Dust Obscuration and Metallicity at z~2: Implications for the Selection and Stellar Populations of High-Redshift

Galaxies

Collaborators: Collaborators:

Chuck Steidel (Caltech), Chuck Steidel (Caltech),

Max Pettini (IoA), Max Pettini (IoA),

Alice Shapley (UCLA),Alice Shapley (UCLA),

Dawn Erb (UCSB)Dawn Erb (UCSB)

Naveen Reddy (Hubble Fellow, NOAO)

Hubble Fellows Symposium, STScI, Baltimore, MD, March 09, 2010

Why z~2-3 is Interesting

Reddy et al 2008a>50% of the stars in the present-day universe formed in the interval 3.5 > z > 1.5; peak of quasar activity

Can watch all of the effects that shaped the present-day universe while they were happening

Galaxies still bright enough for spectroscopic study on 8-10 m class telescopes

Cosmologically interesting epoch where we also have access to a wealth of multi-wavelength indicators of star formation (UV, H-alpha, 24 micron, X-ray, radio)

Advantages of our analysis

- > 2000 spectroscopic redshifts at the bright-end

- modeling of systematic effects

- maximum-likelihood constraints on LF that are robust to non-uniform sources of scatter

- 31000 LBGs in 31 independent fields

Results on the UV LF at z~2-3

Reddy et al 2009 Steep faint-end slope of ~ -1.73, similar to that measured at z~4-6

0.1L*0.07L*

Sub-L* Galaxies Dominate the UV Luminosity Density at z>2

Dust Extinction in Typical High Redshift Galaxies

Tight correlation between observed 24 micron (rest-frame 8 micron) dust luminosity and H luminosity

Correlation between UV slope and dust attenuation in L* galaxies at z~2

What about UV-faint galaxies?

L(8) L(Ha) L(IR)

Correlation between Bolometric Luminosity and Dust Extinction at z~2

Log[L(bol)]=(0.69±0.03)log[L(IR)/L(UV)]+(10.91 ±0.04)Saturation of

UV luminosity

around L*(UV)

Redshift Evolution of Dust Extinction

• Tight correlation between bolometric luminosity and dust attenuation

• But, typical z~2 galaxies are less dusty than local galaxies *at a give bolometric luminosity*

• Evolution in extinction per unit SFR due to increasing dust-to-gas ratio?

In practice, this means that UV selection probes a larger dynamic range at high redshift than one would infer from local

observations

Comparison of Metallicities: Oxygen Abundance vs. Dust Obscuration

Systematics with Stellar Population Age

Do younger galaxies at high redshift exhibit

evidence for a steeper dust extinction curve?

External Constraints on Stellar Populations

Implications for High Redshift Galaxy Selection…• Redshift evolution of

luminosity-obscuration/luminosity-metallicity relations implies an increasing UV transparency with larger look back time (missing dusty galaxies from UV-dropout surveys not as severe as one would have predicted from local observations)

• Luminosity dependence of dust at any given epoch implies lower SFRDs; a more rapid rise in the SFRD to z~2; does it match the evolution of quasars?

• Youngest galaxies (<100 Myr) at high-z consistent with steeper extinction law implying lower dust / bolometric SFR; even lower SFRDs at high redshift?

• Adopting a common luminosity limit (accounting for L* evolution) and the luminosity/redshift dependence of dust allows us to recover the measured SMD from the integrated SFRD

SFRD

AGN LD

Aird et al. (2010)