The multichannel R-matrix method has been used extensively in the past by Ramsbottom et al. (1994, 1995, 1996a, 1997) to produce effective collision strength for electron-impact excitation of astrophysically important NIV, NeVII, SII and ArIV. These calculations have been proved to have a typical accuracy of 10% or better (Keenan et al. 1995, 1996). We would expect the present effective collision strengths to have a similar accuracy, but it is difficult to assess due to the lack of theoretical data available for comparison, and because the astrophysical diagnostic applications have not yet been performed using this new atomic data. In the near future theoretical line intensities, calculated using the present results, will be compared with high resolution optical observations of planetaries, and this will assist in predicting the accuracy more precisely. The most serious error in the tabulated data occurs due to the omission of higher-lying target states in the wavefunction representation of the MgVI ion. To quantify these errors, a much more sophisticated calculation would need to be performed. The present work is the most elaborate to date due to the fact that extensive configuration-interaction wavefunctions were utilized, and a very fine mesh of electron-impact energies was used to properly delineate the detailed resonance structures in the collisional cross sections. The neglect of these autoionizing resonances would have a dramatic effect on the resulting effective collision strengths. In addition the present calculation took care to include correlation terms in the total wavefunction to compensate for the omitted higher-lying levels.
Acknowledgements
The work reported in this paper has been supported by PPARC, under the auspices of a Rolling Grant GR/L 20276, and the computations were carried out on the CRAY J932-SN9522 at Rutherford Appleton Laboratory. We are grateful to Prof. F.P. Keenan for helpful discussions.