Nicolas Andres – Phd’s defense – sep. 9, 2023

The LIGO Virgo collaboration marked the beginnings of gravitational astronomy by providing direct evidence of their existence in September 2015. The detection of gravitationnal wave coming from a binary black holes merger led to the physic’s Nobel price. This field has since experienced a great growth, each discovery of which allows an advance in disciplines such as astrophysics, cosmology and fundamental physics. At the end of each observation period, the detectors are stopped and many aspects are improved. This work is part of the preparation phase between period O3 and O4 beginning in May 2024 to configure interferometers in their advanced states in order to optimize their sensitivities. Calibration then becomes crucial in order to accurately reconstruct the signal containing gravitational wave information, allowing detection and the production of scientific results such as the measurement of the Hubble constant, etc. An instrumentation work has been carried out, allowing an accurate and regular measurement of the time stamp (timing) of the readout sensing chain of the interferometer signal, which must be mastered better than 0.01 ms for the purpose of a joint analysis of the detectors network data. Many devices for the calibration of the interferometer rely on the reading of control signals by photodetectors whose frequency response has been assumed to be flat. In order to avoid any bias introduced in the reconstruction of the signal, a measurement method must be developed for a frequency calibration of each photo detector involved. 

Two methods are compared for use in the O5 period. In addition, the increasing sensitivity of the detectors means more detections. Collaboration analysis chains need to follow instrumental improvements by developing new tools to optimize real-time and off-ligne signal search. The MBTA Low Latency Analysis Chain is one of 4 collaboration analysis pipelines focusing on the search for compact binary coalescences by combining independent data analysis from all 3 detectors. It has many powerful noise rejection tools, but does not take into account any astrophysical information a priori. Through the accumulation of data in previous observation periods, the collaboration was able to establish more accurate mass distribution models for compact binary coalescence populations. During my thesis, a new tool was developed by the MBTA team using this new information, aimed at estimating the probability of origin of events (astrophysics or not) and at classifying the nature of the astrophysical source. This tool finally made it possible to restructure the global analysis chain by using it as the main parameter for classifying events according to their level of significance. The collaboration produces low-latency public alerts for multi-messenger astronomy, providing information related to detected signals common to the different analytical pipelines. 

Not knowing in advance the preferences of the different experiences partners of the LIGO Virgo collaboration to define the optimal parameters allowing multi-messenger detections, it was decided to test another method to implement similar astrophysical information in the MBTA analysis chain. A technique for including astrophysical information directly in the parameter defining the ranking by significance level of candidate events is presented. This method makes it possible to improve research by providing better discrimination between astrophysical and background noise events. By considering the observation period O3 this method makes it possible to increase the number of detection by 10% with MBTA , detections that have been confirmed by the other chains of analysis.