5 Shock excitation in NGC 2818

As noted previously in Sect. 3, the [SII] and [OIII] profiles along the southern part of NGC 2818 appear to be significantly more extended than their higher excitation counterparts, a characteristic which results in enhanced [SII]/H$\alpha$ and NII/H$\alpha$ line ratios towards the nebular outer edge (Fig. 3). This region is also that of strongest (and probable shock excited) H₂ S(1) emission observed by Schild (1995).

By contrast, the more northerly emission counterparts (where H₂ S(1) emission is weaker) possess similar distributions in all transitions, although the ratios [NII]/H$\alpha$ and [SII]/H$\alpha$ again peak at large radial distances.

The correlation between the locations of shocked H₂ S(1) and enhanced low excitation line ratios is at the very least suggestive, and may imply a similar shock origin for both optical and IR transitions. To evaluate this possibility, we have therefore indicated the location of the minor axis ratios [NII]/H$\alpha$ and [SII]/H$\alpha$ in a diagnostic diagram due to Sabbadin et al. (1977; Fig. 6). Although the data base employed by Sabbadin et al. appears not to have been corrected for extinction, the influence of reddening is probably very small.

Inspection of Fig. 6 suggests that the dereddened line ratios for NGC 2818 parallel the trend observed for normal PN, although they are in fact outside of the ranges expected for most emission line nebulae. There is no evidence here to support shock excitation, which might be expected to give ratios similar to those of supernova remnants (SNR).

By contrast, a rather different result is obtained where we compare the [SII]/H$\alpha$ and [OIII]/H$\alpha$ ratios with de-reddened line ratios for "normal" planetary nebulae, and the values predicted for planar and bow-shocks (see Fig. 7, and the models of Hartigan et al. 1987, and Shull & McKee 1979). In constructing this diagram, we have employed a heterogeneous sample of 492 spectra derived from the catalogue of Kaler et al. (1997), together extinctions compiled by Tylenda et al. (1992), and the extinction curve of Savage & Mathis (1979).

It is apparent, from Fig. 7, that whilst the line ratios with log([SII]/H$\alpha) < -0.5$ (observed at the centre of NGC 2818) are similar to those for "normal" PN, the ratios in the outer parts are significantly greater, and extend into the regime defined by the shock modelling solutions. It seems likely, therefore, that these ratios are explicable by shock excitation.

The characteristics of the shocks responsible for this emission are difficult to identify with precision. Where shock excitation dominates, however, then one would expect to observe the trends in [SII]/H$\alpha$ and [OIII]/H$\alpha$ noted in Fig. 8; where the individual curves correspond to differing line ratios, and planar modelling solutions having high (10³ cm^-3) and low (10² cm^-2) pre-shock densities (Hartigan et al. 1987; see labelling to the right). The observed ranges in the line ratios are indicated by vertical arrows in this figure.

Taken together, both sets of line ratio would imply values $V_{\rm s} \geq 110$ km s^-1 or larger, associated (presumably) with some as yet unobserved high velocity wind. Similar high-velocity outflows have previously been noted in NGC 6537 (Cuesta et al. 1995), M 2-9, Hb5 (Phillips & Mampaso 1988), NGC 6302 (Meaburn & Walsh 1980) and several other bipolar sources (cf. Grewing 1988), and appear to constitute a primary mechanism for shaping nebular shells at relatively late stages of PN evolution. In particular, the interaction of such winds with superwind envelopes has been invoked to explain precisely those structures which appear to characterise the NGC 2818 morphology (e.g. Cuesta et al. 1993, 1995; Balick et al. 1987; Icke & Preston 1989; Icke et al. 1992), including the dual arm-like extensions on either side of the major axis (Fig. 1).

  

Figure 6: Location of de-reddened minor axis ratios for NGC 2818 in an H$\alpha$/[NII] - H$\alpha$/[SII] diagram due to Sabbadin et al. (1977). Although the observed ratios parallel the regime for normal PN, they are outside of the ranges expected for most emission line nebulae

  

Figure 7: Location of de-reddened minor axis ratios within the [SII]/H$\alpha$ - [OIII]/H$\alpha$ plane $(\blacksquare)$. It may be noted that core values for NGC 2818 (i.e. those at low radial offsets) occur at low values of [SII]/H$\alpha$ (log([SII]/H$\alpha) < - 0.5)$; those towards the south periphery of the source are located to the upper left; and those corresponding to northerly axial displacements are positioned to the upper right (i.e. typically log([SII]/H$\alpha) \gt -0.5$, log([OIII]/H$\alpha ) \gt 0.35$). We also show, for comparison, de-reddened line ratios for a heterogeneous sample of "normal" PN ($\times$), and results for planar- and bow-shock modelling due to Hartigan et al. (1987) and Shull & McKee (1979) ($\square)$. Note that line ratios towards the periphery of NGC 2818 appear to lie increasingly within the shock modelling regime

  

Figure 8: Predicted variation in [OIII]/H$\alpha$ and [SII]/H$\alpha$ line ratios as a function of shock velocity, together (arrows) with the observed de-reddened ranges in these parameters towards the periphery of NGC 2818. We have indicated, separately, the trends anticipated for low (102 cm-3) and high (103 cm-3) pre-shock densities