6. The 1/4 keV X-ray background
The flattened-halo model of the keV RASS energy band is based on physical considerations which constrain important properties of the keV XRB radiation. In particular, the derived temperature of the galactic halo thermal plasma (kT = 0.13-0.14 keV) is a sensitive function of the intensity ratio between the keV and keV XRB radiation. Therefore, we demonstrate the validity of our 3/4 keV galactic X-ray halo model by the extrapolation to the keV XRB.
One has to take into account that roughly half of the observed keV radiation is caused by local (within, say, 100 pc) X-ray emission (McCammon & Sanders 1990). Within the low-volume-density environment of the Sun, the local interstellar medium is distributed inhomogeneously (Frisch 1995). Welsh et al. (1991) found that beyond a distance of 50 pc the interstellar medium cannot be regarded as a local cavity of regular shape. In consequence, the foreground X-ray emission term in the X-ray radiation-transport equation cannot be regarded as constant across the entire sky; thus a more a careful radiation-transport analysis for the keV band is called for. Such an analysis was recently performed towards selected fields by Herbstmeier et al. (1995), Kerp et al. (1996, 1997). We used their values of the galactic-halo intensity in the keV energy range and compare these values with our modelled keV count rates. As mentioned above, this test constrains the temperature interval of the galactic halo plasma. In the case that the keV and the distant keV diffuse galactic X-ray emission can be attributed to the same plasma we have to find a linear correlation between both intensity distributions.
First, we evaluate the extragalactic XRB intensity in the keV band. Using the spectral fitting results discussed in Sect. 3, the extragalactic XRB contribution to the keV count rate can be estimated as . This value is in quantitative agreement with the extragalactic keV XRB intensity values determined by the analysis of shadowing galaxies carried out by Barber et al. (1996). Second, we subtract this extragalactic XRB count rate from the keV XRB count rate determined by Herbstmeier et al. (1995) and by Kerp et al. (1996, 1997). Third, we plot these modified keV intensities against the keV count rates derived from the flattened-halo model. Fig. 10 shows that, indeed, the keV count rate of the galactic halo is linearly correlated with the distant keV XRB count rate. This result indicates that the keV and the distant keV XRB have the same source distribution. Indicated by the solid, dashed, and dotted lines in Fig. 10 are different plasma temperatures ranging only from kT = 0.13 to 0.14 keV. This narrow temperature range is in agreement with the PSPC spectral-fitting results of Sect. 3.
These findings suggest that a significant fraction of the keV and keV XRB radiation is caused by thermal-plasma emission originating in the galactic halo.
6.1. Modeling the 1/4 keV sky
To validate the above suggestion further, we scaled the keV X-ray halo intensity distribution to the keV energy range and cross-correlated the modelled XRB intensity distribution with the observed one. The angular resolution of the RASS data is not sufficient to allow a detailed 1/4 keV X-ray transport analysis: the angular resolution results in the small-scale structure of the X-ray attenuating interstellar medium being smeared out.
To determine the foreground intensity distribution we subtracted a modelled keV X-ray halo XRB map from the observed RASS keV map. For each latitude range we evaluated a constant foreground intensity (cf. Table 2), representing the minimum offset between the modelled halo and the observed keV distribution. Using the flattened X-ray halo model and a plasma temperature of kT = 0.135 keV and with the constant foreground count rate tabulated in Table 2, we evaluated a second keV XRB intensity distribution (see panels (f) of Figs. 8 and 9) which has to be compared with the observed XRB intensity distribution shown in panels (g) of Figs. 8 and 9. Although the observed keV count rate is dominated by the foreground contribution, we expect that, especially on smaller angular scales, photoelectric absorption of the XRB components produces the observed keV count rate variations.
Utilizing the H I column density maps shown in panels (b) of Figs. 8 and 9, it is possible to identify several areas where the modelled XRB map reproduces the observed keV distribution particularly well. These regions include that of the the inner side of the north celestial pole loop near ), as an X-ray bright spot; the high-density region near , as absorption; the low-density region near , as a bright band; and the complex absorption pattern near . Even the Lockman hole region near is well reproduced.
Although we have neglected a longitude variation of the foreground component, the difference maps (observed minus modelled XRB) shown in panels (h) of Figs. 8 and 9 reveal only two areas other than the excess X-ray emission associated with Loop I and the "Monogem Ring" (), which have in the northern hemisphere; other northern regions are characterized by deviations between observation and model between . These two regions with a deficit of predicted keV emission coincide with the positions of two HVC complexes, namely complex C, near (see Kerp et al. 1996), and complex M, near (see Herbstmeier et al. 1995). The brightest spot of complex M is centered near the HVC cloud M III, which is also detectable in the keV maps shown by Herbstmeier et al. (1995). Since these enhancements towards the two HVC complexes is also visible in the residual map of the X-ray radiation-transport analysis by Marshall & Clark (1984) instrumental effects can be excluded for the ROSAT data. In contrast to the keV situation, the keV residual map reveals no significant excess emission close to the two HVC complexes C and M.
The keV residual emission in the southern hemisphere can be divided into two parts. At , the predicted count rate is too low (); at , the predictions are too high (). Since this behaviour is not positionally correlated with any distinct H I structure and a similar residual feature is not visible in keV data, this behaviour is not produced by the X-ray transport solution. It may be caused by a variation within the foreground X-ray emission or by a dependence of absorption on galactic longitude, e.g. by a large-scale inhomogeneity in the low-volume-density surroundings of the Sun. But since the distribution of this deviation in the southern hemisphere at is parallel to the scanning direction of the RASS data, a residual effect of the RASS data reduction cannot be excluded.
Despite the problem of the unknown distribution of foreground emission, the halo model appears to be applicable to the keV band as well as to the keV bands, down to the accuracy limit of the RASS X-ray data.
In our hydrostatic model the X-ray halo emission is symmetric with respect to the galactic plane (the Sun's z-offset from the plane is negligible compared to ). In the keV band the observed X-ray count rates toward both galactic poles differ significantly (see Fig. 9g). Snowden (1997) concludes that this is due to a higher intensity of the unabsorbed X-ray halo emission in the northern galactic hemisphere. Our analysis suggests that the observed north/south asymmetry is caused by the photoelectric absorption of the foreground gas traced by the H I emission in addition to an asymmetry in the 1/4 keV foreground intensity.
© European Southern Observatory (ESO) 1998
Online publication: March 10, 1998