1. Dynamical signatures of mergers Amina Helmi

2. Dynamical signatures of mergers In collaboration with @Groningen: Facundo Gomez, Maarten Breddels, Laura Sales, Yang-Shyang Li elsewhere: Andrew Cooper, Michelle Wilson, Alvaro Villalobos, Simon White, Anthony Brown, Heather Morrison

3. The team @ Groningen

4. Outline Motivation: mergers and substructures –Why, where and lessons Models: stellar halo in Aquarius –Inner and outer halo –Dynamics of streams and predictions –How to retrieve history Models: thick disk simulations –Tests of formation –RAVE’s latest constraints Summary

5. Streams and Substructures Belokurov et al 2007

6. Lessons from streams: MW’s gravitational potential Streams in halo excellent probes: –groups of stars on parallel orbits moving under influence of the dark halo potential (Eyre & Binney 2009; Jin’s talk) Insight on the nature of dark matter: –density profile, shape and graininess –Also test of e.g. MOND First attempts with Sgr conflicting results (Helmi 2004; Law et al. 2005; Law & Majewski 2010; Penarrubia’s talk) Models of GD-1: strong constraints on global shape (disk + halo) and v LSR (Koposov et al. 2009) Johnston 1998 GD-1 stream in SDSS: dissolved cluster Grillmair & Dionatos 2006

7. Lessons from streams: Merger history Stars retain memory of origin: Orbits; chemical abundances; age distrib –trace internal evolution –mergers leave imprints in remnant: phase-space substructure Helmi, Cooper et al. (2010) Merger history: Tests of cosmological model How many objects? When? What characteristics? Assembly of Milky Way and characteristics of Universe at very early times

8. Where to look for substructure?  Stellar halo  Most metal-poor and ancient stars in the MW  It can form from the superposition of disrupted satellites  Thick disk  Old and metal-weak stars  Disks are fragile: easily heated up by minor mergers thick disk stellar halo bulge thin disk

9. Stellar halo

10. Outer Stellar halo -Substructure common in the halo (SDSS, 2MASS…) -> mergers -> Broad, diffuse streams (large progenitors? …but beware of biases) overdensities -> nature not always clear Belokurov et al 2007

11. Some questions What are the properties of stellar halos and satellites in  CDM? How do they compare to the Milky Way or M31? What are models predictions for future surveys? What do we learn about the history of a galaxy from its stellar halo?

12. Aquarius halos coupled to SA models -1% most bound particles represent stars/stellar pops in these objects -Follow the history, their present-day location and dynamics Springel et al. 2008

13. Helmi, Cooper et al. 2010Cooper et al. 2010 Stellar halo formation in the Aquarius simulations

14. Aquarius on the sky Helmi, Cooper et al. 2010 Inner halo (d < 10 kpc): very smooth (triaxial in shape) Substructure apparent at d > 10 kpc and dominant at d > 30-50 kpc Anisotropically distributed (coherent in dist): infall pattern!

15. Stellar halos at d ~ 10-30 kpc Broad/diffuse features dominant Narrow streams also present Sgr and O-stream visible in the Aq-A sky! Helmi, Cooper et al. 2010

16. The Aquarius “field of streams” Helmi, Cooper et al. 2009

17. The Aquarius “field of streams” Helmi, Cooper et al. 2009 Streams from different progenitors may follow similar paths on sky 24 different objects contribute to this field M * ~ 10 3 M sun (orange) upto 9 x 10 7 M sun (purple)

18. Stellar halo Distribution on the sky: anisotropic at large radii New surveys likely to find (most) substructures in same regions as SDSS Reflects infall pattern of material; relation to large scale environment around galaxy Inner stellar halos are older: by age and dynamically –contain unique clues about early galactic history

19. Inner Galaxy  More interesting:  location of large majority of stars  Mixing timescales are short  Debris no longer spatially coherent  Memory is retained in kinematics Velocity space Helmi et al. 1999 Velocity space A galaxy disrupted very early on in our own backyard

20. Inner/nearby stellar halo in Aquarius Few objects contribute here: 75% of stars near Sun from 3-5 parents Memory in kinematics (despite “chaotic” build-up) –~ 400 streams crossing Solar neighbourhood –2/3 of stars in volume are in “massive” streams –1/3 of stars on chaotic orbits -> streams no longer identifiable Gomez et al. 2010c

