We present radiative transfer simulations for blue kilonovae hours after neutron star (NS) mergers by performing detailed opacity calculations for the first time. We calculate atomic structures and opacities of highly ionized elements (up to the 10th ionization) with atomic number Z = 20–56. We find that the bound–bound transitions of heavy elements are the dominant source of the opacities in the early phase (t < 1 day after the merger) and that the ions with a half-closed electron shell provide the highest contributions. The Planck mean opacity for lanthanide-free ejecta (with electron fraction of Y e = 0.30–0.40) can only reach around at t = 0.1 days, whereas that increases up to at t = 1 day. The spherical ejecta model with an ejecta mass of M ej = 0.05 M ⊙ gives the bolometric luminosity of at t ∼ 0.1 days. We confirm that the existing bolometric and multicolor data of GW170817 can be naturally explained by the purely radioactive model. The expected early UV signals reach 20.5 mag at t ∼ 4.3 hr for sources even at 200 Mpc, which is detectable by the facilities such as Swift and the Ultraviolet Transient Astronomy Satellite (ULTRASAT). The early-phase luminosity is sensitive to the structure of the outer ejecta, as also pointed out by Kasen et al. Therefore, the early UV observations give strong constraints on the structure of the outer ejecta and the presence of a heating source besides r-process nuclei.

We investigate the effect of the presence of lanthanides (Z = 57–71) on the kilonova at t ∼ 1 hr after the neutron star merger for the first time. For this purpose, we calculate the atomic structures and the opacities for selected lanthanides: Nd (Z = 60), Sm (Z = 62), and Eu (Z = 63). We consider the ionization degree up to 10th (XI), applicable for the ejecta at t ∼ a few hours after the merger, when the temperature is T ∼ 105 K. We find that the opacities for the highly ionized lanthanides are exceptionally high, reaching κ exp ∼ 1000 cm 2 g − 1 for Eu, due to the highly dense energy levels. Using the new opacity, we perform radiative transfer simulations to show that the early light curves become fainter by a (maximum) factor of four, in comparison to lanthanide-free ejecta at t ∼ 0.1 days. However, the period at which the light curves are affected is relatively brief owing to the rapid time evolution of the opacity in the outermost layer of the ejecta. We predict that for a source at a distance of ∼100 Mpc, UV brightness for lanthanide-rich ejecta shows a drop to ∼21–22 mag at t ∼ 0.1 days and the UV peaks around t ∼ 0.2 days with a magnitude of ∼19 mag. Future detection of such a kilonova by an existing UV satellite like Swift or the upcoming UV satellite ULTRASAT will provide useful constraints on the abundance in the outer ejecta and the corresponding nucleosynthesis conditions in the neutron star mergers.

A gravitational wave event, S190510g, which was classified as a binary-neutron-star coalescence at the time of preliminary alert, was detected by LIGO/Virgo collaboration on 2019 May 10. At 1.7 hours after the issue of its preliminary alert, we started a target-of-opportunity imaging observation in the Y band to search for its optical counterpart using the Hyper Suprime-Cam (HSC) on the Subaru Telescope. The observation covers a 118.8 deg2 sky area corresponding to $11.6\%$ confidence in the localization skymap released in the preliminary alert and $1.2\%$ in the updated skymap. We divided the observed area into two fields based on the availability of HSC reference images. For the fields with the HSC reference images, we applied an image subtraction technique; for the fields without the HSC reference images, we sought individual HSC images by matching a catalog of observed objects with the PS1 catalog. The search depth is 22.28 mag in the former method and the limit of search depth is 21.3 mag in the latter method. Subsequently, we performed visual inspection and obtained 83 candidates using the former method and 50 candidates using the latter method. Since we only have the one-day photometric data, we evaluated the probability of candidates being located inside the 3D skymap by estimating their distances with photometry of associated extended objects. We found three candidates are likely located inside the 3D skymap and concluded they could be a counterpart of S190510g, while most of the 133 candidates were likely to be supernovae because the number density of candidates was consistent with the expected number of supernova detections. By comparing our observational depth with a light-curve model of such a kilonova reproducing AT2017gfo, we show that early deep observations with the Subaru/HSC can capture the rising phase of the blue component of a kilonova at the estimated distance of S190510g (∼230 Mpc).

Aims. We present extensive energy level and transition data for the Ce IV spectrum. By providing accurate atomic data, we evaluate the impact of atomic data on the opacity in the neutron star merger ejecta. Methods. We performed energy spectra and transition data calculations using the GRASP2018 package, which is based on the multiconfiguration Dirac-Hartree-Fock and relativistic configuration interaction methods, and the HULLAC code, which is based on a parametric potential method. Results. We present energy spectra calculated for the 225 levels for the Ce 3+ ion. Energy levels are compared with recommended values from the NIST Atomic Spectra Database and other available works. The root-mean-square (rms) deviations obtained for the GRASP2018 energy levels of the 5p 6 nl configurations from the NIST data are 1270 cm −1 . The rms deviations for the HULLAC results from the NIST data are 5780 cm −1 . Furthermore, electric dipole (E1) transition data, line strengths, weighted oscillator strengths, and transition rates are computed between the above levels. The computed transition rates are compared with other theoretical computations. We also evaluate the accuracy of the wave functions and transition parameters by analyzing the dependencies of the line strength S on the gauge parameter G. The gauge dependency method also allows us to determine the transitions for which the ratio between the Babushkin and Coulomb gauges shows real agreement between forms and the transitions for which the agreement between both gauges is random. Using the GRASP2018 and HULLAC data, the opacities in the neutron star merger ejecta are also calculated. We find that the opacity of Ce IV is higher than that presented by previous works, which is because of the higher completeness of our atomic data. Although the differences in the energy levels and transition probabilities cause different features in the opacity spectrum, the Planck mean opacities of both data sets agree within 20 %.

No abstract