Astron. Astrophys. 318, 111-133 (1997) 8. Extragalactic contaminationAlthough very low galactic fields such as the one investigated here are unlikely to contain a large fraction of extragalactic sources, our discovered AGN at b = shows that this population is not completely screened out and may account for a significant fraction of the optically unidentified and relatively faint X-ray sources. In order to estimate this 'pollution' we assumed that the un-absorbed extragalactic log N( S)-log S function was essentially that given in Hasinger et al. (1993). At our maximum possible flux sensitivity of 1-2 10^{-13} erg cm^{-2} s^{-1} we are far above the departure from the near Euclidean distribution and we may write the number of sources having X-ray flux within s and as: However, our only measured quantity is the source count rate and for a given intrinsic spectrum the flux over count ratio may depend heavily on the intervening photoelectric absorption. In the Lockmann Hole (Hasinger et al. 1993) the soft X-ray flux is s = where c is the count rate measured from the source and the differential extragalactic count distribution is: The observed count rate from sources dimmed by galactic absorption may be expressed as where is the count absorption coefficient at position . Averaging over the whole area and using the explicit form of the differential count distribution we may write: Therefore, the absorbed extragalactic distribution has the same slope as the un-absorbed one but shifted to lower count rates. We assumed a log N( S)-log S (0.5-2 keV) relation with N = 104, = 2.44 and an average spectrum of extragalactic sources in the form of a power law of energy index 0.96 (Hasinger et al. 1993, Gioia et al. 1990). We computed the count to flux ratios and count absorption coefficients for different values of the galactic absorption by folding the incident absorbed spectra with the detector response curve and by fitting a three degree polynomial to the log(K)/log() relation. Because of the high photoelectric absorption we neglected the count rate resulting from the 0.1-0.5 keV low energy part of the spectrum. The galactic absorption was estimated from the HI maps of Strong et al. (1988) and Dame et al. (1987) interpolated on a grid. The total was taken as + 2 3.5 10^{20} (e.g. Solomon & Rivolo 1987). Finally, the K coefficient was then integrated over the observed area. Our simulation shows that on the average Cygnus field the count rate of extragalactic sources is only 15% of that detected at high galactic latitude and that at a given flux level we only detect 3.5% of the total extragalactic population. Table 7 lists the expected number of extragalactic sources as function of the absorbed count rates. Table 7. Absorbed extragalactic log N( S)-log S function We have only one positive optical identification with a Seyfert 1 nucleus at a count rate of 0.1 cts s^{-1} (RX J2135.9+4728, index 12). Another source (RX J2133.3+4762, index 35) with a count rate of 0.035 cts s^{-1} is likely extragalactic. It has a very hard spectrum similar to that of the identified AGN and deep optical investigations fail to reveal the counterpart. We show in Fig. 17 the position of our two extragalactic candidates on a map together with the investigated area. It is probably meaningful that our two non-galactic candidates fall into a region of very low absorption.
Considering the small number statistics, we believe that the number and count rates of our identified extragalactic candidates (2 sources with 0.03 cts s^{-1}) are in accord with our simulation. We conclude that the extragalactic population is likely to represent only a very small fraction ( 3%) of the 68 sources detected in the 'full' area above our completeness level of 0.02 cts s^{-1}. © European Southern Observatory (ESO) 1997 Online publication: July 8, 1998 |