2 The sample

The sample consists of objects which are classified in NED as type 2 Seyferts and are characterized by low absorption in X-rays ( $N_{\rm H}~\rlap{$<$ }{\lower 1.0ex\hbox{$\sim$ }}10^{22}$ cm^-2).

  

 

Table 1: Unobscured Seyfert 2 galaxies sample.

Nome	RA & Dec (2000)	z	$N_{\rm HGal}^{\S}$	NED
				class.
IRAS 00317-2142	00 34 13.8 -21 26 21	0.02680	1.5	S1.8
IRAS 01428-0404	00 45 25.2 -03 49 36	0.01820	4.28	S2
IC 1631	01 08 44.8 -46 28 33	0.03084	2.17	S2$^\star$
NGC 2992	09 45 42.0 -14 19 35	0.00771	5.26	S2
NGC 3147	10 16 53.6 +73 24 03	0.00941	3.64	S2
NGC 4565	12 36 20.8 +25 59 16	0.00428	1.30	S1.9
NGC 4579	12 37 43.5 +11 49 05	0.00507	2.47	L/S1.9
NGC 4594	12 39 58.8 -11 37 28	0.00364	3.77	L/S1.9
NGC 4698	12 48 23.0 +08 29 14	0.00334	1.87	S2
NGC 5033	13 13 27.3 +36 35 36	0.00292	1.03	S1.9
MRK 273x	13 44 47.4 +55 54 11	0.45800	1.10	S2
NGC 5995	15 48 24.9 -13 45 28	0.02519	10.6	S2
NGC 6221	16 52 46.1 -59 13 07	0.00494	15.0	S2$^\star$
NGC 6251	16 32 31.8 +82 32 16	0.02302	5.49	S2
IRAS 20051-1117	20 07 51.3 -11 08 33	0.03149	6.8	SB/S2
NGC 7590	23 18 55.0 -42 14 17	0.00532	1.96	S2
NGC 7679	23 28 46.8 +03 30 41	0.01714	5.13	S2$^\star$

§ Column density in units of 10²⁰ cm^-2.
$\star$ Transition objects between Seyfert 2 and starburst (see Fig. 1).

The 17 objects of the sample are listed in Table 1 with optical positions in equatorial coordinates for epoch J2000, redshift z as reported in NED, Galactic column density from 21 cm measurement in units of 10²⁰ cm^-2 obtained from the HEASARC (High Energy Astrophysics Science Archive Research Center) on-line service and NED classification ( $\rm S=Seyfert$, $\rm L=Liner$, $\rm SB=Starburst$). In Table 2 we list the main X-ray spectral parameters in addition to the Infrared and [OIII]$\lambda$5007 flux values.

2.1 Optical classification

Figure 1: a) Diagnostic diagram with [OIII]$\lambda$5007/H$_{\beta }$ vs. [NII]$\lambda$6583/H$_{\alpha }$. b) Diagnostic diagram with [OIII]$\lambda$5007/H$_{\beta }$ vs. [OI]$\lambda$6300/H$_{\alpha }$. In both a) and b) diagrams Seyfert 2, Starburst and HII regions are divided by solid lines.

 

 

Table 2: X-ray spectral parameters for the sample.

	Nome	$\Gamma$	$N_{\rm Hint}^{a}$	EW(Fe $_{\rm k\alpha}^b$)	$F_{\rm 2 {-} 10~keV}^c$	$L_{\rm 2 {-} 10~keV}^d$	$F_{\rm IR}$e	$F_{\rm [OIII]}$f
1	IRAS 00317-2142	2.00 +0.07-0.07	1.9 +0.2-0.2	<900	0.08	42.01	25.94	24
2	IRAS 01428-0404	1.95 +1.33-0.73	32 +109-32	-	0.04	41.38	<7.42	-
3	IC 1631	2.10 +0.10-0.10	<31.6	<70	1.00	43.23	<6.85	52.0
4	NGC 2992	1.70 +0.2-0.2	90.0 +3.0-3.0	147 +37-37	7.4	42.30	50.79	680.0
5	NGC 3147	1.94 +0.20-0.19	<2.9	675 +395-328	0.22	41.62	44.39	9.0
6	NGC 4565	1.7 +0.2-0.2	$N_{\rm H,gal}$	-	0.02	39.32	42.03	6.0
7	NGC 4579	1.88 +0.03-0.03	$N_{\rm H,gal}$	-	0.52	41.23	38.37	9.0
8	NGC 4594	1.5 +0.04-0.03	17 +11-9	-	0.16	40.86	31.90	7.0
9	NGC 4698	1.91 +0.14-0.14	8.1 +8.2-7.8	<425	0.10	40.50	<3.14	2.4
10	NGC 5033	1.7 +0.02-0.02	$N_{\rm H,gal}$	100 +100-100	0.28	41.04	82.77	17.0
11	MRK 273x	1.66 +0.15-0.11	14.1 +5.5-5.0	<30	0.01	43.62	-	0.14
12	NGC 5995	1.81 +0.04-0.03	90 +5-3	144 +41-41	2.89	43.52	28.54	66
13	NGC 6221	1.9	110 +8.6-8.3	360 +210-93g	1.4	41.78	-	2.14
14	NGC 6251	1.83 +0.21-0.18	75 +64-58	443 +313-272g	0.14	41.13	-	57.0
15	IRAS 20051-1117	1.92 +0.21-0.14	<40	272 +52-73	0.24	42.63	7.59	15.2
16	NGC 7590	2.29 +0.20-0.13	<9.2	-	0.12	40.79	44.53	17.0
17	NGC 7679	1.75 +0.03-0.06	2.2 +1.5-1.4	<200	0.60	42.53	49.60	108.26

