A Search for Relativistic Binary Pulsars in the Galactic Plane

OF THESIS submitted by Ralph Paul Eatough for the Degree of Doctor of Philosophy and entitled A Search for Relativistic Binary Pulsars in the Galactic Plane. September 2009 Relativistic binary pulsar systems, such as double neutron star and neutron star white dwarf binaries, currently provide the best means by which to test Einstein’s theory of general relativity and alternative theories of gravity in strong gravitational fields. A radio pulsar with a black hole companion would be the model system for such tests because of the immense orbital velocities and gravitational potentials present. Unfortunately finding such systems in pulsar surveys is not easy. Doppler smearing of the pulse frequency due to orbital motion throughout the survey observation renders standard pulsar search techniques ineffective. Binary search algorithms that compensate for these effects are computationally expensive in blind pulsar searches. As such, the analyses of most pulsar surveys are performed with standard ‘un-accelerated’ searches that select against these systems. In this work I have attempted to combat these selection effects in the Parkes multi-beam pulsar survey by reprocessing the entire survey with an efficient ‘coherent acceleration search’. The Parkes multi-beam pulsar survey is the largest and most successful pulsar survey ever undertaken, having discovered 770 pulsars up to the start of this work. It covers a large region along the Galactic plane from l = 260 to l = 50 and |b| < 5. The survey was carried out between 1997 and 2003 using a 13-beam receiver with bandwidth 288 MHz centred on 1374 MHz at the 64-m Parkes radio telescope in Australia. Previous ‘incoherent acceleration searches’ of the survey have had some success at finding relativistic binary systems, however the elusive pulsar black hole binary system has so far evaded detection. Despite efforts to make the coherent acceleration search efficient this reprocessing has required an order of magnitude more computational operations than 11 in previous analyses. As a result reprocessing was performed on a large-scale computing facility via a computational grid designed for the analysis of particle physics data. At the time of writing this is the first time an entire pulsar survey has been processed, over the Internet, using a computational grid. Throughout this work a number of improvements have been made to the search software used for reprocessing. In addition, two innovations in pulsar searching have been developed and are presented. The first is a new technique for removing the harmful effects of terrestrial radio frequency interference on pulsar search observations. This work has been published in Monthly Notices of the Royal Astronomical Society (Eatough et al., 2009). The second is a new technique for selecting credible pulsar candidates from large processing results databases using artificial neural networks. This work is also currently being prepared for publication. Reprocessing has resulted in the discovery of 15 pulsars, including a binary millisecond pulsar in 6 day orbit around a low mass companion, a pulsar with an interpulse, and the possible discovery of a new intermittent pulsar. A number of these pulsars were identified as a direct result of the new techniques described above. All sources now have coherent timing solutions and are also currently being prepared for publication in a paper. To date no more relativistic binary pulsars have been found. The possible reasons for the non-detection and future work that may help in finding these systems is also discussed.