at low redshift o Lower mass galaxies make a larger fraction of stars at low redshift Oser et al. 2010, submitted

15. The two phases of galaxy formation o 45 simulations stacked in mass bins o Early assembly is dominated by in-situ formation, more so in massive galaxies (6 > z > 3) o Low mass galaxies assemble half their mass by in-situ formation o The late assembly of massive galaxies is dominated by accretion (up to 80%) of stellar system (3 > z > 0)

16. The origin of stars in galaxies o Ex-situ stars form at high redshift (z=4) o Ex-situ stars are accreted below z 1 at high rates for massive galaxies o In-situ stars start forming at high redshift and continue to contribute to the growth of low mass galaxies until the present day o Galaxies assemble half their mass at z 1 o More massive galaxies are older downsizing (Keres et al. 2009)

17. … and some more consequences o More massive galaxies had more accretion o Galaxies formed by more dissipation have more concentrated dark matter halos o Galaxies formed by more dissipation have denser galaxies o Mass-size relation is driven by accretion

18. The rapid size evolution of spheroids Naab et al. 2010 Good agreement with observed strong size evolution for massive early-type galaxies proportional to (1+z) α, α=-1.22 (Franx et al. 2008), - 1.48 (Buitrago et al. 2008), -1.17 (Williams et al. 2010)

19. Compact massive ellipticals at z2 Galaxies at higher redshift have higher velocity dispersions but move onto the local correlations – detailed merger analysis is ongoing Naab, Oser, Ostriker, Johansson 2010

20. The fundamental mass plane from strong lensing o 53 field early-type strong gravitational lens galaxies from the Sloan Lens ACS (SLACS) survey (Bolton et al. 2008) o Estimate of the total mass (dark+gas+stars) within r e /2 o Representative for early-type galaxies with M * > 10 11 M (Auger et al. 2009) o Dynamical mass is a good proxy for the true masses Our simulations agree well with the observed lensing mass plane, total mass profile is isothermal! Jesseit, Naab et al. 2010

21. Central dark matter fractions o The average central dark matter fraction agrees with estimates from lensing and dynamical modeling o Reasonable models for the mass distribution in massive early-type galaxies at z=0 (see talk by Barnabe) Auger et al. 2009

22. o Idealized merger simulations have limited predictive power with respect to evolution o The formation of elliptical galaxies is a two phase process o The cores (kpc) of early-type galaxies form at 2 < z < 6 by dissipation/cold gas flows (monolithic collapse) (Keres et al. 2005, Dekel et al. 2009, Hopkins et al. 2009) and by merging of smaller structures of stars/gas at the same time as the halo is building up (e.g. Hopkins et al. 09/10, van Dokkum et al. 2010) o Ellipticals grow at 0 < z < 3 by accretion/mergers (dry mergers) of old stars ( 10 kpc) - all mass ratios, minor mergers dominate, major mergers have a more dramatic effect o Effect of accretion can explain the mass-size relation and can be the key to the observed strong size evolution!!! o Failure of the major merger hypothesis?! o Simulated central galaxies follow the fundamental mass plane with reasonable central dark matter fractions – reasonable models for ellipticals ConclusionsConclusions