Observational Properties of Extragalactic Globular Cluster Systems Duncan A. Forbes Centre for Astrophysics & Supercomputing, Swinburne University

Globular Clusters and Galaxy Formation Duncan A. Forbes Centre for Astrophysics & Supercomputing, Swinburne University

Collaborators - SAGES Project Mike Beasley (Swinburne) Jean Brodie (Lick Observatory) John Huchra (Harvard-Smithsonian) Markus Kissler-Patig (ESO) Soeren Larsen (Lick Observatory)

Telescopes

Globular Clusters Bound homogeneous collections of ~105 M All galaxies with MV < -15 have at least one globular, some have over 10,000 They have a universal luminosity function which gives the Hubble constant accurate to 10% Share the same star formation and chemical enrichment history as their host galaxy. Provide a powerful and unique probe of galaxy formation.

Two sub-populations Most (perhaps all) large galaxies reveal two distinct sub-populations of globular clusters. [Fe/H] = -1.5 -0.5 V-I = 0.95 1.15

Why study extragalactic GCs ? The MW has only 140 known GCs, M31 has ~400, M87 has ~10,000 The Local Group has no giant ellipticals Probe to higher metallicity GCs Same distance and reddening No foreground stars in the GC spectra Discover something new ! Extragalactic Globular Cluster Database: http://astronomy.swin.edu.au/dforbes

Local Group GCs (D ~1 Mpc) 3-6m telescopes Photometry CMD Morphology [0.1” ~ 0.4 pc] Optical colours Infrared colours Sizes Spectra (V ~ 16) Metallicities Abundances Relative Ages System Kinematics Brodie & Huchra 1990, 1991

Virgo/Fornax GCs (D ~15 Mpc) 8-10m telescopes Photometry Optical colours (~50) Infrared colours (~4) Sizes (~20) [0.1” ~ 7 pc] Spectra (V ~ 22) Metallicities (~10) Abundance ratios (~6) Relative Ages (~6) System Kinematics (~4)

Milky Way Globular Cluster System Sub-populations: Metal-rich ~50 Bulge (RGC < 5 kpc) Thick disk (RGC > 5 kpc) Metal-poor ~100 Old Halo (prograde) Young Halo (retrograde) Young halo + 4 Sgr dwarf GCs = Sandage noise

Milky Way Bulge Clusters The inner metal-rich GCs are: spherically distributed similar metallicity to bulge stars similar velocity dispersion to bulge follow the bulge rotation => associated with the bulge Minniti 1995

Bulge Clusters in M31 and M81 The inner metal-rich GCs are: spherically distributed M31 M81 similar metallicity to bulge stars M31 M81? similar velocity dispersion to bulge M31 M81 follow the bulge rotation M31 M81

The Sombrero Galaxy

Metallicities Number of metal-rich GCs scale with the bulge Forbes, Brodie & Larsen 2001

Globular Clusters in M104 M104 M31 MW Sa Sb Sbc 667 100 53 Hubble type 667 100 53 0.80 0.25 0.19 4.2 0.21 0.19 1.1 0.63 0.84 Hubble type Metal-rich GCs Bulge-to-total Disk SN Bulge SN The metal-rich GCs in M104 must be associated with the bulge not disk component.

GCs in Spirals Inner metal-rich GCs in spirals are associated with the bulge not the disk. The number of bulge GCs scales with bulge luminosity (bulge SN ~1). Forbes, Brodie & Larsen 2001

The Elliptical Galaxy Formally Known as The Local Group M31+MW+M33+LMC+SMC+... gE with MV = – 22.0 and N = 700 +/- 125 Universal luminosity function

Local Group Elliptical Metallicity peaks at [Fe/H] = –1.55, –0.64 Ratio 2.5:1 SN = 1.1 -> 2.5 S + S -> gE with low SN Forbes, Masters, Minniti & Barmby 2000

Numbers, Specific Frequency The bulge SN for spirals is ~ 1 The total SN for field ellipticals is 1-3 (Harris 1991) The fraction of red GCs in ellipticals is about 0.5 The bulge SN for field ellipticals is ~1 => Spirals and field ellipticals have a similar number of metal-rich GCs per unit starlight.

