... Seyferts

Based on results collected at the European Southern Observatory, La Silla, Chile and Onsala Space Observatory, Sweden.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... ring

Suspected to exist in all Seyfert galaxies e.g. (Myers & Scoville 1987; Plante et al. 1991; Bergman et al. 1992; Irwin & Sofue 1992; Kohno et al. 1996; Tacconi et al. 1996; Curran et al. 1998; Tacconi et al. 1999a). In fact such molecular rings may in all galaxies (Sofue 1991), including our own (Güsten 1989).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... cone

Dust and ionised gas, which have accreted onto the AGN, being driven back outward in the direction of the jets by radiation pressure.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 

Except in the case of NGC 5033 which they class as a Sy1, cf. Sy 1.9 (NASA/IPAC Extragalactic Database, Maiolino et al. 1997).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

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

For the Sy1s Fig. 1 gives $L_{\rm CO}=10^{-5}L_{\rm FIR}\hspace{0mm}^{0.75}$, and forcing a gradient of 1 gives $L_{\rm CO}=10^{-8}L_{\rm FIR}{\rm ~K~ km~s}^{-1}{\rm ~kpc}^2~L_{\odot}^{-1}$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 273

The one ultra-luminous infrared galaxy (ULIRG) of the Sy2 sample.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... with

Thus ruling out the bias.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... sampled

For example, as observed in NGC 1068 (Planesas et al. 1989) and NGC 1365 Sandqvist et al. (1995).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Seyfert

Regarding the previous section, note that Circinus, the closest lying example of a Seyfert galaxy, has $L_{\rm FIR}=6.2~10^{9}L_{\odot}$ (Lonsdale et al. 1985) and $L_{\rm CO}\approx30~10^{3}$ K km s^-1 kpc² (Curran 2000), thus giving the high CO/FIR luminosity ratio expected from the $L_{\rm FIR}\sim10^{10}~L_{\odot}$ sample. Similar luminosities are found in the next furthest Seyfert NGC 4945 (Curran 2000).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... galaxies)

Where high resolution data is available, it is compared with our results.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... (velocity)

Throughout this paper we assume H₀=75 km s^-1 Mpc^-1.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... distance

Since we compare beam sizes it is not necessary to take into account the main beam efficiencies.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Circinus

The luminosity (Curran et al. 2000) could also contribute to the observed intensity, cf. Circinus (Curran et al. 1998; Curran 2000), and so we assign limits only, Table 2.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... model

A disc model gives a gentler rise.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...$30-50^{\circ}$

Although they do narrow this range down to $30-40^{\circ}$ from their models.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... observed

Applying a 14'' beam to the model and comparing it with the corresponding $2\rightarrow 1$ spectrum, we find that not such a good match is achieved (cf. Fig. 7) between the model profile and that in Fig. 4. For the $2\rightarrow 1$ transition the model spectrum is similar in shape to that in Fig. 7 and with an intensity of 0.47 K (cf. $T_{\rm mb}\approx0.5$ K, Curran et al. 2000) while the observed is quite asymmetric. We believe that this is due to a pointing error (Curran et al. 2000).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... similar

In fact the same for $\eta_{\rm mb}=0.7$at SEST, giving $T_{\rm mb}=0.36$ K for an inclination of $40^{\circ }$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Circinus

NGC 1365 is approximately five times the distance of Circinus and so based on this, a 220'' beam would be required to simulate a 600 pc ring, although a 40'' beam (approximately the $1\rightarrow 0$ beam at SEST) gives the observed intensity.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 

Using the $N_{{\rm H}_{2}}/L_{{\rm CO}}$ Galactic conversion ratio in both cases (Strong et al. 1988).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... factor

Although $\approx 600$ pc is just the bulk ring/disc extent in Circinus. The CO $1\rightarrow 0$ transition has been detected out to nearly 4 kpc along the major axis (Curran et al. 2000a).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... scale

