2.1. The Distribution of NH
Past hard X-ray surveys of Sy2s were strongly biased in favor of X-ray bright sources (according to all-sky surveys), which tend to be the least absorbed ones. More recently, X-ray observations have probed X-ray weaker Sy2s, partly removing the selection against heavily obscured objects (eg. Turner et al. 1997). We (Salvati et al. 1997, Maiolino et al. 1998) used BeppoSAX to observe an [OIII]-selected sample of previously unobserved (X-ray weak) Sy2s. These observations discovered a large fraction of Compton thick objects, a result which confirms the bias against heavily obscured systems affecting previous surveys. However, none of these surveys is actually complete or free from biases and, therefore, none of them is suitable (if taken separately) to derive the distribution of absorbing column densities among Sy2s. To accomplish this goal, we merged all the available hard X-ray observations of Sy2s and extracted a complete subsample limited in intrinsic (i.e. unabsorbed) luminosity as inferred from the [OIII] narrow emission line (Risaliti et al. 1999). This subsample is composed of 45 objects and the corresponding NH distribution, shown in Fig. 1, can be considered the best estimate of the true distribution that can be obtained with the available data. The most interesting result is that this distribution is significantly shifted toward large columns with respect to past estimates: most (~ 75%) of the Sy2s are heavily obscured ( NH > 1023 cm-2) and about half are Compton thick. The NH does not appear to correlate with the (intrinsic) luminosity, at variance with early results. This NH distribution has various implications some of which will be discussed in the following.
Figure 1. The distribution of absorbing column densities among Seyfert galaxies. (from Risaliti et al. 1999, with modifications).