Testing the validity of the phenomenological gravitational waveform models for nonspinning binary black hole searches at low masses

Abstract

The phenomenological gravitational waveform models, which we refer to as PhenomA, PhenomB, and PhenomC, generate full inspiral, merger, and ringdown (IMR) waveforms of coalescing binary back holes (BBHs). These models are defined in the Fourier domain, thus can be used for fast matched filtering in the gravitational wave search. PhenomA has been developed for nonspinning BBH waveforms, while PhenomB and PhenomC were designed to model the waveforms of BBH systems with nonprecessing (aligned) spins, but can also be used for nonspinning systems. In this work, we study the validity of the phenomenological models for nonspinning BBH searches at low masses, m1,2 < 4Msun and m1 + m2 = M < 30Msun, with Advanced LIGO. As our complete signal waveform model, we adopt EOBNRv2, which is a time-domain IMR waveform model. To investigate the search efficiency of the phenomenological template models, we calculate fitting factors (FFs) by exploring overlap surfaces. We find that only PhenomC is valid to obtain FFs better than 0.97 in the mass range of M < 15Msun. Above 15Msun, PhenomA is most efficient in symmetric mass region, PhenomB is most efficient in highly asymmetric mass region, and PhenomC is most efficient in the intermediate region. Specifically, we propose an effective phenomenological template family that can be constructed by employing the phenomenological models in four subregions individually. We find that FFs of the effective templates are better than 0.97 in our entire mass region and mostly greater than 0.99.