Gravitational Waves and Electromagnetic Signals from a Neutron Star Merger

end-to-end physics of NS mergers GRB + afterflow binary stellar evolution (10 6-10 9 years) Final inspiral (minutes) gravitational waves post-merger accretion (seconds) hydrodynamics, gravity, neutrino physics,nuclear reactions, magnetic fields radioactivity (days-weeks) nuclear decay chains (alpha, beta, gamma, fission) thermalization optical/ir kilonova merger dynamics (miliseconds) hydrodynamics, general relativity, nuclear equation of state, neutrino physics, nucleosynthesis (seconds) r-process reaction networks, nuclear data inputs radiation transport (days weeks) Time-dependent spectral Boltzmann transport Atomic microphysics

dynamical t ~ milliseconds S. Rosswog tidal tail ejecta M ~ 10-4 - 10-2 Msun v ~ 0.2c - 0.3c very neutron rich, Ye 0.1 squeezed polar ejecta M ~ 10-4 - 10-2 Msun v ~ 0.2c - 0.3c less neutron rich Ye 0.25 MERGER MASS EJECTION Bauiswein+ 2014 post-merger t ~ seconds Fernandez &Metzger 2013 disk wind ejecta M ~ 10-2 - 10-1 Msun v ~ 0.05c - 0.1c range of Ye = 0.1-0.4

neutrino irradiation of NS merger ejecta weak interactions drive Ye closer to 0.5 (e.g., Metzger & Fernandez 2013) neutrino energy density neutrino mean energy NS $ e + n $ p + e e + p $ n + e + richers, kasen, et al 2015

number of protons Heavy element production via rapid-neutron capture very neutron rich ejecta Ye ~< 0.25 number of neutrons Nuclear reaction network calculations Jonas Lippuner less neutron rich ejecta Ye >~ 0.25

Abundances from r-process nucleosynthesis reaction networks calculations for fixed entropy & expansion time light r-process heavy r-process

Schematic view of NS merger ejecta shocked polar v ~ 0.2c-0.3c M < 0.01 M light r-process disk wind mixed composition v ~ 0.1c M ~ 0.01-0.1 M tidal tails v ~ 0.2c-0.3c M < 0.01 M heavy r-process neutron star + neutron star prompt collapse to black hole

Radioactive decay and thermalization radioactive power rate to Roberts et al. 2011 rate ~ t -1.3 Radioactive power Metzger+2010 Roberts+2011 isotopic contributions time in days fraction Decay products and thermalization efficiencies thermalization efficiency Barnes, DK et al. 2016 efficiency ~ t -1 - t -2 Thermalization efficiency Barnes, Kasen, Wu, Martinez-Pineda 2016 time in days

Radioactive kilonova light curve models Type Ia SN model kilonova models

modeling kilonova light curves and spectra solution to the radiation transport (Boltzmann) equation absorption emission scattering = photon field specific intensity = opacity coefficient = emissivity and set primarily by numerous blended atomic line transitions depends on ionization/excitation state of gas (level populations assumed to be local thermodynamic equilibrium) Transport solved by Monte Carlo methods (Sedona code) e.g,. Kasen+2006, Roth and Kasen (2015)

opacity of r-process kilonova ejecta bound-bound (lines) electron scattering bound-free (photoionization) free-free (bremsstrahlung)

s-shell (g=2) r-process opacity and atomic complexity limited experimental line data requires atomic structure modeling p-shell (g=6) d-shell (g=10) f-shell (g=14)

r-process opacity and atomic complexity Half-filled shells have more complex configurations Atomic line/level data is still sparse (especially in infrared) New atomic-structure calculations cover the statistical properties of all r-process species but uncertainties remain in details (kasen+ in prep) approximate # of atomic levels d p s f atomic number Z lanthanides actinides

Level energy distributions - singly ionized species atomic structure calculations of all r-process species w/ autostructure code thermally populated states Z = 60 Z = 64 Z = 42 Z = 26

lanthanides (f-shell) kilonova opacity from atomic structure modeling (d-shell) kasen+ 2013 T = 5000 K rho = 10-14 g cm -3

Model kilonova spectra dependence on lanthanide fraction kasen, badnell and barnes 2013, barnes & kasen 2013, kasen+2017 SSecific luminoviwy (10 41 eug V 1 µm 1 ) 2.0 1.5 1.0 0.5 UV oswical infuaued X lan =10 5 X lan =10 4 X lan =10 1 X lan =10 1 model spectrum at 4.5 days after merger 0.5 1.0 1.5 2.0 2.5 3.0 WavelengWh (micuon)

Model kilonova light curves: dependence on lanthanide fraction kasen, badnell and barnes 2013, barnes & kasen 2013, kasen+2017 Log 10 bolometrlc lumlnoslty (erg s 1 ) 42.0 41.5 41.0 40.5 40.0 X lan )10 4 X lan )10 1 rddlodctlve hedtlng rdte bolometric light curve 0 2 4 6 8 10 12 DDys slnce merger

Abundances from r-process nucleosynthesis reaction networks calculations for fixed entropy & expansion time light r-process heavy r-process kasen+2015 lanthanides

kilonova SSS17a bolometric light curve radioactivity: Lippuner & Roberts 2015 Q(t) of 0.02 Msun

kilonova AT2017gfo bolometric light curve radioactivity: Lippuner & Roberts 2015 thermalization: Barnes+ 2016 Q(t)*f(t) of 0.04 Msun

Multi-wavelength photometry of SSS17a Villar+2017

kilonova SSS17a bolometric light curve SSS17a bolometric bolometric compilation: Waxman+ 2017 models: Kasen+2017 lanthanide rich lanthanide poor

kilonova AT2017gfo spectrum @ day 1.5

kilonova SSS17a spectrum @ day 2.5 data Pian+2017 x-shooter, models Kasen+2017 light + heavy r-process model rf

Model spectrum dependence on composition features are Doppler-broadened blends of multiple lines 2.5 cerium 10 5elaWive flux F λ plus RfffseW 2.0 1.5 1.0 0.5 nerdymium 0 nerdymium 10 defaulw (srlar) 0.0 0.5 1.0 1.5 2.0 2.5 WavelengWh (micrrn)

Spectral determination of ejecta velocity (consistent with blackbody emitting radius, e.g., Drout+17, Troja+17) b heavy r-prrcess W 4.5 days Kasen+2017 v 0.03c v 0.10c v 0.20c v 0.30c 0.5 1.0 1.5 2.0 2.5 3.0 WavelengWh (micrrns)

GW170817: Some questions Are neutron star mergers a site (the site) of the r-process? Blue kilonova (light r-process) M ~ 0.025 Msun - v ~ 0.3c; Xlan < 10-4 Red kilonova (heavy r-process) M ~ 0.04 Msun - v ~ 0.1c, Xlan ~ 10-2 Merger rate (from LIGO): Rm ~ 1 per 10 4-10 5 years per galaxy Can potentially account for all r-process in galaxy Mgalaxy = 5 x 10 3 Msun = fstar x Mm x Rm x tgal But ejecta masses and rates are certain. Was the 1 event typical?