PhD thesis defense: Novel Strategies for Continuous Gravitational Wave Searches in the Era of the Advanced Detectors.

The 30th June of 2023 Rodrigo Tenorio Márquez presented his PhD thesis: Novel Strategies for Continuous Gravitational Wave Searches in the Era of the Advanced Detectors.

Title: Novel Strategies for Continuous Gravitational Wave Searches in the Era of the Advanced Detectors.
Author: Rodrigo Tenorio Márquez
Directors: Dra. Alicia Magdalena Sintes Olives i Dr. David Benjamin Keitel

Abstract: Continuous gravitational waves (CWs) are long-lasting forms of gravitational radiation whose detection is yet to be achieved. The expected sources of such signals are rapidly-spinning non-axisymmetric neutron stars (NSs) within our galaxy, even though more exotic sources, such the evaporation of boson clouds around spinning black holes, have been also considered in the literature. A direct detection of a CW would expand our knowledge about the galactic NS population, as well as the extreme physics undergoing in these objects. Also, it would allow to further test the validity of General Relativity.

This thesis presents three new methods to post-process and follow-up the results of all-sky CW searches. Two searches for CWs from isolated and binary unknown NSs conducted on Advanced LIGO data are also presented. These methods have been routinely used in CW searches conducted by the LIGO–Virgo–KAGRA collaboration.

First, we introduce a new notion of distance among CW signals to select of interesting candidates resulting from the main stage of a CW search. This distance can be used to compresses into a smaller number of meaningful groups for an easier follow up. This new approach increases the sensitivity of a pilot search on O2 data for CWs from unknown NSs in binary systems with respect to using an ad-hoc Euclidean distance.

The second method evaluates the result of a multi-stage follow-up of an interesting CW candidate produced by a generic CW search. Concretely, it proposes a new Bayes factor to establish whether the behaviour of a candidate throughout several follow-up stages is consistent an astrophysical signal. The signal hypothesis’ distribution follows from first principles, and evaluates the consistency of signal amplitudes across different follow-up stages with different sensitivities; the noise hypothesis’ distribution uses extreme-value-theory results to estimate the expected loudest candidate produced by a background.The effectiveness of this approach is demonstrated by analyzing thirty outliers produced by several open-data searches using the second observing run of the Advanced LIGO detectors. None of these outliers was deemed consistent with an astrophysical source.

Finally, we propose a method to estimate the loudest candidate produced by the background in a gravitational-wave search. The method builds on extreme-value-theory results, which are based on the tail behaviour of probability distributions. Thus, it allows for the construction of meaningful detection thresholds even if the underlying distribution of the detection statistic at hand is unknown. This property is then exploited to re-evaluate the post-processing of a search on O2 open data to use a detection statistic more robust to instrumental artifacts. This statistic was not used in the original search as its distribution under the noise hypothesis is unknown; our method on the other hand, can be applied without any major trouble.

We then present two blind CW searches for unknown NSs using data from the third observing run of the LIGO–Virgo–KAGRA detector, focusing respectively on NSs in binary systems and isolated NSs. These searches made use of data-analysis strategies derived in this thesis to deliver the most sensitive results in the analyzed parameter spaces and an unprecedented precision in the recovery of parameters from artificially-generated signals. Despite the lack of CW signal detection, the methods developed in this thesis represent a step forward towards the effective analysis of broad parameter-space regions. The post-processing and follow-up strategies here presented will serve as a basis for searching wide parameter-space regions in the forthcoming runs of the advanced network of detectors. These approaches may also become relevant as future detectors start to probe lower frequencies of the gravitational-wave spectrum, as then a higher number of systems will produce gravitational-wave signals compatible with the CW model.