Early-type galaxy evolution : insights from the rest-frame UV

1. Early-type galaxy evolution : insights from the rest-frame UV Sugata Kaviraj Oxford/UCL Collaborators: Sukyoung Yi, Kevin Schawinski, Eric Gawiser, Pieter van Dokkum, Richard Ellis Malaysia 2009

2. Spheroidal (early-type) galaxies • Smooth and featureless • Dominated by old population II stars • Dominate the high-mass end of the galaxy LF • Dominate the stellar mass density at low redshift

3. Early-type galaxy formationThe ‘classical’ model Optical colours are red with small scatter (e.g. Bower et al. 1992) High [Mg/Fe] values (e.g. Thomas 1999 but see Smith et al. 2009!) Little or no (optical) luminosity evolution (e.g. Scarlata et al. 2007) Bulk of star formation at high redshift (z>2) Star formation timescale < 1 Gyr Early-types already in place at high redshift? Short high-redshift starburst then mainly passive ageing (‘monolithic evolution’)

4. Making galaxies in the ‘standard’ LCDM model • Build DM ‘backbone’ (N-body sims) • Seed DM halos with gas • Baryonic evolution (gas to star conversion, feedback) using recipes • DM halo mass function  galaxy luminosity function • Early-types form through major (mass ratio < 3) mergers (e.g. Cole et al. 2000, Hatton et al. 2003) • Continuous star formation over a Hubble time • Compatible with early-type properties?

5. The effect of young stars • 99.9% of stars form 10 Gyrs ago • 0.1% of stars form 0.1 Gyrs in the past • Optical spectrum indistinguishable from purely old population • But a strong UV signal!

6. The effect of young stars • 99.5% of stars form 10 Gyrs ago • 0.5% of stars form 0.1 Gyrs in the past • Optical spectrum indistinguishable from purely old population • But a strong UV signal!

7. UV studies of nearby early-type galaxies • UV data requires space-based telescope • GALEX • Far UV (1500 A) and Near UV (2300 A) filters • Unprecedented resolution and FOV • Pre-select ‘spheroidal’ galaxies using SDSS fracdev parameter (crude bulge/disk decomposition): fracdev>0.95 • Cross-match with GALEX, visually inspect for contaminants and remove AGN (which could contaminate UV)

8. Early-type UV colours: GALEX & SDSS • Expected tight relation in optical (g-r) CMR • But NUV CMR shows a spread of 6 mags - strong UV sources present in nearby early-type galaxies Kaviraj et al., 2007, ApJS, 173, 619 Schawinski et al., 2007, ApJS, 173, 512 Yi et al., 2005, ApJ, 619, L111

9. Early-type UV sourcesOld or young? • Old metal-rich HB stars can produce UV (the UV upturn phenomenon) • Study NUV colour of NGC 4552 - any UV ‘left over’ likely to be from young stars • Yi et al. (2005) find only 4/62 galaxies with the UV spectral shape of UV upturn galaxies

10. The ‘star-forming’ early-type fraction • Low-redshift star-forming fraction is at least 30% • Could be as high as 80-90% • 1-5% in young stars with ages 300-500 Myrs (also Martin, O’Connell et al. 2007) • Widespread recent star formation in nearby early-types Kaviraj et al., 2007, ApJS, 173, 619

11. The high-redshift galaxy populationThe Chandra Deep Field South (CDFS) • Optical (U,B,V…) photometry traces rest-frame UV beyond z~0.5 • No old stars! • Need depth (in U and B bands especially), redshifts and morphologies • MUSYC (UBVRIzJK) + COMBO-17 (photo-z) + VVDS (spec-z) + GEMS (HST images) HST ACS images

12. Rest-frame UV colours at high redshift(0.5<z<1) Low z Kaviraj et al., 2008, MNRAS, 388, 67

13. Rest-frame UV colours at high redshift(0.5<z<1) • 0.2% of early-types consistent with purely passive evolution since z=3 • 1.2% of early-types consistent with purely passive evolution since z=2 • SSP assumption, result is robust using either BC2003 or Yi (2003) stellar models Kaviraj et al. 2008, MNRAS, 388, 67

14. Recent star formation in high-z early-types • Luminous galaxies (MV<-21) form up to 10-15% of their mass after z=1 (consistent with SAM predictions) • Low-mass galaxies form 30-60% of their mass after z=1 Kaviraj et al. 2008, MNRAS, 388, 67

15. What drives the recent star formation?Indirect evidence for minor mergers • Numerical simulations of minor mergers (mass ratios 1:3 - 1:10) • Satellite gas fractions > 20% • Monte-Carlo simulation drived by LCDM minor mergers statistics • Good agreement with observed UV and optical CMRs Kaviraj et al. 2009, MNRAS in press (arXiv:0711.1493) see also Bezanson et al. 2009 (arXiv0903.2044)

16. What drives the recent star formation? Relaxed ETGs

17. What drives the recent star formation? Relaxed ETGs Disturbed ETGs (30% of the ETG population)

18. What drives the recent star formation? COSMOS z=0.6 Rest-frame (NUV-g)

19. What drives the recent star formation? COSMOS z=0.6 Rest-frame (NUV-g)

20. What drives the recent star formation? Internal mass loss Not enough gas at low redshift (Kaviraj et al. 2007) Condensation from hot gas reservoir Hot gas fraction of the order of total recent star formation (O'Sullivan et al. 2003) Major mergers Major merger rate (e.g. Conselice 2007) does not seem enough to satisfy disturbed ETG fraction (~35-40%)  Minor mergers Several factors more frequent than major mergers Consistent with disturbed ETGs dominating the blue cloud and mass fractions forming in young stars Not expected to perturb the morphology of the galaxy

21. SummaryarXiv:0710.1311 • Early-types of all luminosities form stars over the last 10 Gyrs, although bulk of the stars do form at high redshift • Negligible fraction of early-types consistent with purely passive ageing since z=2 • Recent star formation mass fraction ~ 1-5% at z~0.1, ~ 7-10% at z~0.7 • Massive early-types form up to 10-15% of their mass after z~1, low-mass early-types form 30-60% of their mass after z~1 • Most likely mechanism driving recent star formation is minor merging – only mechanism that is consistent with all observational data