5 Ionisation structure and abundance ratios

The models for nucleosynthesis yield from massive stars predict that the mass or abundance ratio $R_{\rm X/Si}$ of ejected mass of any element X with respect to silicon varies significantly as a function of the progenitor mass M. We show the observed mean values of $R_{\rm X/Si}$as well as its rms variation in Table 2, together with the predictions for models with a progenitor mass of 11, 12 and 13 $M_{\odot }$.

 

 

Table 2: Mean measured abundance ratios and rms scatter compared with theoretical predictions for progenitor masses of 11, 12 and 13 $M_{\odot }$.

ratio	mean	rms	11 $M_{\odot }$	12 $M_{\odot }$	13 $M_{\odot }$
O/Si	1.69	1.37	0.44	0.16	0.33
Ne/Si	0.24	0.37	0.59	0.12	0.33
Mg/Si	0.16	0.15	0.57	0.12	0.41
S/Si	1.25	0.24	0.87	1.53	0.88
Ar/Si	1.38	0.48	0.65	2.04	0.64
Ca/Si	1.46	0.68	0.63	1.62	6.56
FeL/Si	0.19	0.65	1.37	0.23	0.96
FeK/Si	0.60	0.51	1.37	0.23	0.96
Ni/Si	1.67	5.52	6.89	0.68	1.80

The observed abundance ratios for X equal to Ne, Mg, S, Ar, Ca and Fe-L (the iron of the cool component) are all consistent with a progenitor mass of $12.0 \pm 0.6~ M_{\odot}$, where we used the grid of spherically symmetric models by Woosley & Weaver (1995). In these models, most of the Si, S, Ar and Ca comes from the zones where explosive O-burning and incomplete explosive Si-burning occurs, and indeed as Fig. 6 shows these elements track each other remarkably well. Furthermore the average abundance ratio fits the expectation for a 12 $M_{\odot }$ progenitor. The correlation between Si and S is remarkably sharp but not perfect, the scatter in Fig. 6 is real. These remaining residuals can be attributed to small inhomogeneities. Table 2 also indicates that the rms scatter of the abundances with respect to Si get larger as Z increases, S to Ar to Ca and indeed through to Fe and Ni. This is to be expected since elements close together in Z are produced in the same layers within the shock collapse structure while those of very different Z are produced in different layers and at different temperatures.

The Fe which arises from complete and incomplete Si burning should give rise to iron line emission. For both the Fe-L and Fe-K lines we see that iron abundance varies over the remnant but does not show any straightforward correlation with the other elements (there is a very large scatter in $R_{\rm Fe/Si}$). This is to be expected if most of the iron arises from complete Si burning. We return to the different morphologies of Si and Fe later in the discussion.

Ne and Mg are mostly produced in shells where Ne/C burning occurs, and the relative scatter in terms of $R_{\rm X/Si}$ is indeed much larger than for S, Ar and Ca (Table 2). Furthermore the abundance maps of Ne and Mg in Fig. 5 are similar and very different from the Si, S, Ar and Ca group.

The oxygen abundance is much higher than predicted by theory, contrary to all other elements. We cannot readily offer an explanation for this, but there are at least two complicating factors. As the XMM RGS maps show (Bleeker et al. 2001), oxygen has a completely different spatial distribution to the other elements (it is more concentrated to the North), and it is also much harder to measure due to the strong galactic absorption and relatively poor spectral resolution of the EPIC cameras at low energies.

The map of the ionisation age of the cool component shows a large spread. The average value at the Northern rim (few times 10¹¹ cm^-3s) matches nicely the value derived from ASCA data (Vink et al. 1996). At the SE rim the ionisation age is much larger (cf. Vink et al. $4\times 10^{11}$ cm^-3s). We confirm this higher value, but also see that there is a large spread in ionisation age. It should also be noted that for ionisation ages larger than about 10¹² cm^-3s the plasma is almost in ionisation equilibrium and therefore the spectra cannot be distinguished from equilibrium spectra; the extremely high values of 10¹³ cm^-3s in the easternmost part of the remnant (Fig. 3) are therefore better interpreted as being just larger than 10¹² cm^-3s. There is also a region of very low ionisation age (less than $3\times 10^{10}$ cm^-3s) stretching from East to West just above the centre of the remant. This region also has a very low emissivity (i.e. low electron density) and can be understood as a low density wake just behind and inside of the shocked ejecta.

The hot component has a more homogeneous distribution of ionisation age, centered around 10¹¹ cm^-3s, again consistent with the typical value found by Vink et al. (1996) but in that case integrated over much larger areas. We have now clearly resolved this component spatially.