Astron. Astrophys. 336, L17-L20 (1998)

3. Chemical composition

We performed a chemical analysis of AC Her on the basis of high-resolution (R50000) high signal-to-noise (S/N 130-280) optical spectra obtained with the Utrecht Echelle Spectrograph mounted on the 4.2m William Hershel Telescope at La Palma, Spain. The spectra were obtained on two consecutive nights (23-24/2/1994; or 0.37 in phase behind deep minimum) and consist of three settings covering the entire optical range from 360nm to 1 µm. From 660nm red-wards, the orders do not overlap and the spectral domain is not not completely covered.

Our method of analysis is described in detail in Van Winckel (1997) and Decin et al. (1998), and we refer to these articles for a full description. We used the CDROM-grid of LTE model atmospheres by Kurucz (1993) in combination with his abundance calculation programme WIDTH9. The parameters of an atmospheric model are determined by forcing the computed abundances of Fe to be independent of excitation potential (T_eff determination), reduced equivalent width (microturbulent velocity ) and the ionization stage (gravity). In order not to compromise the result of such analysis, one has to take care that only lines with well determined oscillator strengths (-values) are selected. The best list of such critically compiled -values useful for the temperature domain of AC Her was published by Lambert et al. (1996) and we restricted the use of Fe-lines to those values. This resulted in a list of 78 Fe I and 28 Fe II lines. Our best model has T_eff = 5500K; = 0.5; = 3.5 and an overall metallicity of -1.5. Typical errors are T = 250 K, = 0.5 and = 1 .

Once the model is defined we performed a complete chemical analysis, again restricting ourselves to lines with good oscillator strengths, clear profiles and equivalent widths in-between 5 and 140 mÅ. The results are listed in Table 2 for the individual ions. The individual lines and their atomic data can be obtained from the authors (HVW) upon request.

Table 2. Chemical analysis of AC Her. The solar abundances to compute the [el/H] ratios are taken from Grevesse (1989) except for the C,N and O abundances which are from Biémont et al. (1993). The dust condensation temperatures are from Wasson (1985) and computed using a solar abundance mix at a pressure of 10^-4 atm.

The abundance pattern is very well correlated with the dust condensation temperature (see Fig. 2). AC Her is clearly another RV Tauri star where the photospheric abundance patterns are determined not by internal nucleosynthesis and subsequent dredge-ups but by a fractionation process in which the atmosphere became depleted of chemical elements with a high condensation temperature. Although often referred to as C-rich, AC Her does not have a chemical signature of a carbon star. The [C/Fe] of +1.0 is indeed high but does not reflect the nucleosynthetic history of the object.

Fig. 2. The chemical deficiencies in the atmosphere of AC Her against the dust condensation temperature of the chemical element.

It is interesting to note that also the s-process elements have high dust condensation temperatures. The determination of the s-process elements in post-AGB stars is often used as a good tracer to investigate the 3rd dredge-up effectiveness. In objects where the depletion of the photosphere was efficient, the chemical pattern mimics a dredge-up pattern by increasing the C/Fe ratio. The s-process elements are also depleted; whether the 3rd dredge-up was indeed efficient or not is then not easy to deduce from the photospheric chemical content and the circumstellar dust chemistry and molecular envelope chemistry are much more reliable tracers in AC Her and RV Tauri stars in general.

The referee (G. Gonzalez) made us aware of a preprint (Giridhar et al., 1998) which includes also a chemical analysis of AC Her. Their abundance study give essentially the same [el/Fe] values for all chemical elements in common. The absolute abundance differences reflect the difference in atmospheric model parameters used.

© European Southern Observatory (ESO) 1998

Online publication: July 20, 1998
helpdesk.link@springer.de