2015A&A...581A..99F

Galaxy-scale gravitational lenses are powerful tools, which can be used to address major astrophysical questions that are still open. They can be identified either through imaging or through spectroscopy, which is less efficient than imaging but offers the major advantage of having both source and deflector red shift previously measured at discovery. Spectroscopic identification requires huge data sets of high spectral quality, such as the SDSS, and has so far focused on early-type galaxies, as the most massive galaxies are found among them. We aimed to perform spectroscopic identification of disk galaxies acting as gravitational lenses. We have selected about 300000 galaxy spectra with EW(H_α)≤-10Å from the SDSS DR 8. On these spectra, we ran our original code RES, which is a fast, reliable tool able to provide a red-shift measure and to identify discordant red-shift systems if present. We have required RES to identify only systems based on a minimum number of four emission lines. We have inspected all the (54) SDSS images of the double z systems identified by RES and discarded systems for which z duplicity could be easily ascribed to the presence of two distinct objects. The remaining 25 systems, for which double z is very likely to be due to the gravitational lensing phenomenon, constitute our sample. For each gravitational lens candidate system, we provide SDSS identification and image emission lines detected by RES and activity classification, when derivable. The disky nature of our candidate lenses is confirmed by their images, stellar mass estimates, g-r rest-frame colours and occurrence of star burst phenomena.