Astron. Astrophys. 334, 935-942 (1998)

1. Introduction

The dusty circumstellar disk around the nearby (19.3 pc, Crifo et al. 1997) A5 v star (HD 39060, HR 2020) is a uniquely prominent example of the Vega-type stars, which are characterized by infrared excess emission and which may represent a post-formation stage of planetesimal or planetary systems (for the observed high frequency of the Vega-phenomenon see, e.g., Habing et al. 1996). Our present knowledge of the system has been summarized by, e.g., Backman & Paresce (1993), Lagrange (1995) and Artymowicz (1997). The system is well observed over nearly the entire electromagnetic spectrum, from far ultraviolet (UV) to millimeter (mm) wavelengths. As yet, no spectral features have been detected that are sharp enough to directly provide proof that the elongated dust structure of size 1 (e.g. Kalas & Jewitt 1995) seen at the star is indeed, as is widely believed, an edge-on disk of grains in Keplerian rotation around the central star. As a first step, before actually attempting to map the rotation curve, one needs to identify and detect appropriate spectral tracers of the velocity field. Rotational transitions in molecules, thought to accompany the relatively cool grains, seem a natural option.

In the outer parts of the disk ( AU), considerable amounts of H₂ O and CO molecules from sublimated/sputtered icy grains could be expected. In addition, comets may release CO both far (many AU) from the star and in its vicinity (Beust et al. 1994). From UV absorption lines toward the star a column density of cm^-2 of CO gas at a temperature of 20 to 50 K has been inferred (Vidal-Madjar et al. 1994, Jolly et al. 1996). On the other hand, CO (1-0) observations by Savoldini & Galletta (1994) resulted in an upper limit on the line intensity, K km s^-1. Savoldini & Galletta interpreted their upper limit in terms of a high depletion of CO, as compared to interstellar gas (a view taken quite often in the literature: e.g. Vidal-Madjar et al. 1994, Lagrange et et al. 1995, Dent et al. 1995). However, these results could simply mean that the millimeter observations were suffering from beam dilution effects. These could be reduced by observing at a higher frequency, e.g. that of the (2-1) transition (by a factor of up to four using the same telescope). In addition, at temperatures above 20 K, the J =2 level of CO is preferably populated as compared to J =1, resulting in a line intrisically at least twice as strong. Consequently, CO (2-1) observations are potentially more sensitive and appear comparatively more powerful to probe any CO gas in the circumstellar disk of . In this paper, we report on such observations. CO (2-1) observations of were already reported by Dent et al. (1995), who interpreted their results in terms of thermodynamic equilibrium (LTE) at a single disk temperature. In the present study we are addressing the molecular excitation and line transfer in the disk in a more quantitative way. In addition to CO, we observed also in the CS (2-1) line in order to study to what extent the chemistry in the disk might differ from that in interstellar clouds.

Artymowicz (in preparation) has proposed that SiO is an abundant form of Si-bearing gas produced in evaporative collisions between the silicate dust grains. Frequent, energetic grain-grain collisions in the innermost part of the disk ( AU) could produce significant amounts of SiO gas, although this region appears largely dust depleted (Lagage & Pantin 1994; see also below, Fig. 5). Compared to CO, the significantly larger dipole moment of SiO offers the advantage of detecting relatively smaller molecular concentrations from millimeter observations. Taking an observational approach, we selected a vibrationally excited transition of SiO to trace any hot and very dense gas (v =2, J =2-1) and the low lying (v =0, 2-1) line to probe the more probable, cooler component at moderate densities.

First results were presented by Liseau & Artymowicz (1997), which were based on an analytical analysis using average disk quantities. In this paper, we offer a more detailed approach taking the disk structure fully into account. Our earlier findings are entirely confirmed by the more rigorous treatment of this paper.

© European Southern Observatory (ESO) 1998

Online publication: June 2, 1998

helpdesk.link@springer.de