[Paper Review] Most Luminous z=9-10 Galaxies: A First Determination of the Bright End of the z~9 and z~10 UV Luminosity Functions using all five CANDELS Fields

R. J. Bouwens, Pascal A. Oesch+15 more|arXiv (Cornell University)|Jun 2, 2015

Galaxies: Formation, Evolution, Phenomena7 references16 citations

TL;DR

This study identifies 14 high-probability z≈9–10 galaxies using combined HST/WFC3/IR and Spitzer/IRAC data across all five CANDELS fields, including 4 new sources from a targeted 10-orbit HST follow-up program. It presents the first robust determination of the bright-end UV luminosity function at z~9 and z~10, finding luminosity densities ~2× lower than extrapolated trends from lower redshifts.

ABSTRACT

The deep, wide-area (900 arcmin**2) WFC3/IR + Spitzer/IRAC observations over the CANDELS program represents a significant resource for constraining the bright end of the z~9 and z~10 UV luminosity functions (LFs). We recently reported the discovery of 6 luminous z~9-10 candidates over the GOODS-North+South fields, but extending this search to the full CANDELS program was impeded by the lack of HST-depth 1.05-micron observations in the other 3 CANDELS fields. Here we attempt to significantly realize the potential of CANDELS for z=9-10 science by combining a search over all 5 fields with results from a new HST program (B9-CANDELS) designed to follow up the highest-probability z~9-10 galaxy candidates with observations at 1.05 microns. The targeted z~9-10 candidates are preselected by taking advantage of the full HST, Spitzer/IRAC S-CANDELS observations, and the deepest-available ground-based optical+near-IR observations. With our follow-up program now 91% complete, we identify 4 new high-probability z~9-10 galaxies in just 10 orbits. This brings our total sample of bright z~9-10 galaxies to 14, including several other new sources from the CANDELS GOODS + ERS fields.Through extensive simulations, we replicate the selection process for our sample (both the preselection and follow-up) and obtain an accurate estimate of the volume density of bright galaxies (M_{UV,AB} 0.4L*), the luminosity densities we find at z~9 and z~10 are ~2x lower than the extrapolated trends. We would expect significant additional gains in these results from the on-going 500-orbit BoRG[z910] program and by obtaining additional follow-up observations over the CANDELS-WIDE fields.

Motivation & Objective

• To determine the bright-end UV luminosity function (LF) at z~9 and z~10 using deep, wide-area observations from the CANDELS survey.
• To overcome prior limitations in the CANDELS program by obtaining 1.05-micron HST follow-up observations for high-probability z~9–10 galaxy candidates.
• To improve the statistical reliability of high-redshift galaxy detections by combining multi-wavelength data from HST, Spitzer/IRAC, and ground-based surveys.
• To estimate the volume density of luminous galaxies at z~9 and z~10 using extensive simulations of the selection and follow-up process.

Proposed method

• Preselection of z~9–10 candidates using full HST and Spitzer/IRAC S-CANDELS data across all five CANDELS fields.
• Targeted HST follow-up observations at 1.05 microns for the highest-probability candidates via the B9-CANDELS program.
• Integration of the deepest available ground-based optical and near-IR data to improve photometric redshift estimates and source reliability.
• Use of extensive simulations to model the full selection and follow-up process, enabling accurate estimation of volume density and luminosity function parameters.
• Application of the AB magnitude system to measure UV absolute magnitudes (M_UV,AB) and compare with L* values at high redshift.
• Statistical analysis to derive the UV luminosity function at z~9 and z~10, focusing on the bright end (M_UV,AB ≤ -20.5) and comparing with extrapolated low-redshift trends.

Experimental results

Research questions

• RQ1 What is the true volume density of luminous galaxies at z~9 and z~10, particularly at the bright end of the UV luminosity function?
• RQ2 How do the observed luminosity densities at z~9 and z~10 compare to extrapolations from lower-redshift luminosity functions?
• RQ3 To what extent do the selection and follow-up procedures bias the detection of high-redshift galaxies, and how can this be corrected via simulation?
• RQ4 What is the contribution of the CANDELS-WIDE fields to the overall census of z~9–10 galaxies, and how will future follow-up improve detection rates?
• RQ5 How do the results from the BoRG[z910] program compare with existing CANDELS data in constraining the bright-end LF?

Key findings

• The study identifies a total of 14 high-probability z≈9–10 galaxies, including 4 new sources discovered in just 10 HST orbits through targeted follow-up observations.
• The volume density of luminous galaxies (M_UV,AB ≤ -20.5) at z~9 and z~10 is accurately estimated using simulations that replicate the full selection and follow-up process.
• The luminosity densities at z~9 and z~10 are found to be approximately 2× lower than expected from extrapolations of lower-redshift luminosity function trends.
• The results demonstrate the feasibility and necessity of deep, wide-area surveys with HST and Spitzer to probe the bright end of the high-redshift UV LF.
• The study highlights that significant improvements are expected from the ongoing 500-orbit BoRG[z910] program and future follow-up in the CANDELS-WIDE fields.
• The combination of multi-wavelength data and targeted follow-up enables a robust, simulation-calibrated determination of the z~9–10 UV luminosity function, setting a new benchmark for high-redshift galaxy studies.