Overview and Motivation for the UDF12 Program
Dramatic progress has been made in the past year in charting the assembly history of the earliest galaxies seen at redshifts z > 6, less than a billion years after the Big Bang. The study of such primitive systems is driven by our desire to understand how and when the Universe was reionized, as well as to determine the physical processes which shaped the mass and luminosity distributions of the emerging generation of star-forming galaxies. These studies represent the final frontier in our quest for a coherent picture of galaxy evolution, and motivate the substantial investments in future facilities such as JWST and the next generation of ground-based near-infrared telescopes.
As the community awaits the next generation facilities, it has become very clear that HST with its new infrared imager, WFC3/IR, can meanwhile play a vital role in exploring this frontier (see Robertson et al. 2010 for a recent review). Through publicly-available images in the Hubble Ultra Deep Field (HUDF) and related deep surveys, important results have emerged (Bouwens et al. 2010a,b; McLure et al. 2010,2011; Bunker et al. 2010; Finkelstein et al. 2010). Rapid progress with HST has been complemented with ambitious ground-based surveys, including panoramic narrow-band imaging of Lyman-α emitting galaxies (LAEs) with Subaru (Ouchi et al. 2010), and ultra-deep spectroscopic surveys of Lyman-break galaxies (LBGs) with Keck and VLT (Stark et al. 2010, 2011; Vanzella et al. 2009, 2010; Lehnert et al. 2010, Schenker et al. 2012).
The questions motivating these studies are fundamental. Sometime in the 600 Myr between a redshift of z = 20 and z = 7, the Universe underwent a phase transition whereby neutral hydrogen created at recombination was reionized. This ‘cosmic reionization’ renders the Universe transparent to ultraviolet (UV) photons and is as significant a milestone in cos- mic history as the decoupling of matter and radiation revealed by the microwave background. Reionization required a sustained source of ionizing photons and the popularly-assumed ex- planation is that it arises from energetic photons that escaped from the first generation of star-forming galaxies (Robertson et al. 2010). The UDF12 program is motivated by the conviction that carefully-planned WFC3/IR observations, building on the past investment of HST in the HUDF, can provide an important test of this hypothesis.
Outstanding Questions and Controversies UDF12 Will Address
As reviewed by Robertson et al. (2010), early star-forming galaxies can maintain reionization over the redshift range 6 < z < 10 provided there is a long-lived population with a dominant contribution from low-luminosity sources and a modest fraction of ionizing photons escape into the intergalactic medium. Our proposal is carefully designed to build on earlier HST investment to address three issues arising from these requirements which collectively hold the key to understanding the contribution of early galaxies to cosmic reionization.

  • What is the form of the faint end of the galaxy luminosity function at redshifts z=7 and z=8?  McLure et al. (2010) and Bouwens et al. (2010a) suggest that the faint-end slope may steepen at early times, counteracting a fading in the characteristic luminosity L∗, thereby ensuring a modest decline in the star formation rate density to z ≃8. However, the available data are too shallow for certainty on this point. Continuity in the abundance of star-forming galaxies over 6 < z < 10 is a key requirement to sustain reionization by galaxies. Securing a reliable form for the luminosity function at z ≃ 7 and z ≃ 8 is achievable with the deeper UDF12 program data.


  • What is the abundance of z ≥ 9 galaxies? The properties of the galaxy population at 9 < z < 11 is particularly contentious. Yan et al. (2010) initially claimed to find 20 faint J125−drops in the HUDF data-set, each of which escaped detection by  McLure et al. (2010) , Bouwens et al. (2010a) and Finkelstein et al. (2010).  Bouwens et al. (2011) initially claimed 3 promising J125−drop candidates at z ≃ 10, none of which feature in Yan et al.’s list. Bouwens et al. (2011) later revised their list of 3 candidates, replacing it with a single new object whose photometric signal/noise depends critically on the aperture used. The confusion arises from searches based on detections in a single band (H160) of inadequate depth. Bouwens et al. claim a relative paucity of z ≃ 10 sources c.f. projections from later times but this is highly speculative given the uncertainties. With increased depth and an improved strategy for locating such sources the UDF12 program can build on earlier data and, for the first time, properly address this key issue. 


  • What does the UV continuum slope displayed by galaxies at z ≥ 7 imply about their ages, metallicities, and the escape fraction of ionizing photons? This issue cuts to the heart of the physics of reionization (Robertson et al. 2010). Some authors report that many of the faintest star-forming galaxies at z ≃ 7 display UV continua with a power-law index β ≃ −3 (where fλ ∝ λβ, Meurer et al. 1999). Such UV continua may be incompatible with current simple stellar population (SSP) models, even assuming extremely low metallicities (Bouwens et al. 2010bBunker et al. 2010). However, Dunlop et al. (2011) demonstrate that the current WFC3/IR images are too shallow to properly address this key issue at the faint luminosities of greatest interest (MUV ≃ −18), and that biases toward spuriously steep slopes can arise as a consequence. Moreover, the filter combination adopted in the current HUDF WFC3/IR imaging campaign means that this crucial measurement cannot even be attempted at z ≃ 8 (where the J125 filter is potentially contaminated by Ly-α emission). The observing strategy of the UDF12 program addresses the requirements both at z ≃ 7 and z≃ 8.