Abstract:
The work presented in this thesis addresses both observational and instrumentation
aspects to the study of low-mass X-ray binaries.
The analysis of the low-mass X-ray binary system 4U 1735−44, observed by the
X-ray satellites BeppoSAX and XMM-Newton, is discussed. For the first time, the
broad-band spectrum of the source was investigated which could be modeled by
an absorbed combination of a disk blackbody and Comptonization components.
The primary focus of the study is set on the fluorescence iron Kα line spectral
feature. This line is an important issue since, in the community, it is still debated
whether such an iron line can be affected by the relativistic effects which occurs
close to a compact object. These effects would result in a broad and asymmetric
shape of the line. However, there are also discussions that the shape of the line
can be distorted by the occurrence of pile-up and only seems to be relativistic. In
this work it is shown that for 4U 1735−44, the line can be modeled either with a
broad Gaussian line or with a relativistic line for which the production site is not
close to the neutron star. However, even if pile-up effects are carefully taken into
account, no distortion of the line due to pile-up was found. The results obtained
by the fit of the broad-band spectrum and the detailed study of the iron line led
to the suggestion of a possible geometry of the source, which is a combination of
the competitive Western and Eastern models.
The second part of the thesis is explicitly focussed on the systematic effects
caused by pile-up and their impact on the iron Kα line. Especially, the Timing
Mode of the EPIC-PN camera aboard XMM-Newton was investigated, since the
results obtained by this instrument were doubted by Ng et al. (2010). Therefore, a
variety of simulations were performed using the SIXTE simulation package. It was
found, as expected, that pile-up hardens the spectrum of a low-mass X-ray binary.
However, no distortion of the iron line due to pile-up was found. Moreover, it is
shown in this thesis that relativistic lines reported in literature are robust against
pile-up and that their relativistic origin can be proven.
Finally, the problem of background extraction for the Timing Mode of the
EPIC-PN is discussed. In this readout mode no region on the CCD can be defined
from which an uncontaminated background spectrum can be extracted. It is shown
that the use of blank field observations is a meaningful option to provide a template from which a suitable background spectrum can be extracted from for a given
observation. This method will be a profitable task for further studies of especially
weak and highly obscured sources observed in Timing Mode.