| Abstract
| - Single determinant Møller−Plesset perturbation (MP) theory at second order (MP2), third order (MP3), andfourth order (MP4) with standard basis sets ranging from cc-pVDZ to cc-pVQZ quality predicts the equilibriumgeometry of FOOF qualitatively incorrect. Sixth-order MP (MP6), CCSD(T), and DFT lead to a qualitativelycorrect FOOF equilibrium geometry re, provided a sufficiently large basis set is used; however, even thesemethods do not succeed in reproducing an exact re geometry. The latter can be achieved only by artificiallyincreasing anomeric delocalization of electron lone pairs at the O atoms into the σ*(OF) orbitals by selectivelyadding diffuse basis functions, adjusting exponents of polarization functions, or enforcing an increase ofelectron pair correlation effects via the choice of a rigid basis set. DFT geometries of FOOF can be improvedin a similar way and, then, DFT presents the best cost-efficiency compromise currently available for describingFOOF and related molecules. DFT and CCSD(T) calculations reveal that FOOF can undergo either rotationat the OO bond or dissociation into FOO and F because the corresponding barriers (trans barrier: 19.4 kcal/mol; dissociation barrier 19.5 kcal/mol) are comparable. Previous estimates as to the height of the rotationalbarriers of FOOF are largely exaggerated. Rotation at the OO bond raises the barrier to dissociation becausethe anomeric effect is switched off. The molecular dipole moment is found to be a sensitive antenna forprobing the quality of the quantum chemical description of FOOF.
|
