Failure of the Fisher matrix when including tidal terms: considering construction of template banks of tidally deformed binary neutron stars

Recent gravitational-wave observations have begun to constrain the internal physics of neutron stars. However, current detection searches for neutron star systems assume that potential neutron stars are low-mass black holes, ignoring any affect on the gravitational-wave signal due to the internal neutron star physics. We wish to create a template bank of binary neutron star waveforms including the effect of tidal deformability. However, we find that the Fisher matrix, which is commonly used to approximate match calculations when placing template banks, is unsuitable to predict the match between two binary neutron star waveforms. We find that the Fisher matrix can predict errors on the mismatch that are larger than 100% when attempting to identify waveforms with a match of 0.97. We explore the regime in which the Fisher matrix cannot be trusted and examine why it breaks down. We demonstrate that including higher-order terms in the Taylor series expansion of the match can reliably compute matches for these examples, but that it is prohibitively computationally expensive to do so. Finally, we demonstrate that stochastic placement can still be used to construct a template bank of tidally deformed neutron star waveforms.