Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 327, 909-920 (1997)

Previous Section Next Section Title Page Table of Contents

4. Results and comparison with Circinus

4.1. Comparison with coronal line ratios

OSMM observed lines of very high excitation, mostly found in the infrared, in the Circinus Galaxy. Mo96 recently augmented the dataset with ISO observations out to 45 [FORMULA] m.

We now compare the spectrum of model H (Table 1) with the coronal lines observed in Circinus. As the [Si IX ]3.935 [FORMULA] m line was measured both by ISO (Mo96) and from the ground (OSMM), we combined the two data sets by expressing all line ratios relative to their respective [Si IX ] flux measurement. As a rule, we retain the OSMM ratios only for the lines not measured with ISO (e.g. [Fe X ] [FORMULA] 6374). Note that the optical groundbased observations were corrected for reddening ([FORMULA]) while the ISO data was not. Reddening is of little conseqence in the mid- and far-infrared since the correction is smaller than the observational errors.

We present in Fig. 5 the ratio of observed to model line fluxes as a function of ion- ization energy. This diagram may be compared directly with the corresponding Fig. 4 of Mo96. We have omitted [Ne III ]15.55 [FORMULA] m since this line is dominated by the IB component (cf. Table 1). The degree of agreement between model and observations is similar for our model and that of Mo96. Except for the iron lines (discussed below, but not shown in Fig. 4 of Mo96) and [Mg VII ]5.51 [FORMULA] m, the dispersion in our Fig. 5 is less than a factor of two, which represents good agreement considering the uncertainties in atomic data (see review by Oliva 1997). The discrepancy between our results and those of Mo96 for [Mg VII ] is probably due to the use of different atomic data. We therefore conclude that MB clouds with internal density gradients induced by radiation pressure are equally successful in fitting the observations as the onion ring geometry with a UV bump studied by OSMM and Mo96.

[FIGURE] Fig. 5. Observed divided by modelled line flux ratios, where all ratios are expressed relative to [Si IX ]3.935 [FORMULA] m. The lines plotted include [O IV ]25.90 [FORMULA] m (open square), [Si VI ]1.963 [FORMULA] m, [Si VII ]2.483 [FORMULA] m, [Si IX ]3.935 [FORMULA] m (filled circles), [S IV ]10.54 [FORMULA] m, [S VIII ] [FORMULA] 9913, [S IX ] 1.25 [FORMULA] m (filled squares), [Ne V ]14.32 [FORMULA] m, [Ne V ]24.31 [FORMULA] m , [Ne VI ] 7.66 [FORMULA] m (upper filled triangles), [Mg V ]5.62 [FORMULA] m , [Mg VII ]5.51 [FORMULA] m , [Mg VIII ]3.03 [FORMULA] m (open triangles), [Ca VIII ]2.321 [FORMULA] m (open pentagon), [Fe VII ] [FORMULA] 6086 and [Fe X ] [FORMULA] 6374 (lower filled triangles). The model is a stratified matter-bounded slab with [FORMULA] =0.5 .

4.2. 'Onion' ring vs cloud geometry

As in Korista & Ferland (1989), OSMM considered that the high excitation gas in Circinus has a filling factor close to unity and a geometry analogous to that of a multilayered `onion', with the excitation of the gas decreasing radially as a result of the increasing dilution of the nuclear radiation. We have proposed here a different geometry in which each individual cloud is sufficiently stratified in excitation to emit the entire observed range of high excitation lines. In our model, the basic difference between NLR clouds and the gas emitting the coronal lines is that the latter have a much higher ionization parameter. On the other hand, the alternative `onion' ring geometry presupposes a uniform distribution of gas which differs from the very low volume filling factor believed to characterize the NLR (Osterbrock 1989). In the specific case of Circinus, a uniform spherical distribution of coronal gas at the high density inferred from the [Ne V ] 24.3 [FORMULA] m/14.3 [FORMULA] m ratio implies that the highest excitation species are confined to radii smaller than 0.2" (Mo96); otherwise the coronal line luminosities would exceed the observed level 2. To the extent that the photon luminosity of the source is unconstrained, this limit in radial distance does not apply to our model since our line luminosities are governed by the covering factor, which is a free parameter.

The geometrical depth of our internally stratified cloud in model H is only 1 pc (hydrogen column density of [FORMULA] atoms [FORMULA] @ or [FORMULA] " at the distance of Circinus ([FORMULA] Mpc). Knowing [FORMULA] of the coronal gas, the brightness predicted for [Si IX ]3.935 [FORMULA] m by model H, and the observed [Si IX ]3.935 [FORMULA] m luminosity in Circinus ([FORMULA]   [FORMULA]  erg cm-2  s-1 after correcting for extinction), we infer that the total cross sectional area of the MB clouds responsible for the coronal line emission must be [FORMULA]. This would imply a covering factor [FORMULA] of 0.13 if the clouds were to lie at a mean radius of 10 pc (0.5") from an isotropic nuclear source. The inferred ionizing photon luminosity in this case is [FORMULA] [FORMULA] for an isotropic source, which is similar to the value derived by Mo96.

4.3. The iron lines

Although encouraging, Fig. 5 shows significant discrepancies in the case of the optical Fe lines (which were left out of Fig. 4 of Mo96). One possible explanation is the large uncertainty in the collision strengths of some Fe transitions. Chief among these is [Fe X ] [FORMULA] 6374, for which recent determinations of the collision strength by Mohan et al.(1994) and Pelan & Berrington (1995) differ by an order of magnitude. Another factor is the large extinction. If obscuration is nonuniform over the nuclear regions of the Circinus Galaxy, dereddening of the emission line spectrum is not possible without precise knowledge of the dust distribution. The fact that the Fe lines are the only optical lines in Fig. 5 is probably not a coincidence. It is conceivable that one sees in the optical only a small portion of the emission region observed in the far infrared. Finally, model H cannot account for the luminosity of [Fe XI ] [FORMULA] 7892 as observed by OSMM.

4.4. The energy distribution: UV bump or no UV bump?

Since our fit to the line ratios is of similar quality to that of Mo96, we argue that one cannot uniquely determine the spectrum of the ionizing radiation from the emission-line ratios. We do not exclude the possibility of a UV bump but consider that its existence cannot be proven with the line ratios plotted in Fig. 5. We have explored using various energy distributions, some very similar to that of Mo96, but without obtaining any real improvement. The most likely explanation for this difference between our results and those of Mo96 is the presence of density gradients in our models. Internally stratified slabs produce a significantly `richer' mixture of high and low excitation lines in comparison with the standard isobaric models (Binette & Raga 1990). Even if we lower the X-ray luminosity by joining the two power-laws of Fig. 1 (dash-lines) at 4 000 eV instead of 500 eV, we obtain almost indistinguishable results by simply using a slightly lower [FORMULA].

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

Online publication: April 6, 1998