^a Column density in units of 10²⁰ cm^-2, ^b Equivalent width of the Fe $_{k\alpha}$ in units of eV, ^c 2-10 keV flux in units of 10^-11 erg cm^-2 s^-1, ^d 2-10 keV logarithm of unabsorbed luminosity ( H₀ = 75 km s^-1 Mpc^-1), ^e Infrared flux in units of 10^-11 erg cm^-2 s^-1, ^f [OIII]$\lambda$5007 corrected flux in units of 10^-14 erg cm^-2 s^-1, ^g Line at 6.65 keV. X-ray references: (1) Georgantopoulos et al. (2000); (2) This work, ASCA data; (3) Awaki et al. (1992); (4) Gilli et al. (2000); (5) This work, SAX data; (6) Cappi et al. (2002); (7) Eracleous et al. (2001); (8) Pellegrini et al. (2002); (9) Pappa et al. (2001); (10) Cappi et al. (2002); (11) Xia et al. (2002); (12) This work, ASCA data; (13) Levenson et al. (2001b); (14) Sambruna et al. (1999); (15) This work, ASCA data; (16) Bassani et al. (1999); (17) Della Ceca et al. (2001). All $F_{\rm IR}$ are based on IRAS data taken from NED. [OIII] references: (1) Moran et al. (1996); (2) Pietsch et al. (1998); (3) Sekiguchi et al. (1993); (4) Gilli et al. (2000); (5) Ho et al. (1997); (6) Ho et al. (1997); (7) Ho et al. (1997); (8) Ho et al. (1997); (9) Ho et al. (1997); (10) Ho et al. (1997); (11) Xia et al. (1999); (12) Lumsden et al. (2001); (13) Levenson et al. (2001b); (14) Shuder & Osterbrock (1981); (15) Moran et al. (1996); (16) Vaceli et al. (1997); (17) Kewley et al. (2000).

All objects in Table 1 are reported as pure or composite Seyfert 2s in NED. In order to check the NED Seyfert 2 classification accurately, we have employed the two optical diagnostic diagrams by Veilleux & Osterbrock (1987) which use line-intensity ratios that are relatively insensitive to reddening and are considered good excitation indicators. Line ratios such as [OIII]$\lambda$5007/H$_{\beta }$, [NII]$\lambda$6583/H$_{\alpha }$, [OI]$\lambda$6300/H$_{\alpha }$ delineate the different excitation mechanisms which operate in HII regions, high-excitation AGNs (Seyferts) and low-excitation AGNs (low-ionization nuclear emission-line regions, LINERS; Heckman 1980).

Although the boundaries between these three classes are not rigorously defined, these diagrams represent a valid system to distinguish between various types of narrow emission line objects. The line ratios shown in Fig. 1 are taken from the literature (see the [OIII]$\lambda$5007 flux references of Table 2). For those sources with more than one observation we adopted the most recent reference; note that not in all cases are the data available or complete.

Most of the objects plotted in Fig. 1 show a well defined optical classification: they are classical type 2 sources. As expected, a few objects are located at the boundaries between Seyfert 2 and Liner/Starburst and so are likely to be Composite objects. The composite Seyfert/Liner nature of NGC 4579 and NGC 4594 is clear cut as seen in both diagrams confirming their NED classification; NGC 5033 and NGC 3147 are much less clear examples as they lie at the boundaries in one diagram but not in the other and so we maintain their NED definition.

Transition objects between Seyfert 2 and Starburst are NGC 6221 and NGC 7679; IC 1631 could be similar but unfortunately we lack information on the [OI]/H$_{\alpha }$ ratio to confirm this hypothesis. IRAS 20051-1117 which is classified as composite in NED is confirmed as this type only in one of the two diagrams and even in this case it is a borderline object: we therefore take this as an indication of the predominance of the Seyfert 2 signature. Therefore we substantially confirm the NED classifications (except in three cases which are flagged in Table 1) and conclude that all sources of our sample are characterized by an optical type-2 signature.

2.2 Diagnostic diagrams

The X-ray characteristics of the sample sources (described in Table 2) strongly suggest the presence of an AGN often of low-luminosity in most objects (the photon indices are canonical and Iron lines are sometimes detected). However this evidence is not sufficient to establish the presence of an active nucleus in all objects and in particular in low luminosity sources, where the luminosity does not allow us to discriminate between emission from an active nucleus or a starburst galaxy. Furthermore our objects could be Compton thick ( $N_{\rm H}> 10^{24}$ cm^-2) but since the photoelectric cut-off would not be detectable in the 2-10 keV spectrum, the column density measurements would be too low. However the presence of an AGN and the Compton nature (thin or thick) of each source can be checked by comparing isotropic versus anisotropic properties.

