Automated selection and characterization of emission-line sources in ACS WFC grism data

Abstract

We present complimentary techniques to find emission-line targets and measure their properties in a semi-automated fashion from grism observations obtained with the Advanced Camera for Surveys aboard the Hubble Space Telescope. The first technique is to find all likely sources in a direct image, extract their spectra and search them for emission lines. The second method is to look for emission-line sources as compact structures in an unsharp masked version of the grism image. Using these methods we identify 46 emission-line targets in the Hubble Deep Field North using a modest (3 orbit) expenditure of HST observing time. Grism spectroscopy is a powerful tool for efficiently identifying interesting low luminosity, moderate redshift emission-line field galaxies. The sources found here have a median i band flux 1.5 mag fainter than the spectroscopic redshift catalog of Cohen et al. They have redshift z <= 1.42, high equivalent widths (typically EW > 100Å), and are usually less luminous than the characteristic luminosity at the same redshift. The chief obstacle in interpreting the results is line identification, since the majority of sources have a single emission line and the spectral resolution is low. Photometric redshifts are useful for providing a first guess redshift. However, even at the depth of the state-of-the-art data used here, photometric errors can result in uncertainties in line identifications, especially for sources with i > ~24.5 ABmag. Reliable line identification for the faintest emission-line galaxies requires additional ground-based spectroscopy for confirmation. Of particular concern are the faint high EW [OII] emitters which could represent a strongly evolving galaxy population if the possibility that they are mis-identified lower redshift interlopers can be ruled out. (Slightly abridged)