Astron. Astrophys. 347, 194-202 (1999)

1. Introduction

HD101584 belongs to a class of peculiar stars at moderate to high galactic latitudes with supergiant-like spectra (Waters et al. 1993). They have been intensively studied during the last decade, and there is a growing belief that most of them are low-mass post-AGB objects (Parthasarathy & Pottasch 1986; Trams et al. 1990; van der Veen et al. 1993; Bakker et al. 1996b). HD101584, was originally classified as an F-type supergiant (Humphreys 1976), but more detailed observations, of high-excitation absorption lines in the red part of the spectrum and photometry, revealed a late B-type star of higher surface gravity (Bakker et al. 1996b). The F-type spectrum probably originates in a dense stellar wind of low ionization degree. Bakker et al. (1996a) infer a period from photometry data, and possibly also from Doppler velocities of high-exitation absorption lines. According to the authors, the most likely explanation is the presence of a low-mass companion in a close, eccentric orbit (separation10 ), but this must be regarded as only tentative. Optical emission and low-excitation absorption lines suggest the presence of a circumsystem disk (size100 ) that is seen nearly edge-on (Bakker et al. 1996b). The CO first vibrational overtone bands that are seen in emission may also originate in such a disk if it is dense enough (Oudmaijer et al. 1995).

The P Cygni-profile components of the Balmer lines show that the primary star is still losing mass at a significant rate, but the estimate is uncertain, (Trams et al. 1990; Bakker et al. 1996b). The mass loss occurs in the form of a wind with a maximum outflow velocity of 100. The presence of additional absorption dips may indicate a variable mass loss. The huge far-infrared excess emission may be dominated by dust emission from an escaping circumstellar AGB-envelope. Assuming this Bakker et al. (1996b) estimate a total dust mass of 0.02, and hence a total circumstellar mass probably in excess of 2.

Loup et al. (1990) and Trams et al. (1990) detected a remarkable CO(J=) line profile from HD101584, and van der Veen et al. (1993) detected the CO(J=) line. The CO emission probably comes from a number of separate regions, and part of the molecular material have expansion velocities at least as high as 130, i.e., considerably higher than in normal AGB-winds. A much narrower, double-peaked OH 1667 MHz maser line (maximum expansion velocity of 40) was detected by te Lintel Hekkert et al. (1992). The integrated OH emission is centered on the star (to within 03), but the data show that the gas velocity increases systematically along PA-60^o with the most blueshifted emission 2" to the SE and the most redshifted emission at 2" to the NW. There is no OH emission within 1" from the star. The presence of OH emission and a 10m feature (Bakker et al. 1996b) suggest that the circumstellar material is O-rich (i.e., C/O1).

© European Southern Observatory (ESO) 1999

Online publication: June 18, 1999
helpdesk.link@springer.de