X-ray spectral timing of GX 339-4

Abstract

We study the accretion disc properties of the X-ray binary GX 339-4 using data provided by the XMM-Newton observatory, using a combination of spectral and timing X-ray analysis. A range of spectral models are tested, ranging from a simple power law model to the more advanced relativistic
reflection models relxill. We initially fit each model to the 3.0−10.0 keV spectrum that contains iron reflection features, and then extended our fitting to the spectrally more complex 0.7−10.0 keV band, although the 2.0 − 2.3 keV range is excluded due to instrumental artefacts. We find there is some model degeneracy with some aspects of the accretion disc geometry such as inner disc radius and disc inclination being difficult to determine, with the inner disc radius of the disc being relatively unconstrained for a range of relxill models. Timing analysis is also performed on the Power Spectral Density (PSD) generated from the light curves of the observations. The energy bands used to produce the PSDs are a soft band, 0.3−0.7 keV, and a harder band, 0.7−1.5 keV. These PSD models provide a well constrained disc truncation radius. We then simultaneously fit the spectral and timing data across all observations. An assumption is made that the inner disc radius in the spectral modelling and disc truncation in the PSD modelling are equivalent, so they are tied together for each observation, then fitted simultaneously to constrain the inner disc/truncation radius using both spectral and timing data. It is found that the combination of both methods of analysis provide a well constrained disc
truncation radius that shows a steady outward expansion of the inner disc radius as the system emerges from an X-ray outburst and the flux decreases. The final model of this study shows that a combination of both spectral and
timing analysis is necessary to properly describe the accretion disc properties given the quality of the data at hand. It is concluded that a higher quality of observational data is required to improve modelling of the spectral and timing properties of GX 339-4. Particularly, better iron line spectroscopy to better fit the inner disc truncation and a higher signal-to-noise power spectra at higher frequencies. Furthermore, improved models that self-consistently predict the energy spectra and PSD would better constrain the free parameters of the models in this study.

Date of Award	9 May 2023
Original language	English
Awarding Institution	University of Bristol
Supervisor	Andrew J Young (Supervisor) & Mark Birkinshaw (Supervisor)