4. Application to observations
4.1. The spectra
As a test and first application of the model atom oxygen abundances for three stars are determined: Vega, Leo and HD 92207. High S/N and high resolution spectra are used in this process.
For Leo and HD 92207 Echelle spectra using FEROS (Kaufer et al., 1999) at the ESO 1.52m telescope in La Silla were obtained in January 1999. Nearly complete wavelength coverage between 3600 and 9200 Å was achieved with a resolving power (with 2.2 pixels per resolution element) yielding a S/N of several hundred in V in a 2 and 5 min exposure, respectively. Data reduction was performed using the MIDAS package, as described in the FEROS documentation (http://www.ls.eso.org/lasilla/Telescopes/2p2T/E1p5M/FEROS/docu/Ferosdocu.html ). The spectra were normalised by fitting a spline function to continuum points and finally shifted in wavelength to the rest frame using the radial velocity determined from cross-correlation with an appropriate synthetic spectrum.
An Echelle spectrum of Vega was kindly made available by A. Korn with almost complete wavelength coverage between 3900 and 9400 Å. FOCES (Pfeiffer et al. 1998) at the Calar Alto 2.2m telescope was used in June 1999 to obtain three exposures of 4 s and 210 s, respectively. The spectra were reduced in the standard way using the routines described by Pfeiffer et al. (1998). After merging of the single spectra and rectification a S/N of 750 near was measured at (2 pixels per resolution element).
In general, the observations are of high quality with few spectral regions corrupted by CCD defects. As the data were obtained only as an addendum to the main observing program, no additional spectra of a fast rotator are available at the correct airmass to remove the telluric features properly.
4.2. Abundance analysis
The basic properties and atmospheric parameters of the test stars are summarised in Table 2. Information on the basic properties are taken from the Bright Star Catalogue (Hoffleit 1982). Atmospheric parameters for Vega are adopted from Castelli & Kurucz (1993) except for which we take from our analysis (still compatible with their value). For the two supergiants atmospheric parameters are determined prior to the oxygen abundance analysis applying the model atmosphere method developed by Venn (1995b) for galactic A-type supergiants, discussed in detail therein. In brief, and are derived simultaneously by finding the ionization equilibrium of MgI / II using a non-LTE model atom as comprehensive as the one presented here (Przybilla et al. 2000) and by fitting the wings of the higher Balmer lines (typically from upwards) which are still formed in photospheric regions in contrast to the wind affected and features. For Vega this method leads to the same parameters as derived by Castelli & Kurucz (1993) by comparing observed and computed energy distributions and Balmer profiles. The microturbulent velocity is determined from LTE spectrum synthesis for a large ensemble of FeII , TiII and CrII lines by demanding that there is no relation between abundance and line strength. Rotational velocities and macroturbulence in the radial-tangential model are derived from spectrum synthesis as both broadening mechanisms alter the line profile in different ways (Gray 1992). Error estimates for the stellar parameters of the test stars are also given in Table 2.
The results of the abundance analysis for oxygen are summarised in Table 3 which gives the wavelength, multiplet number, lower excitation potential and the adopted gf value for the observed lines (all from Wiese et al. (1996); multiplet numbering according to Moore 1976) and the measured equivalent widths, derived abundances and non-LTE abundance corrections for the different stars. Blended lines are marked by "S" as long as an analysis via spectrum synthesis is still feasible and for lines originating in the hydrogen line wings the equivalent widths are measured against the local continuum ( in parentheses). Non-LTE and LTE mean values and the line-to-line scatter () from the lines in the visible are also given, the near-infrared lines are omitted for reasons discussed below. For Vega non-LTE abundances for a model with a depth-dependent microturbulence (Gigas 1986, see Sect. 5 for details) are also displayed. Note that the abundances are derived from the detailed spectrum synthesis results and not from an equivalent-width study. We deviate from this general procedure for the near-infrared lines. The LTE abundances and therefore the non-LTE abundance corrections are determined by reproducing the observed equivalent widths as fitting the observed profiles proves to be impossible in LTE. The non-LTE abundances from the individual (unblended) lines of the test stars are plotted versus their equivalent widths in Fig. 12, clearly demonstrating the remarkably small scatter in the line-to-line abundances. In particular, the application of a depth-dependent microturbulence in Vega results in perfect agreement of oxygen abundances derived from the weak and strong lines.
In Figs. 13 to 15 theoretical line profiles for the derived mean non-LTE oxygen abundance are compared with the observations; excellent agreement is found for the lines in the visible. Other elements are included for the spectrum synthesis in LTE in order to disentangle line blends. As some of the oxygen lines are formed in the wings of HI lines, profiles for hydrogen are calculated on the basis of non-LTE level populations. The OI abundances for the test stars are:
We list the values obtained from the non-LTE analysis together with uncertainties from the line-to-line scatter and systematic errors (cf. Sect. 2.3); the number of analysed lines is given in parentheses. Non-LTE shifts the derived oxygen abundances systematically to lower values and the line-to-line scatter is slightly reduced in comparison to LTE.
