Astron. Astrophys. 324, 185-195 (1997)

1. Introduction

In homonuclear molecules, such as H₂, there is no change of dipole moment during rotation or vibration, and such molecules are therefore unable to absorb (or emit) dipole radiation. However, during transient interactions of such non-polar molecules, short-lived "super-molecular" species, as for example H₂ -H₂, are formed, a temporary dipole moment is induced, and a relatively weak dipole absorption becomes possible.

Herzberg (1952) was the first to bring collision induced absorption (CIA) into an astrophysical context, and suggested this mechanism as being responsible for some hitherto unexplained absorption bands in the spectrum of Uranus. By comparing laboratory H₂ spectra with spectra of Uranus and interpreting the bands as due to CIA, Herzberg derived the first limits on the pressure of the thick Uranian cloud top, and produced one of the first pieces of spectroscopic evidence that hydrogen was the primary constituent in Uranus. Today CIA processes are part of the standard modelling of the low-temperature, dense gases in the atmosphere of the giant planets and of Titan and Venus.

Linsky (1969) developed a simplified analytic description of a few bands of H₂ -H₂ and H He CIA as a function of frequency and temperature, and was the first to point out that the pseudo-continuous character of the CIA could cause it to have a large effect also in stellar atmospheres. Here it will block the energy which would otherwise have escaped in the transparent spectral regions between the absorption lines. This will be particularly true when polar absorbers (e.g., H₂ O, H , TiO, etc) are depleted, as for example in cool stars of low metallicity.

In the same year, Tsuji (1969) considered molecular opacities in cool stellar atmospheres, including various sources of continuous opacity. Independently, using his own estimation of CIA intensity, he reached the same conclusion that CIA is an important source of opacity at high pressures, which can not be neglected.

In the following decade CIA in stellar atmospheres attracted more attention. Shipman (1977) computed the first white dwarf model atmosphere composed of pure H₂, which included Linsky's data. He found that at a temperature of 4000 K CIA contributes essentially all the opacity at the wavelengths where the flux is emitted. At higher temperatures other opacity sources were dominant. Also Mould & Liebert (1978) have included CIA due to H₂ -H₂ and H₂ -He pairs, to compute new white dwarf model atmospheres, again using Linsky's data. They reiterated the importance of CIA, but no specific results were shown.

Palla (1985) pointed out, for the first time, that, depending upon gas density, CIA may be an essential, and often even the dominant source of opacity, in primordial protostars at temperatures between 1000 and 7000 K. In Palla's work, atmospheres composed of hydrogen and helium were considered and Linsky's (1969) data were again input.

One year later, Stahler et al. (1986), using Linsky's data, confirmed Palla's findings: the source of opacity due to collision induced absorption cannot be neglected.

Based on the then existing quantum mechanical CIA models due to Borysow and collaborators (Borysow 1994 and references therein; see also summary below), Lenzuni et al. (1991) and Saumon et al. (1994) studied the impact of CIA on hypothetical zero-metallicity, primordial proto-stars and zero-metallicity brown dwarfs, and found very large effects.

We have now extended these computations of CIA due to H₂ -H₂ and H₂ -He into the higher temperature regime appropriate for stellar atmospheres, and studied their impact on realistic atmospheric models of existing stars. In Sect. 2 we summarize the existing quantum mechanical CIA data and in Sect. 3 we present the result of our new calculations. In Sect. 4 we compare our results with previous data, where available, and give some warnings about the limitations of our data and of previous data. In Sect. 5 we describe the impact of our data on stellar model atmospheres, and quantify in which region of the HR diagram (and metallicity) the collision induced absorption is important for the stellar atmospheric structure and for the analysis of photometric and spectroscopic data.

Our data are applicable to stars of arbitrary chemical composition, but in this paper we will restrict the discussion to oxygen-rich stars. We cover the region in fundamental stellar parameters which represents brown dwarfs, all types of K and M dwarfs, as well as the metal-poor sub-giants and giants that could be found in the Galactic halo, in globular clusters, and in metal-poor galaxies such as the Fornax dSph and the SMC. In later papers we will discuss carbon dwarfs and white dwarfs, which will require additional input data. Our CIA data do not include H H pairs, which could also contribute significantly to the opacity, and we have restricted the inclusion of bound-bound molecular line transitions to those molecules for which complete, high-quality quantum mechanical computations exist. In particular, it may be desirable to include additional diatomic hydrides in the discussion. In future work we hope to be able to extend the data toward a higher degree of completeness. Also dust formation may be of relevance for the very coolest objects.

© European Southern Observatory (ESO) 1997

Online publication: May 26, 1998

helpdesk.link@springer.de