Coalescing black-hole–black-hole and neutron-star–black-hole binaries are among the most promising sources for first-generation ground-based interferometers, but the possibility of detecting them with standard matched-filtering techniques is seriously imperiled by our ignorance about the gravitational waveforms expected from these systems, and by their sheer complexity. This is because gravitational-wave searches proceed by systematically correlating the detector output with signal templates (the theoretical models of the waveforms).
Since 2001, in collaboration with Alessandra Buonanno, Yanbei Chen, and Yi Pan, I have been working to devise robust detection methods for these binaries.
The gravitational-wave signals emitted by inspiraling black-hole–black-hole binaries with total masses between 10 and 40 solar masses enter the frequency range of good interferometer sensitivity at a very late stage in the inspiral, when only few orbits are left before the merger, and when different post–Newtonian approximations predict different waveforms. It follows that no template family based on a single approximation can be trusted to represent the physical signals adequately.
In 2001 and 2002, Buonanno, Chen, and I developed a phenomenological template family that embeds all the credible approximated models and extrapolates beyond them. This template family (known as BCV) was used in binary-inspiral searches on LIGO data from runs S2 to S4.
A second complication is introduced by black-hole spins, which play a significant role in the late stages of black-hole–black-hole inspirals. The efficient detection of these systems requires modeling the precession-induced modulations of the signal, which depend on a prohibitively large number of parameters (including the initial directions of the spins, and the position and orientation of the binary with respect to the GW detector).
In 2002, Buonanno, Chen, and I developed a family of modulated detection templates that are functions of very few physical and phenomenological parameters, and that model remarkably well the dynamical and precessional effects of spin. This template family (known as BCV2) was used in spinning-binary–inspiral searches on LIGO data from run S3.
Buonanno, Chen, and I also proposed a separate template family for spinning binaries, consisting of essentially exact waveforms written directly in terms of the physical mass and spin parameters of the binary. In 2003, working with Yi Pan, we devised a fast matched-filtering scheme for these templates. The crucial insight was a new decomposition of detector response into a time-dependent component that encodes the evolution of the binary, and a constant component that represents the relative position and orientation of detector and binary.
I am now implementing this scheme (in collaboration with Diego Fazi, Curt Cutler, and others) for searches in LIGO data from the flagship run S5 (one year of triple coincidence at design sensitivity), and from the runs at enhanced sensitivity that will begin in 2009.
Go back to my research page.