4 Shell density mapping

The complexity displayed in Figs. 1 and 4 is also visible in the density map illustrated in Fig. 5, where we have employed images in [SII] $\lambda 6717$ and $\lambda 6731$ to determine $n_{\rm e}$, and used a mean temperature $T_{\rm e} = 14500$ K (relevant for [OIII], Dufour 1984). While this value may not be appropriate, the derived densities are relatively insensitive to $T_{\rm e}$. The profile for the minor axis is illustrated in Fig. 3. Several features are noteworthy:

1) It is apparent that the variation in density is both marked and irregular, with regions of enhanced density correlating closely with increases in [SII] intensity and/or line ratio [SII]/H$\alpha$. The structure of the higher density regimes, in brief, is similar to the filamentary source structure evident from direct imaging.

2) The density contrast ratio for the filaments (that is, the ratio between densities in the filaments and those of the broader envelope) is at least $\sim 2$. Given that most of the filaments are only barely if at all resolved, the actual contrast ratios are probably much higher.

3) The density peaks at $\sim 1.2\ 10^3$ cm^-3 close to the nucleus. This region may be an unresolved mix of filaments, however, and the value of $n_{\rm e}$ is probably a lower limit.

  

Figure 4: Image of [SII]/H$\alpha$ and [NII]/H$\alpha$ ratios

  

Figure 5: Map of electron density (above) derived from the [SII] $\lambda\lambda 6717$, 6731 Å images, whence it is clear that the shell density structure is rather irregular and complex. The map of [OIII]/H$\alpha$ (below) implies that excitation trends are also complex, with higher excitation in an annular regime having $\Delta\alpha \times\Delta\delta$ $36\times 16$ arcsec2

4) There appears to be a weak but suggestive connection in the northern part between minor axis [OIII]/H$\alpha$ filaments, and some of the higher density ridges and condensations. It is conceivable that we are witnessing the effects of shock interaction between the primary shell and a collimated outflow. Such mechanisms may also explain point-reflection symmetries in other nebulae, and radially extended low excitation features in NGC 6210 and K 1-2 (Phillips & Cuesta 1996; Bond & Livio 1990).

5) There is no clear evidence for density enhancement close to the southerly limit of the ionised shell, as might be expected were appreciable shock compression to occur (see discussion in Sect. 5). This may, however, be a consequence of lower [SII] intensities in this regime, and of the corresponding difficulty in evaluating reliable [SII] line ratios.