The wavelengths in the tables are Ritz-type values derived from experimental energy level values in the NIST Atomic Spectra Database (ASD) [1]. That is, the wave number of a particular transition is found as the difference of the values of the combining energy levels in cm-1, and the wavelength in vacuum is the reciprocal of the wave number. Except for Li-like ions, only transitions are considered for which experimental energies are known for both lower and upper levels. For Li-like ions extremely precise ab initio calculations provide values for some energy levels that have uncertainties comparable to experimental values. Thus, for some levels of Li-like ions that have not been observed experimentally (see compilations [2-4]) we used the level values, and in turn wavelengths, obtained from such precise calculations. Where this has been done, the levels are specially denoted in the tables.
The ionization energies given in the text portion for each ion are taken from
values for the ionization limits in ASD. The values in cm-1 were
converted to electron volts with the factor
In compiling the transition probabilities we selected only values obtained with the most advanced theoretical and experimental methods. Our general evaluation criteria were those that have been developed at NIST [6,7]. We normally list here only values having estimated uncertainties of ±50% or less. Some exceptions have been made for important lines. Because of the limited amount of experimental results available for highly ionized ions, for most transitions we had to rely on theoretical data.
The most extensive source of theoretical data was the Opacity Project (OP) [8], which has produced multiplet f-values for the spectra of many elements. However, since the OP calculations generally do not include spin-orbit interaction they do not provide values for individual lines of a multiplet. Therefore for the present work the average values for LS multiplets were decomposed into their LSJ fine structure components using LS coupling rules [9]. For the present light atoms LS coupling should be a good approximation. For ions such where this is clearly not the case we have used results of calculations that do include spin-orbit and other relativistic effects. Tachiev and Froese Fischer [10] have performed calculations for Be-, B-, C-, N-, O-, F-, Ne-like ions with the multiconfiguration Hartree-Fock (MCHF) method with Breit-Pauli corrections and have made their results available on the World Wide Web. Blackford and Hibbert [11] have carried out extensive calculations for F-like ions [11] with the configuration interaction code-version 3 (CIV3) [12]. The same method was used by Aggarwal for several C-like ions [13]. For the Be- and B-like ions, the data of Safronova and coworkers were found very useful [14-16]. This was done by relativistic many-body perturbation theory (MBPT). Vilkas and coworkers applied many-body perturbation theory including Breit-Pauli corrections to obtain transition probabilities for ions of C, N, and O [17-19]. For comparative analyses data from many other sources were also used in our work.