GGA-Level Subsystem DFT Achieves Sub-kcal/mol Accuracy Intermolecular Interactions by Mimicking Nonlocal Functionals

Abstract

The key feature of nonlocal kinetic energy functionals is their ability to reduce to the Thomas-Fermi functional in the regions of high density and to the von Weizsacker functional in the region of low-density/high reduced density gradient. This behavior is crucial when these functionals are employed in subsystem DFT simulations to approximate the nonadditive kinetic energy. We propose a GGA nonadditive kinetic energy functional which mimics the good behavior of nonlocal functionals, retaining the computational complexity of typical semilocal functionals. Crucially, this functional depends on the inter-subsystem density overlap. The new functional reproduces Kohn−Sham DFT and benchmark CCSD(T) interaction energies of weakly interacting dimers in the S22-5 and S66 test sets with a mean absolute deviation well below 1 kcal/mol.

Plasmon Couplings from Subsystem Time-Dependent Density Functional Theory

“…The ∼ superscript symbol in eq indicates that the KE functional T s is (necessarily) approximated. In fact, the kinetic energy as a functional of the electronic density is not known and different approximations can be used. − ,,,, On the other hand, the XC functional is, in the case of a semilocal functional, already expressed as a functional of the density, so eq can be computed exactly (within the semilocal approximation). Additional approximations are instead required in the case of more advanced XC functionals. − …”

“…The FDE method can reproduce the exact Kohn–Sham (KS) ground-state density of the whole (supermolecular) system, considering only calculations of the subsystems, thanks to the inclusion of an embedding potential. , The accuracy of the embedding potential, however, depends on the accuracy of the kinetic energy (KE) functional as a function of the density, which is unknown and must be approximated. Current approximations for the KE functional are very accurate only for weakly interacting systems. − For subsystems connected by chemical bonds, either reconstructed potentials by the inversion technique − or external orthogonality methods − have been employed.…”

Density Embedding Method for Nanoscale Molecule–Metal Interfaces

“…As expected, comparing the performance of the semilocal revAPBEK and the nonlocal LMGP NAKE functionals, we see that LMGP brings the sDFT result much closer to the KS-DFT benchmark. This is expected because LMGP was shown to be superior to revAPBEK in reproducing weak intermolecular interactions. , …”

“…This is expected because LMGP was shown to be superior to revAPBEK in reproducing weak intermolecular interactions. 46,48 We should remark that sDFT constrains the number of electrons in each subsystem to remain constant throughout the SCF procedure. Therefore, our method is expected to become approximate whenever there is a strong molecule−metal charge transfer.…”

“…For NAXC, one can borrow the same (pure) density functionals employed in KS-DFT implementations (such as local and semilocal functionals). For NAKE, the situation is more delicate, as we discuss later in this manuscript, where generally subsystems can only be weakly interacting in order to keep the simulations quantitatively close to a KS-DFT simulation of the entire system. − We refer the interested reader to several reviews on sDFT − for additional details.…”