2023MNRAS.525.3384K

We study kilonova emission from binary neutron star (BNS) mergers for the case that a remnant massive neutron star (MNS) forms and collapses to a black hole within 20 ms after the onset of the merger (which we refer to as 'a short-lived case') by consistently employing numerical relativity and nucleosynthesis results. We find that such kilonovae are fainter and last shorter than those for BNSs resulting in the formation of long-lived (${\gg} 1\, {\rm s}$) MNSs, in particular in the optical band. The resulting light curves are too faint and last for a too short duration to explain the kilonova observation for the BNS associated with GW170817, indicating that the merger remnant formed in GW170817 is unlikely to have collapsed to a black hole within a short period of time (∼20 ms) after the onset of the merger. Our present result implies that early observation is necessary to detect kilonovae associated with BNSs leading to short-lived MNS formation in particular for the optical blue band as well as that kilonovae could be hidden by the gamma-ray burst afterglow for nearly face-on observation. We provide a possible approximate scaling law for near-infrared light curves with the given reference time and magnitude when the decline power of the z-band magnitude, dM_z/dlog₁₀t, reaches 2.5. This scaling law suggests that the HK-band follow-up observation should be at least 1 mag deeper than that for the z-band reference magnitude and earlier than 4 times the reference time.