We report the discovery of 24 Lyman-break candidates at 7 \lesssim z \lesssim 10.5 , in the Hubble Frontier Fields ( HFF ) imaging data of Abell 2744 ( z = 0.308 ) , plus Spiter /IRAC data and archival ACS data .
The sample includes a triple image system with a photometric redshift of z \simeq 7.4 .
This high redshift is geometrically confirmed by our lens model corresponding to deflection angles that are 12 % larger than the lower-redshift systems used to calibrate the lens model at z = 2.019 .
The majority of our high-redshift candidates are not expected to be multiply lensed given their locations in the image plane and the brightness of foreground galaxies , but are magnified by factors of \sim 1.3 - 15 , so that we are seeing further down the luminosity function than comparable deep field imaging .
It is apparent that the redshift distribution of these sources does not smoothly extend over the full redshift range accessible at z < 12 , but appears to break above z = 9 .
Nine candidates are clustered within a small region of 20 ^ { \prime \prime } across , representing a potentially unprecedented concentration .
Given the poor statistics , however , we must await similar constraints from the additional HFF clusters to properly examine this trend .
The physical properties of our candidates are examined using the range of lens models developed for the HFF program by various groups including our own , for a better estimate of underlying systematics .
Our spectral-energy-distribution fits for the brightest objects suggest stellar masses of \simeq 10 ^ { 9 } M _ { \odot } , star-formation rates of \simeq 4 M _ { \odot } yr ^ { -1 } , and a typical formation redshift of z \lesssim 19 .