Assessment of Gaussian-3 and density functional theories for a larger experimental test set

The G2/97 test set [J. Chem. Phys. 106, 1063 (1997)] for assessing quantum chemical methods used to predict thermochemical data is expanded to include 75 additional enthalpies of formation of larger molecules. This new set, r eferred to as the G3/99 test set, includes enthalpies of formation, ionizat ion potentials, electron affinities, and proton affinities in the G2/97 set and 75 new enthalpies of formation. The total number of energies in the G3 /99 set is 376. Overall, G3 theory has a mean absolute deviation of 1.07 kc al/mol for the G3/99 test set and does about as well for the new hydrocarbo ns and substituted hydrocarbons as it does for those in the G2/97 test. How ever, G3 theory has large deviations for several of the new nonhydrogen sys tems in the G3/99 test set such as SF6 and PF5. Part of the source of error is traced to the inadequate geometries used in G3 theory for these molecul es. Other variations of G3 theory are also assessed such as G3(MP2), G3(MP3 ), and the versions of G3 theory using scaled energy terms instead of the h igher level correction. These variations also do well for the larger hydroc arbons and substituted hydrocarbons, but fail for the same nonhydrogen syst ems as G3 theory. The density functional methods assessed in this study, in cluding the hybrid B3LYP method, all have much larger deviations from exper iment for the new enthalpies of formation in the expanded test set; the mea n absolute deviation more than doubles compared to that for the enthalpies in the G2/97 test set. This is due to a cumulative effect of the errors in the larger molecules in the density functional methods. (C) 2000 American I nstitute of Physics. [S0021-9606(00)30815-7].