The Sloan Digital Sky Survey Quasar Lens Search. I. Candidate Selection Algorithm

Abstract

We present an algorithm for selecting an uniform sample of gravitationally lensed quasar candidates from low-redshift (0.6<z<2.2) quasars brighter than i=19.1 that have been spectroscopically identified in the SDSS. Our algorithm uses morphological and color selections that are intended to identify small- and large-separation lenses, respectively. Our selection algorithm only relies on parameters that the SDSS standard image processing pipeline generates, allowing easy and fast selection of lens candidates. The algorithm has been tested against simulated SDSS images, which adopt distributions of field and quasar parameters taken from the real SDSS data as input. Furthermore, we take differential reddening into account. We find that our selection algorithm is almost complete down to separations of 1'' and flux ratios of 10^-0.5. The algorithm selects both double and quadruple lenses. At a separation of 2'', doubles and quads are selected with similar completeness, and above (below) 2'' the selection of quads is better (worse) than for doubles. Our morphological selection identifies a non-negligible fraction of single quasars: To remove these we fit images of candidates with a model of two point sources and reject those with unusually small image separations and/or large magnitude differences between the two point sources. We estimate the efficiency of our selection algorithm to be at least 8% at image separations smaller than 2'', comparable to that of radio surveys. The efficiency declines as the image separation increases, because of larger contamination from stars. We also present the magnification factor of lensed images as a function of the image separation, which is needed for accurate computation of magnification bias.