3. Pulse phase resolved spectroscopy

To investigate the dependence of the spectral parameters on pulse phase the A0 data were binned into five phase bins and the data from each phase bin were fit to an absorbed power law with one Gaussian emission line at $\sim$ 6.7 keV. No significant variation in the intensity of the iron line was seen. However, due to the poor counting statistics, the data do not rule out the presence of variability. The 90% confidence upper limit on the variation in line strength is 70%, which is consistent with the 50% variation in iron line intensity detected by Day et al. (1993) in Ginga data. There is some evidence for variation of the column density and photon index with pulse phase. The best-fit values of these parameters are plotted against pulse phase in Figure 2.3. The column density appears to vary in phase with the flux and the photon index appears to vary roughly 180 degrees out of phase with it. This apparent flattening of the continuum at the pulse peak has been observed before (White et al. (1983)). These variations may be due to changes in the emission geometry which occur as the neutron star rotates, or, more likely, they may indicate that these models are too simple and that the continuum is more complex than a single absorbed power law.

Figure 7: Variation of spectral parameters with pulse phase. Error bars represent 90% confidence intervals.