A new automatic method to identify galaxy mergers - I. Description and application to the Space Telescope A901/902 Galaxy Evolution Survey

We present a new automatic method to identify galaxy mergers using the morphological information contained in the residual images of galaxies after the subtraction of a smooth Sersic model. The removal of the bulk signal from the host galaxy light is done with the aim of detecting the much fainter and elusive minor mergers. The specific morphological parameters that are used in the merger diagnostic suggested here are the residual flux fraction (RFF) and the asymmetry of the residuals [A(Res)]. The new diagnostic has been calibrated and optimized so that the resulting merger sample is very complete. However, the contamination by non-mergers is also high. If the same optimization method is adopted for combinations of other structural parameters such as the Concentration, Asymmetry, clumpineSs (CAS) system, the merger indicator we introduce yields merger samples of equal or higher statistical quality than the samples obtained through the use of other structural parameters. We investigate the ability of the method presented here to select minor mergers by identifying a sample of visually classified mergers that would not have been picked up by the use of the CAS system, when using its usual limits. However, given the low prevalence of mergers among the general population of galaxies and the optimization used here, we find that the merger diagnostic introduced in this work is best used as a negative merger test, that is, it is very effective at selecting non-merging galaxies. In common with all the currently available automatic methods, the sample of merger candidates selected is heavily contaminated by non-mergers, and further steps are needed to produce a clean merger sample. This merger diagnostic has been developed using the Hubble Space Telescope/ACS F606W images of the A901/902 multiple cluster system (z= 0.165) obtained by the Space Telescope A901/902 Galaxy Evolution Survey team. In particular, we have focused on a mass- and magnitude-limited sample (log M/M > 9.0, R-ega,R-Total <= 23.5 mag) which includes 905 cluster galaxies and 655 field galaxies of all morphological types.