A fast rotating disc is believed to be present within the central $\sim700$ pc (Burbridge & Burbridge 1960; Lindblad 1978), and Sandqvist (1999) has recently observed a $\approx 600$ pc molecular ring in CO $3\rightarrow2$. In the case of the $1\rightarrow 0$ transition, we would expect the molecular gas to be more extended than this; by perhaps a factor of $\approx2$, e.g. as in the Sy2s Circinus and NGC 4945 (Dahlem et al. 1993; Mauersberger et al. 1996; Curran et al. 1998; Curran et al. 2000b).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... disc

For the disc model an inclination of gives a comparable velocity width to the ring model.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... inclinations

Bryant & Scoville (1996) estimate an inclination of providing the $N_{{\rm H}_{2}}/L_{{\rm CO}}$ Galactic conversion applies, although this is questionable in ultra-luminous/Seyfert galaxies (Maloney & Black 1988; Maloney 1990; Shier et al. 1994; Mauersberger et al. 1996; Curran et al. 1998).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... times

A 200'' beam gives the observed intensity as opposed to the 1400'' beam expected from the distance scale.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... around

Since the relatively high luminosity within the beam (7 times Circinus) may lead to an overestimate (Sect. 3.8).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 

As previously mentioned, preliminary results from large scale observations of Circinus suggest a global CO luminosity of $\approx 30~10^{3}$ K km s^-1 kpc² out to $\approx5$ kpc (Curran et al. 2000). It should be noted, however, that we are using a model of the molecular ring in Circinus, which has a luminosity of $\approx5\%$ of this, to model the rings/discs in these galaxies, in which the molecular gas is expected to be very centralised (Curran et al. 2000).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 

Assuming that the rotation curve does not deviate significantly from that of the ring in Circinus, e.g. a Keplerian curve (Curran et al. 1998; Curran 1998). A similar rotation curve was used by Downes & Solomon (1998) to model the molecular gas as a disc in several ultra-luminous galaxies.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 1068

Although our results only agree (at an inclination of $40^{\circ }$) with the favoured models of Tacconi et al. (1994), where streaming motions and disc warping are taken into account.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... torus

This is assuming that it is physically meaningful to distinguish different gas structures. Bands of various gas states do appear to be present, as molecular rings are observed in many Seyfert galaxies (Sect. 1), within which the atomic gas recombines under favourable conditions to form a masing/obscuring body (e.g. Neufeld & Maloney 1995).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Sy1.5s

The remaining sub-classes are estimated accordingly.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 5135

And possibly UGC 03374, NGCs 2273, 5347, 6814 and 7130. Note that these may still be approximately aligned with the molecular ring as geometrically thick tori could obscure at considerably lower inclinations than $70^{\circ}$ (Krolik & Begelman 1988; Wilson & Tsvetanov 1994; Maiolino & Rieke 1995; Capetti et al. 1996).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Sect. 3.13)

Also possibly in NGC 5347, where the model is also uncertain (Sect. 3.12).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... HCN

From Table 1 we find that $L_{\rm HCN}/L_{\rm FIR}({\rm Sy2})\approx L_{\rm HCN}/L_{\rm FIR}({\rm Sy1})\approx10^{-9}{\rm ~K~ km~s}^{-1}{\rm ~kpc}^2~L_{\odot}^{-1}$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... luminosity

We would expect this ratio to be independent of the FIR luminosity were the FIR emission arising mainly from young stars.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Seyferts

Like the FIR, a significant proportion of the HCN emission could come from the AGN (Curran et al. 2000 and references within). Kohno et al. (1999) attribute the HCN luminosity in Seyferts to the obscuration and so a deficit in star-burst activity in type 1 nuclei may not significantly affect the observed HCN luminosity. This possibly explains why the CO luminosities differ while the HCN luminosities are similar between the two classes.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... ring

Tacconi et al. (1999b) have noted several couplings of the 100-pc scale disc to the larger scale gas in the luminous merger NGC 6240.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 

The transformation of a type 2 to a type 1 Seyfert has recently been noted by Aretxaga et al. (1999).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.