Sub-dominant modes of the gravitational radiation from compact binary coalescencesconstruction of hybrid waveforms and impact on gravitational wave searches

Resumo

During the last decades, a worldwide effort leaded by the LIGO, Virgo and GEO600 have pursued without success the first direct detection of the gravitational waves (GW) predicted by Einstein’s general relativity (GR). During this year and the next ones, a new generation of GW detectors up to ten times more sensitive than the previous ones will explore the cosmos searching for GW signals. This makes the scientific community to be confident that we are on the verge of the first direct observation of GW. Among the possible GW sources, one of the most promising ones are the compact binary coalescences (CBC). These consist of couples of inspiraling black holes and/or neutron stars which eventually merge. The detection process of these systems is based on the matched filter technique. This requires to have at our disposal precise models (or templates) of the signals we expect to detect, which are used as filters of the incoming signal. However, the templates used in current searches neglect the higher order mode content of the signal, considering only the contribution from its dominant harmonic. A CBC can be considered to have three stages: inspiral, merger and ringdown. The GW radiation emitted during the first stage can be analytically modeled in the framework of the post-Newtonian (PN) approximation. However, the strong gravitational fields and high velocities present during the late inspiral and merger makes necessary to solve the full Einstein equations. This is only possible in the framework of numerical relativity, with the help of supercomputers. This thesis is focused on the study of the consequences of the neglection of higher order modes in current searches in terms of loss of detections and errors in the measurement of the parameters of the corresponding source. To this end, we will first build GW signals including higher order modes, that we will eventually use as our model of the real signal. This will be addressed by constructing hybrid waveforms, combination of the PN and NR result, including in this process the higher harmonics of the signal. This process will motivate a full study of the accuracy of the PN and NR higher order modes and the corresponding sources of error. Results indicate that the dominant error source in PN those due to the truncation of the PN series while NR errors are dominated by those due to the finitude of the radius at which the signal is extracted by NR codes. In a second step, we will use our hybrid waveforms as model of the real signal emitted by equal spin CBC’s and check the ability of current templates for detecting these signals.Results indicate that for the case of the future Advanced LIGO detector, losses of more than 10% of events will happen due to neglection of higher order modes for systems with mass ratio q ≥ 6 and total mass M > 100M and that parameter estimation is likely to be affected by systematic biases due to neglection of higher order modes for systems with total mass M > 170M for a signal-to-noise ratio of∼ 8. However, the situation is worse for the upcoming early Advanced LIGO, for which losses of 10% happen for q ≥ 4, reaching values of 26% for the worst cases. Also, systematic parameter biases will affect parameter estimation for systems of total mass M > 80M .