5 Discussion

We report for the first time electron impact excitation rates computed in the close-coupling approximation for all the fine-structure transitions within the n=3 complex of Fexv. Important effects that influence the rate accuracy have been studied and taken into account; that is, target representation, the resonance structure, the convergence of the partial wave expansion, top-up procedures, asymptotic long-range potential couplings and the energy-mesh step. This effort has resulted in a considerable improvement over previous work, which had been mainly carried out with the simpler distorted wave approximation. As shown, relevant effects were sometimes overlooked in previous work hence limiting the overall reliability of the available collisional datasets. However, from extensive comparisons we have attempted to evaluate the accuracy of the present results. We are confident that the listed effective collision strengths with magnitudes $\Upsilon(T)\gt 10^{-2}$ are accurate to better than 20% whereas those with smaller values are probably reliable only within a factor of 2. We have also demonstrated that excitation rates for highly ionised systems, e.g. Fexv, obtained from sparsely tabulated collision strengths that do not resolve the resonance structure can be unreliable, and therefore standing inconsistencies in plasma diagnostics at high temperatures in many situations can indeed be due to poor atomic data.

Acknowledgements

Part of the present work was carried out during visits by MEG and CM to the Observatoire de Paris, Meudon, France. The hospitality received is gratefully acknowledged. The visits were funded by the CNRS, IVIC, CONICIT, Fundación Polar, the Observatoire de Paris and the Ministère des Affaires Étrangères. Computations were carried out at the Ohio Supercomputer Center, Columbus, Ohio, U.S.A., and at CeCalCULA, Universidad de Los Andes, Mérida, Venezuela. This research has been supported by CONICIT under contract No. S1-95000521.