A Study of Persistent and Bursting Hard X-ray Emission from Magnetars with the INTEGRAL Satellite

Magnetars are a small class of neutron stars with the highest known magnetic fields, reaching up to 10^15 G. Their emission is mainly powered by the decay and evolution of their magnetic field. This thesis presents a systematic study of hard X-ray emission from magnetars using the large set of archival data from IBIS/ISGRI on board INTEGRAL, the ESA X-/gamma-ray (3 keV to 10 MeV) satellite that operated from 2002 to 2025. The three main types of hard X-ray/soft gamma-ray phenomena from these sources were considered in this work: persistent emission, short bursts, and giant flares. For what concerns the study of long-term persistent hard X-ray emission, I report the detection of eight magnetars. For five of them (4U 0142+614, 1RXS J170849-400910, 1E 1841-045, SGR 1806-20, and PSR J1846-0258), which were detected multiple times over the mission duration, I extracted averaged spectra and long-term light curves. The first three sources showed constant fluxes over 20 years, while the latter two exhibited significant variability. Notably, SGR 1806-20 showed non-monotonic flux evolution after the 2004 outburst, accompanied by spectral hardening as the flux declined. Additionally, I present a broadband study of the 2021 outburst of Swift J1555.2-5402 using Swift, NICER, NuSTAR, and INTEGRAL. The soft X-ray blackbody temperature remained steady at ~1.2 keV while the emitting region radius decreased from ~2.1 km to ~0.3 km, suggesting sustained heating of a contracting hotspot. I carried out a systematic study of short bursts from all known magnetars and candidates. Analysing over 1,100 Ms of data, I detected 1,491 bursts from 21 sources: 934 from SGR 1806-20, 232 from SGR 1935+2154, 146 from 1E 1547.0-5408, and 37 from 18 other sources. Temporal analysis reveals that burst durations follow a log-normal distribution, centred at ~0.1 s, with source-dependent variations. For bursts with sufficient counts, the spectra are well described by an exponentially cutoff power law with peak energies of ~20-60 keV for SGR 1806-20 and SGR 1935+2154, and ~35-100 keV for bursts from 1E 1547.0-5408. I found a significant anticorrelation between Epeak and fluence for SGR 1806-20, which provided the largest number of bursts among the sources in the sample. As this is the only systematic search for magnetar bursts using INTEGRAL data, the sample includes many bursts that have never been reported before, including single bursts from sources that have not exhibited intensive bursting activity. The initial spikes of magnetar giant flares can also be detected in galaxies outside the Local Group. I report the INTEGRAL results on the two most recently discovered events of this type: GRB 231115A, associated with the nearby starburst galaxy M82, and GRB 241107A, likely originating from the galaxy PGC 86046. To constrain the rate of magnetar giant flares, I performed a systematic search using INTEGRAL's extensive observations of the Virgo cluster and other nearby star-forming galaxies. The search covered approximately 35 Ms of Virgo cluster observations and 103 Ms on seven nearby galaxies with high star-formation rates. Beyond M82, no additional giant flares were detected. From these non-detections, I derived rate constraints on the occurrence of MGFs: an upper limit of approximately one giant flare with E > 3×10^45 erg every ~500 yr per magnetar from the Virgo search, and a lower limit of R(>E) > 4×10^-4 yr^-1 magnetar^-1 for energies E < 10^45 erg from the nearby galaxy search that included the M82 detection. These are the first rate constraints derived with a hard X-ray instrument with a large field of view and arcmin location accuracy.