E-MOSAICS: tracing galaxy formation and assembly with globular - - PowerPoint PPT Presentation

E-MOSAICS: tracing galaxy formation and assembly with globular clusters

Joel Pfeffer (LJMU)

Diederik Kruijssen (ZAH), Rob Crain (LJMU), Nate Bastian (LJMU) Marta Reina-Campos (ZAH), Meghan Hughes (LJMU)

Pfeffer et al. 2018 Kruijssen, Pfeffer+ 2018a, MNRAS, subm. Kruijssen, Pfeffer+ 2018b, arXiv:1806.05680 Thob, Crain, Pfeffer+, in prep.

Using globular clusters (GCs) to trace galaxy formation?

Globular clusters are powerful probes of galaxy formation (e.g. review by

Brodie & Strader 06)

Can observe GCs to significantly larger distances than individual stars, and (potentially) obtain metallicities, ages, kinematics

Using globular clusters (GCs) to trace galaxy formation?

Globular clusters are powerful probes of galaxy formation (e.g. review by

Brodie & Strader 06)

Can observe GCs to significantly larger distances than individual stars, and (potentially) obtain metallicities, ages, kinematics But we require a complete model for galaxy and GC formation. . .

Towards a complete model of GC formation

We want to model full star cluster populations in populations of galaxies Currently not possible to simultaneously model the small scales of star/star cluster formation (< pc) and large scales of galaxy formation (∼Mpc) in hydrodynamical simulations

Towards a complete model of GC formation

We want to model full star cluster populations in populations of galaxies Currently not possible to simultaneously model the small scales of star/star cluster formation (< pc) and large scales of galaxy formation (∼Mpc) in hydrodynamical simulations Requires:

• A model for cluster formation (observations of young star clusters)
• Cluster evolution and disruption (N-body simulations)
• Galaxy formation including baryons

Towards a complete model of GC formation

We want to model full star cluster populations in populations of galaxies Currently not possible to simultaneously model the small scales of star/star cluster formation (< pc) and large scales of galaxy formation (∼Mpc) in hydrodynamical simulations Requires:

• A model for cluster formation (observations of young star clusters)
• Cluster evolution and disruption (N-body simulations)
• Galaxy formation including baryons

Enter E-MOSAICS. . .

The E-MOSAICS project: co-formation of galaxies and GCs

MOdelling Star cluster population Assembly In Cosmological Simulations in the context of EAGLE (Pfeffer+ 18; Kruijssen+ 18) Couple sub-grid cluster model (MOSAICS) to EAGLE galaxy formation model

(Schaye+ 15; Crain+ 15)

Using EAGLE Recal (high-res) model (baryonic particle masses ∼ 2 × 105 M⊙) 25 cosmological zoom-ins of Milky Way-mass galaxies (over 200 simulations in total including subgrid model testing) Near future: galaxy groups zooms (Mvir ∼ 1013 M⊙) and 34 Mpc periodic volume currently

• running. . .

With thanks to the Virgo Consortium for DiRAC supercomputing time

MOSAICS: sub-grid model for cluster formation and evolution

Kruijssen+ 11; Pfeffer+ 18:

• On-the-fly modelling
• Each star particle hosts its own sub-grid cluster population

⇒ Form cluster population with local bound cluster formation efficiency (CFE) at each new star particle during simulation

• Schechter initial cluster mass function (power law slope −2, with

truncation Mc,∗), consistent with observations of YSCs

• Cluster formation depends on local (gas/dynamical) properties in
• simulation. Completely described by CFE and Mc,∗
• Cluster mass-loss by stellar evolution, tidal shocks and evaporation

using the evolving local tidal field of each ‘cluster particle’

• Dynamical friction in post-processing

The E-MOSAICS project: co-formation of galaxies and GCs

First self-consistent simulations of the formation and evolution of Milky Way-type galaxies and their GC populations over full cosmic history

(Pfeffer+ 18; Kruijssen+ 18)

Two main goals of E-MOSAICS

Are GCs just evolved versions of young clusters? Yes Can we use GCs to trace the formation and assembly of galaxies?

