Astron. Astrophys. 357, 501-506 (2000)

3. Temporal analysis

The arrival times of the photons were first converted to the solar system barycenter. In Fig. 1 we show the background subtracted X-ray light curves, obtained in three energy ranges, with 300 s time bin size. The maximum intensity variation is 30% in the 1.6-10 keV and 20-60 keV energy ranges, and 60% in 10-20 keV range.

Fig. 1. GS 1843+00 background subtracted light curves in three energy ranges. The gaps are due to South Atlantic Anomaly passages and Earth occultations

The (1.6-10 keV) GS 1843+00 power spectrum, is shown in Fig. 2. An outstanding peak at 0.0339 Hz is clearly observed. The GS 1843+00 pulse period was obtained with an epoch-folding technique using barycentric corrected 1.6-10 keV MECS data, while the (1) uncertainty was determined by fitting the arrival times of sets of 9 averaged profiles, each of 16 phase bins. The best-fit period is . There is no evidence for any change in spin-period during the observation with a upper-limit of  s s^-1. Using this period value, we folded the light curves in different energy bands. The pulse profiles in five energy ranges are shown in Fig. 3. At lower energies the pulse profile is clearly asymmetric with a double peak shape, whilst at higher energies it becomes a simple sinusoid.

Fig. 2. GS 1843+00 power spectrum in the 1.6-10 keV energy range. The peak due to the fundamental frequency at 0.0339 Hz and the aperiodic variability are clearly visible

Fig. 3. GS 1843+00 pulse profiles in five energy ranges. 1 uncertainties are shown

The variation with energy of the pulsed fraction, defined as the semi-amplitude of the modulation divided by the average intensity, is shown in Fig. 4. There is no evidence for an increase in the fractional periodic variation with energy.

Fig. 4. The GS 1843+00 Pulsed fraction versus energy

In the BeppoSAXobservation, GS 1843+00 may have opposite behavior than during Ginga observation, where the pulsed fraction was found to increase with energy and the pulse profile was clearly single peaked a lower energies and more structured at higher energies (Koyama et al. 1990b).

From BATSE  ¹, RXTE (Takeshima 1997) and BeppoSAXdata a clear spin-up trend ( s s^-1) over 30 days is evident (Fig. 5). The mean spin-up timescale, , is a very rapid 24.6 years. However a difference of of 0.01 s is observed between the BeppoSAX period and the one expected from the BATSE data extrapolation. This could be due to a Doppler effect of orbital motion. Actually, the change of the pulse period, , due to orbital motion is constrained to be

where is the orbital period, e the eccentricity, i the inclination angle, the mass of the neutron star, the mass of the companion star, G is the gravitational constant and c the speed of light. Following Corbet (1986) relation to estimate the orbital period (50d) and assuming a mass of 15 for the companion star, typical of Be star, the upper limit of turns out to be for circular motion.

Fig. 5. The period history of the GS 1843+00 outburst between early 1997 March to early April obtained from BATSE , BeppoSAXand RXTE . All the data points are consistent with a smooth rapid spin-up trend

To study the aperiodic variability, firstly reported by Koyama et al. (1990b) the (1.6-10 keV) and (10-37 keV) power spectra were fitted with a power law. The Poisson white noise was subtracted from Leahy normalized power spectra (Leahy et al. 1983). The power indices, 1.440.17 and 0.90.25 respectively, are consistent with those found in the Ginga Observation (Koyama et al. 1990b). The relative amplitude of the aperiodic variation, calculated dividing the root square of the integrated PDS over Hz to 10 Hz by an average intensity, is larger in the lower energy band (18) than in the higher energy band (2).

© European Southern Observatory (ESO) 2000

Online publication: June 5, 2000
helpdesk.link@springer.de