Spectroscopic Confirmation of Faint Lyman Break Galaxies at Redshifts Four and Five in the Hubble Ultra Deep Field

TL;DR

This study spectroscopically confirms the faintest known Lyman break galaxies at redshifts four and five, providing insights into their emission properties and the effectiveness of photometric selection methods.

Contribution

It presents the first spectroscopic confirmation of very faint high-redshift Lyman break galaxies using HST grism data and compares different photometric selection criteria.

Findings

Confirmed 39 high-redshift galaxies spectroscopically.

Traditional V-dropout criteria are 64% complete and 81% reliable.

About 25% of the sample shows Lyman alpha emission.

Abstract

We present the faintest spectroscopically confirmed sample of redshift four and five Lyman break galaxies to date. The sample is based on slitless grism spectra of the Hubble Ultra Deep Field region from the GRAPES (Grism ACS Program for Extragalactic Science) and PEARS (Probing Evolution and Reionization Spectroscopically) projects, using the G800L grism on the HST Advanced Camera for Surveys. We report here confirmations of 39 galaxies, pre-selected as candidate Lyman break galaxies using photometric selection criteria. We compare a "traditional" V-dropout selection to a more liberal one (with V-i > 0.9), and find that the traditional criteria are about 64% complete and 81% reliable. We also study the Lyman alpha emission properties of our sample. We find that Lyman alpha emission is detected in about 1/4 of the sample, and that our broad-band color selected sample includes 55% of previously published line-selected Lyman alpha sources. Finally, we examine our stacked 2D spectra. We demonstrate that strong, spatially extended (arcsecond scale) Lyman alpha emission is not a generic property of these Lyman break galaxies, but that a modest extension of the Lyman alpha photosphere (compared to the starlight) may be present in those galaxies with prominent Lyman alpha emission.