A kilonova as the electromagnetic counterpart to a gravitational wave source - ePESSTO Letter to Nature and data release

Image obtained by ESO’s Gamma-ray Burst Optical/Near-infrared Detector (GROND) attached to the MPG/ESO 2.2-metre telescope at La Silla Observatory. Credit: ESO/S. Smartt & T.-W. Chen.

The first electromagnetic counterpart to a gravitational wave source has been discovered and the ePESSTO team led one of the most extensive studies on optical and near-infrared evolution of the transient. LIGO-Virgo detected a strong signal from a binary neutron star inspiral and merger for the first time (GW170817).

This is a ground breaking result that opens up a new field in astrophysics by combining the study of light with gravitational waves. The ePESSTO observing team of Joe Lyman and David Homan were on the NTT as the counterpart was found in NGC4993. We took the critical classification spectrum on the night of 2017 Aug 18 at 23:20, 24hrs after the discovery (Lyman et al. GCN 21582) of the transient now known with the IAU Name AT2017gfo. Our spectrum, combined with the fast fading from our early, precise Pan-STARRS photometry (Chambers et al. GCN 21617), showed this really was an unprecedented transient discovery.

The ESO press release, with great media material (images and videos) is available here.

Combining photometry and spectra from the UV (provided by the Swift team), through the optical and near-infrared (Pan-STARRS, GROND and Boyden) with physical modelling we published a letter in Nature showing that this matches theoretical models of merging neutron stars that produce an optical and near-infrared transient known as a kilonova. Our models show, that if the power source is radioactive decay, it has a power-law slope of -1.2 ± 0.3, in good agreement with r-process. We identify caesium and tellurium features in the early spectra, and model the rapidly fading and reddening luminosity to determine mass and opacity of the ejecta. Many questions remain, including the opacity velocity and ejecta composition, and if one or two dynamical components are required. Further modelling of the full data set is needed and to aid this effort we release all our calibrated data here.

The Nature letter is available here (ADS link here) and Stephen’s blog for ESO is here.

Here we release all the data (spectra and accessible photometric tables) from the Nature paper, in an easily useable form. When using these data, please cite Smartt S.J. et al. 2017, Nature, 551, 75-79.

This tarball contains all the spectra in plain ascii format, and the photometric data tables in
ascii and csv format : AT2017gfo_ePESSTO_datafiles.tar.gz

Addendum (2018 May 17)
The publication of all data in October 2017, compiled at the Open Kilonova Catalog, indicated our K-band photometry points from GROND were fainter than other published points at the same epochs (e.g Tanvir et al 2017 ApJ 848. L27, Kasliwal et al. 2017 Science 358, 1559, and Drout et al. 2017,Science 358, 1570). This was due to faint residual flux in the image we used as a reference template for GROND data in the image subtraction method (MJD=57994.97369, phase = 12.324 days). We have corrected this photometry by employing the publicly available VISTA Hemisphere Survey (VHS) K-band image as the reference and recalculating the photometry, based on those frames. In the following tarball we provide all the data as above but with two additional files containing the updated near infrared photometry values (see the README_20180517). All other data products remain as they are. This is the recommended data package to use: AT2017gfo_ePESSTO_datafiles_Kcorrected.tar.gz

Our ePESSTO and VLT xshooter spectra with TARDIS radiative transfer models (see Smartt et al. 2017 for more details). All spectra are released here.