Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 336, 648-653 (1998)

Previous Section Next Section Title Page Table of Contents

4. Discussion

The bipolar structure of the Red Rectangle is located very close to the core of the system. This structure suggests the existence of an equatorial disk and is consistent with the current model of a carbon-rich post-AGB star evolving into the planetary nebula phase (Van Winckel et al. 1995; Roddier et al. 1995; Lopez et al. 1995, 1997). Our results reveal clear similarities between images taken at 2.2 µm and at 3.3 µm UIR on the one hand, and between L' and M images on the other hand. In addition, our K image is similar to the H image of Roddier et al. (1995). There is also good correlation between our 3.3 µm UIR image and UIR mid-infrared images of Hora et al. (1996). The north-south extension of the nebula is largest at shorter wavelengths (H, K) and smallest at longer wavelengths (L', M & N). The "spikes" which are so striking at H become progressively less apparent as the wavelength is increased. However, all infrared images of the nebula are north-south elongated. This could be an indication that scattering is more efficient in the polar direction. We assume that at shorter wavelengths the overall feature is dominated by scattering whereas long wavelength images display both scattering and dust emission. The scattering begins to be quite efficient in the near infrared when the grain size distribution function contains particles greater than 1 µm while larger particles are needed to explain the scattering efficiency observed at longer wavelengths. The main issue of the physical interpretation of these results is what causes the apparent X-shaped morphology or at least the rectangular shape present in infrared images of the nebula. Is it due to the spatial distribution of the dust density and/or of the size distribution of the grains or is it due to other physical conditions? We suggest that the observed X-shaped morphology is produced by the scattering of dust grains in a compact equatorial thick disk with evacuated polar regions. Large grains are present not only in the equatorial plane (as proposed by Cohen et al. 1975; Jura et al. 1995,1997 and by Lopez et al. 1995, 1997) but also in the lobes. The increase of the opening angle with wavelength seen in our images could indicate that large grains are mostly located at the outer edges of the cavity walls and supports the scenario proposed by Jura et al. (1997). In fact, according to Jura et al. (1997) large carbon particles up to 0.02 cm in diameter are confined in a long-lived disk around the star. In such an orbiting disk, grain-grain collisions create and eject a whole range of possible species including small and large particles. The X-shaped structure, observed close to the core of the system, could be produced by the dust that is pushed outwards by expanding gas. Material ejected in the outflow or formed by interaction of the outflow with the dust may condense along the cone walls giving the observed lobes (Roddier et al. 1995).

Our 3.3 µm UIR image is elongated along the cone walls of the nebula while Bregman et al. (1993), who performed imaging at 3.3 µm and at 11.2 µm, found the 3.3 µm image to be centrally peaked. However, the lower resolution of their 3.3 µm image did not allow them to detect the morphological features present in our image. UIR images obtained by Hora et al. (1996) at 8.0 µm, 8.6 µm and 11.2 µm display a shape similar to our 3.3 µm image and are more elongated than the mid-infrared continuum emission. According to Hora et al. (1996), the mid-infrared emission appears to be trace material in the bipolar outflow regions, rather than in an equatorial density enhancement.

The hypothesis of a compact equatorial disk with a polar free cavity which collimates a mass outflow driven by radiation is consistent with our 3.3 µm image. Hence, if the dust grains causing the scattering are different from the UIR particles, then UIR images may will be the result of the excitation of these particles by the ultraviolet radiation of the central star.

No binarity in the system is seen in our data. It is more likely that the inner structure of the bipolar lobes of the nebula has been misinterpreted by visual observers as a double star (see Roddier et al. 1995). The hypothesis that the bipolar nebula was created from a binary system by tidal interaction between the secondary and the expanding red giant primary was proposed by Morris (1981) and Roddier et al. (1995) while Yusef-Zadeh et al. (1984) and Lopez et al. (1995, 1997) proposed a model where a single star is surrounded by an axisymmetric dust shell. Models of the Red Rectangle obtained by Lopez et al. (1997) which use radiative transfer simulations, reproduce the spectral energy distribution and both visible (on scales of some arcsec) and 2.2 µm images of the nebula quite well. However Lopez et al. (1997) have difficulty in fitting both the visible and infrared shapes with the same set of model parameters. The main result of their study is the presence of large particles in the circumstellar environment of the nebula, with sizes greater than or equal to 1.2 µm.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1998

Online publication: July 20, 1998