Astron. Astrophys. 344, 333-341 (1999)

4. The Lockman hole

ROSAT PSPC datasets for the Lockman hole region (Lockman et al. 1986) were accessed from the archive with the criterion that they lie within or at the periphery of this lowest column density region. The common feature of these observations is that they are some of the deepest observations performed with ROSAT, thus providing an opportunity to improve upon the quality of the spectral fits than hitherto achieved.

These deep observations, however, initially gave quite poor fits to the data using the previously defined recipes. To investigate this effect further the time series for the total event count rate (TEVS) were examined for the Lockman hole observation centred at =149^o.5, b =53^o.2 (ROSAT observation sequence number rp900029), which has a total exposure time of 122,123 seconds. The entire observation consists essentially of three viewing periods. Upon applying an additional constraint of a maximum total event rate of 14 cts s^-1 the fluctuations in the time series become essentially insignificant, giving us confidence that most of the contaminating events have been removed from the data. The resultant exposure time for the observation was however reduced to just 9,278 seconds, typical of the exposures for the other PSPC fields listed in Table 1. The output spectra resulting from the good time windows used in generating these time series are plotted in Fig. 10. It is evident from the essential differences in the two spectra that the bump-like feature present at 550 eV, before application of the maximum total event rate cutoff, is a contaminant most likely to be associated with particle background enhancements. As the observation under consideration here commenced in April 1991 and completed in April 1992, the enhancements could therefore be connected with the solar maximum.

Fig. 10. Spectral plot for the deep observation rp900029 in the Lockman field. The triangles represent the spectrum generated with just the master veto threshold of 170 cts s-1, while the diamonds represent the spectrum with the additional constraint of the photon event rate being below 14 cts s-1. The best three component spectral fit for this region is shown by the solid line.

The setting of a rather strict threshold on TEVS has improved the spectral fit for the deep observation (see last entry in Table 1) and although the fit is still far from acceptable, the degree of discrepancy is no worse than for some of the other poorly fitted regions in Table 1. Having already explored the halo and foreground components in some detail, it is now reasonable to concentrate on the extragalactic component; it is conceivable that these discrepancies are due to fluctuations in the power law component which are neglected by the simple extrapolation from 5 keV down to 2 keV. To test this tentative hypothesis two more free parameters are introduced into the model, namely the spectral slope and the spectral coefficient. A five parameter fit is thus performed for the deep observation rp900029 and the ranges of the parameters are tabulated in Table 2. The results are summarised in Table 3.

Table 2. Parameter search space for the 5 parameter fit to the rp900029 deep observation in the Lockman field The range of variation for each variable parameter is shown with the number of increments over the indicated range being enclosed in brackets. Fixed parameters are printed in boldface

Table 3. Best fit results () for rp900029 deep observation in the Lockman field

However, an equally acceptable  fit can be obtained by a wider variation in the spectral slope and coefficient than is represented by the differences between the fixed power law and that shown in Table 3. Since both the average halo and extragalactic components are significant in the 1-2 keV energy range, there is no way of separating the two effects in our measurements, especially in view of our limited knowledge about the origin and spectral properties of the extragalactic component. The best estimate one can make is to extrapolate the higher energy power law - our standard model - to the 1-2 keV range, but all our statements about the average halo component are subject to the uncertainty arising from this assumption.

Recent work by Miyaji et al. (1998) supports our results, although they have performed simultaneous spectral fits to this region using both ROSAT PSPC and ASCA data. Their model incorporates a broken power-law for the extragalactic component which steepens at E   1 keV.

© European Southern Observatory (ESO) 1999

Online publication: March 10, 1999
helpdesk.link@springer.de