The electron affinity of oxygen: A systematic configuration interaction approach
Abstract
A sequence of configuration interaction (CI) wave functions, constructed so as to systematically approach the complete basis set, full CI limit, is used to argue that the only alternatives for improving the accuracy of electron affinity calculations are: (1) recovery of a sufficient fraction of the correlation energy of both anion and neutral so that the remaining error in the energy difference is acceptably small, or (2) methodological bias in favor of the more difficult to describe anion. Extended Gaussian basis sets, of the type recently employed in atomic hyperfine spin calculations are capable of recovering 95%96% of the total O (^{3}P) correlation energy. With much greater difficulty this basis can also recover an equivalent fraction of the O^{} (^{2} P) correlation energy. Nevertheless, the calculated electron affinity(1.31 eV) still underestimates the experimental value of 1.46 eV by 10%. Estimates based on multireference second order pertubation theory suggest that another 0.05 eV (EA=1.36 eV) is available from our present basis set, but cannot be variationally recovered due to hardware and software limitations. These results are in accord with the expectation that ΔE will not converge more rapidly than either E(O) or E(O^{}).
 Publication:

Journal of Chemical Physics
 Pub Date:
 January 1989
 DOI:
 10.1063/1.456154
 Bibcode:
 1989JChPh..90.1024F
 Keywords:

 Configuration Interaction;
 Electron Affinity;
 Oxygen Atoms;
 Wave Functions;
 Hartree Approximation;
 Self Consistent Fields;
 Atomic and Molecular Physics