Finding Galactic- halo substructure in the Gaia data Amina Helmi - PowerPoint PPT Presentation

Finding Galactic- halo substructure in the Gaia data Amina Helmi

Stellar halo: treasure trove of merger relics t = 1 Gyr t = 2 Gyr Cosmological model’s characteristic: hierarchical • growth: mergers Disrupted galaxies/debris naturally in a stellar halo: • ! merger signatures: Substructures and tidal streams t = 3 Gyr t = 4.5 Gyr Questions: • Were mergers important for galaxies like MW? • How often and when did they happen? • What were the building blocks? • t = 8 Gyr t = T now today • Stars are “fossils” • Motions, ages, chemical composition trace origin • Substructures pinpoint to merger debris • Probe force field ! mass (gravity) snapshots: J. Gardner

Testing the cold dark matter paradigm Is this “picture” correct? Credit: V. Springel • Are galaxies like the Milky Way and its nearest neighbours embedded in dark matter halos like those predicted by the cosmological model?

Testing the cold dark matter paradigm Is this “picture” correct? • Are galaxies like the Milky Way and its nearest neighbours embedded in dark matter halos like those predicted by the cosmological model?

Testing the cold dark matter paradigm Is this “picture” correct? Credit: V. Springel Are galaxies like the Milky Way and its nearest neighbours embedded in dark matter halos like • those predicted by the cosmological model? • How much dark matter is there? – how is it distributed? – what is the dark matter? • Is Gravity correct?

A stream in a dark halo with substructure Springel et al. 2008 Granularity: Hundreds of thousands dark clumps if dark matter particle is cold

The accretion history unveiled so far: The Galactic halo from SDSS/PanStarrs North Galac?c Cap Belokurov et al. 2006 + Outer halo: R > 20 kpc • Clear evidence of substructure • Limited to high-surface brightness features (progenitors/time of events) • Qualitatively consistent with expectations from Galac?c An?centre Slater et al. 2014 + Λ CDM (Helmi et al. 2011; Deason et al. 2014)

PanSTARRS 3 π survey Many narrow streams mapped/discovered. Bernard et al. (2016)

The relevance of kinema?c informa?on

The relevance of kinema?c informa?on Proper motions from Gaia DR2 (April 2018) vt=200 km/s ! μ ~1 – 5 mas/yr (d ~ 10 – 40 kpc) (G ~ 17) expected error: σ μ ~ 0.1mas/yr ! Trace substructures, outlier removal, and map MW potential

Not all substructure is accreted – does pinpoint to interactions and mergers Li et al 2017 Deason et al 2014 Gomez et al. (2016, 2017) Price-Whelan et al 2015

Nearby halo https://www.astro.rug.nl/~ahelmi/research/gaia/movie.html conserved quantities Memory of origin: retained in the motions Helmi & de Zeeuw 2000 energy 100s of streams should cross Sun’s vicinity " So far.. not much evidence (small samples) " How to find more? ! Clustering in conserved quantities " angular momentum

Jovan Veljanoski Construction of a halo sample: TGAS x RAVE Maarten Breddels • TGAS dataset is significant improvement, but need full ! ~ 200,000 stars in common phase-space information ! cross-match to RAVE survey • RAVE: spectra for 500k stars in southern sky: v los , [M/H], spectrophotometric distance/parallax (with TGAS priors, McMillan et al. 2017) • Metallicity cut [M/H] cal < -1 dex 600 600 to select preferentially halo 400 400 • Remove stars with disk-like 200 200 v y (km/s) v y (km/s) kinematics 0 0 • 2-Gaussian decomposition − 200 − 200 ! sample of 1307 − 400 − 400 genuine halo stars − 600 − 600 − 600 − 400 − 200 0 200 400 600 − 600 − 400 − 200 0 200 400 600 v x (km/s) v z (km/s) Helmi, Veljanoski, Breddels et al. (2017), Veljanoski et al. (in prep)

Statistical tests and searches of substructure Models predict • several hundred moving groups or 20 streams in Solar Neighbourhood ! we search for excess clustering in velocity space with a correlation function 10 0 • substructure to be more easily apparent in Integrals of Motion space ! we characterise the distribution, 10 degree of clustering and establish significance 20

Velocity correlation function 1 . 5 Helmi, Veljanoski et al. (2017) 1 . 4 1 . 3 Helmi et al. (2017) 1 . 2 1+< ξ > 1 . 1 1 . 0 0 . 9 0 . 8 0 100 200 300 400 500 600 700 velocity difference | v i − v j | (km/s) • Very significant excess of pairs in data compared to random/smooth for Δ < 20 km/s, 5.5 σ (120 pairs of stars in excess) – for 20 < Δ < 40 km/s: 8.8 σ (328 pairs in excess) – • Also for very large separations, there is a significant excess

The amount of substructure: comparison to cosmological simulations Helmi et al. (2017) Cooper et al. (2010) • Simulations of halos purely built via accretion show excess on small and large separations of similar amplitude – some variation from halo to halo ! Milky Way halo consistent with being fully built via accretion

Integrals of motion - space retrograde less-bound 0 Helmi, Veljanoski et al. (2017) − 25000 − 50000 E (km 2 /s 2 ) − 75000 − 100000 − 125000 − 150000 − 175000 − 200000 − 4000 − 3000 − 2000 − 1000 0 1000 2000 3000 4000 L z (km/s kpc) ! very retrograde motions: 73% of all stars (for E > -1.3x10 5 km 2 /s 2 ) In randomised (re-shuffled) smooth distributions the probability of having so many loosely bound counter-rotating stars is < 0.1%

