2014MNRAS.439.3986R

This paper presents a technique to search for supermassive black hole binaries (MBHBs) in the Sloan Digital Sky Survey (SDSS). The search is based on the peculiar properties of merging galaxies that are found in a mock galaxy catalogue from the Millennium Simulation. MBHBs are expected to be the main gravitational wave (GW) sources for pulsar timing arrays (PTAs); however, it is still unclear if the observed GW signal will be produced by a few single MBHBs, or if it will have the properties of a stochastic background. The goal of this work is to produce a map of the sky in which each galaxy is assigned a probability of having suffered a recent merger, and of hosting a MBHB that could be detected by PTAs. This constitutes a step forward in the understanding of the expected PTA signal: the skymap can be used to investigate the clustering properties of PTA sources and the spatial distribution of the observable GW signal power; moreover, galaxies with the highest probabilities could be used as inputs in targeted searches for individual GW sources. We also investigate the distribution of neighbouring galaxies around galaxies hosting MBHBs, finding that the most likely detectable PTA sources are located in dense galaxy environments. Different techniques are used in the search, including Bayesian and machine learning algorithms, with consistent outputs. Our method generates a list of galaxies classified as MBHB hosts, that can be combined with other searches to effectively reduce the number of misclassifications. The spectral coverage of the SDSS reaches less than a fifth of the sky, and the catalogue becomes severely incomplete at large redshifts; however, this technique can be applied in the future to larger catalogues to obtain complete, observationally based information of the expected GW signal detectable by PTAs.