The variable cyclotron line of GX 301–2

Abstract

Abstract. We present pulse phase resolved spectra of the hypergiant high mass X-ray binary GX 301−2. We observed the source in 2001 October with RXTE continuously for a total on-source time of almost 200 ksec. We model the continuum with both, a heavily absorbed partial covering model and a reflection model. In either case we find that the well known cyclotron resonant scattering feature (CRSF) at ∼35 keV is -although present at all pulse phases -strongly variable over the pulse: the line position varies by 25% from 30 keV in the fall of the secondary pulse to 38 keV in the fall of the main pulse where it is deepest. The line variability implies that we are seeing regions of magnetic field strength varying between 3.4 × 10 12 G and 4.2 × 10 12 G.

GX 301-2 with INTEGRAL

“…The resulting spectral parameters are mostly consistent with those found in the literature. Any discrepancies between the published values and our results can be attributed to the known variability of the source and that the analyses of [2] and [3] were based on data from different instruments: RXTE and BeppoSAX, respectively. It is therefore not surprising that the fit results are somewhat different.…”

“…The corresponding fit results for the Fermi-Dirac cutoff model are shown in Table 3.2. Most resulting parameters are consistent with the values reported by [2].…”

“…Orbital period may be calculated by comparing our T PA value with that reported by [8] and assuming constant period. It turns out to be significantly less than reported earlier P orb = 41.482 (2), which makes the assumption of a constant period questionable. Our estimate for secular change of orbital period using T PA values by [14,8] …”

“…We discuss the spectral models suitable to describe the spectrum and the impact of the model on the dependence of the cyclotron resonance scattering feature (CRSF) energy with pulse phase. The history of the pulse period since the launch of INTEGRAL is determined and the spin down trend, established by RXTE [1,2], and BeppoSAX [3] observations is confirmed. Additionally we provide an updated orbital period value based on time arrival analysis of archival data which is found to be consistent with the period obtained from the RXTE ASM lightcurve of the source.…”

“…It is clear that it is almost impossible to separate contributions from individual absorption columns such that only a small number of photoelectric absorption components is required. For example [2] and [3] adopted two columns to describe their spectra. We therefore also use only two column densities effectively, but fix one to zero (pcfabs model in XSPEC):…”