Astron. Astrophys. 352, 287-296 (1999)

4. Abundances

4.1. Metals, dust, and

The metal abundances in component A are given in Table 2. Since only a total hydrogen column density for components A and B could be measured, the abundances may be too low by dex. In presence of the relatively large uncertainties this effect seems negligible.

As can be seen in Table 2 the depletion factors of the elements vary between -0.3 and -1.7. This moderate depletion indicates a low abundance of dust along the line of sight.

The total amount of H₂ seen on this line of sight is cm^-2, which gives an abundance relative to H I of dex. With this low value the total amount of hydrogen, cm^-2, is hardly larger than the value given above for just H I .

The extinction towards BD +39 3226 is given by Dworetsky et al. (1982) as , which again indicates a low dust abundance. In addition, the fractional abundance of molecular hydrogen has a low value of . The gas to color excess ratio for our line of sight of  cm^-2 mag^-1 is slightly lower than the typical value for the galactic intercloud gas of  cm^-2 mag^-1 given by Bohlin et al. (1978). This deviation, though, can be explained by uncertainties in the value and should not be overinterpretated. We can conclude that most of the absorption seen on this line of sight arises from intercloud gas with only little contribution from the clouds having  K.

Savage et al. (1977) have generated a plot of vs. for their survey of 70 stars. For lines of sight with , they find only very small fractions of H₂ down to values of -6. This is in line with the tight connection between the existence of dust grains and of molecules forming on their surfaces. Our low value fits well into this group of stars of low fractional abundance.

In their plot, Savage et al. (1977) show theoretical curves from equilibrium models of Black for different hydrogen densities, radiation densities, and molecule formation rates. According to this, our value is described by Black's "Model 2", which assumes a density of 100 cm^-3 at a temperature of about 100 K, in accordance to the temperature derived above. The radiation density at 1000 Å has then a value of  erg cm^{- 3} Å^-1 if the molecule formation rate is  cm³ s^-1.

4.2. Deuterium

From the HI and DI column densities we obtain a deuterium abundance on this line of sight of . We also tried to identify absorption lines of HD, but without success. Taking the H₂ column density of only cm^-2 one would expect HD column densities as low as about cm^-2, which is definitely below the detection limit of cm^-2 for the spectra available. It can be ruled out that any significant amount of deuterium is hidden in HD. Deuterium depletion by interactions with dust grains is unlikely in a warm ISM component with low gas-to-dust ratio.

The abundance is a value for the whole line of sight, possibly an average over more than one cloud with different ratios. The uncertainties are rather large, but the value is reliable within its limitations. There was no need for assumptions about the velocity structure of the absorption (as it is necessary for line modelling) since the hydrogen column density determination is mainly based on fully damped absorption while the deuterium lines lie in the linear part of the curve of growth. In both cases the curve of growth is insensitive to effects arising from clouds of different b-values and column densities. Furthermore, a model of the stellar spectrum was used only for Lyman fitting. The uncertainty arising from modelling is negligible because the centres of the stellar and interstellar absorption lines are well separated due to the high radial velocity of the star.

Varying scattered light may affect the measured equivalent widths, as mentioned in Sect. 2. However, we do not find evidence for systematic background substraction errors beyond the 5% level in all analyzed H and D lines and the scatter of the equivalent widths is not larger than expected for the estimated errors. The effect of systematic errors would be rather small. For example, if all H and D equivalent widths were larger by 10% than measured due to wrong background substraction, the D I column density also would rise by % while the H I column density resulting from curve of growth analysis and Lyman fitting would change even less. Thus the ratio would be increased by less than 10% and would remain well within the given range.

© European Southern Observatory (ESO) 1999

Online publication: November 23, 1999
helpdesk.link@springer.de