A Universal Globular Cluster Luminosity Function Luminosities A Universal Globular Cluster Luminosity Function MV  Ellipticals –7.33 +/- 0.04 1.36 +/- 0.03 Spirals –7.46 +/- 0.08 1.21 +/- 0.05 Ho = 74 +/- 7 km/s/Mpc GCLF Ho = 72 +/- 8 km/s/Mpc HST Key Project Harris 2000

Sizes For Sp  S0 E  cD the GCs reveal a size–colour trend. The blue GCs are larger by ~20%. This trend exists for a range of galaxy types and galactocentric radii. Larsen et al. 2001

Metallicities All large (bulge) galaxies reveal a similar GC metallicity distribution. All ( MV < –15 ) galaxies, reveal a population of GCs with [Fe/H] ~ –1.5. The WLM galaxy has one GC, [Fe/H] = –1.52 age = 14.8 Gyrs (Hodge et al. 1999).

Metallicity vs Galaxy Mass Blue GCs <2.5 V–I ~ 0.95 Pregalactic ? Red GCs ~4 Spirals fit the trend Forbes, Larsen & Brodie 2001

Metallicity vs Galaxy Mass Red GC relation has similar slope to galaxy colour relation. Red GCs and galaxy stars formed in the same star formation event. Forbes, Larsen & Brodie 2001

Colour - Colour Galaxy and GC colours from the same observation. The red GCs and field stars have a very similar metallicity and age. Also NGC 5128 (Harris et al. 1999) Bulge C-T1 colour Forbes & Forte 2001

Spatial Distribution Red GCs are centrally concentrated, have similar azimuthal and density profile to the `bulge’ light. Blue GCs are more extended. Does the blue GC density profile follow the X-ray/halo profile ? Blue Red Red Ellipticals Halo `Bulge’ ? Spirals Halo Bulge Disk

Summary from Photometry Blue GCs = halo common to all galaxies [Fe/H] ~ –1.5 constant colour  Pregalactic ? Red GCs = bulge [Fe/H] ~ –0.5 colour varies with galaxy mass formed in same event as bulge stars

GC Ages NGC 1399 NGC 1399 blue ~3.5hrs on Keck 10m Hß error = +/- 0.2 to 0.3 A NGC 1399 -2.2 < [Fe/H] < 0.3 ages ~ 12 Gyrs (some young GCs) blue 1.6 Gyr 2.3 Gyr red 12 Gyr LRIS spectra, 2hrs Caveat: BHBs

Abundance Ratios High S/N spectra of GCs in ellipticals suggests that all GCs have supersolar alpha abundance ratios, eg [Mg/Fe] ~ +0.3 (similar to the MW). Supersolar ratios indicate the dominance of SN II vs SN Ia products in the GC-forming gas. This is due to: short time formation timescale IMF skewed to high mass stars => important chemical evolution clues [Note: Maraston etal 2001 found solar ratio proto-GCs in the ongoing merger NGC7252]

System Kinematics M49 M49 red low  ~ galaxy blue high  , rotate M87 red ~ blue, rotate No general trends Zepf etal. 2000

GC Formation Scenarios Mergers of spirals (Ashman & Zepf 1992) Two-phase collapse (Forbes, Brodie & Grillmair 1997) In situ plus accretion (Cote, Marzke & West 1998)

Merger Model Ashman & Zepf 1992

Merger Model

Meanwhile, 10 billion years later... Merger Model Meanwhile, 10 billion years later...

Merger Model

Multi-phase collapse model Galactic Phase additional gas collapse formation of metal-rich globulars and bulge stars Pre-galactic Phase clumpy collapse of gas cloud formation of metal-poor globulars and a few halo stars Dormant Phase star formation stopped by SNII gas reheated by SNIa gas cools T = 0 T = 2 Forbes, Brodie & Grillmair 1997

Multi-phase collapse model

Mike Beasley Duncan Forbes Ray Sharples Carlton Baugh

Merging of Dark Matter halos using Monte-Carlo method yielding ‘Merger Trees’. Gaseous fragments form some metal-poor stars and GCs before the main galaxy. Gaseous pre-galactic fragments may merge/collapse forming a burst of stars that results in an elliptical galaxy. The bulk of stars are formed in this burst. Hot gas cools onto the elliptical galaxy, forming a cold gas disk over time and hence a spiral galaxy. Merger Tree

Simply taking a fraction of the stars formed in the model produces a skewed, unimodal distribution in V-I. [Fe/H] V-I Assuming that the efficiency is dependent upon the SFR (e.g. Larsen 2000) produces a sharper unimodal peak. Truncating the blue GC formation at high-redshift produces a bimodal- distribution.

