SpringerLink
ForumSpringerAstron. Astrophys.
ForumWhats NewSearchOrders


Astron. Astrophys. 317, 178-184 (1997)

Previous Section Next Section Title Page Table of Contents

7. Discussion

Table 3 presents the mean non-LTE abundances of nitrogen and sulphur derived in Sect. 5 together with their mean non-LTE abundance corrections. In addition the last three columns give the non-LTE abundances of carbon and iron and the LTE abundance of silicon taken from Holweger & Stürenburg (1993).

[TABLE]

Table 3. Nitrogen and sulphur abundances for all program stars together with carbon, iron and silicon abundances from Holweger & Stürenburg (1993)

According to the accretion/diffusion model (see Sect. 1 ) nitrogen should behave like carbon. Both are volatile elements which remain in the gas phase during any gas-dust separation and diffusion causes both elements to settle. Since carbon is more abundant than nitrogen, saturation effects are stronger for the former.

Fig. 3 shows the anticorrelation of [C/Si] and [Si/H] (asterisks) found in Holweger & Stürenburg (1993) - this figure is comparable to their Fig. 2 without the [FORMULA] Boo stars. Nitrogen and sulphur are now included in this figure (plus signs and diamonds respectively). The [N/Si] ratio follows the anticorrelation of [C/Si]. For all stars of the sample the [S/Si] ratio is larger than the respective [C/Si] and [N/Si] ratio. Nevertheless an anticorrelation similar to that found for carbon and nitrogen is possible.

[FIGURE]Fig. 3. [C/Si] (asterisks), [N/Si] (plus signs) and [S/Si] (diamonds) versus [Si/H]; carbon and silicon abundances from Holweger & Stürenburg (1993). Dotted lines indicate solar abundance ratios. Carbon and nitrogen reveal a tight anticorrelation

[FIGURE]Fig. 4. [C/Fe] versus [Fe/H] for all program stars (Holweger & Stürenburg 1993). Dotted lines indicate solar abundance ratios

[FIGURE]Fig. 5. [N/Fe] versus [Fe/H] for all program stars (iron abundance from Holweger & Stürenburg 1993). Dotted lines indicate solar abundance ratios

[FIGURE]Fig. 6. [S/Fe] versus [Fe/H] for all program stars (iron abundance from Holweger & Stürenburg 1993). Dotted lines indicate solar abundance ratios

For sulphur, and possibly also for carbon and nitrogen, the anticorrelation found above strongly depends on two stars, HR 5798 and HR 4138. Both spectra are free of residual fringes and should yield reliable abundances. The typical uncertainties from measuring the equivalent width are 0.1 dex and after non-LTE corrections are applied, the scatter in the individual lines is in most cases smaller than 0.1 dex. The reality of the observed anticorrelations, therefore seems to be confirmed.

Table 3 shows that the silicon abundance for five of the 15 program stars could not be derived, but iron abundances exist for all of them. Since silicon and iron behave quite similarly in the gas/dust model, the [C/Fe] and [N/Fe] ratios should give trends similar to those seen in silicon and are thus available for the whole sample.

Figs. 4 and 5 show the respective ratios versus iron abundance. Since iron varies across only a small interval, -0.1 dex < [Fe/H] < 0.5 dex, the anticorrelation has a steeper gradient than in the case of silicon. But there is a clear correspondence between the two elements carbon and nitrogen.

The [S/Fe] ratio is larger than the solar one in most program stars. In view of the uncertainties in the abundance determination there is no clear correlation visible in the data (Fig. 6 ). However an anticorrelation as found above for [C/Fe] and [N/Fe] cannot be excluded.

An extensive search for other correlations reveals (apart from a non-systematic scatter) no obvious dependence on effective temperature or gravity, that is evolutionary effects are small or absent.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1997

helpdesk.link@springer.de