The astrophysically rare element gallium (Z=31) has earlier been studied with respect to its abundance in normal and chemically peculiar (CP) stars. In the sun its abundance is listed as log =+2.9 (Anders & Grevesse, 1989) on a scale where log =12, and it is believed to be the lightest element to have a substantial contribution to its abundance arising from the slow neutron caption process (s-process) of nucleosynthesis. Other contributions to its abundance arise from more energetic processes (e and r-processes). Among the CP stars the spectrum of GaII is particularly strong in some members of the HgMn and He-weak subclasses. The line enhancements represent abundances of up to four orders of magnitude over the solar value when simply interpreted by homogeneous model atmospheres. However, as for many other spectrum anomalies exhibited by CP stars the line enhancements are believed to result from atmospheric processes, such as diffusion, rather than from nucleosynthetic origins.
Abundance analyses performed by using spectral lines in the optical region (Adelman 1989; Lanz et al. 1993) have resulted in a different value than that obtained with the GaII and GaIII resonance lines in the UV (Takada-Hidai et al. 1986; Smith 1996). This has recently been referred to as "The Gallium problem" (Dworetsky et al., 1998). The most appealing explanation for the discordant results has been the need for hyperfine structure (hfs) of spectral lines located in the optical wavelength region (Dworetsky et al. 1998).
Smith used the resonance lines of GaII and GaIII in the UV for studies of several objects and found the gallium abundance to be lower using these lines compared to results from other investigations in the visible. For some objects included in Smith's work the abundance obtained from individual spectral lines deviates much from the average value, especially if only a moderate enhancement of gallium is observed compared to the sun. Smith determined an average value for the gallium abundance in Cnc to be log =6.6 from UV lines which is 0.6 dex below the value presented by Adelman (1989) from optical region GaII lines.
Dworetsky et al. attempted to equalize the difference observed by Smith by investigating spectral lines in the optical wavelength region and utilizing a limited incorporation of the hfs. They used the GaII 4d - 4f transitions (4251 - 4262 Å) in addition to the GaII 5s - 5p transition at 6334 Å and were able to match the result by Smith. Their abundance result is an averaged value of five spectral lines, where a systematic decrease in abundance is observed when lines at longer wavelengths are used. The hfs was estimated by including an incomplete set of hyperfine components based on experimentally measured wavelengths (Bidelman & Corliss, 1962), except for GaII 6334 where the structure was adopted from Lanz et al.
Recent laboratory measurements of the GaII spectrum (Karlsson & Litzen, 2000) have provided a complete set of hyperfine and isotope components for the spectral lines corresponding to the 5s - 5p and 5p - 5d transitions, which are located in the red spectral region. The 4d - 4f transitions used by Dworetsky et al. have also been analyzed by Karlsson & Litzén, but due to large configuration interaction in the 4f levels it is at this time only possible to experimentally estimate the isotopic and hyperfine structures using the Fourier Transform Spectrometer technique. The shift between the outermost hyperfine components for the 4d - 4f transitions is reasonably large and without considering the line structure, their use will increase the uncertainty of the result. The GaII lines in the red spectral region have a simpler structure, therefore wavelengths can be derived individually for each component and the hyperfine constants derived unambiguously.
In this study spectral lines from GaII are investigated with a complete set of experimentally determined hyperfine and isotopic components. We have chosen to analyze spectra of Cnc and HR 7775 in the visible region to investigate the abundance and isotopic mixture of gallium in a further attempt to reconcile the "Gallium problem".
© European Southern Observatory (ESO) 2000
Online publication: December 11, 2000