21.  Better spaces to look for clustering -> conserved quantities: Energy vs Lz  Lumps/structures visible for each accreted galaxy “Conserved” quantities/ Integrals of motion Gomez et al. 2010c

22. Frequency space Stars in a portion of a stream have very similar orbital frequencies (periods) As streams cross in space, they may still be recognized by their velocities or their frequencies Johnston 1998

23. - Streams sharply defined on a regular grid - Characteristic separation  depends on time since accretion: N str = [  (t)/ 2  ] = [  t /2  ] and  =   /N str t acc =  /2  - Timescale easily estimated with Fourier analysis techniques - works well (within 15 – 25%) even in time-dependent, or live potentials Frequency space: time since accretion Velocities Energy-L z Frequencies Gomez & AH 2010

24. Summary Accretion likely important in build-up of stellar halo –Objects? how many? properties? masses, star formation and chemical histories? Cosmological models of formation –predict substructure at all radii kinematic at r < 10-20 kpc; hundreds of streams near the Sun Spatially coherent at large distances and anisotropic on sky ->infall pattern –Qualitatively in agreement w/observations Present and Future: many ongoing surveys and Gaia –Substructure from ancient events expected –More difficult to detect (IoM, frequencies)… but early epochs! –Gaia capable of unraveling accretion history of the Galaxy

25. Thick disk

26. Thick disk formation simulations Thick disk can result from heating by minor merger of pre-existent disk Villalobos & Helmi 2008

27. Thick disk: can we find the debris? How to tell red from white “stars”? No spatial correlations (after few Gyr) Villalobos & Helmi 2008 Volume around the Sun: –Velocity distribution distinct from disk –Characteristic “banana” shape

28. Radial velocity trends Pre-Disk stars: V los ~ sin(longitude) –reflects most stars still on circular orbits –Behaviour generic all models w/pre-disk Stars from satellite deviate from this pattern –Can be identified in the wings of the distribution (Gilmore et al. 2003) Villalobos & Helmi 2009 Thick disk “stars” near the Sun (r < 2 kpc)

29. Other formation scenarios for Galactic thick disk Accretion: purely from disrupted satellites –Satellites accreted on preferential directions (Abadi et al. 2003) Gas-rich mergers –Intense star formation phase (gas unsettled; Brook et al. 2004) Radial migration –(Resonance) scattering by spiral arms stars from inner Galaxy to solar neighbourhood (Sellwood & Binney 2002; Schoenrich & Binney 2009; Roskar et al. 2008) Density  (z) Height above the plane

30. A powerful test of formation: orbital eccentricity Stars’ orbits: pre-existing disk: fairly circular from satellite: eccentric Generic test for any model of formation: e-distribution 1. Whole disk by accretion > Flat 2. Pre-existing disk > Pronounced peak at low e Secondary peak at high e (if by merger event) Sales et al. 2009

31. RAVE constraints on the eccentricity distribution for thick disk stars Full phase-space coordinates Sample of stars at 1 z = [1, 3] kpc; R < 3 kpc distance information Modified Breddels et al. (2010) for DR3 (version 08/2008; 135,000 stars) Zwitter et al. (2010); Padova set Red clump stars (0.6 < J-K s < 0.7; 1.5 < log g < 2.7; M V = -1.61) Kinematics: good agreement with literature Wilson et al. 2010

32. Eccentricity distribution e-distribution triangular in shape: peak at e ~ 0.2, tail towards high e –Shape is very similar for all samples (KS consistent) –As expected, slightly lower number of high ecc stars in red clump (no halo) Integrate orbit in Galactic potential to derive e-distr Wilson et al. (2010)

33. Eccentricity distribution and models Prominent peak at low ecc rules out accretion model –Most thick disk stars formed in-situ Shape appears most consistent w/merger model –Heating model shows second peak (not present in data) –Migration model more Gaussian than apparent in data Sales et al. 2009 Wilson et al. (2010)

34. Summary -Comparison of simulations to test formation of thick disk: ecc-test -Dataset use: RAVE stars at 1 < |z| < 3 kpc -RGB stars using Breddels et al (2010); from Zwitter et al. (2010) -Red clump stars -All kinematics consistent, also w/literature values -Ecc-distribution is triangular in shape: prominent peak at low ecc (expected if most stars formed in-situ) -Inconsistent w/pure accretion model -Subtleties to rule out other models: -None made to reproduce MW (thick disk) -Heating via minor merger: second peak location depends on IC -Migration: does not have a stellar halo