Answers are grounded in this paper and cite their sources.
Second‐order density functional methods are used to introduce the electron correlation in Hartree‐Fock ( HF ) ab initio electronic energy calculations of three‐dimensional potential energy surfaces ( PES ). We analyze the behavior of these methods in PES calculations by applying them to the Li + FH reaction, which has been considered a prototype of the elementary atom–diatom reactions. This system has been studied also by the usual techniques, allowing a point‐by‐point (for a total of 317 grid points) comparison for the lowest 2 A ' adiabatic state. In particular, we compare the results obtained using the HF , Møller–Plesset ( MP 3 level), and configuration interaction ( CISD and MRDCI levels) methods with the corresponding results obtained using the Colle–Salvetti ( CS ) and Moscardó‐San Fabián ( MSF ) procedures using the HF results as the starting point. We found that the CS and MSF procedures support the prediction of a shallow well in the entrance channel that deepens slightly away from collinearity and disappears for a bond angle Θ < 74°. We also found that the constrained saddle‐point positions remain essentially constant from Θ = 180°–90° and are clearly in the exit channel as for the MRDCI approach (corresponding to the best results). In conclusion, there is a good overall agreement, but there is a question in which this agreement is less pronounced: the heights of the saddle points including the transition state. In particular, the transition‐state height is about 3 kcal/mol higher than the more accurate value obtained with the MRDCI approach. However, the second‐order density functional methods have been capable of reducing the HF barrier from 18 to 9 kcal/mol (all of these values obtained by spline interpolation), the latter value being very similar to the CISD result.
No takes yet. Share an insight, caveat, or question.
Suárez et al. (1994) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: