Astron. Astrophys. 356, 146-156 (2000)

6. Abundance analysis for HR 1094

The spectrum for HR 1094 is dominated by spectral lines from the iron-group elements. However, the REE are also prominent, since they have rich line spectra and are relatively abundant. Spectral lines from the iron-group elements, preferably iron lines in the red spectral region, are used to determine the metallicity since these lines are less affected by blending. However, in the presence of a magnetic field the Zeeman broadening is more pronounced in the red spectral region and spectral lines little affected by the field should be used. A metallicity analysis was performed by investigating appropriate iron-group lines, using a solar abundance atmospheric model. This preliminary synthetic spectrum analysis yielded HR 1094 to have a metal abundance five times greater than the sun.

Using an ATLAS9 model based on this enhanced metallicity, the spectral lines for the iron-group elements were analyzed for the entire spectral region with results presented in Table 2 and the corresponding line list in Appendix A. The appropriateness of useful lines is difficult to assess since most lines are blended or affected by either IS, hfs or Zeeman broadening, but their credibility can be estimated by investigating the line profile for possible blends. For determining the iron-abundance, FeII 4508 is of special interest since it is less influenced by certain broadening mechanisms and its oscillator strength has been derived experimentally. From FeII 4508 the iron abundance, log , is determined to be 8.05. This is a smaller value compared with results obtained by using other FeII lines, which emphasises the uncertainties the broadening mechanisms introduced into the abundance analysis. Analysis of transitions from high and low states give different results which might be due to the difficulties in deriving oscillator strengths for highly excited states, but can also be a result of non-LTE effects in the stellar photosphere or a non-uniform temperature distribution. We determined 8.46 to be an average value of log , which agrees with the value determined by Sadakane within 0.05 dex, yet caution that the interpretation of the iron abundance, like that of all other elements, is not straightforward.

Table 2. Result of the abundance analysis compared with solar values and the earlier analysis by Sadakane. log =12.

The great overabundance of cobalt is noticeable from the presence of optical region lines. Analysis of the IUE spectrum between 2100 and 2400 Å, which includes the strong CoII 2286 line, confirms this enhancement. Scandium and nickel do not have strong unblended lines in the investigated spectral region and only an upper limit can be determined.

Many chemically peculiar stars are known to be underabundant in helium. HeI 4471 was investigated in order to determine the helium abundance. The analysis used a spectrum provided by S. Adelman (The Citadel) obtained with a resolving power of 35 000, and showed helium to be deficient by 0.9 dex.

Chlorine is a rarely observed element in stellar spectra, and claims of its detection, are therefore always met with caution. Many of the strong spectral lines of ClII in the optical region are present in our data and a few of them in the same order (Fig. 10). We are able to confirm an overabundance of chlorine but at a reduced level from that proposed by Sadakane.

Fig. 10. Due to a large overabundance, spectral lines from chlorine are visible in the spectrum of HR 1094. Solid curve: Observed spectrum. Synthetic spectrum generated with chlorine abundance enhanced by 2.6 dex relative to the sun.

The heavy elements Pt, Au and Hg are present with abundances similar to the non-magnetic HgMn stars. Mercury has been investigated with respect to its isotope mixture by analyzing the HgII 3984 line observed with high resolving power. The isotope structure was analyzed by including the hfs and IS components from Wahlgren et al. (2000) into our synthetic stellar spectrum. The wavelength scale was set relative to CoII 3983 and CrII 3979, and the use of a terrestrial mercury isotope mixture shows a shift in wavelength of approximately 20 mÅ. The wavelength calibration is set with an accuracy of 10 mÅ, based on errors in the determination of the energy levels for chromium (Sugar & Corliss 1985). This shift may be explainable by magnetic broadening in addition to a wavelength shift but the magnetic influence is not expected to be great at this wavelength and another explanation is preferable. One possible explanation of the wavelength shift is if the isotope mixture consists only of the isotopes , and . The HgII 1942 line was investigated in the IUE spectrum, using the isotopic structure presented by Leckrone et al. (1991) for terrestrial mixture, to confirm the presence and the abundance of mercury. The isotopic mixture can not be confirmed due to the low resolving power of the IUE data. A similar isotope investigation for PtII 4046 was made with isotopic structure from Wahlgren et al. (2000), but such information was not possible to extract since the weak PtII 4046 line is blended with spectral lines from REE.

NI 1742 and 1745 indicate a reduced abundance of nitrogen of three orders of magnitude compared to the solar value. Oxygen was mainly analyzed from the OI triplet at 6156 Å and shows an underabundance of approximately 0.8 dex compared to the sun. The strong CI lines in the UV wavelength area were severely blended and the CII lines are too weak to use for an accurate abundance analysis. All carbon lines implied carbon to be underabundant by at least one order of magnitude.

Analysis of the spectrum for HR 1094 showed that the REE were represented by lines from their second spectra. Since HR 1094 is reasonably hot and the REE have low ionisation potentials, it seems likely that third and possibly fourth spectrum lines could also be present. If the presence of different ions is dependent only on the effective temperature and the electron density, then the third spectrum should be the most prominent with spectral lines corresponding to transitions between low excited states. The fourth spectrum might be observable but with present atomic data no lines were identified in the optical region. Most of the strong spectral lines from the REE in the optical region are severely blended. Observations made in the near-UV would increase the accuracy of the abundance analysis, since the REE (e.g. CeIII , PrIII and GdIII ) have many strong spectral lines between 3100 and 3600 Å. But analyses made for some of the REE, in particular gadolinium, show strong unblended spectral lines from both the second and the third spectrum. We have for an abundance analysis usable GdII lines from both of our NOT observations, which show the abundance of gadolinium to differ by approximately one order of magnitude. The observations were made at different rotational phases and the difference in abundance may be due to an inhomogeneous gadolinium abundance over the stellar surface. A contributing factor to the observed gadolinium abundance variation with phase may come from potential systematic errors in the oscillator strength data of Meggers et al. (1975). Since phases, = 0.073 and = 0.677, 0.683 utilize different wavelength intervals.

© European Southern Observatory (ESO) 2000

Online publication: March 28, 2000
helpdesk.link@springer.de