Searching electromagnetic counterparts of gravitational wave signals

Gravitational-wave (GW) were detected on 14 September 2015 by the Laser Interferometer Gravitational-wave Observatory (LIGO). The following challenge was the joint observation of a Compact binary coalescence (CBC) in both GW and Electromagnetic (EM) channels. This is difficult because GW sky location uncertainties are typically tens or hundreds of square degrees. Multimessenger observations of binary system containing a neutron star were expected to answer many open questions of modern astrophysics, from the nature of short GRB to the origin of heavy elements. For this reason the astronomical community worldwide was preparing for this event. The work of this thesis was developed in this context. My work focused on the search of the possible optical counterparts of GW events. For the search of the expected optical transient I tested, implemented and exploited two complementary approaches using the data of the observing facilities available to the collaborations of which I am member: a) The GRAvitational Wave Inaf TeAm (GRAWITA), which performs an optical transient search with the 2.6 meter VLT survey telescope (VST). As part of GRAWITA, I developed the transient detection pipeline based on image difference of the wide field survey. I applied this tool to the follow up of three GW triggers, GW150914, GW151226 and GW170814. b) The Distance Less Than 40 Mpc survey (DLT40), which makes use of 40 cm robotic telescopes for targeting individual galaxies. In this project, I developed a prioritization algorithm to select galaxies inside the GW error-box, with the aim to maximise the detection probability in case of nearby triggers. My algorithm was used to define the strategy and follow-up ten GW triggers. After a few inconclusive attempts, on August 17, 2017 with DLT40, I contributed to the discovery of the first optical counterpart of a GW source, DLT17ck (labelled also AT2017gfo and SSS17a). With GRAWITA we observed this source with an almost daily cadence for two weeks both in imaging and spectroscopy, proving that the binary neutron star (BNS) merging produces r-process elements. I used the previous record of the DLT40 SN search to derive one of the first direct estimate of the BNS rates. I also contributed to a first attempt to measure the Hubble constant from combined GW-EM observations. Finally, I developed a machine learning algorithm with the aim of a more rapid and efficient transient candidate selection. This tool is already implemented in the ongoing DLT40 SN survey and it will be used by GRAWITA in the incoming LIGO-VIRGO collaboration (LVC) O3 run.