Stars at the tip of the AGB lose mass copiously in the form of slow and dense stellar winds. Dust in the winds reddens the central starlight, and in many cases renders the star optically very faint or invisible (e.g. IRC+10216, OH26.5+0.6). The energy distribution of such extreme AGB stars typically peaks in the mid-IR, by virtue of the thermal reradiation by the optically thick circumstellar dust shells. In contrast, planetary nebulae (PN), which are the evolutionary descendants of the AGB stars, show little circumstellar reddening, and have spectral energy distributions which peak in the mid- or far-IR (e.g. NGC7027).
By implication, objects in transition between the AGB star and PN phases of their evolution should simultaneously exhibit rising stellar effective temperatures, and dust excesses which move to longer wavelengths as the dust shells coast outwards and cool, along with decreasing circumstellar reddening. Such transition objects are often referred to as protoplanetary nebulae, since it is assumed that they will evolve into PN. However, the speed with which such objects evolve is not yet observationally established, hence it is not actually certain that all such objects will become PN. We therefore refer to them as post-AGB stars.
One of the first such transition objects to be discovered was AFGL2688, noted by Ney et al. (1975) because of its extreme brightness in the mid-IR and very cool colours ( 200K), and discovered by them to be associated with a spectacular bipolar optical reflection nebula, consisting of two tulip shaped lobes on an axis at a position angle of about 15 . A pair of horns is seen projecting out from each lobe away from the center of the nebula. It had in fact previously been catalogued as a probable galaxy (IV Zw 67), and is known as the Cygnus Egg Nebula because of its egg-like appearance on the Palomar sky survey plates. Distance estimates to AFGL2688 have generally been in the range 1.0 to 1.5kpc (Crampton, Cowley & Humphreys 1975; Cohen & Kuhi 1977); we will adopt in this work a value of 1.2kpc, intermediate in this range. The central star, obscured by its circumstellar dust shell but observable in the reflection nebulosity, has F5Ia spectral type (Crampton et al. 1975) consistent with a post-AGB status. Anomalous bands due to C3 and C2 indicate that AFGL2688 is C-rich, presumably descended from a C-star (Crampton et al. 1975). Single dish molecular line observations of AFGL2688 in the 12 CO J=1-0 line show the classical parabolic line profile that suggests an optically thick, spherical shell of molecular gas coasting outwards at a constant velocity, presumably produced by the progenitor AGB star (Knapp & Morris 1985). Truong-Bach et al. (1990) mapped AFGL2688 in the CO J=2-1 and J=1-0 lines, and showed that the AGB wind remnant is spherical. Observations in a number of mm-wave molecular lines by Young et al. (1992) suggested to them that AFGL2688 has at least three velocity components: a dominant, slow, 23km/sec expansion velocity wind due to the AGB progenitor, a `medium velocity' 45km/sec wind, and a high velocity component with velocities up to 100km/sec, with the highest velocity component closest to the star, and the lowest furthest from the star. Evidence for interacting winds preceded these observations: Thronson (1982) observed shock-excited molecular hydrogen lines in the near-IR spectrum of AFGL2688, taken by him to be produced by the interaction of winds with different velocities. This interpretation was strengthened by the observation by Beckwith, Beck & Gatley (1984) that the molecular hydrogen emission was confined to the bipolar lobes. These authors argued that the shocks are generated where a fast bipolar outflow from the warm central star interacts with a slower wind from the progenitor AGB star. Velocity resolved images in the molecular hydrogen S(1) 1-0 line at 2.122 m by Smith et al. (1990) show a most remarkable and unexpected structure: the molecular hydrogen emission is seen in four blobs spaced equally about the central object, two lying in the optical bipolar lobes and two orthogonal to them. The N and E blobs are tipped towards us, and each have LSR velocities of about -60km/sec, whilst the S and W lobes are pointed away from us and each have LSR velocities of about -30km/sec. Latter et al. (1993) interpreted their narrowband images of H2 2.122 m emission, which also show four blobs, in terms of a model in which the N-S blobs arise from shock excitation in the bipolar lobes where the fast wind punches through the AGB wind remnant, and the E-W blobs arise from shock excitation where the fast wind crashes into a rotating torus observed in HCN J=1-0 line emission by Bieging & Rieu (1988). However, if the H2 velocities of the E-W blobs are actually rotational, then the rotation rate is unprecedentedly large for an AGB envelope.
In an attempt to investigate the morphology of the inner, most active regions of AFGL2688, we have taken high angular resolution images at infrared wavelengths in the continuum from 1.2 to 19 m, and narrowband images of the 2.12 m H2 line emission. We present our images of AFGL2688 in the next section. In Sect. 3 we discuss the standard model of the nebula, and the deficiencies which lead us to propose a revised model of the source. We describe this alternative model for the nebula in Sect. 4, and show how it better explains observations of AFGL2688. In Sect. 5, we develop a detailed radiative transfer model for the spectral energy distribution (SED), quantitatively compare the standard model and our new model, and determine the mass loss history of the star in the later stages of its AGB evolution. In Sect. 6, we discuss the results and their implications for the nature of the progenitor of AFGL2688 and for other post-AGB stars.
© European Southern Observatory (ESO) 1997
Online publication: March 24, 1998