Recently, it has been suggested that the metallicity aversion of long-duration gamma-ray bursts (LGRBs) is not intrinsic to their formation, but rather a consequence of the anti-correlation between star-formation and metallicity seen in the general g
alaxy population. To investigate this proposal, we compare the metallicity of the hosts of LGRBs, broad-lined Type Ic (Ic-bl) supernovae (SNe), and Type II SNe to each other and to the metallicity distribution of star-forming galaxies using the SDSS to represent galaxies in the local universe and the TKRS for galaxies at intermediate redshifts. The differing metallicity distributions of the LGRB hosts and the star formation in local galaxies forces us to conclude that the low-metallicity preference of LGRBs is not primarily driven by the anti-correlation between star-formation and metallicity, but rather must be overwhelmingly due to the astrophysics of the LGRBs themselves. Three quarters of our LGRB sample are found at metallicities below 12+log(O/H) < 8.6, while less than a tenth of local star-formation is at similarly low metallicities. However, our SN samples are statistically consistent with the metallicity distribution of the general galaxy population. Using the TKRS population of galaxies, we are able to exclude the possibility that the LGRB host metallicity aversion is caused by the decrease in galaxy metallicity with redshift. The presence of the strong metallicity difference between LGRBs and Ic-bl SNe largely eliminates the possibility that the observed LGRB metallicity bias is a byproduct of a difference in the initial mass functions of the galaxy populations. Rather, metallicity below half-solar must be a fundamental component of the evolutionary process that separates LGRBs from the vast majority of Ic-bl SNe and from the bulk of local star-formation.
We detect the optical afterglow and host galaxy of GRB 070714B. Our observations of the afterglow show an initial plateau in the lightcurve for approximately the first 5 to 25 minutes, then steepening to a powerlaw decay with index alpha= 0.86 +/- 0.
10 for the period between 1 to 24 hours post burst. This is consistent with the X-ray light-curve which shows an initial plateau followed by a similar subsequent decay. At late time, we detect a host galaxy at the location of the optical transient. Gemini Nod & Shuffle spectroscopic observations of the host show a single emission line at 7167 angstroms which, based on a grizJHK photometric redshift, we conclude is the 3727 angstrom [O II] line. We therefore find a redshift of z=0.923. This redshift, as well as a subsequent probable spectroscopic redshift determination of GRB 070429B at z=0.904 by two other groups, significantly exceeds the previous highest spectroscopically confirmed short burst redshift of z=0.546 for GRB 051221. This dramatically moves back the time at which we know short bursts were being formed, and suggests that the present evidence for an old progenitor population may be observationally biased.
We present spectroscopy of the host of GRB 051022 with GMOS nod and shuffle on Gemini South and NIRSPEC on Keck II. We determine a metallicity for the host of log(O/H)+12 = 8.77 using the R23 method (Kobulnicky & Kewley 2004 scale) making this the hi
ghest metallicity long burst host yet observed. The galaxy itself is unusually luminous for a LGRB host with a rest frame B band absolute magnitude -21.5 and has the spectrum of a rapidly star-forming galaxy. Our work raises the question of whether other dark burst hosts will show high metallicities.
Gemini Nod & Shuffle spectroscopy on the host of the short GRB 070714B shows a single emission line at 7167 angstroms which, based on a grizJHK photometric redshift, we conclude is the 3727 angstrom [O II] line. This places the host at a redshift of
z=.923 exceeding the previous record for the highest spectroscopically confirmed short burst redshift of z=.546 held by GRB 051221. This dramatically moves back the time at which we know short bursts were being formed, and suggests that the present evidence for an old progenitor population may be observationally biased.
