London dispersion forces by range-separated hybrid density functional with second order perturbational corrections: the case of rare gas complexes.

A satisfactory account of the van der Waals (vdW) (London dispersion) forces is, in general not possible by the Kohn-Sham method using standard local, semilocal generalized gradient approximation (GGA), or meta-GGA density functionals. The recently proposed range-separated hybrid (RSH) approach, supplemented by second order perturbational corrections (MP2) to include long-range dynamic correlation effects, offers a physically consistent, seamless description of dispersion forces. It is based on a rigorous generalization of the Kohn-Sham method, where long-range exchange and correlation effects are treated by wave function methods, while short-range electron exchange and correlation are handled by local or semilocal functionals. The method is tested on a series of rare gas dimers in comparison with standard wave function theory and density functional theory approaches. In contrast to the most successful exchange correlation functionals, which describe at best the vdW minimum, the RSH+MP2 approach is valid also in the asymptotic region and the potential curve displays the correct 1/R(6) behavior at large internuclear separations. In contrast to usual MP2 calculations, the basis set superposition error is considerably reduced, making RSH+MP2 an ideal tool for exploring the potential energy surface of weakly bound molecular complexes.