3. Atomic data
Laboratory spectra of gallium were recorded with the Fourier Transform Spectrometer (FTS) at the Atomic Spectroscopy Laboratory in Lund. The spectra were obtained using a hollow-cathode discharge as a light source and covered the region 12 500-50 000 cm-1 (8000-2000 Å). All GaII lines showing significant hfs and isotopic shift (IS) were analyzed by means of a computer code, where a function consisting of co-added gaussians was fitted to the observed feature. Parameters included in the fitting process were the magnetic dipole and electric quadrupole hyperfine constants, A and B respectively, and the central wavelength. The electric quadrupole constant, B, was found to be insignificant during the line fitting process. Wavelengths for the hfs components of a transition were therefore derived from the magnetic dipole constant and the central wavelength. The fitting process is not applicable for the 4d - 4f transitions due to severe perturbations of the hyperfine structure by configuration interaction. For those transitions the hyperfine and isotopic structure were estimated by measuring the wavelengths and the intensities for all resolved features in the FTS spectrum. A more comprehensive report of the laboratory measurements will be published elsewhere (Karlsson & Litzen, 2000).
Oscillator strengths have been taken from Ryabchikova & Smirnov (1994). The values for the 4d - 4f transitions in the blue spectral region are based on lifetimes measured with the beam-foil technique. No experimental data for the lifetimes of the 5p and 5d levels are at the present time available. The oscillator strengths for the spectral lines in the red wavelength region are based on extrapolated lifetimes from the 4f levels. The branching fractions for all levels are, with a few exceptions, set by Ryabchikova & Smirnov to be 1.0, which is too large a value due to the multiplicity of the deexcitation possibilities. The oscillator strengths for the 5p - 5d transitions at 5360 and 5421 Å are determined astrophysically by Ryabchikova & Smirnov and compared in Table 1 with corresponding values from this work. The log gf values in our analysis are obtained with a complete set of hyperfine and isotopic components, which partly can explain the discrepancy between the log gf values observed for the 5421 line.
Table 1. GaII lines used in this analysis.
© European Southern Observatory (ESO) 2000
Online publication: December 11, 2000