Probing the internal composition of neutron stars with gravitational waves

Abstract

Gravitational waves from neutron star binary inspirals contain information about the as yet unknown equation of state of supranuclear matter. In the absence of definitive experimental evidence that determines the correct equation of state, a number of diverse models that give the pressure inside a neutron star as function of its density have been constructed by nuclear physicists. These models differ not only in the approximations and techniques they employ to solve the many-body Schrödinger equation, but also in the internal neutron star composition they assume. We study whether gravitational wave observations of neutron star binaries in quasicircular inspirals up to contact will allow us to distinguish between equations of state of differing internal composition, thereby providing important information about the properties and behavior of extremely high density matter. We carry out a Bayesian model selection analysis, and find that second generation gravitational wave detectors can heavily constrain equations of state that contain only quark matter, but hybrid stars containing both normal and quark matter are typically harder to distinguish from normal matter stars. A gravitational wave detection with a signal-to-noise ratio of 20 and masses around 1.4M⊙ would provide indications of the existence or absence of strange quark stars, while a signal-to-noise ratio 30 detection could either detect or rule out strange quark stars with a 20 to 1 confidence. The presence of kaon condensates or hyperons in neutron star inner cores cannot be easily confirmed. For example, for the equations of state studied in this paper, even a gravitational wave signal with a signal-to-noise ratio as high as 60 would not allow us to claim a detection of kaon condensates or hyperons with confidence greater than 5 to 1. On the other hand, if kaon condensates and hyperons do not form in neutron stars, a gravitational wave signal with similar signal-to-noise ratio would be able to constrain their existence with an 11 to 1 confidence for high-mass systems. We, finally, find that combining multiple lower signal-to-noise ratio detections (stacking) must be handled with caution since it could fail in cases where the prior information dominates over new information from the data.