Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 330, L17-L20 (1998)

Previous Section Next Section Title Page Table of Contents

1. Introduction

As part of the guaranteed time observing program presently carried out with instruments on board the Infrared Space Observatory (Kessler et al. 1996) we use the ISO/SWS (de Graauw et al. 1996) to study the infrared spectra of about 40 bright Asymptotic Giant Branch (AGB) stars (cf. Iben and Renzini 1983). The sample consists of stars with different chemical compositions (O-rich and C-rich) and a wide range of mass loss rates (optical Miras and Carbon stars, OH/IR stars, infrared Carbon stars). A few M-type supergiants are also included. Preliminary results of studies based on data obtained in this program have been reported by Justtanont et al. (1996, 1997), Cami et al. (1997) and Yamamura et al. (1997).

One of the surprising features in the AOT01 SWS spectra (resolution [FORMULA] 250) of AGB stars were unresolved emission lines at 13.87 µm, 14.97 µm and 16.18 µm in the spectra of several O-rich Mira variables. When it turned out that all stars showing these lines also exhibited the 13 µm dust feature in their spectra we asked for and obtained additional observing time to study a larger sample of stars with the 13 µm dust emission feature using the ISO/SWS at higher spectral resolution (AOT06 and Fabry-Perot observing modes).

At the time of writing this Letter our dataset consists of five stars showing in total five emission lines: lines at 13.87 and 16.18 µm observed in emission in all stars, one at 14.97 sometimes seen in absorption (in R Cas and W Hya) and two additional emission lines at 13.48 and 15.40 µm in EP Aqr. None of these lines have ever been seen prior to the launch of ISO. This region of the spectrum is inaccessible from the ground due to atmospheric absorption and the IRAS/LRS could not detect these lines because its resolution was too low.

Initially, since they were unresolved at the AOT01 resolution, we speculated that these lines might be due to ions associated with chromospheric activity in the stellar atmosphere, possibly related to the onset of mass loss (Justtanont et al. 1997). However, based on the recent high-resolution observations which show that the lines are just resolved and based on a more careful analysis we are now able to report the identification of these lines as due to ro-vibrational transitions between different vibrational bending modes of CO2. Recently, Ryde et al. (1997) have independently proposed that the emission of the 14.97 µm line in their spectrum of R CrA is due to CO2.

The correlation of the CO2 lines with the 13 µm dust feature is quite significant. This feature was discovered by Vardya et al. (1986) in the IRAS/LRS spectra of Mira variables with symmetric light curves. Onaka et al. (1989) proposed that the 13 µm dust feature was due to Al2 O3, a high-temperature condensate which is thought to be the first form of dust to condense in the outflow (Tielens 1990). However, optical properties of crystalline Al2 O3 do not produce emission at 13 µm with a comparable width, but at a shorter wavelength and with a narrower profile (Sloan et al. 1996). Another possible candidate suggested by Begemann et al. (1997) is magnesium silicate. Kozasa and Sogawa (1997) have proposed the composite of an aluminum oxide core coated by silicate as a possible explanation for the 13 µm emission. Sloan et al. (1996) studied a large number of these stars and found that stars with the 13 µm emission tend to have bluer colour, i.e. lower mass loss rates ([FORMULA] M [FORMULA] yr-1), and that about 50% of all O-rich AGB stars exhibit this emission feature. Many of the stars with 13 µm emission have been monitored for their light variations. The majority of them displays symmetric light curves indicative of a relatively weak pulsational shock wave.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1998

Online publication: January 8, 1998