4MOST Consortium Survey 1: the Milky Way Halo Low-Resolution Survey

Surveys DOI: 10.18727/0722-6691/5120

4MOST Consortium Survey 1: The Milky Way Halo Low-Resolution Survey

Amina Helmi 1 Scientific context thousand stars out to 30–60 kpc, includ- Mike Irwin 2 ing tangential velocities with errors Alis Deason 3 Halo stars in the Milky Way spend signifi- between 5 and 50 km s–1 depending on Eduardo Balbinot 1 cant amounts of time at large distances their apparent magnitudes. 4MOST will Vasily Belokurov 2 from the Galactic centre, hence their tra- push the radial velocities much further, Joss Bland-Hawthorn 4 jectories are sensitive to the mass distri- down to G ~ 20 magnitudes at 1–2 km s–1 Norbert Christlieb 5 bution of the dark matter halo. Our survey, accuracy. Our survey will also measure Maria-Rosa L. Cioni 6 the 4MOST Consortium Milky Way Halo line-of-sight velocities for stars at the tip Sofia Feltzing7 Low-Resolution Survey, is therefore key of the red giant branch up to a distance Eva K. Grebel 8 for measuring the full mass distribution of of 250 kpc, and the rarer carbon and Georges Kordopatis 9 the Milky Way. In conjunction with the other asymptotic giant branch stars to Else Starkenburg 6 more local Milky Way Halo High-Resolution 1 Mpc. Not only will the estimates of the Nicholas Walton 2 Survey, and the detailed surveys of the total mass of the Galaxy be much more C. Clare Worley 2 Milky Way disc and bulge carried out in precise, but also its 3D distribution the 4MOST MIlky Way Disc And BuLgE (density and shape) will be within reach Low- and High-Resolution Surveys using various dynamical modelling 1 Kapteyn Instituut, Rijksunversiteit (4MIDABLE-LR and 4MIDABLE-HR), it techniques. Groningen, the Netherlands also aims to determine the complete 2 Institute of Astronomy, University of merger and assembly history of the Gal- Our survey will also detect faint stellar Cambridge, UK axy (see Christlieb et al., p. 26; Chiappini streams and any substructures in a com- 3 Department of Physics, Durham et al., p. 30 and Bensby et al., p. 35). bined abundance-kinematic space. University, UK In particular, the follow-up of thin streams 4 Sydney Institute for Astronomy, The Galactic halo contains large amounts will allow us to pin down the lumpiness University of Sydney, Australia of substructure at distances beyond of the dark matter halo and to constrain its 5 Zentrum für Astronomie der Universität 20 kpc, discovered with wide-field photo- nature (Erkal & Belokurov, 2016; Bonaca Heidelberg/Landessternwarte, Germany metric surveys more than 10 years ago et al., 2018). All of these measurements 6 Leibniz-Institut für Astrophysik Potsdam (Belokurov et al., 2006). At these large will thus lead to strong constraints on (AIP), Germany distances, debris is more spatially coher- cosmological models. 7 Lund Observatory, Lund University, ent because the mixing timescales are Sweden long. With the release of the Gaia astro- A significant by-product of determining 8 Zentrum für Astronomie der Universität metric data in April 2018 and the availa- the metallicity distribution function across Heidelberg/Astronomisches Rechen- bility of 6D phase-space information, it the halo will be the discovery of extremely Institut, Germany has been demonstrated that the inner metal-poor stars, which will be the focus 9 Observatoire de la Côte d’Azur, Nice, halo is dominated by merger debris from of further, more detailed follow-up, most France a single object as large as the Small likely using facilities other than 4MOST. Magellanic Cloud at the time of accretion (Helmi et al., 2018; Belokurov et al., 2018). The goal of this survey is to study the Beyond our Galactic neighbourhood, the Specific scientific goals formation and evolution of the Milky chemical characterisation of which is the Way halo to deduce its assembly his- focus of the Milky Way High-Resolution The specific goals of our survey are: tory and the 3D distribution of mass in Survey (Christlieb et al, p.26), there is still – Determining the density profile, shape the Milky Way. The combination of multi- much to learn. To make significant pro- and characteristic parameters of the band photometry, Gaia proper motion gress and pin down the full merger his- dark matter halo of the Milky Way — and parallax data, and radial velocities tory of the Milky Way we require spectro- including testing alternative theories of and the metallicity and elemental abun- scopic data of distant halo stars over a gravity such as Modified Newtonian dances obtained from low-resolution large portion of the sky that can be com- Dynamics (MOND) — and possibly their spectra of halo giants with 4MOST, will bined with the parallax and proper motion evolution in time. yield an unprecedented characterisa- information from Gaia. – Measurement of the perturbations tion of the Milky Way halo and its inter- induced by clumps on the spatial and face with the thick disc. The survey Most current mass estimates of the kinematic properties of cold streams will produce a volume- and magnitude- Milky Way have relied on small numbers leading to constraints on the mass limited complete sample of giant stars of tracers, and hence are likely subject spectrum of perturbers and on the in the halo. It will cover at least 10 000 to bias given the substructures present in nature of dark matter. square degrees of high Galactic latitude, the halo. The existing data sets contain, – Quantifying the amount of kinematic and measure line-of-sight velocities at most, 150 objects beyond 50 kpc substructure as a function of distance with a precision of 1–2 km s–1 as well (dwarf galaxies, globular clusters, halo and location on the sky. This will allow as metallicities to within 0.2 dex. stars; see, for example, the review article the discovery of substructures, new by Bland-Hawthorn & Gerhard, 2016). dwarf galaxies and other low surface Gaia will provide partial data for a few brightness objects, the characterisation

