Negative Atomic Ions. II

Abstract

Electron affinities of negative atomic ions were obtained in a previous paper [H. R. Johnson and F. Rohrlich, J. Chem. Phys. 30, 1608 (1959)] by isoelectronic extrapolation using a four-term approximation formula. Recently, Edlén [B. Edlén, J. Chem. Phys. 33, 98 (1960)] simplified this procedure, using only three terms and improved experimental input data. The theoretical justification of this procedure lies in the assumption that these formulas are the beginning terms of a perturbation series which is at least semiconvergent. Therefore, a four-term formula should give better results than a three-term formula. It is shown here that the experiments indicate that this is not likely to be the case, thereby raising strong doubts about the justification of this method. An alternative method is therefore suggested. Here, the regularity pattern of the ionization potentials as a function of Z (fixed degree of ionization) is extrapolated. This is therefore an extrapolation of the results of a ``horizontal analysis.'' The method requires one experimental affinity for each p shell as long as sufficiently precise data on ionization potentials of highly ionized atoms (fourth and higher spectrum) are not available. The results agree within experimental error, except for the most recent value of C— which is about 10% too low.