5 Discussion

The main goal of this work was also to collect all available data on the collisional ionization, radiative and dielectronic recombination rates data. Most of this data come out from new advanced computational calculations. About the ionization rates we would like to remark that recently Kato et al. (1991) reviewed and compared the rates obtained with several empirical formulae used by Lotz (1967, 1968), by the Belfast group (Bell et al. 1983 and Lennon et al. 1988), by AR85 and AR, and by Pindzola et al. (1987). They conclude that generally the rate coefficient derived on the basis of experimental data of cross sections are in good agreement among the various authors. Otherwise when no experimental data are available only AR85 and AR included the EA process. This reflects in a difference at high temperature for some isosequences (e.g., Na-like, Li-like, and Mg- to Ar-like sequences). Kato et al. (1991) also remark that for some ions the Belfast group rates are lower by a factor 1.626 due to a misprint on the Younger (1982) table. Based on the Kato results, Voronov (1997) refitted the data of the Belfast group with a very simple formula for the total ionization rates. Instead of its simplicity we do not use the Voronov data for the following reason: i) Voronov give a fit only for the total ionization rate. This formula is valid as long as we consider the case of ionization equilibrium plasmas. Under non-equilibrium conditions (such as apply to transient plasmas which are present e.g., in supernova remnants and solar/stellar flares) inner-shell ionization may play an important role, both in the determination of the ionization balance and in the formation of fluorescent lines. Then it can be important in certain cases (e.g., for the Be- and B-like sequences) to know the contribution from different atomic subshells separately (Mewe private communication). ii) we checked the Voronov results and we saw that, instead of his claim, some curves are in agreement with the original data of the Belfast group without the correction factor of 1.626 (which implies that in these cases his results are just a factor of 1.626 below the corresponding AR results, see e.g. Kato et al. (1991) for Fe XXII to Fe XXVI). So, for some ions, the Voronov ionization rates are uncorrected. We would like to remark that the ionization rates for Ni I to Ni X are probably wrong by large factors due to a large underestimate of the Excitation autoionization process in the AR85 (Arnaud private communication). In a forthcoming paper (Kaastra et al. 1998) we intend to update the ionization rate including EA, resonance excitation double autoionization and direct multiple ionization.

We believe that at present our atomic data collection represents the state of the art in such calculations. An important point of this work was to reorganize all the available data with a few and simple fitting formulae in order to use them in a numerical modeling of plasma processes.

Obviously, the differences in the ionization fraction between our work and the previous ones, reflects directly in the computed line power emissions. In particular, our preliminary calculations show sensitive differences, depending on the selected plasma temperature, in the computation of the total line radiation emission with respect to previous similar work. We will fully address these issues in a forthcoming paper.

Acknowledgements

We are grateful to R. Mewe, for helpful comments and suggestions. We are indebted to M. Mattioli who has pointed out to us some mistakes in the data.