The Messenger 175 – March 2019 23 Surveys Helmi A. et al., 4MOST Consortium Survey 1

24h 18 h 12h 6h 0h Figure 1. Input catalogue density distribution for 30° the goal survey area of the Milky Way Halo Low- Resolution Survey. The stars shown here have been extracted from Gaia Data Release 2 (Gaia colla boration et al., 2018) and satisfy the following crite- 0° ria: –10 < G + 5 log10(proper motion) < 10, parallax – 2sparallax < 0.2; 0.55 < G–GRP < 0.8 magnitudes, and 15 < G < 20 magnitudes. This leads to a sample of approximately 2 million objects satisfying the –30° declination range and |b| > 20 degrees, after pruning out 5- and 10-degree radius regions around the Small and Large Magellanic Clouds, respectively. –60° Detailed coordination with the 4MOST Consortium Magellanic Cloud Survey (see Cioni et al., p. 54) will be carried out to ensure a smooth transition 1752 962 targets Equatorial between the surveys, as well as refinement of the target selection criteria. Notice the Sagittarius streams and other halo over-densities in this version 1 10 100 1000 of the input catalogue for our survey. The goal sur- Object density (N/deg2) vey area is –80 < dec < +20 degrees, however, it should be noted that the baseline survey for 4MOST is between –70 < dec < +5 degrees, and hence targets outside this footprint are less likely to be of their properties and their relation to lar clusters. This magnitude range over- observed (see Guiglion et al., p. 17). the build-up of the halo. laps at the bright end with the Milky Way – Characterisation of the metallicity and Halo High-Resolution Survey, which not elemental abundance distribution only provides cross-checks on derived (mostly magnesium and iron) through- stellar properties, but also ensures full out the halo, and also of each of the linkage between local and distant halo individual structures discovered. This populations. Since the halo density is in the number of such streams and this will yield enhanced samples of objects low, a wide-field instrument like 4MOST impacts the determination of the mass with very low metallicities or peculiar is essential to cover a large area in a distribution in the halo (Sanderson et al., elemental abundances for more reasonable time. The depth of our survey 2015). detailed follow-up, complementing the is perfectly suited to the goals, and 4MOST Consortium Milky Way Halo matches exactly the depth reached with Radial velocity estimates will generally High-Resolution Survey (Christlieb et al., Gaia with useful proper motion informa- be obtained from a combination of the p. 26) which focuses on the halo near tion. The strongest constraints on the Mg b triplet (Mgb) and the near-infrared the Sun. Such samples should constrain mass distribution come from streams cov- Ca II triplet (CaT) regions, which both the properties and yields of the first ering large angles on the sky, again push- contain sets of strong absorption lines, generation of stars (Population III). ing for large area surveys. For stars with also easily detectable in stars with low – Characterisation of the stellar halo-thick higher signal-to-noise (S/N > 25 per Å) metallicity. The velocity precision has disc interface from overlap with the useful constraints on the a-element to be 1–2 km s–1 in order to measure the 4MIDABLE-LR survey (Chiappini et al., abundances ([a/Fe] with error ≤ 0.1 dex), mean velocity in a field to approximately p. 30) with the aim of jointly constraining will be obtained and these are very 500 m s–1, which promises excellent con- the temporal assembly and evolution of important to further aid subdividing and straints on the mass distribution in the the thick disc and inner halo. characterising halo substructures (see Milky Way, and the dark matter granular- for example, Hayes et al., 2018; Helmi et ity imprinted in the velocities of stream al., 2018). stars (for example, Bonaca et al., 2018). Science requirements We note that accurate constraints on the The optimal streams for the determination Galactic potential can be obtained even The survey we propose will lead to a of the Milky Way gravitational potential if only limited proper motion information sample comprising on the order of (mass, shape, time evolution and granu- is available. If narrow streams are not 1.5 million giant stars in the halo (mainly larity) are thin and cold, and typically distributed isotropically on the sky, for K giant stars but also including the rarer originate in objects with stellar mass example as a consequence of infall along 5 A stars, particularly blue horizontal smaller than a few times 10 M⊙. Models filaments, it will be important to comple- branch stars, together with M giant stars of galaxy formation (Cooper et al., 2010) ment the kinematic maps of streams with and carbon stars) across the virial volume predict about 50–100 such thin streams those from field stars. of the Milky Way, with kinematics precise observable down to a G magnitude ~ 20, to 1–2 km s–1 and overall metallicities across a 10 000 square degree region precise to ≤ 0.2 dex. This means observ- of the sky, and there may be many more Target selection and survey area ing all giant stars in the halo in the mag from disrupted globular clusters (Bonaca nitude range 15 ≤ G ≤ 20 magnitudes, et al., 2014; and see Malhan et al., 2018 The density of the stellar halo, and hence including those in the lower density for the first detections with Gaia DR2). of the kinematic tracers, is low. Moreover, regions of halo dwarf galaxies and globu- A smaller area leads to a significant loss since the density profile of the halo drops

