paper

Dust Obscuration and Metallicity at High Redshift: New Inferences from UV, H-alpha, and 8 Micron Observations of z~2 Star-Forming Galaxies

arXiv:1002.0837 · doi:10.1088/0004-637X/712/2/1070

Abstract

We use a sample of 90 spectroscopically-confirmed Lyman Break Galaxies with H-alpha and 24 micron observations to constrain the relationship between rest-frame 8 micron luminosity, L(8), and star formation rate (SFR) for L* galaxies at z~2. We find a tight correlation with 0.24 dex scatter between L8 and L(Ha)/SFR for z~2 galaxies with L(IR)~10^10 - 10^12 Lsun. Employing this relationship with a larger sample of 392 galaxies with spectroscopic redshifts, we find that the UV slope can be used to recover the dust attenuation of the vast majority of L* galaxies at z~2 to within 0.4 dex scatter using the local correlation. Separately, young galaxies with ages <100 Myr appear to follow an extinction curve that is steeper than the one found for local starburst galaxies. Therefore, such young galaxies may be significantly less dusty than inferred previously. Our results provide the first direct evidence, independent of the UV slope, for a correlation between UV and bolometric luminosity at high redshift, in the sense that UV-faint galaxies are also on average less infrared and less bolometrically-luminous than their UV-bright counterparts. Further, as the SFR increases, the UV luminosity reaches a maximum value corresponding to L* at z~2, implying that dust obscuration may be largely responsible for modulating the bright-end of the UV luminosity function. L* galaxies at z~2, while at least an order of magnitude more bolometrically-luminous, exhibit ratios of metals-to-dust that are similar to those of local starbursts. This result is expected if high-redshift galaxies are forming their stars in a less metal-rich environment compared to local galaxies of the same luminosity, thus naturally leading to a redshift evolution in both the luminosity-metallicity and luminosity-obscuration relations. [Abridged]

24 pages, 18 figures, 2 tables; accepted for publication in ApJ

References in corpus (19)

Cited by in corpus (292)