3. Comparison of results with previous observations of the iron line

Although the evidence for a 6.4 keV line is weak in this observation, the Ginga observations of 1989 are consistent with lines at both 6.4 and 6.7 keV (Nagase et al. (1992)). However, the resolution of the instrument was not sufficient to determine their widths. As mentioned in Section 5, these lines were observed with Ginga as the pulsar went into eclipse. The 6.7 keV line underwent a gradual, partial eclipse, while that of the 6.4 keV line was abrupt. This indicated that the sizes of the emission sites of the 6.4 and 6.7 keV lines were $\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}3\times10^{10}\rm\ cm$
and $\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}8\times 10^{11}\rm\ cm$
respectively. Also, the assumed stellar wind density ( $\sim$ 10²² cm^-2) was too small to produce the 180 eV equivalent width observed with Ginga for the 6.4 keV line. Thus, it was concluded that the 6.4 keV line was due to fluorescence of relatively un-ionized matter near the accretion disk and the 6.7 keV line was due to fluorescence of the extended ionized stellar wind.

The Alfvén radius may be estimated from Section 3.2 as r_A $\approx$ 4 x 10⁸ cm which is close to the value obtained by Nagase et al. for the size of the 6.4 keV emission region. Line emission would not be expected from inside the magnetosphere because of the extreme ionization conditions there. This suggests that fluorescence of material in the inner accretion disk is responsible for the 6.4 keV emission.

At the same orbital phase as the BBXRT observation, the line luminosities derived from the Ginga observation were 5 x 10³⁵ erg s^-1 for the the 6.4 keV line and $\sim$ 1 x 10³⁵ erg s^-1 for the 6.7 keV line, assuming a distance of 8 kpc. The unabsorbed line luminosities found with BBXRT were 1.3 x 10³⁵ and 5 x 10³⁵ erg s^-1 respectively, assuming the same distance. In order to check this, the A0 data were rebinned so that their resolution matched that of the Ginga LAC (Makino and the ASTRO-C Team (1987)). The iron emission was fit with a single Gaussian. The best fit line energy was 6.63$\pm$0.09 keV. The value obtained with Ginga for the line energy in the high post-egress orbital phase was 6.50$\pm$0.08 keV, reflecting the different relative intensities of the 6.4 and 6.7 keV components. The lower equivalent width obtained for the 6.4 keV line during the BBXRT observation could be produced by reprocessing in a wind of column density $\sim$ 10²² cm^-2 (Makishima (1986)). It is possible that luminosity dependent changes in the angular distribution of emission from the neutron star could alter the relative strengths of the 6.4 and 6.7 keV lines, especially if they are produced at different reprocessing sites.