The Seen & the Unseen: from X-ray Binary Catalogues to Probabilistic Noise-Processing Techniques

X-ray astronomy provides a unique window into the most energetic and extreme processes in the universe. In particular, compact objects such as neutron stars, or black holes accreting matter from a companion star serving as a natural laboratory for stellar evolution, accretion processes, behavior of matter in strong magnetic fields, and much more. An accurate background modeling is essential for extracting astrophysical signals, especially faint signals in X-ray observations. This thesis consists of two separate parts, the first part is about catalogues of X-ray binaries, the second part presents a novel method to deal with the unwanted background. The first part of this thesis deals with the X-ray binaries which are known to be a possible endpoint of stellar evolution. These systems consist of compact objects, e.g. neutron stars or black holes, and a non-degenerated companion. Depending on the mass of the companion, the system is classified into two different class. If the mass of the companion is below one solar mass the systems are classified as Low Mass X-ray Binaries (LMXBs), while above eight solar masses they are classified as High Mass X-ray Binaries (HMXBs). In such highly energetic systems, the compact object accretes from the companion, emitting large amount of energies in the X-rays. A variety of such systems are known within and outside our galaxy. It is therefore crucial to create a catalogue containing such a system, specially for a population study. Numerous known catalogues have been created in the past, and most recently between 2003 and 2007 for galactic HMXBs and LMXBs (Liu et al., 2006, 2007; Ritter and Kolb, 2003). Since then, new systems have been discovered and parameters of known systems have changed. The goal of the first project was to publish new catalogues of HMXBs and LMXBs, in cooperation with Dr. Avakyan. This project resulted in two publications (Neumann et al. (2023), and Avakyan et al. (2023)) and a website containing both catalogues, providing users an opportunity to explore them interactively. The second project deals with new, and alternative method to reduce background in the XMM-Newton observations. Already established methods have their own benefits but also downsides. The simple subtraction of a constant background value might not be computationally demanding; however, in the regime of low count rates, this technique can result in an unphysical negative count. An alternative is to model the source and background, which can yield a more significant result than simple subtraction; however, it is more computationally demanding and requires a prior knowledge of the source which is not always possible, especially for a new object. To deal with such shortcomings of the existing methods, the technique explained in this part uses a probabilistic approach, which prevents a negative source count and does not require prior knowledge of the object. Since the probabilistic technique needs an estimated background, part of the project was to develop images in different energy ranges which can be used to estimate the background in later observations. Then the technique was first tested on the artificial data and then applied to real observations to evaluate its performance.