Atomic data from the IRON Project XXXII. On the accuracy of the effective collision strength for the electron impact excitation of the quadrupole transition in Ar III

Since Burgess et al. (1997) have recently questioned the accuracy of the ef fective collision strength calculated in the IRON Project for the electron impact excitation of the 3s(2)3p(4) D-1 - S-1 quadrupole transition in Ar I II, an extended R-matrix calculation has been performed for this transition . The original 24-state target model was maintained, but the energy regime was increased to 100 Ryd. It is shown that in order to ensure convergence o f the partial wave expansion at such energies, it is necessary to take into account partial collision strengths up to L = 30 and to "top-up" with a ge ometric series procedure. By comparing effective collision strengths, it is found that the differences from the original calculation are not greater t han 25% around the upper end of the common temperature range and that they are much smaller than 20% over most of it. This is consistent with the accu racy rating (20%) previously assigned to transitions in this low ionisation system. Also the present high-temperature Limit agrees fairly well (15%) w ith the Coulomb-Born limit estimated by Burgess et al., thus confirming our previous accuracy rating. It appears that Burgess et al., in their data as sessment, have overextended the low-energy behaviour of our reduced effecti ve collision strength to obtain an extrapolated high-temperature limit that appeared to be in error by a factor of 2.