3 How many unabsorbed Seyfert 2 galaxies are there?

Bassani et al. (1999) presented a column density distribution for a sample of Seyfert 2s for which X-ray data were available. The distribution was shifted toward high $N_{\rm H}$ values and only $\sim$20% of the sources had a column density less than 10²² cm^-2. However, the sample used by Bassani et al. (1999) was a collection of data and so it did not fulfill the requirements of completeness. When a complete sample was considered (Risaliti et al. 1999) the column density distribution found was similar for high column density values: 75% of type 2 Seyferts were found to be heavily obscured ( $N_{\rm H}> 10^{23}$ cm^-2) with 50% of the sources being Compton thick ( $N_{\rm H}> 10^{24}$ cm^-2). However, only $\sim$ 4% of the sample was characterized by low absorption ( $N_{\rm H}<10^{22}$ cm^-2). The sample considered by Risaliti et al. (1999) was derived by the Maiolino & Rieke (1995) sample completed with NGC 1808 and integrated with 18 new sources found by Ho et al. (1997) which would have been included in the Maiolino & Rieke (1995) sample if they had been discovered earlier. This sample has been limited in [OIII] flux ( $F_{\rm [OIII]}> 40 \times 10^{-14}$ erg cm^-2 s^-1) in order to be independent of absorption effects and to have the greatest possible range in column density.

We have reconsidered the total sample reported in Risaliti et al. (1999), considering also the sources with $F_{\rm [OIII]}< 40 \times 10^{-14}$ erg cm^-2 s^-1and we have updated the column density distribution with $N_{\rm H}$ measurements which have become available in the meantime. Finally, we have used the X-ray versus [OIII] flux ratios to assess the possible Compton thick nature of the sample sources, if a column density was not available: sources found to be Compton thick were ascribed to $N_{\rm H}> 10^{24}$ cm^-2 range. This sample consists of 92 sources and for 59 sources we have the value of the column density. We find that the fraction of objects with $N_{\rm H}<10^{22}$ cm^-2 increases to $\sim$12% and this is to be considered a lower limit as not all sources have the column density measured. If we consider only the 59 sources with known $N_{\rm H}$, the fraction is $\sim$18%.

 

 

Table 3: Fractions for three samples.

log($N_{\rm H}$)	Risaliti (99)	Ho et al. (97)	Ho sample
cm-2	total sample	sample	<22 Mpc
<22	>12%	>10%	>15%
22-23	>14%	>4%	>7%
23-24	>16%	>14%	>19%
>24	>22%	>10%	>11%
no data	36%	62%	48%

References for $N_{\rm H}$: Ho (1999), Allen et al. (2000), Terashima et al. (2000), Guainazzi (2002),
Pappa et al. (2001), Levenson et al. (2001b), Cappi et al. (2002).

In order to check this result we also considered all sources which are classified as type 2 Seyferts (including the transition objects indicated as Liner2/Seyfert2 or Transition2/Seyfert2) in the Ho et al. (1997) sample, which is complete in magnitude to $B_{\rm T}< 12.0$ mag. The total number of these objects is 49 and the column density has been determined for 19 of them. For all these sources the lower limit for the fraction of objects with $N_{\rm H}<10^{22}$ cm^-2 still remains around 10%, while for only those with known $N_{\rm H}$ (19) the fraction is $\sim$26%. If we limit the sample in distance to <22 Mpc, in order to limit the number of sources without a column density measurement, we obtain a sample of 27 sources and 14 objects have an estimation of the $N_{\rm H}$. In this case, the lower limit for the fraction increases to 15%, while for the 14 sources this fraction is $\sim$28%. In Table 3 we show the lower limits for the column density distributions derived in the three samples.

Although the number of sources for which there is no information available is high in all samples, the agreement found is reassuring. It is also interesting to note that an important fraction of unabsorbed objects is found only when low luminosity AGNs (LLAGNs) are considered: this is evident in the Risaliti et al. (1999) total sample when objects at low luminosity are included and it is obvious in the Ho et al. (1997) sample which is mainly made up by such sources.

Recently, Mainieri et al. (2002) have presented the results of the X-ray spectral analysis of the deep survey of the Lockman Hole field with the XMM-Newton observatory. Nearly 30% of type 2 Seyfert galaxies show $N_{\rm H}<10^{22}$ cm^-2.

If confirmed by future studies, the fraction of low absorbed Seyfert 2 galaxies obtained (presumably in the range of 10%-30%) could have important implications for synthesis models of the X-ray background and the Unified Theory of Seyfert galaxies.

Figure 4: X-ray column density vs. $L_{\rm 2{-}10~keV}$ for the sources sample. No significant correlation between the two quantities is found. We highlight the region covered by Risaliti et al. (1999) sample sources and show that most of our objects are LLAGNs.

In Fig. 4 we plot the column density, measured from the X-ray spectra, against 2-10 keV luminosity corrected for absorption. As expected we find no significant correlation (at 90% level for a two tailed test) between these two quantities confirming previous findings by Risaliti et al. (1999). Nevertheless it is interesting to note that we have extended the parameter space in Fig. 4 considerably towards lower $N_{\rm H}$ and X-ray luminosities compared to Risaliti et al. (1999) work: this further supports the idea that more unobscured Seyfert 2s are observed going towards low luminosities. Since the LLAGNs constitute the majority of the nearby AGN population, it becomes evident that unobscured type 2 objects may constitute a non-negligible part of the type 2 class.