4. Discussion and conclusions
Out of the original sample of 8 LMC Ofpe/WN9 stars originally investigated by Nota et al. (1996b), and Pasquali et al. (1997) and listed in Table 7, five have been found to have nebular lines in their ground based optical spectra (see also Walborn, 1982; Bohannan & Walborn 1989). These five stars were the subject of subsequent detailed studies aimed at establishing whether the nebular lines were indication of the presence of an associated ejected nebula.
Table 7. The sample of LMC Ofpe/WN9 from Pasquali et al. (1997)
As already mentioned, the nebula around BE 381 has been found to be kinematically associated, but not composed of ejected material, and therefore has most likely been formed by interstellar medium swept up by the fast O stellar wind. Very little can yet be said of either R99 or HDE269927c. Attempts were made to resolve a circumstellar nebula around R99 from the ground with coronography, and with HST (Schulte-Ladbeck et al. 1998), and have failed so far. HDE269927c has also been investigated from the ground with coronography without a detection, and HST observations are still pending (Schulte-Ladbeck et al. 1998). We can safely conclude that out of the original sample, only S119 and S 61 show convincing evidence of an associated circumstellar nebula composed of ejected material.
From the point of view of the stellar properties, S119 and S 61 are quite similar, in terms of terminal wind velocity, luminosity and mass loss rate. In Table 8, extracted from Pasquali et al. (1997), we list their fundamental parameters, derived from HST UV and optical spectroscopic observations: with the possible exception of the temperature, which places the two stars roughly at the two extremes of the Ofpe/WN9 sequence, all parameters are very similar. Neither star is known to be variable: however, very few observations have been obtained for these two stars in recent years, and any possible variability could have been missed.
Table 8. Optical and Near-IR photometry from Nota et al. (1996). Stellar properties from Pasquali et al. (1997)
The two associated nebulae are also very similar, in terms of morphology, size, and kinematical characteristics. Their morphology is an elliptical shell, with well defined boundaries. A very bright lobe can be distinguished in both nebulae, at PA = 45o for S119 (Nota et al. 1994), and PA = 0o for S 61. They have roughly the same nebular spatial extension: for S119, corresponding to a linear scale of 2 2 pc, in diameter for S 61 (1.8 pc). The two nebulae expand with velocities which, within the errors, are identical: 25 km s-1 for S119, 28 km s-1 for S 61. Hence, the dynamical ages are also very similar: 5 104 yrs for S119, 3 104 yrs for S 61. In terms of mass, S 61 (4 ) is a factor of two more massive than S119 (1.7 , Nota et al. 1994), mainly due to a difference in density. In fact, Smith et al. (1998) find for S119 an upper limit to the Te of 6800 K, and a corresponding density of ne = 680, in correspondence to the brightest lobe of the nebula. In comparison, we find for S 61 a value ne = 400 and Te = 6120 K at the pointing position N. However, it is necessary to point out that when we calculate the mass of ionized gas contained in the nebula, we assume an average density. Significant density variations have been noticed in the S119 nebula (Nota et al. 1996b), possibly indicating that the nebula is clumpy, and the determination of the mass might be affected by a high uncertainty. In addition, in order to determine the density, we use the [SII] 6717/6731 ratio in a regime where a small variation in the ratio (10%) can produce a large difference in the determination of the density (a factor two) and, therefore, of the mass.
We can conclude that within the limitations associated with our measurements, S119 and S 61 are in a very similar phase of their evolutionary history. The presence of an ejected nebula also indicates that they have undergone, some 104 yrs ago, a LBV type "giant outburst" where they have ejected a significant fraction of their outer layers in the surrounding medium. We can confidently state this was a true LBV outburst because in both cases the ejected nebulae are also very similar to typical LBV nebulae, such as AG Carinae, in terms of morphology, size, dynamical ages and masses.
Moreover, the similarity extends to the chemical composition, and we do find for these two Ofpe/WN9 ejected nebulae the same abundance anomaly recently found for a number of galactic and LMC LBVs (eg. AG Carinae; Smith et al. 1997). The chemical composition of the nebula around S119 had been already derived by Smith et al. (1998), who used HST/FOS spectra of the brightest nebular region to derive nebular parameters and element abundances. They were not able to derive a secure value for the electron temperature, but with an upper limit of 6800 K they derived N/O = 1.41-2.45, N/H = 8.08-8.41, S/H = 6.18-6.52. They compared the nebular abundances obtained with the expected surface abundances of LBVs and concluded that, if LBV atmospheres consist of CNO-processed material, the event which generated the nebula had taken place before, or at the very start, of the LBV phase. Comparison of these nebular abundances with the abundances, for example, of SN1987A, showed remarkable similarity. The inner ring of SN1987A is thought to be composed of RSG wind material (Fransson et al. 1989; Panagia et al. 1997). This finding, together with other considerations on the nebular expansion velocities, and on the dust content (Waters 1998), suggest that LBV nebulae were once the CN-processed convective envelope of a RSG. For S 61, we also find high N enrichment, comparable to S119. We conclude that such enrichment is also consistent with material which has been CN-processed only, as observed in PN. However, a detailed abundance analysis will be necessary to put on firm quantitative grounds our findings and to establish whether there is any O depletion. The slow expansion velocity of the nebula would also support the RSG origin.
Our new findings on the element abundances for S 61 have a threefold impact:
1) they complement the morphological and kinematical information in establishing a close similarity between the two nebulae around S119 and S 61. This similary is especially interesting considering that these are the only two confirmed associated nebulae around Ofpe/WN9 stars discovered so far, and it implies that not only these two stars have undergone a LBV type outburst, but the nebular composition confirms that this event has occurred approximately at the same time in their evolution;
2) together with the morphological and kinematical information, they add one more point in support to the scenario that Ofpe/WN9 stars are closely related to LBVs. In fact, the measured N/S and N/H ratios for the S 61 nebula (Table 5) are in very good agreement with results already obtained on LBV nebulae such as AG Carinae (Smith et al. 1997) and R127 (Smith et al. 1998), also reported in Table 5.
3) They also support the suggestion by Smith et al. (1998) that these nebulae most likely were the convective envelope of a RSG, which has been gently shed prior to the classical LBV phase.
In order to understand whether this scenario is correct, more work is needed both on the observational and theoretical aspects of the problem. Observationally, it is necessary to perform accurate abundance analyses of all LBV nebulae, and therefore establish accurately the statistical significance of these findings. Theoretically, a consistent evolutionary scenario is needed which explains how these superluminous stars can experience a RSG phase when no such luminous RSG counterpart is actually observed.
© European Southern Observatory (ESO) 1999
Online publication: March 1, 1999