Two main goals of E-MOSAICS

Are GCs just evolved versions of young clusters? Yes Can we use GCs to trace the formation and assembly of galaxies?

Two main goals of E-MOSAICS

Are GCs just evolved versions of young clusters? Yes Can we use GCs to trace the formation and assembly of galaxies?

E-MOSAICS: GCs are good tracers of galaxy formation

Young star clusters are the peaks of star formation in the hierarchical ISM (see

Longmore+ 14 for a review)

If GCs form like young clusters, then GCs trace the enrichment history of their host galaxy

−2.5 −2.0 −1.5 −1.0 −0.5 0.0 0.5 Metallicity [Fe/H] MW09 0.0 0.1 0.2 0.4 0.6 1.0 2.0 3.0 6.0 Formation redshift MW14 0.0 0.1 0.2 0.4 0.6 1.0 2.0 3.0 6.0 Formation redshift −2.5 −2.0 −1.5 −1.0 −0.5 0.0 0.5 Metallicity [Fe/H] MW15 MW18 2 4 6 8 10 12 14 Age [Gyr] −2.5 −2.0 −1.5 −1.0 −0.5 0.0 0.5 Metallicity [Fe/H] MW19 2 4 6 8 10 12 14 Age [Gyr] MW23 106 107 108 109 1010 Galaxy stellar mass M∗ [M⊙]

(Kruijssen, Pfeffer+ 2018, subm.)

E-MOSAICS: galaxy formation from GC age-metallicity relations

Correlate 12 GC age-Z metrics and NGC with 30 quantities describing galaxy formation E.g. Mvir, Vmax, cNFW, formation and assembly timescales, merger histories ⇒ Obtain 20 highly significant correlations (peff = p/Ncorr, Holm 79)

(Kruijssen, Pfeffer+ 2018, subm.)

E-MOSAICS: galaxy formation from GC age-metallicity relations

Correlate 12 GC age-Z metrics and NGC with 30 quantities describing galaxy formation E.g. Mvir, Vmax, cNFW, formation and assembly timescales, merger histories ⇒ Obtain 20 highly significant correlations (peff = p/Ncorr, Holm 79)

(Kruijssen, Pfeffer+ 2018, subm.)

E-MOSAICS: application to the Milky Way

(Kruijssen, Pfeffer+ 2018, MNRAS, arXiv:1806.05680)

E-MOSAICS: application to the Milky Way

MW had ∼15 mergers with galaxies M∗ > 5 × 106 M⊙ The MW assembled early for its halo mass: za ≈ 1.2 (sim. mean ≈ 0.8)

(See also Mackereth+ 18, based on stellar [α/Fe]-[Fe/H] bimodality) (Kruijssen, Pfeffer+ 2018, MNRAS, arXiv:1806.05680)

E-MOSAICS: age-metallicity-mass relation for galaxies

Can we see galaxy accretion events in the GC age-metallicity relations?

(e.g. Forbes & Bridges 10; Leaman+ 13)

Galaxy enrichment history depends on galaxy mass ⇒ higher mass galaxies enrich faster (Median galaxy enrichment histories from EAGLE Recal)

2 4 6 8 10 12 14 Age [Gyr] 2.5 2.0 1.5 1.0 0.5 0.0 0.5 [Fe/H] 7.5 < log10(M * /M ) < 8 8 < log10(M * /M ) < 8.5 8.5 < log10(M * /M ) < 9 9 < log10(M * /M ) < 9.5 9.5 < log10(M * /M ) < 10 10 < log10(M * /M ) < 10.5

E-MOSAICS: MW GC age-metallicity relations

GC relations in age-metallicity space constrain both the galaxy mass evolution and the number of GCs per halo (at z = 0) MW accreted two galaxies M∗ ≈ 109M⊙ with ∼20 GCs and one galaxy M∗ ≈ 108M⊙ with ∼8 GCs (Probably more we can’t distinguish)

