After input from the community and thoughtful discussion, we decided to include the F140W filter for the HST Frontier Fields blank field observations in addition to the cluster observations. Without the F140W filter, galaxies at redshifts z >~ 10 would be detected only in the F160W data and as a result, such candidates would naturally be considered much less robust than those with detections in two filters.
Likewise, inclusion of the F140W filter is very beneficial for deriving the redshift of galaxies at redshifts z >~ 9, and it is critical for deriving the redshift of galaxies at redshifts z >~ 10. This has been clearly demonstrated with the triply-imaged z ~ 10.7 candidate behind MACSJ0647.8+7015 from the CLASH MCT program (Coe et al. 2013, Figure 1) and the possible z ~ 11.9 candidate in the HUDF12 program (Ellis et al. 2013, Figure 2), both of which critically depended on the F140W data in determining their redshifts.
The F140W filter is also very useful in deriving the rest-frame UV slopes of galaxies at z <~ 8 by providing three rest-frame UV measurements completely redward of the Lyman break. The addition of the F140W filter to the blank fields provides a homogeneous WFC3/IR filter set over the entire HST Frontier Fields program and the HUDF12 program.
How many galaxies at z >~ 8.5 might we expect to find in the blank field observations? Starting with the z ~ 8 galaxy UV luminosity function derived by Bradley et al. 2012, we extrapolate to higher redshifts by applying optimistic (dM*/dz = 0.36) and pessimistic (dM*/dz = 1.05) evolution of M* based on constraints from the current sample of eight z >~ 9 candidates (Bouwens et al. 2011, Zheng et al. 2012, Coe et al. 2013, Bouwens et al. 2013, Ellis et al. 2013, Oesch et al. 2013). Integrating down to a 5-sigma limit of 28.7 AB, the limiting F160W depth after adding the F140W filter, we estimate the following numbers of galaxies over six blank fields:
|Blank Field High-z Number Counts|
|7.5 < z < 8.5|||||73.2 (-17.4, +20.1)*|
|8.5 < z < 9.5|||||31.7 (-7.4, +8.8)|||||10.5 (-2.7, +3.2)|
|9.5 < z < 10.5|||||13.0 (-3.3, +3.9)|||||0.60 (-0.25, +0.32)|
|10.5 < z < 11.5|||||4.7 (-1.3, +1.5)|||||0.029 (-0.003, +0.015)|
|* Not based on an extrapolation from lower redshift|
Note that the above numbers do not include the effective selection incompleteness, which is likely significant (>~ 0.5) at the 5-sigma limit. Considering 50% completeness, we expect to find ~5-16 z~9 candidates and perhaps up to ~6 and ~2 candidates at z~10 and z~11, respectively, over the six blank fields. The redshift accuracy and study of these high-redshift galaxies are greatly enhanced by the inclusion of the F140W filter.