24 The Messenger 175 – March 2019 rapidly, these tracers are rare at large 1000 distances. The expected average source density is 100–200 stars per square degree at a G magnitude ~ 20. A large Mgb CaT survey area (minimum 10 000 square 800 degrees; goal 150 000 square degrees) is also needed to find the rare, precious red giant branch stars near the virial radius of the Milky Way. Candidate halo giant 600 stars will be selected on the basis of Gaia s)

photometry, parallax and proper motion (ADU information possibly supplemented with x Flu photometry from ground-based imaging 400 surveys (DES, SDSS, VST, PanSTARRS). Furthermore, we will also specifically target stars in streams known at the time of the survey, potentially down to the 200 main-sequence turnoff to increase the number of objects and provide tighter constraints on the dynamics of the stream. Our aim is to target every star 0 lying in a cold stream area in the available 5000 5200 5400 8400 8600 8800 Wavelength (Å) Wavelength (Å) magnitude range, since we expect (field) contamination at a level of 70%–100%, depending on how the stars are pre- Spectral success criterion and figure of Figure 2. Example of a 1D-extracted spectrum of selected. Although Gaia will yield some merit a halo K giant star (g = 18.73, i = 17.51 magnitudes) from a 4MOST simulated exposure of 3 × 1020 sec- useful prior constraints on the distances onds in dark conditions. The stellar Mgb and CaT of stars in the inner halo, photometric The spectral success criterion of our sur- lines are blueshifted from their reference values by distance estimates will need to be com- vey will not be binary (i.e., passed/failed), the high negative heliocentric velocity (–267 km s–1) bined with the Gaia data both to improve but “fuzzy” since spectra with S/N of the star. The average S/N in the continuum in these regions is around 25 Å–1. According to the the inner halo distances and to provide below the boundary value are still useful study by the Galactic pipeline working group, a distance proxy for the outer halo. Here for deriving radial velocities, albeit with these can reach a precision of [Fe/H] ~ 0.15 and in particular, the 4MOST spectra will be a lower precision. For computing the [a/Fe] ~ 0.1 dex. key. The spectra will be coupled with spectral success value, we will employ a extant broadband photometry to derive non-linear function f(S/N) that maps photometric distances and accurate the S/N of each spectrum onto the value radial velocities for the halo star samples range [0,1] and is defined to be 0.5 if the (for example, Xue et al., 2014). The same S/N = 10 per Å in the continuum in the survey will also make full use of the Mgb and CaT regions (compare Figure 2). combination of kinematic and positional information combined with chemical, The survey figure of merit (FoM) is chosen i.e., [Fe/H] and [a/Fe], signatures to char- to yield a high completeness per field acterise substructures at large radii, be with a large fraction of the stars satisfying they streams or dwarf galaxies. well-defined S/N constraints over specific wavelength regions. The overall survey References The goal survey area ranges from decli- FoM is defined to be nations (dec) of +20 to –80 degrees and Belokurov, V. et al. 2006, ApJL, 642, L137 Y LHM"% H H H Belokurov, V. et al. 2018, MNRAS, 478, 611 covers all right ascensions (RAs) satisfy- %N,LHM^ > @` ing Galactic latitudes |b| > 20 degrees. Bland-Hawthorn, J. & Gerhard, O. 2016, ARA&A, 54, 529 This yields some 4500 square degrees where Ai is the area of 4MOST field i in Bonaca, A. et al. 2014, ApJ, 795, 94 north of the celestial equator and around square degrees and CFi the complete- Bonaca, A. et al. 2018, arXiv:1811.03631 12 500 square degrees south of the ness fraction for a field, i.e., the fraction Cooper, A. P. et al. 2010, MNRAS, 406, 744 equator, giving a total of 17 000 square of stars satisfying the S/N constraints; Erkal, D. et al. 2016, MNRAS, 463, 102 Gaia Collaboration et al. 2018, A&A, 616, A1 degrees. As mentioned previously, galaxy 15 000 square degrees reflects our main Helmi, A. et al. 2018, Nature, 563, 85 formation simulations suggest a minimum survey area goal. Hayes, C. R. et al. 2018, ApJ, 852, 49 requirement for the halo survey area to Malhan, K., Ibata, R. A. & Martin, N. F. 2018, be at least 10 000 square degrees, while MNRAS, 481, 3442 Sanderson, R. E., Helmi, A. & Hogg, D. W. 2015, a desirable goal would be to cover at ApJ, 801, 98 least 15 000 square degrees. Xue, X.-X. et al. 2014, ApJ, 784, 170

The Messenger 175 – March 2019 25