GC Colour Diversity After correcting for magnitude incompleteness and limited areal coverage, the model can reproduce the diversity of GC colour distributions seen with HST (eg Larsen etal. 2001)

GC Number vs Host Galaxy Luminosity The model galaxies (small circles) are well matched to the observations (red symbols). The purple line shows constant GC formation efficiency.

A Massive Elliptical Model Age ‘peak’ of blue GCs is a result of truncation at high redshift. Ages of red GCs show a greater range and appear ‘bursty’. Old ages and particularly bimodal metallicity distributions lead to bimodal colour distributions. Model

What drives SN ? The plot shows model galaxies (blue), data (red) and the spiral-spiral merger model of Bekki etal. (green). High SN galaxies are associated with large numbers of blue GCs, not red GCs as expected in spiral-spiral mergers.

Evolution in SN For most ellipticals there is very little evolution in SN in the last ~10 Gyrs. In the low luminosity elliptical, SN increases from 4.5 to 4.7 as the result of a major merger 4 Gyrs ago. MB = -22.0 MB = -20.4

SN of young ellipticals Galaxy Merger Age SN NGC3156 1 Gyr 0.3 NGC1700 2 Gyr 1.4 -> 2 NGC6702 2 Gyr 2.3 -> 4 NGC1316 3 Gyr 1.5 -> 2 NGC3610 3 Gyr 1.1 Old ellipticals have SN ~ 3 (field) to 6 (cluster)

GC Metallicity vs Mass The red model GCs, are generally consistent with the data, but do not have the same slope. The blue model GCs behave similarly to the red GCs, but with less scatter.

The model predicts that low-L and field galaxies have younger red GCs. GC Age Predictions Age The mean age of red GCs shows a strong dependence on dark matter halo velocity (mass). The model predicts that low-L and field galaxies have younger red GCs. log  N Age

GC Ages NGC 1399 NGC 1399 blue ~3.5hrs on Keck 10m Hß error = +/- 0.25 A NGC 1399 -2.2 < [Fe/H] < 0.3 ages ~ 12 Gyrs (some young GCs) blue 1.6 Gyr 2.3 Gyr red 12 Gyr LRIS spectra, 2hrs

Observational Summary The inner metal-rich GCs in the Milky Way and other spirals have a bulge (not disk) origin. The GC systems of spirals and ellipticals show remarkable similarities. Blue and red GCs have similar ages ~12 Gyrs (but red GCs could be younger by 2-4 Gyrs) Some very young (~2 Gyrs) GCs have been found in `old’ ellipticals. Blue and red GCs have [Mg/Fe] ~ +0.3. Red GCs trace elliptical galaxy star formation.

Model Predictions Our model (Beasley etal. 2002) of GC formation in a Hierarchical Universe assumes truncation of blue GCs at z = 5, and predicts: SN is determined at early epochs; late stage mergers have little effect on SN Blue GCs formed ~12 Gyrs ago in all ellipticals. Red GCs have a mean age of ~8-10 Gyrs in field and low luminosity ellipticals. In general: spirals and ellipticals have similar GC systems.

Future Developments HST+ACS U-band and 8m wide-field K-band imaging studies (eg Puzia etal) Age structure within the red GCs (eg Beasley etal) Nature of the new large (Reff ~ 10 pc) red low luminosity (MV ~ -6) clusters (eg Brodie & Larsen) Importance of stripped nucleated dE (eg Bekki etal) HST+ACS CMDs for Local Group GCs (eg Rich etal) Halo mass estimates: GC kinematics vs X-rays Better SSP grids (eg BHBs, AGB, supersolar)

Late epoch mergers of Sp + Sp  (low SN) E Formation Timeline Blue GCs form in metal-poor gaseous fragments with little or no knowledge of potential well. Halo formation. 12 Gyrs 8-11 Gyrs Clumpy collapse/merger of gaseous fragments form metal-rich red GCs and `bulge’ stars. Late epoch mergers of Sp + Sp  (low SN) E Time http://astronomy.swin.edu.au/dforbes

Number per unit Starlight  = 0.2% McLaughlin (1999) proposed a universal GC formation efficiency  = MGC / Mgas + Mstars = 0.26 % Mgas = current Xray gas mass 2 Ntot ~  L

NGC 5128 Reveals a mostly metal-rich halo, which has similar metallicity to the metal-rich globular clusters. Harris etal 1999