(Kruijssen, Pfeffer+ 2018, MNRAS, arXiv:1806.05680)

E-MOSAICS: MW GC age-metallicity relations

GC relations in age-metallicity space constrain both the galaxy mass evolution and the number of GCs per halo (at z = 0) MW accreted two galaxies M∗ ≈ 109M⊙ with ∼20 GCs and one galaxy M∗ ≈ 108M⊙ with ∼8 GCs (Probably more we can’t distinguish)

(Kruijssen, Pfeffer+ 2018, MNRAS, arXiv:1806.05680)

Formation and assembly of the Milky Way from its GCs

106 107 108 109 1010 1011 Stellar mass [M⊙]

Main progenitor Satellites 1 & 2 Satellite 3 Papovich et al. (2015) Milky Way progenitors

2 4 6 8 10 12 14 Age [Gyr]

• 2.5
• 2.0
• 1.5
• 1.0
• 0.5

0.0 0.5 Metallicity of newly-formed stars [Fe/H]

GCs with 105 < M/M⊙ < 106.3 Haywood et al. (2013) Galactic disc stars Snaith et al. (2015) Galactic enrichment history

0.0 0.1 0.2 0.4 0.6 1.0 2.0 3.0 6.0 Redshift

(Kruijssen, Pfeffer+ 2018, MNRAS, arXiv:1806.05680)

Formation and assembly of the Milky Way from its GCs

(Kruijssen, Pfeffer+ 2018, MNRAS, arXiv:1806.05680)

Formation and assembly of the Milky Way from its GCs

Canis Major = The Sausage/Gaia-Enceladus (?) Most massive galaxy ever accreted = Kraken

(Kruijssen, Pfeffer+ 2018, MNRAS, arXiv:1806.05680)

E-MOSAICS: using GCs to trace shapes of MW-mass DM haloes

Thob, Crain, Pfeffer+ in prep.

(Using the iterative reduced inertia tensor method, Schneider+12)

Accreted GCs trace the DM halo shape poorly. . .

0.0 0.2 0.4 0.6 0.8 1.0 ǫDM 0.0 0.2 0.4 0.6 0.8 1.0 ǫGCs

11 23 12 20 1 14 10 13 19 17 16 21 15 5 8 7 4 18 22 6 3 24 9 2

ExSitu vs DM in 90%GCs 0.4 0.5 0.6 0.7 0.8 0.9 1 Mcentral/MFOF 0.0 0.2 0.4 0.6 0.8 1.0 ǫDM 0.0 0.2 0.4 0.6 0.8 1.0 ǫGCs

11 23 12 20 1 14 10 13 19 17 16 21 15 5 8 7 4 18 22 6 3 24 9 2

Poor Fe/H vs DM in 90%GCs 0.4 0.5 0.6 0.7 0.8 0.9 1 Mcentral/MFOF

Metal-poor GCs ([Fe/H] < −1) trace the shape of the DM halo very well!

Concluding remarks/questions

With E-MOSAICS we can now trace the formation and assembly of galaxies using their GC populations Can we distinguish Satellite 1 & 2 using the orbits of their GCs? Has Satellite 2 been found? Deposited its GCs at 10-20 kpc (Myeong+18;

Helmi+18)

Many “satellite branch” GCs at <10

• kpc. Evidence of Kraken?

106 107 108 109 1010 1011 Stellar mass [M⊙]

Main progenitor Satellites 1 & 2 Satellite 3 Papovich et al. (2015) Milky Way progenitors

2 4 6 8 10 12 14 Age [Gyr]

• 2.5
• 2.0
• 1.5
• 1.0
• 0.5

0.0 0.5 Metallicity of newly-formed stars [Fe/H]

GCs with 105 < M/M⊙ < 106.3 Haywood et al. (2013) Galactic disc stars Snaith et al. (2015) Galactic enrichment history

0.0 0.1 0.2 0.4 0.6 1.0 2.0 3.0 6.0 Redshift