The Brightest of Reionizing Galaxies Survey is a large Hubble Space Telescope Program focused on finding rare and bright galaxy candidates at redshift z~8-10, that is when the Universe was about 500 to 650 million years old. We use Hubble’s Wide Field Camera 3, which is highly efficient in the near-infrared bands (Y, J, H) that are needed to identify these very distant galaxies. The near-infrared filters are complemented by deep optical imaging.


BoRG is a pure-parallel survey. This means that the data are acquired while Hubble is pointing for primary spectroscopic observations (typically distant QSOs or extrasolar planets).

BoRG was originally designed in Cycle 17 (GO 11700) to search for z~8 sources and continued with observations in Cycle 19 (GO 12572) and Cycle 20 (GO 12905), augmented by assimilation of other parallel imaging in the Hubble archive (see Bradley et al. 2012 and Schmidt et al. 2014 for the list of pointings and associated programs). Currently, we are boldly moving to Hubble’s detection frontier at z~9-10, and the BoRG Cycle22 observations (GO 13767) are optimized for identifying J-band dropouts that is galaxies observed when the Universe was about 500 Myr old.

We are releasing all science images through the Mikulski Archive for Space Telescopes.

BALFF: Bayesian Algorithm for Luminosity Function Fitting publicly available

Estimating the BoRG z~8 luminosity functionSchmidt et al. (2014) presented a Bayesian framework for obtaining robust and un-biased luminosity functions of samples of photometrically selected Lyman Break galaxies (or any other samples of galaxies for that matter). Mason et al. (2015) extended this formalism to account for the strong and weak lensing magnification bias.

The Bayesian Algorithm for Luminosity Function Fitting (BALFF) used to obtain the Schechter luminosity function fits in Schmidt et al. (2014) and Mason et al. (2015) shown in the figure above, is now publicly available at:


If you find BALFF useful for any current or future work please cite


First results from the [z9-10] BoRG survey!

BoRG postdoc Valentina Calvi (STSci) published the first results from the z~9-10 BoRG survey in the Astrophysical Journal.

In ~25% of the total area in the survey we find 5 galaxy candidates at z ~ 8.3 – 10 detected at high confidence, when the universe is just ~ 500 Myr old. One of these candidates, if confirmed, would be the brightest galaxy ever found at such early times ($$m_{AB} = 24.5$$ at z ~ 8.4). We hope our awarded Spitzer observations will help to confirm these candidates.


Postage stamps, p(z) and SEDs, of the J – H dropout candidates presented in Calvi et al. 2016

These new Lyman-break galaxy candidates are ideal targets for future follow-up to investigate the growth of the first galaxies, and reionization.

Spitzer follow-up of BoRG z>=9 candidates accepted!

BoRG collaborator Rychard Bouwens (Leiden University) has been awarded up to 25.8 hours of priority 1 observing time with Spitzer/IRAC to follow-up BoRG sources!

Spitzer/IRAC will provide the much needed IR observations to distinguish between the true z ~ 9 – 11 galaxies formed just a few hundred million years after the Big Band and dusty, old z ~ 2 galaxies which have similar spectral signatures in HST near-IR broad band imaging. As the reviewer says “Spitzer IRAC is the only instrument that allows for such a sample to be detected… This is very important for preparing for JWST”.

Using IRAC to weed out the low redshift interlopers, we expect to secure a sample of ~5-10 bright z  ~ 9 – 11 galaxies. This will help us to constrain the bright end of the luminosity function at very high redshift, without being affected by cosmic variance, and provide an exciting sample of galaxies for spectroscopic follow-up with Keck and JWST.


Magnification bias in the BoRG luminosity function

BoRG graduate student Charlotte Mason (UCSB) has developed a Bayesian framework to account for the magnification bias from gravitational lensing in luminosity functions, and used the BoRG z~8 LF as a test case. Magnification bias can distort the bright end of high redshift LFs so it is important to understand how it affects our results.

The work has been accepted for publication in the Astrophysical Journal.


We find that magnification bias is currently not significant for the luminosity range of galaxies in BoRG: the effect is only significant for the very brightest galaxies, with $$M_{UV}<-22$$. However, we also show that the effect will dominate future wide-area surveys such as Euclid and WFIRST, which will detect the rarest very bright galaxies, so it will be crucial to account for magnification bias in these surveys.

BoRG Cycle 22 is approved for 32 days of observing time!

A new BoRG survey has been approved for Cycle 22! We boldly proposed to extend the search for the brightest galaxies that Hubble can detect, focusing on redshift z~10 (just 500 Myr after the Big Bang), and we obtained what might well be the largest allocation of Hubble observing time awarded in this cycle: 480 parallel orbits (more than 750 hours).

Pending a successful Phase II proposal, and scheduling, Hubble will spend more than one month observing for our BoRG program!

While we wait for JWST, this new dataset has the potential of being “transformative” (borrowing the word from the reviewers’ comments), and the BoRG team are looking forward to planning, and then analyzing the upcoming observations.

For further information see the associated KICC press release.


Latest BoRG luminosity function at z~8

BoRG postdoc Kasper Borello Schmidt published the latest determination of the z~8 BoRG luminosity function in the Astrophysical Journal.

The luminosity function measured by combining BoRG data with deeper data from the Hubble Ultradeep Field and ERS observations is shown here:


The luminosity function measure shows a very strong evolution in the few hundred Myr separating z~8 from z~6, and like our previous determination in 2012, it finds a steep faint end slope ($$\alpha$$ ~ -1.87 $$\pm$$ 0.26).

We also used the BoRG z~8 luminosity function to perform inference on the epoch of reionization, and in particular on the neutral fraction of hydrogen when the universe was only ~600 Myr old. The BoRG data prefer a ‘late reionization’ scenario, where a significant fraction of the IGM is neutral at z~8, i.e. the reionization was still underway at that time. This is in very good agreement with the results of our recent MOSFIRE follow-up campaign.

In addition to obtaining the largest area determination of the z~8 luminosity function published to date in a refereed journal, Kasper developed a new rigorous Bayesian approach to fit the luminosity function without resorting to binning of the data, commonly employed by previous studies.