By Dan Coe, ESA/AURA Astronomer at STScI
The Hubble and Spitzer Space Telescopes have observed galaxies over 95% of the way back to the Big Bang. The most distant are observed as they were more than 13 billion years ago. These are the building blocks of galaxies like our own Milky Way, which itself dates back to about 13.2 billion years ago.
Two infrared imaging strategies have yielded the most distant galaxy candidates: 1) deep integrations on relatively blank patches of sky such as the Ultra Deep Field (UDF) [1, 2]; and 2) using galaxy clusters as gravitational lenses to magnify the distant universe [3, 4].
Now, for the first time, these two strategies are being combined in the Frontier Fields program. Hubble and Spitzer are obtaining optical and infrared images nearly as deep as the UDF of six lensing galaxy clusters and six nearby “blank” fields over a 3-year period.
The 6 deep blank fields will mitigate the uncertainties previously associated with having only a single UDF (cosmic variance). And factoring in the lensing magnifications, the 6 Hubble cluster images are revealing the faintest sources ever observed (intrinsically nJy, or AB mag > 31).
Figure 1: The second pair of Frontier Fields observed by Hubble: MACSJ0416.1-2403 (left: ACS + WFC3/IR) and a blank field 6′ away (right: WFC3/IR). Each image is ~50″ x 70″, roughly 1/5 the WFC3/IR FOV. The raw HST images have no proprietary period, and STScI processes the images for weekly public releases.
When complete, we estimate these 12 fields may yield ~70 z > 9 candidates (~6 per field), transforming our understanding of the universe’s first 550 million years . To date, only about a dozen candidates are known at these high redshifts [1, 2, 3, 4, 6, 7].
Figure 2: Estimated number counts at z ~ 8 – 12 from the full Frontier Fields program plotted cumulatively as a function of magnitude in the reddest Hubble filter . For these “optimistic” estimates, we extrapolate from a luminosity function evolving with M* consistent with 4 < z < 8 observations. We show predictions for 4 different redshifts both in the field (solid lines) and lensed according to three different publicly available models (dashed lines). The 5-sigma detection limit is F160W AB < 28.7, just twice as bright as the UDF 5-sigma limit F160W AB < 29.45.
So far, two Frontier Fields have deep infrared imaging from both Hubble and Spitzer. These images reveal many high-redshift candidates but so far nothing convincing beyond z > 9 [5, 8, 9]. The UDF previously exhibited a similar dearth of faint z ~ 10 candidates (1 where 9 were expected ). But wider field imaging from CANDELS revealed an excess of brighter z ~ 9-10 candidates (6 where none were expected ). CLASH observations of 25 lensing clusters yielded a z ~ 11 candidate more consistent with extrapolated expectations  but perhaps an overall deficit of z ~ 9-10 candidates .
Figure 3: Candidate z ~ 7.9 – 8.7 galaxies revealed in deep Hubble WFC3/IR imaging of the blank Frontier Field adjacent to MACSJ0416.1-2403 . Likely due to their high redshifts, these galaxies are detected only faintly if at all in the bluest WFC3/IR filter (1.06 microns). (Alternatively, some may be red early type / dusty galaxies at z ~ 1 – 2.)
Are we witnessing a surprisingly rapid buildup in galaxy numbers during the first 500 million years since the Big Bang? Or are these seemingly surprising results simply a product of cosmic variance and small number statistics? The full Frontier Fields program will mitigate these uncertainties delivering robust population statistics and properties of galaxies in the first 400-550 million years (z ~ 9-11).
1. Ellis, R. S. et al. 2013, ApJL, 763, 7
2. Oesch, P. A. et al. 2013, ApJ, 773, 75
3. Coe, D. et al. 2013, ApJ, 762, 32
4. Zheng, W. et al. 2012, Nature, 489
5. Coe, D. et al. 2014, arXiv:1405.0011
6. Oesch, P. A. et al. 2014, ApJ, 786, 108
7. Bouwens, R. et al. 2012, arXiv:1211.2230
8. Atek, H. et al. 2014, ApJ, 786, 60
9. Zheng, W. et al. 2014, arXiv:1402.6743