Hα SPECTROSCOPY OF GALAXIES AT z > 2: KINEMATICS AND STAR FORMATION

摘要

We present near-infrared spectroscopy of Hα emission lines in a sample of 16 star-forming galaxies at redshifts 2.0 < z < 2.6. Our targets are drawn from a large sample of galaxies photometrically selected and spectroscopically confirmed to lie in this redshift range. We have obtained this large sample with an extension of the broadband U_nGR color criteria used to identify Lyman break galaxies at z ～ 3. The primary selection criterion for IR spectroscopic observation was proximity to a QSO sight line; we therefore expect the galaxies presented here to be representative of the sample as a whole. Six of the galaxies exhibit spatially extended, tilted Hα emission lines; rotation curves for these objects reach mean velocities of ～ 150 km s~(-1) at radii of ～6 kpc, without corrections for inclination or any other observational effect. The velocities and radii give a mean dynamical mass of ≥ 4 x 10~(10) solar mass. We have obtained archival Hubble Space Telescope images for two of these galaxies; they are morphologically irregular. One-dimensional velocity dispersions for the 16 galaxies range from ～50 to ～260 km s~(-1), and in cases in which we have both virial masses implied by the velocity dispersions and dynamical masses derived from the spatially extended emission lines, they are in rough agreement. We compare our kinematic results with similar measurements made at z ～ 3 and find that both the observed rotational velocities and velocity dispersions tend to be larger at z ～ 2 than at z ～ 3. We also calculate star formation rates (SFRs) from the Hα luminosities and compare them with SFRs calculated from the UV continuum luminosity. We find a mean SFR_(Hα) of 16 solar mass yr~(-1) and an average SFR_(Hα)/SFR_(UV) ratio of 2.4, without correcting for extinction. We see moderate evidence for an inverse correlation between the UV continuum luminosity and the ratio SFR_(Hα)/SFR_(UV), such as might be observed if the UV-faint galaxies suffered greater extinction. We discuss the effects of dust and star formation history on the SFRs and conclude that extinction is the most likely explanation for the discrepancy between the two SFRs.