Abstract:

The epoch at z > 7 is thought to be important in the formation of the first galaxies, but the requirement of sensitive, wide-area near-infrared observations makes discovery of these sources difficult. This dissertation presents two surveys for z > 7 galaxies, using parallel imaging with the Near Infrared Camera and Multi-Object Spectrometer (NICMOS) on the Hubble Space Telescope . In this way, we have efficiently accrued significant areas with sensitivities that can not be achieved from the ground. First, we carried out a serendipitous survey in J110 and H160 to search for galaxies at z ~ 9. This survey consists of 135 arcmin 2 of imaging in 228 independent sight lines, reaching typical 5? sensitivities of J110 and H 160 [Special characters omitted.] 25 (AB). Fourteen J-dropout Lyman break galaxy (LBG) candidates are selected with J110 - H160 ? 2.5, and the likelihood that any are genuine z ~ 9 galaxies is considered. We find that if [Special characters omitted.] is significantly brighter than [Special characters omitted.] =6 as measured by Bouwens et al. (2006, 2007; a factor of four), then a few J-dropout LBGs are likely to be detected in our survey. Distinguishing these candidates from dust-reddened star-forming galaxies at z ~ 2 - 3 requires followup observations. Longer wavelength followup observations of five J110 -H160 selected dropouts, using the Infrared Array Camera (IRAC) on the Spitzer Space Telescope , shows that four of the five targets are certainly low redshift interlopers. Remarkably, though, one source has the spectral energy distribution (SED) of a z ~ 9 LBG, with an apparent Balmer break between 3.6 and 4.5?m. The high redshift of this galaxy would imply that the luminous end of the rest-frame UV luminosity function has not evolved substantially from z ~ 9 to z ~ 3. However, more recent followup i ' observations show a weak detection, strongly suggesting that the galaxy is an interloper. While our observations have shown that none of these five targets are viable z ~ 9 candidates, the remainder of the J110 - H160 selected sources are yet unobserved. However, we remain doubtful of their status as possible z ~ 9 candidates.

The second survey is a complementary, smaller-area but deeper survey for galaxies at z ~ 7 in addition to z ~ 9. This survey covers ~ 14.4 arcmin2 in the Great Observatories Origins Deep Survey (GOODS, 5.9 arcmin2 ), the Cosmic Origins Survey (COSMOS, 7.2 arcmin2 ), and SSA22 (1.3 arcmin 2 ). These images increase the amount of deep near-infrared data by more than 60% in fields where the investment in deep optical data has already been made. We find no z > 7 candidates in our survey area, consistent with the Bouwens et al. (2008) measurements at z ~ 7 and 9 (over 23 arcmin2 ), which predict 0.7 galaxies at z ~ 7 and < 0.03 galaxies at z ~ 9. For the case of luminosity evolution, assuming a Schecter parameterization with a typical [straight phi]* = 10-3 Mpc-3 , we find M * > -20.0 for z ~ 7 and M * > -20.7 for z ~ 9 (68% confidence). This suggests that the downward luminosity evolution of LBGs continues at z > 6, although our result is marginally consistent with the z ~ 6 LF of Bouwens et al. (2006, 2007).