D'Onofrio, Luca (2022) Targeted search and ensemble analysis for the detection of gravitational waves from known pulsars. [Tesi di dottorato]

Anteprima

Testo D'Onofrio_Luca_35.pdf Download (5MB) | Anteprima

Abstract

Continuous gravitational waves (CWs) are promising signals targeted by the LIGO and Virgo detectors; possible sources of this type of signals are spinning neutron stars with a non-axisymmetric mass distribution. The "strength" of the signal depends on the ellipticity, the physical parameter that quantifies the mass distribution asymmetry respect to the rotation axis. According to the assumption and the knowledge of the source parameters (sky position, rotation frequency parameters), different strategies can be adopted to search for a CW signal. The simplest, but also the most sensitive strategy, is the targeted search: measuring the electromagnetic emission of the source, as in the case of pulsars, it is possible to infer with high accuracy the sky position and the rotation parameters. So far, there is no evidence of a CW signal in LIGO and Virgo data. The 5n-vector ensemble method is a multiple test for the targeted search of the CW emission from a collection of known pulsars. This method can improve the detection probability combining the results from individually undetectable pulsars if few signals are near the detection threshold. In this thesis, I describe and test the 5n-vector ensemble method showing a significant improvement in the detection probability. Then, I apply this procedure to the O3 dataset from the LIGO and Virgo detectors considering an ensemble of 223 known pulsars. I find no evidence of a signal from the ensemble and set 95% credible upper limit on a global parameter for the ensemble and, assuming a common exponential distribution for the pulsars’ ellipticitites, 95% credible upper limit on the mean ellipticity. Using two independent hierarchical Bayesian procedures, we find 95% upper limit of 2.7 × 10^−9 and 1.8 × 10^−9 on the mean of the assumed exponential distributed ellipticity distribution for the 223 pulsars. This results are more than one order of magnitude below the upper limit of a different search where the authors considered a Bayesian procedure on an ensemble of 92 pulsars and data from the LIGO V6 science run.

Downloads

Actions (login required)

Modifica documento