Vega shows an oxygen deficiency of 0.3 dex. This is slightly less than its general underabundance in the heavier elements by typically 0.5 dex (solar abundances adopted from Grevesse et al. (1996), but see Reetz (1998) for a critical reexamination of the solar oxygen abundance). We find negligible non-LTE corrections for the weak lines in this main sequence star. The quality of the line fit to the strong near-infrared lines is drastically improved by applying a variable microturbulence in the line formation as found by Gigas (1986), while the weak lines are quite insensitive, cf. Sect. 5.
In the case of Leo the oxygen abundance is marginally below the solar value. Only small ( 0.15 dex) non-LTE abundance corrections apply for the weak lines. On the other hand, huge corrections occur in the strong lines. The line profiles can not be reproduced accurately as an unidentified mechanism broadens the line wings (also present in the wings of strong lines from other elements). Nevertheless, the line depths are predicted quite correctly by adopting the oxygen abundance from the weak lines. Generally, we derive an approximately solar metallicity for this star from the LTE elemental abundance analysis.
For HD 92207 we find an oxygen abundance similar to Leo with pronounced non-LTE abundance corrections; from the LTE spectrum synthesis we obtain [Fe/H] dex. Some regions in the spectrum of this star are still of too low S/N to analyse several of the weakest oxygen lines. The strong OI lines in the infrared are definitely affected by the wind outflow velocity field observed for this star, showing asymmetrical profiles with blueward shifted absorption, as are the strong lines of other elements.
4.3. Comparison with other analyses
In the following our results for Vega and Leo are compared with those of other recent analyses (cf. Table 5) To our knowledge abundances for the extreme supergiant HD 92207 are determined for the first time.
Table 5. Comparison of oxygen abundances for Lyr and Leo.
Vega For this star Takeda (1993) derives an oxygen abundance of 8.6 from a non-LTE analysis of 11 OI lines with equivalent widths similar to those from Table 3. The scatter in the abundances from individual lines is larger than ours. Remarkable are the comparatively large non-LTE abundance corrections that he finds - typically two to three times as large as ours, even for the weak lines, cf. Sect. 3.4 for details. His LTE abundance from the weak lines of (5 lines) results partly from using -values dex smaller than our OP data. Some of the discrepancy is also related to the different background opacities used (Kurucz, 1979vs. Kurucz, 1992); his stellar parameters for Vega are almost identical with ours: K and at [Fe/H]0.6 dex.
Venn & Lambert (1990) find an OI abundance of 8.74 from a LTE analysis of 6155-8 for the stellar parameters (9650/4.0) and values of similar to ours. However, their equivalent widths differ by 25% from our measurements.
In the LTE study of Lambert et al. (1982) an abundance of 8.820.12 is derived from 4 weak OI lines being inconsistent with our findings. Their gf values and their equivalent-width measurements are similar to ours. But they use a model atmosphere with solar elemental composition at K and thus compensating the higher temperatures by increasing the oxygen abundance.
Leo Takeda & Takada-Hidai (1998) find an oxygen abundance of 8.70 from a non-LTE study of the OI triplet 6155-8 in the spectrum of Leo; they give a comparatively large non-LTE abundance correction of . Both their non-LTE and their LTE value are just about consistent with our results. The discrepancy can be traced to their stellar parameters (=10200 K, =1.9) - which would also result in a higher LTE oxygen abundance within our approximation - and to their OI non-LTE model, cf. Sect. 3.4.
Venn (1995a) gives an OI abundance of from a LTE study for stellar parameters (9700/2.0). Comparing our LTE results for the analysed lines ( 6155-8, 6453-4) with hers gives excellent agreement. In that work non-LTE effects are estimated to result in abundances 0.2 dex lower, adopting the results of Baschek et al. (1977).
The LTE study of Lambert et al. (1988) finds from 6155-8 for the stellar parameters (10500/2.00). An appropriate reduction of their would bring their OI abundance into better accordance with ours.
Wolf (1971) derives from three oxygen lines in an early LTE study on the basis of an unblanketed model atmosphere for the parameters =10400 K and =2.05 with a depth dependent of 2...10 km s-1. Given the large scatter in the line-to-line abundances, the deviations from our modern gf values (1 dex in one triplet) and the differences in some of the measured equivalent widths the similarity of his results with our findings is coincidental.
Where not mentioned explicitly, the equivalent widths measured in the studies above agree well with ours and the gf values used are almost identical. The higher effective temperatures adopted by Takeda & Takada-Hidai (1998) and Lambert et al. (1988) can most likely be traced back to differences in the line blanketing (Kurucz 1979) for the model atmospheres.
© European Southern Observatory (ESO) 2000
Online publication: July 13, 2000