Searches for very-high-redshift galaxies over the past decade have yielded a large sample of more than 6,000 galaxies existing just 900-2,000 million years (Myr) after the Big Bang (redshifts 6 > z > 3; ref. 1). The Hubble Ultra Deep Field (HUDF09) d
ata have yielded the first reliable detections of z ~ 8 galaxies that, together with reports of a gamma-ray burst at z ~ 8.2 (refs 10, 11), constitute the earliest objects reliably reported to date. Observations of z ~ 7-8 galaxies suggest substantial star formation at z > 9-10. Here we use the full two-year HUDF09 data to conduct an ultra-deep search for z ~ 10 galaxies in the heart of the reionization epoch, only 500 Myr after the Big Bang. Not only do we find one possible z ~ 10 galaxy candidate, but we show that, regardless of source detections, the star formation rate density is much smaller (~10%) at this time than it is just ~200 Myr later at z ~ 8. This demonstrates how rapid galaxy build-up was at z ~ 10, as galaxies increased in both luminosity density and volume density from z ~ 8 to z ~ 10. The 100-200 Myr before z ~ 10 is clearly a crucial phase in the assembly of the earliest galaxies.
We use the ultra-deep WFC3/IR data over the HUDF and the Early Release Science WFC3/IR data over the CDF-South GOODS field to quantify the broadband spectral properties of candidate star-forming galaxies at z~7. We determine the UV-continuum slope be
ta in these galaxies, and compare the slopes with galaxies at later times to measure the evolution in beta. For luminous L*(z=3) galaxies, we measure a mean UV-continuum slope beta of -2.0+/-0.2, which is comparable to the beta~-2 derived at similar luminosities at z~5-6. However, for the lower luminosity 0.1L*(z=3) galaxies, we measure a mean beta of -3.0+/-0.2. This is substantially bluer than is found for similar luminosity galaxies at z~4, just 800 Myr later, and even at z~5-6. In principle, the observed beta of -3.0 can be matched by a very young, dust-free stellar population, but when nebular emission is included the expected beta becomes >~-2.7. To produce these very blue betas (i.e., beta~-3), extremely low metallicities and mechanisms to reduce the red nebular emission are likely required. For example, a large escape fraction (i.e., f_{esc}>~0.3) could minimize the contribution from this red nebular emission. If this is correct and the escape fraction in faint z~7 galaxies is >~0.3, it may help to explain how galaxies reionize the universe.
We provide a systematic measurement of the rest-frame UV continuum slope beta over a wide range in redshift (z~2-6) and rest-frame UV luminosity (0.1-2L*) to improve estimates of the SFR density at high redshift. We utilize the deep optical and infra
red data (ACS/NICMOS) over the CDF-S and HDF-N GOODS fields, as well as the UDF for our primary UBVi dropout sample. We correct the observed distributions for selection biases and photometric scatter. We find that the UV-continuum slope of the most luminous galaxies is substantially redder at z~2-4 than it is at z~5-6. Lower luminosity galaxies are also found to be bluer than higher luminosity galaxies at z~2.5 and z~4. We do not find a large number of galaxies with betas as red as -1 in our dropout selections at z~4, and particularly at z>~5, even though such sources could be readily selected from our data. This suggests that star-forming galaxies at z>~5 almost universally have very blue UV-continuum slopes, and that there are not likely to be a substantial number of dust-obscured galaxies at z>~5 that are missed in dropout searches. Using the same relation between UV-continuum slope and dust extinction as found to be appropriate at z~0 and z~2, we estimate the average dust extinction of galaxies as a function of redshift and UV luminosity in a consistent way. We find that the estimated dust extinction increases substantially with cosmic time for the most UV luminous galaxies, but remains small (<~2x) at all times for lower luminosity galaxies. Because these same lower luminosity galaxies dominate the luminosity density in the UV, the overall dust extinction correction remains modest at all redshifts. We include the contribution from ULIRGs in our SFR density estimates at z~2-6, but find that they contribute only ~20% of the total at z~2.5 and <~10% at z>~4.
We utilize the newly-acquired, ultra-deep WFC3/IR observations over the HUDF to search for star-forming galaxies at z~8-8.5, only 600 million years from recombination, using a Y_{105}-dropout selection. The new 4.7 arcmin**2 WFC3/IR observations reac
h to ~28.8 AB mag (5 sigma) in the Y_{105}J_{125}H_{160} bands. These remarkable data reach ~1.5 AB mag deeper than the previous data over the HUDF, and now are an excellent match to the HUDF optical ACS data. For our search criteria, we use a two-color Lyman-Break selection technique to identify z~8-8.5 Y_{105}-dropouts. We find 5 likely z~8-8.5 candidates. The sources have H_{160}-band magnitudes of ~28.3 AB mag and very blue UV-continuum slopes, with a median estimated beta of <~-2.5 (where f_{lambda}propto lambda^{beta}). This suggests that z~8 galaxies are not only essentially dust free but also may have very young ages or low metallicities. The observed number of Y_{105}-dropout candidates is smaller than the 20+/-6 sources expected assuming no evolution from z~6, but is consistent with the 5 expected extrapolating the Bouwens et al. 2008 LF results to z~8. These results provide evidence that the evolution in the LF seen from z~7 to z~3 continues to z~8. The remarkable improvement in the sensitivity of WFC3/IR has enabled HST to cross a threshold, revealing star-forming galaxies at z~8-9.
