Answers are grounded in this paper and cite their sources.
A new gradient-corrected exchange functional (G96) is introduced. While similar to Becke's B88 functional, it is much simpler and its potential in finite systems is asymptotically unbounded. The mean absolute deviations of the B88 and G96 exchange energies from the corresponding Hartree-Fock values for the atoms H to Ar are 12⋅5 and 8⋅5 mE h, respectively. In combination with the LYP correlation functional, it yields a density functional (G-LYP) that performs similarly to B-LYP on the standard G2 benchmark and we infer that the usefulness of a functional for molecular calculations does not depend critically on its asymptotic behaviour.
No takes yet. Share an insight, caveat, or question.
Peter M. W. Gill (1996) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: