Astron. Astrophys. 363, 815-820 (2000)

4. Spectrum analysis

Our analysis is performed by synthetic spectrum fitting using the program SYNTHE (Kurucz, 1993) with complete IS and hfs of GaII incorporated into the atomic line input. The primary target is  Cnc, whose spectrum has been analyzed for all strong GaII lines in the optical region.  Cnc has been observed on two occasions at the same phase and the results obtained by using different data sets are comparable. The  Cnc stellar system is analyzed with stellar atmosphere parameters adopted from Ryabchikova et al., with = 13200 K and log g= 3.7 for the primary star and = 8500 K and log g = 4.0 for the secondary. The stellar parameters are obtained by fitting spectrophotometry and hydrogen line profiles, considering a mass ratio between the primary and secondary component of =2.20.1. The projected equatorial rotational velocity, , for the primary object was determined to be less than 7 km s^-1 by synthetic spectrum fitting of sharp FeII lines while the corresponding value for the secondary component was derived similarly by using MgII 4481 to be of the order of 40 km s^-1. Fe II lines were used for the primary component since they do not show any significant hfs or IS, while the Mg II 4481 was the only usable feature from the secondary component in our stellar spectrum. Spectral line profiles are generated with different values of and log . The best fit provided the value of the rotational velocity. Line broadening mechanisms such as radiative, Stark and Doppler effects are included in the analysis via the damping constants derived by the SYNTHE program. Stellar models were calculated using the LTE approximation with the ATLAS9 (Kurucz, 1993) program under the assumption of no turbulent velocity.

The spectrum of the HgMn star HR 7775 has been investigated for comparison. However, a detailed analysis of the line profiles is not possible due to the weakness of the GaII spectrum. For HR 7775 we used a stellar atmosphere model with = 10750 K, log g= 4.0 and = 2 km s^-1, where all stellar parameters are from Wahlgren et al. (2000). The effective temperature and the surface gravity are based on Strömgren photometry using indices from Hauck & Mermilliod (1980).

The 5p - 5d transitions correspond to energy levels with small hyperfine constants, consequently we are not able to see any structure in these spectral features and the analysis is limited to abundance determination. Even if a spectral line profile does not show any signs of being affected by hfs or IS, it is of great importance to use line structure data in the analysis. As shown in Fig. 1, a difference of approximately 0.5 dex is observed if the hfs and IS are neglected.

Fig. 1. Abundance analysis using GaII 5416 in  Cnc. Solid curve: observed profile. Synthetic spectra are generated with (dotted) and without (dashed) hfs/IS. Both synthetic spectra are computed for an abundance of log =7.00. The difference in depth for the two spectra correspond to a difference in abundance of 0.5 dex.

The 5s - 5p transitions at 6334, 6419 and 6456 Å have a larger separation between the hyperfine and isotopic components which allows us to do a detailed analysis of structure and intensities of individual components. The wavelength separations for the individual components forms a broad spectral line with a peculiar line profile, which as shown in Fig. 2 might be interpreted as blends from unknown features.

Fig. 2. Analysis of the isotopic mixture using GaII 6419 in  Cnc. Solid curve: observed profile. Synthetic spectrum generated with hfs/IS, log =7.10 (dotted) is compared with a single line profile at 6455.9 Å, generated without hfs/IS, log =6.50 (dashed).

Gallium has two stable isotopes with a terrestrial abundance ratio of 60:40 (:). The separation of the spectral lines from different isotopes is small compared to the shift due to the nuclear moment, therefore it is difficult to achieve information about the isotopic mixture in an object. However, GaII 6419 and 6334 present line profiles indicating a different isotopic mixture than the terrestrial. Since the hyperfine constants for are smaller than corresponding values for , the outermost components will both be members from the latter. When the isotopic mixture is altered from 60:40 to 80:20, the spectral line becomes narrower by approximately 25 mÅ and agrees better with the observed profile. The change in line profile cannot be reproduced by broadening mechanisms such as rotation or turbulent velocity since different isotopic mixtures affect the individual components differently. The effect is noticeable for both GaII 6419 and 6334 in the  Cnc spectrum. Spectra obtained with a very high resolving power (R=200 000) will not give additional information due to the magnitude of the rotational velocity. We have calibrated the wavelength scale in this region with FeII lines, due to their insignificant hfs and IS. The wavelength calibration is set to an accuracy of 10 mÅ, based on errors in the determination of the iron line wavelengths (S. Johansson, private communication) and the synthetic spectrum fitting.

The analysis of the GaII 4d - 4f transitions between 4251 - 4262 Å was performed by including an estimated hyperfine and isotopic structure. The wavelength calibration for these transitions was made with MnII lines with a complete set of hyperfine components (Holt et al., 1999), with the same accuracy as for the iron line calibration. The increased line density compared to the red spectral region complicates the analysis and most GaII lines seem to be blended with other features. The spectra of  Cnc and HR 7775 both present an asymmetric line profile for GaII 4251 and even though the objects are not similar in spectral type or chemical composition this might imply that the line is disturbed by a common unknown feature. The GaII 4255 feature is a mixture of the 4d D₂ - 4f ³F₃ and the 4d ³D₂ - 4f ³F₂ transitions, each with their own set of hyperfine and isotopic components included in our analysis. The log gf values for the two involved transitions have been summed, where for 4d ³D₂ - 4f ³F₂, log gf=-0.32 has been used. This value is calculated from the Multi Configuration Hartree Fock (MCHF) program (Froese Fischer et al., 1997), but due to configuration mixing it is associated with large errors. However, it is in agreement with the value of log gf=-0.30 used by Dworetsky et al. GaII 4262 is blended with CrII 4261.9 in its blue wing, which might explain its deviation from the average gallium abundance.

The average abundance of gallium, log , in  Cnc is determined based on the results for individual GaII lines from the NTT spectrum. The result for each spectral line has been weighted with respect to its credibility in the abundance analysis considering line blending and the accuracy of the line profile fitting. The uncertainty of the oscillator strengths have not been considered during this process. A maximum value (3) for the weighting parameter corresponds to a nicely fitted unblended spectral line. The astrophysical log gf values are based on the result = 7.1 from the NTT data.

The abundance determined from the individual GaII lines is presented in Table 1 together with the newly determined astrophysical log gf values for GaII 5360, 5363 and 5421. A complete list of those lines including wavelengths and effective log gf values for the hyperfine and isotopic components is presented in Table 2 for the case of the terrestrial isotopic mixture. The effective log gf values are scaled with respect to the relative abundance of the isotopes while the values for the individual hyperfine components are calculated assuming LS-coupling.

Table 2. Wavelengths and effective astrophysical log gf values for the hyperfine and isotopic components to GaII 5360, 5363 and 5421 Å, assuming a terrestrial (:; 60:40) isotopic mixture.
Notes:
a) Karlsson & Litzén (2000)

© European Southern Observatory (ESO) 2000

Online publication: December 11, 2000
helpdesk.link@springer.de