Published online by Cambridge University Press: 01 June 2008
We present a study of remarkably luminous and unique dwarf galaxies at redshifts of 0.5 < z < 0.7, selected from the DEEP2 Galaxy Redshift survey by the presence of the temperature sensitive [OIII]λ4363 emission line. Measurements of this important auroral line, as well as other strong oxygen lines, allow us to estimate the integrated oxygen abundances of these galaxies accurately without being subject to the degeneracy inherent in the standard R23 system used by most studies. [O/H] estimates range between 1/5–1/10 of the solar value. Not surprisingly, these systems are exceedingly rare and hence represent a population that is not typically present in local surveys such as SDSS, or smaller volume deep surveys such as GOODS.
Our low-metallicity galaxies exhibit many unprecedented characteristics. With B-band luminosities close to L*, thse dwarfs lie significantly away from the luminosity-metallicity relationships of both local and intermediate redshift star-forming galaxies. Using stellar masses determined from optical and NIR photometry, we show that they also deviate strongly from corresponding mass-metallicity relationships. Their specific star formation rates are high, implying a significant burst of recent star formation. A campaign of high resolution spectroscopic follow-up shows that our galaxies have dynamical properties similar to local HII and compact emission line galaxies, but mass-to-light ratios that are much higher than average star-forming dwarfs.
The low metallicities, high specific star formation rates, and small halo masses of our galaxies mark them as lower redshift analogs of Lyman-Break galaxies, which, at z ~ 2 are evolving onto the metallicity sequence that we observe in the galaxy population of today. In this sense, these systems offer fundamental insights into the physical processes and regulatory mechanisms that drive galaxy evolution in that epoch of major star formation and stellar mass assembly.