I have a couple of issues with the article. First and foremost is using gravitational lensing to measure redshift. Gravitational lensing distorts the light, and therefore distorts the spectrum. Redshift, on its own, is not a good measure of distance. A much better "standard candle," especially for vast distances, are type 1a supernovas. We know the absolute magnitude of all type 1a supernovas. By viewing the type 1a supernova in the visible spectrum we can determine its apparent magnitude. With both the absolute and apparent magnitudes we can calculate distance. Combined with an analysis of the spectrum we get a much better idea of its distance.
Second, they did not mention the name of the galaxy in the article. However, upon further research I believe the galaxy is called "MACS1149-JD1," but if that is the case then this discovery was made in April 2012, not September as the article implies.
Lastly, the article states that "[t]he first luminous stars and their host galaxies emerged a few hundred million years later. The energy released by these earliest galaxies is thought to have caused the neutral hydrogen strewn throughout the universe to ionize, or lose an electron, a state that the gas has remained in since that time." The very first luminous stars would have formed before galaxy formation, about 30 million years after the Big Bang. Galaxies formed around 300 million years after the Big Bang.
Originally they estimated the Epoch of Reionization lasted about one billion years, but recent findings suggest that it may have only taken about 500 million years to reach the ~10% opacity that we see today as a result of ULIRGs.
The article below says that MACS1149-JD1 was discovered using "a Hubble survey called CLASH (Cluster Lensing and Supernova Survey with Hubble) to find the lensing cluster, MACS1149" because the object is too distant to be seen by Hubble in the visible spectrum.
Source:
Infant galaxy offers tantalizing peek at early Universe : Nature News & Comment