2 Statistical studies of the difference in dense gas abundances between type 1 and type 2 Seyferts

In Table 1 we show the results of Curran et al. (2000).

  

 

Table 1: The observed luminosities of the sample. Sy refers to the main Seyfert type, $L_{{\rm CO}~1\rightarrow0}$ and $L_{{\rm HCN}~1\rightarrow0}$ refer to the luminosity over the HPBW for each respective transition, with $1\sigma$ errors and upper limits [ $\times 10^{3}$ K km s^-1 kpc²]. Like Heckman et al. (1989) we have calculated the physical beam area by using the distances given by heliocentric radial velocities. The $\int L_{{\rm CO}~1\rightarrow0}$ refers to the global CO  $1\rightarrow 0$ luminosity [ $\times 10^{3}$ K km s^-1 kpc²]: ^aYoung et al. (1995), ^bSandqvist et al. (1995) (over $204''\times 164''$) and ^cBryant & Scoville (1996). $L_{\rm FIR}$ [ $10^{10}L_{\odot }$] refers to the far infrared luminosity computed using the FIR flux (Lonsdale et al. 1985) and v is the heliocentric radial velocity (NASA/IPAC Extragalactic Database). Adapted from Curran et al. (2000)

Galaxy	Sy	$L_{{\rm CO}~1\rightarrow0}$	$\int L_{{\rm CO}~1\rightarrow0}$	$L_{{\rm HCN}~1\rightarrow0}$	$L_{\rm FIR}$	v [km s-1]
NGC 0034	2	$2.3\pm0.1$	-	$0.56\pm0.07$	14.3	5931
NGC 0931	1	$0.08\pm0.04$	-	<0.1	2.1	5001
NGC 1068	2	$0.42\pm0.01$	$23\pm8^a$	$0.09\pm0.01$	7.4	1134
NGC 1365	2	$1.5\pm0.1$	5.3b	$0.16\pm0.03$	6.8	1636
NGC 1667	2	$1.18\pm0.08$	-	$0.45\pm0.06$	4.2	4547
UGC 03374	1	$0.40\pm0.04$	-	<0.05	2.9	6141
NGC 2273	2	$0.041\pm0.003$	$0.38\pm0.07^a$	$0.008\pm0.004$	0.66	1840
Mrk 10	1	$0.3\pm0.1$	-	<0.1	2.7	8770
NGC 4593	1	$0.08\pm0.02$	-	-	0.79	2698
Mrk 231	1	$6.0\pm0.6$	5c	$1.0\pm0.2$	128	12651
NGC 5033	2	$0.093\pm0.006$	$10\pm3^a$	$0.014\pm0.002$	0.53	875
Mrk 273	2	$2.4\pm0.4$	-	$2.4\pm0.8$	73	11318
NGC 5135	2	$2.0\pm0.1$	-	$0.11\pm0.01$	9.0	4112
NGC 5347	2	$0.03\pm0.01$	-	<0.01	0.28	2336
NGC 5548	1	$0.4\pm0.2$	-	-	0.86	5149
Arp 220	2	$2.6\pm0.1$	-	$0.4\pm0.2$	84	5314
NGC 6814	1	$0.118\pm0.002$	$0.38\pm0.06^a$	$0.011\pm0.002$	0.66	1563
NGC 7130	2	$2.6\pm0.2$	-	$0.16\pm0.02$	11.9	4842
NGC 7172	2	$0.23\pm0.04$	-	-	1.2	2603
NGC 7469	1	$2.00\pm0.09$	$1.9\pm0.3^a$	$0.25\pm0.07$	18.2	4889

The source list is based upon the detections of Heckman et al. (1989) as when planning the survey it was our original intention to test the previous results. In the same way as Heckman et al. (1989), the Seyferts are segregated into two main classes using the classification scheme of Maiolino & Rieke (1995) (and Meurs & Wilson 1984; Edelson 1987; Osterbrock & Shaw 1988); where types 1, 1.2 and 1.5 constitute type 1 Seyferts (hereafter Sy1s) and types 1.8, 1.9 and 2 constitute type 2 Seyferts (hereafter Sy2s).

