Density functional theory (DFT) within standard local density/generalized gradient approximations (LDA/GGAs) have been proved to succesfully predict the properties of a wide class of electronic systems at a reasonable computational time. However there exist a number of situations in wich DFT within LDA/GGAs qualitatively fails. One of such problems is their inability, due their intrinsic local nature, to describe long-range interaction between non overlapping molecular fragments, or weakly bound systems such as molecules about to break during a chemical reaction. The Adiabatic Connection Fluctuation Dissipation (ACFD) formalism tackle this kind of problems at a very foundamental level providing a perfect starting point for the development of truly non-local functionals. We have indeed developed and implemented an efficient scheme for the calculation of the correlation energy via the ACFD theorem going beyond the random phase approximation (RPA) by including the exact-exchange contribution to the kernel. We found that this contribution plays a crucial role for a correct and accurate description of the total energy of an electronic system without compromising the achivements of the original RPA functional.
