The rapid reddening and featureless optical spectra of the optical counterpart of GW170817, at 2017gfo, during the first four days

We present the spectroscopic evolution of AT 2017gfo, the optical counterpart of the first binary neutron star (BNS) merger detected by LIGO and Virgo, GW170817. While models have long predicted that a BNS merger could produce a kilonova (KN), we have not been able to definitively test these models until now. From one day to four days after the merger, we took five spectra of AT 2017gfo before it faded away, which was possible because it was at a distance of only 39.5 Mpc in the galaxy NGC 4993. The spectra evolve from blue (∼6400 K) to red (∼3500 K) over the three days we observed. The spectra are relatively featureless-some weak features exist in our latest spectrum, but they are likely due to the host galaxy. However, a simple blackbody is not sufficient to explain our data: another source of luminosity or opacity is necessary. Predictions from simulations of KNe qualitatively match the observed spectroscopic evolution after two days past the merger, but underpredict the blue flux in our earliest spectrum. From our best-fit models, we infer that AT 2017gfo had an ejecta mass of 0.03 M⊙, high ejecta velocities of 0.3c, and a low mass fraction ∼10-4 of high-opacity lanthanides and actinides. One possible explanation for the early excess of blue flux is that the outer ejecta is lanthanide-poor, while the inner ejecta has a higher abundance of high-opacity material. With the discovery and follow-up of this unique transient, combining gravitational-wave and electromagnetic astronomy, we have arrived in the multi-messenger era.