When we plot the CO luminosity against that of the FIR, Fig. 1,

Figure 1: $\log L_{\rm CO}$ [ ${\rm K ~km~s}^{-1}~{\rm kpc}^2$] versus $\log L_{\rm FIR}$ [$L_{\odot}$] for the observed Sy1s (crosses) and Sy2s (circles) with recessional velocities of  km s-1: This velocity corresponds to a HPBW of $\approx12$ kpc (at SEST, $\approx 10$ kpc at Onsala (OSO); assuming H0=75 km s-1 Mpc-1), beyond which there is expected to exist little molecular gas (Maiolino et al. 1997), making us confident that we have sampled all of the CO in the faster receding sources

we see that apart from an anomalous point at $\approx12$, $\approx3.5$, the two linear fits would be close to parallel, although the intercept would be $\approx0.4$ higher for the Sy2s making $\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy2})\approx3\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy1})$. Taking the average value of $L_{\rm CO}/ L_{\rm FIR}$ for both the main classes, however, we find

$$\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy2})\approx\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy1})$$ (1)

$\approx10^{-8}{\rm ~K~ km~s}^{-1}{\rm ~kpc}^2~L_{\odot}^{-1}$. The anomalous point in Fig. 1 is in fact due to Mrk 273 which has been noted to have an unusually low $L_{\rm CO}/L_{\rm HCN}$ ratio (Curran et al. 2000).

Taking the average $L_{\rm CO}/ L_{\rm FIR}$ ratios for all of the sample (i.e. including the global values), we find that

$$\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy2})\approx3\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy1})$$ (2)

$\approx 4~10^{-8}{\rm ~K~ km~s}^{-1}{\rm ~kpc}^2~L_{\odot}^{-1}$. These results suggest two distinct possibilities:

1. Only use of the sources with  km s^-1 is valid, i.e. Eq. (1) holds. This agrees with the results of Maiolino et al. (1997), who compare the CO with blue luminosities and HI mass;
2. Use of all of the sources is valid, i.e. Eq. (2) holds. This gives a similar result to Heckman et al. (1989), who also use HI mass to obtain the same result as they get from the blue luminosities.

Since our results may be consistent with those of both Heckman et al. (1989) and Maiolino et al. (1997), where CO and blue luminosities are compared, they may suggest that the blue continuum is more isotropic than previously thought, i.e. arising from an additional source to the AGN (e.g. Fernandes & Terlevich 1995).

Our findings could be compounded by the fact that (perhaps due to a selection effect), on average, . From our sample we find that, on average, $L_{\rm FIR}~({\rm Sy2})\approx L_{\rm FIR}~({\rm Sy1})(\approx 2~10^{11}~L_{\odot}$) and so it appears as though the results are bias free, although the average FIR luminosity values are dominated by the ULIRGs Mrk231, Mrk 273 and Arp220 (Table 1). If we exclude these from the sample, we find $L_{\rm FIR}~({\rm Sy2})\approx 1.4L_{\rm FIR}~({\rm Sy1})$and if we take the averages for all of the sample with  km s^-1 the factor remains similar (1.3). Applying this factor to Eq. (2) gives $L_{\rm CO}({\rm Sy2})\approx2L_{\rm CO}({\rm Sy1})$, i.e. the result of Heckman et al. (1989). When we examine their observational results, however, we find that the single position observations have been used to determine the luminosities for all of the observed sample. As mentioned in Fig. 1, we believe that the CO is only fully sampled in sources in which  km s^-1, or in the case of the NRAO 12 m 55'' beam,  km s^-1, which leaves the same "distant'' sample as ours. It is not quite clear whether their method (normalising the beam area to the optical area) takes this effect fully into account when calculating CO luminosities for the  km s^-1 (near-by) sample. Note that we obtain the same result as Eq. (1) when we use the whole sample without considering beam-filling, i.e. using the third column of Table 1 for the CO luminosities regardless of v.

In the case of Maiolino et al. (1997), they have excluded sources in which the beam does not sample out to radii beyond 4 kpc, i.e.  km s^-1 with the NRAO 12 m. Their sample corresponds to our distant sample, suggesting that $\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy2})\approx\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy1})$ for . However, Fig. 1, which uses the same sample, suggests that $\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy2})\approx3\frac{L_{\rm CO}}{L_{\rm FIR}}({\rm Sy1}),$which is also found to apply for the whole sample (according to the literature). Referring to Curran et al. (2000), we see that, asides from the differences in mean $L_{\rm FIR}$ between the near-by and distant galaxies, $L_{\rm CO}/L_{\rm HCN}\approx6$ (distant Seyferts) cf. $L_{\rm CO}/L_{\rm HCN}\approx17$ (all of the Seyferts) and so it appears as though there exists a distinct difference between the near-by and distant samples. This is discussed in Sect. 4.