5. The intermediate and low excitation lines
5.1. Rarity of clouds with coronal line emission
OSMM showed that unlike the coronal lines, the velocities of the optical lines of [O III ] 5007, [N II ] 6583 @ [O I ], [S II ], H and He II 4686 are essentially consistent with systemic. Inspired by the work of Viegas & Prieto (1992), we have discussed in Paper I the possible existence of two populations of clouds to explain the range of gaseous excitation observed in the optical in a small sample of Seyfert and radio galaxies. What are the implications of a much higher excitation component for this kind of model? Although the matter-bounded component of model H does account for the coronal lines, it is probably unsuitable for the bulk of the intermediate excitation lines (e.g. [O III ]), which are at least in part kinematically distinct (OSMM). There must also be gas of lower excitation because the ratios of the coronal lines to the recombination lines are larger in model H than is observed. For instance, adopting a recombination case B ratio of Br /H ( K), we infer that the [Si IX ]3.935 m/H ratio in Circinus is . This value is 7 times smaller than that predicted by the matter-bounded component of model H. Of course part or all of the ionizing radiation escaping the MB slab (65%) can be reprocessed by more distant/denser low excitation clouds. However, even allowing for this (adopting A =2.5 as discussed below), we find that the coronal gas component must have a covering factor of only a quarter of that of other less excited but still matter-bounded clouds. For this reason, we have calculated MB models of lower (Table 1). As the lowest excitation coronal lines (e.g. [Si VI ],[O IV ], [S IV ]) in Fig. 5 can be produced by the lower models, a proper fit of all the coronal lines would entail combining model H with other models of lesser excitation such as the MB component of model M (). We believe, however, that such an attempt is premature given the few constraints available and the uncertainties in atomic data and metallicity.
5.2. The low excitation IB component
Whether or not the observations require a range in excitation for the MB clouds, 65% of the impinging ionizing photons still escape those clouds. We therefore propose, as in Paper I, that some of these unabsorbed photons are absorbed at larger radii and reprocessed into the observed low excitation lines. This component, dubbed IB (ionization-bounded) in Paper I, can be calculated for each MB component in Table 1 once we decide on its ionization parameter U . As we do not have enough information to determine the proper combination of the matter-bounded models H, M and L, we are even less inclined to try to determine which IB component is most appropriate. The IB line spectra listed in Table. 1 are therefore tentative and are intended for comparison with other Seyferts.
As U cannot be constrained from the optical spectrum of Circinus alone, we have adopted as constraints the same sample of line ratios as in Paper I, which are representative of the ENLR of Seyfert and radio galaxies. We found that an ionization parameter of U allows model H (see following sections) to fit reasonably well most of the line ratios covered in Paper I. This U is realized by simply applying a geometrical dilution factor of to the spectrum transmitted through the MB component. For definiteness we have set U constant for all three models. For some of the lines, the intensities are quite similar across the three IB models of Table 1 (e.g. [O II ] 3727, [O I ] 6300, [S II ] 6724 and [N II ] 6583 @ We have not explored a lower density IB component (of equal U , therefore implying a higher dilution of the radiation) as would be required to account for the low density inferred from [S II ] in Circinus ( cm-3). Such a procedure, however, has been successfully applied by Rodriguez-Ardilla et al. (1997) to the two Seyferts A08.12 and F10.01 which are characterized by a much larger density contrast between the high and low excitation emission line regions.
5.2.1. The parametrized A -sequence
The parameter A is used to represent the relative proportion of visible MB and IB clouds. It is a free parameter which alters the level of excitation of the spectra in much the same way as the ionization parameter in traditional photoionization models. A sequence of models as a function of A may be obtained by combining the MB and IB spectra of any of the three models of Table 1 as follows:
where R , R , R are the line ratios relative to H of line i. C represents the luminosity ratio of the H produced by the MB component relative to the IB component and is a constant for a given model.
Our scheme for using matter-bounded components is fairly simple and basic as it recognizes only two populations of clouds. However, as shown in Paper 1, it presents significant advantages, including a resolution of the temperature problem. Other schemes in which a continuous range of cloud opacity is considered for the NLR or ENLR gas can be found in Komossa & Shulz (1997), Nazarova et al. (1997), Moore & Cohen (1994, 1996), Acosta et al. (1996), Morganti et al. (1991) and Binette (1984).
5.2.2. Comparison with Paper I
The A -sequence of model H (solid line) is plotted in Figs. 6 and 7. The most direct measure of A is the He II 4686 @ ratio. In the Circinus Galaxy, the dereddened ratio of 0.32 of OSMM indicates a value of A .
In Fig. 6 we see that model H reaches R , corresponding to Te 15 000 K (at a density of cm-3), in agreement with the high EELR temperatures inferred by Storchi-Bergmann et al. (1996) and Wilson, Binette & Storchi-Bergmann (1997). Fig. 7 indicates that very small [N II ] 6583 @ ratios, as observed in Pks 0349-27 (filled triangles), are not necessarily a signature of star formation but a natural consequence of having a region dominated by matter-bounded clouds (Binette 1984).
© European Southern Observatory (ESO) 1997
Online publication: April 6, 1998