We report the discovery and monitoring of the near-infrared counterpart (AT2017gfo) of a binary neutron-star merger event detected as a gravitational wave source by Advanced LIGO/Virgo (GW170817) and as a short gammaray burst by Fermi /GBM and Integral /SPI-ACS (GRB 170817A). The evolution of the transient light is consistent with predictions for the behaviour of a "kilonova/macronova", powered by the radioactive decay of massive neutronrich nuclides created via r-process nucleosynthesis in the neutron-star ejecta. In particular, evidence for this scenario is found from broad features seen in Hubble Space Telescope infrared spectroscopy, similar to those predicted for lanthanide dominated ejecta, and the much slower evolution in the near-infrared K s -band compared to the optical. This indicates that the late-time light is dominated by high-opacity lanthanide-rich ejecta, suggesting nucleosynthesis to the 3rd r-process peak (atomic masses A ≈ 195). This discovery confirms that neutron-star mergers produce kilo-/macronovae and that they are at least a major -if not the dominant -site of rapid neutron capture nucleosynthesis in the universe.

Extended emission (EE) is a high-energy, early time rebrightening sometimes seen in the light curves of short gamma-ray bursts (GRBs). We present the first contiguous fits to the EE tail and the later X-ray plateau, unified within a single model. Our central engine is a magnetar surrounded by a fall-back accretion disc, formed by either the merger of two compact objects or the accretion-induced collapse of a white dwarf. During the EE phase, material is accelerated to super-Keplarian velocities and ejected from the system by the rapidly rotating (P ≈ 1 − 10 ms) and very strong (10 15 G) magnetic field in a process known as magnetic propellering. The X-ray plateau is modelled as magnetic dipole spin-down emission. We first explore the range of GRB phenomena that the propeller could potentially reproduce, using a series of template light curves to devise a classification scheme based on phenomology. We then obtain fits to the light curves of 9 GRBs with EE, simultaneously fitting both the propeller and the magnetic dipole spin-down and finding typical disc masses of a few 10 −3 M ⊙ to a few 10 −2 M ⊙ . This is done for ballistic, viscous disc and exponential accretion rates. We find that the conversion efficiency from kinetic energy to EM emission for propellered material needs to be 10% and that the best fitting results come from an exponential accretion profile.

We present observations of the optical afterglow of GRB 170817A, made by the Hubble Space Telescope, between February and August 2018, up to one year after the neutron star merger, GW170817. The afterglow shows a rapid decline beyond 170 days, and confirms the jet origin for the observed outflow, in contrast to more slowly declining expectations for 'failed-jet' scenarios. We show here that the broadband (radio, optical, X-ray) afterglow is consistent with a structured outflow where an ultrarelativistic jet, with Lorentz factor Γ 100, forms a narrow core (∼ 5 • ) and is surrounded by a wider angular component that extends to ∼ 15 • , which is itself relativistic (Γ 5). For a two-component model of this structure, the late-time optical decline, where F ∝ t −α , is α = 2.20 ± 0.18, and for a Gaussian structure the decline is α = 2.45 ± 0.23. We find the Gaussian model to be consistent with both the early ∼ 10 days and late 290 days data. The agreement of the optical light curve with the evolution of the broadband spectral energy distribution, and its continued decline, indicates that the optical flux is arising primarily from the afterglow and not any underlying host system. This provides the deepest limits on any host stellar cluster, with a luminosity 4000L (M F606W −4.3).