If a molecular torus is present in Seyfert galaxies, then it should block the X-ray emission coming from the central engine but it shouldn't intercept emission coming from larger scale structures like the Narrow Line Region or a non nuclear starburst region. The column density could then be inferred from the flux ratios of the X-ray fluxes versus various isotropic emission measurements. The [OIII]$\lambda$5007 flux is considered a good isotropic indicator because it is produced in the Narrow Line Region (Maiolino & Rieke 1995; Risaliti et al. 1999; Bassani et al. 1999). Also the Far-Infrared emission seems to be produced over a larger region than that of the molecular torus and it has been used as an isotropic indicator too.

Figure 2: $F_{\rm X}/F_{\rm IR}$ vs. $F_{\rm [OIII]}/F_{\rm IR}$ for the sources of the sample. We separate with dashed lines the Compton thin, Compton thick and Starburst region. The fluxes used are listed in Table 2.

The $F_{\rm X}/F_{\rm [OIII]}$ ratio has been studied in a large sample of Seyfert 2 galaxies: all Compton thin Seyferts show ratios higher than $\sim$1 while Compton thick sources show ratios below this value (Bassani et al. 1999). The $F_{\rm X}/F_{\rm IR}$ ratio has also been largely discussed and used in the literature to investigate the presence of high column densities. Typically type 1 and Compton thin type 2 AGN show ratios of $\sim$0.1, while Compton thick type 2 objects show ratios lower than $5 \times 10^{-4}$(David et al. 1992; Mulchaey et al. 1994; Risaliti et al. 1999). Finally, since infrared emission is associated mainly with star-forming activity while the [OIII]$\lambda$5007 emission is produced by photons generated in the active nucleus, by comparing the $F_{\rm [OIII]}/F_{\rm IR}$ ratio of all the starbursts and AGNs in the Ho et al. (1997) sample, Bassani et al. (in preparation) find that $\sim$90% of the starburst galaxies show a value below 10^-4 while $\sim$88% of AGNs show a value above 10^-4, therefore they suggest this value as a means of discriminating between one galaxy class and the other. In conclusion, these 3 ratios can provide an independent way to establish which is the dominant component between AGN or starburst and at the same time they are a powerful tool in the detection of Compton thick sources when an X-ray spectral analysis is not sufficient.

Figure 3: $F_{\rm X}/F_{\rm [OIII]}$ vs. $F_{\rm [OIII]}/F_{\rm IR}$ for the sources of the sample. We separate with dashed lines the Compton thin, Compton thick and starburst region. The fluxes used are listed in Table 2.

The fluxes used in our diagrams (Figs. 2 and 3) are listed in Table 2. All far-infrared fluxes are based on IRAS data, taken from NED. For this paper we adopt the same definition used in Mulchaey et al. (1994) for the far-infrared flux:

$$% F_{\rm IR}=F_{\rm 25~\mu m} \times \nu_{\rm 25~\mu m} + F_{\rm 60~\mu m} \times \nu_{\rm 60~\mu m}.$$ (1)

Note that due to the IRAS angular resolution the $F_{\rm IR}$ value is likely to be overestimated for the nuclear region of each galaxy; this translates into $F_{\rm X}/F_{\rm IR}$ and $F_{\rm [OIII]}/F_{\rm IR}$ ratios smaller than in reality. The [OIII]$\lambda$5007 flux of each galaxy has been corrected for extinction using the formula given in Bassani et al. (1999).

In both Figs. 2 and 3 the starburst region is empty suggesting that the dominant component is likely to be an AGN; the region populated by our sources is that of the Compton thin regime, indicating that indeed we measure the real amount of absorption in all our objects. This result is confirmed also for those objects for which the data are not complete: the $F_{\rm X}/F_{\rm [OIII]}$ratios for MRK 273x, NGC 6221 and NGC 6251 are 71, 654 and 24 respectively and the $F_{\rm X}/F_{\rm IR}$ ratio for IRAS 01428-0404 is $5.4 \times 10^{-3}$.

The major problem when dealing with low luminosity AGN is the possible contamination from off nuclear sources in the observed galaxy: in this case the measured flux is overestimated by non-imaging/low-angular resolution instruments and the spectrum alterated. This is what is emerging from recent works based on Chandra and XMM-Newton observations: the X-ray fluxes are in most cases lower than those measured by past satellites (see Chandra results by Ho et al. 2001). A consequence of this could be a mistaken evaluation of the column density and/or of the Compton nature of the source. However, for those 5 sources of our sample which have Chandra or XMM-Newton data the spectral parameters are substantially in agreement with the old values and even if the 2-10 keV luminosities are somehow decreased, all these objects are still in the Compton thin region.

In any case, it is important to consider the possibility that more accurate and higher spatial resolution observations could reduce the flux of some of our sources.