4. Conclusions and discussion
In the present study we find evidence for a correlation between the orientation of the long axes of nebulae ejected by massive stars and the global structure of the galaxy to which they belong. The Galaxy and the LMC offer complementary views of the same phenomenon. In the Galaxy, the axes are essentially found aligned with the galactic plane, and then also with the galactic magnetic field. The effect is statistically significant. In the LMC, the axes closely follow the magnetic field lines, supporting the correlation between nebular orientations and galactic structure. The latter result suggests a major role of the magnetic field in the alignment effect as opposed to a purely dynamical mechanism, although the magnetic structure in the LMC is probably intimately related to the dynamical structure as in large spiral galaxies (Klein et al. 1993).
The present sample of LBV-type nebulae is unfortunately small, and a confirmation of the effect will await new discoveries of similar objects, especially in external galaxies. Misaligned nebulae could be important for unambiguously identifying the dominant role of the magnetic field. If it ultimately appears that the misalignment with the galactic plane is real and not due to errors in the interpretation of morphological features, these nebulae could align with more local structures in the interstellar magnetic field, like the large-scale magnetic field loops which extend at high galactic latitudes (e.g. Mathewson & Ford 1970). More detailed studies of these objects are therefore needed to determine their true morphology 4, as well as the direction of the interstellar magnetic field in their environs.
At the origin of the observed alignment, one could invoke the direct effect of the ambient interstellar magnetic field on the nebular morphology. But, as discussed by several authors in the context of planetary nebulae (Heiligman 1980, Phillips 1997), if the ambient magnetic field can induce small modifications to the morphology, it is not able to shape the outflowing material. This is a fortiori true for the more massive nebulae around LBV and WR stars. More likely, the nebular axis is correlated to the symmetry (rotation) axis of the star and/or to the orientation of a preexisting disk (which could be magnetized). These symmetry axes are themselves related to the direction of the ambient magnetic field from the star formation process. Such a scenario, already suggested by Aitken et al. (1995) for Car, is supported by several arguments. Indeed, Schulte-Ladbeck et al. (1993, 1994) and Clampin et al. (1995) have shown from polarization studies that the morphological axes of the nebulae surrounding the LBVs R127, AG Car and HR Car are related to the symmetry axes of the stellar winds, and consequently most probably to those of the stars. From the theoretical point of view, a relic protostellar disk or an equatorially compressed outflow due to stellar rotation, both with relative importance which may differ from one object to another, are welcome to explain the preexisting density contrast needed to explain the shapes of LBV nebulae (Nota et al. 1995). Such a density contrast is expected to be perpendicular to the nebular long axis. Finally, disks around young stars are known to be perpendicular to the interstellar magnetic field, the outflows themselves being parallel. This is apparently a key feature of the star formation process which is reproduced by a variety of models (e.g. Pudritz & Norman 1986, Strom et al. 1988, Appenzeller 1989). In this context, it is remarkable that the large-scale distribution of molecular gas around Car (as shown in Cox 1995) appears to have roughly the same orientation as the homunculus.
Whatever the mechanism which causes it, it is quite striking that an alignment with the galactic plane or with the galactic magnetic field has been reported for several types of objects, from planetary nebulae 5 to supernova remnants (e.g. Gaensler 1998, and references therein), although it is not clear whether all nebula sub-classes do show an alignment effect, and if it also prevails out of the galactic plane (cf. Grinin & Zvereva 1968, Gaensler 1998). The alignment with galactic structures, and more particularly with galactic magnetic fields, could therefore be a generic property of nebulae ejected by evolved stars. Note this does not mean that the same mechanism is at work in all cases. For example, the morphology of supernovae remnants may follow the cavity created by previous ejecta (and namely the LBV/WR stellar winds and nebulae) as suggested by Bisnovatyi-Kogan et al. (1990). In this view, it is interesting to note that the radio remnant of SN1987A seems to follow the morphology of the optical nebula (Gaensler at al. 1997). Also, mechanisms of star formation may substantially differ for massive and low-mass stars (e.g. Shepherd & Churchwell 1996), and more particularly the relative influence of the interstellar magnetic field in the early stages of formation (Mestel & Subramanian 1991).
As a whole, the dynamical effect of such collective non-isotropic gas and dust ejections along preferred directions - and especially from the most massive stars- could be important for a detailed understanding of the interstellar medium evolution, including in starbursts. And one may further speculate that it might provide a natural explanation to some viewing angle effects discussed in the context of stellar ejecta models proposed for active galactic nuclei and broad absorption line quasars.
© European Southern Observatory (ESO) 1999
Online publication: March 10, 1999