An evaluation of the modified 6-31G* basis set for the atoms Ga-Kr using the Gaussian-3 and Gaussian-4 composite methods

Abstract

The new modified 6-31G* basis set of Mitin and Merz for the elements Ga–Kr is evaluated by applying the Gaussian-3 and Gaussian-4 composite methods to the set of 19 atomisation energies, 15 ionisation energies, 4 electron affinities and 2 proton affinities for atoms and molecules containing these elements. The standard 6-31G basis set with additional diffuse and polarisation functions is replaced by the modified basis set. Results are compared for the calculated geometries and energy values against experimental values, the results using the standard 6-31G basis and other composite methods. The calculated geometries are improved in the Gaussian-3 composite method where MP2 is used to determine the geometries, but are slightly worse in the Gaussian-4 composite method where the density functional B3LYP method is used. The energy values with the new basis set have slightly higher errors against experiment for the Gaussian-3 approach, but the errors for the Gaussian-4 approach are essentially unchanged. If the new basis set is widely accepted after further studies, there appears to be no serious disadvantage in using it for these composite methods.

Keywords:

• quantum chemistry
• 6-31G basis set
• composite methods

Acknowledgements

Thanks are due to L. A. Curtiss for providing the B3LYP optimised geometries for G4 theory with the original 6-31G basis set, and to A. V. Mitin for discussions on the minor errors in the modified 6-31G basis set. This paper is dedicated to Professor H. F. Schaefer III, on the occasion of his 65th birthday, for firing up my interest in calculations on the third-row main group elements when visiting his group in 1990.