Barton-Howard empirical surface

The three best cti) initio values for Dc range from 4.3 to 5.0 kcal/mol.26,35,70 xhe best "purely" experimental value is 4.6 0.2 kcal/mol (1622 65 cm ), obtained by combining the value of Do = 3.04 0.01 kcal/mol (1065 5 cm l), obtained Dayton et 165 by photofragment spectroscopy, with the empirical estimate of AZPE = 1.6 0.2 kcal/mol (557 60 cm-1) of Pine and Howard, which they obtained using experimental values for the H-F stretch perturbations and the results of Barton and Howard s surface modelling for the contribution of the van der Waals modes to AZPE. The theoretical values of Do in Table I are consistent with the experimental value within 0.0-0.6 kcal/mol (0.0-0.4 kcal/mol if we ignore the experimental uncertainty). 

The most accurate surfaces for (HF)2 are the Barton-Howard (BH) empirical surface and four surfaces based on correlated electronic stmcture calculations with polarized one-electronic basis sets. The BH surface is defined only for fixed monomer bond lengths. Both the BH surface and the surface of Hancock, the authors, and Dykstra (HTD) are calibrated only in the van der Waals region. The surfaces of Redmon and Binkley (RB),27 Schwenke and the author (RBST), and Bunker, Kofranek,... 

There is little empirical information about the C2h saddle point with which to compare. The most reliable empirical estimates are probably those derived from the model potential surface of Barton and Howard. They obtained itf = 2.70 A, 6hff = 61.5 deg, = 4.0 kcal/mol, and = 0.9 kcal/mol. The fit of Hancock et to their out-of-plane bend potential gives 04 = 522 cm-l at the ACCD/TZ(d,p) saddle point geometry. 

Barton and Howard used their calculation as part of their modelling effort to obtain an empirical potential energy surface. They treated the HF bonds as rigid and obtained the tunneling effect implicitly by diagonalizing a four-dimensional Hamiltonian. 

---