4. A new model for the AFGL2688 nebula
We present in Fig. 8 a schematic view of our model for the AFGL2688 nebula, which is consistent with all the observations published to date. We interpret the set of knots and filaments which comprise the N and E blobs in Fig. 5 as a roughly circular loop of shock-excited emission. The bright knots at the N and E ends of the loop then probably represent limb brightening. The S and W blobs represent the limb brightened edges of another roughly circular loop. The loops represent the location at which the fast bipolar outflow is punching through one of the dense AGB mass loss shells seen in the I-band image of Latter et al. (1993). The loops are not complete, because of the highly clumped nature of the circumstellar environment which is seen very clearly in the E limb of the NE H2 loop. The spacing of the concentric AGB shells in the Latter et al. I-band image is sufficiently close ( 3") that within the radius of the outer, most prominent loop of H2 emission we might expect to see another loop. There is indeed some evidence for this, since the complicated structure of the H2 nebula is most easily explained if each cone in the biconical outflow contains two loops of shock excited H2, spaced by 3", as depicted in Fig. 8.
We interpret the H2 ansae as shock-excited knots of emission arising where the fast bipolar outflow collides with the surrounding AGB wind, perhaps analogous to the FLIERs seen in some PN. The ansae then indicate the distance to which the fast outflow has thus far penetrated. They are approximately 11" from the center of the nebula, in agreement with the result of Smith et al. (1990) who find that the high velocity component of the CO J=3-2 line extends to about 10" from the IR source. The spacing of the rings seen in the I-band image of Latter et al. (1993) is about 3". Since the H2 ansae appear at about 11" from the center of the nebula, we might expect to see three loops of shock-excited H2 emission, where the fast wind punches through each of the three innermost rings (which we take to delineate dense, spherical, AGB mass-loss shells). While we do see two loops in H2 (Fig. 8), a third loop is evidently not yet seen in H2, which could indicate that the density of the AGB shell at this distance from the central star is too low to form a shock at the interface of the two winds. However, some compression is bound to occur at the interface, so that this may form a good scattering medium. The two offset axes of the optical nebula probably result from the overlap of two or three loops of compressed material, which forms a bright lobe of scattered continuum radiation.
If we assume that the two loops are approximately circular, then we can determine the geometry of the nebula from their projected dimensions and positions. The bipolar axis of the nebula in this model lies at a PA 60 connecting the two H2 ansae, and is inclined to the plane of the sky at an angle 30 .
The asymmetry in PA of the nebula observed in the continuum, and the horns so prominent at I and J, are also explained by this model. In Fig. 8, the bipolar axis of the nebula has a 30 inclination to the plane of the sky which causes the two H2 loops to overlap from the observer's point of view. The N and S overlap regions delineate two axes, which run along the loops of H2, at PA 5 , with an E-W offset between them of 2" explaining the PA asymmetry and E-W offset between the symmetry axes of the N-S lobes. The models of Yusef-Zadeh, Morris & White (1984) and Latter et al. (1993) predict a pair of horns which, qualitatively, match those observed. In our model, a beam of radiation escapes from the star through the biconical regions carved out by the high velocity flow. These regions are filled with a low density, dusty medium. Limb brightening of the beam of radiation, scattered off the dust in the cavities, will generate horns, without the need for biconical cavities extending to large distances from the central star, which are needed by the "standard model", and which do not appear to be consistent with mm-wave CO maps.
One further important observation that must be explained is the apparent absence of any sign of scattered radiation from the E and W of the nebula, where we claim there are large openings into the bipolar cavities. One possibility is that, given that we have stated that the circumstellar environment is highly inhomogeneous, there is simply an absence of scattering material in these directions. A second, more likely, possibility is that the nebula is being asymmetrically illuminated. We have shown, by taking deeper images than have been published before, at a variety of wavelengths, that there is at least some scattered radiation from these directions. There is some faint extension at J along the entire western edge of the prominent N-S lobes. At K, even using a narrowband filter, this extension becomes more discernible (Fig. 4b), and there appears extension to the eastern edge as well as the western edge of the lobes. At nbL, at which wavelength some thermal emission from the nebula becomes important, a box-like structure of the shape required to fit our model is evident. In the mid-IR, where there is only thermal emission from dust, and no scattered continuum, the whole nebula is rectangular. As we progress to longer wavelengths, and we pass from the regime of scattered stellar continuum to thermal nebular dust emission, the nebula appears to rotate E of N. This spatial distribution of the thermal IR emission suggests that in fact there is material all around the large bipolar flow cavities, but it is not isotropically illuminated. This suggestion is further strengthened by the large scale of the shock-excited H2 emission loops, which reveal the presence of material all around the flow cavities, albeit with some gaps due to the inhomogeneity of the circumstellar medium. The nebula could be asymmetrically illuminated if dense dusty blobs located close to the openings of the inner, dense torus, partially occlude the openings, shadowing the outer E-W parts of the nebula (Fig. 9). The concept embodied in Fig. 9 implies a significant degree of nebular asymmetry in the inner part of the nebula, which is indeed seen, both in scattered radiation at I and J, and in thermal emission in the mid-IR. In order to explain the observations exactly, the shadowing material would have to be to the NW of the opening on the SW side of the torus, and to the SE of the opening on the NE side, which implies a troublesome symmetry that we regard as a weakness in this model. This shadowing aspect of the model will be further discussed in Sect. 6. The dust temperature map (Fig. 6), as we mentioned earlier, can be interpreted as consistent with our new conceptual model.
© European Southern Observatory (ESO) 1997
Online publication: March 24, 1998