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

Abstract

We present an algorithm for selecting a uniform sample of gravitationallylensed quasar candidates from low-redshift (0.6{\lt}z{\lt}2.2) quasarsbrighter than i=19.1 that have been spectroscopically identifiedin the Sloan Digital Sky Survey (SDSS). Our algorithm uses morphologicaland color selections that are intended to identify small- and large-separationlenses, respectively. Our selection algorithm only relies on parametersthat the SDSS standard image processing pipeline generates, allowingeasy and fast selection of lens candidates. The algorithm has beentested against simulated SDSS images, which adopt distributions offield and quasar parameters taken from the real SDSS data as input.Furthermore, we take differential reddening into account. We findthat our selection algorithm is almost complete down to separationsof 1'' and flux ratios of 10^{-0.5}. The algorithm selects bothdouble and quadruple lenses. At a separation of 2'', doubles andquads are selected with similar completeness, and above or below2'' the selection of quads is better or worse, respectively, thanfor doubles. Our morphological selection identifies a nonnegligiblefraction of single quasars: to remove these we fit images of candidateswith a model of two point sources and reject those with unusuallysmall image separations and/or large magnitude differences betweenthe two point sources. We estimate the efficiency of our selectionalgorithm to be at least 8% at image separations smaller than 2'',comparable to that of radio surveys. The efficiency declines as theimage separation increases because of larger contamination from stars.We also present the magnification factor of lensed images as a functionof the image separation, which is needed for accurate computationof magnification bias.