Understanding neutron star population and binary stellar evolution

evolution

The continued discovery of new pulsars through pulsar surveys will enrich our knowl- edge of the underlying neutron star population and clarify the mechanisms of neutron star formation. The latter is currently an open question due to the so-called birthrate problem (Popov et al.,2006;Keane & Kramer,2008) that arose from the discrepancy between the observed supernovae rate and the number of neutron stars predicted by population models, if the new observational manifestations of neutron stars such as RRATS (see Section 1.6.4) and X-Ray Dim Isolated Neutron Stars (XDINS) (see e.g. Haberl, 2004) are taken into account. To explain this discrepancy, other channels of formation, for example, by accretion-induced collapse of a white dwarf (Dessart et al., 2006;Freire & Tauris,2014)) have been hypothesized. Measuring the binary parame- ters of new discoveries and looking at the population as a whole, such theories can be tested.

1.9

Thesis outline

In this thesis the main focus is given to searching for pulsars with the 100-m Effelsberg radio telescope and the further follow-up of the discoveries made.

InChapter 2I describe the methods and instrumentation used in pulsar searches in general and give special attention to those relevant for projects with the Effelsberg radio telescope.

InChapter 3I report on the latest updates on the High Time-Resolution Universe Pulsar Survey – the North: the all-sky survey being conducted with the 100-m Effelsberg radio telescope. In this chapter I show the recent progress in observing and processing the data, analyze the survey sensitivity and present newly discovered pulsars.

InChapter 4I provide detailed timing solutions and speculate on possible evolu- tionary scenarios for “Two MSPs from the HTRU-North”, PSR J2045+3633 and PSR J2053+4650, two most exciting survey discoveries in the past five of years.

In Chapter 5 I discuss smaller projects conducted with the Effelsberg tele- scope, namely: timing of the unusually eccentric pulsar–He white dwarf system PSR J1946+3417 and searching for repeating radio signals from FRB 150418.

Chapter 2

Pulsar and transient searches

In this chapter I will briefly describe: 1) how radio signals undergo transformation from being weak oscillations of the electromagnetic field gathered by an antenna to becoming a ready-to-work-with set of quantized values recorded in a file; 2) what operations should be performed over these values to extract information about the nature of the source, i.e. to make a conclusion whether the received signal is a pulsar or terrestrial signal, or just noise.

2.1

Instrumentation and data acquisition

In this thesis all the data were obtained with a single-dish telescope, hence, in the following discussion we consider the scheme with this kind of antenna.

Fig.2.1shows a typical system of pulsar data acquisition. Conventionally it consists of two parts: thefrontend and thebackend. The frontend deals with the original radio frequency (RF) signal on its way from the antenna to the down-converting mixer. The electromagnetic waves collected and focused by the antenna’s parabolic reflector are conveyed to the receiver through the feed horn where they are transformed into alternating electric voltage of the same radio frequency. Two orthogonal polarisations are selected by the probes (horizontal and vertical for linear polarisation, or clockwise and counterclockwise for circular) placed at the output of the horn and further follow the same paths in separate channels.

The induced voltage is very tiny, typically just above the thermal noise floor, and must be amplified. Amplification, in general, strengthens both the wanted signal and the background noise, thus, reducing the detectability of the former. An effective approach to minimise the injection of the latter is usage of the so-called low-noise amplifiers (LNA), the ones at the start of the chain that are incased in a cryogenic dewar whose temperature is supported at the level of a few tens of kelvins preventing the growth of the noise temperature. Outside the dewar, the signal undergoes filtering and additional amplification (RF amplifier) within the selected frequency range. Usually the frontend parts are physically situated near the receiver as the transmission losses for radio frequencies are high. In order to transfer the signal after the frontend section to the place of processing and storage with minimal losses and without additional distortions introduced by cascaded amplifier chains, it is down-converted to a lower intermediate frequency (IF). Heterodyning, i.e. mixing with the signal of the local oscillator (LO) (or a chain of local oscillators), produces a number of frequenciesfIF− = fRF−fLO andfIF+ =fRF+fLO though only the differencefIF− calledlower sideband is further used. The down-converted signal passes through a set of IF amplifiers (more

Figure 2.1: A typical data acquisition system used in a pulsar survey.

amplification is always needed!) before leaving the surroundings of the primary focus cabin and then travels a few hundred meters further via coaxial cables to enter the backend.

The backend is all the electronics responsible for signal digitisation, processing and storage. Its key components include an analog-to-digital converter (ADC) producing time-sampled data followed by a spectrometer, a device channelizing this digital data stream into many narrow frequency bands, and a temporary storage machine. Different types of digital spectrometers may be exploited depending on a particular scientific purpose and hardware resources available. The ones that have become widespread during the last decade are the Fast Fourier Transform (FFT) spectrometers1. They are constructed on high-performance field programmable gate array (FPGA) boards config- ured to provide both wide bandwidth and high-frequency resolution. They compute FFT of time-sampled data chunks and transform it into power spectra. To prevent the power from being distributed between adjacent frequency bins (due to the finite length of the FFT), a pre-FFT windowing may be performed2. This is implemented in so- called polyphase filterbank spectrometers accustomed for use in pulsar astronomy. The power spectra are further accumulated over a number of clock cycles and integrated to

1_{Their operation principle is based on the property that an N-point Discrete Fourier Transform} (DFT) acts like a bank ofN finite impulse response (FIR) bandpass filters (see e.g.Price,2016).

2_{Weighting the FFT coefficients reduces sidelobes but widens the main lobe, i.e the resultant} channel bandwidth.

2.2. Working with data: pulsar searching 29

bring the channelized data back in the time domain. Both polarisations are summed since the information about polarisation is not necessary for the purpose of searching. As a last point of the signal path in the data acquisition system, the resultant channels of power time samples are recorded into a file in a raw-telescope format.