Integrals of motion – space L z (km/s kpc) − 2000 − 1000 0 1000 2000 − 100000 • Statistical comparison to smooth − 6 distributions allows identification of − 120000 overdensities in E vs Lz − 7 21 − 140000 E (km 2 /s 2 ) scaled E • Structures at Lz ~-500 km/s kpc 18 − 8 − 160000 could be related to OmegaCen 14 13 debris (Dinescu 2002) 9 − 180000 − 9 7 6 • VelHel-6: stars with disk-like 3 kinematics but counter-rotating − 200000 − 10 − 3 − 2 − 1 0 1 2 3 scaled L z Helmi, Veljanoski, Breddels et al. (2017), Veljanoski et al. (in prep) see also Myuoung et al. (2017)

The retrograde halo in context • Not common in cosmological simulations (e.g. Illustris; Vogelsberger et al. 2014) • Less than 1% of MW-mass galaxies have more than 60% of the less bound stars on Helmi et al. (2017) retrograde orbits (here defined as r > 15 kpc)

Chemical abundances − 0 . 8 − 25000 − 1 . 0 − 50000 − 1 . 2 − 0 . 8 − 130000 − 75000 − 1 . 0 − 1 . 4 − 140000 [M/H] (dex) E (km 2 /s 2 ) − 1 . 2 − 150000 − 100000 − 1 . 6 [Fe/H] (dex) − 1 . 4 E (km 2 /s 2 ) − 160000 − 125000 − 1 . 6 − 1 . 8 − 170000 − 1 . 8 Helmi et al. (2017) − 180000 − 150000 − 2 . 0 − 2 . 0 − 190000 − 2 . 2 − 175000 − 2 . 2 − 200000 − 2 . 4 − 2 . 4 − 200000 − 210000 − 2000 − 1000 0 1000 2000 L z (km/s kpc) − 4000 − 3000 − 2000 − 1000 0 1000 2000 L z (km/s kpc) Veljanoski (in prep) • C. Boeche chemical pipeline, not all stars have detailed abundances (SNR > 20, McMillan sample) • Stars with L z < 0 on average lower metallicity, both [M/H] and [Fe/H] • May be some clumpiness (?)

Chemical abundances: substructures 1 . 0 [Mg/Fe] (dex) 0 . 5 0 . 0 − 0 . 5 Retrograde halo Stars not in substructures Substructure 3 Substructure 6 Substructure 7 − 1 . 0 1 . 0 [Mg/Fe] (dex) 0 . 5 0 . 0 − 0 . 5 Substructure 9 Substructure 13 Substructure 14 Substructure 18 Substructure 21 − 1 . 0 − 3 − 2 − 1 0 − 3 − 2 − 1 0 − 3 − 2 − 1 0 − 3 − 2 − 1 0 − 3 − 2 − 1 0 [Fe] (dex) [Fe] (dex) [Fe] (dex) [Fe] (dex) [Fe] (dex) Probabilities drawn from Helmi et al. (2017) 3 . 0 6 Retrograde halo Stars not in substructures Substructure 3 p < 1% Substructure 6 p < 1% Substructure 7 p < 0.1% p < 0.1% overall population can be 100 8 8 2 . 5 5 Number of stars 80 relatively small Veljanoski (in prep) 6 6 2 . 0 4 60 1 . 5 3 4 4 40 1 . 0 2 2 2 20 0 . 5 1 Similar behaviour in e.g. 0 0 0 0 . 0 0 6 3 . 0 6 2 . 0 1 . 0 p < 1% Substructure 9 Substructure 13 Substructure 14 Substructure 18 p < 6% Substructure 21 [Mg/Fe] 5 2 . 5 5 0 . 8 Number of stars 1 . 5 4 2 . 0 4 0 . 6 3 1 . 5 3 1 . 0 Generally limited by 0 . 4 2 1 . 0 2 number of stars 0 . 5 0 . 2 1 0 . 5 1 0 0 . 0 0 0 . 0 0 . 0 − 2 − 1 0 − 2 − 1 0 − 2 − 1 0 − 2 − 1 0 − 2 − 1 0 [Fe/H] (dex) [Fe/H] (dex) [Fe/H] (dex) [Fe/H] (dex) [Fe/H] (dex)

PI – GA surveys: Vanessa Hill

Clustering in integrals of motion (e.g. actions) maximal for right gravitational potential (DR2) Sanderson et al. (2014, 2016)

Summary • Halo substructure is useful for dynamics (dark matter) and merger history • Photometric surveys mapped large structures in the outer halo • TGAS x RAVE: excess of close velocity pairs and IoM space rich in substructure – at level consistent with cosmological simulations of halos purely built via accretion – Less-bound halo stars predominantly retrograde (significance > 99.9%) – Many overdensities for more bound halo • What’s coming: – DR2 (April 2018) will be fantastic: proper motions and parallaxes for 1 billion stars! – 4MOST and WEAVE: spectroscopic follow – Characterization of the stars in the structures found, e.g. chemical abundances, ages – Numerical simulations for orbits, infall times, link to other structures in the halo – constraints on characteristic mass and scale of Milky Way