2.1. Field selection
The fields have been selected from blank sky fields with both ASCA and ROSAT observations. Severe screening criteria have been applied to the ROSAT data in order to minimize contamination from non-cosmic background. This limits the number of usable fields for our purpose. The fields used for this work, with sufficient amount of screened data, are the Lockman Hole (hereafter LH) and the Lynx-3A (hereafter LX) field. In order to use the same sky fields as much as possible for all instruments, we have selected the PSPC and GIS data from the sky area approximately corresponding to the ASCA SIS FOV (). The log of ROSAT and ASCA observations on these fields are shown in Table 1.
Table 1. Log of ASCA and ROSAT observations
2.2. ASCA observation
The screenings of the ASCA data have been made with the following criteria: avoid SAA; earth elevation angle 5 deg (from night earth) or 25 deg (from day earth); avoid 2 minutes after the satellite day-night transition; magnetic cutoff rigidity 8 GeV/c. The events from hot and flickering pixels of SISs have been removed. The data were taken when the GIS SP-discriminator was disabled, resulting in a high GIS background. Non-cosmic background (NXB) has been collected from night earth observations with the same magnetic cutoff rigidity screening, with 110,300 and 102,066 seconds of exposure for SIS and GIS respectively. The reproducibility of the GIS NXB is about 3% (Ishisaki 1997) and the NXB constitutes about 10% and 30% of the total count for SIS and GIS respectively.
We have created spectral response matrices using a ray-tracing assuming a uniform sky over a large field, taking into account the energy-dependent PSF and stray light problems of the ASCA instruments. For the GIS, Ishisaki (1997) estimates the systematic error of the absolute flux measurements using this procedure at .
2.3. ROSAT observation
In this work, we have used the time intervals with the sun-satellite-zenith angle greater than 120 deg (night time, avoiding contaminations by solar scattered X-rays), the geographic latitude between -30 and 30 degrees (tropic), and the Master-Veto (MV) rate between 40 and 170 (the particle background rate can be estimated using the Plucinsky et al. [ 1997] model). The model particle background (PBG) rates are about 5% of the total events in channels 100 PI 200 (1-2 keV) and 1-2% in 52 PI 99 (0.5-1 keV). Thus possible errors associated with the particle background modeling
does not affect our analysis significantly.
Light curves for the screened data have been further examined to exclude the time range with flux enhancements. In particular, the light curve of low energy channels ( keV or PI 51) for LH data show significant variation due to the Long Term Enhancements (LTE)(e.g. Snowden et al. 1994), while no significant light curve variation has been detected for PI 51. Thus we have used the data from the period where the light curve is close to minimum for the LH PI 51 spectrum. Since the light curve minimum for a particular pointed observation does not necessarily mean no LTE contamination, limiting our analysis. The EXSAS (Zimmermann et al. 1995) package has been used for data screening, particle background subtraction, and spectrum preparation.
© European Southern Observatory (ESO) 1998
Online publication: May 12, 1998